CD28 Costimulation Is Required for Interleukin 2 Secretion and Receptor-Mediated TC

[CANCER RESEARCH 61, 1976–1982, March 1, 2001] T-Cell Activation by Recombinant Receptors: CD28 Costimulation Is Required for Interleukin 2 Secretion and Receptor-mediated T-Cell Proliferation but Does Not Affect Receptor-mediated Target Cell Lysis1

Andreas Hombach, Dagmar Sent, Claudia Schneider, Claudia Heuser, Dimitra Koch, Christoph Pohl, Barbara Seliger, and Hinrich Abken2 Klinik I fu¨r Innere Medizin, Labor Tumorgenetik, Universita¨t zu Ko¨ln, D-50924 Ko¨ln [A. H., D. S., C. S., C. H., D. K., H. A.]; St. Elisabeth-Krankenhaus Ko¨ln-Hohenlind, 50935 Ko¨ln [C. P.]; and Johannes Gutenberg-Universitat, III. Medizinische Klinik, 55101 Mainz [B. S.], Germany

ABSTRACT receptor molecule of choice. Retroviral transduction depends on proliferating cells but is highly inefficient for resting cells. According to Recombinant T-cell receptors with antibody-like specificity are success- the two-signal paradigm, resting T cells can be either preactivated by fully used to direct CTLs toward a MHC-independent immune response incubation with an anti-CD3 mab in addition to exogenous IL-2 or by against target cells. Here we monitored the specific activation of receptor grafted CTLs in the context of CD28 costimulation. Peripheral blood T anti-CD28-specific mabs (6) that mimick signaling mediated by re- cells were retrovirally engrafted with recombinant anti-CD30 and anti- ceptor molecules. CD28 costimulation, in addition to signaling carcinoembryonic antigen receptors, respectively, that harbor either the through the TCR/CD3 complex, is one of the key events required for Fc⑀RI-␥ or the CD3-␨ intracellular signaling domain. Cross-linking of the complete activation of resting T cells, resulting in cellular prolif- recombinant receptors by solid-phase bound ligand, i.e., CD30 and a eration, cytokine secretion, CTL-mediated target cell lysis, and pre- carcinoembryonic antigen receptor-specific anti-idiotypic antibody, re- vention of activation-induced anergy (reviewed in Ref. 7). Studies of spectively, induces IFN-␥ secretion that is further enhanced by CD28 CD28 deficient animals suggest that CD28 preferentially serves to costimulation of grafted T cells. Induction of interleukin (IL)-2 secretion, amplify and sustain a primary T-cell response (8) and to lower the in contrast, requires CD28 costimulation in addition to receptor cross- amount of antigen required to achieve full cellular activation (9). linking, irrespective of T-cell preactivation by anti-CD3 monoclonal an- Resting T cells, however, can be alternatively activated via B7- tibody plus IL-2 or by anti-CD3 monoclonal antibody plus anti-CD28 monoclonal antibody. Accordingly, induction of IL-2 secretion upon re- independent pathways or even without any costimulation (10, 11). In ceptor cross-linking by membrane-bound antigen requires CD28/B7 co- contrast to resting T cells, the role of CD28/B7 costimulation in stimulation whereas IFN-␥ secretion and cell proliferation does not. The completely activated T cells is much less clear. Once T cells are efficiency of cytolysis by receptor-grafted CTLs does not depend on and is activated, the triggering of antigen-specific cytolysis via the TCR/ not affected by CD28 costimulation. The data demonstrate that CTL CD3 complex appears to be independent of CD28/B7 costimulation. proliferation, cytokine secretion, and cytolysis upon receptor cross-linking Moreover, the proliferation of CD8ϩ T cells seems to be uncoupled are differentially modulated by CD28 costimulation and that cytolysis from their cytolytic activity but is substantially enhanced by B7 does not require B7 expression on target cells. costimulation (12). In this report, we explored the role of CD28 costimulation for INTRODUCTION receptor-mediated signaling of T cells grafted with chimeric receptors displaying antibody-like specificity for the CEA and the Hodgkin’s T cells grafted with a recombinant receptor combine the advantages lymphoma-associated CD30 antigen, respectively. The signaling do- of MHC-independent, antibody-based antigen binding with efficient mains of the chimeric receptors are derived from the Fc⑀RI ␥-chain or T-cell activation upon specific binding to the receptor ligand (1–3). the CD3 ␨-chain, respectively. Using this panel of recombinant re- The antigen binding domain of the receptor consists of a single-chain ceptors, we here demonstrate that receptor-mediated target cell lysis is antibody fragment (scFv) that is derived from a mab.3 The intracel- nearly unaffected by CD28/B7 signaling. In contrast, antigen-specific lular signaling domain is derived from the cytoplasmic part of a IL-2 secretion and cellular proliferation of grafted T cells is highly membrane-bound receptor that is capable of inducing cellular activa- dependent on CD28 costimulation and, at least, is partially uncoupled tion, e.g., the Fc⑀RI receptor ␥-chain or the CD3 ␨-chain. T cells from target cell lysis capacity. engrafted with the recombinant receptor molecule are designed to induce a MHC-independent, antigen-specific immune response upon receptor cross-linking by antigen, thus functioning as a valuable MATERIALS AND METHODS model system for the analysis of receptor-mediated T-cell functions (reviewed in Refs. 4 and 5). Cell Lines and Antibodies. LoVo cells (ATCC CCL 229) and LS174T T cells can be permanently grafted with designed, recombinant (ATCC CCL 188) cells are CEA-expressing colon carcinoma cell lines. The cell lines were cultured in RPMI 1640 supplemented with 10% (v/v) FCS. The receptors by retroviral transduction of vector constructs encoding the anti-CEA mab BW431/26, the anti-CD30 mab HRS3, the anti-HRS3 idiotypic mab 9G10, and the anti-idiotypic mab BW2064/36 with specificity for the Received 6/5/00; accepted 12/28/00. anti-CEA mab were described elsewhere (13–15). The human CD30-Fc fusion The costs of publication of this article were defrayed in part by the payment of page protein (16) was purified from supernatants of transfected Chinese hamster charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ovary cells and myeloma cells, respectively, by affinity chromatography on 1 Supported by Grant SFB502 from the Deutsche Forschungsgemeinschaft, Bonn, and antihuman IgG antibody agarose (Sigma Chemical Co., Deisenhofen, Germa- Grants 70-2235-Ab1 and 10-1175-Se4 from the Deutsche Krebshilfe, Bonn. ny). The anti-CD3 mab OKT3 was obtained from ATCC (ATCC CRL 8001), 2 To whom requests for reprints should be addressed, at Laboratory of Tumor Genetics, and the anti-CD28 mab 15E8 was kindly provided by R. van Lier (Department Department I of Internal Medicine, University of Cologne, Josef-Stelzmann-Str. 9, D-50931 Cologne, Germany. Phone and Fax: 49-221-478-4130; E-mail: hinrich.abken@ of Immunobiology, University of Amsterdam, Amsterdam, The Netherlands). medizin.uni-koeln.de. Both mabs were protein A purified (Pharmacia-Amersham, Freiburg, Ger- 3 The abbreviations used are: mab, monoclonal antibody; TCR, T-cell receptor; CEA, many) from murine ascites or cell culture supernatants. carcinoembryonic antigen; IL, interleukin; ATCC, American Type Culture Collection; Generation of B7 Transfectants. The bicistronic expression plasmid pCB/ CMV, cytomegalovirus; PE, phycoerythrin; PBL, peripheral blood lymphocyte; XTT, 2,3-bis(2-methoxy-4-nitro-5-sulfonyl)-5[(phenyl-amino)carbonyl]-2H-tetrazolium hydroxide; neo contains the coding sequences for the B7-1 molecule and the B7-2 ICAM, intercellular adhesion molecule; MuLV, murine leukemia virus. molecule, linked by an internal ribosomal entry site sequence, for simultaneous 1976

