HRP950156A2 - Tumor necrosis factor - Google Patents

Description

The technical field to which the invention belongs

The invention is from the field of genetic engineering.

Technical problem

This application relates to lymphokines. Specifically, it refers to cytotoxic factors secreted by lymph cells and to procedures for making them in recombinant cells.

State of the art

It is known that immune cells such as B cells, T cells, natural killer cells and macrophages create substances that show cytotoxic activity against tumor cells but are not harmful to normal cells. These substances are variously named, for example, lymphotoxin, tumor necrosis factor, NK cell cytotoxic factor, hemorrhagic necrosis factor, macrophage cytotoxin, or macrophage cytotoxic factor. Currently, the identities of the proteins accompanying these names are unclear. The main difficulties are that the biological assays used for the detection of proteins do not distinguish between them, that proteins appear to occur in nature as aggregates or hydrolysis products, and that the quantities obtained so far have been too small so that a high degree of purification was required for complete protein characterization which has not been achieved.

Typically, such cytotoxic substances are found in the sera of intact animals, or in the supernatants of lymphoid cell cultures or cell lines after the animals or cells have been treated with a substance known to stimulate the proliferation of immune cells ("inducer"). After that, the serum or supernatant is regenerated and tested for cytotoxic activity against the desired tumor cell line. This cell line and others used in bioassays of this type are nonspecific in their lytic reaction because different apparently discrete products of the lymph cells are capable of performing termination. Similar non-specific reactions are observed in vitro in tumor necrosis assays. Thus, cytolytic assays that detect disruption of cell lines in vitro or tumor necrosis in vivo are inadequate to distinguish between various cytotoxic lymph products.

Cytotoxic factors are tentatively classified based on the type of cells from which they are induced. For example, lymphotoxin is a name commonly applied to cytotoxic secretory products of B or T lymphocytes, or cell lines derived from them, while tumor necrosis factor is often used to describe cytotoxic products of macrophages or their derived cell lines. However, this classification system has not been developed to the point where there is any certainty that it is only one protein, or proteins that have different names but are actually different.

Attempts have been made to purify and characterize the cytotoxic factors secreted by each cell type. To the extent that reports vary with respect to the property of a cytotoxic factor, or are completely inconsistent with a given property, it can be concluded either that the characterization was erroneous or that many discrete cytotoxic factors are secreted by each cell type. For example, cytotoxic products derived from macrophages, monocytes or monocytic cell lines, although sometimes referred to generically as tumor necrosis factors, have been shown to have properties that appear to be consistent with the single cytotoxic product theory. See, for example, the following literature: R. MacFarlan et al., 1980, "AJEBAK" 58 (pt 5): 489-500 D. Mannel et al., 1980, "Infection and Immunology" 30 (2): 523- 530 H. Ohnishi et al., UK Patent Application 2,106,117A; and J. Hammerstrom, 1982, "Scand J. Immunol." 15: 311-318.

On the other hand, C. Zacharchuk et al., 1983, "Proc. Nat. Acad. Sci. USA", 80: 6341-6345 suggest that guinea pig lymphotoxin and guinea pig macrophage cytotoxic factor are similar, if not identical. Similar conclusions are found in Ruff et al., 1981, Lymphokines Vol. 2 pp. 235-272 on pp. 241-242.

Attempts to characterize immune cytotoxic factors have also focused on using as starting material the serum or peritoneal fluid of animals that have been exposed to immunogenic antigens. These sources contain the entire cornucopia of the stressed immune system, as opposed to the product or products of a single cell type or line. The following should be understood as publications of this type: S. Green et al., 1982, "J. Nat. Cancer Inst." 68(6): 997-1003 ("tumor necrosis induction factor"); M. Ruff et al., 1980, "J. Immunology" 125(4): 1671-1677 ("tumor necrosis factor"); H. Enomoto et al., European Patent Application 86475 ("antitumor substance"); H. Oettgen et. al., 1980, "Recent Results Cancer Res." 75: 207-212 ("tumor necrosis factor"); F. Kull et al., 1981, "J. Immunol." 126(4): 1279-1283 ("Tumor Cell Cytotoxin":); D. Mannel et al., 1980, "Infection and Immunity" 28(1): 204-211 ("cytotoxic factor"); N. Matthews et al., 1980, "Br. J. Cancer": 42: 416-422 ("tumor necrosis factor"); S. Green et al., 1976, "Proc. Nat. Acad. Sci. USA:, 73(2); 381-385 ("serum factor"); N. Satomi et al., 1981, "Jpn. J. Exp. Med." 51(6): 317-322; N. Matthews, 1979, "No. J. Cancer" 40: 534-539 ("tumor necrosis factor"); K. Haranaka et al., 1981, "Jpn. J. Exp. Med." 51(3): 191-194 ("tumor necrosis factor"); and L. Old et al., European Patent Application 90892; T. Umeda et al., 1983, "Cellular and Molecular Biology" 29(5 ): 349-352; and H. Enomoto et al., 1983, European Patent Application 86,475.

Further literature to be consulted is J. Nissen-Meyer et al., 1982, "Infection and Immunity" 38(1); 67-73; J. Klostergaard et al., 1981, "Mol. Immunol." 18(12): 1049-1054; N. Sloan, U.S. patent 4,359,415; and H. Hayashi et al., U.S. Patent 4,447,355; K. Hanamaka et al., 1983, European Patent Application 90,892; and G. Granger et al., 1978, "Cellular Immunology" 38: 388-402.

European Patent Application Publn. But. 100641 describes a cytotoxic polypeptide substantially purified to be free of impurities from a human lymphoblastoid cell culture. This polypeptide has been designated a lymphotoxin, although its relationship to other cytotoxic polypeptides named lymphotoxin is unclear. It was not known whether this was the only cytotoxic polypeptide produced by immune cells, as suggested by Zacharchuk et al. (Id.), or was one of a potential family of cytotoxic factors.

The polypeptide from Application 641 has two amino termini, with the larger variant ending in Leu Pro Gly Val Gly Leu Thr Pro Ser Ala Ala Gln Thr Ala Arg Gln His Pro Lys Met His Leu Ala His Ser Thr... and the smaller variant being blunted with the amino terminus His Ser Thr Leu Lys Pro Ala Ala...

According to earlier literature, interferons, which show some tumor-inhibiting activity, and a poorly characterized protein having an AlaAla amino terminus (U.K. Patent Application Publn. No. 2,117,385A), were candidates for non-lymphotoxin cytotoxic factors. As will be seen, the tumor necrosis factor of the present invention is not an interferon, nor is it a lymphotoxin and does not have an AlaAla amino terminus.

The object of this invention is (a) to conclusively determine whether or not there is another tumor necrosis factor besides lymphotoxin and, if so, to identify it in such a way that it is clearly distinguishable from other such factors; (b) to produce such factor by methods other than induced cell culture, which is expensive and produces a product that is contaminated with the induction agent, or by induction of peripheral blood lymphocytes, which is economically impractical, poorly reproducible, and produces a product that is contaminated with homologous proteins of cells and plasma; (c) to obtain DNA encoding such tumor necrosis factor and to express the DNA in recombinant culture; (d) to synthesize such factor in a recombinant culture in a mature form; (e) to modify the coding sequence or structure of such factor; (f) formulating such factor into therapeutic dosage forms and administering to animals for the treatment of tumors; and (g) to produce diagnostic reagents related to such factor.

Description of the solution to the technical problem with implementation examples

The cytotoxic factor was purified to homogeneity, characterized and expressed in recombinant culture. This factor is designated tumor necrosis factor (TNF) for convenience and is defined below. It is provided in the bulk in a homogeneous form of cell culture at a specific activity greater than about 10 million units/mg of protein, and usually about 100 million units/mg.

Human tumor necrosis factor synthesized in recombinant culture is characterized by the presence of non-human cellular components, including proteins, in amounts and characters that are physiologically acceptable for administration to patients in accordance with the tumor necrosis factor. These components will usually be of yeast, prokaryotic or non-human higher eukaryotic origin and are present in harmless contaminant amounts of the order of less than 1 wt. percentage.

Furthermore, recombinant cell culture enables the production of tumor necrosis factor that is absolutely free of homologous proteins. Homologous proteins are those that normally accompany tumor necrosis factor as found in nature, eg, in cells, cell exudates or body fluids. For example, the homologous protein for human tumor necrosis factor is human serum albumin. Heterologous proteins are the opposite, i.e. they are naturally monitored and are not found in combination with the tumor necrosis factor in question.

DNA that codes for tumor necrosis factor and which, when expressed in recombinant or transformed culture, yields copious amounts of tumor necrosis factor is provided. This DNA is novel because cDNA obtained by reverse transcription of mRNA from an induced monocytic cell line does not contain introns and is free of any flanking regions that code for other proteins of the organism from which the mRNA originates.

Chromosomal DNA coding for THF was obtained by probing genomic DNA libraries with cDNA. Chromosomal DNA is free of flanking regions that code for other proteins but may contain introns.

Isolated tumor necrosis factor DNA is readily modified by nucleotide substitution, deletion, or insertion, leading to new DNA sequences encoding tumor necrosis factor or its derivatives. These modified sequences are used to produce mutant tumor necrosis factor and to directly express mature tumor necrosis factor. Modified sequences are also useful in enhancing the efficiency of expression of tumor necrosis factor in selected host-vector systems, eg, by modification due to adaptation to the preferential codon of the host cells.

These new DNA sequences or fragments thereof are labeled and used in hybridization assays for genetic material coding for tumor necrosis factor.

In procedures for the synthesis of tumor necrosis factor, DNA encoding tumor necrosis factor is ligated into a replicating (reproducing) vector, the vector is used to transform host cells, the host cells are cultured, and tumor necrosis factor is regenerated from the culture. This general procedure is used for the construction of tumor necrosis factor that has the characteristics of monocyte tumor necrosis factor or for the construction of vectors and host cells selected for transformation. Types of tumor necrosis factor that can be synthesized herein include mature (valyl amino-terminal) tumor necrosis factor, pre-tumor necrosis factor ("pre-TNF", as defined herein), and TNF derivatives including (a) fusion proteins wherein TNF or any fragment thereof (including mature tumor necrosis factor) linked to other . proteins or polypeptides by peptide linkage to the amino and/or carboxyl terminal amino acids of TNF or fragments thereof, (b) TNF fragments, including mature tumor necrosis factor or pre-TNF fragments in which any preprotein amino acid is an amino terminal amino acid of the fragment, (c) TNF mutants or fragments thereof (including mature tumor necrosis factor) where one or more amino acid residues are substituted, inserted or omitted, and/or (d) methionyl or modified methionyl (such as formyl methiones or other blocked methionyl species) amino-terminal addition derivatives previous proteins, fragments or mutants.

Typically, if a mammalian cell is transformed with (a) a vector containing the entire structural tumor necrosis factor gene (including the 5' start codon), or (b) a mature tumor necrosis factor gene or a mature tumor necrosis factor derivative that is operably linked to a eukaryotic secretory leader sequence (which may also include a tumor necrosis factor secretory leader presequence), and the cell is then cultured, then mature tumor necrosis factor is regenerated from the culture.

Similarly, if the DNA encoding TNF is operably linked into a vector for a secretory leader sequence that has been properly processed by a host cell for transformation (usually the organism from which the leader sequence was obtained), and the cell is transformed with the vector and cultured, then synthesizes tumor necrosis factor without amino-terminal methionyl or blocked methionyl. Specifically, E. coli transformed with vectors in which DNA encoding mature tumor necrosis factor is linked 5' to DNA encoding STII enterotoxin signal polypeptide will properly process a high percentage of the hybrid preprotein into mature tumor necrosis factor. Secretory leader sequences and host cells can be chosen so that proper transport of the mature protein into the cell periplasm occurs.

Also within the scope of this invention are derivatives of tumor necrosis factor that differ from amino acid sequence or glycolysis variations. Such derivatives are characterized by covalent or aggregative association with chemical species. Derivatives mainly fall into three classes: salts, covalent modifications of the side chain and terminal residue, and adsorption complexes.

If DNA coding for mature tumor necrosis factor is operatively ligated into a vector, and the vector is used to transform a host cell and the cell is cultured, mature tumor necrosis factor is found in the cytoplasm of the cells. Therefore, it is not necessary to develop secretion systems in order to obtain mature tumor necrosis factor. This is unexpected since direct expression normally yields a methionylated protein. Furthermore, the protein is stable and soluble in recombinant cell culture, i.e. it is neither proteolytically interrupted by intracellular proteases nor deposited in the form of refractile bodies. Accordingly, novel fermentations are provided which involve lower eukaryotic or prokaryotic cells having non-methionylated mature tumor necrosis factor present in the cytoplasm of such cells.

Although tumor necrosis factor can be made by culturing animal cell lines, e.g., a monocytic cell line that is induced to grow in the presence of 4-beta-phorbol-12-myristate-13-acetate (PMA) or such immortalized cell lines as hybridomas or EBV transformed cells (U.S. Patent 4,464,465), preferably the tumor necrosis factor is synthesized in recombinant cell culture as further described below.

Once the tumor necrosis factor is made by fermentation, it is mainly purified by regenerating the supersaturated culture fluid or culture with interrupted cells, separating the solid substances, adsorbing the tumor necrosis factor from the supernatant mixture (which contains the tumor necrosis factor and other proteins) onto some hydrophobic substance, elution of the factor of tumor necrosis from the substance, by adsorption of tumor necrosis factor on a tertiary amino-anion exchange resin, elution of tumor necrosis factor from the resin, adsorption of tumor necrosis factor on an anion exchange resin (preferably quaternary amino-substituted) which has an essentially compatible particle size, and elution of the non-goose factor tumors with resin. Optionally, preparations of tumor necrosis factor are concentrated and purified by chromatofocusing in a purification procedure, for example by isoelectric focusing or passage through a molecular gel such as Sephadex G-25.

Purified tumor necrosis factor from recombinant or induced cell culture is combined for therapeutic use with physiologically harmless stabilizers and compounds and formulated into a dosage form, for example, by lyophilization into dosage vials, or stored in stabilized aqueous preparations. Alternatively, tumor necrosis factor is incorporated into a polymer matrix for implantation into tumors or surgical sites from which tumors have been excised, so that a "timed" release of tumor necrosis factor at a localized high concentration is effected.

