CD8 T-cell Responses against Cyclin B1 in Breast Cancer Patients with Tumors Overexpressing p53

Research over the past decade has unraveled important new insight about the interplay between cancer cells and cells of the immune system, in particular T cells. Thus, it is well established that cancer cells express tumor-associated antigens (TAA) that are recognized in the context of human leukocyte antigen (HLA) molecules at the cell surface. A high number of TAA has been characterized and the antigenic peptides identified (1). Obviously, the identification of peptide antigens that are recognized by the host immune system offers the means to use the targeting capacity of the immune system therapeutically in the treatment of metastatic cancers, and various strategies are currently tested in animal models and clinical trials (2–4). Although this strategy has not yet met its promise, it nevertheless represents one of the more promising future treatments of metastatic cancers, emphasizing the need to study peptide-specific T-cell responses against cancer for selecting best-suited candidate peptides for immunologic targeting of cancer. Early attempts to characterize TAA only considered the criteria expression/no expression by the tumor cells, whereas more recent strives in the field has more fully appreciated current knowledge in cancer cell biology and selected target structures from proteins associated with key features of the malignant cells, for example, drug resistance (5), resistance to apoptosis (6), or disruption of cell cycle regulation (7, 8). To the latter, disruption of cell cycle control is one of the hallmarks of cancer, frequently associated with genetic alteration or overexpression of key players that regulate the cell cycle.

Cyclin B1 is a component of the cdc2 complex that regulates transition from G₂ to M phase of the cell cycle. In normal cells, it is expressed at low cell cycle–dependent levels and is mainly found in the nucleus. Overexpression of cyclin B1 (and other cyclins) has been observed in human cancer, for example, cancers of the breast, lung, colon, and melanoma, as well as leukemias and lymphomas (9, 10). Importantly, in cancer cells, cyclin B1 is mainly localized in the cytoplasm. The molecular mechanism of cyclin B1 overexpression has been shown to be due to the mutation of p53 (11). p53 plays a key role, directly or indirectly, for the inhibition of apoptosis, gene regulation, and drug resistance (12). To this end, point mutation of p53 often leads to accumulation of, at least partly, nonfunctional protein. About the expression of cyclin B1, wild-type p53 seem to regulate the cyclin B1 level posttranslationally, for example, by destabilizing cyclin B1. Thus, p53-negative cell lines overexpressing cyclin B1 showed diminished protein level upon transfection of wild-type p53; however, mRNA levels remained similar (11). The net outcome is accumulation of p53, as well as cyclin B1, in the cytoplasm of cancer cells harboring mutated p53.

One of the most frequent genetic alterations in cancer is point mutations in the coding region of the p53 gene (13). This underlines the notion that transforming genetic alterations may lead to an altered immunologic profile of the cancer cell, not only directly by changes in expression levels. Importantly, it has been shown that in vitro recognition of cancer cells by cyclin B1–specific CTLs was the result of nonfunctional p53, in turn leading to overexpression of cyclin B1 (11).

The overexpression of cyclin B1 in many cancers of different origin and the fact that immune escape by down-regulation or loss of expression of this protein would impair sustained tumor cell growth highlights cyclin B1 as a very attractive target for immunotherapy of cancer. To this end, Kao and colleagues identified two HLA-A2–restricted peptide epitopes from the cyclin B1 protein (14). Moreover, these peptides were shown to represent targets for cancer cell recognition by CTL in breast cancer, as well as squamous cell carcinoma of the head and neck, and thus is a shared TAA (3).

In the present study, we have examined peripheral blood mononuclear cells (PBMCs) from a panel of late stage breast cancer patients for T-cell reactivity against two HLA-A2–restricted peptides derived from the cyclin B1 protein. Our data underline cyclin B1 as an important immunologic target for therapeutic intervention in breast cancer and, probably, all cancers that have nonfunctional p53 protein.

Patients. Blood samples were drawn from patients with breast cancer a minimum of 4 wks after termination of any kind of anticancer therapy. PBMCs were isolated using lymphoprep separation, serologically HLA typed (Department of Clinical Immunology, University Hospital Copenhagen, Denmark), and frozen in FCS with 10% DMSO.

