Expression of SART3 antigen and induction of CTLs by SART3-derived peptides in breast cancer patients

We recently reported the SART3 tumour-rejection antigen as possessing tumour epitopes capable of inducing HLA-class I-restricted cytotoxic T lymphocytes (CTLs). This study investigated expression of the SART3 antigen in breast cancer to explore an appropriate molecule for use in specific immunotherapy of breast cancer patients. The SART3 antigen was detected in all of the breast cancer cell lines tested, 30 of 40 (75%) breast cancer tissue samples, and 0 of 3 non-tumourous breast tissue samples. SART3 derived peptides at positions 109–118 and 315–323 induced HLA-A24 restricted CTLs that reacted to breast cancer cells from the peripheral blood mononuclear cells (PBMCs) of breast cancer patients. Therefore, the SART3 antigen and its peptides could be an appropriate molecule for use in specific immunotherapy of the majority of HLA-A24-positive breast cancer patients. © 2001 Cancer Research Campaignhttp://www.bjcancer.com

The incidence of breast cancer has been increasing since the 1960s at the global level. The 5-year survival rate is generally low in stage III breast cancer patients, and extremely low in the stage IV patients, regardless of different treatment modalities including hormone therapy, chemotherapy and radiotherapy. Therefore, development of new treatment modalities is needed, and one of them will be specific immunotherapy (Nestle et al, 1998;Rosenberg et al, 1998;Marchant et al, 1999). However, little information is available regarding tumour-rejection antigens on breast cancers. Many tumour-rejection antigens have been identified from melanomas (van der Bruggen et al, 1991Traversari et al, 1992;Kawakami et al, 1994Kawakami et al, , 1995Robbins et al, 1996Robbins et al, , 1997, but these antigens were rarely expressed in breast cancers (Vincenzo et al, 1995). SART1 259 antigen that was identified from an oesophageal cancer was also rarely expressed in breast cancers (Kawamoto et al, 1999). MUC1 antigen and HER2/neu oncogene products might encode a tumour epitope on HLA-A2 + breast cancers (David et al, 1995;Reddish et al, 1998), but the peptide specificity and HLA-restriction have not been clearly demonstrated.
We recently reported that the SART3 tumour-rejection antigen possesses tumour epitopes capable of inducing HLA-A24restricted CTLs in head and neck cancer patients (Yang et al, 1999). The 140 kD of the SART3 antigen was expressed in the nucleus of all of the malignant cell lines and in the majority of cancer tissues tested, but not in the nucleus of any of the normal tissues except that of the testis. It was also expressed in the cytosol of the majority of proliferating cells, including normal cells and malignant cells.
In this study, we investigated expression of the SART3 antigen in breast cancer tissues to explore for an appropriate molecule useful for specific immunotherapy of breast cancer, and found evidence suggesting that the SART3 could be an appropriate target molecule.
Breast cancer tissue samples (n = 40) and non-tumourous breast tissue samples (n = 3) were obtained at the time of surgery in the Kurume University Hospital. All of the cancers were histologically determined to be adenocarcinoma. In samples, stage I and stage II group numbered 23 cases and stage III and stage IV numbered 17 cases. A section of each sample was minced with scissors and kept at -80˚C until use. The KE -4 oesophageal squamous carcinoma cell line (HLA-A2402/A2601) from which the SART3 was cloned (Yang et al, 1999) was used as a positive control. VA13 fibroblast cells (no expression of HLA-class I alleles), and CIR-A402 cells were used as target cells as negative controls, as reported previously (Yang et al, 1999;Murayama et al. 2000).

Western blot analysis
Expression of the SART3 antigen in samples was investigated by Western blot analysis using the polyclonal anti-SART3 antibody as reported previously (Yang et al, 1999).

Expression of oestrogen and progesterone receptors
Expression of the oestrogen or progesterone receptors in frozen sections was measured using the standard enzyme immunoassay commercially available from SRL (Tokyo). For the assays Abbot ER-EIA mAb or Abbott PgR-EIA mAb (Abbott Park) was used according to the manufacturer's instructions. The international UICC-TNM system was used for staging the 31 breast cancer patients.