The preparations are obtained without contaminating cytotoxic factors such as lymphotoxin, interferons or other cytotoxic proteins known from the literature. However, in therapeutic applications, tumor necrosis factor is conveniently combined with predetermined amounts of lymphotoxin and/or interferon. Preparations containing tumor necrosis factor and interferon, such as interferon gamma, are particularly useful as they have been found to exhibit synergistic cytotoxic activity.

Tumor suppressing factor preparations are administered to animals, especially to patients with malignant tumors, in therapeutically effective doses. Suitable dosages will be apparent to one skilled in the art of dosing therapy as further discussed below.

Brief description of the drawing

Sl. 1 shows the elution profile of tumor necrosis factor from the control porous glass.

Sl. 2 shows the elution profile of tumor necrosis factor from diphenylaminoethyl cellulose.

Sl. 3 shows tumor necrosis factor elution profiles from fast protein liquid chromatography.

Sl. 4 shows the elution profile of tumor necrosis factor after chromatofocusing.

Sl. 5 shows the molecular weight of tumor necrosis factor by SDS PAGE gel electrophoresis.

Sl. 6 shows the molecular weight of tumor necrosis factor after HPLC elution.

Sl. 7 shows the elution profile of tumor necrosis factor from the HPLC C4 column.

Sl. 8 illustrates the separation of tumor necrosis factor in the form of trypsin-digested fragments on HPLC.

Sl. 9 illustrates the cytotoxic effect of mixtures of gamma-interferon and tumor necrosis factor.

Sl. 10 shows the nucleotide and amino acid sequence for pre-human tumor necrosis factor, including the complete tumor necrosis factor secretory leader sequence.

Sl. 11 shows the construction of a tumor necrosis factor expression vector.

Sl. 12 describes the E. coli STII heat-stable enterotoxin gene.

Tumor necrosis factor is defined for the purposes of this invention as a polypeptide distinct from a lymphotoxin that has preferential cytotoxic activity and has a region showing functional amino acid homology to the mature tumor necrosis factor amino acid sequence shown in FIG. 10, its fragment, or with a derivative of such a polypeptide or fragment.

Preferential cytotoxic activity is defined as preferential destruction or growth inhibition of tumor cells compared to normal cells under the same conditions. Preferential cytotoxic activity is detected by the effect of the polypeptide on tumor cells in vivo or in vitro compared to normal cells or tissue. Cell disruption is mainly a diagnostic index in vitro, while tumor necrosis is determined by in vivo experiments. However, cytotoxic activity can be manifested as cytostasis or as antiproliferative activity. Suitable test systems are well known. For example, the cell disruption assay for determining the specific activity of tumor suppressor factor described below is acceptable, as is the assay described in B. Aggarval et al., in Thymic Hormones and Lymphokines, 1983, ed. A. Goldstein, Spring Symposium on Health Sciences, George Washington University Medical Center (A549 cell line discussed in this literature is available from ATCC as CCL18S).

The specific activity of TNF is defined as disruption of the desired cells, rather than cytostasis. One unit of tumor necrosis factor is defined as the amount required for 50 percent disruption of the desired cells plated in each well according to Example 1. However, this does not preclude other assays to measure specific activity, e.g., methods based on the growth rate of the desired cells.

PredTNF is a type of tumor necrosis factor that is involved in the pre-demission of tumor necrosis factor. It is characterized by the presence of signal (or leader) polypeptide molecules that serve to post-translationally direct the protein to a location inside or outside cells.

Basically, the signal polypeptide (which will have no tumor necrosis activity of its own) is proteolytically cleaved from a residual protein that has tumor necrosis activity as part of a secretory process in which the protein is transported into the host cell periplasm or culture medium. The signal peptide can be microbial or mammalian (including the native 76 residue presequence) but is preferably a signal that is homologous to the host cell.

Native tumor necrosis factor from normal biological sources has a calculated molecular weight of 17,000 on sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) as described below, an isoelectric point of about 5.3, and a propensity for hydrolysis by trypsin at multiple sites. Native tumor necrosis factor purified by reverse-phase HPLC is hydrolyzed with trypsin into at least nine fragments under the conditions described below. The precise number of fragments into which tumor necrosis factor is hydrolyzed will depend on such factors as trypsin activity, factor concentration, tumor necrosis factor, and incubation parameters, including ionic strength, pH temperature, and incubation time.

It does not appear that the tumor necrosis factor is a glycoprotein, since it is not retained on the lectin affinity column and the analysis of the deduced amino acid sequence has no likely sites for glycosylation. Also, material produced in recombinant E. coli culture (which lacks the ability to glycosylate) comigrates with natural TNF on SDS-PAGE.

The degree of amino acid sequence homology that brings the polypeptide within the scope of the tumor necrosis factor definition will depend on whether the homology between the candidate protein and the tumor necrosis factor falls within or outside the region of the tumor necrosis factor responsible for cytotoxic activity; regions that are critical for cytotoxic activity should show a high degree of homology in order to fall within the scope of the definition, while sequences that are not involved in maintaining the conformation of tumor necrosis factor or in receptor binding may show relatively weak homology. Further, critical regions can exhibit cytolytic activity yet retain homology, as defined herein, if residues containing functionally similar amino acid flanking sequences are substituted. Functionally similar refers to the dominant characteristics of the side chains such as basic, neutral, or acidic, or to the presence or absence of a sternal mass. However, tumor necrosis factor as defined herein specifically excludes lymphotoxin.

Basically, the polypeptide that is defined as a tumor necrosis factor will contain regions that are mostly homologous with the protein from FIG. 10 or its fragments over a continuous block of about 20 to 100 amino acid residues, especially the blocks that are covered by residues 35-66 and 110-133.

The most significant factor for establishing the identity of the polypeptide as tumor necrosis factor is the ability of antisera capable of substantially neutralizing the cytolytic activity of mature tumor necrosis factor as shown in FIG. 10 to also largely neutralize the cytolytic activity of the polypeptide in question. However, it will be understood that immunological identity and cytological identity are not necessarily coextensive. Neutralizing antibody for tumor necrosis factor from FIG. 10 may not bind the candidate protein because the neutralizing antibody need not target a specific binding site on tumor necrosis factor that is critical for its cytotoxic activity. Instead, the antibody can bind a harmless region and exert its neutralizing effect through a steric effect. Therefore, the candidate protein mutated in this harmless region may no longer bind to the neutralizing antibody, but will still be a tumor necrosis factor due to essential homology and biological activity.

It is important to note that such characteristics as molecular weight, isoelectric point, and the like, for native or wild-type human tumor necrosis factor in FIG. , 10 obtained from peripheral lymphocytes or established cultures of cell lines, only descriptive for natural types of tumor necrosis factor. Tumor necrosis factor as encompassed by the foregoing definition will include other types that will not exhibit all the characteristics of native tumor necrosis factor. While tumor necrosis factor as defined herein includes native tumor necrosis factor, other related cytotoxic polypeptides will fall within the definition. For example, TNF derivatives such as insertion mutants, deletion mutants, or fusion proteins described above will yield tumor necrosis factor (fusion proteins with mature tumor necrosis factor or with preTNF itself as well as insertion mutants will have a higher molecular weight than the native mature of tumor necrosis factor, while mutants from the omission of mature tumor necrosis factor will have a lower molecular weight). Similarly, tumor necrosis factor can be engineered to reduce or eliminate the propensity for hydrolysis by trypsin or other proteases. Finally, post-translational processing of human preTNF in cell lines derived from non-primate mammals can produce microheterogeneity in the amino-terminal region, such that valine is no longer an amino-terminal amino acid.

The amino acid sequence for human tumor necrosis factor derived from its cDNA is described in FIG. 10. Mature or native tumor necrosis factor is represented by amino acid residues 1 to 157. Note that this sequence includes a 76 residue signal sequence that is believed to be cleaved off during normal processing of the translated transcript to produce the mature protein. The sites for trypsin hydrolysis are indicated by arrows.

Note that the language "capable" of cytotoxic activity or of tumor necrosis in vivo means that the term tumor necrosis factor includes polypeptides that can be translated, as by enzymatic hydrolysis, from an inactive state analogous to a zymogen into a polypeptide fragment that exhibits the desired biological activity. Typically, inactive precursors will be fusion proteins in which the mature tumor necrosis factor is peptide-linked at or near its carboxyl peptide bond and is selected to be prone to proteolytic hydrolysis to release tumor necrosis factor, either in vivo or as part of a manufacturing protocol , in vitro.

The tumor necrosis factor generated in this way will then show the cytotoxic activity required by definition.

Although tumor necrosis factor is usually intended to denote human tumor necrosis factor, tumor necrosis factor from such sources as mice, pigs, horses, or oxen is included within the definition of tumor necrosis factor as long as it meets the standards described above for homologous regions and cytotoxic activity. TNF is not type specific, eg, human TNF is active on mouse tumors. That is why TNF from one type can be used in the therapy of another.

Tumor necrosis factor also includes multimeric forms. Tumor necrosis factor spontaneously aggregates into multimers, usually dimers or multiple multimers. Multimers are cytotoxic and therefore suitable for in vivo therapy. Although it is desirable to express and regenerate tumor necrosis factor as a largely homogeneous multimer or monomer, tumor necrosis factor can be used therapeutically as a mixture of different multimers.

Derivatives of tumor necrosis factor are included in the scope of the term tumor necrosis factor. Derivatives include amino acid sequence mutants, glycosylation variants, and covalent or aggregative conjugates with other chemical types. Covalent derivatives are attached to groups found in the TNF amino acid side chains or at the N- or C-terminus, by means known in the art. These derivatives may, for example, include: aliphatic esters or amides of the carboxyl terminus or residues containing carboxyl side chains, eg, asp10; O-acyl derivatives of residues containing hydroxyl groups such as ser52, ser3, ser4 or ser5: N-acyl derivatives of an amino-terminal amino acid or residues containing amino-terminal amino acids or residues containing amino groups, eg, lysine or arginine; and cys101 and cys69 derivatives. The acyl group is selected from the group consisting of alkyl type (including C3 to C10 normal alkyl) to form alkanoyl species, and carbocyclic or heterocyclic compounds to form aroyl species. Reactive groups are preferably difunctional compounds known per se for use in cross-linking proteins into insoluble matrices via reactive side groups. Preferred sites for derivatization are on cysteine and histidine residues.

Covalent or aggregative derivatives are useful as immunoassay reagents as well as for affinity purification procedures. For example, tumor necrosis factor is insolubilized by covalent binding to cyanogen bromide-activated sepharose using known methods or adsorbed to polyolefin surfaces (with or without glutaraldehyde cross-linking) for use in the assay or purification of antiTNF antibodies or cell surface receptors. Tumor necrosis factor is also labeled with a detectable group, e.g., radioiodinated by the chloramine T procedure, covalently bound to rare earth chelates, or conjugated to another fluorescent group for use in diagnostic assays, particularly for the diagnosis of TNF levels in biological samples by competitive immunoassays. species. Such derivatives may fall outside the above definition of TNF because they do not need to show cytotoxic activity, but only anti-TNF cross-linking activity.

Mutant derivatives of tumor necrosis factor involve predetermined, ie, site-specific, mutations of TNF or its fragments. The goal of mutagenesis is to construct DNA that encodes a tumor necrosis factor as defined above, i.e., a tumor necrosis factor that exhibits cytotoxic activity against tumor cells in vitro or that induces tumor necrosis in vivo, and that retains residual sequence homology with Sl. 10, but which also shows improvement in traits and activity. Mutant tumor necrosis factor is defined as a polypeptide that otherwise falls within the definition of homology for tumor necrosis factor as set forth herein but which has a different amino acid sequence from tumor necrosis factor either by omission, substitution, or insertion. For example, tumor necrosis factor lysine or arginine residues (arginine 2, 6, 82, 44, and 131 and lysine 98, 90, and 65) can be mutated to a histidine or other amino acid residue that does not render the protein oroteolytically labile. Similarly, cysteine 101 can be replaced with other residues and cross-linked by kerosene to provide oxidative stability. It is not necessary that the mutants meet the requirements for tumor necrosis factor activity, since even biologically inactive mutants will be useful after labeling or immobilization as reagents in immunoassays. However, in this case the mutants will retain at least one epitope site that is cross-linking reactive with the tumor necrosis factor antibody.

The regions of the tumor necrosis factor molecule within residues 35 to 66 and 110 to 133 conclusively show substantial homology (50 percent) with lymphotoxin. The hydrophobic carboxy-termini (residues 150 to 157 of tumor necrosis factor) of the two molecules are also significantly conserved. Since both proteins demonstrate in vivo tumor necrosis factor cytotoxic activity, these regions are believed to be important in the shared activity of lymphotoxin and tumor necrosis factor.

As such, residues in these regions are desirable for mutagenesis aimed at directly affecting tumor necrosis factor activity in a given cell. A relatively non-conserved region at residues 67-109 of tumor necrosis factor may function to correctly position the two flanking relatively homologous regions into a conformation that is primarily for cytotoxic activity. Such an arrangement, which can be achieved by the Cys69-Cys101 disulfide bond in tumor necrosis factor, may explain the differences in specificity and activity between the two molecules. In this regard, since these residues have been postulated to represent the active core of tumor necrosis factor, they can be synthesized chemically or by deletion mutagenesis as short chain blunted polypeptides having tumor necrosis factor activity.

While the site of the mutation is predetermined, the mutation itself need not be predetermined. For example. in order to optimize the performance of mutants in position 131, arbitrary mutagenesis can be performed on the codon for arginine 131 and the expressed tumor necrosis factor is tested for the optimal combination of cytotoxic activity and protease resistance. Techniques for making substitution mutations at predetermined sites in DNA of known sequence are well known, for example, M13 is an example of mutagenesis.

Tumor necrosis factor mutagenesis is performed by making amino acid insertions, usually at the end of about 1 to 10 amino acid residues, or deletions of about 1 to 30 residues. Substitutions, omissions, insertions or any subcombination can be combined to arrive at the final construction. Insertions include amino or carboxyl-terminal fusions, eg, a hydrophobic extension added to the carboxyl terminus. However, preferably only substitutional mutagenesis is performed. Apparently, mutations in the coding DNA must not place the sequence out of the reading frame and preferably will not create complementary regions that could create a secondary mRNA structure.

