Vaccination with Ep-CAM Protein or Anti-Idiotypic Antibody Induces ...

2 downloads 22 Views 164KB Size Report
Hadassah Medical School, Jerusalem, Israel ... 2) is expressed at a high level by a ... Cancer Society, Cancer Society in Stockholm, King Gustav V Jubilee. Fund ...
Vol. 10, 5391–5402, August 15, 2004

Clinical Cancer Research 5391

Vaccination with Ep-CAM Protein or Anti-Idiotypic Antibody Induces Th1-Biased Response against MHC Class I- and II-Restricted Ep-CAM Epitopes in Colorectal Carcinoma Patients Szilvia Mosolits,1 Katja Markovic,1 Jan-Erik Fro¨din,1 Lena Virving,1 Carl G. M. Magnusson,2 Michael Steinitz,3 Jan Fagerberg,1 and Håkan Mellstedt1 1

Immune and Gene Therapy Laboratory, Department of Oncology (Radiumhemmet), Karolinska Institute, Stockholm, Sweden; 2 ¨ ngelholm Hospital, A ¨ ngelholm, Department of Clinical Chemistry, A Sweden; and 3Experimental Pathology, The Hebrew UniversityHadassah Medical School, Jerusalem, Israel

ABSTRACT Purpose: The tumor-associated antigen Ep-CAM (epithelial cell adhesion molecule) is overexpressed in colorectal carcinoma (CRC). The aim of the present study was to evaluate and compare the safety and immunogenicity of a recombinant Ep-CAM protein and a human anti-idiotypic antibody (anti-Id) mimicking Ep-CAM. Experimental Design: Patients with resected American Joint Committee on Cancer stages II–IV CRC without remaining macroscopic disease received intradermal/subcutaneous injections of Ep-CAM (400 ␮g/dose; n ⴝ 7) or anti-Id (500 ␮g/dose; n ⴝ 6) at weeks 0, 2, and 6 in combination with granulocyte macrophage colony-stimulating factor (75 ␮g/day, for 4 consecutive days). Results: Adverse reactions were mild (grade I–II). All patients immunized with the Ep-CAM protein produced Ep-CAM–specific IgG antibodies, predominantly IgG1 and IgG3 subclasses, whereas no humoral response was induced by the anti-Id vaccine. All patients, with one exception in each group, mounted an Ep-CAM–specific proliferative Tcell response. The immune response was more rapid, potent, and protracted after Ep-CAM in comparison with anti-Id vaccination. Interferon-␥-secreting cells (ELISPOT) were detected in both immunization groups against the Ep-CAM

Received 3/3/04; revised 4/30/04; accepted 5/10/04. Grant support: Grants from the Swedish Research Council, Swedish Cancer Society, Cancer Society in Stockholm, King Gustav V Jubilee Fund, Cancer and Allergy Foundation, Torsten and Ragnar So¨derberg Foundation, Gunnar Nilsson Cancer Foundation, and Karolinska Institute Foundations. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Requests for reprints: Håkan Mellstedt, Department of Oncology, Cancer Center Karolinska, Karolinska Hospital, S-171 76 Stockholm, Sweden. Phone: 46-8-517-74308; Fax: 46-8-318-327; E-mail: hakan. mellstedt@ kus.se.

protein as well as various Ep-CAM– derived MHC class Iand II-restricted peptides. Flow cytometry analysis showed that Ep-CAM–specific interferon-␥- and perforin-producing cells predominantly resided within CD8ⴙCD56ⴚ and CD8dimCD56ⴙ T cells. Conclusions: Ep-CAM protein in combination with granulocyte macrophage colony-stimulating factor induced a long-lasting, Th1-biased humoral and cellular immune response compared with anti-Id. Ep-CAM–specific T cells and natural killer-like T cells responding in a MHC class Iand II-restricted manner were also induced. Vaccination with Ep-CAM protein may warrant further investigation as a novel therapeutic approach to CRC.

INTRODUCTION Fifty percent of patients with colorectal carcinoma (CRC) develop recurrence after successful surgery, presumably due to micrometastases (1). Adjuvant chemotherapy may improve the absolute survival by 5– 6% in stage III CRC, whereas its beneficial effect in stage II disease is controversial (1). Vaccine therapy might be an alternative treatment strategy to improve the prognosis of CRC patients. The tumor-associated antigen (TAA) Ep-CAM (epithelial cell adhesion molecule; Ref. 2) is expressed at a high level by a large number of epithelial neoplasias, including CRC. Normal epithelial cells express Ep-CAM at a lower density restricted to the basolateral cell surface (2). CRC patients, but not healthy donors or patients with inflammatory bowel disease, mount a natural immune response against Ep-CAM (3, 4). Treatment with monoclonal antibody (mAb) 17-1A (anti-Ep-CAM) may improve the survival of patients with resected stage III CRC (5–7). Vaccination with viral vectors encoding Ep-CAM could induce Ep-CAM–specific immune responses in animal models and humans (8, 9). Immunization with Ep-CAM protein produced in a baculovirus expression system and conjugated to alum has also elicited humoral and cellular immunity in CRC patients (10). Anti-idiotypic antibody (anti-Id) mimicking a TAA may serve as a surrogate antigen (Ag) and induce TAA-specific immunity (11). Vaccination with anti-Ids mimicking Ep-CAM has elicited Ep-CAM–specific humoral and cellular immune responses in cancer patients (8, 12). Anti-Id or the Ep-CAM protein has, however, not been tested in combination with the potent adjuvant granulocyte macrophage colony-stimulating factor (GM-CSF), which has been shown to significantly augment both humoral and cellular immune responses against weakly immunogenic Ags (13). It is also controversial whether an anti-Id or the bona fide Ag is more potent as a vaccine (14, 15).

5392 Ep-CAM Induces MHC Class I-Restricted IFN-␥ Response

Optimal antitumor immunity entails the induction of Agspecific Th1-biased humoral and cellular effector functions (16). Various in vitro assays are used to monitor Ag-specific immune responses including proliferation assay, ELISPOT, and multiparameter flow cytometry (17). In humans, type 1 immunity is associated with IgG1 and IgG3, whereas type 2 immunity is associated with IgG4 antibody (Ab) responses (18). Thus, analysis of the Ag-specific IgG subclass response may serve as a surrogate marker for type 1 and 2 immunity. The present study constitutes a basis for future development of a vaccination approach targeting Ep-CAM. The objectives were to evaluate in CRC patients the safety and immunogenicity of an Ep-CAM protein and a human anti-Id vaccine in combination with GM-CSF and to identify MHC-restricted immunogenic epitopes of the Ep-CAM molecule in a genetically diverse population.