All patients were candidates to a clinical dendritic-cell vaccination trial (ref. 15 for further clinical information). All patients had disease dissemination involving 1 to 6 metastatic sites (e.g., lymph nodes, bone, liver, and lung). Patients with brain metastases were not included. The median age of the patients was 58 y (range, 32-71 y). Half of the patients had estrogen receptor–positive breast cancer and approximately one third expressed her-2/neu (HER2). Most breast cancer patients had previously received up to five different chemotherapy regimens, (e.g., epirubicin, docetaxel, cabecitabine), and patients with estrogen receptor–positive tumors had up to three kinds of endocrine therapy. Furthermore, patients with HER2-positive tumors were previously treated with transtuzumab. Blood from anonymized healthy donors (age, 18-70 y) were used as controls.

The study was approved by the Danish Ethics Committee, and informed consent was obtained from patients and healthy individuals before inclusion according to the Declaration of Helsinki.

Immune histochemical staining for p53. Sections of formalin-fixed, paraffin-embedded tissue from the primary breast carcinoma were cut and stained with antibody against p53 DO7 (DAKO; 1:50), as previously described (15). In brief, the slides were, treated with citrate buffer solution for antigen retrieval, boiled in microwave, and, washed in buffer solution (TBS). The staining procedure included incubation with primary antibody for 1 h, wash in TBS, incubation for 1 h with the secondary antibody, and finally, incubation with streptavidin-biotin-complex (DAKO). A multiblock including several different tissues was used for positive and negative controls. For estimation of positive reaction, only the strongly stained nuclei were counted and calculated as percentage of all tumor nuclei. The tumor sample was regarded as p53 negative if <5% of the cells stained positive.

Peptides and assembly assay for peptide binding to MHC class I molecules. The cyclin B1–derived peptides CB9 (AKYLMELTM) and CB-P4 (AKYLMELCC), the modified versions CB9L2 (ALYLMELTM) and CB9M2 (AMYLMELTM), and the p53-derived peptides R9V p53_65-73 (RMPEAAPPV), G11V p53_187-197 (GLAPPQHLIRV), L9V p53_264-272 (LLGRNSFEV), Y9L-L2 p53_103-111 (YLGSYGFRL), K9V-L2 p53_139-147 (KLCPVQLWV), and S9V-L2 p53_149-157 (SLPPPGTRV) were all purchased from Genscript. The binding affinity of the synthetic peptides (GenScript) to HLA-A2 molecules, metabolically labeled with [³⁵S]-methionine, was measured in the assembly assay, as described previously (16). The assay is based on peptide-mediated stabilization of empty HLA molecules released upon cell lysis from the TAP-deficient cell line T2. Stably folded HLA molecules were immune precipitated using the HLA class I–specific, conformation-dependent monoclonal antibody W6/32 and separated by isoelectric focusing gel electrophoresis. MHC heavy-chain bands were quantified using the ImageGauge Phosphoimager program (FUJI Photo Film). The intensity of the band is directly related to the amount of peptide-bound class I MHC complex recovered during the assay. Subsequently, the extent of stabilization of HLA-A2 is directly related to the binding affinity of the added peptide. The recovery of HLA-A2 was measured in presence of 40, 4, 0.4, and 0.04 μmol/L of the relevant peptide. The C₅₀ value was calculated for each peptide as the peptide concentration sufficient for half-maximal stabilization.

Antigen stimulation of PBMCs. To extend the sensitivity of the ELISPOT assay, PBMCs were stimulated once in vitro with peptide before analysis (17). At day 0, PBMCs were thawed and plated in 2 ml/well at a concentration of 2 × 10⁶ cells in 24-well plates (Nunc) in X-vivo medium (BioWhittaker) with 5% heat-inactivated human serum in the presence of 10 μmol/L peptide (GenScript). One day later, 40 IU/mL recombinant interleukin 2 (Chiron) was added to the cultures. The cultured cells were tested in the IFN-γ ELISPOT assay on day 8.

