Frequencies of Tetramer⁺ T Cells Specific for the Wild-Type Sequence p53_264–272 Peptide in the Circulation of Patients with Head and Neck Cancer¹ Free

The gene encoding p53 protein is frequently mutated in many human cancers, including SCCHN,³ which generally results in accumulation (overexpression) of p53 molecules in these tumors (1, 2, 3, 4). As most of these mutations involve an alteration of a single amino acid in p53 molecules, the majority of the accumulating mutant protein resembles the wt p53 (4). Therefore, enhanced presentation of wt epitopes derived from p53 accumulating in tumors is possible, and might lead to their recognition by the immune system and the development of antitumor CTLs (5, 6, 7, 8, 9). For this reason, wt p53 epitopes are considered attractive targets for immunotherapy of cancer.

We reported previously on the generation of CTLs recognizing the HLA-A2.1-restricted wt p53_264–272 epitope from PBMCs obtained from SCCHN patients using autologous peptide-pulsed dendritic cells as antigen-presenting cells (9). Surprisingly, we observed that CTLs reactive against this wt p53 epitope could only be generated from T-cell precursors in PBMCs of patients whose tumors either did not accumulate p53 or accumulated mutant p53 molecules that could not present this epitope (9). We hypothesized that in vivo, the presence of expandable CTL precursors specific for the wt p53_264–272 epitope led to immunoselection, resulting in the elimination of tumors expressing the epitope and favored the outgrowth of “epitope-loss” tumor cells able to evade the host immune system. On the other hand, it was also possible that HPV infection, known to occur in a substantial proportion of SCCHN, could lead to inactivation of wt p53 and enhanced processing of p53 epitopes (10). The consequence of either phenomenon would be the presence in patients with wt p53 tumors of relatively high frequencies of T cells specific for the wt p53_264–272 epitope. To test these hypotheses, we investigated the frequency of p53_264–272-specific precursor T cells in the peripheral circulation of 30 HLA-A2.1⁺ patients with SCCHN and 31 HLA-A2.1⁺ healthy controls, using multimeric peptide-MHC complexes. We also performed PCR analyses of genomic DNA isolated from the patient tumors for p53 and HPV E6/E7.

The availability of multimeric peptide-MHC complexes, which are generically referred to as tetramers, allows for direct identification and phenotyping of antigen-specific T cells in the peripheral circulation. Tetramers bind to more than one TCR on a specific T cell and, therefore, have a relatively slow dissociation rate (11, 12). However, the specificity of tetramer binding to the TCR has to be carefully controlled, particularly when the frequency of peptide-specific precursor T cells in the peripheral circulation is expected to be very low. Anticipating that discrimination between nonspecific and specific tetramer binding to p53_264–272-specific T cells might be difficult, we used a novel four-color flow cytometry assay that simultaneously measures tetramer, CD3, CD8, and CD14 binding (13). Furthermore, binding of the p53 tetramer was compared with that of the influenza virus matrix peptide GILGFVFTL (FLU, a model for recall responses) and the HIV reverse transcriptase peptide ILKEPVHGV (HIV, a model for responses to a new antigen).

In this study, the frequency of wt p53_264–272 peptide-specific CD8⁺ T cells in the circulation of patients with SCCHN was correlated with the p53 expression in each patient tumor, its HPV status, and the presence of p53 antibodies in the serum. Our results provide significant insights into the in vivo interactions that might occur between the developing tumor and immune system in these patients.

Peripheral blood samples or leukapheresis products were obtained from 30 HLA-A2.1⁺ SCCHN patients, 31 HLA-A2.1⁺ healthy donors, and 10 HLA-A2.1⁻ healthy donors. PBMCs were isolated by centrifugation over Ficoll-Hypaque gradients (Amersham Pharmacia Biotech, Piscataway, NJ). Leukapheresis products were obtained from the Institute of Transfusion Medicine, Pittsburgh, PA. The study was approved by the Institutional Review Board at the University of Pittsburgh, and written informed consent was obtained from each participating individual. All of the products were tested and found to be negative for HIV-I antigens and antibodies to HIV. PBMCs were phenotyped for expression of HLA-A2 molecules by flow cytometry using the anti-HLA-A2 mAb BB7.2 (American Type Culture Collection, Manassas, VA) and an IgG isotype as a control. The verification of the A2.1 subtype was performed using polymerase chain reaction with sequence-specific primers as described previously (9, 14). PBMCs were either used fresh or were frozen at a concentration of 50 × 10⁶ cells/ml in the freezing medium consisting of human AB serum (Pel-Freeze, Brown Deer, WI) plus 10% DMSO (Fisher Scientific, Pittsburgh, PA).

