IE85054B1 - Monoclonal antibodies directed to the cytotoxic lymphocyte maturation factor - Google Patents

Abstract

ABSTRACT The present invention relates to the field of cytokines, in particular to those cytokines which synergize with interleukin-2 (IL-2) to activate cytotoxic lymphocytes such as the cytokine Cytotoxic Lymphocyte Maturation Factor (CUE-). The present invention also relates to monoclonal antibodies directed to GET.

Description

Monoclonal Antibodies Directed to the Cytotoxic Lymphocyte Maturation Factor F. HOFFMANN-LA ROCHE AG The present invention relates to the field of cytokines, in particular to those cytokines which synergize with interleukin-2 (IL-2) to activate cytotoxic lymphocytes such as the cytokine Cytotoxic Lymphocyte Maturation Factor (CUE-). The present invention also relates to monoclonal antibodies directed to GET.

'Cytokine' is one term for a group of protein cell regulators, variously called lymphokines, monokines, interleukins and interferons, which are produced by a wide variety of cells in the body. These cytokines play an important role in many physiological responses, are involved in the pathophysiology of a range of diseases, and have therapeutic potential. They are a heterogeneous group of proteins having the following characteristics in common.

They are low molecular weight (580 kDa) secreted proteins which are often glycosylated; they are involved in immunity and inflammation where they regulate the amplitude and duration of a response; and are usually produced transiently and locally, acting in a paracrine or autocrine, endocrine manner- rather than Cytokines are extremely potent. generally acting at picomolar concentrations; and interact with high affinity cell surface receptors specific for each cytokine or cytokine group. Their cell surface binding ultimately leads to a change in the pattern of cellular RNA and protein synthesis, and to altered cell behavior. Individual cytokines have multiple overlapping cell regulatory actions. "it is concomitantly exposed.

The response of a cell to a given cytokine is dependent upon the local concentration of the cytokine, upon the cell type it is acting on and upon other cell regulators to which The overlapping regulatory actions of these structurally unrelated proteins which bind to different cell surface receptors is at least partially accounted for by the induction of common proteins which can have common response elements in their DNA. Cytokines interact in a network by: first, inducing each other; second, transmodulating cytokine cell surface receptors and third, by synergistic. additive or antagonistic interactions on cell function. 299 (l989)].

[Immunology Today lg: The potential utility of cytokines in the treatment of neoplasia and as immunoenhancing agents has recently been demonstrated in studies using human recombinant interleukin—2 (rlL-2). Natural interleukin—2 (IL-2) is a lymphokine which is produced and secreted by T—lymphocytes.

This glycoprotein molecule is intimately involved in the induction of virtually all immune responses in which T—cells play a role. B cell responses in vitro are also enhanced by the presence of IL-2. IL-2 has also been implicated as a differentiation inducing factor in the control of B and T lymphocyte responses. in ameliorating chemotherapy—induced immunosuppression the treatment of certain infectious diseases [J.

[Immunol. Lett. _g:3o7—314 (l985)].

"However. the clinical use of rIL—2 has been complicated "by the serious side effects which it may cause [N. Engl. J.

Kobayashi et al. (J. Exp. Med. (1989) 170, 827-845) relates to the identification and purification of natural killer cell stimulatory factor (NKSF), a cytokine with multiple biologic effects on human lymphocytes.

Thus, the present application describes a novel cytokine protein called Cytotoxic Lymphocyte Maturation Factor (CLMF) which is produced and synthesized by cells capable of secreting CLMF. Examples for such cells are mammalian cells particularly human lymphoblastoid cells. In the presence of low concentrations of IL-2 CLMF synergistically induces the cytolytic activity of Lymphokine Activated Killer (LAK) cells. CLMF is also capable of stimulating T-cell growth.

CLMF can be isolated in a substantially pure form by the following steps: a) stimulating B lymphoblastoid cells such as NC—37 cells to produce and secrete cytokines into a supernatant liquid; b) collecting the supernatant liquid produced by the stimulated cells: c) separating the supernatant liquid into protein fractions: d) testing each protein fraction for the presence of CLMF; e) retaining the protein fractions which are able to stimulate T-cell growth. said fractions containing an active protein which is responsible for the T—cell stimulating activity of the protein fractions: f) isolating said active protein into a substantially pure form, said protein being Cytolytic Lymphocyte Maturation Factor (CLMF).

The CLMF protein obtained in this way is free from other cytokine proteins. The natural CLMF protein is a 75 kilodalton (kDa) heterodimer comprised of two polypeptide subunits, a 40 kDa subunit and a 35 kDa subunit which are bonded together via one or more disulfide bonds. The present application also describes the nucleotide sequence of the CLMF gene and the amino acid sequence of the CLMP protein encoded by the said (,|é:J1(-3 .

Furthermore the present invention relates to isolated polyclonal or monoclonal antibodies capable of binding to CLMF- Monoclonal antibodies prepared against a partially purified preparation of CLMF have been identified and 25I—labelled immunodepletion of CLMF bioactivity. 3: characterized by 1: CLMF. 2: blotting of CLMF, 4: its cellular receptor and 5: immunoprecipitation of western 5I—CLMF binding to neutralization of CLMF inhibition of bioactivity. Twenty hybridomas secreting anti—CLMF antibodies were identified. The antibodies were found to I—labelled CLMF bioactivity as assessed in immunoprecipitate CLMF and to immunodeplete the T—cell proliferation and LAK cell induction assays. western blot analysis showed that each antibody binds to the 70 kDa heterodimer and to one of the subunits. Each of the above—mentioned 20 anti—CLMF monoclonal antibodies were specific for CLMF and in particular for the 40 kDa subunit of CLMF. A CLMF receptor binding assay has been developed to evaluate the ability of individual antibodies to inhibit CLMF binding to its cellular receptor.— The assay measures the binding of 125 I—labelled CLMF to PHA activated PBL blast cells in the presence and absence of each antibody. Of the 20 antibodies tested, 12 antibodies were found to inhibit greater than 60% of the l25I—labelled CLMF binding to the blast cells. Two inhibitory antibodies, viz. 7B2 and 4A1, neutralize CLMF bioactivity while one non-inhibitory antibody, 8E3, does not neutralize CLMT bioactivity. antibodies which block 125I—labelled CLM binding to its cellular receptor will neutralize CLMF bioactivity as These data confirm that assessed by the T—cell proliferation and LAK cell induction assays. The ability of the antibodies specific for the 40 kDa subunit of CLMF to neutralize CLMF bioactivity indicates that determinants on the 40 kDa subunit are necessary for binding to the CLME cellular receptor.

The monoclonal anti—CLMF antibodies of the present invention provide powerful analytical, diagnostic and therapeutic reagents for the immunoaffinity purification of natural and recombinant human CLMF. the development of human CLMF immunoassays, the identification of the active site of the 40 kDa subunit of CLMF and may be used in therapeutic treatments of patients which require selective immunosuppression of cytotoxic T cells, such as in transplantation. Monoclonal antibodies which recognize different epitopes on human CLMF can be used as reagents in a sensitive two—site immunoassay to measure levels of CLMF in biological fluids, cell culture supernatants and human cell extracts.

Thus. monoclonal antibodies against CLMF which exhibit a number of the present invention is also directed to utilities including but not limited to: l. Utilizing the monoclonal antibodies as affinity reagents for the purification of natural and recombinant human CLMF} . Utilizing the monoclonal antibodies as reagents to configure enzyme-immunoassays and radioimmunoassays to measure natural and recombinant CLMF in biological fluids, cell culture supernatants. cell extracts and on plasma membranes of human cells and as reagents for a drug screening assay; . Utilizing the monoclonal antibodies as reagents to construct sensitive two—site immunoassays to measure CLMF in biological fluids, cell culture supernatants and human cell EXCIBCCSZ . Utilizing the monoclonal antibodies as reagents to identify determinants of the 40 kDa subunit which participate in binding to the 35 kDa subunit and which participate in binding to the CLMF cellular receptor; . Utilizing the intact IgG molecules, the Fab fragments or the humanized IgG molecules of the inhibitory monoclonal antibodies as therapeutic drugs for the selective blockade of proliferation and activation of cytotoxic T cells, such as in transplantation. ’ BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a plot of a supernatant solution obtained from cultured NC37 lymphoblastoid cells applied to a Nu-Gel P—SP column showing the protein fraction containing TGF activity being eluted with a salt gradient.

Figure 2 is a plot of the material containing TGF activity obtained from the separation shown in Figure 1 as it was being eluted with a salt gradient through a Blue-B-Agarose Column.

Figure 3 shows the plot of the material containing TGF activity obtained from the separation shown in Figure 2 as it was being eluted with a NaC1 gradient through a Mono Q column.

Figure 4 shows a SDS-polyacrylamide gel electrophoresis (SDS—PAGE) analysis of the fractions 30 to 45, 48 and 50 obtained from the step illustrated in Figure 3. The numbers on the left side, i.e. 44 and 68. refer to the apparent molecular weight of standard proteins of 44 and 68 kDa in lane 8.

Figure 5 shows the elution profile through a Vydac Diphenyl column of fraction 38 from the Mono Q Chromatography separation (reversed—phase HPLC) shown in Figure 3.

Figure 6 shows SDs—PAGE analysis of protein purity of the protein fractions 85-90 recovered from the separation process depicted in Figure 5.

Figure 7 shows a SDS—PAGE analysis of fractions 87 and from the reversed—phase HPLC separation under non—reducing (lane A: without B-mercaptoethanol) and reducing (lane B; in the presence of B~mercaptoethanol) conditions showing the 75,000 molecular weight CLMF remaining lanes in the gel shown in this Figure contain separated into two subunits of 40 kDa and 35 kDa. standard proteins comprising the 44 and 68 kDa marker protein.

Figure 8 shows the elution pattern of the proteins from the supernatant solution from NC—37 cells applied to a Nu- Gel P—SP column and eluted with a salt gradient.

Figure 9 is a B1ue—B—Agarose column salt gradient elution profile of the active fractions obtained from the Nu—Ge1 P—sP column elution shown in Figure 8.

Figure 10 is a Mono—Q column salt gradient elution profile of the active fractions obtained from the elution shown in Figure 9.

Figure 11 is the elution pattern through a Vydac Diphenyl column of active fractions 39 and 40 obtained from the Mono Q Chromatography shown in Figure 10.

Figure 12 shows a SDS—PAGE analysis under reducing conditions of the active fractions obtained from the separation process shown in Figure 11.

Figure 13 is a schematic diagram depicting the separation of the 40 kDa subunit from the 35 kDa subunit of the CLMF cytokine.

Figure 14 is a schematic diagram depicting the determination of the amino acid composition, the N—terminal sequencing, the proteolytic digestion and the complete sequencing of the 40 kDa subunit of the CLMF cytokine.

Figure 15 shows a separation of the tryptic peptides of the digested 40 kDa subunit of the CLMF cytokine.

Figure 16 shows a separation of the proteolytic peptides of the Staphylococcus aureus V8 protease digested 40 kDa subunit CLMF.

Figure 17 is a chart which summarizes the information on the protein structure obtained from the analysis of the proteolytic peptides of the 40 kDa subunit of CLMF. The following abbreviations and symbols are used: N—t — N—terminal sequencing on intact protein Tr — tryptic peptides from map HP2383 numbered by "fraction number V8 — V8 protease peptides from map HP2412 numbered by fraction number - indicates probable glycosylation site: boxes indicate potential sites Figure 18 shows the SDS—PAGE analysis of Fraction 39 from the Mono Q FPLC elution profile shown in Figure 3. Lane A: Standardproteins without B—mercaptoethano1: lane B: Fraction 39 without B—mercaptoethanol; lane C: Fraction 39 with B—mercaptoethano1; lane D: Standard proteins with B-mercaptoethanol.

Figure 19 relates to the purification of the 35 kDa subunit by reversed—phase HPLC and depicts the elution pattern through a Vydac C-18 column of fraction 39 of the Mono Q chromatography which was reduced in 5% B—mercapto- ethanol.

Figure 20 shows a SDS—PAGE gel analysis under non-reducing conditions of the fractions which were fluorescamine positive from the Vydac C-18 column elution S: = profile shown in Figure 19. protein—standard; F: = flow—through; numbers refer to the fraction number.

Figure 21 depicts the elution pattern of a tryptic digest of fractions 36 and 37 of the Mono Q Chromatography through a YMC ODS column.

Figure 22 shows the stained PVDF membrane with the smeared bands comprising the CNBr cleaved CLMF before 22B) and after (Fig. , 14. 12. and 9 kDa, contain the CNBr fragments having the following sequences: (Fig. 22A) excising the regions of about 29, respectively. The regiones _ 11 - I (P?)—P-K—N—L—Q—L—K—P—L—K—N—?—V—(Q?)- (New sequence from 40 kDa protein) ?—Q—K—A—(R?)—Q—T-L-E-F-Y-P-?-T- (New sequence starting at residue no. 30 of 35 kDa protein) I11 V—V—L—T—?—D—T-P—E-E—D—G—1—T- (Starts at residue no. 24 of 40 kDa protein) IV V-D-A—V—(H?)—K-L—K—Y—E—?—Y-T—?-?—F-F-I- (Starts at residue no. 190 of 40 kDa protein) note: it is assumed or known that the above sequences are preceeded by a Met residue. V Figure 23 shows a reverse—phase HPLC separation of the peptide fragments obtained by cleaving CLMF with CNBr.

Figure 24 shows an SDS-PAGE of pure CLMF and "free" unassociated 40 kDa subunit of CLMF purified by affinity chromatography using the monoclonal antibody 7B2 covalently attached to an agarose resin. Lane A: molecular weight marker proteins; lane B: starting material: lane C: flow- through; lane D: acid eluatez lane E: potassium thiocyanate eluate. 4 Figure 25 a. b, c and d show the DNA sequence and the deduced amino acid sequence of the 40 kDa subunit of human CLMF.

Figure 26 a. b and c show the CDNA sequence and the deduced amino acid sequence of the 35 kDa subunit of CLMF Figure 27 depicts the inhibition of CLMF bioactivity by serum from rats immunized with CLMF and from non—immunized rats (control). _ 12 - Figure 28 shows a SDS-PAGE analysis of immunoprecipitates of l25l—CLMF by monoclonal antibodies Al (lane 1), 4Dl (lane 2), 8E3 (lane 3) and 9C8 (lane 4), bY 8 Control 3ntib0dY (lane 5). by immune rat serum (lanes 6 and 8) and by normal rat serum (lanes 7 and 9). On the left side the molecular weight in kDa is indicated.

Figure 29 shows the immunodepletion of CLMF bioactivity (TGF activity) by monoclonal anti—CLMF antibodies (a—CLMF).

Figure 30 shows the immunodepletion of CLMF bioactivity (LAK induction activity) by monoclonal anti—CLMF antibodies (a—CLMF).

Figure 31 shows a Western blot analysis of the reactivity of the monoclonal antibodies (mAbs) 7B2, 4A1 8B3. 6A3, 9F5 and 2A3 and of rat polyclonal anti—CLMF antibodies (RS1) with the CLMF 75 kDa heterodimer. NRS: = normal rat serum.

Figure 32 shows a Western blot analysis of the reactivity of monoclonal and rat polyclonal anti—CLMF antibodies with the CLMF 40 kDa subunit. In lanes 1 to 18 the following mAbs were used: 4A1, 4D1, 7B2, 7A1. 2A3, 1C1, BB4, BA2, BE3, 1B8, 4A6, 6A2. 8C4, 9P5, 6A3. 9C8, 8A1 and 22E7, respectively. In lane 19 a control antibody, in lane a fusion rat serum and in lane 21 a normal rat serum was used.

