Identification of secernin 1 as a novel immunotherapy target for gastric cancer using the expression profiles of cDNA microarray

Cell lines

Human gastric cancer cell lines (MKN1, MKN28, and MKN45) expressing HLA‐A*2402, Kato III, another gastric cancer line expressing HLA‐A*0201/0207, a mouse fibroblast cell line (NIH3T3), a monkey kidney cell line (COS‐7), and a human hybrid between B and T lymphoblastic cell lines (T2) expressing HLA‐A*0201 were purchased from the American Type Culture Collection (Rockville, MD). A human hepatocellular carcinoma cell line, SNU475 (HLA‐A*0201/1101) was obtained from the Korean Cell Line Bank (Seoul National University, Seoul, Korea). A3LCL, an Epstein–Barr virus‐transformed B‐lymphoblastoid cell line (HLA‐A*0301/), was generously provided by Takara Shuzo (Otsu, Japan). All of these cells were cultured as monolayers in appropriate media: Dulbecco's modified Eagle's medium (COS7, NIH3T3), and RPMI‐1640 medium (T2, A3LCL, MKN1, MKN28, MKN45, Kato III and SNU475), supplemented with 10% fetal bovine serum (Cansera International, Ontario, Canada) and 1% antibiotic/antimycotic solution (Sigma‐Aldrich, St Louis, MO, USA). Cells were maintained at 37°C in an atmosphere of humidified air with 5% CO₂. Cancerous tissues and corresponding non‐cancerous mucosae were excised from 20 patients during surgery, after informed consent had been obtained.

RNA preparation and RT‐PCR analysis

Total RNA was extracted using either the RNeasy kit (Qiagen, Dusseldorf, Germany) or the Trizol reagent (Life Technologies, Rockville, MD, USA) according to the manufacturers’ instructions. Ten µg of total RNA was reverse‐transcribed for single‐stranded cDNAs using the poly dT_12–18 primer (Amersham Pharmacia Biotech, Buckinghamshire, UK) with Superscript II reverse transcriptase (Life Technologies). Each single‐stranded cDNA was then diluted for subsequent PCR amplification for target genes along with the GAPDH gene as a quantitative control. All of the reactions involved initial denaturation at 94°C for 4 min followed by 20 cycles at 94°C for 30 s, 58°C for 30 s, and 72°C for 30 s (for GAPDH); or 30 cycles at 94°C for 30 s, 56°C for 30 s, and 72°C for 30 s (for SCRN1). Primer sequences were as follows: for GAPDH forward 5′‐ACAACAGCCTCAAGATCATCAG‐3′; for GAPDH reverse 5′‐GGTCCACCACTGACACGTTG‐3′; for SCRN1 forward 5′‐TGCATAGACTCTGAGTTTTTCC‐3′; and for SCRN1 reverse 5′‐CTTTCTCATTTTACTCAAACAGG‐3′.

Northern blot analysis on multiple tissues

Human multiple‐tissue blots (Clontech, Palo Alto, CA) were hybridized with ³²P‐labeled SCRN1 cDNA as a probe. Pre‐hybridization, hybridization and washing were performed according to the manufacturer's recommendations. The blots were then autoradiographed with intensifying screens at −80°C for 24 h.

Immunocytochemical staining

The entire coding region of SCRN1 was amplified by RT‐PCR using a set of SCRN1‐specific primers, 5′‐GCGGATCCAGGATGGCTGCAGCTCCTCCAAG‐3′ and 5′‐TAGAATTCTTAAAGAACTTAATCTCCGTGTCAACAC‐3′. The PCR products were then cloned into the appropriate cloning site of pcDNA3.1myc/His (Invitrogen, Carlsbad, CA, USA) vector. COS7 cells transfected with pcDNA3.1myc/His‐SCRN1 were fixed with PBS containing 4% paraformaldehyde for 15 min, and were then rendered permeable with PBS containing 0.1% Triton X‐100 for 2.5 min at room temperature. Subsequently, the cells were covered with 2% bovine serum albumin in PBS for 24 h at 4°C to block any non‐specific hybridization. Mouse antimyc monoclonal antibody (Sigma) at 1 : 1000 dilution was used as the first antibody, and the reaction was visualized after incubation with rhodamine‐conjugated antimouse antibody (Leinco, St Louis, MO, USA; ICN, Costa Mesa, CA, USA). Nuclei were counter‐stained with 4′,6′‐diamidine‐2′‐phenylindole dihydrochloride. Fluorescent images were obtained under an Eclipse E800 microscope (Nikon, Tokyo, Japan).

