WO1997035004A1 - Cell stimulation - Google Patents

Abstract

Disclosed is a chimeric polypeptide for display on a cell surface, comprising an effective portion of the CD28 intracellular signalling domain, a transmembrane domain and an extracellular antigen-specific binding domain of an immunoglobulin.

Description

Title: CELL STIMULATION

Field of the Invention

This invention relates to nucleic acid sequences encoding chimeric polypeptides, cells comprising the nucleic acid sequence and to the chimeric polypeptides.

Background of the Invention

T cells recognise antigen as peptide fragments presented in the context of major histocompatibility complex (MHC) molecules at the surface of cells (1, 2). However, while T cell antigen receptor (TCR) signal transduction, may be sufficient to activate antigen-primed T cells (3), TCR ligation alone is not sufficient to initiate an immune response under most circumstances (4). Indeed it may result in T cell unresponsiveness or death (5, 6). For optimal activation of naive T cells, subsequent autocrine-driven clonal expansion and induction of effector functions, additional or costimulatory signals are needed (7). Recent work has demonstrated that a key "costimulatory" signal is provided by interaction of the T cell surface molecule CD28 with members of the B7 family (B7-1, also known as CD80; B7-2, also known as CD86) on antigen presenting cells (APC) (for review see ref. 8). This second signal has no effect alone but enhances the effectiveness of stimulation via the TCR/CD3 complex (9, 10). An important mechanism of action of CD28-mediated costimulation is to promote cell cycle progression by enhancing interleukin-2 (IL-2) production, and regulating programmed cell death (11 , 12). The CD28 costimulatory pathway also seems to be important in the proliferation of NK cells (13). Assessing the role of CD28-mediated costimulation is thus clearly important for our understanding of the immune system as well as for the design of new strategies for immunotherapy.

Costimulatory molecules are being investigated to make modified tumour cell vaccines. A vaccine for cancer has long been a goal of immunotherapy but most cancer cells are poorly immunogenic because the differences from normal tissues are small and because of defects in the antigen presentation pathways of tumour cells (14); the lack of costimulation may further promote the ability of already weakly immunogenic tumours to "sneak through" and evade the immune system at an early stage of their development. The absence of costimulatory molecules may even contribute to the degree of immune tolerance of the tumour (15). Certainly in animal models the presence of B7-1 transfected into tumour cells enhances the immunogenicity of weakly immunogenic tumours (16, 17) although it has no effect on the immunogenicity of non-immunogenic tumours (18). The cellular basis for the rejection of genetically modified tumour cells expressing the B7-1 molecule, and the induction of immunity to the parental tumour appears to be largely a result of direct activation of CD8 + T cells (16, 19, 20) although NK cells are also thought to be important (21).

However such methods have some drawbacks. First, tumour specific CD4+ T cells which might provide a source of exogenous help to sustain cytotoxic T lymphocyte (CTL) responses are unlikely to be generated using this approach (22). Second, another regulatory molecule, CTLA-4, can also bind to B7 molecules and has a 10- to 20-fold higher affinity than CD28 (23). CTLA-4 has been shown to be a critical negative regulator of T cell activation (24-26) and thus, when only small amounts of B7 are expressed, CTLA-4, by virtue of its higher affinity, might be dominant over CD28 (24). Adoptive T cell therapy is an alternative approach and has been used in many diseases - most notably for the therapy of viral (27) or malignant disease (28). Frequent loss of MHC is a problem (14) and chimeric T cell receptors can be used to overcome this problem (29). However, in general the target cells will not express B7 counter-receptors and thus will not be able to provide optimal costimulatory signals for T cell activation and proliferation.

WO 96/13584, published after the priority date of the present application, discloses CTL expressing a chimeric receptor molecule which enables the CTL to become activa_ted by antigen presenting cells which would not normally activate said CTL. This is accomplished by fusing the intracellular signalling domain of CD28 with a transmembrane domain, and an extracellular domain which is derived from a receptor for a ligand that is expressed on the surface of the antigen presenting cell. The specification gives LFA-1 and CD2 as examples of the receptor molecules which can contribute the extracellular domain of the chimeric receptor.

WO 96/23814, also published after the priority date of the present application, claims chimeric co-stimulatory receptor proteins which comprise a single polypeptide having an extracellular ligand binding domain, a transmembrane domain and a cytoplasmic co-stimulatory effector function signalling domain. CD28 is mentioned as one possible source of cytoplasmic signalling domain.

The work leading to the present invention has been published (after the priority date ofthe present application) by Alvarez- Vallina & Hawkins (1996 Eur. J. Immunol. 26, 2304-2309).

Summary of the Invention

In a first aspect the invention provides a nucleic acid sequence encoding a chimeric polypeptide for display on a cell surface, the chimeric polypeptide comprising an effective portion of CD28 intracellular signalling domain, a transmembrane domain and an extracellular antigen-specific binding domain of an immunoglobulin molecule or fragment thereof.

In a second aspect the invention provides a nucleic acid construct, comprising the nucleic acid sequence defined above, the construct being adapted for expression in a eukaryotic cell. In a third aspect the invention provides a chimeric polypeptide molecule, encoded by the nucleic acid sequence defmed above.

The "effective portion" of the CD28 molecule will possess at least the signal activity associated with the intracellular portion of the wild type CD28 molecule and will preferably also comprise the transmembrane poπion of the wild type CD28 molecule, although those skilled in the art will appreciate that the transmembrane portion from some other receptor, or even a transmembrane "consensus" portion, may suffice to locate the chimeric polypeptide within the membrane of a cell in which it is produced. Conveniently, all or most of the extracellular domain of the wild type CD28 molecule will be absent from the chimeric polypeptide, thus accommodating the antigen-specific immunoglobulin binding domain.

The immunoglobulin binding domain preferably has binding affinity for a disease marker, such as a tumour-associated antigen or a viral polypeptide or peptide. Desirably the immunoglobulin binding domain may be a single chain variable fragment (scFv) or, less preferably, other antibody fragment such as Fab, Fab₂, Fv etc.

Preferably the chimeric polypeptide will comprise a spacer region between the immunoglobulin binding domain and the transmembrane domain which anchorsd the protein in the cell membrane. Conveniently, this spacer region may be provided by a fragment of the CD28 extracellular domain. The inventors have found that a spacer region of about 18-25 amino acid residues, preferably about 19-21 residues, confers unexpectedly improved expression characteristics (as judged by FACS analysis) on the chimeric polypeptide. In turn, this appears to provide optimal, or near optimal, co-stimulatory functional characteristics on the chimeric polypeptide.

In CD28 the extracellular end of the transmembrane domain is located at residue 135. Thus, where the spacer region is provided by CD28, the spacer region will generally comprise at least residues 135-117 (numbering from 1 at the N terminal of CD28). In particular, a preferred embodiment of the invention encodes a polypeptide comprising residues 135-114 of CD28. Some of the features of this chimeric protein are described below.

