Vaccination with class I tumor peptides has been performed to induce tumor-reactive CD8+ T cells in vivo. However, the kinds of immune responses that vaccination might elicit in patients are not fully understood. In this study we tried to elucidate the mechanisms by which vaccination of class I binding tumor peptides into an HLA-A2+ lung cancer patient elicited dramatic amounts of IgG1 and IgG2 specific to a nonamer peptide, ubiquitin-conjugated enzyme variant Kua (UBE2V)43–51. The UBE2V43–51 peptide contains cysteine at the sixth position. HLA-DR-restricted and UBE2V43–51 peptide-recognizing CD4+ T cells were induced from postvaccination, but not from prevaccination, PBMCs of the cancer patient. In addition, a CD4+ T cell line (UB-2) and its clone (UB-2.3), both of which recognize the UBE2V43–51 peptide in the context of HLA-DRB1*0403 molecules, were successfully established from postvaccination PBMCs. The peptide vaccination increased the frequency of peptide-specific T cells, especially CD4+ T cells. In contrast, mass spectrometric analysis revealed that the vaccinated UBE2V43–51 peptide contained both monomeric and dimeric forms. Both forms, fractionated by reverse phase HPLC, were recognized by UB-2 and UB-2.3 cells. Recognition by these CD4+ T cells was observed despite the addition of a reduction reagent or the fixation of APC. Overall, these results indicate that vaccination with class I tumor peptides can induce HLA-DR-restricted CD4+ T cells in vivo and elicit humoral immune responses, and that a cysteine-containing peptide can be recognized by CD4+ T cells not only as a monomer, but also as a dimer.
Recent advances in molecular biology and tumor immunology have resulted in the identification of many tumor Ags and antigenic peptides recognized by CD8+ T cells (1, 2). Numerous clinical trials have shown that vaccination of class I binding tumor peptides can increase peptide-specific CD8+ T cells in the periphery of cancer patients (3, 4). Because vaccination with class I binding tumor peptides has been performed to induce tumor-reactive CD8+ T cells in vivo, the analysis of peptide-induced immune responses has focused on CD8+ T cells. Therefore, it remains unclear what kinds of immune responses, other than CD8+ T cells, might be elicited in peptide-vaccinated cancer patients.
We have identified a panel of antigenic peptides capable of inducing tumor-reactive CTLs in patients with epithelial cancers (5, 6, 7, 8) and have used these peptides in peptide-based immunotherapy (9, 10). In conducting our clinical trials, we observed cases in which peptide-specific IgG was dramatically elicited after the peptide vaccination. This observation led us to test the possibility that vaccination with class I binding tumor peptides would induce peptide-recognizing CD4+ Th cells, as in vivo generation of Ag-specific IgG generally requires a cytokine from Th cells as a result of cognate interaction between T and B cells (11). In the present study we provide evidence that vaccination with class I binding tumor peptides induced the in vivo generation of HLA-DR4-restricted CD4+ T cells in a cancer patient, and that a cysteine-containing peptide can be recognized by CD4+ T cells not only as a monomer, but also as a dimer.
Materials and Methods
Patient and peptides
At the time of the study, the patient (male; EBL-101; HLA-A*0201+ and HLA-DRB1*0403/0403) was 69 years old with non-small cell lung cancer and bone marrow metastasis. The following peptides were used for the clinical trial and were prepared under conditions of good manufacturing practice by Multiple Peptide Systems (San Diego, CA): cyclophilin B-derived peptide (CypB172–179; 3 VLEGMEVV), Lck-derived peptide (Lck422–430; DVWSFGILL), ubiquitin-conjugated enzyme variant Kua (UBE2V)-derived peptide (UBE2V43–51; RLQEWCSVI), and Wolf-Hirschhorn syndrome candidate 2 protein (WHSC2)-derived peptide (WHSC2141–149; ILGELREKV). These four peptides have the ability to induce tumor-reactive CTL activity in the PBMCs of HLA-A2+ cancer patients (6, 7, 8). UBE2V was identified as an epithelial cancer-related Ag by the screening of a cDNA library from a pancreatic adenocarcinoma cell line, panc-1 (6). The other HLA-A2-binding peptides used as negative controls were an HIV-derived peptide (HIV gag77–85; SLYNTYATL), an EBV-derived peptide (BMLFI259–267; GLCTLVAML), and a Flu peptide (t-A-MP-MI58–66; GILGFVFTL). These peptides, of >90% purity, were purchased from the Biologica (Nagoya, Japan). All peptides were dissolved with DMSO at a dose of 10 mg/ml.