Not all mutations in the DNA coding for tumor necrosis factor will be expressed in the final secreted product. For example, a major class of DNA substitution mutations are those in which a different secretory leader sequence or signal is substituted for the native human secretory leader sequence, or by deletions within the leader sequence or substitutions, where most or all of the native leader sequence is exchanged for a leader sequence that will the intended host is more likely to understand. For example, in the construction of a prokaryotic expression vector, the human secretory leader sequence is omitted in favor of bacterial alkaline phosphatase or heat-stable enterotoxin II leader sequences, and for yeast, the leader sequence is replaced in favor of yeast invertase, alpha factor, or acid phosphatase leader sequences. However, the human secretory leader sequence can be taken up by hosts other than human cell lines, most likely in cell culture of higher eukaryotic cells. When the secretory leader sequence is "understood" by the host, a fusion protein consisting of tumor necrosis factor and the leader sequence is usually cleaved at the leader sequence-tumor necrosis factor peptide bond in cases leading to secretion of tumor necrosis factor. Thus, even if mutant pre-TNF DNA is used to transform the host, and mutant pre-TNF is synthesized as an intermediate, the resulting tumor necrosis factor is usually the native, mature tumor necrosis factor.

Another major class of DNA mutants that are not expressed as tumor necrosis factor derivatives are nucleotide substitutions that are made to enhance expression, primarily to avoid amino-terminal loops in the transcribed mRNA (see pending U.S.S.N. 303,687, incorporated herein by reference) or to codons are provided that are more easily transcribed by the chosen host, eg the well-known E. coli preferred codons for E. coli expression.

Substantially homogeneous tumor necrosis factor means a tumor necrosis factor that is substantially free of other native proteins with the source from which the tumor necrosis factor was isolated. This means that the homogeneous tumor necrosis factor is substantially free of blood plasma proteins such as albumin, fibrinogen, serine proteases, alpha-globulins, cytotoxic non-tumor necrosis factor polypeptides such as lymphotoxin or interferons, or other proteins of the cell or organism it serves. as a synthetic source for tumor necrosis factor, including whole cells and certain cellular debris. However, homogeneous tumor necrosis factor may include such substances as stabilizers and ingredients described below, predetermined amounts of proteins from a cell or organism that serve as a synthetic source, proteins from other cell sources or organisms than for tumor necrosis factor, and synthetic polypeptides such as poly-L-lysine. The recombinant tumor necrosis factor that is expressed is allogeneic, eg, bacterial, belongs to the host cell and, of course, will be expressed completely without the genetic source of the protein.

Tumor necrosis factor is preferably synthesized in cultures of recombinant organisms. Neither peripheral blood lymphocytes (PBLs) nor cell lines are preferred. It is difficult in practice to obtain PBLs of one class that are free from contamination by cells of other classes, for example, to obtain microphases free of B or T cells. Such contamination will make the separation procedure applied to the products of such cells more difficult because of other potential cytotoxic factors and proteins released by the contaminating cells.

Further, tumor necrosis factor obtained from a non-recombinant culture is expensive and will consist only of native tumor necrosis factor, and such cultures do not have the flexibility of a recombinant culture to improve tumor necrosis factor characteristics.

DNA coding for tumor necrosis factor is obtained by chemical synthesis, by testing mRNA reverse transcripts from PBL or cell line cultures, or by testing genomic libraries from any cell. Suitable cell line cultures include monocytic cell lines such as the promyelocytic cell line designated "HL-60" in the art (one of which is available from ATCC as CCL 240) or a cell line. In 937 histiocytic lymphoma (ATCC CRL 1593). These and other cell lines are induced and expressed by the secretion of tumor necrosis factor by exposing the cells to chemical or physical agents known in the art, mainly tumorigenic or mitogenic agents. Tumor necrosis factor can be effectively induced in certain monocytic cell lines with PMA alone; otherwise, such conventional agents as lipopolysaccharide, staphylococcal enterotoxin B, or thymosin alpha-1 were not as effective for tumor necrosis factor induction in these cell lines.

Since a variable amount of testing is required to locate a cell line that expresses tumor necrosis factor (and therefore contains the desired mRNA) it may be simpler to synthesize the gene. The synthesis is convenient because unique restriction sites can be introduced (thus facilitating the use of genes in vectors containing restriction sites otherwise present in the native sequence) and steps that can be performed to enhance translational efficiency, as discussed below.

This DNA is covalently labeled with a detecting substance such as a fluorescent group, a radioactive atom or a chemiluminescent group by known procedures. It is then used in conventional hybridization tests. Such assays are used in the identification of TNF vectors and transformants as described in the Examples below, or for in vitro diagnosis, such as the detection of TNF mRNA in blood cells.

TNF mRNA is unexpectedly relatively rare even in induced HL60 cells, perhaps due to instability in the messenger that occurs for unknown reasons. Furthermore, the time course of the appearance of TNF mRNA after cell induction is important. TNF mRNA appears in cells after only a short period of about 4 hours post-induction. In this respect, its appearance differs from the appearance of lymphotoxin, which occurs at about 12 hours post-induction. This makes cDNA easy to overlook when one does not know what to expect. However, once its presence is realized and fully complementary DNA is made accessible, as the present description allows, it is routine to screen cDNA libraries of induced HL60 or PBLs for tumor necrosis factor cDNA using probes having sequences of such DNA. The HL60 phage libraries tested in the examples contain a relatively consistent number of positive plaques, and therefore it is clear that routine hybridization assays will identify the phage containing the desired cDNA.

Tumor necrosis factor-synthesizing cells of the HL-60 cell line are initially cultured in a conventional manner until a density of about 8-12 x 10 5 cells/ml is reached. Cells are detached from the culture, washed, transferred to serum-free medium, and cultured in PMA-containing medium. Cultivation is then continued until the desired concentration of tumor necrosis factor accumulates in the culture medium, typically about 400 tumor necrosis factor units/ml. After that, the culture supernatant is preferably clarified by centrifugation or other means for separating cellular debris from dissolved components. Centrifugation should be done at low speed so that only suspended particles are moved. The supernatant is then purified as described below.

Alternatively and preferably, tumor necrosis factor is synthesized in host cells that have been transformed with vectors containing DNA encoding tumor necrosis factor. A vector is a replicating DNA construct. Vectors are used herein either to amplify DNA encoding tumor necrosis factor and/or to express DNA encoding tumor necrosis factor. An expression vector is a replicating DNA construct in which the DNA sequence coding for tumor necrosis factor is operably linked to suitable control sequences capable of expressing necrosis factor, a transcriptional promoter, an optional transcriptional control operator sequence, a sequence coding for a suitable mRNA for ribosomal binding sites, and sequences that control the termination of transcription and translation.

Vectors include plasmids, viruses (including phage), and integrating DNA fragments (ie, which can be integrated into the host genome by recombination). Once in a suitable host, the vector replicates and functions independently of the host's genome, or can, in some cases, integrate into the genome itself. In the present application, "vector" is generic with "plasmid" but plasmids are currently the most commonly used form of vector. However, all other forms of vectors that perform an equivalent function and are, or become, known in the art are suitable for use herein. A suitable vector will contain replicon and control sequences derived from species compatible with the intended host for expression. Transformed host cells are cells that have been transformed or transfected with tumor necrosis factor vectors constructed using recombinant DNA techniques. Transformed host cells usually express tumor necrosis factor. The expressed tumor necrosis vector will be deposited intracellularly or secreted either into the periplasmic space or into the culture supernatant, depending on the chosen host cell.

DNA regions are operatively linked when they are functionally related to each other. For example, DNA for a presequence or secretory leader sequence is linked to DNA for a polypeptide if it is expressed as a preprotein that is deposited in the secretion of the polypeptide; a promoter is operably linked to a coding sequence if it controls transcription of the sequence; or the ribosome binding site is operably linked to the coding sequence if positioned to allow translation. Basically, operably linked means touched and, in the case of secretory leader sequences, also touched in the read-through phase.

Suitable host cells are prokaryotes, yeast cells or higher eukaryotes. Prokaryotes include Gram-negative or Gram-positive organisms, for example, E. coli W3110 (ATCC 27,325) described in the Examples, although other prokaryotes such as E. coli B, E. coli X1776 (ATCC 31,537), E. coli 294 are suitable. (ATCC 31,446), psudomonas species or Serratia Marcesans.

Prokaryotic host-vector systems are preferred for tumor necrosis factor expression. Although the tumor necrosis factor molecule contains two cysteine residues, thus implying the most moderate potential requirement for post-translational processing to form a potential disulfide bond, E. coli, for example, expresses biologically active tumor necrosis factor. A plethora of suitable microbial vectors are available. Generally, a microbial vector will contain an origin of replication that distinguishes the intended host, a promoter that will function in the host, and a phenotypic selection gene, for example, a gene that codes for proteins that confer antibiotic resistance or satisfy some auxotropic requirement. Similar constructions will be produced for other hosts. E. coli is typically transformed using pBR322, a plasmid derived from E. coli species (Bolivar, et al., 1977, "Gene" 2:95). pBR322 contains genes for ampicillin and tetracycline resistance, thus providing an easy means of identifying transformed cells.

Vectors must contain a promoter that recognizes the host organism. This is generally a promoter that is homologous to the intended host. The promoters most commonly used in the construction of recombinant DNA include beta-lactamase (penicillanase) and promoter systems: lactose (Chang et al., 1978, "Nature", 275: 615; and Goeddel et al., 1979, "Nature" 281: 544), the tryptophan (trp) promoter system (Goeddel et al., 1980, "Nucleic Acids Res." 8: 4057 and EPO Appl, Publ.No. 36,776) and the tac promoter /H. De Boer et al., 1980, "Proc. Nat'1. acad. Sci. U.S.A." 80: 21-25 (1983)/ While these are the most commonly used, other known microbial promoters are suitable. Details relating to their nucleotide sequences have been published, enabling one skilled in the art to operably link them to DNA coding for tumor necrosis factor in plasmid vectors (Siebenlist et al., 1980, "Cell" 20: 269) and to DNA coding for tumor necrosis vector. Currently preferred is the pBR322 derivative vector containing the E. coli alkaline phosphatase promoter with the trp ShineDalgarno sequence. The promoter and Shine-Dalgarno sequences are operably linked to DNA encoding TNF, i.e. they are positioned to promote transcription of TNF mRNA from the DNA .

In addition to prokaryotes, eukaryotic microbes such as yeast cultures have been transformed with vectors encoding tumor necrosis factor. Saccharomyces cerevisiae, or common yeast, is the most commonly used of the lower eukaryotic host microorganisms, although various other types are usually available. Yeast vectors will generally contain an origin of replication from a 2 micron yeast plasmid or an autonomously replicating sequence (ARS), promoter, TNF, polyadenylation and transcription termination sequences, and a selection gene. A suitable plasmid for expressing tumor necrosis factor in yeast is Yrp7, (Stinchcomb et al., 1979, "Nature" 282: 39; Kingsman et al., 1979, "Gene" 7 141; Tschemper et al., 1980, "Gene" , 10: 157). This plasmid already contains the trpl gene which provides a selection marker for a mutant type of yeast that does not have the ability to grow in tryptophan, for example, ATCC No. 44076 or PEP4-1 (Jones, 1977, "Genetics", 85: 12). The presence of rtpl damage in the genome of a yeast cell line then provides an efficient environment for detection of transformation by growth in the absence of tryptophan.

Suitable sequences for promotion in yeast vectors include metallothionein promoters. 3-phosphoglycetate kinase (Hitzeman et al., 1980, "J. Biol. Chem.", 255; 2073) or other glycolytic enzymes (Hess et al., 1968, "J. Adv. Enzyme Reg.", 7: 149 ; and Holland et al., 1978, "Biochemistry", 17: 4900), such as enolase, glyceraldehyde-3-phosphate dehydrogenase, hexokinase, pyruvate decarboxylase, phosphofructokinase, glucose-6-phosphate isomerase, 3-phosphoglycerate mutase, pyruvate kinase, triphosphate isomerase, phosphoglucose isomerase and glucokinase. Suitable vectors and promoters for use in yeast expression are described further in R. Hitzeman et al., EPO Publn. But. 73,657.

Other promoters that have the additional advantage of transcription being controlled under growth conditions are the promoter regions for alcohol dehydrogenase 2, isocytochrome C, acid phosphatase, degradative enzymes involved in nitrogen metabolism, and the previously mentioned metallothionein and glyceraldehyde-3-phosphate dehydrogenase, as well as enzymes which are responsible for the utilization of maltose and galactose. When constructing suitable expression plasmids, termination sequences flanking these genes are also ligated into the expression vector 3' of the tumor necrosis factor coding sequence to ensure mRNA polyadenylation and termination.

In addition to microorganisms, cell cultures derived from multicellular organisms can also be used as hosts.

However, this is not desirable because great results have been obtained so far with TNF-expressing microbes. In principle, any higher eukaryotic cell culture can work, whether from a vertebrate or invertebrate culture. However, the interest was greatest in vertebrate cells and propagation of vertebrate cells in culture (tissue culture) became a routine procedure in recent years /Tissue Culture, Academic Press, Kruse and Patterson, editors (1973)/. Examples of useful host cell lines are VERO and HeLa cells, Chinese hamster ovary (CHO) cell lines, and W138, BHK, COS-7, and MDCK cell lines. Expression vectors for such cells typically include (if necessary) an origin of replication, a promoter located upstream of the gene to be expressed, together with a ribosome binding site, an RNA termination site (if intron-containing genomic DNA is used), a site for polyadenylation and a transcription termination sequence.

Transcriptional and translational control sequences in expression vectors to be used in transforming vertebrate cells are often provided from viral sources. For example, commonly used promoters are derived from polyoma, Adenovirus 2 and most preferably from Simian Virus 40 (SV40). The early and late promoters are particularly useful because they are readily obtained from the virus as a fragment thus containing the SV40 viral origin of replication (Fiers et al., 1978, "Nature", 273: 113). Smaller or larger fragments of SV40 can also be used, provided that a sequence of approximately 250 bp extending from the Hind III site to the Bgl I site located at the viral origin of replication is included. Furthermore, it is also possible, and often desirable, to use a human genomic promoter, control and/or signal sequences that normally accompany tumor necrosis factor, provided that such control sequences are compatible with host cell systems.

