- Email: [email protected]
Depletion of activated Vβ8+ T cells disrupts bispecific antibody directed antitumor immunity1
Journal of Surgical Research 122, 103–112 (2004) doi:10.1016/j.jss.2004.06.017
Depletion of Activated V8⫹ T Cells Disrupts Bispecific Antibody Directed Antitumor Immunity 1 Elizabeth J. McConnell, M.D.,*,2 Elisabeth C. McLemore, M.D.,* Robert Talac, M.D.,† Lokesh Joshi, Ph.D,* and Heidi Nelson, M.D.† *Arizona State University, Arizona Biodesign Institute, Tempe, Arizona; and †Mayo Clinic, Department of Colon and Rectal Surgery, Rochester, Minnesota Submitted for publication April 6, 2004
Introduction. Activation of V8ⴙ T cells with superantigen staphylococcal enterotoxin B (SEB) and use of an antitumor, anti-CD3 bispecific antibody (BsAb) leads to tumor protective immunity. We hypothesize that V8ⴙ T-cell activation in combination with BsAb is crucial for tumor protective immunity in this model. Methods. Adolescent C3H/HeN mice were intravenously injected with syngeneic CL62 melanoma to establish pulmonary metastasis. Three days after establishing pulmonary metastasis, predominantly V8ⴙ T cells are activated with 50 g of intraperitoneal superantigen SEB. T cells were depleted at different time points in relation to SEB administration to assess the effect on protective immunity against a second tumor challenge. Results. Protective immunity is significantly (P < 0.008) decreased when V8ⴙ depletion occurs 6 h after SEB injection, as growth of rechallenged CL62 melanoma occurred in 43%. Protective immunity is present at all other time points when mice survive V8ⴙ T-cell depletion. Survival of animals treated with SEB/BsAb (82%) is significantly better (P < 0.002) than with SEB alone (60%) or nontreated control (0%). Survival when V8ⴙ T-cell depletion occurred at 6 h and 48 h postSEB is 72% and 77%, respectfully, and is statistically indistinguishable (P < 0.232 and P < 0.602). If T-cell depletion was conducted before SEB administration, however, the combination of SEB and BsAb did not result in significant protective immunity. T-cell deple-
1 Supported by an intramural grant from the Department of Colon and Rectal Surgery of the Mayo Foundation. 2 To whom correspondence and reprint requests should be addressed at Arizona State University, Harrington Department of Bioengineering, Mail Code 9709, Tempe, AZ 85282-9709. Fax: 480-7276899 . E-mail: [email protected].
tion before the use of SEB alone, without BsAb, failed to result in significant protective immunity. Conclusions. Depletion of V8ⴙ T cells 6 h after activation disrupts the development of protective immunity. © 2004 Elsevier Inc. All rights reserved. Key Words: bispecific antibody; protective immunity; immunotherapy; staphylococcal enterotoxin B; V8ⴙ T cell. INTRODUCTION
Bispecific antibodies (BsAb) and superantigens have both been used individually and in combination in the treatment of solid malignant tumor models resembling metastatic disease because these models are easily targeted with immunotherapy [1–3]. The effector cells responsible for the cytolytic destruction of the tumor are considered to be the primary cells involved in protective immunity [4 –7]. In a metastatic melanoma model, we used a tumor and CD3 T-cell BsAb and the superantigen staphylococcus enterotoxin B (SEB) to determine any functional role in the timing of activation of the effector cells in the development of tumor protective immunity. Using this model in previous studies, we have shown the successful induction of tumor antigen-specific protective immunity in mice [8, 9]. After intravenous injection of CL62 tumor cells, mice developed pulmonary metastasis. Thereafter, superantigen SEB and CL62 tumor x anti-CD3 T-cell BsAb were administered alone or in combination. Surviving mice (superantigen SEB alone or in combination with BsAb) underwent a second tumor challenge with either CL62, K1735, or AG104 tumor cells. Most mice rejected the second tumor challenge with CL62. In addition, most mice rejected the p97-negative parental cell line, K1735, indi-
103
0022-4804/04 $30.00 © 2004 Elsevier Inc. All rights reserved.
104
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
cating an immune response to tumor antigens common to both cell lines that were not the original target of the BsAb. Alternatively, injection of the syngenic AG104 tumor cell line resulted in tumor growth [9]. Evidence of cellular immunity was obtained from the results of delayed-type hypersensitivity, proliferation, and cytotoxicity assays, which revealed the presence of tumorspecific memory in BsAb-treated, CL62-cured mice [9]. The critical effector cell, the V8⫹ T cell, is the predominant cell activated by the superantigen SEB [9]. Whether V8⫹ T-cell activation is critical for the development of immune memory or protective immunity has yet to be determined. The scope of this study is limited to the investigation of the V8⫹ T cell. Superantigen SEB activation of V8⫹ T cells leads to the release of TH1 cytokines, which ultimately activate the entire T-cell repertoire. Activation also occurs in conjunction with the anti-CD3 F(ab=)2 arm of the bispecific antibody. Other studies suggest that SEB is an effective stimulator of both cell-mediated and innate immunity [10]. SEB is known to increase antigen-specific CD4⫹ memory T cells [11]. When used with irradiated B16F10 melanoma cells, formation of effective, but rather short-lived, tumor immune memory occurs [12]. The role of the CD4 and CD8 T cell in the development of protective immunity has been examined in an additional experiment and will not be discussed in this paper. In summary, this study was conducted to determine the effect of depletion of V8⫹ T cells at different time points before and after the administration of SEB alone or in combination with a BsAb to determine the role the effector V8⫹ T cells may play in the development of tumor immunity. METHODS Tumor Models Female C3H/HeN mice (age, 6 to 9 weeks) were purchased from Harlan Sprague Dawley (Indianapolis, IN, USA), and IL-2 knockout C3H/HeJ mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). All animals were maintained in accordance with the animal care guidelines of the National Institutes of Health. Pulmonary metastases were established in C3H/HeN or C3H/HeJ mice by injecting 5 ⫻ 10 5 CL62 melanoma tumor cells suspended in 500 l of phosphate-buffered saline (PBS) intravenously via the tail vein (day 0).
Cell Lines CL62, a murine melanoma cell line, is a clone derived from the K1735 melanoma syngeneic to C3H/HeN mice transfected with a gene encoding p97 human tumor antigen. Both CL62 and K1735 tumor cell lines were provided by Drs. K.E. Hellstrom and I. Hellstrom (Oncogene Corp., Seattle, WA, USA) and maintained as an adherent monolayer culture in complete media, as described previously [5]. Cells were routinely tested by flow cytometry to determine if ⬎90% of CL62 melanoma cells expressed p97 antigen.
Bispecific Antibody Bispecific antibody was prepared by covalently conjugating the F(ab=)2 fragment of a hamster antimouse CD3 antibody 500A2 to the F(ab=)2 fragment of the hybridoma producing murine monoclonal antibody 96.5, as described previously [5]. The 500A2 F(ab=)2 fragment binds to the epsilon chain of the murine CD3 molecule; it was produced by intraperitoneal growth of the 500A2 hybridoma (provided by Dr. D.J. McKean, Mayo Clinic, Rochester, MN, USA). Murine monoclonal antibody 96.5 binds p97 human melanoma antigen expressed on ⬎90% of melanoma cells. The purity and function of the BsAb was confirmed using two-color flow cytometry analyses by a Becton-Dickinson FACscan using the Cell Quest program and chromium release assay [5].
Treatment Protocol The CL62 tumor cells were harvested in the log phase of growth using trypsin and washed. Some tumor cells were subjected to a single dose of irradiation (10 Gy) from a 60Co source (1.08 Gy/min). Half a million live or radiation-killed tumor cells suspended in 200 L of PBS were injected intravenously in the tail vein of the mice. Treatment was started 3 days after tumor inoculation. Tumorbearing mice were injected intraperitoneally with 50 g of SEB (day 3) suspended in 500 l of PBS and 4 days later (day 7) with 5 mg of BsAb injected intravenously. Mice were observed weekly for survival. Treated mice surviving beyond 60 days after inoculation of CL62 tumor cells were considered “cured” and were injected subcutaneous with CL62 tumor with 5 ⫻ 10 6 tumor cells (CL62 or K1735) administered subcutaneous and further observed for tumor growth and survival weekly for the subsequent 300 days.
In Vivo Depletion Studies To determine whether the expansion of V8⫹ T cells plays a role in the production of effector cells that provide tumor-specific immunity, CL62 melanoma-bearing mice were depleted of V8⫹ T cells by injection of 200 g of F23.1 monoclonal antibody (provided by Dr. C. S. David, Mayo Clinic). As a control, we also analyzed depletions of CD4⫹ T cells, by injection of 200 g of GK1.5 monoclonal antibody (ATCC, Rockville, MD, USA); CD8⫹ T-cells, by injection of 200 g of 116-13.1 monoclonal antibody (ATCC); and V7⫹ T cells by injection of 500 g of TR310 (provided by Dr. C. S. David). Treatment with the respective monoclonal antibody leads to ⬎80% depletion of a relevant T-cell subset, as determined by FACscan analyses in more than 15 mice analyzed in perfecting this technique. Depletion of controls CD4⫹, and CD8⫹ T cells were performed at various time points: 7 days before, and 6 h and 48 h after SEB administration.