expression of both B7-1 and B7-2 under control of the CMV early promoter/ (0.5 ␮g/ml; both purchased from Endogen). IL-2 was bound by a solid-phase enhancer (17). The colorectal carcinoma cell lines LoVo and LS174T, respec- antihuman IL-2 antibody (1:250) and detected by a biotinylated antihuman tively, were transfected with pCB/neo DNA construct using the FuGENE IL-2 antibody (1:250; both purchased from PharMingen). The reaction product transfection reagent (Roche Diagnostics, Mannheim, Germany), according to was visualized by a combination of a peroxidase-streptavidin conjugate (1: the manufacturer’s instructions. After culture for 2 days, transfected cells were 10,000) and ABTS (both purchased from Roche Diagnostics) as a substrate. selected in the presence of G418 (2 mg/ml; Sigma) and subsequently sub- Cell Proliferation of PKH26-labeled T Cells. The membrane of receptor cloned by limiting dilution. Simultaneous expression of B7-1 and B7-2 on the grafted and nontransduced PBLs, respectively, was labeled with the red fluo- surface of transfected cells was determined by flow cytometry analysis as rescent dye PKH26 (Sigma) as described recently (22, 23). PKH26-labeled, described below. receptor grafted and nontransduced PBLs, respectively, were cocultured for Generation of Chimeric Receptors and Transduction of Peripheral 72 h with B7-expressing and nontransfected CEAϩ colon carcinoma cells Blood T Cells. Both the generation and expression of the CEA-specific (5 ϫ 104 cells/well), respectively. Nonadherent PBLs were harvested and BW431/26-scFv-Fc-␥ and -␨ receptors and the CD30-specific HRS3-scFv-␥ analyzed by flow cytometry. The analysis was performed in triplicate, and the receptor were described recently in detail (18–20). To express the recombinant cells from individual wells were pooled before analysis. The lymphocyte receptors in peripheral blood T cells, their expression cassettes were inserted population was defined setting forward and side-scatter parameters. Dead cells into the retroviral vector pSTITCH (21). Briefly, the DNA sequences coding were excluded from analysis by staining with propidium iodide. Cell division for the recombinant receptors were flanked with NcoI(5Ј) and BglII (3Ј) is accompanied by reduced intensity of the membrane dye PKH26. Cell restriction sites, respectively, by PCR techniques using the following primer division was monitored by PKH26 fluorescence intensity; histogram markers oligonucleotides: 5Ј-TGA TCC ATG GAC TAG TAC GTA ATG GAT TTT were set with Ͼ97.5% of freshly labeled viable lymphocytes laying inside the CAG GTGCAG ATT TTC-3Ј (sense); 5Ј-GGC AGA TCT GTC GAC CTG defined histogram region. TTA GCG AGG GGG CAG-3Ј (␨ antisense); 5Ј-GGC AGA TCT GAT CAG XTT-based Cytotoxicity Assay. Specific cytotoxicity of receptor grafted TCG ACT CTA AAG CTA CTG TGG TGG-3Ј (␥ antisense; restriction sites T cells to target cells was monitored by XTT-based colorimetric assay accord- are underlined). The PCR products were digested and inserted into the NcoI ing to Jost et al. (24). Briefly, receptor grafted and nontransduced T cells and BamHI sites of the pSP72 vector DNA that also contains parts of the (1 ϫ 105 cells/well) were cocultivated in triplicate in round-bottomed micro- ϩ MuLV splice-acceptor and long terminal repeat sequences (21). Herewith, the titer plates with B7-expressing or nontransfected CEA LoVo and LS174T BamHI site was eliminated by the BamHI/BglII ligation. The resulting DNA cells, respectively. After 48 h, supernatants were harvested and in addition was digested with BglII and partially digested with XhoI. Subsequently, DNA analyzed for cytokine secretion. XTT reagent (1 mg/ml; Cell Proliferation Kit fragments comprising the sequences for the recombinant receptor and parts of II; Roche Diagnostics) was added to the cells and incubated for 90 min at 37°C. the MuLV splice acceptor and long terminal repeat sequences were inserted Reduction of XTT to formazan by viable tumor cells was monitored colori- into the BglII and XhoI sites of the E3qB vector DNA constituting the metrically at an absorbance wavelength of 450 nm and a reference wavelength MuLV-derived retroviral expression vectors pSTITCH/BW431/26-scFv-Fc- of 630 nm. Maximal reduction of XTT was determined as the mean of six wells ␨/␥ and pSTITCH/HRS3-scFv-␥, respectively. To generate GALV- containing only tumor cells, and the background was determined as the mean pseudotyped retrovirus for infection of peripheral blood T cells, the retroviral of six wells containing RPMI 1640 and 10% FCS. The nonspecific formation expression