This portion (residues 135-114) of the CD28 extracellular domain has certain structural properties which may be important for optimal expression and/or co¬ stimulatory activity: it comprises a cysteine residue (position 123) which is thus available for the formation of disulphide bonds, and comprises an α-helical region (residues 122-114) which allows for lateral spacing between the attached immunoglobulin binding domains (which may be required for efficient folding thereof). Those skilled in the art will appreciate that amino acid sequences other than those derived from wild-type CD28 may possess similar structural motifs and thus also have similar desirable properties. For example, conservative amino acid substitutions of the CD28 sequence may be made without disrupting the desirable structural features. Further, homology with other cell surface receptor molecules (such as CD8) may indicate similar amino acid sequences which have similar structural properties. Routine variation of the α-helical portion may be attempted in order to further optimise the functional characteristics of the chimeric protein: for example, one or two extra turns of α-helix may be inserted so as to increase the lateral spacing bewteen the attached immunoglobulin binding domains, if desired. The CD28 molecule naturally forms dimers, but the attached immunoglobulin domains may well not be naturally dimeric (as with scFvs), so that such lateral spacing is important for maximum activity.

A further preferred feature is the inclusion in the chimeric polypeptide of a short peptide linker region between the immunoglobulin binding domain and the spacer region. Typically this will comprise a short flexible peptide of 3-10 amino acid residues, more typically 3-5 residues. Specific examples include an alanine tripeptide, described in the examples below. However, greater flexibility may be conferred by a glycine tripeptide, or a proline tripeptide. In practice, a glycine and/or proline peptide may introduce to great an amount of flexibility and linkers comprising threonine and/or serine may be preferred.

It is also preferred that the nucleic acid sequence will also encode a leader sequence, so as to direct the expressed chimeric polypeptide into the endoplasmic reticulum (ER) and hence to the plasma membrane of a cell in which it is produced. Generally the cell is a eukaryotic cell (e.g. a yeast cell), and desirably a mammalian cell (especially a human cell), preferably an immunocompetent cell, as exemplified below. A number of suitable leader sequences are known to those skilled in the art including, for example, the human V_H1 leader sequence, as utilised in the examples below.

In a fourth aspect, the invention provides a eukaryotic cell into which has been introduced a nucleic acid sequence in accordance with the first aspect of the invention, or the progeny of such a cell. Preferably the cell will be an immunocompetent cell, such as a lymphocyte, or a precursor thereof and advantageously will be a human cell. Typically the cell will be a T cell (especially a cytotoxic T lymphocyte) or natural killer (NK) cell.

In one embodiment, the cell into which the nucleic acid sequence is introduced will also express (before or after introduction of the nucleic acid sequence of the invention) one or more further chimeric polypeptides. Typically these will be chimeric cell surface molecules comprising an extracellular antigen-specific immunoglobulin binding domain fused with an intracellular signalling domain. Conveniently the signalling domain will be derived from a molecule which modulates the immunocompetent status of the cell, such as CD3, CD4 or CD8 and the like. The immunoglobulin binding domains of the one or more further chimeric polypeptides may have binding affinity for the same antigen as the chimeric polypeptide encoded by the introduced nucleic acid sequence, or they may have binding affinity for one or more different antigens. Alternative chimeric molecules which may also be present in the cell are disclosed in WO 96/13584 and WO 96/23814.

In a fifth aspect, the invention provides a method of treating a human patient, the method comprising introducing into the patient a nucleic acid sequence in accordance with the invention. This may be done by introducing the nucleic acid directly into the patient by "gene therapy", but may be accomplished more conveniently by introducing into the patient cells already comprising the nucleic acid sequence. Typically these will be immunocompetent cells, and will preferably be autologous peripheral blood lymphocytes (PBL), conveniently previously obtained from the patient. Further useful teaching as to methods relevant to application of the present invention is found in WO 96/13584 and WO 96/23814, the content of which is incorporated herein by reference. For example, WO 96/13584 discloses at pages 28-34 methods of making nucleic acid constructs, their introduction into suitable host cells and selection thereof, whilst page 36 (lines 4-21) of WO 96/13584 discloses methods of infusion of patients with doses of cloned human lymphocytes.

The invention will now be further described by way of illustrative example and with reference to the accompanying drawings, which are as follows:

Figure 1 is a schematic representation of chimeric scFv constructs (scFv CD28 above, scFvCD3f below). The boxes from left to right represents DNA segments corresponding to human VH1 leader sequence (■) heavy chain variable region (VH), linker (diagonal striped box), light chain variable region (VL) and the truncated human CD28 molecule (dotted box) (starting from amino acid 124, numbering according to ref. 37) or the truncated human CD3f molecule (dotted box) (starting from amino acid 19, numbering according to ref. 38). Sequences (Seq ID No.s 1 & 2) displayed show the junctions between the murine scFv domain (small letters) and the human CD28 or CD3f molecules (capital letters). The first amino acids corresponding to the transmembrane region (TM) are underlined. The last three amino acids of the predicted CD3f leader sequence are boxed.

Figures 2A-2C show FACS analysis of transfected cells illustrating expression of the chimeric gene products on the cell surface. The parental (Jurkat, Fig. 2A) and transfected (JNIPCD28 and JNIPCD3f, Figs. 2B & 2C) cell lines were analysed by flow cytometry after staining with goat antisera to murine λ-light chain FITC-conjugated (solid line). Negative controls (broken line) are overlaid. The fluorescence channel number is plotted along the X-axis, and the y-axis represents the relative cell number.

Figures 3 A-C show IL-2 production by parental and transfected Jurkat cells. Approximately 5xl0⁴ Jurkat (A), JNIPCD28 (B) or JNIPCD3f (C) cells were cultured in the presence of PMA (10 ng/ml) and the indicated stimuli, in medium alone (unshaded bar) or with plastic immobilised NIP₁₀-BSA (shaded bar). Stimuli were used as follow: ionomycin, lμM; anti-CD3e (YTH12.5HL) ascitic fluid, 1 :2,000 final dilution; purified anti-CD28 (YTH913.12), 2.5 μg/ml. MAbs were further crosslinked by addition of polyclonal rabbit anti-rat IgG (20 μg/ml). Cells were cultured in triplicate and after 20 hours cell free supernatants were assayed for IL-2 content. Data are from one representative experiment out of five.

Figure 4 is a graph demonstrating quantitative assessment of costimulation in response to various stimuli. Approximately 2.5xl0⁴ cells were stimulated in the presence of PMA (10 ng/ml) and ionomycin (1 μM) with different concentrations of soluble NIP₁₀-BSA conjugates (solid line with shaded boxes). The level of IL-2 secreted was compared with obtained in JNIPCD28 cells stimulated with PMA and ionomycin alone (solid line without boxes) or in the presence of anti-CD28 mAb (dashed line) in solution at 2.5 μg/ml (S-anti-CD28), further crosslinked (SC-anti-CD28) by addition of polyclonal rabbit anti-rat IgG (20 μg/ml) or plastic immobilised (ianti-CD28). Cells were cultured in triplicate, and SEs were less than 10% at every point. One of three separate experiments is shown.