PATIENTS AND METHODS Patients. Thirteen patients were enrolled in this study. Eligibility criteria included histopathological diagnosis of EpCAM-positive adenocarcinoma of the colon or rectum, American Joint Committee on Cancer stage II–IV disease, and resection of the primary tumor (and metastasis in stage IV) without evidence of remaining macroscopic disease. Patients were required to have a Karnofsky performance status of ⱖ80% and adequate hepatic and renal function. Exclusion criteria were as follows: pregnancy or nursing; HIV seropositivity; alcohol or drug abuse; autoimmune disease; known hypersensitivity to insect cells, baculovirus, or GM-CSF; active infection; chronic systemic disorder; or chemo- or radiotherapy or other immunosuppressive agent within 30 days before study entry. Patients were included before routine adjuvant chemotherapy for stage III colon cancer was introduced in Sweden. The study was approved by the Ethical Committee of the Karolinska Institute. All patients were treated after informed consent. Before entry into the study, a complete case history was obtained, and physical exam and blood tests including hemoglobin, white blood cell count with differentials and platelet counts, serum creatinine and electrolytes, serum protein electrophoresis, tests for liver function, thyroid function, serum tumor markers [carcinoembryonic antigen (CEA), CA19-9, CA50], and standard urine analysis were performed. HLA typing was done at the Tissue Typing Laboratory, Huddinge University Hospital. A chest X-ray and CAT scan of the abdomen were performed as well. During follow-up, patients were regularly checked for performance status and vital signs, routine blood hematology and chemistry analyses, standard urine analysis, thyroid function tests, and serum tumor markers. Adverse events were assessed according to the National Cancer Institute Common Toxicity Criteria Version 2.0.4 Vaccination was discontinued if disease progression occurred, and patients received standard therapy (surgery, chemotherapy, or radiotherapy). Vaccine Preparations. The extracellular domain of the Ep-CAM protein was purified using an immunoaffinity column

coupled with the anti-Ep-CAM mAb GA733 (Centocor, Malvern, PA) from supernatants of cultured Spodoptera frugiperda insect cells infected with a recombinant Ep-CAM baculovirus construct as described elsewhere (19). The only modifications to the previously described method were that serum-free medium SF900II (GIBCO, Paisley, Scotland) was used during the whole culture period and detergent was omitted from the elution buffer in immunoaffinity chromatography. The human anti-Id (IgG1␬) mimicking Ep-CAM (ab2␤) was produced by an Epstein-Barr virus-immortalized human lymphoblastoid cell line derived from a patient treated with murine mAb 17-1A (20, 21). Anti-Id was purified on protein A and mAb Trap G columns (Pharmacia, Uppsala, Sweden) as described previously (20, 21). Purity of the Ag preparations was confirmed by SDSPAGE and electrophoresis. Immunoreactivity of the products was checked by enzyme-linked immunosorbent assay (ELISA) and Western blot (19 –21). Tests for sterility, endotoxin, and viruses were performed according to the protocol described previously (21). Vaccination Schedule. Patients were randomized into one of the two vaccination schedules receiving intradermal/ subcutaneous injections of Ep-CAM protein (400 ␮g/dose; n ⫽ 7) or anti-Id (500 ␮g/dose; n ⫽ 6) at weeks 0, 2, and 6. Patients received concomitant GM-CSF (75 ␮g/day; Leucomax; Schering-Plough, Kenilworth, NJ) for 4 consecutive days at each immunization. The first dose of GM-CSF was given the day before the vaccine. Ags and GM-CSF were injected to the same site in the deltoid region. Patients who exhibited a negative Ep-CAM–specific proliferative T-cell response for a prolonged time and gave consent received additional booster vaccinations (see “Results”). Patients were clinically followed and immune monitored at regular intervals (see “Results”). Monoclonal Antibodies, Control Antigens, Peptides, and Colorectal Carcinoma Cell Lines. The humanized mAb 3622W94 recognizing Ep-CAM was a kind gift from Dr. J. H. Ellis (GlaxoSmithKline, Stevenage, United Kingdom). Mouse antihuman HLA-ABC mAb (clone W6/32; DAKO, Glostrup, Denmark), mouse antihuman HLA-DR mAb (clone G46-6; BD PharMingen, Mountain View, CA), and purified mouse IgG2a as a control (BD PharMingen) were used in blocking experiments (10 ␮g/ml). As controls for the baculovirus recombinant Ep-CAM protein, a baculovirus control protein (BCP), baculovirus recombinant CEA (Protein Sciences Corp., Meriden, CT), and a baculovirus recombinant HIV gp160 (kindly provided by Dr. B. Wahren; Swedish Institute for Infectious Disease Control, Karolinska Institute, Stockholm, Sweden) were used. Production of BCP and CEA has been described elsewhere (22). As controls for the human anti-Id, a human mAb preparation (IgG1␬; Ref. 12) and human polyclonal IgG (Gammaglobulin; Pharmacia, Stockholm, Sweden) were used. The protein sequences of the complete human Ep-CAM (23) and the complementarity determining regions (CDRs) of anti-Id (24) were submitted to online databases5 (25, 26) pre-

5 4

http://ctep.cancer.gov/reporting/ctc.html.

http://SYFPEITHI.de/ and http://bimas.dcrt.nih.gov/cgi-bin/molbio/ ken_parker_comboform.

Clinical Cancer Research 5393

Table 1 Sequence

Ep-CAM– derived and control peptides

Amino acid position

Ep-CAM–derived peptides VLAFGLLLA‡ YQLDPKFITSI ILYENNVITI KAPEFSMQGL‡ GLKAGVIAV AQEECVCENY DLDPGQTLIY

6–14 174–184 184–193 255–264 263–271 23–32 241–250

DVDIADVAYY APEFSMQGL‡ DPKFITSIL‡ TRYQLDPKF RRAKPEGAL TATFAAAQEECVCEN

206–215 256–264 177–185 172–180 80–88 17–31

CENYKLAVNCFVNNN

29–43

CLVMKAEMNGSKLGR TSTCWCVNTAGVRRT

66–80 113–127

Control peptides ILKEPVHGV KEPIVGAETFYVDGA

HIV RT HIV RT II

HLA binding A2.1 A2.1 A2.1 A2.1 A2.1 A1 A1 A26 A26 B7 B7 B27 B27 DR1 DR4 DR1 DR4 DR4 DR1 DR4 A2.1 DR1 DR4

SYFPEITHI score*

Bimas score†

23 IC50 ⫽ 2 ␮mol/L§ 29 18 27 26 36 27 32 22 21 29 27 26 22 28 28 26 24 18

19.4 DT50 ⬎ 4 h§ 535 9.8 5.6 6.7 1250 n.a. n.a. 80 72 5000 6000 n.a. n.a. n.a. n.a. n.a. n.a. n.a.

30 32 26

39 n.a. n.a.

Abbreviation: n.a., not available in the database. * http://SYFPEITHI.de/ (25). † http://bimas.dcrt.nih.gov/cgi-bin/molbio/ken_parker_comboform (26). ‡ Peptides containing identical anchor residues and showing ⬎33% amino acid sequence homology with peptides derived from the CDRs of anti-Id. § Ras et al. (27).

dicting peptides binding to various MHC class I and II alleles. Top-scoring peptides derived from Ep-CAM and anti-Id CDRs were compared for amino acid sequence homology. Based on these analyses, 15 nonamer, decamer, or 15-mer Ep-CAM– derived peptides; 2 control peptides; and an additional 11-mer Ep-CAM– derived peptide described by Ras et al. (27) were selected for synthesis (Table 1). Peptides were synthesized by standard solid-phase chemistry on a multiple peptide synthesizer, analyzed by mass spectrometry (Thermo Hybaid GmbH, Ulm, Germany), and checked for purity (⬎90%) by analytical high-performance liquid chromatography. The CRC cell lines LS180, SW837, and SW948 were purchased from American Type Culture Collection and cultured in recommended American Type Culture Collection medium. BL-41 was kindly provided by Dr. E. Klein (Karolinska Institute, Stockholm, Sweden). Detection of IgG and IgG Subclass Antibodies against Ep-CAM (ELISA). Abs against Ep-CAM were analyzed as described previously (9), with recombinant gp160 protein used as control Ag. Wells were coated with 2.5 ␮g/ml Ep-CAM and gp160 protein. Serum samples were assayed at 1:50 and 1:100 dilutions. The anti-Ep-CAM mAb 3622W94 (0.1–100 ng/ml) was used as a standard. The anti-Ep-CAM IgG concentration [arbitrary ELISA units (EU)/ml] was calculated from the standard curve. Sera of healthy blood donors (n ⫽ 30) were used as controls. The median age of the controls (22 males and 8

