| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Laboratory of Tumor Immunology, Division of Oncology, and
Department of Neurosurgery, University Hospital Geneva, Geneva, Switzerland; and
Laboratory of Immunology, E.F.S. Bourgogne Franche Comté, Besançon, France
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
It was once thought that the immune privileged status of the brain may pose an insurmountable challenge for antitumor immune responses. However, immune privilege is currently interpreted as meaning an immune reactivity that is modified rather than absent (4). Indeed, lymphocyte entry to the brain (5, 6) and a degree of lymphatic drainage from the CNS are now well described (7). Furthermore, spontaneous immune responses have also been documented in human glioma (8), although these are insufficient to eliminate the tumor. More impressive antitumor activity has been demonstrated in different animal models, in which responses induced in the periphery are able to mediate antitumor effects in the brain (9, 10, 11, 12, 13, 14, 15). Many of these approaches have been developed from models for tumors in other sites, but there is not always a direct translation of such therapies to brain tumors. For example, studies in which multiple cytokines have been tested as modulators of immune responses gave different results according to the tumor model (SMA-560, B16) and site of implantation (13, 16). Other attempts to create a cellular vaccine by overexpression of ICAM-1 on a glioma cell line resulted in growth inhibition of glioma cells implanted s.c., but not in the CNS (17). Furthermore, in certain clinical studies, it was suggested that tumors metastasizing to the CNS could be particularly resistant to treatment with adjuvant immunotherapy (18). In most of these studies, the immune (or indeed other) mechanisms responsible for the success or failure of the immunomodulating therapy have been difficult to directly address. In many cases, this is because certain rodent brain tumor models are poorly characterized or are not always totally syngeneic (12, 19, 20).
To facilitate immunological analyses, some recent studies have used better characterized tumor models in the brains of syngeneic mice, such as B16/F10 melanoma (13, 21) and C3 sarcoma cells transfected with the human papilloma virus type 16 (22). These models have provided useful information about ways of inducing efficient antitumor responses and of the different cellular and cytokine requirements for such immunity. Nevertheless, the special requirements for safe, efficient immune responses in the CNS necessitate further information about the specificity of such responses. However, there is very little in vivo data for the fine specificity of immune responses to brain tumors, principally because of the difficulty to date in defining specific T cells. We have therefore chosen to investigate immune responses to a brain tumor using a model that facilitates identification of specific T cells. This is the immune response elicited in syngeneic DBA/2 mice after implantation of P815-CW3 tumor cells. It was demonstrated (in non-CNS sites to date) that P815-CW3 tumor cells are rejected by a potent CTL response focused on an immunodominant determinant from region 170–179 of the CW3 molecule (CW3170–179). These CTLs exhibit highly conserved TCR structural features, including the apparently exclusive usage of a BV10 TCR (23), and can be readily monitored ex vivo by flow cytometry as an expanded BV10+CD62L-CD8+ subset (24, 25) or by TCR spectratyping (26). The utility and accuracy of defining CW3-specific CTLs by either of these two techniques have also been confirmed by MHC/peptide tetramer analysis (26).
In this study, we exploit the unique characteristics of the immune response to P815-CW3 to address crucial issues about the recruitment and function of CTLs specific for a tumor located in the CNS. We show that the majority of CD8 cells infiltrating the brain parenchyma are CW3 specific and retain high functional activity even in the absence of CD4 cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Transfection of the P815 murine mastocytoma with HLA-CW3 has been previously described (27). Clone 444/C9.3.1 was used for all implantations, hereafter referred to as P815-CW3, kindly supplied by Drs. J. L. Maryanski and J.-C. Cerottini (Ludwig Institute for Cancer Research, Lausanne, Switzerland). Cultured cells were washed and resuspended in PBS before implantations. For in vitro cytotoxicity tests, nontransfected P815 cells were also used as targets.
