T-cell receptor Vβ gene usage by T cells reactive with the tumor-rejection antigen SART-1 in oral squamous cell carcinoma

Authors


Abstract

We recently described that the SART-1690–698 peptide could induce HLA-A24-restricted cytotoxic T lymphocytes (CTLs), which recognize the SART-1math image tumor cells from peripheral blood mononuclear cells (PBMCs) of HLA-A24+ cancer patients. In our study, in 5 of 14 HLA-A24+ patients with oral squamous cell carcinomas (SCCs), CTLs could be induced with the SART-1690–698 peptide from the PBMCs. In 2 of the patients from whom the highest CTL activities were induced, the T-cell receptor (TCR) Vβ repertoire expressed by the SART-1690–698-specific CTLs was found to be restricted and multiple Vβ families were predominantly expressed in each patient. Although the predominant Vβ families were different between the 2 patients, Vβ7 was highly and commonly predominant. The same predominant Vβ families were also detected in the tumor-infiltrating lymphocytes (TILs) from each patient, and each Vβ family contained one or more unique T-cell clonotypes. The unique T-cell clonotypes were found to be common between the TILs and SART-1690–698-specific CTLs from each patient, and especially 2 T-cell clonotypes with Vβ7 were identical even in the 2 patients. One of the 2 T-cell clonotypes with Vβ7 was detected in the TILs from 11 of 14 HLA-A24+ patients and another was found in those from 8 of HLA-A24+ patients, while none of 10 HLA-A24 patients demonstrated both T-cell clonotypes. These results strongly suggest that the T-cell clonotypes with Vβ7 are major TCR Vβ genes expressed by SART-1690–698-specific CTLs. Furthermore, autologous tumor cells from one of the HLA-A24+ patients stimulated the PBMCs and regional lymph node cells (LNCs) to expand the same T-cell clonotypes as those in the SART-1690–698-specific CTLs. These results strongly suggest that the SART-1690–698-specific CTLs clearly accumulate in vivo, especially in the TILs, as a consequence of in situ antigenic stimulation by autologous tumor cells. The identification of the unique TCR Vβ genes used by SART-1259-specific CTLs should help to improve the diagnosis of the specific immune response in patients with SART-1math image cancers, especially during anticancer immunotherapy. © 2003 Wiley-Liss, Inc.

Many antigenic peptides recognized by HLA class I-restricted cytotoxic T lymphocytes (CTLs) against melanomas have been identified in the past 10 years.1, 2, 3, 4, 5 Some of them are presently being investigated in clinical trial as cancer vaccines and have led to major tumor regression in some patients with melanomas.6, 7, 8 These tumor-rejection peptides are thus expected to be a new tool for specific immunotherapy to melanomas. In squamous cell carcinomas (SCCs), one of the major cancers in humans, tumor-infiltrating lymphocytes (TILs) are quite commonly observed9 and specific immunotherapy with the tumor-rejection antigens is thus also expected as a new treatment modality for patients with SCCs.10 We recently identified a SART-1 gene encoding both the 43 kD of SART-1259 antigen and the 125 kD of SART-1800 antigen.11, 12 The SART-1259 antigen is preferentially expressed in the cytosol of a majority of SCCs and some adenocarcinomas, but not in other types of cancers or any normal cells except for those of the testis, while the SART-1800 antigen is expressed in the nucleus of all proliferating cells. Therefore, the SART-1259 antigen is expected to be a source of tumor-rejection peptides recognized by T cells, while the SART-1800 antigen is not. More recently, we also identified several nonapeptides from the SART-1259 antigen that are capable of inducing HLA-A26- or A24-restricted and tumor-specific CTLs from peripheral blood mononuclear cells (PBMCs) of patients with esophageal, lung, breast, brain, gynecologic and renal cancers,12, 13, 14, 15, 16, 17 and some of the nonapeptides have already been used in clinical trials for specific immunotherapy of HLA-A26+ or -A24+ patients with SCCs and adenocarcinomas as a cancer vaccine.18

These antigenic peptides complexed with HLA molecules are recognized by specific T-cell receptors (TCRs) on CTLs. The TCR is composed of 2 disulfide-linked chains that are either αβ or γδ chains. Each TCR chain includes a variable (V) and a constant (C) region. During T-cell differentiation, unique V region genes are constructed by a recombination of V, diversity (D) and joining (J) segments for the β and γ chains, and of V and J segments for the α and δ chains, in a way similar to immunoglobulin gene rearrangement.19 In addition to the combinational variability, the addition of extra N nucleotides between the segments during the recombination process increases TCR diversity. The region encoded by the hypervariable V-J or V-D-J junctions is defined as the complementarity-determining region 3 (CDR3) and is considered to be essential for binding to the antigenic peptide. Recently, the availability of nucleotide sequences for many of the TCR Vβ gene segments has facilitated the investigation of the TCR Vβ clonotypes with unique CDR3 by a single-strand conformation polymorphism (SSCP) analysis, as well as TCR Vβ repertoire by a PCR-based analysis. A restricted TCR Vβ gene usage and unique CDR3 sequences have thus been found in the TILs of several cancers (reviewed in reference 20) including SCCs.10

In our study, we examined the TCR Vβ genes expressed by SART-1259-specific CTLs, which were induced from PBMCs of HLA-A24+ patients with oral SCC by in vitro stimulation with the nonapeptide, in order to identify the unique TCR Vβ genes. Secondly, the TCR Vβ genes usage was compared to that by TILs in autologous tumor tissues and CTLs induced from PBMCs and regional lymph node cells (LNCs) by in vitro stimulation with autologous SCC cells, to confirm the presence of SART-1259-specific CTLs in vivo. The identification of the unique TCR Vβ genes used by SART-1259-specific CTLs is expected to make it easier to quantify and therefore to diagnose the SART-1259-specific immune response in patients with SART-1259-positive cancers, especially during SART-1-based anticancer immunotherapy.

