Monoclonal Anti-Idiotype Antibody 6G6.C4 Fused to GM-CSF Is Capable of Breaking Tolerance to Carcinoembryonic Antigen (CEA) in CEA–Transgenic Mice

Research Article

Monoclonal Anti-idiotype Antibody 6G6.C4 Fused to GM-CSF Is Capable of Breaking Tolerance to Carcinoembryonic Antigen (CEA) in CEA–Transgenic Mice

Christian Schwegler,1 Alexandra Dorn-Beineke,2 Stefanie Nittka,2 Carol Stocking,3 and Michael Neumaier2

1Department of Internal Medicine and Clinical Immunology Bad Bramstedt, University of Lu¨beck, Lu¨beck, Germany; 2Institute for Clinical Chemistry, Faculty for Clinical Medicine Mannheim, University of Heidelberg, Heidelberg, Germany; and 3Heinrich-Pette Institute for Virology, University of Hamburg, Hamburg, Germany

Abstract maintained in >95% of the colorectal carcinomas and their Internal image anti-idiotypic antibodies are capable of metastases, making this glycoprotein the most widely used mimicking tumor-associated antigens and thus may serve as human tumor marker (1). Circulating CEA can be detected in surrogate for vaccination strategies in cancer patients. The the serum of patients with CEA-positive tumors and has been monoclonal antibody (mAb) 6G6.C4 mimics an epitope extensively used as a serum marker in the monitoring of tumor specific for the human carcinoembryonic antigen (CEA) and stage in these cancer patients. A number of studies have been generates a CEA-specific response (Ab3) in various experi- reported recently in which monoclonal anti-CEA antibodies have mental animals. In humans, however, 6G6.C4 only yields a been used for antigen-specific passive immunotherapy in cancer very limited humoral anti-CEA reaction presumably due to patients (2, 3). tolerance against the CEA autoantigen. In this study, we In contrast, an active specific humoral immunotherapy would investigated the CEA-specific Ab3 response in mice transgenic aim to sensitize the patient’s immune system against its tumor cell for the human CEA and tested whether the antigen tolerance using tumor-associated antigens (TAA) as vaccines allowing for a could be overcome by fusing a recombinant single-chain self-sustained immunity against the cancer cells. variable fragment of 6G6.C4 (scFv6G6.C4) to the murine According to Niels Jerne’s idiotypic network hypothesis, anti- granulocyte macrophage colony-stimulating factor (GM-CSF). genic epitope structures can be mirrored through an anti-idiotypic Like mAb 6G6.C4, the fusion protein (scFv6G6.C4/GM-CSF) cascade of antibodies. Of these, the so-called internal image anti- h retained binding to the CEA-specific idiotype mAb T84.66. idiotypic (Ab2 ) antibodies can mimic epitopes of self-antigens Also, scFv6G6.C4/GM-CSF was biologically active as measured and serve for tumor vaccination strategies (4–8). by proliferation of the GM-CSF–dependent murine FDC-P1 Clinical studies have shown the induction of TAA-specific h cells in vitro. After immunization with the scFv6G6.C4/GM- antibody (Ab3) responses by Ab2 anti-idiotypes in tumor patients CSF fusion protein, CEA-transgenic animals showed signifi- bearing a variety of tumors (9–14). h cantly enhanced Ab3 antibody responses to scFv6G6.C4 It has been shown that the Ab2 against the monoclonal (P = 0.005) and to CEA (P = 0.012) compared with the antibody (mAb) chT84.66, a chimeric human/mouse mAb recog- scFV6G6.C4 alone. Sera from mice immunized with the fusion nizing a CEA-specific epitope, can serve as a surrogate antigen for protein specifically recognized CEA in Western blot analyses CEA in experimental animals and is capable to generate an anti- with no cross-reaction to CEA-related antigens. Finally, the CEA response with the same specificity as the T84.66 idiotype Ab3 antisera detected single CEA-expressing tumor cells in antibody (15, 16). suspension as shown by flow cytometry. Taken together, these To improve the efficacy of the antitumor response, antibodies data show in a model antigenically related to the human have been fused to cytokines or growth factors including system that vaccination with scFv6G6.C4/GM-CSF improves interleukin (IL)-2, IL-6, granulocyte macrophage colony-stimulating vaccination against an endogenous tumor-associated antigen factor (GM-CSF; refs. 17–21). resulting in a highly specific humoral Ab3 response in vivo GM-CSF is an immunoregulatory cytokine that induces the that is capable of bind single circulating CEA-positive tumor growth of granulocyte and monocyte progenitors and possesses cells. (Cancer Res 2005; 65(5): 1925-33) important functions in the differentiation and maturation of hematopoietic cells, up-regulates class II MHC molecules (22), Introduction and enhances their phagocytic activity and antigen-presenting capabilities (23). GM-CSF stimulates the proliferation and survival The human carcinoembryonic antigen (CEA), originally des- of Langerhans cells and dendritic cells in vitro (24) and has been cribed as a tumor-specific antigen, is a member of the CEA- postulated to be a critical mediator in the primary immune related cell adhesion molecule (CEACAM) family expressed on the response because of its action on these and other antigen- apical surface of colon epithelial cells. CEA expression is presenting cells (25). In vitro and in vivo studies have shown that GM-CSF fused to antibody molecules retains its biological activity and continues Requests for reprints: Michael Neumaier, Institute for Clinical Chemistry, to mediate both antibody-dependent cellular cytotoxicity and University Hospital Mannheim, University of Heidelberg, Theodor-Kutzer-Ufer complement-dependent cytotoxicity (20, 26). 1-3, 68167 Mannheim, Germany. Phone: 49-621-383-2222; Fax: 49-621-383-3819; E-mail: [email protected]. In earlier studies, we have raised the monoclonal anti-idiotype I2005 American Association for Cancer Research. 6G6.C4 against the anti-CEA mAb chT84.66 (15) known to possess www.aacrjournals.org 1925 Cancer Res 2005; 65: (5). March 1, 2005

