T Cell Response + Protective Cd8 Qs-21 and Il-12 Adjuvants Induce a Tcr Vaccines against T Cell Lymphoma

Tumor-specific TCR can serve as an effective target for active immunotherapy of T cell malignancies. Using the murine T cell tumor model C6VL, vaccination with C6VL TCR protected mice from a subsequent lethal dose of tumor cells. This study characterizes the immune mechanisms involved in the tumor protection, and the influence of immunologic adjuvants in inducing a protective immune response. Immune responses induced by TCR vaccines formulated with various adjuvants: QS-21, IL-12, SAF-1, CD40L, and GM-CSF were compared. QS-21, IL-12, and SAF-1 biased the humoral immune response toward Th1-type, reflected by the induction of IgG2a and IgG2b anti-C6VL TCR Abs. CD40L and GM-CSF exclusively produced IgG1 Abs, reflecting a Th2-type immune response. In our tumor model system, only vaccines containing adjuvants that induced a Th1-type immune response favored tumor protection. Furthermore, we demonstrated that CD8+ T cells were necessary and sufficient for tumor protection using anti-CD8 mAb depletion and adoptive cell transfer experiments. Transfer of hyperimmune serum containing anti-C6VL TCR Abs into na ive mice had modest anti-tumor effects and was not sufficient to prevent tumor growth. TCR-vaccinated B cell-deficient mice were not protected against C6VL tumor, and tumor protection was not completely restored after hyperimmune serum transfer. Thus, B cells may serve as important APCs in inducing a protective immune response. Based on these results future TCR vaccines should be designed to maintain native TCR conformation, as well as induce a strong Th1-type immune response.

T he V region of a TCR is encoded by a unique combination of V, (D), and J gene segments.This recombination process gives rise to a TCR with unique determinants (Id) that is expressed on the surface of T lymphocytes.Since most T cell malignancies are derived from the clonal proliferation of T lymphocytes, tumors' TCR can serve as specific targets for antitumor immunotherapy.The Id determinants on TCR have many characteristics that are similar to surface Ig on B lymphocytes.The use of B cell Id in tumor-specific vaccines in the treatment of B cell malignancies has been well described.In murine B cell lymphoma models, active immunotherapy using tumor Id proteins resulted in the induction of protective immune responses that prevented tumor growth (1)(2)(3)(4).Various clinical trials are currently ongoing in which B cell lymphoma patients receive tumor Id vaccines (5)(6)(7).
Given the encouraging results of B cell Id vaccines in the treatment of B cell malignancies and the similarities between B and T cell Ag receptors, we have established and characterized a murine T cell lymphoma model system for the development of TCR vaccines (8).The use of TCR as immunogen, especially the use of V region peptides, has been well documented in the treatment of autoimmune diseases (9 -12).However, the immune response induced by V region peptide therapy resulted in the deletion or suppression of normal T cells in addition to self-reactive T cells hav-ing the same V regions (13)(14)(15), which is undesirable in antitumor immunotherapy.To develop an antitumor vaccine that would induce an immune response specific for tumorigenic T cells without resulting in the deletion of normal T cells, we reported the use of soluble, heterodimeric TCR as immunogens in vaccines for active tumor immunotherapy that was Id specific (8).One of the major obstacles in evaluating the efficacy of tumor-specific TCR as an active immunotherapy vaccine was the production of sufficient quantities of soluble, heterodimeric TCR proteins.Since TCR proteins are not secreted, we have cloned and overexpressed a recombinant TCR derived from the murine T cell tumor C6VL as a phophatidylinositol-linked (PI) 3 protein on transfected cells (8,16).After digestion with PI-specific phospholipase, large quantities of soluble C6VL TCR proteins were obtained and purified.In our previous study, we demonstrated the antitumor effects of TCR vaccine in active immunotherapy of the murine T cell tumor C6VL.Mice immunized with soluble C6VL TCR conjugated to keyhole limpet hemocyanin (KLH) in the presence of syntex adjuvant Formulation-1 (SAF-1) induced a strong C6VL TCR-specific humoral immune response.TCR-vaccinated mice were protected against a lethal dose of C6VL tumor cells.No deleterious effect was observed on normal T cell populations, nor skewing of the TCR repertoire, and the protection was specific against C6VL and not other T cell tumors.
In this study, we have characterized the immune mechanisms involved in tumor protection of vaccinated mice.Different immunologic adjuvants were used in an attempt to improve the efficacy of TCR vaccines, and to understand the correlation between tumor protection and the quality of immune response induced.We have found that adjuvants that bias the immune response toward Th1type, reflected by the induction of IgG2 Abs and effector CD8 ϩ T cells, confer the best protection.We have further found that CD8 ϩ T cells are necessary and sufficient for tumor protection in this model system, and that anti-C6VL TCR Abs provide only mild antitumor effects.Understanding of the mechanisms of tumor protection induced by the TCR vaccines will facilitate the design of future immunotherapies for T cell malignancies.

