CA1340266C - Human pluripotent hematopoietic colony stimulating factor, method of production and use - Google Patents

Abstract

Highly purified pluripotent hematopoietic colony-stimulating factor (pluripotent CSF), a glycoprotein (MW
19,600) constitutively produced by human tumor cells has been highly purified from low serum-containing conditioned medium to apparent homogeneity. Pluripotent CSF supports the growth of human mixed colonies (CFU-GEMM), granulocyte-macrophage colonies (CFU-GM), early erythroid colonies (BFU-E) and induces differentiation of human leukemic cells. The specific activity of the purified pluripotent CSF in the CFU-GM
assay is 1.5 X 10 8 U/mg protein.

Description

This application concerns human pluripotent colony stimulating factor (P-CSF) also known as pluripoietin.
Background This work was done in part with government funding under United States Public Health Service Grants CA-32516, HL-31780, CA-20194, CA-23766 and CA-00966.
Therefore the United States government has certain rights in this invention.
Abbreviations CFU-GEMM:Colony forming unit - granulocyte, erythroid, macrophage, megakaryocyte.
CFU-GM: Colony forming unit - granulocyte-macrophage, BFU-E: erythroid burst forming unit, GM-CSF: Granulocyte-macrophage colony stimulating factor.
Colony-stimulating factors (CSFs) are hormone-like glycoproteins produced by a variety of tissues and tumor cell lines which regulate hematopoiesis and are required for the clonal growth and maturation of normal bone marrow cell precursors in vitro (Burgess, A.W., et al.
(1980) Blood 56:947-958; Nicola, N.A., et al. (1984) Immunology Today 5:76-81). In contrast to the murine system (Nicola, N.A., et al. (1983) ~. Biol. Chem.
258:9017-9021; Ihle, J.N., et al. (1982) ~. Immunol .

X

13S0266 13~ 66 129:2431-2436; Fung, M.C., et al. (1984) Nature 307:233-237; Cough, N.M., et al. (1984) Nature 309:763-767), human CSFs have been less well characteri~ed, both biologically and biochemically (Nicola, N.A., et al.
(1979) Blood 54:614-627; Wu, M.C., et al. (1980) ~.
Clin. Invest. 65:772-775; Golde, D.W., et al. (1980) Proc, Nat'l. acad. Sci. USA 77:593-596; Lusis, A.J., et al. (1981) Blood 57:13-21; Abboud, C.N., et al. (1981) Blood 58:1148-1154; Okabe, T., et al. (1982) J. Cell.
Phys. 110:43-49). Purification to apparent homogeneity has only been reported for macrophage active CSF (CSF-l) (Das, S.K., et al. (1981) Blood 58:630-641; Das, S.K., et al. (1982) ~. Biol. Chem. 257:13679-13684) and erythroid potentiating activity [Westbrook, C.A. et al.
~. Biol. Chem. 259:9992-9996 (1984)] and for granulocyte-macrophage CSF (GM-CSF) [Gasson, J.C., et al. Science 226:1339-1342 (1984)], but not for human pluripotent CSF.
Assays are available to detect human clonogenic precursors that give rise to cells of the erythroid, granulocytic, megakaryocytic, macrophage (CFU-GEMM) (Fauser, A.A., et al. (1978) Blood 52:1243-1248; Fauser, A.A., et al. (1979) Blood 53:1023-1027) and possibly lymphoid (Messner, H.A., et al. (1981) Blood 58:402-405) 13~02~

lineages. CSFs with activities on these multipotential progenitor cells (pluripotent CSF, or P-CSF) are produced by mitogen- or antigen activated T lymphocytes (Ruppert, S., et al. (1983) Exp. Hematol. 11:154-161) and conatitutively by human tumor cell lines such as the SK-HEP-l human hepatoma cell line (J, Gabrilove, K.Welte, Li Lu, H. Castro-Malaspina, M.A.S. Moore, Blood, 66:407-415, August 1985); the 5637 bladder carcinoma cell line (reported herein and in Proc. Nat.
Acad. sci., 82:1526-1530, 1985); and by the HTLV-transformed lymphoid cells (Fauser, A.A., et al.(1981) Stem Cells, 1:73-80; Salahuddin, S.Z., et al., (1984) Science 223:703-707). Pluripotent CSF is involved in the proliferation and differentiation of pluripotent progenitor cells leading to the production of all major blood cell types. This is therefore a broad spectrum CSF. It also induces differentiation of leukemic cells.
Su~nary This application concerns human pluripotent colony stimulating factor CSF for the stimulation of proliferation and differentiation of pluripotent progenitor cells to all major blood cell types which is purified to apparent homogeneity. Its biological effects 13~02~6 include the induction of functional markers of differentiation of normal and leukemic cells.
Brief Description of the Drawings Figure 1 shows ion-exchange chromatography of 5637 conditioned medium (CM) followed by the Gel filtration chromatograph shown in Figure 2.
Figure 3 shows pooled gel filtration eluants on HPLC (reverse phase). Figure 4 shows SDS-PAGE whereas Figure 5 shows preparative SDS-PAGE and figure 6 isoelectrofocusing of the purified pluripotent CSF.
Description o~ the Drawings Figure 1: Ion exchange chromatography One liter dialyzed ammonium sulfate-precipitate of 5637 CM was applied in 0.05M Tris/HCl, pH 7.8, on a 1 L
DEAE cellulose (DE 52) column. Bound proteins were eluted with a linear gradient of NaC1 (0.05 - 0.3 M) in 0.05 - Tris/NCl, pH 7.8, as indicated ( - ). The elution of proteins was monitored by absorption at 280 nm (o-o) and each fraction was tested for CSF activities (GM-CSF
activity: ~). Proteins from the first peak of GM-CSF
activity eluted from the column gave rise to mixed colonies in a CFU-GEMM assay and were used for further purification (pluripotent CSF).

X

1 3 ~

Fiqure 2: Gel filtration chromatography The pluripotent CSF containing concentrated pool of DEAE cellulose chromatography was loaded on an AcA 54 Ultrogel column (2.6x90 cm) and eluted with PBS, Arrows denote the elution points of bovine serum albumin (MW
68, 000), and chymotrypsinogen (MW 25,000). The elution of proteins was monitored by absorption at 280 nm (o-o) and each fraction was tested for pluripotent CSF
activity (GM-CSF activity: ~).
Figure 3: Reverse phase high-performance liquid chromatography ~HPLC) The pooled fractions with pluripotent CSF
activities eluted from the gel filtration column were acidified to pH 4.0 and loaded onto a C 18 (uBondapak, Waters) column. The bound proteins were eluted with a linear gradient of 1-propanol in 0. 9M acetic acid/0. 2M
pyridine, pH 4,0. The elution of proteins were monitored by absorption at 280 nm (-) and each fraction was tested for pluripotent CSF activity (GM-CSF activity: A ).
Figure 4: SDS-polyacryl~ide qel electrophoresis ~SDS-PAOE ) The pluripotent CSF eluted from the HPLC column (200 ng; peak fraction) was lyophilized and treated with 1% SDS in 0. 0625 M Tris/HC1, pH 6 . 8, and 20% glycerol, 13~026 under reducing conditions (5% 2-mercaptoethanol) for one hour at 37~C and then applied to a 15% polyacrylamide gel. After electrophoresis, the protein bands were visualized by the silver staining technique.
Figure 5: Preparati~e SDS-PA~~
Pluripotent CSF eluted from HPLC (Fig. 3) was treated and processed (under non-reducing conditions) as shown in Fig. 4. After electrophoresis, the gel was sliced into 4 mm sections and proteins from each slice were eluted into RPMI 1640 containing 10% FCS. After 18 hours, eluted proteins were assayed for pluripotent CSF
activity (GM-CSF activity: shaded area).
Figure 6: Isoelectrofo~-~ing HPLC purified lyophilized pluripotent CSF was supplemented with 20% (v/v) glycerol and 2% ampholines (pH 3.5-10) and layered onto the isodense region of an 0-60% gradient of glycerol containing 2% ampholines (pH

3.5-10). After isoelectrofocusing (2,000 V, 24 hours), 5 ml fractions were collected and the pH (o) determined in each fraction. All fractions were subsequently dialyzed and tested for pluripotent CSF activity (GM-CSF
activity: ~
We report the purification and biochemical characterization of a human pluripotent CSF, produced and released constitutively by human cells especially . . , ,~ . ..