In 30 patients, histologically verified squamous cell carcinomas originated in one of the following sites: the nose (n = 1), oral cavity (n = 7), oropharynx (n = 4), larynx (n = 16), and hypopharynx (n = 2). Tumors were classified for tumor stage (T₁ = 10; T₂ = 9; T₃ = 3; T₄ = 8; and T_x = 1), nodal stage (N⁰ = 22; N₁ = 2; N₂ = 5; and N₃ = 1), and the presence of distant metastases (n = 0 of 30).

Tetramers were obtained through the National Institute of Allergy and Infectious Diseases Tetramer Facility and the NIH AIDS Research and Reference Reagent Program. Stock solutions contained 0.5 μg monomer/ml. Peptides provided to the National Institute of Allergy and Infectious Diseases Tetramer Facility were either GILGFVFTL, an influenza matrix immunodominant peptide (residues 58–66), the HIV-1 reverse transcriptase peptide (pol 476–484) ILKEPVHGV, or the HLA-A2.1-binding peptide LLGRNSFEV (15, 16), corresponding to wt p53_264–272 peptide. The specificity of the LLGRNSFEV tetramer was confirmed by staining against the anti-p53-specific CTL line as described previously (9) and by the lack of staining with irrelevant CTLs, as well as HLA-A2-negative PBMCs of healthy donors.

To minimize background staining each tetramer was titered and used at the lowest concentration that still gave a clearly discernible positive population in a donor vaccinated for influenza (for FLU_58–66 tetramer) as well as in an HIV-infected individual (for pol_476–484). The final dilution of both preparations during staining, relative to the stock reagent supplied by the NIH, was 1/300. Within a 2-fold range of tetramer concentrations bracketing the concentrations used here, the frequency of tetramer-positive events and competition of CD3 binding (13) were stable, and tetramer fluorescence intensity was within 80% of that obtained at saturating concentrations. At saturating concentrations, CD3 competition decreased, fluorescence intensity of tetramer positive cells increased, and tetramer frequency increased, the latter attributable chiefly to a greater number of tetramer-dim events.

The default panel of antibodies used for these studies was CD14-FITC (RMO52; Immunotech, Miami, FL), Tetramer-PE, CD3 ECD (HIT3a; Beckman Coulter, Miami, FL), and CD8-PC5 [SFCI21Thy2D3(T8); Beckman Coulter]. Additionally, antihuman CD45RA FITC (Immunotech) and anti-CD45RO ECD (Beckman Coulter) were used for the characterization of the CD45 phenotype.

Immediately before staining, cells were washed twice with the staining medium, consisting of PBS +0.1% (w/v) BSA +0.1% (w/v) sodium-azide, and resuspended at a concentration of 5 × 10⁶/ml in a volume of 150 μl. Tetramer (5 μl of 1:10 dilution of stock solution) was added at room temperature for 30 min, followed by a 30 min incubation with antibodies (7.5 μL of each) at 4°C. After two additional washes, the cells were resuspended in ∼1 ml of 0.5% methanol-free formaldehyde in PBS. At least 1 × 10⁶ events were collected using a four-color Coulter Epics XL cytometer set on low or medium flow rate at a maximum of 1000 events/sec. Flow cytometry data were analyzed in real time using Beckman-Coulter System II software. In initial experiments, the region defining tetramer-positive events was determined by evaluating PBMCs stained with the Ab panel but without tetramer. This region was held constant throughout the analysis. Data were saved as FCS 2.0 Listmode files for subsequent reanalysis in System II or WinList (Verity Software House, Topsham, ME).

A wt p53_264–272-specific CTL line (9) was used as a positive control for the flow cytometric evaluation and additionally to visualize tetramer binding to the specific T cells by confocal microscopy. The CTL line was incubated with the p53_264–272 peptide containing tetramer in azide-free PBS followed by incubation with FITC-conjugated anti-CD8 Ab. The cells were then fixed with 1% (w/v) paraformaldehyde, placed on a slide, and analyzed by confocal laser scanning microscopy at ×600 original magnification (Leica TCS NT confocal LSM; Leica Lasertechnik, Heidelberg, Germany). Images were edited using the Adobe PhotoShop software program (Adobe Systems, Mountain View, CA).