One milliliter of the peptide solution at 1 mg/ml was mixed with an equal volume of IFA (Montanide ISA-51; Seppic, Paris, France), emulsified in 5-ml sterilized syringes, and injected s.c. into the lateral thigh. Immediate- and delayed-type hypersensitivity reactions were determined at 20 min and 24 h after the skin test, respectively. The patient received the vaccinations at 2-wk intervals. This protocol was approved by the Kurume University review board and the independent ethical committee. The patient gave written informed consent before the trial.
Cell lines and stable transfectants
HM-LCL is an HLA-DRB1*0403+ EBV-transformed lymphoblast cell line (provided by Dr. N. Emi, Nagoya University, Nagoya, Japan). To prepare HLA-DRA1*0101-expressing 293T cells, the HLA-DRA1*0101 gene was cloned into the pCR3.1 vector (Invitrogen, Carlsbad, CA) by the TA cloning method and was electroporated into 293T cells. To obtain HLA-DRB1*0403, HLA-DRB1*0405, and HLA-DRB4*0103 genes, each HLA-DR gene was cloned into a pcDNA3.1/V5-His TOPO vector (Invitrogen) by TA cloning. These DR genes were cut with HindIII and XbaI enzymes and were inserted into the pcDNA3.1/Hygro (Invitrogen). These genes were electroporated into HLA-DRA1*0101-expressing 293T cells, followed by selection with hygromycin B (Invitrogen) at a dose of 150 μg/ml.
Detection of peptide-specific IgG
Peptide-specific IgG levels in the plasma were measured by ELISA, as previously reported (12). In brief, after a 2-h incubation of the samples in the peptide-immobilized plates, the plates were washed and further incubated for 2 h with 1/1000 diluted rabbit anti-human IgG (γ-chain-specific; DAKO, Glostrup, Denmark). After washing, 100 μl of 1/100 diluted goat anti-rabbit Ig-conjugated HRP (En Vision; DAKO) was added to each well. After washing, 100 μl/well tetramethylbenzidine substrate solution (Kirkegaard & Perry Laboratories, Guildford, U.K.) was added, and the reaction was stopped by the addition of 1 M phosphoric acid. To determine IgG subclasses, rabbit Abs against human IgG1 and IgG2 were used.
Culture of PBMCs and establishment of a UBE2V43–51 peptide-recognizing T cell line and its clone
PBMCs (2 × 106) of the patient EBL-101 before and after peptide vaccination were depleted of CD8+ T cells using Dynabeads M-450 (Dynal, Oslo, Norway). The remaining cells were cultured in 48-well plates with the UBE2V43–52 peptide (10 μg/ml) in RPMI 1640 medium with 5% human serum. On days 3 and 7, IL-2 was added to a final dose of 20 U/ml. On day 10, the cultured cells were harvested and then cultured with autologous PBMCs (1 × 104 cells/well), which were preloaded with or without the UBE2V43–51 peptide (10 μg/ml). After 20-h incubation, the supernatants were collected, and the levels of IFN-γ were determined by ELISA. In some groups, 10 μg/ml of anti-HLA-A2 (BB7.2; mouse IgG2b), anti-HLA-DR (L243; mouse IgG2a), or anti-HLA-DP (B7/21; mouse IgG1) mAb was added to the wells. The UBE2V43–51 peptide-stimulated PBMCs were seeded into 96-well, round plates at a cell dose of 10 cells/well in the presence of irradiated allogeneic PBMCs and PHA (10 μg/ml) with a culture medium consisting of 45% RPMI 1640 medium, 45% AIM-V medium (Life Technologies, Gaithersburg, MD), 10% FCS, 100 U/ml human rIL-2, and 0.1 mM MEM nonessential amino acid solution (Life Technologies). After 2 wk in the culture medium, reactivity to the UBE2V43–51 peptide was examined by ELISA for IFN-γ, and a cell line designated UB-2 was established. Thereafter, limiting dilution at a cell dose of 0.5 cells/well was conducted, and a clone, designated UB-2.3, was established.