C6VL TCR purification
Recombinant C6VL ␣␤-TCR was obtained as previously described (8).Briefly, C6VL TCR ␣ and ␤ genes were isolated and then modified such that the transmembrane regions of the C regions were replaced with sequences from 3Ј region of the decay-accelerating factor (DAF) gene, which encodes for a PI linkage domain (20).BW5147 cells transfected with the modified C6VL ␣␤-TCR/DAF genes expressed C6VL TCR as a PI-linked protein on the cell surface.The C6VL TCR/DAF expression was amplified with methotrexate selection.A high TCR-expressing clone (1D6-480-4) was used for the source of the recombinant C6VL TCR.TCR protein was released from the cell surface by incubating the TCR transfectants with PI-phospholipase C (PLC), a phospholipase specific for the PI linkage.Soluble C6VL ␣␤-TCR was purified from the PI-PLC-digested supernatants with a H57-597 (mAb anti-C␤) Sepharose affinity column.Affinitypurified TCR protein was concentrated using Con A lectin Sepharose affinity column (Sigma Chemical Company, St. Louis, MO).Purified C6VL ␣␤-TCR was dialyzed extensively against PBS and then filter sterilized.TCR concentration was determined by BCA protein assay (Pierce Chemical Co., Rockford, IL).

TCR immunizations
Purified C6V1 ␣␤-TCR were chemically conjugated to KLH (Calbiochem, San Diego, CA) at 1:1 ratio (w/w) using glutaraldehyde as previously described (2).Mice were immunized s.c. with TCR-KLH conjugates containing 35 g (500 pmol) TCR in various adjuvant formulations in a total volume of 200 l PBS.SAF-1 was prepared as previously described (2,21).SAF-1 is composed of 0.2% v/v Tween-80, 5% v/v squalene, 2.5% v/v Pluronic L121, and 100 g/ml Thr 1 -MDP.Thr 1 -MDP was included only in the first immunization.QS-21 (kindly provided by Aquila Biopharmaceuticals, Framingham, MA) is a purified saponin-based adjuvant from Quillaja saponaria Molina extracts (22,23), and was given at 10 g per injection.Murine rIL-12 (kindly provided by Stan Wolf, Genetics Institute, Cambridge, MA) was given at 0.5 g per injection.Both recombinant murine granulocyte/macrophage CSF (GM-CSF) and soluble trimeric murine CD40 ligand (CD40L) were kindly provided by Immunex Corporation (Seattle, WA).CD40L was given at 2.5 g per injection.GM-CSF was given in four 1-g doses per immunization, with 1 g GM-CSF coinjected with TCR-KLH, followed by three 1-g doses given 1, 2, and 3 days after TCR immunization (24).The irrelevant protein vaccine control consisted of Id protein derived from the murine B cell lymphoma 38C13, similarly conjugated to KLH and given in equal molar amount in adjuvants (2).Immunizations were given three times at 2-wk intervals.Serum samples were collected 10 days after each immunization.

Anti-C6VL ␣␤-TCR ELISA assay
ELISA assays were performed as previously described (8).Briefly, 96-well Maxisorb plates (Nunc, Naperville, IL) were coated with anti-C␤ mAb H57-597.Purified C6VL ␣␤-TCR diluted in PBS containing 2% BSA was bound to H57-597-coated plates.Mouse immune serum was diluted and titered over eight wells in twofold dilutions.A standard curve was generated by titering mAb 124-40, a mouse IgG anti-C6VL TCR clonotype encoded on ␣-chain V region (18).Bound mouse Abs were detected using a peroxidase-conjugated donkey anti-mouse Ig Ab (Jackson Immuno-Research, West Grove, PA).After removal of unbound proteins, substrate solution containing 2,2Ј-azinobis(3-ethyl-benzthiazoline sulfonic acid (ABTS) was added to each well.The color reaction was allowed to develop at room temperature.Absorbance at 405-450 mm was measured using a V max microplate reader (Molecular Devices, Menlo Park, CA).Anti-C6VL ␣␤-TCR Ab titers were determined from the linear portion of the standard curve.