Figures 5A & 5B are FACS analysis plots showing cell surface expression of ύie chimeric gene products in double transfected Jurkat cells. Flow cytometry analysis of JNIPCD3f/OxCD28 cells stained with goat anti-mouse λ-light chain FITC-conjugated (Fig. 5A; solid line) or phOX_{[ 8}-BSA conjugates followed by streptavidin-conjugated-FITC (Fig. 5B; solid line). Negative controls (broken line) are overlaid (x-axis, fluorescence intensity; y-axis, relative cell number).

Figures 6A and 6B show signal transduction by double transfected Jurkat cells (JNIPCD3T/OxCD28). In double transfected Jurkat cells (JNIPCD3$7OxCD28) chimeric CD3f and CD28 are functional. Fig. 6 A is a bar chart: Approximately 5x10" JNIPCD3f/OxCD28 cells were cultured in triplicate in the presence of PMA (10 ng/ml) and the indicate stimuli (see Figure legend 3 for details) with medium alone (unshaded bar), plastic immobilised NIP₁₀-BSA (black bar) or plastic immobilised phOx₁₀-BSA (diagonal striped bar). The results shown are from one representative experiment out of three. Fig. 6B shows a graph of results obtained when JNIPCD3f/OxCD28 cells (5xl0⁴) were stimulated with PMA and different concentrations of soluble phOx, ₈-BSA conjugates in the presence of plastic immobilised NIP₁₀-BSA (shaded circle) or soluble crosslinked anti-CD3e mAb (unshaded box). Dashed lines represent the level of IL-2 secreted by JNIPCD3f/OxCD28 cells stimulated in the presence of PMA with soluble crosslinked anti-CD3e mAb, plastic immobilised NIP₁₀-BSA or soluble crosslinked anti-CD3e and anti-CD28 mAbs. SEs were less than 10% of the mean. One of two similar experiments is shown.

Figure 7 is a schematic representation comparing three nucleic acid constructs (J28, J28IIH and J28EKS) in accordance with the invention;

Figures 8A, 8B and 8C are graphs showing the results of FACS (fluorescence- activated cell sorting) performed on cells transfected with a nucleic acid construct in accordance with the invention (A = J28; B = J28EKS; and C = J281TH);

Figures 9A, B and C are graphs showing the amount of IL-2 produced by cells in response to stimulation of chimeric CD28 molecules, in accordance with the invention, expressed on their surface (9A = J28; 9B = J28EKS; 9C = J28IIH); and

Figure 10 shows the DNA and amino acid sequence of expressed CD28.

Examples

Materials and Methods

Cells. Antibodies and Antigens. The Jurkat T cell line (clone E6-1) (Ref. 30) was kindly provided by R. Paneil (Medical Research Council, Laboratory of Molecular Biology, UK). Cells were maintained in RPMI 1640 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin and 25 mM HEPES buffer (all from GIBCO-BRL. Gaitersburg, MD), referred to as complete medium (CM). The monoclonal antibodies used included YTH12.5HL (rat IgG2b), and YTH913.12 (rat IgG2b) (Serotec Ltd, Oxford, UK) specific for human CD3e and CD28 molecules respectively. For direct staining the FITC-conjugated monoclonal antibodies UCHT-1 (anti-CD3e) (mouse IgGl) and YTH913.12 (Serotec) were used. Goat polyclonal antisera to mouse lambda-light chain (FITC-conjugated) was from Southern Biotechnology Associates, Inc. (Birmingham, AL). The optimal titre of all reagents was assessed before their use. Bovine serum albumin (BSA) was conjugated with 4-hydroxy-5-iodo-3-nitrophenyl acetyl (NIP) (Cambridge Research Biochemicals, Northwich, UK) in a molar ratio of 10: 1 (NIP,o-BSA) or with 2-phenyl-2-oxazolin-5-one (phOx) (Sigma Chemical Co., St. Louis, MO) in molar ratios of 10: 1 (phOx₁₀-BSA) or 1.8: 1 (phOx,.₈-BSA), as previously described (31, 32): The phOx_{I 8}-BSA conjugate was subsequently biotinylated using a kit (Amersham International, Amersham, UK) as specified by manufacturers.

Construction ofscFv Fusion Genes. Two single chain Fv (scFv) antibody fragments derived from the hapten-specific mAbs B1.8 (mouse IgGlλ; anti-NIP) (33) and NQ10/12.5 (mouse IgGlk; anti-phOx) (34) were used. Both scFvs were subcloned as Pst l/Not I fragments into the vector pVACl (35) containing a human V_H1 leader sequence (36) for eukaryotic expression. This vector was used for subsequent engineering.

The human CD28 and CD3f cDNAs were derived from mRNA (Oligotex, Qiagen Gmbh, Germany) of the Jurkat cell line by reverse transcription using a oligo (dT),₈ primer, followed by PCR. The truncated human CD28 molecule (residues 124-202) (37) was constructed by PCR with a 5' oligonucleotide (Seq ID No. 3) (ATATAGCGCGGCCGCTCCAAGTCCCCTATTTCCCGGACC) which introduced a Not I site (underlined) eleven amino acids distal to the predicted transmembrane r e g i o n a n d a 3 ' o l i g o n u c l e o t i d e ( S e q I D N o . 4 ) (TACGTCTAGATTATTAGGAGCGATAGGCTGCGAAGTCGC) which introduced an Xba I site (underlined) just after the CD28 stop codon. The final CD28 construct was created by ligation of the Not \IXba I cleaved PCR fragment downstream of both the scFv domain of B1.8 and NQ10/12.5 to form pVACl/NIPCD28 and pVACl/phOxCD28 plasmids.

The sequence of the truncated CD28 molecule was verified using a Taq DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA) and an Applied Biosystems 373 Automated DNA sequencer. The same strategy was used to construct the scFv-CD3f fusion gene. The 5 '-oligonucleotide (Seq ID No. 5) (ATATAGCC GGCCOCCTACAGAGGCACAGAGCTTTGGCCT) introduced a NotI site twelve amino acids distal to the transmembrane region (38).

T h e 3 O l i g o n u c l e o t i d e ( S e q I D N o . 6 ) (TACGICTAGATTATTAGCGAGGGGGCAGGGCCTGCATGT) introduced an Xbal site just after the CD3f stop codon. A diagram of the fmal scFv fusion construct is shown in Figure 1. The final scFv fusion cassette was subcloned as .Xføl-blunted / Hindlll fragments into the Bo/nHI-blunted/HZ/idL-I backbones of the episomal expression EBV-based vectors pCEP4 (Cytomegalovirus immediate early Enhancer/Promoter) and pREP9 (Rous Sarcoma Virus long terminal repeat Enhancer/Promoter) (obtained from Invitrogen Corporation, San Diego, CA) carrying respectively the hygromycin and neomycin genes as selectable markers. The final plasmids were termed pCEP4/NIPCD28, pREP9/phOxCD28 and pCEP4/NIPCD3f.

