WO1996033746A1

WO1996033746A1 - Treatment of tumors with cells expressing interferons, tumor necrosis factors, or other cytokines

Info

Publication number: WO1996033746A1
Application number: PCT/US1996/006054
Authority: WO
Inventors: David L. Ennist; Yawen Chiang; Suzanne Forry-Schaudies
Original assignee: Genetic Therapy, Inc.
Priority date: 1995-04-28
Filing date: 1996-04-24
Publication date: 1996-10-31
Also published as: AU5634696A

Abstract

A method of treating a tumor in a host which comprises administering to the site of the tumor non-tumor cells which express one or more agents selected from the group consisting of an interferon, a tumor necrosis factor, and GM-CSF. When the agent is an interferon or a tumor necrosis factor such method provides for a local, non-immune mediated anti-tumor response. The cell which expresses the interferon or tumor necrosis factor and which is administered to the site of the tumor also may be administered, in one embodiment, in conjunction with (a) a non-tumor cell that is engineered with a polynucleotide encoding a cytokine; and (b) tumor cells of the type of tumor which is being treated, in order to provide an immediate, non-immune mediated anti-tumor response, followed by an immune response against the tumor.

Description

TREATMENT OF TUMORS WITH CELLS EXPRESSING INTERFERONS, TUMOR NECROSIS FACTORS, OR OTHER CYTOKINES

This invention relates to the treatment of tumors . More particularly, this invention relates to the treatment of tumors by administering to the host at the site of the tumor cells (e.g., fibroblasts) , which express an interferon or a tumor necrosis factor or granulocyte macrophage colony- stimulating factor (GM-CSF) . This invention also relates to treating a tumor by administering to a host a non-tumor cell engineered with a polynucleotide encoding a cytokine and a tumor cell. BACKGROUND OF THE INVENTION

Efforts have been undertaken to treat tumors by inducing immune responses against the tumors. Glick, et al . , Neurosurgerv. Vol. 36, No. 3, pgs. 548-555 (March 1995) disclose the administration of fibroblasts genetically engineered to secrete cytokines to mice in order to induce antitumor immunity to a murine glioma in vivo. Tahara, et al . , Cancer Research, Vol 54, pgs. 182-189 (January 1, 1994) disclose the administration to mice of fibroblasts genetically engineered to secrete Interleukin-12 to suppress tumor growth and induce antitumor immunity to a murine melanoma. Lotze, et al . , Human Gene Therapy, Vol. 5, pgs. 41-55 (1994) , discloses a clinical protocol for gene therapy of cancer by administering fibroblasts engineered with the Interleukin-4 gene, and admixed with autologous tumor in order to elicit an immune response.

Dranoff, et al . , Proc. Nat. Acad. Sci .. Vol. 90, pgs. 3539-3543 (1993) , disclose the vaccination of mice with irradiated B16 melanoma cells engineered with a gene encoding murine granulocyte macrophage-colony stimulating factor. Such vaccination stimulates anti-tumor immunity. Gansbacher, et al . , Cancer Research. Vol. 50, pgs. 7820-7825 (December 15, 1990) discloses the injection of mice with CMS-5 tumor cells transduced with retroviral vectors including the Interferon-γ gene. The Interferon-γ producing cells induced a long lasting state of T-cell immunity, as judged by rejection of CMS-5 tumor challenge and persistence of specific cytotoxic activity in the spleen cell population. Ferrantini, et al . , Cancer Research. Vol. 53, pgs. 1107-1112 (March 1, 1993) , teach the injection of Friend Leukemia Cells transduced with a retroviral vector including the Interferon- αl gene into mice. Mice that were given injections of such cells developed a long-lasting tumor-specific immune resistance to subsequent injection with highly metastatic Friend Leukemia Cells.

Although the above disclosures teach that an immune response can be generated against a tumor, such an immune response is not an immediate response, but rather is generated over a matter of days. Thus, it would be desirable to generate an immediate anti-tumor effect independent of an anti-tumor response generated by the immune system. It also would be desirable to combine such an anti-tumor response with an anti-tumor response generated by the immune system. BRIEF DESCRIPTION OF THE DRAWINGS

The invention now will be described with respect to the drawings, wherein:

Figure 1 is a schematic of the construction of plasmid PG1; Figure 2 is the sequence of the multiple cloning site in pGl;

Figure 3 is a map of plasmid pGl;

Figure 4 is a map of plasmid pBg;

Figure 5 is a map of plasmid pN2;

Figure 6 is a map of plasmid pGlNa;

Figure 7 is a map of plasmid pLNSX;

Figure 8 is a map of plasmid pSvNa;

Figure 9 is a map of plasmid pGlXSvNa;

Figure 10 is a map of plasmid pGlFlSvNa;

Figure 11 is a map of plasmid pGlF2SvNa;

Figure 12 is a map of plasmid pGlF31SvNa;

Figure 13 is a map of plasmid pGlmF3SvNa;

Fi-gure 14 is a graph of the survival times of mice inoculated with parental B16F10 cells, or with B16F10 cells transduced with B16F10/GlNa or B16F10/GlF2SvNa;

Figure 15 is a graph of the tumor volumes of mice treated with parental B16F10 cells; B16F10/GlF2SvNa cells; B16F10/GlmF3SvNa cells; or a mixture of B16F10/GlF2SvNa cells and B16F10/GlmF3SvNa cells;

Figure 16 is a graph of the anti-B16F10 CTL response from mice bearing a primary B16F10 tumor, and in mice inoculated with B16F10/GlF2SvNa cells or B16F10/GlmF3SvNa cells prior to secondary exposure with parental B16F10 cells;

Figure 17 is a graph of the survival over a 70-day period of mice inoculated with Renca tumor cells and BALB3T3/GlNa, BALB3T3/GlF2SvNa, or BALB3T3/GlmF3SvNa fibroblasts;

Figure 18 is a graph of the survival over a 180-day period of mice inoculated with Renca tumor cells and BALB3T3/GlNa, BALB3T3/GlF2SvNa, or BALB3T3/GlmF3SvNa fibroblasts;

Figure 19 is a map of pGlmGmSvNa; Fi-gure 20 is a graph of the tumor size in mice injected with Renca cells; Renca cells and PA317/GlNa cells; or Renca cells and PA317 GlmGmSvNa cells;

Figure 21 is a graph of the tumor volume in mice injected with PA317/GlNa cells and non-irradiated B16F10 cells, or with PA317/GlmGmSvNa cells and non-irradiated B16F10 cells;

Figure 22 is a graph of the percent survival of mice challenged with Renca cells subsequent to vaccination with irradiated Renca cells; iradiated Balb 3T3/GlNa cells and irradiated Renca cells; or irradiated Balb 3T3/GlmGmSvNa cells and irradiated Renca cells; and

Figure 23 is a map of plasmid pGlGm. DETAILED DESCRIPTION OF THE INVENTION

In accordance with an aspect of the present invention, there is provided a method of treating a tumor in a host. The method comprises administering to a tumor site cells, preferably non-tumor cells, which express a therapeutically effective amount of one or more agents selected from the group consisting of an interferon, or a tumor necrosis factor.

The term "treating a tumor" as used herein means that one provides for the inhibition, prevention, or destruction of the growth of the tumor cells. The term "treating a tumor" also encompasses preventing recurrence of a tumor which has been resected.

Although the scope of the present invention is not intended to be limited to any theoretical reasoning, Applicants have found that when one administers to a host at a tumor site cells which express an interferon or a tumor necrosis factor, that one obtains a local, non-immune mediated anti-tumor effect; i.e., growth of the tumor cells is inhibited, prevented, or destroyed, and that such effect is not dependent upon an immunological response. Thus, through such administration of the cells which express an interferon or tumor necrosis factor, the inhibition, prevention or destruction of the growth of the tumor is not delayed until the onset of an immune response against the tumor. Thus, a more immediate treatment effect is accomplished. The immediate anti-tumor effect may be a result of the destruction of the vasculature of the tumor, or of direct destruction of the tumor cells, or of inhibition of tumor cell proliferation.

Non-tumor cells which may be administered to the host include, but are not limited to, fibroblasts, keratinocytes, or "universal" cells engineered to be tolerated by all individuals; and retroviral producer cells capable of generating retroviral vectors, including at least one polynucleotide encoding an agent selected from the group consisting of interferons and tumor necrosis factors.

In a preferred embodiment, the non-tumor cell is a fibroblast. Examples of fibroblasts which may be employed include, but are not limited to, autologous-syngeneic fibroblasts, allogeneic fibroblasts, xenogeneic fibroblasts, autologous fibroblasts, human foreskin fibroblasts, mouse BLKcl4 fibroblasts, BALB 3T3 fibroblasts, and NIH 3T3 fibroblasts.

In one embodiment, the agent is an interferon. Interferons which may be expressed include, but are not limited to, interferons of the Interferon-or family, including the Interferon- A/D chimera described hereinbelow, Interferon-/3, and Interferon-γ. In one embodiment, the interferon is an interferon of the Interferon-α family. In another embodiment, the interferon in Interferon-γ.

In another embodiment, the agent is a tumor necrosis factor. Tumor necrosis factors which may be employed include, but are not limited to, TNF-α and TNF-jS. In one embodiment, the tumor necrosis factor is TNF-α.

In one embodiment, the cell which expresses the interferon or tumor necrosis factor, is genetically engineered with at least one polynucleotide which encodes the interferon or tumor necrosis factor.

The term "polynucleotide" as used herein means a polymeric form of nucleotide of any length, and includes ribonucleotides and deoxyribonucleotides. Such term also includes single- and double-stranded DNA, as well as single- and double-stranded RNA. The term also includes modified polynucleotides such as methylated or capped polynucleotides .

Genes encoding the agents are obtainable through sources known to those skilled in the art (e.g., Genbank, ATCC, etc.) , and/or may be isolated from expression vehicles (e.g. , plasmids) obtainable through sources known to those skilled in the art through standard techniques (e.g., PCR) known to those skilled in the art.

The polynucleotide encoding the interferon or tumor necrosis factor is contained within an appropriate expression vehicle which has been transduced into the cell. Such expression vehicles include, but are not limited to, plasmids, eukaryotic vectors, prokaryotic vectors (such as, for example, bacterial vectors), and viral vectors.

