NZ238659A - Mast cell growth factor, recombinant production and isolated sequences - Google Patents

Description

23 8 6 5 9 Pflp' y .^)'. !r ."K*.

H : J e' ,- . ..^.".^c.3!A | ! tfaj.f&llvz, j ^'c7""7 I'M ; P.O. JOI* N-r. !$&?. r j Patents Form No- 5 Vv NEW ZEALAND \f 2 1 JUN1991 PATENTS ACT 1953 V. - COMPLETE SPECIFICATION MAST CELL GROWTH FACTOR WE, IMMUNEX CORPORATION, a Delaware corporation, U.S.A. of SI University Street, Seattle, Washington 98101, U.S.A. hereby declare the invention, for which we pray that a patent may be granted to us, and the method by which it is to be performed, to be particularly described in and by the following statement: (followed by page la) 2 3 8 6 5 la IMMUNEX CORPORATION 0521-NZ TITLE MAST CELL GROWTH FACTOR TECHNICAL FIELD OF THE INVENTION This invention relates to mammalian Mast Cell Growth Factor polypeptides ("MGF'). nucleotide sequences encoding MGF and certain derivatives, analogs and allelic variants thereof, processes for production of MGF polypeptides by recombinant DNA technology or by purification from culture media of cell lines which consritutivelv produce MGF, and pharmaceutical compositions comprising MGF and an additional growth factor. More specifically, the invention provides isolated mammalian nucleotide sequences encoding MGF and derivatives thereof that regulate early events in stem cell proliferation. More specifically, the present invention discloses murine and human nucleotide sequences encoding murine and human MGF, and also nucleotide sequences capable of hybridizing to disclosed and/or deposited nucleotide sequences and encoding MGF and derivatives or analogs thereof having MGF biological activity.

BACKGROUND OF THE INVENTION Hematopoietic growth factors regulate the growth and maturation of various lineages of blood cells. All known blood cells are believed to develop from a single class of precursor cells called stem cells. Each hematopoierin causes specific classes of blood cells to differentiate and proliferate. When a stem cell divides in the bone marrow, it can replicate itself as a stem cell or become committed to a particular developmental pathway. As a result of commitment, a stem cell displavss receptors on its cell surface that enables it to respond to certain hormonal signals. Such signals push the cell further down a pathway leading to terminal differentiation. As differentiation proceeds, the first recognizable precursors appear, including erythroblasts (red cell precursors) and myeloblasts (precursors of granulocytes and monocytes).

Scientific understanding of hematopoiesis has been aided by developments in molecular biology. A number of glycoproteins have been identified which regulate cell development at various levels within the hematopoietic stem and progenitor cell hierarchy. The majority of growth factors that have been identified influence relatively late sages of differentiation and regulate the number and function of mature differentiated hematopoietic elements.

Identification of the factors controlling more primitive elements of the hierarchy has proven difficult, due to the very low frequency of these progenitor cells, and a lack of phenotypic markers restricted to stem cells.

By definition, a "true" pluripotent hematopoietic stem cell (PHSC) is a cell capable of extensive proliferation. self-renewaL and the ability to mediate long-term reconstitution of all hematopoietic lineages. PHSC are extremely rare cells in the hematopoietic system, which are 238659 2 not actively cycling under normal circumstances. In the mouse, attempts to purify PHSC have centered on the purification of spleen colony-forming ceils (CFU-S) (Spangmde et al.. Science 241:58 (1988)). However, others have demonstrated that CFU-S cells are not the "true" PHSC as assessed by their capacity for long-term reconstitution. Other studies have suggested a cell more primitive than a CFU-S which then gives rise to daughter CFU-S cells. Such a cell has been referred to as a "pre-CFU-S." Other cells have certain stem cell-like features, including high proliferative potential colony-forming cells (HPP-CFC), a mixed colony-forming cell (CFU-GEMM). a blast cell CFC. and long-term culture initiating cells. These later cell types rail to fulfill all of the criteria for a "true" stem cell, but point to the complexity and heterogenous functional capability of primitive hematopoietic cells.

Progenitor cells are the immediate progeny of stem cells, but differ in two respects.

First, progenitor cells have a restricted capacity for self-renewal, and secondly, progenitor cells are committed irreversibly to a single lineage of hematopoietic differentiation (or two, in the case of many granulocyte-macrophage progenitor cells). Progenitor cells are the cells responsible for forming the majority of observable cell colonies in experimental cultures of hematopoietic cells. Examples of progenitor cells include, granulocyte-macrophage progenitors, eosinophil progenitors, megakaryocyte progenitors, multipotential progenitors and ervthroid progenitors.

Hematopoiesis in vivo is regulated both by feedback signals generated from peripheral tissues. i.e., long-range regulation, and by interaction between stromal cells in the bone marrow and target hematopoietic stem or progenitor cells. Regulation is likely due to the transfer of regulatory molecules from ceil to celL Stromal cell lines produce a variety of known hematopoietic factors, such as interieukin-7 (IL-7), granulocyte-macrophage colony stimulating factor (GM-CSF), granulocyte CSF (G-CSF), macrophage CSF (M-CSF), interleukin-1 (EL-la and EL-1(3). interleukin-6 (IL-6) and leukemia inhibitory factor (LIF).

There is a continuing need in the an to identify hematopoietic regulatory factors and the cellular receptors that bind to each regulatory factor. One such receptor is a tyrosine kinase receptor encoded by the c-lat proto-oncogene (Yarden et al. EMBO J <5:334-1 1987). This receptor is structurally related to the receptor for M-CSF, (Qui et al. EMBO J. 7:1003 1988) which is encoded by the c-fms proto-oncogene (Shen- et al. Cell 11:665 1985). Prior to the present invention, the ligand for the c-ki: receptor had not yet been identified or characterized in the art.

Mice with mutations at the Steel or W locus display a similar phenotype characterized by a reduction of pluripotent hematopoietic stem cells (PHSC), anemia, a deficiency of mast ceils, and defects in gametogenesis and pigmentation (Russell Adv. Genet. 20:5511979). The hematopoietic defect in W mice is intrinsic to the PHSC, whereas that of the Steel mouse is the result of a microenvironmental aberration. Recent studies have shown that the W locus on murine chromosome 5 encodes the c-ki: proto-oncogene (Chabot e: al. Nature 555:89 1988: 23 8 6 j Gcissler ct al. Cell 55: 185 1988). Previous studies have shown that bone marrow from Sted mice transplanted into W recipients resulted in a hematological^ normal phenocype (Bernstein ct al. Ann. N,Y. Acad. Sci. 149:415 1968) and has led to the hypothesis that the gene product of the Steel locus is the ligand for c-ki: (Chabot et al. Manure 555:88 1988; Gcissler et al. Cell 55: 185 1988).

In mice, mutations at the Steel (SI) locus identify a gene essential for the development of neural crest-derived melanocytes, primordial genn cells and hematopoietic stem cells. These cell populations form in SI homozygous embryos, but fail to proliferate or survive, and thus the animals are black-eved. white, sterile and severely anemic. Several alleles are lethal when homozygous due to failure of erythropoiesis. The lethal period begins at the thirteenth day of gestation after hematopoiesis has commenced in the normal fetal liver. Heterozygote animals are generally viable but differ in severity on three different affected lineages.

Graft experiments have demonstrated that the SI gene acts in the extracellular environment rather than in the affected cells themselves. For example, when SI/SI premigratory neural crest cells (precursors of melanocytes) are grafted alongside wild type (■+■/+) skin, they give rise to pigmented melanocytes that populate and pigment skin hair follicles, whereas the converse experiment results in unpigmented hairs. Similarly, hematopoietic stem cells from anemic S!/SId animals can repopulate stroma and rescue a Iethally-iiradiated syngeneic host. However, -r/-r stem cells fail to rescue a SllSld anemia. Although grafting experiments of primordial germ cells have not been possible, studies of Sl/Sld <-> +/-r chimeras suggest that SI/SId germ cells are able to populate the gonad and give rise to functional sperm if the appropriate environment is present.

In mice, W or dominant white spotting mutations parallel those of the SI series. The W mutation in mice affects the same three (primordial germ cells, hematopoietic stem cells and neural-crest derived melanocytes) stem cell lineages, ervthroid cells and mast cells. Due to severe anemia the W mutation is generally lethal in the homozygous condition. Grafting experiments, much like those carried out for SI mutations, suggest that the site of action for the W gene is within the affected cells themselves rather than within the extracellular environment. Neural crest cells from W/W embryos grafted alongside +/■*■ skin are unable to give rise to melanocytes, whereas neural crest cells form melanocytes that populate and pigment WlW skin hair follicle. Hematopoietic stem cells from severely anemic WlV/v animals fail to rescue Iethally-irradiated syngeneic hosts but the \V!Ww hematopoietic microenvironment efficiently supports +/-r stem cell proliferation. 3v inference, the W gene is thought to act in the same way in primordial germ cells, although no experimental system exists for testing this hypothesis.

Therefore, there is a need in the an to identify and characterize mammalian equivalents and homologies of the gene product of the mouse Steel locus, herein designated MGF. Such potential hematopoietic stromal cell factors are likely to be useful regulatory molecules for 23 3 6 5 controlling the early growth and differentiation of stem cells. The ligand is likely to be useful for seating aplastic anemia when an individual already has erythropoietin (EPO) available. MGF is likely to be further useful to treat bone marrow suppression or ineffective hematopoiesis. MGF should be useful when used in combination with other factors having colony stimulating biological activity, such as GM-CSF. G-CSF. IL-3, CSF-1, EL-6 and IL-1.

SUMMARY OF THE INVENTION We have purified, sequenced and cloned cDNAs encoding several forms of a murine and human (i.e., mammalian) MGF. a novel hematopoietin which stimulates a select group of IL-3 dependent mas: cell lines and hematopoietic progenitor cells. The nucleotide sequence and deduced amino acid sequence of murine MGF ("mMGF) are disclosed in Figures 1 and 2. (Sequence ED Nos. 1 and 2). The nucleotide sequence and corresponding amino acid sequence of the intracellular region, transmembrane region and the extracellular region of human MGF ChMGF") are disclosed in Figures 3 and 4 (Sequence LD. Nos. 3 and 4). Mature hMGF polypeptide begins at a Glu residue L The mature extracellular region of hMGF is 156 amino acids in length as shown in Figure 3, terminating at the Asp residue at amino acid 157. There is another form of hMGF resulting from an alternative mRN'A splicing event comprising an exon encoding an additional 28 amino acids beginning at amino acid 148. This form is illustrated in Figure 4. The alternative form is provided by clone hMGF-2.4 and has 185 amino acids in its full length mature extracellular region.

The present invention further comprises other mammalian MGF polypeptides having MGF biological activity and encoded by nucleotide sequences which hybridize, under conditions of moderate stringency, to a probe defined by clone 10 (deposited with the ATCC on September 11, 1990 under accession number 68396) and shown in Figure 2. or by mMGF-94 (shown in Figure 1). or which hybridize to human clone MGF-2D (Figure 3) or hMGF-2.4 (Figure 4) under conditions of high stringency or to various human sequence probes described herein. Moreover, the present invention comprises hMGF nucleotide sequences and polypeptides, wherein said nucleotide sequences hybridize, under high stringency conditions, to a probe defined by clone MGF-2D (whose sequence is shown in Figure 3) or to the nucleotide sequence encoding the extracellular portion of the human sequence and shown in Figure 3.

Proliferative responses to mMGF correlated with expression of c-kit mRNA.

Moreover, purified, radiolabeled mammalian MGF can be cross-linked to responder cells and subsequendy immunoprecipitated with an antisera recognizing the carboxv-terminus of the c-kit gene product. Therefore. MGF is the ligand for the protein receptor expression product of c-kit. Moreover, binding to the polypeptide receptor encoded by the c-kit proto-oncogene is a measure of MGF biological activity. The c-kit sequences (mouse and human) have been 238659 published (see Background section) and expressed in various ceil types as disclosed herein in the Examples.

MGF has a transmembrane region, an extracellular region and an intracellular region.

Human MGF has a 27 amino acid transmembrane region beginning with a Ser residue at amino acid 158 (Figure 3. clone MGF-2D) or amino acid 186 (Figure 4. clone MGF-2.4). A transmembrane region typically comprises hydrophobic amino acid residues flanked by charged amino acids. Tne intracellular region of hMGF begins with a Lys residue at amino acid 185 (Figure 3, clone MGF-2D) or at amino acid 213 (Figure 4. clone MGF-2.4) and is 36 amino acids in length. Another difference between hMGF-2.4 and hMGF-2D is that amino acid 149 of hMGF-2D is a Gly whereas corresponding amino acid (just after the spliced exon sequence) 177 of hMGF-2.4 is an Arg. This amino acid difference is a result of the spliced exon sequence being insencd within a codon. Figures 1.2.3 and 4 illustrate nucleotide and amino acid sequences of wild type mMGF (Figures 1 and 2) and hMGF (Figures 3 and 4).

Clone hMGF-2D comprises the full coding region of wild type hMGF without the additional 28 amino acid alternative exon region. Clone hMGF-2.4 comprises the full hMGF coding region with the alternative exon sequence. The polypeptide expressed by the extracellular region of clone nMGF-2.4 is called &2S MGF.

Mammalian or human hematopoiesis. including megakaryoevtopoiesis. are influenced by administration of MGF. However, optimal hematopoiesis is achieved by combination therapy with MGF and other cytokines, including but not limited to GM-CSF, IL-3, a fusion protein comprising both GM-CSF and EL-3, EPO. IL-la. EL-lp, G-CSF. UF, 1L-6. and IL-7.

More specifically, we found that MGF has the ability to augment the action of other cytokines for megakaryoevtopoiesis. an activity shared by EL-3 and EL-6. and provides maximal hematopoietic progenitor proliferative stimulus when employed in combination with GM-CSF, EL-3 or GM-CSF/EL-3 fusion proteins.

MGF was found to influence early lymphoid development, early myeloid development, and promote proliferation and sustain hematopoiesis of early stem cells or progenitor cells of the more primitive types.

BRTEF DESCRIPTION' OF THE DRAWINGS Figure 1 snows a cDNA sequence and corresponding amino acid sequence for clone mMGF-94 (Sequence ID No. 1) without the 16 amino acid spliced sequence. Figure 2 shows the nucleotide sequence and corresponding amino acid sequence for murine MGF clone mMGF-10 (Sequence DD No. 2) with the alternative mRNA exon sequence that adds 16 amino acids to the extracellular region of the oolvoentide.

Figure 3 illustrates the wild type hMGF cDNA and amino acid sequence as derived from clone hMGF-2D (Sequence ID No. 3) without the 28 amino acid alternative exon region. 238659 6 This sequence encodes a polypeptide that is also called A2S hMGF. Figure 4 illustrates the wild type hMGF cDNA and amino acid sequence as derived from clone hMGF-2.4 (Sequence ID No. 4). The transmembrane region begins with a Ser Ser sequence following an Asp residue and ends with a Leu Tvr Trp sequence just prior to a Lys Lys sequence that begins (N-terminus) the intracellular region. A portion of the 5* non-coding region is also shown. The mature hMGF polypeptide begins with a GIu residue in the extracellular region.