Assessment of Serum Cytokine Levels To measure serum cytokine levels, C3H/HeN mice were bled via the tail vein at 3, 6, 12, 18, 24, 30, 48, and 72 h after SEB administration. Th-1 type (IL-2, IFN-␥, IL-12) and Th-2 type (IL-4, IL-10) cytokine levels were determined with enzyme-linked immunosorbent assay (ELISA) using commercially available kits (PharMingen, San Diego, CA, USA), according to the manufacturer’s instructions. Serum was separated and stored at – 80°C until the ELISA was preformed. The detection limits of the assays were 40 pg/ml for all cytokines. Positive controls were used in all analyses.
In Vivo Cytokine-Dependent Studies The contribution of IL-2 and IFN-␥ cytokines to antitumor effects and tumor-specific memory was assessed in vivo by depletion of these specific cytokines. For IL-2 cytokine studies, C3H/HeJ IL-2 knockout and C3H/HeJ (control) mice were used. The mice were treated according to the treatment protocol described above and monitored for survival. Because IL-2 knockout mice are susceptible to inflamma-
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
105
RESULTS V8 T-Cell Populations Expand with SEB Treatment
FIG. 1. Characterization of the T-cell activation. Flow cytometric analysis of receptors of T cells found within the spleen is represented. The splenocytes are specifically labeled with antibody bearing fluorescent dyes to evaluate the expression and presence of V8⫹, CD4⫹, CD8⫹ and CD3⫹ T cells. Antibody to CD3⫹ is the prominent antigen expressed on all T cells. Lag phase is from 0 to 12 h, expansion phase from 6 to 24 h, and death phase from 24 h on. Expansion of T cells after administration of staphylococcal enterotoxin B (SEB) doubled based on CD3 labeling (20% to 40%). T cells were also labeled with CD4⫹ and CD8⫹ with a different fluorescent wavelength dye from V8⫹ to determine specific group expansion. CD4⫹V8 ⫹ () and CD8 ⫹V8 ⫹ (Œ) percent expression was reported as a function of the total CD3⫹ T-cell population. The percentage of cells expressing the labeled characteristics is represented. The CD4⫹V8 ⫹ () increased threefold and the CD8 ⫹V8 ⫹ (Œ) population increased fourfold. The total number of V8 ⫹ T cells within a 48-h period rose from 25% to ⬎80% of the total T-cell population within the spleen. Activation-induced cell death occurred in the subsequent 48 h with a dramatic drop of the population almost back to baseline at 72 h.
tory bowel disease, subcutaneous injection of 5 ⫻ 10 6 CL62 and K1735 tumor cells, respectively, was performed once significant survival differences had been observed (on day 45 after tumor inoculation) and only in the absence of any symptoms. The mice were monitored for survival and tumor growth by palpation and caliper measurement. For IFN-␥, the C3H/HeN mice, treated according to the protocol described above, were injected intraperitoneally with antimouse IFN-␥ monoclonal antibody (R4-6A2) suspended in 1 ml PBS (ATCC). The antimouse IFN-␥ monoclonal antibody was injected on day 0, 1, 3, and 8 in doses of 3, 3, 1, and 1 mg, respectively. Mice were bled and serum collected to determine IFN-␥ depletion. Cured mice were injected subcutaneous with 5 ⫻ 10 6 CL62 or K1735 tumor cells and monitored for survival and tumor growth.
Administration of intraperitoneal SEB is not lethal in C3H/HeN mice and results in a well-documented expansion of the V8⫹ T-cell subsets. Production of CD8⫹ and CD4⫹ is directly increased by SEB administration (Fig. 1). Stimulation of the V8⫹ T cells by SEB increased the percentage of the T-cell population within the spleen more than 3.5 times above baseline. This increase is also seen with both the CD4⫹ T cells, which increased three times, and the CD8⫹ T-cells, which increased four times. The naïve control, which is the first point of the graph, defines the baseline. The in vivo responses of CD3, CD4⫹, CD8⫹, V3, V6, V7, V8.1-3, V11, and V12 positive T-cell subpopulations were measured by FAC analysis at different time points (data not shown). During the first 12 h, the CD3⫹ T-cell population declined, whereas the V8⫹ T-cell population remained at a steady state. In the second phase, the V8⫹/CD4⫹ and V8⫹/CD8⫹ T-cell population increased three- to fourfold. The V8⫹/ CD4⫹ and V8⫹/CD8⫹ T cell reached maximal population levels 48 h after stimulation. Thereafter, both CD4⫹ and CD8⫹ T-cell populations declined and rapidly underwent activation-induced cell death. Similar Survival with BsAb Administration at 24 or 72 Hours After SEB
After establishment of CL62 pulmonary metastasis on day 3, T-cells were activated with i.p. administration of SEB. Then, BsAb was administered on either day 4 or day 7. Our previous publications of this model had always used BsAb only at day 4, just
Statistical Methods Survival data were analyzed with the Kaplan-Meier method and displayed as actuarial survival curves. The survival curves were compared with a log-rank test. Probability values ⬍0.05 were considered significant. Statistical significance was assessed using oneway ANOVA model. The margins of error for comparison were obtained by calculating the 95% confidence intervals for the differences between groups. The tumor was determined to be present or absent. All experiments were performed at least three times, and representative data are displayed.
FIG. 2. Survival after CL62 pulmonary metastasis in C3H/HEN mice. Following the intravenous injection of CL62 tumor cells, mice were treated with staphylococcal enterotoxin B (SEB) on day 3 and bispecific antibody (BsAb) on day 4 (□ n ⫽ 40) or day 7 (, n ⫽ 34).
106
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
FIG. 3. Treatment and V8 depletion timeline. Treatments and V8 depletions are depicted in a linear timeline. SEB ⫽ staphylococcal enterotoxin B.
24 h after SEB [5, 8, 9]. Survival is similar in each group, regardless of use at day 4 or day 7 (72% versus 92%) as shown in Figure 2. Administration of the BsAb on day 7 allows time for depletion of the expanding V8⫹ T cells. V8ⴙ T-Cell Depletion at Three Different Time Points
Figure 3 demonstrates the course of treatment and depletion of V8⫹ T cells in this murine tumor model. The timeline demonstrates that the V8⫹ T-cell depletions were performed (200 g of F23.1 monoclonal antibody) either just before, or 6 h or 48 h after SEB administration. Surviving mice were monitored for 300 days.
Significant Survival after V8ⴙ T-Cell Depletion Requires BsAb
Overall, depletion of the V8⫹ T cells did not affect survival if depleted after SEB administration (Fig. 4). Depletion of the expanding V8⫹ T cells before SEB administration resulted in significantly impaired survival when compared with the SEB treated, nondepleted controls (P ⬍ 0.05). Depletion of V8 ⫹ T-cells either 6 or 48 h after SEB, however, was not statistically different (72% versus 77%) when compared with the treated control 82% (P ⬍ 0.232 and P ⬍ 0.602). All mice treated with irradiated tumor cells survived. SEB administration alone and in combination with BsAb provided protective immunity to tumor challenge. Radiation killed tumor cells alone did not result in protective immunity (data not shown). SEB Stimulated T Cells are Cytotoxic and Tumor Targeted with BsAb
The in vitro T-cell cytotoxicity against the CL62 tumor is tested in vivo on V8⫹ and V7⫹ depleted mice. SEB administration predominantly activates V8⫹ T cells, creating cytotoxic T lymphocytes (CTL), but it also activates V7⫹ T cells. As a control, the T cells of mice depleted of either V7⫹ or V8⫹ T cells were tested before and after SEB administration. Depletion of the V8⫹ subset after SEB activation did not change targeted cytotoxicity of V8⫹ T cells in vitro (Figure 5). Decrease in Protective Immunity with V8ⴙ T-Cell Depletion 6 Hours After SEB FIG. 4. Survival curve with V8 depletion. Additional groups of mice were then treated with staphylococcal enterotoxin B (SEB) and bispecific antibody (BsAb), but depleted of V8 T cells, either before SEB (‘, n ⫽ 39), 6 h after SEB (⽧, n ⫽ 22), or 48 h after SEB (’, n ⫽ 9).
As shown in the Figure 3 time line, the “cured” mice were evaluated for immune memory by a second subcutaneous injection tumor cells. Our previous studies demonstrated that tumor-specific immunity existed against targeted (p97⫹) and nontargeted an-
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
107
FIG. 5. In vitro T-cell cytotoxicity is tested on in vivo V8 and V7 depleted mice Depletion of V8⫹ subset after staphylococcal enterotoxin B (SEB) activation does not change targeted cytotoxicity of V8-T Cells in vitro. Splenocytes were taken from (A) untreated (), (B) V8-depleted after SEB (⽧), (C) V8-depleted before SEB (‘), or (D) V7-depleted () C3H/HeN mice, which were all co-cultured with SEB and irradiated CL62 melanoma cells. Splenocytes were then placed over a nylon wool column and T cell-purified and tested for direct CTL activity on CL62 cells (dashed line) or targeted CTL activity with the bispecific antibody (BsAb) 96.5 ⫻ 500A2 (solid line). Group B and C T cells were depleted of V8 T cells and group D (controls) depleted of V7 T cells and treated with SEB. Data shown represent one of three experiments performed with similar results. E/T, effector target ratio.
tigens [9]. These studies were performed on mice that survived 60 days after the treatment had “cured” the mice and were injected subcutaneously with 5 ⫻ 10 6 CL62 or K1735 tumor cells and observed for tumor growth for 100 days. Table 1 demonstrates the results of all mice having a secondary tumor challenge with either the CL62 or K1735 tumor cell line. A large percentage of mice depleted of V8⫹ T cells 6 h after SEB administration through
the lag and expansion (log) phases grew CL62 tumor and K1735 tumor (43% and 64% growth, respectively). A significant absence was noted in tumorspecific immunity to the targeted antigens of CL62 compared with control (P ⬍ 0.008 and P ⬍ 0.134, respectively). Tumor growth occurred within the first 50 days in all of the control CD4⫹ and CD8⫹ T-cell depleted mice (100%). Tumor growth did not occur in the mice that survived SEB/BsAb treatment (0%
108
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
TABLE 1 Tumor Growth in Control and Treated Animals After Subcutaneous Rechallenge Tumor Groups
CL62
Grew (%)
Naive control SEB cured SEB/V8⫹ T cells depleted/BsAb cured –6 h –48 h SEB/BsAb cured: –V8⫹ T cells depleted –CD4⫹ T cells depleted –CD8⫹ T cells depleted
21/22 0/22*
95% 0%
6/14* 0/6 0/42* 0/9* 9/9. 9/9.