Abbreviations

C, constant; CDR3, complementarity-determining region 3; CTL, cytotoxic T lymphocyte; D, diversity; E/T, effector-to-target; J, joining; LNC, lymph node cell; PBMC, peripheral blood mononuclear cell; SCC, squamous cell carcinoma; SSCP, single-strand conformation polymorphism; SSPE, saline sodium phosphate EDTA; TCR, T-cell receptor; TIL, tumor-infiltrating lymphocyte; V, variable.

MATERIAL AND METHODS

Patients and samples

Serologic HLA class I typing of PBMCs was performed in 105 patients with primary oral SCCs who were referred to the Department of Oral and Maxillofacial Surgery, Kyushu University Dental Hospital, and 16 (15.2%) and 51 (48.6%) of the patients were determined to be HLA-A24-homozygous and -heterozygous, respectively. Tumor biopsies and heparinized peripheral blood samples were obtained at the same time from 8, 6 and 10 of the HLA-A24-homozygous, HLA-A24-heterozygous, and HLA-A24 patients, respectively. From the 8 HLA-A24-homozygous and 6 HLA-A24-heterozygous patients, heparinized peripheral blood samples were obtained at least 2 more times. WK2 SCC and fibroblast cell lines were established from a biopsy specimen of one of the HLA-A24-heterozygous patients, and the regional lymph nodes and a heparinized peripheral blood sample were obtained from the patient at the time of a neck dissection. None of the patients had received any anticancer treatment prior to the biopsy.

Cell lines

A KE4 SCC cell line was established from patients with esophageal SCC in our laboratory as reported previously.11 QG56 lung SCC and VA13 fibroblast cell lines were kindly provided by Dr. K. Yasumoto (University of Occupational and Environmental Health, Kitakyushu, Japan) and Dr. H. Saya (Kumamoto University School of Medicine, Kumamoto, Japan), respectively. WK2 SCC and fibroblast cell lines were established from an HLA-A24-heterozygous patient with oral SCC as mentioned above. HLA alleles and the SART-1 expression of the tumor cell lines were determined as reported previously.21

Induction of CTLs by nonapeptides

The method for both CTL induction by the peptide and the estimation of the CTL activity has been described previously.12, 13, 15, 16, 17 In brief, the PBMCs (2 × 106) were incubated with 10 mM of a peptide in 1 well of a 24-well plate containing 2 ml of culture medium (45% RPMI-1640 medium, 45% AIM-V medium (Invitrogen, Carlsbad, CA), 10% FCS (Equitech Bio, Ingram, TX), 100 U/ml of recombinant IL-2 (Shionogi Co., Osaka, Japan)). On days 7 and 14 of culture, the cells were harvested, washed and reincubated with the irradiated (50 gray) autologous PBMCs acting as antigen-presenting cells that had been preincubated with the same peptide at the same dose for 2 hr. The ratio of responder to stimulator cells was 10:1. The peptides used in our study were SART-1690–698 (EYRGFTQDF) peptide, which is one of the most potent peptides reported to induce HLA-A24-restricted and SART-1-specific CTLs, and HIV peptide (RYLRDQQLLGI), which is capable of binding to HLA-A24 as a negative control.13 These peptides were kindly provided by Dr. M. Kanaoka (Sumitomo, Osaka, Japan) and the purity was 95%. Effector cells were harvested on day 21 of culture and were immediately tested for the ability to produce IFN-γ by an ELISA and the cytotoxic activity by trypan blue exclusion, in response to various target cells for 18 hr at various effector-to-target cell (E/T) ratios of 10:1, 5:1 and 2.5:1 in triplicate assays. ELISA was performed using a kit (Japan Immunoresearch Laboratories Co., Gunma, Japan) and the limit of sensitivity was 5 pg/ml. The well was considered positive if it contained effector cells exhibiting higher levels of both IFN-γ production and cytotoxicity in response to KE4 cells (HLA-A2402/A2601, SART-1math image and SART-1math image) compared to those activities in response to QG56 cells (HLA-A2601, SART-1math image and SART-1math image) and VA13 cells (no expression of HLA-A alleles, SART-1math image and SART-1math image), and also compared to those activities of the cells incubated with the control peptide in response to KE4 cells.

Induction of CTLs by autologous SCC cell lines

The PBMCs and LNCs were prepared from an HLA-A24-heterozygous patient with oral SCC from whom WK2 cells had been established, and were incubated in 96-well round-bottomed microplates (5 × 104 cells/well) in culture medium (45% RPMI-1640 medium, 45% AIM-V medium, 10% FCS) supplemented with 10 μg/ml of PHA (Sigma, St. Louis, MO), in the presence of 10 μM of the SART-1690–698 peptide or HIV peptide. Three days after the initial culture, the medium was changed to a culture medium supplemented with 100 U/ml of recombinant IL-2. The proliferating microcultures were thereafter transferred into 96-well flat-bottomed microplates. Stimulation with the same peptide at the same dose was then repeated on days 3, 7, 10, 14 and 17 of culture, and autologous WK2 cells pretreated with 20 μg/ml of mitomycin C at 37°C for 1 hr (2 × 103 cells/well) were added to half of the microcultures on days 7 and 14. Effector cells were harvested on day 21 of culture and then were immediately tested for the ability to produce IFN-γ and the cytotoxicity in response to WK2 and KE4 cells at E/T ratios of 10:1 and 5:1 in triplicate assays. Autologous fibroblast cells were used as a negative control for target cells.