Figure 5 shows a dose-response comparison of BFU-E and CFU-GEMM activities of different concentrations of MGF.

Figure 6a shows the number of cclls recovered from purified stem cell fractions after three days in culture with medium alone, mMGF alone, IL-3 alone or EL-3 plus mMGF for early (dull) and late (bright) stem cells. Figure 6b shows the recovery of CFU-S from dull and bright populations of primitive stem ceils after 3 days of culture with medium alone, mMGF alone, IL-3 alone or IL-3 plus mMGF.

DFTATI.ED DESCRIPTION OF THF. INVENTION We have discovered a novel mammalian polypeptide that is the ligand for die receptor encoded by the c-ki: proto-oncogene. More particularly, we have cloned and sequenced a cDNA encoding purified mMGF. We then cloned and sequenced a cDNA corresponding to hMGF. Both MGFs have an extracellular region that contains the c-ki: binding region, a hydrophobic transmembrane region delineated by charged amino acids on either side and an intracellular region. The present invention further comprises a soluble human MGF polypeptides comprising amino acid sequences corresponding to the extracellular domain or at least a region of the extracellular domain encompassing all four Lys residues.

MGF is a ligand for the gene product of the c-kit proto-oncogene. c-ki: encodes a tyrosine kinase receptor on bone marrow stem cells and other tissues including, for example, primordial germ cells (PGC). melanocytes and developing (embryonic) neural crest cells arising from the neural tube. Tnus, MGF activity comprises signaling proliferation and/or differentiation of precursor bone marrow stem cells and a variety of other germ-type cells for developing central and peripheral nervous system, gametogenesis and pigmentation.

Moreover. MGF is preferably secreted by bone marrow stromal cells in either precursor or mature forms, or both. There is also a natural soluble form of MGF comprising the extracellular region or a fragment thereof.

Assavs for MGF Isolation and characterization of MGF requires a mammalian cell line that produces MGF, and a receptor cell line for assay that proliferates in response to MGF stimulation. A biological assay for mammalian MGF can use, for example, a growth factor-dependent mast 238659 cell line to look for ccil proliferation. Any EL-3 dependent cell line expressing c-kit can be used to provide responder cells for a proliferation assay receptor or binding assay for a mammalian MGF. Moreover, fresh mast cells isolated from bone marrow of normal mice express c-kit and can similarly be used for receptor binding assays or for proliferation assays for mammalian MGF. In addition, fresh mast cells isolated from marrow samples taken from humans or from mammals can be used to assay for mammalian MGF.

A MGF-dependent cell line can be derived from fresh Thy-1~ marrow cells of NFS/N mice. The NFS/N 1 cell line is a mast cell that lacks Thy-1"". ThB. and B220. This cell line responds to purified murine IL-3 and recombinant murine EL-4, but not to natural or recombinant murine GM-CSF. recombinant human G-CSF or murine CSF-1. An EL-3-dependent cell line was obtained from IL-3 responsive cells from an NFS/N mouse strain passaged in a high concentration of WEHI-3 conditioned media ("WEHI-3-CM") containing EL-3 after enrichment of Tny-i~ cells by a fluorescence-activated cell sorter. After passage in viaro for several months, the cell line was cloned. Cell line cloning can be accomplished, for example, by limiting dilution and by isolation of individual colonies growing in methylceilulose in response to a combination of WEHI-3-CM and stromal cell-CM. WEHI-3-CM are collected from supematants of the WEHI-3 murine leukemia cell line (ATCC TIB68) growing in RPMI-16-10 medium supplemented with 2% teal calf serum The supernatant is dialvzed against distilled water and concentrated five-fold prior to use.

The NFS/N 1-MC6 or MC6 cell line provides biological activity assays for murine MGF by proliferating in response to murine MGF as measured by colony formation or by uptake of tritiated thymidine. Similarly, human MGF can be assayed for MGF biological activity by anyone of a number of cell lines that require human EL-3 for growth and express the product of the c-kii gene product. Examples of such cells include TFl cells (Kitamura et aL, J. Cell. Physiol. 140:323.1989) and Mo7e cells (Avanzi et al.. Br. J. Haematol. 69:359, 1988). To conduct a proliferation assay, factor-dependent cells (e.g., MC6 cells. TFl cells, or mast cells) are placed into replicate microwells in a suitable medium, such as RPMI1640 supplemented with 10% fetal calf serum. Supematants from cells secreting putative growth factor are added to the wells and the cells are incubated overnight at 37*C in a 5% CO2 atmosphere. The cells are next pulsed with I u.Ci of tritiated thymidine and incubated for an additional eight hours. The wells are harvested onto glass fiber filter discs for counting by liquid scintillation. Presence of a growth factor is determined by increased tritium in the cells as an indication of cellular division.

MGF Polypeptides MGF refers to a family of mammalian polypeptides which are capable of stimulating EL-3 dependent mast cell lines and hematopoietic progenitor cells, and serve as a ligand for the gene product of the c-kit proto-oncogene.. As used herein, the term MGF includes analogs or 23 8 6 8 subunits of native mammalian polypeptides with substantially identical or substantially similar amino acid polypeptide sequences which bind to the protein expressed by the c-ki: proto-oncogene and which induce proliferarion of mas: cells, for example, the IL-3 dependent murine mast cell line MC6 or human cell line TFl. MGF may consist solely of its extracellular region or a fragment thereof including all four Cys residues of the extracellular domain and may lack a transmembrane region and intracellular domain. The extracellular region of MGF or fragment thereof is a soluble polypeptide. Both human and murine native sequence MGF have been found in two variations, differing by an extra coding region in the extracellular domain of the polypeptide. One human variant (clone MGF 2.4) has an extra 28 amino acids beginning at amino acids 148 in the extracellular domain of the polypeptide. One murine variant (clone mMGF-10) has an extra 16 amino acids beginning with amino acid 148 in its extracellular domain. Similarly, clone hMGF2-D lacks a 28 amino acid sequence in the extracellular domain. The extracellular domain of this polypeptide is called A2S hMGF.

Human MGF polypeptides have a full length 185 amino acid extracellular domain.

This polypeptide has five glycosvlation sites and four Cys residues. The firs: Cys residue at position 3 of the mature human polypeptide sequence binds to the third Cys residue at position 89 of the human MGF polypeptide sequence. Similarly, the second Cys residue at position 43 binds to die fourth Cys residue at position 138. A28 hMGF retains ail four Cys residues, but eliminates die fifth glycosvlation site. When several amino acids are removed from the C-terminus of hMGF extracellular domain, up to the fourth Cys residue at position 138. biological activity is retained. Similarly, one can remove the first three N"-terminal amino acids, up to die firs: Cys residue at position 3 of mature human MGF. and retain biological activity. Therefore, a human MGF polypeptide having only 136 amino acid residues, but retaining all four Cys residues, retains significant biological activity, comparable to full length 185 amino acid human MGF or A28 hMGF. A28 hMGF has significant activity.

Murine and human MGF are $7% homologous at die DNA level and 90% homologous at the protein leveL However, despite 90% homology at the protein level, there are enough differences in protein structure for human MGF to be 2-3 logs less active in a murine cell (MC6) assay and murine MGF to be 2-4 fold less active in a human (TFl cell) proliferation assay.

In order to find important amino acid residues involved in receptor binding and conferring species specificity, we made human-murine chimeric MGF polypeptides incorporating various sequence regions of the extracellular domain. We were able to determine that there were at least one and possibly a group of amino acids involved in conferring murine MGF biological activity between Sspl restriction site (between residues 90 and 91) and an EcoRI (between residues 138 and 189) restriction site in the human sequence. The protein sequence between these restriction sites contains 50 amino acids, 14 of which are different amino acids between the human and murine MGF polypeptides. We made a murine-human- 238659 9 murine MGF chimeric protein having murine sequence from the 5" end to the Ssp I site, human sequence from the Ssp I site to the EcoRl site and murine sequence from the EcoRI site to the 3' end of the extracellular domain. This chimera was acdve in a human assay but inactive in a murine assay. The 50 amino acid region was not sufficient (alone) to confer total biological activity but needs to be configured in the context of additional protein.

We investigated 14- different amino acids within the 50 amino acid region. Chimeric constructs were made by a PCR overlap extension method and transfected into COS cells.

COS cell supematants were assayed in murine and human proliferarion assays. Based upon these data, we found that the amino acid at position 91 is important to confer species specificity of an MGF polypeptide. Murine MGF has a GIu (no charge) residue at position 91, and human MGF has a Lys (positively charged) residue. Thus, the amino acid residue at position 91 of either human or murine MGF is important for receptor binding or protein conformation (tertiary structure).

One can make analogs of the hMGF molecule which retain MGF biological activity. For example, glycosvlation sites can be altered to facilitate expression in yeast or mammalian cell systems. The region surrounding the third Cys at position 89 is important for receptor binding. However, the Val residue at residue 90 can be altered to any other amino acid without affecting biological activity. The adjoining Lys residue at position 91 confers human species specificity to hMGF and requires a GIu residue at this position for murine species specificity.

Human MGF analogs can vary in length from about 135 amino acids to about 185 amino acids constituting the extracellular domain of the hMGF polypeptide. Biologically acdve hMGF analog polypeptides comprise four Cys residues, but can vary at other positions according to the following sequence: Xn Xn Xn Cys Arg Asn Arg Val Thr Asn Asn Vai Lys As? Val Thr Lys Leu "val Ala Asn Leu Pro Lys As? Tyr Met lie Thr Leu Lys Tyr Vai Pro Gly Met Asp Val Leu Pro Ser His Cys Tr? lie Ser Giu Met Vai Val Gin Leu Ser Asp Ser Leu Thr As? Leu Leu As? Lys Phe Ser Asn lie Ser Giu Gly Leu Ser Asn Tyr Ser lie lie As? Lys Leu Val Asn lie Vai As? As? Leu Vai Ri Cys S-2 Q Giu Asn Ser Ser Lys As? Leu Lys Lys Ser Phe Lys Ser Pro Giu Pro Arg Leu Phe Thr Pro Giu Giu Phe Phe Arg lie Phe Asn Arc Ser lie A.s? Ala Phe Lys As? Phe Vai Vai Ala Ser Giu Thr Ser As? Cys Xn Xn Xn Xr. Xn Xn Xn Xn Xn Xn wherein n is 0 or 1, X is any naturally occurring amino acid: Ri is any amino acid except Giu: R? is any amino acid except Val: and Q is Lys or Arg to provide human activity, or is GIu to provide murine activity. 23 8 6 The present invention further comprises cDNA sequences that encode hMGF. Human MGF is encoded, for example, by a nucleotide sequence comprising the sequence from nucleotide 112 to nucleotide 516 in Figure 3, or a DNA sequence that hybridizes to a probe defined by the foregoing sequence under conditions of high stringency.

Moderate stringency hybridization conditions, as defined herein and as known to those of skill in the an, refer to conditions described in. for example, Sambrook et aL Molecular Cloning: A Laboratory Manuel, 2 Ed. Vol. 1 pages 1.101-1.104 (Cold Spring Harbor Laboratory Press 1989). Exemplary conditions of moderate stringency are a prewashing with x SSC, 0.5% SDS, 1 mM EDTA (pH S.0) and overnight hybridization at 50*C in 2 X SSC. Exemplary severe or high stringency conditions are overnight hybridization at about 68"C in a 6 x SSC solution, washing at room temperature with 6 x SSC solution, followed by washing at about 68*C in a 0.6 x SSC solution.

We initially purified and characterized murine MGF. After obtaining an mMGF cDNA. we obtained the cDNA sequence for hMGF by cross-species hybridization. We have further cloned, sequenced and expressed two variant hMGF cDNAs.

The primary amino acid structure of mammalian MGF may be modified by forming covalent or aggregative conjugates with other chemical moieties, such as glycosyl groups, lipids, phosphate, acetyl groups and the like, or by creating amino acid sequence mutants.

Such polypeptides are called derivatives. Covalent derivatives of mammalian MGF are prepared, for example, by linking particular functional groups to mammalian MGF amino acid side chains or at the N-terminus or C-terminus of a mammalian MGF polypeptide. Other derivatives of mammalian MGF within the scope of this invention include fusions of MGF or its fragments with other proteins or polypeptides, such as by synthesis in recombinant culture as N-terminal or C-terminal fusions. For example, the fusion polypeptide may comprise a signal (or leader) polypeptide sequence at the N-terminal region of a mammalian MGF polypeptide which co-translationallv or post-translationally directs transfer of a mammalian polypeptide from its site of synthesis to a site inside or outside of the cell membrane or wall (e.g., the yeast a-factor leader). Mammalian MGF polypeptide fusions can also comprise fusion to other biologically active proteins or immunoglobulins.

The present invention further includes mammalian MGF polypeptides having altered glycosvlation. Mammalian MGF expressed in yeast or mammalian expression systems (e.g.. COS-7 cells) may be similar or significantly different in molecular weight and glycosvlation pattern than a native mammalian MGF polypeptide. This depends upon the choice of expression system. Expression of mammalian MGF polypeptides in bacterial expression systems, such as E. colL provides non-glvcosylated molecules.

Functional mutant analogs of human or mammalian MGF can be synthesized, for example, with inactivated N-glvcosylation sites by oligonucleotide synthesis and ligation or by site-specific mutagenesis techniques. N-glvcosylation sites in eukaryotic polypeptides are 1! 238659 characterized by an amino acid triplet Asn-O-Q where O is any amino acid except Pro and H is Ser or Thr. In this sequence, carbohydrate residues are covalentiy attached at the Asn side chain. Human MGF clone MGF-2D contains four glvcosvlation sites in the extracellular 9 • « region of the polypeptide and hMGF-2.4. further contains a fifth glycosvlation site in the 28 amino acid alternative exon located in the extracellular region. The glycosvlation sites in hMGF-2D begin with the Asn residue at positions 65,72,93 and 120. The fifth glycosyladon site of human MGF is at position 170. The A28 hMGF polypeptide does not contain a fifth glycosvlation site. Surprisingly, yeas: expression of human MGF hvperglycosylates only the fifth glycosvlation site.

Mammalian MGF polypeptides are encoded by multi-exon genes. As exemplified by the various constructs described herein, the present invention includes DNAs and polypeptides corresponding to alternative mRNAs which can be attributed to different mRNA splicing events following transcription, and which share regions of identity or similarity with the cDNAs disclosed herein.

Bioequivalent analogs of mammalian MGF polypeptides or MGF muteins (defined as polypeptides having MGF biological activity and binding to cells expressing c-kii proto-oncogenes) may be constructed, for example, by making various substitutions of amino acid residues or sequences, or by deleting terminal or internal residues or sequences not needed for biological activity. Generally, substitutions are made conservatively by substituting an amino acid having phvsiochemical characteristics resembling those of the replaced residue. Further substitutions may be made outside of the "core" sequence (for human MGF, the core sequence lies between the firs: the founh Cys residues) needed for mammalian MGF binding and biological activity. However, the amino acid residue at position 91 is essential for receptor binding and helps to confer species specificity. Subunits of a mammalian MGF polypeptide may be constructed by deleting terminal or internal residues or sequences.

