Purification and Properties of a Mammary-Uterine-Pituitary Tumor Cell Growth Factor from Pregnant Sheep Uterus*

A mammary-uterine-pituitary tumor cell growth factor has been purified from lyophilized powders of pregnant sheep uteri by a five-step procedure. Uterine-derived growth factor (UDGF) was extracted from the powders with 0.1 M acetic acid, heated at 95 OC, and further purified by sulfopropyl-Sephadex C-25, Sephadex G-50, and carboxymethyl-Sephadex C-25 chro- matography. From 500 g of uterine powder, 40 to 50 mg of UDGF can be isolated at an overall yield of 33%. The degree of homogeneity of the final preparations was estimated by 8 M urea, 0.1% sodium dodecyl sul- fate-polyacrylamide gel electrophoresis (PAGE), and by PAGE under nondissociating conditions at either pH 8.5 or 4.5. In all PAGE experiments, the purified UDGF preparation showed a single Coomassie blue- stained band that directly corresponded to the only area of elution of UDGF activity from duplicate un- stained gels. Molecular sieve high performance liquid chromatography HPLC, reverse phase HPLC on an octylsilyl ((28) column, and hydrophobic chromatography on octyl-Sepharose CL-4B all confirm a similar degree (Le. ~ 9 0 % ) of homogeneity. The M, of UDGF estimated by urea/sodium dodecyl sulfate-PAGE was 4200

A mammary-uterine-pituitary tumor cell growth factor has been purified from lyophilized powders of pregnant sheep uteri by a five-step procedure. Uterinederived growth factor (UDGF) was extracted from the powders with 0.1 M acetic acid, heated at 95 OC, and further purified by sulfopropyl-Sephadex C-25, Sephadex G-50, and carboxymethyl-Sephadex C-25 chromatography. From 500 g of uterine powder, 40 to 50 mg of UDGF can be isolated at an overall yield of 33%. The degree of homogeneity of the final preparations was estimated by 8 M urea, 0.1% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (PAGE), and by PAGE under nondissociating conditions at either pH 8.5 or 4.5. In all PAGE experiments, the purified UDGF preparation showed a single Coomassie bluestained band that directly corresponded to the only area of elution of UDGF activity from duplicate unstained gels. Molecular sieve high performance liquid chromatography HPLC, reverse phase HPLC on an octylsilyl ((28) column, and hydrophobic chromatography on octyl-Sepharose CL-4B all confirm a similar degree (Le. ~9 0 % ) of homogeneity. The M, of UDGF estimated by urea/sodium dodecyl sulfate-PAGE was 4200 2 500 and, by molecular sieve HPLC, 6200 f 1000. The isoelectric point of UDGF was estimated as PI = 7.3. The UDGF isolated showed marked cell-type specificity for established cell lines that were derived from estrogen-responsive tumors; purified sheep UDGF was mitogenic for MTWOPL rat mammary tumor cells (at 10"' to lo-' M concentrations) while showing no mitogenic activity toward normal rat diploid fibroblasts. UDGF also promoted growth of uterine-derived tumor cells and the GH3/C14 rat pituitary line. Measuring growth as an increase in cell number, UDGF supported the logarithmic growth of the MTW9/ PL rat mammary tumor cells over 6 days; other known hormones and growth factors were not able to substitute for the UDGF mitogenic action on MTW9/PL cells. It is concluded that a rapid, high-yield method of purification of a new uterine-derived growth factor activity has been developed. Polypeptide  of various animal species (1)(2)(3)(4)(5)(6), noncellular plasma fractions (7-lo), cellular elements of the blood (11)(12)(13), and cells growing in culture (14)(15)(16)(17); these activities have been implicated in mammalian cell growth in vivo (7)(8)(9)(18)(19)(20)(21) and in vitro (see reviews 22 and 23). Several of these mitogens have been purified to homogeneity by procedures that utilize extraction into acidic pH solutions and/or treatment at high temperatures to remove major impurities. The mitogenic species isolated by these methods show molecular weights ranging from 6,000 to 63,000 (1,6,11) and primary structures consisting of either one polypeptide chain (1) or two polypeptide chains covalently attached by disulfide bonds (11). Major members of the acid-stable family of growth factors are EGF' (l), MSA (mixture of IGF I and IGF 11) (10,17). SmC (IGF I) (9), IGF I1 (8, 9), PDGF (11)(12)(13), TGF (15,16), and SGF (14). While, to some degree, members of this group share common properties (i.e. insulin-like responses of IGF I and IGF II), most are distinct molecular entities which, in some cases, interact with different specific cell-surface receptors (24, 25), and in other cases interact with a common receptor (26).
From data available, it is apparent that several of the purified growth factors have broad cell type and species specificities as demonstrated by their ability to promote growth of a wide variety of cells in culture. Growth factors in general also have not shown marked species specificity, since mitogens isolated from human sources promote growth of mouse cells in culture (11)(12)(13), and vice versa (24). A widely recognized example of this broad cell type and species specificity is mouse EGF, which promotes growth of human fibroblasts (24), mouse 3T3 cells (27), rat liver cells (28), chicken epidermal cells (29), rat testis cells (301, and bovine corneal and lens epithelium (311, as well as many other cell types (see reviews 23 and 32).
In addition to the other cell types stimulated to grow by EGF, it has been shown that this activity promotes normal mouse mammary cell growth in culture (33), and that it is an important component of serum-free medium mixtures of hormones, attachment factors, and growth factors that are now used to maintain continuous growth of human mammary tumor cell lines in culture (34,35), and normal mouse mam-

Growth Factors for Estrogen-responsive Tumor Cells
mary cells cultured in collagen gels (36). Nevertheless, the role of EGF in mammary growth in vivo is not clear. A question remains whether plasma levels of EGF correlate with mammary growth in vivo, since EGF activity is androgen related (37) in submaxillary glands of mice, and androgen treatment has no measurable effect on mouse plasma EGF levels while estrogen treatment causes a depression of plasma levels. However, another investigator (38) has reported that plasma EGF is elevated during pregnancy, a condition accompanied by high plasma levels of estrogens and progestins and rapid mammary growth. Since data from in vivo experiments were contradictory, the effect of EGF on mouse mammary gland development in vitro has been studied (39); results demonstrate a potential role of EGF in mammary gland development. Also, EGF has been reported to be the main mitogenic activity found in human milk, again implying a role of this growth factor in pregnancy-related development and/ or fetal development (40). In another study, an acid-stable mammary-stimulating factor different from EGF was partially purified from porcine serum (41), but the physiological significance of the activity remains to be defined.  (46) to isolate a UDGF from a larger scale source than rat tissue. By extraction and chromatography methods conducted at approximately neutral pH using acetone powders of pregnant sheep uteri, UDGF could be purified partially ( i e . 1,700-fold) by a 6-step procedure to yield a high specific activity (25,000-Da) product that was still only an estimated 3% pure'; the overall yield from this partial purification was less than 1% which was not sufficient to allow continued purification to homogeneity or to provide sufficient material for extensive biochemical studies. At this point, either a source of growth factor richer than pregnant sheep uteri was necessary for the large scale purification, or a substantially modified procedure was required to increase yields. Using new acetic acid-extraction and heattreatment methods followed by three chromatography steps, milligram amounts of activity have now been isolated from pregnant sheep uteri. An apparent M, = 4,200 and 6,200 was estimated by SDS-PAGE and HPLC methods, respectively; by several criteria the UDGF isolated appears to be a highly homogeneous preparation that demonstrates significant growth responses in the 10"' to lo-' M concentration range.