Assay of recognition
The UB-2 and UB-2.3 cells were incubated separately with the indicated stimulator cells, which were pulsed with peptides for 90 min. After a 20-h culture, the level of IFN-γ in the supernatant was determined by ELISA. To block dimerization of the UBE2V43–51 peptide, the peptide was cultured with 2 mm DTT (Molecular Probes, Eugene, OR) in a cysteine-free DMEM medium (Life Technologies, Gaithersburg, MD; catalogue no. 21013-0247) without serum for 2 h in CO2 incubator. HM-LCL cells were then added and cultured for additional 90 min. Thereafter, these peptide-pulsed HM-LCL cells were fixed with 1% paraformaldehyde for 5 min. After thoroughly washing the cells, these HM-LCL cells were cultured with effector cells. In some experiments HM-LCL cells were fixed with 1% paraformaldehyde before pulsation of the peptide to inhibit internalization of the pulsed peptide.
Flow cytometric analysis
To examine the phenotypes of cells, they were stained with anti-CD8, anti-CD4, anti-HLA-A2, or anti-HLA-DR mAb, followed by FITC-conjugated goat anti-mouse IgG. The results were analyzed by the CellQuest program (BD Biosciences, Mountain View, CA).
Assay of peptide-specific T cell precursors
The assay of peptide-specific CD8+ T cell precursors was performed based on a previously reported culture method (13). Briefly, PBMCs (1 × 105 cells/well) were incubated with 10 μg/ml of the UBE2V43–15 peptide in the wells of U-bottom, 96-well microculture plates (Nunc, Roskilde, Denmark) in 200 μl of culture medium. The culture medium consisted of 45% RPMI 1640 medium, 45% AIM-V medium, 10% FCS, 100 U/ml IL-2, and 0.1 μM MEM nonessential amino acid solution. Half the medium was removed and replaced with new medium containing a corresponding peptide (20 μg/ml) every 3 days. After incubation for 14 days, these cells were harvested and tested for their ability to produce IFN-γ in response to T2 cells that were preloaded with either the UBE2V43–51 peptide or the HIV peptide as a negative control. To estimate the frequency of CD4+ T cells, CD8+ T cell-depleted cells (1 × 105 cells/well) were incubated with 10 μg/ml of the UBE2V43–15 peptide in the wells of U-bottom, 96-well microculture plates in 200 μl of culture medium consisting of RPMI 1640 medium with 5% human serum. On day 2, IL-2 was added to a final dose of 10 U/ml. On day 7, half the medium was removed and replaced with new medium containing the UBE2V43–15 peptide (20 μg/ml), and 10 U/ml IL-2 was added on day 8. After incubation for 14 days, these cells were harvested and tested for their ability to produce IFN-γ in response to HLA-DRB1*0403-expressing 293T cells, which were preloaded with either the UBE2V43–51 peptide or HIV peptide, in the presence of 50 U/ml IL-2. Both assays were performed in 12 wells, and cultured cells in one well were divided into four wells. Thereafter, two wells were used for control HIV peptide-pulsed cells, and the other two wells were used for the UBE2V43–51 peptide-pulsed cells. Thus, cytokine release was assessed in two different ELISA wells.
Frequency assay of peptide-specific T cells
CD8+ or CD4+ T cells from the pre- or postvaccination PBMCs were positively isolated using the CD8 Positive Isolation Kit (Dynal, Oslo, Norway) or the CD4 Positive Isolation Kit (Dynal). CD8+ or CD4+ T cells (1600, 400, 100, and 25 cells/well) were then cultured with irradiated allogeneic PBMCs (5 × 104 cell/well), 10 μg/ml PHA, and 100 U/ml IL-2 in 24 wells of 96-well, round-bottom plates. On day 10, these cultured cells were harvested and stimulated with T2 cells (for CD8) or 293TDRB1*0403 cells (for CD4), which were pulsed with either UBE2V43–51 peptide or the HIV peptide. The cultured cells in one well were divided into four wells, with two being used as control HIV peptide-pulsed cells and the other two being used for the UBE2V43–51 peptide-pulsed cells. The level of IFN-γ in the supernatant was determined by ELISA. The wells that produced significantly (p < 0.05) IFN-γ in response to the UBE2V43–51 peptide were judged to be positive.