We have purified mMGF to provide a substantially homogeneous polypeptide preparation from a crude protein solution, such as conditioned medium collected from cells expressing MGF. The purification procedure we used is described in Example 1 herein.

The nucleotide sequence and corresponding amino acid sequence of mMGF are illustrated in Figures 1 and 2. These murine sequences were obtained from mMGF-10 and mMGF-94, construc:ed from -r/-r cDNA. that was Sal 1 linkered into a 5a/1 site of vector HAVEo. These isolates were cloned as described in Example 2 herein.

An oligonucleotide probe was created from the nucleotide sequence of isolate 10 used to screen a human cDNA library under moderate stringency conditions. Similar cross species hybridization techniques may be applied to clone new mammalian MGF nucleotide sequences when a particular mammalian sequence (e.g.. murine) is known and cDNA or genomic DN'A libraries made from cells that would produce MGF polypeptides are available for the desired species. 12 238659 Expression of MGF Polypeptides MGF and especially human MGF and analogs thereof can be expressed in a variety of prokaryotic and eukaryotic host systems. Examples of prokaryoric host systems include £. coli bacteria and Staphylococcus aureus. Examples of eukaryodc host systems include mammalian cells (COS). Chinese hampster ovaiy (CHO) and yeast cells such as S. cerevisiae.

We have expressed various human MGF polypeptides in a yeast (5. cerevisiae) system using a yeast expression vector. pDCY-120 which comprises an ADH2 promoter sequence and the a-factor secretion leader. The ADH2 promoter has been described in Russell et al., J. Biol. Chem. 255:2674. 1982, and Beier et al., Nature 300:724, 1982. The a-factor leader sequence is often inserted between the promoter sequence and the structural gene sequence. The use of an a-factor leader sequence to improve expression of a particular structure gene has been described in, for example. Kurjan et al.. Cell 30:933. 1982, and Bitter et al., Proc. Natl. Acad. Sci. USA 81:5330. 1984.

The pDCY 120 vector further comprises an Asp 718 restriction site at the 5* end and a Bam HI restriction site at the 3* end to facilitate insertion of a structural gene. The pDCY 120 expression vector is cut with Xko I and an insert comprising a MGF cDN'A is Iigated into the pDCY 120 vector with Sal I restriction enzyme.

Recombinant expression vectors include suitable transcriptional or translational regulatory or structural nucleotide sequences which may be derived from mammalian, microbial, viral or insect genes. Examples of regulatory sequences include transcriptional promoters or enhancers, an optional operator sequence to control transcription, a sequence encoding suitable mRNA ribosomal binding sites, and appropriate sequences which control transcription and translation initiation and termination.

Suitable host cells for expression of mammalian MGF or derivatives thereof include prokaryotes. yeast or higher eukarvotic ceils under the control of appropriate promoters. Prokarvotes include gram negative or gram positive organisms, for example E. coli or bacilli. Suitable prokaryotic hosts cells for transformation include, for example. E. coli. Bacillus subalis, Salmonella ryphimuriunu and various other species within the genera Pseudomonas, Streptomyces, and Staphylococcus. Higher eukaryotic cells include established cell lines of mammalian origin as described below. Cell-free translation systems could also be employed to produce mammalian MGF or derivatives thereof using RNAs derived from the DNA constructs disclosed herein. Appropriate cloning and expression vectors for use with bacterial, fungal, veast, and mammalian cellular hosts are described, for example, in Pouwels et aL Cloning Vectors: A Laboratory Manual, Elsevier. New York. 1985.

Promoter sequences are commonly used for recombinant prokaryotic host cell expression vectors. Common promoter sequences include j3-lac:amase (penicillinase), lactose promoter system (Chang et al.. Nature 275:615,1978: and Goeddel et al.. Nature 2S1:544. 13 238659 1979). tryptophan (trp) promoter system (Goeddel et al.. Nucl. Acids Res. 3:4057. 1980: and EP-A- 36.776) and tac promoter (Maniaris. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, p. 412.19S2). A particularly useful prokaryotic host cell expression system employs a phage X ?i_ promoter and a cI857ts thermolabile repressor sequence. Plasmid vectors available from the American Type Culture Collection which incorporate derivatives of the \ Pj_ promoter include plasmid pHUB2 (resident in E. coli strain JMB9 (ATCC 37092)) and pPLc2S (resident in E. coli RR1 (ATCC 53082)).

Yeast transformation protocols are known to those of skill in the an. One such protocol is described by Hinnen et al.. Proc. Nail. Acad. Sci. USA 75:1929. 1978. The Hinnen et al. protocol selects for Trp"1" transformants in a selective medium, wherein the selective medium consists of 0.67% yeast nitrogen base. 0.5% casamino acids. 2% glucose. 10 p.g/ml adenine and 20 jig/ml uracil.

Yeast host cells transformed by vectors containing ADH2 promoter sequence were grown for inducing expression in a "rich" medium. An example of a rich medium is one consisting of 1% yeast extract. 2% peptone, and 1% glucose supplemented with 80 |ig/ml adenine and 80 (J.g/ml uracil. Derepression of the ADH2 promoter occurs when glucose is exhausted from the medium.

Mammalian or insect host cell culture systems could also be employed to express recombinant mammalian MGF polypeptide or derivatives thereof. Examples of suitable mammalian host cell lines include the COS-7 lines of monkey kidney cells (ATCC CRL 1651) (Gluzman et al.. Cell 25:175. 1981). L cells. C127 cells. 3T3 cells (ATCC CCL 163). Chinese hamster ovary (CHO) cells, HeLa cells, and BHK (ATCC CRL 10) cell lines. Suitable mammalian expression vectors include nontranscribed elements such as an origin of replication, a promoter sequence, an enhancer linked to the structural gene, other 5" or 3' flanking nontranscribed sequences, such as ribosome binding sites, a polyadenylarion site, splice donor and acceptor sites, and transcripdonal termination sequences.

Transcriptional and translationai control sequences in mammalian host cell expression vectors may be provided by viral sources. For example, commonly used mammalian cell promoter sequences and enhancer sequences are derived from Polyoma. Adenovirus 2, Simian Virus 40 (SV40), and human cytomegalovirus. DNA sequences derived from the SV40 viral genome, for example, SV40 origin, early and late promoter, enhancer, splice, and polyadenylarion sites may be used to provide the other generic elements required for expression of a structural gene sequence in a mammalian host cell. Viral early and late promoters are particularly useful because both are easily obtained from a viral genome as a fragment which may also contain a viral origin of replication (Fiers et al.. Naaire 273:113. 1978). Smaller or larger SV40 fragments may also be used, provided the approximately 250 bp sequence extending from the Hind III site toward the Bgl I site located in the SV40 viral origin of replication site is included. 14 238659 Purification of Recombinant Mammalian MGF MGF polypeptides may be prepared by culturing transformed host cells under culture conditions necessary to express MGF or derivatives thereof. The resulting expressed polypeptides may then be purified from culture media or cell extracts. MGF or a derivative thereof is concentrated using a commercially available protein concentration filter, for example, an Amicon or Millipore Pellicon ultrafiltration unit. Following the concentration step, the concentrate can be applied to a purification matrix such as a gel filtration medium.

Alternatively, an anion exchange resin can be employed, for example, a matrix or substrate having pendant diethvlaminoethyl (DEAE) groups. The matrices can be acrylamide. agarose, dextran, cellulose or other types commonly employed in protein purification. Alternatively, a cation exchange step can be employed. Suitable cation exchangers include various insoluble matrices comprising sulfopropvl or carboxymethyl groups. Sulfopropyl groups are preferred (e.g.. S-Sepharose® columns).

Finally, one or more reverse-phase high performance liquid chromatography (RP-HPLQ steps employing hydrophobic RP-HPLC media, e.g., silica gel having pendant methyl or other aliphatic groups, can be employed to further purify MGF. Some or all of the foregoing purification steps, in various combinations, can also be employed to provide a substantially homogeneous recombinant protein. Alternatively, some or all of the steps used in the purification procedure described above for murine MGF can also be employed.

Recombinant protein produced in bacterial culture is usually isolated by initial disruption of the host cells, centrifugation, extraction from cell pellets if an insoluble polypeptide, or from the supernatant if a soluble polypeptide, followed by one or more concentration, salting-out, ion exchange or size exclusion chromatography steps. Finally, RP-HPLC can be employed for final purification steps. Microbial cells can be disrupted by any convenient method, including freeze-thaw cycling, sonication. mechanical disruption, or use of cell lysing agents.

Transformed yeast host cells are preferably employed to express MGF as a secreted polypeptide. We mace several expression constructs for yeast host cells to secrete various fragments of MGF as a secreted polypeptide. One construct. pDCY 551, encodes an amino acid sequence comprising the first 148 amino acids of human MGF. pIXY 551 was made by mutagenizing pEXY-481 (full length human MGF extracellular domain) to remove 37 amino acids from the C-terminus. PCR reaction conditions allow one to begin with full length (185 amino acids) sequence of the extracellular domain of human MGF and make shoner fragments. pIXY 551 comprises an ADH2 promoter sequence, an a-factor leader sequence, a FLAG sequence, followed by the coding region for the first 148 amino acids of human MGF extracellular domains. We also made shoner versions of human MGF extracellular domain. For example, pIXY 552 comprises the N-terminal 133 amino acids of human MGF o 15 2 3 8 6 5 extracellular domain. pIXY 553 comprises the first 119 amino acids of human MGF N-terminal, and pIXY 554 comprises the first 104 amino acids of human MGF extracellular domain.

All of die construes were transfecred into yeast (Saccharomyces cerevisiae) strain XV2181 for expression. Yeast were grown to an optimal optical density and supematants were collected for analysis of protein and human MGF activity in a TFl cell proliferarion assay. When normalized for units per microgram of MGF protein, pIXY 551 and pIXY 490 (A28 hMGF) exhibited the highest mean specific activity of 8326 and 6938 unitsAig, respectively. By comparison, full-length 185 amino acid MGF produced 4520 units/ug. and another version comprising the first 164 amino acids of the extracellular domain produced 4978 units/ug- Shoner versions, that eliminate the fourth Cys residue, had significantly less biological activity as determined in a TFl cell proliferarion assay. For example, pIXY 553 comprising the first 133 amino acids of the N-terminal sequence had 1143 units/ug of activity while pIXY 553 (119 N-terminal amino acids) and pIXY 554 (104 N-terminal amino acids) did not produce any proliferation activity in the TFl cell assay when compared to background.

Purification of recombinant human MGF when expressed in yeast supematants involves three chromatographic steps. Yeast broth from transformed yeast making recombinant human MGF was titrated :o pH 3 with a citrate buffer and HCI. The buffered broth is diluted with water and loaded onto a cation exchange column, such as S-Sepharose® (Pharmacia) that has been equilibrated in the citrate buffer. The column is washed with citrate buffer, and MGF is eluted with 50 mM B-alanine pH 4.0 and a solution of 50 mM sodium acetate pH 5.0. The next step is a reverse phase high performance liquid chromatography (RP-HPLC) with a bonded alkyl column (preferably a C4 column). The column is washed with 0.1% TFA (trifluoroacetic acid) until the pH drops to 2.0. MGF is eluted with a linear gradient of 0-100% acetoninile in 0.1% TFA. MGF elutes at approximately 50% acetonitrile.

Fractions containing MGF from the reverse phase column were pooled and buffered with 25 mM citrate pH 3.0. MGF was applied to a cation exchange column such as S-Sepharose®, and equilibrated with 25 mM citrate pH 3.0. The column was washed with citrate buffer and eluted with an elurion buffer containing 0.5 M N:aCl in 50 mM Tris pH 8.0. This step concentrates MGF in the elution buffer.

Administration of Mammalian MGF Polypeptide and Derivative Compositions The present invention provides methods of using therapeutic compositions comprising an effective amount of MGF in a suitable diluent or carrier. For therapeutic use, purified MGF or a biologically active derivative thereof is administered to a patient, preferably a human, for treatment in a manner appropriate to the indication. Thus, for example. MGF compositions administered to suppress a form of anemia can be given by bolus injection, continuous infusion, sustained release from implants, or other suitable technique. Typically, a MGF 23 8 6 5 16 therapeutic agent will be administered in the form of a pharmaceutical composition comprising purified polypeptide in conjunction with physiologically acceptable carriers, excipients or diluents. Such carriers will be nontoxic to patients at the dosages and concentrations employed. Ordinarily, the preparation of such compositions entails combining a mammalian MGF polypeptide or derivative thereof with buffers, antioxidants such as ascorbic acid, low molecular weight (less than about 10 residues) polypeptides, proteins, amino acids. carbohydrates including glucose, sucrose or dextrans, chelating agents such as EDTA, glutathione and other stabilizers and excipients. Neutral buffered saline or saline mixed with conspecific serum albumin are exemplary appropriate diluents.

The cellular events that lead to blood all production are controlled by multiple cytokines. Many of these cytokines are produced by marrow stromal cells or by cells that traffic through the marrow companment. Hematopoietic stem cell development occurs primarily at specific sites within the marrow. Of the hematopoietic stem cells, the human CFU-B1 and HPP-CFC are believed to be the cells responsible for initiating long term hematopoiesis in vivo. Each of these cells possesses similar phenotypic properties and a number of biological behavior patterns associated with human stem cells. A variety of in vitro assays have been utilized to determine cytokine requirements of these primitive progenitor cells.

When administered alone, MGF (extracellular domain) can augment the proliferation of both myeloid and lymphoid hematopoietic progenitor cells. Most importantly. MGF exhibits potent synergistic activity in conjunction with other colony-stimulating factors that can result in increased colony number and size. More specifically. MGF can synergisticallv interact with later acting hematopoietic growth factors such as Interleukin-1 (IL-1). Interleukin-3 (IL-3), granulocyte-macrophage colony-stimulating factor (GM-CSF). Interieukin-7 (IL-7), and a fusion protein (pIXY-321) that comprises active sites of both GM-CSF and IL-3. pDCY-321 is a GM-CSF/EL-3 fusion protein havingLeu-^Asp^GIu^1 hGM-CSF/Gly^SerGly5Ser/Pro^AspI^Asp70hIL-3, and is described in U.S. Patent Aplicarion 07/567.983 filed October 14.1990, the disclosure of which is incorporated by reference herein.

A bone marrow suboooularion of CD34-- DR* CD 15" is a model for a group of Aft V * primitive hematopoietic progenitor cells. When MGF is added to this model of hematopoietic progenitor cells, a modest granulocyte-macrophage colony-stimulating activity but no erythropoietic burst or mixed colony formation activity was evident. Using other models of primitive hematopoietic progenitor cells, MGF has promoted erythropoietic bursts (macroscopic) and mixed colony formation in the presence of EPO.