Affinity elution with ammonium thiocyanate
ELISA assays were performed as described above with the following modification (25,26).Ninety-six-well microtiter plates were coated with 2 g/ml purified C6VL ␣␤-TCR in 50 mM carbonate buffer (pH 9.0).Pooled immune serum from groups of 10 mice each vaccinated with C6VL ␣␤-TCR in various adjuvants was incubated on the TCR-coated plates for 1 h at room temperature at 0.1 g/ml anti-C6VL ␣␤-TCR Ab (titers were determined from previous ELISA assays).The Ab concentration was chosen such that the absorbance readings were near the top of the linear portion on the standard curves, and were equivalent among different adjuvant groups.After washing, 100 l of ammonium thiocyanate (NH 4 SCN) in PBS ranging from 0 to 3 M was added to replicate wells.The plates were incubated for 15 min at room temperature and washed.Bound Abs were detected using a peroxidase-conjugated donkey anti-mouse Ig Ab.The absorbance reading in the presence of 0 M NH 4 SCN was defined as 100% initial binding.Affinity index was defined as the molar concentration of thiocyanate required to reduce the initial absorbance reading by 50% (25,26).

Isotype-specific ELISA
ELISA assays were performed as described above with the following modification.Pooled serum from groups of 10 mice each were captured on TCR-coated plates.After 1-h incubation, bound mouse Abs were detected by using peroxidase-conjugated goat anti-mouse IgG1, IgG2a, IgG2b, and IgG3 Abs (Southern Biotechnology Associates, Birmingham, AL).Relative units were calculated using a pooled serum standard, with the starting dilution of the standard arbitrarily set as 1.All serum samples were calibrated similarly against the serum standard.

Tumor challenge
Two weeks after the third protein immunization, mice were challenged with a lethal dose of C6VL cells.A frozen aliquot of C6VL cells was thawed and was grown in RPMI 1640 containing 10% FCS and 50 m 2-ME for 3 days.The tumor cells were collected, washed three times in HBSS, and diluted to 1 ϫ 10 4 cells/ml.Mice were injected i.p. with 5000 C6VL tumor cells in 500 l HBSS.Survival of mice challenged with tumor was monitored for at least 60 days after tumor injection.Survival curves were generated using the Kaplan-Meier method.Statistical analysis was done using the log-rank test.

Transfer of hyperimmune serum
Hyperimmune serum was obtained from 40 C57BL/6 mice immunized three times with C6VL ␣␤-TCR-KLH in QS-21.Serum was collected and pooled.The hyperimmune globulin fraction was precipitated by adding a saturated ammonium sulfate solution slowly to the serum to a final concentration of 45% v/v, and incubated overnight at 4°C.Precipitated proteins were pelleted by centrifugation at 20,000 ϫ g for 30 min at 4°C.The protein pellet was resuspended in a minimal amount of PBS and dialyzed extensively in PBS.The proteins were further concentrated using Centriplus-30 concentrators (Amicon, Beverly, MA).Final anti-TCR Ab titers as well as total IgG concentrations were determined by ELISA as described above.Recipient mice were injected i.p. with 500 l concentrated hyperimmune globulin protein.One day after serum transfer, mice were challenged with C6VL tumor cells and followed for survival.For negative Ab control, serum from either naı ¨ve mice or mice vaccinated three times with 38C13 Id-KLH in QS-21 was collected and similarly concentrated.Control mice were injected with 500 l concentrated irrelevant globulin protein containing equivalent amount of IgG as compared with the anti-TCR hyperimmune globulin protein.Mice were bled 2 h before tumor challenge and the level of circulating anti-TCR Ab titer was determined by ELISA.

In vivo depletion of CD8 ϩ T cells
C57BL/6 mice were depleted of CD8 ϩ T cells using the anti-CD8 mAb 2.43 (rat IgG2b).Control mice were injected with an irrelevant isotypematched mAb SFR8-B6 (anti-human HLA Bw6).Both 2.43 and SFR8-B6 hybridomas were grown as ascites in pristane-primed nude mice, mAbs were harvested as ascitic fluid, diluted in PBS, and filter sterilized.Concentration of mAbs was determined using a rat IgG2b-specific ELISA assay.Mice were injected i.p. with 250 g mAb in 500 l PBS on days 6, 5, and 4 before tumor challenge.Three weekly injections were given posttumor challenge to maintain the depletion, starting 1 wk after the third mAb injection.The extent of CD8 ϩ T cell depletion in peripheral blood was analyzed by flow cytometry 1 day before tumor challenge and 3 days after the last weekly mAb treatment (22 days post-tumor challenge) using a nonblocking anti-CD8 mAb (PharMingen, San Diego, CA).