Figure 33 shows the binding of l25l—CLMF to PHA~activated peripheral blood lymphocyte (PBL) lymphoblasts.

Figure 34 shows the inhibition of l25l—CLMF binding to PHA—activated PBL blast cells by rat anti—CLMF serum. The data are expressed as amount (% bound) of l2SI—CLMF binding to the cells in the presence of the indicated concentrations of serum when compared to the total specific binding in the absence of serum.

Figure 35 shows the inhibition of the binding of l2SI—CLMF to PHA-activated PBL blast cells by monoclonal antibody supernatants. The data are expressed as % l25I—CLMF to the cells in the presence of a 1:1 dilution of supernatant when compared to inhibition of the binding of the total specific binding in the absence of antibody supernatant.

Figure 36 shows the inhibition of the binding of SI-CLMF to PHA—activated PBL blast cells by various concentrations of purified monoclonal antibodies. The data l25I—CLMF bound to the cells in the presence of the indicated the total are expressed as the amount (% cpm bound) of concentrations of antibody when compared to specific binding in the absence of antibody.

Figure 37 shows a western blot analysis of the antibody with the 75 kDa CLMF (nonreduced) and with the 35 kDa CLMF subunit (reduced). reactivity of a rabbit polyclonal anti—CLMF The antibody was prepared against a synthetic peptide fragment of the 35 kDa CLMF subunit. Lanes 1 to 5 are without B-mercaptoethanol: lanes 6 to 10 with B-mercaptoethanol. -14..

Lane 3 4 6 7 8 9 »l ul CLMF ul CLMF ul CLMF Blank Blank ul prestained molecular weight standards 1 ul CLMF ul CLMF ul CLMF ul prestained molecular weight standards The amino acid sequences of the 35 kDa subunit and the 40 kDa subunit of the CLMF protein is depicted in Figures 25 and 26.

The present invention also relates to antibodies directed to the said CLMF proteins as well as to the processes for the preparation of the said antibodies. -16..

The practice of the present invention will employ. unless otherwise indicated. conventional techniques of molecular biology. microbiology, recombinant DNA and immunology, which are within the skills of an artisan in the field. Such techniques are explained fully in the literature. See e.g.. Maniatis. Fitsch & Sambrook.

MOLECULAR CLONING; A LABORATORY MANUAL (1982); DNA CLONING, VOLUMES I AND II (D.N Glover ed., 1985); OLIGONUCLEOTIDE SYNTHESIS (M.J. Gait ed., 1984): NUCLEIC ACID HYBRIDIZATION (B.D. Hames & S.J. Higgins eds., 1984); TRANSCRIPTION AND TRANSLATION (B.D. Harnes & S.J. Higgins eds., 1984); ANIMAL CELL CULTURE (R.I. Freshney ed., 1986): IMMOBILIZED CELLS AND ENZYMES (IRL Press, 1986); B. Perbal. A PRACTICAL GUIDE TO MOLECULAR CLONING (1984): the series, METHODS IN ENZYMOLOGY (Academic Press, GENE TRANSFER VECTORS FOR MAMMALIAN CELLS (J.H. Miller and M.P. Calos eds., 1987. Cold Spring Harbor Laboratory), Methods in Enzymology vol. 154 Inc.); -17.. and Vol. 155 (Wu and Grossman. end Wu. eds., respectively); IMMUNOCHEMICAL METHODS IN CELL AND MOLECULAR BIOLOGY (Mayer and Walker, l987.

PROTEIN PURIFICATION: PRINCIPLES AND PRACTICE, Edition (1987. N.Y.), and HANDBOOK OF EXPERIMENTAL IMMUNOLOGY, VOLUMES l—IV (D.M. Weir and C.C.

Blackwell eds., 1986). eds., Academic Press, London), Scopes, second Springer—Verlag.

For example, useful vectors may consist of segments of chromosomal, non—chromosomal and synthetic DNA sequences. Examples of such vectors are viral vectors. such as the various known derivatives of sv4o, bacterial vectors, such as plasmids from E. pCRl, pBR322, pMB9 and RP4. phage DNAS, derivatives of phagek. M13 and other filamentous coli including such as the numerous single—stranded DNA phages, as well as vectors useful in yeasts, such as the zu plasmid, vectors useful in eukaryotic cells more preferably vectors useful in animal adenovirus and/or cells, such as those containing SV40. retrovirus derived DNA sequences. Useful vectors may be also derived from combinations of plasmids and phage DNA’s, such as plasmids which have been modified to comprise phage DNA or other derivatives thereof.

Expression vectors which may be used for the preparation of recombinant CLMF proteins are characterized by comprising at least one expression control sequence which is operatively linked to the CLMF DNA sequence inserted in the vector in order to control and to regulate the expression of the cloned CLMF DNA sequence. Examples of useful expression control sequences are the lac system, the trp system. the tac system, the trc system, major operator and promoter regions of phage A, the control region of fd coat protein, the glycolytic promoters of yeast. e.g.. the promoter for —phosphog1ycerate kinase. the promoters of yeast acid _]_8- Pho S, and promoters derived from polyoma phosphatase, e.g., the promoters of the yeast a—mating factors. and simian virus, virus, adenovirus, retrovirus, e.g., the early and late promoters or SV40, and other sequences known to control the expression of genes of prokaryotic or eukaryotic cells and of their viruses as well as combinations of the said promoter/operator sequences.

Furthermore, within each specific expression vector, various sites may be selected for insertion of the CLMF—re1ated DNA sequences of the present invention. These sites are usually designated by the restriction endonuclease which cut them. They are well recognized by those of skill in the art. It is, of course to be understood that an expression vector useful in this invention need not have a restriction endonuclease site for insertion of the chosen DNA fragment. fragment by alternative means.

Instead, the vector could be joined to the The site chosen in the expression vector for the insertion of a selected DNA fragment and the operative linking of the DNA fragment to an expression control sequence is determined by a variety of factors, such as the number of sites susceptible to a particular restriction enzyme, the location of start and stop codons relative to the vector sequence and the desired _ 19 _ selection method for the host transformed with the recombinant vector. The choice of a vector and an insertion site for a DNA sequence is determined by a balance of these factors. not all selections being equally effective for a given case.

The host cell used for the expression of the CLMF—related DNA sequence known hosts. may be selected from a variety of Examples for such hosts are prokaryotic or eukaryotic cells. A large number of such hosts are available from various depositories such as the American Type Culture Collection (ATCC) or the Deutsche Sammlung fur Mikroorganismen (DSM). Examples for prokaryotic cellular hosts are bacterial strains such as E.coli, B.subti1is and others. Preferred hosts are mammalian cells such as the SV40 transformed African Green monkey kidney cell line COS.

Not all host/expression vector combinations function with equal efficiency in expressing a given DNA sequence.

However, a particular selection of a host/expression vector combination may be made by those of skill in the art after due consideration of the principles set forth herein without departing from the scope of this invention. For example, the selection should be based on a balancing of a number of factors, These include, for example, compatibility of the host and vector, susceptibility of the protein to proteolytic degradation by host cell enzymes, possible contamination of the protein to be expressed by host cell proteins difficult to remove during purification, toxicity of the proteins encoded by the DNA sequence to the host, ease of recovery of the desired protein, expression characteristics of the DNA sequence and the expression control sequence operatively linked to them, biosafety, costs and the folding, form or any other necessary post—expression modifications of the desired protein. _ 20 - The host organisms which contain the expression vector comprising the CLMF DNA are usually grown up under conditions which are optimal for the growth of the host organism. Towards the end of the exponential growth, when the increase in the number of cells per unit time decreases, the DNA Coding for the protein is transcribed and the transcribed the expression of the CLMF protein is induced, i.e. mRNA is translated. The induction can be effected by adding an inducer or a derepressor to the growth medium or by altering a physical parameter, e.g. by a temperature change.

The CLMF protein produced in the host organism can be secreted by the cell by special transport mechanisms or can be isolated by breaking open the cell. The cell can be broken open by mechanical means [Charm et al., Meth. Enzmol. gg: 476-556 (l971)], treatment) or by chemical means (e.g. detergent treatment, by enzymatic treatment (e.g. lysozyme urea or guanidine-HC1 treatment, etc.) or by a combination thereof.

In eukaryotes, polypeptides which are secreted from the cell are synthesized in the form of a precursor molecule.

The mature polypeptide results by cleaving off the so—called signal peptide. As prokaryotic host organisms are not capable of cleaving eukaryotic signal peptides from precursor molecules, eukaryotic polypeptides must be expressed directly in their mature form in prokaryotic host which organisms. The translation start signal AUG. corresponds to the codon ATG on the level of the DNA, causes that all polypeptides are synthesized in a prokaryotic host organism with a methionine residue at the N-terminus. In certain cases, depending on the expression system used and possibly depending on the polypeptide to be expressed this N—terminal methionine residue is cleaved off.

The CLMF produced by fermentation of the prokaryotic and eukaryotic hosts ._2]_.. can then be purified to essential homogeneity by known methods such as, for example, by centrifugation at different velocities. by precipitation with ammonium sulphate. by dialysis (at normal pressure or at reduced pressure). by preparative isoelectric focusing, by preparative gel electrophoresis or by various chromatographic methods such as gel filtration, high performance liquid chromatography (HPLC). ion exchange chromatography, reverse phase chromatography and affinity chromatography (e.g. on Sepharose” Blue CL—6B or on carrier—bound monoclonal antibodies directed against CLMF).

The proteins or derivatives prepared in accordance with this invention or pharmaceutical compositions comprising the protein or derivative may be administered to warm blooded mammals for the clinical uses indicated above. The administration may be by any conventional modes of administration of agents which exhibit antitumor activity auch as by intralesional or parenteral application either intravenously, subcutaneously or intramuscularly. Obviously, the required dosage will vary _ 22 _ with the particular condition being treated, the severity of the condition, the duration of the treatment and the method for administration. A suitable dosage form for pharmaceuti- cal use may be obtained from sterile filtered, lyophilized protein reconstituted prior to use in a conventional manner.

It is also within the skill of the artisan in the field to prepare pharmaceutical compositions comprising protein of the present invention by mixing the said protein with compatible pharmaceutically acceptable carrier materials such as buffers, stabilizers, bacteriostats and other excipients and additives conventionally employed in pharmaceutical parenteral dosage forms. The present invention also relates to such pharmaceutical compositions.

The preferred form of administration depends on the intended mode of administration and therapeutic application. The pharmaceutical compositions comprising a protein or peptide derivative of the present invention also will preferably include conventional pharmaceutically acceptable carriers and may include other medicinal agents carriers, excipients. etc.. interleukin-2), adjuvants, (e.g. e.g., Preferably. human serum albumin or plasma preparations. the compositions of the invention are in the form of a unit dose and will usually be administered one or more times a day. The unit dose is preferably packed in 1 ml vials containing an effective amount of the protein and if desired of interleukin—2 in lyophilized form. The vials containing the protein and if desired the interleukin—2 are preferably packed in a container together with written instructions describing the correct use of the pharmaceutical composition. _ 23 _ In order that our invention herein described may be more fully understood, the following examples are set forth. It should be understood that these examples are for illustrative purposes only and should not be construed as limiting this invention in any way to the specific embodiments recited therein. It has to be noted that the specific product names and suppliers mentioned below are not meant to be mandatory. The person skilled in the art is in a position to select alternative products from other suppliers.

EXAMPLE PURIFICATION AND CHARACTERIZATION OF CYTOTOXIC LYMPHOCYTE MATURATION FACTOR (CLMF) Production of Supernatant Liquid Containing CLMF.

Human NC—37 B lymphoblastoid cells (ATCC CCL 214, American Type Culture Collection, Rockville, MD) were used for production of CLMF. These cells were maintained by serial passage in RPMI 1640 medium supplemented with 5% heat—inactivated (56°C, 2 mM min.) fetal bovine serum. and 100 ug/ml streptomycin (all cell culture media were from GIBCO Grand Island, NY).

L—glutamine, 100 units/ml penicillin, Laboratories.

Higher producer sublines of NC—37 cells were derived by limiting dilution cloning in liquid microcultures. Each well of three Costar 3596 microplates (costar Co., Cambridge, MA) received 100 ul of a cell suspension containing five NC—37 cells/ml. The medium used for the cloning was a lzl mixture of fresh passage medium and filtered, conditioned medium from stock cultures of the parent NC-37 cells. One week and two weeks after culture initiation each of the microcultures was fed with 50 ul of the 1:1 mix of fresh and conditioned medium. Between 3 and weeks after culture initiation the contents of wells _ 24 _ containing clones of NC—37 cells were harvested and passed into larger cultures. when the number of cells in a given subline exceeded 1.4 x 106, one million cells were stimulated to produce CLMF in 1 ml cultures containing 3 ng/ml phorbol 12-myristate l3—acetate (PMA) (Sigma Chemical Co., St. MO) and 100 ng/ml calcium ionophore A23l87 (Sigma). Supernatants were Louis, harvested from the cultures after 2 days, dialyzed against about 50 volumes of Dulbecco's phosphate buffered saline (Gibco) using e.g. SPECTROPOR® #1 tubing (Fisher Scientific) overnight with one change of buffer and then for 4 hours against 50 volumes of RPMI 1640 medium with 50 ug/ml of gentamicin (both from Gibco) and tested for CLMF by means of the T cell growth factor assay (see below).

NC-37.89. NC-37.98, and NC—37.102, identified which routinely produced CLMF at titers > 4 times Three sublines. were the titers produced by the parental NC-37 cell line. Since cells from these three sublines produced CLMF at similar titers (3 800 units/ml), culture supernatants derived from the three sublines were pooled for use as starting material for the purification of CLMF.

Bulk production of CLMF was carried out in roller bottle cultures on a roller apparatus set at about 38 rpms (Wheaten Cell Production Roller Apparatus Model II, Wheaton Millville, NJ). Cell suspensions were prepared containing l~l.5 x 106 NC—37.89. NC-37.98 or NC-37.102 cells/ml in RPMI 1640 medium supplemented with 1% Instruments, Nutridoma-SP (Boehringer Mannheim Biochemicals, IN), 2 mM L—glutamine, 100 units/ml ug/ml streptomycin, 10 nq/ml PMA and 20-25 Two hundred fifty to three Indianapolis, penicillin. ng/ml calcium ionophore A23l87. hundred fifty ml aliquots of the cell suspensions were added to Falcon 3027 tissue culture roller bottles (Becton Lincoln Park. NJ) which had been gassed with a 95% air.

Dickinson. mixture of 5% CO , The roller bottles were then _ 25 - capped tightly and incubated at 37°C with continuous rolling for three days. At the end of this time, the culture supernatants were harvested. EDTA and phenylmethylsulfonyl fluoride (both from Boehringer Mannheim) were added to the culture supernatants at final concentrations of 1 mM and 0.1 mM, respectively, to retard proteolytic degradation. The supernatants were stored at 4°C.

Lympokine Activated Killer (LAK) Cell Induction (LCI) Assay.

