We have previously described a tumor model in which the influenza hemagglutinin protein (HA) expressed on the BALB/c-derived MT901 tumor line serves as an immunization-dependent tumor rejection antigen in normal syngeneic mice. Although the HA antigen in this model is clearly foreign to normal BALB/c mice, many tumor antigens recognized by T cells in humans and mice are nonmutated antigens that are expressed on normal tissues as well as on the tumor cells, thereby raising issues of self-tolerance and autoimmunity in attempts to use such antigens therapeutically. To examine these issues, we have applied our tumor model to syngeneic mice that broadly express HA at low levels as a “self”-transgene. Unlike the situation in normal BALB/c mice, immunization of HA-transgenic mice did not result in tumor protection nor did it generate cytotoxic T cell or IgG responses against the HA self-antigen. However, if immunization of HA-transgenic mice was preceded by adoptive transfer of spleen and lymph node cells from normal untreated BALB/c mice, then HA-specific tumor protective immune responses were generated. Despite the self-nature of the HA antigen, no obvious manifestations of autoimmunity were observed. The immunity established in the transgenic mice was notably different from that observed in normal mice in that it was considerably more transient and required CD4 T cells for both successful immunization and subsequent tumor protection, qualities that were not associated with the immunity established in normal BALB/c mice. Collectively our results suggest that transferred cells can be transiently and selectively directed against a tumor-expressed self-antigen before returning to a tolerant state.

The past decade has seen the identification of a large number of tumor antigens and the elicitation of antitumor responses against such antigens in both humans and mice (1, 2). Perhaps surprisingly, many of these antigens, including the P1A antigen (3), the tyrosinase-related protein 2 (4), and the HER-2/neu antigen (5), are expressed not only on the cancerous cells, but also on normal tissues. A number of studies have demonstrated that autoreactive T cells with specificity for such peripherally expressed antigens are present in the normal T cell repertoire and attempts to stimulate these cells to mediate tumor rejection have met with varying degrees of success in several tumor models. Thus, there are instances where antitumor responses directed against a normal differentiation antigen could be generated with little evidence for destructive autoimmunity (6, 7, 8, 9, 10). However, in some cases successful rejection of established tumors could only be achieved in the presence of severe autoimmune responses (11) and in other instances, the immune responses elicited were unable to cause tumor rejection (12). Collectively, these results have served to emphasize that differences between immune requirements for tumor rejection and those mediating autoimmunity can be exploited for tumor rejection in some cases, but that the ability to break tolerance and induce significant autoimmune responses without causing autoimmune disease is highly dependent on the antigen chosen.

Although the majority of self-tumor antigens studied to date are differentiation antigens with a restricted normal tissue distribution, one antigen, p53, differs significantly in that it is broadly expressed at low levels in virtually every nucleated cell but is overexpressed in a substantial portion of cancers (13). Whether broadly expressed self-antigens can serve as targets for cancer immunotherapy has remained an open question. In the case of p53, cytotoxic T cells recognizing the native protein have been generated from murine and human responding lymphocytes and some of these lyse p53-overexpressing tumors in vitro(14, 15). Additionally, adoptive transfer of p53-specific CTLs generated in p53null mice have been shown to mediate tumor rejection without inducing autoimmune disease when transferred into normal mice (16). Lastly, protective anti-p53 responses have been generated by immunization with p53-transduced dendritic cells (17). These data suggest that in the case of p53, despite broad expression on normal tissues, overexpression of p53 on tumor cells may be sufficient to make it a useful target for cancer immunotherapy.

The extent to which p53 is unique as tumor therapy target, as opposed to being representative of a class of broadly expressed self-antigens that are differentially expressed by tumor cells, is unclear. To examine whether a non-p53 antigen that is similarly expressed broadly on self-tissues and at higher levels on tumor tissues can serve as a tumor rejection antigen, we developed a tumor model using the influenza hemagglutinin protein (HA)3 expressed on a tumor cell line and syngeneic mice that have a broad, low-level expression of HA as a transgene. HA has served as an important model antigen in examining issues of peripheral tolerance and autoimmunity. As a model differentiation antigen, HA expressed under the rat insulin promoter (ins-HA) has been used to probe how self-tolerance to a peripheral antigen affects antigen-specific tumor rejection. These studies found that while ins-HA mice were tolerant to the HA protein (18), they were able to generate low-avidity, HA-specific CTLs that protected against HA-expressing tumor challenge without overt autoimmunity (6). Studies using HA-specific T-cell receptor-transgenic mice and HA-expressing tumors have been important in illustrating the ability of established tumors to induce antigen-specific T-cell anergy (19) and have also been useful in demonstrating the importance of the CD4 T-cell response in tumor rejection (20, 21). However, in these studies, the HA antigen was tumor specific and not expressed on normal tissues.

We previously reported that tumor expression of HA in BALB/c mice, where it is perceived as a “neo-antigen,” renders the syngeneic mammary carcinoma line MT901 strongly immunogenic (22). In this report, we show that expression of the HA antigen does not induce tumor immunogenicity in mice that also express HA as a transgenic “self-antigen.” Consistent with these mice having a well-established state of immune tolerance to the HA antigen, we were unable to demonstrate either cytotoxic T- or B-cell responses against the HA antigen in HA-transgenic mice that had been exposed to HA-expressing MT901 tumor cells. Significantly, though these mice demonstrated tolerance to the HA antigen, adoptive transfer of bulk population T cells from transgene negative mice, which contain HA-specific T cells, was able to transiently confer specific protection against the HA-bearing tumor without induction of overt autoimmune disease. Together these results suggest that HA-specific T cells can be expanded and mediate tumor protection even in an environment where the target tumor antigen is expressed at low levels on multiple normal tissues. These results are consistent with a threshold model for self-antigen-specific cancer immunotherapy in which encounter of antigen on the tumor can be distinguished from encounter of antigen on normal self-tissues in a manner that allows tumor rejection without autoimmune destruction of normal tissues.

Animals.

BALB/c mice were purchased from Harlan Laboratories (Madison, WI). HA104 mice expressing the influenza virus A/PR/8/34 (H1N1) hemagglutinin gene as a SV40 promoter-driven transgene and transgene-negative littermate mice were provided courtesy of Dr. Andrew Caton (Wistar Institute, Philadelphia, PA) and have been previously described (23). All mice were maintained in a pathogen-free microisolator environment at the University of Pennsylvania School of Medicine Laboratory Animal Facilities and were maintained and sacrificed in full accordance with federal and state policies on animal research.

Antibodies, Immunofluorescence, and Flow Cytometry.

Tumor cells maintained in culture were harvested and analyzed with biotinylated mAbs specific for H-2Kd, H-2Dd, B7.1, B7.2, I-Ad (BD PharMingen, San Diego, CA), as well as the H36-4-5 mAb specific to the Sb region of the hemagglutinin gene of the influenza virus A/PR/8/34 (H1N1) (24) (provided courtesy of Dr. Walter Gerhard, Wistar Institute). For staining, cells were suspended in PBS supplemented with 5% fetal bovine serum, incubated for 30 min at 4°C with 0.5–1 μg of the relevant mAb per 106 cells, washed twice, and 0.5 μg/test of streptavidin-FITC or streptavidin-phycoerythrin (BD PharMingen) was added to the residual volume. Cells were incubated an additional 30 min at 4°C and washed an additional two times before suspension in PBS or PBS supplemented with 1% formaldehyde and analysis on a FACScan or FACSCalibur (Becton Dickinson, Sunnyvale, CA).