Colony formation assay

The entire coding region of SCRN1 was amplified by RT‐PCR using a set of primers, 5′‐GCGGATCCAGGATGGCTGCAGCTCCTCCAAG‐3′ and 5′‐TAGAATTCTTAAAGAACTTAATCTCCGTGTCAACAC‐3′, and the PCR product was subsequently cloned into an appropriate cloning site of pcDNA3.1myc/His vector (Invitrogen). NIH3T3 cells were transfected transiently with pcDNA3.1myc/His‐SCRN1 or pcDNA3.1myc/His (mock) using FuGENE6 transfection reagent (Roche, Grenzacherstrasse, Switzerland), and incubated with 0.9 mg/mL of geneticin for 14 days. The surviving cells were fixed with 100% methanol and stained using Giemsa solution. All experiments were carried out in duplicate.

Selection and synthesis of candidate peptides derived from SCRN1

From the 9‐mer and 10‐mer peptides derived from SCRN1, candidates for the antigenic epitope were selected based on the predicted binding affinities to HLA‐A*0201 molecules, the most frequently observed HLA‐allele, using the HLA Peptide Binding Predictions program,^{( 26} , ^{27 )} available at the BIMAS (National Institutes of Health, Bethesda, MD) website http://bimas.dcrt.nih.gov/cgi‐bin/molbio/ken_parker_comboform. These peptides were synthesized by Mimotopes (San Diego, CA) using the standard solid‐phase synthesis method and were purified by reverse phase high‐performance liquid chromatography. The purity (>90%) and the identity of the peptides were determined by analytical reverse phase high‐performance liquid chromatography and mass spectrometry analysis, respectively. These peptides were dissolved in dimethylsulfoxide at 20 mg/mL and stored at −80°C. CMVpp65_495−503 (NLVPMVATV),^{( 28 )} an HLA‐A*0201‐restricted epitope peptide derived from cytomegalovirus protein pp65, was used for cytotoxicity assay as a control.

CTL induction by candidate peptides in vitro

DCs were prepared from PBMC donated by HLA‐A*0201 healthy volunteers for the use of antigen presenting cells as described previously.^{( 14} , ^{29 )} Briefly, the adherent monocyte‐enriched population isolated from PBMC in a plastic tissue culture flask was cultured in the presence of 1000 IU/mL of granulocyte macrophage colony‐stimulating factor (generous gift from Kirin Brewery, Tokyo, Japan) and 1000 IU/mL of IL‐4 (Techne, Minneapolis, MN) in AIM‐V (Life Technologies) containing 2% heat‐inactivated AS. After 5 days in the culture, OK‐432 (generously provided by Chugai Pharmaceutical, Tokyo, Japan) was added to the culture at the concentration of 10 µg/mL to induce the maturation of DCs. After 7 days in the culture, the mature DCs were harvested and pulsed with 20 µM of the candidate peptides for 4 h at 20°C, and these peptide‐pulsed DCs were irradiated (55 Gy) and incubated with autologous CD8⁺ T lymphocytes, obtained by positive selection with Dynabeads M‐450 CD8 and Detachabead (both from Dynal, Lake Success, NY) at a 1 : 20 ratio. These cultures were set up in 48‐well plates (Corning, Corning, NY); each well contained 1.5 × 10⁴ peptide‐pulsed DCs, 3 × 10⁵ CD8⁺ T cells and 10 ng/mL IL‐7 (Techne) in 0.5 mL of AIM‐V/2%AS. Three days later, IL‐2 (Chiron, Emeryville, CA) was added to these cultures at a final concentration of 20 IU/mL. On days 7 and 14, the T cells were restimulated with the autologous peptide‐pulsed DCs as described above. Cytotoxicity was tested on day 21 against peptide‐pulsed T2 cells at day 21 as described elsewhere.^{( 29 )}