Tumors of 30 SCCHN patients included in this study were available as paraffin blocks archived at the University of Pittsburgh Medical Center. The histology of each case was reviewed by a pathologist (S. D. F.), and representative tissue sections containing areas of invasive SCCHN were selected for microdissection. Normal-appearing salivary gland tissue or skeletal muscle was microdissected separately to serve as an internal nontumor control. Using 4-μm thick recut, unstained histological sections, normal and malignant tissue samples were removed under stereomicroscopic observation. Sufficient material was collected from a single histological section to afford replicate analysis. Samples were treated with proteinase K at a final concentration of 100 μg/ml for 2 h and then boiled for 5 min to remove protease activity. PCR used sets of amplification primers flanking exons 5–8 of the p53 gene in four separate PCRs (17). Amplified DNA from microdissected tissues also included splice sites. PCR products were electrophoresed in 4% agarose, and the ethidium bromide-stained bands were excised and then isolated with glassmilk. DNA sequencing used antisense PCR primers for each exon with [³³P]dATP as the reporter molecule, and sequence analysis was read from overnight exposed autoradiograms of 6% polyacrylamide gels.

For p53 immunohistochemistry, formalin-fixed, paraffin-embedded tumor tissues were sectioned (3–5 μm), air-dried overnight at 37°C, deparaffinized, and dehydrated and stained with a mAb against p53, D0–7 (Dako, Carpinteria, CA), which recognizes an epitope in the NH₂ terminus between amino acids 35–45, and reacts with wild-type and most mutant forms of the p53 protein. The avidin-biotin-peroxidase method was used to visualize the p53, according to the instructions supplied by the manufacturer (Dako). The immunostained slides were evaluated by light microscopy for p53 accumulation. The tumor was considered p53-positive when >25% of the tumor cells showed staining intensity of 2+ and higher on the scale of 0–4+. IgG isotype mAb used at the same concentration as the primary mAb served as a negative control.

Ab to p53 in the patient and control sera were detected by an Enzyme ELISA purchased from PharmaCell Immunotech Coulter, Miami, FL, using microtiter plates coated with recombinant human wt p53 protein. Peroxidase-conjugated goat antihuman IgG was used for detection of human anti-p53 Ab by a colorimetric reaction. Staining intensity was compared with a standard curve, and anti-p53 levels ≥1.1 units/ml were considered to be positive. Assays were performed twice in triplicates and included sera obtained from seropositive as well as seronegative individuals as internal positive/negative controls.

PCR analysis was performed using amplification primers for HPV-16 E6/E7 (ATGCACCAAAAGAGAACTGC and TGCCCATTAACAGGTCTTCC) and β-actin (GCGAGAAGATGACCCAG and GCCTGGATAGCAACGTA) as control, using tumor DNA isolated as described above. DNA aliquots obtained from 25 of 30 specimens were screened in four separate PCR reactions. A solution (50 μl) consisting of 25 mm MgCl₂, 1.5 μm of each primer, 1.25U of Taq DNA-polymerase (Promega), 2 mm of deoxynucleotide triphosphate, and double-distilled H₂O was added to each amplification tube. Amplification was performed with denaturation at 94°C for 1 min, annealing at 57°C for 1 min, and extension at 72°C for 2 min. The process was repeated for the total of 40 cycles. In all of the PCR reactions, DNA obtained from HPV-16 E6/E7+ Caski and C33 cell lines were included as positive controls. PCR products (700-bp for HPV-16 E6/E7 and 500-bp for β-actin) were electrophoresed in 4% agarose and the bands visualized in the presence of ethidium bromide. Twelve of 25 tumors tested were positive for HPV-16 E6/E7.

Tetramer-positive cells were quantified by flow cytometry and expressed as frequencies (e.g., 1/1000) or reciprocal frequencies (e.g., 1000). We examined raw reciprocal frequency data and log-transformed reciprocal frequency data using normal probability plots (13). For all three tetramers, the log-transformed data were better modeled by the normal distribution. Accordingly, descriptive statistics (means, SDs, and confidence intervals) and statistical analyses (Student’s t test, two-tailed), were performed on log-transformed reciprocal frequencies. For the comparison of multiple parameters, ANOVA (Tukey-Kramer multiple comparison test) was applied. The specificity of tetramer binding to the TCR was measured by competition of CD3 binding (13). Competition of CD3 binding was expressed according to the formula:

where CD3Mnl is the log MFI of CD3 staining of the population designated by the subscript. For tetramer frequencies as well as competition of Ab binding, lower limits of detection were established by estimating the 99^th percentile (geometric mean plus 2.6 SDs) of responses measured in PBMCs from HLA-A2⁻ subjects. Linear regression was performed by the method of least squares. In regression plots the line of best fit and the 95% confidence intervals about the line of best fit are shown. The coefficient of correlation (r²), the slope of the regression line, and the P associated with the slope are reported as regression summary statistics. Multivariate analysis was performed to determine the relationship between the frequency of p53-specific CTL, accumulation of p53 in the tumor, and the HPV status of the tumor. Statistical analysis and statistical graphics were performed using Systat Version 9 (SPSS Inc, Chicago, IL).

