Type 1 and Type 2 CD8⁺ Effector T Cell Subpopulations Promote Long-Term Tumor Immunity and Protection to Progressively Growing Tumor¹

Adoptive T cell immunotherapy has been shown to be a viable modality for the treatment of certain human cancers. Therapeutic efficacy by adoptively transferred tumor Ag-reactive T cell populations is dependent, in part, on the ability of donor cells to persist long-term and retain antitumor function in vivo. Although persistence of long-term antitumor memory responses by adoptively transferred tumor Ag-specific T cell populations can efficiently induce tumor regression and prevent potential tumor relapse in cancer patients, factors contributing to the generation and maintenance of tumor Ag-specific CD8⁺ memory T cells remains poorly defined.

Immunological memory can be defined as an immune response that is faster and more potent than that of the primary response following re-exposure to the same Ag (1). In general, induction of primary antitumor responses by naive CD8⁺ T cells requires TCR recognition of relavent tumor-associated Ag in the context of MHC class I, either expressed endogenously by tumor cells or exogenously as peptide presented by APCs capable of generating effective costimulatory signals (2). Such interaction results in the clonal expansion and differentiation of naive CD8⁺ T cells into activated effector cells that mediate tumor cell killing. It is clear that Ag recognition alone is not sufficient for optimal T cell activation and that secondary costimulatory signals are necessary to aviod tolerance or anergy (3). Most of these activated CD8⁺ effector T cells eventually die via activation-induced cell death with some proportion becoming tumor-reactive memory T cells via an undefined mechanism (1, 4, 5). It has been suggested that generation of memory cells with tumor-specific activity can be viewed as a process whereby responding T cells survive rather than die, presumably through up-regulation or down-regulation of survival (anti-apoptotic) or death (apoptotic-inducing) signals, respectively (4, 6). Cytokines such as IL-2, IL-4, IL-15, and IFN-γ have been shown to enhance survival and, in some cases, promote clonal expansion of activated CD8⁺ T cell subpopulations through the regulation of such signals (4, 6, 7, 8, 9). Although these cytokines may be sufficient for T cell activation, survival, and perhaps memory cell development in vivo, other factors such as the strength of T effector cell-to-tumor target cell interaction, cell surface costimulatory molecules, and the quality, duration, and concentration of recognizable Ag at the site of tumor encounter may be of equal importance (2, 10, 11).

We and others have shown that cytolytic CD8⁺ effector T cells fall into two subpopulations based on their cytokine-secreting profiles following tumor Ag encounter (12, 13, 14, 15). Type 1 CD8⁺ T cells (Tc1)³ secrete IFN-γ and TNF-α, whereas type 2 CD8⁺ T cells (Tc2) characteristically secrete IL-4, IL-5, IL-10, and IL-13. Aside from their direct tumorcidal-inducing potentials, Tc1 and Tc2 effector cell-derived cytokines may further promote and enhance therapeutic efficacy through either autocrine and/or paracrine mechanisms that result in effector cell survival and/or memory CD8⁺ T cell development. Previously, we have shown that adoptive transfer of either Tc1 or Tc2 CD8⁺ effector T cell subpopulations can effectively induce tumor cell regression and subsequently prolong survival times in mice bearing established pulmonary malignancy (12). In the current study, we assessed the therapeutic effects and generation of long-term tumor-Ag specific T cell memory responses by tumor-reactive Tc1 and Tc2 effector cell subpopulations in tumor-bearing mice. Using a previously described OVA-expressing B16 melanoma (B16-OVA) lung metastasis model (12), we show that adoptively transferred Tc1 and Tc2 effector cell subpopulations can induce suppression of established B16 melanoma lung metastases and subsequently establish long-term tumor Ag-specific immunity in a high proportion of mice with established pulmonary tumor. Collectively, effector cell-treated mice exhibiting long-term tumor immunity showed 1) resistance to lethal tumor rechallenge with B16-OVA, 2) heightened T cell-mediated tumor Ag-specific cytolytic responses ex vivo, 3) an elevation in the total numbers of both CD8⁺ and CD4⁺ T cell populations in the lung, and 4) detectable levels of long-lived donor CD8⁺ TCR Vα2 transgene-positive cells, accompanied by an elevation in the total numbers of CD8⁺/CD44^high activated and/or memory T cell subpopulations in both spleen and lungs of treated mice. Moreover, memory responses induced by Tc2 and Tc1 effector cell therapy were dependent, in part, on both tumor burden and effector cell-derived IL-4, IL-5, and IFN-γ, respectively. We discuss the role of Tc1 and Tc2 CD8⁺ effector T cell subpopulations in successful tumor immunity and generation of long-lasting tumor Ag-specific memory responses in mice with established disseminated malignancy.

