Abstract
T cells of tumor bearers often show defective TCR-mediated signaling events and, therefore, exhibit impaired immune responses. As such, patients with heavy tumor burden are often not amenable to adoptive T cell therapy. To overcome this limitation, we have developed a chimeric receptor that joins an extracellular single chain Fv (scFv) of a specific Ab for Ag recognition to an intracellular protein tyrosine kinase (PTK) for signal propagation. Stimulation through the scFv-PTK receptor should bypass defective TCR-proximal events and directly access the T cell’s effector mechanisms. In this study we describe the optimization of a scFv-PTK configuration, leading to complete T cell activation. The cytosolic PTK Syk is superior to its family member, Zap-70, for intracellular signaling. As a transmembrane (TM) domain, CD4 performs better than CD8 when plastic-immobilized Ag serves as a stimulator. However, when APC are used to trigger chimeric receptors, the need for a flexible spacer between the scFv and TM domains becomes apparent. The CD8α-derived hinge successfully performs this task in chimeric scFv-Syk receptors regardless of its cysteine content. A cytotoxic T cell hybridoma expressing chimeric receptor genes composed of scFv-CD8hinge-CD8TM-Syk or scFv-CD8hinge-CD4TM-Syk is efficiently stimulated to produce IL-2 upon interaction with APC and specifically lyses appropriate target cells in a non-MHC-restricted manner.
New advances in cancer immunotherapy are directed toward the enhancement of the patient’s immune response against his own tumor or the manipulation and stimulation of T cells outside of the body. Our attempts to develop an effective cancer immunotherapy have focused on the redirection of T cells using chimeric TCRs (cTCR)4 that confer non-MHC-restricted, anti-tumor specificities of mAbs onto cytolytic T cells to produce “T bodies” (1). We have more recently adopted the single-chain variable domain (scFv) design as the recognition unit of the cTCR, attaching it to the ζ-chain of the TCR/CD3 or the FcεRI γ-chain. These chimeric receptors can trigger cytokine secretion upon encountering Ag and mediate non-MHC-restricted, Ag-positive target cell lysis. The T body approach is a potent one when either anti-hapten or anti-tumor Abs are employed (2, 3, 4, 5, 6, 7, 8). Furthermore, murine T cells expressing the chimeric receptor were proven to be competent effector cells in the prevention of tumor growth in vivo (9).
In human patients many tumors escape immune surveillance either because they do not induce an immune response or because they evade immune recognition. Often the T cells of cancer patients are anergized and thus nonreactive to the tumor markers they could potentially recognize. Recent reports attribute some of these states of unresponsiveness to alterations in the composition of the TCR and defects in early stages of cellular activation (10, 11, 12, 13). Such impaired signaling may limit the use of cTCRs of the scFvRγ/ζ design, which are also dependent on TCR proximal kinases for signal transduction. To bypass these receptor proximal events we have designed a new type of cTCR, in which we combined the extracellular scFv with intracellular protein tyrosine kinases (PTK).
The cytoplasmic PTKs, Zap-70 and Syk, members of the Syk family of PTKs, have both been shown to participate in the T cell signal cascade (14, 15). Zap-70 was originally discovered by virtue of its ability to associate with ζ-chains (14), but was subsequently shown to also associate with other CD3 components of activated T cells (16). Syk, a 72-kDa PTK, is abundant in several hemopoietic lineages (17) and is activated upon association with the cytoplasmic domains of a number of immune recognition receptors, including the B cell (18) and T cell (15) Ag, Fcε (19), and Fcγ (20, 21, 22, 23) receptors.
In an attempt to enhance the efficiency of the T body signal leading to target cell lysis and to determine the optimal configuration for chimeric scFv-PTK receptors, we have attached a scFv via different transmembrane domains to the PTKs Zap-70 and Syk, thus directly accessing the T cell’s own signaling machinery. Our results support the idea that these two kinases, although homologous in structure, have different regulatory restraints and signaling capabilities. Syk, when triggered in this manner, is capable of transmitting signals for T cell activation, while Zap-70 is not. Furthermore, in Syk-based chimeric receptors, we demonstrate the importance of connecting transmembrane and extracellular hinge domains for obtaining maximal responses to cell-derived signals.
Materials and Methods
Abs and cell lines
Sp6 is an anti-2,4,6-trinitrophenyl (anti-TNP) mAb. GK20.5 is an anti-Sp6 mAb. TNP modified fowl γ-globulin (TNP-FγG) or A.20 lymphoma cells (TNP-A.20) were made as previously described (1, 2). Protein G-Sepharose was obtained from Pharmacia. Oligonucleotide synthesis and DNA sequencing were performed at The Weizmann Institute of Science (Rehovot, Israel).