A description of the methods ofpurification and the properties of sheep UDGF are presented next.

EXPERIMENTAL PROCEDURES3
Cell Cultures-The MTW9/PL rat mammary tumor cell line used in this study was established in culture from the estrogen-and The per cent purity of this former Preparation was calculated from the specific activity of the highest purity preparation (CM-Sephadex, Step 5) achieved in this report as summarized in Table I.
Portions of this paper (including Table M1 and Figs. M1-11) are presented in miniprint at the end of this paper. Miniprint is easily prolactin-responsive MTW9A tumor (47) by methods described previously (48). MTWS/PL cell growth in culture is thyroid hormone (49) and pituitary factor (44) responsive, but not estrogen responsive (50). The MTW9/PL cells were assayed periodically for the ability to form estrogen-responsive tumors in vi00 as described previously (48); after more than 6 years in culture, the cells still demonstrate the same hormone-responsive tumor formation properties described in the original report (48).
Stock cultures of MTW9/PL cells were maintained in the formulation of DME medium prepared at the higher (4.5 g/liter) concentration of glucose and supplemented with 2 mM glutamine, 240 pg/ ml of potassium penicillin G, 540 pglml of streptomycin sulfate, 50 pg/ml of sodium ampicillin, 15 mM HEPES (pH 7.2), and 10% (v/v) FCS. The cells were grown at 37 "C in a humid atmosphere of 5% C02 and 95% air and passaged every 3-4 days at a density of 6.0 X 10' cells/78-cm2 plastic tissue culture dish (Corning). The DME was purchased from Grand Island Biological and serum from K. C. Biologicals. All sera were used without heat inactivation.
Growth Factor Assay Methods-The growth factor specific activity in either 0.1 M acetic acid extracts of fresh uteri or during purification from lyophilized powders of pregnant sheep uteri was determined by a bioassay procedure which measures the growth response of MTW9/ PL cells in serum-free DME. During purification, Method 1 was used exclusively and followed the incorporation of tritium-labeled thymidine into DNA in response to varying concentrations of protein.
Stock cultures used to initiate the assays were always used on day 3 after passage (counting day 0 as time of subculture); this procedure was essential since log phase cells were necessary for the assays of UDGF activity. Stock cultures (from 78-cm2 dishes) were trypsintreated briefly (2-4 min at room temperature) and the action of trypsin terminated by the addition of DME containing 10% (v/v) FCS. The detached cells were sedimented at 200 X g for 5 min and resuspended in fresh DME containing 10% FCS. This stock was then diluted to 2.0 X 10' cells/ml in DME containing 10% (v/v) FCS and, with gentle stirring to keep the cells suspended evenly, 1.0-ml aliquota transferred to each of 24 wells in plastic cluster well tissue culture plates (Costar 3524, Cambridge, MA). The cells in the wells were then incubated for 24 h at 37 "C in the C02 incubator to allow complete attachment and healing from the effects of trypsin; this incubation allowed the cells to begin undergoing metabolic processes. It must be noted that the period of incubation in FCS-containing medium is essential to the assay; if the cells are plated directly into serum-free medium in the wells, no UDGF growth factor activity is identified even with the most active preparations.
After 24 h in serum-containing medium, the DME was removed from the cultures by gentle aspiration and replaced immediately (one well at a time) with 0.9 ml of serum-free DME. The plates were again incubated in a C02 atmosphere at 37 "C for exactly 24 h, after which 0.1 ml of UDGF preparation was added. Growth factor dilutions were always made in serum-free DME.
Incorporation of tritium-labeled thymidine was conducted for a 2-h pulse-labeling period exactly at 22-24 h after addition of UDGF. The labeled precursor ([methyl-3H] thymidine, specific activity, 70 Ci/mmol, purchased from Schwarz/ Mann) was added in 0.050-ml portions containing 1.0 pci. Labeling was conducted at 37 "C in the humid COZ incubator. Thymidine incorporation was terminated by addition of 1.0 ml of Carnoy's fixative composed of three parts methanol and one part glacial acetic acid. Addition of this solution fixed the cells to the bottom of the culture wells and, by allowing the plates to stand for 3-5 h, the soluble (non-DNA) pools of thymidine precursors were washed out. This fixing solution was then removed by aspiration, and each well washed twice with 2.0-ml portions of 80% methanol (20% water) to remove all residual labeled thymidine. The methanol rinse was carried out with 24 wells per wash procedure. Then 0.3 ml of trypsin (1-300 hog pancreas preparation purchased from Nutritional Biochemicals) prepared at a concentration of 20 mg/ml in 50 m M HEPES (pH 7.3) was added to each well. After at least 2 h at room temperature, 0.7 ml of 1% SDS prepared in distilled water was added to each well, and the total 1.0 ml from each well transferred to wintillation vials. The radioactivity was determined in a Packard Tri-Carb Liquid Scintillation Spectrometer after addition of 10  tional Diagnostics). The intra-and interassay variations in samples were usually 15% or less. Averages of either duplicate or triplicate wells were used in all cases. We have reviewed recently (45) several other factors that must be controlled during use of this bioassay procedure.
Using Method 1, the specific activity of the various UDGF preparations was estimated as the concentration of protein required to half replace the growth response of MTW9/PL cells to 10% (v/v) FCS. The incorporation of label in response to no additions (zero addition control) was designated C , while the incorporation of label in response to 10% FCS was designated Cl0. The specific activity (G,) was the arithmetic mean of these numbers; decreasing values represent a higher state of purity. A unit of UDGF was defined as that amount of protein which gave half replacement of the 10% FCS growth response. Throughout this study, protein concentrations were measured by the method of Bradford (51), using bovine serum albumin as standard and reagents purchased from Bio-Rad Laboratories, Richmond, CA.
