This invention relates to mamπiamodulin, which is a novel factor not known to exist so far in nature, produced by hormone-independent human mammary tumor cells affecting the morphology, growth and hormone receptor expression of hormone-dependent tumor cells. More specifically the invention relates to mammamodulin in purified form, i.e. at least partially free from compounds with which it is associated in nature, up to high purity, e.g. enabling the determination of at least partial amino acid sequence(s). Means for the production and purification of mammamodulin are described, as are related diagnostic and therapeutic strategies made possible by the discovery of this factor.

Growth of human mammary tumors is a complex process governed by hormones and growth factors of endocrine, autocrine and paracrine origin. An important feature of mammary tumors is their dependency on hormones, particularly estrogen. This hormone dependency is often lost as the cancer develops, rendering endocrine treatment ineffective, and correlates closely with more aggressive behavior by the tumors. The prognosis for breast cancer victims once the tumor enters the hormone-independent stage is extremely poor.

It has now been discovered that a previously unknown factor, which has been named mammamodulin (MM), is expressed by hormone-independent mammary tumor cells. This factor stimulates less aggressive, hormone- dependent tumor cells to grow more rapidly while the expression of hormone receptors is down modulated. It is believed that, long term, MM plays a central role in the phenotypic switch from hormone-dependency to hormone- independency during tumor progression.

This factor has been, for the first time, identified, isolated and purified. Availability of substantially pure MM now enables various diagnostic and therapeutic strategies in the treatment of breast cancer. For example, MM levels may be assayed and used as a diagnostic indicator to determine the stage and seriousness of breast cancer. Substances which are observed to suppress MM's cell activation properties, as well as monoclonal antibodies to MM or other MM binders, may be used as therapeutics to block MM's stimulation of hormone-dependent cells, thereby stabilizing the cancer and preventing progression into the aggressive, hormone-independent stage of tumor growth. Finally, MM analogues may be used therapeutically to block MM receptors on hormone-dependent cells.

MM is a protein having an apparent molecular weight of approximately 52-55 kϋodaltons, as measured by SDS polyacrylamide gel electrophoresis of highly purified preparations on 10-15% gradient gels. MM comprises partial amino acid sequences of

Leu-Val-Leu-Arg-X-X-Glu-Thr (SEQ ID No: 1) Ser-Glu-Leu-Arg-Ile-Asn-Lys (SEQ ID No: 2), and X-Leu-X-Asn-Pro-X-X-Tyr-Leu (SEQ ID No: 3),

wherein "X" represents an unidentified amino acid residue.

It is believed that the protein as a whole may contain allelic variations comprising additions, deletions, insertions or substitutions of residues comprising up to ten percent, but preferably no more than five percent, of the complete amino acid sequence, so long as the biological activity is retained. The protein may also contain various post-translational modifications such as glycosylation and formation of cysteine bridges.

MM is further characterized in that it is heat and acid labile, and trypsin- and mercaptoethanol-sensitive, and easily forms heterogeneous aggregates with other proteins.

MM can be extracted and purified to apparent homogeneity from cell cultures of fast-growing hormone-independent tumor cells, as described in the examples herein. MM may also be obtained by using amino acid sequence information to screen a cDNA library from a producer cell line to obtain the cDNA of the MM gene. The gene may then be cloned and introduced into an appropriate vector for expression in a suitable prokaryotic or eukaryotic system.

Purified MM may be used as an antigen to produce polyclonal or monoclonal antibodies by conventional processes. Antibodies, in particular monoclonal antibodies are useful (i) as therapeutics to block MM and prevent it from stimulating hormone-dependent cells; and (ii) in assays (including bioassays, radioimmunoassays (RIAs), fluoroimmunoassays (FIAs),

Western blot assays, and ELISAs) to detect the presence of MM in biopsies of mammary tumors, thereby giving useful diagnostic information as to stage of development and prognosis of the cancer.

Purified MM may also be used in a binding assay to screen and identify compounds having affinity for MM and in competitive assays to screen and identify compounds having affinity for MM receptors on hormone-dependent cells. Such compounds are useful in inhibiting the biological activity of MM.

