Family of CSF-l Proteins

Hematopoietins, i.e., hematopoietic growth factors, are proteins that promote the survival, growth and differentiation of hematopoietic cells. Colony stimulating factors (CSFs) are a subset of these hematopoietic growth factors that are characterized by the ability to support the growth, in vitro, of colonies of hematopoietic cells arising from progenitor cells of bone marrow, fetal liver and other hematopoietic organs.

The biochemical and biological identification and characterization of certain hematopoietins has been hampered by the small quantities of the naturally occuring factors available from natural sources, e.g., blood and urine. With recombinant genetic engineering techniques, however, some of these hematopoietins have been molecularly cloned, heterologously expressed and purified to homogeneity. [See D. Metcalf, "The Molecular Biology and Functions of the Granulocyte-Macrophage Colony Stimulating Factors," Blood, 67(2):257-267 (1986).] Among these hematopoietins are human and murine GM-CSF, human G-CSF, human CSF-1 and murine IL3. Human IL-3 has also recently been identified [Y. C. Yang et al., Cell, 47(1):3-10 (1986)]. Both human GM-CSF [See, R. Donahue et al., Nature, 321:872-875 (1986)] and murine IL3 [See J. Kindler et al, Proc. Natl. Acad. Sci. U.S.A., 83:1001-1005 (1986)] have a demonstrated effect on hematopoiesis in vivo.

A cDNA sequence purportedly encoding human urinary CSF-1 has been reported by E. S. Kawasaki, et al., Science, 230:291-196 (1985) [hereinafter "Kawasaki"], which, when expressed in a COS cell transient expression system, produced a 26 kD protein that competed with labeled murine CSF-1 in a radioreceptor assay. The protein also reportedly stimulated mouse bone marrow proliferation, resulting in predominantly monocytic lineage type colonies in the mouse bone marrow assay. The protein biological activity was reportedly inhibited by neutralizing antisera for CSF-1.

A. DNAs and Proteins of this Invention

We have discovered a new DNA sequence encoding a protein with CSF-1-like activities which is depicted in FIG. 1, along with the deduced amino acid sequence therefor. The coding region of the DNA sequence of FIG. 1 differs from that reported by Kawasaki primarily by the inclusion of a novel, 894 base pair DNA sequence between nucleotides #689 and #1584 which encodes amino acids #150 through #437 of the depicted peptide sequence. We have observed that mammalian cells expressing that DNA sequence apparently process the precursor protein to yield a homodimer comprising two subunits. Each of the subunits comprises the mature form of the protein containing an amino acid sequence spanning Glu-1 at least through Leu-189 of FIG. 1.

Further research with that DNA sequence led to the discovery of a family of CSF-1-like proteins and DNA sequences encoding them. For instance, one such protein is encoded by nucleotides 1 to 699 of FIG. 1, another by nucleotides 1 to 881, another by nucleotides 1 to 881, another by nucleotides 1 to 1012, and another by nucleotides 1 to 1332. All such proteins have been found to possess CSF-1 activity, as described below.

A common distinguishing structural feature of this family of proteins is that their peptide sequences include part or all of the peptide sequence encoded by the unique region between nucleotides 689 and 1584 of the DNA sequence of FIG. 1.

This invention thus encompasses CSF-1 proteins capable of stimulating the proliferation of monocytic cells in human and mouse bone marrow assays (as discussed below) which are characterized by a peptide sequence encompassing the sequence Glu-1 through Val-152, as shown in FIG. 1. In certain embodiments, the peptide sequence continues further into the unique region encoded by the DNA sequence of FIG. 1. For example, in one embodiment, the protein is characterized by the peptide sequence Glu-1 through at least Leu-189. Other embodiments are characterized by containing the peptide sequences Glu-1 through Asp-213, Glu-1 through Pro-257 and Glu-1 through Arg-363, respectively, and with reference to FIG. 1.

It should be noted that in all cases, the proteins of this family are homodimers of protein subunits comprising the above-mentioned peptide sequences.