RNA extraction and cDNA synthesis

Total RNA was prepared from the PBMCs, LNCs, and tumor specimens by the acidified guanidinium-phenol-chloroform method as previously described.22, 23 Three to 5 μg of the total RNA preparation was used for the synthesis of cDNA. Briefly, RNA was incubated for 1 hr at 37°C with 20 U of RNasin ribonuclease inhibitor (Promega, Madison, WI), 0.5 μg of oligo(dT)12, 13, 14, 15, 16, 17, 18 (Pharmacia, Uppsala, Sweden), 0.5 mM of each dNTP (Pharmacia), 10 mM of DTT and 100 U of RNase H reverse transcriptase (Invitrogen).

Semiquantitative PCR analysis of TCR Vβ gene usage

The cDNA was amplified and quantified according to previously described methods,10, 22, 24 with only slight modifications. Briefly, one-twenty-fifth of the TCR β chain cDNAs were amplified by PCR with the Vβ- and Cβ-specific primers, at a final concentration of 0.4 mM each. The sequences of the specific primers were the same as previously described.10, 22 In the present study, Vβ10- and Vβ19-specific primers were omitted from a panel of primers specific for the Vβ gene segment subfamily (Vβ1– Vβ24), since Vβ10 and Vβ19 have been reported to be nonfunctional genes that cannot be expressed at the protein level.25, 26, 27 The amplification was performed with 2.5 U of Taq DNA polymerase and the T3 Thermocycler (Biometra, Göttingen, Germany) under the following conditions: denaturing at 94°C, annealing at 55°C, and extension at 72°C for 30 sec each. For a quantitation of the Vβ transcripts, the aliquots of PCR samples were harvested every 3 cycles after 24-cycle amplification. Those PCR products were electrophoresed through 1.8% agarose gel, transferred to NYTRAN-N (Schleicher & Schuell, Dassel, Germany) and hybridized with a biotinylated internal Cβ probe (5′-A(AC)AA(GC)GTGTTCCCACCCGAGGTCGCTGTGTT-3′). After hybridization was done for 12 hr at 42°C in 2× saline sodium phosphate EDTA (SSPE), 5× Denhart's solution and 0.5% SDS, the filters were washed for 10 min at 48°C in 0.2× SSPE and 0.5% SDS. The PCR products were visualized by subsequent incubation with streptavidin, biotinylated alkaline phosphatase and a chemiluminescent substrate system (Phototope-Star detection kit, New England Biolab, Beverly, MA). After visualization, the intensities of Vβ-specific bands were quantified by an NIH Image analysis, and the appropriate cycles within an exponential phase of amplification for all the Vβ transcripts were determined. The semiquantitative PCR results were expressed as follows: %Vβ = 100 × (intensity of a Vβ-specific band)/(sum of intensities of all Vβ-specific bands), according to previously described methods.24, 28

SSCP analysis of TCR Vβ genes

An SSCP analysis was performed as previously described.24, 29, 30 Briefly, amplified DNA was diluted (1:20) in a denaturing solution (95% formamide, 10 mM EDTA, 0.1% bromophenol blue and 0.1% xylene cyanol) and heated to 90°C for 2 min. The diluted samples were then applied to nondenaturing 4% polyacrylamide gels containing 10% glycerol and electrophoresed at a 35 W constant power for approximately 2 hr. After electrophoresis, the DNA was transferred to a GeneScreen membrane (NEN™ Lifesciences, Boston, MA). For visualization of the DNA, the membrane was hybridized with the biotinylated internal Cβ probe and detected by subsequent incubation with streptavidin, biotinylated alkaline phosphatase and the chemiluminescent substrate system. The accumulated clone was counted if a clear band was identified on the background smear. The total number of clones was then determined to represent the mean value of triplicate independent counts.

Statistical analysis

The statistical significance of the differences between the groups was determined by either post hoc multiple comparison test by Bonferroni and Dunn or chi-square test. p values less than 0.05 were considered to be significant.

RESULTS

In vitro induction of CTLs from PBMCs by the SART-1690–698 peptide

PBMCs from 14 HLA-A24+ patients with oral SCC were stimulated in vitro 3 times with a SART-1690–698 peptide, one of the most potent nonapeptides in inducing HLA-A24-restricted and SART-1259-specific CTLs,13 and the induction of HLA-A24-restricted and SART-1690–698-specific CTL activity was estimated by IFN-γ production and cytotoxicity (Table I). In 5 (patients 1, 6, 9, 10 and 13) of the 14 patients, higher levels of IFN-γ production and cytotoxicity were significantly induced by the recognition of KE4 (HLA-A24+, SART-1math image and SART-1math image) compared to that of either QG56 (HLA-A24, SART-1math image and SART-1math image) or VA13 (HLA-A24, SART-1math image and SART-1math image). The increased IFN-γ production in these 5 patients was dependent on the increased number of effector cells (Fig. 1). These results indicate that HLA-A24-restricted and SART-1690–698-specific CTLs were satisfactorily induced in these 5 patients. Although not significant, in 7 and 6 of the remaining 9 patients, higher levels of IFN-γ production and cytotoxicity, respectively, were observed in response to KE4 compared to those to QG56 and VA13, thus suggesting the induction of HLA-A24-restricted and SART-1690–698-specific CTLs. Furthermore, even in the PBMCs cultured with the control peptide, higher levels of IFN-γ production and cytotoxicity were observed in 11 and 9 of the 14 patients in response to KE4 compared to those to QG56 and VA13, thus suggesting the preexistence of HLA-A24-restricted and SART-1690–698-specific CTLs, probably due to in situ antigenic stimulation.