The origin of replication can be provided either by engineering the vector to include some exogenous origin, such as derived from SV40 or another viral source (eg, polyoma, Adenovirus, VSV, or BPV). If the vector integrates into the host cell chromosome, the latter is often satisfactory.

When selecting a suitable mammalian host cell for transfection with vectors comprising DNA sequences encoding both tumor necrosis factor and dihydrofolate reductase (DHFR), it is appropriate to select the host according to the type of DHFR protein used. If wild-type DHFR protein is used, it is preferable to select a host cell that is deficient in DHFR so that the DHFR coding sequence can be used as a marker for successful transfection in a medium of choice lacking hypoxanthine, glycine and thymidine. A suitable host cell in this case is a Chinese hamster ovary (CHO) cell line deficient in DHFR activity, constructed and propagated as described in Urlaub and Chasin, 1980, "Proc. Natl. Acad. Sci." (USA) 77: 4216.

On the other hand, if DNA encoding a DHFR protein with low binding affinity for methotrexate (MTX) is used as a control sequence, it is not necessary to use DHFR-resistant cells. Because mutant DHFR is resistant to MTX, media containing MTX can be used as a selection agent provided only host cells are sensitive to MTX. Many eukaryotic cells capable of MTX uptake appear to be sensitive to methotrexate. One such cell line is the CHO line, CHO-K1 (ATCC No. CCL 61).

Tumor necrosis factor is initially regenerated from cultures. Transformed non-secreted cells are disrupted by sonication or other acceptable method and debris is separated by centrifugation, while supernatants from secreted cells (such as induced cell lines) are simply separated from the cells by centrifugation. Then one or more of the following stages can be used, or they can be completely replaced by other procedures. The following procedure was used to purify tumor necrosis factor to a level satisfactory for sequencing. This is not necessarily co-extensive with the purification required for the therapeutic product.

As an initial step in purification, tumor necrosis factor is adsorbed onto a hydrophobic substance from an interrupted culture or saturated culture medium. The hydrophobic substance is preferably such a non-gelatinous surface as silicate or polyolefin, although alkyl Sepharose is also suitable. The preferred embodiment is controlled porous glass. An amount of about 1 volume of controlled porous glass is mixed with 50 volumes of supernatant and allowed to adsorb at about 4°C without stirring for a period of about 30 minutes to 2 hours, preferably about 1 hour, under slightly alkaline conditions. The adsorbent should then be washed with a suitable buffer so that incorporated contaminant proteins are separated. The adsorbed tumor necrosis factor is eluted from the hydrophobic substance by changing the solvation properties of the surrounding medium. Elution can be achieved by running a buffered solution at about pH 7 to 8.5, preferably about 8, containing 1M salt and an effective amount of an aqueous solution of a water-miscible organic phenol, such as, for example, ethylene glycol or glycerin, usually ethylene glycol in the interval 10-30 percent v/v, preferably about 20 percent v/v. Of course, the optimal conditions will depend on the polyol used. Elution fractions containing tumor necrosis factor are detected by an in vitro assay as described below or another suitable assay. Purification and yield at this stage from monocyte cell culture, as well as from later stages, are shown below in Table I.

Further purification is achieved by adsorption of tumor necrosis factor on tertiary or quaternary amino anion-exchange resin.

Preferred resins for this purpose are resins with a hydrophilic matrix such as cross-linked polystyrene, dextran or cellulose substituted with alkyl tertiary or quaternary amino groups. Commercial products of this type are available as DEAE cellulose, QAE Sefadex or under the trade name Mono Q (in each case where the ethyl alkyl substituent is in each of these products). Best results are obtained with the rapid protein liquid chromatography system described by J. Richey, October 1982, "American Laboratory" using macroporous substantially uniform particles from Ugelstad et al., 1983, "Nature" 303: 95-96. This system enables the purification of tumor necrosis factor to a high level.

Purification to substantial homogeneity is achieved only after further separation on SDS PAG electrophoresis or C4-reverse phase high pressure liquid chromatography (HPLC) as described in the Examples below. However, this product is not desirable for therapy because it loses essential activity upon exposure to SDA or an HPLC organic solvent. Protein concentration was determined by the method of M. Bradford, 1976, "Anal. Biochem." 72: 248-254. During the final stages of purification, protein concentration was estimated by amino acid composition and also by amino acid sequence.

Tumor necrosis factor is formulated for administration by mixing tumor necrosis factor having the desired degree of purity with physiologically acceptable carriers, i.e. carriers that are non-toxic to recipients at the doses and concentrations used. Typically, this is accomplished by combining tumor necrosis factor with buffers, antioxidants such as ascorbic acid, low molecular weight polypeptides (less than about 10 residues), proteins, amino acids, carbohydrates including glucose and dextrins, chelating agents such as EDTA, and other stabilizers and ingredients. The carrier should be formulated to stabilize tumor necrosis factor as a dimer and/or preferably as a trimer. This should be followed by avoiding salts or detergents in concentrations that dissociate tumor necrosis factor into monomers. Alternatively, conditions that aggregate tumor necrosis factor into multiple multimers should be avoided. A nonionic surfactant such as Tween 20 is generally used to show excessive aggregation during purification as well as for allylation or storage in water. Tumor necrosis factor to be used for therapeutic administration must be sterile. This is thus achieved by filtration through sterile filtration membranes. Tumor necrosis factor will usually be stored in lyophilized form.

Tumor necrosis factor is optionally combined with other antineoplastic agents such as chemotherapeutic antibiotics (actinomycin-D, adriamycin, aclacinomycin A), or with agents to enhance or stimulate the immune response, for example, immunoglobulins such as gamma-globulin, including immunoglobulins that have an affinity for cell surface antigens of neoplasms. Further, since interferons act synergistically with tumor necrosis factor in cell disruption assays, alpha, beta, or gamma interferon is preferably combined with preparations of tumor necrosis factor or preparations containing tumor necrosis factor and lymphotoxin. A typical formulation comprises tumor necrosis factor and gamma interferon in a unit activity ratio of about 0.1:1 to 200:1, usually 10:1, and may contain lyrophotoxin in place of the tumor necrosis factor portion. These relationships are of course subject to modification as experience in therapy requires.

Tumor necrosis factor preparations are given to animals that have tumors. Administration is by known methods, eg, intravenous, intraperitoneal, intramuscular, intralesional infusion ig and injection of sterile tumor necrosis factor solutions, or timed release systems as indicated below. Tumor necrosis factor is administered intralesionally, eg, by direct injection into solid tumors. In the case of widespread tumors such as leukemia, administration is preferred intravenously and the lymphatic system. Tumors of the abdominal organs, such as ovarian cancer, are conveniently treated by intraperitoneal infusion using peritoneal dialysis and solutions that are peritoneally compatible. However, tumor necrosis factor is usually given by continuous infusion, although a bolus injection is acceptable.

The tumor necrosis factor is preferably administered from a time-release implanted article. Examples of suitable systems for proteins having the molecular weight of tumor necrosis factor dimers or trimers include copolymers of L-glutamic acid and gamma ethyl-L-glutamate (U.Sidam et al., 1983, "Biopolymers" 22 (1): 547-556) , poly (2-hydroxyethyl methacrylate) (R. Langer et al., 1981, "J. Biomed. Mater. Res." 15: 167-277 and R. Langer, 1982, "Chem. Tech." 12: 98 -105) or ethylene vinyl acetate (R. Langer et al., Id.). This item is implanted at surgical sites where tumors have been removed. Alternatively, tumor necrosis factor is encapsulated in semipermeable microcapsules or liposomes for injection into the tumor. This method of administration is particularly useful for tumors that cannot be surgically removed, for example, brain tumors.

TABLE I

PURIFICATION OF HUMAN TUMOR NECROSIS FACTOR FROM HL-60 CELL CULTURE MEDIUM

[image] * corrected for partial destruction of tumor necrosis factor activity induced with SDS or with TFA and propanol.

The amount of tumor necrosis factor administered will depend, for example, on the method of administration, the tumor in question and the condition of the patient. Intralesional injections will require less tumor necrosis factor on a body weight basis than intravenous infusion, although some tumor types, eg, solid tumors appear to be more resistant to tumor necrosis factor than others, eg, leukemic ones. Therefore, it will be necessary for the therapist to titrate the dose and modify the route of administration as necessary to achieve optimal cytotoxic activity against the attacked tumor, as can be determined, for example, by tumor biopsy or diagnostic tests for cancer markers such as carcinoembryogenic antigen , considering any recombinant toxicity found at the elevated dose. Typically, doses of tumor necrosis factor in mice up to 120 micrograms/kg body weight/day by intravenous administration have been found to be substantially non-toxic and efficacious in vivo.

Tumor necrosis factor is not believed to be type-specific in its cytotoxic activity, so tumor necrosis factor other than human tumor necrosis factor, eg, derived from ox or pig, can be used in the treatment of human tumors. However, it is preferable to use tumor necrosis factor from the species being treated to avoid the potential generation of autoantibodies.

In order to simplify the tumor, certain procedures that often appear will be referred to by abbreviated terms.

Plasmids are denoted by a small p preceding and/or following capital letters and/or numbers. The starting plasmids of the invention are commercially available, are publicly available without any limitation, or can be constructed from such available plasmids according to published procedures. Further, other equivalent plasmids are known in the art as will be apparent to those skilled in the art.

"Digesting" DNA refers to the catalytic breaking of DNA with an enzyme that acts only on certain locations in the DNA. Such enzymes are called restriction enzymes, and the sites for which each is specific are called restriction sites. "Partial" digestion refers to incomplete digestion with a restriction enzyme, i.e., such conditions are chosen that lead to some but not complete termination at the sites for a given restriction endonuclease in the DNA substrate. The various restriction enzymes used herein are commercially available and their reaction conditions, cofactors and other requirements are used as specified by the enzyme supplier. Restriction enzymes are usually designated by abbreviations consisting of a capital letter followed by other letters and then, generally, a number representing the microorganism from which each restriction enzyme was originally obtained. Basically, about 1 µg of plasmid or DNA fragment with about 1 enzyme unit in about 20 µl of buffer solution is used. Appropriate buffers and amounts of substrate for certain restriction enzymes are specified by the manufacturer. Incubation times of about 1 hour at 37°C are usually used, but may vary according to the supplier's instructions. After incubation, the protein is separated by extraction with phenol and chloroform, and the digested nucleic acid is regenerated from the aqueous fraction by precipitation with ethanol. Restriction enzyme digestion is rarely followed by hydrolysis with bacterial alkaline phosphatase of the terminal 5' phosphates to prevent the two restriction-broken ends of the DNA fragment from "circularizing" or from forming a closed loop that will slow the insertion of another DNA fragment at the restriction site. Unless otherwise indicated, plasmid digestion was not followed by 5' terminal dephosphorylation. The procedures and reagents for dephosphorylation are conventional (T. Maniatis et al., 1982, Molecular Cloning pp. 133-134).

"Regenerating" or "isolating" a given DNA fragment from a restriction digest means separating the digest by polyacrylamide gel electrophoresis, identifying the fragment of interest by comparing its mobility with that of marker DNA fragments of known molecular weight, and separating the gel from the DNA. This procedure is mostly known. For example, see R. Lawn et al., 1981, "Nucleic Acids Res." 9: 6103-6114, and D. Goeddel et al., 1980, "Nucleic Acids Res." 8: 4057.

"Southern analysis" is a procedure by which the presence of DNA sequences in the digest of a preparation containing DNA is confirmed by hybridization to a known, labeled oligonucleotide or DNA fragment. For purposes herein, unless otherwise specified, Southern analysis shall mean separation of the digest on 1 percent agarose, denaturation and transfer to nitrocellulose by the procedure of E. Southern, 1975, "J. Mol. Biol." 98: 503-517, and hybridization. as described in T. Maniatis et al., 1978, "Cell" 15: 687-701.

"Transformation" means the introduction of DNA into an organism for which the DNA can be replicated, either as an extrachromosomal element or as a chromosomal integrant. Unless otherwise indicated, the procedure used herein to transform E. coli is the CaCl 2 procedure of Mandel et al., 1970, "J. Mol. Biol." 53:154.

"Ligation" refers to a process for forming phosphodiester bonds between double-stranded nucleic acid fragments (T. Maniatis et al., Id., p. 146). Unless otherwise indicated, ligation can be accomplished using known buffers and conditions with 10 units of T4 DNA ligase (ligase) per 0.5 µg of approximately equimolar amounts of DNA fragments to be ligated.

"Making" DNA from transformants means isolating plasmid DNA from a microbial culture. Unless otherwise provided, the alkaline/SDS procedure of Maniatis et al., Id., may be used. p. 90.

"Oligonucleotides" are polydeoxynucleotides of short length with one or two strands that are chemically synthesized by known methods and then purified on polyacrylamide gels.

All literature citations are expressed by reference.

Example 1

Tests

Specific tumor necrosis factor activity was determined by a previously described modified cell disruption assay (B. Spofford, 1974, "J. Immun." 112:2111). Murine fibroblast cells L-929 (ATC CCL-929) were cultured in flat-bottomed 96-well plates (3040, Falcon Plasics, Oxnard, CA) at 30,000 cells (vol. 0.1 ml) per well in the presence of 1 µg/ml actinomycin. D and serially diluted tested sample (0.125 ml.). Cells are incubated in a humidified atmosphere at 37°C with 5 percent CO2. The tested sample is separated after 18 hours, the plates are washed and cell disruption is detected by staining the plates with a 0.5 percent solution of crystal violet in methanol: water (1:4) (v/v). The end point on the microtiter plates is determined using a Microelis autoreader (Dynatech) set for absorbance at 450 nm and transmission at 570 nm. Cells exposed to culture medium alone were set to 0 percent disruption and those exposed to 3M guanidine hydrochloride solution provided a 100 percent disruption endpoint. One unit of tumor necrosis factor is defined as the amount of tumor necrosis factor (when tested in a 0.125 ml volume) required for 50 percent cell disruption.

Tumor necrosis factor was also tested in an in vivo tumor necrosis assay. Briefly, this test is performed by culturing Meth A Sarcoma cells (5 x 105 cells) in CB6F1 female mice (BALB/c x C57BL/6)Fl for 7-10 days and then intratumor injection with a sample of tumor necrosis factor. After 24 hours, mice are sacrificed by cervical dislocation, tumors are dissociated and necrosis is assessed histologically as previously described in E. Carswell et al., 1975, "Proc. Nat. Acad. Sci." 72: 3666-3670.