This invention further encompasses DNA sequences which encode the novel family of CSF-1 proteins described herein, including both the sequences of FIG. 1 and sequences which differ therefrom in codon sequence due to the degeneracies of the genetic code or differ in nucleotide sequence due to cross-species variation or because of induced modifications which also encode the novel growth factors of the family described herein. A number of such modifications are discussed below.

B. Biological Properties of the proteins of this invention

In addition to the DNA sequence homology to the sequences of FIG. 1, the members of this novel family of growth factors are also characterized by having at least one biological property of a CSF-1-like growth factor. Preferably more than one CSF-1-like biological property is demonstrated by any one member of the family of growth factors of the present invention. "CSF-1-like biological property" is defined herein to include one or more of the following biological characteristics and in vivo and in vitro activities. One such property is the support of the growth and differentiation of progenitor cells committed to the monocyte lineages. For example, in a standard human bone marrow assay, a CSF-1-like biological property is the stimulation of macrophage colonies. Another CSF-1-like biological property is complete inactivation by pre-incubation with 2-mercaptoethanol or an excess of rabbit antiserum raised to purified human urinary CSF-1. Additionally, CSF-1-like properties include the ability to stimulate monocytes to produce other CSF-like factors e.g. myeloid CSF, TNF and interfeon activity which interact with accessory and mature cells. Further CSF-1-like biological properties are the stimulation of biological activities of mature macrophages and the cells of continuous macrophage-like cell lines. Yet another CSF-1-like biological property is an apparent molecular weight of about 50 to 90 kd by non-reducing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Other biological properties attributed to CSF-1 have been disclosed in the art.

C. Other Aspects of the Invention

As a further aspect of the present invention there are provided novel cDNA sequences coding on expression for CSF-1-like polypeptides or growth factors. These DNA sequences include those depicted in FIG. 1 in a 5' to 3' direction and those sequences described above. Variations in the DNA sequences of FIG. 1 which are caused by point mutations or by induced modifications to enhance the activity or production of the polypeptides are also encompassed in the invention. Similarly, synthetic polypeptides which wholly or partially duplicate continuous sequences of the amino acid residues of FIG. 1 are also part of this invention. These sequences, by virtue of sharing primary, secondary or tertiary structural and conformational characteristics with naturally-occurring CSF-1-like polypeptides of the invention may possess biological activity and/or immunological properties in common with the naturally-occuring product. Thus, they may be employed as biologically active or immunological substitutes for naturally-occurring primate CSF-1-like polypeptides in therapeutic and immunological processes.

As another aspect of the present invention, there is provided a novel method for producing the novel family of CSF-1-like growth factors. The method of the present invention involves culturing a suitable cell or cell line, which has been transformed with a vector containing a cDNA sequence coding on expression for a novel CSF-1-like polypeptide. Suitable cells or cell lines may be mammalian cells, such as Chinese hamster ovary cells (CHO). Another suitable mammalian cell line, which is described in the accompanying examples, is the monkey COS-1 cell line. A similarly useful mammalian cell line is the CV-1 cell line. Also suitable for use in the present invention are bacterial cells. For example, the various strains of E. coli are well-known as host cells in the field of biotechnology. Various strains of B. subtilis may also be employed in this method. Many strains of yeast cells known to those skilled in the art are also available as host cells for expression of the polypeptides of the present invention. Additionally, where desired, insect cells may be utilized as host cells in the method of the present invention. So long as a non-human host cell is used, the proteins of this invention are per force produced free from other human proteins.

Another aspect of the present invention provides vectors for use in the method of expression of these novel polypeptides. These vectors contain the novel DNA sequences described herein which code for the novel polypeptides of the invention. Alternatively, vectors incorporating modified sequences as described herein are also embodiments of the present invention and useful in the production of these CSF-1-like polypeptides. The vector employed in the method also contains selected regulatory sequences in operative association with the DNA coding sequences of the invention and capable of directing the replication and expression thereof in selected host cells. A variety of such regulatory sequences useful in the vectors of the present invention are well known to those skilled in the art.