Construction of chimeric receptors
Construction of the anti-TNP single chain Fv genes from the Sp6 mAb has been described previously (2). For isolation of the Zap-70 tyrosine kinase gene, reverse transcription-PCR was conducted on Jurkat human leukemia RNA using primers derived from the human sequence with appropriate restriction sites added. Similarly, a DNA fragment was prepared containing the coding sequence for the CD8 transmembrane and hinge regions (residues 116–208). RNA from both Jurkat and anti-CD3-stimulated human T cells was used for reverse transcription-PCR of the human syk gene, using primers derived from the porcine sequence with comparison to the predicted human amino acid sequence (24). A DNA fragment encoding the human CD4 transmembrane region (residues 375–395) was synthesized in both directions, ligated, and inserted between appropriate restriction sites. Constructs were sequenced in their entirety before insertion into the pRSV expression vector. Oligonucleotides for construction of four initial chimeric genes (Z4, Z8/H, S4, and S8/H) were synthesized as follows: zap-70 (PCR was performed to clone halves, and DNA pieces were later joined at the silent EcoRI site): first piece 5′ primer, 5′-CGTCTAGAACCATGCCAGACCCCGCGGCGCACCTGCCCTTCTTCT-3′; first piece 3′ primer, 5′-CCATCTGAATTCAGGGTGTCGATTCG-3′; second piece 5′ primer, 5′-CGAATCGACACCCTGAATTCAGATGG-3′; second piece 3′ primer, 5′-GCGTCGACTTCAGGCACAGGCAGCCTCAGCCTTCTG-3′; Syk (PCR was performed to clone halves, and DNA pieces were later joined at a HindIII site): first piece 5′ primer, 5′-CGTCTAGAACCATGGCAGACAGTGCCAACCACTTGCCCTTCTTCT-3′; first piece 3′ primer, 5′-TTCTTCCCCTCGGGGATGGAAAGCTTCCC-3′; second piece 5′ primer, 5′-AAGGACAAAACTGGGAAGCTTTCCATCCC-3′; second piece 3′ primer, 5′-CGCTCGAGTTTAATTAACCACATCGTAGTAGTA-3′; CD8 transmembrane plus hinge (including J region of heavy chain): 5′ primer, 5′-CCGGTCACCGTCTCTTCAGCGCTGAGCAACTCCATCATGTACTTCAG-3′; 3′ primer, 5′-CGTCTAGAGTGGTTGCAGTAAAGGGTGATAAC-3′; and CD4 transmembrane: sense oligonucleotide, 5′-GTCACCGTCTCTTCAGCGGCCCTGATTGTGCTGGGGGGCGTCGCCGGCCTCCTGCTTTTCATTGGGCTAGGCATCTTCTTCT-3′; antisense oligonucleotide, 5′-CTAGAGAAGAAGATGCCTAGCCCAATGAAAAGCAGGAGGCCGGCGACGCCCCCCAGCACAATCAGGGCCGCCGCTGAAGAGACG-3′.
Generation of subsequent chimeric receptors was performed as follows.
S8 (CD8 transmembrane only)
The CD8 hinge was removed until the EcoRV site, at the junction of the hinge and transmembrane regions. The following oligonucleotides were then ligated to the EcoRV site to provide a BstEII site for subsequent fusion to the scFv: 5′-GTCACCGTCTCTTCAGCG-3′ and 3′-GCAGAGAAGTCGC-5′.
S4cys (CD4 transmembrane with added cysteine residue)
The following oligonucleotides were inserted at the NgoMI site of CD4 to introduce a cysteine residue at the transmembrane/scFv junction: 5′GTCACCGTCTCTTCATGCGCATTGATTGTGCTGGGGGGCGTCGG-3′ and 3′-CAGAGAAGTACGCGTAACTAACACGACCCCCCGCAGCGGCC-5′.
S4/H (CD4 transmembrane with CD8 hinge)
The CD8 hinge, until the EcoRV site, was ligated to the CD4 transmembrane at an FspI site, which was introduced with the oligonucleotides used to make the S4cys construct.
S8/H* (wild-type CD8 transmembrane with mutated CD8 hinge)
The cysteine residues within the hinge and transmembrane domains of CD8 were changed to serines using the Chameleon double-stranded, site-directed mutagenesis kit (Stratagene, La Jolla, CA) and the following primers: 5′-pGCGCCCAGAGGCTAGCCGGCCAGCGGCG-3′, 5′-pGCTGGACTTCGCCAGTGACATCTACATCTGGGCGCCCTTGGCCGGGACTAGTGGGGTCCTTCTC-3′, and 5′-pGTTATCACCCTTTATTCGAACCACAGGAACC-3′. Once sequence changes were verified, the mutated hinge piece was ligated to a StyI site in the wild-type transmembrane to obtain the S8/H* receptor.