vector DNA (6 ␮g of DNA) was cotransfected into 293T cells by of formazan attributable to the presence of effector cells was determined from calcium phosphate coprecipitation with the retroviral helper plasmid DNAs triplicate wells containing effector cells alone but at the same numbers as in the pHIT60 and pCOLT (each 6 ␮g of DNA). pHIT60 encodes the MuLV gag and corresponding experimental wells. The number of viable tumor cells [%] was pol genes, whereas pCOLT encodes the GALV-envelope gene under control of calculated as follows: the CMV promotor/enhancer (21). Peripheral blood lymphocytes from healthy donors were isolated by density centrifugation and cultured for 48 h in RPMI Absorbanceexperimental wells Ϫ corresponding number of effector cells % viability ϭ ϫ 100 1640 supplemented with 10% FCS in the presence of IL-2 (400 units/ml; Absorbancetumor cells without effectors Ϫ medium Endogen, Woburn, MA) and OKT3 mab (100 ng/ml). The cells were harvested, washed, resuspended in medium with IL-2 (400 units/ml), and cocul- To demonstrate the specificity of BW431/26-scFv-Fc-␥/␨ receptor-mediated ϩ tivated for 48 h with transiently transfected 293T cells. T cells were harvested, lysis of CEA tumor cells, the assay was also performed in the presence of the and receptor expression was monitored by flow cytometric analysis. anti-BW431/26 idiotypic mab BW2064/36 and as a control in the presence of Immunofluorescence Analysis. Receptor grafted T cells were identified the anti-HRS3 idiotypic mab 9G10 (each 2 ␮g/ml). by two-color immunofluorescence using BW431/26-scFv- and HRS3-scFv- ␮ specific anti-idiotypic mabs (both IgG1; 10 g/ml), respectively, and the RESULTS anti-CD3 mab OKT3 (IgG2a; 2.5 ␮g/ml). Bound antibodies were detected by Ј ␮ a FITC-conjugated F(ab )2 antimouse IgG1 antibody (2 g/ml) and a PE- Expression of Recombinant Receptors in Peripheral Blood Ј ␮ conjugated F(ab )2 antimouse IgG2a antibody (2 g/ml; both purchased from Lymphocytes. The DNA expression cassettes encoding the recom- Southern Biotechnology, Birmingham, AL), respectively. Expression of B7-1 binant anti-CEA and anti-CD30 T-cell receptors, respectively, were and B7-1 on transfected tumor cells was determined by using FITC-conjugated inserted into the retroviral expression vector pSTITCH. Peripheral anti-B7–1 (MAB104) and PE-conjugated anti-B7–2 mabs (HA5.2B7; both blood T cells were isolated and transduced as described in “Materials purchased from Coulter-Immunotech, Hamburg, Germany), respectively. Ex- pression of CEA was monitored by incubation of tumor cells with an anti-CEA and Methods.” T cells expressing the anti-CEA or anti-CD30 receptor mab (CEJ065; Coulter-Immunotech) and subsequently a FITC-conjugated were identified by two-color fluorescence using BW431/26-scFv- and Ј ␮ HRS3-scFv-specific anti-idiotypic mabs, respectively, and the anti- F(ab )2 antihuman IgG1 antibody (2 g/ml). Immunofluorescence was analyzed by using a FACScan cytofluorometer equipped with the CellQuest CD3 mab OKT3 (Fig. 1). research software (Becton Dickinson, Mountain View, CA). CD28-mediated Costimulation Is Required for IL-2 Secretion Stimulation of Receptor Grafted Peripheral Blood T Cells. Microtiter but not for IFN-␥ Secretion by Grafted T Cells. Stimulation of plates were coated with the anti-CD28 mab 15E8, the anti-CD3 mab OKT3, anti-CD3/IL-2 preactivated peripheral T cells with 0.5–2 ␮g/ml of the anti-BW431/26 idiotypic antibody BW2064/36, or the CD30-Fc fusion solid-phase bound anti-CD28 mab induced secretion of IFN-␥ but not ϫ 5 protein. Transduced and nontransduced peripheral blood T cells (1 10 of IL-2 (data not shown). We here monitored IFN-␥ and IL-2 secre- cells/well), respectively, were incubated for 48 h at 37°C in coated microtiter tion by grafted T cells upon stimulation via the recombinant receptor plates. Alternatively, receptor grafted and nontransduced T cells (1.25 ϫ 104 to 10 ϫ 104/well) were cocultivated for 48 h with B7-1- and B7-2-expressing in the presence or without CD28 costimulation. Thus, the anti-CD3 ϩ ϩ ␮ CEA colon carcinoma or nontransfected CEA colon carcinoma cells mab OKT3 (2 g/ml), the anti-BW431/26 idiotypic mab BW2064/36 (5 ϫ 104/well). The culture supernatants were analyzed for secretion of IFN-␥ (4 ␮g/ml), and the CD30-Fc fusion protein (4 ␮g/ml), respectively, and IL-2 by ELISA. Briefly, IFN-␥ was bound by a solid-phase antihuman were coated without or together with the anti-CD28 mab15E8 (2 IFN-␥ mab (1 ␮g/ml) and detected by a biotinylated anti-human IFN-␥ mab ␮g/ml) onto microtiter plates, and HRS3-scFv-␥ receptor and BW431/ 1977