Adoptive transfer of T cell-enriched lymphocytes
Fifteen C57BL/6 donor mice were immunized three times with C6VL ␣␤-TCR-KLH in QS-21 as described above.Control donor mice were similarly immunized with 38C13 Id-KLH in QS-21.Ten days after the last immunization, donor mice were sacrificed to harvest spleens and inguinal lymph nodes.Immune lymphocytes from donors were pooled.The lymphocyte preparation was depleted of contaminating RBCs by resuspending the cells in 0.144 M NH 4 Cl, 0.017 M Tris-HCl, pH 7.2.To remove B cells, lymphocytes were resuspended in ice-cold PBS/2% FCS and added onto sterile 15-cm petri dishes that were previously coated with 125 g goat anti-mouse Ig Ab (BioSource International, Camarillo, CA) and 500 g irrelevant Ig (anti-human Id mAb LC4) diluted in PBS.After incubation for 1 h at 4°C, the plates were gently swirled and nonadherent cells were collected.The panning procedure was repeated once, and the B cell-depleted lymphocytes were pooled from each plate, washed twice in HBSS, and used for adoptive transfer.Lymphocytes were analyzed pre-and post-B cell panning by flow cytometry to determine B and T lymphocyte proportion, as determined by CD3 ϩ and CD19 ϩ cell populations.Twenty million T cell-enriched immune lymphocytes were injected i.v. in 200 l HBSS into naı ¨ve recipients which were exposed to a sublethal dose of 400 rad of whole body gamma irradiation (RT250 X-ray Unit, Philips Medical Systems, Scarborough, Canada).Mice were challenged with C6VL tumor cells 4 days after cell transfer.

Proliferation assay against KLH
Four days after adoptive transfer of lymphocytes, one recipient mouse from each adoptive transfer group not used in tumor challenge was sacrificed.Spleens were harvested and made into single cell suspension and depleted of RBCs as described earlier.Splenocytes were washed and resuspended in serum-free AIM-V medium (Life Technologies, Grand Island, NY) supplemented with 50 M 2-ME.Fifty thousand cells were seeded per well in 96-well U-bottom tissue culture plates.KLH was added into each well at a final concentration of 100, 10, and 1 g/ml and the cells were cultured at 37°C in a humidified incubator for 4 days.On day 3, 1 Ci of [ 3 H]thymidine was added to each well.Cells were harvested onto glass fiber filters and counted on a scintillation counter (Wallac Micro Beta 1450, Turku, Finland).

Humoral immune responses induced by TCR vaccines in various adjuvants
To determine whether the efficacy of TCR vaccines could be improved beyond what was previously reported (8), mice were immunized with TCR vaccines formulated with various adjuvants.The magnitude and quality of humoral immune response stimulated by TCR vaccines composed of SAF-1, QS-21, IL-12, CD40L, or GM-CSF were compared.After three vaccinations, serum samples were collected and analyzed by ELISA for anti-C6VL TCR Ab titers.Control mice were immunized with an irrelevant protein 38C13 Id-KLH given in matching adjuvants.Mice immunized with TCR-KLH in QS-21 and SAF-1 were very effective in inducing high amounts of anti-C6VL TCR Abs, with average titers of 499 g/ml and 329 g/ml, respectively.In contrast, mice immunized with TCR-KLH in CD40L, GM-CSF, and IL-12 had 5-to 10-fold lower Ab titers, with an average of 32 g/ml, 86 g/ml, and 46 g/ml, respectively (Fig. 1).The Ab were specific for C6VL TCR, as serum from mice immunized with an irrelevant protein in matching adjuvant did not make any anti-C6VL Ab response (data not shown).
The isotypes of the anti-C6VL TCR Abs were determined to evaluate for qualitative differences among the Abs induced by the various adjuvants (Table I).The levels of mouse anti-C6VL TCR IgG1, IgG2a, IgG2b, and IgG3 present in hyperimmune serum stimulated by different adjuvants were measured using a pooled hyperimmune serum standard.Distinct Ab isotype profiles were produced by the different adjuvants.IgG1 Abs were induced by all five adjuvants tested.IgG2a Abs were detected only in mice immunized with IL-12 and QS-21.IgG2b Abs were induced by IL-12, QS-21, and SAF-1.In contrast, no IgG2a or IgG2b Abs were detected in the CD40L and GM-CSF groups.IgG3 Abs were not detected in any of the serum samples tested.Data shown are representative of at least two independent experiments.Each IgG subclass required the use of a different isotype-specific Ab detectors, thus the absolute quantities of IgG1, IgG2a, and IgG2b within each adjuvant group may not be reflected by the relative units measured in isotype-specific ELISAs.
In addition to measuring Ab titers and isotypes, the relative avidity of anti-C6VL TCR Abs was compared among the various vaccinated groups using thiocyanate elution in a modified ELISA assay (Fig. 2).Antisera induced by the difference adjuvants were adjusted to equal anti-C6VL TCR Ab concentrations and captured on TCR-coated ELISA plates.Bound Abs were exposed to increasing concentrations of NH 4 SCN.Disruption of Ag-Ab binding by chaotropic ions has been reported to correlate well with the affinity of mAbs (25), and has been used to determine the avidity of polyclonal Ab as well (26,27).Resistance to thiocyanate elution was used to determine relative Ab avidity in this assay and compared with the mouse anti-C6VL clonotypic mAb 124-40.The  avidity among the different adjuvant groups were not significantly different compared with the mAb 124-40, as there was an approximately 10% variation in interplate absorbance reading (25,26).In addition, the dissociation rate of bound anti-C6VL TCR Abs from C6VL TCR was measured using the surface plasmon resonance biosensor.No significant difference was detected in the rates of Ab dissociation among the different adjuvant groups (data not shown).