Expression of chimeric receptors
Transfection into the 27J cytotoxic murine hybridoma was performed by electroporation as previously described (2). Expression of receptors on the surface of transfected cells, selected in G418, was evaluated by immunofluorescence staining using the GK20.5 anti-Sp6 Id and FITC-labeled anti-mouse Fab′ Ab. For Western blotting, cells were solubilized in lysis buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 1% Triton X-100, 2 mM EDTA, 2 mM EGTA, 50 mM NaF, 2 mM Na3VO4, 1 mM PMSF, 0.4% aprotonin (24.4 Trypsin Inhibitor Units/ml), 5 μg/ml leupeptin, and 10 μg/ml soybean trypsin inhibitor) and 50 μg of total lysate protein under partially reducing (lysates plus reducing sample buffer were incubated for 20 min at room temperature) or nonreducing conditions were separated on 10% SDS-polyacrylamide gels. Proteins were then transferred to nitrocellulose membranes and reacted with the GK20.5 Ab. Recognition of the primary Ab was visualized with the ECL (enhanced chemiluminescence) Western blotting detection system (Amersham, Aylesbury, U.K.).
For biotinylation studies, 3 × 107 cells were surface labeled with biotinamidocaproate N-hydroxysuccinimide ester (Sigma Chemical Co., St. Louis, MO) in sodium borate buffer for 15 min at room temperature. After extensive washing, cells were lysed as described above, and chimeric receptors were immunoprecipitated with the GK20.5 Ab bound to protein G-Sepharose. After SDS-PAGE and transfer to nitrocellulose, detection of biotinylated receptors was performed with peroxidase-labeled Extravidin (Sigma) and ECL.
Kinase assays
Kinase assays were performed on anti-Sp6 immunoprecipitates that were washed three times in a buffer containing 50 mM HEPES (pH 7), 150 mM NaCl, 0.1% Triton, 10% glycerol, 1 mM Na3VO4, and 5 mM NaF. Sepharose-bound proteins were then incubated in the same buffer containing 10 mM magnesium acetate, 10 mM MnCl2, and 20 μCi of [γ32P]ATP (3000 Ci/mmol; DuPont-New England Nuclear, Boston, MA) for 10 min at room temperature and washed three more times in wash buffer before resuspension in reducing sample buffer. After separation by SDS-PAGE, proteins were transferred to polyvinylidene difluoride (MSI, Westboro, MA), and the membrane was exposed to x-ray film both before and after treatment with 1 N KOH for 1 h at 55°C.
Functional assays
To measure specific IL-2 production, transfectants were incubated with TNP-modified A.20 cells or plastic-immobilized TNP-fowl γ-globulin (TNP-FγG) in DMEM containing 10% FCS for 18 to 24 h. Supernatants were collected, and IL-2 was assayed as previously described (25) using an IL-2-dependent CTL line and methyltetrazolium acid staining. To evaluate the ability of transfectants to specifically lyse TNP-labeled target cells, transfectants were preincubated with immobilized TNP-FγG for 4 h and then tested in an 8-h 51Cr release assay (25). Both IL-2 and cytotoxicity assays were performed in triplicate. The difference between replicate measurements did not exceed 10% of the indicated value. Determination of the ability of the different cells to produce IL-2 was repeated at least six times. The cytotoxicity assay was performed at different E:T ratios and was repeated at least twice for each cell. The figures presented here depict representative experiments.
Results
Construction and expression of scFv-PTK chimeric receptors
As a model system, the scFv of an anti-TNP Ab, Sp6, was attached to the PTKs Zap-70 or Syk via a transmembrane region. For this purpose, the hinge and the transmembrane domains of CD8α, which normally exist as a dimer on the cell surface, or the transmembrane portion of CD4, which normally exists as a monomer on the cell surface, were used. It has been shown that the CD8α hinge region is important in the expression and extension of Ig-like domains (7, 26). The four signaling molecules (Z4 and S4 for Zap-70 or Syk-derived receptors with CD4 transmembrane; Z8/H and S8/H for Zap-70 or Syk-derived receptors with CD8 transmembrane and hinge) were constructed as shown in Figure 1⇓A and cloned into an expression vector that contains the Rous sarcoma virus long terminal repeat promoter and the neomycin resistance gene (27). The DNA was transfected into a mutant of the murine T cell hybridoma MD45. This mutant, 27J, lacks the TCR as a result of defective production of its α-chain (28). Stable transfectants grown in media containing G-418 were assayed for cell surface expression of cTCRs by reacting cells with an anti-Id Ab (GK20.5) to the scFv. Figure 2⇓ shows representative staining patterns of transfectants after FACS analysis. Chimeric receptors were expressed at varying degrees, yet, in general, the CD8 transmembrane and hinge-containing receptors (S8/H, Z8/H) were more efficiently brought to the cell surface and/or were more accessible to recognition by the GK20.5 Ab (see Discussion).