A wt p53_264–272-specific CTL line established earlier (9) was stained with the FITC-conjugated anti-CD8 (green) and PE-conjugated tetramer (red). Its confocal microscopy mid-plane image is shown in Fig. 1. On binding to the TCR, tetrameric p53_264–272-MHC class I complexes were internalized in the absence of sodium azide. This p53_264–272-specific T-cell line served as positive control in subsequent flow cytometry analyses of precursor T cells in the peripheral circulation of human subjects.

For the detection of unstimulated precursor T cells specific for the wt p53_264–272 peptide, PBMCs of SCCHN patients and healthy donors were directly stained with the tetramer and analyzed by flow cytometry. To assure the specificity of tetramer binding, we developed previously a gating strategy to eliminate CD14⁺ monocytes as well as apoptotic and necrotic cells, all of which could bind tetramer and/or Ab nonspecifically (13). Furthermore, CD3-negative cells were eliminated by compound gating, which finally resulted in discriminative dot plots showing CD8⁺ tetramer⁺ T cells. Fig. 2 shows the representative dot plots of stained PBMCs obtained from a healthy individual and from 2 patients with SCCHN. The cases shown are representative of relatively low (Fig. 2, left; 1 of 5757), intermediate (Fig. 2, middle; 1 of 3063) and high (Fig. 2, right; 1 of 1140) p53_264–272 tetramer binding frequencies, respectively.

To establish the lower detection limit for tetramer binding in HLA-A2.1⁺ individuals, we stained PBMCs obtained from 10 HLA-A2⁻ individuals. Despite the application of the gating strategy described above, low levels of p53_264–272 tetramer⁺ CD3⁺ CD8⁺ T cells were detected in PBMCs of HLA-A2⁻ donors (geometric mean = 1/23,397). Because these events were nonspecific by definition, we established a cutoff for the lower detection limit of this assay at the upper 99^th percentile of tetramer⁺ CD8⁺ T cells in HLA-A2⁻ individuals. This cutoff frequency of 1 of 7,805 was applied to all of the data obtained from testing of the 10 HLA-A2⁻ subjects, 30 HLA-A2.1⁺ SCCHN patients, and 31 HLA-A2.1⁺ healthy controls. As shown in Fig. 3, none of the HLA-A2⁻ individuals had frequencies of p53_264–272 tetramer⁺ CD8⁺ T cells exceeding the established cutoff. On the other hand, 23 of 30 HLA-A2.1⁺ SCCHN patients and 25 of 31 healthy controls had frequencies of p53_264–272 tetramer⁺ CD8⁺ T cells above the cutoff point (geometric means = 1/3,533 and 1/5,207, respectively).

We have reported previously that tetramer-positive T cells stained dimmer for CD3 than did tetramer-negative T cells in PBMCs obtained from HLA-A2.1⁺ subjects. However, this is not the case for the spurious tetramer-positive events seen in PBMCs obtained from HLA-A2⁻ subjects (13). We demonstrated that this phenomenon results from a competition between tetramer and anti-CD3 Abs binding to TCR. This competition was subsequently quantified and introduced as a marker for the specificity of tetramer binding to the TCR complex (13). Because there was no detectable competition between tetramer and anti-CD3 Ab binding for CD8⁺ T cells in 9 of 10 HLA-A2.1⁻ individuals, we were able to define a cutoff based on the 99^th percentile of CD3 competition in these HLA-A2⁻ subjects (3.2%). Competition by anti-CD3 Ab in excess of this cutoff was considered to be significant. The mean of competition for the wt p53_264–272 tetramer was 10.0 ± 1.4% (mean ± SE) in T cells obtained from patients and 5.3 ± 1.5% in normal controls. PBMCs of 23 of 30 SCCHN patients but only 10 of 31 normal controls exceeded both the cutoff for competition as well as the cutoff for frequency (see above), and, thus, were considered to exhibit specific binding of the p53_264–272 tetramer (Fig. 4).