The members of the novel family of CSF-1-like growth factors may be used in the treatment of diseases characterized by a decreased level of hematopoietic cells, particularly those of myeloid, and monocyte lineages. These factors may be used to directly stimulate monocyte and macrophage production and may indirectly stimulate other hematopoietic lineages. Among conditions susceptible to treatment with the polypeptides of the present invention is leukopenia, a reduction in the number of circulating leucocytes (white cells) in the peripheral blood. Leukopenia may be induced by exposure to certain viruses or to radiation. It is often a side effect of various forms of cancer therapy, e.g., exposure to chemotherapeutic drugs. Therapeutic treatment of leukopenia with these CSF-1-like polypeptide compositions may avoid undesirable side effects caused by treatment with presently available drugs.

In addition these polypeptides may serve as activators of mature white cells in cases of serious infection. These factors may be employed to treat infectious diseases characterized by intracellular parasitism, e.g., viral infections (herpes, cytomegalovirus), bacterial infections (Mycobacterium, Listeria), fungal infections (Candida) and parasitic infections (Malaria) and the like.

Alone, or in combination with other hematopoietins, these factors enhance macrophage function causing the activated macrophages to kill tumor cells, to release alphainterferon, to kill parasites or to release and enhance other CSFs which may stimulate the proliferation and activation of other blood cells.

The polypeptides of the present invention may also be employed, alone or in combination with other hematopoietins, in the treatment of other blood cell deficiencies, including thrombocytopenia (platelet deficiency), or anemia (red cell deficiency). Other uses for these novel polypeptides are in the treatment of patients recovering from bone marrow transplants, enhancing host defense during surgery and in burn patients. These factors may also be employed to develop monoclonal and polyclonal antibodies generated by standard methods for diagnostic or therapeutic use.

Therefore, as yet another aspect of the invention are therapeutic compositions for treating the conditions referred to above. Such compositions comprise a therapeutically effective amount of one or more of the members of the family of CSF-1-like polypeptides of the present invention in admixture with a pharmaceutically acceptable carrier. This composition can be systematically administered either parenterally, intraveneously or subcutaneously. When systematically administered, the therapeutic composition for use in this invention is, of course, in the form of a pyrogen-free, parenterally acceptable aqueous solution. The preparation of such a parenterally acceptable protein solution, having due regard to pH, isotonicity, stability and the like, is within the skill of the art.

The dosage regimen will be determined by the attending physician considering various factors which modify the action of drugs, e.g. the condition, body weight, sex, and diet of the patient, the severity of any infection, time of administration and other clinical factors. Generally, the daily regimen should be in the range of 1-1000 micrograms of polypeptide or 50 to 5000 units (i.e., a unit being the concentration of polypeptide which leads to half maximal stimulation in a standard murine bone marrow assay) of polypeptide per kilogram of body weight. This therapeutic composition may also be administered in conjunction with other human factors. A non-exclusive list of other appropriate hematopoietins, CSFs and interleukins for interaction with the polypeptides of the present invention includes GM-CSF, G-CSF, Meg-CSF, erythropoietin (EPO), IL-1, IL-3, other CSF-1-like polypeptides, H-1, IL-4, IL-2, B-cell growth factor, B-cell differentiation factor and eosinophil differentiation factor. The dosage recited above would be adjusted to compensate for such additional components in the therapeutic composition. Progress of the treated patient can be monitored by periodic assessment of the hematological profile, e.g. white cell count and the like.

FIG. 1 shows the DNA sequence encoding a protein with CSF-like activity along with the deduced amino acid sequence.

The present invention provides a family of CSF-1-like growth factors characterized by amino acid sequences substantially homologous to the amino acid sequences illustrated in FIG. 1 as described herein. These sequences may be encoded by the DNA sequences depicted in the FIGURE or variously modified sequences as described above. These polypeptides are also characterized by CSF-1-like biological properties.