females) was 57 years (range, 48 –70 years), and the IgG titers against Ep-CAM were 0.871 ⫹ 0.817 EU/ml (mean ⫹ 2 SD). Based on these results, the threshold level for a positive response was set to ⬎1.688 EU/ml. IgG titers against gp160 in healthy controls and patients before immunization were 1.056 ⫾ 0.066 and 0.411 ⫾ 0.162 (mean ⫾ SE), respectively. Ep-CAM–specific IgG1, IgG2, IgG3, and IgG4 Abs were determined as described previously (28). The coating concentration for Ep-CAM and gp160 was 2 ␮g/ml. All serum samples were initially assayed at 1:40 dilution, except in the case of IgG4, for which a 1:8 dilution was used. High-titered sera were further assayed at higher dilutions. The Ab concentrations were calculated from standard curves established from chimeric IgG1, IgG2, IgG3, and IgG4 anti-5-iodo-4-hydroxy-3-nitrophenacetyl acid hapten Abs using bovine serum albumin (BSA)5-iodo-4-hydroxy-3-nitro-phenacetyl acid conjugate (10 ␮g/ml) as coating Ag (28). Taking serum dilution into account and correcting for background, the sensitivity of the assay was 8 (IgG1), 4 (IgG2 and IgG3), and 0.4 EU/ml (IgG4). This procedure permits comparison of results in a semiquantitative way, where 1 EU approximately corresponds to 1 ng of Ab. Western Blot. Proteins (20 ␮g/ml) and molecular mass standards (SeeBlue; Novex, San Diego, CA) were separated by SDS-PAGE on 4 –12% Bis-Tris NuPAGE precast gradient gels (Novex) under reducing or nonreducing conditions according to the manufacturer’s instructions. Proteins were electroblotted for 1 h onto a nitrocellulose membrane (0.45 ␮m; Novex), which

5394 Ep-CAM Induces MHC Class I-Restricted IFN-␥ Response

was subsequently blocked (1 h) with PBS containing 4% milk powder (Semper, Stockholm, Sweden), 2% BSA (Sigma-Aldrich, St. Louis, MO), and 0.05% Tween 20. Membranes were incubated overnight at 4°C with either humanized anti-Ep-CAM mAb 3622W94 (IgG1; 1 ␮g/ml), a control human mAb preparation (IgG1␬; 1 ug/ml), or serum samples (1:100) diluted in blocking solution. After washing, antihuman IgG1 mAb (1:5000 dilution; Ref. 28) was added for 6 h at room temperature, followed by alkaline phosphatase-conjugated rabbit F(ab⬘)2 antimouse IgG Ab overnight at 4°C (28). After washing in 0.9% NaCl, the membranes were developed with alkaline phosphatase color buffer (Bio-Rad Lab, Stockholm, Sweden). All washing steps (3 ⫻ 5 min) were performed with blocking solution diluted (1:10) in PBS containing 0.05% Tween 20. Lymphoproliferative Assay. Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized venous blood by Ficoll gradient centrifugation, and proliferative T-cell response assessed in a standard [3H]thymidine incorporation assay as described previously (9). Protein Ags were added at concentrations 1, 10, and 100 ng/ml. Twenty-three Ep-CAM– derived 18-mer peptides (with 6-amino acid NH2- and COOHterminal overlap; Ref. 3) were also used as stimulators (1–10 ␮g/ml). Phytohemagglutinin (PHA; 10 ␮g/ml; Sigma), concanavalin A (10 ␮g/ml; Amersham Bioscience, Uppsala, Sweden), purified protein derivative of tuberculin (PPD; 2.5 ␮g/ml; Statens Seruminstitut, Copenhagen, Denmark), and tetanus toxoid (TT; 50 ng/ml; SBL Vaccine, Stockholm, Sweden) were used as positive controls. A stimulation index (SI) was calculated for each triplicate by dividing mean cpm in experimental wells by that of the background value (cells in medium alone). To control for anti-baculovirus response, the highest SI value of cells stimulated with either BCP or CEA was subtracted from that of cells stimulated with Ep-CAM in each test. SI values of IgG1␬ or gammaglobulin were subtracted from that of the anti-Id in a similar manner. In healthy control donors, the highest SI values induced by Ep-CAM (n ⫽ 34) and anti-Id (n ⫽ 29) were 1.7 ⫹ 1.2 and 1.6 ⫹ 1.2 (mean ⫹ 2 SD), respectively. The corresponding value (n ⫽ 21) for the 18-mer Ep-CAM– derived peptides was 2.1 ⫹ 1.4. Based on these results, the threshold level for a positive proliferative T-cell response was set to ⬎3.0 for protein Ags and ⬎3.5 for peptides. SI values against PHA, concanavalin A, PPD, and TT in patients were 295 ⫾ 74, 84 ⫾ 23, 85 ⫾ 23, and 9 ⫾ 6 (mean ⫾ SE), respectively. The corresponding figures in healthy controls were 94 ⫾ 24, 101 ⫾ 39, 192 ⫾ 50, and 40 ⫾ 16, respectively. The differences in results between patients and controls were statistically not significant, except for TT (P ⫽ 0.004). Enzyme-Linked Immunospot Assay (ELISPOT) for Detection of Interferon-␥-Producing Cells. Cryopreserved PBMCs were thawed and allowed to recover overnight at 37°C in humidified air with 5% CO2. ELISPOT assay was performed as described previously (9). Briefly, interferon (IFN)-␥ secretion was assessed on nitrocellulose membrane bottomed-plates (Millipore, Bedford, MA) using a mouse antihuman IFN-␥ mAb pair, clone 1-D1K (10 ␮g/ml) and 7-B6-1 (1 ␮g/ml; Mabtech AB, Stockholm, Sweden). PBMCs (105 cells/ well) were incubated for 20 h in the presence of Ep-CAM, BCP and CEA proteins (100 and 1000 ng/ml) as well as Ep-CAM–