Female C57BL/6 mice, 6–10 wk of age, were obtained from the Animal Breeding Facility at the Trudeau Institute. The OT-I mouse strain, on a C57BL/6 background (H-2^b), was originally obtained from Dr. Michael Bevan (University of Washington, Seattle, WA). These mice express a transgenic TCR Vα2 specific for the SIINFEKL peptide of OVA in the context of MHC class I, H2-K^b (16). Homozygous IL-4^−/− (OT-I.IL-4), IL-5^−/− (OT-I.IL-5), and IFN-γ^−/− (OT-I.IFN-γ) knockout mice expressing the TCR Vα2 transgene were generated by backcrossing OT-I mice onto designated cytokine gene knockout mice (H-2^b). Animals were maintained and treated according to animal care committee guidelines of the National Institutes of Health (Bethesda, MD) and the Trudeau Institute.

The weakly immunogenic B16-OVA and parental B16 tumor cell lines that are syngeneic to the C57BL/6 background were kindly provided by Drs. Edith Lord and John Frelinger (Rochester, NY). EL4 and the derivative OVA-expressing EG.7-OVA cell lines were obtained from the American Type Culture Collection (Manassas, VA).

Spleens were collected from mice and single-cell suspensions were prepared, washed twice in HBSS, and resuspended in RPMI 1640 (Life Technologies, Grand Island, NY) supplemented with 2 mM pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, 10 mM HEPES, and 10% heat-inactivated FCS (Life Technologies). CD8-enriched T cells were obtained by treating with anti-CD4 (RL172.4), anti-heat-stable Ag (J11D), and anti-MHC class II (D3.137, M5114, and CA4) mAbs for 30 min at 4°C. Cells were washed and incubated with rabbit (Pel-Freeze, Rogers, AR) and guinea pig (Harlan, Indianapolis, IN) complement for 30 min at 37°C (12). For preparation of single-cell suspensions from lung parenchyma, lungs were flushed in situ with HBSS via cannulation of the heart to remove residual intravascular blood pools. Minced lung tissues were incubated for 1 h at 37°C on a rocker platform in 1.5 ml/lung of RPMI 1640 supplemented with DNase I (50 U/ml; Sigma, St. Louis, MO); collagenase I, type 4197 (250 U/ml; Sigma); and 5% FCS. After incubation, digested lung tissues were mechanically dispersed through stainless-steel mesh screens in RPMI 1640-5% FCS. After three washes in RPMI 1640-5% FCS, lymphoid cells were resuspended in RPMI-10% FCS to attain a cell concentration of 1 × 10⁷ viable cells/ml. Cytospin preparations of cells from lung homogenates were fixed with methanol and stained with eosin and methylene blue (Fisher, Pittsburgh, PA). Cell differential counts were performed on a total of 200–300 cells on coded slides.

To obtain effector cells to OVA peptide, single-cell suspensions from spleen and lymph nodes of designated OVA Ag-specific TCR Vα2 transgene-positive mice were washed twice in HBSS and resuspended in RPMI 1640-10% FCS. CD8-enriched T cells were obtained by passing lymphoid cell suspensions through nylon wool columns and treating with anti-CD4 (RL172.4), anti-heat-stable Ag (J11D), anti-MHC class II (D3.137, M5114, and CA4) mAbs, and complement as previously described (12). Small resting CD8 T cells were harvested from Percoll Gradients (Sigma) and resuspended to appropriate cell concentrations in culture media. Naive CD8 cells were typically 90% pure as demonstrated by immunofluorescent Ab staining. APCs were enriched from spleens of normal C57BL/B6 mice by anti-Thy1.2 (HO13.14 and F7D5), anti-CD4 (RL172.4), and anti-CD8 (3.155) mAbs and complement. T cell-depleted APCs were pulsed with OVA peptide (10 μM) for 30 min at 37°C and treated with mitomycin-C (50 μg/ml; Sigma) for an additional 30 min at 37°C. For Tc1 effector cell generation, naive CD8 T cells from OT-I trangenic mice (2 × 10⁵ cells/ml) were stimulated with mitomycin C-treated, OVA peptide-pulsed APCs (6 × 10⁵ cells/ml) in the presence of IL-2 (20 U/ml; X63.IL-2 supernatants (17)), IL-12 (2 ng/ml; kindly provided by Dr. Stanley Wolf, Genetics Institute, Cambridge, MA), and anti-IL-4 mAb (200 U/ml; X63.Ag.IL-4 supernatants). Alternatively, for Tc2 effector cell generation, naive CD8 T cells from OT-I trangenic mice (2 × 10⁵ cells/ml) were stimulated with mitomycin C-treated, OVA peptide-pulsed APCs (6 × 10⁵ cells/ml) in the presence of IL-2 (20 U/ml), IL-4 (200 U/ml; X63.IL-4 supernatants (17)), and anti-IFN-γ mAb (20 μg/ml; XMG1.2). Effector cell cultures were incubated for 4 days with additional IL-2 (20 U/ml) added to the cultures on day 2 to promote CD8 cell expansion of Tc1 or Tc2 populations.