Mass spectrometric analysis and peptide fractionation
The vaccinated UBE2V43–51 peptide was analyzed by liquid chromatography (Gold Nouveau; Beckman, Palo Alto, CA) with a Vydac C18 (218TP54) column. Solution A was 0.1% trifluoroacetic acid (TFA), and solution B was 0.1% TFA/acetonitrile. The flow rate was 1 ml/min. To separate monomeric and dimeric forms, the UBE2V43–51 peptide was fractionated by a 625LC liquid chromatography system (Waters, Milford, MA) using reverse phase HPLC with a Finepak SIL300 C18 T-7 column (Nippon Bunkou Kogyo, Tokyo, Japan). Solution A was 0.1% TFA, and solution B was a mixture of solution A and acetonitrile at a ratio of 2:3. The peptide was eluted at a flow rate of 1 ml/min.
The statistical significance of the data was determined by two-tailed Student’s t test, and p < 0.05 was considered statistically significant throughout the study.
Marked elicitation of IgG against the vaccinated UBE2V43–51 peptide
The patient EBL-101 was vaccinated with four different HLA-A2-binding peptides, CypB172–179, Lck422–430, UBE2V43–51, and WHSC2141–149, with a 2-wk interval at a dose of 3 mg after confirmation that the peptide-specific CTL precursors were present in the PBMCs (data not shown). The patient’s disease remained stable for 18 mo after the peptide vaccination. The serum levels of IgG specific to the administered peptides were measured, as IgG reactive to class I-binding tumor peptides can be detected in peptide-vaccinated cancer patients (9). As a result, IgG specific to the UBE2V43–51 peptide, but not to the CypB172–179 peptide, drastically increased after the third vaccination (Fig. 1,A). No IgG against the other two peptides was detected, and the patient developed a strong delayed-type hypersensitivity reaction only against the UBE2V43–51 peptide (data not shown). IgG specific to the UBE2V43–51 peptide was absorbed by culturr in peptide-coated plates in an Ag-specific manner (Fig. 1,B). In addition, IgG specific to the UBE2V43–51 peptide consisted of both IgG1 and IgG2 subclasses (Fig. 1 C), indirectly suggesting the in vivo elicitation of both Th1-type and Th2-type T cell responses.
Induction of UBE2V43–51 peptide-recognizing CD4+ cells after peptide vaccination
To investigate the mechanisms of elicitation of IgG against the UBE2V43–51 peptide after the peptide vaccination, we tested the possibility of the participation of peptide-recognizing CD4+ Th cells. Then we determined whether UBE2V43–52 peptide-recognizing CD4+ T cells could be detected after the peptide vaccination (Fig. 2). To exclude the influence of UBE2V43–51 peptide-recognizing CD8+ T cells, CD8+ T cells were depleted before the in vitro culture. In the prevaccination PBMCs, the percentages of CD4+ T cells were 24.9 and 32.6% before and after CD8+ T cell depletion, respectively. In the PBMCs after the third vaccination, the percentages of CD4+ T cells were 17.5 and 23.8% before and after CD8+ T cell depletion, respectively. In both cases the percentage of CD8+ T cells after CD8+ T cell depletion was <1% (data not shown). As a result, IFN-γ-producing CD4+ T cells in response to the UBE2V43–51 peptide were detected in the PBMCs after the third peptide vaccination. IFN-γ production was inhibited by the addition of anti-HLA-DR mAb, but not by the addition of anti-HLA-DP mAb. Peptide-specific IL-6 production was not observed in this assay (data not shown). These results suggest that the vaccine-induced elicitation of UBE2V43–51 peptide-specific IgG was the result of in vivo induction of peptide-specific CD4+ Th cells.