In one embodiment, the vector is a viral vector. Viral vectors which may be employed include RNA viral vectors (such as retroviral vectors) , and DNA virus vectors (such as adenoviral vectors, adeno-associated virus vectors, Herpes Virus vectors, and vaccinia virus vectors) . When an RNA virus vector is employed, in constructing the vector, the polynucleotide encoding the interferon, or tumor necrosis factor, is in the form of RNA. When a DNA virus vector is employed, in constructing the vector, the polynucleotide encoding the interferon, or tumor necrosis factor, is in the form of DNA.

In one embodiment, the viral vector is a retroviral vector. Examples of retroviral vectors which may be employed include, but are not limited to, Moloney Murine Leukemia Virus, spleen necrosis virus, and vectors derived from retroviruses such as Rous Sarcoma Virus, Harvey Sarcoma Virus, avian leukosis virus, human immunodeficiency virus, myeloproliferative sarcoma virus, and mammary tumor virus. The vector is generally a replication incompetent retrovirus particle.

Retroviral vectors are useful as agents to mediate retroviral-mediated gene transfer into eukaryotic cells. Retroviral vectors are generally constructed such that the majority of sequences coding for the structural genes of the virus are deleted and replaced by the gene(s) of interest. Most often, the structural genes (i.e., gag, pol, and env) , are removed from the retroviral backbone using genetic engineering techniques known in the art. This may include digestion with the appropriate restriction endonuclease or, in some instances, with Bal 31 exonuclease to generate fragments containing appropriate portions of the packaging signal .

These new genes have been incorporated into the proviral backbone in several general ways. The most straightforward constructions are ones in which the structural genes of the retrovirus are replaced by a single gene which then is transcribed under the control of the viral regulatory sequences within the long terminal repeat (LTR) . Retroviral vectors have also been constructed which can introduce more than one gene into target cells. Usually, in such vectors one gene is under the regulatory control of the viral LTR, while the second gene is expressed either off a spliced message or is under the regulation of its own, internal promoter. Alternatively, two genes may be expressed from a single promoter by the use of an Internal Ribosome Entry Site.

Efforts have been directed at minimizing the viral component of the viral backbone, largely in an effort to reduce the chance for recombination between the vector and the packaging-defective helper virus within packaging cells. A packaging-defective helper virus is necessary to provide the structural genes of a retrovirus, which have been deleted from the vector itself.

Examples of retroviral vectors which may be employed include, but are not limited to, Moloney Murine Leukemia Virus vectors such as those described in Miller, et al . , Biotechniques, Vol. 7, pgs. 980-990 (1989) , and in Miller, et al., Human Gene Therapy, Vol. 1, pgs. 5-14 (1990) .

In a preferred embodiment, the retroviral vector may include at least four cloning, or restriction enzyme recognition sites, wherein at least two of the sites have an average frequency of appearance in eukaryotic genes of less than once in 10,000 base pairs; i.e., the restriction product has an average DNA size of at least 10,000 base pairs. Preferred cloning sites are selected from the group consisting of NotI, SnaBI, Sail, and Xhol . In a preferred embodiment, the retroviral vector includes each of these cloning sites. Such vectors are further described in U.S. Patent Application Serial No. 08/340,805, filed November 17, 1994, and in PCT Application No. W091/10728, published July 25, 1991, and incorporated herein by reference in their entireties.

When a retroviral vector including such cloning sites is employed, there may also be provided a shuttle cloning vector which includes at least two cloning sites which are compatible with at least two cloning sites selected from the group consisting of NotI, SnaBI, Sail, and Xhol located on the retroviral vector. The shuttle cloning vector also includes at least one desired gene which is capable of being transferred from the shuttle cloning vector to the retroviral vector.

The shuttle cloning vector may be constructed from a basic "backbone" vector or fragment to which are ligated one or more linkers which include cloning or restriction enzyme recognition sites. Included in the cloning sites are the compatible, or complementary cloning sites hereinabove described. Genes and/or promoters having ends corresponding to the restriction sites of the shuttle vector may be ligated into the shuttle vector through techniques known in the art .

The shuttle cloning vector can be employed to amplify DNA sequences in prokaryotic systems. The shuttle cloning vector may be prepared from plasmids generally used in prokaryotic systems and in particular in bacteria. Thus, for example, the shuttle cloning vector may be derived from plasmids such as pBR322; pUC 18; etc.

The vector includes one or more promoters. Suitable promoters which may be employed include, but are not limited to, the retroviral LTR; the SV40 promoter; and the human cytomegalovirus (CMV) promoter described in Miller, et al. , Biotechni-crues, Vol. 7, No. 9, 980-990 (1989), or any other promoter (e.g., cellular promoters such as eukaryotic cellular promoters including, but not limited to, the histone, pol III, and β-actin promoters) . Other viral promoters which may be employed include, but are not limited to, adenovirus promoters, TK promoters, and B19 parvovirus promoters. The selection of a suitable promoter will be apparent to those skilled in the art from the teachings contained herein.

The vector then is employed to transduce a packaging cell line to form a producer cell line. Examples of packaging cells which may be transfected include, but are not limited to, the PE501, PA317, ψ- 2 , iV-AM, PA12, T19-14X, VT- 19-17-H2, ψ CRE, ψ CRIP, GP+E-86, GP+envAml2, and DAN cell lines, as described in Miller, Human Gene Therapy, Vol. 1, pgs. 5-14 (1990) . The vector containing the polynucleotide encoding the interferon, tumor necrosis factor, or GM-CSF may transduce the packaging cells through any means known in the art. Such means include, but are not limited to, electroporation, the use of liposomes, and CaP0₄ precipitation. The packaging cells thus become producer cells which generate retroviral vectors which include a polynucleotide encoding the interferon, tumor necrosis factor, or GM-CSF. Such retroviral vectors then are transduced into cells, which may be non-tumor cells (such as, for example, fibroblasts) , or tumor cells, whereby the transduced cells will express the interferon or tumor necrosis factor.

The cells are administered to the host at the site of the tumor in an amount effective to inhibit, prevent, or destroy the growth of the tumor in the host. Such administration may be non-systemic, and wherein the cells are not administered at a site remote from the tumor. For example, such administration may be by direct, non-systemic injection of the cells to the site of the tumor. The host may be a mammalian host, including human and non-human primate hosts. In general, such cells are administered to the host in an amount of at least 10⁵ cells, and in general such amount does not exceed 10¹⁰ cells. Preferably, the cells are administered in an amount of from about 10^β cells to about 10^β cells. The exact dosage of cells which is to be administered is dependent upon a variety of factors, including the age, weight, and sex of the patient, and the type and severity of the tumor to be treated.

The cells are administered to the site of the tumor in conjunction with an acceptable pharmaceutical carrier. Suitable pharmaceutical carriers include, li-quid carriers such as a saline solution, aqueous buffers such as phosphate buffers and Tris buffers, or these buffers containing Polybrene (Sigma Chemical, St. Louis, Missouri) . The selection of a suitable pharmaceutical carrier is deemed to be apparent to those skilled in the art from the teachings contained herein.

Tumors which may be treated in accordance with the present invention include malignant and non-malignant tumors. Malignant (including primary and metastatic) tumors which may be treated include, but are not limited to, those occurring in the adrenal glands; bladder, bone; breast; cervix; endocrine glands (including thyroid glands, the pituitary gland, and the pancreas) ; stomach; small intestine; peritoneal cavity; colon; rectum; heart; hematopoietic tissue; kidney; liver; lung; muscle; nervous system; brain; eye; oral cavity; pharynx; larynx and other head and neck cancers; ovaries; penis; prostate; skin (including melanoma, basal cell carcinoma, and s-quamous cell carcinoma) ; testicles; thymus; and uterus.

The method of the present invention also is particularly applicable to the treatment of surgically inaccessible tumors.

In another embodiment, the non-tumor cells which express an interferon, or a tumor necrosis factor, may be administered to a host whose tumor has been resected (for example, head and neck tumors) . In such as embodiment, a tumor mass is removed from a host, and the cells expressing an interferon or a tumor necrosis factor are administered to the site of the tumor. Such administration of the cells expressing the interferon or tumor necrosis factor or GM-CSF provides for the inhibition, prevention, or destruction of any residual tumor cells remaining in the host subse-quent to removal of the tumor mass.

In one alternative, the cell which is administered to the site of the tumor is a retroviral producer cell, such as those hereinabove described. Upon administration of the producer cells to the site of the tumor, the producer cells generate retroviral vectors including a polynucleotide encoding an interferon or tumor necrosis factor or GM-CSF. Such retroviral vectors then transduce the tumor cells, which then produce the interferon or tumor necrosis factor in an amount which inhibits, prevents, or destroys the growth of the tumor. The producer cells may be administered in an amount of from about 10⁵ cells to about 10¹⁰ cells, preferably from about 10^€ cells to about 10^θ cells. The producer cells may be administered in combination with an acceptable pharmaceutical carrier, such as those hereinabove described.

In another embodiment, the viral vector is an adenoviral vector.

In one embodiment, the vector is an adenoviral vector.

The adenoviral vector which is employed may, in one embodiment, be an adenoviral vector which includes essentially the complete adenoviral genome. (Shenk, et al. , Curr. To . Microbiol. Immunol. , (1984); 111 (3) :1-39) . Alternatively, the adenoviral vector may be a modified adenoviral vector in which at least a portion of the adenoviral genome has been deleted.

In one embodiment, the vector comprises an adenoviral 5' ITR; an adenoviral 3' ITR; an adenoviral encapsidation signal; a DNA sequence encoding an interferon or tumor necrosis factor; and a promoter for expressing the DNA sequence encoding the interferon or tumor necrosis factor or GM-CSF. The vector is free of at least the majority of adenoviral El and E3 DNA sequences, but is not necessarily free of all of the E2 and E4 DNA sequences, and DNA sequences encoding adenoviral proteins transcribed by the adenoviral major late promoter.