Culture supernatants and chromatographic fractions were tested for their ability to synergize with rIL—2 to induce the generation of cytolytic LAK cells as follows. Human peripheral blood mononuclear cells (PBMC) were isolated by the following method. Blood from normal volunteer donors was drawn into syringes containing sufficient sterile preservative-free heparin (Sigma) to give a final The blood was balanced salt solution (HBSS) The diluted blood was then layered over 15 ml aliquots of Ficoll/sodium concentration of approximately 5 units/ml. diluted 1:1 with Hanks‘ without calcium or magnesium (GIBCO). diatrizoate solution (Lymphocyte Separation Medium, Organon Teknika Corp., Durham, NC) in 50 ml Falcon 2098 centrifuge The tubes were centrifuged for 30 minutes at room the cells tubes. temperature at 500 x g. Following centrifugation, floating on the Ficoll/sodium diatrizoate layer were collected and diluted by mixing with 3 2 volumes of HESS without calcium or magnesium. The resulting cell suspension was then layered over 15 ml aliquots of 20% sucrose (Fisher) in RPMI l640 medium with 1% human AB serum (Irvine Scientific, Santa Ana, CA) in Falcon 2098 centrifuge tubes.

The tubes were centrifuged for 10 minutes at room temperature at 500 x q, and the supernatant fluids were discarded. The cell pellets were resuspended in 5 ml of HBSS without calcium or magnesium, repelleted by centrifugation, and finally resuspended in the appropriate _ 25 _ culture medium. Accessory cells were removed from the PBMC by treatment with 5 mM L—glutamic acid dimethyl ester (Sigma) using the same conditions as described by Thiele et al. J. Immunol. ;;;:2282—229O (1983) for accessory cell depletion by L-leucine methyl ester except that the glutamic acid ester was substituted for the leucine ester.

The accessory cell—depleted PBMC were further fractionated by centrifugation on a discontinuous Percoll density gradient (Pharmacia, Piscataway, NJ) as described by Wong et al.. Cell Immunol. ;;;:39—54 (1988). cells recovered Mononuclear from the 38, 41. 45. and 58% Percoll layers used as a source of LAK cell precursors in cells recovered from the Percoll gradient were pooled and The were washed and the assay. suspended in tissue culture medium (TCM) composed of a 1:1 mixture of RPMI 1640 and Dulbecco's modified Eagle's medium, supplemented with 0.1 mM nonessential amino acids. 60 pg/ml arginine HCl, 10 mM HEPES buffer. ’ ug/ml streptomycin (all available from GIBCO), 5 x l0_5 mM L—g1utamine, 100 units/ml penicillin, 100 M 2—mercaptoethanol (Fisher Scientific, Fair Lawn, NJ), l mg/ml dextrose (Fisher), and 5% human AB serum (Irvine CA). 24-well tissue culture plates (costar, Cambridge, MA) in 1 ml cultures (7.5 x 105 Scientific, Santa Ana, These cells were incubated in cells/culture) to which lo'4 M hydrocortisone sodium succinate (Sigma) was added to minimize endogenous cytokine production. some cultures also received human rIL—2 (supplied by Hoffmann-La Roche, Inc., Nutley, NJ) at a final concentration of 5 units/ml and/or supernatants to be assayed for CLMF activity. All cultures were incubated for 3-4 days at 37°C in a humidified atmosphere of 5% CO2. 95% air.

At the end of this incubation, the contents of each culture were harvested. and the cells were pelleted by centrifugation and resuspended in 0.5 ml of fresh TCM. One tenth ml aliquots of these cell suspensions were mixed with _ 27 _ . 51 H 0.1 ml aliquots of Cr-labelled K562 or Rajl cells (both cell lines may be obtained from the ATCC) and tested for their lytic activity in 5 hour 51c: release assays. method for labelling target cells with 51 . 51 the percentage of Cr released from target cells lcr released spontaneously from target cells incubated alone. incubated with lymphocytes and g is the percentage of The total releasable 51C: was determined by lysis of the al., All lymphocyte populations LAK Cell Induction Microassay. The microassay for measuring synergy between rIL—2 and CLMF-containing solutions in the induction of human LAK cells was similar to the LAK cell induction assay described above but with the following modifications. Human peripheral blood mononuclear cells which had been depleted of accessory cells and fractionated by Percoll gradient centrifugation as described above were added to the wells of Costar 3596 microplates (5 x 104 cells/well). (5 units/ml final concentration) and/or purified CLMF or Some of the wells also received rIL—2 immunodepleted CLMF-containing solutions. All cultures contained 1O—4 M hydrocortisone sodium succinate (Sigma) and were brought to a total volume of 0.1 ml by addition of TCM with 5% human AB serum. The cultures were incubated for 3 days at 37°C, after which 0.1 ml of 5lCr—labelled K562 cells (5 X 104 cells/ml in TCM with 5% human AB serum) were added to each well. The cultures were then incubated overnight at 37°C. Following this, the cultures were centrifuged for 5 minutes at 500 X g, and the supernatant solutions were harvested by use of a Skatron supernatant collection system (Skatron, Sterling, VA). The amount of . .

Cr released into each supernatant $0lL1tlOI'l W85 measured _28_ with a gamma counter (Packard, Downer's Grove, IL). and the . . Sl _ % specific Cr release was calculated as described above. All samples were assayed in quadruplicate.

Qytolytic T Lymphocyte (CTL) Generation Assay.

Methods used for generating and measuring the lytic activity of human CTL have been described in detail by Gately et al. in J. Immunol. 13;: 1274-1282 (1986) and by Wong et al. in Cell. Immunol. 111: 39-54 (1988). peripheral blood mononuclear cells were isolated from the Human blood of normal volunteer donors, depleted of accessory cells by treatment with L—glutamic acid dimethyl ester. and fractioned by Percoll gradient centrifugation as described above. High density lymphocytes recovered from the interface between the 45% and 58% Percoll layers were used as responder lymphocytes in mixed lymphocyte-tumor cultures (MLTC). culture plates (Costar #3424) by incubation of Percoll CTL were generated in MLTC in 24-well tissue gradient—derived high density lymphocytes (7.5 x 105 uv—irradiated melanoma cells e.g. HTl44 (obtainable from ATCC) or with 5 x 104 gamma»irradiated melanoma cells e.g. HTI44 in TCM with 5% . 5 culture) together with 1 x 10 human AB serum (1.2 ml/culture). For uv—irradiation, HTl44 cells/ml in Hanks‘ balanced salt solution without phenol red (GIBCO) cells were suspended at a density of 1-1.5 x 106 containing 1% human AB serum. One ml aliquots of the cell suspension were added to 35 x 10 mm plastic tissue culture dishes (Falcon #3001), and the cells were then irradiated (960 uw/cmz for 5 min) by use of a 254 nm uv light (model UVG-54 MINERALIGHT® lamp, Ultra—violet Products, CA). were suspended at a density of 1-5 X 106 lnc., For gamma irradiation. HTI44 cells cells/ml in TCM with 5% human AB serum and irradiated (10,000 rad) by use of san Gabriel. a cesium source irradiator (model 143, J.L. Shepherd and CA).

HT144 were centrifuged and resuspended in TCM with 5% human Associates, San Fernando, Uv— or gamma—irradiated AB serum at the desired cell density for addition to the MLTC. In addition to lymphocytes and melanoma cells. some MLTC received human rIL-2 and/or purified human CLMF at the concentrations indicated. Hydrocortisone sodium succinate (Sigma) was added to the MLTC at a final concentration of -4 M (cultures containing uv—irradiated melanoma cells) or 10—5 M (cultures containing gamma-irradiated melanoma cells) to supress endogenous cytokine production [S. Gillis et al., J. Immunol. ;;;: 1624-1631 (l979)] and to reduce the generation of nonspecific LAK cells in the cultures [L.M.

Muul and M.K. Gately. J. Immunol. lggz 1202-1207 (l934)]. at 37°C in a humidified for 6 days. At the end of lymphocytes from replicate cultures were pooled, The cultures were incubated atmosphere of 5% CO2 in air this time, centrifuged. resuspended in 1.2 ml TCM containing 5% human AB serum, and tested for their ability to lyse HTI44 melanoma cells, and. as a specificity control, K562 erythroleukemia cells (obtainable from ATCC) in overnight Cr release assays.

Melanoma cells and K562 cells were labeled with 51Cr sodium chromate as described by Gately et al. [JNCI Q3: 1245-1254 (1982)). Likewise, measurement of lympocyte— mediated lysis of 5lCr—labeled melanoma cells was performed in a manner identical to that described by Gately et al. (ibid.) for quantitating lysis of glioma target Slcr-labeled K562 cells, 0.1 ml aliquots of lymphocyte suspensions were mixed with 25 lCr—labeled K562 (2 x 105 TCM with 5% human AB serum) in the wells of costar 3696 cells. For assaying the lysis of ul aliquots of cells/ml in "half—area" microtest plates. After overnight incubation at 37°C, the plates were centrifuged for 5 min at 1400 x g, and ul of culture medium was aspirated from each well. The Cr in each sample was measured 1 and the % 5 Cr All assays amount of with a gamma counter (Packard), specific release was calculated as described above. were performed in quadruplicate, and values in the table (see below) represent -30.. the means 1 1 S.E.M. of replicate samples.

T cell growth factor (TGF) assay.

The ability of culture supernatants and chromatographic fractions to stimulate the proliferation of PHA—activated human T lymphoblasts was measured as follows. Human PBMC were isolated by centrifugation over discontinuous Ficoll and sucrose gradients as described above for the LCI assay.

The PBMC (5 x 105 cells/ml) were cultured at 37°C in TCM containing 0.1% phytohemagglutinin-P (PHA—P) (Difco Laboratories. Detroit. MI). After 3 days, the cultures were split 1:1 with fresh TCM, and human rIL—2 was added to each The cultures were then incubated for an additional 1 to 2 days, culture to give a final concentration of S0 units/ml. at which time the cells were harvested, washed. and cells/ml. To this cell suspension was added heat—inactivated goat anti—human rIL-2 . 5 resuspended in TCM at 4 x 10 antiserum (final dilution: 1/200) to block any potential IL—2—induced cell proliferation in the assay. This antiserum may be prepared using methods well—known in the art or may be obtained from Genzyme Co., Boston, MA. The antiserum used was shown to cause 50% neutralization of 2 units/ml rIL-2 at a serum dilution of l/20.000.

Fifty ul aliquots of the cell suspension containing anti-IL-2 antiserum were mixed with 50 ul aliquots of serial dilutions of culture supernatants or chromatographic The cultures were incubated for 1 day at 37°C in a humidified fractions in the wells of costar 3596 microplates. . . 3 . . atmosphere of 5% CO2 in air, and 50 ul of H—thymid1ne (New England Nuclear, Boston, MA), 10 uCi/ml in TCM, were then added to each well. The cultures were further incubated overnight. subsequently. the culture contents were harvested onto glass fiber filters by means of a cell harvester (Cambridge Technology Inc., Cambridge, MA), and . . . . .

H—thym1d1ne incorporation into cellular DNA was measured _ 31 _ by liquid scintillation counting. All samples were assayed in triplicate.

In purifying CLMF it was necessary to define units of activity in order to construct chromatographic elution profiles and to calculate the percent recovery of activity and the specific activity of the purified material. To do this. a partially purified preparation of human cytokines produced by coculturing PHA—activated human PBMC with NC—37 cells was used as a standard. The preparation was assigned an arbitrary titer of 2000 units/ml. Several dilutions of this preparation were included in each TGF or LAK induction assay. The results obtained for the standard preparation were used to construct a dose—response curve from which could be interpolated units/ml of activity in each unknown sample at the dilution tested. Multiplication of this value by the dilution factor yielded the activity of the original sample expressed in units/ml.

For antibody neutralization studies, the TGF assay was modified as follows. Twenty-five ul aliquots of CLMF-containing medium were mixed with 50 ul aliquots of serial dilutions of antiserum or antibody solutions in the wells of COSTAR 3S96® microplates. incubated for 30 minutes at 37°C, and 25 ul aliquots of a The mixtures were suspension of FHA-activated lymphoblasts (8 X 105/ml in TCM plus 1:100 anti—rIL—2) were then added to each well.

The cultures were further incubated, pulsed with H-thymidine, harvested, and analyzed for 3H—thymidine incorporation as described above.

Natural killer (NK) cell activation assay.

Purified CLMF was tested for its ability to activate NK cells when added alone or in combination with rIL—2 as follows. Human PBMC were isolated by centrifugation over discontinuous Ficoll and sucrose gradients as described _32_ above and were suspended in RPMI 1640 medium supplemented with 10% heat—inactivated fetal bovine serum, penicillin. 100 ug/ml streptomycin, units/ml and 2’mM L—glutamine.

The PBMC were incubated overnight at 37°C in 1 ml cultures C 6 (5 X 10 purified CLMF at various concentrations. cells/culture) together with rIL-2 and/or After 18-20 hours, the contents of the cultures were harvested and centrifuged, and the cells were resuspended in the same medium used for The cytolytic activity of the cultured PBMC was then assessed in 51 the overnight cultures.

Cr release assays as described above.

Concentration of cell supernatant solutions Stored, frozen crude human CLMF supernatant solutions totaling 60 liters prepared from several batches of induced NC—37 cells were pooled and concentrated 30-fold using the Pellicon Cassette System (30,000 NMWL PTTKOOO05; Corp., Bedford, MA). volume of approximately 1.9 liters, performed with 10 mM MES, NaOH.

Millipore After concentrating to the desired a buffer exchange was pH adjusted to 6.0 with 10 N The concentrate was centrifuged at 10.000 x g for minutes at 4°C and the precipitate discarded.

Ion—Exchanqe Chromatography on NuGel P-SP Column The concentrated supernatant solution was applied at a flow rate of 120 ml/hr to a Nu—Gel P—SP (Separation Industries, Metuchen, NJ) column (5 x 5 cm), lOmM MES, pH 6.0. absorbance monitoring at 280 nm was obtained. equilibrated in The column was washed until baseline Absorbed proteins were then eluted with a 500 ml salt gradient from O to 0.5 M NaCl/10 mM MES, 1). growth factor (TGF) activity. pH 6.0 at a flow rate of 2 ml/min (Fig. Aliquots of fractions were assayed for T cell Fractions containing TGF Fisher pH 7.5 in activity were pooled and dialyzed (Spectra/Por 7, Scientific) against 50 volumes 20 mM Tris/HC1, - 33 _ order to reduce the salt concentration of the 50-fold. preparation by Qye—Affinity Chromatography on Blue B—Agarose Column The dialyzed sample was centrifuged at 10,000 x g for 10 The supernatant solution was applied at a flow rate of 20 ml/hr minutes at 4°C and the precipitate discarded. to a Blue B—Agarose (Amicon, Danvers. MA) column (2.5 x 10 cm) equilibrated in 20 mM Tris/HCl, pH 7.5. The column was washed with this same buffer until baseline absorbance monitoring at 280 nm was obtained. Absorbed proteins were then eluted with a 500 ml salt gradient from 0 to 0.5 M NaCl/20 mM Tris/Hcl, pH 7.5 at a flow rate of 15 ml/hr (Fig. 2). Aliquots of fractions were assayed for TGF activity. Fractions containing TGF activity were pooled and dialyzed (Spectra/Por 7. Fisher Scientific) against 100 volumes 20 mM Tris/Hcl, pH 7.5 in order to reduce the salt concentration of the preparation by l00—fold.

Ion—Exchange Chromatography on Mono Q Chromatography The dialyzed sample was filtered through a 0.45 um NY) and the filtrate applied at a flow rate of 60 ml/hr to a Mono Q cellulose acetate filter (Na1gene Co., Rochester, HR 5/5 (Pharmacia LKB Biotechnology, Inc., Piscataway, NJ) column (5 x 50mm) equilibrated in 20mM Tris/HCI, pH 7.5.

The column was washed with this same buffer until baseline Absorbed proteins were then eluted with a 1 hr linear salt gradient from O to 0.25 M NaCl/20 mM Tris/HCl. pH 7.5 at a flow rate of 60 ml/hr (Fig. 3).