Table I. In Vitro Induction of CTL Activities from PBMCS by the SART-1690–698 Peptide1
PatientHLA-AIFN-γ (pg/ml) production in response to% specific cytotoxicity toSignificance
KE4QG56VA13KE4QG56VA13
  • 1

    PBMCs from 14 HLA-A24+ patients with oral SCC were incubated with 10 μM of the SART-1690–698 peptide. At days 7 and 14 of culture, the cells were collected, washed and stimulated with irradiated autologous PBMCs as APCs, which had been preincubated with the same nonapeptide for 2 hr. The cells were then harvested at day 21 of culture and tested for their ability to produce IFN-γ and cytotoxic activity in response to KE4 (HLA-A24+, SART-1math image and SART-1math image), QG56 (HLA-A24, SART-1math image and SART-1math image) and VA13 (HLA-A24, SART-1math image and SART-1math image) cells at E/T ratio of 5:1 in triplicate assays. Mean values of triplicate assays are shown, and background IFN-γ production and cytotoxicity by the effector cells alone were subtracted. Background IFN-γ production and cytotoxicity by the cells incubated with control peptide are shown in parentheses. Deviation of the triplicate determinations were usually <20% of mean values. The CTL induction was considered significant if the cells incubated with the SART-1690–698 peptide contained effector cells showing significantly higher levels of both IFN-γ production and cytotoxicity in response to KE4 cells, compared to those activities of all the following cells; the cells incubated with the SART-1690–698 peptide in response to QG56 and VA13 cells and the cells incubated with the control peptide in response to KE4, QG56 and VA13 cells (post-hoc multiple comparison test by Bonferroni and Dunn).

  • NS, not significant.

124166 (60)49 (34)37 (25)37 (14)10 (6)5 (4)p < 0.01
22447 (84)<0 (5)<0 (14)<0 (10)<0 (0)<0 (10)NS
32449 (50)29 (37)33 (15)27 (24)21 (25)18 (13)NS
42453 (128)9 (68)1 (92)25 (36)<0 (21)<0 (25)NS
5241 (1)12 (2)1 (1)<0 (0)<0 (0)<0 (0)NS
624235 (170)144 (74)104 (139)90 (78)72 (56)71 (75)p < 0.05
724106 (76)34 (18)136 (96)82 (70)44 (28)57 (43)NS
824739 (465)577 (311)576 (333)76 (52)63 (40)48 (36)NS
92/24454 (264)148 (157)171 (142)67 (38)30 (33)35 (26)p < 0.02
102/24134 (61)16 (7)10 (23)29 (15)4 (0)<0 (5)p < 0.05
112/2412 (11)<0 (0)<0 (11)<0 (0)<0 (0)<0 (0)NS
122/2445 (28)<0 (3)4 (6)<0 (0)<0 (0)<0 (0)NS
1311/24327 (128)103 (93)121 (117)57 (31)29 (26)35 (32)p < 0.01
1411/2474 (98)42 (72)32 (66)23 (37)9 (15)6 (24)NS
Figure 1.

Dose dependence of CTL activities induced from PBMCs by the SART-1690–698 peptide. PBMCs from patients 1, 9 and 13 were stimulated with the SART-1690–698 peptide, and the induced CTL activities were studied at different E/T ratios in triplicate assays, as described in Table I. Mean values of the triplicate assays are shown. Similar results were obtained from patients 6 and 10.

TCR Vβ genes used by SART-1690–698-specific CTLs and TILs in SCCs

To identify TCR Vβ genes used by HLA-A24-restricted and SART-1690–698-specific CTLs, we first examined TCR Vβ gene usage in the PBMCs stimulated with the SART-1690–698 peptide and the control peptide from 2 patients (patients 1 and 13). Representative visualized bands and %Vβ in patient 1 are shown in Figure 2a and b, respectively. The Vβ families expressed in the fresh PBMCs were diverse in both patients 1 and 13, and the number of detected Vβ families was 22 and 21, respectively. After cultivation with the control peptide, the PBMCs expressed slightly restricted numbers of Vβ families, and the numbers of detected Vβ families were 14 and 20, respectively. In contrast, the Vβ repertoire in the PBMCs stimulated with the SART-1690–698 peptide was apparently restricted in both patients, and the number of detected Vβ families was 10 and 19, respectively. Stimulation with the SART-1690–698 peptides thus induced more restricted Vβ repertoire than that with the control peptide. The extent of Vβ gene usage in each family was then compared between the PBMCs stimulated with the SART-1690–698 peptide and those with the control peptide. The families in which the %Vβ increased by more than 25% in the PBMCs stimulated with the SART-1690–698 peptide in comparison to those with the control peptide were defined as predominant because the PCR-based analysis performed in this experiment permits only a semiquantitative assessment. As shown in Table II, Vβ6, Vβ7 and Vβ13.2 in patient 1 and Vβ2, Vβ3, Vβ6, Vβ7, Vβ12 and Vβ14 in patient 13 were predominant in the PBMCs stimulated with the SART-1690–698 peptide. These results thus suggest that SART-1690–698-specific CTLs express these multiple Vβ families.