Syngeneic B6 mice were injected i.v. with either 5 × 10⁵ or 1 × 10⁵ B16-OVA cells to establish pulmonary metastases. Seven days after tumor challenge, mice were treated i.v. with 2 × 10⁶ Tc1 or Tc2 OVA-specific effector T cells, and survival times were monitored daily (12). Control groups of mice received no treatment. At weekly intervals after therapy, designated mice were sacrificed for enumeration of pulmonary metastatic nodules. Metastases on freshly isolated lungs appeared as discrete black-pigmented foci that were easily distinguishable from normal lung tissue. Alternatively, mice surviving primary tumor challenge were rechallenged with similar cell numbers of either B16-OVA or parental B16 tumor cells, and survival times were monitored as previously described.

Cytolytic T cell activity was determined by a standard ⁵¹Cr-release assay as described previously (12). Briefly, syngeneic EG.7-OVA or EL4 target cells were radiolabeled with 200 μCi Na₂⁵¹CrO₄ (ICN Radiochemicals, Irvine, CA) for 1 h at 37°C, washed, and resuspended in RPMI 1640-10% FCS. CD8⁺ T cells were combined with tumor target cells (1 × 10⁴ cells/well) at various E:T cell ratios in 96-well U-bottom plates (Costar, Cambridge, MA) and incubated for 4 h at 37°C with 5% CO₂. Culture supernatants were harvested and counted in a Wizard automatic gamma-counter (Wallac, Gaithersburg, MD). Spontaneous release of ⁵¹Cr was determined by incubation of targets in the absence of effectors, whereas maximum release of ⁵¹Cr was determined by incubation of targets in 1% Triton X-100. Results are expressed as the percent specific release and were calculated as follows: % specific release = [(experimental − spontaneous)/(maximum − spontaneous)] × 100. Results are also expressed as LU/10⁶ effector cells. One lytic unit was defined as the number of effector cells required to cause 30% lysis of 10⁴ target cells.

Detection of secreted cytokines in supernatants of T cell cultures after restimulation has been described previously (12, 13). Briefly, freshly generated effector cells (2 × 10⁵/ml) were harvested, washed, and restimulated with either mitomycin C-treated EG.7-OVA or EL4 tumor cells (6 × 10⁵/ml) for 24 h in 1-ml volumes. Culture supernatants were harvested and assessed for cytokine content by cytokine-specific ELISA. Murine IL-5 and IL-4 were measured with anti-IL-5 (TRFK5) and anti-IL-4 (TRFK4) mAbs, respectively. IFN-γ was detected by anti-IFN-γ mAbs R46A2 and XMG1.2. Standard curves were constructed with purified IL-4 (X63.IL-4 supernatants), IL-5 (X63.IL-5 supernatants), and IFN-γ (X63.IFN-γ supernatants). Values for T cells or stimulator cells cultured in media alone were negligible.

Single-cell suspensions of either spleen or processed murine lung were washed three times in a fluorescent Ab buffer consisting of 1% BSA and 0.02% sodium azide in 0.01 M PBS (pH 7.2). CD8 lymphocytes expressing the TCR Vα2 transgene were phenotyped by their expression of surface markers using direct immunofluorescence staining techniques. Lymphocytes (10⁶) were mixed with 100 μl of fluorescent Ab buffer containing 1 μg of both Cy-chrome-conjugated anti-CD8 (PharMingen, San Diego, CA) and fluorescein-conjugated anti-Vα2 (PharMingen; clone B20.1), PE-conjugated CD44 (PharMingen; clone IM7), or CD4 (PharMingen; clone GK1.5) mAbs and incubated for 20 min on ice. Stained cell preparations were than washed three times in fluorescent Ab buffer and analyzed by multiparameter flow cytometry using a Becton Dickinson FACScan (San Jose, CA). Ten thousand cells were analyzed per sample with dead cells excluded on the basis of forward light scatter. Surface-marker analysis was performed using CellQuest Software (Becton Dickinson), and the percentages and absolute cell numbers of positive cells were determined.

For statistical analysis, the two-tailed Student t test or nonparametric Mann-Whitney rank sum test was used.

CD8⁺ Tc1 and Tc2 effector T cells were generated in vitro from OVA-specific TCR transgenic OT-I mice as described in Material and Methods. As shown in earlier studies (12), both Tc1 and Tc2 effector cells demonstrated tumor Ag-specific cytolytic activity to OVA Ag-expressing tumor cell targets (EG.7-OVA) with the latter being nearly 9-fold greater than that of the former (100 LU/10⁶ effector cells vs 11 LU/10⁶ effector cells). Furthermore, Tc1 effector cell populations produced substantial amounts of IFN-γ with no detectable levels of IL-4 or IL-5 when restimulated with OVA-expressing EG.7 tumor cells. In contrast, Tc2 populations released considerable amounts of IL-5 and IL-4 with low, yet detectable, levels of IFN-γ upon restimulation with OVA-expressing tumor cells. Flow cytometric analysis showed that both Tc1 and Tc2 effector cell populations expressed similar patterns of cell-surface Ag markers that are characteristic of effector cell phenotype (12). Both effector cell populations were TCR Vα2⁺, CD8⁺CD4⁻ and expressed up-regulated levels of both CD44 and CD25 and down-regulated levels of CD62L.