derived and control peptides (10 ␮g/ml; Table 1). In some tests, irradiated (6000 rads) CRC cells (104 cells/well) were used as stimulators. Cells stimulated with PHA (10 ␮g/ml) and PPD (2.5 ␮g/ml) served as positive controls. Streptavidin-alkaline phosphatase (1:1000; Mabtech), followed by 5-bromo-4-chloro3-indolyl phosphate/nitroblue tetrazolium (Sigma) was used for development of the color reaction. Results were quantified using an automated computer-assisted video imaging analysis system (Axioplan 2; Carl Zeiss Vision). Results are expressed as the number of spot-forming units (SFU) per 106 cells in 6 experimental wells (mean ⫾ SE) after subtraction of the background value (cells in medium alone). In some tests, the numbers of replicates were reduced due to lack of cells. Results were considered positive if the number of SFU in experimental wells was significantly (P ⬍ 0.05) greater than the background value and was at least twice that of the background. SFU of healthy control donors against Ep-CAM (n ⫽ 24), BCP (n ⫽ 14), and CEA (n ⫽ 14) were 11 ⫾ 3, 6 ⫾ 3, and 11 ⫾ 5 per 106 PBMCs (mean ⫾ SE), respectively. The corresponding figures (n ⫽ 10) for anti-Id, IgG1␬, and gammaglobulin were 14 ⫾ 4, 16 ⫾ 5, and 19 ⫾ 9 per 106 PBMCs, respectively. No positive response against any of these proteins was detected in healthy controls. HLA-typed healthy controls (n ⫽ 10) were also tested against Ep-CAM– derived peptides (Table 1). A positive response was detected against p29 – 43 in one control and against p113–127 in another donor. The other peptides were negative. Intracellular Interferon-␥ and Perforin Detection by Four-Color Flow Cytometry. PBMCs (1 ⫻ 106 cells/ml) were incubated in the presence of Ep-CAM or BCP (1,000 ng/ml) for 16 h at 37°C in humidified air with 5% CO2. PBMCs with or without phorbol 12-myristate 13-acetate (50 ng/ml; Sigma) and ionomycin (250 ng/ml; Sigma) served as positive and negative controls, respectively. Brefeldin A (Sigma) was added (10 ␮g/ml), and incubation continued for an additional 4 h. Cells were washed and incubated in 4% paraformaldehyde (Sigma) for 10 min on ice followed by washing in PBS containing 1% BSA and 0.1% NaN3. Surface staining was carried out by incubation for 30 min on ice in dark with mAb (BD PharMingen) mixtures as follows: (a) anti-CD3-APC, antiCD8-PerCP-Cy5.5, and anti-CD4-R-PE; and (b) anti-CD3APC, anti-CD8-PerCP-Cy5.5, and anti-CD56-R-PE. Isotypematched Abs served as negative controls. After washing, cells were incubated in FACS Permeabilizing Solution (BD Immunocytometry Systems, San Jose, CA) for 10 min at room temperature. Cells were washed and stained with anti-IFN-␥-FITC or anti-perforin-FITC or isotype control-FITC mAbs (all from BD PharMingen) for 30 min on ice in the dark. Anti-vimentinFITC Ab (kindly provided by Dr. R. Lenkei, CALAB ResearchNova Medical, Stockholm, Sweden) was used to test for the efficiency of permeabilization. Cells were washed and resuspended in 0.5% paraformaldehyde in PBS, and data acquisition was performed using a flow cytometer (FACScalibur). A total of 200,000 events gated on the basis of forward and side scatter for lymphocytes were collected. Data were analyzed using CellQuest software (BD Immunocytometry Systems). Statistical Analysis. The nonparametric WilcoxonMann Whitney two-tailed rank-sum test was used for compar-

Clinical Cancer Research 5395

Table 2

Patient characteristics and vaccination schedules

Patients

Age (yrs)/Gender

AJCC stage

Tumor site

Prior therapy*

Immunization† (wks)

A1 A2 A3 A4 A5 A6 A7 B1 B2 B3 B4 B5 B6

66/F 76/M 63/F 73/M 72/M 78/M 77/M 73/M 51/M 59/F 77/F 65/M 71/M

II II II III II II II III IV‡ II III II II

Coli caeci Recti Coli descendens Coli sigmoidei Coli descendens Coli sigmoidei Coli sigmoidei Coli caeci Coli caeci Coli transversi Coli ascendens Coli caeci Recti

No RT No No No No No No No No No No RT

Ep-CAM (0, 2, 6) Ep-CAM (0, 2, 6, 152, 154, 156) Ep-CAM (0, 2, 6) Ep-CAM (0, 2, 6) Ep-CAM (0, 2, 6) Ep-CAM (0, 2, 6, 34, 37) Ep-CAM (0, 2, 6) Anti-Id (0, 2, 6, 81, 83) Anti-Id (0, 2, 6, 56, 58) Anti-Id (0, 2, 6, 54, 56) Anti-Id (0, 2, 6, 54, 56, 80, 82) Anti-Id (0, 2, 6) Anti-Id (0, 2, 6)

Abbreviations: AJCC, American Joint Committee on Cancer; RT, preoperative radiotherapy. * Therapy before immunization except for surgical resection of the tumor. † All patients received three intradermal/subcutaneous injections of Ep-CAM (400 ␮g/dose) or anti-Id (500 ␮g/dose) at weeks 0, 2, and 6 together with GM-CSF (75 ␮g/day, for 4 consecutive days at each immunization). Six patients received booster vaccinations at time points indicated within the parentheses. ‡ Localized liver metastasis was completely resected before immunization.

ison of the number of SFU (ELISPOT) in replicates of experimental versus background wells.

RESULTS Patient Characteristics. Thirteen patients who underwent surgical resection for adenocarcinoma of the colon (n ⫽ 11) or rectum [n ⫽ 2 (American Joint Committee on Cancer stages II–IV)] with no evidence of remaining macroscopic disease were included in the study (Table 2). Median time from surgery to start of immunization was 4 months (range, 1– 44 months). The patients were immunized with Ep-CAM protein (n ⫽ 7) or human anti-Id mimicking Ep-CAM (n ⫽ 6) at weeks 0, 2, and 6 in conjunction with GM-CSF. In addition to the three initial immunizations, two patients in the Ep-CAM vaccine group and four patients in the anti-Id vaccine group received booster doses (see “Patients and Methods;” Table 2). Adverse Events. Vaccination was well tolerated (Table 3). Each patient developed injection site reactions, which were of grade 2 in 80% of the patients and grade 1 in 20% of the

Table 3

Local Systemic

patients. These resolved within a week. Multiple injections (⬎5) were associated with prolonged (up to 42 weeks) redness and swelling. Only mild (grade 1) systemic adverse events were noted. One patient, 2.5 years after vaccination with the Ep-CAM protein, developed a mild autoimmune thyroiditis. The patient received a low-dose thyroid hormone substitution. The symptoms subsequently resolved, and no further treatment was required. Ep-CAM–Specific IgG Response. None of the patients had Abs against Ep-CAM before vaccination. All patients immunized with the Ep-CAM protein elicited an Ep-CAM–specific IgG response (Fig. 1A). The highest Ab titers were seen after the third vaccine dose, which was followed by a gradual decrease. No specific anti-Ep-CAM Abs were detected in any of the patients immunized with anti-Id. The IgG Abs induced by Ep-CAM vaccination were predominantly of IgG1 type with a moderate IgG3 subclass response (Fig. 1B). A weak IgG2 response was also elicited. Only the IgG1 subclass was detected after 1 year. No IgG4 Abs were

Frequency of adverse events in Ep-CAM- or anti-Id–immunized patients Adverse event

Ep-CAM ⫹ GM-CSF (n ⫽ 7)

Anti-Id ⫹ GM-CSF (n ⫽ 6)

Injection site reaction* Injection site blister Hypoesthesia Fatigue Chills Fever Flu-like symptoms Myalgia Diarrhea Irritability Autoimmune thyroiditis†

7 (100) 2 (29) 0 (0) 0 (0) 1 (14) 1 (14) 1 (14) 0 (0) 1 (14) 0 (0) 1 (14)

6 (100) 3 (50) 1 (17) 1 (17) 0 (0) 2 (33) 0 (0) 1 (17) 0 (0) 2 (33) 0 (0)

NOTE. All values expressed as n (%). * Pain, redness, heat, swelling, and itching. † Diagnosed 2.5 years after immunization and resolved within the subsequent 2 years.