Table II. TCR Vβ Gene Usage in PBMCS Stimulated with the SART-1690–698 Peptide and TILS1
PatientEffector cellsVβ families
12345.15.26789111213.113.214151617182021222324
  • 1

    PBMCs and TILs were obtained from 2 HLA-A24+ patients with oral SCC (patients 1 and 13) and the stimulation of the PBMCs with the SART-1690–698 peptide was as mentioned in Table I. The relative amounts of Vβ transcripts were estimated as described in Figure 2. The %Vβ of each family in the PBMCs stimulated with the SART-1690–698 peptide and TILs was compared with that of the same family in the PBMCs cultured with the control peptide and fresh PBMCs, respectively.–equation image, The increase of %Vβ in the PBMCs stimulated with the SART-1690–698 peptide and TILS was more than 50% when compared with that of the PBMCs cultured with the control peptide and fresh PBMCs, respectively.–↑, The increase of %Vβ in the PBMCs stimulated with the SART-1690–698 peptide and TILS was 25% to 50%.–No mark, The increase of %Vβ in the PBMCs stimulated with the SART-1690–698 peptide and TILS was less than 25%.

Patient 1TIL    equation image equation image              
PBMC with SART-1690–698      equation imageequation image     equation image          
Patient 13TIL equation imageequation image    equation image       equation image    
PBMC with SART-1690–698    equation image     equation image         

TCR Vβ gene usage in the tumor specimens containing TILs from the 2 patients was also examined and compared to that in the fresh PBMCs (Fig. 2 and Table II). The Vβ repertoire in the TILs was apparently restricted in both patients, and the numbers of detected Vβ families were 19 and 18, respectively. Vβ5.1, Vβ5.2, Vβ7, Vβ12 and Vβ16 in patient 1 and Vβ2, Vβ3, Vβ7, Vβ9, Vβ12, Vβ13.1, Vβ14 and Vβ20 in patient 13 were predominant in the TILs in comparison with the fresh PBMCs. As a result, Vβ7 in patient 1 and Vβ2, Vβ3, Vβ7, Vβ12 and Vβ14 in patient 13 were predominant in both the TILs and PBMCs stimulated with the SART-1690–698 peptide. These results thus suggest that SART-1690–698-specific CTLs had been accumulated in TILs as a consequence of in situ antigenic stimulation. Interestingly, in these predominant families in both the TILs and PBMCs stimulated with the SART-1690–698 peptide, Vβ7 was common between the 2 patients.

Figure 2.

The expression of TCR Vβ gene transcripts in PBMCs stimulated with the SART-1690–698 peptide and the control peptide, fresh PBMCs and TILs. (a) Visualized bands of PCR products for each Vβ family. PBMCs and TILs were obtained from an HLA-A24+ patient with oral SCC (patient 1), and the stimulation of the PBMCs with the SART-1690–698 peptide was performed as described in Table I. The Vβ transcripts amplified by PCR through 27 cycles were loaded onto 1.8% agarose gel, transferred to a filter and then hybridized with a biotinylated internal Cβ probe, as outlined in Material and Methods. (b) The relative amounts of Vβ transcripts are indicated as a percentage, calculated as described in Material and Methods.

We next examined the TCR Vβ clonotypes in the PBMCs stimulated with the SART-1690–698 peptide, those with the control peptide, TILs and fresh PBMCs from the 2 patients by an SSCP analysis, and representative visualized bands are shown in Figure 3. In both of the patients, most of the PCR products from the fresh PBMCs developed as a smear, whereas many distinct bands, which indicate either expanding or accumulating T-cell clonotypes, are detected in the PBMCs stimulated with the SART-1690–698 peptide, those with the control peptide and TILs. The total number of distinct bands in all Vβ families from the PBMCs stimulated with the SART-1690–698 peptide was smaller than that from the PBMCs with the control peptide, while that from TILs was higher than that from the fresh PBMCs. The total numbers of distinct bands in all Vβ families from the PBMCs stimulated with the SART-1690–698 peptide, those with the control peptide, TILs and fresh PBMCs were 33, 100, 84 and 62 in patient 1 and 49, 67, 109 and 85 in patient 13, respectively. Distinct bands whose migration patterns are identical in the TILs and PBMCs stimulated with the SART-1690–698 peptide were found in Vβ5.2, Vβ6, Vβ7, Vβ12 and Vβ14 transcripts from patient 1 and Vβ2, Vβ3, Vβ6, Vβ7, Vβ12 and Vβ14 transcripts from patient 13 (Fig. 4), thus suggesting that these unique T-cell clonotypes are involved in the recognition of the SART-1690–698 peptide and expand in the TILs as a consequence of in situ antigenic stimulation. Interestingly, 1 or 2 of these unique T-cell clonotypes in each of Vβ6, Vβ7, Vβ12 and Vβ14 transcripts showed the identical migration pattern in the 2 patients, suggesting that these T-cell clonotypes are common between the 2 patients.

Figure 3.

An SSCP analysis of TCR Vβ gene transcripts in PBMCs stimulated with the SART-1690–698 peptide and the control peptide, fresh PBMCs and TILs. PBMCs and TILs were obtained from an HLA-A24+ patient with oral SCC (patient 1), and the stimulation of the PBMCs with the SART-1690–698 peptide was performed as described in Table I. PCR products of Vβ transcripts were denatured into single strands and electrophoresed in a nondenatured polyacrylamide gel by the SSCP method. See Material and Methods for a detailed description of all procedures.