IFN-γ ELISPOT assay. The ELISPOT assay was used to quantify peptide epitope–specific IFN-γ–releasing effector cells, as described previously (18). Briefly, nitrocellulose-bottomed 96-well plates (MultiScreen MAIP N45, Millipore) were coated with anti–IFN-γ Ab (1-D1K, Mabtech). The wells were washed and blocked by X-vivo medium, and the effector cells were added in duplicates at different cell concentrations, with or without 10 μmol/L peptide. The plates were incubated overnight. The following day, medium was discarded, and the wells were washed before addition of biotinylated secondary Ab (7-B6-1-Biotin, Mabtech). The plates were incubated at room temperature for 2 h and washed, and avidin-enzyme conjugate (AP-Avidin, Calbiochem/Invitrogen Life Technologies) was added to each well. Plates were incubated at room temperature for 1 h, and the enzyme substrate NBT/BCIP (Invitrogen Life Technologies) was added to each well and incubated at room temperature for 5 to 10 min. Upon the emergence of dark purple spots, the reaction was terminated by washing with tap water. The spots were counted using the ImmunoSpot Series 2.0 Analyzer (CTL Analyzers), and the peptide-specific CTL frequency could be calculated from the numbers of spot-forming cells. The number of peptide-specific CTL is calculated as the average number of peptide specific IFN-γ spots formed in the IFN-γ ELISPOT done in duplicates or triplicates. Responders are defined as the mean number of spots from triplicates +/− 1/2 SD > 25 spots per 10⁵ cells.

Flow cytometry. CD8 stainings were done with PBMCs from cancer patients. For tetramer stainings, PBMCs from cancer patients were stimulated once in vitro with CB9L2 peptide. The CD8 cells were isolated from PBMCs using the Dynal CD8 negative isolation kit (Dynal Biotech) at day 7. The resulting T-cell cultures were stained with a phycoerythrin (PE)-coupled tetramer (ALYLMELTM), followed by antibody staining with the flourochrome-coupled monoclonal antibodies CD8-allophycocyanin and CD3-FITC (BD Immunocytometry Systems). Tetramer was prepared using the MHC-peptide exchange technology (19), as described (20). Tetramer stainings were done in PBS + 2% FCS, for 15 min at room temperature, in the dark, whereas antibody stainings were done in PBS + 2% FCS, at 4°C for 20 min in the dark. The MHC tetramer complexes used were HLA-A2/CB9L2 (ALYLMELTM) and HLA-A2/HIV-1 pol476-484 (ILKEPVHGV). The samples were analyzed on BD fluorescence-activated cell sorting (FACS) aria, using DIVA software (BD Biosciences).

Immune histochemistry. In 14 of 36 patients, a biopsy from the primary breast tumor stained positive (>5% positive cells) for p53 by immune histochemistry (Fig. 1). The percentage of p53 positive tumor cells ranged from 30% to 100%.

Binding of wild-type and modified cyclin B1–derived peptides to HLA-A2. Initially, two previously characterized cyclin B1–derived peptides were synthesized and examined for binding to HLA-A2 (Table 1). C₅₀ values were estimated for each peptide as the peptide concentration needed for half-maximal stabilization of HLA-A2. Both peptides displayed very low binding affinity for HLA-A2 (Table 1). Consequently, we synthesized analogue peptides in which the natural amino acids were replaced with theoretically more optimal anchor residues. We modified the CB9 peptide introducing leucine or methionine in position 2 instead of lysine. Both modified peptides (CB9L2 and CB9M2) displayed high affinity to HLA-A2 (Table 1).

CTL responses against cyclin B1–derived peptides in breast cancer patients. Blood samples from 36 HLA-A2–positive breast cancer patients were analyzed by ELISPOT for reactivity against peptide epitopes restricted to HLA-A2, derived from the cyclin B1 protein. As control, PBMCs from 10 healthy HLA-A2⁺ individuals were examined. Immune responses against CB-P4 and/or CB9 were detected in 12 of the patients. Two patients showed responses against CB-P4 and CB9 peptides. These responses occupied ∼90 to 210 and 90 to 150 peptide-specific CTL per 3 × 10⁵ in vitro–stimulated PBMCs, respectively. Immune responses against the modified peptides showed increased responses compared with the native CB9 peptide. The responses against the modified peptides CB9L2 and CB9M2 occupied 80 to 240 and 120 to 230 peptide-specific CTL per 3 × 10⁵ in vitro–stimulated PBMCs, respectively. The strongest responses were observed for the CB9L2 modified peptide (Fig. 2A and B). No responses against the cyclin B1–derived peptides were detected in control individuals (Fig. 2B).