A schematic representation of the chimeric scFv-PTK receptors constructed in this study. Receptors contain either Zap-70 (Z) or Syk (S) for intracellular signaling, CD4 and/or CD8α segments for transmembrane (4 or 8) and hinge (H) portions, and an anti-TNP scFv for extracellular recognition.
Cell surface detection of scFv-Zap-70 and scFv-Syk receptors. FACS analysis of stable transfectants of the murine T cell hybridoma 27J (designed as in Fig. 1⇑A), stained with irrelevant (solid line) or anti-Id (dashed line) Abs and FITC-labeled anti-mouse Fab′.
Protein expression and kinase function of chimeric receptors were also verified in cell lysates of transfectants. As shown in Figure 3⇓A, immunoblotting with anti-Sp6 idiotypic Ab confirms the expression of Z4 and Z8/H receptors, at apparent molecular masses of 104 and 110 kDa, respectively, in cells that displayed the receptor on the cell surface as well. Similarly, the S4 and S8/H receptors were detected in cell lysates at 106 and 112 kDa, respectively (Fig. 3⇓A). To test for kinase activity, receptors were immunoprecipitated with the GK20.5 anti-Id Ab from equivalent amounts of lysate protein and subjected to in vitro kinase assays. As shown in Figure 3⇓B, all four chimeric receptors could effectively incorporate [γ-32P]ATP. Zap-70-based molecules appear to be less active in this assay (data not shown), consistent with recent results describing this kinase’s inferior intrinsic enzymatic activity compared with that of Syk molecules (29).
Protein expression and kinase activity of scFv-Zap-70 and scFv-Syk receptors. A, Total cell lysates from 27J and its transfectants were separated by SDS-PAGE and immunoblotted with anti-Id Abs. B, Anti-Id immunoprecipitations from 27J and its transfectants were subjected to in vitro kinase assays with [γ-32P]ATP. Washed immunoprecipitates were separated by SDS-PAGE and transferred to a polyvinylidene difluoride membrane, and the membrane was treated with 1 N KOH for 1 h at 55°C.
Functionality of scFv-PTK receptors
To assess their ability to initiate IL-2 production, cells expressing the various Sp6-scFv-PTK receptors were stimulated with plastic-immobilized TNP-modified carrier protein (TNP-FγG) or TNP-modified murine B lymphoma cells (TNP-A.20 or TNP-L1210). Here and throughout this study assays of IL-2 secretion were performed on a fairly large number of clones per transfection, sometimes even multiple transfections. Wherever possible, the analyses were performed on cells expressing similar levels of chimeric receptors on their surface. Yet, it should be noted that the clonal nature of these cells, in fact, prohibits a precise quantitative comparison between them. The data shown are representative experiments of each transfectant’s particular behavior.
Despite effective cell surface expression, both Zap-70-based receptors, Z4 and Z8/H, were unable to trigger the T cell hybridoma to secrete cytokine when stimulated with either TNP-FγG or TNP-A.20 (Fig. 4⇓). Soluble TNP-FγG or additional carrier proteins modified with TNP (such as BSA or keyhole limpet hemocyanin) were also ineffective in triggering cytokine release in these cells (data not shown). This inability to generate IL-2 cannot be attributed to an intrinsic defect in IL-2 gene induction, since treatment with TPA and calcium ionophore effectively stimulates IL-2 production in the scFv-Zap-70 transfectants (not shown).
Ag stimulation of scFv-PTK receptors. 27J and its Zap-70-based (A) or Syk-based (B) transfectants were incubated with either plastic-immobilized TNP-modified FγG or TNP-modified A.20 or L1210 cells. IL-2 production in supernatants was assayed using an IL-2-dependent CTL line.