Cell Lines and Transfections.

MT901, subsequently referred to as MT-WT (for MT901 wild-type), is a weakly immunogenic, BALB/c-derived mammary carcinoma line (25), provided courtesy of Dr. Alfred Chang (University of Michigan, Ann Arbor, MI). The MT-HA and MT-CT lines were generated by transfection of MT-WT with an HA expression plasmid (using the cytomegalovirus promoter) or the empty control vector, respectively, as previously described (22). P815 mouse mastocytoma cells were provided by Yvonne Paterson (University of Pennsylvania) and CT26, a BALB/c-derived murine carcinoma line, was provided courtesy of Dr. William Lee (University of Pennsylvania).

Viruses, Peptides, and Viral Immunizations/Infections.

Initial stocks of allantoic fluid containing influenza virus strain PR8 [A/Puerto Rico/8/34(H1N1)] (HA) as well the J1 virus (a cross between PR8 and HK in which the PR8-HA has been replaced with the serologically non-cross-reactive H3 subtype [A/Hong Kong/1/68(H3N2)]) were provided by Dr. Andrew Caton. Influenza virus was grown in the allantoic cavity of embryonated hen’s eggs, frozen in aliquots as allantoic fluid, or further purified by sucrose gradient centrifugation as previously described (26). Virus titers were determined by hemagglutination as previously described (26). The H-2Kd binding peptide antigen corresponding to positions 518–526 of the PR8-HA protein was produced by the Protein Chemistry Laboratory at the University of Pennsylvania. For immunizations, allantoic fluid mixed with PBS (20 μl, total volume) containing 20 HAU of influenza virus was delivered to mice intranasally. Influenza infection of cells was accomplished by culturing cells in 400 HAU/ml of influenza virus in serum-free RPMI 1640 for 1 h followed by washing to remove free virus particles. Peptide pulsing of cells was accomplished by culturing of cells at 106 cells/ml in 1 nm HA peptide for 1 h prior to washing and use in assays.

Tumor Cell Inoculations and Immunizations.

For primary tumor challenges, tumor cells suspended in 100 μl of PBS were inoculated s.c. into either the right or left hindquarter of the mouse. Onset of tumor growth was scored by the presence of a palpable lump near the site of inoculation. Tumor growth was monitored at 2- to 5-day increments after inoculation by measurement in two perpendicular dimensions with a microcaliper, and the size was recorded as an area. For tumor immunization studies, mice were injected s.c. as indicated with 106 irradiated (5000 rad) MT-HA, MT-CT, or MT-WT tumor cells in 100 μl of PBS as described above. Fourteen to 42 days after immunization, mice were rechallenged by s.c. inoculation of nonirradiated tumor cells to the opposite flank from the immunization site. Onset of tumor and tumor growth were measured as described above.

Production of Cytotoxic Effector Cells.

BALB/c or HA104 mice were immunized with 106 irradiated (5000 rad) MT-HA tumor cells or influenza virus (as a positive control). Fourteen days later, tumor-immunized mice were challenged in the opposite flank with 106 viable MT-HA tumor cells. Five days after tumor challenge, mice were sacrificed and spleen and draining lymph node cells were harvested, processed to single-cell suspension, and cultured for 5 days at a 1:1 ratio with 3 × 106 irradiated (3000 rad) influenza-infected syngeneic spleen cells/ml or at a 3:1 ratio with 1 × 106 irradiated (12,000 rad) HA peptide-pulsed P815 cells. For cytotoxicity assays, syngeneic P815 and CT26 cells were HA peptide pulsed or influenza infected, respectively, or mock treated to serve as negative controls and labeled at 37°C for 2 h using 200 μCi of sodium 51chromate/106 cells. Targets were washed and seeded into 96-well round-bottomed plates at 5000 cells/well in 100 μl of complete RPMI 1640, to which effector cells were added at 2.5–5 × 105 cells/well to establish 50:1 and 100:1 E:T ratios. After 6 h at 37°C, plates were centrifuged and 100 μl of supernatant were removed from each well to assess isotope release using a scintillation counter. Percent specific lysis was calculated by the formula:

\(percent\ specific\ lysis\ {=}\ (sample\ release\ {-}\ spontaneous\ release)/(maximal\ release\ {-}\ spontaneous\ release)\ {\times}\ 100\)
⁠.

Detection of Serum Antibody Production.

ELISA plates were coated overnight with 50 μl of sodium carbonate buffer containing either pelleted or purified virus at approximately 103 HAU/ml. Plates were washed twice in 0.5% PBS-Tween 20 (Sigma, St. Louis, MO), and serum samples or anti-HA mAb H36-4-5 standards diluted in PBS with 5% fetal bovine serum were added as indicated. Plates were again incubated overnight at 4°C and then washed three times as described above. Biotin-labeled goat antimouse IgG (γ-chain specific; Sigma) was added at 100 ng/ml in PBS with 10% FCS and incubated for 45 min at room temperature prior to four washes and addition of avidin-labeled alkaline phosphatase. Plates were incubated for another 45 min and washed four times prior to addition of substrate. Plates were read on an EL-311sx plate reader (Biotek Instruments, Winooski, VT) at least 1 h after addition of substrate.

Cell and Antibody Transfer Experiments.

For cell transfers, spleen and lymph node cells were harvested from naive BALB/c or HA104 mice as indicated, mechanically processed to single-cell suspensions, and transferred i.v. into naive HA104 mice by retro-orbital injection prior to tumor cell immunizations and/or challenges. For antibody transfers, H36-4-5 mAb was administered at 1 mg/ml in PBS by i.v. injection (retro-orbital) at the doses and times indicated. For serum transfers, BALB/c mice were immunized with irradiated MT-HA tumor cells, then challenged with viable MT-HA tumor cells as described above. Five days after tumor challenge, sera were harvested, pooled, diluted in an equal volume of PBS, and administered to naive HA104 mice by i.p. injections as indicated.

CD4 T Cell Depletions.

Depletion of CD4 cells from mice was accomplished as previously described (22). Briefly, mice were given i.p. injections of 150 μg of purified anti-CD4, GK1.5 mAb (obtained from the American Type Culture Collection, Manassas, VA) in approximately 100 μl of PBS for 3 consecutive days followed by administration of an additional 150 μg of mAb every 4 days. Mice treated with this regimen typically showed <2% of CD4+ or CD8+ T cells, respectively, in their spleens for at least 4 days after cessation of treatment as measured by antibody staining and flow cytometry analysis. To generate CD4-depleted cells for use in adoptive transfer experiments, freshly harvested spleen and lymph node cells were incubated in GK1.5 hybridoma supernatant (5 ml/108 cells) and washed with PBS before addition of anti-rat immunoglobulin-coated magnetic beads (Perspective Biosystems, Farmingham, MA). Coated cells were then removed by magnetic separation. A sample of the purified cells were stained for fluorescence-activated cell sorting analysis with anti-Thy1.2-allophycocyanin, CD4-FITC, and CD8-phycoerythrin (BD PharMingen) using rat and hamster immunoglobulins (BD PharMingen) as controls to verify that samples contained <3% of the depleted cell population.