Peptides and CTL induction
The SART3 109-118 (VYDYNCHVDL) and SART3 315-323 (AYID-FEMKI) peptides were used for CTL induction as described previously (Yang et al, 1999). PBMCs from breast cancer patients (n = 4) were incubated with 10 µM of a peptide in one well of a 24-well plate containing 2 ml of culture medium with 100 units ml -1 of IL-2 (Shiongi Pharm Co, Osaka). At days 7, 14 and 21 of culture, the cells were harvested, washed and re-incubated with the irradiated (45 Gy) autologous PBMCs acting as antigenpresenting cells, which had been pre-incubated with the same peptide at the same dose for 3 h. These peptide-stimulated PBMCs (2-3 × 10 3 cells well -1 ) were harvested at day 28 of culture, and further cultured in a 96-well U-bottom microculture plate in the presence of feeder cells (irradiated HLA-A24 + PBMCs) that had been pre-pulsed with a corresponding peptide in order to obtain large numbers of effector cells. Seven to 10 days later, the expanded cells were transferred to wells of a 24-well plate and were incubated in the absence of either a peptide or feeder cells for an additional 21-28 days.
The surface phenotypes and CTL activity of these cells was tested by an IFN-γ production assay (limit of sensitivity = 10 pg ml -1 ), and 51 Cr-labelled assays at different E:T ratios at the triplicate assays. For preparation of CTL sublines, the peptide-stimulated PBMCs were incubated at 1 or 10 or 100 cells per well in the presence of irradiated HLA-A24 PBMCs (2-3 × 10 5 cells well -1 ) that were pre-loaded with a relevant peptide, followed by testing of their surface phenotype and CTL activity to produce IFN-γ in response to CIR-A2402 pre-loaded with a peptide.

Expression of SART3 antigen
The 140 kD of the SART3 antigen was expressed in both the cytosol and nuclear fractions of all eight breast cancer cell lines tested. This band was also detectable in the cytosol fraction of 28 of 40 (70%) breast cancer tissue samples, and in the nuclear fraction of 30 of 40 (75%) breast cancer tissue samples. The bands other than the 140 kD, which were observed in cancer samples, were evaluated as non-specific since the polyclonal antibody was used for the detection.
The expression of SART3 antigen in the group of stage I and II was 18 of 23 samples (78.3%) in the cytosol, and 19 of 23 samples (82.6%) in the nucleus, while that in the group of the stage III and IV was 10 of 17 samples (58.8%) in the cytosol, and 11 of 17 samples (64.7%) in the nucleus. However, it was not detectable in either the cytosol or nuclear fraction of any of the normal breast tissue samples tested. Examples of results are shown in Figure 1, and a summary of these results is presented in Table 1.

Expression of oestrogen or progesterone receptors
The expression of oestrogen receptor (ER) or progesterone receptor (PR) in these tumours was investigated in view of potential combined immunotherapy and hormone therapy. Twenty-three and 17 of the 31 breast cancer tumours expressed ER and PR, respectively (Table 1). Among the 31 tumours, 16 tumours were SART3 + and ER + , two tumours were SART3 + and ER -, seven tumours were SART3and ER + . The remaining six tumours were SART3and ER -. Similarly, 11 tumours were SART3 + and PR + , seven tumours were SART3 + and PR -, six tumours were SART3and PR + , and seven tumours were SART3and PR -.  Figure 1 Expression of the SART3 antigen in the breast cancer cell lines, breast cancer tissue samples, and non-tumourous breast tissue samples was investigated by Western blot analysis with polyclonal anti-SART3 antibody (Yang et al, 1999).
One of six or three of 20 sublines from the PBMCs of patient 1, when stimulated with the SART3 109-118 peptide or SART3 315-323 peptide, produced a significant amount of IFN-γ by recognition of C1R-A2402 cells pulsed with a corresponding peptide, respectively ( Figure 3A and B). Similarly, one of three or four of 20 sublines from PBMCs of patient 2, when stimulated with the SART3 109-118 peptide or SART3 315-323 peptide, produced a significant