Cytotoxicity assay

Cytotoxicity of the stimulated T cells was measured with ⁵¹Cr release assay. The T2 cells (HLA‐A*0201‐positive) and tumor cells were used as target cells. The T2 cells were pulsed with 20 µM peptide for 16 h before the labeling. The target cells were subsequently labeled with 100 µCi of Na₂ ⁵¹CrO₄ (Perkin‐Elmer Life Sciences, Boston, MA) for an additional 1 h at 37°C. The labeled cells (target) were then washed with PBS and mixed with the pretreated T cells (effector) at various E/T ratios in 96‐well plates. Half of the supernatant was collected from each well after 4 h of incubation, and its radioactivity was measured in a gamma scintillation counter. The percentage of specific cytotoxicity was determined by calculating the percentage of specific ⁵¹Cr‐release by the following formula: [(⁵¹Cr‐releasing activity of each experiment‐background activity)/(maximum ⁵¹Cr‐releasing activity‐background activity)]× 100 (%).

Background release was determined by incubating the target cells alone, in the absence of effectors, and the maximum release was obtained by incubating the targets with 1 mol/L HCl.

Establishment of CTL clones

The lymphoid cells induced on 48‐well plates and tested for positive peptide‐specific lysis as described above were expanded to establish CTL lines, and subsequently subjected for limiting dilution to establish CTL clones. To establish CTL lines, a total of 5 × 10⁴ CTLs were re‐suspended and expanded in 25 mL of AIM‐V/2%AS, along with 2.5 × 10⁷ irradiated (3300 rads) allogenic PBMCs and 5 × 10⁶ irradiated (8000 rads) A3LCL cells, in the presence of 30 ng/mL of anti‐CD3 mAb (BD Bioscience‐PharMingen, San Diego, CA) and 120 IU/mL of IL‐2 as reported by Walter et al.^{( 30 )} One day after initiating the cultures, 120 IU/mL IL‐2 was added to the cultures, and the cultures were fed with fresh AIM‐V/5%AS containing 30 IU/mL IL‐2 on days 5, 8, 11 and 14. The cytotoxicity of the expanded CTLs was tested on day 14, and we carried out limiting dilution of the CTL lines showing significant cytotoxicities against their corresponding peptides were carried out to establish CTL clones. Limiting dilution^{( 14 )} was performed with 7 × 10⁴ cells of allogenic PBMC, 1 × 10⁴ cells of A3LCL cells, 30 ng/mL of anti‐CD3 mAb and 125 IU/mL of IL‐2 per well in 96 round‐bottomed microtiter plates and each microculture was tested for cytotoxicity. Cells from CTL lines were added 0.3, 0.5, 1, 3, and 5 cells per 96 well (0.3 = 1 cell per 3 wells). Three plates were prepared for each cell number. We picked up the cells that grew well. The CTLs in a positive well were expanded once again as ‘CTL clones’ for further examination.

Cold target inhibition assay

Antigen specificity was confirmed by the cold target inhibition assay. SNU475 cells expressing both HLA‐A*0201 and SCRN1 were labeled with Na₂ ⁵¹CrO₄ (hot target) and subsequently mixed at various cold target : hot target (C/H) ratios with unlabeled T2 cells (cold target) pulsed with the peptide (20 µM for 16 h at 37°C) to compete for the recognition of ⁵¹Cr‐labeled SNU475 cells. After preparation of target cells, CTLs (effector) were added and incubated for 4 h. The number of target cells was fixed to 1 × 10⁴/well. The cytotoxicity of each C/H ratio was measured using ⁵¹Cr release assay. The percentage of specific inhibition (% inhibition) was calculated using the following formula: [(%Lysis (C/H ratio 0) – %Lysis (each experiment))/%Lysis (C/H ratio 0)] × 100 (%).