~ 13iO26~

tumor cells such as bladder carcinoma cell line 5637 (ATCC HTB-9) and hepatoma cell SK-HEP-l: (ATCC HTB52).
The cell line (5637) was obtained from Jorgen Fogh at Sloan-Kettering Institute, 1275 York Avenue, New York, New York 10021.
Pluripotent CSF biological properties :include differentiation of progenitor cells to all major blood types as well as differentiation of leukemic cells.
As~ay for GM-CSF Activity GM-CSF activity was tested on human bone marrow (BM) cells cultured with serial dilutions of test samples in semi-solid agar. BM from healthy human volunteers, who gave informed consent, was diluted 1:5 in phosphate buffered saline (PBS) and separated by density gradient centrifugation on Ficoll-Hypaque. 105 separated cells were plated in l ml of 0.3% agar culture medium that included supplemented McCoy's 5A medium and 10~ heat inactivated fetal calf serum (FCS), as described (Broxmeyer, H.E. et al. (1977) Exp. ~ematol.
5: 87-102). To this mixture serial dilutions of a laboratory standard or test samples (10%;v/v) in RPMI
1640 with 10% FCS were added. Cultures were scored for colonies (greater than 40 cells/aggregate) and morphology was assessed after 7 and 14 days of incubation. GM-CSF units were determined from dose 1:~4~26~

response curves and expressed as U/ml, where 50 U is the CSF concentration stimulating half-maximal colony number to develop (Nicola, N.A., et al. (1983) ~. Biol. Chem.
258: 9017-9021).
Assay for Colony Stimulating Factor for BFU-E and CFU-OE MM
The colony assay for human BFU-E and CFU-GEMM was performed as previously described (Li Lu, et al. (1983) Bl ood 61: 250-256). Human bone marrow cells were subjected to a density cut with Ficoll-Hypaque~ (density 1.077 gm/cm3; Pharmacia Fine Chemicals, Piscataway, N.J.) and the low density cells were suspended in RPM1 1640 containing 10% FCS at 2 x 107 cells/ml and placed for adherence on Falcon tissue cultures dishes (#3003, Becton Dickinson and Co., Cockeysville, MD) for 1~ hr.
at 37~C. The nonadherent cells were depleted of T
lymphocytes by rosetting with neuraminidase-treated sheep erythrocytes. Medium conditioned by leukocytes from patients with hemochromatosls in the presence of 1%
(v/v) phytohemagglutinin (PHA) (Li Lu, et al., (Bl ood 61:250-256, 1983) as positive control or serial dilutions of test samples were then added at 5~ (v/v) to x 104 of these low density, non-adherent and T
lymphocyte depleted bone marrow cells in a 1 ml mixture X - g _ 134026n of Iscove's modified Dulbecco medium (GIBCO Grand Island, NY), 0.8% methylcellulose, 30% FCS, 5x10 5 M 2-mercaptoethanol, 0.2 mM Hemin, and one unit of erythropoietin {Hyclone, or Connaught Labs., Willowdale, Ontario, Canada). The addition of Hemin is necessary to obtain optimal cloning efficiency (Li Lu, et al. (1983) Exp. Hematol . 11:721-729). Dishes were incubated in a humidified atmosphere of 5% CO2 in air at 37~C. After 14 days of incubation, colonies were scored and morphology was assessed.
As shown above, a single protein stimulates colony formation by CFU-GEMM, BFU-E, and CFU-GM progenitor cells.
This protein we termed "pluripotent CSF". Due to the low numbers of mixed colonies per dish attainable in this assay system, titration of test samples for determination of pluripotent CSF activity meets with considerable difficulties. Therefore, we used the GM-CSF assay units as described above to measure the GM-CSF
aspect of the pluripotent CSF activity in those samples that supported growth of BFU-E and CFU-GEMM for calculating the specific activity throughout the purification procedure.

~ X

13qO2fi~
.A.

Differentiation Induction Assay Titrated samples of purified pluripotent CSF were assayed for differentiation induction of WEHI 3B ~D+) or HL-60 leukemic cells described (Metcalf. D, (1980) Int.
5 J. Cancer 25:225-233; Fibach, E., et al., ~. Cell Physiol. 113:(1) 152, (1982) ).
Rossette Assays for Fc Roceptor, OKM1 and LQU M2 antigens Cell receptors for immunoglobulin Fc were assayed with IgG (Cappel Laboratories, West Chester, PA) coated sheep erythrocytes as desribed elsewhere (Ralph, P., et al. (1983) Blood 62:1169). OKM1 (Ortho Diagnostic Systems Inc., Raritan, NJ) or Leu M2 (Beckton Dickinson, Mountain View, CA) reactive antigens were detected by 15 incubating 106 cells/O.1 ml phosphate buffered saline containing. 1.0 ug/ml monoclonal antibody for 20 min at 24~C, washing, incubating 20 min at 24~C with a a:100 dilution of rabbit anti-rat (Leu M2) or anti-mouse (OKMl) IgG serum (Cappel Laboratories, Cochranville, PA), washing and rosetting with protein A-coated erythrocytes as described previously (Ralph, P. et al., Blood 62:1169, 1983).

X

_ .. . . . , . ... _~_ . ... ~_ . , . ... _ .. . . ...

13~02S~

Assays for fMLP Receptor Receptors for chemotactic peptide, formyl-Methionyl-Leucyl-Phenylalanine (fMLP), were assayed as follows: 2 X 106 cells were incubated with 15 nM 3H-fMLP
(New England Nuclear, Boston, MA) in a total volume of 0.2 ml in the presence or absence of 10 uM unlabelled fMLP (Sigma Corp. St. Louis, MO). After three hours at 4~C the cell suspensions were rapidly filtered onto glass fiber discs (Whatman Inc., Clifton, N.J.), which were then washed with 30 ml of 4~C phosphate buffered saline (Harris, P. et al. (1985) Cancer Res. 45:9).
Radioactivity on the discs were counted by liquid scintillation spectrophotometry.
Measurement of PMA-Stimulated Hydrogen Peroxide Release The production of hydrogen peroxide in response to PMA stimulation was assayed by horse radish peroxidase (HRPO) (Sigma) mediated H202 dependent oxidation of homovanillic acid (HVA) (Sigma), as described (Harris, P., et al. (1985) (Cancer Res., 45:9, (1985)). Briefly, cells (1 X 106) were suspended in 2 ml of a solution containing 100 micromolar HVA 5 U/ml HRPO in the absence or presence of 30 ng/ml PMA. Following 90 min.
incubation at 37~C, the incubation was centrifuged and 0.25 ml of 25 mM EDTA, 0.1 M glycine-NaOH, pH 12 was added to the supernates. A 30% stock solution of X

134~261i hydrogen peroxide (Sigma) was used to prepare H202 standards (0.001 to 50 nmoles/assay) for the construction of a standard curve. The HVA oxidation product was measured on a Perkins-Elmer Model MPF-44A
fluorescence spectrophotometer. Excitation and emission were set at 312 nm and 420 nm, respectively.
Prostaglandin Measurements Cells for prostaglandin production assay were washed three times in phosphate-buffered saline and placed in fresh RPMI 1640 media (without FCS) in the presence or absence of 10 micrograms/ml Concanvalin - A
(Con-A). Cells were cultured for 24 hours, centrifuged and the supernates harvested. Supernates were stored at -20~C until assayed.
Prostaglandin standards PGE2 6-keto-PGFla and TBX2 were kindly supplied by Dr. J. Pike (Upjohn Company, Kalamazzo, Mch.). Tritium labelled compounds were purchased from New England Nuclear (Boston, MA). Rabbit antisera to PGE2 were obtained from the Pasteur Institute (Paris, France). Antibodies to 6-keto PGFla were raised in the laboratory (Rashida Karmali). The cross reactivity of these antibodies for the non-targeted PGs were to greater than 4% except for the PGE2 antisera which cross reacted 10% with PGE1 standard. The procedure for extracting the prostaglandins has been 13~02~