TGF activity and protein purity was assessed without reduction by SDS~PAGE [Laemmli. Nature (London) g;1:680—685 (l970)] using 12% slab gels.

[Morrissey, Anal. Biochem. ;;1:307—3lO (l98l)] to visualize absorbance monitoring at 280 nm was obtained.

Aliquots of fractions were assayed for Gels were silver stained protein (Fig. 4). Fractions 36 and 37 were of greater than -34.. % purity and revealed a major band at 75,000 molecular weight. Fractions 38 through 41 containing TGF activity, revealed the 75 kDa protein by SDS—PAGE with major contaminants at 55,000 and 40,000 molecular weight.

Therefore, to eliminate these contaminating proteins, fraction 33 of the previous Mono Q chromatography was diluted 1:1 vol/vol with B M urea and pumped onto a Vydac diphenyl column using a reversed-phase HPLC enrichment technique. The column was then washed with 5 ml of 0.1% trifluoroacetic acid. Elution of the proteins was accomplished with a gradient of 0~70% acetonitrile over 7 hrs in 0.1% trifluoroacetic acid (Fig. 5). Aliquots of fractions were assayed for TGF activity. Protein purity of the fractions containing TGF activity was assessed by SDS—PAGE under non—reducing conditions using a 10% slab gel. 6). and revealed protein of 75.000 molecular weight.

The gel was silver stained to visualize protein (Fig.

Fractions 86 through 90 were of greater than 95% purity Fractions 87 and 88 were pooled and aliquots were analyzed by SDS—PAGE under reducing (in the presence of B—mercaptoethanol) and non—reducing conditions (in the absence of B—mercapto- ethanol). the 75,000 molecular weight CLMF was separated into two subunits of 40.000 and 35.000 daltons (Fig. 7). Thus it was concluded that CLMF is a 75 kDa heterodimer composed of disu1fide— —bonded 40 kDa and 35 kDa subunits.

Under the reducing conditions.

The overall purification of CLMF that was achieved is shown in Table 1. The protein content of the Mono Q— and Vydac diphenyl—purified material was calculated on the basis DD+N.® :o~..uUcu: noa x ~.m oHo.o moo.o meg x mw.m meg x vn.m H.~ Hacmzmflo mvfixam =owuum.: pea K m.o moe.o ~ao.o mod N nv.m wofi K om.o m 0 0:02 um zomuumum nod x m.o m»o.o mno.o ooa H v.9 cog K ov.u H 0 0:02 nofi x m.H Ha vm.o moa H v.~ ooa K HH.m me mmouou oofi K m.m no o>.o coa K a.~ mom N oo.~ om mmum H0032 _ 3 mumuucmocoo _ vofi x m.m ommm mm.H mod N o.m moa x bm.H ovm.fi cmumpflflumuufia muzmumcuwmzm Q2 az Q2 mofi x e.q moa K mm.~ ooo.oo fidmo umfioom Aaexav Amev .fis\msV flay A~E\:V AHEV »afl>«uo< cmmuoum cwwuoum mums: xuw>Huo< ofiuwommm Hauoe cmaoom Hmuoe UOHOOQ ®EDHO> swam H mqm _ 35 _ of amino acid analysis. A specific activity of 8.5 x 107 . 7 units/mq and 5.2 x lo units/mg for Mono Q- dipheny1—purified material respectively, fact and Vydac was obtained. that the diphenyl—purified protein due to inactivation or denaturation of some of the molecules of CLMF in the HPLC elution solvents (i.e., 0.1% trifluoroacetic acid). acetonitrile in Chemical Characterization The ability to prepare homogeneous CLMF allowed for the first time the determination of the amino acid composition and a partial sequence analysis of the naturally occurring CLMF protein. Between 10 and 20 picomoles of Mono—Q—purified CLMF was subjected to hydrolysis. amino acid composition was and its determined (Table 2). cysteine and tryptophan were not determined (ND).

Quantitation of histidine was not possible due artifact peak. associated with Tris.

Proline. to a large coeluting with His (*).

Between 5 and 30 picomoles of dipheny1_puri£ied CLMF was subjected to hydrolysis with and without pre-treatment with performic acid. Complete amino acid composition was thus obtained (Table 3) with the exception of tryptophan.

Amino—termina1 sequence determination was attempted by automated Edman degradation on 100 pmol of the Mono Q-purified CLMF. two sequences present.

Data from the first 22 cycles indicated as would be expected from the heterodimeric structure of CLMF. These results may be summarized as follows: - 35 _ TABLE 3 Amino acid Aspartic acid or esfiaragine Tnreonine serine Glutamic acid or glucamine Praline Glycine Alanine cysteine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Histidine Lysine Acqinine Tryptophan _ Dreyer.

Reversed—Pnase HPLC phic system has been de A.S. and Lewis.

Neurocnemistry, Eds. Marks.

York) Vol. 6. scribed previously (1985) in Research Methods in N. and Rodnight. R.

(Plenum. New An automated fluorescence detection 'ne (Polysciences.

PA) monitored the protei -193.

Hesperia. CA). gradient in 0.1% TFA.

Egptein Analvsis instrument which n with fluorescamine for detection .. and Stein. 5. (1986) in Methods of Protein Microcnaracterization (shively. J.E., Ed.). pp. 105-119.

Humana Press. Clifton. NJ].

CA) [Hewick, W.J., , Hood, L.E..

J, ;;g:799o—7997 (1931)).

Phenylthionydantoin (PTH) amino acid derivatives were identified "on—1ine“ with an ABI Model 120A PTH analyzer, DETERMINATION OF PARTIAL -—j Purification of the 40 KDa subunit of CLMF AMINO ACID SEQUENCES OF THE :: P 2.4 liters using the Pellicon Cassette System and stored at To clarify this concentrate after thawing, the preparation was centrifuged and the precipitate discarded.

The supernatant solution was applied to a Nu—Ge1 P—SP column and protein was eluted with a 8). salt gradient (Fig.

Peak TGF activity was determined and the active fractions were pooled and dialyzed in order to reduce the by 50-fold. This after centrifugation to remove particulates, salt concentration of the preparation material, was applied to a Blue~B—Agarose column. a salt gradient (Fig. 9).

Protein was eluted with Peak TGF activity was determined and the active fractions were pooled and dialyzed in order to reduce the salt concentration of the preparation by lOO—fold.

Mono Q column. ). activity.

This material, after filtration. was applied to a Protein was eluted with a salt gradient (Fig. Aliquots of fractions were assayed for TGF Fractions 39 and 40 of the previous Mono Q chromato- graphy were pooled and diluted 1:1 vol/vol with 8M urea and pumped onto a Vydac diphenyl column using an enrichment technique. The column was then washed with 5 ml of 0.1% trifluoroacetic acid. Elution of the proteins was A accomplished with a gradient of O—70% acetonitrile over 7 hrs in 0.1% trifluoroacetic acid (Fig. 11). Aliquots of fractions were assayed for TGF activity. Protein purity of the fractions containing TGF activity was assessed by SDS—PAGE under reducing conditions, i.e. in the presence of B—mercaptoethanol (Fig. 12). contained the 40,000 dalton subunit >90% pure.

Fractions 94 through 97 Determination of the amino—terminal sequences of the subunits of CLMF _________________ _ 41 _ The ability to prepare a highly enriched preparation of the 40,000 dalton subunit of CLMF allowed for its partial sequence analysis.

Amino terminal sequence determination was attempted by automated Edman degradation on 20 pmol of the diphenyl- purified 40,000 dalton subunit. The results may be summarized as follows: Cycle 1 2 3 4 5 6 7 Amino -------------- -- Acid I W E L K K D Cycle 8 9 10 ll 12 13 14 Amino Acid V Y V V E L D Cycle 15 16 17 18 19 20 21 Amino Acid W Y P D A P G Cycle 22 23 Amino Acid E M with regard to the sequence analysis of 75,000 dalton CLMF and the sequence analysis of the 40,000 dalton subunit of CLMF, one can deduce the amino terminal sequence of the ,000 dalton subunit of CLMF. The amino terminal sequences of the 35,000 dalton subunit and the 40,000 dalton subunit can be summarized as follows: .000 dalton subunit: NH —?-?—Leu—Pro-Va1—Ala—Thr(?)—Pro-Asp—Pro—Gly— 2 Met—Phe—Pro—?—Leu—His—His—Ser(?)—Gln— ,000 dalton subunit: NH —I1e—Trp—Glu—Leu—Lys—Lys—Asp—Val—Tyr—Va1—val—G1u 2 23 Leu—Asp—Trp—Tyr—Pro—Asp—A1a—Pro—Gly—Glu—Met— where ? represents an undetermined or Determination of internal amino acid sequence segments of the 40 kDa subunit of CLMF CLMF was purified as described above. The 40,000 dalton subunit was separated and purified from the 35,000 dalton subunit by the method described by Matsudaira (J. Biol- Chem. gggz 10035-10038 (1987)). Fifty micrograms of CLMF (in 500 ul of 20 mM Tris, pH 7.5; 0.15 M Nacl) was diluted with 200 ul of a 2 x concentrate of sample buffer 680-685 (1970)). The sample was concentrated to 400 ul and disulfide bonds broken by the [Laemmli, Nature 227: addition of 18 ul B—mercaptoethanol followed by exposure to 105°C for 6 minutes.

The sample was loaded onto a minigel (1.0 mm thick) containing 12% polyacrylamide and electrophoresed according to Laemmli (supra). After electrophoresis, the gels were soaked in transfer buffer (10 mM 3-cyclohexylamino—1—pro— % methanol. reduce the amount of Tris and glycine. panesulfonic acid. pH 11.0) for 5 minutes to During this time, a polyvinylidene difluoride (PVDF) membrane (Immobilon; Millipore: Bedford, MA) was rinsed with 100% methanol and stored in transfer buffer. The gel, backed with two sheets "best—guessed" residue. of PVDF membrane and several sheets of blotting paper, was assembled into a blotting apparatus and electroeluted for 30 buffer. The PVDF membrane was washed in deionized H20 for 5 minutes. The edge of the .1% Coomassie Blue R-250 in 50% methanol for 5 minutes. and then destained in 50% methanol, 10% acetic acid for 5-10 The 40,000 dalton stained band was then matched to the corresponding region of the minutes at room temperature. matched the one previously determined (see above). method, the 40,000 dalton protein 40,000 subunit of CLMF.

By this was identified as the and collected in an Eppendorf They were then immersed in 300 ul of a 2% polyvinyl- After 30 tube. pyrrolidone (40,000 dalton) solution minutes. containing 2 ug of trypsin was added. 4 hours at 37°C.

The sample was shaken and incubated for The supernatant solution was then transferred into a second Eppendorf tube and the membrane pieces were further washed once with 100 ul of 88% _ 44 _ (vol/vol) formic acid and three times with 100 ul of deionized water. digestion mixture in the second Eppendorf tube. The resultant peptides contained in the pooled digest were separated a YMC C-18 column (2.6 x 50 mm: Morris Plains, NJ).

TABLE 4 Amino Acid Residue No.

Aspartic acid or asparagine 27.9 Threonine V 20.7 Serine 24.6 Glutamic acid or glutamine 44.6 Proline ND Glycine 16.3 Alanine 16.2 Cysteine ND Valine 20.9 Methionine 2.5 Isoleucine 10.3 Leucine 22.9 Tyrosine 12.9 Phenylalanine 9.9 Histidine 5.2 Lysine 24.5 Arginine 12.5 Tryptophan ND ygtg: The results represent the mean of two analyses.

Proline, cysteine, and tryptophan were not determined All washing solutions were added to the by narrow bore HPLC (HPl09oA, Hewlett Packard) on (28) (23) (34) (35) (14) (15) (14) (10) (23) (2) (12) (22) (12) (9) (26) (12) (10) (ND).

Values in parentheses represent the theoretical amino acid composition of the 40,000 dalton subunit based upon the primary structure of the protein deduced from sequence analysis of cloned 40,000 dalton subunit. _ 45 - The above described procedure is shown schematically in Figures 13 and 14.

The tryptic peptide map of the digested 40,000 dalton subunit is shown in Figure 15. Peptides were eluted with a linear gradient of acetonitrile. The peaks which were sequenced are numbered according to their fraction number.

The amino acid sequence of these peptides is shown in Table Many tryptic peptides were recovered from all regions of the intact 40.000 dalton subunit (Table 5). The N-terminal hexapeptide (fraction no. 60) was recovered in high yield.

The carboxy-terminal peptide (fraction no. 72) was recovered and is the full length of the predicted C—termina1 peptide although the last two amino acids were not positively confirmed by sequencing. This is probably due to the fact that Cys and Ser residues are not detected well, especially when they occur at the end of a peptide. Four potential Asn—linked carbohydrate sites may be predicted from the CDNA sequence. Two peptides containing two of these sites were sequenced. When peptide 196-208 (fraction no. 70) was sequenced, no peak was detected at residue 200 indicating that this Asn (predicted by the CDNA) is indeed Peptide 103-108 (fraction no. 52) yielded Asn Therefore, this site is not glycosylated. glycosylated. at residue 103.

An unknown peak seen in the phenylisothiocyanate (PTH) sequence analysis [Hewick et al., J. Biol. Chem. 256: 7990 (l98l)] of fraction no. 55 was detected at the position corresponding to residue no. 148. The site is predicted to be a Cys residue which is normally not detected by sequence analysis unless it is modified.

The above PVDF transfer procedure was repeated on a second 50 ug aliquot of CLMF (see Figure 13 and 14 for TABLE 5 Trvptic 40kDa CLMT peptides off PVDF fraction residue N—termina1 sequence £2; £2; -108 N—K—T—F-L—R -157 G—S—S—D—P—Q-G—V-T-*-G—A—A—T—L—S—A—E—R .55 & 57 267-279(?) V-F—T—D—K-T-S-A—T—V—I-?—R -58 T-L-T—I—Q—V—K -228 N-L—Q—L—K-P—L-K-N-S—R -6 I-W—E—L-K-K -? A-Q-D—R—Y-Y—S—S— -102(?) K-E-D—G-I—W-S-T—D-I—L—K—D—Q—K—E—P— 70 196-208 L—K-Y—E-?-Y-T—S-S—F—F-I-(R?) -96(?) K-E-D—G—I—?-S-T—D—I—L-K —306(?) A—Q-D—R-Y-Y-SS—W—E—?-A—S—V-P-?-? 78 71-85 (G?)-G—E-V-L—S—H—S—L—L-L-(L?)—H—K—K -47 _ procedure outline). However, the blotted 40,000 dalton subunit was fragmented with the proteolytic enzyme, Staphylococcus aureus VB protease (Endoproteinase Glu-C, IN). were digested for 6 hours at 37°C with 20 ng of V8.

Boehringer Mannheim, Indianapolis, Membrane pieces peptides were extracted with 88% (vol/vol) formic acid and separated on a Phase Separations column (2 X 150 mm, C8 83, Queensferry, England, UK) (Figure 16). Peptides were eluted with a linear gradient of acetonitrile. The peaks which were sequenced are numbered according to their fraction number. The amino acid sequence of these peptides is shown in Table 6.

TABLE 6 V8 (Glu—C) 40kDa peptides Off PVDF N—termina1 sequence fraction no. residue no. -3 I-W—E -12 L-K-K-D—V—Y—V-V—E -22 L—D—W—Y-P-D-A-P-G-E -59 V—L—G—S—G-K—T—L—T-I—Q—V-K-(E?) Three major peaks of peptide (fraction nos. 47. 54 and ) containing four peptides were sequenced. All four peptides were from the amino—termina1 region of the 40 kDa subunit indicating that the N-terminus of the protein is most susceptible to V8-digestion.