Mice and P815 cell implantations
Adult female DBA/2 mice (RCC, Füllingsdorf, Switzerland, or IFFA Credo, L’Arbresle, France) were implanted with viable P815-CW3 cells from culture. For s.c. implantations, 2 x 107 cells were injected in the flank. For intracerebral (i.c.)3 implantation, mice were first anesthetized by inhalation of isoflurane (Abbott, Baar, Switzerland), followed by i.p. injection of a mixture of xylazine (Bayer, Leverkusen, Germany) and ketamine (Warner-Lambert, Baar, Switzerland). P815-CW3 cells were then implanted into the pallidum with the aid of a syringe mounted in a stereotaxic instrument (Stoelting, Indulab, Gams, Switzerland), 2.5 mm lateral to bregma and 3.5 mm below the surface of the skull. Preliminary dose-response experiments in which between 5 x 103 and 5 x 105 P815 cells were implanted i.c. indicated that immune responses and clinical symptoms were more uniformly apparent when 5 x 105 cells were implanted, and so this dose was used for all subsequent experiments. Higher numbers of cells could not be implanted because the limit of weight loss permitted by the local animal welfare authorities had been reached.
Cell preparations
PBL were purified by Ficoll-Hypaque centrifugation (Pharmacia, Uppsala, Sweden), and single cell suspensions from lymph nodes and spleen were prepared by standard procedures. For isolation of brain-infiltrating leukocytes (BILs), mice were sacrificed by CO2 asphyxia, then immediately perfused through the left cardiac ventricle with isotonic Ringer’s solution. Brains were removed and BILs were isolated by enzymatic digestion and modified Ficoll-Hypaque centrifugation, as previously described (28). The enzymes used in this procedure do not degrade the principal surface receptors on T cells, allowing ex vivo functional and phenotypic analyses to be performed without any preculture of cells.
Flow cytometry
The CW3-specific immune response was assessed by quantifying BV10+CD62L-CD8+ cells (24). PBLs, splenocytes, lymph node cells, or BILs were triple stained with FITC-conjugated CD62L (Mel-14; ImmunoKontakt, Frankfurt, Germany), PE-conjugated CD8 (53-6.7; PharMingen, San Diego, CA), and biotinylated anti-BV10b (B21.5; PharMingen) revealed with streptavidin-tricolor (Caltag Laboratories, Burlingame, CA). Additional analyses were performed using CD4 PE (KT6; Serotec, Oxford, U.K.); CD11b FITC (M1/70; PharMingen); CD4 biotin (RM4-4; PharMingen), or CD11b biotin (M1/70, PharMingen), followed by streptavidin-tricolor (Caltag). Samples were analyzed using a FACScan equipped with CellQuest software (Becton Dickinson, Mountain View, CA). Where means for pooled data are given, SEM are also indicated.
Immunohistology
Brains (perfused as described above) were snap frozen in 2-methylbutane (Merck, Wertheim/Main, Germany) and stored at -80°C until used for sectioning. Coronal sections (7 µm) were made and fixed with 2% paraformaldehyde. After quenching of endogenous peroxidase, sections were stained with primary Abs. The following rat mAb were used as tissue culture supernatants prepared in our laboratory: anti-BV10 (B21.5) (29), CD4 (GK1.5, ATCC TIB207), CD8 (H35-17.2) (30), and antimacrophage/microglia (F4/80, ATCC HB 198). The secondary Ab was goat anti-rat Ig coupled to HRP (BioSource International, Camarillo, CA). Purified polyclonal rabbit antisera were used to detect glial fibrillary acidic protein (Dako Diagnostics, Zug, Switzerland) and P1Ap (kindly supplied by Dr. A. Amar-Costesec) (31), followed by donkey anti-rabbit Ig coupled to HRP (Jackson ImmunoResearch, West Grove, PA). Peroxidase activity was revealed by adding a freshly prepared solution of AEC substrate (3-amino-9-ethyl-carbazole; Sigma, Buchs, Switzerland) and H2O2. Sections were counterstained with hematoxylin and eosin.