Immunization of HA104 Mice with Irradiated MT-HA Tumor Cells or Influenza Virus, either Alone or in Combination, Fails to Generate Tumor-protective Responses or HA-specific Cytotoxic Responses.

We previously reported the generation of an HA-expressing transfectant of the MT901 cell line (MT-HA) and an empty vector control line (MT-CT) (22). These lines are MHC class II, B7, intercellular adhesion molecule 1, and CD40 ligand negative and express similar levels of H-2Kd. Although the HA antigen has little effect on the growth of primary MT tumor cell inoculations in naive immunocompetent mice, it serves as a target for CD8+ T cell-mediated tumor rejection in mice immunized with either influenza virus or irradiated MT-HA tumor cells. Rejection responses only occurred against the HA-expressing line, demonstrating that the HA antigen could serve as a tumor rejection antigen if appropriate and sufficient immune responses were present (22). In this study, to determine whether the HA antigen could also serve as a model tumor rejection antigen in mice that ubiquitously express the HA antigen as a self-antigen, we used HA104 mice. The HA104 line of transgenic mice express HA on most tissues, including the thymus, and broad-based T cell tolerance to HA has been reported for identically derived HA-transgenic mice with similar expression levels and patterns (23).

To examine whether immunization approaches that were effective in normal BALB/c mice, where the HA antigen is perceived as foreign, could be effective in HA104 mice, HA104 and transgene-negative mice were immunized with influenza virus, irradiated MT-HA cells, or MT-CT cells given coincident with influenza infection. As shown in Fig. 1, whereas these maneuvers were all protective in transgene-negative littermates, they were completely ineffective in protecting HA104 mice from a subsequent MT-HA tumor cell challenge, indicating that if immune responses were raised against the HA antigen in these mice, they were insufficient to mediate tumor protection.

Absence of Antibody and Cytotoxicity Responses in MT-HA Tumor-immunized HA104 Mice.

Although immunization approaches that were successful in BALB/c or transgene-negative mice were not able to protect HA104 mice, it remained possible that presentation of the HA antigen under immunizing conditions could generate HA-specific responses, but that the magnitude or quality of these responses were insufficient to generate tumor protection. Therefore, we examined the HA antigen-specific cytotoxic and B-cell responses after immunization of HA104 mice with MT-HA tumor cells. As shown in Table 1, whereas immunization of normal BALB/c mice with irradiated MT-HA tumor cells elicited cytolytic responses against either influenza virus-infected targets or targets pulsed with the HA-derived H-2Kd epitope peptide (518–526), similar immunization of HA104 mice failed to elicit detectable cytotoxic responses. These results suggest that HA-specific cytotoxic responses, if raised at all in HA104 mice, are extremely weak and remain below threshold for detection by this method. Similarly, examination of HA-specific IgG production in the serum of MT-HA tumor-immunized HA104 and BALB/c mice revealed relatively high titers of HA-specific IgG antibodies in BALB/c mice, but no measurable IgG responses in HA104 mice as examined by an HA-specific ELISA (Fig. 2). Since HA104 mice are known to generate a modified IgG HA-specific antibody response upon challenge with influenza virus (27), the absence of such responses here suggest that IgG responses against the HA antigen expressed on the tumor are dependent on T cell help that is lacking in the HA104 mice. Further supporting this conclusion, we found that treatment of normal mice with anti-CD40 ligand mAb, a treatment which abrogates the development of MT-HA-specific tumor immunity in normal BALB/c mice (22), also blocks primary HA-specific IgG responses following MT-HA tumor immunization in these mice (data not shown). Importantly, although antibody responses were present in tumor-protected normal mice, these responses are unlikely to mediate tumor rejection, as treatment of BALB/c mice with anti-HA mAb alone or in conjunction with MT-WT tumor immunization had no effect on subsequent growth of MT-HA tumor challenges (Fig. 3).

Tumor Protection Can Be Adoptively Transferred with T Cells from Syngeneic Normal Transgene-Negative Mice to HA Transgene-expressing Mice.

Although we were unable to identify HA-specific cytotoxic or antibody responses from HA104 mice induced under conditions of MT-HA tumor cell immunization, these findings did not address the question of whether HA-reactive T cells could mediate protection in HA104 mice. Rather than further examine the function of different immunization approaches on the activation of such HA-specific responses (if any) as might be inducible in these mice, we chose to focus instead on whether HA-specific cells could be transferred into an environment in which they are autoreactive and differentiate functionally between the self-antigen expressed on the tumor and that on self-tissues. To do this, HA104 mice were given adoptive transfers of spleen and lymph node cells from naive BALB/c mice. Recipient HA104 mice were then immunized with irradiated MT-HA tumor cells as described above. This coordination of adoptive transfer with immunization of mice using irradiated tumor cells protected the majority of mice against subsequent tumor challenges (Fig. 4,A). However, unlike protection mediated by irradiated MT-HA tumor cell immunization of transgene-negative mice, where none of five mice challenged 41 days after tumor immunization developed tumors (data not shown), the immunity provided by adoptive transfer coupled with MT-HA tumor immunization was not long-lived. Mice challenged with MT-HA tumor cells 14 days after adoptive transfer and immunization were well protected (Fig. 4,A), but only one of five mice challenged 21 days after adoptive transfer and immunization was protected (Fig. 4,B). If tumor immunization was delayed and given 7 or 26 days after adoptive transfer, only 2 of 11 total mice were protected against a challenge with MT-HA tumor cells 14 days later (Fig. 4, C and D). Collectively, these data suggest that there is a window in time in which adoptively transferred cells with specificity for a broadly expressed self-antigen may be induced to mediate tumor protection against that antigen expressed on tumor cells. The observed protection appeared to be HA specific as adoptive transfer of HA104 cells into HA104 mice provided no protection to subsequent MT-HA tumor challenges. Despite our findings indicating that protection afforded was dependent on the HA antigen, none of the adoptively transferred mice showed signs of autoimmune distress (ruffled fur, weight loss, lethargy, or agitation). Collectively these results indicate that HA-specific T cells present in the unimmunized repertoire of a BALB/c mouse can transiently participate in protective immune responses in self-tolerant HA104 mice and that they can be directed against HA without the induction of overt autoimmune disease despite the expression of the target antigen on most cells.

Depletion of CD4+ T Cells either before Transfer or after Transfer and Immunization Abrogates the Transfer of Immunization Competence.