Figure 17 summarizes the protein structural determination of the 40,000 dalton subunit of CLMF. ._48 ..

Direct determination of the amino-terminal sequence of the ,000 dalton subunit of CLMF SDS—PAGE analysis of the Mono Q fraction 39 (see Fig. 3) under reducing (in the presence of B-mercaptoethanol) and non—reducing (in the absence of B—mercaptoethanol) conditions (Fig. 18) demonstrated that the 40,000 dalton .000 dalton CLMF subunit (i.e. unassociated with the 35,000 dalton subunit). molecular weight "contaminant" is "free" The evidence which points to this deduction is that without reduction (lane B. Fig. 18) mainly 75.000 dalton CLMF is present with some 40,000 dalton protein. After reduction (lane C, Fig. 18), the 75,000 dalton CLMF is gone yielding the 35,000 dalton subunit and an enriched 40,000 dalton band.

Fraction 39 of the previous Mono Q chromatography was reduced in 5% B-mercaptoethanol in the presence of 4 M urea and heated for 5 minute at 95°C. The sample was pumped onto a Vydac C-18 column using an enrichment technique and the column was then washed with 5 ml of 0.1% trifluoroacetic acid. Elution of the proteins was accomplished with a gradient of O—70% acetonitrile over 5 hrs in 0.1% trifluoroacetic acid (Fig. 19). Protein purity of the fractions which were fluorescamine positive was assessed by SDS—PAGE under non—reducing conditions using a 10% slab gel. ). ,000 molecular weight which was greater than 95% pure.

The gel was silver stained to visualize protein (Fig.

Fractions 112 through 117 revealed a diffuse band at The 40,000 dalton subunit and any other proteins present in fraction 39 remained bound to the C-18 column. These proteins (including the 40,000 dalton subunit) were finally eluted with a solution of 42% formic acid/40% l—propanol.

The ability to prepare homogeneous 35,000 subunit allowed for the determination of the amino acid composition and partial sequence analysis of the lower molecular weight subunit of the CLMF protein. Approximately 1 ug of 35 kDa . _49 _ subunit was subjected to hydrolysis, and its amino acid Amino—terminal sequence determination was attempted by automated Edman degradation on 100 pmol of the C-18 purified kDa subunit. Data from the first 20 cycles confirmed the sequence obtained by deduction as described above.

Furthermore, the second amino acid was obtained in addition to amino acids 21 through 26. These results may be summarized as follows: Cycle 1 2 3 4 5 6 7 Amino -------- -- Acid 7 N L P V A T Cycle 8 9 10 11 12 13 14 Amino ----------- -- Acid P D P G M F P Cycle 15 16 l7 18 19 20 21 Amino __‘_ Acid 9 L H H s Q N Cycle 22 23 24 25 26 Amino Acid L L R A V Therefore, the amino terminal sequence of the 35,000 dalton subunit can be summarized as follows: ,000 dalton subunit: ‘ 1o NH2—?—Asn—Leu—Pro—Val—Ala—Thr—Pro—Asp—Pro—Gly—Met— 26 Phe—Pro—?—Leu—His-His—Ser—Gln—Asn—Leu—Leu-Arg—Ala—Va1 where ? represents an undetermined residue.

Determination of the sequence of a tryptic fragment of CLMF Mono Q fractions 36 and 37 from the initial purification of CLMF were pooled (approximately 100 pmol/1.7 ml). A ul sample was removed and the volume of the rest was reduced to 200 ul under a stream of helium. One hundred microliters of 0.1 M ammonium bicarbonate was added.

Trypsin (Worthington Biochemical Corp., Freehold, NJ) cleavage was performed at a substrate—to-enzyme ratio of 2:1 (w/w) at 37°C for 20 hours. The resultant peptide fragments were reduced and carboxymethylated. This was accomplished by addition of 160 ul of 0.1 M Tris—HCl, pH 8.5/6 M guanidine—HCl. The volume was reduced to 200 ul under a stream of helium, and 4 ul of dithiothreitol (50 mg/ml) was added. The mixture was incubated at 37°C for 4 hrs.

After reductive cleavage of the disulfide bonds, [lqcjiodoacetic acid (4 umol) was added and the resulting solution was incubated in the dark at room temperature for 10 minutes.

The resultant peptide reversed—phase HPLC (Fig. column (2.6 x 50 mm, YMC, fragments were isolated by ) on an s—5 120 fingstrom ons Inc., Morris Plains. NJ).

Peptides were eluted with a l—propanol gradient in 0.9 M acetic acid, pH adjusted to 4.0 with pyridine. The amino acid sequence of the peptide found in fraction 46 was found to be: Asp—Ile—Ile-Lys-Pro-Asp—Pro—Pro—Lys (determined by automated Edman degradation).

Determination of internal amino acid sequence segments of CLMF CLMF was purified as previously described. Approximately 80 ug of protein was precipitated with 10% trichloroacetic acid. The precipitate was dissolved in 70% (v/v) aqueous formic acid at room temperature. An approximately 50-fold molar excess over methionine residues of cyanogen bromide (CNBr) in a small volume of 70% formic acid was added, with stirring, and the mixture was incubated in the dark under oxygen—free helium at room temperature for 48 hrs. The mixture was diluted with 15 volumes of water, divided into two equal portions and dried under a stream of helium. For complete removal of the acid and by-products, the drying was repeated after further addition of water. _ 52- one of the portions (approx. 40 ug) of fragmented CLMF was dissolved with 50 ul Laemmli sample buffer [Laemmli, Nature ggzz 680-685 (1970)] containing 4% B—mercaptoethanol followed by exposure to 105° C for 6 minutes. The sample was loaded into 3 wells of a minigel (1.0 mm thick) containing 17.5% polyacrylamide and electrophoresed according to Laemmli (supra).

After electrophoresis, the gels were soaked in transfer buffer (10 mM 3-cyc1ohexy1amino—l—propanesulfonic acid, 10% pH 11.0) for 30 min. polyvinylidene difluoride (PVDF) membrane (Immobilon: methanol, During this time, a Millipore: Bedford, MA) was rinsed with 100% methanol and The gel. of PVDF membrane and sandwiched with blotting paper, was stored in transfer buffer. backed with two sheets into a blotting apparatus and electroeluted for 30 The PVDF assembled min. at 0.5 Amps in transfer buffer. membrane was deionized H20 for 5 min and stained with 0.1% Blue R-250 in 50% methanol for 5 min, washed in Coomassie and then destained in 50% methanol, 10% acetic acid for 5-10 min at room temperature. A number of smeared bands were observed (see Fig. 22B).

Five regions of the membrane were excised across the These A summary of the sequences obtained A. three last lanes containing the CLMF CNBr digest. regions were sequenced. from the CNBr fragments of CLMF is shown on Fig.

The second portion (approx. 40 pg) of fragment CLMF 400-500 pl 88% formic acid .1 M Tris/HC1, 0.5 M NaOH, pH The sample was pH adjusted to pH 4.0 with formic was dissolved in approx. containing 6 M guanidine HC1. 8.0. acid. The peptide fragments were isolated by reversed—phase HPLC (Fig. 23) on a Vydac C4 column (4.6 x 20 mm, The CA).

Sep/a/ra/tions Group, Hesperia, Peptides were eluted with a 4.5 hours linear gradient of acetonitrile in 0.1% TFA. One of these peaks was sequenced and the amino acid sequence of this peptide was: Fraction No. N-Terminal Sequence V-D—A-VI—K—L-K—Y—E.‘—?—Y—T-S—(S?) —F—F—I—R-D-I-I—K—P— (Starts at residue number 190 of 40 kDa subunit) It is assumed or known that the above sequence is preceded by a Met residue. The residue marked with "?" represents a "best—guessed" residue.

PURIFICATION OF CLMF AND THE 40,000 DALTON SUBUNIT THEREOF USING AFFINITY CHROMATOGRAPHY An affinity chromatography resin was prepared by covalently attaching the monoclonal antibody 7B2, the preparation of which is described below, to activated agarose. Similarly, the below outlined purification could also be carried out by covalently coupling the antibody to silica or thin microporous membranes. The activated agarose was prepared as follows: . 100 ml Sepharose CL—6B was washed three times with 100 ml H20. 2. 100 ml of 1% sodium meta—periodate in H20 was added to the resin and the suspension shaken at room temperature for 60 min.

. The resin was washed with cold H thoroughly.

The covalent attachment of 7B2 to the activated agarose was carried out as follows: . 9 ml of the activated agarose prepared as described above was suspended in 7 ml of 7B2 (approx. 3.9 mg/ml) in phosphate buffered saline, pH 7.4. . 50.2 mg of cyanoborohydride was added to the gel suspension which was shaken overnight at 4°C.

. The gel suspension was filtered and added to 7 ml of 1.0 M ethanolamine, pH 7.0 containing 50.2 mg of cyanoborohydride. one milliliter of the above described resin (approx. 2.6 mg lgG/ml gel) was packed in a column and washed extensively with phosphate buffered saline. Fractions from the Mono Q chromatography containing the 75 kDa CLMF protein and additional major contaminating proteins were pooled (approx. 3.5 x 106 U TGF activity) and dialyzed extensively against PBS. This preparation was applied to the 7B2—Sepharose column at a rate of 5 ml/hr at room temperature. The column was washed with phosphate buffered saline (pH 7.4) until baseline absorbance monitoring at 280 nm was obtained. Adsorbed proteins were then eluted with 0.2 N acetic acid, 0.15 M NaCl. at approx. pH 3. Aliquots of fractions were assayed for TGF activity. Approximately 76% of the starting activity was recovered in the acid eluate.

Protein purity was assessed without reduction by sDS—PAGE [Laemmli, Nature ggzz 680-685 (l970)] using a 10% slab gel. Gels were silver stained [Morrissey. Anal.

Biochem. ;;1:307—3lO (l9Bl)] to visualize protein. The acid eluant contained pure CLMF and the "free" unassociated 40 kDa subunit of CLMF (Fig. 24).

DETERMINATION OF THE PI OF CLMF Thirty microliters of the pooled Mono Q fractions 36 and 37 (see Fig. 3) were spotted onto a precast ampholine PAGplate gel. pH 3.5-9.5 (Pharmacia LKB Biotechnology) to determine the pl of CLMF. Based on pl standard markers, a major band was observed at pl 4.8 and a minor band at pl .2.

Based on pH determination, the pl of these bands are .2 and 4.6 respectively.

BIOLOGIC ACTIVITIES OF PURIFIED CLMF Purified CLMF stimulated the proliferation of human PHA—activated lymphoblasts in the T cell growth factor assay (Table 8). The T cell growth factor activity of the purified CLMF recovered from the Mono Q column was compared to that of a standard preparation of human lymphokines in five separate experiments, and the specific activity of the purified CLMF was found to be 8.5 1 0.9 X 107 protein. units/mg In one experiment in which purified CLMF obtained from diphenyl HPLC was compared to the standard lymphokine preparation in the TGF assay, a specific activity of 5.2 X 107 units/mg protein was observed. when suboptimal concentrations of purified CLMF and human rIL-2 were tested in combination in the TGF assay. additive proliferation was observed (Table 8), up to the maximum proliferation caused by rIL-2 could be that the alone. However, proliferation caused by rIL-2 distinguished from proliferation due to CLMF in former was totally inhibited in the presence of a neutralizing goat anti—human IL-2 antiserum but the latter was not affected.

The ability of purified CLMF to activate cytotoxic effector cells was examined both in a 4-day LAK cell induction assay and in an overnight NK cell activation assay. In the LCI assay, purified CLMF at concentrations as high as 800 units/ml had little activity in the absence of IL-2 (Table 9). concentrations of human rIL-2 in causing LAK cell induction However, CLMF synergized with low in as much as the lytic activity generated in the presence of both cytokines was significantly greater than the sum of the lytic activities observed in cultures containing either cytokine alone (Table 9). In the presence of rIL-2, purified CLMF was active at concentrations as low as 3 units/ml. -56 _ TABLE 8 Purified Human CLMF Stimulates the Proliferation of Human PHA-Activated Lymphoblasts gxtokine Added: 3H-Thymidine Incorporated by Human CLMFC Human rIL-2 PHA-Activated Lymphoblasts Expt. (u/ml) (u/m1) (mean cpm + 1 S.E.M.) la 0 0 10,607 1 596 500 0 70.058 1 1,630 100 0 60,377 1 1,927 0 36,018 1 321 4 0 24,996 1 669 0.8 0 17,765 1 790 213 o 0 9,976 1 374 200 0 60,980 1 1,713 50 0 38,817 1 884 12.5 a 18,885 1 2,132 3.1 0 13,648 1 731 0 16 80,041 1 5,835 0 4 21.282 1 1,145 O 1 11,241 1 898 S0 4 62.050 1 2,408 12.5 4 40,628 1 2,196 3.1 4 31,144 1 3.754 All cultures in experiment 1 contained goat anti-human rIL—2. b None of the cultures in experiment 2 contained goat anti-human rIL-2.

C Purified human CLMF from Mono Q FPLC. - 57 ..

TABLE 9 Purified Human CLMF Synergizes with Human rIL-2 in the Generation of Lymphokine-Activated Killer (LAX) Cells in 4-Day Cultures Qytokine Added: % Specific 51Cr Releasea from: Human CLM1-"b Human rIL-2 11ju/ml) (u/ml) K562 Raji 0 0 3 1 1.7 -1 1 0.5 800 0 7 1 0.3 1 1 0.1 200 0 5 1 1.1 1 1 0.4 50 0 4 1 3.0 0 1 0.9 0 5 10 1 2.4 2 1 0.8 300 5 41 1 4.0 11 1 0.0 200 5 42 1 1.9 11 1 0.3 50 5 36 1 2.7 9 1 0.8 12.5 5 28 1 2.1 7 1 0.7 3.1 5 19 1 0.8 5 1 0.3 0-8 5 14 1 1.2 3 1 0.8 a Values represent the means 1 1 S.E.M. of quadruplicate determinations. The spontaneous Cr release values for K562 and Raji were 16% and 14%, respectively.

Purified human CLMF from Mono Q FPLC.

In contrast to the results in the 4-day LAK induction assay, purified CLMF was effective by itself in activating human NK cells in an overnight assay (Table 10). In this assay, CLMF was active at concentrations as low as 1.6 units/ml. when CLMF was tested in combination with human rIL-2, the two cytokines together had, at best, additive effects in enhancing NK activity (Table 10).

TABLE 10 Purified Human CLMF Causes Activation of Natural Killer (NK) Cells in Overnight Cultures % Specific 51Cr Releasea from Raji Cells at Effector/Target Ratio=: Cytokine Added: Human CLMFb Human rII.-2 (u/ml) (u/m1; 20/1 5/1 0 0 10 1 0.6 5 1 0.4 40 0 31 1 0.4 14 1 0.5 8 0 23 1 2.1 12 1 0.4 1.6 0 15 1 0.3 10 1 0-6 0.3 0 12 1 1.2 9 1 0.2 0 1 13 1 0.4 6 1 0.5 40 1 33 1 2.0 17 1 0.5 s 1 26 1 0.0 13 1 1.9 1.6 1 19 1 1.1 11 + 2.1 0.3 1 16 1 1.0 10 1 1.5 0 5 20 1 1.3 13 1 0.6 40 5 23 1 2.0 12 1 1.5 8 5 29 1 1.1 16 1 0.7 1.6 5 27 1 1.2 13 1 0.0 0.3 5 24 1 1.0 13 1 1.2 0 25 33 1 1.4 19 1 0.7 Each value represents the mean 1 1 S.E.M. of quadruplicate determinations. The spontaneous 5lCr release was 9%. b Purified human CLMF from Mono Q FPLC. _ 59 _ In addition to its ability to enhance the lytic activity of nonspecific NK/LAK cells, CLMF also facilitated specific human cytolytic T lymphocyte (CTL) responses in vitro. CLMF increased the specific allogeneic CTL response to weakly immunogenic, gamma—irradiated HTl44 melanoma cells (Table 11). also facilitated specific allogeneic human CTL responses to In combination with a low concentration of rIL—2, CLMF uv-irradiated HT144 melanoma cells, which did not elicit any detectable CTL response in the absence of added cytokines (Table 11). The specificity of the cytolytic effector cells generated in these studies was demonstrated by their ability 51Cr—labeled HT 144 melanoma cells but little or no lysis of K562 cells.