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2005 American Association for Cancer Research. Cancer Research a high affinity and specificity for CEA. It has been shown that CAGTCT and VL3.2_XhoI (32 nt) 5V-AGCTCGAGGCCACGTTTGATTTC- 6G6.C4 is capable to elicit a CEA-specific response in experimental CCAGCTTGGT were used to amplify the light chain variable (VL) region animals and thus may have the potential to function as a vaccine in gene. After purification by phenol/chloroform extraction and precipitation, active immunotherapy against CEA-positive carcinomas (15, 16). the VH and VL amplicons were digested with BamHI and BglII, respectively, and subsequently ligated using T4-ligase (New England Biolabs GmbH, In this study, we report the design of the single-chain (scFv) fusion Frankfurt am Main, Germany) overnight at 16jC. The ligation products protein scFv6G6.C4/GM-CSF by genetic engineering and character- were purified from 1.5% agarose gels using the QiaEx method according to ize its capability to elicit a specific humoral anti-CEA response the manufacturer’s recommendation (Qiagen, Hilden, Germany) and (Ab3 response) in CEA-tolerant transgenic mice. We also show that reamplified by Pfu Polymerase (Stratagene, La Jolla, CA) using VH5.1_NdeI Ab3 antibodies generated by vaccination with scFv6G6.C4/GM-CSF and VL3.2_XhoI as primers. After phenol/chloroform extraction and will detect single CEA-expressing tumor cells by flow cytometry. precipitation, the scFv fragment was restricted with NdeI and XhoI and ligated into pET27b(+) plasmids (EMD Biosciences, Inc., Novagen Brand, Madison, WI) previously cut with NdeI and XhoI. Clones with correct scFv Materials and Methods genes were verified by their resistance against BamHI and BglII and their Antigens and Antibodies. Purified CEA and murine mAbs T84.1 and sensitivity against BstYI. Finally, the integrity of the recombinant scFv was mT84.66 were kindly provided by J.E. Shively (Division of Immunology, verified by DNA sequencing. Beckman Research Institute of the City of Hope, Duarte, CA). The chimeric The GM-CSF cassette was amplified from pUC18-BBG39 (British Biotech, human/mouse mAb chT84.66 has been described previously (27). The Oxford, United Kingdom) harboring the murine GM-CSF as a designer gene murine mAbs BY114 and 29H2 were purchased from InnoGenex (San with E. coli codon usage for optimal expression. PCR amplification was done Ramon, CA) and Novocastra Laboratories (Newcastle upon Tyne, United using GM-CSFL1_NdeI (29 nt) 5V-TACATATGGCTCCGACGCGAGCCCGATC Kingdom) respectively. and GM-CSFR1_SfiI(33nt)5V-GGGCCGACTGGGCCTTTTG- Cell Lines and Protein Extracts. Jurkat wild-type cells, the murine GACTGGTTTTTTGC, and the product was cloned into pCRII (Invitrogen colon carcinoma cell line MC38 and its CEA-transfected subline C15-A.3 GmbH, Karlsruhe, Germany). Clones with the GM-CSF cassette in the and its CEACAM1-transfected subline MC38-BGP, and CEACAM6- orientation of the lacZa transcription were verified and further processed transfected HeLa-NCA cells were kindly provided by J.E. Shively. The by SfiI and XhoI digestion. The 6G6.C4 scFv cassette was PCR-amplified human colon carcinoma cell line HT29 was purchased from American using VH5.2_SfiI (38nt) 5V-CCAAGGCCCAGTCGGCCCAGGTCCAACTA- Type Culture Collection. Human granulocytes were obtained from CAGCAGCCT and VL3.2_XhoI (see above) using PfuI Polymerase (Stratagene) healthy donors after informed consent. Human normal colon tissue was and subsequently restricted with SfiIandXhoI. After phenol/chloroform obtained from surgical specimens. All cell lines were grown in DMEM extraction and precipitation, the digested DNA fragment was ligated or RPMI (Life Technologies, Karlsruhe, Germany) containing 10% FCS, 4 into the recombinant pCRII/mGM-CSF plasmid previously digested with mmol/L glutamine (PRA International, Lenexa, KS), and penicillin/ SfiIandXhoI. streptomycin (100 units/mL and 10 Ag/mL, Biochrom, Berlin, Germany). Bacterial Expression and Preparation of Protein Fractions. Protein Cell extracts were prepared for Western blot analyses in TNE buffer expression was done using BL21DE3 cells previously transformed with the [50 mmol/L Tris (Life Technologies), 150 mmol/L NaCl, 2 mmol/L EDTA recombinant pET27(+) plasmids (EMD Biosciences, Inc., Novagen Brand). (Merck, Darmstadt, Germany), 1% Triton X-100 (pH 7.6, Sigma-Aldrich, Bacteria were grown in Luria-Bertani medium with 100 Ag/mL kanamycin

Taufkirchen, Germany)]. As proteinase inhibitors, 1 mol/L Na3OV4, 1 mg/ (Life Technologies) to an absorbance of 0.8 (E = 600 nm) after which mL aprotinin/leupeptin, and 0.5 mol/L NaF were included. isopropyl-L-thio-B-D-galactopyranoside (MBI Fermentas GmbH, St. Leon- Granulocyte membrane extracts were prepared as follows: Granulocytes Rot, Germany) was added to a final concentration of 1 mmol/L. Culture was were isolated from buffy coats following lysis of erythrocytes in a buffer continued at 40jC for additional 4 hours. j containing 155 mmol/L NH4Cl, 10 mmol/L KHCO3, and 0.1 mmol/L EDTA Subsequently, the cells were centrifuged (12,000 Â g,4 C, 20 min), and (pH 8.0). The cell pellet was treated with 1% NP40 diluted in PBS the pellet was resuspended in 50 mmol/L Tris-HCl (pH 8.0), 2 mmol/L containing 5 mmol/L benzamidine, 10 mmol/L EDTA, 100 mmol/L 6- EDTA, 0.35 mmol/L Lysozyme (Sigma Aldrich, Steinheim, Germany), and aminohexanoic acid, and 2 mmol/L phenylmethanesulfonylfluoride for 0.1% Triton X-100 for 15 minutes at 30jC. The cells were then cooled on 12 hours at 4jC. Following centrifugation (10,000 Â g;30min),the ice, disrupted by ultrasound, and centrifuged (12,000 Â g,4jC, 20 min). NP40-soluble supernatant was concentrated 10-fold using Centricon Aliquots of both the soluble and the insoluble protein fraction were 30 membranes (Amicon, Witten, Germany). Buffer was adjusted to a examined by SDS-PAGE and Western blot analysis. The yield of final concentration of 0.1% NP40 in PBS with proteinase inhibitor recombinant proteins per milligram wet weight of the crude bacterial concentrations as mentioned above. extract was assessed semiquantitatively in SDS-PAGE gels by comparing Human colon tissue was kept on wet ice at all times. After mechanical the intensities of stained fusion protein bands to known amounts of the disruption to pieces of f5mm3, the tissue was homogenized using an benchmark molecular weight marker proteins (Life Technologies). Ultra-Turrax homogenizer (IKA Labortechnik, Staufen, Germany). Glyco- Refolding and Purification of the Insoluble Protein Fractions. The proteins were extracted from the homogenates using 1.6 mol/L perchloric insoluble inclusion bodies were extensively purified by five cycles of acid. After centrifugation (3,000 Â g, 20 min), the supernatant was passed washing with PBS and subsequent centrifugation. The insoluble protein through gauze and the pellet was discarded. The supernatant was fraction was then denatured in a buffer containing 8 mol/L urea, neutralized at pH 7 to 8 using 1 and 5 mol/L NaOH and was subsequently 5 mmol/L reduced glutathione, 0.5 mmol/L oxidized glutathione, and dialyzed extensively against tap water and finally against distilled water. 0.05 mol/L DTT (Sigma-Aldrich, Taufkirchen, Germany) for 12 hours on The colon extract was then centrifuged (3,000 Â g, 15 min) to remove an overhead shaker. The denatured proteins were then diluted 1:10 (v/v) precipitates before concentrating it using a Minitan ultrafiltration system in 8 mol/L urea and incubated for additional 60 minutes while shaking. (Millipore, Schwalbach, Germany). Subsequently, the proteins were extensively dialyzed against PBS (pH 8.5) Construction of scFv6G6.C4 and scFv6G6.C4/GM-CSF Fusion Pro- at 4jC with frequent change of dialysis buffer and then centrifuged tein. The 6G6.C4 scFv fragment was constructed from the cloned heavy and (12,000 Â g, 20 min). The soluble renatured proteins were purified by gel light chain genes of the monoclonal anti-idiotype 6G6.C4 by PCR (16). filtration using a 120 mL HiPrep Sephacryl S-200 column (Pharmacia/ Briefly, the primers VH5.1_NdeI (29 nt) 5V-TACATATGCAGGTCCAACTA- Biotech, Freiburg, Germany) equilibrated in PBS (pH 8.5) on a fast CAGCAGCCT and VHLIN_BamHI (55 nt) 5V-ACCACTCTCACAGTCTCCT- protein liquid chromatography system (Pharmacia, Freiburg, Germany) at CATCAGGAGGAGGCTCAGGTGGTGGTGGATCCAAG were used to amplify 0.5 mL/min. The chromatography was monitored at 280 nm. Five- the heavy chain variable (VH) region gene. The primers VLLIN_BglII (53 nt) milliliter fractions were collected and analyzed for the recombinant 5V-AAAGATCTGGTGGAGGAGGATCTGGGGAGAGAGACATTGTGATGATC- proteins by SDS-PAGE.