We previously reported that tumor immunity to MT-HA could be established in BALB/c mice that had been depleted of CD4+, but not CD8+, T cells even when depletion occurred prior to immunization. These results suggested that naive CD8+ T cells did not require CD4+ T cell help to generate protective antitumor immune responses in this model. Therefore, we were interested in determining whether naive CD8+ T cells could transfer immunization competence to HA104 mice in the absence of CD4+ T cells. To examine this, spleen and lymph node cells from naive BALB/c mice were processed to single-cell suspension and depleted of CD4+ T cells using antibody-coated magnetic beads. This method consistently depleted CD4+ T cells to <5% of the total cell population. Resultant CD4-depleted populations were adoptively transferred into naive HA104 mice by retro-orbital inoculation, and mice were immunized with 106 irradiated MT-HA tumor cells as described above. Fourteen days later, mice were challenged with viable MT-HA tumor cells and followed for tumor onset and progression. As shown in Fig. 5 A, mice receiving CD4-depleted cell transfers were unable to generate successful immunity against the MT-HA tumor. These results indicate that in the context of the cell transfers to HA104 mice, CD4+ T cells were important in establishing tumor immunity.

One possible explanation for the observed dependence of protective immune responses on the presence of CD4+ T cells in the adoptive transfer population could be that sufficient priming and/or expansion of CD8+ T-cell responses required CD4+ T cell help. Certainly, fewer HA-specific CD8+ cells were transferred into HA104 mice than were present in normal BALB/c mice. Additionally, whereas in our previous study (22) depletion of CD4+ T cells was not obligatory for development of MT-HA tumor immunity in BALB/c mice, animals depleted of CD4+ T cells before immunization were somewhat less likely to be protected against the subsequent tumor challenge, suggesting that CD4+ T cells could enhance the development of protective immune responses. In contrast, depletion after immunization had no effect on immune protection, suggesting that once CD8+ T cells had been sufficiently expanded and activated, CD4+ T cells were not necessary for adequate protective immunity. To determine whether the presence of CD4+ T cells during the immunization phase the response would be sufficient to enable tumor rejection in HA104 mice, HA104 mice received adoptive transfers of naive BALB/c splenocyte and lymph node populations followed by immunization with irradiated MT-HA tumor cells as before. Twelve days later, mice were treated for 3 consecutive days with CD4-depleting mAb as before. On the 3rd day of mAb treatment, mice were challenged with viable MT-HA tumor cells as before. As shown in Fig. 5 B, none of the mice depleted of CD4+ T cells after tumor immunization, but prior to challenge, was able to reject the MT-HA tumor challenge. Thus, under the conditions studied here and unlike the case for transgene-negative BALB/c mice, CD4+ T cells were absolutely required for both successful immunization and subsequent tumor rejection in HA104 mice.

T-cell responses against antigens expressed both on cancerous and noncancerous cells occur naturally in cancer patients. In fact, such shared antigens currently comprise the bulk of identified tumor antigens in human cancers. Whether tumor-autoantigen-specific T cells represent an active response directed against the cancer, as opposed to reflecting the normal autospecific T-cell repertoire of natural immunity, is unclear, but the presence of such cells has led to a growing momentum behind approaches aimed at targeting shared tumor antigens for cancer immunotherapy. Mobilizing the immune system against antigens that are expressed on normal and tumor cells is potentially problematic for several reasons. First, it is well established that the immune system can, and frequently does, exhibit profound tolerance to antigens expressed on normal tissues. Second, even where responses can be elicited against a self-antigen, the response may be insufficient or inappropriate to mediate cancer rejection. Finally, where immunity to such self-antigens is induced, there remains the issue of autoimmune destruction of normal tissues expressing the same antigen.

Despite these limitations, several experimental models have had success in impeding the growth of certain tumors by raising immune responses against lineage-specific or tissue-differentiation antigens that are shared between the normal tissue and the tumor cell (6, 7, 8, 9, 10, 12). In mice that express HA as a transgene on pancreatic islet cells, immunization with influenza virus or a vaccinia construct expressing a dominant class I (Kd) HA peptide activated low-avidity CTL cells that were able to mediate specific rejection of HA-expressing Renca tumor cells without inducing appreciable autoimmune destruction of islet cells (6). Similarly, using vaccinia viruses expressing murine homologues of five human melanocyte-differentiation antigens, Overwijk et al.(9) were able to generate tumor-protective immune responses targeted against one of the five antigens examined (TRP-1). However, in this case vitiligo was induced, suggesting that autoimmune destruction of normal cells was also ongoing. Although these experiments have suggested that differences between immune requirements for tumor rejection and those mediating deleterious autoimmunity can exist, a recent study examining the requirement for destruction of established tumors sharing an antigen with either β islet cells or arterial smooth muscle and cardiomyocytes found that tumor destruction targeted against the self-antigen could not be dissociated from severe autoimmune responses (11). Thus, the conditions that permit selective destruction of tumor cells without unacceptable autoimmune destruction remain unclear.

We have previously reported the generation of a tumor model in which the influenza HA expressed on a tumor cell line serves as an immunization-dependent tumor rejection antigen in normal mice. In this report, we extend this model into mice that express the HA antigen at low levels on most tissues as a transgene. In these mice, neither viral nor tumor immunization was effective at raising detectable responses against the HA autoantigen. However, HA-specific tumor protection could be generated in these mice if they received spleen and lymph node cells from untreated, normal BALB/c mice prior to immunization. In normal mice, immunization against the HA antigen generates a long-lived HA-specific tumor immunity. This immunity can be established in the absence of CD4+ T cells either at immunization or subsequent tumor challenge. In marked contrast, the tumor immunity established by adoptive transfer in the HA-Tg mice is comparatively short-lived and requires both CD4+ and CD8+ T cells to be present at the time of immunization and tumor challenge. Importantly, tumor-rejecting mice displayed no signs of overt autoimmunity. Of course, lack of clinically evident autoimmunity does not preclude a low level of T- or B-cell autoreactivity, but it does suggest that any such responses in this model are self-limiting.

In some respects, our results are similar to those observed in treatment of tumors that overexpress p53. Like HA in our model, p53 is broadly expressed in normal tissues, but is overexpressed in around 50% of human cancers (13). Vierboom et al.(16) showed that adoptive transfer of wild-type p53-specific CTL cells raised in p53null mice could discriminate between p53-overexpressing tumor cells and normal tissues to mediate tumor rejection in the absence of significant infiltration of normal tissues by the autoreactive cells (16). Most likely, these results reflect differences in the levels of antigen expression by the tumor cells and normal tissues. However, it is also possible that the tumor environment provides necessary immune-enabling “danger” signals that are not present in the normal tissue such as has been proposed by Matzinger (28). Our results differ significantly from Vierboom et al.(16) in that, rather than transfer activated self-tumor antigen-specific effector cells, we transferred a presumably naive population of lymphocytes that included self-tumor antigen-specific progenitors and were able to transiently activate protective immune responses from these cells by immunization in situ, despite broad expression of the HA antigen on normal tissues.