43% 0% 0% 0% 100% 100%
K1735 21/22 5/22* 9/14 T 0/6 1/42* 0/9* 9/9. 9/9.
Grew (%) 95% 23% 64% 0% 0% 0% 100% 100%
TP ⬍ 0.134 compared with naive control.SEB ⫽ staphylococcus enterotoxin B; BsAb ⫽ bispecific antibody. * P ⬍ 0.008 compared with naive control.
growth). Mice depleted of V8⫹ T cells 48 h after SEB rejected tumor (0% growth). Cytokines INF-␥ and IL-2 Increase in Serum after SEB Administration
The administration of SEB resulted in a rapid increase in the serum levels of Th-1 cytokines, IL-2, and INF-␥ as shown in Figure 6. IL-2 serum levels peaked at 3 h, as shown in Figure 6, correlating with the lag phase as shown in Figure 1. INF-␥ levels peaked at 12 h, as shown in Figure 6, which correlates with the time period just before the expansion phase as shown in Figure 1. IL-10 serum levels rose, as shown in Figure 6, during the expansion phase of V8⫹ T cells, as shown in Figure 1. Levels of IL-12 and IL-4 were not detectable. This experiment, which was repeated with depletion of V8⫹ T cells before SEB administration, resulted in no detectable level of cytokines. Serum cytokine levels were also measured after BsAb administration alone, which alone did not result in enhanced cytokine production nor did it enhance SEB-induced cytokine production (data not shown). Cytokine Depleted Mice Survive and Demonstrate Antitumor Memory
To confirm the role of IL-2 and INF-␥ in survival and tumor-specific immunity in this treatment model, the same studies were preformed with IL-2 knockout or INF-␥ monoclonal antibody-depleted mice. The transgenic IL-2 knockout mice (Jackson Laboratories) and INF-␥ monoclonal antibody-depleted mice were administered SEB and BsAb after establishment of pulmonary metastasis as described above. Survival patterns for the IL-2 knockout (49%) or INF-␥ depleted (79%) mice were statistically the same as the SEB/BsAb treated wild-type control (Figs. 7 and 8). By testing the immune memory with tumor rechallenge in these
cytokine-deficient mice, we demonstrated the same rejection pattern as in the “cured” wild-type controls. All mice were challenged at day 45 and rejected both CL62 and K1735 tumors (data not shown). The life span of both the IL-2 and INF-␥ depleted control mice was shorter than the nondepleted naïve control. Mice in both these groups were only observed for 60 days before sacrifice. The number of cytokine-depleted mice was significantly smaller than any other group so differences may have gone undetected. DISCUSSION
This study demonstrates that the early recognition of tumor antigen by activated T cells is critical to the formation of protective immunity. No previous studies have investigated the role of time point depletion of effector V8⫹ T cells to determine if the T-cell directed protective immunity is dependent on the presence of the initially activated T cells. A detailed evaluation of T-cell exposure to tumor antigen and the time frame of subsequent memory T-cell formation is critical to the understanding of immune function and the development of protective immunity [5, 8, 9]. The T-cell immune memory is long-lasting and the effect is tumor antigen-specific [5, 8, 9]. By combining the use of superantigen SEB with the depletion of the V8⫹ T-cell population, we studied the effect of tumor rechallenge on the T-cell memory response. We provided direct evidence that the timing of V8⫹ T-cell depletion is vital to sustaining memory T-cell protective immunity. Additionally, we confirmed that significant survival was dependent on the administration of BsAb. Initially, we were interested in determining if the T-cell protective immunity was dependent on V8⫹ T cells, the predominant T cell activated by SEB. We have previously reported that immune memory is T-cell mediated and is dependent on both the tumor
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
FIG. 6. Staphylococcal enterotoxin B (SEB) administration increased serum levels of IL-2 and interferon (INF)-␥. L-2 serum levels peaked at 3 h, INF-␥ levels peaked at 12 h, and IL-10 serum levels rose over 24 h.
antigen targeted by the BsAb, and the nontargeted antigens present on the tumor cell surface [8]. Mice
109
treated with SEB alone demonstrated increased survival after tumor introduction [5, 8, 9]. Survival, however, was not as dramatic when compared with the survival of mice treated with both the SEB and BsAb [5, 8, 9]. We concluded that two mechanisms are involved in T-cell memory production: BsAb activation of the anti-CD3 T-Cell receptor and SEB promotion of signaling in naïve T cells [9]. Superantigens, which have been used as T-cell activators, are effective stimulators of both cell-mediated and innate immunity [10]. Investigators have found that when superantigen is administered at the time of OX40 ligation, a 60 times increase occurs in the antigen-specific CD4⫹ memory T cells [11]. The superantigen activation of tumor-specific CD4⫹ T-helper lymphocytes is pivotal because CD4⫹ T cells are regulatory cells in the generation of long-term immune memory. Others have demonstrated that the formation of effective tumor immune memory is possible with the use of irradiated B16F10 melanoma cells with superantigen [12]. SEB has recently been shown to have a role in the activation of dendritic cells in the lymphoid organs of mice [13]. SEB-induced T-cell activation is well documented [14], and other investigators have shown that SEB directly activates V8⫹ cytotoxic T cells. SEB induces cytotoxicity by several mechanisms, including superantigen-mediated, antibody-dependent T-cell cytotoxicity [15]. Our research has shown that immunization with SEB alone, in this model and in a transgenic spontaneous pancreatic tumor model, activates and matures tumor antigen-specific cytotoxic T cells [16]. All of these studies focus on the role of the superantigen. Superantigen alone, however, is not sufficient for survival when V8⫹ T cells are depleted in this model, which implies a significant role for protection of the animal by the BsAb. Use of a BsAb in our model makes a significant contribution to enhancing survival and we conclude also protective immunity. The BsAb provides an effective stimulator and modulator of the immune system [17]. Recent studies have shown that a BsAb made by the combination of a superantigen and antitumor C215 monoclonal antibody improved survival and resulted in CD4⫹ T-cell– dependent cytokine release in a C215transfected syngeneic B16 melanoma model [18]. We recognize the limitations to a full understanding of how immune memory is established, even in this model. The independent and unique role of T-cell expansion, particularly CD4⫹T cells, in the absence of BsAb could not be explored in this study. Treatment of any of the mice with the T-cell activator and depletion with no BsAb resulted in a low survival rate, and the number of mice needed to establish immune memory without BsAb would have been prohibitively large. The correlation between the time-specific depletion of expanding T cells and the loss of tumor protective immunity could only be controlled for by a consistent administration of
110
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
FIG. 7. Survival with interferon (IFN)-␥ depleted C3H/HEN mice. Before intravenous injection of CL62 tumor cells, C3H/HEN mice were treated intraperitoneally with 1 mg of R4682 monoclonal antibody (mAb). The depletion was carried out over an 8-d period while mice where treated with either saline control (, n ⫽ 5); with staphylococcal enterotoxin B (SEB) on day 3 and bispecific antibody (BsAb) on day 4 (not shown, n ⫽ 5, survival 10%); SEB day 3 and BsAb day 7 (, n ⫽ 5); SEB alone on day 3 (⽧, n ⫽ 5). Tumor-bearing, IFN-␥ depleted mice were also treated with SEB on day 3 and BsAb on day 7 (□, n ⫽ 4).