Method 2 of the growth factor assay was estimating growth directly by an increase in cell number. F6r this assay the same type of stock cultures and medium were used as described for Method 1, except that the cells (3.0 X 104/35-mm diameter culture dish) were plated in a total volume of 3.0 ml of DME containing 10% FCS for 24 h, after which this medium was removed by gentle aspiration and replaced with 2.5 ml of serum-free DME and 0.5 ml of UDGF or other preparation diluted with serum-free DME. Either on designated days, or on day 6 only, triplicate plates were harvested by trypsin treatment and total cell numbers determined with a Coulter Counter (Model ZB1).
Polyacrylamide Gel Electrophoresis-The degree of homogeneity of the purified UDGF was estimated by either 8 M urea, 0.1% SDS-PAGE, by the standard pH 8.5 non-SDS-PAGE, or by pH 4.5 non-SDS-PAGE, followed by localization of proteins by Coomassie brilliant blue R-250 staining.
The urea-SDS-PAGE system used was that of Swank and Munkres (52). Electrophoresis was performed with 6 X 90-mm gels formed in glass tubes. The final acrylamide and bisacrylamide concentration of the gels was 12.5 and 1.25%, respectively, and each gel contained final concentrations of 8 M urea, 0.1% SDS, and 0.1 M phosphoric acid adjusted to pH 6.8 with solid Tris base. Running buffer contained 0.1% SDS in 0.1 M phosphate/Tris, pH 6.8. The samples were run at 3 mA/gel for 5-7 h at room temperature. Unless otherwise stated, reducing agents such as 2-ME or DTT were omitted from the running buffer and from the sample boiling solution. All samples were pretreated by heating in 1% SDS for 2 min at 100 'C. The proteins in the gels were fixed and stained in one step with a solution of 0.2% Coomassie blue dye dissolved in methanol/acetic acid/water (5:1:5 v/ v/v). Staining was conducted for 2 h at room temperature and destaining was accomplished over several days with 7% (v/v) acetic acid and 12% (v/v) methanol in water. When required, stained protein band intensity was estimated by densitometer scan at 595 nm using a Gilford recording spectrophotometer. The molecular weight markers used to construct calibration curves were either intact or cyanogen bromide-generated fragments of horse heart myoglobin of known M,, which were myoglobin 111, 2,512; myoglobin 11, 6,214; myoglobin I, 8,159; myoglobin I + 11, 14,404; and myoglobin, 16,949. These standards were purchased from BDH.
Localization of the UDGF activity in the urea-SDS gels was accomplished by eluting the activity from unstained gels. The frozen gel columns were sliced into 2-mm discs and each incubated separately in 1.0 ml of 10 mM sodium phosphate buffer (pH 6.0) for 24 h at 4 "C. The activity eluted was estimated by diluting 0.020 ml of each disc eluate into 5.0 ml of serum-free DME, and using 0.10 ml of this solution in duplicate assays by Method 1 described above. This 2500fold dilution eliminated the toxic effects of residual SDS and urea on the MTW9/PL cells.
Analytical electrophoresis under non-SDS conditions at pH 8.5 was conducted as described by Davis (53) using gels composed of 15, 12,10, and 7.5% acrylamide. Staining and destaining were done under conditions identical to those used with the urea-SDS gels. The stained protein band localization was performed by densitometer scan at 595 nm; UDGF activity localization from unstained frozen gels was performed by slicing the gel column into 2-mm thick discs and eluting each for 5 h at 45 "C into 0.5 ml of 0.1 M acetic acid. The activity eluted from each disc was estimated as described above for the urea-SDS gels. Analytical non-SDS-PAGE also was conducted a t pH 4.5 by the method of Reisfeld et al. (54). The conditions fer staining these gels, as well as localization of UDGF activity, were identical to those just described for the non-SDS electrophoresis at pH 8.5.
Other Analytical Methods-The isoelectric point of the UDGF, as well as the degree of homogeneity, waa estimated by isoelectric focusing. The procedure was conducted in 6 X 90-mm glass tubes with 4.75% acrylamide and 0.25% bisacrylamide gels containing 0.03% (w/v) ammonium persulfate, 0.05% (v/v) TEMED, 8 M urea, and 2.0% ampholytes. The ampholytes used were purchased from Pharmacia as Pharmalyte pH 3-10. The anode solution was 0.01 M phosphoric acid and the cathode solution was 1.5% ethylenediamine. Samples were dissolved in 10 M urea solution in either anode or cathode buffer and applied from either electrode. The pH at application time did not influence the eventual position of focusing. The focusing was continued at 4 "C for 12 h at 200 V; gels were incubated with 12.5% trichloroacetic acid at 4 "C for 2 days with 4 changes in trichloroacetic acid solution to fix the proteins into the gel matrix and to remove the ampholytes which stain intensely with Coomassie blue. The protein bands were then localized with Coomassie blue stain as described above with the urea-SDS gels. Densitometer scans of the stained gels were performed at 595 nm as above. The final pH gradient of the experiments was estimated by slicing an unstained frozen gel into 1-mm discs, eluting each for 12 h into 0.5 ml of degassed distilled water, and determining the pH of the resulting solution. The isoelectric point of the UDGF activity was estimated from unstained frozen gels by slicing into 1-mm discs, eluting each at 4 "C for 24 h in 0.5 ml of water, diluting 0.01 ml of eluate into 5.0 ml of serum-free DME, and using 0.10 ml to assay for mitogenic activity by Method 1 described above. An estimated 20% of the total activity applied could be recovered by these methods.
The HPLC methods employed were performed with a Waters apparatus equipped with a dual wavelength detector which allowed monitoring at 280 and 210 nm. Analysis by molecular weight sieve was accomplished with two tandem arranged Bio-Rad TSK-125 columns (7.5 X 300 mm) equilibrated with 0.1% trifluoroacetic acid and eluted with the same acid. Fractions were diluted to sufficient degree to eliminate the cytotoxic effects of the fluoroacetic acid on MTW9/ PL cells used in the UDGF bioassay. Reverse phase HPLC was conducted with the same instrument using a Brownlee Laboratories octylsilyl (C8) 10-pm pore column (4.6 X 150 mm). This column was equilibrated and the UDGF eluted as described in the text.
Sources of Sheep Tissues for Purification-Fresh sheep uteri were obtained from ewes sacrificed on the day of the experiment. Timed pregnant ewes for these procedures were obtained from The University of Texas Science Park at Bastrop, TX. Timing of pregnancy was estimated by x-ray of the fetus size in utero. Before use of any uteri, all fetal and placental materials were removed and the uterine tissue washed extensively, first with running tap water and then with saline to remove residual blood. Fresh uteri were cut into 4-cm3 pieces, passed through a meat grinder, and homogenized with a Tissumizer (Tekmar) in 0.10 M acetic acid for 15 min at 4 "C at low to medium speed in the ratio 1 g of tissue to 3.0 ml of acetic acid. The extraction was then continued overnight without homogenization, but with stirring at 4 "C. The active supernatant was collected at 25,000 x g for 60 min and stored frozen at -20 "C until used. The lyophilized sheep tissue was obtained as a custom-made product of the Waitaki Refrigerating Limited, Christchurch, New Zealand. These preparations began with removing uteri from early pregnant sheep, washing out the fetal contents and placenta, and placing uteri on ice during the day-long collection process. These tissues were then rinsed again thoroughly to remove residual blood, passed through a meat grinder, frozen, and finally at a later date lyophilized. The lyophilized materials were then converted to a powder by treatment for 1 to 2 h in a stainless steel ball mill. The powders represent 20-25% of the original wet weight of the uteri. The UDGF activity of these powders was stable at -20 "C for 1 year.