Additionally, compounds which inhibit the expression of MM in tumor cells, in particular mammary tumor cells, e.g. hormone-independent mammary tumor cells and/or have affinity for mammamodulin receptors on tumor cells, in particular on mammary tumor cells, e.g. hormone-independent mammary tumor cells are novel and useful. The only compound known to be a potent inhibitor of MM is heparin.

Flτmτnττte 1

Production of MM-containing condition medium from hormone inrieτ>Pτ.riRτ.t. cell cultures.

The human, breast cancer cell line MDA-MB-231 was grown in HL medium supplemented with 5 % fetal calf serum. The optimized HL medium was developed for the human leukemia cell line HL60 as described in European Patent AppHcation Publication No. 417563. The cells were grown in an adherent fashion with a doubling time of about 22 - 24 hours in T-flasks under 5 % Cθ2-balance air and 96 % water saturation. After achieving the confluent state, the growth medium was replaced by the serum-free HL medium. Mammamodulin (MM) is released into the culture medium during growth as well as in the non-growth confluent state.

A slightly modified 23 1 airlift fermentor (Chemap AG, Switzerland) was used for the scale-up production of MM . The outer volume of the draft tube was filled with 8.5 kg Raschig rings (8 8 5 mm, stone ware) and 10 Siraf 25 rings (25 x 25 x 2 mm, porous glass Raschig rings, Schott, Germany) giving about 20 πfi available surface area. The fermentor was inoculated with tryp- sinized MDA-MB-231 cells from 2 Cell Factories units (Nunc, 6000 cm ² per unit). Two medium exchanges were performed during the growth phase of 15 days. The airlift fermentor was aerated with 1 normal liter /min air and the solved oxygen content (30 %), pH (7.2) and temperature (37°C) were continuously monitored.

MM-containing spent medium was collected by changing 17 - 18 1 serum-free fortified HL medium (glucose from 5 g to 7.5 g, glutamine from 5 mM to 6.5 mM, Primatone RL from 0.25 % to 0.3 %) daily. The metabolic state of the cells was monitored during the whole fermentation process by measuring glucose and glutamine consumption and lactate and ammonia formation as described (Schumpp B. and Schlaeger E.-J., J. Cell. Sci. 97: 639- 647 [1990]).

The MM titre remained constant during the production campaign.

The MM containing cell-free fermentor volume was concentrated 20-fold (100 1 to 5.5 1) by using an Amicon SP20 ultrafiltration unit (Amicon, Switzer¬ land) equipped with two S10Y10 spiral-wound cartridges (1.8 m^ membrane

area, MW cutoff 10 kDa). The MM-containing concentrates were frozen at -80°C.

l mτmτlft 2

Pnrific-flt-ion of MM from the supernatant of MDA-MB-231 cells to partial purity bv 4 chromatogranhic. stens

All operations were done at room temperature (RT) if not otherwise stated. Fractions of purification runs were characterized by determination of the absorbence at 280 nm, by SDS-PAGE reduced according to Laemmli, U. Nature 227, 680-685 (1970) and by an assay for biological activity as described in Example 6.

5,5 1 of frozen concentrated supernatant (prepared as described under Example 1, corresponding to 100 1 of orignal supernatant) were thawn and centrifuged at 3000 g for 15 min (4°C).The absorbence of the supernatant at 280 nm at 1cm path length multiplied by the volume in ml (= OD280) was 35000. The total activity in the assay was 5,8 .IO ⁶ U (100%), the specific activity 165 U/OD280.

The supernatant was then diluted by 121 of

8 M urea

0,3 mM 3-{(3-Cholamidopropyl)- dimethylammonio}- <Buffer A>

1-propanesulfonate (CHAPS) 30 mM Tris/HCl . pH 7,5

and 1,7 1 of dest. water. The solution was then applied to a Q-Sepharose FF column (Pharmacia, Dύbendorf, Switzerland). The column (1,2 1, 4°-8°C) had been pre-equilibrated by

6 M urea

0,2 mM CHAPS

20 mM Tris HCl <Buffer B>

The column was washed by 1 column volume of buffer A and was then eluted (10 ml/min, RT) by a linear gradient from 100 % buffer B to 40 % buffer B and 60 % of