Because Tc1 and Tc2 effector cell populations showed highly tumor Ag-specific cytolytic and cytokine-releasing activities in vitro, the therapeutic efficacy of Tc1 and Tc2 effector cell populations were assessed in an experimental B16-OVA lung metastases model. Normal syngeneic C57BL/6 mice (n = 6/group) were injected i.v. with 5 × 10⁵ B16-OVA tumor cells. Seven days later, when disseminated micrometastases were established, 2 × 10⁶ OVA Ag-specific Tc1 or Tc2 effector cell populations were systemically transferred into tumor-bearing mice, and survival times were monitored daily. As shown in Fig. 1, pulmonary tumors grew progressively in untreated mice, whereas tumor growth among corresponding animals treated with either Tc1 or Tc2 effector cell therapy showed a marked reduction in the number of lung-associated tumor colonies for up to 30 days after tumor challenge. Concomitantly, survival times of tumor-bearing mice receiving Tc1 or Tc2 effector cell therapy were substantially prolonged compared with those of untreated animals (Fig. 2,A). Delays in the onset of mortality among both groups of Tc1 and Tc2 effector cell-treated mice ranged between 35 and 45 days after tumor challenge with a mean survival time of 46.2 ± 1.5 and 44.8 ± 2.1 days, respectively. In contrast, the onset of mortality among untreated tumor-bearing control mice occurred as early as 25 days with a mean survival time of 27.1 ± 0.9 days after tumor challenge. Furthermore, treatment of tumor-bearing mice with similar numbers of naive pre-effector CD8⁺ T cells did not substantially influence survival times compared with those of corresponding untreated tumor-bearing control mice (Fig. 2,A). Next, we assessed the immunological specificity of OVA Ag-specific Tc1 and Tc2 effector cell populations. As shown in Fig. 2 B, transfer of cell numbers as high as 5 × 10⁶ of either Tc1 or Tc2 effector cells into mice challenged with the non-OVA Ag-expressing B16 parent line showed no detectable therapeutic effect in survival times compared with those of control groups of untreated B16 tumor-bearing mice.

Because large tumor-cell burdens may qualitatively and/or quantitatively interfere with effective immunotherapy (2), we investigated the ability of Tc1 and Tc2 effector cells to induce long-term protective immunity in mice with lower levels of established pulmonary metastases. Pulmonary tumors were induced by i.v. injection of 1 × 10⁵ B16-OVA cells into syngeneic B6 mice. Seven days later, when disseminated micrometastases were established, 2 × 10⁶ systemic OVA Ag-specific Tc1 or Tc2 effector cells were injected i.v., and survival times were monitored as previously described. As shown in Fig. 3, treatment with either single-dose Tc1 or Tc2 effector cell therapy prolonged survival times among 75% (6 of 8) and 87% (7 of 8) of tumor-bearing mice, respectively. In contrast, the mean survival time of untreated tumor-bearing control mice was 45 ± 5.1 days after tumor challenge. Sixty-five days after initial tumor cell challenge, surviving mice were rechallenged with 1 × 10⁵ B16-OVA tumor cells. All mice, initially receiving Tc1 or Tc2 effector cell therapy, exhibited protection and long-term survival for >120 days after tumor cell rechallenge (Fig. 3). Corresponding groups of untreated control mice receiving B16-OVA tumor cells died within 40 days after tumor challenge, suggesting that tumor cell preparations used for rechallenge experiments were lethal in naive mice (data not shown). Furthermore, surviving mice challenged with the non-OVA-expressing B16 parent line showed no evidence of effective tumor rejection (data not shown). This indicated that protective immunity to lethal tumor-cell challenge was established in vivo among tumor-bearing mice treated with adoptively transferred Tc1 and Tc2 effector-cell populations and that antitumor responses were highly Ag specific. Similar results were obtained in two other independent experiments.

Local and systemic T cell populations from lungs and spleens of mice receiving effector cell therapy nearly 180 days earlier were enumerated by multicolor flow cytometric analysis. As shown in Table I, the absolute cell numbers and percentages of systemic CD8⁺ T cells from mice (n = 4/group) receiving either Tc1 or Tc2 effector-cell therapy were similar but significantly enhanced (p < 0.01) compared with those of corresponding cell populations in normal mice of similar age. In spleens of mice receiving either Tc1 or Tc2 effector-cell therapy, the absolute cell numbers of total CD8⁺ T cells were 18.6 ± 1.6 × 10⁶ (16.1%) and 23.3 ± 2.7 × 10⁶ (16.3%), respectively, whereas the absolute cell number of corresponding CD8⁺ T cells among normal age-related control mice was 11.6 ± 3.8 × 10⁶ (11.7%). In contrast, CD4⁺ T cell numbers from these same effector cell-treated animals showed no significant difference when compared with those of normal control mice (18.6 ± 1.6 and 30.1 ± 2.6 × 10⁶ vs 23.8 ± 3.8 × 10⁶; p < 0.30). Moreover, the CD4/CD8 T cell ratios in spleens of mice receiving either Tc1 (1.00) or Tc2 (1.29) effector-cell therapy were comparatively lower than normal control (2.05) mice in that effector cell-treated mice had a greater proportion of systemic CD8⁺ T cells after therapy (Table I).