In vivo depletion of CD4 T cells
In some experiments, mice were depleted of CD4 T cells by i.p. injection of 0.2 mg of CD4 mAb (GK1.5, ATCC TIB 207) on days -4, -2, and +3, relative to P815-CW3 implantation. Depletion to at least 95% was confirmed by flow cytometry of PBL the day before P815-CW3 implantation using a noncompeting Ab (RM4-4, described above).
Cytolytic assay
P815 cells (or transfected P815 cells) were labeled with 150 µCi of sodium [51Cr]chromate, as previously described (27) for 1 h at 37°C and washed three times. For peptide pulsing, 10 µM of peptide CW3170–179 was added during the labeling procedure. A total of 1200 51Cr-labeled target cells was mixed with varying numbers of freshly isolated effector cells in round-bottom microplates, the cells being resuspended in DMEM supplemented with 5% FCS and HEPES. 51Cr release in supernatants was measured after 4 h of incubation at 37°C. The percent specific lysis was calculated as: 100 x ((experimental - spontaneous release)/(total - spontaneous release)).
Size analysis of complementarity-determining region 3 (CDR3) of the TCR
The CDR3 region of the PCR-amplified TCR BV4 and BV10 transcripts was analyzed using a run-off procedure, as previously described (8, 32, 33). Briefly, total RNA was prepared from perfused brain using TRIzol (Life Technologies, Paisley, U.K.) and converted to cDNA by standard methods using reverse transcriptase and an oligo(dT) primer. These cDNAs were amplified using validated 5' sense primers specific for either BV4 or BV10 and one 3' antisense primer specific for the BC gene segment (33). Aliquots (2 µl) of BV4-BC or BV10-BC PCR products were subjected to a five-cycle run-off reaction, using a dye-labeled oligonucleotide primer specific for the BC segment. The run-off products were then run on an automated sequencer in the presence of fluorescent size markers. The length of the DNA fragments and the fluorescence intensity of the bands were analyzed with Immunoscope software (developed by C. Pannetier, Pasteur Institute, Paris, France).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Preliminary experiments in which different numbers of P815-CW3 cells were implanted i.c. indicated that there were few external clinical signs until 5 x 105 cells were used. At this cell dose, most mice exhibited some transient weight loss and mild clinical symptoms (ruffled fur, hunched posture). This cell dose was used for all subsequent experiments. Neither mice injected i.c. with PBS, nor mice injected s.c. with P815-CW3 exhibited weight loss or other symptoms.
Using flow cytometry, we screened blood and lymphoid tissue of mice
implanted with P815-CW3 cells for evidence of CW3-specific CTLs
(BV10+CD62L-CD8+
cells). As previously reported for i.p. immunized mice, i.c.
implantation of P815-CW3 also resulted in a significant expansion of
CW3-specific CTLs, in both spleen and PBL (Fig. 1
A). Lymph nodes were also
examined (axillary, inguinal, and cervical), but generally no
significant proportions of CW3-specific CTLs were detectable (data not
shown).
|
B) as well as the spleen (not shown). Indeed, as early as 7
days postimplantation (d.p.i.), i.c. implanted mice showed an elevated
proportion of CW3-specific CTLs in PBL that was significantly higher
than in s.c. implanted mice (p = 0.001,
Mann-Whitney rank sum test). However, the response did not reach the
same magnitude and declined faster than the slower response induced
after s.c. implantation. Intraperitoneal implantation shows similar
kinetics and magnitude to s.c. implantation (data not shown and Ref.
24), and i.v. injection results in either an undetectable
(in most mice) or a very low level response (data not shown). Regression of i.c. P815-CW3 tumor correlates with leukocyte infiltration of the brain parenchyma
Immunohistology of brain cryosections 5 days postimplantation
revealed significant tumor growth in the ipsilateral hemisphere
relative to the site of tumor implantation; this could be readily
visualized by the morphology of the P815-CW3 cells and by positive
staining with antiserum specific for the P815 tumor Ag P1Ap (Fig. 2A). At this time point, there
was also significant staining in the ipsilateral hemisphere with F4/80
mAb, specific for macrophages, monocytes, and microglial cells (not
shown). Strongest F4/80 staining was in the peritumoral region, but
some F4/80+ cells also infiltrated the tumor
mass. Only rare CD4 or CD8 T cells were detected at 5 d.p.i. (not
shown).