Example 2

Use of PBLs or monocytic cell lines for the synthesis of tumor necrosis factor

An inoculated culture of the HL-60 human promyelocytic cell line having a cell density of 1 x 10 5 cells/ml is cultured in 2 liter (890 cm 2 ) roller bottles using 500 ml of RPMI 1640 medium (Irvine Scientific, Santa Ana, CA) containing 10 mM HEPES, 0.05 mM beta-mercaptoethanol, 100 units/ml penicillin, 100 µg/ml streptomycin and 10 percent fetal calf serum. After three days at 37°C when the culture has reached a cell density of 8-12 x 105 cells/ml, the cells are harvested by centrifugation at 1000 g for 10 min., washed twice with serum-free RPMI 1640 medium and transferred to the same medium as described above without serum at a cell density of 15-20 x 105 cells/ml. Cells are cultured in 2-liter roller bottles in the presence of 10 ng/ml PMA. After 16-24 hours, the cells are separated by filtration through a 3 µm Sealleen filter (Pall Trinity Micro Corp. Cortland, NY). The clear filtrate is tested for tumor necrosis factor activity and used for subsequent purification and characterization. This procedure produces about 400 units of tumor necrosis factor units/ml of saturated culture medium.

Human peripheral blood monocytes have also been used to produce tumor necrosis factor. Platepheresis debris was obtained from the American Red Cross, Boston, MA and was used within 24 hours of collection. Initial separation of monocytes from erythrocytes is achieved by centrifugation on Ficoll-Hypaque gradients at 1000 g for 30 minutes. Cells collected at the interface were washed three times with phosphate buffered saline. Monocytes derived from each donor were cultured separately in 2-liter roller bottles in serum-free RPMI 1640 medium at a cell density of 2.5 x 106 cells/ml. 1 µg/ml each of staphylococcal enterotoxin B (SEB) and recombinant thymosin alpha-1 were added to the culture and cells were incubated in a humidified atmosphere at 37°C with 10% CO2. After 24-72 hours, depending on the donor, the cell supernatants are harvested and processed in a similar manner to that for the HL-60 cell line. Tumor necrosis factor yields from PBL cultures vary widely, depending on the induction agents used. Addition of PMA to the induction system described above increases the cytolytic activity of cell supernatants. However, the cell supernatants contained both tumor necrosis factor and lymphotoxin (determination of lymphotoxin or tumor necrosis factor in mixtures of tumor necrosis factor and lymphotoxin was performed by performing a cell disruption assay with test samples preincubated with a rabbit neutralizing antibody for tumor necrosis factor or lymphotoxin and determining the residual activity in the L-929 cell disruption test).

Example 3

Controlled chromatography on porous glass beads

Tumor necrosis factor activity from cell culture was discontinuously absorbed on controlled porous glass beads (Catalog No. CPG 00350, Electro-Nucleonics, Fairfield, NO) equilibrated with 10 mM sodium phosphate buffer, pH 8.0, with constant stirring at 4°C. 100 ml of glass beads were used per 5 liters of substrate. After stirring for 1 hour, the beads are allowed to settle and the supernatant is decanted. The beads are then loaded into a 5 x 50 cm column at room temperature and washed with 10 mM sodium phosphate buffer, pH 8.0, containing 1M NaCl. Tumor necrosis factor activity is eluted from glass beads with 20 percent ethylene glycol in 10 mM sodium phosphate buffer, pH 8.0, containing 1 M NaCl. The elution profile of HL-60 supernatant from the column is shown in Fig. 1.

Example 4

Chromatography on DEAE cellulose

The eluate from Example 3 was applied directly to a column with DEAE cellulose 53 (Whatman) equilibrated with 10 mM sodium phosphate buffer at pH 8.0 and 0.01 percent Tween 20, at a flow rate of approximately 500 ml/hour. After the column flow rate is set to 100 ml/hour, the column is loaded with 4.2 x 10 6 units of tumor necrosis factor in 1,080 ml of sample, the column is washed with equilibration buffer and eluted with stepwise gradients of 75 mM, 150 mM and 500 mM sodium. -chloride in 10 mM phosphate buffer (pH 8.0). The eluate is monitored for absorbance at 280 nm and tumor necrosis factor activity as a function of eluate fractions. The results are shown in Figure 2.

Example 5

Fast liquid chromatography of proteins

The active protein necrosis factor fraction from Example 4 is concentrated and dialyzed against 20 mM Tris HCl, pH 8.0, containing 0.01 percent Tween 20 and 1 mM sodium azide (buffer A) in an Amicon mixing cell using a YM-membrane or other membrane for dialysis for a molecular weight fraction below that of TNF. The membrane is washed twice with buffer A. A column of Sepharose beads with a substituted ammonium group (9.8 µm beads in a 5 x 0.5 cm column, sold as Mono Q resin, Pharmacia) in a fast protein liquid chromatography (FPLC) unit (Pharmacia ) which is equipped with a gradient programmer (GP250) and two pumps (P-500), is pre-equilibrated with dialysis buffers via a supercircuit at a flow rate of 1 ml/min as further described in J. Richey "American Laboratory" October 1982, page 1. The collected washing fluids and dialysis concentrate are loaded onto the column, the column is washed with buffer A and then eluted with a linear gradient of 40-75 mM buffer A. The linear gradients are programmed as follows: 0-5 min equilibration buffer, 5.1-15 min, 25 mM NaCl, 15.1-25 min, 40 mM NaCl, 25-60 min, 40-75 mM NaCl lined gradient, 60-65 min, 75 mM NaCl, 65.1-70 min, 100 mM NaCl, 70-80 min, 100-1000 mM NaCl linear gradients, 80-90 min, 100 mM NaCl, 90.1-110 min, equilibrium pu fair. The effluent is collected in 2 ml fractions and recorded for absorbance at 280 nm, conductivity and necrosis factor activity. The results are shown in Figure 3.

Example 6

Chromatofocusing

Chromatofocusing is performed using a Pharmacia Mono P column (20 x 0.5 cm) in an FPLC system as in Example 5. The biologically active fraction (tumor necrosis factor) eluted in fractions numbers 37 to 45 from Example 5 is concentrated and dialyzed on an Amicon mixed cell with YM -10 membrane opposite the column equilibration buffer, i.e., 0.025 M bis. Tris HCl, pH 6.7. The sample is loaded onto a Mono P column at room temperature via a supercoil at a flow rate of 1 ml/min. The column is washed with equilibration buffer until the absorbance at 280 is brought to baseline and then eluted with a linear pH gradient established by washing the column with pH 7.5 percent polybuffer at pH 4.7 (Pharmacia).

Fractions of 1 ml are pooled and the absorbance at 280 nm and pH of the effluent are recorded. The results are shown in Figure 4. As can be seen from Figure 4, the isoelectric point of tumor necrosis factor is about 5.3.

Example 7

Preparative electrophoresis on SDS-polyacrylamide gel

Fifteen percent polyacrylamide gels (11 x 16 cm) 1.5-3.0 mm thick are made according to a modification of the procedure of U. Laemmli, 1970, "Nature" 227: 680-685. Both digestion and accumulation gels contained 0.1 percent SDS and 0.5 percent Tween 20. Other buffers and concentration of cross-linking reagents were the same as for analytical SDS-PAGE gels. Active tumor necrosis factor fractions from steps of Example 5 or 6 are pooled, concentrated, and dialyzed against 6.25 mM Tris Hcl, pH 7.0, containing 0.005 percent SDS on an Amicon mixing cell using a YM-10 membrane. After separation of the dialyzed concentrate, the membrane is washed three times with a small volume of sample buffer (0.2 percent SDS, 0.02 percent Tween 20, 30 percent glycerol, 0.03 M Tris Hcl, pH 6.8, 0.005 percent staining agent). The dialyzed concentrate and wash fluids are pooled (total volume 1-4 ml), mercaptoethanol is optionally added to establish SDS PAGE reducing conditions, and the sample is loaded into a large well cast in the accumulation gel. Small wells adjacent to the sample well were used to pre-stain the molecular weight marker phosphorylase (94K), bovine serum albumin (67K), ovalbumin (43K), carbonic anhydrase: (30K), soy trypsin inhibitor (20K) and lysosome (14.4 K). The gels were run in a Biorad vertical electrophoresis system cooled to 12°C, with a constant current of 20 mA per mm of gel thickness, until the staining agent reached the bottom of the gel.

After electrophoresis, one of the glass plates is separated from the gel and the positions of molecular weight markers are noted. The strip containing the applied tumor necrosis factor sample is then cut into 0.25 cm sections according to the molecular weights of the marker proteins. These gel pieces are then placed in polypropylene tubes containing 1-2 ml of 10 mM ammonium bicarbonate and 0.01 percent Tween 20, pH 8, and allowed to elute for 16 hours at 4°C. The eluates are then tested for tumor necrosis factor activity and the results are shown in Figure 5. The molecular weight of tumor necrosis factor on the SDS gel was about 17,000, either under reducing or non-reducing conditions, indicating one molecular array.

The protein was regenerated from eluates of gel pieces free of salts and substances of low molecular weight by the following treatment: Small columns were made containing 0.2 ml of Sep-pak C18 resin, which was previously washed with acetonitrile, 1-propanol, 1 percent trifluoroacetic acid (TFA ) and distilled water and then equilibrated with 10 mM ammonium bicarbonate containing 0.01 percent Tween 20, pH 8.0. The gel eluate is loaded onto the column and the effluent is collected. The resin is then washed with approximately 5 ml each of distilled water and 0.1% TFA, so that free amino acids and buffer salts are separated. Tumor necrosis factor is eluted from the resin with 1 ml of 50 percent 1-propanol in 0.1 percent TFA; Further elutions are also performed with 1 ml each of 50% 1-propanol in 0.1% TFA and 99% 1-propanol in 1% TFA, but the protein is usually eluted with the first buffer. In this phase, approximately 80 percent of tumor necrosis factor bioactivity is inactivated. Although the tumor necrosis factor thus obtained can be used for sequence analysis, it is preferable to use the HPLC run-off described in Example 8 below for this purpose.

Example 8

Liquid chromatography under high pressure

The molecular weight of native intact tumor necrosis factor was determined using gel permeation chromatography under high pressure. The latter was performed at room temperature using a TSK G2000 SW gel HPLC column (Alltech Associates, Deerfield, IL) (7.5 x 60 mm). A 1 ml sample of purified tumor necrosis factor from Example 5 containing approximately 1 µg of protein and 15,600 units of activity is isocratically eluted from the gel column at a flow rate of 0.5 ml/min, with 0.2 M sodium phosphate buffer, pH 7.0.

The column is calibrated with bovine serum albumin (MT 66,000), ovalbumin (MT 45,000), bovine uric acid B (MT 29,000) and lysosome (MT 14,300). Fractions of one ml are obtained and tested for tumor necrosis factor activity. Fractions showing tumor necrosis factor activity consistent with a molecular weight of 45,000 ± 6,000 were eluted (Figure 6).

Example 9

HPLC with reverse phase

Tumor necrosis factor was also purified by reverse-phase HPLC using Synchropak columns on Warer's Associates. Inch. chromatographic system as described previously (W. Kohr et al., 1982, Anal. Biochem, 122: 348-3~9). Protein peaks were detected at 280 nm after elution with a linear gradient of 1 to 23 percent v/v 1-propanol in 0.1 percent TFA for the first 15 minutes and 23-30 percent v/v 1-propanol in 0.1 percent TFA for the next 15 minutes at flow rate of 1 ml per minute. Peaks are tested for cytolytic activity. The organic solvents used to elute tumor necrosis factor from the C4 column reduced tumor necrosis factor activity by about 80 percent. Tumor necrosis factor purified by this procedure was vacuum dried, and then processed for amino acid analysis and sequencing.

The results are recorded in Fig. 7. Fig. 7 shows that the tumor necrosis factor obtained in the effluent from Example 5 contained biologically inactive protein contaminants by elution with retention times of about 16 and 19 minutes. The bioactive yield from C4-RPHPLC was largely homogeneous, based on criteria for the amino-terminal sequence.

Example 10

Determination of partial amino acid sequence for tumor necrosis factor

Tumor necrosis factor was digested with trypsin as follows: Homogeneous tumor necrosis factor from Example 9 was dissolved, dried and redissolved in 100 mM ammonium bicarbonate buffer containing 5 percent w/w TPCK trypsin (Worthington Biochemicals) 1 mM CaCl2 and 0.01 percentage of Tween 20 at a ratio of enzyme to substrate of 1:20, incubated for 6 hours at 37°C. The reaction mixture was applied to C4 HPLC as described above in order to separate the peptide fragments. The results are shown in Fig. 8. A total of 9 fragments are observed (fragments 2 and 2' elute together at the peak labeled T2 in Fig. 8). It is believed that the additional tenth column fragment does not retain. The amino acid sequences for the intact tumor necrosis factor of Examples 8 and 9 and the trypsin hydrolysis fragments obtained in this Example were determined by automated sequential Edman degradation using a modified Beckman model 890B sequencer equipped with cold switches. Polybrene (1.25 mg) is used as a carrier in the cup. Based on the amino acid composition of the intact molecule, the molecular weight of intact tumor necrosis factor is 17,000. This figure is consistent with the SDS-PAGE data and confirms the absence of glycosylation.

Example 11

Synergistic effect of tumor necrosis factor and gamma-interferon

Murine melanoma cells B16, (Mason Research, Worcester, MA.) cell line C57B1/6 origin, are seeded in a microtitre plate at 5,000 cells/well and incubated for 4 hours at 37°C in a 5% CO2-humidified incubator before the addition of lymphocytes . The tumor necrosis factor obtained from Example 1 was purified to substantial homogeneity by HPLC and quantified by its activity in the L-929 cell cytolysis assay described above. Similarly, purified recombinant mouse gamma interferon (P. Gray et al., 1983, "Proc. Natl. Acad, Sci. U.S.A." 80: 5842-5846) was tested for its antiviral activity against EMC-infected L cells (D Goeddel et al., 1980, "Nature" (London) 287: 411-416). Murine gamma interferon and human tumor necrosis factor are separately diluted to the dilutions shown in FIG. 10. First, gamma-interferon is added to the indicated wells, and immediately after that, diluted tumor necrosis factor is added, to a final volume of 0.2 ml/well. At the end of the 72-hour incubation, the cells are stained with 0.5 percent crystal violet in 20 percent methanol. The results are shown in Fig. 10. B16 is relatively resistant to either tumor necrosis factor or IFN-gamma alone, no detectable cytolysis was observed at 1,000 units/ml of tumor necrosis factor. However, the addition of very small amounts of gamma interferon (as little as 5 units/ml) led to cytolysis.