6 M urea 0,2 mM CHAPS

20 mM Tris/HCl <Buffer C>

500 mM NaCl pH 7,5

within 7 h and 36 min. Fractions between 30 and 50 % of buffer C (1550 ml) contained 2630 OD280. 7,5 X10 ⁶ U (130%) and the specific activity was 2852 U/OD280 _' The chromatogram is shown in Figure 1. The fractions were pooled and concentrated (4°-8°C) to 210 ml by a MϋUpore Minitan System equipped with 4 PTGC cassettes (Millipore, Volketswil, Switzerland). The concentrate was 1 : 15 diluted with

1 mM N-Dodecyl-N,N-d_methyl-3-ammonio- propansulfonate (Zwittergent 3-12) 50 μM Ethylene glycol-bis(beta- aminoethyl ether)

N,N,N',N' - tetraacetic acid (EGTA) <Buffer D>

150 mM NaCl 0,02 % NaN3 pH 7,6

to give a volume of 3150 ml. This volume was applied to a 80 ml - column of Heparin-Sepharose CL-6B (Pharmacia, 4°-8°C), which had been preequiUbrated by buffer D. The colunm was washed by 24 column volumes of buffer D and then eluted (3 ml/min, RT) by a linear gradient from 100 % buffer D to 100 % of

1 mM Zwittergent 3-12

50 μM EGTA

50 mM Tris/HCl <Buffer E>

1350 mM NaCl

0,02 % NaNβ pH 7,6

Fractions eluting between 39 and 64 % of buffer E (240 ml) contained 19,1 OD280, 7,7Xlθ6 U (134 %) and the specific activity was 403 000 U/OD28O (for chromatogram see Fig.2).The fractions were pooled and concentrated (4°-8°C) to 80 ml using a 400 ml Amicon stirring cell equipped with an YM10 membrane (Amicon, Zurich, Switzerland). 80 ml of

6 M Guanidinium hydrochloride

2,5 mM Zwittergent 3-12

50 μM EGTA <Buffer F>

50 mM Tris/HCl pH 7,5

were added and the concentration/dilution procedure was 3 times repeated. The solution was finally concentrated to give 2,6 ml. This volume was applied to a Superdex 200 column (Pharmacia, 1,6 X 60, 120 ml, 1 ml/min). The column had been preequiUbrated with buffer F. Fractions eluting between 47 and 63 % of the column volume (20 ml) contained 3,8 OD28O, 4X10 ⁶ U (69 %) and the specific activity was 1.050.000 U/OD28O (f° ^r chromatogram see Fig. 3). The fractions were pooled and concentrated using a 50 ml Amicon stirring cell equipped with an YM10 membrane to give 5 ml, which were then diluted by 5 ml of

6 M urea

0,3 mM Zwittergent 3-12 <Buffer G>

20 mM Tris/HCl pH 7,5

Concentration/dilution was once repeated and the solution was concentrated to 1,6 ml and then 1:30 diluted to give a final volume of 48 ml. This volume was applied to a Mono S column (1 ml, Pharmacia) preequiUbrated with buffer G. The column was washed with 12 column volumes of buffer G and then eluted by a linear gradient (0,5 ml/min) from 100 % buffer G to 100 % of

6 M urea

0,3 mM Zwittergent 3-12

20 mM Tris/HCl <Buffer H>

500 mM NaCl pH 7.5

Fractions eluting between 36 and 56 % of buffer H contained 0,27 OD280. 4,4X10 ⁶ U and the specific activity was 16.000.000 U/OD280 (for chromatogram see Fig.4). The total yield based on activity was 76 % and the piuification factor was 97000. Figure 5 demonstrates the protein composition of investigated fractions by SDS-PAGE. The biological activity correlates with the intensity of a protein band, which shows a molecular weight between 41 and 45 kD, when analyzed by SDS-PAGE (reduced) of 12,5 % T stained by Coomassie R 250. Part of the fractions was further purified (see Example 4).