|
B), but
there was significant infiltration of BV10+ T
cells (Fig. 2, C and D), CD8 T cells (not shown),
and CD4 T cells (Fig. 2, E and F). The presence
of F4/80+ cells also persisted at day 8, with
similar localization to that found for infiltrating lymphocytes (not
shown). The vast majority of infiltrating T cells was found in the
ipsilateral hemisphere, in perivascular locations as well as
infiltrating the brain parenchyma. In some sections, the site of tumor
implantation could be localized and this region was usually well
infiltrated (sometimes intensely) with leukocytes.
Brain sections were also examined at later time points, at 11, 14, 17,
21, and 22 d.p.i.; tumor cells could not be detected at these time
points, but the inflammatory infiltrate persisted, albeit at lower
levels. The other notable difference to the brains taken 8 d.p.i.
was that infiltrating cells were widely distributed throughout both
cerebral hemispheres; this is illustrated for
BV10+ T cells (Fig. 2, G and
H), and comparable results were found for cells positive for
CD4, CD8, and F4/80 (at time points from day 11 onward). Cells were
found in most regions of the brain: intraparenchymal, perivascular,
periventricular, and sometimes in the lateral ventricles.
Dominant oligoclonal expansions of BV10+ T cells expressing a 6-aa CDR3 region are detected in brains of mice implanted i.c. with P815-CW3
Two of the principal features of the T cell repertoire specific
for the H-2Kd-restricted immunodominant epitope
CW3170–179 expressed by P815-CW3 cells are the
BV10 selection and the usage of a 6-aa CDR3 region (23, 34). The flow cytometry and immunohistochemistry data indicated
that BV10+ cells had been expanded; the CDR3 size
was examined by TCR spectratyping (Fig. 3). Isolation of RNA was conducted on
brains that had been well perfused to eliminate leukocytes present in
vessels. RT-PCR was performed using primers specific either for BV10 or
BV4 TCR gene segments, the latter having no known preferential usage in
the response to P815-CW3. The curves obtained for BV4 and BV10 for
spleen from control mice (Fig. 3) showed a bell-shaped form, consistent
with polyclonal populations of T cells, with most CDR3 sizes
represented. However, for brains of mice previously implanted i.c. with
P815-CW3, curves obtained for BV10 (Fig. 3) showed a single predominant
peak corresponding to 6 aa, consistent with an expanded population of
BV10+ T cells expressing an appropriate TCR for
recognition of CW3170–179, whereas the profile
for BV4 was similar to that found for control mice.
|
BILs were isolated from mice implanted i.c. 8 days previously with
P815-CW3, a time point at which immunohistochemistry data indicated
that there were high numbers of leukocytes infiltrating the brain
parenchyma and at which a high level systemic immune response (>50%
BV10+ cells in the
CD62L-CD8+ subset) was
detectable in the PBL of most (36/47) mice tested. An average of
106 leukocytes was isolated from the brain of
each mouse; isolations from control mice that had not been implanted
i.c. yielded too few cells to count accurately, but probably no more
than 5 x 104 cells per brain. The enzymes
(collagenase D and DNase I, together with the trypsin inhibitor
N
-p-tosyl-L-lysine chloromethyl
ketone) and method of dissociation had no effect on expression levels
of CD4, CD8, BV10 TCR, CD62L, or CD11b (as judged by preliminary
experiments with control cells), nor did enzyme treatment affect
cytotoxic function of a control CTL clone. Therefore, the phenotype and
function of isolated BILs were tested immediately after isolation,
allowing conclusions to be made regarding the likely in vivo
characteristics of these cells.
Among mice showing a high level systemic response to CW3, flow
cytometric analysis showed that the immune infiltrate consisted of
mainly CD8 cells, CD4 cells, and CD11b+ cells
(macrophages or microglial cells). Only very low numbers of B cells, NK
cells, and granulocytes were found in preliminary experiments, so in
view of the limited number of cells available for analysis, these
populations were not routinely stained. Consistent with the notion that
only activated T cells can enter the CNS (35), more than
95% of both CD4 and CD8 cells were negative for CD62L expression.