Effectiveness of adjuvants in tumor protection
To test whether the immune response induced by TCR vaccines in various adjuvants would inhibit tumor growth, we challenge mice vaccinated with TCR vaccines in various adjuvants with a lethal dose of C6VL cells.Only mice immunized with TCR vaccines given in SAF-1, QS-21, and IL-12 were significantly protected against C6VL tumor compared with PBS control (Fig. 3), with p values of 0.01, 0.003, and 0.009, respectively.Mice given TCR vaccines in CD40L and GM-CSF were not protected and died at rates similar to control mice given PBS.The three protected groups were not statistically different from one another.To rule out a nonspecific protective effect due to adjuvants, additional mice were vaccinated in subsequent experiments with an irrelevant protein, 38C13 Id-KLH, in SAF-1, QS-21, and IL-12 and challenged with C6VL cells.TCR-vaccinated mice had significantly prolonged median survival (Ͼ60 days) compared with mice vaccinated with control protein (median survival of ϳ30 days) (Table II).Survival of control mice was not statistically different from mice injected with PBS only.The tumor protective effect of TCR vaccines in SAF-1, QS-21, and IL-12 adjuvants had been confirmed in at least three independent experiments.The level of tumor protection by TCR vaccines would varying different experiments, but statistical significance was always reached (p Ͻ 0.05) when compared with negative control groups.

Role of B cells and Abs in tumor protection
The antitumor effect of anti-C6VL TCR Abs was tested by injecting naı ¨ve mice with a concentrated globulin fraction of hyperimmune serum obtained from C57BL/6 mice vaccinated with TCR-KLH in QS-21.The average anti-TCR Ab titer in recipient mice 1 day after globulin transfer was 231 Ϯ 25 g/ml, comparable with actively vaccinated mice, which had an average anti-C6VL TCR Ab titer of 247 Ϯ 54 g/ml (data not shown).Negative control mice were injected with an equivalent amount of similarly prepared globulin obtained from mice vaccinated with 38C13 Id-KLH in QS-21.One day after the Ab transfer, mice were challenged with C6VL cells and followed for survival.Mice transferred with anti-C6VL TCR hyperimmune globulin had significantly prolonged survival compared with mice given control globulin (Fig. 4).However, tumor protection was not restored to the level achieved in actively vaccinated mice.
The importance of B cells and Abs in the protective immune response against C6VL was further evaluated by vaccinating B cell-deficient J H D mice with TCR-KLH vaccines formulated in either IL-12 or QS-21.These two adjuvants were chosen for comparison because they had a dramatic difference in the quantity of anti-C6VL TCR Abs induced, and both adjuvants stimulated equivalent protective effect in C57BL/6 mice.After three immunizations, mice were challenged with C6VL cells.Neither the IL-12 nor QS-21 vaccinated groups were protected against C6VL tumor as compared with the survival of negative control J H D mice that were vaccinated with 38C13 Id-KLH in IL-12 (Fig. 5).The loss of tumor protection in J H D mice could be due either to the   b Median survival of mice vaccinated with SAF-1 adjuvant represents results from an independent experiment.absence of Abs, or to a deficiency in B cell Ag presentation.We tested whether tumor protection can be restored by transferring hyperimmune globulin (obtained from C57BL/6 mice vaccinated with TCR-KLH in QS-21) into J H D mice that were actively vaccinated with TCR-KLH in QS-21.The anti-tumor effect of Abs alone was evaluated by transferring hyperimmune globulin into J H D mice vaccinated with 38C13 Id-KLH in QS-21.The average anti-TCR Ab titer in recipient J H D mice 1 day after Ab transfer was similar to the Ab titers obtained in C57BL/6 mice in Fig. 4. Negative control J H D mice were vaccinated with TCR-KLH in QS-21, and injected with control globulin obtained from naı ¨ve C57BL/6 mice.One day after serum transfer, mice were challenged with C6VL tumor.J H D mice given anti-C6VL TCR hyperimmune globulin had prolonged survival compared with mice given control globulin (Fig. 6).However, the degree of tumor protection in mice first vaccinated with TCR-KLH and then given anti-TCR Abs was not significantly different from mice that were injected with anti-TCR Abs alone.