Example 12

Isolation of messenger RNA

Total RNA from HL-60 cell cultures (4 hours after PMA induction) or peripheral blood monocytes was cultured as described in Example 2 and extracted in gavnin as described in Ward et al., 1972, "J. Virol. " 9: 61. Cells were pelleted by centrifugation and then resuspended in 10 mM NaCl, 10 mM Tris-HCl pH 7.5, 1.5 mM MgCl2. Cells are disrupted by addition of NP-40 (1% final concentration), and the nucleus is pelleted by centrifugation. The supernatant contained total RNA, which was further purified by multiple extractions with phenol and chloroform. The aqueous phase is made to be 0.2 M in NaCl and then the total RNA is precipitated by adding two volumes of ethanol. A typical yield from 1 gram of cultured cells was about 6 milligrams of total RNA. Polyadenylated mRNA (about 100 pg) was obtained on oligo(dT)cellulose by the procedure of H. Aviv et al., 1972, "Proc. Natl. Acad. Sci. U.S.A." 69: 1408-1412.

Example 13

cDNA library

7.5 µg of poly(A)+ mRNA from Example 12 is translated into double-stranded cDNA by successive action of reverse transcriptase, Klenow fragment DNA polymerase and S1 nuclease (P. Gray et al., 1982, "Nature" 295: 503-508, M (Wickers et al., 1978, J. Biol. Chem. 253: 2483-2495). About 80 ng of cDNA that is more than 600 bp (bp=base pairs) in length was isolated from polyacrylamide gel.

Synthetic DNA adapter sequence 5'AATTCATGCGTTCTTACAG 3'

3'GTACGCAAGAATGTC 5'

was subjected to cDNA ligation to create EcoRI. terminus. As is conventional in the art, the adapter is chemically synthesized as two separate strands, the 5' end of one strand being phosphorylated with a polynucleotide kinase and the two strands annealed. The cDNA (20 ng) is then reisolated from a polyacrylamide gel, inserted by ligation into ECORI-digested lambdagt-10, packaged into phage fragments and propagated in E. coli type C600 hfl (Huynh et al., 1984, Practical Approaches in Biochemistry, IRL Press Ltd., Oxford England) or other known type suitable for lambda phage propagation. A cDNA library of about 200,000 independent clones is obtained.

Example 14

Creation of a deoxyoligonucleotide probe for tumor necrosis factor cDNA

A DNA hybridization probe of 42 nucleotides, based on the preliminary amino acid sequence of the tryptic peptide of tumor necrosis factor TD-6 (E-T-P-E-G-A-E-A-K-P-W-Y-E-K-) was constructed based on published codon usage frequencies (R. Grantham et al., 1981 "Nucleic Acids Res." 9: 43-74), and codon bias of human IFN-gamma (P. Gray et al., 1982, "Nature" 295: 503-508), and human lymphotoxin. The preliminary sequence was wrong (final K should be P). Regardless, this sequence led to a successful probe. The synthetic probe had the sequence 5' dGAAA CCCC TGAAGGGGGCTGAAGCCAAGCCAAGCCCTGGTATGAAA-AG 3' and was synthesized by the procedure of R. Crea et al., 1980, "Nucleic Acids Res." 8: 2331-2348- The probe was phosphorylated with (gamma-32P) ATP and T4 polynucleotide kinase as described previously (Goeddel et al., 1979, "Nature" 281: 544).

Example 15

Identification of a cDNA clone containing tumor necrosis factor coding sequences

About 200,000 recombinant phages from the lambdagt 10 cDNA library are screened by DNA hybridization using the 32 P-labeled 42-mer of Example 14 under high stringency conditions according to A. Ullrich et al., 1984, "EMBO J." 3: 361-364 (or alternatively P. Gray et al., 1983, "Proc. Nat. Acad. Sci. U.S.A." 80: 5842-5846, S. Anderson et al., 1983, "Proc Natl. Acad. Sci . U.S.A." 80: 6836-6842 and M. Jaye et al., 1983, "nucleic Acids Res." 11: 2325-2335). Nine distinct clones hybridize with the probe and the plaque is purified. 32P-labeled cDNA is then made using mRNA from uninduced HL-60 cells. DNA from seven of these nine phage clones did not hybridize with this "uninduced" probe and were therefore judged to be candidate tumor necrosis factor cDNA sequences. The cDNA clone containing the largest insert was designated lambda42-4. This insert is sequenced using the dideoxy strand termination procedure (A. Smith, 1980, "Methods in Enzymology" 65: 560580 and F. Sanger et al., 1977, "Proc. Natl. Acad. Sci. U.S.A." 74:5463-5467 ) after subcloning into the Ml3mp8 vector (J. Messing et al., 1981, "Nusleic Acids Res." 9: 309-321).

The cDNA sequence obtained for lambda42-4 contained the entire coding region for mature tumor necrosis factor plus part of its signal peptide. The correct orientation and reading frame of the DNA was deduced by comparison with the amino acid sequence of the T4 tryptic peptide of tumor necrosis factor. The amino terminal valine residue of tumor necrosis factor was determined by protein sequencing as indicated for amino acid 1, and was followed by 156 additional amino acids before encountering a stop codon in the read-through. The calculated molecular weight is 17,356 daltons.

Example 16

Identification of a cDNA clone containing the complete coding sequences for PreTNF

cDNA clone lambda42-4 contains the entire region coding for mature TNF but lacks the complete sequence coding for the signal peptide as demonstrated by its lack of an initiation codon. To obtain the missing information in the sequence, the hexadecanucleotide example dTGGATGTTCGTCCTCC (complementary to nucleotides 855 to 870, Fig. 10) was synthesized chemically. This example is tested for the mRNA of Example 12 and then cDNA is synthesized using the procedure of Example 12 and then cDNA is synthesized using the cDNA clone in lambdagt10 according to the procedure described in Example 13. This library is tested by hybridization analysis as a lambda42-4 cDNA probe. insert that is 32P-marked. 16 positive clones were obtained, the longest of which (lambda16-4) contains a cDNA insert that extends 337 bp further 5' from the lambda42-4 insert. The complex sequence of TNF cDNA inserts lambdai6-4 (nucleotides 1-870) and lambda42-4 (nucleotides 337-1648) is shown in Figure 10.

Example 17

Construction of an expression vector for direct expression of tumor necrosis factor

The procedure used to express the tumor necrosis factor sDBA sequence obtained in Example 15 is shown in FIG. 11. Phage lambda42-4 from Example 15, which contains the entire coding sequence for mature tumor necrosis factor and part of the putative tumor necrosis factor secretory leader sequence is digested with EcoRI and a fragment of approximately 800 bp containing the tumor necrosis factor coding region is regenerated . This fragment is digested with Ava I and Hind III and a 578 bp fragment is regenerated (marked "C" in Fig. 11). This fragment codes for amino acids 8-157 of tumor necrosis factor.

Two synthetic deoxyoligonucleotides (labeled fragment "B" in Fig. 11) were made (see adapter sequence construction in Example 13) incorporating an Xba I cohesive terminus at the 5' end, an Ava I cohesive terminus at the 3' end, a met initiation codon and codons for the first seven amino-terminal amino acids for tumor necrosis factor. The codons for these amino acids were chosen based on preference for E. coli. The AATT sequence upstream of the start codon is selected for the proper distance of the start codon from the trp ribosome binding sequence and, in combination with the amino acid codons, to elyroinate the potential messenger RNA circuit.

Segroents B and C are then joined in triple ligation with the pBR322 derivative containing the trp promoter sequence with the Shine-Dalgarno sequence of the trp leader peptide (European Patent Application Publn. No. 36776). A derivative was obtained or engineered to contain unique Xba I and Hind III sites between the trp promoter and the TetR gene. Either pLTtrpl (Gray et al., 1984, "Nature" 312: 721-724) or prepETA (Gray et al., 1984, "Proc. Natl. Acad. Sci. U.S.A." 81: 2645-2649) are suitable starting vectors. of this species, although others can be constructed from pBR322, the trp promoter and any required synthetic linker sequences. Both pBR322 and plasmids containing the trp promoter are publicly available. The pBR322 portion of the selected vector may have the AvaI-PvuII segment from bp 1424 to 2065 omitted (indicated by "XAP" in the plasmid name). Either of the preceding plasmids is simultaneously digested with Xba I and Hind III and a large fragment of the vector is regenerated . This fragment, and fragments B and C are ligated with T4 DNA ligase and the ligation mixture is used to transform E. coli 294 (ATCC 31446). Colonies that are resistant to ampicillin are selected, plasmid DNA is regenerated and characterized by restriction mapping endonuclease and DNA sequencing, resulting in pTrpXAPTNF containing inserts B and C.

Example 18

Expression of tumor necrosis factor in E. coli

E. coli ATCC 31446 transformed with pTNFtrp is cultured in M9 medium containing 20 µg/ml ampicillin and the culture is cultured to A550=0.3. Indoleacetic acid is added to obtain a final concentration of 20 µg/ml and the culture grows to A550=1.10 ml cells are concentrated and resuspended in phosphate buffered saline. Cells are sonicated and diluted for determination of tumor necrosis factor as in the assay of Example 1. An activity of approximately 105 units per ml of culture is obtained. This activity is neutralized by preincubation with rabbit antisera from rabbits immunized against human tumor necrosis factor.

Example 19

Expression of tumor necrosis factor in E. coli

This procedure is preferable to that of Example 18.

The preferred host for use with the above vectors is the irreversible tonA E. coli type. such types are resistant to bacteriophages and are therefore far more suitable for large-scale cultivation than wild-type types. The following is a description of a convenient procedure for generating such a type. E. coli W3110 is transduced with lambda:: Tn10, a lambda bacteriophage containing a Tn10 transposable element, to generate the Tn10 hop circuit of E. coli W3110. (N. Klecker et al., 1977 "J. Mol. Biol." 116: 125).

E. coli W3110:: Tn10 hop culture is cultivated in L broth at 37°C to a cell density of about 1 x 109/ml. 0.5 ml of the culture is centrifuged and the pellet is resuspended in 0.2 ml of lambdaphi80 (or T1) lysate containing 7.0 x 109 pfu. Allow the phage to absorb for 30 minutes at 37°C. The suspension is then sprayed onto EMB plates supplemented with tetracycline (15 µg/ml). After overnight incubation at 37°C, faint pink colonies are collected in 3 ml L broth, cultured overnight at 37°C, washed twice and resuspended in L broth. This culture is infected with bacteriophage. P1 kc. and the phage lysate is regenerated (J. Miller, 1972, Experiments in Molecular Biology, Cold Spring Harbor Laboratory, p 304).

E. coli AT982 (no. 4546, E. coli Genetic Stock Center, New Haven, Conn.) is transduced for tetracycline resistance using this P1 kc. lysate. Transductants are selected on plates of L broth supplemented with tetracycline (15 µg/ml) and (40 µg/ml) dap (diaminopimelic acid). The resulting transductants are tested for resistance to tetracycline and regeneration of the dap gene gene (dap+, tet R). dap+, tetR transductants are treated for tetracycline resistance and regeneration of the dap gene gene (dap+, tetR). dap+, tetR transductants are then tested for lambdahi80 (or Tl) resistance.

PK kc licks are then made on several dap+, tetR, lambdaphi80 (or T1) resistant types. Lysates are used to transduce E. coli W3110 to tetracycline resistance. Transductants are screened and selected for resistance to lambdaphi80 (or T1).

Tetracycline-susceptible isolates are selected from W3110 fhuA:Tn10-lambdaphi80R transductants (S. Naloy et al. 1981 "J.Bact." 145: 1110) These isolates are screened for phage lambdaphi80 resistance and tetracycline sensitivity after one colony purification.

DNA is isolated from several tetracycline-sensitive and -resistant lambdaphi80 phage mutants and digested with SstII. SstII-digested DNA is characterized by the Southern procedure using radiolabeled and SstII-digested lambda:Tn10 DNA as a probe to determine whether Tn10 is cleaved (R. Davis et al., 1980, Advanced Bacterial: Genetics, Cold Spring Harbor Laboratories). One of the tetracycline-susceptible isolates was shown to have lost two of the Tn10 hybridization bands when compared to hybridization between DNA from lambda:Tn10 and parenteral W3110 fhuA:Tn10lambdaphi80R. The third hybridization band had altered mobility indicating omission caused by imprecise cleavage of Tn10.

SDS-gel electrophoresis of outer membrane preparations from the missense type Tn10 reveals that the band predicted to be the fhuA protein has altered electrophoretic mobility compared to the wild-type fhuA protein. The resulting protein is non-functional as the lambdaphi80 receptor phage protein. Another independent strain also subjected to imprecise Tn10 excision shows no fhuA protein on the SDA gel.

None of these types demonstrated reversion to tetracycline resistance or lambdaphi80 susceptibility indicating that there is imprecise excision of all or part of the Tn10 transposon together with either partial or complete deletion of the fhuA gene. Preferably, one of such W3110 strains (NL106) is used as a host for TNF-encoding carriers as described elsewhere herein.