Tfrnπmlft ft

A) Purification of MM out of the supernatant o£MDA-MB-231 cells to high purity bv five chromatograph c stβna

In a similar experiment as described under Example 2 MM was purified from 130 1 of supernatant. In difference to Example 2 more MM was found in the flow through of the Q-Sepharose and this chromatography was therefore repeated with the flow through. Also further chromatography via Heparin-Sepharose and Superdex 200 was done via 2 separate runs. The gradient of the Mono S chromatography was improved as follows:

0 % buffer H to 40 % buffer H in 48 min 40 % bufferH to 60 % bufferH in 120 min 60 % buffer H to 100 % buffer H in 46 min

The eluted fractions between 42 and 45,5 % of buffer H contained 0,61

OD280. 4,1X10 ⁶ U and the specific activity was 6,7X10 ⁶ U/OD280.

For further purification, these fractions were concentrated using an Amicon stirring cell of 10 ml equipped with an YM10 membrane. Volume was reduced from 11 ml to 0,7 ml. 2,5 ml of buffer F were added and the volume was reduced to 0,7 ml. Dilution and concentration were once repeated and the sample of 0,7 ml was applied to Superdex 200 column (1 X 30 cm, 0,5 ml/min) preequlibrated in buffer F (for chromatogram see Fig. 6). 50 μl -aliquots of the resulting fractions were transfered into buffer G by the use of

a Fast Desalting column (Pharmacia) in connection with the Pharmacia SMART system. The resulting samples were investigated by SDS-PAGE (reduced, 12,5 % T, Coomassie stain, see Fig. 7). The major fraction (m) exhibits only one protein band in the range from 41 to 45 kD. The yield in this fraction is 0,12 OD280, the activity 4X10 ⁶ U and the specific activity is 33X10 ⁶ U/OD280. The total yield based on activity is 36 % and the purification factor 170.000.

The fraction n eluting behind the major fraction (m) was further o purified (see Example 4).

B) Purification of MM out of the supernatant of MDA-MB-231 cells to h fh purity bv eight chromatographic steps 5

In a similar experiment as described in Example 2, MM was purified from 146 1 of supernatant with the following steps (omitting chromatography on Q-Sepharose): 6.6 1 of concentrate were diluted and loaded on a column of Heparin Sepharose (80 ml) using 1 mM CHAPS as detergent. The active 0 fractions eluted were diluted and loaded on a column of Heparin Sepharose (50 ml) using 1 mM CHAPS as detergent and 2 mM spermidine as an additive in all buffers. The active fractions were pooled, diluted with a buffer containing 4 mM zwittergent 3-12 as detergent and loaded on a column of Heparin Sepharose (10 ml). The active fractions (8 ml) were pooled and, after 5 supplementation with 6 M guanidinium hydrochloride, purified in 4 runs with 2 ml each on a Superdex 200 column (120 ml). The active fractions were pooled, diluted and concentrated in two steps using a 10 ml and a 2 ml column of Heparin Sepharose. The fraction with the highest activity (2 x IO- ⁶ U/ml) of the final step was used for growth experiments as described in 0 Example 5C). This fraction was run on a SDS acrylamid gradient gel (10-15%, Phast Gel, Pharmacia). After silver staining it showed a major band of approximately 52 kDa.

5

Eϊmm te 4

Purification of MM to apparent homogeneity

The material from Example 2 ( fractions f to h) was purified according procedure A and fraction n from Example 3 was concentrated according pro¬ cedure B, because of the high salt concentration (Buffer F) present in the sample.

Procedure A: The ionic strength of the material from Example 2 was re¬ duced by fourfold dilution and then concentrated on a Mono S (PC 1.6/5) (Phar¬ macia, Uppsala, Sweden) column on a SMART chromatography system (Pharmacia).

The column was developed with a flow rate of 100 μl / min and a linear gradient (12.5%/min) from buffer G to H (for buffer composition see Example 2). 50 μl fractions were collected and active fractions (for activity test see Ex¬ ample 6) were pooled. A typical chromatogram of the concentration of MM is shown in Figure 8.