The specific sequences illustrated in FIG. 1 are exemplary members of the growth factor family of the present invention. The 4 kb DNA of FIG. 1 codes on expression for a novel CSF-1-like protein, originally named CSF-69. It was isolated from poly A+ mRNA of the SV40 transformed trophoblast cell line TPA30-1 [ATCC #CRL-1583]. The sequence of FIG. 1 contains a long open translational reading frame of 1662 nucleotides, encoding a 554 amino acid polypeptide. The protein coding region of the 4 kb sequence extends from nucleotide #146 (methionine at -32) to nucleotide #1807 which is followed by a TAG stop codon. There are four potential asparagine-linked glycosylation sites illustrated by the characteristic sequences, Asn-X-Ser or Asn-X-Thr. The remaining 2200 nucleotides of 3' non-coding sequence of the 4 kb region may have a regulatory role in transcription in the natural host. The 3' end of the sequence also contains an AT-rich segment including several repeats of the sequence ATTTA, which is believed to be related to the RNA message stability [See, G. Shaw and R. Kamen, Cell, 46(5):659-677 (1986)].

This sequence contains three portions, i.e. from nucleotide #1 to #415, from nucleotide #419 to #689 and from nucleotide #1584 to #1823, which are found in the sequence of Kawasaki et al., supra. The coding region of the CSF-1 described by Kawasaki et al., differs from that described herein by the absence of 894 bp between nucleotides #689 and #1584 recited in FIG. 1.

This approximately 4 kb DNA sequence is harbored in plasmid p3ACSF-69 in E. coli HB 101, which was deposited in the American Type Culture Collection, 12301 Parklawn Dr., Rockville, Md. on Apr. 16, 1986 and given accession number ATCC 67092.

The 4 kb sequence of FIG. 1 codes on expression for one novel CSF-1-like protein of the present invention, CSF-69. CSF-69 is characterized by an apparent molecular weight of approximately 70-90 kD when analyzed by polyacrylamide gel electrophoresis under non-reducing conditions. However, if this analysis is performed after reduction of the CSF-69, the protein is characterized by an apparent molecular weight of 35-45 kD suggesting that CSF-69 is a disulfide linked homo-dimer of 35-45 kD subunits. In p3ACSF-69 conditioned media, CSF-69 demonstrated CSF-1-like activity in in vitro mouse and human bone marrow assays.

The approximately 61 kD precursor encoded by the sequence of FIG. 1 is processed at the amino terminus by removal of a 32 residue signal peptide and in the carboxy terminal region by removal of about 333 residues to yield a subunit of approximately 189 amino acids with a predicted molecular weight of 21 kD. Thus the mature CSF-1 monomer has Glu at its amino terminal and extends at least through to amino acid Leu at position #189 (See FIG. 1). This subunit retains two of the four potential sites for addition of asparagine-linked carboyhydriate that are present in the sequence of FIG. 1. Glycosylation of the 21 kD polypeptide at these two positions is presumed to account for most of the remaining mass of the 35-45 kD subunit of CSF-69.

As previously mentioned, other novel CSF-1-like proteins of the present invention are encoded by only a portion of the sequence of FIG. 1. One such CSF-1-like protein is encoded by nucleotides 1 to 1332 of the cDNA sequence of FIG. 1 and has activity, upon expression, in the murine bone marrow assay described below. Similarly, cDNA sequences of from nucleotides 1 to 699, from nucleotides 1 to 881 and from nucleotides 1 to 1012, when blunted and inserted into the expression vector described below also produced active protein. Additionally, another full length cDNA clone had a cytosine at position 1678 of FIG. 1, rather than a thymidine. This change conserved the amino acid, alanine, in the sequence.

The family of CSF-1-like growth factors provided herein also includes factors encoded by the sequences of FIG. 1 into which nucleotide modifications have been deliberately engineered. Such modifications in the DNA sequences can be made by one skilled in the art using various techniques. Specific modifications of interest in these CSF-1-like related sequences include the replacement of one or more of the cysteine residues in the coding sequences with other amino acids, e.g. serine, to eliminate a disulfide bridge. The 699 nucleotide factor, for example, has eliminated several of the cysteines entirely. Mutagenic techniques for such replacement are well known to one skilled in the art. [See, e.g., U.S. Pat. No. 4,518,584.]