In contrast to the Zap-70-derived clones, Syk-based transfectants, containing either the CD4 transmembrane (S4) or the CD8 transmembrane and hinge regions (S8/H), secreted IL-2 when reacted with plastic-immobilized TNP-FγG (Fig. 4⇑), and this response was dependent on the molar ratio of TNP to FγG molecule (not shown). Neither type of receptor was stimulated with soluble TNP-FγG or unmodified FγG (data not shown). Interestingly, when chimeric Syk transfectants were incubated with TNP-modified B lymphoma cells (TNP-A.20 or TNP-L1210), only the S8/H clones responded by secreting cytokine (Fig. 4⇑). Even the most potent CD4-based IL-2 secretors upon TNP-FγG stimulation were unable to respond to TNP in cell-bound form. Increasing the concentration of TNP used for modification of cells did not trigger a response in S4 clones, nor did stimulation with other TNP-modified target cells including the B lymphoma Baf, a murine fibrosarcoma (WEHI-164), or IGROV, a human ovarian carcinoma (data not shown). S8/H clones, however, responded in all these cases. Likewise, when Ag was presented to transfectants following cellular processing of TNP-modified protein, only the S8/H receptors were activated (not shown).
The CD8 hinge is responsible for the response to cell-bound Ag
To map the source of the differential response to cell-derived Ag in S4 and S8/H receptors, we constructed two additional chimeric receptors (Fig. 1⇑B). First, we attached the CD8 extracellular hinge piece onto the CD4 transmembrane portion (S4/H). Transfection of this receptor DNA into 27J cells resulted in clones resembling the S8/H transfectants. They showed increased surface expression (Fig. 5⇓A) and could respond to both immobilized protein- and cell-bound Ag at levels similar to S8/H clones (Fig. 6⇓). As shown previously, the S4 clone responded well to immobilized protein Ag only.
Cell surface detection and expression of scFv-Syk receptors. A, FACS analysis of the T cell hybridoma 27J and its transfectants S8 and S4/H, stained with irrelevant (solid line) or anti-Id (dashed line) Abs and FITC-labeled anti-mouse Fab′. B, 27J, S4, and S8 cells were surface biotinylated, and the scFv-Syk receptors were immunoprecipitated with anti-Id Abs. After separation of proteins on SDS-PAGE and transfer to nitrocellulose membranes, biotin-labeled receptors were detected with horseradish peroxidase-streptavidin.
The hinge region is important for stimulation by cell-derived Ag. 27J and its scFv-Syk transfectants were incubated with either plastic-immobilized TNP-modified FγG or TNP-modified A.20 or L1210 cells. IL-2 production in supernatants was assayed using an IL-2-dependent CTL line.
Second, to produce a chimeric receptor in which the scFv is connected to Syk via the CD8 transmembrane domain only (S8), we removed the CD8 extracellular hinge from the original S8/H construct (Fig. 1⇑B). Transfection of this DNA produced clones similar to those that contained the CD4 transmembrane domain, in that both receptors were undetectable by FACS analysis but were, in fact, present on the cell surface, as seen after surface biotinylation and anti-Id immunoprecipitation (Fig. 5⇑, A and B). Interestingly, removal of the hinge domain impaired the ability of this receptor to signal for T cell activation. Thus, the S8 chimeric receptor differed from its S4 counterpart in that it was entirely unable to transfer signals for production of IL-2, whether stimulated with immobilized protein or cell-bound Ag (Fig. 6⇑). From these studies we concluded that for scFv-Syk receptors to trigger IL-2 secretion when encountering cell-bound Ag, CD8 hinge sequences play an important role. Furthermore, for Syk-derived chimeric receptors, the CD8 transmembrane domain alone is insufficient, as it requires the presence of the CD8 hinge region for response to either type of stimulant.
Cysteine residues within the CD8 hinge are not necessary for its effect
CD8α normally exists on the cell surface as disulfide-linked homo- or heterodimers (with CD8β); cysteine residues within the CD8 hinge are responsible for this dimerization state (30, 31). Accordingly, the scFv-Syk receptor containing the CD8 transmembrane and hinge regions is detected in dimeric and multimeric forms when total lysate protein is separated under nonreducing conditions and then immunoblotted with anti-Id Ab (Fig. 7⇓A). Also, since conditions required for recognition of the Sp6 idiotope by Ab GK20.5 do not allow for complete reduction, we detected a larger protein band at 200 kDa in partially reduced lysates of S8/H cells, corresponding to the dimeric form of the chimeric receptor (Fig. 7⇓A). In contrast, the CD4-based receptor (S4) was found only in monomeric form, whether reduced or not, as predicted from its amino acid sequence (Fig. 7⇓A).
Cysteine residues within the hinge confer a dimeric state. A, Total cell lysates from 27J and the original two receptors, S4 and S8/H, were separated by SDS-PAGE under partially reducing (see text) or nonreducing conditions and then immunoblotted with anti-Id Abs. B, As in A, but with lysates from receptors containing either wild-type or mutated (Cys to Ser) hinge sequences.