Next, frequencies of wt p53_264–272-specific CD8⁺ T cells were compared with those obtained with the HIV tetramer or the FLU tetramer. These comparisons were performed to evaluate p53-specific responses in the context of those to known novel and recall antigenic peptides. The frequencies of FLU- and HIV-specific T cells are displayed in Fig. 5 as normal distribution curves. The frequencies of wt p53_264–272-specific T cells are shown as individual circles. For healthy individuals (Fig. 5, left), the majority of p53 frequencies fall within the HIV normal distribution curve. In contrast to PBMCs from the healthy control group, PBMCs from SCCHN patients seem to have a bimodal distribution, with a majority of the frequencies located within the normal HIV distribution curve, and six frequencies shifted to the right (Fig. 5, note the log scale) toward FLU frequencies.

On more careful examination, when the frequencies of wt p53_264–272 peptide-specific CD8⁺ T cells in the circulation of patients (see Table 1) were compared with those obtained in normal controls, it appeared that the patients could be divided into two different groups. The first subset consists of patients with wt p53 tumors that do not accumulate p53 or those with tumors unlikely to present the wt p53_264–272 epitope because of the type of mutation they harbor. These patients have significantly higher frequencies (P ≤ 0.0005) of wt p53_264–272-specific CD8⁺ T cells in the peripheral circulation than the patients in the second subset, whose tumors accumulate p53 and, in theory, could have a higher potential for presentation of this epitope (Fig. 6). The SCCHN patients in the second subset, those with tumors accumulating p53, have lower frequencies of CD8⁺ T cells specific for this p53 epitope, which do not significantly differ from those obtained for normal controls (Fig. 6). Confirming our initial observations, this result suggests that accumulation of p53 in the tumor does not positively correlate with the frequency of the epitope-specific CD8⁺ T cells detectable in the circulation of patients with SCCHN.

Expression of CD45 isoforms on the surface of T cells is routinely used to discriminate between naive (CD45RA⁺) and memory (CD45RO⁺) T-cell subsets (18). To determine whether tetramer⁺ CD8⁺ T cells detected in the circulation of patients with SCCHN belong to the memory or naive T-cell subsets, multicolor flow analysis including anti-CD45 Abs was performed. The gates for CD45RO and CD45RA expression were set on CD8⁺ tetramer⁻ T cells, as shown in the left panel of Fig. 7. We have shown previously that the majority of CD8⁺ tetramer⁺ cells for the recall antigen, FLU, were CD45RO⁺/CD45RA⁻ memory T cells, whereas those recognizing the novel HIV antigen were predominantly CD45RO⁻/CD45RA⁺ naive T cells (13). This analysis was performed on samples obtained from two groups of representative patients: one with relatively high frequencies (average ∼1/2700), the other with low frequencies (average ∼1/5500) of wt p53_264–272-specific CD8⁺ T cells as well as from normal donors (see Table 2). Normal donors (Fig. 7, middle panel) and patients with low frequencies of p53_264–272 tetramer⁺ T cells had similar percentages of memory (∼10%) and naive (∼70%) T cells in the peripheral circulation. In contrast, a significantly higher percentage of memory cells (36.5%) was found in patients with high frequencies of wt p53_264–272-specific T cells (Fig. 7, right panel).

Immunohistochemistry of p53 protein and sequencing of genomic PCR products of p53 exons 5–8 in patient tumors were done to evaluate the potential of these tumors to present the wt p53_264–272 epitope, and ultimately relate this information to the frequencies of tetramer⁺ CD8⁺ T cells detected in the peripheral circulation of these patients (Table 1). Although exceptions have been noted in the literature, most tumor cells lines sensitive to CTL recognizing this epitope have been shown to accumulate p53. Of the 30 primary tumors analyzed, more then half (17 of 30) showed accumulation of p53. The primary tumor of one patient, #28, scored negative, but a lymph node metastasis was positive. From this total of 31 tumors (18 of 31 with p53 accumulation), 28 underwent sequencing of genomic PCR products of exons 5–8 (3 cases were not available). Of the 16 available tumors (2 of 18 not available) showing p53 accumulation, 13 were found to express missense mutations within p53 exons 5–8. In two tumors (patients #7 and #10), p53 missense mutations were located within or directly next to the wt p53_264–272 peptide sequence. Such a mutation (R273H) was shown previously to prevent presentation of the epitope (19). For subsequent analysis, these tumors were considered as unlikely to be presenting the wt p53_264–272 epitope, and they are designated by brackets for p53 accumulation in Table 1.