Role of T cells in tumor protection
The importance of T cell-mediated immune response in protection against C6VL tumor in vivo was determined by two different approaches.The role of effector CD8 ϩ T cells in conferring tumor protection was assessed in the first approach by using an anti-CD8 mAb to deplete CD8 ϩ T cells in mice immunized with TCR-KLH in QS-21.Mice were treated with either an anti-CD8 mAb, or an isotype-matched mAb after the third vaccination and prior to tumor challenge.The absence of CD8 ϩ T cells in peripheral blood was confirmed by flow cytometry using a non-cross-blocking anti-CD8 Ab.In the anti-CD8 mAb-treated group, greater than 98% of CD8 ϩ T cells were depleted (data not shown), and depletion was maintained throughout the observation period with weekly injection of depleting mAbs.The isotype matched irrelevant mAb had no effect on CD8 ϩ T cells population.TCR-vaccinated mice that were not depleted of CD8 ϩ T cells were significantly protected against C6VL tumor as compared to mice vaccinated with PBS or a control protein.Depletion of CD8 ϩ T cells completely abrogated tumor protection in QS-21 vaccinated mice (Fig. 7).Anti-C6VL TCR Abs were present in both CD8 ϩ T cell-depleted and nondepleted groups, with an average anti-C6VL TCR Ab titer of 235 Ϯ 181 g/ml, and 318 Ϯ 192 g/ml, respectively.Similar result was    observed when mice immunized with TCR vaccine in IL-12 were depleted of CD8 ϩ T cells (data not shown).The importance of CD4 ϩ T cells cannot be tested in this experiment, as anti-CD4 Abs would bind to the CD4 ϩ C6VL tumor cells in addition to normal CD4 ϩ T cells.
To evaluate whether tumor protection against C6VL could be solely conferred by immune T lymphocytes, an adoptive cell transfer experiment was performed.Spleens and lymph nodes were harvested from donor mice vaccinated with C6VL TCR-KLH in QS-21.This cell population contained 57% CD19 ϩ B cells and 30% CD3 ϩ T cells as determined by flow cytometry (data not shown).The remaining cell population consisted mostly of monocytes.After panning to remove B cells, less than 3% of B cells remained in the total population, and T cells were enriched to 78%.A total of 2 ϫ 10 7 of these T cell-enriched immune lymphocytes were injected i.v.into sublethally irradiated naı ¨ve recipients.Negative control mice were injected with 2 ϫ 10 7 of similarly treated immune lymphocytes from donors vaccinated with 38C13 Id-KLH in QS-21.To confirm the transfer of immune T cells, a T cell proliferation assay was performed.Representative recipient mice were sacrificed 4 days after cell transfer, and splenocytes were used in a T cell proliferation assay against KLH (Fig. 8).Both adoptive transfer recipient groups had comparable KLH-specific proliferative responses, indicating transfer of equivalent numbers of immune competent cells.The proliferative responses of cell transfer recipients were less than the response of TCR vaccinated donor mice, particularly in the higher KLH doses.Some background proliferation in the absence of KLH stimulation was observed in both cell transfer groups.The mice were challenged with C6VL tumor 4 days after cell transfer.The positive control group was injected three times with TCR-KLH in QS-21.Recipient mice that received immune lymphocytes from TCR vaccinated donors were equally protected against C6VL tumor challenge as compared with TCR protein vaccinated group (Fig. 9).Lymphocytes from irrelevantly vaccinated donors showed a slight degree of nonspecific protection compared with the PBS group.