Procedure B: The material from Example 3 was concentrated on a re¬ versed phase column Poros R H (Perseptive Biosystems, Cambridge, MA USA) (inner diameter 0.8mm, length 10 cm) packed by LC Packings, Amster¬ dam, Netherlands on the SMART chromatography system from Pharmacia using the following buffers:

0.1% TFA in water <Buffer J>

0.09% TFA in acetonitrile <Buffer K>

The column was developed with a flow rate of 50 μl / min and a linear gradient (5%/min) from buffer J to K. 25 μl fractions were collected and ac¬ tive fractions were pooled.

The samples from procedures A and B were applied to agarose gels (Pro- Sieve Gel System from FMC Bioproducts, Rockland , ME, U.S.A) and run as described by the manufacturer under non-reducing conditions using high quality grade SDS.

After completion of the electrophoresis the agarose gel was cut into 2 τmτι slices from the anode to the cathode and the protein was eluted at am¬ bient temperature for at least four hours in

6 M Urea

20 mM Tris-HCl, pH 7.5

0.3 mM Zwittergent 3-12 <Buffer I>

0.3 M NaCl

Activity could be measured in the gel eluates corresponding to the 45 to

55 kDa slices. Active eluates were pooled.

The pooled material was applied to a reversed phase column Poros R H (Perseptive Biosystems, Cambridge, MA, U.S.A) (inner diameter 0.8mm, length 10 cm) packed by LC Packings, Amsterdam, Netherlands) on a

SMART chromatography system (Pharmacia).

The column was developed with a flow rate of 50 μl / min and a linear gradient (2.5%/min) from buffer J to K. 12 μl fractions were collected and ac- tive fractions (see Example 6) were pooled. A typical chromatogram for the separation of MM on a reversed phase column is shown in Figure 9.

Aliquots of the purified product were loaded on a SDS gel (Phast Gel 10- 15%; Pharmacia) under nonreducing and reducing conditions together with low molecular weight protein standards (available from, e.g. Bio-Rad, Her¬ cules, CA USA, 8ng/lane/protein, lane 1), and stained with silver (Fig. 10a). Scans of lanes 1-3 were also performed and are shown in Fig. 10b.

230 pmoles of the purified product were used for sequencing. MM was reduced, S-carboxy-amidomethylated and digested with endoproteinase Lys-C

(Wako, Neuss, Germany) for 16 hours at 37°C. Enzyme to substrate ratio was 1:100. The digest was collected on a Mini S cation exchanger column (3x0.5 cm) equilibrated with 20 % acetonitrile in potassium phosphate, pH 2.5. The column was developed with a NaCl gradient and the peak-containing fractions were collected. Peaks were further purified on Vydac C4 (0.1x15cm) or Brownlee RP-300 (0.1x10cm) reversed phase columns, developed with a water/acetonitrile-0.1 % TFA gradient and sequenced on a 475 A or 477A sequenator (Applied Biosystems, Foster City, CA). Sequences of three peptides were obtained having the amino acid sequences of Leu-Val-Leu-Arg-X-X-Glu-Thr (SEQ ID No: 1),

Ser-Glu-Leu-Arg-He-Asn-Lys (SEQ ID No: 2), and

X-Leu-X-Asn-Pro-X-X-Tyr-Leu (SEQ ID No: 3).

"X" represents a sequencing cycle in which the amino acid was not possible to identify.

Effects of MM on hormone-dependent cells

A) Mor ^p holo ^gy : The morphological changes induced by exposing ZR-75-

1 hormone-dependent cells to conditioned medium from hormone- independent human mammary tumor cells MDA-MB-231 are depicted in the photomicrographs of Fig. 11. 50,000 hormone-dependent cells (ZR-75-1, available, e.g., from American Type Culture Collection (ATCC), Rockville, Maryland, USA) per cm^ were seeded on cell culture dishes precoated with collagen type IV and grown for three days under serum-free conditions. Then the culture medium was replaced by serum-free medium conditioned by hormone-independent MDA-MB-231 cells. Consecutive phase contrast photomicro-graphs were taken immediately before the medium exchange (Fig. 11a) and 15 minutes (Fig. lib), 35 minutes (Fig. lie) and 3 hours (Fig. lid) after stimulation. The photomicrographs show that unstimulated hormone-dependent cells grow in culture as epithelial-like patches with tight intercellular contacts (Fig. 11a). The peripheral cells show ruffling membranes and start to produce lamellipodia (Fig. lib) within minutes of stimulation by MM-containing conditioned medium from the producer cells or by purified protein preparations of all purification stages. The colonies enlarge considerably due to cell flattening and cell-to-cell contacts get weaker in the following time (Figs, lie and lid). Activation of the responding cells by