In another embodiment, the gene in the E2a region that encodes the 72 kilodalton binding protein is mutated to produce a temperature sensitive protein that is active at 32°C, the temperature at which viral particles are produced, but is inactive at 37°C, the temperature of the animal or human host. This temperature sensitive mutant is described in Ensinger, et al . , J.Virology, 10:328-339 (1972) ; Van der Vliet, et al. , J.Virology. 15:348-354 (1975) ; and Friefeld, et al., Virology, 124:380-389 (1983) ; Englehardt, et al . , Proc.Nat.Acad.Sci. , Vol. 91, pgs. 6196-6200 (June 1994) ; Yang, et al. , Nature Genetics, Vol. 7, pgs. 362-369 (July 1994) .

An adenoviral vector is constructed preferably first by constructing, according to standard techniques, a shuttle plasmid which contains, beginning at the 5' end, the "critical left end elements," which include an adenoviral 5' ITR, an adenoviral encapsidation signal, and an Ela enhancer sequence; a promoter (which may be an adenoviral promoter or a foreign promoter) ; a tripartite leader sequence, a multiple cloning site (which may be as herein described) ; a poly A signal; and a DNA segment which corresponds to a segment of the adenoviral genome. Such DNA segment serves as a substrate for homologous recombination with a modified or mutated adenovirus, and such segment may encompass, for example, a segment of the adenovirus 5 genome no longer than from base 3329 to base 6246 of the genome. The plasmid may also include a selectable marker and an origin of replication. The origin of replication may be a bacterial origin of replication. A desired DNA sequence encoding an interferon or tumor necrosis factor may or GM-CSF be inserted into the multiple cloning site of such plasmid for production of a vector for use in accordance with the invention.

The plasmid is used to produce an adenoviral vector by homologous recombination with a modified or mutated adenovirus in which at least the majority of the El and E3 adenoviral DNA sequences have been deleted. Such homologous recombination may be effected through co-transfection of the plasmid vector and the modified adenovirus into a helper cell line, such as 293 cells, by CaP0₄ precipitation. The homologous recombination produces a recombinant adenoviral vector which includes DNA sequences derived from the shuttle plasmid between the Not I site and the homologous recombination fra-gment, and DNA derived from the El and E3 deleted adenovirus between the homologous recombination fragment and the 3' ITR. In one embodiment, the homologous recombination fragment overlaps with nucleotides 3329 to 6246 of the adenovirus 5 genome (ATCC VR-5) .

Through such homologous recombination, a vector is formed which includes an adenoviral 5' ITR, an adenoviral encapsidation signal; an Ela enhancer sequence; a promoter; a tripartite leader sequence; a DNA sequence encoding an interferon or tumor necrosis factor or GM-CSF; a poly A signal; adenoviral DNA free of at least the majority of the El and E3 adenoviral DNA sequences; and an adenoviral 3' ITR. This vector may then be transfected into a helper cell line, such as the 293 helper cell line (ATCC No. CRL1573) , which will include the Ela and Elb DNA sequences, which are necessary for viral replication, to generate replication defective viral vector particles.

The vector hereinabove described may include a multiple cloning site to facilitate the insertion of the DNA sequence encoding the interferon or tumor necrosis factor or GM-CSF into the cloning vector. In general, the multiple cloning site includes "rare" restriction enzyme sites; i.e., sites which are found in eukaryotic genes at a frequency of from about one in every 10,000 to about one in every 100,000 base pairs. An appropriate vector is thus formed by cutting the cloning vector by standard techniques at appropriate restriction sites in the multiple cloning site, and then ligating the DNA sequence encoding an interferon or tumor necrosis factor or GM-CSF into the cloning vector.

The DNA sequence encoding the interferon, tumor necrosis factor, or GM-CSF is under the control of a suitable promoter, which may be selected from those hereinabove described, or such DNA may be under the control of its own native promoter.

In one embodiment, the adenovirus may be constructed by using a yeast artificial chromosome (or YAC) containing an adenoviral genome according to the method described in Ketner, et al . , PNAS. Vol. 91, pgs. 6186-6190 (1994) , in conjunction with the teachings contained herein. In this embodiment, the adenovirus yeast artificial chromosome is produced by homologous recombination in vivo between adenoviral DNA and yeast artificial chromosome plasmid vectors carrying segments of the adenoviral left and right genomic termini . A DNA sequence encoding an interferon or tumor necrosis factor or GM-CSF then may be cloned into the adenoviral DNA. The modified adenoviral genome then is excised from the adenovirus yeast artificial chromosome in order to be used to generate adenoviral vector particles as hereinabove described.

The adenoviral vector particles, which include the DNA sequence encoding an interferon or tumor necrosis factor, then are transduced into non-tumor cells such as those hereinabove described, whereby such transduced cells will express the interferon or tumor necrosis factor. The transduced cells then are administered to the site of the tumor in the host in amount hereinabove described.

In another embodiment, a tumor is treated by administering non-tumor cells which express GM-CSF to elicit a systemic immune response against the tumor, wherein said non-tumor cells are administered at the site of the tumor or are administered at some other site in combination with such tumor cells or a tumor antigen of the type of tumor to be treated. In a preferred embodiment, a systemic immune response may be elicited by administering such cells to a host in combination with tumor cells of the type of tumor to be treated or a tumor antigen of the type of tumor to be treated. Thus, in accordance with another aspect of the present invention, there is provided a method of treating a tumor in a host which comprises administering to a host a mixture of (a) non-tumor cells engineered with a polynucleotide encoding granulocyte macrophage colony stimulating factor (GM-CSF) ; and (b) an agent selected from the group consisting of (i) tumor cells of the type of tumor which is to be treated and (ii) at least one tumor antigen of the type of tumor which is to be treated.

Non-tumor cells engineered with a polynucleotide encoding GM-CSF, and which may be administered to a host, include, but are not limited to, fibroblasts, keratinocytes, "universal" cells, and producer cells such as those hereinabove described.

The polynucleotide encoding GM-CSF is contained in an appropriate expression vehicle, such as those hereinabove described, which has been transduced into the non-tumor cell. In one embodiment, the expression vehicle is a retroviral vector including a polynucleotide encoding GM-CSF, which is transduced into the non-tumor cell.

In one embodiment, the agent is a tumor cell.

In one embodiment, prior to administration of the tumor cells, the tumor cells are treated (such as, for example, by irradiation) such that the tumor cells are incapable of forming a new tumor when administered to the host.

The non-tumor cells engineered with a polynucleotide encoding GM-CSF and the tumor cells are administered in amounts effective to inhibit, prevent, or destroy the growth of the tumor, or residual tumor, or tumor metastaseε. Although the scope of this aspect of the present invention is not to be limited by any theoretical reasoning, the administration of the non-tumor cells engineered with a polynucleotide encoding GM-CSF and the tumor cells elicit an immune response (such as, for example, a systemic immune response) against the tumor. In one embodiment, the non- tumor cells and the tumor cells are administered non- systemically at a subcutaneous site. For example, the cells may be delivered intradermally or subcutaneously, or intramuscularly at a site distal to the primary tumor or any metastases. In general, the non-tumor cells are administered in an amount of at least 10^s cells, and in general does not exceed 10¹⁰ cells. Preferably, the non-tumor cells are administered in an amount of from about 10⁶ cells to about 10^β cells. The tumor cells are administered in an amount of at least 10^s cells, and in general such amount does not exceed 10^αo cells. Preferably, the tumor cells are administered in an amount of from about 10⁶ cells to about 10⁸ cells. Tumors which may be treated include those hereinabove described.

The non-tumor cells engineered with a polynucleotide encoding GM-CSF and the tumor cells against which an immune response is to be elicited may be administered to a host in conjunction with the cells expressing an interferon or a tumor necrosis factor, which are administered at a tumor site. Thus, in accordance with yet another aspect of the present invention, there is provided a method of treating a tumor in a host which comprises administering to a tumor site cells which express an agent selected from the group consisting of an interferon and a tumor necrosis factor, in conjunction with administering to a host (such as, for example, by subcutaneous non-systemic administration) a mixture of (a) non-tumor cells engineered with a polynucleotide encoding a cytokine which enhances or promotes an immunogenic response against tumor cells which are being treated, and (b) tumor cells and/or antigens of tumor cells against which an immune response is to be elicited. The non- tumor cell and tumor cell and/or antigens of tumor cells may be selected from those hereinabove described.

Cytokines encoded by the polynucleotide with which the non-tumor cell is engineered include, but are not limited to, GM-CSF, Interleukin-2, Interleukin-4, and Interleukin-12. In one embodiment, the cytokine is GM-CSF. The polynucleotide encoding the cytokine may be contained in an expression vehicle such as those hereinabove described, which is transduced into the non-tumor cell. By treating a tumor in accordance with this method, one obtains an immediate anti-tumor response or effect, which is not an immune response, through the administration of a cell expressing an interferon or tumor necrosis factor at the site of the tumor. Such response is followed by an immune response against the tumor obtained through the administration of the non-tumor cell engineered with a polynucleotide encoding a cytokine and tumor cells and/or antigens of tumor cells of a tumor which is the subject of the treatment. Thus, one is able to treat a tumor in accordance with this method through non-immune and immune mechanisms.

EXAMPLES

The invention now will be described with respect to the following examples; however, the scope of the present invention is not intended to be limited thereby.

Example 1 Construction of GIFlSvNa, GlF2SvNa. GlF3SvNa. and GlmF3SvNa

Plasmids pGlFlSvNa, pGlF2SvNa, pGlF31SvNa, and pGlmF3SvNa were derived from plasmid pGl. Plasmid pGl was constructed from pLNSX (Figure 7) (Palmer, et al. , Blood, Vol. 73, pgs. 438-445) . The construction strategy for plasmid pGl is shown in Figure 1. The 1.6kb EcoRI fragment, containing the 5' Moloney Murine Sarcoma Virus (MoMuSV) LTR, and the 3. Okb EcoRI/Clal fragment, containing the 3' LTR, the bacterial origin of replication and the ampicillin resistance gene, were isolated separately. A linker containing seven unique cloning sites was then used to close the EcoRI/Clal fragment on itself, thus generating the plasmid pGO. The plasmid pGO was used to generate the vector plasmid pGl (Figure 3) by the insertion of the 1.6kB EcoRI fragment containing the 5' LTR into the unique EcoRI site of pGO. Thus, pGl (Figure 3) consists of a retroviral vector backbone composed of a 5' portion derived from MoMuSV, a short portion of gag in which the authentic ATG start codon has been mutated to TAG (Bender, et al . 1987) , a 54 base pair multiple cloning site (MCS) containing, from 5' to 3 ' the sites EcoRI, NotI, SnaBI, Sail, BamHI, Xhol, Hindu, Apal, and Clal and a 3' portion of MoMuLV from base pairs 7764 to 7813 (numbered as described (Van Beveren, et al. , Cold Spring Harbor, Vol. 2, pg. 567, (1985)) (Figure 2) . The MCS was designed to generate a maximum number of uni-que insertion sites, based on a screen of non-cutting restriction enzymes of the pGl plasmid, the neo^r gene, the β-galactosidase gene, the hygromycin^r gene, and the SV40 promoter.