described earlier (Karmali, R.A., et al. (1982) Prostagl. Leukotr. Med. 8:565). Briefly, a trace of [3H]-PG was added to aliquots of standard and samples before being extracted once with petroleum ether. After acidification to pH 3.5, the samples were extracted twice with diethyl ether, dried under nitrogen and reconstitution in assay buffer. The efficiency of this extraction procedure to this point was 85-95~. Standard quantities of each prostaglandin (0-1000 pg) or the extracted sample to be measured were prepared in 0.1 ml aliquots of assay buffer. Antisera and label were added successively in 0.1 ml aliquots and incubated at 4~C for 8-12 hours. Bound and free [3H]-PG were separated by 0.5 ml dextran-coated charcoal (0.5-1.0% w/v) to estimate the amount of each compound in the unknown sample. The detection limit of this assay has been found to be lOpg.
The intra-assay coefficient of variation was 9.0%.
Alkaline and Acid pho~rh~tase~ b-Glucuronidase and N-Acetyl Glucuronidase Assays Cell extracts were prepared in 0.5 ml of PBS 1% NP-40, incubated 5 min at 24~C, then spun for 10 min at 3X104g. Supernates were collected and assayed. Extracts were assayed of their alkaline and acid phosphatase, b-glucuronidase and N-acetyl glucuronidase activity using the respective Sigma kits. Activities of extracts are ~' X

~ , .... . _ . ......... ... ..

02~6 expressed as change in absorbance per unit time per unit sample volume divided by the cell concentration in the culture or in the extract and compared to control activity. Measurements were made in a Beckman ACTA-CV
spectrophotometer.
Glycoconjugate assay Cytokine preparations were assayed in a [3H]-Glucosamine incorporation assay. Replicate wells were plated with 100 microliters of inducing agent to be tested. Previously washed (3X in PBS) HL-60 or U937 cells (1 X 107 cells/ml) in RPMI 1640 without FCS were added (50 microliters). After a four hour incubation 20 microliters of 25uCi/ml [3H]-Glucosamine in 1~ BSA (w/v) in PBS was added to the culture and plates were incubated for an additional 16 hours. Cells were harvested (Mini-Mash, Microbiological Associates, MD) onto glass filter paper with water wash (x4, 0.1 ml each), followed by 0.4N Perchlorate wash (x4, 0.1 ml) and water (2x, 0.1 ml). Radioactivity on glass discs was determined by liquid scintillation spectrophotometry.
Statistical Analysis Students' T test to compare means was carried out using the significance limits of a two tailed test.

X

~ 13~026~

Pr~paration of 5637 cell line conditioned mQdium (5637 CM) The human bladder carcinoma cell line 5637 has been reported to produce a colony stimulating factor for granulocytes and macrophages (Svet-Moldavsky, G.J., et al. (1980) Exp. Hematol. 8 (Suppl. 7):76). The cell line has been maintained at Sloan-Kettering Institute (New York, N.Y.) for several years. It is serially passaged by trypsinization in the presence of EDTA and grows rapidly to form an adherent monolayer in plastic tissue culture flasks. Routinely, cells are cultured in RPMI 1640, supplemented with 2 mM L-glutamine, antibiotics and 10% FSC. For purification of pluripotent CSF activity from 5637 conditioned medium (5637 CM), confluent cell cultures were intermittently cultured in medium containing 0.2% FSC. After 48-72 hours, 5637 CM was harvested, cells and cell debris removed by centrifugation (20 min, 10,000 x g), and stored at -20~C until use.
5637 cells also contain a multitude of subclones which either produce p-CSF in better yield and/or have less inhibitor present. Over 120 subclones have been isolated. One such subclone lA6 was found to produce at least twice as much as the parent cell line and possibly 5-10 fold more resulting in a range of between 2-10 X

13~0~6S

times more p-CSF from the lA6 subclone than from the parent 5637 cell line as determined by the assay methods outlined. This subclone or the parent cell line 5638 can be used to isolate p-CSF in good yield. Subclones are isolated by limiting single dilution techniques to produce a single cell per well in order to grow up a pure cell line from each well. Best results are obtained if the cells are distributed such that 37% of the wells (one out of every three) show growth at a certain dilution. There is then a good mathematical chance of obtaining subcloning to obtain outgrowth of only one cell from the one of three wells showing growth.
Subclone lA6 cell line is on deposit and available at Sloan-Kettering Institute for Cancer Research 1275 York Avenue, New York, N.Y. 10021. We refer to use of the lA6 in yielding sequence data on the protein p-CSF with subsequent preparation of the recombinant p-CSF from such a sequenced probe (P11 - top P14) as follows:
sequence data on the protein p-CSF with subsequent preparation of recombinant p-CSF from such a sequenced probe (P11 - top P14) as follows:
"(B)Sequencing of Materials Provided by Revised Methods In order to obtain a sufficient amount of pure material to perform suitably definitive amino acid X

134~2~fi sequence analysis, cells of a bladder carcinoma cell line 5637 (subclone lA6) as produced at Sloan-Kettering were obtained from Dr. E. Platzer. Cells were initially cultured Iscove's medium (GIBCO, Grand Island, New York) in flasks to confluence. When confluent, the cultures were trypsinized and seeded into roller bottles (1-1/2 flasks/bottle) each containing 25 ml of preconditioned Iscove's medium under 5% CO2. The cells were grown overnight at 37~C at 0.3 rpm.

- 17a -.. . . . ..

Cytodex-l* beads (Pharmacia, Uppsala, Sweden) were washed and sterili~ed using the following procedures. Eight grams of beads ~ere introduced into a bottle and 400 ml of Pss was added. Beads were su~pend~d by swirling gen~ly fo~
3 hours. ~fter allo~inq the bead8 to set~le~ the PBS was drawn off, the bead~ were rinsed in PBS and fresh pBS was added. The ~eads ~ere autoclaved for 15 minutes, Prior to use, the bead~ wer~ wa~hed in Iscove'S mediu~ plus 10~ fe~al calf serum (FCS) before adding fresh medium plu~ 10% FCS to obtain trea~ed bead6.

~ f~er removing all but 30 ml of the medium from each roller bo~tle, 3~ ml of fre~h medium plus 10~ F~S and 40 ml of treated bead6 were added to the bottles. The bottles were ~assed with 5% Co2 ~nd all bubbles were removed by ~uction. The bottles were pl~ced in rolle~ racks at 3 rpm for 1/~ hour before reducing the speed to 0.3 rpm.
After 3 hour~, an additional flask ~a~ trypsinized and added to each roller bottle oontaining ~eads.

~ t 40~ to 50~ of confluence the roller bottle ~ultures were washed with 50 ml PBS and rolled for 10 min.
before re~oving the PB~. The cells were cultured for 48 hours in medium A [Iscove I 8 medium containin~ 0.2% FCS, 10 8~ hydrocortisone, ~mM glutamine, 10~ units/ml Trademark . , .

13~02E~

penicillin, and 100 ug/~l ~trepto~ycin]. ~ext, the culture supernatant was harve~ted by centrifugation at 3000 rpm for 15 ~in., and stored at -70~C, The cùltures were refed with ~edium A ~ontaining 10% FCS and were cultured for 48 hour~.
After discarding the medium, the cells were washed with PBS
as above and cultured ~or 4 8 hou~s in medium A . The supernatant was again har~e~ted and ~reated as pr~viously described, ~ pproximately 30 liters of medium condi~ioned by lA6 cel~s were concentrated to abou~ 2 liters on a Millipore Pellicon ~ni~ equipped with 2 cassettes ha~ing 10,0~0 M.W.
cutoffs at a filtra~e rate of about ~00 ml/min. and a~ a ~etentate rate of about 1000 ml/min. The concentrate was diafiltered with about 10 liters ~f 50mM Tris (pH 7.8) u~n~
the same apparatus and some flow rates. The diafiltered concentrate was loaded ~t 40 ml/mi~. onto a 1 liter DE
cellulose colu~n equilibrated in 50 m~ Tris (pH 7.8). After loadinq, the column was washed at the same rate with 1 liter of 50mM Tris (pH 7.8) and then wit~ 2 liters of 50 ~M Tri~
(pH 7.8) with 50 mM Na~1, The col~mn was then sequentially eluted with s~x 1 liter solutio~s of 50~M Tris (pH 7 . 5~
containing the following concentrations of NaCl: 75~M;
lOOmM; 125mM; 150~M; l~OmM; and 300m~. Fractions ~50 ~1) were collected, and active fractions were pooled and .. .