Analysis of BILs triple stained with CD8, CD62L, and anti-BV10 TCR
Abs (Fig. 4A) showed that a
high proportion of the
CD62L-CD8+ cells used a
BV10 TCR (generally more than 70%), whereas BV10 usage by CD4 cells
was always less than 10%. This result (also consistent with the
immunohistochemistry and TCR molecular analysis) encouraged us to test
the specific cytotoxic function of BILs ex vivo against P815 cells
pulsed with a synthetic peptide corresponding to the immunodominant CTL
epitope CW3170–179 (Fig. 4B). Freshly
isolated BILs were highly and specifically cytotoxic, killing P815
cells only when the specific peptide was present. Furthermore, when E:T
ratios were calculated based on the numbers of
BV10+CD62L-CD8+
cells (thus encompassing the vast majority of cells defined in many
previous studies as being able to recognize
CW3170–179), this high cytotoxic activity was
manifested at even modest E:T ratios. These results are therefore
consistent with the BV10+
CD8+ T cells infiltrating the brain parenchyma
(detected in vivo by various techniques) as being fully
differentiated CTLs.
|
It is not possible to generalize about the CD4 dependence of CTL
responses in the brain, because very few studies have looked at this
aspect in systems enabling the analysis of specific CTL responses. We
therefore investigated this issue in the response to P815-CW3,
depleting mice of CD4 T cells before tumor cell implantation and
maintaining the depleted state throughout the response. Intraperitoneal
injection of CD4 mAb (GK1.5) resulted in CD4 T cell depletion to more
than 95% of normal CD4 T cell numbers (in PBL), as tested by flow
cytometry the day before i.c. implantation of P815-CW3. A further i.p.
injection of depleting CD4 mAb was administered 3 d.p.i., and at
sacrifice virtually no CD4 T cells were detected in the brain
parenchyma, by flow cytometry or immunohistochemistry (data not shown).
However, the absence of CD4 T cells did not inhibit the infiltration of
BV10+ cells into the brain parenchyma, as
observed by immunohistochemical staining of day 8 brain cryosections
(Fig. 5A). Furthermore, BILs
could be isolated from brains of depleted mice, exhibiting comparable
phenotype and function as for nondepleted mice (Fig. 5, B
and C). The majority of mice tested (9/13, 69%) showed a
comparable infiltration of CTLs to nondepleted mice.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The most basic question we address in this study is that of immune
function in the CNS. Although P815-CW3 cells efficiently induce
CW3-specific CTLs in other sites (24, 25, 36), it was
unknown whether this would be the case in the CNS. Indeed, the many
reports of immunosuppression associated with brain tumors (37, 38) and even with normal brain microenvironment (19, 39, 40) made this question highly relevant. The short-term growth of
i.c. tumors followed by their regression (Fig. 2), accompanied by a
systemic specific immune response (Fig. 1) and minor clinical symptoms,
suggested that tumor-specific immunity was occurring. What was
surprising was the rapidity of the specific response compared with that
achieved after s.c. implantation (Fig. 1B) or after i.p.
implantation (24), as measured by flow cytometry of PBL
and splenocytes. However, these data only reflect the tissue that is
tested and may not detect expanded populations of cells at the tumor
site or in draining lymphoid tissue. Nevertheless, for i.c. implanted
mice, a strong leukocyte infiltration of the brain parenchyma was
detected by immunohistochemistry at about 8 d.p.i. (Fig. 2). We
sequentially refined the identification of the infiltrating cells by 1)
staining with mAb to BV10 (Fig. 2); 2) TCR molecular analysis of
perfused brain tissue that revealed expanded populations of
BV10+ T cells with a 6-aa CDR3 region (Fig. 3);
3) flow cytometric analysis of leukocytes isolated from perfused brain,
indicating large numbers of
BV10+CD62L-CD8+
T cells (Fig. 4A); and finally 4) ex vivo functional
analysis of leukocytes isolated from perfused brain that efficiently
killed P815 cells pulsed with peptide CW3170–179
or P815-CW3 (not shown), but not unmodified P815 cells without peptide
(Fig. 4B). Taken together, these data convincingly indicate
that large numbers of CW3-specific CTLs were induced after i.c.
implantation of P815-CW3 and that they infiltrated the brain and were
specifically cytotoxic.