To construct pBg (Figure 4) the 3.0 kb BamHI/EcoRI lacZ fragment that encodes /3-galactosidase was isolated from pMC1871 (Pharmacia) . This fra-gment lacks the extreme 5' and 3' ends of the 3-galactosidase open reading frame. Linkers that would restore the complete lacZ open reading frame and add restriction sites to each end of the lacZ gene were synthesized and ligated to the BamHI/EcoRI lacZ fragment. The structure of the 5' linker was as follows: 5' - 1/2 Ndel - SphI - NotI - SnaBI - Sail - SacII - AceI - Nrul - Bglll - III 27 bp ribosomal binding signal - Kozak consensus sequence/Ncol - first 21 bp of the lacZ open reading frame - 1/2 BamHI - 3' . The structure of the 3' linker was as follows: 5' - 1/2 mutated EcoRI - last 55 bp of the lacZ open reading frame - Xhol - Hindlll - Smal - 1/2 EcoRI - 3' . The restriction sites in the linkers were chosen because they are not present in the neomycin resistance gene, the S-galactosidase gene, the hygromycin resistance gene, or the SV40 promoter. The 27 bp ribosomal binding signal was included in the 5' linker because it is believed to enhance mRNA stability (Hagenbuchle, et al . , Cell 13:551-563, 1978 and Lawrence and Jackson, J. Mol. Biol. 162:317-334, 1982) . The Kozak consensus sequence (5' -GCCGCCACCATGG-3' ) has been shown to signal initiation of mRNA translation (Kozak, Nucl.Acids Res. 12:857-872, 1984) . The Kozak consensus se-quence includes the Ncol site that marks the ATG translation initiation codon. pBR322 (Bolivar et al. Gene 2:95, 1977) was digested with Ndel and EcoRI and the 2.1 kb fragment that contains the ampicillin resistance gene and the bacterial origin of replication was isolated. The ligated 5' linker - lacZ - 3' linker DNA described above was ligated to the pBR322 Ndel/EcoRI vector to generate pBg. pBg has utility as a shuttle plasmid because the lacZ gene can be excised and another gene inserted into any of the restriction sites that are present at the 5' and 3' ends of the lacZ gene. Because these restriction sites are reiterated in the pGl plasmid, the lacZ gene or genes that replace it in the shuttle plasmid construct can easily be moved into pGl.

The "backbone" vector pGlNa was constructed from pGl and pN2 (Armentano, et al., J. Virology, Vol. 61, pgs. 1647-1650 (1987)) . pGlNa was constructed by cutting pN2 (Figure 5) with EcoRI and AsUII, filling in the ends of the EcoRI/AsuII fragment containing the neo^Rgene, and ligating the fragment into SnaBI digested pGl to form pGlNa (Figure 6) . pGl was cut with Hindlll and Sail. pSvNa (Figure 8) , which contains the SV40 promoter from pLNSX (Figure 7) and the neo^Rgene from pN2, also was cut with Hindlll and Sail, and a Hindlll-Sall fragment containing an SV40 promoter and a β- galactosidase gene was ligated into Hindlll/Sall digested pGl to form pGlXSvNa (Figure 9) . pGlFlSvNa (Figure 10) includes the human Interferon-α gene under the control of the retroviral LTR. pGlF2SvNa (Figure 11) includes a Bglll chimera of the human Interferon- a A and D genes under the control of the retroviral LTR. This chimera is active in human and mouse cells. (Rehberg, et al . , J . Biol. Chem.. Vol. 251, pg. 11497 (1992)) . pGlF31SvNa (Figure 12) includes the human Interferon-γ gene under the control of the retroviral LTR. pGlmF3SvNa (Figure 13) includes the murine Interferon-γ gene under the control of the retroviral LTR.

The cDNA for murine IFNγ (Genbark Accession No. K00083) was obtained by PCR from plasmid pGEMmuIFNγ (Sandoz, East Hanover, New Jersey) using a 5' primer (GCA GAT CTT TCC GCA GCA GCC GCC ACC ATG AAC GCT ACA CAC TGC ATC) that added a Bglll site, a ribosome binding site and a Kozak initiation sequence and a 3' primer (GTC CTC GAG TCA GCA GCG ACT CCT TTT CC) that added a Xhol site. The resulting PCR product was inserted into the plasmid backbone of the Bglll/Xhol cut pBg plasmid. The result, pMF3, was cut with NotI and Xhol and ligated into Notl/Sall digested pGlXSvNa to create pGlmF3SvNa. The cDNA for human IFNα2 (EMBL Accession No. A15695) was obtained on plasmid pL31 from Hoffman-LaRoche, Nutley, New Jersey. It was digested with Ddel and PstI to remove the IFNα2 cDNA. It was inserted into BamHI/PstI digested pUC18 (ATCC No. L08752) with the aid of a linker composed of the BamHI and Ddel sites. The result, pFl, was digested with SnaBI and Dral and the insert was ligated into the SnaBI site of pGlXSvNa to obtain pGlFlSvNa. The cDNA for IFNα A/D was obtained on plasmid pLIF-AD from Hoffman- LaRoche, Nutley, New Jersey. The insert was cut internally with Bglll and the Bglll fragment of pFl containing the 5' end of the IFN 2 gene was inserted. The result of the ligation, plasmid pF2.Temp, was digested with Ncol and PstI and the insert was ligated into the backbone of pFl following removal of the IFNα2 cDNA by digestion with Ncol and PstI to create plasmid pF2. Plasmid pF2 was SnaBI/EcoRI digested and the chimeric IFNcr A/D cDNA was ligated into the SnaBI site of pGlXSvNa to create pGlF2SvNa. The cDNA for human IFNγ (EMBL Accession No. V00536) was obtained by RT PCR of human tumor infiltrating lymphocyte (TIL) cell mRNA using a 5' primer (GGT ACG TAG CCG CCA CCA TGA AAT ATA CAA GTT ATA TCT TGG) that added a SnaBI site and a Kozak initiation sequence and a 3' primer (CTG TCG ACT CGA GTT ACT GGG ATG CTC TTC G) that provided a Sail site. The PCR product was cloned into pGlXSvNa that had been cut with SnaBI and Sail to create pGlF31SvNa.

Each of plasmids pGlFlSvNa, pGlF2SvNa, pGlF3SvNa, and pGlmF3SvNa was transfected into the murine PE501 ecotropic packaging cell line. (Miller, et al . , 1990) Transient supernatants from these transfectants were used to transinfect murine PA317 (ATCC No. CRL 9078) amphotropic packaging cells . The amphotropic producers were cloned and characterized by Southern Blot analysis and retroviral titer production. Clones having the expected retroviral insert sizes that produced high retroviral titers were picked for further analysis. All producer cells were negative for contamination by mycoplasmas and replication competent retroviruses . The retroviruses generated from the producer cells are hereinafter referred to as GIFlSvNa, GlF2SvNa, GlF31SvNa, and GlmF3SvNa.

Example 2

Murine CT26 (Brattain, et al . , Cancer Research. Vol. 40, pg. 2142 (1980)) colonic tumor, Renca (Murphy, et al . , J. Nat. Can. Inst .. Vol. 50, pg. 1013 (1973)) renal carcinoma, and B16F10 (Fidler, et al . , Cancer Research, Vol. 37, pg. 3945 (1977)) melanoma cells and the non-tumorigenic BLKcl4 (ATCC No. T1B81) and BALB3T3 (ATCC No. CCL163) fibroblast cells were transduced with GlF2SvNa, GlmF3SvNa, or GINa retroviral supernatants. In each case, 2.5xl0⁵ cells were transduced with retroviral supernatant in the presence of 8μg/ml Polybrene at multiplicities of infection from 0.5 to 10. Following selection in G418, the cells were characterized by Southern Blot analysis and interferon secretion. Bioassays were performed on GlF2SvNa trausduced cells using HS68 cells (ATCC No. CRL1635) and encephalomyocarditis virus (ATCC No. VR129B) by the method in Unit 6.9.5 of Current Protocols in Immunology, Coligan, et al . , eds . (1994) . Mouse Interferon-α secretion by GlmF3SvNa transduced cells was determined with a mouse Interferon-α ELISA kit (Gibco catalog no. 32«0SA) . Additionally, all cell lines were tested and found to be negative for contamination by mycoplasma and replication competent retroviruses . A B16F10/GlF2SvNa mixed population was cloned by limiting dilution to obtain high, medium and low Interferon-α A/D secreting cells. Southern Blot analysis and interferon secretion results are given in Table I below. In Table I below, ND means not detected; and "expected" means that a retroviral band was detected through Southern Blot analysis.