DISCUSSION
This study showed that the SART3 antigen encoding tumour epitopes capable of inducing CTLs was expressed in both the cytosol and nuclear fractions of all breast cell lines and the majority of breast tumours tested. In contrast, this antigen was undetectable in non-tumourous breast tissue samples. We recently obtained evidence that this is a RNA-binding protein involved in splicing regulation of RNA (Harada, unpublished results). These results suggest that the SART3 antigen plays an important role in cellular proliferation. This issue is now under investigation in our laboratory. The present study also showed that the SART3 109-118 and SART 315-323 peptide induced HLA-A24-restricted CTLs recognizing the SART3 + tumour cells including breast cancer cells in the PBMCs of all HLA-A24 + breast cancer patients tested. These CTLs failed to lyse either HLA-A24tumour cells (Figure 2) or HLA-A24 + normal T cells (PHA-blastic cells) (data not shown). These results indicate the presence of CTL precursors reacting to the SART3 epitope on breast cancer cells in the circulation of the majority of breast cancer patients.
In contrast to the SART3 antigen, the other tumour antigens (MAGE-1, MAGE-4, SARTI 259 , MUC1) were expressed in only a portion of breast cancers (Vincenzo et al, 1995;Reddish et al, 1998;Kawamoto et al, 1999). Further, no peptides were proven to have the ability to induce HLA-class I restricted CTLs reacting to breast cancer cells. Administration of HER2/neu peptide capable of binding to HLA-A2 molecules failed to induce the CTL activity against cancer cells (Zaks and Rosenberg, 1998). Therefore, the SART3 antigen and its peptides could be one of the most appropriate molecules for use in specific immunotherapy of breast cancer patients.
There are few treatment modalities available for patients with breast cancer which is lacking ER or resistant to the existing hormone therapy (Slamon et al, 1989;Teixeira et al, 1995;Yamauchi et al, 1997). Among 31 tumours tested, 16 (52%) expressed both SART3 and ER, two (7%) were SART3 + and ER -, seven (23%) were SART3and ER + , and six (19%) were SART3and ER -. These results suggest that a substantial proportion of breast cancer patients would be appropriate candidates for specific immunotherapy with the SART3 peptides, either administered independently or in combination with hormone therapy. Similar results were obtained with PR. Further, SART3 expression in the group of stage I and II was higher than that in the group of stage III and IV. A similar trend was observed in the expression of ER and PR. These results suggest that both immunotherapy and hormone therapy are more effective for breast cancer patients with early stage rather than the patients with advanced stage, although details involved in this phenomenon is not clear at the present time.
The HLA-A24 allele is found in 60% of Japanese, 20% of Caucasians and 12% of Africans (Imanishi et al, 1992). We have recently identified tumour-epitopes of the SART3 antigen on HLA-A0207 molecules, which are able to induce HLA-A2restricted and tumour-specific CTLs in the PBMCs of epithelial cancer patients (Ito et al, unpublished results). The HLA-A2 allele is found in 40% of Japanese, 50% of Caucasians and 12% of Africans (Imanishi et al, 1992).

C D A B
INF-C production (pg ml -1 ) INF-g production (pg ml -1 ) INF-g production (pg ml -1 ) INF-g production (pg ml -1 ) * * * * Figure 3 Peptide specificity of CTL sublines was established by incubation of the PBMCs of patients 1 and 2, which had been stimulated with the SART3 109-118 or SART3 315-323 peptide. HLA-A2402-binding HIV peptide (RYLRDQQLLGI) was used as a negative control. One of six or three of 20 sublines from the PBMCs of patient 1, when stimulated with the SART3 109-118 peptide or SART3 315-323 peptide, produced a significant amount of IFN-γ by recognition of C1R-A2402 cells pulsed with a corresponding peptide, respectively (A, B). Similarly, one of three or four of 20 sublines from the PBMCs of patient 2, when stimulated with the SART3 109-118 peptide or SART3 315-323 peptide, produced a significant amount of IFN-γ by recognition of C1R-A2402 cells pulsed with a corresponding peptide, respectively (C, D). The other sublines failed to show peptide specificity. Values represent the mean of duplicate assays. The two-tailed Student's t-test was employed for statistical analysis (*P ≤ 0.05) Subsequently, the SART3 antigen and its peptides could be an appropriate molecule for use in specific immunotherapy of HLA-A24 + or -A2 + breast cancer patients. These findings could provide important information for the development of a specific immunotherapy for a relatively large number of breast cancer patients all over the world.