5396 Ep-CAM Induces MHC Class I-Restricted IFN-␥ Response

After vaccination, all but two patients (A7 and B5) mounted an Ep-CAM–specific proliferative T-cell response, which appeared more rapidly, was stronger, and was of longer duration in Ep-CAM- as compared with anti-Id–immunized patients (Fig. 4A). Booster vaccination induced no significant amplification of the Ep-CAM–specific proliferative T-cell response (data not shown). All anti-Id–immunized patients developed a proliferative T-cell response against anti-Id (Fig. 4B), which appeared at the same time as the anti-Ep-CAM T-cell response (Fig. 4A). The response against anti-Id was, however, sustained for a longer time period. Patients in both immunization groups also mounted a proliferative T-cell response against 18-mer Ep-CAM– derived peptides. A strong antipeptide response was usually associated with a potent proliferative T-cell response against the Ep-CAM and anti-Id proteins (data not shown). The proliferative T-cell response against the Ep-CAM protein could be significantly inhibited by mAbs against MHC class I and II molecules (data not shown), suggesting the induction of

Fig. 1 A, anti-Ep-CAM Abs (mean ⫹ SE; ELISA) of patients immunized with Ep-CAM (F) or anti-Id (䡺). B, anti-Ep-CAM IgG subclass Abs (mean ⫹ SE) of patients immunized with Ep-CAM. Results are shown as a ratio between post- and pre-immune anti-Ep-CAM IgG concentration (EU/ml) after subtraction of the IgG concentration against the control protein. Arrows indicate vaccination time points.

induced by the three initial vaccine doses. However, booster vaccination resulted in the emergence of all four IgG subclasses (Fig. 2). In one patient (patient A2), there was a switch in the IgG1:IgG4 ratio, whereas in another patient (patient A6), IgG1 remained the predominant subclass. To verify the specificity of anti-Ep-CAM Abs, sera of patients immunized with Ep-CAM were tested in Western blot against the Ep-CAM and control proteins, respectively. Postimmune sera specifically bound to the Ep-CAM protein under both nonreducing and reducing conditions, suggesting recognition of both conformational and linear epitopes. The anti-EpCAM mAb 3622W94 only reacted with nonreduced Ag. Preimmune sera were negative (Fig. 3). Ep-CAM–Specific Proliferative T-Cell Response. Patients were regularly monitored for an Ep-CAM–specific proliferative T-cell response until the test was negative at two consecutive time points (median, 2 years; range, 7– 69 months). None of the patients had a specific proliferative T-cell response against the Ep-CAM or anti-Id proteins before immunization.

Fig. 2 Effect of booster vaccination on the anti-Ep-CAM IgG subclass Ab titers in patients immunized with Ep-CAM (ELISA). Results are shown as a ratio between post- and pre-immune anti-Ep-CAM IgG subclass concentration (EU/ml) after subtraction of the corresponding IgG subclass concentration against the control protein. Arrows indicate vaccination time points. f, IgG1; ‚, IgG2; Œ, IgG3; E, IgG4.

Clinical Cancer Research 5397

anti-Id protein before vaccination. However, two HLA-A2⫹ patients had a positive response against the Ep-CAM– derived peptides, p174 –184 (patient B2) and p255–264 (patient B3). These responses disappeared after vaccination. Four patients (patients A6, A7, B3, and B5) had a positive response against HLA-DR1/DR4restricted p113–127 before immunization. After vaccination with Ep-CAM or anti-Id, specific IFN␥–producing cells against the Ep-CAM protein as well as against various Ep-CAM– derived MHC class I- and II-restricted peptides were detected (Table 4). Both the frequency and magnitude of the Ep-CAM–specific IFN-␥ response were similar in the two groups (data not shown). As an example, results for patient A6 after three immunizations with Ep-CAM are shown in Fig. 5A. IFN-␥ response was detected against an HLA-A2- and two HLA-DR1-restricted peptides. However, no response against the whole Ep-CAM protein was noted. As another example, results for patient A2 are shown in Fig. 5B. After the three initial Ep-CAM vaccinations, a weak IFN-␥ response was seen against an HLA-A26 –restricted peptide, but not against the whole Ep-CAM protein. After three booster

Fig. 3 Western blotting for anti-Ep-CAM IgG Abs of patient A3 vaccinated with the Ep-CAM protein. A, nonreducing conditions; B, reducing conditions. Lanes 1 and 2 show reaction against Ep-CAM and CEA used as a negative control Ag, respectively. mAb 3622W94 (humanized anti-Ep-CAM mAb) and an anti-CEA-positive serum were used as positive controls for Ep-CAM and CEA, respectively.

Ep-CAM–specific T cells within both CD4⫹ and CD8⫹ T-cell subsets. Interferon-␥-Producing Cells against the Ep-CAM Protein and Ep-CAM–Derived Peptides (ELISPOT). PBMCs were tested before and at an average of four different time points after vaccination for IFN-␥–producing T cells. None of the patients had a specific IFN-␥ response against the whole Ep-CAM or

Fig. 4 Proliferative T-cell response against Ep-CAM (A) and anti-Id (B) of patients immunized with Ep-CAM (F) or anti-Id (䡺). Results are shown as a ratio between post- and pre-immune SI value (mean ⫹ SE) against the antigens (Ep-CAM and anti-Id) after subtraction of the SI value against the control proteins. Arrows indicate vaccination time points.

5398 Ep-CAM Induces MHC Class I-Restricted IFN-␥ Response

Table 4

IFN-␥–producing T cells of patients vaccinated with the Ep-CAM protein (A) or anti-Id (B) recognizing the Ep-CAM protein and Ep-CAM– derived peptides (ELISPOT) Protein antigens

Patient Ep-CAM BCP CEA A1 A2 A3 A4 A5 A6 A7 B1 B2 B3 B4 B5 B6

⫹ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫺ ⫹ ⫹ ⫺ ⫹

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

Ep-CAM–derived peptides

HIV RT peptides

MHC class I

MHC class II

p174–184, p184–193 p80–88, p206–215, p241–250 n.a. p6–14, p174–184, p184–193 p23–32, p184–193 p6–14 p172–180 p6–14, p174–184, p184–193, p255–264 p6–14 p6–14 p23–32

p113–127 n.a. n.a. n.a. p17–31, p113–127 p29–43, p113–127* p113–127† n.a. n.a. p29–43, p66–80, p113–127* n.a. p113–127† n.a.

p6–14, p177–185, p255–264, p256–264, p263–271

MHC class I MHC class II ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ n.a. n.a. n.a. ⫺ ⫺ ⫺ n.a. n.a. ⫺ n.a. ⫺ n.a.

HLA haplotype A

B

DR

2, 9 24, 26 3, 24 2, 24 1, 2 2, 28 2 2, 19 1, 2 2, 3 1, 25 2 2, 3

5, 40 27 35, 60 7, 60 8, 15 44 27, 62 12, 40 7, 37 44, 62 5, 44 60, 62 7, 62

1, 53 9 11, 13 13, 14 3, 4 1, 13 1, 8 6, 7 12, 15 4 7, 15 4, 13 13, 15

NOTE. A positive response at any tested time-point is indicated by ⫹, and by listed peptides (see Patients and Methods for criteria of positivity). A negative response is indicated by ⫺. Abbreviation: n.a., no peptides were available that matched the binding motif of the patients’ HLA type. * Specific IFN-␥ response before and after vaccination. † Specific IFN-␥ response before vaccination, which was significantly boosted by vaccination.