13~02~

c~ncentrated to 65 ~1 on an Amicon ultrafiltration stirred cell unit equipped ~ith a YM5 membrane~ This ~oncentrate was loaded onto a ~ liter AcA54 g~l filtra~ion column equilibrated in PBS. The column was run at 80 ml/hr. and 10 ml fractions were collected. Activ~ fraction~ were pooled and loaded directly onto a C4 high performance li~uid chromatography (HPLC) column.

Sa~ples, ranging in volume from 125 ml to 850 ml and containin~ 1-8 ~g of protein, about 10% of which ~as hpG-CSF. Sample~ were lo~ded onto the column at a flo~ rate ran~ing fro~ 1 ml to 4 ml per minute. After loading and an initial washin~ w~th 0.lM am~oniu~ ace~ate (p~ 6~0-7.0) in 80~ 2-propanol at a flow rate of 1/ml~min. One ~illilite~
fraction~ were collected and monito~ed for protein~ at 2~0nm, 2~0nm and 280 n~.

As a re~ult Of purification, fractions containing hp~-CSF were ~learly separated la~ fractions ~2 and 73 of B0) from other protein-containing fraction~. HpG-CSF wa~
isolated (1~0-30 ~) at a purity of about 85~5~ and at a yield of about 50~. From thic purifi~d materlal 9~4g was ~ed in Run #4, an a~ino a¢~d sequence analysis wherein the protein sample was ~pplied ~o a TFA-activated ~lass fiber dis~ without polybrene, Se~uence analysi~ wa~ c~rried out . . _ ~

13402fi~

wi~h an AB 470A sequencer according to the methods of Hewick, et al., J. Biol. Chem., 256, 7990-7Y97 ~1981) and Lai, Anal. Chem. Acta. 163, 243-248 (19~4). The results of Run #4 appear in Table ~II.

TABLE III

Thr - P~o - ~eu - Gly - Pro - Ala - Ser - Ser - Leu - P~o-~0 Gln - Ser - Phe - Leu - Leu - Lys - ~Lys) - Le~l - (Glu) - Glu -~0 Val - Arg - Lys - Ile -(Gln~- ~ly - Val - Gly - Ala - Ala-~eu - X - X

In Run #~, ~eyond 31 cycle~ (~orresponding to residue 31 in Ta)~le III, no further significant se~uen~e infor~ation was obtained. In order to obtain a longer unambiguous sequence, in a Ru~ ~5, 14~g of hpG-CSF purified fro~ conditioned medium were reduced with 10~ of -~ercaptoethanol for one hour at 45~C, then thoroughly dried under a vacuum. ~he protein residue was then redi~solved in 5~ formi~ acid before being applied to a polybrenized glass fiber disc. Se~uenoe analysis was carried out as for Run ~4 ._, .

13~02~

a~ove. The results of ~un #5 are given ~n Table IV.

TABLE IV

~hr-Pro-Leu-Gly-Pro-Ala-Ser-Ser- Leu - P~o - Gln -Ser-Phe-Leu-Leu-Lys-Cys-Leu-Glu-Gln- Val - Arg - ~y6 -Ile-Gln-~ly-~sp-Gly-Ala-Ala-Leu-Gln- Phe - ~ys - Leu -Gly-~5 Ala-Thr-Tyr-Lys-Val-Phe-Ser-Thr-(Arg)-(Phe)-(~et)-X-The amino acid sequence given in Table IV W~8 sufficiently long ~44 residues) and unambiguous to conRt~uct probes for obtaining hpG-CSF cDNA." (end quote).
quote) A~monlum Sulfate Precipitation, Ion-exchange-chromatography, Gel filtration The firfit three puri~ication ~eps ((NH4)2SO4~precipitat~on, Ion-exchange-Chromatography on DE~E cellulose, DE 52, Whatman, Clifton, ~.J., and gel filtrati~n on A~A 54 Ultrogel, LKB, Inc. Rockland, MD) were performed as described (Wel~e, K., et al. ~1982) J. Exp.
Med, 156;4$4-464) except that AcA 54 was used instead of A~

, . ... ,. ,, ,, ~ ,. ~

134026~

44 (see also Descriptions of Fi~. 1 and 2).

Re~erse Pha~e High-Performance ~iquid ~hromatography ~RP-HPLC) RP-HPLC was performed with a Waters HPLC system (M

6, 000 solvent del~ very pump~, Inodel 400 variable wa~elength detector, data module and data p~ocessor, ~aters, Associates, Milford, MA). The separation was performed on a ~Bondapak ~18 column (Waters). The buffers used ~ere:
Buffer ~: 0.~ M acetic acid/0.2M pyridine, p~ 4.0; buffer ~:
buffe~ A in S096 l-p~opanol (Burdick and ~c)cson, IJab., Muskegon, MI). Acetio acid and py~idine were purcha~ed from Fisher, Scientifi~ Co.. T~e plurip~tent CSF con~aining pool obtained f~om gel ~iltration was acidified with aeetie acid to pH 4 . O and injected onto the uBond~pak C1~ col~mn without regard to sample volume. The column was was~ed with buffer A (10 min) and bound proteins were eluted using a steep gradient 0-40~ buffer B within the fir~t 20 mln and ~ 40 -10~ gradient of buffer B in 120 Tnin. THe flow ~te wa~
adjusted to lml/nlin and 3 ml fraction~ were collected. ~rom each f~action a 0. 5 ml aliquot wa~ supplemented with 10%
FCS, dialy~ed ~gainst PBS and tested for pluripotent CSF
activity .

.

13~26~

Isoelectrofoc~ing One ml of the purified pluripotent CSF wa~
supplemented wit~ 20~ glycerol (vol/vol) and 2% Ampholines (vol/vol), pR 3.5-10 (L~ Produets, Inc.). A 5-60% glycerol density gradient containing ~ A~pholines, p~ 3.5-10, ~as layered in~o a isoelectrofocusing colum~ (LR~ ~100). The pluripoten~ CSF sample was applied onto the isodense region of the gradient, followed by isoelecSrofocusing (2,000 V, 24 hour~). Fi~e ml fra~tions were collected and the p~I
determined in each fraetion~ The ~ractions were dialyzed against PBS and subsequently tested for pluripotent CSF
a¢tivity, Sodium Dodeeylsulfate-Polyacrylamide ~el Electrophoresis (SDS-PAGE) The diQcontinuous Tris-glyCine ~ystem of Laemmli (Laem~ .K. (1970) Nature 227:6~0-685) wa~ used for 1.5 mm slab gel~ of 15% a~rylamide. The samples (200 ng lyophilized protein eluted from HPLC) were treated with 1 SDS in 0.0625 M Tris-~Cl, p~ 6,8 at ~7~C for 1 hour under both reducin~ (5~ 2-mercaptoethanol~ and non-reducing condition~ and then loaded on the gel. After elec~rophoresis, gel~ were ~tain~d ~y the Biorad silver staining method ~Biorad Labora~ories, Ro~kville Centre, NY).
Apparent molecular weight~ ~ere determined using protein standards ovalbumin (Mw 43,000), chymotrypsinogen ~MW
~5,700~, beta-lactoglobulin (MW 18,400), lysozy~e ~
14,300) and cytochrome C (M~ 12,3001 (Bethesda Research Laboratories, Inc. ~a~ther~burg, MP~ or from Pharmacia Fine Chemical~, Pis~ataway, New Jersey. After treatment (see above) of lyophilized pluripotent CSF under non-reduced conditions and subsequent electrophoresis, parallel gels were sliced in ~ mm or 2 mm sections, respeetively and proteins from ea~h ~ e eluted either in~o 0.5 ml RPMI 1~40 containin~ 10% FCS or into phosphate bu~fered saline (P~S;
20 mM phosphate, O.lS M ~aCl). Af~er exten~ive dialysis, the eluted material was a~sayed for pluripotent CSF
a~tivity.