Cancer Res 2005; 65: (5). March 1, 2005 1926 www.aacrjournals.org

SDS-PAGE and Western Blot Analysis of Recombinant Proteins. follows: Cell extracts from human normal colon tissue, human granulocytes, Proteins were separated by PAGE on 7.5% polyacrylamide gels under human HT29 colon carcinoma cells, and Jurkat wild-type cells were denaturating and reducing conditions. Gels were either developed with separated by SDS-PAGE on 7.5% gels under denaturating and reducing Silver Stain Plus (Bio-Rad Laboratories GmbH, Munich, Germany), Brilliant conditions and subsequently blotted onto nitrocellulose BA45 membranes Blue G Colloidal protein stain (Sigma-Aldrich, Taufkirchen, Germany) or (Schleicher & Schuell). After blocking with 5% nonfat dry milk/PBS, the blotted onto BA45 nitrocellulose membranes (Schleicher & Schuell, Dassel, membranes were incubated with the sera harvested from immunized mice, Germany) using a semidry blotter (Biometra, Go¨ttingen, Germany) for 60 diluted 1:100 in blocking buffer. MAbs mT84.66 (specific for CEA), T84.1 minutes at 100 V. For Western blot analysis, the membranes were blocked (anti-CEA, cross-reactive with CEACAM1 and CEACAM6), BY114 (specific with 5% nonfat dry milk/PBS and incubated with mouse herpes simplex for CEACAM6), and 29H2 (specific for CEACAM1) were used as control virus–tag monoclonal IgG1 antibody (1:5,000, EMD Biosciences, Inc., antibodies. Sera from nonimmunized CEA-transgenic mice were used as Novagen Brand) for 60 minutes at room temperature. Following three negative controls. Bound antibodies were detected with peroxidase- washes with PBS/Tween, bound HSV-tag antibody was detected using conjugated goat anti-mouse IgG (Fc-specific; diluted 1:20,000, Dianova). peroxidase-conjugated goat anti-mouse IgG (Fc-specific; 1:1,000, Dianova, Bound antibodies were visualized by chemiluminescence using the ECL+ Hamburg, Germany) and diaminobenzidine. Western Blotting Detection System (Amersham Biosciences). Functional Characterization of scFv6G6.C4/GM-CSF Fusion Protein. CEA-specific binding of Ab3 antibodies in the sera of immunized mice The binding of scFv6G6.C4 and the scFv6G6.C4/GM-CSF fusion protein to was tested by flow cytometry using CEA-expressing murine and human the T84.66 idiotype antibody was determined by sandwich ELISA. Chimeric colon carcinoma cells. In brief, 1 Â 106 cells/mL of each murine MC38 colon human/mouse T84.66 (chT84.66) was used as the catcher antibody. Ninety- carcinoma cells, the transfected MC38 subline C15-A.3, and MC38-BGP as six-well plates were coated with chT84.66 at 2 Ag/mL overnight, then well as CEACAM6-transfected human HeLa and human HT29 cells were blocked with blocking buffer containing 5% bovine serum albumin/PBS for incubated with the immune sera (1:50 diluted). MAbs mT84.66 and T84.1 2 hours. The scFv6G6.C4 and the scFv6G6.C4/GM-CSF were adjusted to were used as positive controls at 2 and 1.25 Ag/mL, respectively, for 1 hour equimolar concentrations and incubated for 1 hour at room temperature at 4jC. After washing, cells were stained with phycoerythrin-labeled goat with the chT84.66 catcher. After three washes, the plates were incubated anti-mouse IgG (DAKO, Hamburg, Germany) for 30 minutes at 4jC. with HSV-tag antibody (1:1,000, EMD Biosciences, Inc., Novagen Brand) for Preimmune sera (1:50 diluted) were used as negative controls. Fluores- 1 hour. Following washing, the immune complexes were detected with cence-activated cell sorting analyses were done using a FACS Calibur peroxidase-conjugated goat anti-mouse IgG (Fc specific; 1:1,000, Dianova) benchtop flow cytometer (Becton Dickinson, Heidelberg, Germany). and OPD. Measurements were done at 492 nm using a microplate ELISA Statistical Analysis. The statistical analyses were done with the Mann- reader (SLT, Crailsheim, Germany). Whitney U test using the WinSTAT software version 1999.2 (Robert K. To determine the biological activity of the growth factor part of the Fitch software, Staufen, Germany). Findings were regarded as significant at scFv6G6.C4/GM-CSF fusion protein, we did a cell growth assay using P < 0.05. GM-CSF-dependent FDC-P1 cells (28). Before cell culture, the soluble scFv6G6.C4/GM-CSF and scFv6G6.C4 preparations were dialyzed against tissue culture medium for 12 hours. Briefly, FDC-P1-cells were Results maintained in MEM supplemented with 10% FCS and 2% glutamine Construction of Recombinant Genes. A scFv fragment gene (Life Technologies). scFv6G6.C4/GM-CSF fusion protein, murine GM-CSF, coding for the monoclonal anti-idiotype antibody 6G6.C4 was or scFv6G6.C4 were added to the cells at different concentrations and generated by engineering an 18 amino acid linker between VH the cell culture was continued for 48 hours at 37jCin5%CO. 2 and VL gene via PCR amplification (Fig. 1A). As defined by the Thereafter, [methyl-3H]thymidine (Amersham Biosciences, Little Chalfont, Buckinghamshire, United Kingdom) was added for another 12 hours of primers VHLIN_BamHI and VLLIN_BglII, the linker sequence incubation, after which the FDC-P1-cells were harvested using a Combi- was included in the amplification of the VH and the VL region 12 Cell Harvester (Molecular Devices GmbH, Munich, Germany) and gene, respectively. Upon ligation of the PCR fragments, the transferred to a Filter (Ready Filter, Beckman Coulter GmbH, Krefeld, BamHI- and BglII-compatible ends were replaced by a BstYI site. Germany). Thymidine incorporation was determined using a liquid Restriction with BstYI of insert-positive clones was successfully scintillation h-counter (LS 1701, Beckman Coulter GmbH). All measure- used to screen for the correct assembly of the scFv fragment. ments were done in triplicates. The GM-CSF cassette was positioned amino terminally to the Immunization Studies. Nine female 6-week-old C57BL/6 mice scFv6G6.C4 gene by directional cloning using a SfiI site included transgenic for CEA were divided into three groups for immunizations into the PCR primers GM-CSFR1_SfiI and VH5.2_SfiI (Fig. 1B). (29). As adjuvant, QS21 was used (kindly provided by Antigenics, Inc., New Final DNA sequencing of the construct confirmed the sequences York, NY; ref. 30). All animals were immunized s.c. four times at 3-week intervals. Animals in group 1 received 5 Ag QS21 adjuvants only; group 2 published for mAb 6G6.C4 (16) and also for the GM-CSF gene animals received 5 Ag QS21 together with 13 Ag scFv6G6.C4; and animals cassette as provided by the vendor (data not shown). The in group 3 received 5 Ag QS21 together with 24 Ag scFv6G6.C4/GM-CSF. expression of the recombinant proteins in E. coli BL21DE3 Sera were collected 1 week after the final immunization and investigated resulted in polypeptides of predicted molecular masses of 47 and for specific Ab3 antibody responses against 6G6.C4 and CEA using ELISA, 26 kDa for scFv6G6.C4/GM-CSF and scFv6G6.C4, respectively Western blot analysis, and flow cytometry. (Fig. 2A). Up to 15 Ag recombinant protein per milligram wet Detection of CEA- and scFv6G6.C4-Specific Antibody (Ab3) weight was obtained as judged from comparing known amounts Responses. Briefly, 96-well microtiter plates CovaLink (Nalge Nunc of marker proteins with serial dilutions of bacterial extracts on International, Wiesbaden, Germany) were coated with 200 ng/mL SDS-PAGE gels after Coomassie Blue staining. The presence of scFv6G6.C4 or 2 Ag/mL CEA for 2 hours. After washing and blocking with the HSV-tag was shown in the recombinant proteins by the HSV- 5% bovine serum albumin/PBS, the plates were incubated with the sera of the immunized mice previously diluted 1:20 to 1:640 (v/v) for 60 minutes. tag antibody in Western blot analyses (data not shown) and by Following three washing steps with PBS/Tween, the bound murine ELISA (see Materials and Methods). immunoglobulins were incubated with peroxidase-conjugated goat anti- Preparation of Protein Fractions. The recombinant proteins mouse IgG (Fc-specific, 1:1,000, Dianova). After three washing steps with were found as insoluble bacterial inclusion bodies and were PBS/Tween, the assays were developed with OPD measured at 492 nm (SLT). processed as described (31). The repeated washing/centrifugation To evaluate the cross-reactivity of the Ab3 response with CEA-like steps of the inclusion bodies already resulted in a substantial proteins of the CEACAM family, Western blot analyses were carried out as purification. After denaturation under reducing conditions, refolding www.aacrjournals.org 1927 Cancer Res 2005; 65: (5). March 1, 2005

