Immunotherapy of Melanoma: A Dichotomy in the Requirement for IFN-γ in Vaccine-Induced Antitumor Immunity Versus Adoptive Immunotherapy

Mice

Female C57BL/6J (wt), GKO (C57BL/6-IFN-γ^tm1Ts), GRKO (129/SV-IFNγR^tm1), and PKO (C57BL/6-PFP^tm1Sdz) mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained in a specific pathogen-free environment. Mice were generally 8–12 wk old at the time of experimentation. Recognized principles of laboratory animal care were followed (Guide for the Care and Use of Laboratory Animals, National Research Council, 1996), and all animal protocols were approved by the Earle A. Chiles Research Institute animal care and use committee.

Tumor cell lines

B16BL6-D5 (D5) is a poorly immunogenic subclone of the spontaneously arising B16BL6 melanoma (18, 22) (provided by S. Shu, Cleveland Clinic Foundation, Cleveland, OH). An early passage of the original BL6 tumor was provided by E. Gorelick and was subcloned by limiting dilution culture in S. Shu’s laboratory. D5 exhibits low to undetectable class I (H-2 D^b and K^b) expression and no class II expression and does not secrete detectable amounts of IFN-γ. D5-G6 is a stable clone of D5 that was originally transduced with a murine GM-CSF retroviral MFG vector (provided by M. Arca, University of Michigan, Ann Arbor, MI) (23). D5-G6 cells secrete ∼200 ng GM-CSF /ml/10⁶ cells/24 h. MCA-101 (H-2^b) is a methylcholanthrene-induced sarcoma that exhibits undetectable levels of MHC class I and class II and low levels of Fas (provided by Nick Restifo, Surgery Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD). MPR-4 is a transformed prostate tumor cell line (provided by Thompson, Baylor College of Medicine, Houston, TX) (24) that has very low MHC class I (H-2^b) expression and no detectable MHC class II expression.

Reagents

The 145-2c11 hybridoma (anti-CD3) was a gift from J. A. Bluestone (University of Chicago, Chicago, IL). Recombinant human IL-2 was provided by M. Giedlin (Chiron, Emeryville, CA). The anti-CD4 (GK1.5, TIB-207), anti-CD8 (2.43, TIB-210), anti-NK1.1 (PK136, HB-191), anti-Mac-1 (M1/70, TIB-128), and rat anti-mouse IFN-γ (IgG1, R4-6A2) hybridomas were obtained from American Type Culture Collection (Manassas, VA). Ascites were prepared in DBA/2 mice primed with pristane and immunosuppressed by injection with 200 mg/kg cyclophosphamide. Purified anti-granulocyte Ab, Gr-1, FITC- and PE-labeled isotype control rat IgG, hamster IgG, and mAb against CD3, CD4, and CD8 were purchased from PharMingen (San Diego, CA). Freshly isolated TVDLN cells were blocked with anti-mouse Fc receptor hybridoma 2.4G2 (HB-197, American Type Culture Collection) culture supernatant before incubation with directly labeled specific Abs.

Culture conditions

Lymphocytes and tumor cells were cultured in complete medium (CM), which consisted of RPMI 1640 containing 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 2 mM l-glutamine, and 50 μg/ml gentamicin sulfate (BioWhittaker, Walkersville, MD). This was further supplemented with 50 mM 2-ME (Aldrich, Milwaukee, WI) and 10% FBS (Life Technologies, Grand Island, NY). Tumor cells were harvested two or three times per week by brief trypsinization (trypsin; BioWhittaker) and maintained in T-75 or T-150 culture flasks.

Tumor vaccination

D5-G6 tumor cells were harvested by trypsinization, washed twice with HBSS, and resuspended at 2 × 10⁷ cells/ml. One million D5-G6 (GM-CSF-transduced) tumor cells were injected s.c. into both hind and fore flanks of wt, GKO, or PKO mice. Eight days following vaccination, the draining superficial inguinal and axillary LN were harvested. TVDLN were resuspended at 2 × 10⁶ cells/ml in CM and cultured in 24-well plates with 50 μl of a 1/40 dilution of 2C11 ascites (hamster anti-mouse CD3 ε chain) as described previously (20). After 2 days of activation the T cells were harvested and expanded in CM containing 60 IU recombinant human IL-2/ml for 3 additional days. T cells were then harvested, washed twice in HBSS, counted, and used in adoptive transfer, cytotoxicity, and cytokine release assays.

