CA2178950A1 - Tumor cell fusions and methods for use of such tumor cell fusions - Google Patents

Abstract

The invention features products and methods for inducing an immune response against a tumor cell by providing to a patient a cell fusion product. By "cell fusion product" is meant a cell membrane from a tumor cell fused to a cell membrane from a second cell that has a greater immunogenic potential than the tumor cell.

Description

~ Wo 95/16775 2 1 7 8 9 5 0 PCT/US94/14297 TIJMOR CEL~ FUSIONS AND METIIODS
FOR USE OF SUC~ TI~MOR CELL FUSIONS
FIELD OF TI~E INVENTION
The present invention relates to products and methods useful for tumor immunotherapy.
BA~cvuNJ OF T/IE lNV~~
A major objective in the field of tumor immunotherapy is the development of strategies for enhancing tumor immunogenicity, with potential application~ for both tumor prevention and cure. To date, various product~ and methods useful for enhancing tumor immunogenicity have been reported. ~owever, the methods described below are not admitted to represent prior art to the pending claims.
In general, tumors that arise de novo are poorly immunogenic, thereby escaping host antitumor responses (Hewitt et al., 33 Br. J. Cancer 241, 1976) .
Methods that have been described f or enhancing tumor immunogenicity include: (1) using mutagen or drug - treatment (Van Pel and Boon 79 Prnc~ ~atl. Acad. Sci. USA
4718 , 1982 and Frost el al . , 159 !:r. Ex~ . Med. 1491, 1984);

(2) transfecting with a foreign gene encoding an exogenous antigen E:uch as influenza hemagglutinin (Fearon et al, 38 Cancer Res. 2975, 1988); (3) reducing the expression of certain molecules in a tumor that regulate its diiferentiation state (Tyk~ ;nqk; ~i Ilan, 259 Science 94, 1993); (4) transferring a gene expressing a lymph~-k;n,o into a tumor, for example, interleukin-2 (Fearon et al., ... . .. ~

Wo 95/16775 PCTiU594/l4297 60 Cell 397, l9g0), interleukin-4 (Tepper et al. , 57 Cell 503, 1989, Golumbek et al., 254 Science 713, 1991), interleukin-6 (Mullen et al./ 52 Cance~ Re~. 6020, 1992), interleukin-7 (McBride et al., 52 Cancer Res. 3931, ls92);
(5) transferring a gene e~ressing a cell surface-associated costimulator into a tumor, for example, B7 (Chen et al., 71 Ç~Ll 1093, 1992; and Townsend et al., 259 Science 368, l9g3); (6) transferring a gene expressing major histocompatibility complex protein into a tumor cell; ~7) transferring a gene eæpressing a protein that ,onh~nr~F: the expression =of ~ a ma~or histocompatibility complex protein in a tumor cell; and (8) transferring a gene expressing a heat shock protein into a tumor cell (Luckacs et al ., 178 J. EXP . Med . 343, 1993 ) .
~ ~Antigen-presenting cells (APC) provide molecular signals including signals mediated by APC-derived soluble cytokines and APC-derived cell surface costimulators such as: (1) B7 (Linsley et al_, 87 Proc. Natl. Acad. Sci.
U.S.A. 5031, lg90); (2) ICAM-I (van Seventer et al., 144 .J. Immunol. 4579, 1990); (3) VCAM-I (van Seventer et al., 174 ~. EYr~. Med. 901, 1991); (4) L~A-3 (van Seventer et al., 21 Eur. J. Immunol. 1711, 1991); and (5) fibronectin (Shimizu et al., 145 ~. Tmml~nnl.l 59, 1990; Nojima et al, 172 J. EXP . Med. 1185 , 1990 ; and Davis et al ., 145 ~.
Tmml-nnl. 785, 1990) .
A tumor~ cell, once appropriately modif ied through genetic manipulation, can itself function as an APC (Chen et al., 71 Cell 1093, 1992; Townsend et al., 259 Science 368, 1993; Tykocinski & Ilan, 259 Science 94, 1993).
S~IARY OF TIIE LNV~N~L~)N
The pre~ent invention provides products and methods useful ~or~ ~nh~nci nr tumor immunogenicity. More specifically, the present invention provides a cell fusion ~ Wo 95/16775 2 1 7 8 9 5 0 PCT/US94/14297 i product, and methods of using the cell fusion product to enhance tumor immunogenicity.
The general usefulness of this technology relates to the prevention and treatment of various diseases, including cancer. The disea~e may be present in any animal, including a human. This technology further embraces a wide range of utilities ;nrll~tlin~ ;n~llrln~ the production of antibodies in vitro. The present invention may also be used to induce the production of antibodies in a variety of animals, including humans.
It has surprisingly been discovered that the immunogenicity of a tumor cell can be remarkably enhanced by fusing a membrane of a tumor cell to a membrane from a another cell with greater immunogenic potential than the tumor cell. ~rhe phrase "cell fusion partner" is used to describe any other cell with a greater immunogenic potential than the tumor cell. Membrane extracts or whole cells may be fused in the pre~ent invention.
It is believed that fusion of a membrane from a tumor cell with a membrane from a cell fusion partner changes the capacity of the tumor cell to activate specific T-cell responders in the host immune system so that an immune response can be mounted against that tumor cell. In support of this, it has been discovered that the introduction of a hybrid cell comprising a tumor cell fused to a cell fusion partner not only reduces tumor growth, but also the growth of normal tumor cells (i . e . .
non-fused tumor cells) within the same host.
It is further believed that the cell fusion 30 partner contributes relevant tumor-specific antigens to the hybrid cell and that the cell fusion partner contributes cell sur~ace costimulators, soluble cytokines, MHC molecules, and other undefined molecular factors to the hybrid cell These contributions in aggregate result in a highly antigenic and immunogeni~ phenot~pe for the hybrid cell. In turr~, the hybrid cell can be used a~ an WO 95116775 PCTrUSs4/14297 ~
21 7895~

effective cellular vaccine. Xowever, this proposed theory is not meant to act as a limit to any alternative theories or mechanisms of carrying out the invention.
The pre~:ent inventlon is based upon the finding that enha~ced ~ n; city can be induced by fusing a cell membrane from~a tumor cell with a cell membrane from a cell fusion partner, preferably one capable of effective antigen presentation and T-cell activation. Through such cell fusion, the full complement of molecular factors~that are normally produced by APC' s and that are required for effective T-cell activation are combined with the full set of potential tumor antigens associated with a particular tumor. Moreover, through the choice of fusion partners, it is possible to tailor the=nature of the anti-tumor T-cell immune response. ~ :
Thus, a broad means has been discovered by which the antigenicity or immunogenicity of any tumor ce~l can be increased by fusion with a membrane from another cell with greater immunDgenic potential than the tumor cell.
Given this discovery and the methodology provided herein (in which examples :of this discovery are provided), it is now straightforward for those in the art to screen any particular target tumor cell~to determine whether fusion with any particular cell fusion partner will enhanc~ the tumor cell' 8 1 n~n; city.
Thus, in a first aspect, the invention features methodæ for inducing immunity against a tumor cell by providing to a patient a cell fusion product.
By "cell fusion product" is meant a cell membrane from a tumor cell ~fused to a cell membrane~from another cell that has a greater immunogenic potential than the tumor cell. O~e type of=~ cell fusion product ~is a hybrid cell. .3y 'Thybrid cell" is meant a cell which is derived from the fusion of two parental cells, and it is either the direct fusion product or a daughter cell that is derived by cell division from the original fusion Wo 95/16775 ~ 1 7 ~ q 5 PCTIUS94/14297 product. A hybrid cell rrnti:l;nq one or more molecular components of each of its parental cells . By "parental "
cell is meant either c~ t of the hybrid cell and includes a tumor cell or another cell with greater immunogenic potential than a tumor ce ~ l.
By ~provide~ is meant any method that resulte in the presence of a cell fusion product in a patient. The step of ~'providing" can be performed in a variety of ways including either administering a cell fusion product that was formed ex vivo or fusing the mem~oranes in vivo.
By " immunity" is meant the state of being refractory to a specific disease, which is mediated by the immune system or a state of not being susceptible to the invasive or pathogenic affects of potGnt;~lly infective microbes or to the affects of potentially toxic antigenic substances . By " immune response" is meant the response of the whole or part of an immune system of an organism.
This response could include the activation of r-~ r or humoral systems, including B-cells, and T-cells.
By "tumor" is meant a collection of cells, usually dysfunctional, due to abnormal proliferation.
Benign tumors are not life threatening, e.g., warts.
Malignant tumors are potentially lethal cancers. All tumor types may be treated using the methods of the present invention, since cell fusion is not dependent upon any particular cell phenotype. The tumor cell may be autologous, heterologous, cultured, primary, or metastatic .
By ~autologous" is meant that a tumor cell is - 30 from the patient to be treated, or from another patient having a common major histocompatibility phenotype. sy "primary~ is meant that a tumor cell from the organ of tumor origin in the patient to be treated is used. It also means a primary cultural cell, as distinct from a cell line. By ~metastatic" is meant that the tumor cell is prolif erating at sites distant ~rom the organ o~ tumor Wo 9~16775 PCTnJSs4/l4297 2 ~ 78 q50 origin. By "heterologous" is meant that a tumor cell~from another patient is used. Clinicians and others skilled in the art are able to identify patients in need of treatment using procedur~s that are well known and routine in the art. Procedures for~ obtaining a tumor cell from such a patient and for culturing such a cell are also well known and routine in the art. By "patient" is meant any animal, including a human, with a tumor By " fusing" is meant a process whereby the cell membranes are combined into a single membrane and a cell fusion product is ~ormed. The fusing may be performed by directly in~ecting ~ the cell fusion partner into a tumor mass in a patient. This may further involve identifying a patient in need of tumor therapy, obtaining a tumor cell f rom the patient, and culturing the tumor cell . The fusion step may involve more than two fusion partners.
The fusion step may also involve the use of a chemical fusogen, such as polyethylene glycol, or electrofusion.
By "immunogenic potential" is meant the capacity to activate specific T-cell responders in the immune system or the ability to raise an immune response in an animal, preferably a human. It is commonly found that tumor cells have poor immunogenic potential relative to APCs. By "anti-tumor response" is meant any response that measurably reduces the size of a tumor, int-lllrl;ng the complete destruction of the tumor ~
By "membrane extract" is meant an extract of a cell enriched for membranes, but not necessarily c~nl-il;n;n~ only membranes. Such an extract is chosen 3 0 because it will have the antigenic and immunogenic properties necessary to induce an immune response to the tumor cell in vivo or ex vivo. A cell membrane extract may be derived fro~L an autologous or heterologous tumor cell or another cell with greater immunogenic potential than the tumor cell. The tumor cell or cell fusion product may be treated, for~ example by irradiation, to Wo 95/16775 2 1 7 5 0 PCT/lDS94/14297 reduce it6 ability to proliferate. By nmembrane~ is meant a sheet, usually about lO nm thick and normally composed of a bimolecular layer of lipid and protein, enclosing or partially enclosing a cell, organelle, or vacuole. Cell fusion products may be formed using less than an entire membrane, for example, portions of membranes may be used.
In other aspects the invention features methods for ;n~ ;ng immunity against a tumor cell by providing to a patient either a tumor cell fused to an APC or tl~e fused membranes of such cells. Examples of a conventional APC
include an activated B-cell, a dendritic cell, a macrophage, an activated T-cell, or an endothelial cell.
Methods for identifying other cell ~usion partners that can confer enhanced immunogenicity to a tumor cell are defined herein. Preferred candidate cells are those for which there is evidence of immunogenic potential.
In yet other aspects the invention f eatures methods for inducing immunity against a tumor cell by providing to a patient a tumor cell fused to an activated B-cell with an artificial adhesin. By ~artificial adhesin" is meant a genet cally engineered molecule that is expressed through gene transfer, or through protein transfer is exogenously coated, on the surface of a cell and thereby promotes adhesion to another cell expressing a molecule on its surface that can bind to the engineered molecule . Examples of artif icial adhesins include a glycosyl -phoshatidylinositol -modif ied polypeptide or a biotin-lipid con~ugate, or other compounds with equivalent properties .
In other aspects, the invention features methods for ;n~ -;n~ immunity against a tumor cell by providing to a patient a T-cell activated by contact with a cell fusion product. The T-cell may be part of an immunoselected subset such as CD8-positive T-cells. T-cells are known to be critical mediators of anti-tumor immunity. In general, T-cells are activated by cells collectively referred to as Wo 95/16775 PCr/USs4/14297 ~