Antibody blocking assay

SNU475 cells pretreated with anti‐HLA‐A/B/C mAb or anti‐HLA‐DR/DB mAb for 1 h were co‐cultured with CTL clones that were pretreated with anti‐CD4 mAb or anti‐CD8 mAb (all mAbs from Dako, Carpinteria, CA) for 1 h, at an E/T of 20. Isotype antibodies, antimouse IgG2a mAb for anti‐HLA‐A/B/C mAb and antimouse IgG1 mAb for anti‐HLA‐DR/DB mAb, anti‐CD4 mAb, and anti‐CD8 mAb were used as the controls. The percentage of specific inhibition (% inhibition) was calculated with the following formula: [(%lysis (isotype) – %lysis (experimental))/%lysis (isotype)] × 100 (%).

Identification of SCRN1, a human gene commonly upregulated in gastric cancer

Using a genome‐wide cDNA microarray containing 23 040 genes, we previously examined expression profiles of 20 gastric cancers and their corresponding normal gastric mucosae.^{( 31 )} Among the commonly upregulated genes in the cancers, we identified a gene with an in‐house accession number of C8121. Its signal intensity ratios of cancer to non‐cancerous gastric tissues were greater than two. Microarray examination was carried out in nine of 11 gastric cancers that passed our cut‐off filter ( Fig. 1a). A homology search with the nucleotide sequence of C8121 revealed that it corresponded with the SCRN1 gene. Semiquantitative RT‐PCR experiments confirmed its increased expression in five of the seven gastric cancers compared with their corresponding non‐cancerous tissues (Fig. 1b). Additional examination of other types of malignancies showed that SCRN1 was upregulated in 22 of 24 colon cancers examined and one of six hepatocellular carcinomas examined (data not shown).

Expression of SCRN1 in normal vital organs

Multi‐tissue Northern blot analyses using SCRN1 cDNA as a probe showed a 4.9‐kb transcript was expressed abundantly in the testis and ovary and marginally in brain, prostate, thymus and small intestine (Fig. 2a). Additional semiquantitative RT‐PCR analysis of SCRN1 in gastric cancer cell lines revealed that it was expressed abundantly in MKN‐1, ‐7, ‐28, ‐45, and ‐74 cells but not in Kato III (Fig. 2b).

Effect of SCRN1 overexpression on cell growth

To analyze the effect of SCRN1 overexpression on cell growth, we carried out a colony‐formation assay on NIH3T3 cells transfected with pcDNA3.1myc/His‐SCRN1 that express myc‐tagged SCRN1 protein. Expression of myc‐tagged SCRN1 was confirmed by Western blot analysis with antimyc antibody (data not shown). Compared to the NIH3T3 cells transfected with a control plasmid (pcDNA3.1myc/His‐LacZ), those with pcDNA3.1myc/His‐SCRN1 showed markedly more colonies in NIH3T3 cells (Fig. 2c). This result was repeated and confirmed in three independent experiments. We did not observe morphological changes in cells using normal microscopy.

Successful induction of CTLs with a candidate peptide derived from SCRN1

Table 1 shows the predicted peptide sequences binding to HLA‐A*0201 in the order of predicted binding affinities, indicated as the binding score. The 40 peptides were tested as candidates whether they could induce peptide‐reactive CTLs or not. SCRN1‐196 (KMDAEHPEL) induced the proliferation of lymphoid cells. The lymphoid cells showed significantly higher cytotoxicity against T2 cells pulsed with SCRN1‐196 peptide compared to those pulsed with the control peptide CMVpp65_495−503 (data not shown). We carried out further limiting dilutions in order to establish CTL clones from the CTL lines.