130 ng/mL for scFv6G6.C4 and 230 ng/mL for scFv6G6.C4/GM- CSF, the average binding seemed slightly higher for scFv6G6.C4/ GM-CSF (0.86 F 0.09) than for scFv6G6.C4 (0.81 F 0.09), but not statistically significant (P = 0.11). scFv6G6.C4/GM-CSF Exhibits GM-CSF Activity. We tested the GM-CSF bioactivity of the 6G6.C4/GM-CSF fusion protein using GM-CSF- and IL3-dependent FDC-P1-cells compared with murine GM-CSF used as a positive control. As shown in Fig. 3B, the fusion protein sustained proliferation of FDC-P1-cells in a dose-dependent manner. The half-maximal proliferation was achieved at a concentration of 2.6 ng/mL scFv6G6.C4/GM-CSF. Murine GM-CSF used in identical protein concentration as for the fusion protein scFv6G6.C4/GM-CSF sustained a comparable proliferation of the FDC-P1-cells. As expected, scFv6G6.C4 alone had no effect in proliferation. This experiment clearly shows that the GM-CSF fusion partner of scFv6G6.C4/GM-CSF has retained growth factor properties to stimulate cell growth of factor- dependent cells. Figure 1. Cloning strategy for the construction of recombinant scFv6G6.C4 (A) Immunologic Responses after Immunization with the and the fusion protein scFv6G6.C4/GM-CSF (B). A, the VH region and the VL Recombinant Proteins scFv6G6.C4/GM-CSF and scFv6G6.C4. region were amplified using primer pairs 1 & 2 and 3 & 4, respectively. Ligation of the products eliminated the BamHI and BglII sites generating the We analyzed the Ab3 response in C57BL/6 mice transgenic for scFv fragment with a diagnostic BstYI restriction site. B, after reamplification the human CEA after repeated immunizations with the adjuvant of scFv6G6.C4, GM-CSF cassettes were assembled using the SfiI restriction included in primers 5 and 7. QS21 (control group), the recombinant scFv6G6.C4, or the scFv6G6.C4/GM-CSF fusion protein. For this study, three animals in each group were immunized using equimolar amounts of the step, and fast protein liquid chromatography, the purity and stability recombinant proteins. Importantly, none of the animals showed of the recombinant proteins was examined in a 7.5% SDS signs of illness or weight loss as indicators of autoimmune polyacrylamide gel. The purity of the recombinant proteins was reactions (not shown). Figure 4A shows the response directed judged to be higher than 90% by silver staining (Fig. 2B). No against the idiotopes of the scFv fragment (i.e., the anti-6G6.C4 degradation was observed after 14 days of storage of 4jC. response). A significant difference in immunogenicity can be Functional Properties of scFv6G6.C4/GM-CSF and observed between the single-chain fragment alone and the single scFv6G6.C4. As shown in Fig. 3A, the immunologic capability to chain as molecular part of the GM-CSF growth factor–bearing bind to its idiotype mAb chT84.66 was shown for both scFv6G6.C4 fusion protein. Indeed, dilutions of the hyperimmune sera and for scFv6G6.C4/GM-CSF. Under equimolar conditions of against the scFv alone did not result in a regular ELISA

Figure 2. SDS-PAGE analyses of recombinant proteins. A, unpurified protein fractions achieved after culture of transfected BL21DE3 E. coli cells separated in a 7.5% SDS polyacrylamide gel and stained with Brilliant Blue G Colloidal protein stain. Lane 1, reference standard; lane 2, insoluble protein fraction containing scFv6G6.C4/GM-CSF with an apparent mass of f47 kDa; lanes 3 and 5, insoluble protein fraction containing scFv6G6.C4 with an apparent mass of f26 kDa; lane 4, soluble protein fraction containing scFv6G6.C4. B, purified protein fractions of scFv6G6.C4/GM-CSF and scFv6G6.C4 were electrophoretically separated in a 7.5% polyacrylamide gel and silver stained. Lane 1, reference standard; lane 2, 1.5 AL soluble scFv6G6.C4/GM-CSF; lane 3, 1.5 AL soluble scFv6G6.C4.