Cell Transfections. Jurkat cells were transfected with lOμg of pCEP4/NIPCD28 or pCEP4/NIPCD3r using Lipofectin (GIBCO-BRL) as previously described (39). Transduced cells were cultured overnight in CM and then selected for resistance to hygromycin B (Calbiochem, La Jolla, CA), initially at lOOμg/ml for another overnight period. The concentration of hygromycin B was then increased in lOOμg steps every 24 hours up to 400μg/ml. Hygromycin-resistant cells were visible at 2 weeks and screened by FACS analysis for stable expression of the relevant construct. Cloned cell populations were obtained by limiting dilution. To generate double gene transfectants, cells expressing the chimeric B1.8-CD3 molecule (JNIPCD3f) were transfected, using the conditions described above, with 10 μg of pREP9/phOxCD28. After transfection, cells were grown for 2 days in CM before plating out in CM supplemented with 400 μg/ml of hygromycin B and 2 mg/ml of G-418 (GIBCO-BRL).

Cell Surface Analysis. Cell surface staining was performed by standard direct immunofluorescence as described (40) using saturating amounts of FITC-conjugated antibodies. To study the expression of chimeric NQ10/12.5-CD28 molecules cells transfected with pREP9/phOxCD28 were incubated with biotinylated phOx, _g-BSA conjugates at 10 μg/ml for 30 minutes at 4°C. After two washes the cells were incubated with streptavidin-conjugated FITC (Sigma) at an appropriate dilution for 30 minutes at 4°C. The cells were washed three times and resuspended in PBS- 1 % FCS with 0.5 % paraformaldehyde, and 0.02% NaN₃. Propidium iodide (PI) (lμg/ml) was included in the final cell suspension to exclude dead cells and debris on the basis of forward scatter and PI staining. Appropriate FITC isotype-matched irrelevant Abs were used in all experiments. The samples were analysed with a FACScan^(R) (Becton Dickinson, Mountain View, CA). A minimum of 10,000 cells was analysed for each sample. Subsequent re-analysis was performed using the CELLQuest software (version 1.2) (Becton Dickinson).

IL-2 Release Assay . Cells (5 or 2.5 x 10⁴/well) were incubated with pharmacological agents such as phorbol 12-myristate 13-acetate (PMA) (10 ng/ml)(Sigma) either alone or with ionomycin (lμM) (Calbiochem) and the effects of receptor specific stimuli investigated. Where indicated, wells were coated for 2 hours at 37 °C with 100 μl of a 100 μg/ml solution of NIP₁₀-BSA or phOxi₀-BSA in PBS, washed 3 times with PBS and then blocked for 1 hour at 37 °C with 200 μl of CM. Plates where coated in the same conditions with anti-CD28 mAb at 10 μg/ml in PBS. Soluble anti-CD28 mAb was added at 2.5 μg/ml, and anti-CD3e was added at a final 1:2,000 dilution of ascites. Polyclonal rabbit anti-rat IgG (mouse adsorbed) (Serotec) was used at 20 μg/ml. Soluble NIP_I0-BSA or phOx₁₀-BSA conjugates were added at the specified concentration. All additives were included at the start of the culture period in a final culture volume of 200 μl. Round-bottomed 96-well Microtiter plates (Corning, New York, NY) were used for all assays. The plates were incubated at 37°C in 5% C0₂. After 20 hours cell-free supernatants were harvested and assayed for IL-2 activity using a commercially available ELISA (Genzyme Diagnostics, Cambridge, MA).

Results

Example 1

Design of Chimeric scFv Gene Constructs. A chimeric gene consisting of an antigen specific recognition unit and a truncated CD28 molecule was constructed (Fig. 1). The recognition function is contributed by a scFv domain derived from the hapten-specific monoclonal antibodies either B1.8 (anti-NIP) (Ref. 33) or NQ10/12.5 (anti-phOx) (34). The transmembrane and cytoplasmic domain of the fusion gene are contributed by the CD28 molecule. To distance the scFv from the cell membrane we left the last eleven amino acids of the predicted CD28 extracellular region intact (Fig. 1) (Ref. 37). However, the construct does not include the membrane-proximal cysteine residue (position 123) thought to be involved in the formation of intermolecular disulphide bonds (8). We have also constructed chimeric NlP-specific scFv molecules incorporating the f chain of the TCR/CD3 as signalling component, including all the extracellular domain and the last three hydrophilic aminoacids (TEA) of the predicted leader sequence (Fig. 1) (Ref. 38).

Expression of the Chimeric scFv Genes. Selection of Jurkat cells stably transfected with the CD28 and CD3f chimeric constructs (JNIPCD28 and JNIPCD3f respectively) resulted in the expression of the scFvBl .8 derived fusion molecules on the cell surface, as shown by flow cytometry using a FITC-conjugated polyclonal goat anti-mouse lambda chain antisera (Fig. 2). In the resultant uncloned hygromycin-resistant cell populations, the scFvB1.8-CD28 molecule was expressed with a more heterogeneous pattern, and with a lower average level, than the scFvB1.8-CD3f (Fig. 2). The transfected Jurkat cells showed a normal pattern of expression of TCR/CD3 (as judged by staining with an anti-CD3e mAb) and CD28 wild-type molecules (data not shown). JNIPCD28 and JNIPCD3J^" uncloned cells were cultured with continued selection for more than four months with no apparent change in the level of B1.8 expression or function over this time suggesting very stable expression of these episomal vectors in human T cell lines.

Assessing the Signal Transduction Pathways in Cells Bearing Chimeric Receptors. IL-2 production in response to various stimuli was determined to assess whether the expressed chimeric scFvCD28 and scFvCD3f molecules can function as active receptor molecules in a physiological manner and if they can act synergistically with signals mediated through the TCR/CD3 and/or CD28 wild-type receptors. We studied in detail the ability of the transduced cells to express the IL-2 gene in response to antigen specific stimulation and/or crosslinking with anti-CD3e and/or anti-CD28 monoclonal antibodies. Representative data are shown in Figure 3. It is interesting to note that in our system treatment with PMA and anti-CD28 mAb (soluble or crosslinked) does not result in a significant levels of IL-2 secretion. This seems to vary with different sublines of Jurkat and several differences in the signal transduction requirements for IL-2 expression in variants have been noted (8).

In the absence of antigen, JNIPCD28 cells showed a pattern of response similar to non-transduced Jurkat cells (Fig. 3 A and 3B). CD28 triggering with anti-CD28 mAb in solution (S-anti-CD28) or further crosslinked on the cell surface (SC-anti-CD28), by addition of mouse adsorbed polyclonal rabbit anti-rat IgG, enhances IL-2 secretion five to sevenfold in response to stimulation with PMA and ionomycin or anti-CD3e mAb (41) (Fig. 3 A and 3B). Thus the expression of scFv-CD28 molecules does not interfere with the normal CD28 (and TCR CD3) signalling pa way which remain intact in JNIPCD28 cells. In the absence of TCR signalling stimulation of JNIPCD28 cells with plastic-immobilised NHVBSA (iNIP₁₀-BSA) had no effect on IL-2 secretion. However, when combined with PMA and SC-anti-CD3e mAb or ionomycin. iNIP₁₀-BSA enhanced IL-2 secretion, although only about two- to three-fold (Fig. 3B). These results suggested the signal elicited through chimeric CD28 (chCD28) molecules, although qualitatively similar, might be somewhat reduced compared to that elicited through wild-type CD28 (wtCD28) receptors.