Depletion of CD4⁺ or CD8⁺ T cells

Shortly before vaccination with D5-G6, in vivo depletion of T cell subsets was performed by i.v. administration of anti-CD4 (GK 1.5) or anti-CD8 (2.43) ascites (0.5 ml of 1/10 dilution). These doses were shown to effectively deplete the appropriate T cell populations in vivo for at least 13 days. Rat IgG (200 μg, I-4131, lot 086H8910, Sigma, St. Louis, MO) administered as indicated above was used as a control treatment. The same dose of Ab was administered 3 days after vaccination. After harvesting the TVDLN, the depletion of the T cell subsets was confirmed by flow cytometry. The depletion of CD4⁺ and CD8⁺ T cell subsets was >95%, respectively (data not shown). These TVDLN were used to generate effector T cells for cytotoxicity assays. The same regimen was used to deplete CD4 or CD8 T cells after adoptive transfer.

Vaccination and rechallenge experiments

To determine a possible role for IFN-γ in the induction of protective immunity upon vaccination with D5-G6, wt or GKO mice were injected s.c. with 10 × 10⁶ irradiated D5-G6 tumor cells (10,000 rad). The mice were rechallenged 14 days after vaccination with 2 × 10⁴ D5 (this is 10 × TD₁₀₀ for D5 tumor). The TD₁₀₀ is the dose at which 100% of the injected animals will develop tumor. The wt and GKO mice that survived for longer than 100 days after adoptive transfer of effector T cells were also rechallenged with 2 × 10⁴ D5 tumor cells s.c. In all rechallenge experiments naive wt and/or GKO mice were included as controls and were challenged with the same tumor dose. Tumor size was determined by measurement of two perpendicular diameters using a digital caliper.

Adoptive immunotherapy

Experimental pulmonary metastases were established by i.v. inoculation of 2 × 10⁵ D5 tumor cells. Three days later T cells were adoptively transferred i.v. Starting on the day of T cell infusion, mice received 90,000 IU IL-2 i.p. once daily for 4 days. Animals were sacrificed 11–13 days following tumor inoculation by CO₂ narcosis, and their lungs were harvested and fixed in Fekete’s solution. The number of pulmonary metastases was counted in a blinded fashion. Metastases that were too numerous to count accurately were known to be >250 metastases and were assigned a value of 250. For survival experiments, animals were followed for at least 100 days.

IFN-γ depletion

Mice with 3-day established pulmonary metastases were injected i.v. with 0.5 ml HBSS containing 200 μl rat anti-mouse IFN-γ (R4-6A2 hybridoma) ascites immediately before adoptive transfer with effector T cells and once daily for the following 3 days. Previously it was shown that the depletion of IFN-γ lasted at least 48 h (14, 15, 25). Control rat IgG (PharMingen) was injected i.v. according to the same protocol.

Measurement of cytokines

After activation and expansion, TVDLN were washed, resuspended in CM and IL-2 (60 IU/ml), and seeded at 4 × 10⁶/2 ml/well in a 24-well plate. The cells were either cultured without further stimulation or stimulated with 2 × 10⁵ D5, MPR-4 tumor cells, or immobilized anti-CD3 (positive control). Supernatants were harvested after 24 h and assayed for the release of IFN-γ, IL-10, and IL-4 by ELISA using commercially available reagents (IFN-γ, PharMingen or Genzyme (Cambridge, MA); IL-4, Genzyme; IL-10, PharMingen). The concentrations of cytokines in the supernatant were determined by regression analysis.

Cytotoxicity assay

Target cell lysis was assessed by 6-h ⁵¹Cr release assays. Tumor cells were incubated with 100 μCi Na₂⁵¹CrO₄ (NEN, Boston, MA) for 1 h, washed twice, and plated into round-bottom 96-well plates with 1 × 10⁴ target cells/well in triplicate. The target cells were incubated with effector T cells at the indicated E:T cell ratios in a total volume of 200 μl CM at 37°C in a CO₂ incubator. The supernatant was harvested and counted, and the percent specific lysis was calculated as previously described (1). Maximum lysis was achieved by incubating target cells with 2% Triton X-100 detergent.