BsAb at the same time point in all depletion experiments. Use of radiation-killed tumor cells would provide no further benefit in answering this question because all mice demonstrated protective immunity with the use of SEB alone. The role of T-cell expansion in tumor protective immunity, specifically CD4⫹ T cells, has been theorized in numerous reviews. Reports on animal models of tumor protective immunity are limited, however. Tumor protective immunity was not completely absent in V8⫹ T-cell depleted mice. This suggests that the overall immune response elicited by the superantigen with the BsAb, not the V8⫹ T-cell activation alone, is sufficient for the recall of the effector T cell. IL-2 cytokine stimulation occurred within 24 to 48 h after superantigen SEB administration. Depletion of V8⫹ T cells after SEB administration did not affect survival, with the exception of the group that did not receive BsAb. Selection of the superantigen that predominantly expands the V8⫹ T-cell population (Fig. 1) followed by depletion of these T cells after SEB administration during the entire expansion phase (6 h after SEB administration) did affect the development of tu-
mor protective immunity (Table 1). Depletion of other T cells, such as V7⫹, which are moderately increased with superantigen administration, has no effect (data not shown). Depletion of other larger subsets of T cells, such as CD4⫹ and CD8⫹ T-cells, reveals the importance of both helper and cytotoxic T cells in this cellmediated process as no mice survived with these T-cell depletions. Further evidence that the tumor protective memory is dependent on T-cell expansion rather than a specific cytokines is exemplified in the cytokine depletion experiments that show survival and serum cytokine levels correlate with, but do not determine the different phases of T-cell activation (Figs. 6, 7, and 8). Exploring the role of IL-2 and INF-␥ provides limited additional information because the life span of the cytokine-depleted mice is short and, therefore, the time frame for detecting the establishment of immune memory is not ideal. Others have demonstrated no specific surface marker or cytokine expressed that distinguished naïve from tumor protective T cells. Regardless of stimulation, CD8⫹ T cells produced IFN-␥ twice as much as CD4⫹ T cells, whereas, IL-2 production was predominantly a CD4⫹ T cell event that increased
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
FIG. 8. Survival with IL-2 C3H/HEJ knockout mice. Following the intravenous injection of CL62 tumor cells, tumor-bearing, IL-2 knockout mice were treated with staphylococcal enterotoxin B (SEB) on day 3 and bispecific antibody (BsAb) on day 7 (□, n ⫽ 4). As controls, immunocompetent C3H/HEJ mice were treated with either saline (control , n ⫽ 5), with SEB on day 3 and BsAb on day 4 (not shown, n ⫽ 5, survival 10%), or SEB on day 3 and BsAb day 7 (, n ⫽ 5); SEB alone on day 3(⽧, n ⫽ 5).
with cell division [19]. Transgenic IL-2 knockout mice were used to help decipher if cytokines played a role in protective immunity. The background of the transgenic mouse is C3H/JeN, which is different from the mouse model in which the pattern of tumor growth had been established. The life span of both the control and transgenic mouse in this model was shorter than predicted. This made the establishment of survival with SEB and BsAb treatment a challenge. Our methods of analysis on the survival differences in the cytokine-depleted mice, which may or may not exist, would have gone undetected with the limited number of animals used in this study. In addition, the depletion by the monoclonal antibody in all of the experiments is not 100% complete. Our findings reveal novel observations that provide a more complete understanding of the timing and events in the formation of protective immunity. We hypothesized that three effectors of protective immunity exist in this model: superantigen, BsAb, and tumor cells. The timeline of events for both the activation of T cells and the generation of protective immunity in this model implicates the role of both CD4⫹ and CD8⫹T cells. Initial investigations into T-cell memory implicated the CD4⫹ T cell as the instigator of two distinct stages of cytotoxic T-lymphocyte development during immune response: an early primary stage dependent on the presence of CD4⫹ cells and a later, CD4independent stage operative during the secondary response, which decays with time [20]. Decline of in vivo
111
antigen-activated CD8⫹ T cells or a poor T-cell expansion response is suggested to result from a reduction in CD4⫹ T cells [21]. Others have confirmed that the interaction between antigen-infected dendritic cells and CD8⫹ T cells through CD40/CD40L was crucial for the expansion and maturation of CD8⫹ memory T cells, the process of which was CD4-independent and independent of IFN-␥ production [22]. Our tumor model may be limited in that it is an injected tumor model. We have previously shown that the immune response of a mouse with a spontaneous cancer is unlike an immune response of a mouse with an injected tumor [16]. Given the limitations of the model, we will continue to explore the role of each effector in the current mouse model. In summary, we have shown that superantigen SEBactivated V8⫹ T cells in the expansion phase are a potent source of immune modulators in an immune competent mouse model of an injectable melanoma. In combination with a bispecific antibody, the SEB activated V8⫹T cells leads to tumor protective immunity. Naïve T-cell responses to both SEB and BsAb are neither independent nor uniform in nature [19]. Naïve T cells respond to different stimuli in different ways, and predictably behave differently in a spontaneous tumor model, which proves difficult to study [23]. The immediate downstream effect of V8⫹ T-cell expansion is an increase in serum IL-2 and INF-␥. Despite the correlation of loss of protective immunity with early depletion of V8⫹ T cells, the depletion had no effect on survival. We presume that BsAb administration, which is crucial for survival, also has a role in assisting SEB activation in the development of long-lasting protective immunity. REFERENCES 1.
Talac, R., and Nelson, H. Current perspectives of bispecific antibody-based immunotherapy. J. Biol. Regul. Homeost. Agents 14: 175, 2000.
2.
Penna, C., Dean, P. A., and Nelson, H. Pulmonary metastases neutralization and tumor rejection by in vivo administration of beta glucan and bispecific antibody. Int. J. Cancer 65: 377, 1996.
3.
Penna, C., Dean, P. A., and Nelson, H. Antitumor x anti-CD3 bifunctional antibodies redirect T-cells activated in vivo with staphylococcal enterotoxin B to neutralize pulmonary metastases. Cancer Res. 54: 2738, 1994.
4.
Ramsey, P. S., Dean, P. A., Tasimowicz-Alpini, A., Donohue, J. H., and Nelson, H. Interferon-gamma-induced intercellular adhesion molecule-1: Expression on human tumor cells enhances bifunctional antibody mediated lysis through a lymphocyte function-associated antigen-1 dependent mechanism. Cancer Res. 53: 4652, 1993.
5.
Chapoval, A. I., Nelson, H., Thibault, C., Penna, C., and Dean, P. Bifunctional antibody retargeting in vivo-activated T lymphocytes: Simplifying clinical application. J. Hematother. 4: 571, 1995.
6.
Chapoval, A. I., Nelson, H., and Thibault, C. Anti-CD3 x anti-
112
7.
8.
9.
10.
11.
12.
13.
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004 tumor F(ab=)2 bifunctional antibody activates and retargets tumor-infiltrating lymphocytes. J. Immunol. 155: 1296, 1995. Thibault, C., Nelson, H., and Chipoval, AI. Tumor-infiltrating lymphocytes can be activated in situ by using in vivo activants plus F(ab=)2 bispecific antibodies. Int. J. Cancer 67: 232, 1996. Porter, L. E., Nelson, H., Ethem Gecim, I., et al. T cell activation and retargeting using staphylococcal enterotoxin B and bispecific antibody: an effective in vivo antitumor strategy. Cancer Immunol. Immunother. 45: 180, 1997. Rice, D. C., Chapoval, A. I., Porter, L., and Nelson, H. Induction of antitumor immunity after cure of pulmonary metastases, using staphylococcal enterotoxin B and bispecific antibody. Cancer Immunol. Immunother. 48: 230, 1999. Litton, M. J., Dohlsten, M., Hansson, J., et al. Tumor therapy with an antibody-targeted superantigen generates a dichotomy between local and systemic immune responses. Am. J. Pathol. 150: 1607, 1997. Maxwell, J. R., Weinberg, A., Prell, R. A., and Vella, AT. Danger and OX40 receptor signaling synergize to enhance memory T cell survival by inhibiting peripheral deletion. J. Immunol. 164: 107, 2000. Kominsky, S. L., Torres, B. A., Hobeika, A. C., et al. Superantigen enhanced protection against a weak tumor-specific melanoma antigen: Implications for prophylactic vaccination against cancer. Int. J. Cancer 94: 834, 2001. Yoon, S., Bae, K. L., Shin, J. Y., et al. Analysis of the in vivo dendritic cell response to the bacterial superantigen staphylococcal enterotoxin B in the mouse spleen. Histol. Histopathol. 16: 1149, 2001.
14.
Marrack, P., and Kappler, J. The staphylococcal enterotoxins and their relatives. Science 248: 705, 1990.
15.
Holzer, U., Bethge, W., Krull, F., et al. Superantigen-
16.
17.
18.
19.
20.
21.
22.
23.
staphylococcal-enterotoxin-A-dependent and antibody-targeted lysis of GD2-positive neuroblastoma cells. Cancer Immunol. Immunother. 41: 129, 1995. McConnell, E. J., Pathangey, L. B., Madsen, C. S., et al. Dendritic cell-tumor cell fusion and staphylococcal enterotoxin B treatment in a pancreatic tumor model. J. Surg. Res. 107: 196, 2002. Kodama, H., Suzuki, M., Katayose, Y., et al. Specific and effective targeting cancer immunotherapy with a combination of three bispecific antibodies. Immunol. Lett. 81: 99, 2002. Rosendahl, A., Kristensson, K., Hansson, J., et al. Repeated treatment with antibody-targeted superantigens strongly inhibits tumor growth. Int. J. Cancer 76: 274, 1998. Brenchley, J. M., Douek, D. C., Ambrozak, D. R., et al. Expansion of activated human naive T-cells precedes effector function. Clin. Exp. Immunol. 130: 432, 2002. Jennings, S. R., Bonneau, R. H., Smith, P. M., et al. CD4positive T lymphocytes are required for the generation of the primary but not the secondary CD8-positive cytolytic T lymphocyte response to herpes simplex virus in C57BL/6 mice. Cell Immunol. 133: 234, 1991. Jin, X., Wills, M., Sissons, J. G., and Carmichael, A. Progressive loss of IL-2-expandable HIV-1-specific cytotoxic T lymphocytes during asymptomatic HIV infection. Eur. J. Immunol. 28: 3564, 1998. Tsunetsugu-Yokota, Y., Tamura, H., Tachibana, M., et al. Selective expansion of perforin-positive CD8⫹ T cells by immature dendritic cells infected with live Bacillus Calmette-Guerin mycobacteria. J. Leukoc. Biol. 72: 115, 2002. Picker, L. J., Singh, M. K., Zdraveski, Z., et al. Direct demonstration of cytokine synthesis heterogeneity among human memory/effector T cells by flow cytometry. Blood 86: 1408, 1995.