In our model, we did not observe HA-specific antibody or cytotoxic responses in HA104 mice that received MT-HA tumor immunizations. In contrast, p53-specific responses have been raised in normal mice. Because we limited our immunization approach to that of antigen-expressing tumor cells, it is unclear whether this represents a significant difference from the p53 model where p53-specific responses have been generated in response to peptide, viral construct, and dendritic cell vaccinations, but have not been reported in immunizations with irradiated tumor cells (13, 17). Furthermore, the tolerant state of the HA-specific immune repertoire in the HA104 mice has not been clearly defined. In several models, including models examining immune responses to p53 and HA (in ins-HA mice), MHC peptide tetramer staining has shown that autospecific T cells that persist are of low T-cell receptor affinity (29, 30, 31). Our inability to induce HA-specific, IgG antibody responses or cytolytic responses from these mice by MT-HA tumor immunization suggests that T-cell responses to the HA protein, if they exist in these mice, are similarly likely to be of low avidity.

Tumor protection mediated by adoptively transferred cells in HA104 mice was HA specific as it protected against challenges with the MT-HA tumor but did not protect against the control tumor. Additionally, transfer of HA104 spleen and lymph node cells did not provide similar protection against the HA-expressing tumor in HA104 mice, demonstrating that the phenomenon was not a result of increased numbers of lymphocytes but rather was dependent on the specificity of the transferred lymphocytes. Thus, HA-specific T cells present in the naive repertoire of BALB/c mice could be transferred into the HA104 mice where their target antigen is a self-antigen, and upon immunization be appropriately activated to participate in tumor rejection without inducing unacceptable levels of autoimmunity. Unlike tumor protection established in transgene-negative mice, where CD8+ T cells could be sufficient, both CD4+ and CD8+ T cells were required to establish immunity in the HA104 mice. This could reflect an increased requirement for CD4+ T cell help in the expansion of the HA-specific effector cells, the transferred population being significantly smaller than that present in normal mice. It could also be anticipated that higher avidity CD8+ T cells might be tolerized by antigen expressed on normal tissues within the HA-transgenic mouse and that either breaking of this tolerance or effective activation of those CD8+ T cells with lower avidity for HA peptides might be more dependent on the presence of CD4+ T cell help. These hypotheses are not mutually exclusive and are consistent with a number of findings indicative of active cooperation between CD4+ and CD8+ T cells in the generation of effective immune responses, including the fact that CD4+ T cells are required for the survival of adoptively transferred CD8+ T cells (32), and that helper epitope-specific CD4+ T cells can synergistically increase the antitumor activity of CD8+ T cells even where the tumor is MHC class II negative (20, 21, 33).

Although we were unable to find significant HA-specific immune responses within the HA104-transgenic mice, our finding that tumor-protective HA-specific immune responses could be generated after adoptive transfer of normal peripheral T cells from BALB/c mice extends the scope of potential tumor rejection antigens to include antigens that are broadly expressed on multiple tissues. The most likely explanation for our findings is that the level of antigen expression on the tumors, being greater than that on normal tissues, allowed these cells to serve as targets for T cells that were either ignorant or tolerant of the same antigen expressed on the context of self-cells. That the tumor protection afforded by immunization associated with adoptive transfer into HA104 mice was considerably less long-lived than that occurring upon immunization of transgene-negative mice could be a reflection of the lower numbers of HA-reactive cells transferred versus those present in transgene-negative mice, but is also consistent with the presence of an autoimmune expansion in adoptively transferred mice that was enhanced by the additional presence of immunogen supplied by the immunizing tumor but was nonetheless followed by a return to an antigen-tolerant state. Supporting this, Rocha and von-Boehmer (34) found that adoptive transfer of antigen-specific T cells into mice endogenously expressing that antigen on antigen-presenting cells resulted in an initial expansion of the autoreactive cells followed by a subsequent decline of autoreactive cell numbers and ultimately reestablishment of the tolerant state. In the examination of HA-specific responses, Adler et al.(35, 36) found that HA-specific CD4+ T cells transferred into mice with broad expression of the HA antigen on nonlymphoid cells became activated and proliferated, but ultimately were rendered tolerant by cross-presentation of the antigen on autologous bone marrow-derived cells that did not themselves express HA. Similarly, HA-specific CD8 cells have been shown to undergo initial proliferation and subsequent tolerance induction when transferred to mice with β-cell expression of HA (37).

Because the MT901 tumor is weakly immunogenic and our model required tumor immunization to see significant protective effects, we cannot rule out that other tumor-specific immune responses that in themselves were insufficient to mediate tumor rejection were nonetheless required to cause rejection in the transgenic mice. In such a situation, the additive effects of responses against the HA antigen and other tumor antigens could cause rejection. Alternatively, inflammation as a result of other antitumor responses might have been required to target the HA-specific response against the tumor. It is quite possible that in the absence of such responses, the HA antigen on the tumor in the HA104 mice might have been ignored or even tolerizing to HA-specific T cells. In this report, we have shown that even in the absence of notable responses to a model self-antigen that is overexpressed on a tumor cell line, such responses could be transferred from mice in which the antigen was not perceived as self and function in the short term to selectively protect against subsequent tumor challenges without inducing autoimmunity. These results suggest that differential expression of self-antigens by normal tissues and tumor tissues may serve to allow therapeutic responses to be directed against cancerous cells with acceptable levels of autoimmune destruction of normal tissues.

Fig. 1.

HA104 mice are not protected against MT-HA tumor challenges by immunization with MT-HA tumor cells or influenza virus. HA104 mice or transgene-negative littermates were immunized either by s.c. injection (left flank) with 106 irradiated (5000 rad) MT-HA cells, intranasal inoculation with 20 HAU of influenza virus, or by intranasal influenza inoculation (20 HAU) coupled with injection of 106 irradiated (5000 rad) MT-CT cells. Mice were challenged 14 days later with 106 viable MT-HA cells and followed for tumor cell growth by palpation of a visible lump near the site of inoculation. Number of mice examined for each data point is shown in parentheses.

Fig. 1.

HA104 mice are not protected against MT-HA tumor challenges by immunization with MT-HA tumor cells or influenza virus. HA104 mice or transgene-negative littermates were immunized either by s.c. injection (left flank) with 106 irradiated (5000 rad) MT-HA cells, intranasal inoculation with 20 HAU of influenza virus, or by intranasal influenza inoculation (20 HAU) coupled with injection of 106 irradiated (5000 rad) MT-CT cells. Mice were challenged 14 days later with 106 viable MT-HA cells and followed for tumor cell growth by palpation of a visible lump near the site of inoculation. Number of mice examined for each data point is shown in parentheses.

Close modal
Fig. 2.

HA104 mice exposed to MT-HA tumor cells do not develop HA-specific IgG responses. On day 0, HA104 or BALB/c mice, respectively, were given a s.c. (left flank) injection of 106 viable MT-CT (A and D), MT-HA (B and E), or 106 irradiated (5000 rad) MT-HA (C and F) cells in 100 ml of PBS. C and F, mice were challenged with 106 viable MT-HA cells to the opposite flank 14 days after immunization. Serum was obtained by retro-orbital bleeding on the days indicated (all mice were not bled on all days), frozen, and subsequently developed for HA-specific IgG antibody titers by ELISA as described in “Materials and Methods.” Y axis, 1/dilution at which serum samples gave a response similar to 30 ng/ml of H36-4-5, HA-specific mAb (at least twice background). Points in gray shaded area indicate lack of detectable antibody at 1/100 serum dilution (the most concentrated dilution tested). Each line represents results with an individual mouse.