Cancer Res 2005; 65: (5). March 1, 2005 1928 www.aacrjournals.org

(anti-scFv6G6.C4 response: P = 0.005; anti-CEA response: P = 0.012) or the control group (anti-scFv6G6.C4 and anti-CEA response: P = 0.002). Ab3 titers induced by scFv6G6.C4 were also significantly different compared with the control group (anti- scFv6G6.C4 response: P = 0.012; anti-CEA response: P = 0.002). Figure 4C shows the ratios of the Ab3 titers against 6G6.C4 and CEA in animals immunized with either the single-chain fragment or the fusion protein. Up to a dilution of 1:160, the strength of the anti-CEA response (white columns)iscomparable with the anti-scFv6G6.C4 response (striped columns) with a

drop in A492 nm readings to 50% at 1:640 dilution. However, over all dilutions of mice sera, the difference was not significant (P = 0.212). These results clearly show that scFv6G6.C4 functions as an internal image of CEA as has been described for the full murine monoclonal 6G6.C4 (15). Also, these results show that a sufficient anti-CEA response in an immunotolerant (i.e., CEA- expressing host can be achieved using a GM-CSF fusion to anti- idiotypic antibodies). We tested the specificity of the scFv6G6.C4/ GM-CSF-induced anti-CEA (Ab3) response using Western blot analyses of tissues expressing CEA-related antigens. As shown in Fig. 5A, CEA is absent in granulocyte extracts as shown by lack of reactivity with the CEA-specific mAb chT84.66, whereas the CEA-related antigens CEACAM1 and CEACAM6 are readily detected by specific mAbs 29H2 and BY114, respectively. In colon extracts and human HT29 colon cancer cells, CEA and Figure 3. Functional characterization of scFv6G6.C4/GM-CSF fusion protein. A, antigen-binding assays using a 96-well sandwich ELISA system. CEACAM1 can be shown by mAbs chT84.66 and 29H2, White columns, A 492 values using scFv6G6.C4/GM-CSF as primary antibody respectively. The slight differences in molecular mass of (1. mAb); striped columns, A 492 values using scFv6G6.C4 as primary molecules of the CEACAM family observed with both antibodies antibody. Shown are the binding capacities of scFv6G6.C4/GM-CSF and scFv6G6.C4 compared with the following negative controls: measurements are due to known glycosylation differences between materials without using primary mAb, measurements without using detection antibody obtained from tissue versus from cell culture. (2. mAb, HSV-tag), and measurements without T84.66 coating. Without using Hyperimmune sera from transgenic mice immunized with peroxidase-conjugated goat anti-mouse IgG, no A 492 was measured for scFv6G6.C4/GM-CSF and scFv6G6.C4 (0.00 F 0.00, data not shown). Bars, the scFv6G6.C4/GM-CSF fusion protein specifically bound to SD. Results are from nine duplicates (scFv6G6.C4-GM-CSF fusion protein and 180 kDa CEA antigen extracted from HT29 human colon scFv6G6.C4) and three duplicates (all negative controls). B, GM-CSF bioactivity of scFv6G6.C4/GM-CSF fusion protein. GM-CSF- and IL3-dependent cancer cells (Fig. 5B) but did not recognize CEA-related FDC-P1 cells were incubated for 48 hours in the presence of increasing amounts antigens present in the granulocyte extracts (32) or HT29 cells (0.026-2600 ng/mL, X-axis) of the scFv6G6.C4/GM-CSF fusion protein (n), (compare Fig. 5A). Extracts from Jurkat cells were not murine GM-CSF as positive control (5), and scFv6G6.C4 as negative control (E). Proliferation of the cells was determined by [methyl-3H]thymidine recognized by the hyperimmune sera. Expectedly, preimmune incorporation assay as described under Materials and Methods. sera of C57BL/6 transgenic mice or C57BL/6 wild-type mice Y-axis, proliferation of FDC-P1 cells in counts per minute (cpm). Points, mean of triplicate measurements; bars, SD. Proliferation of FDC-P1-cells was showed no cross-reactivity with CEA-related antigens. Taken sustained by scFv6G6.C4/GM-CSF and mGM-CSF, not by scFv6G6.C4. together, these results clearly show that the Ab3 response after vaccination with the scFv6G6.C4/GM-CSF fusion protein is highly specific and capable to distinguish between CEA and titration curve, indicating that very little antibody is being the highly homologous antigens. produced against 6G6.C4. In contrast, a pronounced anti-6G6.C4 The ability of Ab3 antibodies from scFv6G6.C4/GM-CSF (Ab3) response was readily detectable using dilutions of the vaccinated CEA-transgenic mice to detect CEA-expressing cells immune sera of well below 1:640 (v/v). Immunization with QS21 in whole blood was analyzed by flow cytometry. The results in Fig. resulted in no detectable anti-6G6.C4 response (Fig. 4A). 6A show specific binding to CEA in the CEA-positive cell lines Sera collected after the third boost at the end of week 10 were C15-A.3 and HT29 compared with the preimmune sera that analyzed for the presence of anti-CEA-specific Ab3 response showed no shift in immunofluorescence. Figure 6B shows the (Fig. 4B). No immune response was noted in mice immunized results for the IgG1 isotype–matched controls to the CEA-mAb with QS21 alone, whereas a very limited CEA-specific binding T84.1 (1.25 Ag/mL) that is known to cross-react with CEA-related was observed in animals immunized with scFv6G6.C4 antibody antigens or the CEA-specific mAb mT84.66 (2 Ag/mL). No binding alone. Again, the titers were very low and seemed to resemble to CEA-negative cell lines was observed in the preimmune sera or unspecific interaction below dilutions of 1:40. In contrast, when the hyperimmune sera. scFv6G6.C4/GM-CSF was used under identical conditions and equimolar amounts in the transgenic animals, anti-CEA (Ab3) Discussion titers of well below 1:640 were easily detectable. Anti-scFv6G6.C4 The use of internal image anti-idiotypic antibody vaccines or anti-CEA responses were significantly higher in immune mimicking TAAs aims to elicit effective immune responses that sera raised against the fusion protein compared with the will be capable to eradicate tumor cells and prevent tumor immunizations with equimolar amounts of scFv6G6.C4 alone recurrence. However, TAAs like the human CEA are self-antigens www.aacrjournals.org 1929 Cancer Res 2005; 65: (5). March 1, 2005

