Increased Tumorigenicity, but Unchanged Immunogenicity, of Transporter for Antigen Presentation 1-deficient Tumors

INTRODUCTION

Deficiencies of components of the MHC class I-antigen-processing and -presentation machinery have been found in a variety of human tumor tissues and tumor cell lines, and have been associated with tumor progression in melanoma and breast carcinoma as well as in colon carcinoma (1 , 2) . One such component is the peptide transporter associated with antigen processing, TAP, 3 which consists of two subunits (3 , 4) . The heterodimeric TAP1 and TAP2 complex selectively transports peptides from the cytosol into the endoplasmic reticulum (5) . TAP-deficient cells lack peptide transport, which results in a reduced supply of peptides for binding to MHC class I antigens. This is associated with reduced stability and surface expression of MHC class I antigens (6 , 7) . TAP deficiencies have been shown in various tumor cell lines (8, 9, 10, 11, 12) . Depending on the models, they are associated with either reduced sensitivity of tumors to T-cell-mediated immune response or increased tumor susceptibility to natural killer cell-mediated lysis. Wild-type TAP1 gene transfer reconstitutes transporter function and MHC class I surface expression on tumor cells and decreases the growth of tumors in mice (10 , 13 , 14) . However, the underlying mechanisms of the antitumor effect of TAP1 expression by tumor cells is still largely unknown.

The antitumor immune response is a multistep process that can be divided into two phases. The priming phase includes the processing and presentation of tumor-associated antigens by antigen-presenting cells, the activation and proliferation of antigen-specific T cells. In the effector phase, the activated immune cells migrate to the tumor site and exert their effector functions such as cytokine production and cytotoxic activity (15 , 16) . Although most of the tumor cells are MHC class I⁺ and class II⁻, because of their lack of costimulatory molecules, they may not be suitable for the direct activation of CD8⁺ T cells in vivo. Rather, bone marrow-derived antigen-presenting cells have been shown to be important in initiating a MHC class I-restricted antitumor response (17, 18, 19) .

To address the effect of TAP deficiencies in tumor cells on either the priming phase or the effector phase of an antitumor response, we established TAP1-positive (TAP1⁺) and TAP1-negative (TAP1⁻) ras-transformed murine fibroblast cell lines. The tumorigenicity and immunogenicity of these cells, and the T-cell-dependent antitumor mechanism in this model system, were investigated. The results showed that loss of TAP1 expression by tumor cells did not change their immunogenicity; however, it enhanced tumor growth in immunized mice.

MATERIALS AND METHODS

Cells and Animals.

Ras-transformed mouse NIH3T3 fibroblasts (NIHpEJcl3; parental cells) were cultured in RPMI 1640 supplemented with 10% FCS, 100 units/ml penicillin and 100 μg/ml streptomycin. The TAP1 (TAP1⁺) and mock-transfected NIHpEJcl3 cells (TAP1⁻) were cultured in the same medium containing additionally 400 μg/ml G418, which ensures stable expression of the transgene (12) . Six-to-8-week-old female DBA-1 and T-cell-deficient nude (BALB/c nu/nu) mice were purchased from Bomholtgard, Ry, Denmark.

Flow Cytometry.

Tumor cells were stained with the mAbs antimouse H-2L^d/q [clone 30-5-7S; purchased from Cedarlane (Hornby, Canada)] and subsequently with the FITC-conjugated goat-antimouse immunoglobulin (FITC-GAM; Beckman/Coulter, Krefeld, Germany) as secondary mAb. Stained samples were analyzed on a flow cytometer (Coulter Beckman, Krefeld, Germany).

In Vivo Experiments.

Exponentially growing tumor cells were harvested, washed with Dulbecco’s PBS and s.c. injected into the left abdominal region of DBA-1 or nude mice in 0.2 ml of Dulbecco’s PBS. Tumor growth was monitored 2–3 times every week. Mice bearing a tumor larger than 10 mm in diameter were recorded as tumor positive. In this model, tumors larger than 10 mm always continued to grow. Tumors smaller than 10 mm in diameter were rejected or, if not, grew progressively. To analyze the immunogenicity of the tumor, DBA-1 mice were left untreated as naïve control or were immunized by s.c. inoculation of different doses of irradiated (100 Gy) tumor cells. Two weeks later, mice were contralaterally challenged with living tumor cells in variable numbers as indicated in figure legends. Mice that rejected living tumor cells were also used as immunized mice for challenge in one experiment.