To ascertain whether these dimeric forms or disulfide-forming cysteine residues are important for the observed effect of the CD8 hinge on response to cell-derived Ag, we constructed yet another chimeric receptor in which the two hinge cysteines were mutated to serines (S8/H* in Fig. 1⇑C). This receptor type was effectively detected on the cell surface by FACS analysis at levels comparable to S8/H receptors (not shown), but was present in monomeric form only, as seen by Western blotting of lysates under nonreducing conditions (Fig. 7⇑B). Apparently, the two cysteines within the CD8 transmembrane region of this chimeric molecule do not form disulfide bonds. Yet, despite its monomeric state, this chimeric Syk receptor could effectively stimulate the cells to secrete IL-2 upon encountering both immobilized protein and cell bound Ag (Fig. 8⇓).
Cysteine residues within the hinge are not important for the response to cell-derived Ag. 27J and its scFv-Syk transfectants were incubated with either plastic-immobilized TNP-modified FγG or TNP-modified A.20 or L1210 cells. IL-2 production in supernatants was assayed using an IL-2-dependent CTL line.
A final construct was then prepared in which a cysteine residue was added to the CD4 transmembrane portion (S4cys, Fig. 1⇑C) at a position analogous to the native cysteine residue present in the TCR ζ-chain or the FcεR γ-chain, which has been shown to be responsible for dimerization of these two membranal signaling proteins (32, 33). This receptor variation behaved similarly to the original CD4-based construct in that it could only respond to immobilized protein Ag (Fig. 8⇑). From the observed signaling capabilities of the last two chimeric Syk receptors, we conclude that the preformation of disulfide-driven multimeric receptor complexes is not a requirement for effective response to Ag presented on a target cell surface.
Cytolytic activity of T cells expressing chimerc Syk receptors
The six chimeric Syk receptors described above were tested for the ability to initiate a signal leading to cytolysis of Ag-coated target cells. To this end, representative transfectants of the cytotoxic hybridoma were prestimulated with immobilized TNP-FγG and then incubated with 51Cr-labeled, TNP-modified A.20 cells. Figure 9⇓ shows the percentage of 51Cr released from target cells. The ability of chimeric receptors to signal for specific target cell lysis was, in general, comparable to their ability to signal for specific IL-2 production. That is, cells containing receptors with the CD8 hinge, whether wild type or mutated (clones S4/H, S8/H, and S8/H*), could effectively lyse Ag-modified target cells in a non-MHC-restricted fashion. Transfectants bearing the mutated hinge (S8/H*), which, on the whole, did not respond as well to cell-derived Ag in the IL-2 assay as their S8/H (wild-type hinge) counterparts (Fig. 8⇑), were also somewhat less cytolytic (Fig. 9⇓). Interestingly, a representative clone of the S4 series, which could not secrete IL-2 upon incubation with Ag-modified cells (see Fig. 4⇑), was able to recognize TNP-labeled target cells under these experimental conditions and showed a low level of specific cytolyis.
Ag stimulation of scFv-Syk receptors triggers specific MHC-independent cytolysis of target cells. Specific cell lysis of TNP-modified A.20 cells by 27J or its scFv-Syk transfectants as measured in a ratio of 2:1 in an 8-h 51Cr release assay.
Discussion
For T body-based immunotherapy, there are several possible advantages to directly combining the tumor-recognizing Ab with a signal-producing PTK. Lytically deficient T lymphocytes from tumor-bearing mice exhibit altered patterns of tyrosine phosphorylation, decreased protein levels of Src family kinase members and of the CD3 ζ-chain, and blunted calcium responses (10, 34). Similarly, reduced expression of ζ-chains and p56lck or altered tyrosine phosphorylation patterns have been reported in tumor-infiltrating lymphocytes taken from patients with colorectal or renal cell carcinoma or non-Hodgkins B cell lymphoma (11, 12, 35). These receptor-proximal defects may restrict the use of cTCRs of the scFv-γ/ζ design, which are also dependent on TCR-associated kinases for signal transduction. Yet, with direct and immediate kinase activation upon Ag recognition, as offered by the scFv-PTK design, this impaired signaling may be overcome.