Protein Assay T~e protein content of samples we~e measured using the ~owry technique (Lo~ry, O,~., et al. ~19~1) J. Biol Chem. 193:265-275)~ For protein ~on~entration$ lower than 2 microgra~/ml, samples were subjected to S~S-PAGE, the protein bands were visualized by the silver staining technique and the protein concentration es~ima~ed ~y ~omparison wi~h a serial dilution of known amounts of proteins .

The example~ shown serve to illustrate the invention without limiting s~me.

.

13402~

Example I
Pluri~tent ~SF activity in 5~37 CM

Confluent layers of 5~37 human bladder carcinoma cells, when cultured for 48-72 hour~ in the presence of 10 ~CS, released into the culture medium 3,0~0 - 10,000 uni~s/ml of G~-CSF activity. ~edia conditioned in the presence of 0.~ FC5 ~till con~ained 10 - 30% ~f ~his ac~i~ity, ~hereas in ~erum free 5637 CM the activity drop6 below 5~ of the activity o~tained in the pre~ence of 10%
~CS. Although GM-C~F activity in 5~37 CM i~ readily detec~able in soft ~gar bone marrow cultures, no~ all batches o~ unfractionated 5637 CM support in vitro grow~h of BFU-E and CFU-GEMM. Four to ten times concentrated 5~37 CM
support in vitro growth of BFU-E and CFU-GE~M, Four to ten ~imes concentrate~ 5637 ~M reduced colony formation ~y CFU-~ 30-70~ indica~ing the presen~e of inhibitor (s) in 5637 CM. In~ibitor~ were removed after ion-exchange chromatography.

Example II
Puri~ication of pluripotent CS~

~ ~0-fold concentration of proteins from the 5637 CM was achieved by precipitation with ~m~onium culfate at - 2~ -~ ~ . .. ... .... . ..

1~3~2~

80~ ~at~ration. The dialyzed precipitate was loaded on to a DEAE cellulose (DE 52) column~ Bound protein~ were eluted with a salt gradient from 0.05-0~3M NaCl in 0.05 M Tris-~Cl, p~ 7.8. GM-C~F acti~ity eluted as peak l between 0.075 M
and O.lM NaCl and with a ~econd peak at 0.13 M NaCl (Fig.
1). Since only peak 1 revealed pluripotent CSP activity, was used only this pool for further purifications, Pea~ 2 included proteins with only ~-CSF activity. We calculated the "fold" purific~tion by measurin~ the GM-CSF ~ctivity of pluripotent CSF. In the unfractionated CM we could not di~criminate between GM-CSF activity a~ part of pluripotent CSF activity and GM-CSF activity without pluripotent properties. Therefore we considered the GM-CSF activity contained in peak 1 from DE 52 as the starting a~tivity (Table l).

Since in the subsequent purification schedule GM-CSF, BFU-E and CFU-GE~M activities copurified in all steps, we named these combined activitie~ "pluripotent CSF"
and have used this ~er~ thereafter. The protein~ of peak 1 of DE 52 chromatography ~including pluripotent C~F activlty) were co~centrated by dialyzing against ~0~ (w/v) Polye~hylenglyCol in PBS and purified further by ACA 54 Ultrogel gel filtration. The pluripotent CS~ activity eluted in fraction~ 42 -4~ as a single peak corresponding to _ 3"~_, ~" C.3 _ . _ _ , . . . _ . , . , . . _ _ , , , " ~ . , _ _ _ , . , _ 13 ~02 G~

a molec~lar weight of 32,000 daltons (Fig. 2). This ~tep resulted in a 65~ recovery of activ~tie~ and a 15 fold increase of cpecific activities (Table 1).

The final ~tep involved chromatography on a reverse phase HPLC column (ubondapak* C 18). Tlle majority of proteins did not bind to this col~mn or eluted at low l-propanol con¢ent~ation~ (less th~n 20% l-propanol; Fig.
3). A minor peak of GM-CSF activity without activity in the CFU-GEMM and BFU-E assays but differentation inaucing activi~y on HL-60 leukemic cells was eluted at around ~0 l-propanol. Pluripotent CSF ~ctivity eluted as a single sharp peak at 42 ~ l-propanol (Fi~, 3). This purifica~ion step resulted in a 600-fold increase of specific ~ctivity and a 25 % recovery of activity. The protein content of the H~LC fra~tion was measured by comparing the density in silver stained SDS-PAG~ with protein standards of known concentrationS. Using this meas~rement, we obtained a specific~ activity of 1.$ x 10~ U/mg protein and a final purification of ~,00~-fold, calculated from the fir~t peak of DEAE cellulo$e chromatography. The overall yield ~a~
6.2%. The puriflc-ation table ~ith the degree of purification of pluripotent CSF as measured by GM-CS~
act~vi~y, protein content, specific activity and y~eld is detailed in Table 1.
* Trademark - 28 -,~ ,.

13~026~
The final preparation obtained after HPLC(pluripotent CSF activity peak fraction) was analyzed on a 15~ SDS-PAGE gel followed by the sensitive silver staining technique (Fig. 4). Only one major protein band with a molecular weight of 18,000 was seen under both, reducing (5% 2-mercaptoethanol) (Fig. 4) and non-reducing conditions. Since the buffer system used for HPLC did not allow monitoring the protein elution pattern by measuring the optical density at 280 nm, we applied proteins of all active fractions on SDS-PAGE. The density of the stained protein band at 18,000 MW in the peak and side fractions was proportional to the amount of biological pluripotent CSF activity. After electrophoresis under non-reducing conditions, a parallel gel was sliced into 4 mm sections and proteins eluted from each slice into RPMI 1640 containing 5~ FCS.
Pluripotent CSF activity was found to be localized in the slice number corresponding to 18,000 MW (Fig. 5).
In three additional, independent purification runs, pluripotent CSF had the same properties and specific activity as described above. In all three runs, parallel gels were sliced into 2 mm sections, proteins eluted into PBS and tested for pluripotent CSF activity.
Re-electrophoresis of the proteins eluted from the slices with pluripotent CSF activity again revealed one 1~02~

single band in a silver stained gel with a molecular weight of 18,000, identical to that shown in Fig. 4.
However, further work using markers from Pharmacia shows the molecular weight of the glycosylated p-CSF to be 19,600. The unglycosylated recombinant protein shows a MW of 18,800.
The purified CSF was also subjected to isoelectrofocusing analysis using a 5-60% glycerol gradient in an IEF column and 2% Ampholines, pH 3.5 - 10.
Pluripotent CSF activity was localized in one fraction (5 ml) with an isoelectric point of 5.5 (Fig. 6). The total recovery of pluripotent CSF activity applied to the column was approximately 20%.
Pluripotent CSF activity did not b:ind to a Concanavalin A agarose. Treatment with neuraminidase did not abolish the biological activity and did not change the IEP. However, the isoelectrofocusing under our conditions did not allow judgment of minor changes of the IEP. These findings suggest that glycosylation might not be a major structural feature.
The partial amino acids sequence was determined by Applied Molecular Genetics (Thousand Oaks, California) on an AB 1470A- Beatrice microsequencer. From the amino terminal end the sequence is Thr, Pro, Leu, Gly, Pro, X

134~2~
.

Ala, Ser, Ser, Leu, Pro. Also see the extended 44 residue sequence above.
Example III
Biological activity of pluripotent CSF: Progenitor cell stimulation and effect on leukemic cells.
1. Progenitor cells:
Fifty units of GM-CSF activity, enough to support the half m~x;m~l growth of CFU-GM, had no clear effect in a CFU-GEMM assay; however, 500 U/ml (GM-CSF activity) of pluripotent CSF clearly supported the growth of human mixed colonies (CFU-GEMM), megakaryocytic colonies, and early erythroid colonies (BFU-E) under our experimental conditions (Table IIA & IIB).
Pluripotent-CSF supports the growth of colony forming progenitors of the granulocyte, mixed granulocyte, macrophage, eosinophil and megakaryocytic cell types. These results can be seen for example in vitro.
We show the results of comparison of 5637-CM and lA6-CM in Table IIC at dilutions of 1/10 through 1/1600.
The 1/10 dilution of lA6 shows an inhibitor to be present in the CM. Essentially this table serves as an example that the lA6 subclone of 5637 has 8.7 times more p-CSF in U/ml under growing conditions containing FCS.
When purifying p-CSF the FCS is reduced to 0.2%.