Identification of Pregnant Sheep Uteri as a Large Scale
Source of UDGF Before initiating a large scale growth factor purification from species other than rat, the new source was evaluated from hormone responsiveness of the UDGF activity. As presented in Fig. 1, 0.1 M acetic acid extracts of fresh uteri from pregnant sheep showed a significantly elevated specific activity compared to those extracted from fresh uteri of ovariectomized ewes, or from fresh uteri from intact ewes that were not in estrus. A series of other assays of individual fresh uteri from 10 animals at random stages of pregnancy showed G,, values ranging from 2 to 10 pg/ml (data not shown), while those from the other two nonpregnant groups continued to  """""-"""""""""""" yield values of G,, = 100 pg/ml or more.
To better define whether stage of pregnancy has a marked effect upon UDGF activity, uteri pooled from 10 or more animals at either early pregnancy (i.e. <49 days), mid-pregnancy (i.e. 50 to 110 days) or late pregnancy ( i e . >110 days) were lyophilized and extracted with 0.1 M acetic acid. These extracts showed G, values ranging from 0.75 pg/ml for early pregnant uteri to 1.3 pg/ml for mid-pregnant uteri (Fig. 2). While these differences may be significant, it was clear that the UDGF activity did not vary greatly during pregnancy and that the process of lyophilization caused the extracted activities to vary less than found with fresh tissue. Further, the extraction of lyophilized powders with 0.1 M acetic acid yielded a 2-4-fold higher specific activity than extraction from fresh uteri by the same methods, suggesting that purification started from lyophilized powders may be more effective than from fresh tissue. Finally, the total amount of extractable growth factor activity was assayed from these same three groups of lyophilized uteri (Fig. 3). Clearly, the amount of total protein extracted, and hence, the total activity extracted was 2-fold greater from the uteri pooled from early pregnant sheep than from uteri of animals at later stages of pregnancy. For this reason, purifications were planned using only lyophilized powders of early pregnant sheep uteri.

Purification of UDGF from Early Pregnant Sheep Uteri
Unless otherwise noted, the purification of UDGF was carried out at 4 "C. The basic five-step procedure is summarized as follows.
Step I-A total of 500 g of lyophilized sheep uterine powder was extracted as described in the legend of Fig. 2. These same extracts used in Fig. 2 were then assayed for total protein by the method of Bradford (51), and for total units of UDGF by Method 1 under "Experimental Procedures." Closed bars represent total 100,000 X g supernatant protein isolated from the acetic acid extract of 10 g of lyophilized powder. Open bars represent total UDGF activity expressed as units isolated in the same extract supernatant.
by guest on July 9, 2020 http://www.jbc.org/ Downloaded from was stirred for 24 h with 5.0 liters of 0.1 M acetic acid. The suspension was then clarified by centrifugation at 25,000 X g for 60 min and the supernatant filtered through glass wool to remove lipids that floated to the surface during centrifugation. This crude supernatant was stored at -20 "C. After thawing, a large precipitate was removed by an additional centrifugation at 25,000 X g for 30 min, and the clear brown supernatant adjusted to pH 4.5 by dropwise addition of glacial acetic acid to a total volume of 4,210 ml.
Step 2"Aliquots (200 ml) of extract obtained in Step 1 were heated (in a boiling water bath) to 94 "C for 10 min, and then rapidly cooled to 0 "C in a propanol/dry ice bath. The large inactive precipitate was removed by centrifugation at 13,000 x g for 30 min, and the active supernatant frozen at -20 "C. After thawing, the supernatant was acidified to pH 3.5 with glacial acetic acid and the inactive insoluble material removed by centrifugation as described above.
Step 3-A total of 55 g of dry sulfopropyl-Sephadex C-25 (Pharmacia) was suspended in 2 liters of 0.1 M acetic acid and sufficient glacial acetic acid added to lower the pH to 3.1. This gel was then washed by decantation five times with 2 liters of 0.1 M acetic acid each time. The settled beads (260 ml) were then added to the acidified supernatant from Step 2, and the slurry was stirred slowly overnight. The sulfopropyl Sephadex beads were then allowed to settle for 2-5 h and the inactive supernatant poured off and discarded. The gel was then transferred to a 5.2-cm diameter glass column and washed successively, first with 1.2 liters of 0.1 M acetic acid, then with 500 ml of 0.001 M acetic acid. These washings contained little or no UDGF activity and were discarded. The growth factor activity was eluted from the sulfopropyl-Sephadex with 0.3 M ammonium acetate, pH 7.2, at a flow rate of 17 ml/h. The active fractions (21 ml/tube) were collected, pooled and lyophilized; the dried residue was then dissolved in 50 ml of 0.1 M acetic acid and any precipitate removed by centrifugation.
Step 4-The 50 ml of active UDGF from sulfopropyl-Sephadex was then applied to a Sephadex G-50 column (5.2 x 135 cm) equilibrated in 0.1 M acetic acid. The column was eluted with the same acetic acid. Fractions (21 ml/tube) were collected at a flow rate of 126 ml/h. As can be seen in Fig. 4, the UDGF eluted as a single active peak which was pooled and lyophilized. The dried residue of UDGF from the Sephadex G-50 column was then dissolved in 20 ml of 10 mM sodium phosphate, pH 6.0, and any precipitate removed by centrifugation.