"antigen-presenting cells~ (APC) . The diverse cell types that comprise this category share the ability to present antigens, via their major histocompatibility molecules, to the T-cell receptors on antigen-specific T-cells.
The preæent invention tliscloses the capacity of a cell fusion product to stimulate anti-tumor T-cells.
This capacity can be utilized not only for ;n vivo stimulation of T-cells, but also for çx vivo stimulation of T-cells. Ex vivo 6timulation of a T-cell using a cell fusion product can be used as a means of amplifying T-cells with tumor specificity prior to infusion of such T-cells into patie~ts. Methods are. well known in the art for delivering T-cells into patients. Methods are well known in the art for derivirLg T-cçlls from the peripheral bloDd of cancer patients or lsolating infiltrating T-cells directly from tumors and nonspe~ lly amplifying their cell numbers using reagents such as interleukin-2.
Contacting such T-cells with a cell fusion product offers a means for selectively amplifying the tumor-specific T-cells out of the mixed T-cell populations at these~sites.
Once amplified, the T-cells can be re-infused into a patient. It should be eviderLt frDm this that the same patient can be coordinately treated with a cell fusion product, as an active vaccine, along with ex vivo amplified T-cells, as a passive vaccine. This combined treatment maximizes therapeutic effects and is advantageous for- immunosuppressed patients.
In still- other aspects, the invention features a method for identifylng a cçll fusion partner that can be fused to a tumor cell to generate a hybrid cell with greater immunogenicity tha~ the tumor cçll. This method involves fusing ~a tumor cell with a candidate cell and determining the immunogenicity of the resulting cell fusion product. Such methods can be based in vivg or ex vivo. Clinicians and others skilled in the art are able to determine the immunogenicity of a cell using procedures that are well known and routine in the art.
In other aspects, the invention f eatures a method for fusing a tumor cell with another cell. This method involve6 expressing an artificial adhesin on either the tumor cell, the other cell, or both. The method also involvee combining the cells with a fusogenic agent. By "fusogenic agent" is meant any compound that increases the ability of a membrane from a tumor cell to fuse with a membrane from another cell that has greater immunogenic potential than said tumor cell.
The present invention discloses that selectivity can be conferred to a cell fu~ion process by inducing relevant paired cells to adhere to each other prior to addition of a fusogenic agent. By combining such a "pre-adhesion" step with subsequent fusion at low cellular densities (generally below lOs cells/ml), more efficient fusion can be achieved. Xowever, use of low cell density is not a required parameter in this invention. As described, a preferred method for achieving such pre-adhesion is through the use of an artificial adhesin that has been delivered to a relevant cell surface by any one of a number of gene and/or protein transfer methods.
However, alternative methods that do not involve artificial adhesins can be used to achieve pre-adhesion.
For example, a heterobifunctional antibody can be used to adhere a tumor cell and an APC . The adherence - inducing - method should not perturb the antigenicity and immunogenicity of the hybrid cell to be used as a membrane source for immunization.
In still other aspects, the invention features an immunogenic c~ll fusion product. In this product a membrane from a tumor cell is fused to a membrane from a another cell with greater immunogenic potential than the tumor cell.
.

Wo 95ll6775 2 1 7 ~ 9 5 o PCTrU594/l4297 The immunogenic cell fusion products of the present invention are distinct from hybridomas used for monoclonal antibody production where a cultured myeloma cell is fused with a splenocyte. The immunogenic cell fusion product of the present lnvention is also distinct f rom non- immunogenic hybrid cells used in routine laboratory experiments. However, the methods described above for inducing immunity against a tumor cell can utilize these hybridoma and hybrid cells. In the present invention, the membrane of a tumor cell isolated from a patient is fused with a membrane from another cell, such as an APC, with greater ~ immunogenic potential than the tumor cell. Thus, unlike monoclonal antibody production, a primary tumor cell is involved in the fusion process.
Furthermore, unlike myelomas which have the ability to produce antibodies, the tumor cell oi~ the present invention need not be able to make antibodies. At any rate, the cell ~1~sion products of the present invention generally exclude the use of a myeloma cell fused: to a splenocyte One or laore of the cells that generate the cell fusion product, or alternatively the cell fusion product itself, may be molecularly modified prior to administration to a patient. Molecular modifications can be directed toward any one of a number of functional endpoints, including ~nh~nr;nrj the fusion process, promoting selective ~ ;rn between the parental cells, altering the in ~ rQ tissue targeting properties of the cell fusion products, and further ~nl~nr;ni the immunogenicity of ~the cell fusion product above and beyond the F~nl~nrF~ immunogenicity that would otherwise be conferred to the hybrid cell by the parental cell.
A cell fusion partner may be modi~ied prior to fusion with a tumor cell in order to increase the immunotherapeutic ef icacy of the resulting hybrid cell .
This modification can be efiected by alternative methods, Wo 9~/16775 ~ 1 7 8 9 5 0 PCTIUS94/14297 - including gene or protein transfer. Examples of proteins that can be expressed or inhibited in an APC include, a cell surface costimulator, a soluble cytokine, a selectin, an adhesin, a major histocompatibility complex protein, and a coinhibitor (e.g., CD8). By "foreign protein" is meant a protein that is not normally expressed in a particular cell. By "natural cell surface molecule" is meant any molecule ~hat is naturally found on the surface of a particular ell.
One molecular modification entails the coating of one or more of the parental cells with artif icial adhesins that promote selective adhesion between the two cell types prior to the fusion event. A tumor cell or another cell may be modified to enhance its fusion potential. ~he tumor cell can be modified in vivo or in vitro prior to cell fusion. When another cell is directly inj ected into a tumor mass, prior to inj ection the tumor mass may be modiiied in vivo, or the other cell may be modified in vitro, or both.
As will be readily apparent to one skilled in the art, the useful n vivo dosage to be administered and the particular mode of administration will vary depending upon the age, weight and liAn species treated, the particular cellular compositions employed, and the specific use for which these cellular compositions are employed . The determination of ef f ective dosage levels, that is the dosage levels ne--~A~Ary to achieve the desired result, will be within the ambit of one skilled in the art .
The present invention also provides kits including materials used in cell fusion.
As can be seen from the above description, the invention generally features generation of a fusion product, e.g., through cell fusion, to provide a reagent suitable for ; ; 7Atiorl against a tumor, either in a prophylactic or treatment procedure, and features methods -. . !
'~' WO 95/16775 2 1 7 ~ PCT/US94/14297 for ~identifying the optimal cell fusion partner~ to be fused with a tumor cell derived from such a tumor.
Prior to this invention, it is believed that no art described how tumor cell immunogenicity may be enhanced by fusion with another cell with greater immunogenic potential than the tumor cell. The ~use of cell fusion provides GubGtantial advantages for practicing tumor cell engineering and ~nh~nc;ng tumor cell immunogenicity. Advantages of cell fusion over gene transfer include, but are not limited to, the following:
First, the pre~ent invention obviates the need for decoding the precise molecular signaling systems for individual tumor~ cell:T-cell combinations. Thus, unlike current methods which iocus upon individual defined molecules, the present invention bypasses the general lack of understanding of the composite set of antigenic peptides and CoGt;~ t~ry molecules required for effective anti-tumor T-cell responses.
Second, cell fuGion is applicable to diverse tumor cell types. Unlike~ gene transfer, cell fusion is not dependent upon the proliferative potential of the tumor cell, and hence, can be applied to a variety of tumor types which grow poorly in primary culture.
Third, cell fusion is a relatively rapid process and does not impose a burden of excessive cell culturing, shortening the interval between biopsy and treatment.
Certain gene transfer-based immunotherapeutic strategies re~uire selection for stable transfectants. This can be a time consuming process and complicates the clinical practice of such methods and imposes a delay period between biopsy and treatment.
Fourth, no biosafety hazards are known to be associated with the cell~ fusion method of the present invention. Gene transfer is dependent in most instances upon genetic vectors comprising viral components which carry with them some degree~ of biosafety hazard.