highly purified MM lasted several hours. Then, the cells reassumed the original patch-like formation. The duration of the morphological changes was dependent on the concentration of MM and lasted longer at higher concentrations. Cell activation was observed in a wide variety of hormone- dependent cells as shown in Table I. All cell lines in Table I are publicly available from, e.g., American Type Culture Collection (ATCC) Rockville, Maryland, USA.

Hormone-independent human mammary tumor cells MDA-MB-453 did not respond with morphological changes upon stimulation by purified MM or by conditioned medium from MDA-MB-231 cells nor did they secrete a factor activating membrane ruffling of ZR-75-1 or T-47D cells.

B) Motilitv: The responding hormone-dependent cells showed a temporary activation of random motility associated with a fibroblast-like cell shape upon exposure to medium of fast growing .hormone-independent cells or purified MM. Due to this induced motility, the cell patches formed after reassoάation may not be composed by the same cells as the patches before stimulation with MM.

C) Proliferation:

1. Effects of MM on cell proliferation of estrogen receptor-positive human mammary tumor cells T47-D

Hormone-dependent T47-D cells were seeded into 96 well cell culture plates previously coated with collagen type IV at a density of 10,000 cells per well and grown in serum- and phenol red-free media. Purified mammamodulin, prepared as described in Example 3B), was used at concentrations of 20 U/ml. For comparison, the effects of the mitogens insulin-like growth factor type I (IGF-I) and estradiol (E2) and the growth inhibitors tumor necrosis factor -alpha (TNF-alpha) and Interleukin-lα (IL- 1) in the absence as well as in the presence of mammamodulin were determined. After 5 days, the cell numbers in the wells were determined as described by Kϋng et al., Analyt. Biochem. 182, 16-19, 1989. The O.D. at 590 nm of the solubilized dye (crystal violet) which was taken up by the fixed cells during staining correlated linearly with cell numbers.

The purified mammamodulin stimulated the proliferation of hormone- dependent cell line T47-D as shown in Fig. 12. In this figure, bars represent means of quadruplicates and standard deviations. IGF-I at the concentration used stimulated the cells maximally and mammamodulin had no additional effect on cell proliferation. Estradiol alone had little effect on cell growth but mammamodulin stimulated these cells to grow faster. The inhibitory effects of TNF-alpha and E -1 on T47-D proliferation were, at least partially, reversed by simultaneous addition of mammamodulin.

2. Effects of MM on cell proliferation of estrogen receptor-positive human mammary tumor cells MCF-7

7,000 MCF-7 cells were seeded per well into cell culture microwell plates (Falcon) and grown under defined serum- and phenol red-free conditions [Kϋng et al., Contr. Oncol. 22, 26-32 (1986)1 without and with 25 U/ml mammamodulin or 1 x IO ^*10 M estradiol. The MM preparation used for this experiment was material of highest purity from reversed phase column as shown in Fig. 9, (pool material). Aliquots of this material were diluted immediately after elution from the column in serum-free cell culture medium. After 3 days, the cultures were refed. After 7 days, the cells were fixed and their numbers measured by the crystal violet assay as described by Kung et al. (1989), supra. The O.D. at 590 nm of the solubilized dye (crystal violet) which was taken up by the fixed cells during staining correlated linearly with cell numbers.

The growth of MCF-7 was strongly stimulated by 25 U/ml MM as shown in Fig. 17. Cells were also stimulated by estradiol (E2) which was used as a control for normal function of the cells. In Fig. 17 the increases in cell numbers of stimulated cultures are expressed in percent of data obtained from the control cultures. The values represent means and standard deviations of quadruplicates.