vaccinations, the response against the HLA-A26 –restricted peptide was amplified, and a response against another HLA-A26 – restricted peptide as well as against the complete Ep-CAM protein emerged. As an example of anti-Id immunization, results for patient B6 are shown in Fig. 5C. A response against three different HLA-A2–restriced peptides and one HLA-B7restricted peptide as well as the Ep-CAM protein and anti-Id was noted. Three of these peptides (p6 –14, p255–264, and p256 – 264) contain identical anchor residues and show ⬎33% amino acid sequence homology with peptides derived from the CDRs of anti-Id. Anti-Id booster vaccination did not have a significant impact on the IFN-␥ response, except for one patient (patient B1; data not shown). Significant IFN-␥ production was induced by Ep-CAMpositive irradiated CRC cell lines matched for HLA-A2 (LS180) or HLA-DR1 (SW837) as compared with HLA-unmatched Ep-CAM-positive CRC (SW948) and HLA-DR1⫹ Ep-CAMnegative Burkitt lymphoma (BL-41) cells (Fig. 6). A significant reduction (P ⫽ 0.016) of the number of IFN-␥-secreting cells was observed when LS180 cells (MHC class I⫹ and class II⫺) were incubated with anti-MHC class I mAb, whereas the control mAb had no significant effect (data not shown). Phenotypes of Ep-CAM–Specific Interferon-␥- and Perforin-Producing Cells. Intracellular IFN-␥ and perforin were assessed in phenotypically characterized cells by fourcolor flow cytometry after in vitro stimulation with the Ep-CAM protein. The results for patient A1 are shown in Table 5. EpCAM–specific IFN-␥–producing cells were predominantly CD8dimCD56⫺ T or CD8dimCD56⫹ natural killer (NK)-like T cells. Practically no IFN-␥ production was detected in CD4⫹ T cells. Perforin-producing Ep-CAM–specific cells resided within the CD8brightCD56⫺ classical cytotoxic T cells and CD8dimCD56⫹ NK-like T cells. Perforin was also induced in CD8dimCD56⫺ T and CD4⫹CD56⫹ NK-like T cells. Furthermore, a CD8⫺ T-cell subset, either CD56⫺ or CD56⫹, as well as NK cells

(CD3⫺CD8dimCD56⫹) also produced IFN-␥ and perforin after Ep-CAM stimulation. Overall Immune Response and Survival. A summary of the overall immune response induced by vaccination and clinical status is shown in Table 6. The design of the study does not allow us to draw any conclusion with regard to the clinical efficacy of an Ep-CAM–specific immune response.

DISCUSSION In this study, the safety and immunogenicity of vaccination of CRC patients with minimal residual disease using Ep-CAM as a target structure were evaluated. A recombinant Ep-CAM protein was compared with a human anti-Id vaccine mimicking Ep-CAM. As an adjuvant, GM-CSF was used. Both vaccines were well tolerated. Systemic adverse events consisted mainly of mild constitutional symptoms. Local injection site reactions were common but not severe. The present vaccine formulations were better tolerated than Ep-CAM expressed in avipox vector (ALVAC; Ref. 9). One patient immunized with Ep-CAM developed a mild and transient autoimmune thyroiditis, which is occasionally associated with immunotherapy (29). Immunization with Ep-CAM, but not anti-Id, induced anti-Ep-CAM IgG Abs. Ab reactivity to both conformational and linear epitopes was noted. The subclass responses were predominantly IgG1 and IgG3. In humans, a Th1-polarized immune response is associated with IgG1 and IgG3 (18), which are the most effective subclasses mediating phagocytosis, complement-dependent cytolysis, and Ab-dependent cellular cytotoxicity (30). After booster vaccination, IgG4 also emerged, most likely reflecting prolonged antigenic stimulation (31). The IgG subclass profile can be influenced by several factors. The choice of adjuvant might be important. Previous studies have shown that Ag precipitated in

Clinical Cancer Research 5399

alum induced a Th2-polarized response (32). Alum-precipitated recombinant CEA combined with GM-CSF induced mainly IgG1 and IgG2 but no IgG3 responses after three initial vaccinations. Repeated vaccinations, however, resulted in the increase of IgG4 titers, which is in line with the present study (33). Immunization with the Ep-CAM protein induced a rapid and long-lasting Ep-CAM–specific proliferative T-cell response as compared with the anti-Id vaccine, comprising both MHC class I- and II-restricted T cells. Baculovirus-derived Ep-CAM protein precipitated in alum has previously been used for vaccination of CRC patients (10). However, the present immunization regimen induced a stronger proliferative T-cell response, which might be due to the use of GM-CSF as an adjuvant as compared with alum. Although Abs and type 1 CD4⫹ T cells have an important

Fig. 6 IFN-␥–producing cells (ELISPOT) recognizing CRC cell lines of an Ep-CAM–immunized patient (patient A6, HLA-A2⫹ and DR1⫹). The number of SFU (mean ⫾ SE) is shown. Asterisks indicate a significantly (P ⬍ 0.05) greater number of SFU of cells stimulated with CRC cell lines as compared with control (BL-41). Table 5 Phenotypic characteristics of Ep-CAM–specific IFN-␥- and perforin-producing cells after vaccination (patient A1; four-color flow cytometry) Cell subset (gated region) Fig. 5 IFN-␥–producing cells (ELISPOT) recognizing the Ep-CAM protein as well as Ep-CAM– derived MHC class I- and II-restricted peptides in patients vaccinated with Ep-CAM (A2 and A6) or anti-Id (B6). A, patient A6 before (䡺) and 20 weeks after (f) primary immunization. B, patient A2 before (䡺) and 20 weeks after (f) primary immunization, before booster vaccine (s; see Table 2), and 5 weeks after the third booster vaccination (o). C, patient B6 2.5 years after primary immunization. The number of SFU (mean ⫾ SE) is shown after subtraction of the background value (cells in medium alone). Asterisks indicate a significantly (P ⬍ 0.05) greater number of SFU in experimental wells as compared with background.

CD3⫹ CD3⫹ CD3⫹ CD3⫹ CD3⫺ CD3⫺

dim

CD8 CD8bright CD4⫹ CD8⫺ CD8dim CD8⫺

IFN-␥

Perforin

CD56⫺

CD56⫹

CD56⫺

CD56⫹

4.2 1.0 0 3.0 0.3 0.9

4.0 1.2 0.6 1.9 10.3 1.4

2.5 4.9 0 1.1 0 0

6.6 0 2.7 2.6 2.4 0

NOTE. Results are shown as the percentage of cells within each gated region of Ep-CAM protein-stimulated cells after subtraction of the value of control protein (BCP)-stimulated cells. When similar comparison was made between cells stimulated with BCP and medium alone, the variation was 0.4 ⫾ 0.6% (mean ⫾ SD). The variation between Ep-CAM- and BCP-stimulated cells in a healthy control was 0.2 ⫾ 0.2% (mean ⫾ SD).

5400 Ep-CAM Induces MHC Class I-Restricted IFN-␥ Response

Table 6 Summary of Ep-CAM–specific immune response in patients vaccinated with Ep-CAM protein or anti-Id Ep-CAM–specific response Humoral Patient

IgG

A1 A2 A3 A4 A5 A6 A7 B1 B2 B3 B4 B5 B6

⫹ ⫹ ⫹ ⫹ ⫹ ⫹ ⫹† ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

Cellular Proliferative IFN-␥ Clinical status ⫹ ⫹ ⫹ ⫹ ⫹† ⫹ ⫺ ⫹† ⫹ ⫹ ⫹ ⫺ ⫹†

⫹ ⫹ ⫺ ⫹ ⫹ ⫹ ⫹† ⫹ ⫹† ⫹ ⫹† ⫹† ⫹

NED NED LTR (11 m) NED LTR (7 m) NED NED NED DM (13 m) NED DM (9 m) NED NED‡

Survival (mo)* 74⫹ 54⫹ 12.5 30⫹ 10 47⫹ 16⫹ 39⫹ 48 28⫹ 39 20⫹ 39

Abbreviations: NED, no evidence of disease; LTR, local tumor recurrence; DM, distant metastasis. * Survival is shown from start of immunization until follow-up in patients alive (⫹) or until death. † A positive response was detected at one time point only. ‡ Died of disease (pneumonia) not related to CRC.