against which clinically relevant humoral or cellular immune responses are not easily achieved. Specifically, clinical studies with anti-idiotype antibodies mimicking CEA have shown Ab3 immune responses (11, 33–35) but the benefit for the patients thus far has been limited. Similarly, we have observed only marginal CEA-specific responses following vaccination with the full murine anti-idiotypic mAb 6G6.C4 in patients with R0- resected CEA-positive pancreatic cancers.4 Because previous studies had shown that 6G6.C4 effectively functions as a surrogate antigen for CEA in mice, rats, and rabbits (15), we have attributed this failure in humans to an insufficient breaking of tolerance toward CEA. Different immunostimulatory cytokines, growth factors, or co-stimulatory antigens have been used either after fusion or by co-administration with antigens or antibodies to improve immune responses (17, 18, 21, 26, 36, 37). For example, most recently, Reinartz et al. (13) have reported on the superiority of a fusion between IL6 and the anti-idiotype antibody (ACA125) mimicking the CA125 TAA compared with the co-administration of the respective single components. However, because this study was conducted in a nontolerant animal model, the immunogenicity of the IL-6 fusion antigen is unknown in the immunotolerant situation. To address this question, we have previously generated a CEA-transgenic mouse expressing CEA as a self-antigen in a tissue-specific manner comparable with humans (29) and have used this system to investigate a recombinant fusion vaccine that consists of the anti-idiotype and murine GM-CSF. We have chosen GM-CSF as a fusion partner, because it effectively stimulates professional antigen-presenting cells including dendritic cells. We speculated that the fusion protein of the 6G6.C4 anti-idiotype with GM-CSF would effectively ‘‘address’’ to the GM-CSF receptor, thereby enhancing the contact of the internal image of the CEA-specific T84.66 epitope to antigen-presenting cells for an improved antigen presentation. Indeed, the protein efficiently stimulates growth of the follicular dendritic cell line FDC-P1 (Fig. 3B) and achieves an effective anti-CEA response not seen with the anti- idiotype 6G6.C4 alone (Fig. 4B). Saha et al. (38) have shown that the pulsing of dendritic cells with an internal image CEA anti- idiotype can result in breaking of humoral and cellular tolerance in CEA-transgenic mice and protect immunized animals from a lethal tumor challenge. These data emphasize the importance of dendritic cells for active specific immunotherapy and confirm our assumption regarding the potential of fusion proteins consisting of an anti-idiotypic antibody and GM-CSF. Work is in progress in the laboratory to show whether the convenient administration of an anti-idiotype fusion vaccine will be an attractive alternative to the ex vivo pulsing of dendritic cells. Taken together, the results presented here are, to the best of our knowledge, the first that show the efficiency of an anti- idiotype/GM-CSF fusion vaccine against a human TAA and the effect on humoral immune response in an immunotolerant Figure 4. Antibody titers after vaccination of CEA-transgenic C57BL/6 mice. A and B, Mice were immunized (three animals per group) by s.c. injection of 24 in vivo system. Ag scFv6G6.C4/GM-CSF (n)or13Ag scFv6G6.C4 (E), which is the equimolar It is known that co-administration of GM-CSF with antibodies amount. All groups received 5 Ag QS21 (one control group 5). After the third against TAAs leads to a strong increase of human anti-mouse boost, the serum titers of anti-scFv6G6.C4 and anti-CEA antibodies were determined by the CovaLink ELISA system coated with scFv6G6.C4 antibody (39). To avoid such unwanted immunoreactivity in the and CEA. A, anti-scFv6G6.C4 Ab3 response. B, anti-CEA Ab3 response. GM-CSF fusion protein, we have converted the full mAb to the Points, mean titers of three identically treated animals measured in triplicates; scFv fragment to yield the scFv6G6.C4/GM-CSF vaccine. Recently, bars, SD. C, comparison of immune responses of mice immunized with the scFv6G6.C4 or the fusion protein. Normalized ratios of the Ab3 titers of scFv have been successfully used to generate specific Ab3 animals immunized with the fusion protein and mice immunized with scFv6G6.C4 were built. White columns, anti-scFv6G6.C4 titers; striped columns, anti-CEA titers. A column indicates the ratio of three identically immunized animals each. 4 Unpublished data.

Cancer Res 2005; 65: (5). March 1, 2005 1930 www.aacrjournals.org

Figure 5. Ab3 antibodies from sera of CEA-transgenic C57BL/6 mice immunized with scFv6G6.C4/GM-CSF bind to CEA antigen. A, detection of CEA-related molecules in various human tissues by mAbs by Western blot. Lanes 1-3, crude extract of human granulocytes. Lanes 4 and 6, protein extracts from normal human colon mucosa. Lanes 5 and 7, crude extract from HT29 colon cancer cells. Samples were separated by 7.5% SDS-PAGE under denaturating and reducing conditions and blotted onto nitrocellulose membranes. Antibodies used were mAb mT84.66 (T), mAb BY114 (B), and mAb 29H2 (H). B, specificity of Ab3 antibodies from sera of CEA-transgenic mice before and after vaccination with scFv6G6.C4/GM-CSF fusion protein. Crude extracts of HT29 cells (HT), Jurkat wild-type cells (J), and human granulocytes (G) were separated by 7.5% SDS-PAGE under denaturating and reducing conditions and blotted onto nitrocellulose membranes. All antisera were diluted 1:100 (v/v) before probing of the blots. Lanes 1-3, detection with sera from CEA-transgenic animals vaccinated with scFv6G6.C4/GM-CSF; lanes 4 and 5, pools of preimmune sera of CEA-transgenic C57BL/6 (n = 7) and wild-type mice (n = 3); lanes 6 and 7, preimmune sera of a CEA-transgenic and a wild-type animal.

antibody (7, 18, 40). The binding to the T84.66 idiotype in vitro anti-CEA antibody is somewhat weaker than the anti-6G6.C4 and the capability to elicit a CEA-specific response shows that response (Fig. 4C). This is not surprising as the anti-CEA the fusion construct retains its image-bearing immunologic antibodies are generated against the confined paratope features. Compared to the scFv6G6.C4 alone, the scFv6G6.C4/ structure, which represents only a part of the total number GM-CSF greatly improved the immune response in general of 6G6.C4 idiotopes. Also, the paratope may be less through its contribution from the GM-CSF fusion partner. immunogenic. Taken together, these results clearly show that Specifically, the internal image anti-idiotypic scFv6G6.C4 fails breaking of tolerance against endogenous CEA can be achieved to yield significant anti-CEA titers in the CEA-tolerant transgenic in transgenic mice and results in a strong humoral response, mice (Fig. 4A and B). While Ab3 antibodies are readily detected when the internal image scFv6G6.C4 is coupled to a potent against 6G6.C4 idiotopes (Fig. 4A), a very weak anti-CEA immunostimulatory molecule. antibody response can only be shown in dilutions up to 1:40 Other studies have investigated a different approach to (v/v; Fig. 4B). In contrast, we found that equimolar amounts of overcome the humoral and cellular tolerance toward CEA in the the fusion protein scFv6G6.C4/GM-CSF used under identical same CEA-transgenic mouse model by using vaccines comprising experimental conditions improved the antibody response in full CEA in recombinant viral vectors with or without additional general against both the 6G6.C4 idiotopes and the CEA internal stimulatory cytokines (41–44). These reports have shown that image. Indeed, anti-CEA antibodies were detectable in dilutions protection against tumor challenges can be achieved in the of 1:640 (v/v; Fig. 4B). immunized animals. However, it remains to be seen if full antigen The reactivities against the 6G6.C4 idiotopes and the vaccines capable of breaking tolerance lead to anti-CEA anti- internal image of CEA were analyzed as the normalized ratios bodies that will cross-react with other members of the CEACAM of the antisera against scFv6G6.C4 or scFv6G6.C4/GM-CSF. family of glycoproteins and are significant for autoimmunity in These ratios show that the two vaccines generate similar humans. Indeed, such cross-reactive immune response has antibody responses up to dilutions of 1:160 beyond which the resulted from the use of a CEA-expressing cell vaccine in CEA www.aacrjournals.org 1931 Cancer Res 2005; 65: (5). March 1, 2005