Such signaling defects not only reflect the unresponsiveness of these individuals’ T cells toward their own tumors, but are also reminiscent of the unusual ζ phosphorylation states and defective kinase recruitment/activation seen in altered peptide ligand-induced T cell anergy (36, 37). Similar modified biochemical phenotypes are seen when comparing signaling events elicited by alloantigen alone (leading to anergy) or by alloantigen and B7-mediated costimulation; the anergic state is characterized by a lack of CD3ε and ζ hyperphosphorylation and subsequent Zap-70 recruitment (38). Furthermore, recent findings have also attributed the induction of T cell anergy to reduced mitogen-activated protein kinase activities and a block of Ras activation (39, 40). Signals downstream of Ras seem to be intact, as is mitogen-activated protein kinase activity upon stimulation by an alternative route, suggesting a defect at the level of specific coupling between the TCR-CD3 complex and Ras (40). Since our current knowledge of T cell activation places the Syk family of tyrosine kinases as proximal signal transducers of the TCR signal (41) and, via Cbl, on a pathway that regulates guanine nucleotide exchange on small G proteins (42, 43, 44), it is tempting to suggest that their direct activation may enable an anergic cell to overcome its signaling block. Thus, although two signals are normally required for complete T cell activation (45), direct activation of a Syk family kinase may bypass the need for the second costimulatory signal, such as that delivered by the CD28 receptor (46).
Collectively, this information prompted us to generate new scFv chimeric receptors, similar in design to those we have described previously (2), yet replacing the FcεRγ or TCRζ intracellular signaling regions with the early tyrosine kinase Zap-70 or Syk. In general, this configuration resembles the chimeric molecules described by Kolanus et al., in which PTKs were tethered to the plasma membrane via CD7 transmembrane and CD16 extracellular domains (24). We reasoned that Ag binding to the extracellular scFv portion of these chimeric molecules should allow for direct triggering of the attached PTKs and immediate stimulation of the T cell’s own signaling machinery. Indeed, Syk-based receptors can efficiently trigger IL-2 production and target cell lysis in an Ag-dependent manner ( Figs. 5–8⇑⇑⇑⇑). Although difficult to compare between cloned transfectants, the scFv-Syk receptors are certainly as effective, if not more so, than chimeras containing γ or ζ intracellularly (C. Fitzer-Attas, and T. Waks, unpublished observations). Interestingly, as in the double chain cTCR (1) and the scFvγ/ζ (2) designs, the scFv-Syk receptors are not activated by soluble Ag, even in multivalent form (such as the TNP-FγG used in this study; data now shown). Activation will only take place with immobilized or cell-bound Ag, suggesting the need for receptor clustering for signal transduction through the intracellular kinase moiety as well. The magnitude of the IL-2 response from scFv-Syk chimeras is quantitatively comparable to that produced from anti-CD3 stimulation of the TCR/CD3-expressing hybridoma, MD45 (data not shown). A more rigorous test of the superior response of kinase-based chimeras will require expression of the different receptors in freshly isolated T lymphocytes or in a transgenic environment.
In our T body approach we have also demonstrated the differences in regulatory restraints imposed upon Zap-70 and Syk. The Zap-70-based chimeric receptors, although efficiently expressed at the cell surface (Fig. 2⇑) and able to autophosphorylate in in vitro kinase assays (Fig. 3⇑), were unable to elicit IL-2 production when exposed to Ag (Fig. 3⇑). These results, in fact, concur with those of others who have described a requirement for Src family tyrosine kinases in the activation of Zap-70 in both Jurkat cells (24) and heterologous systems (47, 48). Stimulation of scFv-Zap70 or scFv-Syk receptors with Ag bypasses the initial activation of Src family kinases, an event that is apparently essential for signal propagation through Zap-70. It has recently been suggested that this dependence involves a transient association between the Src family kinase and Zap-70, as well as their functional cooperativity in the phosphorylation of substrates such as HS1 or Cbl (48, 49). The inability of aggregated chimeric Zap-70 molecules to initiate T cell functions (Fig. 3⇑) (24) may be due to the lack of Src family kinase help in the phosphorylation of these or other substrates. In contrast, we have previously shown that activation of chimeric Syk molecules causes an increase in Syk-Cbl association and a rise in Cbl phosphorylation in a manner that is apparently Src kinase independent (42). Others have also demonstrated that Syk is not strictly dependent on Src family kinases for its activation in Jurkat T cells (15, 24). Our data thus support the idea that Syk, by virtue of its superior kinase activity (29) and constitutive association with the TCR (15), may be in a position to initiate T cell signals without the prior activation of Src family kinases.