To assess the quantitative signalling ability of chCD28 receptors in more detail, JNIPCD28 cells were stimulated with soluble NIP₁₀-BSA conjugates, using a broad range of concentrations, in the presence of PMA and ionomycin. The production of IL-2 in response to these stimuli is represented in Fig. 4.

Optimal levels of costimulation were obtained at low concentrations of multivalent antigen (6.25-100 ng/ml). In these conditions, the enhancement of IL-2 secretion was similar to that observed after stimulation of the wtCD28 molecule with S-anti-CD28 or SC-anti-CD28 mAb (Fig. 4). Furthermore, stimulation of Jurkat (not shown) or JNIPCD28 cells (Fig. 4) with plastic immobilised anti-CD28 mAb in the presence of PMA and ionomycin, revealed that IL-2 expression is comparable to that produced in response to iNIP₁₀-BSA. These results suggest that the reduced CD28 signalling function observed in JNIPCD28 cells, after antigen specific stimulation with plastic-immobilised NIPι₀-BSA probably results from engagement of insufficient chCD28 receptors. At high levels of multivalent antigen a reduced amount of IL-2 is produced by JNIPCD28 cells stimulated with PMA and ionomycin (Fig. 4). This is presumably a result of reduced crosslmking and receptor clustering (42).

The results indicate that in conditions of optimal scFv-CD28 crosslinking (i.e. at appropriate concentrations of soluble multivalent antigen) there is a clear correlation between the intensity of CD28-mediated signalling and the antigen concentration (Fig. 4). Interestingly, from the threshold level there is a very steep rise to a plateau of maximal IL-2 secretion. Engagement of additional CD28 receptors did not result in a further increase in the level of costimulation (Fig. 4). This is further supported by the finding that after combined stimulation of chimeric and unmodified CD28 molecules in JNIPCD28 cells, with iNIP₁₀-BSA and SC-anti-CD28 mAb, the response was not greater than that to SC-anti-CD28 mAb alone (Fig. 3B).

The inventors also assessed the functional ability of the scFvB(1.8)-CD3f expressing Jurkat cells to produce IL-2 in presence or absence of i.e. iNIP_l0-BSA conjugates. In the absence of antigen, JNIPCD3f cells showed a pattern of response similar to non-transduced Jurkat cells (Fig. 3 A and 3C). However the response to PMA and ionomycin was 4-fold less in JNIPCD3f cells than in control Jurkat cells. Likewise the level of IL-2 secretion after stimulation with PMA and SC-anti-CD28 mAb, was reduced in the presence of ionomycin compared to that obtained in the same conditions with SC-anti-CD3e mAb (Fig. 3C). This observation was reproducible in all experiments and suggests an impaired response to calcium ionophores in JNIPCD3f cells. The significance of this difference is not clear and further studies are needed to clarify this apparently defective function. Stimulation of JNIPCD3J" cells with plastic immobilised NIP₁₀-BSA conjugates in the presence of PMA induced IL-2 secretion that was similar to that observed in untransfected Jurkat cells or JNIPCD3J- cells in response to SC-anti-CDe mAb and PMA (Fig. 3 A and 3C). Combined stimulation of JNIPCD3f cells with iNIP₁₀-BSA and SC-anti-CD3e mAb in response to PMA resulted in enhancement of IL-2 production compared to that obtained with iNIP₁₀-BSA or SC-anti-CD3e mAb alone (Fig. 3C), indicating that chimeric CD3f-transduced signals represent an additional stimulus to that using anti-CD3e mAbs alone. Most importantly this antigen specific signal synergizes with signals mediated by the wild-type CD28 molecule for optimal IL-2 production (Fig. 3C).

The results given are for the uncloned population but similar results were also derived from individual clones obtained after limiting dilutional cloning (data not shown). Finally, plastic immobilised or soluble phOx₁₀-BSA did not activate JNIPCD28 or JNIPCD3f cells (not shown) thus demonstrating the specificity of the response towards NIP. Non- transduced parental Jurkat cells could not be stimulated to secrete IL-2 by hapten-protein conjugates (Fig. 3A). Likewise the rabbit anti-rat IgG antisera (mouse adsorbed) used in crosslinking assays does not crossreact with the murine hapten-specific domains used (B1.8 and NQ10/12.5) (not shown).

Simultaneous Expression of scFv-CD28 and scFv-CD3ζ Fusion Proteins. To investigate whether the chimeric receptors could mimic the binary nature of the lymphocyte activation process, and thus to probe the possibility of inducing full activation of T cells in an antigen specific manner, we generated double transfectants. These cells simultaneously express chimeric CD28 and chimeric CD3 ξ molecules on the cell surface. In order to allow assessment of the contribution by each chimeric molecule to the T cell activation process we employed two different antigen specific scFv domains B1.8 (anti-NIP) and NQ10/12.5 (anti-phOx). JNIPCD3f cells were transfected with a chimeric CD28 construct in which the antigen-binding domain was derived from an anti-phOx specific mAb (34).

As shown in Figure 5 double transfectants resistant to both hygromycin and G418 (JNIPCD3f/OxCD28) expressed the chimeric B1.8-CD3f at similar levels to the parental single transfectant JNIPCD3f cells. The NQ10/12.5-CD28 chimeric molecule was expressed in a functional form, as judged by using biotinylated phOxi ₈-BSA conjugates and streptavidin-conjugated FITC (Fig. 5). The staining pattern was heterogeneous as in JNIPCD28 cells (Fig. 2 and 5). Likewise the expression of TCR/CD3 and CD28 wild-type molecules was unaffected (data not shown). The cells showed a normal growth pattern and were stable, without the need for selective subcloning, for a long period.

In absence of the antigens, uncloned JNIPCD3f/OxCD28 cells showed a response pattern similar to JNIPCD3f cells (Fig. 3C and 6A). Stimulation with iNIP₁₀-BSA conjugates elicited a TCR/CD3-like response as in the parental single transfected cells (Fig. 6A). Stimulation with iphOx₁₀-BSA conjugates induced a signal qualitatively similar to that elicited by crosslinking of the wtCD28 molecule (Fig. 6A). As observed in the single transfected JNIPCD28 cells, maximal levels of costimulation were obtained with soluble phOx₁₀-BSA conjugates at low concentrations (in this case ranging from 3.12 to 12.5 ng/ml - data not shown). These results indicate that in Jurkat T cells the simultaneous expression of two different antigen specific CD28 and CD3f functional chimeric molecules does not perturb the physiologic CD28 and TCR/CD3 signalling pathways. In addition, signals generated through these chimeric molecules can synergize independently with those generated through the wild-type molecules. Most importantly, in the presence of PMA. combined stimulation of JNIPCD3f/OxCD28 cells with both antigens (immobilised NIP₁₀-BSA and soluble phOx₁₀-BSA) resulted in enhanced IL-2 production compared to that observed after stimulation of JNIPCD3j7OxCD28 cells with PMA and SC-anti-CD3e mAb or iNIP₁₀-BSA (Fig 6B). The level of IL-2 secretion was quantitatively similar to that produced by wild type Jurkat cells (not shown) or JNIPCD3f/OxCD28 cells stimulated with PMA, SC-anti-CD3e and SC-anti-CD28 mAbs (Fig. 6B). These results show that signals triggered upon specific crosslinking of chimeric CD28 and CD3$^" molecules via extracellular binding domains with different antigenic specificities act synergistically for maximal IL-2 secretion. Similar levels of IL-2 were observed after crosslinking of chCD28 molecules with soluble phOx₁₀-BSA in the presence of SC-anti-CD3e and PMA (Fig. 6B), indicating that chCD28 receptors can provide optimal levels of costimulation for signalling via chCD3f or unmodified TCR/CD3 receptors.