Depletion of CD4⁺ and CD8⁺ Cells.

DBA-1 mice were immunized with 4 × 10⁵ irradiated parental tumor cells and were challenged with the same cells 2 weeks later. Three days before challenge, mice were depleted of CD4⁺ or CD8⁺ cells by i.p. injection of 200 μg of rat antimouse mAb GK1.5 (anti-CD4) or 2.43 (anti-CD8) in a volume of 0.5 ml. Depletion of the respective T-cell subpopulation was controlled by flow cytometric analysis of peripheral blood cells using PE-labeled anti-CD4 (RM4-5) and PE-labeled anti-CD8 mAbs (53-6.7; BD PharMingen) and lasted for at least 4 weeks.

IFN-γ Production by Immune Spleen Cells.

DBA-1 mice (three to five per group) were either left untreated or immunized twice by s.c. injection of 1 × 10⁵ irradiated TAP1⁺ or TAP1⁻ tumor cells/mouse. Seven days after the second immunization, spleen cells were isolated and cultured at a concentration of 2 × 10⁶ cells/ml in RPMI medium. For restimulation of the antigen-specific T cells, irradiated TAP1⁺ or TAP1⁻ tumor cells were added to the spleen cell cultures at a ratio of 1:40 (mixed tumor:spleen cell culture). Culture supernatants were collected at day 3 and day 5 after in vitro restimulation. For determination of the IFN-γ production by immune spleen cells, a commercially available kit (OptEIA Mouse IFNγ Set; PharMingen, Hamburg, Germany) was used. The detection limit was 31 pg/ml.

Immunohistochemistry.

DBA-1 mice (three to five per group) were either left untreated or immunized twice by s.c. injection of 1 × 10⁵ irradiated TAP1⁻ tumor cells. Seven days after the second immunization, mice were s.c. challenged with 1 × 10⁶ TAP1⁺ or TAP1⁻ living tumor cells. Tumors were isolated 4 and 6 days after the challenge. Preparation of cryostat tissue sections and alkaline phosphatase immunostaining were done as described previously (20) . The mAbs used here were anti-CD8 (53–6.7) and isotype-matched control mAbs (PharMingen, San Diego, CA). Alkaline phosphatase-conjugated goat antirat IgG and rabbit antigoat IgG were purchased from Jackson Immunoresearch Laboratories, Inc., West Grove, PA.

RESULTS

Increased MHC Class I Expression on TAP1 Gene Transfer in Tumor Cells.

We have previously reported that transformation of murine NIH3T3 fibroblasts with oncogene ras led to reduced expression of TAP1 and MHC class I molecules by the transformed cells, NIHpEJCl3 (12) . To evaluate the effect of TAP1 expression by tumor cells on antitumor immunity, we established TAP1⁺ and TAP1⁻ cells by transfection of NIHpEJCl3 (parental) cells with a TAP1 expression vector and an empty vector containing only the neomycin resistance gene, respectively (12) . TAP1⁺, but not the TAP1⁻ control cells, expressed high levels of TAP1 protein (12) . Correspondingly, TAP1⁺, but not the TAP1⁻ and parental cells, expressed increased levels of MHC class I surface antigens (Fig. 1) ⇓ . The MHC class I expression was stable during the time of the study.

TAP1 Expression Decreases Tumorigenicity only in T-Cell-Competent Mice.