NL106 is transformed with ptrpXAPTNF and inoculated into 10 liters of ph 7.4 medium that has the following formula:

Component g/l

(NH4)2SO4 5.0

K2HPO4 6.0

NaH2PO4 3.0

On Citrate 1.0

L-Tryptophan 0.2

NZ Amin AS 4.0

Yeast extract 4.0

MgSO4 1.2

Glucose 25.0

Trace element solution

(Fe, Zn, Co, Mo, Cu, C and Mn ions) 0.5 ml

Tetracycline 1.0 ml

Glucose is added to the culture at a rate of 1 g/minute when the A550 of the culture reaches about 20. Fermentation is carried out at 37°C until the A550 reaches 136 (about 20 hours). The culture is centrifuged to form a cell paste and the paste is then extracted for 30 min. at pH 8.0 and at room temperature with a buffer containing 50 mM Tris, 10 mM EDTA, 1000 mM NaCl, 2000 mM carbimide and 0.1 percent beta-mercaptoethanol. The extract was diluted and tested as described in Example 1. 1x108 units of tumor necrosis factor activity were demonstrated in this assay as equivalent to 1 mg of tumor necrosis factor. Up to about 2 grams of tumor necrosis factor is obtained per liter of culture. Amino-terminal sequencing demonstrated that about 75 to 86% by weight was valyl amino-terminal (mature) tumor necrosis factor, while the remainder was met-TNF. Further, in addition, the protein was not present at a high level of expression in the refractile bodies, nor did it otherwise appear to be toxic to the cells as demonstrated by the extremely high cell densities obtained.

Example 20

Construction and expression of a mutant gene for renal necrosis factor

In this provided example, Examples 17-18 were repeated except that oligonucleotide fragment B was synthesized with the histidine codon CAT instead of the arginine 6 codon CGT. Mutant tumor necrosis factor was expressed.

Example 21

Construction and expression of another tumor necrosis factor mutant gene

In this example, the procedure from Examples 17-18 was repeated with oligonucleotide fragment B coding for leucine (CTT) instead of residues 2 of the arginine codon. About 1200 mg of mature TNF activity per liter of culture was obtained in initial tests. Unprocessed TNF was undetectable in culture.

Example 22

Construction of a vector encoding tumor necrosis factor fusion with a secretory signal sequence

The sequence of the E. coli heat-stable enterotoxin STII gene is described in FIG. 12. In this Example, a fragment containing the STII secretory signal and the Shine-Dalgarno sequence is ligated downstream of the E. coli alkaline phosphatase promoter. The STII signal was followed in the 3' direction with a synthetic oligonucleotide providing codons for the initial seven amino-terminal amino acids of tumor necrosis factor and the rest of the tumor necrosis factor coding sequence. All of the above were collected in the pBR322 vector.

pWM501 (Picken et al., 1983, "Infection and Immunity" 42 (1): 269-275) contains the STII gene described in FIG. 12. pW501 is digested with Xba1 and NsiI and a fragment of approximately 90 bp is isolated. This fragment could also be synthesized by organic methods known per se (fragment A).

The pBR322-Trp plasmid as described in Example 17 (p20kLT is digested with XbaI and HindIII, and a large vector fragment (fragment B) is regenerated. This fragment contains the E. coli origin of replication and the gene conferring the ampicillin resistance phenotype.

The synthetic oligonucleotide is synthesized as two strands and annealed to give the following structure (the sticky ends of the restriction site and the amino acids encoded by the oligonucleotide are also indicated.

VAL ARG SER SER SER ARG THR

5' GTA CGT TCT TCT TCT CGA AGT 3'

ACGT CAT ACG AGA AGA AGA GCA TGA GGCT

NsiI AvaI

This is labeled fragment C.

pTNFtrp from Example 18 was digested with AvaI and HindIII. A 578 bp AvaI-HindIII fragment is regenerated

(fragment D). It contains the entire TNF coding sequence except for the first seven amino acids.

A DNA sequence comprising the E. coli alkaline phosphatase (AP) promoter linked to a heterologous Shine-Dalga~io (S.D.) sequence (trp) and having EcoRI and XbaI termini was constructed as follows. A DNA fragment containing part of the AP promoter was isolated from the pHI-1 plasmid. (H. Inouye et al., 1981, "J.Bacteriol." 146: 668-675), although any suitable sources containing the AP promoter isolated from plasmid pHI-1 may be used. (H. Inouye et al., 1981, "J. Bacteriol." 146 :668-675), although any suitable sources containing AP promoter DNA can be used. pHI-1 is digested with HpaI to open the plasmid.

GAATTCGAATTC

The synthetic linker sequence CTTAAGCTTAAG is ligated to the plasmid and the ligated plasmid is digested with excess EcoRI to disrupt all EcoRI sites and deficient RsaI activity in order to only partially disrupt all RsaI sites (EcoRI and RsaI steps can also be performed sequentially and not simultaneously). A 420 bp fragment containing the AP promoter was regenerated from a partial EcoRI-RsaI digest.

Trp S.D. the sequence is obtained as follows. A plasmid or organism containing the trp promoter (pIFN-beta 2, D. Leung et al., 1984,"Biotechnology" 2:458-464) is digested with XbaI and RsaI and a 30 bp fragment containing trp is regenerated S.D. sequence. This fragment is ligated to the 420 bp AP promoter fragment to give the 450 bp EcoRI-XbaI fragment E. Fragment E has the nucleotide sequence:

EcoRI

GAATTCAACTTCFCCATAGTTTGGATAAGGAAATACAGACATGAAAAATCCATTGCTGAGTTGIIATTT AAGCTTGCCCAAAAAAGAAGAGTCGAAAGAACFGTGTGCGCAGGTAGAAAGCTTTGGATTATCGICA CTGCAATGGTTCGCAATATGGCGCAAAATGACCAACAGCGGTTGATTGATCAGGTAGAGGGGGCGTGTA CGAGGTAAAGCCCGATGCCAGCATTCTGACGACGATAGTGTGCGGGCGATTACGTAAGCCGAACCITA TTGAAGCATCCTCGTAGTAAAAGTAATTTTTCAACAGCTGTCATAAAGTTGICCGGTCGCCCGAGACTT

trpS.D.XbaI

ATAGTCGCCTTTGTTTTTATTITTAATGTATTGTACGAAGTTCACGTAAAAAGGGTATCTAGA

Fragments A, B, C and D are ligated in a four-part ligation and the ligation mixture is used to transform E. coli 294. Transformants are identified by growth on LB plates containing amicillin. Plasmid trpSTIITNF is isolated from a colony of transformants. This plasmid is digested with XbaI and EcoRI to cleave the trp promoter, then ligated to the 450 bp EcoRI-XbaI fragment E containing the E. coli alkaline phosphatase promoter. The resulting plasmid was called pAPSTITTNF.

Example 23

Expression and processing of a tumor necrosis factor fusion with a secretory signal sequence

E. coli NL106 is transfected with pAPSTIITNF and inoculated into 10 liters of pH 7.0 medium which has the following formula:

Component g/l

(NH4)2SO4 5.0

K2HPO4 2.6

NaH2PO4 1.3

On Citrate 1.0

KCl 1.5

NZ Amin AS 5.0

Yeast extract 2.0

MgSO4 1.2

Glucose 25.0

Trace element solution

(Fe, Zn, Co, Mo, Cu, B and Mn ions) 0.5 ml

Tetracycline 20.0

Cultivation is carried out in the same manner as described above in Example 19, except that A550140 is reached. The culture at this point contained about 400 mg of tumor necrosis factor/liter, about 70-80% by weight of this was properly processed into mature protein as estimated from electrophoresis gels. Approximately the same activity of tumor necrosis factor is regenerated after extraction of all cells using the procedure used in Example 19 and is regenerated by osmotic shock of the cells.

Example 24

Construction and expression of complementary derivatives of tumor necrosis factor

Mutant derivatives of the TNF amino acid sequence from FIG. 10 to satisfy one or more of at least the following objectives:

To increase half-life in vivo, to increase cytolytic activity, to increase net differential cytolytic activity on tumor cells versus normal cells, to make TNF immunogens to make anti-TNF antibodies for diagnostic examples, to construct unique sites for covalent modification (for example, when enzyme markers are covalently attached in making EMIT or ELISA diagnostic reagents), and to alter the physical properties of TNF, eg, its solubility, pI and the like. The assessment of the desired activity in the selected derivatives was determined by routine testing in suitable tests in a known manner.

Preferably, TNF and its derivatives will not include TNF having a sequence corresponding to non-primates, also preferably will not have non-primate amino termini, e.g., the desValArg amino terminus which is characteristic of a substance called rabbit tumor necrosis factor, nor the amino terminus of the TNF gene which is identified as having redwood

Val-Arg-Ser-Arg-Thr-Pro-Ser-Asp-Lys-ProVal-Ala-Val-Ser-Val-Ala-Asn-Pro-Aln-Ala-Gly-Gly-(Wang et al., “Science ” 228: 14-154, 1985).

The following procedure, based on J. Adelman et al., "DNA" 2(3): 163-193, is generally applicable to the construction and expression of any mutant TNF DNA sequence containing silent mutations or a mutant coding for TNF. In order to avoid redundancy, representative derivatives are illustrated in which Arg 32 is converted to histidyl (substitution), His 73 is omitted (deletion mutation) and leucyl is fused to Leu 157 (insertion). However, it should be clear that all other mutant species are generated in the same way.

Other methods for creating silent or pronounced mutations in the DNA coding for tumor necrosis factor are known to those skilled in the art. For example, mutant DNA is constructed by simple chemical synthesis of the entire sequence or by synthesis of part of the sequence and the fragment is ligated into the rest of the required DNA. Chemical synthesis of DNA is suitable when the expert wants to make a mutant directly without first obtaining from natural sources the DNA that codes for the tumor necrosis factor. However, usually the DNA will encode the natural sequence of amino acids, including allelic variants thereof, and it will be desirable to make certain mutant derivatives.

It is desirable in making the DNA that codes for the mutant TNF derivatives that no codon changes are made that create an opportunity for the mRNA transcribed from them to form strong stem-and-loop structures. Avoiding such structures will generally lead to higher returns. Furthermore, for the same reason, a codon transformant that is desirable for the host should be used.

A suitable starting DNA is the EcoRI-HindIII fragment of pTrpXAPTNF (Example 17) obtained by sequential digestion with EcoRI and HindIII, and then isolation of the fragment that codes for the TNF gene. This fragment includes the trp promoter and the structural gene for tumor necrosis factor methionyl.

To obtain a single-stranded copy of this gene suitable for mutagenesis, an EcoRI-HindIII fragment was cloned into the polybinding site of phage M13 mp8 RF-DNA (J. Messing et al., 1982, "Gene" 19: 269-276, "RF " means the replicative form of the phage: this phage is commercially available). An aliquot of the EcoRI-HindIII digestion mixture was added to a ligation reaction containing 10 ng of M13 mp8 RF-DNA previously digested with EcoRI and HindIII. After incubation at room temperature for 2 hours, the ligation mixture is used to transform E. coliJM103 (commercially available type, JM101 can also be used). Transformed cells are coated with top agar containing X-GAL (dibromo-dichloro-indolyl-galactoside) and IPTG (isopropylthiogalactoside). Bacterial cultures (1 ml) . infected with phage harvested from colorless plaques are used to isolate M13 mp8/TNF RFDNA using a microassay procedure (Birnoboil and Doly, 1980, "Nuc. Acids Res." 7: 1513-1523). The obtained recombinant phage M13 mp8/T'NF carries the coding strand for the TNF gene.

For site-directed mutagenesis, oligodeoxyribonucleotides (mutagenesis oligomers) are synthesized with a complementary sequence with directions of 15 bases on either side of the mutation site as shown in the following diagrams, where N denotes the complementary bases and M denotes the nucleic acid to be inserted , omit or substitute. Insertions or deletions are usually made in groups of 3 in order to keep the downstream gene in phase.

[image]

M1 denotes a base or oligomer that is not complementary to the base or oligomer M2. Here, M1 is the desired mutant sequence. Typically, oligomers are also made to act to make more than one type of mutation at a time.

The antisense oligomer for Arg-his 32 is a mutant

p CAG GAG ATT GGC ATG GCG GTT CAG CCA HTG-OH.

The antisense oligomer for His 73 is omitted

p GTG GGT GAG CAC GGT GGA GGG GCA GCC-OH.

The antisense oligomer for Leu 158 is an insertion

p TGT TCG TCC TCC TCA AAG CAG GGC AAT GAT CCC-OH.

These examples are synthesized by conventional procedures. For use in the mutagenesis procedure, 10 pmol of oligomer or lac example 5'-GT'ITTCCCAGTCACGAC-3' are each phosphorylated for 30 min at 37°C in 10 µl of 50 mM Tris-HCl (pH 7.5), 0.1 mM EDTA, 10 mM MgCl2 . 10 mM dithiothreitol, 0.1 mM ATP, containing 2 U of T4 polynucleotide kinase. For use as probes (see below) 2 p moles of synthetic polyoligonucleotides are phosphorylated as above except that 0.1 mM ATP is replaced by 1 µM gamma-32P ATP (Amersham). Specific activities are routinely greater than 5 x 106 cpm/pmole of oligonucleotide array.

Hybridization of each oligomer and lac example to single-stranded DNA from phage M13 mp8/TNF, and then extension of the example, leads to the formation of partial heteroduplex DNAs, one strand of which contains the mutant DNA.

To form a partial heteroduplex, single-stranded M13 mp8/TNF DNA (300 ng) was heated at 80°C (2 min), 50°C (5 min) and at room temperature (5 min) in 20 µl 10 mM Tris- HCl (pH 7.5), 0.1 mM EDTA, 50 mM NaCl, containing 1 pmol each of phosphorylated oligomer and example (added as aliquots from the kinase reaction). Example extension begins by adding 30 µl of 50 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 12 mM MgCl2, 10 mM dithiothreitol, 0.7 mM ATP, 0.07 mM dATP, 0.2 mM each of dGTP, dTTP, and dCTP containing 2 U E. coli DNA polymerase I, large fragment and 20 U T4 DNA ligase. After 30 minutes at room temperature, the reaction mixtures are incubated at 4°C. Aliquots are extracted with phenol and the DNA is precipitated with ethanol and dissolved in 15 µl of water. The DNA in these aliquots is used to transform E. coli JM103.

lac example hybridizes to phage at the 5' oligomer location. Example elongation stabilizes the heteroduplex structure. The oligomer and example is enzymatically phosphorylated to allow T4 DNA ligase to join the DNA strands.