Analysis of the immunohistochemistry data gives further insight into the likely sequence of events leading to tumor elimination. The tumor is clearly visible at day 5 after implantation, but at this stage there are very few T cells infiltrating the brain. Moreover, at this stage in blood or spleen, there is no detectable expansion of BV10+CD62L-CD8+ T cells by flow cytometry or of BV10+ cells with a 6-aa CDR3 by TCR spectratyping (data not shown). However, there is evidence for innate immune reactions, because many F4/80+ cells are present. Cells of this phenotype can correspond to resident parenchymal microglial cells as well as perivascular and recruited peripheral macrophages (41, 42). The presence of these microglia/macrophages is probably a critical step for lymphocyte infiltration, because in macrophage-depleted mice, activated leukocytes can extravasate from blood across CNS endothelium, but cannot penetrate the parenchyma. Instead, they accumulate in the perivascular and subarachnoid spaces (43). However, microglia/macrophage activation and recruitment may not be the exclusive mechanism responsible for the development of the early immune response, because astrocytes may also play a role (44, 45, 46). Indeed, glial fibrillary acidic protein expression by astrocytes was high in the ipsilateral hemisphere at 5 d.p.i. (data not shown), one of the signs of reactive gliosis that accompanies many forms of CNS stress (47). Astrocytes in such an activated state can participate in immune reactions by up-regulation of MHC and adhesion molecules (44, 48, 49) and liberation of inflammatory chemokines and cytokines (45, 46, 50, 51). Analysis of events even nearer to the day of implantation is difficult to interpret in a transplanted tumor model because of the inevitable physical trauma associated with implantation, but these issues of the initiating events for an antibrain tumor response are evidently important to address in future studies.
The significant lymphocyte accumulation at the tumor site by 8
d.p.i. also corresponds to the peak of weight loss and the appearance
of mild clinical symptoms in some mice. Because by this time the tumor
burden is reduced compared with day 5, the immune response rather than
the tumor may be responsible for the symptoms. Very few lymphocytes are
present in the normal brain (or indeed in the nontumor-laden
contralateral hemisphere of the implanted mice investigated in this
study; Fig. 2), but activated lymphocytes can traverse the blood brain
barrier (35), for example, as may occur after initial T
cell priming at a peripheral site. However, whether Ag or intact tumor
cells (19) reach secondary lymphoid organs is far from
clear and is not an issue that has been directly addressed in this
study. Once an inflammatory response is initiated in the CNS, naive T
cells may also contribute to the immune infiltrate
(52).
The initially restricted localization of infiltrating T cells is consistent with retention at the site of highest Ag concentration. The MHC class I-restricted CTLs can presumably interact directly with H-2Kd-expressing P815 cells, whereas the abundant CD4 cells may recognize Ags expressed by Ia+ microglial cells or macrophages, although the specificity of the CD4 cell component of this response has not been characterized. It was also suggested that there is not only recruitment and retention of specific T cells, but also the possibility of local expansion (53). The more widespread distribution of the immune infiltrate at 11 d.p.i. or later, after most of the tumor cells have been eliminated (as judged by immunohistochemistry), may be a consequence of the lack of retention of cells in the ipsilateral hemisphere due to the less frequent encounters with Ag. Studies of viral immune responses have indicated the relatively long-term persistence of CD8 T cells in the CNS, even when only noninfectious viral Ags remain (54), although these CTLs had lost their cytolytic capacities. In the model we describe in this work, it is possible that once P815-CW3 cells have been eliminated, tumor Ag might persist on APC in the brain even 3 wk after implantation. However, by this stage there were too few CTLs for us to isolate to examine their function.