Immunohistochemical analysis of tumor-bearing lungs

Mice with 3-day established pulmonary metastases received 90,000 IU IL-2 alone or together with the adoptive transfer of wt or GKO effector T cells. Twenty-four hours later mice were killed, and frozen sections of lungs were prepared. Tissue sections were blocked with avidin and biotin and then stained with a control rat IgG, anti-CD4, anti-CD8, anti-NK1.1, anti-Mac-1, or anti-Gr-1 Abs (specified above). Sections were washed and incubated with biotin-labeled goat anti-rat IgG, washed, and incubated with the Vectastain ABC reagent (Vector Laboratories, Burlingame, CA). Slides were developed using diaminobenzidine solution (Vector) and counterstained with hematoxylin. Images were acquired and processed with a Power Mac G3 computer equipped with a Pixera digital camera (model PVC 100c; Pixera, Los Gatos, CA).

Statistical analysis

The statistical significance of differences in the number of metastases between experimental groups was determined using the Wilcoxon rank sum test. Two-sided p < 0.05 was considered significant. Each treatment group consisted of at least five mice, and no animal was excluded from the statistical evaluations. The significance of differences in cytokine secretion was determined using Student’s paired t test. Two-sided p < 0.05 was considered significant. Statistical analysis of tumor growth in naive and vaccinated wt and GKO mice was determined by nonparametric (distribution-free) tumor growth analysis performed on medians and rank order statistics.

IFN-γ is required for vaccine-induced protective immunity

To examine the role of IFN-γ in the development of protective immunity we determined whether vaccination with irradiated D5-G6 tumor cells primes a protective immune response in GKO mice as effectively as it does in wt mice. GKO and wt mice were vaccinated with 10 × 10⁶ irradiated D5-G6 cells and challenged 14 days later with an s.c. dose of 2 × 10⁴ live D5 tumor cells. As a control, naive wt and GKO mice were also injected with the same number of D5 tumor cells. The growth of D5 tumor cells was markedly enhanced in naive GKO mice compared with naive wt mice (Fig. 1⇓, A and C). Subcutaneous tumor nodules appeared ∼2 days earlier in GKO mice and grew faster, reaching 10 mm² significantly earlier than wt mice (p < 0.05). After vaccination with D5-G6, all but one (9 of 10) of the vaccinated wt mice developed immunity to D5, but none of the GKO mice (0 of 10) was resistant to a D5 challenge (Fig. 1⇓, B and D). However, tumors in vaccinated GKO mice grew more slowly than those in naive GKO mice. These results are in accord with an earlier observation by Hung et al., who were unable to provide protection by vaccinating GKO mice with the parental B16 melanoma transduced with the gene for GM-CSF (1).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

IFN-γ is required for tumor immunity induced by active vaccination. Ten wt and 10 GKO mice were injected s.c. with 10 × 10⁶ irradiated D5-G6 cells (10,000 rad). Fourteen days later 10 naive wt (A) and GKO (C) mice or the vaccinated wt (B) and GKO (D) mice were rechallenged with 2 × 10⁴ viable D5 tumor cells. Tumor size was measured 5 days after D5 challenge. Each line represents a single mouse, and the number of mice that developed tumor/total number of mice challenged are presented in each panel.

TVDLN T cells in GKO mice exhibit an activated phenotype

The inability to induce protective immunity in GKO mice following vaccination with D5-G6 could be explained by the failure of D5-G6 to prime tumor-reactive T cells in GKO TVDLN or, if priming did occur, by a requirement for IFN-γ in the effector phase of tumor regression. To determine whether T cells in TVDLN from D5-G6-vaccinated GKO mice were primed, we examined the percentage of CD69-positive T cells in the LN of both naive and vaccinated wt and GKO mice. Fig. 2⇓ shows comparable numbers of CD69⁺ CD4⁺ and CD8⁺ T cells in the LN of naive wt and GKO mice. Following vaccination there was a similar increase in the percentage of CD69⁺ T cells in the CD4 and CD8 subsets in both wt and GKO mice. Furthermore, no substantial differences in wt or GKO mice were observed for the expression of two other T cell activation markers, OX-40 and CD44 (data not shown). This documents that despite the absence of IFN-γ, T cells in TVDLN of GKO mice were activated following vaccination with D5-G6.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

TVDLN from wt and GKO mice have similar activation phenotypes. CD69 expression on both CD4⁺ and CD8⁺ T cells in TVDLN from naive or D5-G6-vaccinated wt and GKO mice was analyzed by flow cytometry with FITC-labeled anti-CD4 or anti-CD8 and PE-labeled anti-CD69 mAb. Ten thousand gated events were collected. The data presented are the mean and SE for three independent experiments.