~178950 95/16775 Pl~rlUS94/14297 It is believed that fusing a tumor cell with another cell with greater immunogenic potential than the tumor cell effectively r~ ~ ;nF~ the antigenic repertoire of such a tumor cell with the multiple molecular factors, including cell surface costimulators, soluble cytokines, and major histocompatibility complex (MHC) molecules, of such other cell Different potential other cells (including different well known APCs such as an activated B-cell, a macrophage, and an activated endothelial cell) differ in their molecular factor composition. Thus, by pairing a particular APC with a particular tumor cell, one can confer to such a tumor cell desired properties.
Cell fusion, as a means of ~nh~n~-;n~ tumor immunogenicity according to the present invention, may be combined with other known methods for ~nh~nf-; ng tumor immunogenicity, including those cited herein, for example, expreæsion of an exogenous gene in a tumor cell or inhibition of an endogenous gene in a tumor cell. Indeed, it is believed appropriate to combine the therapies or methods described herein with other methods to enhance the immunogenicity of a tumor cell.
The summary of the invention described in detail above is not 1ntF~n~ in any way to limit the scope of the present invention, which 18 defined in the appended claims. Other features and advantages of the invention will be apparent from the following description of the preferred embodiments thereof, and from the claims.
sRIEF DESCRIPTION OF T~iE DRaWINGS
Fig. 1 shows schematic diagrams of the pM-CSF-GPI/R~P4~ and pM-CSFR/REP7~ expression constructs which encode members of an artif icial adhesion pair . For the pM-CSF-GPI/REP4c~ expression construct (panel a), the M-CSF-DAF chimeric sequences (thicker box) are depicted in the multiple cloning site of pREP4Om For the pM-35 CSFR/REP7~ expression construct (panel b), the M-CSF

WO 9~/1677~ ~ ~ 7 ~ ~ ~Q PCTnJSs4/14297 receptor coding ser~uence ~thicker box) is depicted in the pREP7~ vector. Abbreviations: EB~7 oriP, Epstein-Barr virus origin of replication; EB~A-l, EBV nuclear antigen-l; RSV 3 ' LTR, Rous sarcoma virus 3 ' long terminal repeat promotor; PA, SV40 polyadenylation/termination sequences;
hph, 11yyl~ y~in-B resistance gene; Amp, ampicillin-resistance gene.
DESCRIPTIO~ OF TIIE ~R~ v ENBODIMENTS
Preferred embodiments of the present invention are described in detail below. However, the following description of the preferred embodiments i8 not intended to limit, in any way, the scope of the present invention which is def ined in the appended claim~ s .
The present invention addreeses the need for conferring a complex phenotype, immunogenicity, to diverse types of tumor cells. Methods are provided for conferring this complex phenotype to tumor cells ex vivo and in vivo.
In addition, methods are provided for using modified tumor cells to generate ~T-cells to~ be used for cell transfer.
Yet other methods are provided for ~nh~nr;ng adhesion between a tumor cell and another cell, as a prelude to fusion of the two cells.
The preaent invention entails the fusion ~ of a tumor cell to another cell Methods ~1~F; rn~ to promote interc~ 1^ adhesion can be combined with conventional non-selective cell fusion methodæ in order to selectively target and enhance the cell iusion process.
More specifically, when tumor cells and APCs are combined, nonselective fusion agents that are routinely used to fuse cells will not only induce tumor cell:APC
fusions, but also tumor cell:tumor cell and APC:APC.
fusions as undesired byproducts. Herein disclosed is a cell fusion method for minimizing these undesired fusion events and simultaneously m~;m; 7lng the desired tumor 35 cell:APC fusion events. This is accompanied by Wo 9511677~ ~ 4/
~ l 7 8 9 5 ~ I Cr/Uss 14297 artificially promoting adhesion between a tumor cell and an APC, without influencing self-adhesiveness, and adding fusing agents or performing electrofusion when adherent tumor cell :APC conjugates are at low cell densities . The lower cell densities would be less 'avorable to fusion between n~ rent cells of the same cell type.
According to the present invention, a combined adhesion/fusion method for generating tumor cell :APC
hybrids can comprise any one of a number adhesion and fusion component methods. A method ~or altering the adhesive propertie6 of cells that is particularly well-suited for the adhesion/fusion method has been developed.
This method is based upon the uE:e of a class of molecules that can be designated by the term ~artificial adhesins. "
Specifically, in one example, a glycosyl-phosphatidylinositol (GPI)-modified variant oi the cytokine macrophage colony stimulating factor (M-CSF), designated M-CSF-GPI, was expressed on the surface o~
human bone marrow stromal cells. A chimeric M-CSF:decay-2 0 accelerating f actor expression construct was used ~or M-CSF-GPI expression . Cell: cell binding assays established that this artificially membrane-tethered cytokine functioned as a potent cellular adhesin, allowing for enhanced binding to M-CSF receptor-expreeeing cellular transfectants. Antibody blocking analyses confirmed the M-CSF :M-CSF-receptor dependence ~ of the enhanced intercellular binding. This capacity to direct the cellular interactive repertoire of selected cells can in principle be applied to other cell types and other 3 0 molecular pairs to be used in cell-based therapies .
Intercellular adhesion is mediated by homotypic and heterotypic molecular interactions at membrane inter~aces. There is a growing compendium of cell sur~ace molecules that have been assigned functions as natural adhesins in regulatory interactions between cells. Since natural adhesins have frequently been found to be W095/1677~ 2 1 ~8 95~ PCTIUS94/14297 multifunctional, they often cannot be used as neutral molecules for altering cellular adhesive properties.
Moreover, most natural adhesins, by virtue of being tr~n! ' d,~e hydrophobic peptide-anchored molecules, can only be expressed on cell membranes by gene trangfer.
Therefore, methods are needed to artificially modify adhesiveness between cells in a neutral way and without necessarily using gene transfer.
A known method for accomplishing this is through the use of a palmitate-conjugated antibody as an artificial adhesin for cross-linking cells (Colsky et al., 124 ~ l. Methods 179, 1989). However, this approach is limited by a number of factors including the re~uirement f or working with multichain antibody u~its .
Soluble antibodies with bifunctional specificities offer:
another potential approach ~or cross-linking cells, but there is still the potential of signal transduction mediated by the antibodies.
Ther~ is herei~ disclosed an alternative approach for art;f;fi~lly e~hancing adhesiveness between cells. Genetically engineered variants of known ligand receptor pairs can be used as art;~;r;;3l adhesins. The strategy is to genetically alter a soluble polypeptide ligand so that it incorporates into the surface of one cell via a carboxy-terminal anchoring domain and yet still retains its capacity to bind its receptor on another~cell.
The results described below document the feasibility of such an approach.
The use o ~ a GPI moiety f or membrane anchorage offers special potential advantages in the context of cellular engineering. Notably, it builds upon the demonstration (Tyk ~inqki et al., 85 ~roc. ~atl. ~ Acad.
Sci. ~SA 3555, 1988) that any polypeptide can be readily produced as a GPI -modif ied variant through the use of chimeric coding sequences ~n~ sing both the coding se~uence for the protein of interest and the GPI signal _ _ _ ~Wo 95/1677~ 2 1 7 8 ~ 5 0 PCT/US94/1.1297 sequence from a naturally GPI-modified protein such as decay accelerating factor (DAF) .
Another potential advantage of the use of GPI
anchors for adhesins stems from the fact that purified GPI-modified proteins, by virtue of their amphophilic properties, can be readily reincorporated into cell membranes in the presence of low, non-lytic concentrations of non-ionic detergents such as NP-40 ~Medof et al., 160 J. Ex~. Med. 1558, 1984) or even in the absence of detergent. Hence, cells can be coated with purified GPI-modified polypeptides, representing a form of "protein paint, " bypassing the rerluirement for gene transfection into the cell whose surface is being molecularly engineered. Delivery of exogenous polypeptides to cells by such a protein transfer approach circumvents problems associated with gene transfer, particular in the case of primary, nontransformed cells which in general cannot be easily transfected.
Numerous applications for this artificial adhesin technology can be envisioned, some with therapeutic implications. For example, the immunostimulatory and/or effector properties of cells used in cell-based therapies, such as tumor cell:APC hybrids, APCs, immunogenic tumor cells, or T-cells, could be selectively .-nl~nre~l by increasing their adhesive properties in a selective way.
A preferred protein transfer method for coating - cells with artif icial adhesins involves the use of GPI-modif ied proteins . Methods f or perf orming protein transfer using GPI-~odified proteins have been described.
In all instances, this has entailed the use of dilute, concentrations of non-ionic detergent, for example, . oo496 NP-40, in the solution containing the GPI-modified proteins. It has been discovered that protein trans~er can be accomplished even more effectively in the absence of any detergent. By leaving out the detergent, higher W095/16775 7~7895a PCr~ss4/1~297 ~