Discussion
Tumor-specific immunity can be induced by vaccinating with idiotypic proteins derived from B and T cell lymphomas.Several studies using murine B cell lymphoma models have demonstrated the anti-tumor effects of vaccination using tumor Igs (1,3,28,29).Our group first reported the use of tumor-derived TCR in the active immunotherapy against T cell lymphoma.Vaccination using a soluble, heterodimeric TCR derived from C6VL tumor induced idiotype-specific immunity which protected mice against a lethal dose of C6VL cells (8).The lack of other published studies that involved the use of soluble TCR protein as antitumor vaccines is likely due to the difficulty in obtaining a sufficient amount of recombinant TCR proteins for vaccination.Consequently, little is known about the immune mechanisms involved in tumor protection using TCR vaccines.
Tumor immunity is often considered to be solely mediated by the cellular immune response, in particular, the activation of cytotoxic CD8 ϩ T cells (30 -32).Humoral immune response is usually thought to have minimal contribution.In fact, some have reported that a nonprotective humoral immune response has an inhibitory effect on the induction of protective T cell-dependent tumor immunity (33).In contrast, studies in a melanoma model (34), as well as in B cell lymphoma models (3,4) suggest that humoral immunity alone is sufficient to mediate tumor immunity.Thus, the requirement and the relative contributions of cellular and humoral immunity are different depending on the tumor system being studied.
In our present study, we characterized the immune mechanisms involved in the tumor protection of TCR-vaccinated mice, and the influence of immunologic adjuvants in inducing a protective immune response.Several adjuvants have been previously reported to induce protective immune responses in other tumor model systems and were chosen for study.Both IL-12 and QS-21 can induce strong Th1-type immune responses, particularly in the generation of CTL (22,23,35,36).Soluble CD40L interacts with CD40, which stimulates both B and T cell responses (37)(38)(39)(40).GM-CSF enhances Ag presentation by APCs such as dendritic cells, and has been shown to be an effective adjuvant in murine tumor model systems (24,28,41).In our study, GM-CSF and CD40L did not stimulate a protective immune response (Fig. 3).This may be due  to differences in the tumor models and Ags being studied.In contrast, mice vaccinated with TCR in SAF-1, QS-21, and IL-12 were well protected against C6VL tumor (Fig. 3 and Table II).Analysis of Abs induced by the various adjuvants showed that humoral responses differed quantitatively as well as qualitatively, reflected by the total Ab titers (Fig. 1) and the Ab isotype profiles (Table I).The magnitude of anti-C6VL TCR Ab response, however, did not correlate with tumor protection, as CD40L and GM-CSF induced anti-TCR Abs to similar levels compared with those of IL-12.Analysis of Ab avidity by thiocyanate elution also showed no significant differences among the Abs induced by various adjuvants (Fig. 2).In contrast, we found a correlation between tumor protection and the induction if IgG2 Abs (Table I).A mixture of IgG1 and IgG2 Abs was produced in the protected groups.In the nonprotective groups, IgG1 Abs were produced exclusively.IgG2 Abs could either participate directly in ADCC as well as complement fixation, or simply reflect the induction of a Th1-type cellular immune response.
Our data demonstrate that Th1-type cellular immune response is important in mediating tumor protection in our tumor model system, and that immune CD8 ϩ T cells are necessary and sufficient for conferring protection.In an adoptive cell transfer experiment, we showed that a single transfer of 2 ϫ 10 7 T cell-enriched immune lymphocytes from TCR-vaccinated donors was sufficient to prevent tumor growth, and the degree of protection was equal to mice actively vaccinated with the TCR protein (Fig. 9).In addition, the tumor protection in TCR-vaccinated mice was completely abrogated by in vivo depletion of CD8 ϩ T cells (Fig. 7), despite high level of circulating anti-C6VL TCR Ab.
Although immune competent CD8 ϩ T cells are necessary and sufficient for tumor protection, B cells are likely important for the induction of a protective immune response.TCR immunization of B cell-deficient J H D mice did not stimulate a protective immune response against C6VL (Fig. 5).The antitumor activity of B cells may be attributed to the production of anti-C6VL TCR Abs, or the result of processing and presentation of tumor Ags to T cells.The antitumor effect of passive infusion of anti-TCR Abs in the treatment of T cell tumors, including the C6VL model, has been reported (42,43).However, in our tumor model system, it is unlikely that the lack of protection in vaccinated J H D mice was due entirely to the absence of anti-C6VL TCR Abs.Transfer of hyperimmune globulin into J H D mice prolonged tumor survival (Fig. 6).The protection was not completely restored, however, as hyperimmune globulin transfer into naı ¨ve C57BL/6 mice was not sufficient in conferring full tumor protection when compared with actively vaccinated mice (Fig. 4).Furthermore, there is no correlation between survival and anti-C6VL TCR Ab titers, as CD8 ϩ T cell-depleted C57BL/6 mice that have high levels of anti-C6VL TCR Ab were not protected against tumor.Interestingly, we were not able to reconstitute a complete protective immune response by transferring hyperimmune serum into J H D mice actively vaccinated with TCR protein, and the degree of tumor protection was not improved beyond the effect of hyperimmune serum alone (Fig. 6).The lack of tumor protection in TCR-vaccinated J H D mice is thus likely due to inadequate T cell response, caused by the absence of Ag processing and presentation by B cells.Although J H D mice were reported to be capable of generating T cell response to highly immunogenic Ags such as viral proteins (44), it is likely that cellular immune response to weakly immunogenic Ags such as TCR is suboptimal.Studies have shown that B cell-deficient mice have impaired priming to protein Ags, leading to defective T cell response (45,46).In our tumor model system, B cells may serve as critical APCs in the processing and presentation of the TCR proteins.This suggests that the native conformation of C6VL TCR must be maintained in order to induce a protective immune response.Indeed, vaccination using a bacterial-derive single chain TCR protein encoding for the C6VL TCR V regions did not protect mice against tumor challenge (unpublished observation).
Protein vaccines are thought to be processed and presented via the exogenous protein processing pathway, leading to class II MHC Ag presentation and the induction of strong humoral immune response.Yet in this study we have clearly demonstrated that TCR protein vaccines can activate CD8 ϩ T cells that are sufficient for tumor protection.The activation of CD8 ϩ T cells requires class I MHC Ag presentation, implying that tumor-derived TCR proteins in vaccinated mice were able to cross over into the endogenous protein-processing pathway.TCR vaccines can induce Idspecific immune responses, thus peptides derived from the CDR3 region that encode for unique determinants of C6VL TCR should be important in inducing a tumor-specific immune response.Immunogenic CDR3 peptides derived from clonotypic TCR have been identified from malignant T cells of human patients (47,48).However, C6VL TCR ␣ and ␤ proteins do not contain any obvious peptide candidates that fit the classical class I binding motifs in the H-2 b haplotype.This was determined by searching the TCR sequence for classical peptide anchor residues for class I MHC (49), as well as by entering the TCR sequence into a peptide-binding motif database (50).The lack of dominant class I epitopes may explain the inability to induce in vitro CTL response against C6VL tumor cells using cytotoxicity and cytokine release assays (data not shown).Nevertheless, the lack of peptide candidates using computer searches does not preclude the presence of class I epitopes.Our data clearly show that a protein TCR vaccination strategy does not limit the type of immune response induced, as TCR proteins can be effectively processed and presented on both class I and class II MHC molecules.It would be of great interest to elucidate the identity of TCR peptides that are being recognized by immune CD8 ϩ T cells that mediate tumor protection.
In summary, we have characterized the tumor protection mechanisms of TCR protein vaccines in the treatment of a murine T cell lymphoma model.We found that adjuvants that induced Th1-type immune responses generated tumor protection, and CD8 ϩ T cells are necessary and sufficient for tumor protection in this model.Based on our results presented here, future experiments should aim at biasing TCR vaccines even further to induce Th1-type immune responses.