Depletion of Activated V8⫹ T Cells Disrupts Bispecific Antibody Directed Antitumor Immunity 1 Elizabeth J. McConnell, M.D.,*,2 Elisabeth C. McLemore, M.D.,* Robert Talac, M.D.,† Lokesh Joshi, Ph.D,* and Heidi Nelson, M.D.† *Arizona State University, Arizona Biodesign Institute, Tempe, Arizona; and †Mayo Clinic, Department of Colon and Rectal Surgery, Rochester, Minnesota Submitted for publication April 6, 2004
Introduction. Activation of V8ⴙ T cells with superantigen staphylococcal enterotoxin B (SEB) and use of an antitumor, anti-CD3 bispecific antibody (BsAb) leads to tumor protective immunity. We hypothesize that V8ⴙ T-cell activation in combination with BsAb is crucial for tumor protective immunity in this model. Methods. Adolescent C3H/HeN mice were intravenously injected with syngeneic CL62 melanoma to establish pulmonary metastasis. Three days after establishing pulmonary metastasis, predominantly V8ⴙ T cells are activated with 50 g of intraperitoneal superantigen SEB. T cells were depleted at different time points in relation to SEB administration to assess the effect on protective immunity against a second tumor challenge. Results. Protective immunity is significantly (P < 0.008) decreased when V8ⴙ depletion occurs 6 h after SEB injection, as growth of rechallenged CL62 melanoma occurred in 43%. Protective immunity is present at all other time points when mice survive V8ⴙ T-cell depletion. Survival of animals treated with SEB/BsAb (82%) is significantly better (P < 0.002) than with SEB alone (60%) or nontreated control (0%). Survival when V8ⴙ T-cell depletion occurred at 6 h and 48 h postSEB is 72% and 77%, respectfully, and is statistically indistinguishable (P < 0.232 and P < 0.602). If T-cell depletion was conducted before SEB administration, however, the combination of SEB and BsAb did not result in significant protective immunity. T-cell deple-
1 Supported by an intramural grant from the Department of Colon and Rectal Surgery of the Mayo Foundation. 2 To whom correspondence and reprint requests should be addressed at Arizona State University, Harrington Department of Bioengineering, Mail Code 9709, Tempe, AZ 85282-9709. Fax: 480-7276899 . E-mail: [email protected].
tion before the use of SEB alone, without BsAb, failed to result in significant protective immunity. Conclusions. Depletion of V8ⴙ T cells 6 h after activation disrupts the development of protective immunity. © 2004 Elsevier Inc. All rights reserved. Key Words: bispecific antibody; protective immunity; immunotherapy; staphylococcal enterotoxin B; V8ⴙ T cell. INTRODUCTION
Bispecific antibodies (BsAb) and superantigens have both been used individually and in combination in the treatment of solid malignant tumor models resembling metastatic disease because these models are easily targeted with immunotherapy [1–3]. The effector cells responsible for the cytolytic destruction of the tumor are considered to be the primary cells involved in protective immunity [4 –7]. In a metastatic melanoma model, we used a tumor and CD3 T-cell BsAb and the superantigen staphylococcus enterotoxin B (SEB) to determine any functional role in the timing of activation of the effector cells in the development of tumor protective immunity. Using this model in previous studies, we have shown the successful induction of tumor antigen-specific protective immunity in mice [8, 9]. After intravenous injection of CL62 tumor cells, mice developed pulmonary metastasis. Thereafter, superantigen SEB and CL62 tumor x anti-CD3 T-cell BsAb were administered alone or in combination. Surviving mice (superantigen SEB alone or in combination with BsAb) underwent a second tumor challenge with either CL62, K1735, or AG104 tumor cells. Most mice rejected the second tumor challenge with CL62. In addition, most mice rejected the p97-negative parental cell line, K1735, indi-
103
0022-4804/04 $30.00 © 2004 Elsevier Inc. All rights reserved.
104
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
cating an immune response to tumor antigens common to both cell lines that were not the original target of the BsAb. Alternatively, injection of the syngenic AG104 tumor cell line resulted in tumor growth [9]. Evidence of cellular immunity was obtained from the results of delayed-type hypersensitivity, proliferation, and cytotoxicity assays, which revealed the presence of tumorspecific memory in BsAb-treated, CL62-cured mice [9]. The critical effector cell, the V8⫹ T cell, is the predominant cell activated by the superantigen SEB [9]. Whether V8⫹ T-cell activation is critical for the development of immune memory or protective immunity has yet to be determined. The scope of this study is limited to the investigation of the V8⫹ T cell. Superantigen SEB activation of V8⫹ T cells leads to the release of TH1 cytokines, which ultimately activate the entire T-cell repertoire. Activation also occurs in conjunction with the anti-CD3 F(ab=)2 arm of the bispecific antibody. Other studies suggest that SEB is an effective stimulator of both cell-mediated and innate immunity [10]. SEB is known to increase antigen-specific CD4⫹ memory T cells [11]. When used with irradiated B16F10 melanoma cells, formation of effective, but rather short-lived, tumor immune memory occurs [12]. The role of the CD4 and CD8 T cell in the development of protective immunity has been examined in an additional experiment and will not be discussed in this paper. In summary, this study was conducted to determine the effect of depletion of V8⫹ T cells at different time points before and after the administration of SEB alone or in combination with a BsAb to determine the role the effector V8⫹ T cells may play in the development of tumor immunity. METHODS Tumor Models Female C3H/HeN mice (age, 6 to 9 weeks) were purchased from Harlan Sprague Dawley (Indianapolis, IN, USA), and IL-2 knockout C3H/HeJ mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA). All animals were maintained in accordance with the animal care guidelines of the National Institutes of Health. Pulmonary metastases were established in C3H/HeN or C3H/HeJ mice by injecting 5 ⫻ 10 5 CL62 melanoma tumor cells suspended in 500 l of phosphate-buffered saline (PBS) intravenously via the tail vein (day 0).
Cell Lines CL62, a murine melanoma cell line, is a clone derived from the K1735 melanoma syngeneic to C3H/HeN mice transfected with a gene encoding p97 human tumor antigen. Both CL62 and K1735 tumor cell lines were provided by Drs. K.E. Hellstrom and I. Hellstrom (Oncogene Corp., Seattle, WA, USA) and maintained as an adherent monolayer culture in complete media, as described previously [5]. Cells were routinely tested by flow cytometry to determine if ⬎90% of CL62 melanoma cells expressed p97 antigen.
Bispecific Antibody Bispecific antibody was prepared by covalently conjugating the F(ab=)2 fragment of a hamster antimouse CD3 antibody 500A2 to the F(ab=)2 fragment of the hybridoma producing murine monoclonal antibody 96.5, as described previously [5]. The 500A2 F(ab=)2 fragment binds to the epsilon chain of the murine CD3 molecule; it was produced by intraperitoneal growth of the 500A2 hybridoma (provided by Dr. D.J. McKean, Mayo Clinic, Rochester, MN, USA). Murine monoclonal antibody 96.5 binds p97 human melanoma antigen expressed on ⬎90% of melanoma cells. The purity and function of the BsAb was confirmed using two-color flow cytometry analyses by a Becton-Dickinson FACscan using the Cell Quest program and chromium release assay [5].
Treatment Protocol The CL62 tumor cells were harvested in the log phase of growth using trypsin and washed. Some tumor cells were subjected to a single dose of irradiation (10 Gy) from a 60Co source (1.08 Gy/min). Half a million live or radiation-killed tumor cells suspended in 200 L of PBS were injected intravenously in the tail vein of the mice. Treatment was started 3 days after tumor inoculation. Tumorbearing mice were injected intraperitoneally with 50 g of SEB (day 3) suspended in 500 l of PBS and 4 days later (day 7) with 5 mg of BsAb injected intravenously. Mice were observed weekly for survival. Treated mice surviving beyond 60 days after inoculation of CL62 tumor cells were considered “cured” and were injected subcutaneous with CL62 tumor with 5 ⫻ 10 6 tumor cells (CL62 or K1735) administered subcutaneous and further observed for tumor growth and survival weekly for the subsequent 300 days.
In Vivo Depletion Studies To determine whether the expansion of V8⫹ T cells plays a role in the production of effector cells that provide tumor-specific immunity, CL62 melanoma-bearing mice were depleted of V8⫹ T cells by injection of 200 g of F23.1 monoclonal antibody (provided by Dr. C. S. David, Mayo Clinic). As a control, we also analyzed depletions of CD4⫹ T cells, by injection of 200 g of GK1.5 monoclonal antibody (ATCC, Rockville, MD, USA); CD8⫹ T-cells, by injection of 200 g of 116-13.1 monoclonal antibody (ATCC); and V7⫹ T cells by injection of 500 g of TR310 (provided by Dr. C. S. David). Treatment with the respective monoclonal antibody leads to ⬎80% depletion of a relevant T-cell subset, as determined by FACscan analyses in more than 15 mice analyzed in perfecting this technique. Depletion of controls CD4⫹, and CD8⫹ T cells were performed at various time points: 7 days before, and 6 h and 48 h after SEB administration.
Assessment of Serum Cytokine Levels To measure serum cytokine levels, C3H/HeN mice were bled via the tail vein at 3, 6, 12, 18, 24, 30, 48, and 72 h after SEB administration. Th-1 type (IL-2, IFN-␥, IL-12) and Th-2 type (IL-4, IL-10) cytokine levels were determined with enzyme-linked immunosorbent assay (ELISA) using commercially available kits (PharMingen, San Diego, CA, USA), according to the manufacturer’s instructions. Serum was separated and stored at – 80°C until the ELISA was preformed. The detection limits of the assays were 40 pg/ml for all cytokines. Positive controls were used in all analyses.