Fig. 1. Two-color immunofluorescence of receptor-grafted peripheral blood T cells. Peripheral blood T cells grafted with the HRS3- scFv-␥ (left), the BW431/26-scFv-Fc-␥ (middle), and the BW431/26- scFv-Fc-␨ (right) receptors were simultaneously incubated with the anti-CD3 mab OKT3 (IgG2a) and the anti-HRS3-scFv-specific idiotypic mab 9G10 (IgG1, top row) or the anti-BW431/26-scFv specific anti-idiotypic mab BW2064/36 (IgG1, bottom row) and analyzed by flow cytometry. Bound antibodies were detected simultaneously by a PE-conjugated antimouse IgG2a antibody and a FITC-conjugated antimouse IgG1 antibody.

26-scFv-Fc-␥ receptor grafted lymphocytes, respectively, were incu- IFN-␥ secretion that was furthermore enhanced by CD28 costimula- bated for 48 h. Analyses of the supernatants revealed that receptor- tion (Fig. 3, A and B). In contrast, IL-2 secretion was only recorded grafted T cells secrete high amounts of IFN-␥ upon specific upon receptor cross-linking in the presence of CD28 costimulation, stimulation with the receptor ligand as well as upon nonspecific irrespective of T-cell preactivation by anti-CD3 mab plus IL2 or by stimulation via anti-CD3 or in the presence of anti-CD28 mab alone anti-CD3 mab plus anti-CD28 mab. Furthermore, T cells grafted (Fig. 2A). The amount of secreted IFN-␥ upon receptor-mediated either with the recombinant ␥-or␨-chain receptor require both CD28 stimulation by binding to ligand is furthermore enhanced by CD28 costimulation for induction of specific IL-2 secretion. costimulation. Receptor-mediated IL-2 secretion, however, was only We also tested whether the requirement of CD28 costimulation can observed in the presence of CD28 costimulation. In contrast, no IL-2 be substituted by high concentrations of the receptor ligand. T cells secretion was detected upon cross-linking of the recombinant receptor grafted with the BW431/26-scFv-Fc-␥ receptor were incubated for or of CD3 in the absence of CD28 costimulation (Fig. 2B). We 48 h in microtiter plates coated with increasing amounts of either the conclude that T cells grafted with the anti-CD30 receptor and the anti-BW431/26 idiotypic mab BW2064/36 or, for comparison, the anti-CEA receptor, respectively, both require CD28 costimulation in anti-CD3 mab OKT3 in the presence/absence of a constant amount of addition to antigen-specific stimulation for the induction of IL-2 the anti-CD28 mab 15E8. Analysis of the culture supernatants re- secretion. vealed that the immobilized anti-idiotypic mab BW2064/36 as well as We asked whether CD28 costimulation of lymphocytes prior to the anti-CD3 mab alone efficiently induced IFN-␥ secretion (Fig. 4, B retroviral transduction or the signaling domain of the recombinant and D), whereas no IL-2 secretion was observed even after stimulation receptor modulates IL-2 secretion upon antigen-specific stimulation with antibodies in coating concentrations of 20 ␮g/ml (Fig. 4, A and of receptor-grafted T cells. We used two recombinant receptors C). CD28 costimulation, on the other hand, resulted in both recom- (BW431/26-scFv-Fc-␥; BW431/26-scFv-Fc-␨) that have the same ex- binant receptor- and CD3-mediated secretion of high amounts of IL-2. tracellular antigen binding domain but different intracellular signaling Moreover, CD28 costimulation lowered the threshold for IFN-␥ se- domains, either the Fc⑀RI ␥-chain or the CD3 ␨-chain. Peripheral cretion dramatically (Fig. 4, B and D). Stimulation of receptor-grafted blood T cells were stimulated with the anti-CD3 mab OKT3 plus T cells with low concentrations of the anti-CD28 mab without sec- either IL-2 or the anti-CD28 mab 15E8 prior to engraftment with the ondary signaling via the recombinant receptor or CD3 induced only recombinant receptor. Antigen-specific receptor cross-linking induced low amounts of IFN-␥ and no IL-2 secretion (Fig. 4, E and F).

Fig. 2. Cytokine secretion by peripheral T cells grafted with HRS3-scFv-␥ (■) and BW431/26-scFv-Fc-␥ (ᮀ) receptors, respectively, upon receptor cross-linking and CD28 costimulation. Micro- titer plates were coated with the anti-CD3 mab OKT3 (2 ␮g/ml), the CD30-Fc protein (4 ␮g/ml), or the anti-BW431/26-scFv (anti-CEA- scFv) idiotypic mab BW2064/36 (4 ␮g/ml) in the presence or absence of the anti-CD28 mab 15E8 (2 ␮g/ml). Peripheral blood T cells (1 ϫ 105/ml), grafted with the HRS3-scFv-␥ receptor (24% transduced T cells; ■) and BW431/26-scFv-Fc-␥ receptors (14.4% transduced T cells; ᮀ), respectively, were incubated for 48 h on coated microtiter plates, and the supernatants were analyzed by ELISA for the secretion of IFN-␥ (A) and IL-2 (B), respectively. The assay was performed in triplicate; bars, SE.