In the lung, both CD4⁺ and CD8⁺ T cell numbers were significantly elevated in mice receiving therapy when compared with those of normal control mice. As shown in Table II, the absolute cell number of CD4⁺ cells among mice receiving Tc1 effector-cell therapy was 9.9 ± 1.6 × 10⁵ (4.0%), whereas that of CD8⁺ cells was 11.8 ± 0.7 × 10⁵ (6.9%). In mice receiving Tc2 effector-cell therapy, the corresponding absolute cell numbers were comparatively greater at 26.6 ± 7.9 × 10⁵ (3.7%) and 43.8 ± 15.3 × 10⁵ (6.4%), respectively. In contrast, the absolute cell numbers of both CD4⁺ and CD8⁺ T cell populations among normal control mice were significantly lower (p < 0.05) at 3.6 ± 0.4 × 10⁵ (3.0%) and 3.5 ± 0.5 × 10⁵ (3.1%), respectively. Consequently, CD4/CD8 ratios among groups of Tc1 and Tc2 effector cell-treated mice were comparatively lower than those of normal control mice (0.84 and 0.61 vs 1.05) in that mice receiving effector-cell therapy manifested greater proportions of CD8⁺ T cells at the site of tumor challenge after successful tumor immunity.

Because up-regulated CD44 expression is indicative of both recently activated and long-lived T cells thought to be responsible for T cell memory (1), we assessed the numbers of CD8⁺ T cell subpopulations coexpressing elevated levels of CD44 surface Ag both at distal (spleen) and at local (lung) sites of tumor growth. As shown in Fig. 4,A, the absolute cell numbers and percentages of systemic CD8⁺/CD44^high cells in spleens of mice receiving either Tc1 (11.1 ± 0.5 × 10⁶ (65.2%)) or Tc2 (15.7 ± 2.6 × 10⁶ (66.8%)) effector-cell therapy nearly 180 days earlier were significantly higher (p < 0.05) than those of normal age-related control mice (6.0 ± 1.5 × 10⁶ (40.7%)). Similarly, the absolute cell numbers of corresponding cell subpopulations in lungs of these same mice were nearly 4- to 6-fold greater than those of normal control mice (Fig. 4 B). CD8⁺/CD44^high cell numbers for Tc1 and Tc2 effector cell-treated mice were 7.9 ± 0.4 × 10⁵ (74.4%) and 9.7 ± 2.3 × 10⁵ (78.6%), respectively, whereas corresponding tissue cell numbers from lungs from normal control mice of similar age were significantly lower (p < 0.05) at 1.7 ± 0.1 × 10⁵ (43.8%).

OVA Ag-specific CD8⁺ T effector cells from OT-I mice that express the TCR transgene Vα2 were generated in vitro and transferred into tumor-bearing mice as previously described. To assess the local and systemic presence and distribution of such effector cells after tumor rechallenge, multicolor flow cytometry was performed on lungs and spleens of effector cell-treated mice. Although equal numbers of Tc1 and Tc2 effector cells were transferred into mice nearly 180 days earlier, the percentages and frequencies of TCR Vα2 transgene-positive cells in both spleens and lungs remained markedly greater than those of normal age-related control mice. As shown in Fig. 5, A–C, the percentages of systemic CD8⁺ T cells expressing the TCR Vα2 transgene were nearly 5- to 10-fold greater in spleens of mice receiving either Tc1 or Tc2 effector-cell treatment compared with those of normal age-related control mice. Similarly, the frequency and percentage of CD8⁺ TCR Vα2 transgene-positive cells in lungs of these same animals were comparatively greater than those of normal control mice (Fig. 5, D–F), suggesting that both donor Tc1 and Tc2 effector cell populations are detectable in the spleen and lungs for extended periods of time after successful long-term tumor immunity to B16-OVA tumor rechallenge.

To assess tumor-specific T cell functional responses in mice exhibiting long-term protection, cytolytic activity of systemic CD8⁺ T cell populations of effector cell-treated mice after tumor rechallenge were assessed ex vivo in a standard 5-h chromium-release assay. As shown in Table III, enriched CD8⁺ T cells from spleens of surviving mice that received either Tc1 or Tc2 effector-cell therapy over 180 days earlier exhibited OVA Ag-specific CTL responses to EG.7-OVA tumor-cell targets. Tumor Ag-specific lytic activity among mice treated with either Tc1 or Tc2 effector-cell therapy ranged from 12 to 45% over different E:T ratios. In contrast, lytic activity to OVA-non-expressing control EL4 target cells was negligible, confirming that killing was Ag specific and most likely not associated to NK cell activity (Table III). Normal age-matched control mice showed no detectable lytic activity to either target-cell population (data not shown). These results suggest the presence of systemic long-lived antitumor CTL memory responses in tumor-bearing mice receiving adoptively transferred Tc1 or Tc2 effector-cell therapy over 180 days earlier.