In Vivo Cytokine-Dependent Studies The contribution of IL-2 and IFN-␥ cytokines to antitumor effects and tumor-specific memory was assessed in vivo by depletion of these specific cytokines. For IL-2 cytokine studies, C3H/HeJ IL-2 knockout and C3H/HeJ (control) mice were used. The mice were treated according to the treatment protocol described above and monitored for survival. Because IL-2 knockout mice are susceptible to inflamma-
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
105
RESULTS V8 T-Cell Populations Expand with SEB Treatment
FIG. 1. Characterization of the T-cell activation. Flow cytometric analysis of receptors of T cells found within the spleen is represented. The splenocytes are specifically labeled with antibody bearing fluorescent dyes to evaluate the expression and presence of V8⫹, CD4⫹, CD8⫹ and CD3⫹ T cells. Antibody to CD3⫹ is the prominent antigen expressed on all T cells. Lag phase is from 0 to 12 h, expansion phase from 6 to 24 h, and death phase from 24 h on. Expansion of T cells after administration of staphylococcal enterotoxin B (SEB) doubled based on CD3 labeling (20% to 40%). T cells were also labeled with CD4⫹ and CD8⫹ with a different fluorescent wavelength dye from V8⫹ to determine specific group expansion. CD4⫹V8 ⫹ () and CD8 ⫹V8 ⫹ (Œ) percent expression was reported as a function of the total CD3⫹ T-cell population. The percentage of cells expressing the labeled characteristics is represented. The CD4⫹V8 ⫹ () increased threefold and the CD8 ⫹V8 ⫹ (Œ) population increased fourfold. The total number of V8 ⫹ T cells within a 48-h period rose from 25% to ⬎80% of the total T-cell population within the spleen. Activation-induced cell death occurred in the subsequent 48 h with a dramatic drop of the population almost back to baseline at 72 h.
tory bowel disease, subcutaneous injection of 5 ⫻ 10 6 CL62 and K1735 tumor cells, respectively, was performed once significant survival differences had been observed (on day 45 after tumor inoculation) and only in the absence of any symptoms. The mice were monitored for survival and tumor growth by palpation and caliper measurement. For IFN-␥, the C3H/HeN mice, treated according to the protocol described above, were injected intraperitoneally with antimouse IFN-␥ monoclonal antibody (R4-6A2) suspended in 1 ml PBS (ATCC). The antimouse IFN-␥ monoclonal antibody was injected on day 0, 1, 3, and 8 in doses of 3, 3, 1, and 1 mg, respectively. Mice were bled and serum collected to determine IFN-␥ depletion. Cured mice were injected subcutaneous with 5 ⫻ 10 6 CL62 or K1735 tumor cells and monitored for survival and tumor growth.
Administration of intraperitoneal SEB is not lethal in C3H/HeN mice and results in a well-documented expansion of the V8⫹ T-cell subsets. Production of CD8⫹ and CD4⫹ is directly increased by SEB administration (Fig. 1). Stimulation of the V8⫹ T cells by SEB increased the percentage of the T-cell population within the spleen more than 3.5 times above baseline. This increase is also seen with both the CD4⫹ T cells, which increased three times, and the CD8⫹ T-cells, which increased four times. The naïve control, which is the first point of the graph, defines the baseline. The in vivo responses of CD3, CD4⫹, CD8⫹, V3, V6, V7, V8.1-3, V11, and V12 positive T-cell subpopulations were measured by FAC analysis at different time points (data not shown). During the first 12 h, the CD3⫹ T-cell population declined, whereas the V8⫹ T-cell population remained at a steady state. In the second phase, the V8⫹/CD4⫹ and V8⫹/CD8⫹ T-cell population increased three- to fourfold. The V8⫹/ CD4⫹ and V8⫹/CD8⫹ T cell reached maximal population levels 48 h after stimulation. Thereafter, both CD4⫹ and CD8⫹ T-cell populations declined and rapidly underwent activation-induced cell death. Similar Survival with BsAb Administration at 24 or 72 Hours After SEB
After establishment of CL62 pulmonary metastasis on day 3, T-cells were activated with i.p. administration of SEB. Then, BsAb was administered on either day 4 or day 7. Our previous publications of this model had always used BsAb only at day 4, just
Statistical Methods Survival data were analyzed with the Kaplan-Meier method and displayed as actuarial survival curves. The survival curves were compared with a log-rank test. Probability values ⬍0.05 were considered significant. Statistical significance was assessed using oneway ANOVA model. The margins of error for comparison were obtained by calculating the 95% confidence intervals for the differences between groups. The tumor was determined to be present or absent. All experiments were performed at least three times, and representative data are displayed.
FIG. 2. Survival after CL62 pulmonary metastasis in C3H/HEN mice. Following the intravenous injection of CL62 tumor cells, mice were treated with staphylococcal enterotoxin B (SEB) on day 3 and bispecific antibody (BsAb) on day 4 (□ n ⫽ 40) or day 7 (, n ⫽ 34).
106
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
FIG. 3. Treatment and V8 depletion timeline. Treatments and V8 depletions are depicted in a linear timeline. SEB ⫽ staphylococcal enterotoxin B.
24 h after SEB [5, 8, 9]. Survival is similar in each group, regardless of use at day 4 or day 7 (72% versus 92%) as shown in Figure 2. Administration of the BsAb on day 7 allows time for depletion of the expanding V8⫹ T cells. V8ⴙ T-Cell Depletion at Three Different Time Points
Figure 3 demonstrates the course of treatment and depletion of V8⫹ T cells in this murine tumor model. The timeline demonstrates that the V8⫹ T-cell depletions were performed (200 g of F23.1 monoclonal antibody) either just before, or 6 h or 48 h after SEB administration. Surviving mice were monitored for 300 days.
Significant Survival after V8ⴙ T-Cell Depletion Requires BsAb
Overall, depletion of the V8⫹ T cells did not affect survival if depleted after SEB administration (Fig. 4). Depletion of the expanding V8⫹ T cells before SEB administration resulted in significantly impaired survival when compared with the SEB treated, nondepleted controls (P ⬍ 0.05). Depletion of V8 ⫹ T-cells either 6 or 48 h after SEB, however, was not statistically different (72% versus 77%) when compared with the treated control 82% (P ⬍ 0.232 and P ⬍ 0.602). All mice treated with irradiated tumor cells survived. SEB administration alone and in combination with BsAb provided protective immunity to tumor challenge. Radiation killed tumor cells alone did not result in protective immunity (data not shown). SEB Stimulated T Cells are Cytotoxic and Tumor Targeted with BsAb
The in vitro T-cell cytotoxicity against the CL62 tumor is tested in vivo on V8⫹ and V7⫹ depleted mice. SEB administration predominantly activates V8⫹ T cells, creating cytotoxic T lymphocytes (CTL), but it also activates V7⫹ T cells. As a control, the T cells of mice depleted of either V7⫹ or V8⫹ T cells were tested before and after SEB administration. Depletion of the V8⫹ subset after SEB activation did not change targeted cytotoxicity of V8⫹ T cells in vitro (Figure 5). Decrease in Protective Immunity with V8ⴙ T-Cell Depletion 6 Hours After SEB FIG. 4. Survival curve with V8 depletion. Additional groups of mice were then treated with staphylococcal enterotoxin B (SEB) and bispecific antibody (BsAb), but depleted of V8 T cells, either before SEB (‘, n ⫽ 39), 6 h after SEB (⽧, n ⫽ 22), or 48 h after SEB (’, n ⫽ 9).
As shown in the Figure 3 time line, the “cured” mice were evaluated for immune memory by a second subcutaneous injection tumor cells. Our previous studies demonstrated that tumor-specific immunity existed against targeted (p97⫹) and nontargeted an-
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
107
FIG. 5. In vitro T-cell cytotoxicity is tested on in vivo V8 and V7 depleted mice Depletion of V8⫹ subset after staphylococcal enterotoxin B (SEB) activation does not change targeted cytotoxicity of V8-T Cells in vitro. Splenocytes were taken from (A) untreated (), (B) V8-depleted after SEB (⽧), (C) V8-depleted before SEB (‘), or (D) V7-depleted () C3H/HeN mice, which were all co-cultured with SEB and irradiated CL62 melanoma cells. Splenocytes were then placed over a nylon wool column and T cell-purified and tested for direct CTL activity on CL62 cells (dashed line) or targeted CTL activity with the bispecific antibody (BsAb) 96.5 ⫻ 500A2 (solid line). Group B and C T cells were depleted of V8 T cells and group D (controls) depleted of V7 T cells and treated with SEB. Data shown represent one of three experiments performed with similar results. E/T, effector target ratio.
tigens [9]. These studies were performed on mice that survived 60 days after the treatment had “cured” the mice and were injected subcutaneously with 5 ⫻ 10 6 CL62 or K1735 tumor cells and observed for tumor growth for 100 days. Table 1 demonstrates the results of all mice having a secondary tumor challenge with either the CL62 or K1735 tumor cell line. A large percentage of mice depleted of V8⫹ T cells 6 h after SEB administration through
the lag and expansion (log) phases grew CL62 tumor and K1735 tumor (43% and 64% growth, respectively). A significant absence was noted in tumorspecific immunity to the targeted antigens of CL62 compared with control (P ⬍ 0.008 and P ⬍ 0.134, respectively). Tumor growth occurred within the first 50 days in all of the control CD4⫹ and CD8⫹ T-cell depleted mice (100%). Tumor growth did not occur in the mice that survived SEB/BsAb treatment (0%
108
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
TABLE 1 Tumor Growth in Control and Treated Animals After Subcutaneous Rechallenge Tumor Groups
CL62
Grew (%)
Naive control SEB cured SEB/V8⫹ T cells depleted/BsAb cured –6 h –48 h SEB/BsAb cured: –V8⫹ T cells depleted –CD4⫹ T cells depleted –CD8⫹ T cells depleted
21/22 0/22*
95% 0%
6/14* 0/6 0/42* 0/9* 9/9. 9/9.