LAK cells which were generated in the same experiments by to cause substantial lysis of In contrast, incubating low density lymphocytes with rIL-2 in the absence of hydrocortisone lysed the K562 cells to a much greater extent than HTI44 melanoma cells. For further discussion of the specificity and identity of the cytolytic effector cells generated in assays such a those shown in Table ll [see Gately et al., J. Immunol. 136: 1274-1282 (1986)). .>Ho>_uuvawou .oouamcmuum1aEEma no =eucm©auuw1>: swan Cm: :om:3 maawu MEOEQHQE Vv~H= u:mmo.am. >H: ccm >:H:u .muomu:umum ammo a .c:m: uwsuo vzu :0 uwuowusumum xm~ Haoouma IHV can awn mzu cwwauon Ucm wan czm awn mzu cmmzumn wmomuuoucm oz» Eouu nmumw>pa: mwuxuozmexfi xumwcmn uwzom owcmmucov ~ + H comuomuu unouuom mm0uo£3 .muwam~ aaouumm ram ccn Imv mzu cmoxumn moouumucm w:u Eouu ©wuo>oomu mwuxoosmexa xumwcwc £DM£ Uwcmmucou v comuowum HAOUHULQ .xHa:uuoumwmu \~ U:m H mucoemuomxv cm IFN can awn mm: mom: you ccm .>~o>muommmwu .m one u wgcoeauomxo cm ran ccm Imw um: maamo vvah: Eouu wmmwmou uuum msoucnucoaw ozh .oanmu vzu cm :zo=m mum mama omozu ecu .>nmnm umuam 0:» cu Umusumflcs vocvo mum: mouxoosmsmm cmsz aaco coon mm: mmmaa ucmomumcmmu .~ ucmemuomxo cu .H«v xamumsmxoummn uo omumu uooumu uouauosmexm m ou ocqccomwwuuou .>nmna umuxuouxo wzu cm mwuauozmexn no :omu:amc m_~ oz» mcmaa omcmmuno ouwz c3o:m mama ozu ~a ucwsmuwmxw :H .:omu:~mc mum um vsm cousamccs xum>quom omuxm you cmammmm ccm .:uH as ~.H cm cvczomusmou .co:wo: .©o~oom mus: mwuzuaao wumuqumac uo mucwucou mzu mu=QEmuwmKfl Luca :_m HMS. flu? nut. «HQ 3 ~+_ m u ... e u an ~ « T m u on S J: .n.-S _.

A u m H H S H u T n a 2 >2 m-S _. _ I I .| >5 ~ + e- ... + Z a ; 3 n u S S m... H: .13 _.

W mu a mug an m Hum m.» 2.5. T3 _. _ H u T m u 2 d u 7 w u .. 3 >=..:_ T3 ..

H u ~ ~ u w: a u 7 ~ u o >::_ T3 v H» .. Hum. a n7 an 7 S m... .13 _. «um ~«~- H»..- mum T3 .. 3...: ..Z.:_ 35. 33: 75:4 S..__ N .uaxu a .Maxm mxau wnqnu mfiocnuwl ozonmuuououux: uwuxoonmeaq .|mIEmMH umuvuox uumw umumomaw I nwmpzuazu no mucoucou O=HH> zn magma <:o:<4mz u~m=mooqq< OH wmmzommmm Hawqoawu z<flD= uHm~umgw.wmuz<:zm uzgo z<::: omuuamsm ad uam _ 61..

Our results demonstrate that purified human CLMF by itself caused proliferation of activated human T lymphocytes. enhanced the cytolytic activity of human NK cells, and augmented human CTL responses. These activities of CLMF, which are similar to those of IL-2, suggest that CLMF, like IL-2. should have immunoenhancing and antitumor effects when used as a single therapeutic agent in vivo.

Clearly, CLMF may also have utility in stimulating the growth in vitro of NK/LAK cells and of activated T cells, such as may be derived from tumor infiltrating lymphocytes Immunol. lgg: 3714-3725 (l989)]. addition, purified CLMF synergized with low concentrations [Topalian et al., J. In of rIL—2 in causing the generation of human LAK cells in culture and acted additively or synergistically with rIL—2 These results suggest that the use of CLMF in combination with in facilitating specific CTL responses in vitro. rIL—2 might constitute a more optimal antitumor therapy.

CLONING OF A CDNA CODING FOR THE 40 kDa SUBUNIT OF HUMAN CLMT ) Cell culture and isolation of polyA+ RNA NC 37 cells (subclone 98) were grown in roller bottles as described above and induced with PMA and calcium ionophore for 15.5 hours. The cells were harvested, resulting in a frozen cell pellet of 1.11 grams comprising about 5.25 x 108 cells. A portion of the culture was continued for 3 days, at which point the bioassay titer for CLMF activity read 2200 harvested for isolation of RNA had indeed produced the CLMF units/ml, indicating that the cells activity. Total RNA was isolated from the frozen cells by standard procedures and The yield of polyA+ RNA was 2.5% polyA+ RNA was obtained by affinity chromatography. (w/w) relative to the total amount of RNA input.

) Establishment of a CDNA library ug of the above polyA+ RNA were reverse transcribed into cDNA using 150 ng random hexamers as primers. A library in lambda gtlo was established. and 1.5 x 10 clones were amplified for the screening.

) Use of PCR to Generate a DNA Probe Specific for the 40 kDa CLMF Subunit CDNA The partial N-terminal sequence of the purified 40 kDa protein is IWELKKDVYVVELDWYPDAP.... mixed primer PCR were designed and synthesized by standard Two primers for use in procedures. The forward primer was designed as the coding strand corresponding to amino acids ELKKD in the above sequence. containing all possible codons for that sequence and having an extension at its 5' end including an EcoRl The ctc gaa ttc gaa/g site and three additional bases adding stability. sequence of the forward primer is thus 5' a 23mer with 64 different c/ttn aaa/g aaa/g ga. i.e. sequences. The reverse primer was designed in the same manner. to represent the antisense strand corresponding to the amino acid sequence YPDAP in the partial N—terminal 40 kDa sequence. The reverse primer thus has the sequence 5' ctc gaa ttc ngg ngc a/gtc ngg a/gta and is a 24 mer containing 256 different sequences. The symbol n stands for any one of the four possible bases a,g,c or t. The primers thus define an amplicon of 72 basepairs in length. After cutting with EcoRl for generating cohesive ends for subcloning, the amplicon size drops to 64 basepairs.

Single-stranded CDNA was generated for use in the PCR as described in section 2 above, using polyA+ RNA from induced and, as a control, uninduced cells. 40 ng of either one of those cDNAs were amplified with forward and reverse primers in 100 pl of 10 mM Tris-HCl pH 8.3/50 mM KC1/1.5 mM Mgclz/0.01 % gelatine/200 uM each of the four nucleotides/10 units Tag-polymerase/250 pmoles of each initial primer. The PCR parameters were as follows: denaturation was at 95°C for 7 minutes. Low stringency _ 53 _ annealing was performed by cooling to 37°C over 2 minutes, incubating 2 minutes at 37°C, heating to 72°C over 2.5 minutes, extending at 72°C for 1.5 minutes, heating to 95°C over 1 minute and denaturing at 95°C for 1 minute: this low stringency annealing cycle was repeated once. Afterwards, standard cycles were run as follows: 95°C for 1 minute, 72° was performed at 72°C for 10 minutes.

°C for 2 minutes, C for 2 minutes. A final extension % of the total samples were run on a 4% agarose gel, stained and analyzed.

The amplicon of the expected size was only detectable in the sample where induced CDNA had been amplified. The remainder of the sample was extracted with phenol, concentrated by precipitation with ethanol and redissolved in 42 ul of water. The sample was digested with 60 units of the restriction enzyme EcoRl in 50 ul at 37°C for 2 hours.

The sample was subsequently run on a 6% polyacrylamide gel and the 64 bp amplicon was cut out of the gel and eluted by standard procedures. The DNA amplicon was subcloned into the EcoRl site of the bluescript SK+ plasmid by standard La Jolla, CA). the E.co1i strain DH5 (obtainable procedures (Stratagene, Colonies obtained from the transformation of from ATCC) were picked and analyzed for the presence of the 64 bp insert (other E.coli strain compatible with bluescript SK+ plasmids may be used). Two positive candidates were sequenced to determine the sequence of the cloned amplicon.

It is clear from this analysis that the correct fragment was amplified, since the deduced amino acid sequence matches exactly the partial amino terminal amino acid sequence from the purified 40 kDa protein. This information was subsequently used to design a 54 bp long oligonucleotide probe that could be used for screening of the CDNA library.

Two oligos were designed, with the following sequence: 5' gag cta aag aaa gat gtt tat gtc gta gaa ttc gat and 5' agg ggc atc cgg ata cca atc caa ttc tac gac ata. These two oligos are partially complementary to form the following SCILICCUIES _ 54 _ ' gagctaaagaaagatgtttatgtcgtagaattggat 3' ' atacagcatcttaacctaaccataggcctacgggga 5' Such a structure can be labelled by using Klenow fragment and labelled nucleotides such that a high specific activity probe results for the screening of CDNA libraries.

) Screening of CDNA Libraries A total of 3 x 105 clones from the amplified library were screened on 6 duplicate filters under the following conditions: 50 ml of 5 x SSC/10 x Denhardts/100 ug/ml denatured calf thymus DNA/20% formamide/0.1% SDS/1.5 x 106 cpm of labelled 54 mer at 37°C for 16 hours. The filters were subsequently washed in 2 x SSC at 42°C for 30 minutes, dried and exposed to X-ray film. After overnight exposure with an intensifying screen. 16 possible positives were picked and further analyzed by a second screening round. 10 rehybridizing phage were isolated and their DNA prepared. 8 of those 10 isolates looked identical, upon BcoRl cutting releasing two fragments of 0.8 kb and 0.6 kb length. indicating a possible internal EcoRl site. Upon blotting and hybridization with the screening probe, only the 0.6 kb fragment showed hybridization. The two fragments were subcloned separately into the EcoR1 site of the bluescript SK+ plasmid as described above and were completely sequenced. This analysis showed that both fragments align in one contiguous CDNA of about 1.4 kb in length with a naturally occurring internal EcoRl site, since both fragments upon translation showed the presence of reading frames coding for tryptic peptides that had actually been isolated from purified 40 kDa protein. The complete sequence of the 40 kDa subunit as deduced from the CDNA is shown in Figure 25. The CDNA codes for one open reading frame of 328 amino acids. The protein starts with the initiating Met, followed by another 21 amino acids that make - 55 _ up a classical hydrophobic signal peptide. of mature purified 40 kDa subunit, The N—terminus IWELKKD....

The mature protein i.e. immediately after the signal sequence. follows thus consists of 306 amino acids. The deduced protein sequence contains 4 possible N—linked glycosylation sites, two of which are present in isolated and sequenced tryptic peptides. one of these two sites is used in vivo for the attachment of a carbohydrate side chain. The calculated molecular weight of the mature unglycosylated protein is 34699. the pl is 5.24. length and is detectable in a northern blot in steady state The corresponding mRNA is 2.4 kb in RNA only from induced cells.

CLONING OF A CDNA CODING FOR THE 35 kDa SUBUNIT OF HUMAN CLMF Cell culture, isolation of mRNA and establishment of a CDNA library were carried out as described earlier for the cloning of the 40 kDa subunit. .

Use of mixed primer PCR to generate a DNA probe specific for the 35 KDa subunit CDNA The partial N—terminal sequence of the purified 35 kDa subunit is ?NLPVATPDPGMFP?LHHSQNLLRAV.... Two primers for use in mixed primer PCR were generated by standard procedures. The forward primer was designed as the coding strand corresponding to the amino acids DPGMF in the above sequence, containing all possible codons for that sequence and having an extension at its 5' end including an EcoRl site and three additional bases adding stability. The CTC GAA TTC a 23 mer with 32 different sequence of this forward primer was thus 5' GAT/C CCN GGN ATG TT —3‘. sequences . i.e.

The reverse primer was designed in the same manner, to represent the antisense strand corresponding to the amino acids NLLRA in the partial N—terminal sequence.

The reverse primer had the sequence 5' CTC GAA TTC NGC NCG/T NAA/G NAA/G A/GTT. i.e a 24mer with 4096 different _ 55 _ sequences. In both primer sequences, N stands for all 4 bases.The two primers thus defined an amplicon 69 bases long. After cutting with EcoRl for generating cohesive ends for subcloning, the amplicon size drops to 61 bases. About 3 ug of human genomic DNA were amplified with forward and reverse primers in 50 ul of 10 mM Tris—HCl pH 8.3 / 50 mM KCl / 1.5 mM MgCl2 / 0.01% gelatine / 200 uM each of the four nucleotides / 2.5 units of Taq polymerase / 64 pmoles of forward and 2048 pmoles of reverse primer (to compensate for the greatly differing complexities of the two primers).

The PCR cycling parameters were as follows: initial denaturation was at 95°C for 7 minutes. Low stringency annealing was performed by cooling to 37°C over 2 minutes, incubating at 37°C for 2 minutes, heating to 72°C over 2.5 minutes, extending at 72°C for 1.5 minutes, heating to 95°C over 1 minute and denaturing at 95°C for 1 minute; this low stringency annealing cycle was repeated once. Afterwards, 40 standard cycles were run as follows: 95°C for 1 minute, 55°C for 2 minutes and 72°C for 3 minutes. A final extension was performed at 72°C for 10 minutes. About 20% of the samples were run on a 6% polyacrylamide gel and an amplicon of the expected size was detected after staining the gel with ethidium bromide. The remainder of the sample was extracted with phenol, concentrated by precipitation with ethanol and redissolved in 17 ul of water. The sample was digested with 20 units of EcoRl enzyme in 20 ul for 60 minutes at 37°C The sample was subsequently fractionated on an 8% polyacrylamide gel and the 61 basepair amplicon was cut out of the gel and eluted by standard procedures. The DNA amplicon was subcloned into the EcoR1 site of the Bluescript plasmid SK+ (Stratagene, La Jolla, CA) by standard procedures. Colonies obtained from the transformation of the E.coli strain DH5 were analyzed for the presence of the 61 basepair insert. Two candidates were sequenced to determine the sequence of the subcloned amplicon. one of the two clones contained the correct sequence, since translation of that sequence resulted in the amino acid sequence expected _ 57 _ from the purified protein. Based on this information, two synthetic oligonucelotides were designed, with the following sequences: 'gatccgggaatgttcccatgccttcaccactccc 3' 3'gtacggaagtggtgagggttttggaggatgcccga 5' Such a structure can be labelled using radiolabelled nucleotides with the Klenow fragment to very high specific activities for library screening.