Because Tc2 effector-cell populations were found to secrete low, yet detectable, levels of IFN-γ upon restimulation with tumor Ag, we assessed the role of effector cell-derived IFN-γ in both Tc1 and Tc2 effector cell-mediated antitumor responses. Pulmonary tumors were induced by i.v. injection of 1 × 10⁵ B16-OVA cells into syngeneic C57BL/6 mice. One week later, 2 × 10⁶ OVA Ag-specific Tc1 or Tc2 effector-cell subpopulations, generated from OT-I.IFN-γ knockout mice, were systemically transferred, and survival times were monitored as previously described. As shown in Fig. 6,A, treatment with either IFN-γ-deficient or wild-type Tc1 effector cells equally prolonged survival times among 4 of 6 (67%) tumor-bearing mice. However, after B16-OVA tumor cell rechallenge on day 60 after therapy, the former was comparatively less effective than that of the latter. All surviving mice, initially receiving wild-type Tc1 effector-cell therapy exhibited tumor protection and long-term survival for >75 days after tumor rechallenge. In contrast, tumor protection among all surviving mice receiving IFN-γ-deficient Tc1 effector-cell therapy was completely eliminated within 45 days after tumor rechallenge. This suggests that effector cell-derived IFN-γ may play a substantial role in establishing effective long-term Tc1 effector cell-mediated tumor immunity. In contrast, heightened tumor-recipient survival times among groups of mice treated with IFN-γ-deficient Tc2 effector-cell populations were not comparatively different to those of corresponding groups of tumor-bearing mice receiving wild-type Tc2 effector-cell therapy, suggesting that Tc2 effector cell-derived IFN-γ does not play a significant role in the development of Tc2 effector cell-mediated long-term tumor immunity (Fig. 6 B). Corresponding groups of untreated control mice receiving similar doses of B16-OVA tumor cells died within 40 days of tumor challenge, suggesting that tumor-cell preparations used for rechallenge experiments were lethal in naive mice (data not shown). Results are representative of two similar experiments.

To determine whether the type 2 cytokines IL-4 and IL-5 were required for the induction of effective Tc2 effector cell-mediated long-term protection, OVA-Ag specific Tc2 effector-cell subpopulations were generated from either OT-I.IL-4 or OT-I.IL-5 knockout mice and transferred into syngeneic C57BL/6 mice bearing 7-day established B16-OVA tumors. As shown in Fig. 7, survival times among tumor-bearing mice receiving either IL-5- or IL-4-deficient Tc2 effector cells were significantly prolonged (p < 0.001) compared with those of untreated control mice. The mean survival times of mice receiving either IL-5- or IL-4-deficient Tc2 effector cells were 45.0 ± 4.5 and 53.2 ± 2.2 days, respectively. In contrast, the mean survival time of untreated tumor-bearing control mice was 31.8 ± 0.5 days after tumor challenge. However, both cytokine gene-deficient effector-cell populations were less effective in prolonging survival times than those of corresponding groups of tumor-bearing mice receiving similar doses of wild-type Tc2 effector-cell therapy. In the latter, effective protection and long-term survival were exhibited in 6 of 8 tumor-bearing mice (75%) for >100 days after tumor challenge (Fig. 7). In contrast, Tc1 effector-cell populations derived from either IL-4 or IL-5 knockout mice were comparable to those of corresponding wild-type Tc1 effector-cell populations (data not shown). This suggests that Tc2 effector cell-derived IL-4 and IL-5 may play a significant and unique role in Tc2 effector cell-mediated long-term tumor immunity and protection.

In the present study we analyzed the therapeutic effects of adoptively transferred OVA Ag-specific Tc1 and Tc2 effector-cell subpopulations in mice bearing established B16-OVA pulmonary malignancy. We show that single-dose immunoadoptive therapy with either effector-cell population induces considerable suppression of pulmonary metastases that subsequently prolong survival times among tumor-bearing mice. This therapeutic effect appeared dependent, in part, on the level of tumor burden. Effector cell-treated mice exhibiting high (5 × 10⁵) tumor burdens experienced significant delays in mortality when compared with those of untreated tumor-bearing control mice, whereas high proportions of mice with low (1 × 10⁵) tumor burdens survived indefinitely. More importantly, the latter were resistant to subsequent tumor rechallenge, indicating induction of long-term tumor Ag-specific immunological memory. Because recurrence of metastatic disease is a major cause of death among cancer patients, this observation suggests that adoptive immunotherapy with either Tc1 or Tc2 effector-cell populations could provide and induce long-term protection to cancer patients with minimal residual disease.