Most notably. MGF was capable of synergizing with other cytokines, such as EL-3. in promoting hematopoietic cell proliferation. For example, the combination of MGF and the fusion protein pIXY-321 produced a significant increase in the number of CFU-GM and HPP-CFC when compared to the colonies formed with just pIXY-321 alone. Table I below presents data from a comparison of CFU-GM. BFU-E. CFU-GEMM. HPP-CFC. and BFU- 23 8 6 5 9 MK colony activities stimulated by various cytoidnes in a population of CD34- DR- CD 15-human progenitor cells.

Table 1 Cytokines Colonies per 1 Ql CD?— PR- CD 1 y Cells CFU-GM BFU-E CFU-GEMM HPP-CFC BFU-MK None 0.5-/- 0.7 12.5 -/- 3.6 040 0+/-0 0 ♦/- 0 IL-3 23.0 -/- 1.4 31.5-/- 5.0 2.5 -/- 0.7 0+/-0 .S -/0 0.3 GM-CSF 56.0 -/- 7.1 24.0 */- 7.1 3.5 »/- 0.7 1.5 -/- 0.7 16+/- 0.9 EL-3 + GM-CSF ■iSo -/- 5.0 .0 -/- U.2 2.5 -/- 0.7 1.5 -/- 0.7 7.9 -/- 1.1 pIXY-321 •12.5 W- 3.6 o o 1 1 o 6.0 -/- 1.4 3.5 -/- 3.1 9.8 +/- 0.7 MGF 18.5 W- 2.1 14.0 -/- 43 0-/-0 040 0+/-0 MGF * EL-3 - - - - 9.0 ♦/- 0.4 MGF + pIXY-321 100.5 -/- 3.6 36.5 -/- 3.6 -/- 2.1 23 J */- 10.7 9.8 +/- 0.5 CFU-GM. BFU-E. CFU-GEMM. and HPP-CFC were assayed with about 103 CD34- DR-CD!5~ marrow cells in methvl cellulose containing one unit of human recombinant * 9 erythropoietin and enumerated at 21 days, or 23 days for HPP-CFC. Concentrations of cytokine included 2 ug/ml for EL-3 (specific activity 3.5 x 10s CFU/MG protein). 1 ng/ml for GM-CSF (specific activity 2.S x 10$ CFU/MG protein). 10 ng/ml forpDCY-321 (specific activity 1-2 x 109 CFU/MG protein), 50 ng/ml for murine MGF (specific activity 10^ CFU/MG protein). These data were performed in duplicate to obtain the standard deviations indicated. The study was repeated, and similar data were obtained.

MGF alone was capable of sustaining long term myelopoiesis but had a limited affect on megakaryoevtopoiesis and erythropoiesis. However, when either pIXY-321 or EL-3 was added with MGF, a profound synergistic effect in promoting progenitor cell expansion of all three hematopoietic lineages was found. The most pronounced effect was seen with the combination of MGF and pIXY-321, which led to a 50-1000 fold increase in progenitor numbers over a 10 week period. MGF was capable of synergizing with EL-3 in promoting megakaryocyte progenitor cell expansion, but was incapable of substantially augmenting maximal proliferative stimulus supplied by pIXY-321 or a combination of GM-CSF and EL-3. Further analyses of these data comparing the effects of MGF alone and in combination with other cytokines showed that although MGF alone had a limited effect upon cellular proliferation of hematopoietic progenitor cells, the presence of MGF within a combination of a later-acting cytokine (e.g.. EL-lct. IL-1S, IL-6. IL-7 . IL-4, IL-3. EPO. LIF, GM-CSF, and a EL-3/GM-CSF fusion protein such as pIXY-321) favored the presence of MGF for the most primitive hematopoietic progenitor ceil expansion. Thus, combinations including MGF and other later- IS 238659 acting cytokines are necessary for maximal proliferation and differentiation of hematopoietic progenitors.

When various cytokines were evaluated in long term cultures of hematopoietic progenitor cells, hematopoiesis was sustained for six to ten weeks in the presence of either IL-3, GM-CSF, pIXY-321, IL-3 plus GM-CSF. or MGF. The combination of EL-3 and GM-CSF was additive, but less than the effect of pIXY-321. Only the combination of MGF plus EL-3 or MGF plus pIXY-321 provided a dramatic synergy, resulting in a 103 - lCP-fold increase in cell numbers over an original 5 x iO3 cellular inoculum. Therefore. MGF in combination with a later-acting cytokine such as IL-3 or a fusion protein of EL-3 and GM-CSF, promotes cellular proliferation and the presence of MGF in that cytokine combination favors the persistence of a more primitive hematopoietic cell type.

Human megakaryoevtopoiesis is regulated by a group of cytokines, including GM-CSF, EL-la, IL-3. IL-6. and MGF. Megakaryoevtopoiesis can be measured by CFU-MK and BFU-MK. Each in vivo measurement determines a level of progenitor cell development leading to platelet formation. Defects in megakaryoevtopoiesis, on the level of BFU-MK or CFU-MK, ultimately lead to various bleeding disorders due to a failure of proper platelet formation or recruitment. MGF affects multiple sages of megakaryocyte development, and synergistically augments colony stimulating activity of various cytokines, including pIXY-321, GM-CSF, EL-3 and EL-6. MGF in combination with IL-3 potentiates the ability of EL-3 to promote colony formation at each level of progenitor cell development (BFU-MK and CFU-MK). Optimal concentrations of GM-CSF, EL-3, pIXY-321, EL-lex, and EL-3 were roughly equivalent when measuring BFU-MK in vitro, and were not further enhanced by the addition of MGF. However, at the level of CFU-MK, MGF synergistically enhanced the ability of pIXY-321, GM-CSF. IL-3 and IL-6 to promote megakaryocyte colony formation in vivo. Maximal human megakaryoevtopoiesis was achieved with the triple combination of EL-3, GM-CSF and MGF, or the combination of MGF plus a GM-CSF/IL-3 fusion protein such as pDCY-321.

MGF also stimulates erythroid proliferation in the presence of erythropoietin (EPO). In an in vivo model with BFU-E derived colonies from patients with sickle cell anemia (a hemolytic anemia) and Diamond-BIackfann anemia (a hypoproiiferative anemia), MGF increased the mean number of BFU-E derived colonies with both types of anemia. MGF at a concentration of 50-100 ng/ml can increase fetal hemoglobin synthesis in sickle cell anemia. The stimulation effect of MGF was decreased by concurrent addition of Interferon-y. The stimulation of BFU-E caused by MGF in the presence of erythropoietin, was evident by measuring both colony number ana size, anc was significantly increased in a dose-dependent manner according to concentration of MGF. Therefore, the combination of MGF and EPO is useful for treating hemolytic anemias and hypoproiiferative anemias. 23 8 6 5 19 MGF can also be effective for stimulating lymphoid precursor cells, and influencing early lymphoid development. MGF stimulated thymidine incorporation in day 13 fetal thymus cells. Fetal thymus cells are a source of lymphoid precursor cells. In addition to MGF. both Interleukin-2 (EL-2) and IL-7 can stimulate significant tritiated thymidine incorporation in day 13 fetal thymus cells. The effects of the combination of MGF and IL-2 were additive, while the effects of MGF plus IL-7 were greater-than-additive. When compared with background, day 13 fetal thymus cell expansion after 9 days of incubation with cytokine, were three-fold higher than background for MGF, three-fold higher than background for IL-7, and 103-fold higher than background for the combination of MGF and EL-7. Fetal thymus cells that were expanded by the combination of MGF and EL-7 were CD3- CD4* CD8* cells. The combination of MGF plus EL-7 stimulated pre B cell colonies two to four-fold above that seen with IL-7 alone. These data illustrate the importance of combining MGF with IL-7 to augment and proliferate early lymphoid cells.

MGF may have further function in organization of the neural tube and brain. More importantly. MGF mediates proliferation and differentiation of fetal neural tube cells to promote development of fetal brain and brain in newborns. Moreover. MGF effects melanocyte function to regulate differentiation and proliferation of melanocytic functions.

The following examples are for purposes of illustration and not by way of limitation. 238659 EXAMPLE 1 PURIFICATION OF MURINE MGF This example describes purification of murine MGF to homogeneity for amino acid sequencing. 45 liters of conditioned medium was obtained from normal WCB6F1 -r/r cells grown in RPMI1640 supplemented with 10% fetal calf serum The conditioned medium was first acidified to pH 2.75 by adding concentrated HC1 and buffered with Na citrate. pH 3.25. to a final concentration of 5 mM Na citrate. The acidified conditioned medium was applied to an S-Sepharose® Fast Flow (Pharmacia) column equilibrated with 0.1 M NaCl 5 mM Na citrate, pH 3.25. After non-bound proteins cleared the column, bound protein was eluted with 0.1 M EPPS (N [2-hydroxyethyl]-piperazine-N'-[3-propane-sulfonic acid] Sigma Chemical Co.). pH 8-5. The eluted protein pool was diluted 1:3 with water and loaded on a DEAE Sephacel (Pharmacia) column equilibrated with 10 mM EPPS. pH 8.5. Protein bound to the column was eluted with a salt gradient (0 - 0.5 M NaCI in 10 mM EPPS. pH 8.5). Fractions were collected and bioassayed using MC6 cells in a biological assay as described herein. Fractions having an activity of at least about 10 units per ml were deemed "active" fractions and were pooled for further purification procedures.

The protein pool was diluted with 2M ammonium sulfate 20 mM Tris HQ, pH 7.5. and centrifuged to remove precipitated proteins. Tne pooled protein was then applied to a Phenyl Sepharose® CL-4B (Pharmacia) column equilibrated with 2 M ammonium sulfate 20 mM Tris HCI, pH 7.5. Bound proteins were eluted using a reverse gradient of 2 M - 0.5 M ammonium sulfate in 20 mM Tris HCI. pH 15. Fractions were collected, dialyzed against PBS (using a 6000-8000 dalton cut-off membrane), and assayed for MGF biological activity using the MC6 cell assay described herein. Fractions containing MGF were pooled for further purification procedures.

Pooled active fractions from the Phenyl Sepharose purification were buffer exchanged with 15 mM CaCl2 20 mM Hepes, pH 7.5. and concentrated about 10 to 12 rimes using an Amicon stirred cell with a YM 10 membrane. The concentrated Phenyl Sepharose pool was passed over a Matrex® gel Blue A (Amicon Corp.) column washed with 5 mM CaCl2 20 mM Hepes, pH 7.5, to remove additional contaminating proteins that bind to the Blue A column. The Blue A pool was acidified by addition of trifluorocetic acid (TFA) while vortexing to a final TFA concentration of about 0.1% (v/v) and filtered through a .022 ll cellulose acetate filter. The acidified protein pool was pumped directly onto a Radial-Pak Vvdac® C-4 column (15 jo. particle size) pre-eauilibrated with 0.1 % TFA in water at a flow rate of about 0.8 ml/min. Bound proteins were eluted using a 1%/min linear gradient of acetonitrile (0.1%, v/v TFA). One minute fractions were collected, neutralized with a Tris buffer, and bioassayed for MGF biological activity.

Active fractions were pooled, and the protein pool was concentrated by rotary evaporation to about 10% of its original volume. The concentrated protein pool was applied to 258659 21 an Aquapore® Butyl C-4 column (7 u. particle size) pre-equilibrated with 0.1% TFA in water, at a flow rate of about 0.2 ml/min. Protein bound to the column was eluted with a .05%/min linear gradient of acetonitrile (0.1%, v/v TFA). Fractions were collected, neutralized, and bioassayed for MGF activity.

Active fractions were pooled for SDS-PAGE. Protein fractions exhibidng MGF activity were collected from 2 runs of the purification protocol described above and concentrated to dryness by rotary evaporation. The protein was resuspended in gel sample buffer and heated to 110*C for five minutes. The protein was then separated using a 10% Laemli gel and transferred onto a PVDF membrane (Pro Blot, Applied Biosystems) using constant current, 60V setting, for 1.0 hour. Three protein bands were visualized by staining the PVDF membrane for five minutes with 0.1% Coomassie Blue in 10% acetic acid. 50% methanol.

The membrane was destained in the acetic acid/methanol solution without Coomassie blue stain, washed extensively with HPLC water, and dried. Protein bands were excised with a razor and placed onto a pre-cvcled filter for a model 477 protein sequencer (Applied Biosystems. Foster City. CA). for N-terminal amino acid sequencing.

EXAMPLE 2 CLONING A MURINE MGF cDNA Polyadenylated RNA was prepared from -r/+ cells (described in Boswell et al. Blood 70: 167a 1987) and cDNAs were prepared using standard techniques. The +/■?• cell line produces murine MGF. cDNA ends were adapted with Sal 1 adapters (Haymerle et aL Nucleic Acid Res. 14:8615-24 1986): *-TCGACTGGAACGAGACGACCTGC7-3' 3'-GACCTTGCTCTGC7GGACGA-5' and cloned into vector HAVEo. A pool consisting of approximately 5000 individual plasmid-containing isolates was plated and screened by hybridization to a DNA fragment. The DNA fragment was prepared using polymerase chain reaction (PCR) amplification of murine MGF sequences from +/+ and LDA11 cell line cDN'A as follows.

The sequence of the N-terminal 28 amino acids of purified murine MGF was used to design synthetic oligonucleotide primers for PCR amplification of murine MGF-cDNA clones from a murine library described below. Tne first five amino acids of the N-terminus (Lys Giu lie Cys Gly) were used to design one primer, 5'- CGCCCGGGAA(G/A)GA(G/A)AT(A/C/T)TG(T/C)GG-3'. a degenerate mixture coding for all possible codon usages of the first five amino acid residues, omitting the third position of Gly and containing an additional eight bases coding for a Smal recognition site and a 5' CG 236659 22 clamp. The amino acid sequences of the mouse mature N-terminus 23-28 (Pro Asn Asp Tyr Met He) were used to design a second primer, a degenerate mixture coding for a complement of all possible codon usages of amino acids 23-28. omitting position 3 of lie and containing an Asp 718 recognition site and a 5' CGC clamp: 5*-CGCGGTACCA.TCA.T(G/A)TA(G/A)TC(G/A)TT(G/A/C/T)GG-3'.

Polyadenyiated RNA from mouse secretor cell lines and LDA11. stimulated under a variety of conditions (LDA11,45% McCoy's medium. 45% HL-1 medium. 10% PBS. 6 hrs; LDA11, HL-1 medium, 28 hrs; LDA11, HL-1 medium, 5.5 hrs; +/-K Opti-MEM medium 9 hrs), was used as separate templates for first strand cDNA synthesis. A portion of first strand cDNA reactions was added to commercially available PCR reaction mixes containing the oligonucleotide primers. This mixture was subjected to 30 cycles of PCR amplification.

Following amplification, samples were purified and subjected to agarose gel electrophoresis. This yielded a 100 base pair DNA fragment thai was excised from gel lanes from four separate reactions involving -rh cells and LDA11 cells. The 100 base pair DNA fragment was purified using an Elutip-D® column (Schleicher & Schuell, Keene. NH), cloned into pBluescript® SK (Stratagene, La Jolla. CA) and used for dideoxy DNA sequencing.