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2005 American Association for Cancer Research. Cancer Research transgenes (45). Interestingly, breaking of immunotolerance therefore provide a considerable advantage over full antigen against CEA does not lead to autoimmune colitis in the vaccines in complex antigen systems like the human CEACAMs transgenes (21, 43–45). In contrast, it is possible that cross- featuring a greater number of highly homologous and cross- reactive immune responses may lead to an autoimmune reaction reacting epitopes. As shown here and by the work of Saha in human tissues that physiologically express CEA-related et al. (38), the efficient addressing of the antigen to antigen- antigens (e.g., against granulocytes) because these do express presenting cells is sufficient to overcome immunotolerance large amounts of CEACAMs 1, 6, and 7. We feel strongly that the and will stimulate the desired immune response. Thus, the question of autoimmunity is particularly important for the full use of strong xenogenic immunostimulants may not be antigen vaccines but cannot be answered appropriately in the necessary. current transgenic model systems. Unfortunately, no transgenic In addition, the highly specific mimicry of a CEA epitope has animals exist that harbors the complete human chromosome 19 resulted in a specific immune response (i.e., no binding of Ab3 CEA gene cluster. antibodies was noted against the major CEA-like antigens that In contrast, anti-idiotype vaccination strategies can be directed are expressed on the surface of granulocytes and in the human against a single non-cross-reacting TAA epitope and may colon cancer cells HT29; Fig. 5B). As suggested by flow

Figure 6. Ab3 antibodies from sera of CEA-transgenic C57BL/6 mice immunized with scFv6G6.C4/GM-CSF bind to CEA+ C15-A.3 and HT29 cells. A, CEAÀ BGPÀ NCAÀ MC38 cells, CEA+ C15-A.3 cells, BGP+ MC38-BGP cells, NCA+ HeLa-NCA cells, and CEA+ BGP+ NCA+ HT29 cells were incubated with 1:50 diluted preimmune sera and immune sera of C57BL/6 Han TgN (CEAgen) HvdP mice. B, all cells, excluding C15-A.3, were incubated with IgG1 isotype control and mAbT84.1 (1.25 Ag/mL). C15-A.3 cells were incubated with IgG1 isotype control and mT84.66 (2 Ag/mL). Bound Ab3 antibodies were detected with phycoerythrin-labeled goat anti-mouse IgG. The cells were analyzed by flow cytometry and the data presented in histograms were overlaid. Conjugate controls of each cell line (data not shown) showed no background staining. Binding of immune sera (A), mAb T84.1, and mT84.66 (B) is indicated by solid lines; binding of preimmune sera (A) and isotype control IgG1 (B) is shown by dotted lines.

Cancer Res 2005; 65: (5). March 1, 2005 1932 www.aacrjournals.org

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2005 American Association for Cancer Research. Anti-CEA Response in CEA-Transgenic Mice cytometry, the humoral immunity conferred by scFv6G6.C4/GM- Acknowledgments CSF is able to recognize, and may presumably eliminate, Received 10/6/2004; revised 12/13/2004; accepted 12/23/2004. circulating tumor cells in cancer-bearing animals. This fusion Grant support: Deutsche Krebshilfe grant to MN W56/94/NE2. protein may therefore be a valuable vaccine for an active specific 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 immunotherapy of CEA-positive human cancers. Further studies with 18 U.S.C. Section 1734 solely to indicate this fact. are now under way to investigate the tumor protective We thank John Shively for kindly providing purified CEA and transfected cell lines, Dr. Weiss for statistical advice, and Kerstin Reher and Raika Gimmini for their properties of scFv6G6.C4/GM-CSF. excellent technical skills.