Figure 4.

Accumulation of common T-cell clonotypes in several Vβ family transcripts from PBMCs stimulated with the SART-1690–698 peptide and TILs. PBMCs and TILs were obtained from 2 HLA-A24+ patients with oral SCC—(a) patient 1; (b) patient 13—and the stimulation of the PBMCs with the SART-1690–698 peptide was as mentioned in Table I. PCR products of Vβ transcripts were analyzed by the SSCP method as described in Figure 3. Visualized bands from PCR products of Vβ transcripts that contain common T-cell clonotypes between the TILs and PBMCs stimulated with the SART-1690–698 peptide in each patient (arrowheads), and identical T-cell clonotypes even in the 2 patients (open arrowheads) are shown. Lanes: lane 1, fresh PBMCs; lane 2, TILs; lane 3, PBMCs cultured with the control peptide; lane 4, PBMCs stimulated with the SART-1690–698 peptide.

Accumulation of common T-cell clonotypes with Vβ7 in TILs

Taken together with the results mentioned above, Vβ7 is supposed to be one of the major Vβ families containing the unique T-cell clonotypes expressed by SART-1690–698-specific CTLs. We thus examined Vβ7 gene usage in the TILs from the 14 HLA-A24+ and 10 HLA-A24 patients with oral SCC (Fig. 5). Vβ7 was more predominant in the TILs in comparison with the fresh PBMCs in 12 of the 14 HLA-A24+ patients and 3 of the 10 HLA-A24 patients (p < 0.05; chi-square test), although data are not shown. Two of the 3 bands shown in Figure 4, each of which showed the identical migration pattern, were frequently detected in the HLA-A24+ patients (p < 0.05; chi-square test). One was visible in 11 and another was in 8 of the 14 HLA-A24+ patients, while both bands were visible in none of the 10 HLA-A24 patients. These results suggest that the T-cell clonotypes with Vβ7, which are possibly involved in the recognition of the SART-1690–698 peptide, are often accumulations in the TILs of HLA-A24+ patients with oral SCC.

Figure 5.

Accumulation of common T-cell clonotypes in Vβ7 transcripts from TILs. TILs were obtained from 14 HLA-A24+ and 10 HLA-A24 patients with oral SCC, and PCR products of Vβ transcripts were analyzed by the SSCP method as described in Figure 3. Representative results from 7 HLA-A24+ and 3 HLA-A24 patients are shown, and 2 distinct bands from the PCR products of Vβ7 transcripts with identical migration patterns (open arrowheads) are frequently visible; bands 1 and 2 are visible in 5 and 4 of the 7 HLA-A24+ patients, respectively. Lanes: lanes 1–7, in order, patients 1, 4, 6, 9, 10, 11 and 13 of HLA-A24+ patients; lanes 8–10, HLA-A24 patients.

In vitro induction of CTLs from PBMCs and LNCs with the SART-1690–698 peptide and autologous SCC cells

In order to confirm that SART-1690–698-specific CTLs expand as a consequence of in situ antigenic stimulation, we established a WK2 SCC cell line from the biopsy of patient 13 and examined the CTL induction from the PBMCs and LNCs by in vitro stimulation with autologous WK2 cells (Table III). The stimulation of the PBMCs and LNCs with either the SART-1690–698 peptide or WK2 cells induced CTL activities in response to both WK2 and KE4 cells, and a combination of WK2 cells and the SART-1690–698 peptide induced the highest CTL activities. The stimulation with WK2 cells induced higher levels of CTL activities in response to WK2 cells, while it induced lower activities in response to KE4, than that with the SART-1690–698 peptide. These results suggest that the stimulation with WK2 cells induced CTL activities in response to other SCC-related antigens, of which either some are not expressed on KE4 cells or such expression is greatly reduced, as well as SART-1259-specific CTL activities. It is also interesting to note that the LNCs developed higher CTL activities in response to either WK2 or KE4 cells than the PBMCs.

Table III. In Vitro Induction of CTLS from PBMCS and LNCS by the SART-1690–698 Peptide and Autologous SCC Cells1
Effector cellsIFN-γ (pg/ml) production in response to% specific cytotoxicity to
WK2KE4FibroblastWK2KE4Fibroblast
  • 1

    PBMCs and LNCs from a HLA-A24+ patients with oral SCC were incubated with PHA in the presence of 10 μM of SART-1690–698 peptide or the control peptide. At days 3, 7, 10, 14 and 17 of culture, the medium was changed to that containing IL-2 and the same nonapeptide at the same dose. At days 7 and 14 of culture, mitomycin C-treated autologous WK2 cells were added into a half of the cultures. The cells were then harvested at day 21 of culture and tested for their ability to produce IFN-γ and cytotoxic activity in response to WK2, KE4 and autologous fibroblast cells at E/T ratio of 10:1 in triplicate assays. Mean values of triplicate assays are shown, and background IFN-γ production and cytotoxicity by the effector cells alone were subtracted. Deviation of the triplicate determinations were usually <20% of mean values. Results at E/T ratio of 5:1 were shown in parenthesis.