We prepared a hybridization probe by random prime labeling of the purified 100 base pair DNA fragment and used the hybridization probe to screen a portion of the plasmid library containing cDNA inserts prepared from ^!~ polyadenyiated RNA. Several positive clones were further characterized, including MGF isolate 4, isolate 10. and isolate 94. Isolate 4 contained sequence information for the coding region minus the last four amino acids of the C terminus. Isolate 10 contained the entire coding region for the mouse polypeptide and is illustrated in Figure 2. Isolate 94 has a 16 amino acid deletion in the extracellular domain compared to isolates 4 and 10.

EXAMPLE 3 IDENTIFICATION OF MGF AS THE LIGAND FOR C-Kit In order to demonstrate that MGF is the ligand for c-kit, murine MGF purified as described in Example 1 was bound to cells expressing c-kit. The resulting c-kit]MGF complexes were covalentlv cross-linked and immunoprecipitated using antibody specific for c-kir. Radioactivity was recovered in the immunoprecipitated fraction.

We first cloned murine c-&: cDNA using polymerase chain reacnon from poly A"8" selected mRNA isolated from MC6 cells (an MGF responder cell line). A 3.0 kilobase cDNA was sequenced and compared to the published murine c-kit sequence (Qui et al. EMBO J. 7:1003 1988). Our clone had two amino acid changes from the published sequence, including an alanine instead of a glutamic acid at position 207 and a glycine instead of an alanine at amino acid 777 (numbered according to Qui et al.). 23 8 6 5 23 We generated anti-pepace antisera to our murine c-kit protein. C-terminaJ peptides corresponding to the last 10 amino acids of the published sequences, amino acids 76-89 and amino acids 730-743 were synthesized. The synthesized pepddes were conjugated with ovalbumin and injected into rabbits. Serum used for immunoprecipitadon experiments was obtained from the rabbits after an initial injection and four subsequent boose. All three antisera, corresponding to the three different pepddes. precipitated a protein of 145 kDa from transfected Rat-2 cells stably expressing c-kit. whereas no proteins were precipitated with preimmune serum or from non-transfec:ed control Rat-2 cells.

Radiolabeled purified mouse MGF was added to MC6 (responder cell line expressing c-ifeir as described herein) and 32D cells (non-responder cell line that docs not express c-kiz) with and without excess unlabeled MGF. MGF was cross-linked to its specific cell surface receptor by the cross-linking agent BS3. Cell lysates were immunoprecipitated with die specific antisera described above, which recognize the C-terminal portion of c-kit, then separated on an SDS-PAGE gel and subjected to autoradiography. Cross-linked receptor-ligand complexes were specifically immunoprecipitated by anti-c-&r antisera only from MC6 cell lysates incubated in the absence of cold-comperitor and had an approximate molecular weight of 175-180 kDa as determined by an autoradiogram of cross-linked receptor (c-kit) ligand (in MGF) complexes (See Figure 6). Preimmune sera did not immunoprecipitate this complex and binding in the presence of excess unlabeled MGF provided no cross-linked complexes detectable in autoradiography. No evidence of MGF binding to 32D cells was seen.

When one subtracts the molecular weight of the c-kit protein from the total molecular weight of the complex, the suggested molecular weight of the ligand is from about 30 kDa to about 35 kDa. These data are consistent with the molecular weight determination of recombinant MGF.

EXAMPLE 4 CROSS HYBRIDIZATION OF A MOUSE MGF PROBE TO A HUMAN LIBRARY A probe was prepared from a cDN:A insert of MGF isolate 10 (an approximately 2 kb Sali fragment) by random prime labeling with a-32p.<iCTP to a specific activity of greater than 5 x l(P dpm/ng. The probe was hybridized to human RNAs on Northern blots in buffer (50% formamide, 0.8 M NaCI, 0.02 M PIPES pH 6.4, 2 mM EDTA, 0.5% sodium dodecyl sulfate (SDS), 100 p.g/mi salmon sperm DNA and 5x1 (P dpm/mi of labelled DNA) at approximately 42"C for greater than 20 hours. Hybridization was followed by extensive washing in 1 X SSC (15 mM trisodium citrate. 165 mM NaCI), 0.1% SDS at approximately 55*C.

We found specific bands by autoradiography of probed and washed blots containing polyadenyiated tnRNA from the human cell lines RAJI (B cell lymphoma), WI26-VA4 (SV40-transformed lung fibroblast), HeLa (ovarian carcinoma), T24 (bladder carcinoma). IMTLH 238659 24 (SV40-transformed stromal line). IE9 (a subclone of IMTLH), peripheral blood T cells (PBT) stimulated with PHA and PMA. PBT grown 6 days in IL-2 and OKT3 MAb then stimulated with ionomycin and PMA. and PBT grown for six days in IL-2 and OKT3 then stimulated with concanavalin A and PMA.

We observed cross hybridization to specific bands in human genomic DNA (Southern blots) when the MGF isolate 10 probe was added in hybridizarion buffer containing 30% formamide. with hybridization conditions as described in this example at approximately 42*C. followed by extensive washes in 0.5 X SSC. 0.1% SDS at approximately 63*C.

We have used an MGF-10 probe in cross hybridization in a buffer containing 40% formamide with hybridizarion conditions as described in this example at approximately 37*C followed by extensive washes in 1 x SSC, 0.1% SDS at approximately 55*C to screen a PBL library. We isolated MGF-positive human clones from this library.

These cDNAs contain only a portion of the 3' MGF coding region, based on a comparison with mMGF. Full length hMGF coding cDN'As were isolated by polymerase chain reaction (PCR) amplification of hMGF cDNAs from HeLa mRNA using oligonucleotide primers designed to specifically amplify hMGF coding cDNAs. The partial cDNA sequence derived from the PBL library clone was used to design a 3' non-coding region oligonucleotide primer (5*-AATGTTACCAGCCAATGTACG-3') which is complementary to sequences located 30 bp downstream of the end of translation in the hMGF cDNAs (nucleotides 878-898 of hMGF-2.4, Figure 3). The 5* non-coding oligonucleotide primer (5'-GGGCTGGATCGCAGCGC-3") was derived from the mMGF 5" non-coding region (nucleotides 122-138 of MGF-10. Figure I). These oligonucleotides were used to amplify hMGF cDNAs from HeLa in a reaction buffer containing 50 mM KCI, 10 mM Tris pH 8.3, 0.01% gelatin. 1.5 mM MgCb 200 uM dNTPs, 1 uM each oligonucleotide primer, 2.5 units Taq polymerase, and approximately 20 ng first strand HeLa cDNA. Reactions were performed on an Ericomp TwinBlock® temperature cycler (Ericomp, San Diego, CA) for 35 cycles of 94 *C for 30 seconds. 45 *C for 45 seconds and 72 *C for 60 seconds, followed by an additional 5 minutes of extension at 72 *C. Agarose gel electrophoresis of the products of PCR amplification snowed that at least four distinct sizes of cDNA were isolated. The most abundant size class migrated in the gel at approximately 850 bp. Tne second most abundant size class migrated at approximately 800 bp. Tnese cDNA products were subcloned and sequenced. Isolate hMGF-2.4 corresponds to the most abundant cDNA from the PCR amplification. Isolate hMGF-2D corresponds the the next most abundant cDNA.

Analysis of mouse MGF ana human MGF shows approximately 89% nucleotide sequence identity and approximately 83% amino acid sequence identity between human and murine MGF sequences.

We can also use the isolate 10 clone (mMGF-10) as a cross-hybridization reagent at approximately 63"C in buffers containing 50% or less formamide. followed by washes in 2 X 238659 or less SSC at 50*C to 63*C. DNA probes can also be prepared by nick translation of mouse MGF cDNA by polymerase chain reaction (PCR) amplification using MGF-specific primers and radiolabeled d NTPs, and by preparation of radiolabeled cDNA using synthetic MGF RNA as a template- The DNA probes can be hybridized in hybridization solutions containing 50% or less formamide at temperatures ranging from 37*C to 50*C. followed by washes at 50*C or above in 2 X SSC or less.

EXAMPI.F 5 EXPRESSION OF MGF This example illustrates the expression of full length murine MGF in COS cells (derived from African Green Monkey kidney cells). Isolate 10 and isolate 4 were transfected into COS cells using standard techniques, such as those described in Cosman et al. Nature 312:768 (1984). The transfected COS cells wex seeded into 96 well plates. Responder cells (e.g.. MC6 cells) that proliferate in response :o MGF were added to the wells and incubated overnight Tririaied thymidine was added to the wells at a concentration of 0.5 jiCi/well. COS cells transfected with HAVEo empty vector were incubated in parallel as a control. A positive control (not shown) consisted of COS cells transfected with HAVEo engineered to express a murine IL-4 cDNA. The MC6 responder cell line also proliferates in the presence of IL-4. Proliferative activity significantly increased (p < 0.001) in the presence of MGF when compared to empty vector controls. For example, clone 4 induced proliferative activity in MC6 cells of about 28.000 cpm compared :o about 13,000 for an empty vector control Both clones 10 and 4 provided mMGF polypeptides with about 13,000 cpm compared to about 4,000 for empty vector control.

We also expressed the extracellular region of murine MGF in COS cells. We prepared a Sst 1 - Bgl 1 fragment encoding the extracellular portion of mMGF through Ala 179 (Figure 2). This is five amino acids N-terminai from the transmembrane region. This fragments was blunted with T4 polymerase and inserted into a vector pDC402 to generate the plasmid PDC402/sMGF-1. Fifteen additional amino acids (Val Asp Gly Pro Cys Gly Arg Tyr Arg Ser Thr Arg Phe Asp Val) and a termination codon were provided by the vector downstream of the coding site or toward the C-terminus of the truncated polypeptide. COS-7 cells were transfected with either the vector containing the truncated cDNA, with vector containing a murine interleukin-4 cDNA (pDC402/mIL-4), or with vector lacking an insert. The transfected COS cells were metabolicallv labeled with ^S-methionine and 35s-cysteine. Supematants were analyzed by SDS-PAGE for size determination of the secreted polypeptide and found a 33 kDa protein (Extracellular region plus glycosvlation). Supematants were also tested for MGF biological activity using the MC6 proliferation assay described herein. The secreted extracellular portion of murine MGF was found to possess MGF biological activity. ,z66tNnrJ o (B9c e5sd Aq pswoxio^) utsj^s jstsa -uiruiop jciniporaxs £9K ununu gjx $9V"U * pspoous Q5t> "AXF srsjsu/ft 'urcmop .rqnipon^s jr)j\ uruinu tpsus] ipj ®9YId * f»poow SZC~AXId •uoprouund jo; sousnbss sppcsdrcoo §£)vh c ?L~ 's^snbss jsprsi jojdcj-s ur 'sousnbss jssouioad cHQYr «u: suaaioa joiosa qh -AXI^ r °-u: YNQD dDK insnnu r Suwssui Xq sprra sjsm soiuisuod ipog "06r-AX3C pirc SIC-AXI^ "sioonsuoo uoissajdxs 2srsA o*u pssn ?A\ 'UF1120? JCinnsoraxs sip u: sousnbas ppc ouiuit g- r s>p~ udiua\ 'JOI'W ScV P^ dOK uraaop jcpip3ci:x? I]n: ununq jo uoisssjcxs 2srsA ssquossp sidnirxs siuj. 1SV3A *a iOK KVKflH iO M01SS3H«IX3 y aidlWXH (?> -sppcsdApd £9K rr:rucu suunm psgund sip rncuj psApsp inp 03 popuspi stm up2Qjd uoisn: psausp-2srsa juruiquioss: sip 03 sssusaisuocssj 10 uisurd sip nncsy "Sipo 93KJO prsisui ai*ccOCH P"= IK ssu!i IF^ SAisuodssj-^j^ 'apnea ipia-. pssrsca: sis/u rarp sssiq. 'JQK supnm pjruru ipi/*\ psaissoo irip 03 jropuspi scaa. sjpo 93K ag uopuodjooui sinpiuiXip uinuiprcu; 3U2 'psyund seat. qscuq 2srsX Pip usuj^. 'ipojq 2srsX sip in uoprjsinojd jqnjpo jo sjojiqiuui orjiDsds-uou uo sousssjc sip 03 snp ussq sarq avw. siip mq '£DK 3uruiquioos.: 20 uoprausouoo 2ssqfiij sip iz usss uoprjodioom suiprmAtp :o uoponpsj sujos stm sjsqi -uoiusrj ssuodssj-ssop r u: sjpo 93ft :o uopr»iflajd ps3rpun:s £3K suimra psuunc par J9J\ lurutauioss-* tpog acsst uoptusinoad 93]^ sip in psicdcuoo sr«\ supraa psuunc pur sprai-2srsa 3iiruiqiac»: 10 ajiaudx: fcoisoioig -rutins "jx:? adou ui psquossp sampsooad Suisn audtug02rtuaiq0 Aiiuyjr Xq ipojq istsa ujcu: psuund srm uisscwd uoisiu £9k suuucn Supjnssj sqx 'uopri/CsooAjS pssjini-fsj jo: ssiis icncjjod jno: sq3 :o suo Supriraui]s 'suuss 02 691 suiScJcdsr psusAUOO qoiqA\ 'v.njq psjrjsuso-'ij^d sip 10 sousnbss sq3 ui uoprjsjjB suo skm a»uj, *(1861) LSZ:££ ?'a?Q "F » 3^Pd) uinipsu: istsa ojui ^jnpojc jucuiqmoosj sip 30 uocsjdss sidsjip jxnp joidsa uoisssjdxs isrsA t oju] ps3r§n P^ p23cj3us§--g3j srm jusmSrjg VMa s^JL "uoiSsj sircjquisuisun: sip uioj; un:sj2sdn 2sn[ '(z ajnSia) csi-dsv Supwonoj uopoo uoncuiui»2 r psjrss C£-JJJLDODDDUJL3DDDODVOJLLDXDVJL33D1VDD1VJ.V - c) sppospnuoSii0 .£ 'ilL "£DK i° snuruus3 -n sjiuctu sip ipiM smcij-u: pssnj (SS6I) rOZI:9 £So]Oiw3?jjO]S -jc 33 ddon x®ovti) sppdsdcoo oiusSpur ur par jsprs] Jo:or: rqdp: ur pspoou? (.£-0191331WDDDDDIO IVOVDOWDWOVOIVODVODVODWOVJDVOVDVWVIVODULLDDVIIIOIVV - c) jsuiud c v "HOK 30 uraaop Supoo icpipornxs sjpus sip Aiiidun: 03 pssn sjs.w s:suiud sppospnuo§n° apsipuXg *tu33sas uoisssadxs2srsA r u: p320iU2SU0D srm '(z sjrviii^ 'egpi sppr ounur) spncsdA]od sq3 jo urcmop jrjniporaxs sip ajuo Suiurauoo'£9k 3° uoisjsa sjanps jsqjouy 6£98£Z 9z * 26a XV2181 is a diploid strain prepared by mating strain X2181-1B (available from the Yeast Genetic Stock Center, Department of Biophysics and Medical Physics, University of California at Berkeley, California, USA) and strain XV617 (available from the Department of Genetics, University of Washington, Seattle, Washington, USA). Strain YNN281 is available from the Yeast Genetic Stock Center, Department of Biophysics and Medical Physics, University of California at Berkeley, California, USA. Both strains are standard laboratory strains of S. cerevisiae. Yeast (SaccAaroznyces cerevisiae) host strains XV2181 and YNN281 were transformed with pIXY-528 or pIXY-490 (followed by page 27) * 238659 Material made from pIXY-528 showed a high degree of hypcrglycosvlanon and a lower level of protein expression than the material made from pIXY-^90. A controlled experiment was conducted utilizing host strain XV2181 and an ELISA assay for total MGF expression in yeast supematants. The yeast were grown to the same optical density and supematants were purified. pEXY-528 produced approximately 50 mg/liter of human MGF whereas pDCY-490. grown under identical conditions with the same expression vector in the same yeast host strain, produced five times more protein, or about 250 mg/L Funher examination of the material produced by pDCY-490 shows virtually no hyperglycosylacon judging by two bands visible on a FLAG Western blot. Tne material produced by pIXY-52S was hvperglycosylated with at least five bands visible on a FLAG Western blot.