1978

ribosomal entry site sequence under control of the CMV promotor/ enhancer. Isolated cell clones simultaneously express B7-1 and B7-2 (termed as B7 positive), as demonstrated by flow cytometry (data not shown). We cocultivated BW431/26-scFv-Fc-␨ receptor grafted and nontransduced T cells with B7-expressing and nontransfected LoVo and LS174T cells, respectively, and recorded the amount of IFN-␥ and IL-2 secreted into the culture supernatant. Cocultivation of receptor-grafted T cells with B7-positive tumor cells resulted in the secretion of substantially higher amounts of IFN-␥ compared with cocultivation with B7 negative tumor cells (Fig. 5, B and D). The effect was similar for both B7-expressing LoVo cells and LS174T cells, respectively. Cocultivation of nontransduced T cells with B7-positive or -negative target cells induced no IFN-␥ secretion. Cocultivation of receptor-grafted T cells with B7-negative LoVo cells resulted in IL-2 secretion, although in low amounts (Fig. 5A), whereas cocultivation with B7-negative LS174T cells did not (Fig. 5C). In contrast, costimulation of recombinant receptors with membrane-bound B7 resulted in secretion of high amounts of IL-2 (Fig. 5, A and C). Cocultivation of nontransduced lymphocytes with B7-positive or -negative tumor cells, however, did not induce IL-2 secretion. These data demonstrate that B7-mediated signaling alone is not sufficient for cytokine secretion. Recombinant receptor-mediated low IL-2 secretion in the presence of nontransfected LoVo cells is in contrast to the strict requirement of CD28 costimulation on solid-phase bound receptor ligand. This effect may be attributable to the expression of ICAM-1 on tumor cells, which also provides costimulatory activity (25). Flow cytometric analysis of LoVo and LS174T tumor cells revealed that LoVo cells express constitutively ICAM-1 in high amounts, whereas LS174T cells express only very low amounts of ICAM-1 (data not shown), suggesting that high levels of ICAM-1 delivered the costimulatory Fig. 3. Influence of T-cell preactivation and the intracellular signaling domain on CD28-modulated cytokine secretion of recombinant receptor-grafted T cells. Microtiter plates were coated with the anti-BW431/26-scFv idiotypic mab BW2064/36 or an isotype- matched control mAb (each 5 ␮g/ml) in the absence or presence of the anti-CD28 mab 15E8 (2 ␮g/ml). Peripheral blood lymphocytes were either preactivated by incubation with anti-CD3 mab plus IL-2 (A and C) or by anti-CD3 mab plus anti-CD28 mab (B and D) and grafted with BW431/26-scFv-Fc-␨ and BW431/26-scFv-Fc-␥ receptors, respectively. The transduction efficiency of T cells preactivated by anti-CD3 mab plus IL-2 was 25.6% (BW431/26-scFv-Fc-␨ receptor) and 26.8% (BW431/26-scFv-Fc-␥ receptor), respectively; the transduction efficiency of T cells preactivated by anti-CD3 mab plus anti-CD28 mab was 29.2% (BW431/26-scFv-Fc-␨ receptor) and 30.4% (BW431/26-scFv- Fc-␨ receptor), respectively. Receptor grafted T cells (1 ϫ 105 cells/well) were incubated for 48 h in coated microtiter plates, and the supernatants were analyzed by ELISA for the secretion of IFN-␥ (A and B) and IL-2 (C and D), respectively. The assay was performed in triplicate; bars, SE.

Taken together, the data demonstrate that: (a) both recombinant receptor-mediated signaling and signaling via the endogenous CD3/ TCR complex induce IFN-␥ secretion but no IL-2 secretion; (b) CD28 costimulation lowers the threshold for IFN-␥ secretion; (c) CD28 costimulation is required for both CD3 and recombinant receptor- mediated IL-2 secretion, irrespective of the signaling chain of the receptor (Fc⑀RI-␥ or CD3-␨) and of T-cell preactivation with or without CD28 costimulation; and (d) the requirement of CD28 costimulation cannot be substituted by increasing amounts of the receptor ligand. IFN-␥ and IL-2 Secretion of Receptor-grafted T Cells upon Coculture with B7-expressing Tumor Cells. Because specific induction of IFN-␥ and IL-2 secretion differs in terms of their require- Fig. 4. Activation of BW431/26-scFv-Fc-␥ receptor-grafted T cells by increasing ments for CD28 costimulation, we asked whether B7 expression on amounts of anti-CD3 mab and of a receptor-specific anti-idiotypic mab in addition to target cells modulates: (a) cytokine secretion of receptor grafted T CD28 costimulation. Microtiter plates were coated with increasing amounts of the cells; (b) specific cytolysis of target cells; and (c) T-cell proliferation anti-BW431/26-scFv idiotypic mab BW2064/36 (A and B), the anti-CD3 mab OKT3 (C and D), or an IgG1 control mab (E and F; each 0.01–20 ␮g/ml) alone or in addition to the upon receptor cross-linking. Two colorectal cancer lines (LoVo and anti-CD28 mab 15E8 (1 ␮g/ml). Peripheral blood T cells were grafted with the BW431/ LS174) that express similar amounts of CEA (data not shown) were 26-scFv-Fc-␥ receptor (26.8% transduced T cells) and were incubated at 1 ϫ 105 cells/well for 48 h in coated microtiter plates. The supernatants were analyzed by ELISA transfected with the expression vector pCB7neo that contains a cas- for IL-2 (A, C, and E) and IFN-␥ (B, D, and F) secretion. The assay was done in triplicate; sette for the expression of both B7-1 and B7-2 linked via an internal bars, SE. 1979