Depletion of CD4 T cells appeared to have little effect on the
infiltration and function of CW3-specific CTLs (Fig. 5). This aspect of
the P815-CW3 response has previously been investigated in other sites,
with the response dependent upon CD4 cells when P815 cells were
injected i.p., but CD4 independent when they were injected
intradermally (25). It was suggested that the abundant
epidermal Langerhans cells and dermal dendritic cells (DC) in the skin
(compared with the less efficient peritoneal macrophages predominant in
the peritoneum) were able to take up Ag from P815-CW3 and prime
specific CTLs even in the absence of CD4 cells; however, such
cross-priming has not yet been directly demonstrated for P815-CW3. As
far as the CNS is concerned, we can only speculate whether a local
professional APC can capture Ag and migrate to lymphoid tissue in the
absence of CD4 activation (or whether intact P815 cells exit the brain,
as discussed previously). Classical DC are not usually found in the
brain, although histological reports have noted their presence in the
choroid plexus (55, 56). Perhaps a more probable candidate
for capturing CNS Ag is the perivascular cell, which can take up Ag via
scavenger receptor-mediated endocytosis and subsequently migrate toward
lymphoid tissue (recently discussed in Ref. 57), although
any CD4 dependence of these mechanisms has not been reported.
Because we depleted CD4 cells before tumor implantation and maintained the depleted state throughout the tumor rejection phase, CTL effector function in the brain parenchyma was also independent of CD4 cells. This is not unprecedented, because in one interesting recent report, the CTL response against B16 melanoma implanted i.c. was also CD4 independent, although only the effector stage of this response was studied (13). Moreover, a CD4 independence at both induction and effector stages of a CD8 T cell antitumor response was also found in an immunization model against an i.c. sarcoma, but in this system, endogenous APC may have been bypassed because the vaccine consisted of cultured, Ag-pulsed DC (22). However, antiviral CTL responses in the CNS are often strictly CD4 dependent, and CTL function and/or viability have been reported to be lost in the absence of CD4 T cells (58, 59). Maintenance of CTL function in the brain is clearly dependent upon many factors that differ according to the model used. In the P511 mastocytoma model, CD8 T cells were unable to differentiate into effector cells in the brain microenvironment (19), subsequently attributed to a sensitivity to TGF-ß present in the cerebrospinal fluid and the interstitial fluid (40). However, most of the experiments were performed in outbred or nonsyngeneic mice, making it difficult to compare with the syngeneic response to a defined peptide Ag that we describe in this work.
In view of the advances in our understanding of antitumor immune responses, it is now tempting to contemplate the development of immunotherapies for spontaneous cerebral malignancies in humans, particularly for tumors such as glioblastoma, for which no adequate treatment exists. Typically, these approaches can be expected to modify both specific and nonspecific components of the immune response. This has particular significance for the CNS, which has probably acquired mechanisms to limit inflammation, essential in a site that has severe physical limitations within the confines of the skull to accommodate tissue swelling. Uncontrolled inflammation can lead to severe neurological dysfunction and may contribute to the initiation or perpetuation of autoimmune disease. Indeed, early tumor immunology studies noted that lethal allergic encephalomyelitis was induced in different species by immunization with human glioma tissue (60). Furthermore, very recent evidence in a rodent gene therapy model for glioblastoma suggested that i.c. immune responses may directly or indirectly perpetuate reactive gliosis and demyelination (61). Taken together, such observations argue for a thorough understanding of not only the cellular subsets required for antitumor effects, but also for the specificity of these responses. The data from the P815-CW3 model that we describe in this work suggest that a highly focused response can mediate potent specific cytotoxicity against tumor cells in the brain, even in the absence of CD4 T cells. Moreover, this antitumor activity occurred without untoward clinical symptoms. This model opens the possibility to explore the factors that preclude such an efficacious response against spontaneous glioblastoma, and how they may be overcome in future immunotherapies.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Paul R. Walker, Division of Oncology, University Hospital Geneva, rue Micheli-du Crest 24, 1211 Geneva 14, Switzerland.