PBMCs53 (35)43 (25)32 (27)15 (11)11 (9)12 (7)
PBMCs with SART-1690–698162 (86)129 (72)38 (31)35 (20)29 (13)14 (8)
PBMCs with WK2223 (156)103 (66)37 (32)55 (32)26 (10)5 (4)
PBMCs with WK2 + SART-1690–698277 (192)165 (91)38 (29)67 (40)42 (26)3 (3)
LNCs98 (42)78 (32)45 (29)30 (17)19 (10)14 (8)
LNCs with SART-1690–698226 (153)176 (83)48 (33)57 (31)30 (13)15 (7)
LNCs with WK2311 (195)111 (77)31 (26)77 (39)27 (11)<0 (<0)
LNCs with WK2 + SART-1690–698356 (224)234 (121)34 (30)89 (48)55 (38)<0 (<0)

To confirm that autologous WK2 cells induced SART-1690–698-specific CTLs, we examined TCR Vβ gene usage in the PBMCs and LNCs stimulated with the SART-1690–698 peptide and WK2 cells (Table IV). The Vβ repertoire in the PBMCs stimulated with WK2 cells and/or the SART-1690–698 peptide was more restricted than that in the PBMCs cultured with the control peptide (data not shown). Regarding predominant Vβ families in comparison to the PBMCs cultured with the control peptide, the stimulation with WK2 cells induced a larger number of predominant Vβ families than that with the SART-1690–698 peptide. The stimulation with a combination of WK2 cells and the SART-1690–698 peptide induced almost the same Vβ repertoire as that induced by the stimulation with WK2 cells. Vβ2, Vβ3, Vβ7, Vβ12 and Vβ14, which were predominant in the PBMCs stimulated with the SART-1690–698 peptide, were also predominant in the PBMCs stimulated with WK2 cells. These Vβ families were more predominant in the PBMCs stimulated with a combination of WK2 cells and the SART-1690–698 peptide. Similar results were obtained with the LNCs. These results thus suggest that autologous WK2 cells induced SART-1690–698-specific CTLs as well as CTLs against other SCC-related antigens, and that T cells expressing Vβ2, Vβ3, Vβ7, Vβ12 and Vβ14 are responsible for the recognition of the SART-1690–698 peptide.

Table IV. TCR Vβ Gene Usage in PBMCS and LNCS Stimulated with the SART-1690–698 Peptide and Autologous SCC Cells
Effector cellsVβ families
12345.15.26789111213.113.214151617182021222324
  1. PBMCs and LNCs from a HLA-A24+ patient with oral SCC (patient 13) were stimulated with the SART-1690–698 peptide and autologous WK2 cells, as described in Table III. The relative amounts of Vβ transcripts were estimated as described in Figure 2. The %Vβ of each family in the PBMCs stimulated with the SART-1690–698 peptide and/or WK2 cells was compared with that of the same family in the PBMCs cultured with the control peptide.–equation image, The increase of %Vβ in the PBMCs stimulated with the SART-1690–698 peptide and/or WK2 cells was more than 50% when compared with that of the PBMCs cultured with the control peptide and fresh PBMCs, respectively.–↑, The increase of %Vβ in the PBMCs stimulated with the SART-1690–698 peptide and TILs was 25% to 50%.–No mark, The increase of %Vβ in the PBMCs stimulated with the SART-1690–698 peptide and TILs was less than 25%.

PBMCs with SART-1690–698     equation image              
PBMCs with WK2 equation image  equation image  equation image equation image    equation image    
PBMCs with WK2 + SART-1690–698 equation imageequation image   equation imageequation image  equation image equation image       
LNCs with SART-1690–698 equation image    equation image              
LNCs with WK2 equation image  equation imageequation image equation image equation image equation image       
LNCs with WK2 + SART-1690–698 equation imageequation image  equation image  equation imageequation image equation image       

We finally examined the T-cell clonotypes with these Vβ families induced in the PBMCs and LNCs stimulated with WK2 cells and/or the SART-1690–698 peptide. As expected, the T-cell clonotypes in Vβ7 transcripts were commonly observed in the PBMCs and LNCs stimulated with WK2 cells and/or the SART-1690–698 peptide (Fig. 6). Although the results are not shown, the T-cell clonotypes in Vβ2, Vβ3, Vβ12 and Vβ14 transcripts were also commonly observed.

Figure 6.

Accumulation of common T-cell clonotypes in Vβ7 transcripts from PBMCs and LNCs stimulated with the SART-1690–698 peptide and/or autologous SCC cells. PBMCs (lanes 3–6) and LNCs (lanes 7–10) from an HLA-A24+ patient with oral SCC (patient 13) were stimulated with the SART-1690–698 peptide and/or autologous WK2 cells, as described in Table III. PCR products of Vβ transcripts were analyzed by the SSCP method as described in Figure 3. Distinct bands from PCR products of Vβ7 transcripts that contain 3 common T-cell clonotypes in the PBMCs and LNCs stimulated with the SART-1690–698 peptide and/or WK2 cells (arrowheads) are shown. For easier comparisons, the results shown in Figure 4 were also included in this figure (lanes 1 and 2). Lanes: lane 1, fresh PBMCs; lane 2, TILs; lanes 3 and 7, cells cultured with the control peptide; lanes 4 and 8, cells stimulated with the SART-1690–698 peptide; lanes 5 and 9, cells stimulated with WK2 cells; lanes 6 and 10, cells stimulated with WK2 cells and the SART-1690–698 peptide.