signal to induce low IL-2 secretion upon cocultivation of receptor grafted T cells with B7-negative LoVo cells (Fig. 5A). Specific Cytolysis of B7-expressing Tumor Cells by Receptor- grafted T Cells. We studied the cytolytic activity of T cells grafted with the BW431/26-scFv-Fc-␨ receptor toward B7-expressing tumor cells by applying a tetrazolium salt (XTT)-based cytotoxicity assay. This assay is suitable for monitoring of target cell lysis over an extended period of time, even with low numbers of effector cells (24). Coincubation of BW431/26-scFv-Fc-␨ receptor grafted T cells with CEAϩ LoVo and LS174T tumor cells, respectively, resulted in highly efficient lysis of CEAϩ target cells, whereas T cells lacking the CEA-specific receptor were poorly cytolytic (Fig. 6, A and B). Co- stimulation via B7, however, only slightly amplified the cytolytic activity of receptor grafted T cells against B7-expressing LoVo cells but not against B7-positive LS174T cells. Target cell lysis is restricted to the recombinant receptor because incubation of grafted T cells in the presence of the anti-BW431/26 idiotypic mab BW2064/36 abol- ished specific cytolysis of tumor cells, whereas an isotype-matched control mab (9G10) did not (Fig. 6, C and D). Proliferation of Receptor-grafted T Cells upon Incubation with B7-expressing Tumor Cells. To monitor T-cell proliferation upon receptor triggering in the context of B7-CD28 costimulation, we labeled the cell membrane of receptor grafted lymphocytes with the Fig. 5. T cells grafted with an anti-CEA receptor specifically secrete IL-2 and IFN-␥ red fluorochrome PKH26. Labeled lymphocytes grafted with BW431/ upon incubation with B7-expressing CEAϩ tumor cells. Nontransduced T cells (open 26-scFv-Fc-␥ and BW431/26-scFv-Fc-␨ receptors, respectively, were symbols) and T cells grafted with the BW431/26-scFv-Fc-␨ receptor (closed symbols; 1.25–10 ϫ 104/well) were cocultured for 48 h with B7-negative, CEAϩ LoVo (A and B) incubated in the presence of B7-positive and -negative LS174T tumor and LS174T (C and D) cells (5 ϫ 104/well), respectively, and their B7-positive deriva- cells, respectively. After 72 h, nonadherent cells were harvested and tives, respectively. The supernatants were harvested and analyzed by ELISA for IL-2 (A analyzed by flow cytometry as described in “Materials and Methods.” and C) and IFN-␥ (B and D) secretion, respectively. The cell ratio indicates lymphocytes: tumor cells; the number of receptor-grafted T cells was 10% of the total number of Incubation with LS174T cells induced proliferation of lymphocytes lymphocytes used for the assay. The assay was done in triplicate; bars, SE. grafted with ␥- and ␨-chain receptors, respectively (Fig. 7). Receptor-

Fig. 6. Specific cytolysis of B7-transfected and nontransfected CEAϩ tumor cells by T cells grafted with the anti-CEA receptor. A and B, PBL (circles) and BW431/26- scFv-Fc-␨ receptor-grafted lymphocytes (squares) were cocultured for 48 h (1.25–10 ϫ 104/well) with B7-positive (closed symbols) and B-7-negative (open symbols) CEAϩ LoVo (A) and LS174T (B) tumor cells (5 ϫ 104/well), respectively. C and D, PBLs and lymphocytes grafted with the BW431/26-scFv-Fc-␨ receptor were cocultured for 48 h (1 ϫ 105/well) with B7-positive and B7-negative CEAϩ LoVo (C) and LS174T (D) cells (5 ϫ 104/well), respectively, in the presence of the anti-HRS3-scFv idiotypic mab 9G10 (f) and the anti-BW431/26 idiotypic mab BW2064/36 (Ⅺ; each 2 ␮g/ml), respectively. The viability of tumor cells was determined by a XTT assay as described in “Materials and Methods.” The cell ratio indicates lymphocytes:tumor cells; the number of receptor-grafted T cells was ϳ10% of the total number of lymphocytes used for the assay. The assay was done in triplicate; bars, SE.

1980

Fig. 7. Specific proliferation of T cells grafted with anti-CEA-␥ and anti-CEA-␨ receptors upon cocultivation with B7-positive and B7-negative LS174T tumor cells. PBLs and lymphocytes grafted with the BW431/26-scFv-Fc-␥ receptor (18% receptor-expressing T cells) or the BW431/26-scFv-Fc-␨ receptor (10% receptor-expressing T cells) were labeled by the red fluorochrome PKH26 as described in “Materials and Methods.” Labeled cells (1 ϫ 105 cells/well) were cocultivated for 72 h with B7-positive and B7- negative CEAϩ LS174T cells (5 ϫ 104/well). Nonadherent cells were harvested and analyzed by flow cytometry; the number of proliferating lymphocytes was determined as described in “Mate- rials and Methods.”