A lack of agreement between the mutated p53 exon 5–8 genotype and p53 accumulation was reported for patients with SCCHN (20). In general, the results of our study indicated a good correlation between p53 missense mutations and accumulation of the altered protein in tumors (Table 1). In 3 cases (patients #19, #20, and #27) a discrepancy between the p53 genotype and p53 expression was observed (Table 1). All three represented tumors with a wt p53 genotype and with p53 accumulation, which could possibly be because of a mutation outside the exons 5–8 or in genes impacting on the stability of p53, such as mdm2 (21).

Stratification of the frequencies of wt p53_264–272-specific T cells detected in the circulation of patients with SCCHN identified three groups of patients, those displaying high, intermediate, or low frequencies of these cells (Table 1). The cutoff for the high frequency designation was based on the upper 95^th percentile (>1/2128) frequency of tetramer⁺CD8⁺ T cells established in HLA-A2⁺ healthy controls. The tumors obtained from the 6 patients with the highest frequencies of these T cells were wt p53 and did not accumulate p53 (Table 1; Fig. 8).

In the second patient group, the intermediate reciprocal frequencies of wt p53_264–272-specific CD8⁺ T cells were lower than 1/2128, yet they exceeded the geometric mean of the patient group as a whole (1/4767). This group of 7 patients was heterogeneous in respect to p53 expression in the tumor: 3 tumors were p53 wt, 2 had p53 mutations and accumulation, and 2 tumors (#7 and #10) expressed missense mutations within or next to the p53_264–272 epitope (positions 273 and 271, respectively) and were probably unable to present this epitope to T cells (Fig. 8).

The third and largest group of the patients (n = 17) had the lowest frequencies of wt p53_264–272-specific T cells. In this group, 14 of 17 primary tumors (>80%) accumulated p53 and had the potential to present this epitope. The frequencies of wt p53_264–272-specific T cells exceeded the lower limit of detection (1/7805) in 7 patients. The mean frequency for the group was significantly lower than that for the other two groups of patients. As indicated in Table 2, wt p53_264–272-specific CD8⁺ T cells present in low frequencies in the peripheral circulation of these patients predominantly expressed a naive phenotype (CD45RA⁺/CD45RO⁻). On the other hand, in patients with high frequencies of wt p53_264–272 epitope-specific T cells, memory T cells (CD45RA⁻/CD45RO⁺) significantly increased in proportion (Fig. 7). Therefore, it would appear that in the peripheral circulation of patients whose tumors have a low potential for presenting the epitope, the frequency of wt p53_264–272-specific T cells with a memory phenotype is high. Therefore, it is likely that these T cells had a previous interaction with targets capable of presenting the p53_264–272 epitope.

PCR analysis indicated that 12 of 26 (46%) tumors we examined contained HPV-16 E6/E7 DNA (Table 1). Among 12 tumors with p53 mutations, 4 (33%) were E6/E7⁺, whereas 8 of 12 (66%) wt p53 tumors were E6/E7⁺. One tumor (#28) was heterogeneous, with cells expressing either wt or mutated p53, and it was also E6/E7⁺. The p53 genotype was not available for 3 tumors analyzed for HPV. We found that 3 of 5 patients with the highest frequencies of wt p53_264–272-specific T cells and wt p53 in the tumor were HPV-16⁺. The patient group with the intermediate frequencies (Fig. 8), contained 7 patients of whom 2 could not be tested for HPV and 1 was not genotyped for p53. Of the remaining 4, 2 were wt p53 and HPV⁺, whereas the other 2 had mutated p53 and were HPV negative. The cohort of 17 patients with low CTL frequencies contained 13 informative cases (tumor #28 was excluded from analysis), with 4 of 9 mutated p53 tumors and 2 of 4 wt p53 tumors positive for HPV-16. By fitting loglinear models to the frequencies of each variable in the 3 × 2 × 2 contingency table, we determined that the frequency of p53_264–272-specific CTL was significantly correlated with the p53 status of the tumor (P = 0.016). In contrast, both the frequency of CTL and p53 status of the tumor were found to be independent of HPV E6/E7 positivity (P = 0.9260 and P = 0.2924, respectively).