References human carcinoembryonic antigen. J Biol Chem carcinoembryonic antigen (CEA) gene family. J Immu- 1993;268:14138–45. nol 1993;150:4978–84. 1. Hammarstrom S, Baranov V. Is there a role for CEA in 17. Penichet ML, Morrison SL. Antibody-cytokine fusion 33. Birebent B, Mitchell E, Akis N, et al. Monoclonal anti- innate immunity in the colon? Trends Microbiol proteins for the therapy of cancer. J Immunol Methods idiotypic antibody mimicking the gastrointestinal 2001;9:119–25. 2001;248:91–101. carcinoma-associated epitope CO17-1A elicits antigen- 2. Abe H, Kuroki M, Tachibana K, et al. Targeted 18. Reinartz S, Hombach A, Kohler S, et al. Interleukin-6 specific humoral and cellular immune responses in sonodynamic therapy of cancer using a photosensitizer fused to an anti-idiotype antibody in a vaccine colorectal cancer patients. Vaccine 2003;21:1601–12. conjugated with antibody against carcinoembryonic increases the specific humoral immune response 34. Gruber R, van Haarlem LJ, Warnaar SO, et al. The antigen. Anticancer Res 2002;22:1575–80. against CA125+ (MUC-16) ovarian cancer. Cancer Res human antimouse immunoglobulin response and the 3. Behr TM, Liersch T, Greiner-Bechert L, et al. Radio- 2003;63:3234–40. anti-idiotypic network have no influence on clinical immunotherapy of small-volume disease of metastatic 19. Tao MH, Levy R. Idiotype/granulocyte-macrophage outcome in patients with minimal residual colorectal colorectal cancer. Cancer 2002;94:1373–81. colony-stimulating factor fusion protein as a vaccine for cancer treated with monoclonal antibody CO17-1A. 4. Baral R, Sherrat A, Das R, et al. Murine monoclonal B-cell lymphoma. Nature 1993;362:755–8. Cancer Res 2000;60:1921–6. anti-idiotypic antibody as a surrogate antigen for 20. Tripathi PK, Qin H, Bhattacharya-Chatterjee M, et al. 35. Staib L, Birebent B, Somasundaram R, et al. Immuno- human Her-2/neu. Int J Cancer 2001;92:88–95. Construction and characterization of a chimeric fusion genicity of recombinant GA733-2E antigen (CO17-1A, 5. Schlebusch H, Wagner U, Grunn U, et al. A monoclonal protein consisting of an anti-idiotype antibody mim- EGP, KS1-4, KSA, Ep-CAM) in gastro-intestinal carcino- antiidiotypic antibody ACA 125 mimicking the tumor- icking a breast cancer-associated antigen and the ma patients. Int J Cancer 2001;92:79–87. associated antigen CA 125 for immunotherapy of cytokine GM-CSF. Hybridoma 1999;18:193–202. 36. Hoffmann P, Mueller N, Shively JE, et al. Fusion ovarian cancer. Hybridoma 1995;14:167–74. 21. Xu X, Clarke P, Szalai G, et al. Targeting and therapy proteins of B7.1 and a carcinoembryonic antigen (CEA)- 6. Spendlove L, Li L, Potter V, et al. A therapeutic human of carcinoembryonic antigen-expressing tumors in specific antibody fragment opsonize CEA-expressing anti-idiotypic antibody mimics CD55 in three distinct transgenic mice with an antibody-interleukin 2 fusion tumor cells and coactivate T-cell immunity. Int J Cancer regions. Eur J Immunol 2000;30:2944–53. protein. Cancer Res 2000;60:4475–84. 2001;92:725–32. 7. Tripathi PK, Qin H, Deng S, et al. Antigen mimicry by 22. Fischer HG, Frosch S, Reske K, et al. Granulocyte- 37. Holliger P, Manzke O, Span M, et al. Carcinoem- an anti-idiotypic antibody single chain variable frag- macrophage colony-stimulating factor activates macro- bryonic antigen (CEA)-specific T-cell activation in colon ment. Mol Immunol 1998;35:853–63. phages derived from bone marrow cultures to synthesis carcinoma induced by anti-CD3 Â anti-CEA bispecific 8. Uemura H, Beniers AJ, Okajima E, et al. Vaccination of MHC class II molecules and to augmented antigen diabodies and B7 Â anti-CEA bispecific fusion proteins. with anti-idiotype antibodies mimicking a renal cell presentation function. J Immunol 1988;141:3882–8. Cancer Res 1999;59:2909–16. carcinoma-associated antigen induces tumor immunity. 23. Coleman DL, Chodakewitz JA, Bartiss AH, et al. 38. Saha A, Chatterjee SK, Foon KA, et al. Dendritic cells Int J Cancer 1994;58:555–61. Granulocyte-macrophage colony-stimulating factor pulsed with an anti-idiotype antibody mimicking 9. Alfonso M, Diaz A, Hernandez AM, et al. An anti- enhances selective effector functions of tissue-derived carcinoembryonic antigen (CEA) can reverse immuno- idiotype vaccine elicits a specific response to N-glycolyl macrophages. Blood 1988;72:573–8. logical tolerance to CEA and induce antitumor sialic acid residues of glycoconjugates in melanoma 24. Inaba K, Inaba M, Romani N, et al. Generation of immunity in CEA transgenic mice. Cancer Res 2004; patients. J Immunol 2002;168:2523–9. large numbers of dendritic cells from mouse bone 64:4995–5003. 10. Foon KA, Oseroff AR, Vaickus L, et al. Immune marrow cultures supplemented with granulocyte/mac- 39. Hjelm Skog A, Ragnhammar P, Fagerberg J, et al. responses in patients with T-cell lymphoma treated rophage colony-stimulating factor. J Exp Med 1992; Clinical effects of monoclonal antibody 17-1A combined with an anti-idiotype antibody mimicking a highly 176:1693–702. with granulocyte/macrophage-colony-stimulating fac- restricted T-cell antigen. Clin Cancer Res 1995;1: 25. Steinman RM, Witmer-Pack M, Inaba K. Dendritic tor and interleukin-2 for treatment of patients with 1285–94. cells: antigen presentation, accessory function and advanced colorectal carcinoma. Cancer Immunol 11. Foon KA, John WJ, Chakraborty M, et al. Clinical and clinical relevance. Adv Exp Med Biol 1993;329:1–9. Immunother 1999;48:463–70. immune responses in resected colon cancer patients 26. Batova A, Kamps A, Gillies SD, et al. The Ch14.18- 40. Coelho M, Gauthier P, Pugniere M, et al. Isolation treated with anti-idiotype monoclonal antibody vaccine GM-CSF fusion protein is effective at mediating and characterisation of a human anti-idiotypic scFv that mimics the carcinoembryonic antigen. J Clin Oncol antibody-dependent cellular cytotoxicity and comple- used as a surrogate tumour antigen to elicit an anti- 1999;17:2889–5. ment-dependent cytotoxicity in vitro. Clin Cancer Res HER-2/neu humoral response in mice. Br J Cancer 12. Foon KA, Lutzky J, Baral RN, et al. Clinical and 1999;5:4259–63. 2004;90:2032–41. immune responses in advanced melanoma patients 27. Neumaier M, Shively L, Chen FS, et al. Cloning of the 41. Kass E, Schlom J, Thompson J, et al. Induction of immunized with an anti-idiotype antibody mimicking genes for T84.66, an antibody that has a high specificity protective host immunity to carcinoembryonic antigen disialoganglioside GD2. J Clin Oncol 2000;18:376–84. and affinity for carcinoembryonic antigen, and expres- (CEA), a self-antigen in CEA transgenic mice, by 13. Reinartz S, Kohler S, Schlebusch H, et al. Vaccina- sion of chimeric human/mouse T84.66 genes in immunizing with a recombinant vaccinia-CEA virus. tion of patients with advanced ovarian carcinoma myeloma and Chinese hamster ovary cells. Cancer Res Cancer Res 1999;59:676–83. with the anti-idiotype ACA125: immunological res- 1990;50:2128–34. 42. Kass E, Panicali DL, Mazzara G, et al. Granu- ponse and survival (phase Ib/II). Clin Cancer Res 28. Dexter TM, Garland J, Scott D, et al. Growth of factor- locyte/macrophage-colony stimulating factor pro- 2004;10:1580–7. dependent hemopoietic precursor cell lines. J Exp Med duced by recombinant avian poxviruses enriches 14. Wagner U, Kohler S, Reinartz S, et al. Immunological 1980;152:1036–47. the regional lymph nodes with antigen-presenting consolidation of ovarian carcinoma recurrences with 29. Eades-Perner AM, van der Putten H, Hirth A, et al. cells and acts as an immunoadjuvant. Cancer Res monoclonal anti-idiotype antibody ACA125: immune Mice transgenic for the human carcinoembryonic 2001;61:206–14. responses and survival in palliative treatment. See The antigen gene maintain its spatiotemporal expression 43. Hodge JW, Grosenbach DW, Aarts WM, et al. Vaccine biology behind: K. A. Foon and M. Bhattacharya- pattern. Cancer Res 1994;54:4169–76. therapy of established tumors in the absence of Chatterjee, Are solid tumor anti-idiotype vaccines ready 30. Kensil CR, Patel U, Lennick M, et al. Separation and autoimmunity. Clin Cancer Res 2003;9:1837–49. for prime time? Clin Cancer Res 7:1112–1115, 2001. Clin characterization of saponins with adjuvant activity 44. Zeytin HE, Patel AC, Rogers CJ, et al. Combination of Cancer Res 2001;7:1154–62. from Quillaja saponaria Molina cortex. J Immunol a poxvirus-based vaccine with a cyclooxygenase-2 15. Gaida F-J, Fenger U, Wagener C, et al. A monoclonal 1991;146:431–7. inhibitor (celecoxib) elicits antitumor immunity and anti-idiotypic antibody bearing the image of an epitope 31. Rudolph R, Lilie H. In vitro folding of inclusion body long-term survival in CEA.Tg/MIN mice. Cancer Res specific to the human carcinoembryonic antigen. Int J proteins. FASEB J 1996;10:49–56. 2004;64:3668–78. Cancer 1992;51:459–65. 32. Stoffel T, Neumaier M, Gaida F-J, et al. Monoclonal, 45. Mizobata S, Tompkins K, Simpson JF, et al. Induction 16. Gaida FJ, Pieper D, Roder UW, et al. Molecular anti-domain and anti-peptides antibodies assign the M r of cytotoxic T cells and their antitumor activity in mice characterization of a cloned idiotypic cascade contain- 160,000 antigen of the CD66 cluster to a mRNA encoded transgenic for carcinoembryonic antigen. Cancer ing a network antigenic determinant specific for the by the biliary glycoprotein (BGP) gene, a member of the Immunol Immunother 2000;49:285–95.

www.aacrjournals.org 1933 Cancer Res 2005; 65: (5). March 1, 2005

Downloaded from cancerres.aacrjournals.org on September 27, 2021. © 2005 American Association for Cancer Research. Monoclonal Anti-idiotype Antibody 6G6.C4 Fused to GM-CSF Is Capable of Breaking Tolerance to Carcinoembryonic A ntigen (CEA) in CEA−Transgenic Mice

Christian Schwegler, Alexandra Dorn-Beineke, Stefanie Nittka, et al.

Cancer Res 2005;65:1925-1933.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/65/5/1925

Cited articles This article cites 45 articles, 25 of which you can access for free at: http://cancerres.aacrjournals.org/content/65/5/1925.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/65/5/1925.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/65/5/1925. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.