FACS analyses of cyclin B1–specific T cells. The presence of CB9L2-specific CTL in PBMCs from breast cancer patients was further evaluated using FACS analyses and MHC tetramer staining. CD8 T cells from 10 breast cancer patients were stained with the HLA-A2/CB9L2 tetramer complex after a short in vitro stimulation. FACS analyses revealed readily detectable populations of tetramer-positive T cells constituting 0.1% of the CD8 T cells in three of the patients. As a control, the same cultures were stained with a HLA-A2/HIV-1 pol_476-484 tetramer. In comparison, the same samples of PBMCs showed ∼0.3% CB9L2-specific, IFN-γ–secreting CD8 T cells when analyzed using the ELISPOT assay. No HLA-A2/CB9L2 tetramer positive cells could be detected in PBMCs from five healthy donors. Figure 2C illustrates tetramer stainings on CD8 T cells from a breast cancer patient and a healthy donor.

Cyclin B1 reactivity correlates with accumulation of p53. Interestingly, T-cell reactivity against the cyclin B1–derived peptides correlated with elevated p53 protein expression in the biopsy. Accordingly, most cyclin B1–responding T cells were found in PBMCs from patients with tumor cells having increased levels of p53. In fact, 9 patients of 14 with overexpressed p53 in the tumors harbored cyclin B1–specific responses, whereas only 3 patients of 22 patients without p53 overexpressing tumors harbored cyclin B1–reactive T cells at detectable levels (Fig. 3). p53 mutation, which often leads to accumulation of the p53 protein in the nucleus and/or cytoplasm because of decreased degradation, is correlated with an up-regulation of cyclin B1, and thus, the immune reactivity toward this protein is induced in these patients (Fig. 3). This prompted us to analyze for reactivity against peptides derived from the p53 protein (R9V p53_65-73, G11V p53_187-197, L9V p53_264-272, Y9L-L2 p53_103-111, K9V-L2 p53_139-147, and S9V-L2 p53_149-157) in PBMCs from patients with cancers overexpressing p53. Five patients that did not host an immune response against the cyclin B1–derived peptides were examined, together with five patients that did show reactivity against at least one of the cyclin B1–derived peptides. However, no responses were observed against any of the six p53 peptide epitopes (data not shown).

Tumorigenesis is characterized by genetic and epigenetic changes that lead to an abnormal protein profile of the cancer cell compared with the normal counterpart, in turn endowing cancer cells with advantages in growth and survival (21). These tumor-specific alterations, playing an important role for the malignant properties of the cancer cell, also offer targets for therapeutic intervention, for example, antisense therapy, small molecule inhibitors, and immunotherapy (22). To the latter, several groups of proteins sharing the functional property of being important for tumor cell growth and/or survival are recognized by host T cells. Thus, immunogenic CD8 restricted peptide epitopes have been characterized in proteins associated with inhibition of apoptosis (6), cell cycle regulators (8, 23, 24), and proteins associated with drug resistance (5, 25). These target structures seem attractive targets because of the global expression among a variety of cancers. In addition, because of the functional properties of these groups of proteins for growth and survival of cancer cells, in some instances, even more pronounced during treatment with conventional therapeutic regimens, tumor escape induced by a selective pressure seems less probable because these escape variants would have a lower growth and survival potential.

The cyclin B1 protein seems highly suited for targeting in cancer therapy. Kao et al. (7) described immunogenic peptides derived from cyclin B1 restricted by the HLA-A2 molecule. These peptides were characterized by elution of peptides from cancer cells; thus, the expression of the peptides on the cell surface of cancer cells is firmly established. However, as shown in the present study, the peptides bind with quite low affinity to the HLA molecule. Many TAA-derived peptides are characterized by low or intermediate affinity to the HLA molecule and still elicit a potent immune response; thus, this may not in itself be problematic (1). On the other hand, it may be advantageous to modify the peptides in anchor residues, for example, for use in vaccination strategies that take advantage of minimal peptides although the advantage of such peptides were recently questioned (26, 27). Nonetheless, for experimental studies, the use of high-affinity modified peptides may increase the numbers of T cells detectable by ELISPOT and also offers the opportunity to use recombinant HLA-peptide complexes in the study of these cells (6).