concentration6 of a GPI-modif ied protein can be used and problems a6sociated with cell ly6i6 by detergent are avoided. The optimal time for co-incubation of a cell with a GPI-modified protein, such a3 an adhe6in-GPI
chimeric polypeptide, i3 two hour3 at either room temperature or 37C. 25 microgram6/ml of the GPI-mQdified polypeptide in the coating reaction with the cell6 i6 generally 6ufficient for aderiuate adhe6in coating, although higher rrn~ ~ntrati-on6 can be u6ed with increa6ed ef f icacy .
Another preferred protein tran6fer method for coating cell3 with artif icial adhe6in6 entail6 the u6e of a chimeric polypeptide in= which an adhe6in polypeptide 6equence, for example, M-CSF, i6 linked to a 6treptavidin 6equence. Method6 for~ u6ing prokaryotic expre66ion 6y6tem6 to quantitatively produce 6uch polypeptide-6treptavidin chimera6 are~ own to tho6e familiar with the art. In u6ing such an adhe6in-6treptavidin chimera, the cell of intere6t i6 pre-coated with biotin. A u6eful method for pre-coating cell6 with biotin i6 through the use of biotin-lipid conjugates which can be used to pre-coat cells to high bioti~ densitie6 (up to 107 biotin molecules/cell). Alternatively, the cell6 can be chemically biotinylated u6ing 6tandard cellular biotinylation prQcedureE. A chimeric adhesin-6treptavidin polypeptide i6 added to a pre-biotinylated cell, in order to generate a cell expre6sing the artificial adhe6in at it6 6urface. It is generally not avidin on one cell and biotin on a 6econd cell that are being u6ed to bring two cell6 together, but instead avidin and biotin are usually 6imply being u6ed on the 6ame cell 6urface to deliver an ;rn to that cell~6 3urface. Then, the 6ame proce66 may be applied to a 6e~con~ cell again, potentially u6ing an avidin-biotin comple~ at that cell 3urface. Advantage6 of thi6 method; include the high 6urface den6itie6 of art;f;~;~l adhesin that can be achieved, the feasibility Wo 95/1677s 2 1 7 8 q 5 a PCTiUS94/l4297 - of producing large ~uantities of the chimeric polypeptide using a prokaryotic expression system, and the biocompatibility aesociated with biotin-lipid conjugates in vivo.
According to the present invention, a tumor cell can be coated with one member of an artificial adhesin pair, and a conventional APC can be coated with the other member of the pair. When cell populations of each of the cell types are combined, intercellular conjugates form, pairing a tumor cell with an APC. ~ When an electric current is applied, or a chemical fusing agent is added, to the mixed cell population, adherent cells within conjugates preferentially fuse. Preferential fusion can be further promoted by keeping the mixed cell population at a low cell density. This will minimize the formation of extraneous tumor cell:tumor cell and APC:APC hybride with no immunotherapeutic potential.
In the example3 provided below, a preferred method for fusing cells comprises the use of polyethylene glycol as a fusing agent. Other methods of fusing cells can be used, including electrofusion or use of viruses or viral components, for example, Sendai virus, that promote cell fusion.
The invention will be more fully understood with reference to the examples which follow. The following examples are intended to illustrate the invention, but not to limit its scope which is defined in the claims appended hereto . The f ollowing examples are presented to illustrate the advantages of the present invention and to - 30 assist one of ordinary skill in the art in making and using the same, but are not intended, in any way, to otherwise limit the scope of the disclosure or the protection granted by letters patent hereon.
;Bxam~le l: He~atocarcinoma fused to ~Tl activated B-cell lose their t~ ri~enicity W0 95/16775 2 1 7 ~ PCT/US9~/14297 ~ctivated B-cells are effective APCs (hanzavechia, 140 ~ature 1985, 1985; Ron, 138 ~ rLmunol.
2848, 1987; and Kurt-Jones, 140 ~. Immunol. 3773, 1988).
Con3equently, this cell type provides an excellent APC
fusion partner for the present invention. BERH-2 is a chemical carcinogen-induced rat hepatocellular car~cinoma cell line from the Wistar rat. BERH-2 grows rapidly and forma tumors in the liver of syngeneic animals.
In ~periments described below, BERH-2 cells were fused with activated B-~cells in an attempt to enhance the immunogenicity of BERH-2 cells. The data provided herein indicates that the hybrid cells, designated BERH-2-B, became more immunogenic ~and less tumorigenic than the tumor cell.
ExamPle lA: Fusion and selection of BERH-2-B cells Activated B-cells were obtained from the spleens of rats injected 14 days earlier with bovine serum albumin in complete Freud' 8 adjuvant. BERX-2 cells were fused with these purified activated B-cells using polyethylene glycol, using a standard B-cell hybridoma fusion protocol.
The fused cells, designated BERH-2-B, were selected by panning, first with a rabbit anti-BERH-2 antiserum and second with a rabbit anti-rat B-cell antiserum.
The rabbit anti-BERH-2 and rabbit anti-rat B-cell antisera : were prepared by immunizing rabbits subcutaneously with either~BERE~-2 hepatoma or purif ied B-cells from Wistar rats in complete Freud' 8 adjuvant .
~ctivated B-cells were purified by panning with plates coated with purified goat_ anti-rat Ig antibody. After repeat boosting during two_months, antiserum was collected and purified by protein G-sepharose chromatography.
Finally, the antiserum was repeatedly absorbed with either BERH-2 hepatoma - cells or rat B-cells .

~WO 95tl6775 ~ l 7 8 9 PCTruS94rl4297 ExamPle lB: ExPression ~n~ characterization of antiqens Expression of class I MHC, class II MHC, B7, ICAM-1 and LFA-1 on BERH-2 cells, activated B-cells and BERH-2-B hybrid cells were assessed. Cells were washed with phosphate-buffered saline (PBS) and stained with monoclonal antibody to rat MHC class I (OX-18), MHC class II (OX-6), ICAM-1 (LA 29) or LFA-1 (WT.1) as primary antibody . To stain f or rat B7, we used a chimeric protein, CTLA-4-Ig. Cells were incubated with the antibodies or chimeric protein for 30 minutes on ice. A
mouse anti-human CD3 monoclonal antibody (GH3, IgG2b) and a chimeric human CD44-Ig protein were used as llegative controls. Cells were washed three times. FITC-goat anti-mouse Ig or FITC-labeled rabbit anti-human Ig was used as a secondary antibody and added for another 30 minutes on ice. After washing, samples were fixed and analyzed in a FACScan f low cytometer .
Parental and hybrid tumor cells were phenotyped by immunocytochemical staining and f low cytometry . The parental BERH-2 cells expressed low levels of class I MHC
antigen and ICAM-1, but lacked class II MHC antigen, LFA-1 and the costimulator B7. All four hybrid BERH-2-B cell lines displayed increased class I MHC expression. In addition, BER~-2-B hybrid cell lines expressed MI~C class II antigen, ICAM-1, LFA-1 and B7. These BERH-2-B cell lines have stably expressed both tumor and B-cell antigens for more than five months.
- Exam~le lC: Com~arison of tumoriq~n; city of ~arent and hvbrid cells - 3 0 The tumorigenicity of parental BERH- 2 and hybrid BERH-2-B cells were compared, and the survival data shows e3lhanced animal survival for syngeneic anlmals injected with BERH-2-B hybrid tumor cells as compared to BERH-~
tumor cells. Two groups of female Wistar rats (ten/group) were injected intrahepatically with 2 x 106 BERH-2 cells or 2 x 106 BERH-2-B hybrid cells.