Receptor aggregation, rather than dimerization, is critical for T cell activation through endogenous TCR/CD3 receptors (50, 51). The same requirement exists when chimeric scFv-ζ or scFv-γ have been used for signaling (2). The highly homologous CD3 ζ-chain and Fcε γ-chain both exist as disulfide-linked homodimers in their respective receptor complexes due to transmembrane cysteine residues. Chimeric receptors based on these molecules are also free to form homodimers or heterodimers with the cell’s endogenous chains (2). In the construction of chimeric receptors with cytosolic PTKs, it was necessary to attach the extracellular Ab and intracellular signaling domains via a transmembrane segment. We exploited this need by incorporating dissimilar transmembrane segments, which can exist as monomers (CD4) or dimers/multimers (CD8α), and studying their signaling potential.
Our results indeed emphasize the importance of the transmembranal and hinge moieties in this chimeric receptor design. Interestingly, while both CD4 transmembrane (S4)-containing and CD8 transmembrane and hinge (S8/H)-containing receptors could trigger IL-2 secretion in response to immobilized protein Ag, only the latter were stimulated when incubated with Ag-modified target cells (Fig. 4⇑). Although CD8-based receptors were, in general, more highly expressed than those with CD4 sequences, comparison of several transfectants expressing similar amounts of cell surface receptors indicates that their phenotypic disparity is not due solely to quantitative differences (data not shown). After generating additional chimeric scFv-Syk receptors (Fig. 1⇑B), we were able to map this differential response to the hinge region of CD8α (Fig. 6⇑). Furthermore, after mutating or adding cysteine residues to two receptors (Fig. 1⇑C), we concluded that hinge elements other than disulfide bonds are essential for its effect (Fig. 8⇑).
The membrane-proximal hinge region of CD8α is heavily decorated with O-linked glycosylation, and these sugars are extensively sialylated (52). These two features are responsible for the extended and flexible structure of native CD8α in its reach for class I MHC molecules (53). By incorporating hinge sequences into our chimeric receptors, we may have conferred upon them a similar extended topology. This is in agreement with our ability to detect the scFv by flow cytometry with anti-Id Ab only on cells expressing receptors with hinge moieties (Figs. 2⇑ and 5⇑A and data not shown). These correctly oriented or extended receptors are also those that can respond to cell-derived Ag, most likely because of this same far-reaching configuration. Supporting this theory is our finding that only those receptors with extended hinges show efficient rosetting of TNP-coated SRBC (data not shown). Similarly, in the construction of single-chain FcεRIγ or ζ-chain receptors, it was found that a membrane-proximal spacer region has a crucial role in their interaction with cell-bound Ag, most likely via its ability to correctly orient the receptors on the cell surface (7, 54). However, this same requirement for a hinge region for recognition of cell-bound Ag is less obvious when the anti-TNP scFv is connected to ζ or γ signaling molecules (C. J. Fitzer-Attas and T. Waks, unpublished data). Apparently, attachment via the native transmembrane portion of ζ or γ in these cTCRs is sufficient for this type of activation.
Contrary to these results, in the response of scFv-Syk chimeras to another form of Ag, plastic-immobilized protein, the hinge is not a necessity, since CD4 transmembrane-derived receptors functioned quite well (Fig. 4⇑). Yet, interestingly, the CD8 transmembrane portion alone is not able to transmit any signals upon stimulation (Fig. 6⇑) despite effective surface expression (Fig. 5⇑B). Thus, the composition of the transmembrane region is also a critical element for receptor activation, and in the construction of kinase-based chimeric molecules it is essential to evaluate the effect of individual transmembrane segments on signaling. Most importantly, the effector functions stimulated by kinase chimeras upon recognition of cell-derived Ag are absolutely dependent on the presence of a connecting hinge region. These studies should assist in development of the optimal chimeric T cell transducing molecule for redirection and activation of unresponsive T lymphocytes in the cancer patient.
Footnotes
-
↵1 This work was supported by Baxter Healthcare Corp., in part by research grants from The Charles and David Wolfson Charitable Trust and The Forchheimer Center for Molecular Genetics at The Weizmann Institute of Science, and a Fellowship Grant from the Israel Cancer Research Fund (to C.J.F.-A.).
-
↵2 Present address: G. W. Hooper Foundation, University of California, San Francisco, CA 94143.
-
↵3 Address correspondence and reprint requests to Dr. Zelig Eshhar, Department of Immunology, The Weizmann Institute of Science, Rehovot 76100, Israel.
-
↵4 Abbreviations used in this paper: cTCR, chimeric T cell receptor; scFv, single chain of the variable domain; PTK, protein tyrosine kinase; TNP, 2,4,6-trinitrophenyl; TNP-FγG, 2,4,6-trinitrophenyl modified fowl γ-globulin.
- Received June 30, 1997.
- Accepted September 22, 1997.
- Copyright © 1998 by The American Association of Immunologists
References
In this issue