Screening of a CDNA library: A total of 106 library were screened on 40 duplicate filters with the above 400 ml of S X SSC / % formamide / 10 X Denhardts / 100 ug/ml denatured calf thymus DNA / 0.1% SDS / 3.8 x 107 37°C overnight. The filters were subsequently washed in 2 X clones from the amplified 16 hours probe under the following conditions: cpm labelled probe at SSC at 40°C and exposed to X-ray film with a screen overnight. Six potential positives were picked from this first round of screening and analyzed by a second round of as above. plaque hybridization, One clone was picked for the final analysis. Upon preparing phage DNA, the clone was found to contain two EcoRl fragments of about 0.8 Kb and 0.3 Kb in size. The two fragments were subcloned separately This analysis showed that the two fragments align into one into the Bluescript SK+ plasmid and sequenced. contigous sequence of about l.1 kb total length with a naturally occurring internal EcoRl site. The complete sequence of the CDNA and the deduced amino acid sequence for the 35 KDa CLMF subunit are shown in Figure 26. The CDNA codes for an open reading frame of 219 amino acids, starting with the initiating Met at position 1. The following 21 amino acids constitute a classical hydrophobic signal the sequence. Immediately following the signal peptide, _ 68 - N—terminus of the mature 35 kDa protein starts with the sequence RNLPVAT.... Purified 35 kDa protein had yielded the sequence ?NLPVAT.... The mature 35 kDa protein thus consists of 197 amino acids, containing three possible The calculated molecular weight of mature unglycosylated protein is 22513. and the pl is 6.09.

N—linked glycosylation sites and 7 Cys—residues.

The corresponding mRNA is 1.4 kb in length and is only detectable in RNA from cells that had been induced for CLMF for at least 6 hours.

EXPRESSION OF BIOLOGICALLY ACTIVE RECOMBINANT CLMF IN COS CELLS The two subunits for CLMF were engineered for expression in mammalian cells as follows: kDa subunit The two EcoRl fragments constituting the full length CDNA for the 40 kDa CLMF subunit were ligated to an expression vector similar to pBC12 [see B. Cullen, Meth. 684-703 (l987)], expression is driven off the SV4O early promoter/enhancer.

Enzymology lggz except that the CDNA Clones containing the two inserts in the proper orientation to each other were selected by colony hybridization with a synthetic oligonucleotide that spans the internal EcoRl site in the 40 kDa CDNA. This oligonucleotide has the following CTG AAG CCA TTA AAG AAT TCT CGG CAG GTG 3‘.

It was labelled by kinasing using standard procedures. sequence: 5' Clones were subsequently analyzed for proper orientation of insert to vector by the polymerase chain reaction procedure, using as forward primer a primer specific for sequences in the SVQO early promoter and as reverse primer an oligonucleotide corresponding to the 40 kDa CDNA sequence 851-868. will give a PCR amplicon of 885 bp. Eight out of 20 clones positions no. Clones with the correct orientation tested gave the predicted fragment, and one was chosen for further study. kDa subunit The full length cDNA for the 35 kDa subunit was amplified out of the original lambda phage by PCR. using primers situated to the left and right of the EcoRl site in lambda gtlo (primers were New England Biolab Articles No. 1231 and 1232). The resulting PCR amplicon was blunt—end ligated into the EcoRV site of the bluescript plasmid SK+ and the DNA propagated. DNA sequencing showed that the orientation of the cDNA insert within the plasmid was such that the end of the cDNA corresponding to the 5' end of the mRNA was close to the Clal site in the polylinker. The insert was thus released by cutting with Clal, filling out this end with T4 DNA polymerase and cutting secondarily with Not 1. The resulting fragment was gel-purified and subcloned into an expression plasmid based on the bluescript vector and containing the SV40 early promoter to drive expression of inserted cDNAs. The sites in the expression plasmid used were a blunt—ended Pstl site at the 5' end and a Not I site at the 3' end of the CDNA. One clone was chosen for further study after ascertaining its structure by PCR as above for the 40 kDa construct.

Expression of the two cDNAs in COS-cells with standard tissue—culture medium containing 1% Nutridoma hours after the transfection, the cells were fed instead of the fetal bovine serum and the supernatants were collected after 40 hours and filtered through a 0.45 n filter.

Supernatant fluids from cultures of CO8 cells which had been transfected with cDNA encoding the 35 kDa CLMF subunit or the 40 kDa CLMF subunit or with both cDNAs were tested for CLMF activity in the T cell growth factor assay (Table 13). As shown in the table, COS cells which had been transfected with only one of the subunit cDNAs did not release biologically active CLMF into the culture fluid.

However, COS cells which had been transfected with both BY comparing the amount of lymphoblast proliferation induced by the culture fluid from doubly transfected COS cells to the subunit cDNAs produced biologically active CLMT. amount of proliferation induced by purified NC—37—derived CLMT, the concentration of CLMF activity in the culture fluid was estimated to be 374 units/ml. Assuming a specific activity of 8 x 107 units/mg CLMF protein, this result suggests that the fluid from cultures of doubly transfected COS cells contained approximately 4.7 ng/ml of recombinant CLMF. .uqmm o ocoz sou“ mzqu co>q.on->m,u: ommuqgzma mud n o~v.v~ comusfiflu m~\H nmuooumzouu xooz awe N mon~vH =o_u:Hwc mxu Uouooumznuu xooz m- u n~a.»v comgsqmc m~\~ n can H a-.ec comuzqmc mxu w oan H oaa.n~ :oma:__c m~\H <:oo umcsasm mzqu max ov amoa H naa.qH comuaamw mxa osn H ovH.- zoflusamc mmxa Bop u ~mm.mH coflusafio mxfl «zoo uficsnsw mzqu max mm mzuaz cmuuouwcmuw maamu woo uo wousuaau Eouu uwzau gznummwmmmw me» u oao.m~ asxwumca o.~ mzau Hopsamz new H omH.an Aaxnumca m uzqu mmuzumz a~o~ H nao.om .a\nuq:= oe mzqu magnum: one u mvm.m~ dexmumcz com men H n=m.v~ 0:02 ~uI.muw ml!“ ado cmos. mcomwauuzoocou «madam ocwxoaxu wummflpozmesg oaum>muo<-<:m an omumuomuoocm w:«u~E»:H:=n wmuou woo cm oommupmxm mzqu ucmcwaeoooz no »u«>muo< pouomm nuzouo “Hem H ma mJm ANTI~CLMF HYBRIDOMAS AND ANTIBODIES Preparation, Characterization and Purification of anti-CLMF Hybridomas Lewis rats (Charles River Laboratories, Wilmington. MA) were initially immunized by the intraperitoneal route (i.p.) with partially purified CLMF mixed with an equal volume of Freund's complete adjuvant (Gibco).

The rats were injected i.p. with booster immunization of CLMF mixed with Freund's incomplete adjuvant (Gibco) according to the schedule in Table 14.

For preparation of activated spleen cells, one rat was injected i.v. with partially purified CLMF on two successive days starting 4 days prior to the cell fusion (Table 14).

Spleen cells were isolated from this rat and fused with N80 cells [Calfre et al.. Meth. Enzymol. Z;:3—46 (1981)) at a ratio of 1:1 (spleen cells % polyethylene glycol (PEG 4000, E. Merck).

NSO cells) with Other cells suitable as fusion partners in hybridoma cell fusion may be used instead of N80 cells. The fused cells were plated at a . 4 . . density of 5 x 10 cells/well/ml in 48 well plates in IMDM [Iscove et al.. J. mented with 15% fetal bovine serum (FBS), beta—mercaptoethanol (0.1 mM).

Exp. Med. l47:923—933 (1978)) supple- glutamine (2 mM), gentamycin (50 ug/ml), HBPES (10 mM) and 15% P388Dl cell supernatant (P388Dl cells are obtainable from ATCC). Hybridoma supernatants were screened for specific CLMF antibodies in 4 assays: l) immunoprecipitation of immunodepletion of CLMF bioactivity, I—labelled CLMF. 2) ) western blotting . . 125 . . with CLMP and 4) inhibition of I—CLMF binding to PHA—activated PBL blast cells.

Hybridoma cell lines secreting anti—CLMF antibodies were cloned by limiting dilution.

Antibodies were purified from large scale hybridoma cultures or ascites fluids by affinity chromatography on protein G bound to cross—1inked agarose according to the manufacturer's protocol (Gammabind G, GQIIGX , Gaithersburq. MD).

Date /28/89 /10/89 /3/89 /18/89 /7/89 /29/89 7/21/89 /2/89 /19/89 /20/89 /23/89 TABLE 14 Immunization Schedule: CLMF (log units/mg) Total Protein Spec. Activity Purity 1 14.1 112.12 gu/mg ) (_%1 4 5 1x10 0.1 pg 15 6.7x10 6.7 4 5 l.2x10 0.1 pg ? 6x10 0.6 lst bleed 6 2.2x10 2 pg 75 2.9xl0 2.9 2nd bleed 4 5 6.3x10 0.63 pg 83 7.5x10 0.75 6 l.2x105 1.2 pg 24 5x10 5.0 3rd bleed 6 . 2.lx1D (1.v.) .lxl0 (i.v.) Fusion -74 _ Isolation and Identification of Monoclonal Antibodies initially screened for CLMF—specific antibodies by I—1abelled CLMF. immunoprecipitation of The radioiodinated partially purified CLMF preparation contains predominantly the CLMP 75 KDa heterodimer, a small amount of the free CLMF 40 kDa subunit and two other proteins of approximately 92 kDa and 25 kDa (Fig. 28). 1251-labelled CLMF preparation retained CLMP bioactivity in the TGF assay, indicating that the labelling procedure did not significantly alter the configuration of the CLMF molecule. The CLMF immunized rat serum immunoprecipitated the 75 kDa heterodimer and the free 40 kDa subunit (Lanes 6 and 8. Fig. 28) whereas normal rat serum did not immunoprecipitate these radiolabelled proteins (Lanes 7 and 9, Fig. 28). immunoprecipitated the 75 KDa heterodimer and the free 40 Four individual monoclonal antibodies also kDa subunit (Fig. 28) but did not immunoprecipitate the 92 KDa or 25 kDa labelled proteins. The immunoprecipitation assay identified twenty hybridomas which secreted anti—CLMF antibodies (Table 15). All the antibodies immunoprecipitated the radiolabelled 75 kDa heterodimer and the free 40 kDa subunit as determined by SDS/PAGE and autoradiography (data shown for 4 representative antibodies in Fig. 28).

Table 15: Monoclonal Anti—CLMF Antibodies (40 kDa Subunit Specific) Inhibitory/Neutralizing B2 — ++ 95 + 2A3 — ++ 99 + 1B8 - +/- 60 ND lCl — ++ 81 ND 4A1 — + 98 + 4C8 ND ND 68 ND 4Dl — + 100 ND 6A2 - +/- 75 ND 7A1 +/- ++ 94 No 8A1 - + 99 ND 8A2 — ++ 83 ND 9C8 — + 62 ND 22E7 ++ ++ 91 ND Non-Inhibitory/Non—Neutralizing BE3 + ++ 35 ~ 9P5 + ++ 18 ND 4A6 - - 17 ND 6A3 ~ + 20 — BC4 - ++ 33 ND BE4 — ++ 1 ND 3983 ND ND 46 ND Control - : western blots: N.R. is non—reduced and Red. is reduced SDS/PAGE For the western blots. a CLMF sample containing 5% 75 kDa heterodimer and 95% free 40 kDa subunit were separated on 10% SDS/PAGE and western blots prepared as described in methods. The blots were scored as strongly positive (++), positive (+), weakly positive (+/-) and negative (—).

. CLMF receptor binding assay: An antibody was considered inhibitory if it would block more than 60% of radiolabelled CLMF binding to the PHA activated PBL blasts.

. Neutralization of CLMF bioactivity as assessed by the TGF assay: An antibody was considered neutralizing if it would block more more than 50% proliferation at 200 pg/ml. The results are presented as positive (+) or negative (—).

: ND: Not Determined _ 76 _ After initially identifying specific CLMF antibodies in the immunoprecipitation assay, the antibodies were tested for their ability to immunodeplete CLMF bioactivity as assessed by the TGF and LAK cell induction assays.

Increasing amounts of CLMF cause a dose dependent increase in the proliferation of PBL blasts in the TGF assay as measured by the incorporation of 3H—thymidine into the dividing blast cells (Fig. 29). Immunodepletion of CLMF activity by immobilized anti—CLMF antibodies occurs in a dose dependent manner (Fig. 29). Aliquots (0.4 and 0.1 ml) of hybridoma supernatant solution will completely deplete 50 and 200 units/ml of CLMF activity from the culture medium. 0.025 ml of supernatant solution will completely deplete 50 0.0062 ml of hybridoma supernatant shows even less depletion of SO and 200 units/ml of CLMF. An aliquot (0.4 ml) of an anti—lL—l receptor antibody supernatant solution shows no immunodepletion of CLMF bioactivity. units/ml but only approximately 50% of 200 units/ml.

Increasing amounts of CLMT also cause a dose dependent increase in the lysis of target cells by LAX cells as Slcr in the LAK cell induction The immobilized anti—CLMF antibodies measured by the release of microassay (Fig. 30). also deplete in a dose dependent manner the CLMF activity in the LAK cell induction assay (Fig. 30). These data confirm that the antibodies which immunoprecipitate the 75 kDa labelled protein from the radiolabelled partially purified CLMF preparation are specific for CLMF. The data also demonstrate that the radiolabelled 75 kDa protein is the protein responsible for CLMF bioactivity.

Methods for Assay of Monoclonal Antibodies . . . . . 12 Purification of CLMF and Labelling of CLMF with 51.

CLMF was partially purified from cell supernatants harvested from human peripheral blood lymphocytes (PBLs) or NC—37 cells as described previously. Partially purified CLMF was _ 77 _ by a modification of the Iodogen labelled with method. Iodogen (Pierce Chemical Co.) was dissolved in chloroform at a concentration of 0.5 mg/ml and 0.1 ml aliquots were added to 12x75 borosilicate glass tubes. The chloroform was evaporated under a stream of nitrogen and the Iodogen was dried in the center of the bottom of the glass tube. The coated tubes were stored in a dessicator at room temperature (RT) under vacuum. For radiolabeling, 0.5-1.0 mci 5I—Na (Amersham) was added to an Iodogen coated tube containing 0.50 ml of Tris-Iodination Buffer (25 mM Tris—HCl pH 7.5. 0.4 M Nacl. at RT. The activated 12 1.5 ml tube containing 0.05—O.l ml CLMF (approximately 5 ug in 0.125 M NaCl, 20 mM Tris—HC1 pH 7.5) and the reaction mixture was further incubated for 8 minutes at RT. mM EDTA) and incubated for 4 minutes .

I solution was transferred to a At the end of the incubation, 0.05 ml of Iodogen Stop Buffer (10 mg/ml tyrosine, 10% glycerol in Dulbecco‘s phosphate buffered saline (PBS) pH 7.4) was added and reacted for 30 seconds. The mixture was then diluted with 1.0 ml Tris-Iodination Buffer and applied to a BioRad BioGel PlODG (BioRad Laboratories) desalting column for chromatography.

The column was eluted with Tris—Iodination Buffer and fractions (1 ml) containing the peak amounts of labelled protein were combined and diluted to 1 x 108 cpm/ml with The TCA precipitable radioactivity (10% trichloroacetic acid final .25% gelatin in Tris-lodination buffer. concentration) was typically in excess of 95% of the total radioactivity. The radiospecific activity ranged from 6000 cpm/fmol to 10,000 cpm/fmol.