DISCUSSION

We previously demonstrated that the SART-1690–698 peptide could frequently induce HLA-A24-restricted CTLs, which recognize the SART-1259-positive tumor.13 In our study, the HLA-A24-restricted and SART-1690–698-specific CTLs could be induced from the PBMCs of HLA-A24+ patients with oral SCCs. In 2 of the patients from whom the highest CTL activities were obtained, TCR Vβ genes used by the CTLs were found to be apparently restricted and to contain multiple unique T-cell clonotypes with different Vβ families. The unique T-cell clonotypes were commonly found in the TILs in tumor tissues presumably as a consequence of in situ antigenic stimulation, and also in the CTLs induced from the PBMCs and LNCs by in vitro stimulation with autologous tumor cells. Although the unique T-cell clonotypes differed between the 2 patients, the major T-cell clonotypes with Vβ7 were commonly observed in the 2 patients. Furthermore, 2 of the T-cell clonotypes with Vβ7 were frequently found in the TILs from other HLA-A24+ patients with oral SCC, but not in those from HLA-A24 patients with oral SCC. These results strongly suggest that major HLA-A24-restricted and SART-1690–698-specific CTLs express these T-cell clonotypes with Vβ7. To obtain further evidence, it will therefore be necessary to establish SART-1690–698-specific CTL clones and to determine their TCR gene sequences. The identification of the unique TCR genes expressed by SART-1259-specific CTLs will make it possible to achieve a genetic diagnosis of the CTL response to SART-1259, for example, by performing clonotypic PCR to quantify the TCRs expressed by the CTLs.31 These kinds of examination will provide an essential biologic parameter reflecting the specific T-cell response in patients with SART-1259+ cancer, especially during SART-1259-based anticancer immunotherapy.

In our study, the T-cell clonotypes with Vβ7 were focused, since the Vβ7 was predominantly expressed in samples obtained from 2 examined patients and was supposed to be one of the major Vβ families containing the unique T-cell clonotypes expressed by SART-1690–698-specific CTLs. T-cell clonotypes with other Vβ families should also be examined without being neglected. For further study, we should make every effort to establish SART-1690–698-specific CTL clones and then examine any expressed TCR Vβ genes, as mentioned above. The CTL clones will eventually provide useful information regarding the CTL activities of each clone, frequencies of the precursor and so on.

Shimizu et al.31 previously reported that CTLs induced with the SART-1690–698 peptide from an HLA-A24+ healthy donor express Vβ18 transcripts with the conserved CDR3 sequence. However, no predominant expression of Vβ18 transcripts by SART-1690–698-specific CTLs was observed in the 2 HLA-A24-positive patients with oral SCC examined in our study. Some explanations for such discrepancies in these results include, for example, the fact that SART-1690–698-specific CTLs are likely to utilize multiple TCR Vβ genes but not only one. Even in the 2 patients examined in our study, SART-1690–698-specific CTLs showed a different TCR Vβ repertoire and expressed multiple TCR Vβ genes. This kind of variation may be due to environmental differences from patient to patient. The TCR repertoire in an individual patient is generally affected by various genetic factors, including HLA haplotypes and environmental factors such as infections with microorganisms. Interestingly, the SART-1690–698 peptide shared 7 amino acids with a nonapeptide from a membrane protein of Saccharomyces, a well-characterized nonpathogenic yeast.32 As suspected, this Saccharomyce-derived nonapeptide could induce HLA-A24-restricted CTLs to react to SART-1259-positive tumor cells.13Saccharomyce is present in many different fermented foods and beverages, and therefore, many humans may well be exposed to this membrane protein on a daily basis. As a result, the mechanism of epitope mimics might also be involved in the induction of CTLs by the SART-1690–698 peptide. In any case, we found the T-cell clonotypes with Vβ7, which were possibly expressed by SART-1690–698-specific CTLs from HLA-A24+ patients with oral SCC.

The SART-1259 antigen has originally been identified as an antigen recognized by a tumor-specific CTL cell line. Although SART-1259-specific CTLs could be induced in vitro from PBMCs, the presence of SART-1259-specific CTLs in TILs has not yet been examined. In our study, T-cell clonotypes induced in vitro from the PBMCs with the SART-1690–698 peptide were also found to accumulate in TILs. Furthermore, these T-cell clonotypes were induced from either the PBMCs or LNCs by in vitro stimulation with autologous tumor cells. These results strongly suggest that SART-1259-specific CTLs expand in vivo as a consequence of in situ stimulation by tumor cells. Interestingly, SART-1690–698-specific CTLs could be more strongly induced from the LNCs than from PBMCs, thus suggesting that LNCs contain higher numbers of the CTL precursor than PBMCs. Furthermore, although not significant, the PBMCs and LNCs showed slight SART-1690–698-specific CTL activities without any in vitro stimulation. These results suggest that SART-1259-specific CTLs are present in LNCs and PBMCs, even if less frequently, as well as TILs, probably as a consequence of in situ antigenic stimulation.

Autologous tumor cells successfully induced CTL activities in response to themselves in our study. These CTL activities were higher and accompanied with a larger number of different T-cell clonotypes than those induced by the SART-1690–698 peptide. These results suggest that autologous tumor cells induced CTL activities in response to other SART-1269-derived peptides or other SCC-related antigens as well as SART-1690–698-specific CTL activities. We recently identified new SCC-related antigens, SART-233 and SART-3.34, 35 Further identification of SCC-related antigens will greatly help to increase our understanding of the molecular basis of a host defense against SCCs as well as to establish a basis for developing specific immunotherapies for SCCs.

Acknowledgements

We express our gratitude to Dr. B. Quinn for his critical review of the article.

Ancillary