Table I

IFN Secretion

Cell Retroviral Southern IFN-α A/D mIFN-Υ line Vector Result I.U./24 ng/24 hrs. hours,

10⁶ cells 10⁶ cells

CT26 none no band ND —

CT26 GINa expected ND —

CT26 GlF2SvNa expected > 4,000 —

CT26 GlmF3SvNa expected — 3.8

Renca none no band ND ND

Renca GINa expected ND ND

Renca GlF2SvNa expected > 5,600 —

Renca GlmF3SvNa rearranged -- ND

B16F10 none no band ND —

B16F10 GINa expected ND --

B16F10 GlF2SvNa expected 180-360 -- mix

B16F10 GlF2SvNa expected 15 -- clone 1

B16F10 GlF2SvNa expected 160 -- clone 4

B16F10 GlF2SvNa expected 2,240- -- clone 14 4,480

B16F10 GlmF3SvNa expected — 20.9

BLKC14 none no band ND ND

BLKC14 GINa expected ND ND

BLKC14 GlF2SvNa expected 731

BLKC14 GlmF3SvNa expected 4.1

BALB3T3 none no band ND ND

BALB3T3 GINa expected ND ND

BALB3T3 GlF2SvNa expected 5,829 --

BALB3T3 GlmF3SvNa expected -- 78 Example 3 63 mice were divided into three groups of 16 and one group of 15. The mice of Group I (16 mice) were injected intradermally with 5 x 10⁴ CT26 cells. The mice of Group II (16 mice) were injected intradermally with 5 x 10* CT26 cells transduced with GINa. The mice of Group III (15 mice) were injected intradermally with 5 x 10⁴ CT26 cells transduced with GlF2SvNa. The mice of Group IV (15 mice) were injected intradermally with CT26 cells transduced with GlmF3SvNa. Tumor onset was noted and measurements were taken at regular intervals. 11 days after injection, four mice from each group were taken for histology. The tumor onset at 11 and 34 days after injection and average tumor volume at 34 days after injection are given in Table II below.

Table II

Group Day 11, animals Day 34 , animals Average with tumor with tumor tumor volume

I 10/16 (62.5%) 4/12 (33%) 742 mm³

II 5/16 (31.3%) 4/12 (33.3%) 1,676 mm³

III 0/16 (0%) 0/12 (0%) 0 mm³

IV 0/15 (0%) 4/11 (36.4%) 81 mm³

CT26 tumor shows an unusually high percentage of spontaneous regressions. The difficulties of this model notwithstanding, none of the animals in Group III developed a tumor during the 73 day course of the study. Also, the tumors in Group IV were substantially smaller than the tumors in control Groups I and II.

In a similar experiment, mice were inoculated intradermally with parental B16F10 cells, B16F10 cells transduced with GINa, B16F10/GlF2SvNa clones 1, 4, or 14 or the mixed population of B16F10 cells transduced with GlF2SvNa (See Table I above) , or the mixed population of B16F10 clones transduced with GlF2SvNa. (See Table I.) Ten mice in each group were injected with 10^s cells per mouse. The survival curve is shown in Figure 14.

As shown in Figure 14 , mice inoculated with clone 1 of B16F10/GlF2SvNa cells, which secreted approximately 15 International Units of Interferon-α A/D per 10⁶ cells in 24 hours, had a survival curve that was very similar to the animals inoculated later with control parental B16F10 cells or B16F10/GlNa cells. An increase in survival was seen when animals were inoculated with the mixed population of B16F10/GlF2SvNa calls (producing about 200 I.U./10⁶ cells/24 hrs.) , or clones secreting greater than 160 I.U./10⁶ cells/24 hrs. Among those animals, 6 out of 27 (22.2%) were long-term (i.e., 6 months) survivors.

In another experiment, the in vivo growth of 5xl0⁵ B16F10 cells transduced with GlF2SvNa or GlmF3SvNa, as well as a mixture of 2.5x10^s B16F10 cells transduced with GlF2SvNa and 2.5x10^s B16F10 cells transduced with GlmF3SvNa was compared with the growth of 5x10^s parental B16F10 cells in C57BL/6J mice following intradermal inocculation (10 animals per group) . The tumor volumes of the mice are shown in Figure 15. During the 34 day course of the experiment, two of the mice treated with B16F10/GlmF3SvNa cells, 5 of the animals treated with the mixture, and 7 of the mice treated with B16F10/GlF2SvNa cells failed to develop a palpable tumor, and tumors in the three treated groups showed a delay in the progression of the tumor volume.

Example 4

10⁶ irradiated B16F10/GlF2SvNa, or mixed populations, or clones 1, 4, or 14 , of B16F10/GlmF3SvNa cells were implanted intradermally into syngeneic mice. 9 or 10 mice received each type of cell. Twelve days later, the mice were challenged with 10⁵ live, non-irradiated parental B16F10 cells, and the onset and progression of the tumor was monitored. No significant differences between animals vaccinated with interferon-producing cells, naive animals, or animals vaccinated with parental B16F10 cells or B16F10/GlNa cells were observed. This was found to be true whether the animals were vaccinated once or boosted an additional two times.

In another experiment, 10* Renca cells transduced with GlF2SvNa were implanted intradermally into syngeneic mice. 5 to 10 mice were in each group of mice. 14 days after the mice received the transduced Renca cells, the mice were challenged with 10^s, or 5xl0⁵; or 10⁶; or 5xl0⁶; or 10⁷ parental Renca cells. The mice were not protected from the challenge of the parental Renca cells.

In yet another experiment, mice were vaccinated intradermally with 5xlO^β irradiated BALB 3T3 fibroblasts transduced with GlF2SvNa plus 10⁶ parental Renca cells. On day 12, the mice were challenged with 3xl0⁶ parental Renca cells. No vaccine effect was observed as a result of the administration of the transduced fibroblasts.

In a further experiment, 10^s live, non-irradiated B16F10, Bl6F10/GlNa, a mixed population of B16F10/GlF2SvNa cells, B16F10/GlF2SvNa clone 4 cells, or B16F10/GlmF3SvNa cells were injected intradermally into immunodeficient nude mice. The interferon-transduced tumor cells grew as poorly in the nude mice as they did in the immunocompetent strains.

In another experiment, two C57B1/6J mice were inoculated with 5 x 10⁵ non-irradiated Bl6F10/GlF2SvNa cells, and two C57B1/6J mice were inoculated with 5 x 10^s non-irradiated B16F10/GlmF3SvNa cells. On day 28, animals that remained free of tumor were challenged intradermally with 10⁶ parental B16F10 cells. As a control, a previously naive C57B16J mouse was inoculated at the same time with 5 x 10⁵ B16F10 cells. Six days later, the spleens were removed, and single cell suspensions were prepared and sensitized in vi tro an additional 8 days using irradiated B16F10 cells. At the end of this time, the ability of the sensitized spenocytes to lyse ¹Cr labeled B16F10 or syngeneic MCA205 cells (Wexler, et al., J. Nat. Cancer Inst. , Vol. 63, pg. 1393 (1979); Asher, et al., J. Immunol.. Vol. 146, pgs. 3227-3234 (1991)) was determined, according to the procedure described in Martz, The ⁵¹Cr-Release Assay for CTL-Mediated Target Cell Lysis in Cvtotoxic T-Cells. Sitkovsky, et al. , eds . , pgs. 457-467 (1993) . As a negative control, the ability of two week old lymphokine activated killer cells (old LAK) to lyse the same targets was determined. As shown in Figure 16, the old LAK cells, as expected, failed to lyse the target cells, and the CTL response was directed specifically against the B16F10 cells. The anti-B16F10 CTL response of the mouse injected with B16F10 cells after only a primary exposure to the tumor was as good or better than the responses of animals that had been exposed to B16F10 following an initial suppression of interferon-transduced B16F10 cells. If the initial suppression of the tumor cells had been due to the stimulation of the CTL, a reaction more vigorous than the primary response of the control mouse should have occurred.

In addition to the minimal role of specific T-cell mediated immunity, the "natural" immune system may not be involved in interferon-mediated suppression of tumor growth. The natural immune system is presented and ready to react in a naive animal, it lacks memory for a particular antigen, but it can be stimulated in a transient sense during and shortly after the presence of the agent causing the reaction.

In another experiment, ten immunocompetent C57B1/6J mice were inoculated on the left flank with either 10^s B16F10/GlF2SvNa, B16F10/GlmF3SvNa, B16F10/GlNa, or parental B16F10 cells, and were challenged simultaneously on the right flank with 10⁵ parental B16F10 cells. Tumor sizes were measured on day 18. If the interferon transduced ceils on the left flank had stimulated natural immunity, there should have been a suppression of parental tumor growth on the opposite flank. Average tumor sizes in the right and left flanks are given in Table III below. Table III

Cells inocul ated into 1 eft flank

B16F10/ B16F10/ B16F10/ GlF2SvNa GlmF3SvNa GINa B16/F10 size of 79±31 80±38 1,103±748 1,705± tumor on 1,184 left flank (mm³) size of 1,576±646 1,725±1,310 1,570±501 1,740± parental mm³ mm³ mm³ 746mm³ B16F10 tumor on right flank (mm³)

As shown in Table III, there was no suppression of parental tumor growth on the opposite flank. The parental tumors grew as well in mice bearing interferon-transduced B16F10 cells as in animals bearing control GINa-transduced or parental B16F10 cells.

In a further experiment, the role of the natural immune system in natural killer cell deficient C57Bl/6J-bg/bg mice also was tested. Mutant C57Bl/6J-bg/bg and wild-type C57B1/6J mice were inoculated with 10⁵ GINa, GlF2SvNa, or GlmF3SvNa transduced B16F10 cells or parental B16F10 cells, and the mice were monitored for tumor onset, progression, and survival. The results are given in Table IV below.

Table IV

Day tumor Day of Day of first first last

Wild type Overall seen death death Average or NK tumor in in in survival deficient incidence crroup qroup qxoup (days)

Parental wt 10/10 7 17 28 25 B16F10

Parental bg/bg 10/10 7 17 28 23

B16F10

B16F10/ wt 10/10 7 20 28 2-4 GINa B16/F10 bg/bg 10/10 7 20 28 24 GINa

B16F10/ t 8/10 7 38 61 45 GlF2SvNa

B16/F10 bg/bg 9/10 7 38 56 50 GlF2SvNa

B16F10/ t 10/10 7 41 56 50 GlmF3SvNa

B16F10/ bg/bg 10/10 7 38 52 47 GlmF3SvNa

As shown in Table IV, the wild-type and natural killer cell deficient mice had similar tumor incidences, days of tumor onset, progression and survival.

Example 5

The experiments described above show that interferon- transduced cells themselves fail to induce a systemically protective response. Rather, the primary effect is a local one. That is, the interferon must be present in the immediate anatomic vicinity of the tumor cell. Furthermore, tumors separated by either space or time are not affected by the interferon-secreting cells.