An analysis of IgG antibodies to p53 in the sera of SCCHN patients by Bourhis et al. (22) identified nearly 20% as seropositive. Because the presence of IgG antibodies to p53 implies a T-cell mediated response, it was of interest to determine whether the frequencies of wt p53_264–272-specific T cells present in the peripheral circulation of our seropositive patients were higher than the mean for all of the SCCHN patients. As indicated in Table 1, 3 of the 30 patients (#9, #10, and #24) scored p53 seropositive. The mean frequency (1/4538) of wt p53_264–272-specific T cells in these 3 patients was not significantly higher than the geometric mean frequency for all of the SCCHN patients (1/4767). In 2 of these patients the tumor accumulated mutant p53, whereas in the third (#9), it did not. Interestingly, the tumor of this patient contained ∼2% of cells positive for p53 (Fig. 9), and it was HPV-16 E6/E7 positive. The presence of p53 autoantibodies in the serum, which is usually associated with p53 accumulation (22), and the relatively high frequency of wt p53_264–272-specific T cells detected in this patient’s circulation (1/2746), suggest that the virus-related enhanced processing of p53 might contribute to effective CTL generation in vivo.

SCCHN, which arise at or in close proximity to mucosal surfaces, interact closely with the host immune cells during tumor initiation, promotion, and progression. As a result of this interaction, tumor cells, which are recognized by immune effector cells, can be eliminated, whereas tumor cells able to evade immune recognition can grow and become resistant to the host immune cells. Tumors can evade immunodetection by a general down-regulation or loss of antigen-presenting molecules, or, more specifically, the loss of immunogenic epitopes (23, 24, 25). An outgrowth of epitope-loss tumor variants, which are resistant to immune effector cells, gives the tumor a “competitive edge” for growth in a hostile environment. Another general mechanism of tumor evasion may involve tumor-associated factors, which can cause dysfunction or even death of immune effector cells (26).

It has been generally assumed that p53 accumulation provides an opportunity for presentation to T cells of immunogenic wt p53 epitopes (7, 27) and generation of wt p53 epitope-specific T cells in tumor-bearing hosts. The expected result would be the presence of relatively high frequencies of wt p53 epitope-specific T cells in the circulation of patients with tumors accumulating p53. However, contrary to expectations, the results of our earlier study indicated that CTL could be generated only from PBMCs of the patients whose tumors either did not accumulate p53 or accumulated, but could not present, the p53_264–272 epitope (9). To confirm these unexpected results, we recruited a larger group of HLA-A2.1⁺ patients with SCCHN and using tetramer technology, determined frequencies and phenotypic characteristics of T cells specific for the wt p53_264–272 peptide in the peripheral circulation of these patients and a group of healthy controls.

We found the highest frequencies of wt p53_264–272-specific CD8⁺ T cells in a subset of patients with SCCHN whose tumors did not accumulate p53 protein and had a wt p53 genotype. Furthermore, in a subset of patients with tumors accumulating mutant p53, the mean frequency of p53_264–272-specific T cells did not differ significantly from that in healthy controls. In principle, p53 accumulating in the tumor has an increased opportunity to be presented to immune cells (28). However, it is known that some tumors with mutated p53 are unable to process the wt p53_264–272 epitope. The precedent for blocking of the epitope processing by tumor with a missense mutation at the hotspot codon 273, flanking wt p53_264–272 epitope, has been described by Theobald et al. (19). It is possible that additional instances of blocked, altered, or incomplete processing of this as well as other p53 epitopes exist. Another plausible explanation for the observed low frequency of the epitope-specific T cells in patients with mutated p53 could be that recognition of tumors by CTL depends not on p53 accumulation alone but on its turnover and processing by malignant cells, as recently reported by Vierboom et al. (29). Thus, it is possible that processing of mutated p53 by the tumor cell proteasome may not lead to optimal presentation of the wt p53_264–272 epitope and effective generation of specific CTL. On the other hand, it has been shown that even when no accumulation of p53 is evident, wt p53 epitope presentation can occur, rendering the tumor susceptible to wt p53-specific CTL (8, 29, 30, 31). Thus, accumulation of mutated p53 is not the only criterion associated with the presentation of wt p53 epitope by the tumor and generation of CTL with p53 specificity.