Effector T cells generated from D5-G6-vaccinated GKO mice exhibit tumor-specific cytotoxicity

To confirm that the T cells in the TVDLN of GKO mice were indeed primed to the D5 tumor, the effector T cells generated from the TVDLN were examined for their ability to kill tumor cells in cytotoxicity assays. Previously, we and others have shown that effector T cells generated from TVDLN following polyclonal anti-CD3 stimulation are tumor specific (18, 20, 26, 27). This specificity has been documented by tumor-specific IFN-γ release in vitro and by tumor-specific regression in vivo upon adoptive transfer into tumor-bearing mice. Recently, we observed that tumor-specific cytotoxicity could be detected more reliably if effector T cells were enriched for CD8⁺ T cells (20). To determine whether GKO mice had developed D5-specific cytolytic activity, CD8⁺ effector cells (CD4 depleted) were generated from wt and GKO mice as specified in Materials and Methods and used in standard cytotoxicity assays. CD8⁺ effector T cells generated from GKO mice exhibited low, but significant, cytotoxicity (p < 0.05) against D5 (Fig. 3⇓A). Because the D5 tumor exhibits a low level of H-2 K^b and D^b, D5 tumor was pretreated for 24 h with 10 ng/ml IFN-γ to up-regulate class I expression. We have previously shown that this increases the ability to detect tumor-specific CTL in this model (20). The cytolytic activity of GKO and wt effector T cells was much higher against the IFN-γ-pretreated D5 tumor (Fig. 3⇓B). In contrast, CD4⁺ effector T cells from either wt or GKO mice did not exhibit cytolytic activity against D5 or D5 pretreated with IFN-γ (data not shown). The cytotoxicity of CD8⁺ effector T cells was tumor specific, because the syngeneic third-party sarcoma, MCA-101, was not lysed by CD8⁺ wt or GKO effector cells (Fig. 3⇓C). Lymphokine-activated killer cells were included in all assays to confirm the lysability of tumor targets. Because tumor-specific CTL were generated from TVDLN of GKO mice, it is clear that IFN-γ is not required for the sensitization of tumor-specific T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

CD8⁺ GKO effector T cells exhibit specific cytoxicity against D5 tumor. CD8⁺ effector T cells generated from either wt (□) or GKO mice (⋄) were incubated with D5 (A), D5 tumor cultured with IFN-γ to up-regulate class I expression, D5/IFN (B), or MCA-101 (C) tumor targets for 6 h at different E:T cell ratios (100:1, 20:1, and 4:1). Lymphokine-activated killer (LAK) cells were included as a positive control for target lysability in each experiment. The results shown are representative of three independent experiments .

GKO effector T cells are therapeutic in vivo

Having shown that effector T cells generated from vaccinated GKO mice were able to exhibit tumor-specific lysis, we then examined whether these T cells could mediate tumor regression. Surprisingly, the adoptive transfer of 35 or 70 × 10⁶ effector T cells generated from GKO as well as wt mice mediated significant (p < 0.05) regression of pulmonary metastases in wt mice (Table I⇓). No difference was observed between groups treated with effector T cells generated from wt or GKO mice. To exclude the possibility that effector T cells generated in GKO mice indirectly induced the secretion of IFN-γ at the tumor site through activation of wt host lymphoid cells, effector T cells from GKO mice were adoptively transferred into GKO recipient animals with established pulmonary metastases. The control GKO mice that were treated with IL-2 developed >250 pulmonary metastases (Table II⇓). Adoptive transfer of 70 × 10⁶ GKO effector T cells into GKO mice resulted in complete regression of pulmonary metastases in these mice as well, demonstrating that T cell-mediated tumor regression in this model is IFN-γ independent.