In order to test this concept further and to provide practical means of delivering interferon to the site of a tumor, BLKcl4 and BALB3T3 fibroblast cell lines were transduced with GINa, GlF2SvNa or GlmF3SvNa as hereinabove described.

100 C57B1/6J mice were divided into 10 groups with 10 mice in each group. Each mouse in one group of mice was inoculated intradermally with 10^s B16F10 cells alone. Each mouse in each of the other groups of mice was inoculated with a mixture of 10⁵ B16F10 cells and (i) 5 x 10"; or (ii) 10⁵; or (iii) 10⁶, (a) BLKcl4/GlNa cells; or (b) BLKcl4/GlF2SvNa cells; or (c) BLKcl4/GlmF3SvNa cells. The average survival time in days for the mice in each group was determined. The results are given in Table V below. Table V

10* B16F10 cells plus

No BLKC14/ BL C14/ B KC14/ othe GINa cells GlF2SvNa cells GlmF3SvNa cellε r cell s

Average 5x10' 10' 10' SX10' 10* 10* 5X10' 10' 10' survive

1 (days) 20ι 19.1 20 19.9 22 21.5 25.9 22.5 IB.6 21.4 3.1 ±2.2 ±3-4 ±2.6 ±2.1 ±3.8 ±4.5 ±5.2 ±2.4 ±5.4

P- -- >.10 >.10 >.10 >.10 >.10 .013 >.10 >.10 >.10 va ue

As shown in Table V, when 10-fold greater numbers of BLKcl4/GlF2SvNa cells were mixed with the tumor cells, a significant increase in average survival was seen (p=.013) .

In another experiment, 100 C57B1/6J mice were divided into 10 groups with 10 mice in each group. Each mouse in one group of mice was inoculated intradermally with 10^s B16F10 cells alone. Each mouse in the other groups of mice was inoculated with 10^s B16F10 cells and (i) 5 x 10⁴; or (ii) 10⁵; or (iii) 10⁶, (a) BALB3T3/GlNa cells; or (b) BALB3T3/GlF2SvNa cells; or (c) BALB3T3/GlmF3SvNa cells. The average survival times for each group of mice were determined. The results are given in Table VI below.

Table VI

10' B16F10 cells plus

No BALB3T3/ BALB3T3/ BALB3T3/ othe G Na cells GlF2SvNa ce 11s GlmF3SvNa cells r cell s

Average 5x10' 10' 10* 5X10* 10' ic 5X10' 10* 10' βurviva

1

<days) 20.7 21.2 19 lβ.e 20.4 21.6 24.7 22.7 22 23.2 ±2.9 ±3.0 ±2.1 ±3.4 ±3.6 ±3.3 ±2.6 ±3.1 ±3.1 2.9

P- -- >.10 >.10 >.10 >.10 >.10 .026 >.10 >.10 .094 value

As shown in Table VI, and as observed also with the syngeneic BLKcl4 fibroblasts, the Interferon-α A/D secreting BALB3T3 cells induced a significant increase in survival when administered in an amount of 10⁶ cells (p=.026) . In addition, secretion of murine Interferon-γ by BALB3T3 cells approached significance (p=.094) at an amount of 10⁶ cells, possibly because the BALB3T3 cells secreted more murine Interferon-γ than the BLKcl4 cells. (See Table I.) The above experiments show that cells that were not derived from the tumor can be used to deliver interferon efficaciously to the local environment of the tumor site . Also the experiment with the allogeneic cells indicate that the cells used to deliver the interferon need not be of the same genetic composition as the tumor-bearing host animal .

Example 6

40 BALB/c AnN mice were divided into 4 groups of 10 mice each. Group I received an intradermal inoculation of 10⁵ Renca cells alone. Group II was inoculated with 5xl0^€ BALB3T3/GlNa cells and 10⁵ Renca cells at a fibroblast-to- tumor cell ratio of 50:1. Group III was inoculated with 5x10* BALB3T3/GlF2SvNa cells and 10⁵ Renca cells at a fibroblast-to-tumor cell ratio of 50:1. Group IV was inoculated with 5xl0⁶ BALB3T3/GlmF3SvNa cells and 10^s Renca cells at a fibroblast-to-tumor ratio of 50:1. Survival curves of each group were plotted and are shown in Figure 17.

When only the animals that have died are used in the calculations, a 22 day increase in survival was seen in the BALB3T3/GlF2SvNa treated animals, and 3 out of 10 animals remained alive at day 63. After 63 days, 28 days after all twenty of the control animals had died, only one out of 10 mice in the BALB3T3/GlmF3SvNa group had died. The nine remaining animals were healthy and completely free of palpable tumor.

Example 7

45 BALB/c AnN mice were divided into i group of 15 animals and 3 groups of 10 animals. The 15 animals in group I received an intraperitoneal injection of 5x10^s Renca cells alone. Groups II, III and IV were inoculated intraperitoneally with 5x10^s Renca cells on day 1 followed by 3xl0⁶ irradiated BALB3T3/GlNa, BALB3T3/GlF2SvNa, or BALB3T3/GlmF3SvNa fibroblasts on day 4 and survival was monitored for 6 months. In this experiment, there was a slight survival advantage in animals that received the BALB3T3/GlmF3SvNa cells and 6 of the 10 animals that were treated with BALB3T3/GlF2SvNa cells were alive after 6 months. Survival curves of each group are shown in Figure 18.

Example 8

A murine GM-CSF expression plasmid, pXMT2-muCSF was obtained from Sandoz Pharmaceuticals (East Hanover, NJ) . The mGM-CSF cDNA in pXMT2-muCSF was made by looping introns out of a phage clone containing the murine GM-CSF gene and reassembling them into a cDNA. The 1.2 kb cDNA was cloned into the Xhol site of pSM using an Xhol/Sall fusion. The restriction enzyme Kpnl was used to excise the 1.2 kb fra-gment containing the mGM-CSF cDNA out of pXMT2-muCSF. The 1.2 kb fragment that contains the mGM-CSF (Genbank accession number X02333 and Gough et al. , EMBO J. , Vol. 4, pgs. 645-653 (1985) ) was restriction enzyme digested with StuI and Plel to generate a 619 bp fragment containing 13 bp of 5'- untranslated sequence, the 462 bp mGM-CSF open reading frame, and 149 bp of 3' -untranslated sequence. The ends of the 619 bp fragment were blunted with DNA polymerase I large (Klenow) fragment.

To construct the retroviral vector pGlmGmSvNa, pGlXSvNa was digested with SnaBI, which cuts in the multicloning region. The 619 bp StuI-Plel fragment containing the mGM-CSF cDNA was ligated to the SnaBI-digested pGlXSvNa using T4 DNA ligase. Ligated plasmid DNA was transformed into E. coli DH5α competent cells (Gibco/BRL, Gaithersburg, MD) and DNA from ampicillin-resistant colonies was screened by restriction enzyme digestion. Plasmids that appeared to contain the 619 bp mGM-CSF insert in the forward orientation [i.e., ATG translation initiation codon adjacent to the retroviral 5' long terminal repeat (LTR)] were termed pGlmGmSvNa (Figure 19) . The dideoxy sequence of one plasmid clone encompassing the region from 300 bp upstream of the mGM-CSF ATG through 350 bp downstream of the TGA translation termation codon indicated that the mGM-CSF cDNA was cloned in the correct orientation and that its se-quence matched that found in pXMT2-muCSF. The sequence showed that 8 bp at the 5' end and 33 bp at the 3' end of the 619 bp mGM-CSF fragment had been removed from the fragment, probably prior to ligation into pGlXSvNa. Because these lost bases are not part of the protein open reading frame, they should not have an effect on the expression of mGM-CSF in the retroviral construct . pGlmGmSvNa DNA was tranfected into the ecotropic packaging cell line, PE501 (Miller, 1990) , using standard calcium phosphate precipitation techniques (Chen et al . , Biotechniσues. Vol. 6, pgs. 632-638 (1988)) . Vector- containing supernatants from the transfected cultures were used to transinfect the amphotropic packaging cell line PA317. Transinfected PA317 were selected under G418 (800 μg/ml) and 38 single cell clones were isolated. The relative number of retroviral particles in the supernatants of the 38 clones was compared by RNA titering and the 8 clones with highest RNA titer were further analyzed by biological (G418) titering, Southern analysis, and ELISA (enzyme linked immunosorbant assay) for mGM-CSF expression (Endogen) . Clone 5 had the highest G418 titer (7.2 x 10⁵ cfu/ml) . Clone 25 was found to express high levels of mGM-CSF (about 200 ng mGM- CSF/10⁶ cells/24 hours) and therefore was chosen for studies re-quiring high mGM-CSF expression.

Example 9

Balb/3T3 clone A31 (ATCC#CCL163) is a murine embryonic fibroblast cell line developed from disaggregated Balb/c mouse embryos (Aaronson et al . , J. Cell. Phvsiol.. Vol. 72, pg. 141 (1968)) . NIH/3T3 (ATCC#CRL1658) is a line of contact-inhibited mouse fibroblasts developed from disaggregated NIH Swiss mouse embryos. Both cell lines were grown in culture and transduced with 10 ml of GlmGmSvNa.5 supernatant. Transduced cells were selected in G418 and single cell clones were isolated. For the Balb/3T3 transduced cells (Balb/GlmGmSvNa) , 60 clones were isolated, and after a preliminary test of mGM-CSF expression by ELISA (Kurstak, Enzyme Immunodia-gnosis, Acadmic Press, Orlando Florida) , 25 clones were further expanded and tested by ELISA for mGM-CSF expression. The Balb/GlmGmSvNa mixed cell population (the G418-selected, transduced Balb3T3 cells prior to cloning) secreted 190 ng mGM-CSF/10⁶ cells/24 hours, whereas the clones secreted 60-833 ng/10⁶ cells/24 hours. Balb/GlmGmSvNa.5 secretes approximately 240 ng mGM-CSF/10⁶ cells/24 hours and has been characterized by Southern and Northern analyses. 26 NIH3T3/GlmGmSvNa clones were isolated and tested for mGM-CSF expression by ELISA. NIH3T3/GlmGmSvNa clone 1 secretes approximately 240 ng mGM-CSF/10⁶ cells/24 hours and was chosen for further studies because this level of mGM-CSF production closely matches that of PA317/GlmGmSvNa.25, and Balb/GlmGmSvNa.5. All transduced clones used for in vivo studies were found to have intact proviral DNA when analyzed by Southern analysis. In addition, Northern analyses showed long and short proviral messages of the expected sizes.