Step 5-The concentrated fractions from Step 4 were desalted by gel filtration on a Sephadex G-25 fine grade column (2.5 X 120 cm), equilibrated, and eluted with 10 mM sodium phosphate, pH 6.0. The active desalted fractions were then applied to a CM-Sephadex C-25 column (2.5 X 31 cm) equilibrated in the same buffer. After washing with 160 ml of the same pH 6.0 buffer, the majority ( i e . 70%) of the growth factor activity was eluted with a linear salt concentration gradient (400-ml total volume) formed by using equal amounts of 10 mM sodium phosphate, pH 6.0, and the same buffer containing 0.3 M sodium chloride. The flow rate was 25 ml/h and 2.0-ml fractions were collected. The active fractions (see   Fig. 4. The CM-Sephadex was equilibrated in 10 mM sodium phosphate buffer, pH 6.0. The active UDGF fractions from Sephadex G-50 were desalted and equilibrated in the same buffer as described in the text. The CM-Sephadex column was eluted sequentially with 10 mM sodium phosphate, pH 6.0, and then with a linear sodium chloride gradient, also as described in the text. Fractions were monitored for protein content at 280 nm, and activity was monitored by the identical method used to measure activity for the Sephadex G-50 column shown in Fig. 4. Sodium chloride concentration was estimated by conductance. yield of the five-step purification procedure was 33%, and the final amount of UDGF isolated was 42 mg of protein which showed a specific activity of G, = 8 ng/ml. Based on the specific activity of the UDGF in the crude acetic acid extract, the overall purification was 162-fold, although if the purification is calculated from our previously reported (46) specific activity of UDGF in crude phosphate-buffered saline extracts of fresh sheep uteri (G5,, = 450 pg/ml), the overall purification achieved here is greater than 56,000-fold. Assays of aliquots  of each pooled fraction are shown in Fig. 6, which not only shows the calculations of Gw for each step of the purification, but also demonstrates that the dose-response curves become a parallel series as purification proceeds. These data suggest that the purification of a single growth factor was being accomplished.

Growth Factors for Estrogen-responsive Tumor Cells
It should be noted that the data presented in Fig. 6    It must be noted that when evaluated by the methods used in this report, the effective molar concentration of UDGF required to half replace the mitogenic response of MTW9/PL cells to 10 (v/v) FCS was in the lo-* M range. These assays were conducted in the absence of serum supplementation, and, in the absence of all other known hormones, attachment factors and growth factors already described as requirements for mammary cell growth in serum-free medium (34, 35). Our reason for using an assay for UDGF based on cell response to addition of a single factor was to ensure the isolation of a true mitogen, rather than an agent which facilitates or augments the action of the other known growth factors. This assay method is different in this regard from the procedures used for measuring the mitogenic responses of cells in cultures to PDGF (11-131, EGF (681, and TGF (15).
In one case (PDGF), the assays of the growth factor were conducted in the presence of 5% (v/v) platelet-poor plasma in the culture medium (12). This supplementation of plasma provided nutrients, attachment factors, and the other known growth factors such as EGF and SmC that potentiate or complete the action of this purified factor (69, 70), and allowed measurement of half-maximal responses (Gm) to PDGF in the IO-" M range. Without plasma supplementation, considerably higher concentrations of PDGF were required to achieve Gw responses. In the case of EGF, low level serum supplementation (Le. 1% v/v calf serum) was required for maximal target cell responses (68). Under these conditions, EGF half-maximal biological activity fell into the 10"' M range. However, as was evident from the data presented in Fig. 6 (Miniprint), when target cell responses to purified EGF were measured in the absence of the serum supplementation, G, was not achieved at concentrations of >I X IO-' M.

Growth Factors for Estrogen-responsive Tumor CelLs
However, it is apparent from these data that the final preparations were able to stimulate MTW9/PL rat mammary tumor cell growth well beyond the level of 10% FCS.
The final product from Step 5 was assayed for the effect of trypsin on the activity. As shown in Table 11, trypsin treatment caused a 96% loss in activity of UDGF within 3 h a t 37 "C. Incubation with the equivalent amount of growth factor with boiled trypsin had no effect on activity.

Evaluation of Homogeneity of the UDGF Preparation after Various Isolation Steps
The purification of UDGF as summarized in Table I (Fig. 8). The range of possible M , originates from the width of the band, which may be related to the properties of the urea-SDS PAGE system used to analyze this sample, since the myoglobin fragments of exact known molecular weights also exhibit relatively broad bands under these same conditions. Also, it is possible that since these proteins are of low M , they may Similar circumstances arose when comparisons were made between UDGF assay methods and those used to measure TGF activities. The tumor and normal tissue-derived TGF activities were effective in the 10"' M range (65,66) only when assayed as colony-formation stimulating factors, and in the presence of serum and a supporting concentration of a second growth factor (EGF). Also, while TGF has been designated a growth factor, its activity does not promote incorporation of labeled precursors into target cell DNA in serum-free culture. Hence, on the basis of the assay differences, a direct comparison of the biological potencies of TGF and UDGF is not possible. From the information presented, as well as other studies not reviewed here (ie. cell growth responses to fibroblast growth factor in 0.4% serum-containing medium), we conclude that the assay methods employed clearly dictate the values of G , calculated for each of the four growth factors discussed and that biological potency is a measure of several supporting interactions which lead to cell division.
To further define the Gso of UDGF under conditions similar to the other factors, serum supplementation was attempted. This method proved ineffective (see Fig. 2, Miniprint) since UDGF added to serum did not stimulate over serum-only controls. These data provided another feature which distinguishes purified UDGF from PDGF, EGF, and TFG. In additional studies aimed at potentiating the effect of UDGF, and hence aimed at lowering the concentration of uterine factor necessary to reach G, with MTWS/PL cells in serum-free defined medium, we have demonstrated that addition of 10 ng/ml of EGF to the cultures causes an 8-10-fold reduction of UDGF Gso to 1-2 X 10"' M (T. Ikeda and D. A. Sirbasku, unpublished results). Therefore, as is the case for the other cell growth responses cited above, addition of multiple factors reduces the effective concentration of each growth factor required for a standard biological response.  Further experiments were conducted to confirm that the Coomassie blue-stained band coincided with the mobility of the UDGF activity. Parallel urea-SDS gels were run as described under "Experimental Procedures"; one set of gels was stained with Coomassie blue to localize the proteins, while the other set was immediately frozen at -20 "C, sliced into discs and the eluant of each disc assayed for UDGF activity.
The results of this experiment are shown in Fig. 9. The migration position of the Coomassie blue-stained component exactly corresponded to the migration position of the recoverable UDGF activity. It must be noted that while the recovery of UDGF activity from urea-SDS gels was only partial, the heating at 100 "C in 8 M urea and 1% SDS caused.an initial 60% inactivation before electrophoresis. Hence, the fraction of activity found after the elution from acrylamide discs represents a 20-25% recovery of the residual activity applied. Attempts were made to increase the yields of the activity eluted' by using several different methods, including removal of residual SDS bound to eluted UDGF by a -18 "C ethanol/ albumin coprecipitation. None of the other methods employed proved more effective than that used to obtain the data in Fig. 9. If boiling of UDGF in 1% SDS was omitted before electrophoresis, two minor Coomassie-stained bands (each approximately 10% of the total stain intensity) appear at M , = 12,000 and 16,000, suggesting aggregation of the UDGF prior to or during electrophoresis at pH 6.8 in phosphate buffers. This aggregation of UDGF at near neutral pH in sodium phosphate buffers is thought to be responsible for the fraction of 30% of the UDGF activity that elutes through the CM-Sephadex column during Step 5 of the purification (Fig.  5). Analysis of the CM-Sephadex flow-through fractions by urea-SDS gel electrophoresis shows that, after boiling in SDS, the protein present in this fraction migrates on urea-SDS gels as a single band equivalent to that of the UDGF eluted by the sodium chloride gradient. These data suggest that the CM-Sephadex flow-through may be an aggregated form of UDGF.