Wo 9S/16~75 2 ~ ~ ~ 9 ~ ~ PCTIUS94/1~297 ~

All animals injected with BERH-2 parental~ cells developed liver tumors - ~nd dled within 6 0 days . In contra~t, the BERH-2-B injected rats remained tumor-free for more than 180 days. While the four hybrid ~11 1;nl~q lost their ability to form tumors in syngeneic rats, they were able to grow and f orm tumors in nude mice .
In rats injected~ with hybrid BERH-2-B ~cells, there were abundant lymphocytic infiltrates at the site of inj ection . The tumor inf iltrating lymphocytes present at the site of BE~-2-B injection at two weekq were a combination of CD4~ and CD8~ T-cells. As shown by immunocytochemistry, moqt of the infiltrating cells were T-cells. Seventy percent were CD8' T-cells, and thirty percent were CD4~ T-cells. There was no inflammatory response in animals injected with parental BERH-2 tumors.
Exam~le 2: HePatocarcinoma:activated B-cell hvbrids can be used as a cellular vaccine to ~revent and cure he~atocarcinoma;
The experimental data cited above established that hepatocarcinoma cells lost tumorigenicity when fused to activated B-cells. Moreover, the findir,g that hybrid cells elicited a prolific T-cell response was consistent with an immunologic explanation for the loss of tumorigenicity. To substantiate this conclusion, tumor prevention and cure experiments were performed.
The f;nll;n~q detailed below show that rats injected with BERE-2-B hybrid cells became resistant to subsequent cha~lenge with parental BERH-2 cell6.
Furthermore, these experiments est~hl; ql'~d that BERH-2 hepatomas were cured by injection of BERH-2-B hybrid cells. Both CD4~ ar,d CD8~T cells were essential ior the induction of protective immunity. However, only CD8 T-cells were re~uired for ~:he eradication of pre-existing BERH- 2 tumors .
The rats immuni2ed with BERH-2-B hybrid cells were able to prevent tumor f ormation by parental BERH- 2 ~1 789~0 ~WO 95/16775 PCT/US94/14297 - cells. Protective immunity was induced with BERH-2-B
hybrid tumor cells.
Groups of female Wistar rats (8/group) were immunized with 2 x 106BERH-2-B or BERH-2 cells subcutaneouæly. Two weeks later, both groups of the rats were challenged with 5 x 106BERH-2 cells intrahepatically.
All ratæ pre-injected with BERH-2-B and subsequently challenged with BERH-2 remained tumor-free for more than 150 days. In contrast, all rats pre-injected with BERH-2 cells and then challenged with the same BERH-2 cells died within 60 days.
Exam~le 2A: Tumor cure ex~eriments A æeries of tumor cure experiments were next performed to show that ; i i~tion with BERH-2-B cells could also eradicate an established hepatoma. One set of fourteen rats were injected intrahepatically with 2 x 106 parental BERH-2 cells. Ten days later, eight of the injected rats were immunized with a subcutaneous injection of 5 x 106 BERH-2-B hybrid cellæ. Theæe ratæ survived for more than 120 days.
In contrast, ratæ injected both times with parental BERH-2 cells all died within 42 days. A second set of rats were surgically implanted with a small fragment of BER~1-2 hepatoma intrahepatically. Fourteen rats were intrahepatically irnr)l~nt~tl with a small fragment (0.3 mm x 0.5 mm) of BERH-2 tumor. Two days later, eight of the animals were inj ected subcutaneously with 5 x 106 - BERH-2-B hybrid cells. The other six rats were injected subcutaneously with same number of BERH-2 cells. Ten days later, a subset of the tumor-implanted animals were injected with BERH-2-B cells, the ~ ;n;ng control rats were injected with parental BERH-2 cells. Whereas all 6 rats injected with BER~-2 cells died within 50 days, only 2 of 8 rats injected with BERH-2-B hybrid cells developed tumors; the latter died at 71 and 74 days after tumor WO 95~16775 PCTiU594/14297 21 78q53 implantation, respectively. Six of the animals lived for more than 180 day8 after tumor implantation.
Example 2B: Determ;n~tion of type of T-cell mediation It was ne~t determined whether the rejection of BERH-2-B cells is mediated by CD4~ ar.d/or CD8~ T-cells.
Rats were depleted of CD4~ cells or CD8 ' cells by antibody treatment prior to injection of BERH-2-B cells. BERH-2-B
cells were able to form tumors in both CD4-depleted and CD8-depleted rats. The effects of depletion of CD4~ or CD8+ cells on the growth of BERH-2-B and BERH-2 tumor cells in vivo are discussed below.
~xi~m le 2C: Grou~s A-D
Female Wistar rats (Groups A, B, C, D) were treated with purified anti-rat CD4 (OX38), or anti-CD8 (OX-8) or ;~ control mouse anti-diethyl~ m;n~ pentaacetic acid monoclonal antibody. Each animal received 500 /lg of the purified antibody intravenously twice per week for three weeks.
Two days before injection of tumor cells, peripheral blood lymphocytes were obtained from indiviaual treated rats and stained with monoclonal antibodies to CD4 or CD8 to verify the depletion of CD4~ or CD8~ cells, respectively. Treatment with anti-CD4 monoclonal antibody depleted more than~ 9596 of the CD4 ' cells, and treatment with anti-CD8 monoclonal antibody depleted close to 9596 of the CD8 ' cells; treatment with control antibody did not alter the number of CD4~ and CD8~ cells. Three days after the last injection of the antibodies, all rats were injected intrahepatically with 5 x lO6 BERH-2-B tumor cells.
Exam~le 2D: Gro~lns E-G
Rats were first immuni_ed with BERH-2-B cell6 and then depleted of CD4* or CD8+ cells 14 days later.
Female Wistar rats (Groups E, F, G) were first; ;7f"l with 2 x lO6 BERH-2-B cells subcutaneously. Two weeks after; ;7i~tion, animals were treated with anti-CD4, or ~17895~
~Wo 95/16775 PCT/US9 ~/14297 anti-CD8 or control antibody, and the efectiveness of the depletions was verified by immuno1uorescence and flow - cytometry. Three days after the last injection of monoclonal antibody, all animals received 5 x 106 BERH-2 cells intrahepatically. These exp~riments have been repeated twice with comparable results.
Table 1 Effects of Depletion of CD4' or CD8+ Cells on the Growth of BERH-2-B and ~ -2 Cells In Vivo Treatment Protocol Number of Antibody Animals Speci~icity immunize f lrst With Ab treat f irst then treat Tumors then immunize with Ab None - 0/6 CD4 + 4 / 6 CD8 + 5/6 15Control Ab + 0/6 CD4 + 0 / 5 CD8 + 5/5 Control Ab + /5 These CD4-or CD8-depleted rats were then challenged with BERH-2 cells. Tumors developed in CD8-depleted, but not in CD4-depleted rats. This indicates that whereas both CD4' and CD8+ cells are necessary for the induction o~ protective anti-tumor immunity, once the immune response has been induced, CD8+ cells are sufficient alone to mediate tumor cell destruction alone. These results contrast with those reported previously for murine m.~l An~mA cells transfected with the B7 costimulator gene where CD4+ cells were not required for inductiol~ of the anti-tumor immune response (Chen, 71 5~1 1093, 1992;
Townsend, 259 Science 368, 1993).

WO 95/16775 PcrNS94114297 2~ 78950~ ~

ExamE~le 2E: ~umor-specificity of immunity It was next determined whether the immunity induced by BERH-2-B cells is tumor-specific. NBT-II is a bladder carcinoma that grow6 rapidly in sYngeneic Wistar rats. Immunization with BERH-2-B hyb~id cells prevented the growth of the parental BERH-2 cells. However, immunization with BERH-2-B was unable to inhibit~ the growth of NBT- II cells .
The specificity of the anti-BERH-2 immune response ~1 i r; tP~l ~ by BERH-2-B hybrid tumor cells was documented. Female Wistar rats were injected with 2 x lO6 BERH- 2 -B cells subcutaneously . Two weeks af ter l7~tjon, one -group of t=he rats were injected with 5 x lO6 BERH-2 cells intrahepatically. Another group oi rats were injected with 5 x 106 ~;IBT-2 rat bladder carcinoma cells (obtained frDm the American Tissue Type Collection) .
Tumor developed locally in the inj ected site in rats ; ~ 7C.(l with BERH-2-B tumor cells and challenged with NBT-2 tumor cells in all eight animals. All ahimals in this group died within 45 days aiter tumor challenge.
In addition, CD8~ T-cells from BERH-2-B-immunized rats killed BERH-2 cells but did not kill NBT-II cells vitro .
Table 2 Specif icity of the Immune Response Elicited bY BERH-2-B Hvbrid Tumor Cells Number oi Animals Immunization ~ ~h~ nge With Tumors BERH-2-B ~ NBT-II 8/8 3 o Exam~le 2F, NecessitY of in vitrQ 8election It was-determined whether the ;n vitro 8election step for hybrid cells is rer~uired for effective induction of anti-tumor immunity.
_ _ ~Wo 95/16775 2 1 7 8 9 5 ~ Pc~r/usg4/l4297 Tumor protective immunity inducea with BERE~-2 tumor cells fused with activated B cells does not require in vitrQ selection. Three groups of rats ~8/group) were injected subcutaneously with 5 x 106 BERE~-2 cells, 5 x 106 BERH-2 cells mixed with 5 x 106 activated B cells, or 5 x 106 BERH-2 cells fused with 5 x 106 activated B cells in the presence of PEG. Fused cells were washed three times with PBS, resuspended in PBS and inj ected subcutaneously .
Two weeks later, all groups of rats were challenged with 5 x 106 BERH-2 tumor cells ; ntr~hf~ratically BERX-2 tumor cells were fused with activated B-cells. After fusion, cells were washed and injected into syngeneic rats with in ~itro selection. The efficiency of the fusion in such experiments ranged form 30% to 50%. As controls, BERX-2 tumor cells mixed with activated B-cells without PEG were injected subcutaneously. All animals were then challenged with the parental BERX-2 cells intrahepatically .
The data indicated that fused cells were immunogenic in the absence of in vitro selection. Only animals ; ; 7f"l with tumor cells fused with activated B-cells were able to reject the parental tumor cells;
simply mixing tumor cells with activated B-cells was not effective. This finding that protective immunity can be induced by tumor cells fused with activated B-cells without an in vitro selection step simplifies the clinical therapeutic application of this method.
- These findings indicate that an effective BERX-2 hepatocarcinoma-specific vaccine can be generated by fusing tumor cells with syngene~c, activated B-cells. It is believed that in addition to class II MHC and B7 costimulator, hybrid BERH-2-B cells may express oth~=
molecules that are essential for the activation of anti-tumor T-cells. This may include, but is not limited to, soluble cytokines. Production of cytokines by hybrid _ _ _ _ _ _ _ _ _ _ _ _ _ . . .. . . . _ r WO 9~/16~75 ~ ~ 7 ~ 5 ~ PCTIUS9.1/14297 tumor cells may be important in the elicitation of host immune responses.~
Example 3: GlYcos~l~hos~hatid~linositol ~"GPI")-modified c~tokine can function as an- 2trtificial ~lh~cin For purpoæes of achieviny high level stable expression of cell surface molecules on human cellula~
transfectants, self-replicating EBV episomal expression vectors were employed (GrQger et al, 81 Gene 285, =1989).
In the present study, two EBV vector variants designated pREP4~ and pREP7,B were used, both of which share a transcriptional cassette in which the RSV 3' LTR promoter, a multiple cloning site, and the SV40 late polyadenylation/termination signal are linked in tandem~
An expression construct for M-CSF-GPI was generated as f ollows ~ The 1. 8 kb M- CSF coding region fragment (XhoI - EcoRI) of p3ACSFR1 (Fig. 1) was inserted into the corresponding sites of pBluescript (pBT, Stratagene, Inc. ) . This generated a GPI-anchored variant of M-CSF, pM-CSF/BT was cut with ~L, filled-in with the pollk, and subsequently cut with BamHI. The 3 ' AvaII
(filled-in) to BamHI from the DAF subclone pDF2.1/BT was subcloned into this vector~ to generate an in-frame M-CSF-DAF chimeric se~uence. The KpnI-BamXI fragment of the resultant plasmid containing the chimeric sec~uence was subcloned into- the corresponding sites of pREP4Y to generate pM-CSF-GPI/REP4~.
An M-CSF receptor EBV episomal expression construct was generated as follows: A 4 ~ Okb EcoRI
fragment of pc-~f~ 102 was subcloned into the EcoRI site of ~Bluescript (Stratagene, Inc~) to generate pM-CSFR/BT~
The 3.6 kb ~XI fragment=of this subclone, containing the entire M-CSFR ~coding region, was subcloned in a sense orientation into the ~HI site of pREP7~3 to generate pM-CSFR/REP7 ~ .
The overall experi~Lental strategy of t~is study was to use stable gene transfer to modify the ~=adhes`ive . . .