43% 0% 0% 0% 100% 100%
K1735 21/22 5/22* 9/14 T 0/6 1/42* 0/9* 9/9. 9/9.
Grew (%) 95% 23% 64% 0% 0% 0% 100% 100%
TP ⬍ 0.134 compared with naive control.SEB ⫽ staphylococcus enterotoxin B; BsAb ⫽ bispecific antibody. * P ⬍ 0.008 compared with naive control.
growth). Mice depleted of V8⫹ T cells 48 h after SEB rejected tumor (0% growth). Cytokines INF-␥ and IL-2 Increase in Serum after SEB Administration
The administration of SEB resulted in a rapid increase in the serum levels of Th-1 cytokines, IL-2, and INF-␥ as shown in Figure 6. IL-2 serum levels peaked at 3 h, as shown in Figure 6, correlating with the lag phase as shown in Figure 1. INF-␥ levels peaked at 12 h, as shown in Figure 6, which correlates with the time period just before the expansion phase as shown in Figure 1. IL-10 serum levels rose, as shown in Figure 6, during the expansion phase of V8⫹ T cells, as shown in Figure 1. Levels of IL-12 and IL-4 were not detectable. This experiment, which was repeated with depletion of V8⫹ T cells before SEB administration, resulted in no detectable level of cytokines. Serum cytokine levels were also measured after BsAb administration alone, which alone did not result in enhanced cytokine production nor did it enhance SEB-induced cytokine production (data not shown). Cytokine Depleted Mice Survive and Demonstrate Antitumor Memory
To confirm the role of IL-2 and INF-␥ in survival and tumor-specific immunity in this treatment model, the same studies were preformed with IL-2 knockout or INF-␥ monoclonal antibody-depleted mice. The transgenic IL-2 knockout mice (Jackson Laboratories) and INF-␥ monoclonal antibody-depleted mice were administered SEB and BsAb after establishment of pulmonary metastasis as described above. Survival patterns for the IL-2 knockout (49%) or INF-␥ depleted (79%) mice were statistically the same as the SEB/BsAb treated wild-type control (Figs. 7 and 8). By testing the immune memory with tumor rechallenge in these
cytokine-deficient mice, we demonstrated the same rejection pattern as in the “cured” wild-type controls. All mice were challenged at day 45 and rejected both CL62 and K1735 tumors (data not shown). The life span of both the IL-2 and INF-␥ depleted control mice was shorter than the nondepleted naïve control. Mice in both these groups were only observed for 60 days before sacrifice. The number of cytokine-depleted mice was significantly smaller than any other group so differences may have gone undetected. DISCUSSION
This study demonstrates that the early recognition of tumor antigen by activated T cells is critical to the formation of protective immunity. No previous studies have investigated the role of time point depletion of effector V8⫹ T cells to determine if the T-cell directed protective immunity is dependent on the presence of the initially activated T cells. A detailed evaluation of T-cell exposure to tumor antigen and the time frame of subsequent memory T-cell formation is critical to the understanding of immune function and the development of protective immunity [5, 8, 9]. The T-cell immune memory is long-lasting and the effect is tumor antigen-specific [5, 8, 9]. By combining the use of superantigen SEB with the depletion of the V8⫹ T-cell population, we studied the effect of tumor rechallenge on the T-cell memory response. We provided direct evidence that the timing of V8⫹ T-cell depletion is vital to sustaining memory T-cell protective immunity. Additionally, we confirmed that significant survival was dependent on the administration of BsAb. Initially, we were interested in determining if the T-cell protective immunity was dependent on V8⫹ T cells, the predominant T cell activated by SEB. We have previously reported that immune memory is T-cell mediated and is dependent on both the tumor
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
FIG. 6. Staphylococcal enterotoxin B (SEB) administration increased serum levels of IL-2 and interferon (INF)-␥. L-2 serum levels peaked at 3 h, INF-␥ levels peaked at 12 h, and IL-10 serum levels rose over 24 h.
antigen targeted by the BsAb, and the nontargeted antigens present on the tumor cell surface [8]. Mice
109
treated with SEB alone demonstrated increased survival after tumor introduction [5, 8, 9]. Survival, however, was not as dramatic when compared with the survival of mice treated with both the SEB and BsAb [5, 8, 9]. We concluded that two mechanisms are involved in T-cell memory production: BsAb activation of the anti-CD3 T-Cell receptor and SEB promotion of signaling in naïve T cells [9]. Superantigens, which have been used as T-cell activators, are effective stimulators of both cell-mediated and innate immunity [10]. Investigators have found that when superantigen is administered at the time of OX40 ligation, a 60 times increase occurs in the antigen-specific CD4⫹ memory T cells [11]. The superantigen activation of tumor-specific CD4⫹ T-helper lymphocytes is pivotal because CD4⫹ T cells are regulatory cells in the generation of long-term immune memory. Others have demonstrated that the formation of effective tumor immune memory is possible with the use of irradiated B16F10 melanoma cells with superantigen [12]. SEB has recently been shown to have a role in the activation of dendritic cells in the lymphoid organs of mice [13]. SEB-induced T-cell activation is well documented [14], and other investigators have shown that SEB directly activates V8⫹ cytotoxic T cells. SEB induces cytotoxicity by several mechanisms, including superantigen-mediated, antibody-dependent T-cell cytotoxicity [15]. Our research has shown that immunization with SEB alone, in this model and in a transgenic spontaneous pancreatic tumor model, activates and matures tumor antigen-specific cytotoxic T cells [16]. All of these studies focus on the role of the superantigen. Superantigen alone, however, is not sufficient for survival when V8⫹ T cells are depleted in this model, which implies a significant role for protection of the animal by the BsAb. Use of a BsAb in our model makes a significant contribution to enhancing survival and we conclude also protective immunity. The BsAb provides an effective stimulator and modulator of the immune system [17]. Recent studies have shown that a BsAb made by the combination of a superantigen and antitumor C215 monoclonal antibody improved survival and resulted in CD4⫹ T-cell– dependent cytokine release in a C215transfected syngeneic B16 melanoma model [18]. We recognize the limitations to a full understanding of how immune memory is established, even in this model. The independent and unique role of T-cell expansion, particularly CD4⫹T cells, in the absence of BsAb could not be explored in this study. Treatment of any of the mice with the T-cell activator and depletion with no BsAb resulted in a low survival rate, and the number of mice needed to establish immune memory without BsAb would have been prohibitively large. The correlation between the time-specific depletion of expanding T cells and the loss of tumor protective immunity could only be controlled for by a consistent administration of
110
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004
FIG. 7. Survival with interferon (IFN)-␥ depleted C3H/HEN mice. Before intravenous injection of CL62 tumor cells, C3H/HEN mice were treated intraperitoneally with 1 mg of R4682 monoclonal antibody (mAb). The depletion was carried out over an 8-d period while mice where treated with either saline control (, n ⫽ 5); with staphylococcal enterotoxin B (SEB) on day 3 and bispecific antibody (BsAb) on day 4 (not shown, n ⫽ 5, survival 10%); SEB day 3 and BsAb day 7 (, n ⫽ 5); SEB alone on day 3 (⽧, n ⫽ 5). Tumor-bearing, IFN-␥ depleted mice were also treated with SEB on day 3 and BsAb on day 7 (□, n ⫽ 4).
BsAb at the same time point in all depletion experiments. Use of radiation-killed tumor cells would provide no further benefit in answering this question because all mice demonstrated protective immunity with the use of SEB alone. The role of T-cell expansion in tumor protective immunity, specifically CD4⫹ T cells, has been theorized in numerous reviews. Reports on animal models of tumor protective immunity are limited, however. Tumor protective immunity was not completely absent in V8⫹ T-cell depleted mice. This suggests that the overall immune response elicited by the superantigen with the BsAb, not the V8⫹ T-cell activation alone, is sufficient for the recall of the effector T cell. IL-2 cytokine stimulation occurred within 24 to 48 h after superantigen SEB administration. Depletion of V8⫹ T cells after SEB administration did not affect survival, with the exception of the group that did not receive BsAb. Selection of the superantigen that predominantly expands the V8⫹ T-cell population (Fig. 1) followed by depletion of these T cells after SEB administration during the entire expansion phase (6 h after SEB administration) did affect the development of tu-
mor protective immunity (Table 1). Depletion of other T cells, such as V7⫹, which are moderately increased with superantigen administration, has no effect (data not shown). Depletion of other larger subsets of T cells, such as CD4⫹ and CD8⫹ T-cells, reveals the importance of both helper and cytotoxic T cells in this cellmediated process as no mice survived with these T-cell depletions. Further evidence that the tumor protective memory is dependent on T-cell expansion rather than a specific cytokines is exemplified in the cytokine depletion experiments that show survival and serum cytokine levels correlate with, but do not determine the different phases of T-cell activation (Figs. 6, 7, and 8). Exploring the role of IL-2 and INF-␥ provides limited additional information because the life span of the cytokine-depleted mice is short and, therefore, the time frame for detecting the establishment of immune memory is not ideal. Others have demonstrated no specific surface marker or cytokine expressed that distinguished naïve from tumor protective T cells. Regardless of stimulation, CD8⫹ T cells produced IFN-␥ twice as much as CD4⫹ T cells, whereas, IL-2 production was predominantly a CD4⫹ T cell event that increased
MCCONNEL ET AL.: T CELLS IN PROTECTIVE IMMUNITY
FIG. 8. Survival with IL-2 C3H/HEJ knockout mice. Following the intravenous injection of CL62 tumor cells, tumor-bearing, IL-2 knockout mice were treated with staphylococcal enterotoxin B (SEB) on day 3 and bispecific antibody (BsAb) on day 7 (□, n ⫽ 4). As controls, immunocompetent C3H/HEJ mice were treated with either saline (control , n ⫽ 5), with SEB on day 3 and BsAb on day 4 (not shown, n ⫽ 5, survival 10%), or SEB on day 3 and BsAb day 7 (, n ⫽ 5); SEB alone on day 3(⽧, n ⫽ 5).