Long-term tumor immunity can be defined as a subset of persistent Ag-specific memory T cells that are capable of mediating and coordinating a faster, stronger, and more prolonged response to tumor Ag re-encounter. In the current study, induction of persistent long-term tumor immunity by adoptively transferred tumor-reactive Tc1 and Tc2 effector cells in tumor-bearing mice was supported by 1) successful resistance to lethal tumor rechallenge, 2) heightened levels of systemic OVA-Ag-specific CTL responses ex vivo, 3) elevation in the total numbers of both CD4 and CD8 T cell populations at the site of tumor growth, 4) detection of long-lived transgene-positive donor cells at sites both proximal (lung) and distal (spleen) to tumor growth, and 5) an elevation in the total numbers of activated/memory CD8⁺ T cell subpopulations coexpressing elevated levels of CD44 surface Ag in both spleen and lungs of treated mice. These results suggest that adoptively transferred tumor-reactive Tc1 or Tc2 effector cells not only persisted as memory and/or activated T cells, but also initiated and potentially modulated host antitumor immune responses in recipients with established pulmonary metastases. Transfer of either polarized Tc1 or Tc2 tumor-reactive CD8 effector-cell populations may effectively regulate established tumor through either direct lytic interaction or by effector cell-derived cytokines that may initiate a cascade of events in the host that result in the recruitment and activation of host immune cells. Subsequently, tumor-associated Ags may be released and re-expressed by host APCs that may aid, in part, with the generation and maintenance of both long-lived donor and host-derived memory T cells. Currently, the mechanisms by which CD8 lymphocytes differentiate into memory cells are unclear. However, it has been suggested that memory T cells can arise directly from progeny of fully differentiated effector T cells or through naive T cells that differentiate along separate developmental stages or pathways (11, 18, 19). Although we are not in a position to preferentially support either view, our data indicate that adoptive transfer of either Tc1 or Tc2 effector-cell subpopulations have the capacity to be long lived and subsequently promote functionally effective T cell memory responses in tumor-bearing mice that are derived, in part, from both host and donor T lymphocytes in vivo. Investigations to further elucidate the role of adoptively transferred Tc1 and Tc2 CD8 effector T cells in the induction and generation of host antitumor immune responses during progressive tumor growth are currently underway.

The mechanisms involved in the development and maintenance of immunologic memory remain unknown. However, it has been suggested that successful generation of long-lived tumor-reactive memory T cells is highly dependent on several factors including the cytokine environment at the site of tumor challenge (2). Therefore, we assessed the roles of Tc1 and Tc2 effector cell-derived cytokines in mediating tumor regression and establishing long-term tumor immunity. Because Tc2 effector-cell populations were found to secrete low, yet detectable, levels of IFN-γ, we assessed the role of effector cell-derived IFN-γ in both Tc1 and Tc2 effector cell-mediated long-term antitumor responses. The adoptive transfer of CD8 effector cells from IFN-γ-deficient OT-I mice showed that these Tc1 effector cells were therapeutically less effective over time than those of corresponding cell populations from wild-type mice. Although survival times in mice treated with the former were substantially prolonged, antitumor effects were short-lived with all mice experiencing mortality upon tumor rechallenge. The data suggest that secretion of IFN-γ by Tc1 effector cells is a necessary component in the mechanism of tumor eradication and long-term antitumor protection by these cells. The local secretion of IFN-γ by transferred effector cells at the site of tumor growth may lead to tumor regression by several mechanisms. It has been shown that IFN-γ can directly inhibit tumor cell growth (9, 20, 21), enhance recruitment and activation of cells that promote innate antitumor immune responses (9, 22, 23, 24), and promote elimination of transformed cells either directly or indirectly through nonimmune mechanisms such as those involving inhibition of angiogenesis by mechanisms that are incompletely understood but involve the IFN-γ-induced angiogenesis-inhibitory chemokines IP-10 and Mig (25, 26). Moreover, IFN-γ has been shown to affect tumor immunogenicity by enhancing tumor Ag presentation that may subsequently promote immune cell survival and maintain long-term antitumor immunity (27, 28, 29). Despite the potential role of effector cell-derived IFN-γ in Tc1 effector-cell therapy, IFN-γ-deficient Tc2 CD8 effector-cell populations were not comparatively different than those of corresponding groups of tumor-bearing mice receiving wild-type Tc2 effector-cell therapy. These data suggested that Tc2 effector cells must act by an effector cell-derived IFN-γ-independent mechanism.