To determine whether the increased MHC class I surface expression after TAP1 gene transfer alters the tumorigenicity of the oncogene-transformed fibroblasts, various doses of TAP1⁺ and TAP1⁻, as well as of parental cells, ranging from 6 × 10³ to 1 × 10⁵ cells/mouse, were injected into syngeneic DBA-1 mice. At an injected dose of 1 × 10⁵ cells/mouse, TAP1⁻ and parental tumors grew out with similar kinetics in all cases (Fig. 2a) ⇓ , whereas 18% of the TAP1⁺ tumors were rejected. On administration of a lower dose of TAP1⁻ and parental cells, such as 2.5 × 10⁴ (Fig. 2b) ⇓ or 6 × 10³ cells/mouse (Fig. 2c) ⇓ , 55–64% or 42–50% of mice, respectively, developed tumors within 50 days on tumor cell inoculation. In contrast, all of the mice that were given injections with 2.5 × 10⁴ or 6 × 10³ TAP1⁺ cells/mouse were tumor free. There exists no significant difference of tumor growth between the TAP1⁻ and parental cells at both of the injected cell concentrations (P > 0.05). It is interesting to note that during the first 10–14 days, TAP1⁺ cells grew faster than the control cells, but the TAP1⁺ cells were eventually rejected, whereas TAP1⁻ cells grew progressively (Fig. 2d) ⇓ .

To ask whether the reduced growth of TAP1-expressing tumor cells was caused by T cells, TAP1⁺ and TAP1⁻ cells were injected into nude mice. As shown in Fig. 2e ⇓ , all of the mice developed tumors with similar kinetics independently of the expression of TAP1. Thus, the TAP1-mediated increase of MHC class I surface expression of tumor cells decreased the tumorigenicity in a T-cell-dependent manner. Because TAP1⁻ cells were not completely MHC class I deficient (Fig. 1) ⇓ , their growth in syngeneic mice was slower than in T-cell-deficient nude mice (by comparison of Fig. 2d ⇓ and 2e ⇓ ). These results indicated that the growth of TAP1⁻ tumors was also to some extent controlled by T cells in immune-competent mice.

TAP1 Expression by Tumor Cells Is Not Required for Immunization.

High levels of MHC class I surface expression on tumor cells could affect tumor immunity by increasing direct antigen presentation to immune cells in the priming phase or, alternatively, by improving the recognition of tumor cells by CD8⁺ T cells in the effector phase. To ask whether TAP1 and, therefore, MHC class I expression on tumor cells increase their immunogenicity, DBA-1 mice were immunized with various concentrations of irradiated TAP1⁺ or TAP1⁻ tumor cells. Two weeks after immunization, mice were challenged with 1 × 10⁵ parental tumor cells. As shown in Fig. 3 ⇓ , immunization with 1 × 10⁵ or 4 × 10⁵ irradiated tumor cells protected mice from tumor challenge in a dose-dependent manner. However, there existed no significant difference between tumor growth in mice immunized with TAP1⁺ and TAP1⁻ cells irrelevant of the cell number used for vaccination (in all cases, P > 0.05). In conclusion, TAP1 expression on tumor cells is not necessary for immunization, which is compatible with the assumption that T cells are mainly activated by “cross-priming.” Because TAP1⁻ cells still expressed low levels of MHC class I surface antigens, the role of direct antigen presentation by tumor cells cannot be excluded.

TAP1 Expression by Tumor Cells in the Effector Phase Increases Tumor Rejection.

Because TAP1 expression by tumor cells is not required for immunization, the growth of TAP1⁺ and TAP1⁻ cells in previously immunized mice was analyzed. DBA-1 mice were left untreated or were immunized with 1 × 10⁵ irradiated parental tumor cells, and, 2 weeks later, mice were challenged with either TAP1⁺ or TAP1⁻ tumor cells. All of the mice challenged with TAP1⁺ cells were tumor free, whereas 50% of the mice challenged with TAP1⁻ cells developed tumors (Fig. 4) ⇓ . The difference of tumor growth between TAP1⁺ and TAP1⁻ cells in immunized mice was statistically significant (P < 0.01). In an additional experiment, mice that had rejected the primary TAP1⁺ tumors as shown in Fig. 2b ⇓ were used. They were randomly separated into two groups and challenged either with TAP1⁺ or TAP1⁻ tumor cells. Three of three mice rejected the TAP1⁺ tumor, whereas one of three mice rejected the TAP1⁻ tumor.