with cell division [19]. Transgenic IL-2 knockout mice were used to help decipher if cytokines played a role in protective immunity. The background of the transgenic mouse is C3H/JeN, which is different from the mouse model in which the pattern of tumor growth had been established. The life span of both the control and transgenic mouse in this model was shorter than predicted. This made the establishment of survival with SEB and BsAb treatment a challenge. Our methods of analysis on the survival differences in the cytokine-depleted mice, which may or may not exist, would have gone undetected with the limited number of animals used in this study. In addition, the depletion by the monoclonal antibody in all of the experiments is not 100% complete. Our findings reveal novel observations that provide a more complete understanding of the timing and events in the formation of protective immunity. We hypothesized that three effectors of protective immunity exist in this model: superantigen, BsAb, and tumor cells. The timeline of events for both the activation of T cells and the generation of protective immunity in this model implicates the role of both CD4⫹ and CD8⫹T cells. Initial investigations into T-cell memory implicated the CD4⫹ T cell as the instigator of two distinct stages of cytotoxic T-lymphocyte development during immune response: an early primary stage dependent on the presence of CD4⫹ cells and a later, CD4independent stage operative during the secondary response, which decays with time [20]. Decline of in vivo
111
antigen-activated CD8⫹ T cells or a poor T-cell expansion response is suggested to result from a reduction in CD4⫹ T cells [21]. Others have confirmed that the interaction between antigen-infected dendritic cells and CD8⫹ T cells through CD40/CD40L was crucial for the expansion and maturation of CD8⫹ memory T cells, the process of which was CD4-independent and independent of IFN-␥ production [22]. Our tumor model may be limited in that it is an injected tumor model. We have previously shown that the immune response of a mouse with a spontaneous cancer is unlike an immune response of a mouse with an injected tumor [16]. Given the limitations of the model, we will continue to explore the role of each effector in the current mouse model. In summary, we have shown that superantigen SEBactivated V8⫹ T cells in the expansion phase are a potent source of immune modulators in an immune competent mouse model of an injectable melanoma. In combination with a bispecific antibody, the SEB activated V8⫹T cells leads to tumor protective immunity. Naïve T-cell responses to both SEB and BsAb are neither independent nor uniform in nature [19]. Naïve T cells respond to different stimuli in different ways, and predictably behave differently in a spontaneous tumor model, which proves difficult to study [23]. The immediate downstream effect of V8⫹ T-cell expansion is an increase in serum IL-2 and INF-␥. Despite the correlation of loss of protective immunity with early depletion of V8⫹ T cells, the depletion had no effect on survival. We presume that BsAb administration, which is crucial for survival, also has a role in assisting SEB activation in the development of long-lasting protective immunity. REFERENCES 1.
Talac, R., and Nelson, H. Current perspectives of bispecific antibody-based immunotherapy. J. Biol. Regul. Homeost. Agents 14: 175, 2000.
2.
Penna, C., Dean, P. A., and Nelson, H. Pulmonary metastases neutralization and tumor rejection by in vivo administration of beta glucan and bispecific antibody. Int. J. Cancer 65: 377, 1996.
3.
Penna, C., Dean, P. A., and Nelson, H. Antitumor x anti-CD3 bifunctional antibodies redirect T-cells activated in vivo with staphylococcal enterotoxin B to neutralize pulmonary metastases. Cancer Res. 54: 2738, 1994.
4.
Ramsey, P. S., Dean, P. A., Tasimowicz-Alpini, A., Donohue, J. H., and Nelson, H. Interferon-gamma-induced intercellular adhesion molecule-1: Expression on human tumor cells enhances bifunctional antibody mediated lysis through a lymphocyte function-associated antigen-1 dependent mechanism. Cancer Res. 53: 4652, 1993.
5.
Chapoval, A. I., Nelson, H., Thibault, C., Penna, C., and Dean, P. Bifunctional antibody retargeting in vivo-activated T lymphocytes: Simplifying clinical application. J. Hematother. 4: 571, 1995.
6.
Chapoval, A. I., Nelson, H., and Thibault, C. Anti-CD3 x anti-
112
7.
8.
9.
10.
11.
12.
13.
JOURNAL OF SURGICAL RESEARCH: VOL. 122, NO. 1, NOVEMBER 2004 tumor F(ab=)2 bifunctional antibody activates and retargets tumor-infiltrating lymphocytes. J. Immunol. 155: 1296, 1995. Thibault, C., Nelson, H., and Chipoval, AI. Tumor-infiltrating lymphocytes can be activated in situ by using in vivo activants plus F(ab=)2 bispecific antibodies. Int. J. Cancer 67: 232, 1996. Porter, L. E., Nelson, H., Ethem Gecim, I., et al. T cell activation and retargeting using staphylococcal enterotoxin B and bispecific antibody: an effective in vivo antitumor strategy. Cancer Immunol. Immunother. 45: 180, 1997. Rice, D. C., Chapoval, A. I., Porter, L., and Nelson, H. Induction of antitumor immunity after cure of pulmonary metastases, using staphylococcal enterotoxin B and bispecific antibody. Cancer Immunol. Immunother. 48: 230, 1999. Litton, M. J., Dohlsten, M., Hansson, J., et al. Tumor therapy with an antibody-targeted superantigen generates a dichotomy between local and systemic immune responses. Am. J. Pathol. 150: 1607, 1997. Maxwell, J. R., Weinberg, A., Prell, R. A., and Vella, AT. Danger and OX40 receptor signaling synergize to enhance memory T cell survival by inhibiting peripheral deletion. J. Immunol. 164: 107, 2000. Kominsky, S. L., Torres, B. A., Hobeika, A. C., et al. Superantigen enhanced protection against a weak tumor-specific melanoma antigen: Implications for prophylactic vaccination against cancer. Int. J. Cancer 94: 834, 2001. Yoon, S., Bae, K. L., Shin, J. Y., et al. Analysis of the in vivo dendritic cell response to the bacterial superantigen staphylococcal enterotoxin B in the mouse spleen. Histol. Histopathol. 16: 1149, 2001.
14.
Marrack, P., and Kappler, J. The staphylococcal enterotoxins and their relatives. Science 248: 705, 1990.
15.
Holzer, U., Bethge, W., Krull, F., et al. Superantigen-
16.
17.
18.
19.
20.
21.
22.
23.
staphylococcal-enterotoxin-A-dependent and antibody-targeted lysis of GD2-positive neuroblastoma cells. Cancer Immunol. Immunother. 41: 129, 1995. McConnell, E. J., Pathangey, L. B., Madsen, C. S., et al. Dendritic cell-tumor cell fusion and staphylococcal enterotoxin B treatment in a pancreatic tumor model. J. Surg. Res. 107: 196, 2002. Kodama, H., Suzuki, M., Katayose, Y., et al. Specific and effective targeting cancer immunotherapy with a combination of three bispecific antibodies. Immunol. Lett. 81: 99, 2002. Rosendahl, A., Kristensson, K., Hansson, J., et al. Repeated treatment with antibody-targeted superantigens strongly inhibits tumor growth. Int. J. Cancer 76: 274, 1998. Brenchley, J. M., Douek, D. C., Ambrozak, D. R., et al. Expansion of activated human naive T-cells precedes effector function. Clin. Exp. Immunol. 130: 432, 2002. Jennings, S. R., Bonneau, R. H., Smith, P. M., et al. CD4positive T lymphocytes are required for the generation of the primary but not the secondary CD8-positive cytolytic T lymphocyte response to herpes simplex virus in C57BL/6 mice. Cell Immunol. 133: 234, 1991. Jin, X., Wills, M., Sissons, J. G., and Carmichael, A. Progressive loss of IL-2-expandable HIV-1-specific cytotoxic T lymphocytes during asymptomatic HIV infection. Eur. J. Immunol. 28: 3564, 1998. Tsunetsugu-Yokota, Y., Tamura, H., Tachibana, M., et al. Selective expansion of perforin-positive CD8⫹ T cells by immature dendritic cells infected with live Bacillus Calmette-Guerin mycobacteria. J. Leukoc. Biol. 72: 115, 2002. Picker, L. J., Singh, M. K., Zdraveski, Z., et al. Direct demonstration of cytokine synthesis heterogeneity among human memory/effector T cells by flow cytometry. Blood 86: 1408, 1995.