A UBE2V43–51 peptide-recognizing Th1-type CD4+ T cell line and its clone
To further investigate the UBE2V43–51 peptide-recognizing CD4+ T cells, we tried to establish T cell lines recognizing the UBE2V43–51 peptide. Because the patient homozygously expressed HLA-DRB1*0403 molecules, an HLA-DRB1*0403-expressing B lymphoblastoid cell line, HM-LCL, was used for the screening. Although the UBE2V43–51 peptide has the ability to bind HLA-A2 molecules (6), the HM-LCL cells were negative for HLA-A2 molecules (Fig. 3). An established T cell line, designated UB-2, was positive for CD4, but not for CD8. In addition, UB-2 cells produced IFN-γ in response to the UBE2V43–51 peptide when pulsed on the HM-LCL cells or on autologous PBMCs, whereas they failed to produce IFN-γ when pulsed on T2 cells (Fig. 4,A). We further confirmed their HLA-DR restriction using 293T cells expressing specific HLA-DRB1 molecules (Fig. 4 B). UB-2 and its clone, UB-2.3, produced a significant level of IFN-γ in response to the UBE2V43–51 peptide only when pulsed on 293TDRB1*0403 cells; neither UB-2 nor UB-2.3 produced IL-4 or IL-6 (data not shown). These results indicate that peptide recognition of these Th1-type UB-2 and UB-2.3 cells was restricted to HLA-DRB1*0403 molecules.
We next determined whether the cloned UB-2.3 cells could recognize 293TDRB1*0403 cells, which were pulsed with various doses of the UBE2V43–51 peptide (Fig. 5). As a result, UBE-2.3 cells produced IFN-γ when 293TDRB1*0403 cells were pulsed with a relatively high dose (>1 μg/ml) of the UBE2V43–51 peptide. In addition, we examined the expression of the UBE2V gene in 293T cells using the RT-PCR method. The result was that the 293T cells were positive for the UBE2V gene (data not shown).
Increased frequency of the UBE2V peptide-specific T cell precursors after peptide vaccination
It is important to kinetically estimate the frequency of CD8+ or CD4+ T cells reactive to the UBE2V43–51 peptide after the peptide vaccination. We first estimated the frequency of peptide-specific CD8+ T cells by using our previously reported culture protocol (13). Whole PBMCs from the patient were repeatedly stimulated with the UBE2V43–51 peptide in the presence of 100 U/ml IL-2. Although CD4+ T cells were not depleted, the reactivity of T cells against peptide-pulsed T2 cells could be judged to depend on HLA-A2-restricted T cells because HLA-DRB1*0403-restricted CD4+ T cells showed no reactivity against UBE2V43–51 peptide-pulsed T2 cells (Fig. 4,A). As shown in Fig. 6,A, the reactivity of HLA-A2-restricted CD8+ T cells against the UBE2V43–51 peptide was detected in one of 12 wells in the prevaccination PBMCs, but it was detected in two and three of 12 wells after the 6th and 12th peptide vaccinations, respectively. With regard to CD4+ T cells, CD8+ T cell-depleted cells were in vitro cultured with the UBE2V43–51 peptide and were examined for their IFN-γ production in response to 293TDRB1*0403 cells that were prepulsed with the HIV peptide or the UBE2V43–51 peptide. The result was that the HLA-DRB1*0403-restricted and UBE2V43–51 peptide-reactive CD4+ T cells were induced only after the peptide vaccination. The reactivity of HLA-DR4-restricted CD4+ T cells against the UBE2V43–51 peptide was not detected in the prevaccination PBMCs, but was detected in two, one, and two of 12 wells after the 3rd, 6th, and 12th peptide vaccinations. We performed a limiting dilution assay to accurately determine the frequencies of peptide-specific CD4+ or CD8+ T cells in the pre- and postvaccination PBMCs (Fig. 6 B). Positively isolated CD8+ or CD4+ T cells from the pre- or postvaccination PBMCs were used for the assay. The positive percentage of CD8+ or CD4+ T cells was >90% (data not shown). As a result, the frequency of peptide-specific CD8+ T cells slightly increased after the 6th and 12th vaccinations, and the increase in the frequency of peptide-specific CD8+ T cells after the 12th vaccination was ∼2-fold. In contrast, the frequency of peptide-specific CD4+ T cells increased prominently after the 3rd, 6th, and 12th vaccinations, and the increase in the frequency of peptide-specific CD4+ T cells after the 12th vaccination was >12-fold. These results indicate that the peptide vaccination increased the frequency of peptide-specific T cells, especially CD4+ T cells.