FIGURE 2 .
FIGURE 2. Relative Ab avidity induced by various adjuvants.Groups of 10 C57BL/6 mice were vaccinated three times s.c. with C6VL ␣, ␤-TCR-KLH vaccines in QS-21, SAF-1, GM-CSF, IL-12, and CD40L.Serum was collected, pooled, and analyzed by ELISA for resistance to thiocyanate elution.Equal amounts of these Abs were applied on microtiter plates, and different concentrations of ammonium thiocyanate were added.mAb 124-40 (anti-C6VL TCR clonotype) was added at the same concentration as the polyclonal sera.100% binding represents absorbance reading in the absence of thiocyanate ions.Each data point was calculated using the average of triplicate values.

a
Values of p were determined using Kaplan-Meier method comparing survival curves of control vaccines and TCR vaccines in matching adjuvants.

FIGURE 6 .
FIGURE 6. Survival of hyperimmune globulin-transferred J H D mice challenged with C6VL tumor.Groups of 13 J H D mice were vaccinated three times s.c. with C6VL ␣, ␤-TCR-KLH in QS-21 or irrelevant protein-KLH in QS-21.Thirteen days after the last vaccination, mice were injected i.p. with the concentrated globulin fraction of anti-C6VL TCR hyperimmune serum, or normal serum obtained from C57BL/6 mice.One day after serum transfer, mice were injected i.p. with 5000 C6VL cells and monitored for survival.

FIGURE 8 .
FIGURE 8. Proliferative response to KLH in adoptive transferred recipient mice.Naı ¨ve C57BL/6 mice were injected i.v. with 2 ϫ 10 7 immune T lymphocytes from donor mice vaccinated with C6VL ␣, ␤-TCR-KLH in QS-21 (dotted bar); or 2) irrelevant protein-KLH in QS-21 (hatched bar).Spleens were harvested from recipient mice 4 days after cell transfer and tested for their ability to proliferate to KLH at 100, 10, and 1 g/ml.Positive control proliferative responses were assayed using a donor mouse vaccinated with C6VL ␣, ␤-TCR-KLH in QS-21 (white bar).Naı ¨ve mouse was used as negative control (black bar).The data are represented as average [ 3 H]thymidine incorporation in cpm Ϯ SD.All data points were calculated from triplicate values.

Table I .
Heavy chain isotype profile of anti-C6VL TCR Abs induced using different adjuvants a Determined by isotype-specific ELISA.Values represent relative units, calculated using a pooled hyperimmune serum standard.IgG3 Abs were not detected in the tested sera.