In the present study, we used the previously characterized native cyclin B1 peptides CB-P4 and CB9, as well as modified versions of these peptides, to screen for reactivity among PBMCs from advanced breast cancer patients. By the ELISPOT assay, we could detect cyclin B1–specific responses in approximately one third of the patients. Thus far, only HLA-A2–restricted peptides have been characterized from the cyclin B1 protein. Thus, considering the probable presence of peptides restricted by other HLA alleles, overall cyclin B1 reactivity would be expected to be even more frequent. Moreover, taking advantage of the modified high-affinity peptide CB9L2, we were able to use tetramer complexes for detection of these T cells among PBMCs, showing a small but distinct population of cyclin B1–specific T cells. The detected responses were only detectable upon in vitro stimulation. However, with few exceptions (e.g., the Mart-1 melanoma-associated antigen; ref. 28), most tumor-specific T-cell responses are of too low frequency for direct ex vivo detection. Unfortunately, because of the low frequency of cyclin B1–specific T cells, we were not able to address the phenotype of the cells. However, cyclin B1–specific T cells were not detectable in healthy donors, pointing to an antigen experienced phenotype elicited by tumor-driven immune activation (29).

p53-derived peptides have been characterized as targets for CD8 T cells in cancer; of note, these peptides are of wild-type origin (8, 30), and recognition is supposedly due to accumulation of the protein due to point mutation. Thus, we speculated that patients with mutated tumors could harbor T-cell responses against p53-derived peptides as well. However, no such responses were detected when we screened for reactivity against six previously described peptides (data not shown; refs. 15, 30). The p53 protein is relatively large, and to our knowledge, T-cell reactivity against p53-derived peptides has been carried out based on prediction of peptides in the protein (31). Obviously, by such a strategy, immune dominant epitopes may be left undetected because the main component of these predictions is capacity to bind and stabilize the HLA molecule. However, as also underscored by the low affinity of the CB-P4 and CB9 peptides, dominant TAAs may in fact be of quite low affinity (7, 32).

Because of the previously described relationship between mutation in p53 and overexpression of cyclin B1, we looked for correlation between p53 accumulation and response to cyclin B1. These data clearly showed that responses against cyclin B1 were predominantly found in patients harboring cancer cells with high levels of cytoplasmic p53. In fact, 9 patients of 14 with accumulation of p53 in the tumors harbored cyclin B1–specific responses, whereas only 3 of 22 patients with p53-negative tumors harbored cyclin B1–reactive T cells at detectable levels. Not alone does this support previous data that p53 mutation leads to overexpression of cyclin B1, it also emphasize the notion of “immune editing” (33, 34), at least in the sense that the immune system responds to changes in protein profiles of the cancer cells, although it is questionable whether the a low-frequency T-cell response confers any substantial selection pressure on the cancer cell population. About cyclin B1 reactivity observed in the three patients with apparent wild-type p53, the most likely cause could be that p53 in these tumors is absent, for example, because of mutation generating a premature stop codon. Another possibility is that the mutation changes the determinant recognized by the antibody. Both these notions are purely speculative.

The high frequency of p53 mutations among human cancers underscores cyclin B1 as a suitable target for vaccination against cancer. Moreover, mutated p53 is associated with bad prognosis, for example, by poor response to chemotherapy, which implies that targeting of cyclin B1 could synergistically be combined with conventional chemotherapy (22). Importantly, although p53 seems to be an obvious target and induced immune responses seem to be clinically relevant (15), it seems that targeting of cyclin B1 is more broadly applicable because cyclin B1 is overexpressed also when tumor cells are devoid of p53 protein. A tempting approach would be to induce and maintain a cyclin B1 CD8 memory response in patients carrying p53 wild-type cancer cells to study whether such a response could prevent or delay further progression of the disease, that is, prevent the outgrowth of p53-negative cancer cells. Moreover, the use of cyclin B1–specific T-cell clones or lines may be applicable to a broad range of cancer patients for adoptive transfer.

No potential conflicts of interest were disclosed.

We thank Merete Jonassen for the excellent technical assistance, Prof. A. Svejgaard and B.K. Jakobsen (Department of Clinical Immunology, University Hospital, Copenhagen, Denmark) for HLA typing, and Tobias Wirenfeldt Klausen for helping with the preparation of figures.