Recognition of both monomeric and dimeric forms of the UBE2V43–51 peptides
The UBE2V43–51 peptide carries cysteine at the 6th position. Chromatography and mass spectrometric analysis were conducted on the vaccinated UBE2V43–51 peptide (Fig. 7). Two peaks, with retention times of ∼8.52 and ∼10.33, were observed in chromatographic analysis. With regard to the mass spectrometric analysis, the peak at retention time ∼8.52 had (M+H)+ = 1133.6 and (M + 2H)2+/2 = 567.3. The peak at retention time ∼10.33 had (M2 + 2H)2+/2 = 1133.0 and (M2 + 3H)3+/3 = 755.6, and the presence of the latter fragment indicated that the peak at the later retention time (∼10.33) was a dimer. We next investigated whether replacement of the 6th cysteine with alanine or serine would have any influence on recognition by UB-2 and UB-2.3 cells and found that the substitution resulted in loss of recognition by these cells (data not shown).
To directly investigate which forms of UBE2V43–51 peptide could be recognized by the UBE2V43–51 peptide-recognizing CD4+ T cells, monomeric and dimeric peptides were fractionated using reverse phase HPLC (Fig. 8,A). As shown in Fig. 8,B, two peaks were observed in IFN-γ production by UB-2 and UB-2.3 cells in response to the HM-LCL pulsed with each fraction. Next, the effect of a reduction reagent, DTT, on recognition of the fractionated monomeric UBE2V43–51 peptide (fractions 15–17 in Fig. 8,B) by UB-2 and UB-2.3 cells was examined to exclude the possibility that the fractionated monomeric UBE2V43–51 peptide became dimeric during the in vitro culture. However, both UB-2 and UB-2.3 cells recognized the DTT-treated monomeric UBE2V43–51 peptide (Fig. 9,A). There remains yet another possibility, i.e., that the dimeric UBE2V43–51 peptide was internalized into HM-LCL cells, reduced, and subsequently presented in monomeric form in the context of HLA-DRB1*0403 molecules, as intracellular processing can induce reduction (14). However, both UB-2 and UB-2.3 cells produced IFN-γ in response to the fractionated dimeric UBE2V43–51 peptide (fractions 25–27 in Fig. 8,B) even when they were pulsed on the fixed HM-LCL cells (Fig. 9 B). These lines of evidence indicate that these CD4+ T cells recognized the UBE2V43–51 peptide in either the monomeric or dimeric form.
A notable finding of the present study was that vaccination with tumor peptides binding to HLA-A2 molecules resulted in the in vivo induction of HLA-DR4-restricted CD4+ T cells in an HLA-A2+ lung cancer patient. After observing the dramatic induction of IgG specific to the UBE2V43–51 peptide after peptide vaccination, we investigated the possibility that CD4+ Th cells participated in this induction based on the following. First, in vivo generation of Ag-specific IgG generally requires a cytokine from Th cells (11). Second, marked elicitation of IgG specific to the UBE2V43–51 peptide was observed, but there was no apparent induction of IgG specific to the other three peptides. This finding indicates that the dramatic induction of IgG specific to the UBE2V43–51 peptide was not the result of nonspecific stimulation of B cells. In this study we found that HLA-DRB1*0403-restricted and UBE2V43–51 peptide-recognizing CD4+ Th cells were induced in the patient. Although peptides binding to MHC class II molecules have been suggested to be 12–25 aa in length, the core sites anchored to MHC class II molecules are sufficient even at a length of only about nine amino acids (15). The amino acid sequence of the UBE2V43–51 peptide could conform to the motif for HLA-DRB1*0403 molecules. Judging from the binding motif of HLA-DRB1*0406 molecules (16), although this motif is different by one amino acid from that of HLA-DRB1*0403 molecules, the leucine at the firstt position and serine at the sixth position are thought to be the amino acids anchored to the HLA-DRB1*0403 molecules. Several researchers have reported that CD4+ T cells can recognize nine or 10 aa in the context of MHC class I molecules (17, 18), and long peptides recognized by both CD8+ T cells and CD4+ T cells have been identified (19, 20). Nevertheless, the present study provides the first evidence that vaccination with class I binding peptides can induce HLA-DR4-restricted CD4+ T cells in vivo, and that a nonamer peptide can be recognized by both Ab and T cells with either CD8 or CD4 molecules.