The UDGF preparations from Step 5 were submitted to non-SDS-PAGE at both basic and acidic pH. Analysis carried out at pH 8.5 with gels containing 7.5, 10, 12, and 15% acrylamide is shown in Fig. 10. Protein bands were again localized by Coomassie blue staining; it can be seen that when equal amounts ( i e . 20 pg) of UDGF were applied to the gels Step 5 UDGF preparation and, after running, was stained with Coomassie blue; the migration position of the single band found is shown above. A duplicate gel received 340 pg of the same UDGF preparation and, after running, was frozen and sliced into 2-mm discs. Each disc was incubated as described in the text to release UDGF, and the eluant assayed by Method 1. The activity found for each disc is presented in the bars. The migration position of bromphenol blue (BPB) is shown. composed of different acrylamide concentrations, a single band is identified in all four concentrations, although the stain intensity was clearly the greatest in 15% gels; this observation of lower staining intensity in lower concentrations of acrylamide gels suggests that the protein in lower concentration gels may diffuse out during the staining and/ or destaining processes. In additional experiments, submitting the UDGF Step 5 samples to longer times of electrophoresis in 15% polyacrylamide gels caused the band to move further toward the cathode as expected (data not shown), but this did not further resolve Coomassie blue staining bands.
Another series of experiments were conducted to correlate the position of the Coomassie-stained band with the elution of UDGF from sliced discs of an unstained 15% acrylamide gel run at pH 8.5. The results of one of these experiments are shown in Fig. 11. The migration of the single stained band corresponded exactly to the elution position of the UDGF activity. In these experiments, the total recovery of activity was 20-30% of the amount applied.
In a final series of acrylamide gel experiments, UDGF Step 5 preparation was analyzed by non-SDS-PAGE at pH 4.5. The conditions of the experiments were similar to those of Fig. 11 with the exception of using a pH 4.5 electrophoresis buffer (54) and the exception of using 12.5% acrylamide gels at pH 4.5 and 15% acrylamide gels at pH 8.5. The UDGF migration into the 12.5% pH 4.5 gels was less than had been observed at pH 8.5. Nevertheless, in these experiments, the Step 5, preparation activity from non-SDS (pH 8.6) 15% acrylamide gels. Electrophoresis was conducted as described in Fig. 10 and "Experimental Procedures." One gel was loaded with 20 pg of UDGF and, after running, was stained with Coomassie blue to establish the migration position of UDGF. A duplicate gel received 75 pg of UDGF and, after running, was frozen and sliced into 2-mm thick discs, and the activity in each disc eluted and assayed as described in the text. The anode and cathode positions are shown as is the migration position of bromphenol blue ( E P E ) .
only protein band identified by Coomassie blue staining exactly corresponded to the UDGF activity eluted from companion unstained 12.5% acrylamide gels prepared and run at pH 4.5 (data not shown).

Evaluation of the State of Homogeneity of Purified UDGF by
Hydrophobic Chromatography and HPLC The homogeneity of the UDGF Step 5 preparation was assessed by hydrophobic chromatography and by HPLC methods.
Hydrophobic chromatography on octyl-Sepharose CL-4B (Pharmacia) is shown in Fig. 12. Before chromatography, the UDGF was equilibrated with the same 6.0 M ammonium acetate in 10 mM sodium phosphate buffer, pH 6.8, used to wash and equilibrate the column. After initiating chromatography, a total of 90% of the applied protein could be accounted for in the 6.0 M ammonium acetate wash with the remaining 10% found in the reverse linear gradient fractions running 6.0-0 M ammonium acetate in 10 mM phosphate buffer (pH 6.8) (fractions 16 through 32), and in the final 50% (v/v) ethylene glycol wash. The UDGF activity (85%) eluted in the major protein fraction of the ammonium acetate wash, and no other activity could be identified in any other region of the elution.
In parallel experiments, to confirm that the UDGF was not highly hydrophobic and to further confirm homogeneity, UDGF analysis was conducted using HPLC reverse phase chromatography on octylsilyl (C8) columns washed first with 1-propanol (immobile phase), and later equilibrated with a mobile phase of 0.1% trifluoroacetic acid (pH 2.2) just before applying the UDGF. The UDGF sample was equilibrated in the same concentration of trifluoroacetic acid by passage through Sephadex G-25 fine grade, equilibrated, and eluted with 0.1% trifluoroacetic acid. From the HPLC C8 column elution times found, UDGF eluted at the same time as a marker ( i e . molecular iodine) that did not interact with this hydrophobic matrix. Approximately 90% of the UDGF activity and 85% of the total protein Step 5, preparation on octyl-Sepharose CL-4B. A 2 X 8-cm column was equilibrated with 6.0 M ammonium acetate in 10 mM sodium phosphate buffer, pH 6.8. The UDGF sample was equilibrated in the same buffer and the total sample of 6.5 mg in 12 ml was applied to the column at a flow rate of 30 ml/h. Fractions (2.0 ml) were collected. The column was first washed with the same ammonium acetate containing buffer (20 ml), and then eluted with a reverse linear gradient of 30 ml, each, of the ammonium acetate/sodium phosphate buffer, and sodium phosphate buffer without the ammonium acetate. Finally, the column was washed with 50% ethylene glycol (EG) to remove residual protein. Fractions were monitored at 280 nm and activity was measured by dilution of 10 pl into 1.0 ml of 0.1 M acetic acid followed by 20-pl dilution of the acetic acid solution into 1.0 ml of serum-free DME. A volume of 100 pl of this DME was used in assay Method 1. v) 1-propanol in 0.1% trifluoroacetic acid (data not shown).
In another HPLC approach to evaluating homogeneity, a sample of UDGF Step 5 was equilibrated with 0.1% trifluoroacetic acid and applied to an in-line series of two HPLC molecular sieve TSK-125 columns equilibrated in the same acid. The activity and protein elution profile are shown in Fig. 13. In each fraction with significant absorption at 210 nm parallel UDGF activity was found. When these tandem columns were calibrated with components of known M,, an estimated UDGF M, = 6200 f 1000 was obtained (data not shown).