Of the CTL clones established from the CTL lines against SCRN1‐196, CTL Clone‐23 showed higher cytotoxicity against T2 cells pulsed with SCRN1‐196 compared to those pulsed with the control peptide (Fig. 3a). Additionally, CTL Clone‐23 was examined for its ability to recognize and lyze the tumor cells endogenously expressing SCRN1. CTL Clone‐23 showed specific cytotoxicity against SNU475, human hepatocellular carcinoma cells that express both HLA‐A*0201 and SCRN1, but it did not show cytotoxicity against HLA‐A*0201‐negative MKN‐45, which expressed SCRN1, nor against SCRN1‐negative Kato III, which expressed HLA‐A*0201 (Fig. 3b). These data suggest that CTL Clone‐23 can recognize SCRN1‐196 peptide processed and presented by SCRN1‐positive tumor cells with HLA‐A*0201 restriction.

Specificity of established CTL clones against SCRN1 peptide

Cold target inhibition assay was performed to confirm the specificity of CTL Clone‐23 against SCRN1‐196 peptide. The SNU475 cells labeled with ⁵¹Cr were used as hot targets, and T2 cells pulsed with peptides without ⁵¹Cr labeling were used as cold targets. In this assay, the specific lysis of ⁵¹Cr‐labeled SNU475 cells was dramatically reduced when they were mixed with unlabeled T2 cells pulsed with the SCRN1‐196 peptide, but not when they were mixed with unlabeled T2 cells that were pulsed with the control peptide (Fig. 3c). To confirm the HLA‐restriction, we examined the cytotoxic activity after the incubation with blocking mAbs. Significantly decreased cytotoxicities were observed when SNU475 cells were pretreated with anti‐HLA‐A/B/C mAb, or when the CTL Clone‐23 cells were preincubated with anti‐CD8 mAb (Fig. 3d). Taken together, the CD8‐positive cells in CTL Clone‐23 recognize and lyze SNU475 cells in the MHC‐class I restricted manner. These results indicate that SCRN1‐196 peptide is a HLA‐A*0201 restricted epitope derived from SCRN1.

CTLs induced with peptides modified at anchoring positions

According to the HLA‐A*0201 antigen motif previously reported by Smith et al.^{( 32 )} the binding affinities of the peptides to the HLA are determined by amino acid residues at position 2 and 9. The newly identified epitope peptide SCRN1‐196 has a relatively low binding score. We synthesized the peptides SCRN1–9 V (KMDAEHPEV) and SCRN1–2 L (KLDAEHPEL), which have modified amino acids only at anchoring positions and are predicted to have much higher binding affinity to HLA‐A*0201 (Table 2). The leucine at position 9 in SCRN1‐196 was changed to valine in SCRN1–9 V, and the methionine at position 2 in SCRN1‐196 was changed to leucine in SCRN1–2 L. These two peptides were predicted to have the capability of binding to the T cell receptor, recognizing native SCRN1‐196 peptide. We again prepared CTLs pulsed with SCRN1–9 V or SCRN1–2 L peptides, and successfully established CTL clones with SCRN1–9 V (Fig. 4a). CTL Clone‐16 killed not only T2 cells pulsed with SCRN1‐196 peptide but also SNU475 cells expressing SCRN1, implicating that SCRN1–9 V can also induce CTLs specific to SCRN1 (Fig. 4b). However, CTL Clone‐16 had only modest cytotoxicity against tumor cells expressing SCRN1. We further compared the efficiency of CTL induction of SCRN1–9 V with that of SCRN1‐196. When we used SCRN1–9 V, CTL lines showing significant cytotoxicity against the native SCRN1‐196 peptide were successfully established in three out of four individuals, whereas the CTL lines were established in only one out of four individuals using SCRN1‐196 peptide. Moreover, positive wells were more frequently observed in the initial cultures when SCRN1–9 V was used (data not shown).

A homology search in protein databases revealed that neither SCRN1‐196 nor SCRN1–9 V has homologous peptide sequences derived from any other gene products (data not shown), indicating that these peptides are likely to induce SCRN1‐specific cytotoxicity.