Example 10 To test whether mGM-CSF-secreting cells could suppress local tumor growth, PA317/GlmGmSvNa.25 cells were mixed with the murine renal carcinoma cell line Renca (Murphy, et al . , J. Nat. Cancer Inst. , Vol. 50, pgs. 1013-1025 (1975)) . In the Renca experiment, PA317/GlNa.40 or PA317/GlmGmSvNa.25 producer cells and nonirradiated Renca (1x10^s cells per mouse) were mixed at either 1:1 or 10:1 ratios and implanted intradermally in the right flanks of Balb/cJ mice (Jackson Laboratories, Bar Harbor, Maine) . Tumor onset and progression was monitored at the injection site and is plotted in Figure 20. These results suggest that animals receiving producer cell - Renca mixtures (control and GM-CSF groups) developed tumors with similar onset and progression profiles. In a similar experiment, PA317/GlNa.40 or PA317/GlmGmSvNa.25 producer cellr were mixed with nonirradiated B16F10 cells (IxlO⁴ cells per mouse) at a 10:1 ratio and implanted intradermally in the right flanks of C57B1/6J mice (Jackson Laboratories) . Observation of tumor onset and progression at the injection site showed that the presence of PA317/GlmGmSvNa.25 in the mixture significantly delayed or prevented (1/10 mice) tumor development (Figure 21) . The two experiments cited above demonstrate that allogeneic fibroblasts that secrete mGM-CSF (PA317/GlmGmSvNa.25) can delay, but generally do not prevent, tumor progression.

Bhr-annpl t. \_

NIH3T3 fibroblasts were derived from NIH Swiss mice and therefore are allogeneic to C57B1/6 mice. The following experiment tests whether allogeneic fibroblasts that secrete mGM-CSF are capable of functioning in a tumor vaccine setting, or whether they would be rapidly rejected. This question is pertinent to a clinical setting in which autologous fibroblasts may not be available. Mixtures of mGM-CSF-secreting, allogeneic fibroblasts and irradiated B16F10 tumor were tested for their ability to elicit systemic immunity to wild-type tumor challenge. In a first experiment, 1x10^s PA317/GlmGmSvNa.25 or PA317/GlNa.40 fibroblasts were mixed with an equal number of irradiated (5000 rads) B16F10 tumor cells. The mixture was implanted intradermally in the flanks of C57B1/6J mice. All mice were challenged on the contralateral flank . with IxlO⁵ nonirradiated, wild-type B16F10 9 days after the vaccination and tumor onset and progression at the challenge site were monitored. The results shown in Table VII below indicate that vaccination with mixtures of PA317/GlmGmSvNa.25 with irradiated B16F10 generates systemic immunity to wild-type tumor challenge.

Table VII

In another experiment, mixtures of 10⁶ irradiated or non¬ irradiated PA317 or NIH3T3 fibroblasts (transduced with GlmGmSvNa, or GINa, or non-transduced) and 10^€ irradiated B16F10 cells were implanted intradermally into C57B1/6J mice. 5xl0⁴ wild-type B16F10 cells were injected in the contralateral flank 11 days later. The number of mice in each group with tumor at 109 days post-challenge was determined. The groups of mice receiving each vaccination mixture and the number of mice in each group with tumor at 109 days post-challenge are given in Table VIII below.

Table VIII

Vaccination Mixture # with tumor 109 days

Fibroblast type ( 10⁶) Tumor cell ( 10⁶) post - challenge

Nonirr PA317 /GlmGmSvNa . 25 IrrB16F10 0/10

IrrPA317 /GlmGmSvNa .25 IrrB16F10 0/10

I rrPA31"V GlmGmSvNa . 25 — 4 / 5

Nonirr PA317/GlNa . 40 IrrB16F10 3 /10

IrrPA317/GlNa .40 IrrB16F10 9/10

Irr PA317 /GINa . 0 — 5 /5

IrrPA317 IrrB16F10 8/10 IrrPA317 — 5/5

IrrNIH3T3/GlmGmSvNa.1 IrrB16F10 0/10

IrrNIH3T3/GlmGmSvNa.1 2/5

IrrNIH3T3/GlNa IrrB16F10 6/10

IrrNIH3T3/GlNa -- 2/5

IrrNIH3T3 IrrB16F10 4/10

IrrNIH3T3 -- 5/5

-- IrrB16F10 3/5

-- -- 4/5

The results given in Table VIII above demonstrate that 100% of mice vaccinated with mixtures of either PA317/GlmGmSvNa.25 and B16F10 cells or NIH 3T3/GlmGmSvNa.1 and B16F10 cells reject wild-type tumor challenge. These results indicate that vaccination with mixtures of irradiated tumor cells and either PA317 or NIH 3T3 fibroblasts elicits systemic immunity to tumor challenge in all animals.

Example 12

The example describes experiments performed to test the efficacy of syngeneic murine GM-CSF secreting fibroblast- tumor cell mixtures as tumor vaccines. In a first experiment lOBalb/cAnN mice were injected intradermally in the flanks with 10⁶ irradiated Renca cells; 20 mice were injected with 10⁶ irradiated Balb 3T3/GlNa cells and 10⁶ irradiated Renca cells; 20 mice were injected with a mixed population of 10⁶ irradiated Balb 3T3/GlmGmSvNa cells and 10* irradiated Renca cells; and 20 mice received 10⁶ irradiated Balb 3T3/GlmGmSvNa.5 (clone 5) cells. Most mice rejected a low dose challenge of 1x10^s or 5xl0⁴ Renca cells administered 10 days after vaccination. A second, high dose (2x10* cells) Renca challenge was injected into the contralateral flank 48 days post-vaccination, and tumors were monitored. The survival data are shown in Figure 22.

As shown in Figure 22, 100% of the mice vaccinated with the murine GM-CSF vaccine are surviving 70 days after the high dose challenge. In comparison, 40 to 50% of the control mice are surviving at this time point. This experiment demonstrates, although vaccination with control mixtures or with tumor cells alone generates some systemic immunity to tumor challenge, this effect is enhanced significantly by the inclusion of GM-CSF-secreting syngeneic fibroblasts into the vaccination mixture.

A similar experiment then was conducted in which Balb/cAnN mice were injected intradermally with (i) 10* irradiated Renca cells; or (ii) 10* irradiated Renca cells and 10* irradiated Balb/GlmGmSvNa.5 cells; or (ii) 10* irradiated Renca cells and 10* irradiated Balb/GlNa cells . 11 days after vaccination, the mice received Renca cells intradermally in amounts of 1x10^s; or 5x10^s; or 1x10*; or 5x10*; or IxlO⁷ cells. The number of mice in each group with tumors at 57 days post- challenge is given in Table IX below.

Table IX

Vaccination Number o£ mice with tumors at Composition various Renca challenge doses

Tumor Fibroblast 10⁵ 5X1 10' 5xl0« 10⁷ 0⁵

10⁶ irr. Renca 0/5 3/5 6/10 9/10 6/6

10⁶ irr. Renca lO'irr.BJ B/ 0/5 1/5 2/10 5/10 2/5 GlmGmSvNa.5

10⁶ irr. Renca 10⁶ irr.BALB/GINa 0/5 2/5 7/10 8/10 5/5

As shown in Table IX, 60% of the mice receiving a GM-CSF vaccination reject a IxlO⁷ Renca cell challenge, where as none of the control animals remain tumor-free at 57 days post- challenge .

Example 13

The human GM-CSF cDNA was originally derived by RTPCR of RNA from human Mo T cell leukemia cells (ATCC CRL 8066) because these cells are known to express GM-CSF. The GM-CSF coding sequence was subsequently cloned into retroviral vectors to create pGlGmSvNa and pGlNaSvGm. pGlGmSvNa and pGlNaSvGm were constructed as follows: PCR of the human GM-CSF from Mo cell reverse transcribed RNA is carried out using the following primers: 5' -GCA CAG CTT TCC GCA GCA GCC GCC ACC ATG TGG CTG CAG AGC CTG which includes a Bglll site (bp3-8) , a ribosome binding site (bp4- 18; see also reference to ribosome binding site in the pBg description) , a Kozak sequence (bpl9-30) , and GM-CSF open reading frame sequence from the ATG translation initiation codon (bp 28-45) ; and 5' -GGC AAG CTT GTC GAC TCA CTC CTG GAC TGG CTC where bp 4-9 are a Hindlll site, bp 10-16 are a Sail site, and bp 17-33 are the antisense se-quence to the last 16 bases of the human GM-CSF open reading frame. The PCR product is purified on an agarose gel, and restriction enzyme digested with Bglll and Hindlll. The fragment is ligated to the retroviral vector pGlNaSvX that has been digested with Bglll and Hindlll. (pGlNaSvX is a plasmid derived from pGlNaSvBg (Mc Lachlin, et al . , Virology, Vol. 195, pgs. 1-5 (1993)) , in which the lacZ gene was excised by digestion of pGlNaSvBg with Bglll and Hindlll.) Ligated plasmid DNA was transformed into E. coli DH5α competent cells and DNA from ampicillin-resistant colonies is screened by restriction enzyme digestion. Plasmids that give the expected restriction fragments are termed pGlNaSvCm. The dideoxy sequence of the GM-CSF cDNA in one such clone is determined and aligned to the expected sequence of human GM-CSF (Wong et al . , Science, Vol. 228, pgs. 810-814 (1985)) to confirm correctness. To generate pGlGmSvNa, pGlNaSvGm is restriction enzyme digested with Bglll and end-filled with Klenow. The cut plasmid is then digested with Sail to isolate the fragment containing the GM-CSF open reading frame. This fragment is ligated to pGlXSvNa that has been digested with SnaBI and Sail (both cut in the polycloning "X" region) . Ligated plasmid is transfomred into DH5α and clones are screened as described above for pGlNaSvGm. The human GM-CSF coding sequence used in the construct pGIGm (Figure 23) was subcloned from the construct pGlNaSvGm. The GM-CSF open reading frame was excised from pGlNaSvGm using restriction enzymes Bglll and Hindlll. This fragment was ligated into pGl that had been digested with restriction enzymes BamHI and Hindlll (these enzymes cleave pGl in the multicloning region) . Ligated plasmid DNA was transformed into E.Coli DH5α competent cells and DNA from ampicillin- resistant colonies was screened by restriction enzyme digestion. Plasmids that appeared by restriction mapping to contain the GM-CSF insert were termed pGIGm. The dideoxy sequences of the cloning junctions (i.e. the regions where the fragments were joined by ligation) were determined and found to be correct. The sequence of the GM-CSF cDNA in pGlNaSvGm had previously been determined to be correct.