The presence of HPV E6 in tumor cells could also influence p53 processing and CTL generation (29). HPV-16 E6/E7 expression has been reported in a substantial proportion of oral SCCHN (32). Expression of wt p53 in HPV E6-transformed cells is compatible with p53 inactivation, its proteolytic degradation, and enhanced processing, as well as presentation of its epitopes to T cells (29). For this reason, we examined the tumors studied for HPV-16 E6/E7 expression, and sought to correlate it with p53 expression and the frequency of CTL specific for wt p53_264–272 epitope. Multivariate analysis indicated that the frequency of wt p53-specific CTL depended on the p53 status of the tumor and not on its positivity for HPV. Nevertheless, it is interesting to note that in 5 of 7 patients with a relatively high frequency of wt p53-specific CTL, who did not accumulate p53, HPV-16 DNA was detected.

Another explanation for the observed low frequency of wt p53_264–272-specific T cells in patients with tumors accumulating p53 is that wt p53 epitopes are “self” determinants, and, thus, tolerance to them has to be overcome to induce an immune response. Studies by Theobald et al. (33) and Hernández et al. (34) demonstrated that tolerance to “self” p53 epitopes in mice is associated with deletion of high avidity T cells and retention of low to intermediate affinity T cells. We have consistently generated comparable anti-wt p53 CTL in humans following in vitro sensitization in the presence of epitope-pulsed dendritic cells (8, 9). Others have also reported generation of such CTL (35, 36). Furthermore, the current study shows that precursors of tetramer-positive anti-p53_264–272 T cells exist, albeit with low frequencies, in PBMCs of patients with tumors accumulating p53. Therefore, it is unlikely that CTL precursors of wt p53_264–272-specific T cells are deleted in cancer patients, as suggested previously. The mechanisms responsible for the failure of these precursor cells to expand ex vivo are presently unknown. It is possible that a certain threshold frequency of these precursor cells is needed to overcome anergy to self or immunosuppressive effects of the tumor microenvironment.

This study emphasizes the complexity of tumor-host interactions relevant to anti-wt p53 responses and to the development of wt p53-based vaccines. Our findings suggest that immune precursor cells of the wt p53_264–272 epitope are present in the circulation of HLA-A2⁺ patients with SCCHN and that in a subset of these patients, this epitope is immunogenic, results in CTL development, and contributes to shaping immunological memory. In this subset of SCCHN patients, CTL specific for wt p53_264–272 epitope might well have been responsible for elimination of tumors presenting the epitope and the outgrowth of epitope-loss tumor cells able to avoid these effector cells. On the other hand, it is also possible that tumors accumulating mutant p53, which are associated with poor prognosis (37), actively participate in elimination of tumor-specific effector cells, as suggested by studies reported from our laboratories.⁴ Tumor-associated apoptosis of such epitope-specific T cells might account for low frequencies of tetramer-positive CD8⁺ T cells in patients with tumors accumulating p53.

A complex interplay of factors, which might determine tumor survival or regression, is best illustrated in patient #9 (Table 1; Fig. 9). A relatively high frequency of the epitope-specific T cells in the circulation, the presence of anti-p53 Abs, accumulation of p53 in a small proportion (∼2%) of tumor cells, and tumor positivity for HPV in this patient, suggest that the patient’s immune system is actively modulating tumor growth. Delivery of wt p53-based vaccines to patients such as this one could result in a rapid expansion of CTL, which might drive the selection of epitope-loss tumor variants. On the other hand, patients with tumors harboring p53 mutations and a low frequency of wt p53-specific CTL are unlikely to benefit from wt p53-based vaccines, because the expected postvaccine expansion of CTL is unlikely. Strategies that might help to overcome these difficulties include the use of an altered peptide ligand of the p53_264–272 epitope (38), identification of other wt p53 epitopes, which might be more immunogenic than p53_264–272, as well as MHC-class II restricted p53 epitopes to provide help for generation of antitumor effector cells. We and others are in the process of evaluating several other known HLA-A2.1-restricted wt p53 epitopes in hope of identifying those that may be able to support generation of high-affinity CTL (39). In planning for future p53 vaccines, the use of individualized or personalized vaccines targeting mutant p53 also needs to be revisited, particularly in light of newer evidence that a considerable number of p53 mutations occur within known CTL-defined epitopes in HLA-A2⁺ SCCHN patients. However, the efficiency of such vaccines depends on the demonstration that a given mutated p53 epitope can be processed and is immunogenic. Therefore, the selection of an optimal wt p53 peptide for vaccination must await additional studies to define characteristics of other available p53 epitopes.

We thank Dr. Saleem Khan, University of Pittsburgh School of Medicine, for performing HPV-16 E6/E7 analysis and William Gooding, Biostatistics Center, University of Pittsburgh Cancer Institute, for providing statistical support.