Because others have reported that either IL-4 or IL-5, when secreted by cytokine gene-modified tumor cells, can effectively suppress tumor growth in vivo (30, 31, 32), we assessed the therapeutic roles of such type 2-related cytokines in inducing and maintaining effective Tc2 effector cell-mediated long-term protection in tumor-bearing mice. Therapy with either IL-4 or IL-5 cytokine-deficient effector-cell populations was markedly less effective than corresponding wild-type Tc2 effector-cell populations in prolonging survival times, with all tumor-bearing mice experiencing mortality within 70 days of primary tumor challenge. This suggested that Tc2 effector cell-derived IL-4 and IL-5 may play a substantial role in Tc2 effector cell-mediated tumor immunity and protection. Local release of IL-4 and IL-5 have been shown to mediate the selective recruitment, localization, and activation of nonspecific antitumor responses, such as macrophages, NK cells, and granulocytes that may facilitate tumor growth inhibition and/or enhance tumor Ag presentation (24, 33, 34, 35). Alternatively, IL-4 and IL-5 may indirectly modulate local antitumor responses by enhancing select expression of various chemokines, such as macrophage inflammatory protein-1α and monocyte chemotactic protein-1, that may affect both donor and host effector-cell trafficking and activation at the site of tumor encounter (36). In either instance, differential utilization of discrete cytokine profiles by Tc1 or Tc2 effector-cell populations appear to induce tumor rejection with different mechanisms that can potentially affect the nature and outcome of effective antitumor responses to progressively growing tumors. Although our observations suggest the existence of a correlation between effector cell-derived cytokines and heightened in vivo therapeutic efficacy and survival among mice with established tumor, other mechanisms such as direct cytolysis by either perforin-mediated or Fas-mediated pathways must be considered. Investigations to elucidate cognate effector cell-mediated lytic mechanisms and their contributions to tumor eradication are in progress.

Survival studies in our pulmonary tumor model showed a marked decrease in mortality among mice with low vs high levels of established tumor burden, suggesting that effective single-dose Tc1 and Tc2 effector-cell therapy is most effective in mice with low-level residual malignancy. Lower effector-cell efficacy among mice previously receiving a higher B16-OVA tumor cell burden may be attributed to the inability of transferred effector cells to keep up with heightened numbers of tumors that rapidly proliferate and lead to enhanced tumor outgrowth in vivo (37). Persistence of such high levels of tumor Ag may lead to tumor-induced tolerance or anergy of adoptively transferred tumor-reactive effector T cells (37, 38). Alternatively, tumor-reactive effector T cells may efficiently eradicate tumor cells, but may select for tumor Ag-negative variants in vivo which would diminish effector-cell recognition and enhance variant tumor outgrowth and progression (39, 40). Additionally, other tumor-related mechanisms contributing to the active immunosuppression of tumor-reactive effector cells have been described and involve release of tumor-derived soluble suppressor factors such as TGF-β and prostaglandins (2), generation of host T cells with suppressor function (41, 42), discordant structural alterations in TCR-mediated signal-transduction pathways (particularly in late tumor-bearing hosts) (37, 43), and induction of CD95-mediated T cell apoptosis by Fas ligand-expressing tumor cells (44, 45).

Although metastatic lesions in effector cell-treated mice were greatly reduced and in fact appeared eliminated in lungs of “cured” mice receiving lower doses of tumor, minimal residual disease among these animals was still apparent. Cytospins from single-cell suspensions of lung homogenates showed low, yet detectable, cell numbers of B16 tumor present in lungs of both Tc1 and Tc2 effector cell-treated mice at times >120 days after tumor rechallenge (data not shown). These observations were in agreement with those described by Vitetta et al. (46), suggesting that a functional immune system can potentially induce a state of cancer dormancy whereby tumor cells would be present but substantially fail to expand for overly extended periods of time. The underlying mechanisms involved in establishing and maintaining tumor dormancy appear to involve an interplay between the immune system and cancer growth. In the case of the latter, tumor variants may arise that carry mutations responsible for transient alterations in the surface expression of select growth-factor receptors and/or tumor-cell signaling pathways that result in a delay in tumor cell growth and responsiveness (46). Alternatively, others have shown that CD8⁺ T cells, but not CD4⁺ T cells, are required for the maintenance of dormancy (47). Moreover, IFN-γ has been shown to act in collaboration with CD8⁺ T cells to mediate, in part, tumor growth arrest and dormancy in vivo (47). Although little is known about the changes that occur in both the host and tumor-cell population that allow tumor escape and regrowth from the dormant state, our observations suggest that, in the course of malignancy, tumor-cell presence is not inconsistent with long-term survival or cure. Further experimentation is warranted to define the potential roles of various tumor-reactive effector-cell subpopulations after Tc1 and Tc2 effector-cell therapy in promoting tumor-cell cytostasis and growth regulation in vivo.

In summary, we show that adoptively transferred Tc1 and Tc2 CD8 effector-cell subpopulations can effectively regulate established B16 melanoma lung metastases and subsequently establish long-term tumor Ag-specific immunity in large proportions of mice with established pulmonary tumor. Antitumor responses induced by Tc2 and Tc1 effector cell therapy were dependent, in part, on both the level of tumor burden and effector cell-derived IL-4, IL-5, and IFN-γ, respectively. Immunoadoptive therapy with such polarized cytokine-secreting effector-cell subpopulations may offer a new strategy for successful tumor immunotherapy and provide a practicable means to elicit more effective T cell-mediated immune responses against various established tumors in primary and metastatic disease.

We thank David Niederbuhl for excellent assistance with the generation of OVA TCR transgene-positive knockout mice. We are particularly grateful to Drs. Edith Lord and John Frelinger for providing the B16-OVA cell line.