Discussion

Signal transduction through CD28 plays a critical role in regulating the initial response of a T cell to antigen (10). The CD28 molecule is a 44-kD glycoprotein member of the immunoglobulin superfamily and is found as a disulphide-linked homodimer on the surface of a major population of human T cells (8). Previous reports indicate that the CD28 extracellular domain may be replaced by the extracellular domain of the CD8 molecule which is structurally similar (43). These CD8/CD28 chimeric molecules demonstrated that the cytoplasmic domain of CD28 is sufficient for costimulation of IL-2 secretion and association with phosphatidylinositol 3 '-kinase. In this study we demonstrate that a scFv of an antibody molecule fused to the transmembrane and cytoplasmic portions of the CD28 molecule can be expressed in a CD28-expressing leukemic T cell line as an functional antigen specific receptor. Intracellular signalling is triggered upon crosslinking of chCD28 molecules through their extracellular antigen specific domain, and this signal is quantitatively similar to that elicited by specific crosslinking of the wtCD28 molecule with mAbs.

Assessment of the response through chCD28 to different concentrations of soluble antigen suggests possible insights into the control of CD28 mediated costimulation. At appropriate concentrations of soluble multivalent antigen optimal crosslinking of chCD28 should be possible. There is a clear correlation between the strength of CD28-mediated signalling and the antigen concentration: the steep dose-response relationship along with a clear ceiling of the effect (Fig. 3) suggesting very tight regulation of CD28-mediated costimulation which could be relevant in CD28-B7 immunobiology. Certainly, the soluble multivalent antigen used here should efficiently promote CD28 aggregation as well as crosslinking and this appears to have important functional consequences in T cell responses (42, 9). The apparent clustered distribution of B7 molecules on human Langerhans cells (44) and the low avidity of CD80 and CD86 molecules for CD28 leads to speculation that serial triggering of the CD28 receptors present on the APC-T cell contact space may be occurring (45). As the affinity of scFv-B1.8 (33) is comparable to that of CD28 and CD80 (45) one may speculate that such mechanisms may also be occurring here. The very steep dose response seen in response to increasing antigen may be as a result of effective crosslinking and aggregation occurring as soon as a critical frequency of receptors are occupied.

It is interesting to note that the chCD28 construct does not include the proximal cysteine residue involved in the formation of intermolecular disulphide bonds. In spite of the absence of covalent dimerization and the inclusion of a scFv as the extracellular ligand binding domain, the chCD28 molecules have functional properties that are similar to the wtCD28. Although we cannot exclude that non- covalent dimerization, with chimeric or unmodified CD28 molecules, may still occur in the cell membrane, our results indicate that in absence of covalent dimerization CD28 molecules are still functional in human T cells. Certainly, the chicken CD28 homologue is expressed as a non-disulphide-linked receptor and yet it appears to have functional properties that are remarkably similar to mammalian CD28 (46), although data are presented below (in example 2) which suggeest that the cell membrane- proximal cysteine residue (and thus the possibility of disulphide bond formation) may be important for maximal activity in the chimeric molecule.

These chimeric scFv-CD28 molecules may be of value in adoptive cellular immunotherapy protocols using peripheral blood lymphocytes or naturally occurring antigen specific T cells as effector cells. At least a subset of NK cells express CD28 (47, 13) and the CD28 costimulatory pathway has been shown to be important in the proliferation and cytokine production of NK cells (13). In addition, the participation of CD28/B7-1 interactions in direct recognition and killing of B7-1 expressing tumour cells by NK cells has been reported (21) so these chimeric scFv-CD28 molecules may also be useful for therapy based on NK cells. The potential benefits of using chimeric scFv-CD28 molecules are to improve the targeting ability (by the scFv moiety) and the activation of modified T or NK cells by target cells that do not express natural B7 counter-receptors.

In the proposed system the antigen specific CD28 costimulatory pathway will be unchecked, avoiding the possibility that CTLA-4-mediated inhibitory signals might be dominant over CD28-derived signals during the inductive phase of a T cell response (24). This could also have disadvantages as it has been recently shown that CTLA-4 has a critical role to maintain immunologic homeostasis (25, 26). If the target cells persist this could lead to uncontrolled proliferation of modified lymphocytes. Animal studies are clearly needed, but even in the absence of CTLA-4-mediated inhibitory effects the senescence of activated scFv-CD28 expressing T cells should occur provided the specific target cells are effectively eradicated.

Previous reports have shown that cells expressing chimeric antibody-TCR/CD3 genes undergo stimulation upon encountering specific antigen and can lyse target cells (29, 48). In this study, we show that chimeric scFv-CD3s^* molecules trigger IL-2 secretion upon encountering the antigen, but not at maximal levels. The restricted expression pattern of the B7 family of molecules (8, 49) means that, in general, redirected T cells expressing chimeric TCR/CD3 molecules will not receive costimulatory signals on binding target cells. It has been demonstrated recently (19), that in absence of exogenous help, a CD28-B7 interaction is required during the initial phase of a CTL response. This costimulatory interaction results in the secretion of IL-2 by CD8* CTL precursors, which will allow the subsequent clonal expansion of antigen specific CTLs. Clearly the presence of CD28-mediated signalling could be valuable to improve the efficacy of adoptive immunotherapy protocols and our preliminary data certainly suggest that scFv-CD28 chimeric receptors with anti-tumour specificity are functional on binding tumour cells (unpublished observations).

The generation of double chimeric T cells expressing antigen specific CD28 and CD3f molecules reveals that both receptors are functional and that chimeric CD28-mediated signals can synergize with unmodified TCR/CD3 or chimeric CD3 ^"-mediated signals to secrete maximal levels of IL-2. The therapeutic use of genetically modified T cells has been demonstrated in animal models (50) and the double chimeric T cells as described here should allow more effective activation on encountering the target. Certainly studies using a pair of bispecific antibodies each designed to bind to a tumour-associated antigen on the tumour and either CD3 or CD28 molecules on the T cell showed that the combination of both antibodies and pre-stimulated T cells resulted in cures of established tumours (51). In the system described above, MHC- independent recognition of target/s antigens in tumour cells by high affinity T lymphocytes, through chimeric antigen-specific CD28 and CD3f molecules, should induce not only the specific lysis of target cells by transduced CTLs, but also the full activation of transduced CD4⁺ T cells. In addition, the cells should be able to proliferate efficiently and their progeny will remain modified which would be a clear advantage over the use of bispecific antibodies.