triggered proliferation was significantly enhanced by incubation with (a) specific target cell lysis by receptor-grafted T cells is independ- B7-expressing LS174T cells. As controls, lymphocytes with and ent of CD28 costimulation, allowing efficient cytolysis of those an- without specific receptor did not proliferate significantly in the ab- tigen-expressing tumor cells that lack costimulatory molecules. ϩ sence of CEA tumor cells. Proliferation of receptor grafted lympho- (b) IL-2 secretion by receptor-grafted T cells requires CD28 co- cytes is specifically mediated by the anti-CEA receptor because T stimulation. Because IL-2 plays a key role for Th1-based cellular cells without specific receptor are not induced to proliferate upon immunity (26), targeting of tumor cells lacking costimulatory mole- ϩ incubation with CEA tumor cells, irrespective of their B7 expres- cules by receptor-grafted T cells will only result in a limited cellular sion. These data indicate that CD28 costimulation is not required, but immune response. The acquisition of additional cellular effectors, e. g. substantially enhances, antigen-specific T-cell proliferation. natural killer cells, and the maintenance of a prolonged antitumor cell reactivity are likely to require CD28 costimulation, despite high DISCUSSION IFN-␥ secretion levels of receptor-grafted T cells. On the other hand, the restricted capacity of receptor-grafted T cells to secrete IL-2 in the In this report, we monitored specific T-cell activation via recom- absence of CD28 costimulation will limit the immune response, thus binant receptors in the context of CD28 costimulation using both preventing autoaggression of activated receptor T cells against normal solid-phase bound receptor ligands as well as ligand-expressing target tissues. From a practical viewpoint, CD28 costimulation may be cells. Our data demonstrate that receptor-mediated target cell lysis does not require CD28 costimulation; however, receptor-mediated useful to enhance receptor-mediated cytolysis in the immunotherapy cytokine secretion and T-cell proliferation depend on CD28 costimu- of malignancies. Blocking of CD28 costimulation may be the strategy lation. Specific receptor signaling upon binding to solid-phase ligand of choice to limit the immune response. without CD28 costimulation induces IFN-␥ secretion but no IL-2 (c) Cytokine secretion upon cross-linking of the recombinant re- secretion, despite T-cell preactivation by either anti-CD3/IL-2 or ceptor and of the CD3/TCR complex, respectively, was found to be anti-CD3/CD28. This partial inefficiency in receptor-mediated cellu- modulated in a similar fashion by CD28 costimulation. In accordance lar activation could not be overcome by increasing the amount of to our data, target cell lysis and effector cell proliferation upon receptor ligand. Receptor cross-linking and simultaneous CD28 co- signaling via the endogenous T-cell receptor in preactivated T cells stimulation is required for the efficient induction of IL-2 secretion. are differentially modulated by CD28 costimulation (12). We con- The data reported here have substantial consequences for the conclude that cellular activation of grafted T cells via the recombinant cept of MHC-independent cellular targeting by recombinant receptor receptor is regulated and integrated in a similar fashion as T-cell molecules: activation via the common CD3/TCR complex. Recombinant T-cell 1981