Determination of Isoelectric Point of UDGF
The isoelectric point of UDGF from Step 5 was determined in glass tube gels as described under "Experimental Procedures." A typical Coomassie blue-stained gel is shown in Fig.  14, in which the sample was applied at the pH 3.0 electrode. The stained gel shows a broad major peak focusing near the midpoint of the gel, and another minor peak focusing at an apparently slightly more basic position. Densitometry scanning (Fig. 15) shows clearly the presence of the minor component, although from these data it is not possible to calculate the exact percentage of the total stained material this band represents. Nevertheless, the data show that the sample of UDGF Step 5 is highly homogeneous. In a parallel experiment, an unfixed and unstained gel was sliced into 1-mm thick discs, and the UDGF activity eluted and assayed. As can be seen in Fig. 16, the UDGF activity elutes over the whole region of both the major and minor band. The estimated PI of the major band is 7.3.

DISCUSSION
Both the results presented in the main text and in the Miniprint will be discussed here.
We have reported previously (42) an estrogen-inducible growth factor activity in extracts of rat uteri. When prepared at neutral pH, this activity was a relatively specific mitogen for tissue culture cell lines which were established from estrogen-responsive or -dependent mammary, pituitary, and kidney tumors from rats and hamsters; when correlations were made between this mitogenic activity and hormone-responsive tumor growth in uiuo, it was confirmed that the level of total extractable mitogenic activity in rat uteri correlated well with the rate of mammary and pituitary tumor growth in rats (42). In attempting further characterization of this mitogen as a product of the uterus, a similar growth factor was identified in estrogen-induced accumulations of rat uterine luminal fluid (55,56); this fluid is well known to contain uterine origin components secreted under estrogen stimulation (57, 58). Attempts to further characterize the rat uterine tissue activity showed that when extracted with phosphate buffers at pH 7.2, the mammary cell growth factor obtained was a heat-labile, trypsin-labile substance that did not show properties related to either steroid hormones or other lipid-like components expected in uterine extracts (43). The apparent M, = 70,000 of this activity also suggested protein properties

(43).
However, from the results available with other growth factors, it was apparent that tissue-derived or plasma-borne activities initially may be identified as high M, factors (i.e. >70,000) in neutral pH extracts, but later shown by treatment with acid, or other dissociating agents, to be relatively low M, (6,000 to 10, OOO) species. Examples of growth factors known to demonstrate these different properties dependent upon pH FIG. 14. Coomassie blue-stained isoelectric focusing 4.75% acrylamide gel. The UDGF, Step 5, preparation (150 pg) was submitted to analysis by isoelectric focusing by the methods described under "Experimental Procedures" and in the text. After focusing was complete, the gel was fixed and the ampholytes removed with 12.5% trichloroacetic acid, the gel stained with Coomassie Blue, and destained by the same method used with 8 M urea, 0.1% SDS gels. The major band is designated by the larger arrow, while the position of migration of the minor band is shown by the smaller arrow.  Step 5, preparation and focusing carried out as described in Fig. 14. At the end of 12 h the gels were removed from the tubes, frozen, and sliced into 1-mm thick discs. The pH of the distilled, degassed water eluate from each disc was the measure of the focusing gradient pH, and aliquots of this eluant diluted into serum-free DME for UDGF activity assay by Method 1. acid, the rat activity was heat-stable (95 "C for 5 min)? We therefore decided to approach the purification of UDGF by methods based on acetic acid extraction and treatment at high temperatures.
The isolation of a high specific activity mammary-uterinepituitary tumor cell growth factor from lyophilized sheep uteri has been accomplished using a high yield five-step procedure, the results of which are summarized in Table I These data support the conclusion that a potent effector of mammary tumor cell growth in culture has been isolated. Furthermore, not only can a high specific activity UDGF be prepared by the methods described, but from each 500 g of lyophilized early pregnant sheep uteri, 40-50 mg can be obtained. This amount calculates to a tissue level of 60-75 mg of UDGF/kg of fresh (wet) uteri, which is an amount that exceeds the known abundance of other acid-stable growth factors such as PDGF (ll), SmC (8), TGF (15,16), and SGF (14). Of the acid-stable group of polypeptide mitogens isolated, only EGF is present in tissues ( i e . mouse submaxillary gland) in such abundance (1).
This preparation of UDGF has been evaluated for degree of homogeneity by several independent methods. During the course of these studies, the properties of the growth factor have been partially elucidated. These studies are summarized as follows: (i) At each of the last two steps of the purification, which are Sephadex G-50 chromatography (Fig. 4) and CM-Sephadex chromatography (Fig. 5), the growth factor activity coincides with a single, well defined protein (ie. 280 nm absorbing) peak. only one Coomassie blue-stained band was found even when as much as 150 pg of UDGF preparation was applied per gek further, elution of individual discs of parallel unstained gels showed UDGF recovery only at the migration position of the single Coomassie-stained band. (iii) Using non-SDS, pH 8.5, PAGE to separate by both M, and net charge, again only a single Coomassie blue-stained band was identified, even at high concentrations (i.e. 150 pg of UDGF per gel), and parallel elutions of sliced discs of unstained gels showed UDGF activity corresponding only to the position of the migration of the stained band. (iv) Repeating the non-SDS-PAGE at pH 4.5 yielded the same results as those found at pH 8.5; the migration position of the UDGF activity and stained protein band were identical. However, the polarity of the electrophoresis at pH 4.5 was opposite that of the pH 8.5, indicating that the PI of UDGF was between these pH values. Also, despite the differences in the polarity of electrophoresis and in the distances migrated by UDGF at pH 4.5 and pH 8.5, no additional Coomassie-stained bands could be identified. (v) Molecular sieve HPLC (Fig. 13) of the UDGF preparation revealed only a single 210 nm absorbing peak containing >90% of the total protein and >85% of the total UDGF activity applied. However, the M , = 6200 ? lo00 estimation by HPLC is significantly higher than that estimated by urea-SDS-PAGE. The reasons for these differences are not yet apparent, but the fact that UDGF heated in 8 M urea plus 1% SDS at 100 "C retains 40% of the initial activity suggests that the growth factor does not assume the expected rigid rod structure in SDS solutions, but instead, may retain some secondary structure. This retention of activity, and hence possible secondary structure, may cause the apparent M, to be estimated lower by SDS than by HPLC methods. This latter method separates on the basis of the Stokes radius of the undenatured UDGF and may more accurately reflect the M, of UDGF. In other work to be reported', no evidence for amino sugars has been found by chemical analysis, suggesting that UDGF is not a glycopeptide; if UDGF were a glycopeptide, this would cause discrepancies in M, estimation by the different methods used in this report. (vi) Using different properties of the UDGF than examined before, hydrophobic chromatography on octyl-Sepharose and reverse phase HPLC on octylsilyl (C8) columns showed the activity to be a polar species that eluted in the void volumes of both chromatographic systems. In both cases, the protein elution profiles and the UDGF activities corresponded in the same fractions. (vii) By isoelectric focusing between pH 3.0 and 10.0 in acrylamide gels (Fig. 14), the UDGF preparation showed a major band focused at average PI = 7.3 while the minor band focused at what initially appeared to be a somewhat more basic PI (Fig. 15). However, when determinations of the pH of the focusing gradient were done, there appears a pH discontinuity at the basic side of the major PI 7.3 band (Fig. 15), and this discontinuity corresponds to the final focusing position of the minor band. Hence, the PI of the minor band is currently estimated as 7.1.