Another notable finding is that a dimeric form of the nonamer peptide was recognized by HLA-DR4-restricted CD4+ T cells. Because the UBE2V43–51 peptide has a cysteine at the sixth position and cysteine can be easily oxidized as a result of disulfide bondage, we investigated the influence of dimerization on recognition by CD4+ T cells. MHC class I binding peptides containing cysteine have been reported to decrease immunogenicity for T cells as a result of either cysteinylation or dimerization (17, 21). In the present study we demonstrated that both monomeric and dimeric peptides could be recognized by HLA-DR-restricted CD4+ T cells. Although the precise structure of the complex formed by binding of dimeric peptides to HLA-DR molecules is unknown, it may be that the dimeric peptides lie in parallel on HLA-DR molecules. Thus, recognition of peptide-class II complexes by CD4+ T cells might be not as rigid as recognition of peptide-class I complexes by CD8+ T cells.
Although it was important to determine whether UBE2V43–51 peptide-recognizing CD4+ T cells or UBE2V peptide-reactive IgG can respond to tumor cells, HLA-DRB1*0403-expressing epithelial cancer cell lines were not available. Therefore, we determined whether the UBE2V43–51 peptide-recognizing CD4+ T cells could recognize autologous dendritic cells, which were pulsed with lysates from panc-1 cells, from which the UBE2V gene was cloned (6). However, no response was observed (data not shown). In regard to Ab, we attempted to determine, by two different methods, whether the IgG contained in the postvaccination serum of this patient could bind to the UBE2V43–51 peptide-pulsed HM-LCL cells. In flow cytometric analysis, the HLA-DRB1*0403-expressing HM-LCL cells, which were prepulsed with the UBE2V43–51 peptide, then cultured with postvaccination serum, followed by anti-human IgG second Ab, were not positively stained (data not shown). In the cytotoxicity assay, the UBE2V43–51 peptide-pulsed HM-LCL cells were not killed by freshly isolated PBMCs in the presence of the heat-inactivated serum, suggesting the absence of Ab-dependent cell-mediated cytotoxicity (data not shown). We further investigated the possibility that vaccine-induced IgG reactive to the UBE2V43–51 peptide merely recognized cysteine-based modification of the UBE2V peptide. We determined whether substitution of the sixth cysteine for alanine or serine could have any effect on recognition by IgG, and we observed no remarkable change in Ab titers against these two modified UBE2V peptides (data not shown). This indicates that the UBE2V43–51 peptide-reactive IgG did not recognize cysteine-based modification of the UBE2V43–51 peptide. We obtained no evidence that anti-UBE2V43–51 peptide IgG was tumor reactive. We speculate that the induction of IgG reactive to the UBE2V43–51 peptide was occurred because the UBE2V peptide is a good epitope not only for B cells, but also for CD4+ T cells.