Other specific mutations of the sequences of the CSF-1-like factors described herein involve modifications of one or both of the glycosylation sites. The absence of glycosylation or only partial glycosylation results from amino acid substitution at one or both of the asparagine-linked glycosylation recognition sites present in the sequences of the CSF-1-like factors shown in FIG. 1. These tripeptide sequences are either asparagine-X-threonine or asparagine-X-serine, where X is usually any amino acid. A variety of amino acid substitutions at one or more of the three amino acid positions of a glycosylation recognition site, especially the first and/or third such positions, result in non-glycosylation at the modified tripeptide sequence. By way of example, Asn.sub.122 of the sequence of FIG. 1 can be replaced with glutamine in one such modified CSF-1-like factor. The resulting factor (Gln.sub.122) should contain only one asparagine-linked carbohydrate moiety rather than two such moieties. Those skilled in the art will appreciate that analogous glycoproteins having the same monoglycosylation may be prepared by substituting another amino acid at position 122, and/or by substituting other amino acids at the other positions within the glycosylation recognition sites, e.g., inserting valine at Thr.sub.124. Similarly, the Asn at position 122 and/or Thr at position 124 may be altered by a mutagenic technique to other amino acids to deglycosylate the factor at that site. Alternatively, both of the sites may be altered as above. [See, e.g. A. Miyajima et al. , EMBO J., 5(6):1993-1197 (1986)].

The following examples illustrate the method of the present invention employing the DNA sequence of FIG. 1 to produce CSF-69.

Construction of an exemplary mammalian expression vector p3ACSF-69

To construct a mammalian vector for expression of a novel CSF-1-like protein, the cDNA sequence depicted in FIG. 1 above was adapted with restriction endonuclease enzyme XhoI linkers (New England Biolabs) and ligated into XhoI-digested, phosphatased COS cell expression vector pXM. pXM contains the SV40 enhancer, adenovirus major late promoter, DHFR coding sequence, SV40 late message poly A addition site and VaI gene. pXM further contains a linker sequence with restriction endonuclease sites for KpnI, PstI and XhoI. The plasmid resulting from the XhoI digestion of pXM and the insertion of the linker and the XhoI adapted DNA sequence of FIG. 1 coding for a CSF-like protein was designated p3ACSF-69. p3ACSF-69 (ATCC #67092) can be transformed by conventional techniques into a suitable mammalian host cell for expression of the CSF-like protein. Exemplary host cells are mammalian cells and cell lines, particularly primate cell lines, rodent cell lines and the like.

A similar expression vector may also be prepared containing the other, CSF-1-like sequences identified above, or containing only the amino acid coding regions of those sequences of FIG. 1, with the 5' and 3' non-coding regions deleted. One skilled in the art can construct other mammalian expression vectors comparable to p3ACSF-69 by cutting the DNA sequence of FIG. 1 from the plasmid with XhoI and employing well-known recombinant genetic engineering techniques and other known vectors, such as pCD [Okayama et al., Mol. Cell Biol. 2:161-170 (1982)] and pJL3, pJL4 [Gough et al., EMBO J. 4:645-653 (1985)]. The transformation of these vectors into appropriate host cells can result in expression of a CSF-1-like protein.

Similarly, one skilled in the art could manipulate the CSF-1-like sequences by eliminating or replacing the mammalian regulatory sequences flanking the coding sequence with yeast, bacterial or insect sequences to create non-mammalian vectors expressable in yeast, bacterial or insect host cells. For example, the coding sequence of FIG. 1 could be cut from p3ACSF-69 with XhoI and further manipulated (e.g., ligated to other known linkers or modified by deleting non-coding sequences therefrom or altering nucleotides therein by other known techniques). The modified CSF coding sequence could then be inserted into, for example, a known bacterial vector using procedures such as described in T. Taniguchi et al, Proc. Natl. Acad. Sci U.S.A., 77:5230-5233 (1980). This exemplary bacterial vector could then be transformed into bacterial host cells and the CSF-69 protein expressed thereby.