There is also the potential for the creation of an appropriate helper environment, by cytokine secretion, which could enable non-transduced CTLs expressing the appropriate TCR to respond against the tumour in a MHC -dependent manner; this could help to avoid the out growth of tumour cells that do not express the selected target antigen. Similarly transduced NK cells with chimeric CD28 molecules could be effective especially against tumour cells which have defects in MHC expression (14). Such modified cells clearly provide an exciting prospect for a new therapeutic modality for the treatment of cancer and some viral diseases. Combined with other developments in cellular therapy such chimeric receptors should facilitate the development of this novel therapeutic modality as well as improving the understanding of the CD28 pathway.

Example 2 Two further nucleic acid constructs, termed J28IIH and J28EKS respectively (named after the first three amino acids, using the single letter amino acid code, of CD28 present in the chimeric polypeptide) were prepared. The constructs were made as previously described for J28 by PCR cloning the CD28 fragments using primers identical to the indicated region (in figure 7) of CD28 and encoding a 5' Norl site for cloning and to introduce the Ala- Ala- Ala tripeptide linker (as shown in figure 1). The constructs with the scFv(Bl .8 - anti-ΝIP) were sequenced and cloned into pCEP4 as described in example 1. Figure 7 is a schematic illustration comparing the J28, J28IIH and J28EKS constructs. In relation to the figure, the numbers indicate the amino acid residues of CD28 (counting from 1 at the Ν-terminal). To assist understanding of the nature of the constructs, Figure 10 shows the DΝA and amino acid sequence of expressed CD28. The fusion sites at residues 108 (for J28EKS), 114 (for J28IIH) and 124 (for J28) are shown in bold. The numbering is of amino acid residues, with the Ν-termiiial as number 1.

The constructs described for CD28-CD4 fusions in WO 96/23814 are fused at position 135 (CH28-2) and 122(CH28-3). It is believed that J28IIH, fused at 114, encodes a substantially superior polypeptide in functional terms to those described in the prior art, as will be apparent from the discussion below.

Jurkat cells were transfected as previously described using lipofection and the population of hygromycin resistant clones subsequently analysed. FACS analysis employed an anti-lambda light chain, again as described in example 1. The results are shown in Figures 8 A, B and C. The data clearly show significantly improved cell surface expression of the chimeric J28IIH (8C) relative to J28 (8A) or J28EKS (8B).

Without wishing to be bound by any particular theory, the inventors hypothesise that the following may be the explanation for the better results observed in connection with J28IIH: the construct J28 originally described is heterogeneously expressed and does not have a disulphide bond, which may be important. The construct J28IIH maintains the disulphide bond closest to the transmembrane region and also the full helix (residues 114-122) to allow lateral spacing. This is probably important in allowing the efficient folding of the light chains of the two scFvs. Unlike the CD28 domains they do not dimerise and to avoid clashes should be maintained at a reasonable distance apart. The Ala-Ala-Ala spacer (arising from the use of a NotI cloning site) is useful in mamtaining this distance and flexibility between the CD28 helical region and the V-regions. The construct J28EKS also retains these features and an additional region which normally packs into the CD28 immunoglobulin fold. When taken out of context this is likely to result in poorly folded protein. The FACS analysis shows poor expression (Fig. 8B) tending to confirm this.

The above translates into improved funtionality for the J28IJH chimeric protein. The functional analysis of the chimeric polypeptides was performed using solid phase ΝIP₁₀-BSA coated on plastic to provide costimulation. The signal through CD3 was provided by PMA and ionomycin as previously described in example 1. Optimal costimulation (positive control) was provided by soluble cross-linked anti-CD28 as described in example 1. The results are shown in Figures 9A-C, which are graphs of IL-2 production (in ng/ml) against concentration of iNIP₁₀-BSA (in μg/ml) for J28 (9A), J28EKS (9B) and J28IIH (9C) respectively. The J28IIH construct is essentially optimally functional at concentrations of iNIP₁₀-BSA > 3 μg/ml but neither of the other constructs can provide optimal costimulation even with concentrations of lOOμg/ml. Interestingly the production of IL2 is also higher in these cells (note that the y-axis scale for the three graphs in figure 9 is not identical) - this suggests that the suboptimal constructs may be generally detrimental to the cells (perhaps due to presence of misfolded protein).

These data show that the precise design of the construct is very important for optimal function, in a manner which could not have reasonably been predicted from the prior art, and computer-assisted modelling performed by the inventors suggests (with the benefit of the present data) that J28IIH is probably very close to the optimal construct of this type (within an amino-acid or so). The modelling also suggests the value of the helical spacer when using molecules such as scFvs instead of dimerising molecules such as CD4 or CD8. REFERENCES