Fig. 2.

HA104 mice exposed to MT-HA tumor cells do not develop HA-specific IgG responses. On day 0, HA104 or BALB/c mice, respectively, were given a s.c. (left flank) injection of 106 viable MT-CT (A and D), MT-HA (B and E), or 106 irradiated (5000 rad) MT-HA (C and F) cells in 100 ml of PBS. C and F, mice were challenged with 106 viable MT-HA cells to the opposite flank 14 days after immunization. Serum was obtained by retro-orbital bleeding on the days indicated (all mice were not bled on all days), frozen, and subsequently developed for HA-specific IgG antibody titers by ELISA as described in “Materials and Methods.” Y axis, 1/dilution at which serum samples gave a response similar to 30 ng/ml of H36-4-5, HA-specific mAb (at least twice background). Points in gray shaded area indicate lack of detectable antibody at 1/100 serum dilution (the most concentrated dilution tested). Each line represents results with an individual mouse.

Close modal
Fig. 3.

Anti-HA antibody infusion does not affect progression of HA tumor growth in BALB/c mice. A, BALB/c were treated with serum collected from MT-HA tumor-treated mice (as described in “Materials and Methods,” 200 μl of 50% serum administered by i.p. injection on days −2, 0, and +2) or HA-specific mAb H36-4-5 (100 μg in 100 μl administered by i.v. injection on days −2 and 0). On day 0, all mice were given a s.c. 5 × 104 MT-HA tumor cell challenge and followed for tumor onset and growth. Control mice received normal mouse serum (□) or mouse IgG fraction (○) given in like fashion. B, BALB/c mice received 106 irradiated (5000 rad) MT-WT cells on day −14 and were subsequently treated with H36-4-5 or control mAb on days −2 and 0 as described above. On day 0, mice were challenged with 106 MT-HA tumor cells and followed for tumor onset as before. n, number of mice evaluated for each data point.

Fig. 3.

Anti-HA antibody infusion does not affect progression of HA tumor growth in BALB/c mice. A, BALB/c were treated with serum collected from MT-HA tumor-treated mice (as described in “Materials and Methods,” 200 μl of 50% serum administered by i.p. injection on days −2, 0, and +2) or HA-specific mAb H36-4-5 (100 μg in 100 μl administered by i.v. injection on days −2 and 0). On day 0, all mice were given a s.c. 5 × 104 MT-HA tumor cell challenge and followed for tumor onset and growth. Control mice received normal mouse serum (□) or mouse IgG fraction (○) given in like fashion. B, BALB/c mice received 106 irradiated (5000 rad) MT-WT cells on day −14 and were subsequently treated with H36-4-5 or control mAb on days −2 and 0 as described above. On day 0, mice were challenged with 106 MT-HA tumor cells and followed for tumor onset as before. n, number of mice evaluated for each data point.

Close modal
Fig. 4.

Adoptive transfer of naive BALB/c cells renders HA104 mice HA-specific immunization competent against the MT-HA tumor. HA104 mice received spleen plus lymph node cells harvested from untreated BALB/c or HA104 mice or received no transfer as indicated in the figure legends. On the same day (A and B) or 7 (C) or 26 days later (D), mice were immunized with 106 irradiated (5000 rad) MT-HA cells by s.c. inoculation as described in the legend to Fig. 1. Fourteen (A, C, and D) or 21 days (B) after immunization, mice were challenged with 106 MT-HA cells and followed for tumor progression. A and C, pooled data from five and two experiments, respectively, and B and D, data from individual experiments. Number of mice examined for each data point is shown in parentheses. BALB/c transfer treatment of HA104 mice in A was significantly more protective than that in B (P = 0.045), C (P = 0.05), and D (P = 0.0037) by χ2 analysis.

Fig. 4.

Adoptive transfer of naive BALB/c cells renders HA104 mice HA-specific immunization competent against the MT-HA tumor. HA104 mice received spleen plus lymph node cells harvested from untreated BALB/c or HA104 mice or received no transfer as indicated in the figure legends. On the same day (A and B) or 7 (C) or 26 days later (D), mice were immunized with 106 irradiated (5000 rad) MT-HA cells by s.c. inoculation as described in the legend to Fig. 1. Fourteen (A, C, and D) or 21 days (B) after immunization, mice were challenged with 106 MT-HA cells and followed for tumor progression. A and C, pooled data from five and two experiments, respectively, and B and D, data from individual experiments. Number of mice examined for each data point is shown in parentheses. BALB/c transfer treatment of HA104 mice in A was significantly more protective than that in B (P = 0.045), C (P = 0.05), and D (P = 0.0037) by χ2 analysis.

Close modal
Fig. 5.

CD4+ cells are required to generate protective immunity against the MT-HA tumor in HA104 mice. Spleen and lymph node cells were harvested from untreated BALB/c mice and suspended in PBS. A, CD4-expressing cells were removed by treatment with anti-CD4 mAb followed by immunomagnetic sorting as described in “Materials and Methods.” The resulting CD4-depleted cells were delivered (3.3 × 107 cells) by retro-orbital injection into HA104 mice. HA104 mice receiving whole-pooled spleen and lymph node cells or no cell transfers served as controls. Later the same day, mice were immunized with 106 irradiated (5000 rad) MT-HA cells administered s.c. as described in the legend to Fig. 1. Mice were challenged 14 days later with 106 viable MT-HA cells and followed for tumor progression. B, mice were adoptively transferred as described above using nondepleted spleen and lymph node cells from untreated BALB/c mice. These mice were subsequently treated with anti-CD4 mAb for 3 days beginning 12 days after immunization to deplete CD4 T cells as described in “Materials and Methods.” On the third day of antibody treatment, mice were challenged with 106 viable MT-HA cells as in A and B and followed for tumor progression. Number of mice examined for each data point is shown in parentheses.

Fig. 5.

CD4+ cells are required to generate protective immunity against the MT-HA tumor in HA104 mice. Spleen and lymph node cells were harvested from untreated BALB/c mice and suspended in PBS. A, CD4-expressing cells were removed by treatment with anti-CD4 mAb followed by immunomagnetic sorting as described in “Materials and Methods.” The resulting CD4-depleted cells were delivered (3.3 × 107 cells) by retro-orbital injection into HA104 mice. HA104 mice receiving whole-pooled spleen and lymph node cells or no cell transfers served as controls. Later the same day, mice were immunized with 106 irradiated (5000 rad) MT-HA cells administered s.c. as described in the legend to Fig. 1. Mice were challenged 14 days later with 106 viable MT-HA cells and followed for tumor progression. B, mice were adoptively transferred as described above using nondepleted spleen and lymph node cells from untreated BALB/c mice. These mice were subsequently treated with anti-CD4 mAb for 3 days beginning 12 days after immunization to deplete CD4 T cells as described in “Materials and Methods.” On the third day of antibody treatment, mice were challenged with 106 viable MT-HA cells as in A and B and followed for tumor progression. Number of mice examined for each data point is shown in parentheses.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

Supported by grants (AI37691 and AI41521) from the NIH. L. A. T. is an Established Investigator of the American Heart Association.