Full length human MGF has five potential N-linked glycosyladon sites, whereas A28 human MGF has four. Funher. multiple serine residues have been removed from the A2S version. It is believed that die series of serine residues function as O-linked glycosyladon sites. Removal of the fifth N-linked glycosvlation site and serine-rich domain in the A2S version of human MGF. surprisingly, resulted in reduced glycosvlation, greater MGF biological activity, and approximately five times greater amounts of protein expression when grown under identical conditions.

This example illustrates that MC-F is the ligand for c-.iir. The experiment described in Example 3 was repeated, except, the c-!cit clone was transfected into COS ceils (along with an empty vector control) and the transfected COS cells were used instead of MC6 responder ceils. The procedures described in Example 5 were followed. We found precipitation of iodinated MGF cross-linked to c-ki: only in those COS cells transfected with c-kit. Tnese data provide funher evidence that MGF is a ligand for c-kit.

This example compares coiony formation induced by various cytokines and controls. Two assays were employed. In the first assay, granulocyte-macrophage colony- and cluster-fonning cells (CFU-GM) were assayed by a procedure substantially similar to that described in Williams et aL Exp. Hematol. 75:2-3 (1987). Briefly, bone marrow ceils were suspended in 0.5 ml of 0.3% agar (Difco, Detroit MI) or 0A% agarose (FMC. Rockland. ME) culture medium containing McCoy's 5A medium supplemented with essential and nonessential amino acids, siutamine, serine, asoaragine, and sodium ovruvate (Gxbco) with 20% fetal bovine^ <& * t j * * VJ k V EXAMPLE 7 BINDING TO EXPRESSED c-A'fr EXAMPLE * MGF STIMULATES COLONY FORMATION IN VITRO 23 8 6 5 28 scrum. Quadruplicate cultures were incubated for seven days in a fully humidified atmosphere of 5% CO2 in air. Colonies (> 50 ceils) and clusters (3-49 cells) were counted with an inverted microscope at 32 X.

In the sccond assay, erythroid burst-forming units (BFU-E) and muldpotennal colony-forming cells (CFU-GEMM) were assayed by a procedure described in Williams ct al.. supra. Briefly, duplicate 35 x 10 mm cultures were stimulated with 2 units per ml of recombinant human erythropoietin (Hyclone). 0.1 mM hemin (Kodak), and 1000 U/ml EL-3 as a source of burst promoting activity. Cultures were incubated for seven days in a fully humidified atmosphere of 5% CC>2.5% O2 and 90% NT2- Maximal colony formation by murine BFU-E and CFU-GEMM was observed on day 7 using the foregoing culture conditions. BFU-E and CFU-GEMM were scored using an inverted stage microscope at 80 X on the basis of hcmoglobinizadon of erythroid elements, producing a characteristic red color, and the absence or presence of myeloid elements for the former and latter, respectively.

Using the foregoing assay, we compared the colony-forming activity of murine EL-3. human EPO, murine MGF at a 1:100 dilution. IL-3 plus EPO. and MGF plus EPO. The results are set forth in the following table: Table 2 Colonv Forming Activity of Cytokines Including MGF Cytokine BFU-E/lte cells CFU-GEMM/102. cell none 0 0 IL-3 0 0 EPO 13 +/-1 1+/-1 MGF (1:100) 0 0 IL-3 + EPO 21 1 8+/-1 MGF + EPO 16 -r/- 2 8+/- 1 MGF + IL-3 16 +/- 2 8+/- 1 These data indicate that MGF stimulates erythroid and primitive mixed colony formation. In another experiment we compared CFU-GM activity of MGF and GM-CSF versus no cytokine control. MGF had significant CFU-GM stimulatory activity (more than GM-CSF) with a stimulation index over background above 50. Moreover, as shown in Figure 5, the BFU-E and CFU-GEMM activities of MGF increase with increasing doses of MGF. This is in contrast to EPO (erythropoietin) whose 3FU-E activity remains constant across the dose range. 238659 29 EXAMPLE <> MGF PLUS IL-3 SYNERGISTICALLY STIMULATES PRIMITIVE STEM CELLS Soncd hematopoietic stem cells can be further soned for metabolic activity. A mitochondrial stain rhodamine 123 stains mitochondria and allows sorting of "bright" cells that arc metabolically active and "dull" cells that are not metabolically active. The sorting procedure leading to bright and dull cells is a modification of the procedure described in Visser et aL Blood Cells 14:369 (1988). Briefly, a low density (< 1.055 g/cnP) bone marrow cell fraction from untreated animals (e.g.. mice) is isolated by spinning the cells over a discontinuous density gradient of meirizamide. During centrifugarion, the cells are stained with WGA/Tx Red. The WGA/Tx Red stained low density cells are next stained with 15-1.4.1/FTTC. Stem cells are selected on a cell sorter by detecting for Tx Red"*" cells, setting a FLS window on the WGA"*" cells to exclude the smallest and largest cells and cell aggregates, selecting ceils having the lowest SSC when looking ar the SSC intensity distribution of the cells in the TxR/FLS window, and selecting for negative cells when looking at the 15-1A1 (FITC) fluorescence distribution of cells in the Tx Rea/FLS/SSC window. After the sort, WGA/Tx Red conjugates on removal and the cells are stained with Rhodamine 123. The cells arc again sorted in the same FLS and SSC windows as before on the basis of their Rhl23 distribution into Rhl23 dull and Rhl23 bright fractions. Dull cells have marrow rcpopulating activity and brights do not Bright cells contain most of the in vitro colony forming cells.

Bright cells and dull cells were compared for proliferative activity in a tritiated thymidine incorporation assay. Approximately 1000 cells were incubated in the presence of mMGF, mIL-3 or mMGF plus mIL-3. Table 3 below shows the synergistic proliferation activity of MGF plus EL-3.

Tnhle 3 CPM GROWTH FACTOR Medium control mMGF (33 U/ml) mIL-3 (100 ng/ml) IL-3 ♦ MGF DULL 668(139) 2569 (874) 9371 (770) 35132 (3098) BRIGHT 380(87) 5166 (748) 34472 (13226) 91112 (33475) These data show the synergistic effect on stem cell proliferation in both the bright and dull populations with the combination of MGF plus IL-3. These data were initially obtained with purified murine MGF as described herein and later repeated with purified recombinant yeast-derived murine MGF as described herein. Both sources of murine MGF provided the same results. 23 8 6 5 The proliferarion assay was repeated with MGF. DL-3 and the combination of MGF ana IL-3 as described in this example, except proliferation was measured by viable cell counting with trypan blue exclusion. The same synergistic activity of the combination was seen with both bright and dull populations of stem cells. These data are shown in Figure 6a.

We examined the activity of mMGF. mEL-3 and the combination of both in a spleen-colony forming assay with primitive stem cells (CFU-S). The assay looks for functional activity of a population of primitive hematopoietic cells. The procedure is described in Visser et al. supra. Briefly, mice are lethallv irradiated to wipe out bone marrow ceils. Cells (i.e., dulls or brights) are injected into each irradiated mouse. After 14 days, the mice are sacrificed and the spleens are examined and counted for macroscopic colonies on the surface of the spleen. Each colony is derived from a single primitive hematopoietic colony-forming cell (CFU-S). The injected primitive cells were either control (freshly isolated), or treated for three or seven days with mMGF. mIL-3. or the combination of mMGF and mEL-3. The concentrations of mMGF (50 U/ml) and mEL-3 (250 ng/ml) were plateau amounts in colony assays and proliferation assays. A Unit (U) of murine MGF is the amount of mMGF that stimulates half maximal tritiated thymidine incorporation by the MC6 ceil line.

Both mMGF and mIL-3 when used alone maintained CFU-S activity for their three days of incubation. When mMGF and mEL-3 are incubated for seven days with the primitive dull ceils, only mMGF but not mEL-3 maintains CFU-S activity of the primitive dull cells.

When using bright cells. mMGF cannot maintain CFU-S activity at either 3 or 7 days, but mEL-3 can maintain CFU-S activity for both 3 and 7 days. Surprisingly, the combination of mMGF and mEL-3 achieved an expansion of CFU-S with the primitive dull cells. The activity of the combination to cause an expansion of the population of the most primitive stem cells has never been seen before with this model. These data are shown in Figure 6b.

Moreover, we and others have examined all possible combinations of known hematopoietic factors without achieving an expansion before. Accordingly, the combination of MGF and EL-3 provides a synergistic effect to cause activation and proliferation of the most primitive population of stem cells. Thus the therapeutic combination of MGF and EL-3, preferably conspecific MGF and IL-3. is useful for treating anemias and other forms of bone marrow failure that require activation and proliferation of the most primitive stem cell populations. This combination will be most effective for treating aplastic anemia, bone marrow transplantation, and reduced hematopoietic counts caused by any reason. Moreover, as MGF has demonstrated proliferative activity in a population of the most primitive stem cells, it will be effective in combination with another one or combination of hematopoietic factors have have proliferative activity for more developed cells in the hematopoietic system, these other hematopoietic factors include, for example. GM-CSF. G-CSF, IL-3, EPO, EL-la. EL-ip, EL-7, LEF, EL-6 and combinations thereof. 238659 31 EXAMPLE 10 PREPARATION OF MONOCLONAL ANTIBODIES TO hMGF Preparations of hMGF are used to generate monoclonal antibodies against hMGF using conventional techniques. Examples of sources of hMGF for use as an immunogen include, for example, purified recombinant MGF, human MGF, or transfected COS cells expressing high amounts of MGF as described in Example 5 herein. An example of a convention technique to generate monoclonal antibodies is described in United Sates Patent 4,411,993. Monoclonal antibodies generated by these techniques are likely to be useful in establishing convenient assays for MGF, or for blocking the effects of MGF in vitro or in vivo.

Mice are immunized with an MGF immunogen emulsified in complete Freund's adjuvant and injected subcutaneouslv into Balb/c mice at amounts ranging from 10-100 ug.

Ten to twelve days later the immunized animals are boosted with additional immunogen emulsified in incomplete Freund's adjuvant. The animals are periodically boosted on a weekly or biweekly immunization schedule. Serum samples are taken periodically for assay of antibody titer, such as by dot-blot or ELISA (enzyme-linked immunosorbent assay). Serum samples may be obtained by retro-orbital bleeding or by tail-tip excision. Once an appropriate antibody titer is detected in positive animals, the positive animals a injected intravenously with MGF in saline. Three or four days after the saline injection, the animals are sacrificed and spleens harvested to splenocvtes.

Splenocvtes are fused to a murine myeloma cell line, such as NS1 to generate hybridoma cell lines. The hybridoma cell lines are plated in multiple microliter plates in a HAT (hypoxanthine. aminopterin and thymidine) selective medium to inhibit proliferation of non-fused cells, myeloma hybrids and spleen cell hybrids.

An ELISA techniques can be used to screen for positive hybridoma clones by assaying for reactivity with MGF. Such techniques are described, for example, in Engvall et al. Immunochemisrry 8:871 1971 and in United States Patent 4.703,004. Anti-MGF monoclonal antibodies are made, for example, in mouse ascites by injecting positive hybridoma clones into peritoneal cavities of syngeneic Balb/c mice. The ascites fluid is purified by standard techniques, such as ammonium sulfate precipitation followed by gel exclusion chromatography, or by affinity chromatography by binding the monoclonal antibody to Protein A (Staphylococcus aureus).

EXAMPLE II GENERATION OF MONOCLONAL ANTIBODIES TO MURINE MGF BY IMMUNIZATION WITH MGF-BEARING CELLS A cDNA encoding mMGF is prepared and inserted into C127 cells (ATCC No. CRL 1616) as described in Dower et al. J. Immunol. 142:4314-20 1989. Briefly, this procedure 2 3 8 6 32 adds the entire bovine papilloma vims genome linearized at a Bam HI site to pDC201 containing mMGF cDNA to form a plasmid BX8. BX8 is transfected into C127 cells along with plasmid PSV2 Neo at a ratio of 10:1 (BX3 to PSV2 N'eo). Transformed cells expressing mMGF are selected.

Approximately 10^ selected transformed CI 27 cells are used :o immunize each Lewis rat Immunization is by intraperitoneal injection three times at three week intervals. Antibody titer is determined by an appropriate assay. Once an appropriate antibody titer is achieved, the rat is boosted by approximately 2 x 10^ selected transformed C127 cells and sacrificed three days later. Spleen cells are harvested from the rat and fused with mouse myeloma cells (NS-1, ATCC No. TIB 18) at a 1 ratio with 50% polyethylene glycol (PEG MW 1500, EM Reagents. Stuttgard. Germany) using standard fusion procedures. Spleen cells are harvested from the rat and fused with mouse myeloma cells (NS-1. ATCC No. TIB 18) at a 4:1 ratio with 50% polyethylene glycol (PEG MW 1500. EM Reagents. Stuttgard. Germany) using standard fusion procedures. The fused cells are called hybridoma cells.

Hybridoma cells are seeded into well plates at a density of 2 x l(P cells per well in a volume of 200 pi media. Hybridoma cells are screened for monoclonal antibody production to MGF by an assay procedure, such as an ELISA assay.

EXAMPLE 12 GENERATION OF MONOCLONAL ANTIBODIES TO MGF BY IMMUNIZATION WITH VACCINIA BEARING MGF cDNA containing the entire coding region of MGF or the extracellular region of MGF is prepared as described herein. The cDNA is inserted into the Srnal site of vaccinia virus (W) plasmid expression vector pSCl 1 (available by license from the U.S. Dept. of Commerce, National Technical Information Service, 5285 Port Royal Rd. Springfield. VA 22161) utilizing a known methods, such as one described in Chakrabarti et al. Mol. Cell Biol. 5:3403-09 1985. Transformed W are visualized by blue plaques. W from blue plaques are selected and used to infect a host cell, such as CV-1 (ATCC CCL 70) or HeLa (ATCC CCL 2). Infected cells are tested for expression of MGF. Recombinant VV from a positive plaque was purified using conventional techniques, such as those described in Chakrabarti et al. supra and Elango et al. Proc. Nazi. Acad. Sci. USA 83:1906-10 1986.