In most protocols involving peptide-based vaccines, the induction of peptide-specific CTLs has been examined after vaccination, without consideration of whether the peptide-specific CTL precursors were pre-existing in cancer patients. We have recently conducted clinical trials in which cancer patients were vaccinated with peptides (maximum of four) to which pre-existing CTL precursors in the periphery were confirmed before vaccination (22, 23). In this protocol the profiles of the vaccinated peptides varied among patients, and the UBE2V43–51 peptide was vaccinated into three HLA-A2+ cancer patients (one with thyroid cancer, one with mastocarcinoma, and one with seminoma) as one of four peptides (22). As a result, anti-UBE2V43–51 peptide IgG was induced in all three patients after the third vaccination. This means that the dramatic induction of IgG reactive to administered peptides is not limited to the present case. In addition, peptide vaccination resulted in the induction of IgG reactive to several tumor peptides other than the UBE2V43–51 peptide (22, 23). This indicates that the elicitation of IgG reactive to administered peptides is not limited to the UBE2V43–51 peptide. In addition, we found that vaccine-induced, peptide-specific IgG tended to detectable in advanced lung cancer patients with a long progression-free survival (23). Although many clinical trials of peptide-based immunotherapy have been conducted to induce CD8+ T cell-mediated tumor regression, no apparent correlation has been observed between in vivo tumor regression and the induction of peptide-specific CTLs in the periphery (3, 24). Our findings imply that the participation of class II-restricted CD4+ Th cells should be considered in immunotherapy with class I binding tumor peptides, and that kinetic measurement of peptide-specific IgG could help to understand immune responses in vaccinated patients.
It is important to kinetically estimate the frequency of CD8+ T cell or CD4+ T cell precursors reactive to the UBE2V43–51 peptide after the peptide vaccination. We estimated the frequency of peptide-specific T cells. The results showed that HLA-A2-restricted CD8+ T cell precursors against the UBE2V43–51 peptide were present in the prevaccination PBMCs, but their frequency increased after the 6th and 12th peptide vaccinations. In contrast, HLA-DRB1*0403-restricted CD4+ T cell precursors against the UBE2V43–51 peptide were detected only after the peptide vaccination. In our clinical trial, vaccination with four peptides, including the UBE2V43–51 peptide, did not increase but rather abolished the pre-existing CTL precursors in the periphery in three HLA-A2+ cancer patients (22), although the peptide vaccination induced UBE2V43–51 peptide-specific IgG in these three patients. We have no explanation for the abolishment of peptide-specific CTL precursors in these patients, but we speculate that the concomitant induction of Th cells contributed to the increase in HLA-A2-restricted T cell response to the UBE2V peptide, at least in this particular patient.
The dominant HLA-A2 subtype in Caucasians is HLA-A*0201, whereas there is no dominant HLA-A2 subtype in Asians (25). The above-described patient was positive for HLA-A*0201 molecules. The UBE2V peptide was originally identified using HLA-A2-restricted and tumor-reactive CTLs from a colon patient (HLA-A*0207/3101) by cDNA library screening (6). The UBE2V peptide was prepared based on the binding motif to the HLA-A*0201 molecules. This CTL line produced IFN-γ in response to the T2 cells (HLA-A*0201) that were prepulsed with the UBE2V peptide and showed cytotoxicity against carcinoma cells, which express either HLA-A*0201 or HLA-A*0207 molecules (6). These results indicate that the UBE2V43–51 peptide could be presented by several HLA-A2 subtypes.
There remains the question of why vaccination with the UBE2V43–51 peptide induced DR4-restricted CD4+ T cells in the patient. We propose the following explanations. First, the amino acid sequence of this peptide just matched the binding motif to HLA-DRB1*0403 molecules. Alternatively, the cysteine-based modification may have made the UBE2V43–51 peptide resistant to degeneration by protease, and the peptide may have maintained its immunogenicity in vivo. Finally, the cysteine-based modification rendered the UBE2V43–51 peptide an artificial Ag and strongly stimulated the immune responses against it. Although we currently have no clear explanation for our observations, our findings should contribute to our understanding of what kinds of immune response might be elicited in vaccinated patients.
This work was supported in part by grants-in-aid from the Ministry of Education, Science, Sports, and Culture of Japan (no. 12213134, to K.I.); the Ministry of Health and Welfare, Japan (no. H14-trans-002, 11-16, and H12-cancer-004, to K.I.); the Mochida Foundation of Japan; the Naito Foundation of Japan; and the Novartis Foundation of Japan for the Promotion of Science.
Abbreviations used in this paper: CypB, cyclophilin B; TFA, trifluoroacetic acid; UBE2V, ubiquitin-conjugated enzyme variant Kua; WHSC2, Wolf-Hirschhorn syndrome candidate 2 protein.