21 7~950 o 95116775 PCTIUS94/14297 - properties of cells. The paired cellular targets chosen were the human SV40 large T-immortalized bone marrow stromal cell line KM-102 and the human myeloid leukemia cell li~e K562. Previous work with both the KM-102 and K562 lines have shown them to be eficient transfection targets with EBV expression vectors.
With the goal of ~nh~nr; n,r adhesion between KM-102 and K562 cells, we stably transfected them with episome-based expression constructs for an artificial GPI-modified variant of M-CSF, designated M-CSF-GPI, in which natural M- CSF coding sequence is linked in- f rame to the GPI signal sequence of human DAF, into KM-102 stromal cells . Indirect immunof luorescence staining demonstrated a high level surface expression of M-CSF epitopes on the pM-CSF GPI/REP4o! hyg~KM-102 transfectants. No M-CSF
epitope was detectable on the surface of nontransfected KM-102 cells or on KM-102 cells t-ansfected with the irrelevant EBV episome pRSVCAT~/22 0 . 2 .
pM-CSFR/REP7,(~ and pM-CSF GPI/REP4~r were introduced into K562 and U937 cells, respectively, by lipofection. The KM-102 stromal cell line was kindly provided by K. Harigaya and m~;n~{l;n~r~ in McCoy's 5a medium ~Gibco, Inc. ) supplemented with 109~ heat-inactivated fetal bovine serum (FBS) (M.A. Bioproducts) /10 mM HEPES/40 ~lg/ml gentamycin sulfate in a humid 5% CO~
atmosphere at 37C. pM-CSF-GPI/REP4(Y was introduced into KM-102 cells by lipofection. Briefly, cells were grown to 5096 confluence in six-well plates, and washed twice with PBS and once with opti-MEM (Gibco) . Cells were then incubated for 5-8 hours at 37C with 1 ml opti-MEM, cnn~;ning 10 ~g DNA and 30 /lg lipofectin, before adding 1 ml of complete medium rnnt~ining 2096 FBS.
Seventy-two hours post-transfection, selection for stable transfectants was begun by replacing the medium with fresh medium cr,n~in;ng 75 ~Lg/ml hygromycin B
(Calbioehem, Inc. ) Stably hygR transfeeted eolonies were _ _ . _ _ _ . _ _ _ _ _ _ WO 95/16775 PCTr~594/14297 picked at 2-3 weeks using cloning rings (Bellco, Inc. ) n~ d and main~ained in 100 llg/ml hygromycin B.
Surface M-CSF expression was established by indirect immunofluorescence using a polyclonal rabbit anti-M-CSF
antibody (Genzyme) and a FITC-anti~rabbit IgG (BMB) secondary antibody. M-CSFR expre~sion was confirmed by indirect i~munofluorescence .and flow cytometry (FACS), using a rat monoclonal anti-M-CSFR primary antibody (Oncogene Sciences) and FITC-conjugated rabbit anti-rat IgG secondary antibody (Miles ICN) .
In parallel, X562 leukemic cells were transfected with human M-CSF receptor (M-CSFR or c-fms) episomal expression construct, designated pM-CSFR/REP7~.
Abundant sur~ace expression o~ natural human M- CSFR was demonstrated by indirect immunostaining and ~low cytometry. To control for ~ episomal transfectants demonstrating non-specific up-regulation of ~ their endogenous adhesion molecules, K562 cells were transfected with the irrel evant episome pRSVCATo!/220 . 2 .
To con~Eirm that pM-CSF-GPI/REP4~Y yields a GPI
membrane anchored product, phosphatidylinositol-specific phospholipase C (PIP~C), an enzyme which specially cleaved GPI moieties made by certain cells f rom their surf aces, was used. ~IPLC treatment o~ KM-102 cells transfected with pM-CSF-GPI/REP4c~ did not result in a significant release of either M-CSF or DAF, another GPI-anchored protein serving as control, f rom the surf ace, as determined indirect immunofluorescence. This suggested that KM-102 is similar to some other cell types that are known to express~ a GPI anchor variant that is resistant to PIP~C cleavage.
PIPI.C cleavage was performed by incubating 1 x 106 cells with 1 unit of PIP~C ~Boehrin~er Mannheim Biochemicals) at 37C for one hour in RPMI 1640 medium containing 10~6FBS and 0 . 01!'s sodium azide . Cleavage was assessed using an anti-DAF antibody and ~low cytometry.

~Wo 95/16775 3 PCr/US94/1429~
In light of this f inding, pM- CSF-GPI/RBP4~ was additionally transfected into the U937 cell line which is known to produce PIPLC-sensitive GPI anchors. PIPLC
susceptibility of M-CSF and endogenous DAF on these cells was assessed. Nontransfected U937 cells did not express M-CSF on their surface. The pM-CSF-GPI/REP40! U937 transfectants expressed high levels of cell surface GM-CSF . The tethered M- CSF could be specif ically cleaved with PIPIIC to an extent ~similar to that seen with endogenous GPI-anchored DAF protein. The PIP~C cleavage yielded removal of surface M-CSF epitope to the same extent as DAF. This indicates that the pM-CSF-GPI/RBP4 construct generates a GPI-anchored form of M-CSF.
It was next de~rm;nf~fl whether the pM-CSF-GPI/RBP40~ KM-102 transfectants would bind preferentially M-CSFR' c~ r targets. Cell:cell binding between adherent KM-102 cells and M-CSFR~ n-~n~r~h~orent K562 targets cells was enhanced approximately three-fold when M-CSF-GPI
was presen~ on the surface of the KM-102 cells.
Intercellular adhesion between adherent KM-102 transfectants and n-~n~ rent K562 transfectants was measured using a cell:cell binding assay. Normal KM-102 cells (nontransfected), a stable transfectant expressing GPI-anchored M-CSF ~pM-CSF-GPI/REP40~), a control transfectant expressing the lymphoid cell surface molecule (pCD8/RBP2 . 1), or an irrelevant transfectant (pcYIL-6/RBP5.1) were separately seeded into wells of a polyvinyl, flat-bottom 96-well plate. 35S-labeled K562 target cells, eIther nontransfected (none) or expressing the M-CSF receptor (pM-CSFR/RBP7~) or Cl~T (pRSVCAT/220.2), were added to the wells and allowed to bind to the KM-102 cells. The number of K562 target cells which remained adherent following an - inverted centrifugation was calculated by measuring the speclf ic activity of the target cells added. A signi~icant increase in adherence between pM-CSFR-REP4~ transfected KM-102 and pM-CSFR-REP,B7 _ _ . .. . . . . .. . . ..