3

The abbreviations used are: HA, influenza hemagglutinin; ins-HA, insulin promoter HA; mAb, monoclonal antibody; HAU, hemagglutinin units.

Table 1

HA104 mice immunized with MT-HA tumor cells do not develop HA-specific cytotoxic responses

Cytotoxic effector cells were generated as described in “Materials and Methods.” Briefly, individual BALB/c and HA104 mice were immunized with irradiated MT-HA tumor cells, or influenza virus as indicated, and challenged 14 days later by s.c. inoculation with viable MT-HA cells to the opposite flank (except for influenza-infected mice that were not challenged). Five days after challenge, mice were sacrificed and spleen and lymph node cells were cultured in vitro with influenza-infected spleen cells. After in vitro priming, effector cells were tested for cytotoxicity against 51Cr-labeled CT26 or P815 target cells. For CT26 targets, cells were either influenza infected to serve as controls (CT26-flu) or mock treated to serve as HA nonexpressors (CT26-mock). For P815 targets, cells were either pulsed with the H-2Kd binding peptide corresponding to positions 518–526 from HA or mock peptide pulsed. Results are expressed as percent specific lysis for the mean of two duplicate wells using effector cells of individual mice and are representative of two experiments.
Strain/immunizationTarget cells
P815 mockP815-HA peptideCT26-mockCT26-flu
 100:1a (%) 100:1 (%) 50:1 (%) 100:1 (%) 50:1 (%) 100:1 (%) 
HA104/MT-HA 15 20 18 14 
 18 28 20 12 
BALB/c/MT-HA 18 42 24 ndb 64 nd 
BALB/c/Influenza 12 65 27 nd 82 nd 
Cytotoxic effector cells were generated as described in “Materials and Methods.” Briefly, individual BALB/c and HA104 mice were immunized with irradiated MT-HA tumor cells, or influenza virus as indicated, and challenged 14 days later by s.c. inoculation with viable MT-HA cells to the opposite flank (except for influenza-infected mice that were not challenged). Five days after challenge, mice were sacrificed and spleen and lymph node cells were cultured in vitro with influenza-infected spleen cells. After in vitro priming, effector cells were tested for cytotoxicity against 51Cr-labeled CT26 or P815 target cells. For CT26 targets, cells were either influenza infected to serve as controls (CT26-flu) or mock treated to serve as HA nonexpressors (CT26-mock). For P815 targets, cells were either pulsed with the H-2Kd binding peptide corresponding to positions 518–526 from HA or mock peptide pulsed. Results are expressed as percent specific lysis for the mean of two duplicate wells using effector cells of individual mice and are representative of two experiments.
Strain/immunizationTarget cells
P815 mockP815-HA peptideCT26-mockCT26-flu
 100:1a (%) 100:1 (%) 50:1 (%) 100:1 (%) 50:1 (%) 100:1 (%) 
HA104/MT-HA 15 20 18 14 
 18 28 20 12 
BALB/c/MT-HA 18 42 24 ndb 64 nd 
BALB/c/Influenza 12 65 27 nd 82 nd 
a

E:T ratios.

b

nd; no data.