Phenol-extracted heteroduplex DNA from aliquot C (10 µl) is added to 10 µl of 0.06 M Na-acetate pH 4.5, 0.6 M NaCl. 0.6 mM ZnCl2 and containing 200 U Sl nuclease. After incubation at 37°C for 5 min. yeast tRNA (5 µg) is added and nucleic acids are regenerated by phenol extraction and precipitation with ethanol. Using the same Sl conditions, 30 ng of single-stranded M13 mp8 DNA (about 10,000 plaque-forming units) produced less than 100 plaques in the DNA-transforming assay, while the same amount of RF-DNA retained more than 80 percent of its transforming properties. Sl-treated DNA is used to transform E. coli JM103 and the resulting phage is analyzed by in situ plaque assay.

Bacterial plates (15 cm diameter) containing several hundred plaques of recombinant M13 phage were seeded by plaque in situ hybridization (Benton et al., 1977, "Science" 196: 180-182) for both parental and mutant genotypes using appropriately labeled oligomers on special filter sets (about 106 cpm per filter). Hybridization is overnight at 50°C, with 40% formamide, 5X SSC. Filters were washed at 45°C, 2 x SSC, 0.02% sodium dodecyl sulfate, air dried, and exposed to X-ray film at -70°C using an intensifying screen.

It will be necessary to vary the stiffness of the hybridization procedure (by changing the concentration of SSC) in order to degrade the oligomer hybridization to the mutant DNA strand (perfect complement) versus the hybridization to the starting DNA, each mutant will vary in its ability to hybridize depending on the nature and number of substituted bases, omitted or inserted bases. For example, single base mutation detection will require more rigorous conditions to discriminate between mutant and nonmutated parental DNA when the mutation is minor, eg, codon omission or 1-3 base substitution, and then the hybridization probe should be smaller than the mutating oligomer. The task of testing for omission in a mutant is facilitated by using probes containing or constituting the omitted sequence to test for sequence loss. This probe will not hybridize to DNA from the selected plaque and it can be concluded that the desired loss of the target sequence has occurred.

The plaque that hybridizes with the labeled oligomer is collected and inoculated into E. coli JM103. Single-stranded DNA (ss) is made from the supernatant and double-stranded (ds) DNA is made from the cell pellet. The ssDNA is used as a footprint for dideoxy sequencing of clones using the M13 universal primer or complementary synthetic oligomer sequences located 3' of the mutated region of tumor necrosis factor DNA. Dideoxy sequencing confirms that the regenerated plaque contains mutant DNA. As such, the phage is labeled M13 mp8/TNF mutant.

M13 mp8/TNFmutant is digested with EcoRI and HindIII and the TNF-coding fragment is regenerated. pTrpXAPTTNF is digested with EcoRi and HindIII and the vector fragment is regenerated. The mutant fragment is then ligated to the vector fragment and the ligation mixture is used to transform E. coli W3110, NL106 or 294 (ATCC 31446). Mutant TNF is regenerated by the procedure of Examples 18 or 19. Supplementary information relating to M13 mutagenesis is provided in U.K. patent application 2,130,219A.

Mutants made according to this procedure are divided into three classes: substitutions, deletions and insertions, further subdivided as shown in the following Table. Unless otherwise indicated, the mutants are from the mature tumor necrosis factor of FIG. 10.

[image] [image]

Of importance are mutations in which the trypsin hydrolysis sites at arg 2, arg 6 (see Examples 20 and 21), arg 32 and arg 131 are omitted or modified so that they are no longer sensitive to trypsin. This may prolong the half-life of tumor necrosis factor in vivo while reducing the possibility of fermentative termination. the arg 2 and arg 6 sites are not critical because biological activity is retained even after the regions in question are separated. However, truncation at arg 32 and 131 sites leads to loss of activity.

Therefore, arg 32 and/or arg 131 are preferably replaced with histidinyl, or, less preferably with gln. This separates or reduces the accessibility of the site for the enzyme but retains the basicity of the side chain. Also, arg 31 is replaced with histidinyl or, less preferably, with glutamine for the same reasons.

Enzyme hydrolysis sites are also inserted between fusion polypeptides and TNF sequences to create sites for the intended release of mature or mutant TNF, or such sites replace residues within the preTNF leader sequence.

Substitution of asn at asp 45 and expression in a host cell (eg, yeast or mammalian cell) capable of glycosylation is expected to produce glycosylated tumor necrosis factor.

Example 25

Expression of tumor necrosis factor in yeast under the control of the ADHP promoter

Plasmids TrpXAPTNF is constructed as described in Example 17 except that p20KLT or ptrpETA (or pBR322) is cut with EcoRI and HindIII rather than XbaI and HindIII and synthetic fragment B is made with an EcoRI cohesive terminus rather than an EcoRI cohesive terminus. The ligation mixture is then used to transform E. coli ATCC 31446 and plasmid pTNFRI is identified by restriction analysis with DNA containing the tumor necrosis factor coding sequence linked to EcoRI sites. Plasmid pTNFRI is isolated, digested with EcoRI and the T1 fragment containing tumor necrosis factor DNA is regenerated.

Plasmid pFRPn (EP 60,057A) is digested with EcoRI, treated with alkaline phosphatase to prevent recirculation, ligated to a fragment of tumor necrosis factor using T4 DNA ligase and the ligation mixture is then used to transform E. coli ATCC 31446. Colonies which resistant to ampicillin yield two sets of plasmids having the T-1 insert in opposite orientations as determined by restriction analysis on agarose electrophoresis gels. Plasmids are purified from E. coli transformants and used to transform yeast harboring a trp1 mutation (for example, yeast strain RH218, unrestricted ATCC Deposit No. 44076) for a trp+ phenotype. The plasmids are oriented so that the start codon of the T-1 segment is located near the alcohol dehydrogenase promoter fragment for transforming yeast to express tumor necrosis factor. Plasmid stability in large-scale fermentations is improved by using an expression plasmid containing a 2 micron origin of replication instead of the pFRPn chromosomal origin of replication (ars 1) and a compatible host type (J. Beggs, 1978, "Nature" 275: 104-109).

Example 26

Expression of tumor necrosis factor in mammalian cells

Plasmid pEHER (EP 117,060A) is digested with EcoRI, treated with intestinal alkaline phosphatase, ligated to fragment T-1 from Example 25 and the ligation mixture used to transform E. coli ATCC 31446. Two plasmids are isolated (designated pEHERTNF I and pEHERTNF II) containing TNF DNA in opposite orientations as determined by restriction analysis on polyacrylamide gels. These plasmids are used for transfection and selection of CHO DHFR-DUX-B11, CHO and Ltk+ cells.

Tissue cell cultures are transfected by mixing 1 µg of pEHERTNF I or pEHERTNF II as made above with 10 µg of carrier rat DNA in a volume of 250 µl, 0.25 M CaCl2, and then adding dropwise 250 µl of HEPES-buffered saline (289 nM NaCl, 1.5 mM Na2PO4, 50 mM HEPES, pH 7.1). After 30 min at room temperature, the solution is added to tissue culture cells growing in 60 mm plastic tissue culture dishes. CHO 1, CHO DHFR-DUX-B11 and Ltk+ cells are used. The dishes contain 3 ml of culture medium corresponding to the host cells.

For CHO 1 and CHO DHFR-DUX-B11 cells, the medium was Ham F-12 medium (Gibco) supplemented with 10 percent calf serum, 100 µg/ml penicillin, 100 µg/ml streptomycin, and 2 µM L-glutamine. For the Ltk+ cell line, the medium was Dulbecco's modified Eagle's medium (DMEM) supplemented as above.

After 3-16 hours, the medium is separated and the cells are washed with 20% glycerol in phosphate buffered saline. Fresh medium is added to each plate and the cells are incubated as before for another two days.

Selection of transfected host cells is performed by trypsinizing the cells after 2 days of growth (which includes treating the cells with ethereal trypsin 0.5 mg/ml containing 0.2 mg/ml EDTA) and adding about 3x105 cells to 10 mm tissue culture plates with selective media.

For the DHFR- medium, there is a formulation (F-12 GIBCO) of the medium that does not contain glycine, hypoxanthine and thymidine (GHT- medium). For DHFR+ host cells, methotrexate (100-1000 nM) is added to normal growth medium. Controls are performed using transfection conditions without plasmid and with plasmid pFD-11 (EP 117,060A) containing normal DHFR. Colonies arising from cells that collect and express the DHFR plasmid are clear in 1-2 weeks. Transformants are identified and those that express tumor necrosis factor.

Claims (33)

1. The procedure for preparing tumor necrosis factor from a mixture with other proteins, indicated by the fact that it includes the adsorption of tumor necrosis factor from such a mixture onto a substance selected from: (i) silicate, (ii) controlled porous glass, (iii) alkyl sepharose, (iv) particles of anion exchange resin of uniform particle size, and then elution of tumor necrosis factor. 2. The method according to claim 1, characterized in that the tumor necrosis factor is eluted from the controlled porous glass with ethylene glycol. 3. The method according to claim 1, characterized in that quaternary amino substituted cross-linked polystyrene is used as anion exchange resin. 4. Human tumor necrosis factor, characterized in that (a) non-glycosylated; (b) has a molecular weight of about 17,000 as determined by SDS-PAGE; (c) capable of preferentially killing or inhibiting the growth of tumor cells compared to normal cells under the same conditions; (d) completely homogeneous as determined by SDS-PAGE; has a specific activity greater than about 10 million units/mg of protein; and has been purified to a degree sufficient for sequencing of intact tumor necrosis factor, or for trypsin hydrolysis of its fragments using sequential Edman digestion on spinal sequencing equipment using cold separators and using Polybrene® as a substrate in the sequencing part; you (e) to include: (i) the sequence of amino acids 1 to 157 shown in Figure 10; or (ii) amino acid sequence Glu Thr Pro Glu Gly Ala Glu Ala Lys Pro Trp Tyr Glu Pro; or (iii) the amino acid sequence of residues 35 to 66 shown in Figure 10; or (iv) the amino acid sequence of residues 110 to 133 shown in Figure 10; or (v) the amino acid sequence of residues 1 to 6, 7 to 44, 45 to 82, 83 to 90 or 99 to 157 shown in Figure 10. 5. Tumor necrosis factor according to claim 4, characterized in that it is free from other cytotoxic proteins. 6. Tumor necrosis factor according to claim 5, characterized by the fact that it is free of proteins selected from lymphotoxicity, α-globulin, serine proteases and interferon. 7. A tumor necrosis factor capable of preferentially destroying or inhibiting the growth of tumor cells compared to normal cells under the same conditions, characterized in that it consists mainly of the sequence of amino acids 1 to 157 shown in Figure 10; possibly with an N-terminal methionyl residue; or from variants (except rabbit TNF) of the mentioned order in which the residue of Val1 and/or Agr2 is retained. 8. Tumor necrosis factor according to claim 7, characterized in that the amino acid residue is substituted, deleted or inserted in the sequence of amino acids 1 to 157 shown in Figure 10. 9. Tumor necrosis factor capable of preferentially destroying or inhibiting the growth of tumor cells compared to normal cells under the same conditions, characterized in that it comprises a variant of the amino acid sequence 1 to 157 shown in Figure 10 (except rabbit TNF) in which the amino acid sequence is substituted, deleted or inserted on the side where amino acid residues 35 to 66, 110 to 133 or 150 to 157 are shown in Figure 10. 10. Tumor necrosis factor capable of preferentially destroying or inhibiting the growth of tumor cells compared to normal cells under the same conditions, characterized in that it comprises a variant of the sequence of amino acids 1 to 157 shown in Figure 10 (except rabbit TNF) in which the sequence of amino acids is substituted, deleted or inserted on the side where amino acid residues 67 to 109 are shown in Figure 10. 11. Tumor necrosis factor according to any claim 8, 9 and 10, characterized in that the substitution of a residue from the class of neutral, acidic or basic amino acid residues is for a residue of the amino acid sequence of the tumor necrosis factor that is not a member of the class to which the substituted amino acid belongs. 12. Tumor necrosis factor according to any one of claims 8, 9 and 10, characterized in that by substituting a residue that has a steric massive side chain, a residue that does not have such a chain is replaced. 13. The tumor necrosis factor according to 12, characterized in that the residue that does not have a sterically bulky side chain is glycine or serine. 14. Tumor necrosis factor according to 12 or 13, characterized in that the residue having such a side chain is phenylalanine or tryptophan. 15. Tumor necrosis factor according to any of claims 8, 9 and 10, characterized in that the lysine or arginine residue is omitted or substituted with an amino acid residue other than lysine or arginine. 16. Tumor necrosis factor according to any claim from 4 to 15, characterized in that it is in a mixture with a physiologically acceptable buffer, amino acid or non-ionic surfactant or their mixture. 17. Tumor necrosis factor according to claim 16, characterized in that it is lyophilized. 18. The DNA encoding a tumor necrosis factor according to any one of claims 7 to 15. 19. DNA according to claim 18, characterized in that the DNA is synthesized. 20. DNA according to claim 18 or 19, characterized in that it is a replicating vector. 21. DNA according to claim 20, characterized in that the vector is a plasmid or a virus. 22. A cell, characterized in that it has been transformed by DNA according to claim 20 or 21. 23. The method, indicated by the fact that it comprises the cultivation of a cell transformed by DNA according to claim 18, so that the secretion of said DNA is achieved, enabling the accumulation of tumor necrosis factor in the culture and its recovery from the culture. 24. A preparation characterized in that it comprises a tumor necrosis factor according to any of claims 4 to 15 and a component from non-human cells that is physiologically acceptable for administration to patients. 25. The preparation according to claim 24, characterized in that said component is a prokaryotic protein. 26. A preparation comprising cells containing a heterologous tumor necrosis factor according to any of claims 7 to 15. 27. The preparation according to claim 26, characterized in that the cells are prokaryotic or lower eukaryotic. 28. A preparation suitable for the treatment of tumors, characterized in that it comprises tumor necrosis factor according to any claim from 4 to 17 or that obtained by the process according to any claim from 1 to 3, and interferon. 29. The preparation according to claim 28, characterized in that it is free of other cytotoxic peptides. 30. The preparation according to claim 28 or 29, characterized in that it is free from other plasma proteins. 31. A cell-free preparation suitable for tumor treatment, characterized in that it comprises a tumor necrosis factor according to any claim from 4 to 17 or one obtained by a process according to any claim from 1 to 3, and interferon, and the preparation is free of lymphotoxin. 32. The preparation according to any one of claims 28 to 31, characterized in that the interferon is gamma-interferon. 33. The preparation according to any of claims 28 to 32, characterized in that it is lyophilized.