WO 95/16775 PCrlUS94/14297 217&950 transfected KM-102 was noted whereas no increase in adhesion was noted for other combinations.
This was specific for M-CSF expression, since control KM-102 transfectants rrnt~;n~ng either pCD8/REP2.1 encoding the irrelevant surf ace protein CD8 or alternatively p~Y~I,-6/REP5.1~ driving antisense II.-6 RNA
expression, demonstrated no ,~nll~n~rPrl binding to M-CSFRt targets. Moreover, no significant augmentation of adhesion was evident when nontransfected K562 cells, or K562 cells transected with pRSVCAT~/220.2 were used as cellular targets. Xowever, ~K562 transfectants overall do appear to be slightly more~ adhesive than nontransfected cells .
To def initively establish that it is the membrane-associated M-CSF, and not secondary expressed surface molecules on the pM-CSF-GPI/REP4~x KM102 transfectants, that was specifically responsible for the enhanced adhesion, antibody blocking analysis was performed. Prior incubation of M-CSFRt K562 target cells with antibodies directed against the M-CSFR or alternatively, o~ surface M-CSFt KM102 cells with polyclonal anti-M-CSF ~n~;ho~ , each partially inhibited this specif ic cellular interaction .
The ef ~ects o~ blocking antibodies upon the bindiny of M-CSF-GPI and M-CSFR-positive cells to each other was assessed in cell:cell binding ~assays.
Nontransfected K562 cells (none) or K562 cells stably expressing the M-CSFR (pM-CSFR/REP7~) or CAT
(pRSVCAT/220.2) were allowed to bind to KM-102 cells expressing GPI-anchored M-CSF (pM-CSF-GPI/REP40~) .
Blocking antibodies were added to the appropriate cells, as indicated, 3 0 min . prior to adding the cells to the wells. Anti-M-CSF, polyclonal anti-M-CSF antibody; anti-M-CSFR, monoclonal anti-M-CSF receptor (c-fm~) antibody;
an anti-TRF, monoclonal anti-transferrin receptor . . . _ _ ~17~S~
~WO 95/16775 PcrNss4/l4297 antibody. Normal rabbit serum (NRS) was added, as indicated, to prevent F receptor cross-linking.
The simultaneous addition of these two blocking antibodies, directed against both members of the ligand:receptor pair, completely blccked the specific binding. Normal rabbit serum, which did not inhibit binding, was included in all eXpAr;rAntA to prevent the cross-linking of cells through Fc receptors expressed on the K562 cells. The antibody-mediated inhibition observed was specific for the M-CSF :M-CSFR pair, since antibodies against the human transferrin receptor - (TFR), known to be expressed on K562 cells, had no blocking effect.
Cell:cell binding assays employed a modification of a published method (McClay et al., 78 Proc. Natl. Acad.
Sci ~SA 4975, 1981). Briefly, 3 x 10~ nontransfected or transfected KM-102 cells were placed in wells of a polyvinyl, flat-bottom 96-well plate (Dynatech ~aboratories) with 0.1 ml complete medium per well, and the cells were incubated at 37C for two days. Wells were pretreated with fetal bovine serum for two hours to promote attachment of the KM-102 ce~ l 8 . The KM-102 cells were generally 60-809~ confluent at the time of the cell:cell binding assay. The K562 target cells, labeled with 35S-methionine, were washed, resuspended in complete medium at 5 x 105 cells per ml, and 0.1 ml was added directly to the wells.
The plates were incubated at 37C for 2.5 hours to allow for maximal binding. Medium was added to each well to produce a positive meniscus, and then plates were - 30 carefully sealed with adhesive plate sealers (Dynatech I-abs, Inc. ) . The plates were inverted and centrifuged for 10 minutes at room temperature using Sorvall micro-plate carriers. A relative centrifugal force (RCF) of 900 x g was used for most experiments. Post-centrifugation, still inverted plates were flash-frozen at -80C, and the bottoms of each well, containing the stromal cells and WO g5116775 PCTIUS94114297 2 1 7gq50 34 bound targets, were cut off and placed in s~;nt;ll;3tion vials for counting. The number of K562 cells bound per well was calculated as follows:
CPM BOUND TO STROMAL CELLS
CELLS BOII~D/WELL = x (5 x l0~) TOTAL CPM ADDED/WELL
Each group represents the means of at least triplicate samples. Representative KM-102 stromal cell wells were harvested and counted to ensure that e~[uivalent l0 cell numbers were present in each well.
For antibody blocking experiments, antibodies were added directly to cell suspensions or to wells (as indicated) and incubated ~at room temperature ior 3 0 minutes just prior to thç addition of target cells to the wells. The final concentratiors of antibodies used in the blocking studies were: rabbit anti-human M-CSF polyclonal antibody (Genzyme), l011g/ml ; rat anti-c-~ma/CSF-l receptor (Oncogene Science, Inc. ), 2 ~Lg/ml; and mouse monoclonal anti-human transferrin receptor (Hybritech, Inc.) 8 ~Lg/ml.
Xeat-inactivated normal rabbit serum (4 ~l/l x l05 cells) was added in each case to prevent the cross-linking of cells through Fc receptors.
Hence, in this example, a method for selectively altering inter~ r adhesion has been demonstrated.
Spe~;f;~lly, it has shown that an artificial GPI-modified variant of a model cytokine M-CSF, when anchored at the cell surface, can augment c~ r binding to M-CSF
receptor-bearing~ tumor cells. High level expression of both the M-CSF-GPI and M-CSFR molecules on their respective cells could be efficiently obtained via gene transfer using- episomal_ expression vectors. The M-CSF-dependence of the effect was verified by anti-M-CSF
and anti-M-CSF receptor antibody blocking. Clearly, the invention is nct limited to these variants since GPI-anchored variants of multiple ligands other than M-CSF
could be used as alternative artif icial adhesins .

~WO95/16775 2 l 7 8 9 50 PCT/US9~/14297 In the above experiment, only the M-CSF
component of the M-CSF:M-CSFR pair was GPI-anchored. The M-CSFR cl~prn~nt could also be GPI-anchored. Together, this would permit coating one cell with a ligand:GPI
(e.g., M-CSF:GPI) chimera and a second cell with a receptor:GPI (e.g., M-CSF~:GPI) chimera and thereby selectively ~nl~nr;ng adhesiveness between the respective cells via protein transfer. Such a use of a ~'disabled~
GPI-modified receptor, as part of an artificial adhesin pair, would obviate the possibilities for unwanted signaling through the receptor. Although there is evidence for cis signaling through certain natural GPI-anchored proteins, it seems likely that most cytokine receptors, when artificially GPI anchored, will not be functionally responsive to their corresponding cytokines due to distortions in molecular topology.
The pre-adhesion step entails the mixing of three components, namely, a tumor cell, a second cell with greater immunogenic potential, and artif icial adhesin molecules. After a co-incubation period, generally lasting greater than 2 0 minutes, the mixture is centrifuged at 400g for 5-l0 minutes in d~L~ ate tissue culture medium supplemented with antibiotic and resuspended in the same medium. The cell suspension is kept at room temperature for 30 minutes with gentle stirring. Subsequently the mixture i8 spun through a sterilized isotonic sucrose ~320 mOsm) solution r~-nti:lining 2 mM sodium phosphate buffer, pH 7.2 (400g for 5-l0 minutes). The cell pellet comprising tumor cell:second cell conjugates is gently suspended in l ml of isotonic sucrose buffer.
Following pre-adhesion using artifical adhesins and subsequent enrichment f or heterologous cell conjugates, fusion is carried out. A preferred method for fusing the heterologous cells in conjugates is electrofusion. Methods for performing electrofusion are Wo 95/16775 PCTIUS94/14297 21 78q50 ~ -well-known to those f;~ r with the art. Of note, most electrofusion protocols comprise two-steps: the induction of membrane-membrane contact followed by application o~
the fusogenic pulse. The dielectrophoresis method is u3ed in most experiments to achieve t~e f irst step of congregating cells jbefore fusion-;n~ f ;n~ electric pulses are applied. Xowever, it is known that the two steps can be dissociated, that is, one need not congregate cells by dielectrophoresis in order to achieve fusion (Sowers, 220 Methods in Enzvmoloqv 196, 1993). A preferred method according to the present invention entails omitting the dielectrophoresis step for congregating cells non-selectively, and instead applying the fusogenic pulse to heterologous cell l conjugates that have been pre-adhered with artificial i~hF~C;nR The ~usogenic pulse can be applied with a high voltage generator, and altenlative instruments are coTnmercially available. Though it is preferable to use con~ugates at low conjugate densities (less than 10~ per ml), one can readily go up to 107 per ml or higher in practicing this invention. Typical electrofusion coIlditions use five square wave pulses of 2 . 5 kV/cm and 5 microsecond duration, at a controlled temperature of 35C and with intervals of 15 sec between pulses to permit dissipation of Joule heat. Optimization of fusion conditions icr particular cell types can be readily performea. (~ess optimal ~ ;nc are~those described by Tsong and Tomita, 220 Methods in EnzYmoloq~
238, 1993).
~hile the present invention has been described in conjunction with the preferred embodiments and examples, one of ordinary skill, af ter reading the foregqing specification, will be able to effect various changes, substitutions ~or e~uivalents and = other alterations to the invention provided herein. It is therefore intended that the protection granted by letters patent hereon be limited only by the definitions c-~nt~;n,o-l .

WO 95/16775 PCl'NS94/14297 in the appended claims, and equivalents thereof. It will be understood that changes may be made and the details o~
~ormulation without departing :~rom the spirit o~ the invention as de~ined in the ~ollowing claims.

Claims (93)

CLAIMS:

1. A method for inducing an immune response against a first tumor cell comprising the step of providing to a patient a cell fusion product, wherein said cell fusion product comprises a membrane from a second tumor cell which can be derived from the same tumor as said first tumor cell fused to a membrane from a conventional antigen-presenting cell.

2. The method of claim 1, wherein said cell fusion product comprises a hybrid cell, wherein said hybrid cell comprises a second tumor cell fused to said conventional antigen-presenting cell.

3. The method of claim 1, wherein said step of providing comprises:
forming said cell fusion product ex vivo; and administering said cell fusion product to said patient.

4. The method of claim 1, wherein said step of providing comprises forming said cell fusion product in vivo.

5. The method of claim 1, wherein said step of providing comprises directly injecting into a tumor in said patient said conventional antigen-presenting cell.

6. The method of claim 1, further comprising the step of obtaining said second tumor cell from said patient.

7. The method of claim 6, further comprising the step of culturing said tumor cell.

8. The method of claim 7, further comprising the step of fusing said membrane from said second tumor cell with said membrane from said conventional antigen-presenting cell.

9. The method ofz claim 1, wherein said second tumor cell is autologous.

10. The method of claim 1, wherein said second tumor cell is a primary tumor cell.

11. The method of claim 1, wherein said second tumor cell is a metastatic tumor cell.

12. The method of claim 1 wherein said antigen-presenting cell is selected from the group consisting of an activated B-cell, a dendritic cell, a macrophage, an activated T-cell and an endothelial cell.

13. The method of claim 1, wherein said antigen-presenting cell is modified to alter its immunogenicity.

14. The method of claim 13, wherein said modification comprises expression of a foreign protein in said antigen-presenting cell.

15. The method of claim 13, wherein said modification is accomplished by gene transfer.

16. The method of claim 13, wherein said modification is accomplished by protein transfer.

17. The method of claim 13, wherein said modification comprises inhibition of a protein in said antigen-presenting cell.

18. The method of claim 17, wherein said protein is selected from the group consisting of a major histocompatibility complex protein and a coinhibitor.