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 2001 American Association for Cancer Research. T-CELL ACTIVATION BY RECOMBINANT RECEPTORS receptor molecules as applied in this study may therefore represent a 9. Cai, Z., and Sprent, J. Influence of antigen dose and costimulation on the primary ϩ valuable tool for the analysis of T-cell activation in peripheral blood response of CD8 T cells in vitro. J. Exp. Med., 183: 2247–2257, 1996. 10. Dubey, C., Croft, M., and Swain, S. L. Costimulatory requirements of naive CD4ϩ lymphocytes. T cells. J. Immunol., 155: 45–57, 1995. (d) Target cell lysis via a recombinant TCR was demonstrated 11. Wang, B., Maile, R., Greenwood, R., Collins, E. J., and Frelinger, J. A. Naive CD8ϩ T cells do not require costimulation for proliferation and differentiation into cytotoxic recently to be coregulated by the adhesion molecule ICAM-1 (27). effector cells. J. Immunol., 164: 1216–1222, 2000. Here we demonstrate that CD28 costimulation has a minor effect on 12. Krummel, M. F., Heath, W. R., and Allison, J. Differential coupling of second signals the efficiency of target cell lysis but substantially influences other for cytotoxicity and proliferation in CD8ϩ T cell effectors: amplification of the lytic potential by B7. J. Immunol., 163: 2999–3006, 1999. activation parameters, such as cell proliferation and cytokine secre- 13. Kaulen, H., Seemann, G., Bosslet, K., Schwaeble, W., and Dippold, W. Humanized tion. On the other hand, target cells with high expression of ICAM-1, anticarcinoembryonic antigen antibody: strategies to enhance human tumor cell as demonstrated for LoVo cells, can induce receptor grafted T cells to killing. Year Immunol., 7: 106–109, 1993. 14. Pohl, C., Renner, C., Schwonzen, M., Schobert, I., Liebenberg, V., Jung, W., Wolf, secrete IL-2 without CD28 costimulation but to a much lower extent J., Pfreundschuh, M., and Diehl, V. CD30 specific AB1-AB2-AB3 internal image than CD28 costimulation. antibody network: potential use as anti-idiotype vaccine against Hodgkin’s lym- Taken together, the data suggest that in the case of LoVo cells, phoma. Int. J. Cancer, 54: 418–425, 1993. 15. Engert, A., Martin, G., Pfreundschuh, M., Amlot, P., Hsu, S. M., Diehl, V., and CD28 costimulation is, at least in part, substituted by the ICAM-1 Thorpe, P. Antitumor effects of ricin A chain immunotoxins prepared from intact costimulation pathway. Costimulatory and adhesion molecules obvi- antibodies and FabЈ fragments on solid human Hodgkin’s disease tumors in mice. ously comodulate cell activation parameters, i.e., cytokine secretion, Cancer Res., 90: 2929–2935, 1990. 16. Renner, C., Jung, W., Sahin, U., Van-Lier, R., and Pfreundschuh, M. The role of proliferation, and cytolysis, independently of each other. The high lymphocyte subsets and adhesion molecules in T cell-dependent cytotoxicity medi- variability in the expression pattern of these molecules on the target ated by CD3 and CD28 bispecific monoclonal antibodies. Eur. J. Immunol., 25: 2027–2033, 1995. cells, moreover, substantially affects the method and the efficiency of 17. Jung, D., Hilmes, C., Knuth, A., Jaeger, E., Huber, C., and Seliger, B. Gene transfer mounting cellular immune responses against virus-infected or neo- of the co-stimulatory molecules B7-1 and B7-2 enhances the immunogenicity of plastically transformed target cells using cytotoxic T cells equipped human renal cell carcinoma to a different extent. Scand. J. Immunol., 50: 242–249, 1993. with either the native MHC-restricted or a grafted MHC-independent 18. Hombach, A., Heuser, C., Sircar, R., Tillmann, T., Diehl, V., Pohl, C., and Abken, H. TCR. A chimeric T cell receptor recognizing the CD30 antigen converts cytotoxic T cells to specificity for Hodgkin and Reed Stemberg cells. Cancer Res., 58: 1116–1119, 1998. ACKNOWLEDGMENTS 19. Hombach, A., Koch, D., Sircar, R., Heuser, C., Diehl, V., Kruis, W., Pohl, C., and Abken, H. A chimeric receptor that selectively targets membrane-bound carcinoem- We thank Dr. R. Bolhuis, Daniel den Hoed Cancer Center, Rotterdam, The bryonic antigen (mCEA) in presence of soluble CEA. Gene Ther., 6: 300–304, 1999. Netherlands, for supplying us with the retroviral vector system pSTITCH. 20. Hombach, A., Schneider, C., Sent, D., Koch, D., Willemsen, R. A., Diehl, V., Kruis, W., Bolhuis, R. L., Pohl, C., and Abken, H. An entirely humanized CD3 ␥g-chain signaling receptor that directs peripheral blood T cells to specific lysis of carcino- REFERENCES embryonic antigen (CEA) positive tumor cells. Int. J. Cancer, 88: 115–120, 2000. 1. Eshhar, Z., Waks, T., Gross, G., and Schindler, D. G. Specific activation and targeting 21. Weijtens, M. E., Willemsen, R. A., Hart, E. H., and Bolhuis, R. L. A retroviral vector of cytotoxic lymphocytes through chimeric single chains consisting of antibody- system “STITCH” in combination with an optimized single chain antibody chimeric binding domains and the ␥ or ␨ subunits of the immunoglobulin and T-cell receptors. receptor gene structure allows efficient gene transduction and expression in human T Proc. Natl. Acad. Sci. USA, 90: 720–724, 1993. lymphocytes. Gene Ther., 5: 1195–1203, 1998. 2. Moritz, D., Wels, W., Mattern, J., and Groner, B. Cytotoxic T lymphocytes with a 22. Allsopp, C. E., Nicholls, S. J., and Langhorne, J. A flow cytometric method to assess grafted recognition specificity for ERBB2-expressing tumor cells. Proc. Natl. Acad. antigen-specific proliferative responses of different subpopulations of fresh and Sci. USA, 91: 4318–4322, 1994. cryopreserved human peripheral blood mononuclear cells. J. Immunol. Methods, 214: 3. Hwu, P., Yang, J. C., Cowherd, R., Treisman, J., Shafer, G. E., Eshhar, Z., and 175–186, 1998. Rosenberg, S. A. In vivo activity of T cells redirected with chimeric antibody/T-cell 23. Hombach, A., Heuser, C., Gerken, M., Fischer, B., Lewalter, K., Diehl, V., Pohl, C., receptor genes. Cancer Res., 55: 3369–3373, 1995. and Abken, H. T cell activation by recombinant Fc⑀RI ␥-chain immune receptors: an 4. Eshhar, Z., Bach, N., Fitzer-Attas, C. J., Gross, G., Lustgarten, J., Waks, T., and extracellular spacer domain impairs antigen dependent T cell activation but not Schindler, D. G. The T-body approach: potential for cancer immunotherapy. Springer antigen recognition. Gene Ther., 7: 1067–1075, 2000. Semin. Immunopathol., 18: 199–209, 1996. 24. Jost, L. M., Kirkwood, J. M., and Whiteside, T. L. Improved short- and long-term 5. Abken, H., Hombach, A., Reinhold, U., and Ferrone, S. Can combined T cell- and XTT-based colorimetric cellular cytotoxicity assay for melanoma and other tumor antibody-based immunotherapy outsmart tumor cells? Immunol. Today, 19: 2–5, cells. J. Immunol. Methods, 147: 153–165, 1992. 1998. 25. Deeths, M. J., and Mescher, M. F. ICAM-1 and B7-1 provide similar but distinct 6. Greenfield, E. A., Nguyen, K. A., and Kuchroo, V. K. CD28/B7 costimulation: a costimulation for CD8ϩ T cells, while CD4ϩ T cells are poorly costimulated by review. Crit. Rev. Immunol., 18: 389–418, 1998. ICAM-1. Eur. J. Immunol., 29: 45–53, 1999. 7. June, C. H., Bluestone, J. A., Nadler, L. M., and Thompson, C. B. The B7 and CD28 26. Morel, P. A., and Oriss, T. B. Crossregulation between Th1 and Th2 cells. Crit. Rev. receptor families. Immunol. Today, 15: 321–331, 1994. Immunol., 18: 275–303, 1998. 8. Ku¨ndig, T. M., Shahinian, A., Kawai, K., Mittrucker, H. W., Sebzda, E., Bachmann, 27. Weijtens, M. E., Willemsen, R. A., van Krimpen, B. A., and Bolhuis, R. L. Chimeric M. F., Mak, T. W., and Ohashi, P. S. Duration of TCR stimulation determines scFv/␥ receptor-mediated T-cell lysis of tumor cells is coregulated by adhesion and costimulatory requirement of T cells. Immunity, 5: 41–52, 1996. accessory molecules. Int. J. Cancer, 77: 181–187, 1998.

1982

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 2001 American Association for Cancer Research. T-Cell Activation by Recombinant Receptors: CD28 Costimulation Is Required for Interleukin 2 Secretion and Receptor-mediated T-Cell Proliferation but Does Not Affect Receptor-mediated Target Cell Lysis

Andreas Hombach, Dagmar Sent, Claudia Schneider, et al.

Cancer Res 2001;61:1976-1982.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/61/5/1976

Cited articles This article cites 26 articles, 8 of which you can access for free at: http://cancerres.aacrjournals.org/content/61/5/1976.full#ref-list-1

Citing articles This article has been cited by 18 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/61/5/1976.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/61/5/1976. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.