role in antitumor immunity, CD8⫹ cytotoxic T lymphocyte (CTLs) are crucial (16). IFN-␥–producing T cells might be considered as a surrogate marker for identification of CTL precursors (34, 35). In this study, a long-lasting IFN-␥ T-cell response against the Ep-CAM protein and Ep-CAM– derived MHC class I- and II-restricted peptides was detected after immunization with either Ep-CAM or anti-Id. An Ep-CAM– specific T-cell response after anti-Id vaccination indicates crossreactivity between the bona fide and internal image Ags. The fact that IFN-␥–producing T cells against MHC class Irestricted peptides were induced suggests that Ep-CAM and anti-Id as exogenous protein Ags were cross-presented in vivo. Antigen-presenting cells can capture and deliver exogenous Ags into the MHC class I processing pathway, and GM-CSF may facilitate this process (36). Interestingly, vaccination of patients with alum-precipitated Ep-CAM without GM-CSF did not induce ex vivo IFN-␥ production (10). It has previously been suggested that immunization with Ep-CAM protein maybe superior to anti-Id (10, 14, 37). However, the authors used pooled delayed-type hypersensitivity and proliferative responses as a basis for comparison of cellular immune responses, despite the fact that delayed-type hypersensitivity was only assessed in Ep-CAM–immunized patients. The response rate in the two immunization groups seemed comparable in terms of both anti-Ep-CAM Abs (6 of 12 versus 7 of 13 patients) and proliferative T-cell response (3 of 12 versus 2 of 13 patients; Refs. 10 and 37). Thus, the published data do not support the conclusion made by the authors. Our data suggest that the bona fide Ag was more effective in inducing a humoral response as compared with anti-Id (7 of 7 versus 0 of 6 patients). Although there was no difference in the proliferative response rate between the two groups, the duration of the response was superior in Ep-CAM–immunized patients. However, an IFN-␥ response of a similar frequency and magnitude was induced by Ep-CAM and anti-Id.

Limited information is available with regard to immunogenic MHC class I-restricted Ep-CAM epitopes, and there is no report on class II-restricted Ep-CAM epitopes. Three HLA-A2– restricted Ep-CAM– derived peptides (p263–271, p184 –192, and p184 –193) have been shown to be immunogenic (4, 27, 38). In the present study, a number of additional functional MHC class Iand II-restricted Ep-CAM– derived peptides were identified in the context of various HLA alleles. The results also corroborate the notion that a polyclonal T-cell response was induced within both the CD4⫹ and CD8⫹ subsets. The immunogenicity of some peptides (p174 –184 and p113–127) was also confirmed in patients immunized with ALVAC-Ep-CAM (9).6 The high frequency of patients (4 of 6 patients) mounting an IFN-␥ response against the MHC class II-restricted p113–127 before immunization might imply the involvement of this epitope in a natural T-cell response against Ep-CAM. However, a larger number of HLA-matched patients and healthy donors must be tested to verify this potentially interesting observation. An IFN-␥ response to this peptide was also seen in a healthy donor (1 of 5 healthy donors), which is not surprising because it has been indicated that T cells capable of reacting against Ep-CAM epitopes are not fully deleted from the immune repertoire (38). CTL effector function of CD8⫹ T cells may correlate with the expression of CD56, and such cells express high levels of IFN-␥ and perforin (39, 40). In response to Ep-CAM protein, we could show that various T-cell subsets secreted IFN-␥ and perforin. Both cytokines were produced primarily by CD8⫹ T cells with or without the expression of CD56. Only CD4⫹ T cells expressing CD56 produced perforin. There was also induction of IFN-␥ and perforin producing CD4⫺CD8⫺ T cells, either CD56⫺ or CD56⫹, most likely representing ␦␥ T cells (41). NK cells were also activated, probably as a bystander effect. Generation of a phenotypically and functionally diverse Ag-specific T-cell repertoire has also been reported after HER2/neu-based vaccination (42). A detailed characterization of Ep-CAM–specific CD56⫹ T cells is of great interest and warranted in future studies. These cells may represent CD1drestricted NKT cells (43). Tumor cells engineered to secrete GM-CSF have been shown to induce expansion of CD1drestricted T cells, which were critical for antitumor immunity in a murine model (44). In conclusion, the present study demonstrates that immunization with a recombinant Ep-CAM protein or anti-Id in combination with GM-CSF was safe and generated a polyclonal functional repertoire of Ag-specific T cells secreting IFN-␥ and/or perforin. The bona fide Ag, in contrast to anti-Id, induced an Ep-CAM–specific Th1-biased Ab and a long-lasting proliferative T-cell response. Thus, the Ep-CAM protein seems to be preferable, as compared with anti-Id, for vaccine development. Moreover, several immunogenic MHC class I- and II-restricted Ep-CAM epitopes were identified that might be of importance for future rational vaccine design.

6

Unpublished data.

Clinical Cancer Research 5401

ACKNOWLEDGMENTS We thank Dr. E. D. Rossmann and R. Lenkei for valuable advice in flow cytometry.

REFERENCES 1. Tebbutt NC, Cattell E, Midgley R, Cunningham D, Kerr D. Systemic treatment of colorectal cancer. Eur J Cancer 2002;38:1000 –15. 2. Balzar M, Winter MJ, de Boer CJ, Litvinov SV. The biology of the 17-1A antigen (Ep-CAM). J Mol Med 1999;77:699 –712. 3. Mosolits S, Steinitz M, Harmenberg U, et al. Immunogenic regions of the GA733-2 tumour-associated antigen recognised by autoantibodies of patients with colorectal carcinoma. Cancer Immunol Immunother 2002;51:209 –18. 4. Nagorsen D, Keilholz U, Rivoltini L, et al. Natural T-cell response against MHC class I epitopes of epithelial cell adhesion molecule, her-2/neu, and carcinoembryonic antigen in patients with colorectal cancer. Cancer Res 2000;60:4850 – 4. 5. Riethmuller G, Holz E, Schlimok G, et al. Monoclonal antibody therapy for resected Dukes’ C colorectal cancer: seven-year outcome of a multicenter randomized trial. J Clin Oncol 1998;16:1788 –94. 6. Punt CJ, Nagy A, Douillard JY, et al. Edrecolomab alone or in combination with fluorouracil and folinic acid in the adjuvant treatment of stage III colon cancer: a randomised study. Lancet 2002; 360:671–7. 7. Fields A, Keller A, Schwartzberg L, et al. Edrecolomab (17-1A antibody) in combination with 5-fluorouracil (FU) based chemotherapy in the adjuvant treatment of stage III colon cancer: results of a randomized North American phase III study. Proc Am Soc Clin Oncol 2002; 21:128a. 8. Basak S, Eck S, Gutzmer R, et al. Colorectal cancer vaccines: antiidiotypic antibody, recombinant protein, and viral vector. Ann N Y Acad Sci 2000;910:237–52; discussion 252–3. 9. Ullenhag GJ, Frodin JE, Mosolits S, et al. Immunization of colorectal carcinoma patients with a recombinant canarypox virus expressing the tumor antigen Ep-CAM/KSA (ALVAC-KSA) and granulocyte macrophage colony-stimulating factor induced a tumor-specific cellular immune response. Clin Cancer Res 2003;9:2447–56. 10. Staib L, Birebent B, Somasundaram R, et al. Immunogenicity of recombinant GA733-2E antigen (CO17-1A, EGP, KS1-4, KSA, Ep-CAM) in gastro-intestinal carcinoma patients. Int J Cancer 2001; 92:79 – 87. 11. Bhattacharya-Chatterjee M, Chatterjee SK, Foon KA. Anti-idiotype vaccine against cancer. Immunol Lett 2000;74:51– 8. 12. Fagerberg J, Steinitz M, Wigzell H, Askelof P, Mellstedt H. Human anti-idiotypic antibodies induced a humoral and cellular immune response against a colorectal carcinoma-associated antigen in patients. Proc Natl Acad Sci USA 1995;92:4773–7. 13. Mellstedt H, Fagerberg J, Frodin JE, et al. Augmentation of the immune response with granulocyte-macrophage colony-stimulating factor and other hematopoietic growth factors. Curr Opin Hematol 1999; 6:169 –75. 14. Birebent B, Somasundaram R, Purev E, et al. Anti-idiotypic antibody and recombinant antigen vaccines in colorectal cancer patients. Crit Rev Oncol Hematol 2001;39:107–13. 15. Bhattachary-Chatterjee M, Nath Baral R, Chatterjee SK, et al. Counterpoint. Cancer vaccines: single-epitope anti-idiotype vaccine versus multiple-epitope antigen vaccine. Cancer Immunol Immunother 2000;49:133– 41. 16. Jager E, Jager D, Knuth A. Antigen-specific immunotherapy and cancer vaccines. Int J Cancer 2003;106:817–20. 17. Clay TM, Hobeika AC, Mosca PJ, Lyerly HK, Morse MA. Assays for monitoring cellular immune responses to active immunotherapy of cancer. Clin Cancer Res 2001;7:1127–35. 18. Widhe M, Ekerfelt C, Forsberg P, Bergstrom S, Ernerudh J. IgG subclasses in Lyme borreliosis: a study of specific IgG subclass distri-