Therapeutic effects of GKO-TVDLN cells are mediated by CD8⁺ cells

In previous studies we have demonstrated that regression of pulmonary metastases in wt mice is mediated by the adoptive transfer of CD8⁺ effector T cells. To determine the effector cell among T cells generated in GKO mice, we depleted CD4⁺ and CD8⁺ cells in vivo by i.p. injection of the appropriate mAb before and 3 days after vaccination with D5-G6. We used a dose of Ab that was known to deplete the corresponding T cell subsets effectively. T cell depletion persisted throughout the course of the experiment (data not shown). As shown in Table III⇓, CD8⁺ T cells (CD4-depleted) from both wt and GKO mice mediated a significant (p < 0.05) reduction in pulmonary metastases, while CD4⁺ T cells (CD8-depleted) failed to exhibit therapeutic activity. These results demonstrate that a CD8⁺ GKO effector T cell plays a critical role in the destruction of pulmonary metastases in this model.

GKO effector T cells cure mice with systemic tumor

Because GKO effector T cells induced regression of established D5 lung metastases, we examined whether that they would also improve long term survival. Tumor-bearing GKO mice were treated with effector T cells generated from GKO mice. Whereas all IL-2-treated control GKO mice were dead by 25 days, all animals receiving GKO effector T cells survived longer than 100 days (Fig. 4⇓). This is similar to the efficacy of wt effector T cells, following which 85% of treated animals survived longer than 100 days.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

GKO effector T cells cure mice with systemic D5 tumor. Six wt mice and 10 GKO mice were inoculated i.v. with 2 × 10⁵ D5 tumor cells. Three days later 70 × 10⁶ wt effector T cells were adoptively transferred into six wt mice with experimental metastases (□) or 70 × 10⁶ GKO effector T cells were transferred into 10 GKO mice with experimental metastases (▴). Ten untreated wt mice (▪) were included as a control. Starting on the day of T cell infusion, mice received 90,000 IU IL-2 i.p. once daily for 4 days and were observed regularly for signs of tumor progression.

Animals treated with GKO effector T cells are immune to subsequent tumor challenge

To determine whether surviving mice developed protective immunity, animals that survived longer than 100 days were rechallenged with an s.c. injection of 2 × 10⁴ D5 tumor cells. As a control, 10 naive GKO mice were challenged with the same tumor dose. As expected, all the naive mice experienced progressive tumor growth and were sacrificed within 20 days (Fig. 5⇓). In contrast, none of the 10 GKO mice cured by adoptive immunotherapy developed a tumor. All rechallenged animals remained tumor free 150 days following rechallenge.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

Mice surviving long term after treatment with GKO T cells develop protective antitumor immunity. Ten GKO mice that were cured of systemic tumor by the adoptive transfer of GKO effector T cells and survived longer than 100 days were rechallenged with 2 × 10⁴ viable D5 tumor cells s.c. (▴). Ten naive mice were included as a control and were challenged with 2 × 10⁴ tumor cells (□). The mean tumor size and SD for the tumors that developed in naive mice are presented.

Neutralization of IFN-γ does not block therapeutic efficacy of wt and PKO effector T cells

Although the experimental results using GKO mice clearly indicated that tumor regression occurred independent of IFN-γ production, it is difficult to exclude any contribution of IFN-γ in a situation where other compensatory effector mechanisms might be operational. One could argue that perforin-mediated cytotoxicity may compensate for the loss of IFN-γ by GKO effector T cells and vice versa. To address this issue we neutralized IFN-γ activity with multiple administrations of Abs to IFN-γ using a dose and scheme that were shown to be effective in neutralizing IFN-γ in vivo (14). The Ab was given before and after adoptive transfer of wt and PKO effector T cells into tumor-bearing wt or PKO mice, respectively. The presence of neutralizing Abs to IFN-γ did not inhibit the efficacy of T cells from wt or PKO mice (Table IV⇓). The lack of inhibitory effect of the neutralization of IFN-γ further supports the nonessential role of IFN-γ in T cell-mediated tumor regression in this model.