3 Abbreviations used in this paper: i.c., intracerebral; AEC, 3-amino-9-ethyl-carbazole substrate; BIL, brain-infiltrating leukocyte; CDR, complementarity-determining region; DC, dendritic cell; d.p.i., days postimplantation.
Received for publication February 23, 2000. Accepted for publication June 22, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
8-, Vß2- or Vß10-bearing T cells. Eur. J. Immunol. 21:3035.[Medline]
. J. Immunol. 149:2358.[Abstract]
This article has been cited by other articles:
|
|
 |
|
  M. Fujita, X. Zhu, R. Ueda, K. Sasaki, G. Kohanbash, E. R. Kastenhuber, H. A. McDonald, G. A. Gibson, S. C. Watkins, R. Muthuswamy, et al. Effective Immunotherapy against Murine Gliomas Using Type 1 Polarizing Dendritic Cells--Significant Roles of CXCL10 Cancer Res., February 15, 2009; 69(4): 1587 - 1595. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Fujita, X. Zhu, K. Sasaki, R. Ueda, K. L. Low, I. F. Pollack, and H. Okada Inhibition of STAT3 Promotes the Efficacy of Adoptive Transfer Therapy Using Type-1 CTLs by Modulation of the Immunological Microenvironment in a Murine Intracranial Glioma J. Immunol., February 15, 2008; 180(4): 2089 - 2098. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Masson, T. Calzascia, W. Di Berardino-Besson, N. de Tribolet, P.-Y. Dietrich, and P. R. Walker Brain Microenvironment Promotes the Final Functional Maturation of Tumor-Specific Effector CD8+ T Cells J. Immunol., July 15, 2007; 179(2): 845 - 853. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Eguchi, M. Hatano, F. Nishimura, X. Zhu, J. E. Dusak, H. Sato, I. F. Pollack, W. J. Storkus, and H. Okada Identification of Interleukin-13 Receptor {alpha}2 Peptide Analogues Capable of Inducing Improved Antiglioma CTL Responses Cancer Res., June 1, 2006; 66(11): 5883 - 5891. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Nishimura, J. E. Dusak, J. Eguchi, X. Zhu, A. Gambotto, W. J. Storkus, and H. Okada Adoptive Transfer of Type 1 CTL Mediates Effective Anti-Central Nervous System Tumor Response: Critical Roles of IFN-Inducible Protein-10. Cancer Res., April 15, 2006; 66(8): 4478 - 4487. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Kuwashima, F. Nishimura, J. Eguchi, H. Sato, M. Hatano, T. Tsugawa, T. Sakaida, J. E. Dusak, W. K. Fellows-Mayle, G. D. Papworth, et al. Delivery of Dendritic Cells Engineered to Secrete IFN-{alpha} into Central Nervous System Tumors Enhances the Efficacy of Peripheral Tumor Cell Vaccines: Dependence on Apoptotic Pathways J. Immunol., August 15, 2005; 175(4): 2730 - 2740. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Akasaki, G. Liu, N. H. C. Chung, M. Ehtesham, K. L. Black, and J. S. Yu Induction of a CD4+ T Regulatory Type 1 Response by Cyclooxygenase-2-Overexpressing Glioma J. Immunol., October 1, 2004; 173(7): 4352 - 4359. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Calzascia, W. Di Berardino-Besson, R. Wilmotte, F. Masson, N. d. Tribolet, P.-Y. Dietrich, and P. R. Walker Cutting Edge: Cross-Presentation as a Mechanism for Efficient Recruitment of Tumor-Specific CTL to the Brain J. Immunol., September 1, 2003; 171(5): 2187 - 2191. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) or P815 cells pulsed with peptide CW3170–179 (•)
as targets. The E:T ratio was calculated either using the total number
of lymphocytes added to the wells (All Cells), or the number of
BV10+CD62L-CD8+ cells, estimated
from the flow cytometric data. Similar levels of specific cytotoxicity
were obtained in nine independent assays.