bution in an interferon-gamma-predominated disease. Scand J Immunol 1998;47:575– 81. 19. Strassburg CP, Kasai Y, Seng BA, et al. Baculovirus recombinant expressing a secreted form of a transmembrane carcinoma-associated antigen. Cancer Res 1992;52:815–21. 20. Steinitz M, Tamir S, Frodin JE, Lefvert AK, Mellstedt H. Human monoclonal anti-idiotypic antibodies. I. Establishment of immortalized cell lines from a tumor patient treated with mouse monoclonal antibodies. J Immunol 1988;141:3516 –22. 21. Fagerberg J, Frodin JE, Ragnhammar P, et al. Induction of an immune network cascade in cancer patients treated with monoclonal antibodies (ab1). II. Is induction of anti-idiotype reactive T cells (T3) of importance for tumor response to mAb therapy? Cancer Immunol Immunother 1994;38:149 –59. 22. Samanci A, Yi Q, Fagerberg J, et al. Pharmacological administration of granulocyte/macrophage-colony-stimulating factor is of significant importance for the induction of a strong humoral and cellular response in patients immunized with recombinant carcinoembryonic antigen. Cancer Immunol Immunother 1998;47:131– 42. 23. Szala S, Froehlich M, Scollon M, et al. Molecular cloning of cDNA for the carcinoma-associated antigen GA733-2. Proc Natl Acad Sci USA 1990;87:3542– 6. 24. Fagerberg J. Structural and immunological characterization of human anti-idiotypic antibodies used for immunization of colorectal carcinoma patients. In: Fagerberg J. The idiotypic network response in cancer patients treated with monoclonal antibodies (Doctoral dissertation). Stockholm, Sweden: Karolinska Institutet; 1995. ISBN 91-6281625-X. 25. Rammensee H, Bachmann J, Emmerich NP, Bachor OA, Stevanovic S. SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics 1999;50:213–9. 26. Parker KC, Bednarek MA, Coligan JE. Scheme for ranking potential HLA-A2 binding peptides based on independent binding of individual peptide side-chains. J Immunol 1994;152:163–75. 27. Ras E, van der Burg SH, Zegveld ST, et al. Identification of potential HLA-A*0201 restricted CTL epitopes derived from the epithelial cell adhesion molecule (Ep-CAM) and the carcinoembryonic antigen (CEA). Hum Immunol 1997;53:81–9. 28. Magnusson CG, Nelson DF, Magnusson MA. Disproportional distribution of isotype and non-isotype-specific IgG subclass anti-IgE autoantibodies in human cord serum. Int Arch Allergy Immunol 1996; 110:31– 40. 29. Vial T, Descotes J. Immune-mediated side-effects of cytokines in humans. Toxicology 1995;105:31–57. 30. Papadea C, Check IJ. Human immunoglobulin G and immunoglobulin G subclasses: biochemical, genetic, and clinical aspects. Crit Rev Clin Lab Sci 1989;27:27–58. 31. Tomee JF, Dubois AE, Koeter GH, et al. Specific IgG4 responses during chronic and transient antigen exposure in aspergillosis. Am J Respir Crit Care Med 1996;153:1952–7. 32. Brewer JM, Conacher M, Hunter CA, et al. Aluminium hydroxide adjuvant initiates strong antigen-specific Th2 responses in the absence of IL-4- or IL-13-mediated signaling. J Immunol 1999;163: 6448 –54. 33. Ullenhag GJ, Frodin JE, Strigard K, Mellstedt H, Magnusson CG. Induction of IgG subclass responses in colorectal carcinoma patients vaccinated with recombinant carcinoembryonic antigen. Cancer Res 2002;62:1364 –9. 34. Derby EG, Reddy V, Nelson EL, et al. Correlation of human CD56⫹ cell cytotoxicity and IFN-gamma production. Cytokine 2001; 13:85–90. 35. Ghanekar SA, Nomura LE, Suni MA, et al. Gamma interferon expression in CD8(⫹) T cells is a marker for circulating cytotoxic T lymphocytes that recognize an HLA A2-restricted epitope of human cytomegalovirus phosphoprotein pp65. Clin Diagn Lab Immunol 2001; 8:628 –31.

5402 Ep-CAM Induces MHC Class I-Restricted IFN-␥ Response

36. Arina A, Tirapu I, Alfaro C, et al. Clinical implications of antigen transfer mechanisms from malignant to dendritic cells. Exploiting crosspriming. Exp Hematol 2002;30:1355– 64. 37. Somasundaram R, Zaloudik J, Jacob L, et al. Induction of antigenspecific T and B cell immunity in colon carcinoma patients by antiidiotypic antibody. J Immunol 1995;155:3253– 61. 38. Trojan A, Witzens M, Schultze JL, et al. Generation of cytotoxic T lymphocytes against native and altered peptides of human leukocyte antigen-A*0201 restricted epitopes from the human epithelial cell adhesion molecule. Cancer Res 2001;61:4761–5. 39. Pittet MJ, Speiser DE, Valmori D, Cerottini JC, Romero P. Cutting edge: cytolytic effector function in human circulating CD8⫹ T cells closely correlates with CD56 surface expression. J Immunol 2000;164: 1148 –52. 40. Santin AD, Hermonat PL, Ravaggi A, et al. Expression of CD56 by human papillomavirus E7-specific CD8⫹ cytotoxic T lymphocytes cor-

relates with increased intracellular perforin expression and enhanced cytotoxicity against HLA-A2-matched cervical tumor cells. Clin Cancer Res 2001;7(suppl):804s–10s 41. Carding SR, Egan PJ. Gammadelta T cells: functional plasticity and heterogeneity. Nat Rev Immunol 2002;2:336 – 45. 42. Knutson KL, Disis ML. Clonal diversity of the T-cell population responding to a dominant HLA-A2 epitope of HER-2/neu after active immunization in an ovarian cancer patient. Hum Immunol 2002;63: 547–57. 43. Taniguchi M, Harada M, Kojo S, Nakayama T, Wakao H. The regulatory role of Valpha14 NKT cells in innate and acquired immune response. Annu Rev Immunol 2003;21:483–513. 44. Gillessen S, Naumov YN, Nieuwenhuis EE, et al. CD1d-restricted T cells regulate dendritic cell function and antitumor immunity in a granulocyte-macrophage colony-stimulating factor-dependent fashion. Proc Natl Acad Sci USA 2003;100:8874 –9.