T cells from PKO and wt donors cure GRKO mice with established pulmonary metastases

Given the limitations of neutralizing Ab to IFN-γ, we next examined the effect of IFN-γ by transferring wt or PKO effector cells into GRKO mice bearing 3-day D5 pulmonary metastases. Although the transferred T cells can produce IFN-γ, host macrophages, neutrophils, NK cells, and other elements of the host animal lack the corresponding receptor and are unable to respond to the cytokine. One limitation of this model is that the GRKO mice (129/SV) were only available on a different genetic background from C57BL/6 mice. This would cause a problem if the allogeneic differences were sufficient to prevent establishment of pulmonary metastases or if cellular elements of the host’s immune system needed to interact with the transferred T cells. However, the development of D5 pulmonary metastases in GRKO mice was indistinguishable from that of metastases in C57BL/6 mice (Table V⇓). Complete tumor regression was observed in these 129/SV tumor-bearing GRKO mice following adoptive transfer of effector T cells from D5-G6 vaccinated wt or PKO C57BL/6 mice (Table V⇓). These data further support the contention that IFN-γ is not critical for T cell-mediated tumor regression even in the absence of perforin-mediated cytotoxicity.

TVDLN cells from GKO mice are not polarized toward a T2 cytokine profile

To date we have shown that T cell-mediated tumor regression is independent of IFN-γ production by T cells or IFN-γ receptor signaling in host immune cells. This unexpected finding conflicts with our original hypothesis that IFN-γ is critical for T cell-mediated tumor regression. Using CD8 TCR transgenic mice, Dobrzanski et al. have shown that in vitro polarized Tc1 and Tc2 cells can mediate tumor regression in adoptive transfer experiments, albeit the Tc2 are substantially less effective (28). Therefore, we questioned whether TVDLN T cells generated from D5-G6-vaccinated GKO mice were polarized to Tc2-type T cells. First, we measured IFN-γ, IL-4, and IL-10 in the supernatant of the TVDLN cells after polyclonal activation with anti-CD3 for 48 h. Whereas wt TVDLN cells secreted high amounts of IFN-γ, as expected there was no secretion of IFN-γ from GKO TVDLN cells. Both wt and GKO TVDLN cells secreted similar amounts of IL-4 and IL-10 upon stimulation with anti-CD3 over 48 h with no significant difference (p > 0.05) compared with wt TVDLN (Fig. 6⇓). This suggests that the TVDLN from GKO mice were no more polarized toward a type 2 cytokine profile than wt T cells.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 6.

The inability to produce IFN-γ in GKO TVDLN does not skew the T cells to a type 2 cytokine profile. D5-G6 TVDLN from wt (□) or GKO (▪) mice were assessed for IFN-γ, IL-4, and IL-10 secretion following activation with anti-CD3. T cells (4 × 10⁶/ml) were cultured in the presence of anti-CD3 for 48 h. Cytokine concentration in the supernatant was determined by ELISA. Data are presented as the mean of three independent experiments ± SE. The limit of detection is 10 pg/ml for IFN-γ and IL-4, and 40 pg/ml for IL-10.

To determine the cytokine profile of the tumor-specific effector T cells generated from wt and GKO mice, they were restimulated with tumor cells or anti-CD3 for 24 h. The supernatant was collected, and the concentrations of IFN-γ, IL-4, and IL-10 were determined by ELISA. Effector T cells generated from wt or GKO mice stimulated with anti-CD3 secreted similar levels of IL-4 and IL-10 (Fig. 7⇓). As we have previously reported, effector cells from wt mice secreted substantial amounts of tumor-specific IFN-γ (p < 0.05). IFN-γ was not produced when T cells were cultured alone or in the presence of the unrelated tumor, MPR-4. Effector T cells generated from wt or GKO mice did not secrete IL-4, the prototypical type 2 cytokine, in response to stimulation with D5 or MPR-4 tumor cells. Similarly, no tumor-specific IL-10 secretion was observed for T cells generated from either wt or GKO mice (Fig. 7⇓). Thus, these results demonstrated that even in the absence of IFN-γ, TVDLN from GKO mice were not skewed toward a T2 profile, as they did not show any increased release of tumor-specific IL-4 or IL-10. In contrast to tumor-infiltrating lymphocytes, TVDLN T cells generated from D5-G6-vaccinated wt or GKO mice did not produce detectable tumor-specific TNF-α, another type 1 cytokine produced by T cells that potentially mediates tumor regression (data not shown). Therefore, although GKO effector T cells mediate tumor regression in an IFN-γ-independent fashion in our experiments, it is unlikely to be mediated by IL-4 as proposed by other investigators (28).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 7.