Lewis rats are immunized by intradermal injection with approximately 10^ plaque forming units (pfu) of recombinant MGF VV. After two weeks, immunized rats are boosted with approximately 10® primary rat fibroblasts infected with recombinant MGF W (greater than 5 pfu per cell). After another two weeks, the rats are boosted with approximately 2 x 10^ Ci27 cells that express MGF. Three days later the rats are sacrificed and its spleen cells recovered. The spleen cells are fused with P3X63-Ag8.653 mouse myeloma cells as described 238659 in Examples 10 and 11. Positive hybridoma clones are selected by an assay, such as an ELISA assay.

EXAMPLE H ASSAYS FOR DETECTING ANTIBODIES ELISA Assav An ELISA assay for screening cells for MGF secretion is conducied by seeding cells into a well plate at a concentration of 4. x 10^ cells per well in a 200 pJ/well volume. This involves preparing a suspension of 2x10° ceils/ml in media. Hybridoma supematants and diluted antisera controls are added to each cell for a 30 minute incubation at room temperature in the wells. Following incubation, the wells are washed and approximately 50 jil/well of antispecies-specinc antisera (e.g.. Goat anti-rat Peroxidase Bio Rad) are added (diluted 1:1000 in 5% fetal calf serum/PBS). The next incubation step is for approximately 30 minutes at room temperature. After washing, substrate solution (for example O-Phenylenediamine 1 mg/ml Zymed, hydrogen peroxide in citrate buffer or TMB substrate. Kirkegaard and Perry) is added to each well. Positive wells are identified by development of color as determined by absorbance in excess of control wells.

ABC Assav Wells are coated overnight with 10 ug/ml of goat antispecies-specific IgG (Zymed S San Francisco. G\), and then blocked with 5% non-fat dry milk for one hour. Hybridoma supematants or diluted monoclonal antibody to MGF is added for one hour. After washing, iodinated MGF is added at approximately 2000 cpm/ul for one hour. After washing, the wells are exposed to film. Positive spots indicate the presence of precipitated MGF protein. Cross-blocking Assav to Determine Eoirooes The relative epitopes of isolated monoclonal antibodies to MGF are determined by a cross-blocking assay. The assay determines is a particular monoclonal antibody inhibits binding of MGF to other specific antibodies. This assay is conducted by binding approximately 2 ml of specific antibody or hybridoma supernatant preincubated with radiolabeled (iodine) MGF (2000 cpm/ml) to a nitrocellulose sheet (Schleicher & Schuell Keene NH). The sheet is dried and then blocked for one hour in 3% BSA (bovine serum albumin) in PSA (phosphate buffered saline). The sheet is thoroughly washed with PBS and exposed to film. A diminished signal indicates inhibition of one antibody's binding by another.

Dot Blot Immunoassay A sheet of nitrocellulose membrane (Schleicher & Schuell Keene NH) is marked into squares and approximately 25 ng of MGF is placed onto the membrane within each square.

After the MGF is dried, the sheet is blocked with 39c BSA in PBS for one hour. After washing and drying the sheet, specific antibody in the form of hybridoma supernatant or 34 antibody is placed onto the membrane for 30 minutes. Next, the membrane is washed and excess antibody removed by incubating with PBS. The membrane is next incubated with diluted labeled antisera which is species specific for the particular antibody. For example, if the specific antibody is a mouse monoclonal antibody, the labeled antisera is a 1:2000 of goat anti-mouse antibody conjugated with, for example, a horse radish peroxidase. After incubation, the membrane is washed and color as an indicator is developed with a substrate for the label, such as 4-chioro-l-napthoI with an oxidizing agent such as hydrogen peroxide. Positive tests show color and indicate the presence of an antibody specific for MGF coated onto the membrane. This assay is useful for screening of multiple hybridoma colonies. Radioirnmunoprecipiration This assay determines the presence of an immunoprecipitating antibody. The assay is conducted by adding 50 pi of PBSTA (PBS containing 50 mg/ml BSA and 10 pl/ml Triton X100.2 pi of rabbit anti-species specific antibody (i.e.. rabbit anti-mouse IgG, IgM and IgA. specific antibody (such as 50 pi of hybridoma supernatant, or 2 pi serum, or 2 pi ascites fluid from a mouse innoculated with a hybridoma). 50 ul of 20% Protein A Sepharose solution (Sigma), and 25 pi of radioiodinated MGF (about 2000 cpm/pl). The sample is spun down and then incubated overnight. Following incubation, the sample is washed three rimes by adding PBS and spinning down the pellet. The pdlet is counted with a gamma counter. A positive indicates the presence of an immunoprecipitating antibody.

From the foregoing it will be appreciated that other sequences, embodiments, activities and purification processes can produce polypeptides with MGF biological activities and are included within the scope of the present invention. 238659 sequence listing r>i s>jr* (1) GENERAI. INFORMATION: {i.) APPLICANT: ni.lli.azis, Douglas E.

Lyman, Stewart (ii) TITLE O? INVENTIONt Mast Cell Growth Fact (iii) NUM3ER OF SEQUENCES: 4 (iv) CORRESPONDENCE ADDRESS: (A) ADDRESSEE: Issiunex Corporation (S) STREET: SI University Street (C) CITY: Stattle (D) STATE: Washington (E) COUNTRY: USA (F) ZIP: 98101 (v) COMPUTER READA3LE FORM: (A) MEDIUM TYPE: Flopoy disk (3) COMPUTED: I3M PC*eoaoatible (C) OPERATING SYSTEM: PC-DOS/MS-DOS CD) SOFTWARE: PatentIa Release #1.24 (vi> CURRENT APPLICATION DATA: {A) APPLICATION NUMBERi (3) FILING DATE: (C) CLASSIFICATION: (viii) ATTORNEY/AGENT INFORMATION; (A) NAME: Oster, Jeffrev 3. (3) REGISTRATION NUMBER: 325S5 (C) REFERENCE/DOCKET NUM3SR: CS21D (ix) TELECOMMUNICATION INFORMATION: (A) TELEPHONE: 206S370430 (2) INFORMATION FOR SEQ ID NO:l: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 1533 base pairs (3) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA (iii) KY?< "ICAL: N (iv) ANTI-SENSE: N (vi) ORIGINAL SOURCE: (A) ORGANISM: Murine (vii) IMMEDIATE SOURCE: (3) CLONE: aMGF-94 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:I: 36 S \ » •">. /■ £ J J 0 D GCGCAACGGC CAAGGACCGG GCGCTGCGTT CGAGCTACCC AATGCTGGGA CTA7CTGCAG SO CCGCTGCTGC 7GCAATATGC 7CGAGC7CCA GAACAGCTAA ACGGAGTCGC CACACCGCTG 120 CCTGCGCTGG ATCGCAGCGC ZGCCTTTCCT rATGAAGAAG ACACAAACT? GGATTATCAC 130 TTGCATTTAT "CTTTCTCT TCAACTGCTC CTATTTAATC C7CTTGTCAA AACCAAGGAG 240 ATCGAATG7G CGGGAATCCT GTGACTGATA ATGTAAAACA CATTACAAAA CTGCGAATAA 300 AA7GTGGCAA ATCTTCCAAA TGACTATATG A7AACCCTCA ACTATGTCAA AAAATTTATA 350 GCCGGGATGG ATGTTTTGCC TAGTCATTGT 7GGCTACGAG ATATCAAGTA ACTAGATGTA 420 ATACAATTAT CACTCAGCTT GACTACTCTT CTGGACAAGT TCAGTTATCA AATATTTCTG 430 AAGGCTTGAG TAATTACTCC ATCATAGACA AAAGGATACT TGGGAAAATA GTGOATGACC 540 TCG7GTTATG CA7GGAAGAA AACGAAAACT GGAGCACCGA AGAATATAAA AGAATCTCCG 500 AACAGGCCAG AAACTAGAAA AGACCTAGTC CTTTACTCCT GAAGAATTCT TTAGTATTTT 550 CAATAGATCC ATTTGGAAGG ATCCCTTTAA GGACTTTATG GTGGCATCTG ACACTAGTGA 720 CTGTAAAATA AAATACGTGC TCTCTTCAAC ATTAGGTCCC GAGAAAGCCA GCTCCCTTAG 780 GATGGAAAGA A7GACAGCAG TACCAGTAAT AGGAAACCCG CAAAGGCCCC TGAAAAAAGA 840 AAAAAGGACT CGGGCCTACA ATGGACAGCC ATGGCATTGC CGGCTCTCAT TAGGTTAATA 900 AAATCGCTTG TAATTGGCTT TGCTTTTGGA GCCTTATACT GGAAGAAGAG AAGAATTAAA 950 CAGTCAAGTC TTACAAGGGC AGTTGAAAAT ArACAGATTA ATGTAGAAAG AGAGAAGAGG 1020 ATAATGAGAT AAGTATGTTG CAACAGAAAG AGAGAGAAGG AAGTGGGAGG TTTCAAGAGG 1080 TGTAATTGTG GACGTATCAA CATTGTTACC TTCGCGGAAC AGTGGCTGGT AACAGTTCAT 1140 GTTTGCTTCA TAAATGAAGC AGCCTTAAAC AAATTCCCAT TCTGTCTCAA GTGACAGACC 1200 TCATCCTTAC CTGTTCTTGC TACCCGTGAC CTTGTGTGGA TGATTCAGTT GTTGGAGCAG 1250 AGTGCTTCGC TGTGAACCCT GCACTGAATT ATCATCTGTA AAGAAAAATC TGCACGGAGC 1320 AGGACTCTGG AGGTTTTGCA AGTGATGATA GGGACAAGAA CATGTGTCCA GTCTACTTGC 1330 ACCGTTTGCA TGGCTTGGGA AACGTCTGAG TGCTGAAAAC CCACCCAGCT TTGTTCTTCA 1440 GTCACAACCT GCAGCCTGTC GTTAATTATG GTCTCTGCAA GTAGATTTCA GCCTGGATGG 1500 TGGGGGGAAT TTTTTTTTTC ACAAAAGAGG GAG 1533 (2) INFORMATION FOR SEQ ID NO:2: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 1589 base pairs (3) TYPE: nucleic.acid (C) STSAJTOEONESS: single (D) TOPOLOGY: linear (ii) MOLSCjLE TYPE: CDNA (iii) HYPOTKETICAL: N m 238659 (vi) ORIGINAL SOURCE: (A) ORGANISM: MURING MAST CELL GRO«TK FACTOR (vii) IMMEDIATE SOURCE: (3) CLONE: 1C.GF-I0 (xi) SSQUSNC2 DESCRIPTION: SZQ ID NO:2: CCCAACCCCC AACGACGGGG CGCTGCGTTC GAGCTACCCA ATGCTGGGAC TATCTGCAGC 50 CGCTGCTGCT GCAA7A7GC7 CGAGCTCCAG AACAGCTAAA CGGAGTCGCC ACACCGCTCC 120 CTGGGCTGGA 7CGCAGCGC7 GCCTTTCCTT ATGAAGAAGA CACAAACTTG GA7TA7CACT 180 TGCAT7TATT 7GTTTCTCTT CAACTGCTCC TATTTAA7CC TCTTGTCAAA ACCAAGGAGA 240 TCGAA7GTGC GGGAA7CC7G 7GACTGATAA TGTAAAAGAC A77ACAAAAC 7GCGAATAAA. 300 ATG7GGCAAA 7CTTCCAAAT GACTATATGA TAACCCTCAA C7A7GTCAAA AAATT7ATAG 3 SO CCGGGATGGA 7GT7T7GCCT AGTCAT7G77 GCCTACGAGA 7A7GAAG7AA CTAGA7GTAA 420 TACAATTATC ACTCAGCT7G ACTACTCT7C TGGACAAGTT CAGTTATCAA ATATT7C7GA 430 AGGCTTGAGT AATTACTCCA TCATAGACAA AAGGATAC7T GGGAAAATAG TGGATGACCT 540 CSTGTTATGC ATGGAAGAAA ACGAAAAC7G GAGCACCGAA GAA7ATAAAA GAATC7CCGA 500 AGAGGCCAGA AAC7AGAAAA GAGGTAGTCC 7TTACTCCTG AACAATTCTT TACTA7TTTC 550 AATAGATCCA TTTGGAAGGA TGCCTTTAAG GACTTTA7GG TGGCATCTGA CACTAGTGAC 720 TCTAAAATAA AATACSTGCT CTCTTCAACA TTAGGTCCCG AGAAAGA7TC CAGAG7CAGT 7 SO ATGGAAGAGT CACAAAACCA TTTATG7TAC CCCCTGT7GC AGCCAGC7CC CT7A77AAAA 840 AAGGAATGAC AGCAG7ACCA GTAATAGGAA AGCCGCAAAG GCCCCTAGAA AAGAAAAAAG 900 AAGAC7CGGG CCTACAATGC ACAGCCATGG CATTGCCGGC TCTCGAGCT7 AA7AAAAATT 950 TCGCTTG7AA TTGGCTTTGC TTTTGGAGCC TTATACTGGA AGAGAAGAA7 TAAGAAACAG 1020 TCAAGTCTTA CAAGGGCAGT TCAAAATA7A CAGATTGTAG AAAGAGAATG AAGAGGATAA 1080 TGAGATAAGT ATGTTGCAAC AGAAAGAGAG AAGGAAGTGG GAGGAATTTC AAGAGGTGTA 1140 ATTGTGCACC TA7CAACA7T GTTACCTTGG GACGCACAGT GGCTGG7AAC AGTTCATGTT 1200 TGCTTCATAA ATGAAGCAGC CTTAAACAAA 7TCCCATTCT G7CTCAAGTG ACAGACCTCA 1250 TCCT7ACCTG TTCT7GC7AC CCGTGACC7T G7GTGGA7GA 77CAGTTGT7 GGAGCAGAG7 1320 GCTTCGCTGT GAACCCTGCA CTGAATTA7C A7CTGTAAAG AAAAATCTGC ACGGAGCAGG 13S0 ACTC7GGAGG TTTTGCAAGT GATGATAGGG ACAAGAACAT GTGTCCAGTC TACT7GCACC 1440 GTTTGCATGG CTTGGGAAAC GTCTGAG7GC TGAAAACCCA CCCAGC7TTG TTCT7CAGTC 1S00 ACAACCTGCA GCCTG7CGTT AA7TA7GGTC 7CTGCAAGTA GATTTCAGCC TGGA7GGTGG 1560 GGGGAATTTT TTTTTTCaCA AAGGGATGTA GAAAACAATT 7AAAAAAACA AAACAAAACA 1520 2 3 36 ATAAACCGCA ACCGCCAACC ACGGGGCGCT GCGTTCGAGC TACCCAATGC TCGGACTATC 1S30 TGCAGCCGC 1589 (2) INFORMATION FOR SEQ ID NO:3: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 1069 base pairs (3) TYPE: nucleic aei£ (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA (iii) HYPOTKETICAL: N (iv) ANTI-SENSE: N (vi) ORTCINAL SOURCE: (A) ORGANISM: HUMAN MAST CELL GROWTH FACTOR (vii) IMMEDIATE SOURCE: (3) CLONE: HMGF-2D (xi) SEQUENCE DESCRIPTION: SEQ 13 NO:3: ACGCCTCGAT CCCAGCGCTG CCTTTCCTTA TGAAGAAGAC ACAAACTTTG TTGGATTCTC 50 ACTTCCATTT ATCTTCAGCT GCTCCTATTT AATCCTTTCT GACTCGTCAA AACTGAAGGG 120 ATCTGCAGGA ATCGTGTGAC TAATAATATG GCAGAAGATA AGTAAAAGAC GTCACTAAAT 130 TGGTGCCAAA TCTTCCAAAA GACTACAAAT AAAAATATGA TAACCCTCAA ATATGTCCCC 240 GGCATGGATG TTTTGCCAAG TTTTAGTAAC ATTGTTGGAT AAGCGAGATG GTAGTACAAT 300 TGTCAGACAG CTTGCTGTAA GAACTGATCT TCTGGACAAG TTTTCAAATA TTTCTGAAGG 3SO CTTGAGTTAA AGGAATTATT CCATCATAGA CAAACTTGTG AATATAGTGG ATGACCTTAT 420 AAAAAAG7GG AGTGCGTGAA AGAAAACTCA TCTAAGGATC TAAAAAAATC AAGCAGAATT 430 CAAGAGCCCA GAACCCAGGC TCTTTACTCC TGAAGAATTC TTTGAGTGGA GAATTTTTAA 540 TAGATCCATT GATGCCTTCA AGGACTTTGT AGTGAGAAGA AAAAAGCATC TGAAACTAGT 500 GATTGTGTGG TTTCTTCAAC ATTAAGTCCT AAGTACAATG AGAAAGGGAA GGCCAAAAAT 550 CCCCCTGGAG ACTCCAGCCT ACACGGAAAG ATGGCCACCC ATGGCATTGC CAGCATTGTT 720 TTCTCTTATA ATTGGCTAAA ATAAAAGTTT GCTTTTGGAG CCTTATACTG GAAGAAGAGA 7S0 CAGCCAAGTC TTAAGAATTA GGACAAGGGC AGTTGAAAAT ATACAAATTA ATGAAGAGGA 340 TAATGAGTAG AAAGAGAGGA AGATAAGTAI GTTGCAAGAG AAAGAGAGAG AGTTTCAAGA 900 AGTGTAATGG GAGGGGATTG TGGCTTGTAT CAAGTTTCTC TTCAACTGCT CCTATTTAAT 950 CCTCTTGTCA AAACCAAGGA GATCGAATGC GGCTGGATCG CAGCGCTGCC TTTCCTTATG 1020 AAGAAGACAC AAACTTTGTA TCCAGGAACC ATAGATTCAA CTTTACTTT 1059 39 238659 (2) INFORMATION FOR SEQ ID NO:4: (i) SEQUENCE CHARACTERISTICS: (A) LENGTH: 1050 base pairs (3) TYPE: nucleic acid (C) STRANDEDNESS: single (D) TOPOLOGY: linear (ii) MOLECULE TYPE: cDNA (iii) HYPOTHETICAL: N (vi) ORIGINAL SOURCE: (A) ORGANISM: HUMAN MAST CELL GROWTH FACTOR (vii) IMMEDIATE SOURCE: (3) CLONE: HMGF-2.4 (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4: CSCCTCGATC GCACCCCTCC CTTTCCTTAT GAAGAAGACA CAAACTTTGT TGGATTCTCA 50 CTTGCATTTA TCTTCAGCTG CTCCTATTTA ATCCTTTCTG ACTCGTCAAA ACTGAAGGGA 120 TCTGCAGGAA TCGTGTGACT AATAATATGG CAGAAGATAA GTAAAAGACG TCACTAAATT 130 GGTCGCAAAT CTTCCAAAAG ACTACAAATA AAAATATGAT AACCCTCAAA TATGTCCCCG 240 GGATGGATGT TTTGCCAAGT TTTAGTAACA TTGTTGGATA AGCGAGATGG TAGTACAATT 300 GTCAGACAGC TTGCTGTAAG AACTGATCTT CTGGACAAGT TTTCAAATAT TTCTGAAGGC 360 TTGAGTTAAA GGAATTATTC CATCATAGAC AAACTTCTGA ATATAGTGGA TGACCTTATA 420 AAAAAGTGGA GTGCGTGAAA GAAAACTCAT CTAAGGATCT AAAAAAATCA AGCAGAATTC 430 ' AAGAGCCCAG AACCCAGGCT CTTTACTCCT GAAGAATTCT TTGAGTGGAG AATTTTTAAT 540 AGATCCATTG ATGCCTTCAA GGACTTTGTA GTGAGAAGAA AAAAGCATCT GAAACTAGTG 500 ATTGTGTGGT TTCTTCAACA TTAAGTCCTA AGTACAATGA GAAAGATTCC AGAGTCAGTG 550 TCACAAAACC ATTTATGTTA CCCGAAGAAT TCCTGTTGCA GCCAGCTCCC TTAGGAATGA 720 CAGCAGTAGC AGTAATAAAA AAGAAAAGGA AGGCCAAAAA TCCCCCTGGA GACTCCAGCC 730 TACACTGGGC AAGAAAGATA AGCCATGGCA TTGCCAGCAT TGTTTTCTCT TATAATTGGC 340 TTTGCTAATA AAAGAATTTG GAGCCTTATA CTGCAAGAAG AGACAGCCAA GTCTTACAAG 900 GGAATTAGGT AGGCAGTTGA AAATATACAA ATTAATGAAG AGGATAATGA GATAAGTAAA 960 GAGAGGAAGA TGTTGCAAGA GAAAGAGAGA GAGTTTCAAG AAGTGTAATT GTGGTGGGAG 1020 GGGACTTGTA TCAAGACACA AACTTTGTTG 10SO