Similar manipulations can be performed for the construction of an insect vector [See, e.g., procedures described in published European patent application 155,476] or a yeast vector [See, e.g., procedures described in published PCT application WO 86 00639] for expression of the CSF-1-like proteins in insect or yeast cells. To clarify our nomenclature, it may be noted that p3ACSF-69 cut with EcoRI and religated simply to remove 3' untranslated DNA sequence has been variously designated p3ACSF-69R1 and more formally as p3aMCSF R1 #1. Because of its close similarity to the parent vector, that vector has at times been loosely referred to simply as p3ACSF-69, for example in Example IV, below.

Expression of a CSF-1-like Protein

Plasmid DNA, prepared from E. coli HB101 containing p3ACSF-69 (ATCC 67092) as described in Maniatis et al., Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory, (1982) was purified by conventional methods involving equilibrium centrifugation in cesium chloride gradients containing ethidium bromide. COS cells (ATCC CRL 1650) were transfected with the purified DNA at a concentration of approximately 5 ug plasmid DNA per 10.sup.6 COS cells and treated with chloroquine according to the procedures described in G. G. Wong et al., Science, 280:810-815 (1985) and R. J. Kaufman et al., Mol. Cell Biol., 2:1304 (1982). 72 hours following transfection, p3ACSF-69 medium can be harvested containing a protein which demonstrates CSF-1-like activity in standard bone marrow assays, as discussed in Example III below.

CSF Activity in In Vitro Assays

A. Mouse assay

Mouse bone marrow assays were conducted as described in D. Metcalf, The Hemopoietic Colony Stimulating Factors, Elsevier Press, New York (1984) with the following modifications:

(a) 2.times.10.sup.5 bone marrow cells per ml were employed in the assay;

(b) final assay volume was 100 ul; and

(c) assays were set up in standard 96 well microtitre plates.

Bone marrow was obtained from the femurs of 6-25 week old female Balb/c mice (Jackson). Using WEHI 3 conditioned medium [J. C. Lee et al., J. Immunol., 128:2393-2398 (1982)] which contains mouse L cell conditioned medium as a standard control, one dilution unit of activity was defined as that concentration of protein which results in a maximal response in this bone marrow assay, i.e., approximately 15 to 20 colonies per 2.times.10.sup.4 mouse bone marrow cells.

p3ACSF-69 conditioned medium was found to be active to at least 1:10.sup.4 dilution in a mouse bone assay and produced mainly monocytic lineage type colonies. The number and type of cells in a maximal response will vary with the strain and age of the mouse donors.

B. Human Assay

Human bone marrow assays, employing non-adherent bone marrow cells, were performed as described in G. G. Wong, et al, supra. p3ACSF-69 conditioned medium was active to a 1:50 dilution in human bone marrow assays and produced predominantly monocytic lineage-type colonies.

Construction of CHO cell lines expressing high levels of CSF-69

One method for producing high levels of the CSF-1-like polypeptides from mammalian cells involves the construction of cells containing multiple copies of the heterologous CSF-1-like gene. The heterologous gene can be linked to an amplifiable marker, e.g., the dihydrofolate reductase (DHFR) gene for which cells containing increased gene copies can be selected by propagation in increasing concentrations of methotrexate (MTX) according to the procedures of Kaufman & Sharp, J. Mol. Biol., supra. This approach can be employed with a number of different cell types.

p3ACSF-69 (see Example I) and the DHFR expression plasmid pAdD26SV(A) 3 (Kaufman & Sharp, Mol. Cell Biol., supra) were co-transfected into DHFR-deficient CHO cells, DUKX-BII, by calcium phosphate coprecipitation and transfection. The initial DHFR expressing transformants were selected for growth in alpha media with dialyzed fetal calf serum, and subsequently selected for amplification by growth in increasing concentrations of MTX as described in Kaufman et al., Mol. Cell Biol. 5:1750 (1983).