CD8⁺ T-Cell Infiltration in TAP1⁺ and TAP1⁻ Tumors in Immunized Mice.

To assess whether the down-regulated expression of MHC class I molecules on tumor cells is associated quantitatively or qualitatively with different tumor infiltrating cells, DBA-1 mice were immunized and challenged with TAP1⁺ or TAP1⁻ cells. At day 4 after challenge, consecutive cryostat sections of tumor tissues were analyzed by immunohistological staining. As shown in Fig. 5 ⇓ , both TAP1⁺ and TAP1⁻ tumors isolated from immunized, but not naïve, mice were evenly infiltrated by CD8⁺ cells. A few CD4⁺ cells also infiltrated tumors and many Gr-1⁺ and Mac-1⁺ cells were found in the tumor center of immunized mice (data not shown). Considering the quantity and the distribution of tumor infiltrating CD8⁺ cells, we detected no significant difference between the TAP1⁺ and TAP1⁻ tumors.

Tumor Rejection in Immunized Mice Is CD8⁺ T-Cell Dependent.

Because both CD4⁺ and CD8⁺ cells infiltrated tumors, we asked which type of T cells was required for tumor rejection. DBA-1 mice were immunized and challenged with the parental tumor cells. The CD4⁺ or CD8⁺ cells were depleted 3 days before challenge. As shown in Fig. 6 ⇓ , immunization of mice with TAP1⁻ parental cells increased their tumor resistance from 10 to 90%. Depletion of CD8⁺ cells completely prevented tumor rejection. In contrast, depletion of CD4⁺ cells did not significantly change the tumor protection rate (P > 0.05). These results indicated that tumor rejection required CD8⁺ but not CD4⁺ T cells in the effector phase.

TAP1⁺ Tumor Cells Induced IFN-γ Production by Spleen Cells from Immunized Mice.

It is known that CD8⁺ T cells can contribute to the antitumor immunity by producing cytokines, like IFNγ or tumor necrosis factor (21) . Therefore, DBA-1 mice were immunized and spleen cells were isolated and restimulated with irradiated TAP1⁺ or TAP1⁻ tumor cells. The supernatants of mixed tumor:spleen cell cultures were collected and analyzed for IFN-γ production by ELISA. As shown in Table 1 ⇓ , spleen cells of naive mice did not secrete detectable amounts of IFN-γ on stimulation with TAP1⁺ or TAP1⁻ cells. However, spleen cells of mice immunized either with TAP1⁺ or TAP1⁻ tumor cells secreted large amounts of IFN-γ. Importantly, the immune spleen cells produced 4 times more IFN-γ, if they were restimulated with TAP1⁺ in comparison with TAP1⁻ mock-transfected tumor cells. There was no significant difference in IFN-γ production between groups immunized with TAP1⁺ and TAP1⁻ cells. Together, TAP1⁺ and TAP1⁻ tumor cells are equally effective for IFN-γ induction, if they were used as vaccines, but not if they were used as restimulators of specific immune cells.

DISCUSSION

Gene-modified tumor cells have been intensively investigated for the development of effective cancer vaccines (22, 23, 24) . One group of genes, such as MHC class I (25) or class II (26) , or costimulatory molecules (27) , was investigated to increase the antigen presentation by tumor cells. TAP1, an important component for the presentation of MHC class I-restricted antigens, has recently been demonstrated to inhibit tumor growth in several tumor models (10 , 13 , 14) . However, whether TAP1 expression by tumor cells leads to a higher antigen expression by tumor cells during the induction of an immune response or a better recognition of tumors by T cells in a preimmunized host is not clear. We demonstrated that reduced TAP1 expression by tumor cells did not interfere with the induction, but with the effector phase, of tumor immunity.