19. The method of claim 14, wherein said foreign protein is selected from the group consisting of a cell surface costimulator, a soluble cytokine, a selectin, an adhesin, and a major histocompatibility complex protein.

20. The method of claim 1, wherein said second tumor cell or conventional antigen-presenting cell is modified to enhance its fusion potential by expressing an adhesin on either of said cells.

21. The method of claim 20, wherein said adhesin is expressed on both of said cells.

22. The method of claim 1 or 21, wherein said adhesin is an artificial adhesin.

23. The method of claim 22, wherein said artificial adhesin is a glycosyl-phosphatidylinositol-modified polypeptide.

24. The method of claim 22, wherein said artificial adhesin comprises a biotin-lipid conjugate and a chimeric streptavidin polypeptide.

25. The method of claim 1, wherein said cell fusion product is made using a chemical fusogen.

26. The method of claim 25, wherein said chemical fusogen is polyethylene glycol.

27. The method of claim 1, wherein said cell fusion product is made using electrofusion.

28. The method of claim 1, wherein said second tumor cell is a heterologous tumor cell.

29. The method of claim 28, wherein said heterologous tumor cell is obtained from a cultured cell line.

30. The method of claim 28, wherein said heterologous tumor cell is a primary tumor cell.

31. The method of claim 28, wherein said heterologous tumor cell is a metastatic tumor cell.

32. The method of claim 1, wherein said membrane from said second tumor cell is derived from an autologous tumor cell.

33. The method of claim 1, wherein said membrane from said second tumor cell is derived from a heterologous tumor cell.

34. The method of claim 1, wherein said second tumor cell is treated prior to fusion to reduce its ability to proliferate.

35. The method of claim 3, wherein said cell fusion product is treated to reduce its ability to proliferate prior to administration to said patient.

36. The method of claim 34 or 35, wherein said treatment is irradiation.

37. The method of claim 5, wherein said second tumor cell of said tumor is modified in vivo prior to injection of said conventional antigen-presenting cell.

38. The method of claim 37, wherein said modification comprises expressing on said second tumor cell a molecule that promotes adhesion to said antigen-presenting cell.

39. The method of claim 5, wherein said conventional antigen-presenting cell is modified in vitro prior to injection.

40. The method of claim 39, wherein said modification comprises expressing on said conventional antigen-presenting cell a molecule that promotes adhesion to said second tumor cell.

41. The method of claim 5, wherein said second tumor cell or said conventional antigen-presenting cell is modified by gene transfer.

42. The method of claim 5, wherein said second tumor cell or said conventional antigen-presenting cell is modified by protein transfer.

43. The method of claim 1, wherein said cell fusion product comprises more than two cells.

44. A method for inducing an immune response against a first tumor cell, comprising the step of providing to a patient a T-cell activated by contact with a cell fusion product, wherein said cell fusion product comprises a membrane from a second tumor cell which can be derived from the same tumor as said first tumor cell fused to a membrane from a conventional antigen-presenting cell.

45. The method of claim 44, further comprising the step of immunoselecting a T-cell subset to provide to a patient.

46. The method of claim 45, wherein said T-cell subset consists of CD8-positive T-cells.

47. A method for inducing an immune response against a first tumor cell comprising the step of providing to a patient a second tumor cell which can be derived from the same tumor as said first tumor cell fused by an artificial adhesin to an activated B-cell.

48. A method for identifying a cell that fuses to a tumor cell and thereby provides a cell fusion product with immunogenicity greater than said tumor cell, comprising the steps of:
fusing said tumor cell with a candidate cell;
and determining the immunogenicity of the resulting cell fusion product.

49. The method of claim 48, wherein determining the step of immunogenicity of said cell fusion product is performed by administering said cell fusion product to an experimental animal.

50. A method for identifying a cell that fuses to a tumor cell and thereby provides a cell fusion product that induces an anti-tumor immune response comprising the steps of:
fusing said tumor cell with a candidate cell;
and determining the anti-tumor immune response of the resulting cell fusion product.

51. A method for fusing a tumor cell with a conventional antigen-presenting cell comprising the steps of:
expressing on either one of said tumor cell or said conventional antigen-presenting cell an artificial adhesin that increases adhesion between said tumor cell and said conventional antigen-presenting cell; and fusing said cells.

52. The method of claim 51, wherein said artificial adhesin on one of said cells contacts a natural cell surface molecule on the other of said cells.

53. The method of claim 51, wherein said artificial adhesin on one of said cells contacts an artificial adhesin on the other of said cells.

54. The method of claim 51, wherein expression of said artificial adhesin is accomplished by protein transfer.

55. The method of claim 51, wherein expression of said artificial adhesin is accomplished by gene transfer.

56. The method of claim 51, wherein said step of fusing said cells is performed at cell densities below 108 cells per milliliter.

44a

57. An immunogenic cell fusion product comprising a membrane from a tumor cell fused to a membrane from a conventional antigen-presenting cell, but wherein said cell fusion product does not comprise a hybridoma cell.

58. The method of claim 57 wherein said antigen-presenting cell is selected from the group consisting of an activated B-cell, a dendritic cell, a macrophage, an activated T-cell and an endothelial cell.

59. The immunogenic cell fusion product of claim 57, wherein said fusion is achieved using a heterobifunctional antibody.

60. The method of claim 1, wherein said conventional antigen-preventing cell is a syngeneic antigen-presenting cell.

61 The method of claim 44 wherein said conventional antigen-preventing cell is a syngeneic antigen-presenting cell.

62. The cell of claim 57 wherein said conventional antigen-preventing cell is a syngeneic antigen-presenting cell.

63. A pharmaceutical composition comprising a cell fusion product which comprises a membrane from a tumor cell fused to a conventional antigen-presenting cell, and a pharmaceutically acceptable diluent or carrier.

64. A pharmaceutical composition according to claim 63 wherein the conventional antigen-presenting cell is selected from the group consisting of an activated B-cell, a dendritic cell, a macrophage, an activated T-cell and an endothelial cell.

65. A pharmaceutical composition according to claim 64 or claim 65 wherein the cell fusion product is a hybrid cell comprising a tumor cell fused to the antigen-presenting cell.

66. A pharmaceutical composition according to claim 65 wherein the tumor cell is autologous.

67. A pharmaceutical composition according to claim 65 wherein the tumor cell is heterologous.

68. A pharmaceutical composition according to claim 67 wherein the heterologous tumor cell is obtained from a cultured cell line.

69. A pharmaceutical composition according to any of claims 65 to 67 wherein the tumor cell is a primary tumor cell.

70. A pharmaceutical composition according to any of claims 65 to 67 wherein the tumor cell is a metastatic tumor cell.

71. A pharmaceutical composition according to claim 63 or claim 64 wherein the tumor cell membrane is derived from an autologous tumor cell.

72. A pharmaceutical composition according to claim 63 or claim 64 wherein the tumor cell membrane is derived from an heterologous tumor cell.

73. A pharmaceutical composition according to any of claims 63 to 72 wherein the antigen-presenting cell is modified to alter its immunogenicity.

74. A pharmaceutical composition according to claim 73 wherein the modification comprises expression of a foreign protein in the antigen-presenting cell.

75. A pharmaceutical composition according to claim 74 wherein the expressed foreign protein is selected from the group consisting of a cell surface costimulator, a soluble cytokine, a selectin, an adhesin, and a major histocompatibility complex protein.

76. A pharmaceutical composition according to any of claims 73 to 75 wherein the modification is accomplished by gene transfer.

77. A pharmaceutical composition according to any of claims 73 to 75 wherein the modification is accomplished by protein transfer.

78. A pharmaceutical composition according to claim 73 wherein the modification comprises inhibition of a protein in said antigen-presenting cell.

79. A pharmaceutical composition according to claim 78 wherein the protein is selected from the group consisting of a major histocompatibility complex protein and a coinhibitor.

80. A pharmaceutical composition according to any of claims 69 to 79 wherein either or both of the tumor cell or the conventional antigen-presenting cell is modified to enhance its fusion potential by expressing an adhesin.

81. A pharmaceutical composition according to claim 80 wherein the adhesin is an artificial adhesin.

82. A pharmaceutical composition according to claim 81 wherein the artificial adhesin is a glycosyl-phosphatidylinositol-modified polypeptide.

83. A pharmaceutical composition according to claim 81 wherein the artificial adhesin comprises a biotin-lipid conjugate and a chimeric streptavidin polypeptide.

84. A pharmaceutical composition according to any of claim 63 to 83 wherein the cell fusion product is treated to reduce its ability to proliferate.

85. A pharmaceutical composition according to claim 84 wherein the cell fusion product is irradiated.

86. A pharmaceutical composition according to any preceding claim wherein the antigen-presenting cell is syngeneic.

87. A pharmaceutical composition comprising T-cells activated by contact with a cell fusion product, wherein said cell fusion product comprises a membrane from a second tumor cell which can be derived from the same tumor as said first tumor cell fused to a membrane from a conventional antigen-presenting cell.

88. A pharmaceutical composition according to claim 87 wherein the T-cells are a T-cell subset.

89. A pharmaceutical composition according to claim 88 wherein the T-cell subset consists of CD8-positive T-cells.

90. A pharmaceutical composition according to any of claims 87 to 89 wherein the antigen-presenting cell is syngeneic.

91. The use of a conventional antigen-presenting cell modified to express a molecule that promotes adhesion to a tumor cell in the manufacture of a medicament for the treatment of cancer.

92. The use according to claim 91 wherein the conventional antigen-presenting cell is modified by gene transfer.

93. The use according to claim 91 wherein the conventional antigen-presenting cell is modified by protein transfer.