In view of the pH gradient discontinuity in the isoelectric focusing experiment, it may be possible that the minor band is artifactually generated by this gradient abnormality. This possibility is supported by the fact that UDGF growth factor activity is found throughout the region of the focusing gel corresponding to both major and minor peaks (Fig. 16). Another possibility is that there are two forms of UDGF which are very closely related, but which cannot be distinguished by any method applied other than isoelectric focusing. If this is the case, it may be that one form of UDGF is a degradation product of another and that this occurs during the focusing experiment since the relative amounts of these two bands vary between different focusing experiments. It is equally Growth Factors for Estrogen-responsive Tumor Cells possible that the two forms of UDGF may be formed by partial proteolytic degradation which occurs during the time required for conversion of the 2-10-kg amounts of fresh uteri to lyophilized powders, or partial proteolytic degradation occurring during the isolation process. At present, these possibilities cannot be distinguished, but from the sum of the data available by all methods applied, we conclude that a highly ( i e . >go%) homogeneous preparation of UDGF has been achieved.
In an effort to establish whether the UDGF purified from sheep uteri possesses the same cell type specificity as the activity identified in crude extracts of rat uteri, experiments were carried out which are presented in the Miniprint. The G, of sheep UDGF was determined for key cell lines that form estrogen-responsive tumors in rodents. These cell lines are the GH3/C14 rat pituitary tumor cells that form estrogen-(61) and thyroid hormone-(62) responsive tumors in Wistar Furth rats, the H-301 hamster kidney tumor cells that form estrogen-dependent tumors in Syrian hamsters (63), ULMS smooth muscle tumor line developed from an estrogen-and androgen-induced leiomyosarcoma of hamster uterus (49), and the UCS endometrial tumor cells developed from an estrogen-induced haqster uterine carcinosarcoma (4%. When assayed by standard methods, UDGF is a potent mitogen with the five cell lines in decreasing order of responsiveness of MTW9/PL mammary cells > uterine UCS cells > kidney tumor cells > uterine ULMS smooth muscle cells > pituitary tumor cells. The Gso of this series of cell tests range from 8-10 ng/ml (MTWS/PL cells) to 1400 ng/ml (GH3/C14 pituitary cells). No mitogenic activity was found with normal diploid rat fibroblasts even at concentrations of UDGF > 50 pg/ml (Fig. 5, Miniprint).
One very interesting new fact brought to light by these studies is that UDGF is a relatively potent mitogen for uterine tumor cells of endometrial origin (G, = 36 ng/ml), and to a lesser degree, for uterine smooth muscle tumor cells (GsO = 176 ng/ml) (see Fig. 4, Miniprint). These observations raise the possibility that the activity isolated from uterus may be an endogenous autostimulatory growth factor (autocrine type activity) that is involved in estrogen-responsive or pregnancyrelated uterine growth or, alternatively, in uterine tumorigenesis. At present, the role in normal uterine growth is being further explored, and preliminary results confirm6 that short term monolayer cultures of normal rat uterine cells respond to UDGF at concentrations of <50 ng/ml in serum-free medium.
In addition to the cell type specificity assays, one other control study was essential. We have purified and characterized UDGF by following its activity as a function of the ability to stimulate incorporation of labeled precursor into DNA. This measure of cell growth is used routinely to purify and otherwise measure mitogenic activities, although it is recognized widely that this measure may not correspond to cell growth as measured by logarithmic increase in cell number. Hence, the experiments (see Fig. 3, Miniprint) confirm that UDGF is a true growth factor promoting a logarithmic rate of cell growth in culture. From the data obtained, it is apparent that UDGF promotes a continuous rate of growth of MTW9/ PL cells in culture and that the UDGF must be consumed or otherwise degraded in culture relatively rapidly since MTWS/ PL cell growth ceases at 48 h after a single addition of growth factor; if additional factor is not provided, the cells decrease in number sharply. These data confirm that not only is UDGF a promoting growth factor, but that it is required for continued survival of the MTWS/PL cells in culture. One additional important feature about the mitogenic activity of UDGF when measured by cell number increase is that a basal (nongrowth supporting) concentration of phosphate-buffered saline (pH 7.2) extract of lyophilized sheep uteri is required for the growth response. This may indicate that either additional factors not yet recognized are present in crude uterine extracts, or that this extract provides essential nutrients and other components required for cell growth.
After isolation and determination of the cell type specificity of UDGF, the next major consideration was whether UDGF isolated from pregnant sheep uteri is a new growth factor activity not previously characterized, or instead, an isolation of a known growth factor from another source. Studies aimed at resolving this problem are presented in the Miniprint. A number of known hormones and purified growth factors were assayed for the ability to replace or supplement the mitogenic effect of UDGF with MTWS/PL mammary cells. From the data presented in Table I (Miniprint), none of the known steroid hormones, polypeptide hormones, or purified growth factors were mitogenic for MTW9/PL cells under conditions where UDGF showed a potent (i.e. GsO = 10 ng/ml) cell growth effect. An example of how these data were obtained is shown in Fig. 11 (Miniprint).
Purified mouse EGF alone was assayed for mitogenic effect on MTWS/PL cells at concentrations of up to 20 pg/ml. No effect was found. Similar assays with other growth factors such as MSA, SmC, insulin, and bovine fibroblast growth factor showed no mitogenic effects. In another type of specificity assay, a cell line was selected that was EGF-responsive ( i e . the Balb/c 3T3 cells) and it was shown that UDGF was at least 150 times less potent a stimulator of cell growth than EGF with these cells, again confirming that the action of at least one of the well known growth factors (EGF) cannot be replaced by UDGF. The UDGF was also treated with the reducing agents DTT and 2-ME, and it was shown that the apparent M , of the activity did not change as estimated by the 8 M urea, 0.1% SDS, 12.5% PAGE (Fig. 9, Miniprint) and that the biological activity was retained completely (Fig. 10, Miniprint). These data further suggest that the activity isolated from sheep uteri is not PDGF-like (ll), EGF-like (64) or TGF-like (65, 66) since these factors lose activity after reduction.
From the data reported here, the UDGF isolated from pregnant sheep uteri appears to be a unique activity. Final confirmation of this conclusion is now being sought by determining the complete amino acid sequence using standard procedures (67). By determining the sequence, data should be obtained which will establish conclusively the molecular identity of UDGF.