Immunodepletion of CLMF. Hybridoma culture supernatants or purified monoclonal antibodies were tested for their Goat anti—rat M0) were washed three times with 10 ml of PBS (Gibco) supplemented with 1% bovine serum albumin (BSA) (Sigma) (PBS/BSA solution). ability to immunodeplete CLMF as follows.

IgG—agarose beads (Sigma Chemical Co., St. Louis, After washing. the beads were resuspended in -78..

PBS/BSA at a final concentration of 50% vol/vol. Aliquots (0.2 ml) of the bead suspension were added to 1.5 ml Eppendorf tubes, together with the indicated amounts of monoclonal antibodies or hybridoma supernatant solutions.

. The volume of each mixture was brought to 1.4 ml by the addition of hybridoma maintenance medium [Iscove's modification of Dulbecco's medium (IMDM) with 0.1% fetal bovine serum (FBS), 10% Nutridoma-SP (Boehringer—Mannheim), and 2 mM L—glutamine], and the mixtures were then incubated for 2 hours at room temperature on a hematology/chemistry mixer.

Following this incubation, the tubes were centrifuged in a Beckman microfuge 12 (1.5 minutes at The beads were again washed three times with PBS/BSA and then resuspended in 1 ml of setting 5), and the supernatants were discarded. tissue culture medium (TCM) containing 5% human AB serum and the indicated concentration of purified human CLMF. The tubes were subsequently incubated overnight at 4°C on the mixer. Following this, the beads were removed by centrifugation in the microfuge, and the resulting immunodepleted supernatant solutions were assayed for residual CLMF activity in the TGF assay or in the microassay for LAK cell induction.

Immunoprecipitation Assay. For the immunoprecipitation reaction, 0.05 to 0.5 ml of hybridoma supernatant. diluted antisera or purified IgG was added to a 1.5 ml microfuge tube containing 0.1 ml of a 50% suspension of goat-anti—rat IgG coupled to agarose (Sigma Chemical Co.). The assay volume was brought up to 0.5 ml with RIPA Buffer (50 mM NaPO4 pH 7.5, 150 mM NaCl, 1% Triton—X 100, 1% deoxycholic 0.1 % SDS, 1% BSA, and 5 mM EDTA) and the mixture was The beads were pelleted by centrifugation for 1 minute at 12,000 x g acid, incubated on a rotating mixer for 2 hours at RT. . . 125 and then resuspended in 1 ml RIPA Buffer containing I CLMF (l x 10 cpm). rotating mixer for 16 hours at 4°C.

The mixture was then incubated on a Following this incubation, the beads were pelleted by centrifugation and _ 79 - washed twice in RIPA without BSA.

The beads were then washed once with 0.125 M Tris—HCl pH 6.8-and 10% glycerol.

The I—CLMF bound to the solid phase antibodies was released by adding 10 ul of 2 X Sample Buffer (Laemmli, supra) with and without 5% B—mercaptoethanol and heating for 3 minutes at 95°C. The immunoprecipitated 125I—CLMF was analyzed by SDS—PAGE on a 10% or 12% polyacrylamide gel and visualized by autoradiography.

CLMF Receptor Binding Assay. The ability of hybridoma supernatant solutions. purified IgG or antisera to inhibit l25I—CLMF to PHA-activated human T lymphoblasts was measured as follows: the binding of 0.1 ml aliquots of serial dilutions of culture supernatants, purified IgG or antisera were mixed with 0.025 ml aliquots of Binding Buffer (RPMI—l640, 5% FBS, 25 mM HEPES pH 7.4) containing 5I—CLMF (1 x 105 cpm). The mixture was incubated on an orbital shaker for 1 hour at RT, then 0.025 ml of activated blasts (5 x 107 cells/ml) was added to each tube. The mixture was further incubated for 1 hour at RT.

Non-specific binding was determined by inclusion of 10 nM unlabelled CLMF in the assay. in duplicate or triplicate.

Incubations were carried out Cell bound radioactivity was separated from free 5I—CLMF by centrifugation of the assay contents through 0.1 ml of an oil mixture (1:2 mixture of Thomas Silicone Fluid 6428—Rl5: A.H. Thomas, Silicone Oil AR 200: Gallard-schlessinger) at 4°C for 90 seconds at 10,000 X g.

The tip containing the cell pellet was excised and cell bound radioactivity was determined in a gamma COUHIEI .

SDS Polyacrylamide Gel Electrophoresis (SDS/PAGE) and western Blotting. 25I—1abelled proteins and partially purified CLMF were treated with Laemmli sample buffer (2% SDS, 125 mM Tris—HCl, pH 6.8, 10% glycerol, 0.025% bromphenol blue) with and without 5% Immunoprecipitated B—mercaptoethanol, heated at 95°C for 3 minutes and separated by SDS/PAGE on 7.5% or l2% precast gels (BioRad Laboratories). 51_1abe11ed proteins. the gels were stained with 0.2% Coomassie For the immunoprecipitated Brilliant Blue in 25% isopropyl alcohol and 10% acetic acid, destained in 10% methanol and 10% acetic acid, dried and analyzed by autoradiography. For western blotting. the proteins separated by SDS/PAGE were transferred to nitrocellulose membrane (0.2 u) for 16 hours at 100 volts in 10 mM Tris—HC1 PH 8.3, 76.8 mM glycine, 20% methanol and 0.01% SDS. The nitrocellulose membrane was blocked for 1 hour at 37°C in 3% gelatin, Tris—HCl pH 7.5, 0.15 M Nacl and then probed with hybridoma supernatant solutions or purified antibody diluted in AB buffer (1% bovine serum albumin, 50 mM sodium phosphate pH 6.5. 0.5 M Nacl, 0.05% Tween 20) for 16 hours at 4°C. After washing with wash buffer (PBS, 0.05% Tween 20), the nitrocellulose strips were incubated for 2 hours at room temperature with goat anti-rat IgG antibody coupled to peroxidase (Boehringer Mannheim Biochemicals) diluted in AB buffer. The nitrocellulose membrane was washed with wash buffer and the bound antibody visualized by incubation for 30 minutes at RT with 4-chloro—1—naptho1 (0.5 mg/ml in 0.15% H202, 0.5 M Nacl, 50 mM Tris—HC1, pH .5). The reaction was stopped by extensive washing with distilled water.

Identification of the CLMF Subunit Bound by the Monoclonal Antibodies. CLMF is a 75 kDa heterodimer protein composed of 40 KDa and 35 kDa subunits. Western blot analysis was used to determine if the monoclonal anti—CLMF antibodies recognized the 40 kDa or the 35 kDa subunits.

Highly purified 75 kDa CLMF heterodimer was separated by non—reducing SDS/PAGE and transferred to nitrocellulose 31). purified CLMF, composed of approximately 95% free 40 kDa subunit and 5% 75 membrane (Fig. In addition. which was kDa heterodimer, was separated by both non—reducing and reducing SDS/PAGE and the proteins were transferred to nitrocellulose membrane (Fig. 32). Individual nitrocellulose strips containing the non-reduced 75 kDa CLMF heterodimer (Fig. 31), the non—reduced 40 kDa subunit (top panel Fig. 32) and the reduced 40 kDa subunit (bottom panel Fig. 32) were probed with monoclonal anti-CLMF antibodies, control monoclonal antibody, rat anti-CLMF serum and control rat serum. The monoclonal anti—CLMT and rat polyclonal anti—CLMF antibodies bind specifically to an approximately 75 kDa heterodimer on the strips containing non—reduced 75 kDa CLMF while the control antibody preparations do not show this binding activity (Fig. 31). All the monoclonal and rat polyclonal anti-CLMF antibodies recognize the non—reduced 40 kDa subunit (top panel, Fig. 32).

However, only the rat polyclonal antiserum and three monoclonal antibodies, BB3, 9F5 and ZZE7, bind to reduced 40 kDa subunit protein (bottom 32). monoclonal antibodies were specific for the 40 kDa subunit of CLMT. panel. Fig. These data demonstrated that all the Identification of a CLMF Receptor on PHA—Activated Lymphoblasts. monoclonal anti—CLMF antibodies immunoprecipitated . . . .

I—label1ed CLMF, immunodepleted CLMF bioactivity and bound to the 40 kDa subunit of CLMF. the antibodies present in the hybridoma supernatant solutions The previous data demonstrated that the However. could not be directly tested for their ability to neutralize CLMF bioactivity in the TGF or LAK cell induction assays due to non—specific inhibitory effects of supernatant solutions with IL-2 monoclonal antibodies demonstrated that antibodies which would block 125 I—IL—2 binding to IL-2 receptor bearing cells would also neutralize IL-2 bioactivity. containing control antibodies. Our previous work Since receptor binding assays are usually unaffected by addition of hybridoma supernatant solutions or other substances. a CLMF receptor binding assay was developed to evaluate the anti—CLMF antibodies for inhibitory/neutra1i- zation activity. A CLMF receptor binding assay was configured with 1251-labelled CLMF and the PHA—activated - 82 _ peripheral blood lymphoblasts (Fig. 33). The binding of I—CLMF to the PHA—activated lymphoblasts was saturable and specific (Fig. 33). Scatchard plot analysis [See Scatchard; Ann. N.Y. Acad. Sci. git 660-672 (1949)) of the equilbrium binding data indicated that the apparent 5I—CLMF binding to the receptor is approximately 200 pM and that each lymphoblast has about 700-800 receptors. dissociation constant for Since the serum from the rat immunized with CLMF showed neutralization of CLMF I—CLMF The rat immune serum bioactivity, it was tested for inhibition of binding to the lymphoblasts (Fig. 34). blocks 50% of l25I—1abelled CLMF binding at approximately a l/500 dilution, while the control rat serum does not show any inhibition at this dilution. with the specificity of the receptor binding assay established, hybridoma supernatant solutions were tested for antibodies which would inhibit l25I—CLMF binding to lymphoblasts.

I—CLMF binding to the lymphoblasts was determined at a 1/2 dilution of each The degree of inhibition of hybridoma supernatant solution (Fig. 35). Twelve hybridoma supernatant solutions inhibited by greater than 60% . .

I—CLMF binding to the lymphoblasts. present in these supernatant solutions have been classified The antibodies as inhibitory/neutralizing antibodies. Six hybridoma . .

I—labelled CLMF binding by less than 40% and were classified as supernatant solutions inhibited non-inhibitory/non—neutralizing antibodies. Control antibody inhibited by approximately 10% the I—CLMF binding to the lymphoblasts.

Three inhibitory antibodies, 7B2, 2A3 and 4A1, 6A3 and 8E3, and two non-inhibitory antibodies. were purified from ascites fluid by protein G affinity chromatography on GammaBind G (Genex, Gaithersburg, M) columns. Antibodies 4A1, 2A3 and 7B2 inhibit in a dose dependent manner l25I—CLMF binding to the lymphoblasts with IC30 _ 33 _ concentrations of 0.7 ug/ml. 7 uq/ml and 9.5 ug/ml, respectively (Fig. 36).

Antibodies 6A3 and 8E3 do not block I—CLMF binding at concentrations of 100 ug/ml (Fig.

). These data demonstrated that the original classification of each antibody as either inhibitory or non—inhibitory was correct.

Direct Neutralization of CLMF Bioactivity by Antibodies. To determine if the antibodies classified as inhibitory by the CLMF receptor binding assay would directly neutralize CLMF bioactivity, each inhibitory antibody was tested for neutralizing activity in the TGF assay (Table 16). Two inhibitory antibodies, 4A1 and 7B2, demonstrated a dose dependent neutralization of CLMF bioactivity (40 units/ml) from 0.03 to 100 ug/ml. with IC50 concentrations of approximately 1 ug/ml and 80 ug/ml, respectively. These data confirmed that antibodies inhibiting 125I—CLMF binding to the CLMF receptor would also neutralize CLMF bioactivity. -34 _ Table 16: Neutralization of CLMF Bioactivity by Monoclonal Anti-CLMF.

RSS8! Contents: CLMFa Antibodyb Total 3H-Thymidine % Neutralizationc Incorporation none none 9923 i 439 CLMF none 25752 1 592 CLMF‘ 41-11 100 pg/ml 12965 i 938 81 12215 i 663 86 4 12985 i 269 81 .8 19932 1 1016 37 .16 22379 i 410 21 .03 25405 t 1093 2 CLMF 7E2 200 pg/ml 10763 t 878 96 100 15083 t 406 67 23690 i 1228 13 4 25849 i 1408 0 CLMF Control 200 pg/ml 27654 1 1086 0 100 22221 1 381 22 27335 t 620 0 Purified CLMF was used in the TGF D Purified antibodies were added at table. assay at a concentration of 40 units/ml. the concentrations indicated in the C Reduction of 3H-thymidine incorporation to the level seen in the absence of added cytokines was considered to be 100% neutralization.

Preparation of Antibodies Against a Synthetic Peptide Fragment of the 35,000 dalton Subunit of CLMF A peptide, comprising amino acids 3-13 of the NH2—termina1 sequence of the 35 kDa CLMF subunit and a COOH—termina1 cysteine (L—P-V-A—T-P-D-P-G—M-F—C), was synthesized by solid—phase peptide methodology, purified by HPLC. and conjugated to keyhole limpet hemocyanin via the methylated bovine serum albumin procedure. Two rabbits were immunized intradermally with the conjugated peptide in Freund‘s complete adjuvant (300 ug peptide/rabbit). Six weeks after immunization, rabbits were boosted with free peptide (100 ug, intravenously) and KLH—conjugated peptide (150 uq. Serum samples were prepared from bleedings taken 7 days later. The boosting and bleeding schedule was repeated every 4-5 weeks. subcutaneously) dissolved in PBS.

Serum samples from the first and second bleedings from each rabbit were evaluated for reaction with the synthetic peptide in a direct ELISA assay. free The synthetic, peptide was coated on microtiter plates at 4 ng/ml and 20 ng/ml. and the plates were washed and blocked with bovine serum albumin. Serum samples were tested at various dilutions (Table 17), and antibody reactivity was detected with the use of a second antibody (HRP—conjugated goat anti—rabbit IgG) with o—phenylenediamine as substrate.

Absorbance values were read at 490 nm after addition of H2804 to stop the reaction. The results indicate that antibody was produced in both rabbits against 35,000 dalton CLMF peptide (Table 17). In separate experiments, we verified that the antibody was specific for the peptide since (a) serum from non-immunized rabbits does not react with the peptide in ELISA. with the (b) sera from rabbits immunized synthetic peptide do not react with a peptide fragment from the 40,000 dalton CLMF subunit and (c) purified IgG from the serum samples also reacts with the — 86 _ synthetic peptide.

A serum sample from one of the rabbits (first bleed) was tested by Western blot analysis for reactivity with 75 kDa CLMF and with the 35 kDa CLMF subunit (Fig. 37). purified CLMF (approximately 120 ug/ml) was run on SDS~PAGE.

Partially transferred to nitrocellulose. and treated with a :500 dilution of the rabbit anti—CLMF peptide antiserum.

Antibody reactivity was detected by use of biotinylated goat anti—rabbit IgG and alkaline phosphatase—conjugated streptavidin. The anti—CLMF peptide antibody was found to react both with nonreduced 75 kDa CLMF protein and with the reduced 35 kDa CLMF subunit (Fig. 37).

Although the antibodies produced in this example were polyclonal, a similar approach could be used to prepare monoclonal antibodies to the 35 kDa subunit of CLMF. The synthetic peptide used in this example or other synthetic peptides based on the amino acid sequence of the 35 kDa CLMF subunit (Fig. 26) could be used to immunize rats. Fusions could be performed and hybridoma cultures screened for the production of monoclonal anti—CLMF antibodies as described above.