We previously observed that irradiation of tumor cells led to a substantial loss of the vaccine effect (28) . Tumor cells transfected to express different cytokine genes or B7 lost the vaccine efficacy if they were irradiated (29 , 30) . An explanation for this is that the enhanced vaccine efficacy of these gene-modified tumor cells is basically caused by the increased amount of antigens to which the host was exposed. Live cells proliferate in vivo, and antigen amounts are accumulated before the tumor is rejected, whereas irradiated cells do not proliferate. Similarly, as shown in Fig. 2d ⇓ , TAP1⁺ cells grew faster than TAP1⁻ cells during the first few days in naïve mice, although they were eventually rejected. If such tumor-free mice were used for challenge experiments, mice that had rejected TAP1⁺ tumors could have effectively controlled the challenge tumor, whereas mice that had rejected the TAP1⁻ tumors did not (data not shown). This result cannot be interpreted as simply increased immunogenicity of the TAP1⁺ cells, because mice immunized with TAP1⁺ tumors have been exposed to a higher amount of tumor antigens than mice immunized with TAP ^- tumors. Notably, this problem was not discussed in several similar studies with TAP1⁻ or MHC class I^- tumor cells (13 , 14 , 31 , 32) . Therefore, the key result is that immunization with irradiated TAP1⁺ tumor cells was not more effective than immunization with irradiated TAP1⁻ tumor cells. Here it is worthy to note that irradiation with a dose of 100 Gy did not change MHC class I expression of either TAP1⁺or TAP1⁻ cells (data not shown). Therefore, our results strengthen the concept of T-cell activation via cross-priming. By using of a MHC class I loss variant of a UV-induced murine tumor, Seung et al. (33) showed that this tumor could still induce a MHC class I-restricted antigen-specific CD8⁺ T-cell response, even though the variant could not be lysed by the antigen-specific CTLs. These results, together with other two recent publications (34 , 35) demonstrated that cross-presentation is effective enough for induction of T-cell response. However, direct antigen presentation by tumor cells is required for their recognition by CD8⁺ T cells during the effector phase of an antitumor response.

The presence of CD8⁺ T cells in tumor infiltrates may reflect an important role of these cells in an antitumor response. Depletion of CD8⁺ T cells totally eliminated antitumor immunity in this tumor model (Fig. 6) ⇓ . However, the dispersed distribution and the limited numbers of TILs in both TAP1⁺ and TAP1⁻ tumors (Fig. 5) ⇓ do not suggest that tumor growth is mainly controlled by a T-cell-mediated killing mechanism. To attribute the increased tumorigenicity of TAP1⁻ cells to impaired T-cell cytotoxicity, a typical chromium release assay was performed. However, no significant T-cell-mediated killing of either TAP1⁺or TAP1⁻ tumor cell targets could be detected (data not shown). IFNγ production by T cells may be more relevant than the cytotoxicity for their antitumor effector functions (21 , 36) . It is important to note that the IFNγ production by immune spleen cells was increased when they were restimulated with TAP1⁺ in comparison with TAP1⁻ tumor cells. It is likely that CD8⁺ T cells produced IFNγ, because the depletion of CD8⁺, but not CD4⁺, T cells in the effector phase prevented tumor rejection (Fig. 6) ⇓ . Treatment of TAP1⁻ cells with IFNγ up-regulates the expression of various antigen-processing components as well as of MHC class I surface molecules (12 , 13) . Because of the possible up-regulation of MHC class I on TAP1⁻ challenge tumors, it is likely that the different outcome of tumor growth between TAP1⁺ and TAP1⁻ cells is because of the different time points of the activation of specific effector T cells in immunized mice. Barth et al. (21) found that the effectiveness of adoptively transferred CD8⁺ T cells in suppressing tumor growth in vivo correlated with their ability to produce IFNγ and TNFα rather than with their cytotoxicity. IFNγ-specific antibodies abrogated the ability of different CD8⁺ TIL cultures to mediate tumor regression, which indicated that IFNγ secretion is an essential part of the mechanism of TIL function (21) . The requirement of IFNγ in our tumor model has, however, still to be demonstrated. Recently, Becker et al. (37) showed that adoptive transfer of IFNγ secreting, but not nonsecreting, T cells, isolated from tumor-immunized mice, could inhibit tumor growth in recipient mice. Together, the finding that the TAP1 expression by tumor cells did not affect the immunogenicity of tumors could be important for the design of cancer vaccines and clinical immunotherapeutical protocols in the future.