One clone, formally designated "5/9 m.alpha.3-18 (0.2)" [ATCC No. CRL 10154] (and also referred to as CHO-3ACSF-69), which was selected for growth in 0.2 micromolar MTX was found to express high levels of biologically active CSF-69. This cell line consistently generated conditioned medium which was active in supporting murine macrophage colony formation at a .about.1:60,000 final dilution. These cells (one 10 cm dish) as well as the parent CHO cells were labeled with 1 mCi of .sup.35 S-met (NEN) in 4 ml of Minimal Essential Medium (MEM) for 4 hours at 37.degree. C. The resulting conditioned media samples were incubated with antiserum raised in rabbits with purified urinary CSF-1. The antigen-antibody complexes were precipitated by adsorption to Staphylococcus aureus cells (Cal Biochem). The complexes were solubilized in a loading buffer lacking reducing agent according to U. K. Laemmli, Nature, 227: 680-685 (1970). To reduce samples they were brought to 100 mM 2-mercaptoethanol and incubated at 37.degree. C. for 30 minutes. Following electrophoresis in 10% polyacrylamide gel, the pattern of labeled proteins were visualized by fluorography (Enhance, NEN) using Kodak XAR film.

Analysis of these immunoprecipitates by SDS polyacrylamide gel electrophoresis under non-reducing conditions revealed that the conditioned medium from the CSF-1-like protein producing CHO cells contained two heterogeneous CSF-1-like protein species of apparent sizes 70-90 kD and greater than 150 kD. The observed size heterogeneity of these CSF-1-like proteins is typical of many glycoproteins. Analysis of the same samples following reduction revealed that the mobility of the 70-90 kD species of CSF-1-like protein shifted to a position consistent with a molecular weight of 35-45 kD while the relative mobility of the larger species (greater than 150 kD) was unaffected by the treatment. Thus at least two different CSF-1-like proteins are expressed by the 5/9 m.alpha.3-18 (0.2) cells: a 70-90 kD protein comprising a disulfide linked dimer of a 35-45 kD subunit, and a much larger species.

Purification of CSF-69

The CHO-cell conditioned media containing 0.5% fetal bovine serum and DMEM-F12 is diluted 1:1 with water. The diluted media is then applied to a QAE `Zeta-Prep` cartridge (LKB) which is equilibrated in 40 mM Tris pH 7.4. The flowthrough containing unbound protein was discarded. Bound protein was washed with 40 mM Tris, pH 7.4 and eluted with 40 mM Tris, pH 7.4 and 0.75M NaCl. The eluate is then diluted with water to a concentration of 0.5M NaCl. Tween 20 was added to 0.05% and this mixture loaded at approximately 1 column volume/hour on to a lentil lectin. Sepharose 4B column [Pharmacia] which had been equilibrated in 20 mM Tris, pH 7.4, 0.5M NaCl and 0.05% Tween 20. The column was washed with 2-5 cv, 20 mM Tris, Ph 7.4, and 0.5M NaCl. Specifically-bound protein was eluted with 20 mM Tris, 0.2M alpha methylmannopyranoside, 0.5M NaCl and 0.05% Tween 20, and then acidified with 10% trifluoracetic acid [TFA]. The eluate was subjected to reverse phase liquid chromatography on a column equilibrated in 30% acetonitrile and 0.1% TFA. Protein was eluted with ascending acetonitrile in 0.1% TFA. Protein collected between 45 and 50% acetonitrile was neutralized in tubes with Tris, pH 8.5 and analyzed.

The preliminary analysis of CSF-69 reveals a specific activity of approximately 10.sup.6 bone marrow units per milligram [see bone marrow assay in Example III].

Numerous modifications and variations in practice of the invention are expected to occur to those skilled in the art upon consideration of the foregoing descriptions of its preferred embodiments. Such modifications and variations are believed to be encompassed in the appended claims.