13~2~

2. Pluripotent CSF also induces the differentiation of leukemic cells. For example, leukemic cell lines HL-60 and WEHI-3B (D+) are induced to differentiate along the granulocytic and/or macrophage pathway. The human leukemic cell line KG-1 responds to pluripotent CSF by increased colony formation in agar and proliferation in liquid suspension culture.
As for mature cells, pluripotent CSF induces increased protein content, for example, in macrophages, whereas IL-3 is not reported to be active on macrophages ~Table III). 50 U/ml and 200 U/ml of GM-CSF activity of the pluripotent CSF were needed to induce half-maximal differentiation of the leukemic cell lines WEHI-3B (D+) and HL-60, respectively. These cells were used in a test system (Metcalf Int. J. Cancer (1980) 25:225 and Fibach et al. (1982) J. Cell. Physiol. 113:152) Table IV(A&B) to show the effect of pluripotent-CSF on leukemic cells.
U937 was obtained from Dr. Nilsson and HL-6() from Dr.
Gallo as freeze-backs of early passages. HL-60 is a myeloid cell line from an acute promyelocytic leukemia [Gallagher et al. Blood 54:713 (1979)]. U937 is a histiocytic lymphoma cell line (Sundstrom and Nilsson (1976) Int. J. Cancer 17:565).
Differentiation of leukemic cells lines in vitro can be achieved by a variety of nonphysiologic (e.g.

... . _ .. .

13iO~6 DMSO, phorboldiesters) and physiologic (e.g. retinoic acid, vitamin D3) inducers (Koeffler et al. (1983) Blood 62:709). Murine G-CSF is known to be a potent inducer of differentiation of WEHI-3B (D+) murine myelomonocytic leukemia cells, whereas Interleukin 3 lacks this activity (Nicola et al. (1984) Immunol Today 5:76) See Table V).
Pluripotent-CSF was tested for leukemia differentiating activity (GM-DF) in a clonal assay system described by Metcalf (1980) Int. J. Cancer 25:~25; Fibach E., et al., J. Cell. Physiol. 113:152, (1'382) using murine WEHI-3B (D+) and human HL-60 promyelocytic leukemia cell lines (Platzer et al., J. Exp. Med.
_ :1788-1801, (1985). Quantitation of GM-DF was obtained by incubation of leukemic cells in agar with serial dilutions of pluripotent CSF. Pluripotent CSF had GM-DF activity on both cell lines. However, HL-60 required approximately 2.5-5X higher concentrations of Pluripotent CSF to achieve 50~ differentiated, spreading colonies versus undifferentiated tight blast cell colonies, than did WEHI-3B (D+) (Platzer et al., J. Exp.
Med. 162:1788-1801, (1985).
Morphological and cytochemical analysis of HL-60 colonies were performed using alpha-naphthylacetate esterase (ANAE) and luxol fast blue (LFB) stains, as X

l~4o2~

markers of the monocyte, macrophage and eosinophil granulocyte lineage respectively. In the presence of pluripotent CSF there is observed an increase in the number of colonies containing polymorphonuclear cells (by hematoxylin stain), LFB cells and in intensit;y of ANAE

stain. Therefore, pluripotent CSF induces differentiation along the macrophage as well as granulocyte pathway. The human leukemia cell line KGl (courtesy Dr. H.P. Koeffler) responded to pluripotent CSF in a dose dependent manner with increased colony formation in agar and increased 3H-thymidine incorporation after 24-48 hours in suspension culture. This might indicate that the GM-DF activity of pluripotent CSF reflects the differentiating capacity of a given leukemic cell line rather than an intrinsic property of the factor.
CM from SK-HEP and cell line 5637 containing pluripotent CSF (free of Interferon) has also shown acquisition of immunoglobulin Fc receptor, growth inhibition, increased expression of monocyte related surface antigens and an increase in lysosomal enzyme content as well as (to distinguish P-CSF from Interferon-gamma) increased receptors for chemotactic peptide, increased hydrogen peroxide release in response to phorbol myristic acetate (PMA) stimulation and the X

1302~

release of prostaglandins (PGE2 and 6-keto PGF1A) as features of differentiation of human promyelocytic leukemia cell line HL-60 and monoblastic leukemia cell line U937. These broad range differentiation factors are thus different from Interferon and conventional colony stimulating activity (CSA). Highly purified Fluripotent CSF increased the receptors for chemotactic peptide and increased glycoconjugate synthesis as a feature of differentiation in both the human promyelocytic leukemia cell line HL-60 and monoblastic leukemia cell llne U937.
3. Pluripotent CSF shows species cross-reactivity on normal murine bone marrow and leukemic cells.
Normal mouse bone marrow cells cultured in agar for 7 days in the presence of saturating concentrations of Pluripoietin formed approximately 10% of the colonies supported by WEHI-3B conditioned media as standard source of CSF('S). All colonies formed in the presence of Pluripotent CSF were of similar morphology, not staining for alpha-naphtyl-acetate esterase or Kaplow's myeloperoxidase; this suggests that a subpopulation of murine colony forming progenitors is responsive to Pluripotent CSF. Weak cross-species activity was found on continuous murine mast cell lines, established as described from murine long-term bone marrow cultures (Tertian et al. (1980) J. Immunol . 127:788). 5,000 A

13~2fi~

cells/well of a mast cell growth factor (MCGF) dependent murine mast cell line were incubated for 24 hours at 37~C
in 96 well plates with serial dilutions of growth factors, and then assayed for H-thymidine uptake as described (Yung et al. (1981) J. Immunol . 127:794).
Results demonstrate little more than 10% murine MCGF
activity of Pluripotent CSF as compared to ConA-LBRM CM, which was used as a standard preparation of murine MCGF.
The murine Interleukin-3 dependent cell line FDC-P2 (courtesy Dr. M. Dexter) did not respond with increased H-thymidine uptake to concentrations of Pluripoietin as high as 2,000 U/ml.
We herein describe the purification of a pluripotent CSF, which is constitutively produced by the human bladder carcinoma cell line 5637, its lA6 subclone or SK-HeP-1. This protein is capable of stimulating the in vitro growth of mixed colony progenitor cells (CFU-GEMM), early erythroid progenitor cells (BFU-E), megakaryocytic (CFU-Mega), granulocyte- macrophage progenitors (CFU-GM) and in addition induces differentiation of both the murine myelomonocytic (WEHI-3B (D+)) and the human promyelocytic (HL-60) leukemic cell lines. The purified pluripotent CSF had a specific activity in the GM-CSF assay of 1.5 X 10 U/mg protein.

X

. . . .... ~

13~02~fi To our knowledge, this is the highest specific activity for a human pluripotent CSF reported to date.
Pluripotent CSF has a molecular weight of 32,000 by gel filtration and 18,000 by SDS-PAGE under both, reduced and non-reduced conditions and an isoelectric point of 5.5.
Pluripotent CSF activities could be eluted from gel slices representing the same molecular weight range as the stained protein band.
The purified protein shown in SDS-PAGE is consistent with pluripotent CSF because:
1) the profile of protein elution visualized in SDS-PAGE and elution of pluripotent CSS
activity (Fig. 3) from reverse phase HPLC
columns is equivalent in the major fraction and side fractions;
2) additional chromatography of the purified protein on diphenyl or octyl reverse phase HPLC
coll~mns using acetonitrile or ethanol as organic solvents for elution did not lead to a separation of protein and pluripotent CSF
activity;
3) identical localization of protein band and pluripotent CSF activity in a preparative SDS-PAGE;