1. Townsend & Bodmer 1989 Annu. Rev. Immunol. 7:601-624.

2. Rόtzschke & Falk 1994 Current Op. Immunol. 6:45-51.

3. Croft et al. , 1994 J. Immunol. 152:2675-2685.

4. Jenkins 1994 Immunity 1 :443-446.

5. Schwartz 1990 Science 248: 1349-1356.

6. Tan et al., 1993 J. Exp. Med. 177: 165.

7. Mueller et al. , 1989 Annu. Rev. Immunol 7:445-480.

8. June et al. , 1994 Immunol. Today 15:321-331.

9. Linsley & Ledbetter 1993 Annu. Rev. Immunol. 11 : 191-212.

10. Allison & Krummel 1995 Science 270:932-933.

11. Harding et al., 1992 Nature 356:607-609.

12. Boise et al., 1995 Immunity 3:87-98.

13. Nandi et al., 1994 J. Immunol. 152:3361-3369.

14. Garrido et al. , 1993 Immunol. Today 14:491-499.

15. Ostrand-Rosenberg 1994 Current Op. Immunol. 6:722-727.

16. Chen et al., 1992 Cell 71 : 1093-1102.

17. Townsend & Allison 1993 Science 259:368-370.

18. Chen et al. , 1994 J. Exp. Med. 179:523-532.

19. Harding & Allison 1993 J. Exp. Med. 177: 1791-1796.

20. Azuma et al. , 1993 J. Immunol. 150:2091-2101.

21. Wu et al., 1995 J. Exp. Med. 182: 1415-1421.

22. Allison et al. , 1995 Current Op. Immunol. 7:682-686.

23. Linsley et al. , 1991 J. Exp. Med. 174:561-569.

24. Krummel & Allison 1995 J. Exp. Med. 182: 459.

25. Waterhouse et al. , 1995 Science 270:985-988.

26. Tivol et al., 1995 Immunity 3:541-547.

27. Riddel & Greenberg 1995 Annu. Rev. Immunol. 13:545-586.

28. Rosenberg 1992 J. Clin. Oncol. 10: 180-199.

29. Eshhar et al. , 1993 Proc. Natl. Acad. Sci. USA. 90:720-724.

30. Weiss et al., 1984 J. Immunol. 133: 123-128.

31. Hawkins & Winter 1992 Eur. J. Immunol. 22:867-870. 32. Figini et al. , 1994 J. Mol. Biol. 239: 68-78.

33. Hawkins et al., 1992 J. Mol. Biol. 226: 889-896.

34. Hoogenboom et al. , 1991 Nucl. Acids Res. 19: 4133-4137.

35. Stevenson et al., 1995 Immunol. Rev. 145:211-228.

36. Pascual & Capra 1991 Adv. Immunol. 49: 1-74.

37. Aruffo & Seed 1987 Proc. Natl. Acad. Sci. USA. 84:8573-8577.

38. Weissman et al. , 1988 Proc. Natl. Acad. Sci. USA. 85:9709-9713.

39. Strathdee et al. , 1992 Nature 356:763-767.

40. Alvarez- Vallina et al., 1993 J. Immunol. 150:8-16.

41. Weiss et al., 1986 J. Immunol. 137:819-825.

42. Ledbetter et al. , Blood 1990 75: 1531-1539.

43. Stein et al.. 1994 Mol. Cell. Biol. 14:3392-3402.

44. Symington et al., 1993 J. Immunol. 150:1286-1295.

45. Linsley et al., 1994 Immunity 1:793-801.

46. Young et al.. 1994 Immunol. 52:3848-3851.

47. Azuma et aL, 1992 J. Immunol. 149: 1115-1123.

48. Hwu et al. , 1993 J. Exp. Med. 178:361-366.

49. Thompson 1995 Cell 81:979-982.

50. Hwu et al.. 1995 Cancer Res. 55:3369.

51. Renner et al.. 1994 Science 264:833-835.

SEQUENCE L ISTING

( 1 ) GENERAL INFORMATION

(l) APPLICANT

(A) NAME Medical Research Council

(B) STREET 20 Park Crescent

(C) CITY London

(E) COUNTRY United Kingdom

(F) POSTAL CODE (ZIP) WIN 4AL

(G) TELEPHONE. (0171) 6365422 (H) TELEFAX (0171) 3231331

(n) TITLE OF INVENTION Improvements in or Relating to Cell Stimulation

(in) NUMBER OF SEQUENCES 6

(iv) COMPUTER READABLE FORM

(A) MEDIUM TYPE Floppy disk

(B) COMPUTER IBM PC compatible

(C) OPERATING SYSTEM PC-DOS/MS-DOS

(D) SOFTWARE PatentIn Release #10. Version #130 (EPO)

(2) INFORMATION FOR SEQ ID NO 1

(i) SEQUENCE CHARACTERISTICS

(A) LENGTH 17 amino acids

(ii) MOLECULE TYPE peptide

(xi) SEQUENCE DESCRIPTION SEQ ID NO 1

Ala Ala Ala Pro Ser Pro Leu Phe Pro Gly Pro Ser Lys Pro Phe Trp 1 5 10 15

Val

(2) INFORMATION FOR SEQ ID NO 2

(l) SEQUENCE CHARACTERISTICS

(A) LENGTH 18 amino acids

(B) TYPE amino acid

(C) STRANDEDNESS

( D) TOPOLOGY l i near

( n ) MOLECULE TYPE pepti de (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 2:

Ala Ala Ala Thr Glu Ala Gin Ser Phe Gly Leu Leu Asp Pro Lys Leu 1 5 10 15

Cys Tyr

(2) INFORMATION FOR SEQ ID NO: 3:

(l) SEQUENCE CHARACTERISTICS.

(A) LENGTH: 39 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 3: ATATAGCGCG GCCGCTCCAA GTCCCCTATT TCCCGGACC 39

(2) INFORMATION FOR SEQ ID NO: 4

(l) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 39 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 4: TACGTCTAGA TTATTAGGAG CGATAGGCTG CGAAGTCGC 39

(2) INFORMATION FOR SEQ ID NO: 5-

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 40 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 5: ATATAGCGCG GCCGCCTACA GAGGCACAGA GCTTTGGCCT 40

(2) INFORMATION FOR SEQ ID NO: 6-

(ι) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 39 base pairs

(B) TYPE- nucleic aciα

(C) STRANDEDNESS: single

(D) TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO: 6: TACGTCTAGA TTATTAGCGA GGGGGCAGGG CCTGCATGT 39

Claims

1. A chimeric polypeptide for display on a cell surface, comprising an effective portion of the CD28 intracellular signalling domain, a transmembrane domain and an extracellular antigen-specific binding domain of an immunoglobulin.

2. A chimeric polypeptide according to claim 1, comprising the transmembrane portion of CD28.

3. A chimeric polypeptide according to claim 1 or 2, comprising a spacer region between the immunoglobulin binding domain and the transmembrane domain.

4. A chimeric polypeptide according to claim 3, wherein the spacer region comprises part of the extracellular domain of CD28.

5. A chimeric polypeptide according to claim 3 or 4, comprising a spacer region of 18-25 amino acid residues.

6. A chimeric polypeptide according to any one of claims 3-5. comprising residues 114-135 of the CD28 extracellular domain (numbering from 1 at d e N-terminal) as a spacer region.

7. A chimeric polypeptide according to any one of the preceding claims, wherein the immunoglobulin binding domain has binding affinity for a disease marker.

8. A chimeric polypeptide according to any one of the preceding claims, wherein the immunoglobulin binding domain has binding affinity for a tumour-associated antigen or a viral polypeptide or fragment thereof.

9. A chimeric polypeptide according to any one of the preceding claims, wherein the immunoglobulin binding domain is a single chain variable fragment (scFv).

10. A chimeric polypeptide according to any one of claims 3-9, comprising a short, flexible peptide linker between the spacer region and the immunoglobulin binding domain.

11. A nucleic acid sequence encoding a chimeric polypeptide in accordance with any one of the preceding claims.

12. A nucleic acid sequence according to claim 11 , adapted for the expression of the chimeric polypeptide in a eukaryotic cell.

13. A nucleic acid sequence according to claim 11 or 12, directing the expression of a leader sequence at the N terminal of the chimeric polypeptide.

14. A nucleic acid sequence according to claim 13, directing the expression of the human V_H1 leader sequence.

15. A nucleic acid construct, comprising a sequence in accordance with any one of claims 11-14.

16. A host cell into which has been introduced the nucleic acid sequence of any one of claims 11-14, or the progeny thereof.

17. A host cell according to claim 16, wherein the host cell is eukaryotic.

18. A host cell according to claim 16 or 17, wherein the host cell is an immunocompetent human cell.

19. A host cell according to any one of claims 16, 17 or 18, wherein the host cell is a T lymphocyte or a natural killer (NK) cell, or a precursor thereof.

20. A host cell according to any one of claims 16-19, expressing one or more further chimeric. surface-displayed polypeptides.

21. A method of treating a human patient, comprising introducing into the patient a nucleic acid sequence according to any one of claims 11-14.

22. A method according to claim 21 , wherein the sequence is introduced into the patient within immunocompetent cells.

23. A method according to claim 21 or 22, wherein the sequence is introduced into the patient within cells previously obtained from the patient into which the sequence has been introduced.