GKO effector T cells do not produce tumor-specific IL-4 or IL-10. Effector T cells generated from D5-G6 TVDLN of either wt or GKO mice were assessed for tumor-specific cytokine release at the time cells were adoptively transferred into tumor-bearing hosts. T cells (4 × 10⁶/ml) were cultured alone, with anti-CD3, with a syngeneic but unrelated tumor (2 × 10⁵/ml), MPR-4, or with D5 (2 × 10⁵/ml). Supernatants were harvested 18–24 h later for quantification of IFN-γ, IL-4, and IL-10. Cytokine release into the supernatant was determined by ELISA using standard kits. Data are presented as the mean of three independent experiments ± SE. The limit of detection is 10 pg/ml for IFN-γ and IL-4, and 40 pg/ml for IL-10.

Effector T cells from GKO, but not PKO, mice induced depigmentation

During the course of these experiments, we observed that 75% of the cured wt mice (9 of 12) and 40% (4 of 10) of the GKO mice that were rechallenged with D5, developed depigmentation (vitiligo) of the skin (Fig. 8⇓). The extent of vitiligo varied from mouse to mouse. This observation is in striking contrast to that seen in PKO mice that were cured of systemic disease by adoptive transfer of PKO T cells and were immune to subsequent tumor challenge. In our studies these PKO mice (0 of 14 mice) never developed vitiligo (data not shown). A similar observation was recently published by Browne et al., who showed that autoimmunity induced by active immunization of mice against the tyrosinase-related protein-2 was dependent on CD8⁺ cells and mediated by perforin (29).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 8.

GKO mice develop depigmentation after adoptive immunotherapy. The picture shows one GKO mouse (top) and one wt mouse (bottom) that developed vitiligo after treatment with GKO or wt TVDLN cells, respectively. Mice had been rechallenged with 2 × 10⁴ viable D5 tumor cells. One naive wt mouse is shown in the middle for comparison.

Immunohistochemistry of tumor-bearing lungs after adoptive immunotherapy

Previously we have shown that regression of pulmonary metastases after adoptive transfer of wt effector T cells is independent of host T cells and NK cells (30); however, contributions from host macrophages and granulocytes has not been ruled out. In fact, Hung et al. demonstrated that CD4⁺ T cell-dependent immunity elicited by vaccination with a GM-CSF gene-modified B16 tumor is dependent on both macrophages and granulocytes (1). To examine whether macrophages and granulocytes infiltrate the lungs after adoptive transfer in this model, we stained frozen sections of lungs with Abs to CD4, CD8, NK1.1, Mac-1, and Gr-1. Lungs were obtained 24 h after treatment with IL-2 alone or IL-2 plus 35 × 10⁶ effector cells from either wt or GKO mice. Lungs treated with low dose IL-2 alone were essentially negative for signs of infiltrating cells (left panel). However, there were marked increases in the numbers of CD4-, CD8-, Mac-1-, and Gr-1-positive cells in the tumor-bearing lungs of mice that received either wt (middle panel) or GKO (right panel) effector T cells (Fig. 9⇓). No staining with anti-NK1.1 Ab was found in any lung section, whereas the Ab did stain a positive control tissue on the same slide (data not shown). These results document that both wt and GKO effector T cells were able to attract macrophages and granulocytes into the tumor-bearing lung and provide a possible explanation for the cells involved in tumor destruction.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 9.

Immunohistochemistry of tumor-bearing lungs after immunotherapy. Twenty-four hours after immunotherapy with either IL-2 alone (IL-2 only, left panel) or IL-2 and effector T cells from wt mice (wt, middle panel) or GKO mice (GKO, right panel), lungs were recovered and frozen, and sections were cut and stained with control IgG, anti-CD4, anti-CD8, anti-NK1.1, anti-MAC-1, or anti-Gr-1 Abs.

Note added in proof.

Our data concur with a recent report by Peng et al. (40) that anti-CD3-activated T cells from TVDLN of GKO mice could mediate regression of experimental pulmonary, intracranial, or s.c. MCA-205 or peritoneal EL-4 tumors in GKO recipient animals. This extensive analysis of T cell-mediated regression at multiple anatomic sites further supports the contention that although IFN-γ may play a critical role in tumor surveillance (41), it is not essential for T cell-mediated tumor regression following adoptive immunotherapy when other immune effector mechanisms are intact.