X

. . . ~ ., 1~026~

4) high specific GM-CSF activity (1.5 X 10 U/mg protein).
Since purified pluripotent CSF is apparently homogeneous, amino acid sequence analysis of the purified protein has been initiated and is partially determined.
Based on the molecular weight of pluripotent CSF as 18,000 it could be calculated that 1 U of pluripotent CSF
was equivalent to 6.7 pg protein or 3.7 X 10 moles. A

pluripotent CSF concentration of 50 U/ml or 1.85 X 10 M

was required for half maximal colony formation for CFU-GM
activity in normal human bone marrow cells.
A ten-fold increase in the amount of pluripotent CSF (500 U/ml GM-CSF activity) was required for clear detection of human CFU-GEMM and erythroid BFU-E
activities (Table II); a 1-2 or 1-2.5 fold increase in pluripotent CSF (e.g. 50-200 U GM-CSF) was needed to induce the differentiation of either WEHI-3B (D+) or HL-60 leukemic cells respectively. These data suggest that the particular action(s) of pluripotent CSF are determined by its concentration as first suggested by Burgess and Metcalf (Blood, 56:947-958) in the murine system. The fact that human pluripotent CSF is able to induce differentiation of leukemic cell lines makes it a protein with unique properties, since for the murine 13~2~fi multi CSF (Interleukin 3) no differentiation activity on leukemic cells has been reported (Ihle, J.N. et al., J.
Immunol. 129:2431-2436 (1982); Nicola, et al.r Immunol.
Today, 5:76, (1984); Watson, et al., Immunol. Ioday, 5:76 (1983); and Fung et al., Nature, 307:233, (1984). (Table V compares the two entities). The murine IL-3 dependent cell line FDC-P2 (Dr. M. Dexter) did not respond with increased 3H-thymidine uptake to Pluripotent-CSF as high as 2,000 U/ml.
Several human CSFs (GM-CSF, G-CSF, eosinophilic CSF, erythroid potentiating activity) have molecular weights between 30,000 and 40,000 on gel filtration (Nicola, N.A. et al., Blood, 54:614-627, (1979); Gold, D.W. et al., Proc. Natl. Acad. Sci. U. S.A., 77:593-596, (1980); Lusis, A.J. et al., Blood 57:13-21, (1981);
Abboud, C.N. et al., Blood, 58:1148-1154, (19~1); Okabe, T. et al., J. Cell. Phys., 110:43-49, (1982), which is similar to the native molecular weight of the pluripotent CSF described here. However, only partially purified erythroid-potentiating activity has been report:ed to have activity in a CFU-GEMM assay (Fauser, A.A., et al., Stem Cells 1: 73-80, (1981).
Constitutive production of pluripotent CSF by the bladder carcinoma cell line 5637 and its lA6 subclone or X

- 134026~

other 5637 subclones suggests that these are valuable sources for large scale production and for isolation and cloning of the gene which codes for pluripotent CSF. The availability of purified human pluripotent CSF has important and far reaching implications in the management of clinical diseases involving hematopoietic derangement or failure, either alone or in combination with other lymphokines or chemotherapy. Such disorders include leukemia and white cell disorders in general. It is useful in transplantation, whether allogeneic or autologous, to augment growth of bone marrow progenitor cells. It can be used in induced forms of bone marrow aplasia or myelosuppression, in radiation therapy or chemotherapy-induced bone marrow depletion, wound healing, burn patients, and in bacterial inflammation.
Here the action of pluripotent-CSF may possibly be due to enhancement of chemotactic peptide receptors or by functioning as a chemo-attractant. It is also found in saliva so may prevent tooth decay and oral infection.
p-CSF may be used alone or together with recombinant material or in conjunction with erythropoietin for treatment in hematopoietic disorders.

X

134026~

Table I: Purification of human pluripotent CSF
Fraction Protein Total Specific Purifica- Yield activity activity tion (UX10 )(U/mg)(fold) (9~) 5637 CM 2g 12 6 X 103 - 100 DEAE 300 mg 5 4 1b 42 cellulose 1.7X10 AcA 54 13 mg 3.1 14 26 Ultrogel 2.4X10 RP--HPLC 5 ~lg 0.74 9 000 6.2 1 . 5X108 GM-CSF activity of pluripotent CSF; U=Units;

estimate of fold purification based on starting activity of peak 1 of DEAE cellulose chromatography.

TABLE IIA
Table IIA: Comparison of CFU-GEMM and BFU-E activities of pluripotent CSF (500 U/ml GM-CSF activity) CFR-GEMM BFU-Ea (Colonies + 1 SEM) (Colonies + 1 SEM) Experiment 1 2 3 1 2 3 #

Medium 0.3+0.3 0 0 42+6 17+3 17+2 PHA--LCM7 + 1 3+0 3.3+0.3 67+1 65+3 34+3 7 7+2.1 4+0. 82.3+0.9 85+6 31+1 28+2 Plurlpotent ' -- -- -- -- -- --CSF

aTarget cells were 5 X 104/ml low density, non-adherent and T cell depleted normal human bone marrow cells.
Experiment 3 was done in the absence of Hemin.

X

13~2~6 bMedium conditioned by leukocytes from patients with hemochromatosis in the presence of 1% PHA.
(positive control) TABLE IIB
Table IIB: Activity of Pluripoietin on pre-CFU

Pluripoietin Exp.1 Exp. 2 Exp. 3 concentration 7 days in 7 days in 5 days in 9 days in suspension suspension suspension suspension culture culture culture culture 1000 U/ml416il8 20i4 32i5 80i8 500 U/ml 367i57 39i4 n.t. n.t.
100 U/ml n.t. 29i6 73i4 30i5 10 U/ml n.t. 12i3 52i3 34i4 Control 200il6 8i2 26i5 20i4 medium ~,.

TABLE llC
COMPARISON OF ACTIVITY OF 5637 AND lA6 IN GM CFU ASSAY

Dilution Cells 1/10 1/100 1/200 1/400 1/800 1/16000 U/ml Colonies 190 + 17 75 + 1 18 + 6 0 0 0 2750 5637-CM +
%100~ (max)39~ of 9% 350 max.
Colonies 0 + 0 93 + 7 130 + 11 124 + 0 64 + 0 30 + 6 24,000 lA6 0~ 49 69 65 34 16 3,000 2'3 cr~

02~
Legend Table II
Normal human bone marrow cells were separated by Ficoll, adherence to plastic and depletion of T cells by rosetting with neuraminidase treated sheep red blood cells. Quadruplicate cultures of 25,000 cells in 100 microliters/well were incubated in 96 well flat bottom tissue culture plates in Iscove's modified Dulbecco's medium supplemented with 30% fetal bovine serum (FBS), 5X10 M 2-mercapto-ethanol and serial dilutions of purified Pluripoietin or control medium for 5, 7 or 9 days at 37~C in 5% CO2 in air. Contents of each well were then resuspended and incorporated into 1 ml agar system in supplemented McCoy's with saturating concentrations (10% v/v) of 5637 CM. Colonies were scored after 7 days of incubation at 37~C in a humidified atmosphere of 5% CO2 in air. Results are expressed as mean colony number per well + 1 standard deviation. CFU input on day 0 were 79 + 5 (exp. 1), 26 + 1 (exp. 2) and 22 + 3 (exp. 3) per well. Bone marrow cells from the donor for experiment 1 grew high numbers of CFU-GM in two unrelated experiments;
no pathophysiological situation was recognized.

A

134026~

TABLE III
Influence of Pluropoietin on protein content in cultures of human macrophages Time in culture Adherent cell Adherent cell protein in response protein in response to control medium to Pluripoietin ~g/coverslip ~g/coverslip Day 1-210.0 + 2.0 28.6 + 7.7 Day 1-320.4 + 1.6 26.8 + 2.5 Day 1-428.4 + 1.6 41.2 + 1.9 Day 4-528.8 + 1.6 28.1 + 3.6 Day 4-643.1 + 4.7 28.1 + 3.6 Day 4-738.2 + 6.1 44.~ + 0.7 Legend Table III
Normal human monocytes/macrophages were isolated from peripheral blood mononuclear cells by adherence to glass surfaces . Two X 10 cells were plated per 13 mm diameter coverslips in 0.1 ml of supplemented RPMI 1640 containing 25% fresh frozen human serum. After 2 hours at 37~C, nonadherent cells were removed by rinsing, and coverslips transferred to 24 well tissue culture trays (day 0). On day 1 and 4, supernatants were replaced by fresh culture medium containing 500 U/ml of purified Pluripoietin or control medium. Protein content was determined 1 to 3 days thereafter by rinsing coverslips 13~q26~

free of culture medium, solubilizing adherent cell protein in 0.5 N NaOH and measuring protein concentration according to the method of Lowry . Results are expressed as mean + 1 standard deviation, from triplicate cultures.