1
Houghton A. N. Cancer antigens: immune recognition of self and altered self.
J. Exp. Med.
,
180
:
1
-4,  
1994
.
2
Rosenberg S. A. A new era for cancer immunotherapy based on the genes that encode cancer antigens.
Immunity
,
10
:
281
-287,  
1999
.
3
Brandle D., Bilsborough J., Rulicke T., Uyttenhove C., Boon T., Van den Eynde B. J. The shared tumor-specific antigen encoded by mouse gene P1A is a target not only for cytolytic T lymphocytes but also for tumor rejection.
Eur. J. Immunol.
,
28
:
4010
-4019,  
1998
.
4
Wang R. F., Appella E., Kawakami Y., Kang X., Rosenberg S. A. Identification of TRP-2 as a human tumor antigen recognized by cytotoxic T lymphocytes.
J. Exp. Med.
,
184
:
2207
-2216,  
1996
.
5
Peoples G. E., Goedegebuure P. S., Smith R., Linehan D. C., Yoshino I., Eberlein T. J. Breast and ovarian cancer-specific cytotoxic T lymphocytes recognize the same HER2/neu-derived peptide.
Proc. Natl. Acad. Sci. USA
,
92
:
432
-436,  
1995
.
6
Morgan D. J., Kreuwel H. T., Fleck S., Levitsky H. I., Pardoll D. M., Sherman L. A. Activation of low avidity CTL specific for a self epitope results in tumor rejection but not autoimmunity.
J. Immunol.
,
160
:
643
-651,  
1998
.
7
Naftzger C., Takechi Y., Kohda H., Hara I., Vijayasaradhi S., Houghton A. N. Immune response to a differentiation antigen induced by altered antigen: a study of tumor rejection and autoimmunity.
Proc. Natl. Acad. Sci. USA
,
93
:
14809
-14814,  
1996
.
8
Overwijk W. W., Tsung A., Irvine K. R., Parkhurst M. R., Goletz T. J., Tsung K., Carroll M. W., Liu C., Moss B., Rosenberg S. A., Restifo N. P. gp100/pmel 17 is a murine tumor rejection antigen: induction of “self”-reactive, tumoricidal T cells using high-affinity, altered peptide ligand.
J. Exp. Med.
,
188
:
277
-286,  
1998
.
9
Overwijk W. W., Lee D. S., Surman D. R., Irvine K. R., Touloukian C. E., Chan C. C., Carroll M. W., Moss B., Rosenberg S. A., Restifo N. P. Vaccination with a recombinant vaccinia virus encoding a “self” antigen induces autoimmune vitiligo and tumor cell destruction in mice: requirement for CD4+ T lymphocytes.
Proc. Natl. Acad. Sci. USA
,
96
:
2982
-2987,  
1999
.
10
Bowne W. B., Srinivasan R., Wolchok J. D., Hawkins W. G., Blachere N. E., Dyall R., Lewis J. J., Houghton A. N. Coupling and uncoupling of tumor immunity and autoimmunity.
J. Exp. Med.
,
190
:
1717
-1722,  
1999
.
11
Ludewig B., Ochsenbein A. F., Odermatt B., Paulin D., Hengartner H., Zinkernagel R. M. Immunotherapy with dendritic cells directed against tumor antigens shared with normal host cells results in severe autoimmune disease.
J. Exp. Med.
,
191
:
795
-804,  
2000
.
12
Speiser D. E., Miranda R., Zakarian A., Bachmann M. F., McKall-Faienza K., Odermatt B., Hanahan D., Zinkernagel R. M., Ohashi P. S. Self antigens expressed by solid tumors do not efficiently stimulate naive or activated T cells: implications for immunotherapy.
J. Exp. Med.
,
186
:
645
-653,  
1997
.
13
Offringa R., Vierboom M. P., van der Burg S. H., Erdile L., Melief C. J. p53: a potential target antigen for immunotherapy of cancer.
Ann. NY Acad. Sci.
,
910
:
223
-233,  
2000
.
14
Theobald M., Biggs J., Dittmer D., Levine A. J., Sherman L. A. Targeting p53 as a general tumor antigen.
Proc. Natl. Acad. Sci. USA
,
92
:
11993
-11997,  
1995
.
15
Ropke M., Hald J., Guldberg P., Zeuthen J., Norgaard L., Fugger L., Svejgaard A., Van der Burg S., Nijman H. W., Melief C. J., Claesson M. H. Spontaneous human squamous cell carcinomas are killed by a human cytotoxic T lymphocyte clone recognizing a wild-type p53-derived peptide.
Proc. Natl. Acad. Sci. USA
,
93
:
14704
-14707,  
1996
.
16
Vierboom M. P., Nijman H. W., Offringa R., van der Voort E. I., van Hall T., van den Broek L., Fleuren G. J., Kenemans P., Kast W. M., Melief C. J. Tumor eradication by wild-type p53-specific cytotoxic T lymphocytes.
J. Exp. Med.
,
186
:
695
-704,  
1997
.
17
Ishida T., Chada S., Stipanov M., Nadaf S., Ciernik F. I., Gabrilovich D. I., Carbone D. P. Dendritic cells transduced with wild-type p53 gene elicit potent anti-tumour immune responses.
Clin. Exp. Immunol.
,
117
:
244
-251,  
1999
.
18
Lo D., Freedman J., Hesse S., Palmiter R. D., Brinster R. L., Sherman L. A. Peripheral tolerance to an islet cell-specific hemagglutinin transgene affects both CD4+ and CD8+ T cells.
Eur. J. Immunol.
,
22
:
1013
-1022,  
1992
.
19
Staveley-O’Carroll K., Sotomayor E., Montgomery J., Borrello I., Hwang L., Fein S., Pardoll D., Levitsky H. Induction of antigen-specific T cell anergy: an early event in the course of tumor progression.
Proc. Natl. Acad. Sci. USA
,
95
:
1178
-1183,  
1998
.
20
Marzo A. L., Lake R. A., Robinson B. W., Scott B. T-cell receptor transgenic analysis of tumor-specific CD8 and CD4 responses in the eradication of solid tumors.
Cancer Res.
,
59
:
1071
-1079,  
1999
.
21
Marzo A. L., Kinnear B. F., Lake R. A., Frelinger J. J., Collins E. J., Robinson B. W., Scott B. Tumor-specific CD4+ T cells have a major “post-licensing” role in CTL-mediated antitumor immunity.
J. Immunol.
,
165
:
6047
-6055,  
2000
.
22
Eck S. C., Turka L. A. Generation of protective immunity against an immunogenic carcinoma requires CD40/CD40L and B7/CD28 interactions but not CD4+ T cells.
Cancer Immunol. Immunother.
,
48
:
336
-341,  
1999
.
23
Caton A. J., Stark S. E., Shih F. F., Cerasoli D. M. Transgenic mice that express different forms of the influenza virus hemagglutinin as a neo-self-antigen.
J. Clin. Immunol.
,
15
:
106S
-112S,  
1995
.
24
Staudt L. M., Gerhard W. Generation of antibody diversity in the immune response of BALB/c mice to influenza virus hemagglutinin. I. Significant variation in repertoire expression between individual mice.
J. Exp. Med.
,
157
:
687
-704,  
1983
.
25
Aruga E., Aruga A., Arca M. J., Lee W. M., Yang N. S., Smith J. W., II, Chang A. E. Immune responsiveness to a murine mammary carcinoma modified to express B7-1, interleukin-12, or GM-CSF.
Cancer Gene Ther.
,
4
:
157
-166,  
1997
.
26
Groth F. d. S., Webster R. G. Disquisitions of original antigenic sin. I. Evidence in man.
J. Exp. Med.
,
124
:
331
-345,  
1966
.
27
Caton A. J., Swartzentruber J. R., Kuhl A. L., Carding S. R., Stark S. E. Activation and negative selection of functionally distinct subsets of antibody-secreting cells by influenza hemagglutinin as a viral and a neo-self antigen.
J. Exp. Med.
,
183
:
13
-26,  
1996
.
28
Matzinger P. An innate sense of danger.
Semin. Immunol.
,
10
:
399
-415,  
1998
.
29
Nugent C. T., Morgan D. J., Biggs J. A., Ko A., Pilip I. M., Pamer E. G., Sherman L. A. Characterization of CD8+ T lymphocytes that persist after peripheral tolerance to a self antigen expressed in the pancreas.
J. Immunol.
,
164
:
191
-200,  
2000
.
30
Hernandez J., Lee P. P., Davis M. M., Sherman L. A. The use of HLA A2.1/p53 peptide tetramers to visualize the impact of self tolerance on the TCR repertoire.
J. Immunol.
,
164
:
596
-602,  
2000
.
31
Targoni O. S., Lehmann P. V. Endogenous myelin basic protein inactivates the high avidity T cell repertoire.
J. Exp. Med.
,
187
:
2055
-2063,  
1998
.
32
Walter E. A., Greenberg P. D., Gilbert M. J., Finch R. J., Watanabe K. S., Thomas E. D., Riddell S. R. Reconstitution of cellular immunity against cytomegalovirus in recipients of allogeneic bone marrow by transfer of T-cell clones from the donor.
N. Engl. J. Med.
,
333
:
1038
-1044,  
1995
.
33
Ossendorp F., Mengede E., Camps M., Filius R., Melief C. J. Specific T helper cell requirement for optimal induction of cytotoxic T lymphocytes against major histocompatibility complex class II negative tumors.
J. Exp. Med.
,
187
:
693
-702,  
1998
.
34
Rocha B., von Boehmer H. Peripheral selection of the T cell repertoire.
Science (Wash DC).
,
251
:
1225
-1228,  
1991
.
35
Adler A. J., Marsh D. W., Yochum G. S., Guzzo J. L., Nigam A., Nelson W. G., Pardoll D. M. CD4+ T cell tolerance to parenchymal self-antigens requires presentation by bone marrow-derived antigen-presenting cells.
J. Exp. Med.
,
187
:
1555
-1564,  
1998
.
36
Adler A. J., Huang C. T., Yochum G. S., Marsh D. W., Pardoll D. M. In vivo CD4+ T cell tolerance induction versus priming is independent of the rate and number of cell divisions.
J. Immunol.
,
164
:
649
-655,  
2000
.
37
Morgan D. J., Kreuwel H. T., Sherman L. A. Antigen concentration and precursor frequency determine the rate of CD8+ T cell tolerance to peripherally expressed antigens.
J. Immunol.
,
163
:
723
-727,  
1999
.