Claims (26)

238659 40 WHAT }lWE CLAIM 1S:-

1. An isolated DNA sequence encoding a polypeptide exhibiting human MGF composed o £ activity, saidsequence Comprising a DNA sequence selected from the group consisting of: (a) a DNA sequence encoding a human MGF polypeptide comprising the sequence: Xn Xa X- Cys Arg Asn Arg Val Thr Asn Asn Val Lys Asp Val Thr Lys Leu Val Ala Asn Leu Pro Lys Asp Tyr Met lie Thr Leu Lys Tyr Val Pro Glv Men Asp Val Leu Pro Ser His Cys Trp lie Ser Giu Mec Val Val Gir. Leu Ser Asp Ser Leu Thr Aso Leu Leu Aso Lvs Phe Ser Asn lie Ser Giu Glv Leu Ser * * Asn Tyr Ser lie He As? Lys Leu Val Asn lie Val Asp Asp Leu Vai Ri Cys 3-2 Q Giu Asn Ser Ser Lys Asp Leu Lys Lys Ser Phe Lys Ser Pro Giu Pro Arg Leu Phe Thr Pro Giu Giu Phe Phe Arg lie Phe Asn Arg Ser lie Asp Ala Phe Lys Asp Phe Val Val Ala Ser Giu Thr Ser As? Cys X- xR Xn Xn Xn Xr, Xa Xn Xn wherein n is an integer 0 or 1, X is a natural amino acid, and Ri and R2 are any amino acid, and Q is Lys, Arg or Giu; and (b) DNA sequences that hybridize to the DNA sequence of (a) under conditions of high stringency, which code on expression for a polypeptide with human MGF activity; and induce proliferation of human hematopoietic progenitor cells in combination with human interieukin-3 (hIL-3).

^ 2. An isolated DNA sequence according to claim 1, wherein Ri is Giu, R2 is Val and Q is Lys.

3. An isolated DNA sequence according to claim 1, wherein the DNA sequence - includes encoding a polypeptide exhibiting human MGF activity comprises a DNA sequence selected •6^ from the group consisting of: (a) the DNA sequence of Sequence I.D. No. 3 (herein) frcm nucleotide 103 to nucleotide 546; (b) DNA sequences that hybridize under high stringency to the DNA sequence of 3 (a) and which code on expression a polypeptide with human MGF activity; and (c) DNA sequences that encode polypeptides encoded by any of the foregoing DNA sequences but vary in nucleotide sequence as a result of degeneracy of the genetic code.

4. An isolated DNA sequence according to claim 3, wherein the DNA sequence f E /w ^nco^es human A28 MGF. s. 9" f i Ii J '38659 41

5. An isolated DNA sequence according to claim 1, wherein the DNA sequence encoding a polypeptide exhibiting human MGF actjvity/gompnscf a DNA sequence selected from the group consisting of: (a) the DNA sequence of Sequence I.D. No. 4 (herein) frctn nucleotide 113 to nucleotide 657; (b) DNA sequences that hybridize under high stringency to the DNA sequence of (a) and which code on expression a polypeptide with human MGF activity; and (c) DNA sequences that encode polypeptides encoded by any of the foregoing DNA sequences but vary in nucleotide sequence as a result of degeneracy of the genetic code.

6. An isolated DNA sequence according to claim 5. wherein the DNA sequence encodes a 1S5 amino acid extracellular domain of hMGF.

7. An isolated DNA sequence encoding MGF or a derivative, analog or allelic variant thereof displaying at least one biological activity selected from the group consisting of: (a) binding to a receptor protein expressed by a mammalian c-kit proto- oncogene; (b) inducing proliferation of mammalian mast cells; (c) inducing proliferation of mammalian hematopoietic stem cells; and (d) inducing proliferation of an IL-3 dependent cell line having the characteristics of MC6. having

8. An isolated MGF poiypeptidcwrriprising the ammo acid sequence: Xn Xr_ x<< Cys Arg Asa Arg Val Tiir Asa Asn Val ivs As? Val Thr Lys Leu Val Ala Asn Leu Pro Lys Asp Tyr Met He Thr Leu Lys Tyr Val Pro Gly Met As? Val Leu Pro Ser His Cys Tr? lie Ser Giu Met Val Val Gin Leu Ser Asp Ser Leu Thr Asp Leu Leu As? Lys Phe Ser Asn He Ser Giu Gly Leu Ser Asn Tvr Ser He He Asp Lys Leu Val Asn He Val Asp Asp Leu Val Cys R2 Q Giu Asn Ser Ser Lys As? Leu Lys Lys Ser Phe Lys Ser Pro Giu Pro Arg Leu Phe Thr Pro Giu Giu Phe Phe Arg He Phe Asn Arg Ser He Asp Ala Phe Lys As? Phe Val Val Ala Ser Giu Thr Ser Asp Cys Xn Xn Xr. Xn Xn Xn Xa Xn Xn xn wherein n is an integer 0 or 1, X is a natural amino acid, Rj and R2 are any amino acid, and Q is Lys, Arg or GIu, or an amino acid sequence encoded by an isolated DNA sequence according to claim 1. /' \ j, - -# 23865 42

9. An isolated MGF polypeptide according to claim 8, wherein Rj is Giu, R2 is Val and Q is Lys.

10. An isolated MGF polypeptide according to claim 8 wherein the amino acid sequence is Giu Gly He Cys Arg Asn Arg Val Thr Asn Asr. val Lys Asp val Thr Lys Leu Val Ala Asn Leu Pro Lys Asp Met He Thr Leu Lys Tyr Val Pro Gly Met Asp val Leu Pro Ser His Cys Trp He Ser Giu Met val val Gin Leu Ser Asp Ser Leu Thr Asp Leu Leu Asp Lys Phe Ser Asn He Ser Giu Gly Leu Ser Asn TV2T Ser He lie Asp Lys Leu val Asn He Val Asp Asp Leu Val Giu Cys Val Lys Giu Asn Ser Ser Lys As? Leu Lys Lys Ser Phe Lys Ser Pro Giu Pro Arg Leu Phe Thr Pro Giu Giu Phe Phe Arg He Phe Asn Arg Ser AS? Ala Phe Lys Asp Phe Val val Ala Ser Giu Thr Ser As? Cys val val Ser Ser Thr Leu Ser Pro Giu Lys [Asp Ser Arg val Ser val Thr Lys Pro Phe Met Leu Pro Pro Val Ala Ala Ser Ser Leu Arg Asn Asp Ser Ser Ser Ser Asn lx * Lys Ala Lys Asn Pro Pro Gly Asp, ► wherein x is 0 or 1 and Y is Arg or Giy.

11. An isolated MGF polypeptide according to claim 10, wherein the sequence is A28 MGF. which includes

12. A recombinant expression vector corr.grising an isolated DNA sequence according to claim 1. including

13. A host cell a recombinant expression vector according to Claim 12.

14. The host cell of claim 13 wherein the host cell is selected from the group consisting of E. coli, Pseudomonas. Bacillus, Sireptomyces. yeast, fungi, insect cells and mammalian cells.

15. The host cell of claim 14 wherein the host cell is a yeast cell of the species Saccharomyces cerevisiae strains X V2181 or YNN281. consisting of

16. A process for purifying recombinant MGF from yeast supematantsicomprising: (a) cation exchange fractionating yeast supernatant containing MGF at an acid pH of from 2.0-4.0 and eluting with acetate at a pH of 4.5-5.5; ; . ?3865 43 (b) reverse phase high performance liquid chromatography (RP-HPLC) fractionating the anion exchange eluate with a linear gradient of acetonitrile; and (c) cation exchange fractionating the RP-HPLC eluate in a physiological pH buffer.

17. A pharmaceutical composition for stimulating proliferation of hematopoietic cells'compnsinf a MGF polypeptide according to claim 8 in combination with a growth factor wherein the growth factor is selected from the group consisting of GM-CSF, G-CSF, EL-3, EPO, IL-6, IL-la, IL-1 (5, IL-7. LIF, pIXY-321, and combinations thereof.

18. The pharmaceutical composition of claim 17 wherein the growth factor is EL-3 or pIXY-321.

19. An antibody immunoreaccive with MGF according to claim 8.

20. A process for using the DNA sequence of claim 1 comprising expressing the DNA sequence of claim 1 in an expression vector contained within a host cell and recovering the recombinant polypeptide exhibiting human MGF activity.

21. An isolated DNA sequence as claimed in claim 1, substantially as herein described with reference to any one of the Examples.

22. An isolated MGF polypeptide as claimed in claim 8, substantially as herein described with reference to any one of the Examples.

23. A recombinant expression vector as claimed in claim 12, substantially as herein described with reference to any one of the Examples.

24. A host cell as claimed in claim 13, substantially as herein described with reference to any one of the Examples.

25. A process as claimed in claim 16, substantially as herein described with reference to any one of the Examples.

26. A pharmaceutical composition as claimed in claim 17, substantially as herein described with reference to any one of the Examples. . L < IMMUNEX CORPORATION * By Their Attorneys J BALDWIN SON & CAREY 55 11JUH]992