X

TABLE IVA

Table IVA. Leukemia differentiation (GM-DF) activity of purified Pluripoietin Purification CM-CSF activity GM-DF activity GM-DF activity WEH1-3B (D+) HL-60 Specific U/ml U/ml Ratio U/ml Ratio activity DF/CSF DF/CSF
U/mg protein I 1.5X1084,000 246,000 2.9 54,000 0.6 II 201,000 502,000 2.5 80,000 0.4 1.25X10 C~
~3 13ql~266 TABLE IVB
Glycoconjugate Synthesis Inducer HL-60 U937 CPM/5X10 cells media 465 210 gIFN500 U/ml 1029a :1500a 100 U/ml 800a 537a 50 U/ml 410 258 LK 50%
(500 U/ml gIFN) 427 910a 5637 CM (GM-CSA) 2 kU/ml 1828a 1200a 1 kU/ml 980a 780a 500 U/ml 670a 490a Pluripoetin 1 kU 4235a 2400a pp aCSF 1 kU
1439a 604a SK-Hep CM 50% 420 200 GCT-CM100% 490a 250 PMA 3.0 ng/ml 2000a 1700a 50.0 ng/ml 420 240 aIFN5000 U/ml 425 230 IL-2 100 U/ml Glycoconjugate synthesis was measured as follows, cells (5X105) were incubated with inducers for 4 hours then - 48 ~
X

... .. , .. . _ _ . .. . . .. .. . . .

13~026~

glucosamine incorporation was evaluated after an additional 16 hours.
Results are mean values from three or more experiments.

a, Significantly different from control, p less than 0.05 by Students T test; b, human P-CSF, units assigned by CFUC
activity; c, partially purified aCSF-like activity, units assigned by CFUc activity.
Legend Table IVa For determination of specific activity, protein concentration of purified Pluripoietin was estimated by comparison with serial dilutions of known amounts of protein in SDS-PAGE, visualized by silver stain. Due to the low frequency of CFU-GEMM in normal human bone marrow cells, the biological activity of Pluripoietin had to be measured using the GM-CSF assay. We compared the ability of serial dilutions of Pluripoietin and a previously determined laboratory standard of 5637 CM to support GM-colony formation in 1 ml semi-solid agar cultures containing 10 low density, normal human bone marrow cells. Fifty units of GM-CSF activity were arbitrarily defined as inducing 50% of maximal colony growth on day 7 of culture. Concentrations of 500 U/ml of P~uripoietin were sufficient to stimulate colony growth from CFU-GEMM
and BFU-E comparable to that supported by optimal amounts of phytohemagglutinin-activated lymphocyte conditioned media. Two independent purifications (I and II) resulted y 13~266 in very similar specific activity. Due to different amounts of starting material, the final concentration of biological activity differs between I and II, but is useful for comparison of GM-CSF and leukemia differentiating activity (GM-DF) of Pluripoietin. GM-DF
activity was determined by incubating 3X10 /ml WEHI-3B

(D+) or 10 /ml HL-60 leukemic cells in 0.3% agar in McCoy's medium containing 12,5% FBS with serial dilutions of Pluripoietin. Cultures were scored on day 7 (WEHI-3B) and day 14 (HL-60) for induction of dispersed, differentiated colonies vs. tight, blast cell colonies (Metcalf, et al. (1980) Int. J. Cancer 25:225 and Fibach, et al., (1982) J. Cell. Physiol. 113:152). Fifty units of GM-DF activity were defined as inducing 50%
differentiated colonies.

, . . ~ ... . . .. . . . . . ..

13~6~

Table V. Biological activities of purified human Pluripoietin and murine Interleukin-3.

Activity Pluripoietin ) Interleukin-3 Clonal growth of hemopoietic progenitors:
CFU-GEMM + +
BFU-E + +
CFU-G,M,GM + +
CFU-EOS + +
CFU-MEG n.t. +
pre-CFU-c (~GPA) + n.t.
stem cell multi-plication (CFU-s) +
c ) +
Species crossreactlv1ty Leukemia differentiating activity (GM-DF) on: +
WEHI-3B (D+) H-TdR uptake in cell +
lines: _ +

Murine mast cell + +
lines (MCGF activity) n.t. +
Histamine production n.t.
Protein synthesis of mature macrophages +
Induction of 2D-SDH
Growth of: +
natural cytotoxic cells n.t.
pre-B cell clones a) Pluripoietin was tested on human target cells, if not noted otherwise.
b) Interleukin-3 activity on murine target cells, if not noted otherwise.
X

, . . . .. ~ . . ......

- i3~0~
Data derived from literature, except GM-DF and acvivity on KG1.
c) Activity on bone marrow derived colony formation in agar cultures.
No human test system available.
n.t. Not tested.

- 51a -X

" , . _ . . . ..

Claims (16)

1. A purified glycoprotein human pluripotent colony stimulating factor that is characterized by:
a) a molecular weight of 19,600 daltons under reducing and non-reducing conditions as determined by SDS-PAGE;
b) a molecular weight of 32,000 daltons as determined by gel filtration;
c) having the ability to stimulate in vitro growth of early hematopoietic progenitor cells as mixed colony progenitor cells, early erythroid progenitor cells, megakaryocytic cells and granulocyte-macrophage progenitors d) an isoelectric point of 5.5;
e) having the pharmacological activity to induce differentiation of leukemic cells; and f) having a partial amino acid composition as determined from the amino-terminal end as Thr, Pro, Leu, Gly, Pro, Ala, Ser, Ser, Leu, Pro, Gln, Ser, Phe, Leu, Leu, Lys, Cys, Leu, Glu, Gln, Val, Arg, Lys, Ile, Gln, Gly, Asp, Gly, Ala, Ala, Leu, Gln, Phe, Lys, Leu, Gly, Ala, Thr, Tyr, Lys, Val, Phe, Ser, Thr, (Arg), (Phe), (Met), X.

2. The factor of claim 1, wherein the factor is purified to homogeneity.

3. The factor of claim 1, wherein the factor is derived from human cells.

4. The factor of claim 1, wherein the factor is derived from human tumor cells.

5. The factor of claim 1, wherein the factor is derived from human bladder cell line 5637, subclone 1A6 from 5637, and hepatoma cell line SK-HEP-1.

6. The factor of claim 1, wherein factor is further characterized by having a specific activity of at least 1.5 x 108U/mg as measured in the GM-CSF activity assay.

7. The factor of claim 6, wherein the leukemic cells induced are leukemic cell lines.

8. The factor of claim 1 having the ability to induce the acquisition of increased receptors for chemotactic peptide and increased glycoconjugate synthesis.

9. Method for preparing the factor of claim 1, wherein the factor is prepared by:
a) high-salt precipitation of protein from a cell-free medium;
b) ion exchange chromatography of the precipitate from a) above;
c) gel filtration of active fractions from step b); and d) reverse-phase high performance liquid chromatography of active fractions from step c) above.

10. Use of pharmacologically active doses of the factor of Claim 1 for inducing differentiation of human leukemic cells.

11. Use of claim 10 to treat leukemia.

12. Use of therapeutically effective doses of the factor of claim 1 for enhancing bone marrow recovery in allogeneic or autologous transplantation and in treatment of radiation, chemically, or chemotherapeutically induced bone marrow aplasia or myelosuppression.

13. Use of therapeutically effective doses of the factor of claim 1 for treating conditions requiring optimum neutrophil or macrophage function.

14. Use of claim 13, wherein the conditions are wounds, wound infection, or burn wounds.

15. Purified subclonal cell line 1A6 isolated from a parent human tumor bladder cell line 5637.

16. Cell line 1A6 of claim 15, wherein 1A6 produces between 2-10 fold higher amounts of p-CSF than the parent cell line.