Because the plasmid pGIGm does not contain a selectable marker such as the neomycin resistance gene, 30 micrograms of pGIGm were cotransfected with 3 micrograms of pSV₂Neo (Powels, e al., Cloning Vectors. Elsevier, pg. VIII-B-b-6-1-9 (1986)) into the PA317 producer cell line. Standard calcium phosphate transfection methods were used (Chen, et al . , 1988) . Transfected cells were diluted into media containing C .6mg/ml G418. Sixty-six PA317/GlGm clones were isolated and the cell supernatants collected. The relative number of retroviral particles in the clone supernatants was compared by RNA titering and the 9 clones with highest RNA titer were further analyzed by ELISA for human GM-CSF (R&D Systems) . Standard biological titering methods cannot be employed for retroviral vectors that do not include a selectable marker such as neomycin. Instead, a Southern analysis titering method was used to titer GlGm clones the were positive by ELISA.

30 micrograms of pGlGmSvNa or pGlNaSvGm were transfected into the PA317 producer cell line to form PA317/GlGmSvNa or PA317/GlNaSvGm producer cells. Clones were screened using a human GM-CSF ELISA assay from R&D Systems. PA317/GlGmSvNa clone 9 and PA317/GlNaSvGm clone 22 were found to have titers of about 5 x 10^s cfu/ml, and were chosen for further studies.

Equal amounts (3ml) of PA317/GlGm, PA317/GlGmSvNa and PA317/GlNaSvGm viral supernatants then were used to transduce 3T3 cells; DNA extracted from cell pellets was subsequently analyzed by Southern Blot. Concurrently, GIGmSvNa and GlNaSvGm supernatants were biologically (G418) titered. Band intensities on Southern corresponded to transduction efficiency (i.e. titer) . Band intensities were measured on a Molecular Dynamics Phosphorlmager:SF. Since the G418 titers of GIGmSvNa and GlNaSvGm were known, titers of the GlGm clones that were co-run on the Southern could be extrapolated. PA317/GlGm clone 55 was chosen for further studies based on RNA titer, ELISA, and Southern titer data. The approximate titer of PA317/GlGm.55 is 1x10* cfu/ml. PA317/GlGm.55 makes approximately 44 ng GM-CSF/10* cells/24 hours as measured by ELISA. The amount of biologically active human GM-CSF in PA317/GlGm.55 supernatant was quantitated utilizing the GM-CSF-sensitive cell line M07e (Avanzi et al. , J.Cell.Phvsiol.. Vol. 145, 458-464 (1990)) . The bioassay (IIT Research Institute, Chicago, 111.) demonstrated that PA317/GlGm.55 produced approximately 57 ng GM-CSF/10* cells/24 hours. Southern analysis of PA317/GlGm.55 DNA has demonstrated only one proviral band of the expected size. The retroviral long message of the expected size is made by PA317/GlGm.55, as tested by Northern analysis.

To create human GM-CSF-producing fibroblasts for clinical therapy, NIH 3T3 fibroblasts are transduced with GlGm.55 supernatant and diluted for cloning. Supernatants from clones are assayed by ELISA for expression of GM-CSF. Clones expressing GM-CSF are assayed for the presence of the intact provirus by Southern. A clone expressing high levels of GM-CSF and bearing the intact GlGm provirus is chosen for clinical applications. Example 14

Patients with surgically inaccessible head and neck tumors are treated by conventional chemo- and radiation therapies. In cases where past experience indicates a better than 50% likelihood that the cancer will recur, 10^s to 10¹⁰ irradiated human fibroblasts or universal cells which express human Interferon-α, and which are engineered by transduction of the retrovirus GIFlSvNa into the fibroblasts, or universal cells (preferably 10* to 10^β) , are injected directly into the tumor site in order to suppress the local regrowth of the tumor. At the same time or shortly thereafter, the same number of irradiated fibroblasts or universal cells which express GM-CSF, which are engineered by transduction of the retrovirus GIGmSvNa into the fibroblasts or universal cells are injected in a similar manner directly into the tumor site in order to suppress the tumor by stimulating a systemic response from the immune system. Additionally, the injections with the universal cells can be repeated at 1 to 3 week intervals. In patients with surgically accessible tumors, the preferred method is to resect surgically the tumor, treat with conventional chemo- and radiation therapies then to suppress the local regrowth with irradiated universal PA317/GlFlSvNa cells (10^s to 10¹⁰, preferably 10* to 10^β) . The tumor is homogenized to a single cell suspension by standard methods and mixed with an equal number of universal PA317/GIGmSvNa cells (10^s to 10¹⁰, preferably 10* to 10^β) . The cell mixture is then irradiated prior to vaccination intramuscularly, intradermally or subcutaneously at a site distal to the tumor to stimulate a systemic immune response. It is anticipated that the combination therapy with IFNα2 and IFNγ will result in an increase in patient survival greater than the use of either therapy by itself.

The disclosure of all patents, publications (including published patent applications) , and database accession numbers, and depository accession numbers referenced in this specification are specifically incorporated herein by reference in their entirety to the same extent as if each such individual patent, publication, and database accession number, and depository accession number were specifically and individually indicated to be incorporated by reference.

It is to be understood, however, that the scope of the present invention is not to be limited to the specific embodiments described above. The invention may be practiced other than as particularly described and still be within the scope of the accompanying claims .

Claims

WHAT IS CLAIMED IS:

1. A method of treating a tumor in a host, comprising: administering to a tumor site non-tumor cells which express a therapeutically effective amount of one or more agents selected from the group consisting of an interferon, and a tumor necrosis factor.

2. The method of Claim 1 wherein said cell is a fibroblast.

3. The method of Claim 1 wherein said non-tumor cell is a retroviral producer cell capable of generating a retroviral vector particle including a polynucleotide encoding an agent selected from the group consisting of an interferon and a tumor necrosis factor.

4. The method of Claim 1 wherein said non-tumor cells were transduced with a vector including at least one polynucleotide encoding an agent selected from the group consisting of an interferon and a tumor necrosis factor.

5. The method of Claim 4 wherein said vector is a retroviral vector.

6. The method of Claim 1 wherein said agent is an interferon.

7. The method of Claim 6 wherein said interferon is Interferon-γ.

8. The method of Claim 6 wherein said interferon is Interferon-α.

9. The method of Claim 1 wherein said cells are administered in an amount of from about 10^s cells to about 10¹⁰ cells.

10. The method of Claim 1 wherein said agent is a tumor necrosis factor.

11. The method of Claim 10 wherein said tumor necrosis factor is TNF-α.

12. A method of treating a tumor, comprising: administering to a tumor site non-tumor cells which express a therapeutically effective amount of GM-CSF to produce a systemic immune response against the tumor.

13. A method of treating a tumor in a host comprising: administering to a host a mixture of (a) non- tumor cells engineered with a polynucleotide encoding GM-CSF; and (b) an agent selected from the group consisting of tumor cells of the type of tumor which is being treated and a tumor antigen of the type of tumor which is being treated.

14. The method of Claim 13 wherein said tumor cells are irradiated prior to administration of said tumor cells to said host .

15. A method of treating a tumor in a host, comprising:

(i) administering to a tumor site non-tumor cells which express a therapeutically effective amoung of one or more agents selected from the group consisting of an interferon, and a tumor necrosis factor; and

(ii) administering to a host a mixture of (a) non-tumor cells engineered with a polynucleotide encoding a cytokine; and (b) an agent selected from the group consisting of tumor cells of the type of tumor which is being treated and a tumor antigen.

16. The method of Claim 15 wherein said non-tumor cells are fibroblasts.

17. The method of Claim 15 wherein said non-tumor cells are retroviral producer cells capable of generating a retroviral vector particle including at least one polynucleotide encoding an agent selected from the group consisting of an interferon, and a tumor necrosis factor.

18. The method of Claim 15 wherein said non-tumor cell which expresses one or more agents selected from the group consisting of an interferon, and a tumor necrosis factor includes an expression vehicle including a polynucleotide encoding an agent selected from the group consisting of an interferon, and a tumor necrosis factor.

19. The method of Claim 18 wherein said expression vehicle is a retroviral vector.

20. The method of Claim 15 wherein said agent is an interferon.

21. The method of Claim 20 wherein said interferon is Interferon α.

22. The method of Claim 20 wherein said interferon is Interferon γ.

23. The method of Claim 15 wherein said agent is a tumor necrosis factor.

24. The method of Claim 23 wherein said tumor necrosis factor is TNF-α.

25. The method of Claim 15 wherein said cells which express one or more agents selected from the group consisting of an interferon and a tumor necrosis factor are administered in an amount of from about 10^s to about 10^1C cells.

26. The method of Claim 15 wherein said cytokine is selected from the group consisting of GM-CSF, Interleukin-2, Interleukin-4, and Interleukin-12.

27. The method of Claim 26 wherein said cytokine is GM- CSF.

28. The method of Claim 15 wherein said non-tumor cell engineered with a polynucleotide encoding a cytokine has been transduced with a retroviral vector including said polynucleotide encoding cytokine.

29. A kit for treating a tumor, comprising:

(a) non-tumor cells expressing an agent selected from the group consisting of an interferon and a tumor necrosis factor;

(b) non-tumor cells engineered with a polynucleotide encoding GM-CSF; and

(c) tumor cells of the type of tumor which is being treated.