D) Mitogenic activity: Cell cycle analysis by flow cytometry demonstrated that MM induced a marked increase in the proportion of hormone-dependent cells in the synthetic (S) phase. Cultures of hormone- dependent MCF7, ZR-75-1 and T47-D cells were washed twice with serum- and phenol red-free medium two and one day before starting the experiments for cell cycle analysis. Then the media were supplemented with 25 units of . pure MM preparation (from pool of reversed phase column in Fig. 9). The

cells were harvested 24 hours after stimulation and analyzed using a standard protocol for the FACScan analyzer (Becton Dickinson, CellFIT program). The results showed that treatment with pure MM considerably increases the percentage of S-phase cells of MCF-7, ZR-75-1 and T47-D cells. MCF-7 cells showed an increase of 186%, ZR-751 cells of 243% and T47-D cells of 161% above control cells. To provide a basis for comparison between MM and known mitogens for these cells, cells were treated with high concentrations of insulin-like growth factor (IGF-I, IO ^*8 M) or estradiol (3 xlO" ⁹ M). The values for S-phase increase induced by IGF-I or estradiol were with MCF-7 cells 176% and 158%, with ZR-75-1 cells 280% and 180% and with T47-D cells 203% and 105%, respectively.

E) Conversion of hormone-dependent cells: MM-containing conditioned media from fast proliferating hormone-independent cell line MDA-MB-231 demonstrably suppressed estrogen receptor (ER) levels of hormone-dependent cells. In MCF-7 cells, exposure to MM-containing media (30 units/ml) for 2 days reduced the number of ERs to about 55% +/- 10% (S.D. from 5 determinations), and in ZR-75-1 cells, the number of ERs was about 32% +/- 20% in comparison to untreated cells, as measured by a binding assay with intact cells using tritiated estrogen (17-β estradiol).

It was also shown that MM suppresses the ER mRNA expression in MCF-7 hormone-dependent cells. MCF-7 cells (2.5 x IO ⁶ ) in serum- and phenol red-free cultures were stimulated with 20 units ml of highly purified MM (Example 2, MonoS fraction between g and h), or with 1 x 10" ^d M estradiol in the absence or presence of MM. The cells were harvested 5 hours after treatment. RNA was then extracted from the cells and purified for mRNA on pre-packed spun columns (Pharmacia). The mRNA concentrations were measured and then 2.5 μg of each extract was loaded and resolved on an agarose gel. After transfer to a membrane (Northern blot), the mRNA was probed for ER message by hybridization with a human ER probe (Fig. 13). ER mRNA was expressed in MCF-7 control cells (Fig. 13, lane 1). IO ^*9 M estradiol stimulated ER mRNA expression (Fig. 13, lane 2). MM reduces the mRNA expression considerably in the absence (Fig. 13, lane 3) as well as in the presence of IO" ⁹ M estradiol (Fig. 13, lane 4).

In a separate experiment, 2.5 x IO ⁶ MCF-7 cells in serum and phenol red-free cultures were stimulated with highly purified MM (Example 3, fraction "m" of Fig. 6), lxl 0' ¹⁰ M estradiol or combination thereof for 24 hours. At this time

point, mRNA was extracted from the cells and purified using a QuickPrep micro mRNA preparation kit from Pharmacia. The mRNA preparations were resolved on 1.1 % agarose gels and transferred to nitrocellulose membrane by Northern blotting. The blot was hybridized with a radioactive ER probe. The results (Fig. 14) demonstrate that estradiol inhibited the expression of mRNA for the ER. MM markedly reduced the mRNA level of ER. This effect was accentuated when estradiol was coincubated with MM.

ER and progesterone receptor (PR) proteins were measured by extraction of stimulated MCF-7 cells and appUcation of enzyme immuno assays (EIA). MCF-7 ceUs were seeded into 12-weU cell culture plates (Costar) in the presence of 5 % FCS at a density of 150,000 cells per weU. The next day, the medium was replaced by serum- and phenol red-free medium. The foUowing day, ceUs in dupUcate wells were stimulated with MM (200 U/ml) or estradiol (1 x IO ^-9 ). After 6 and 48 h, the media were removed and the ceUs frozen at -70°C for at least 12 hours. After thawing, the cells were extracted according to the method described by Madeddu et al. [Eur. J. Cancer Clin. Oncol. 24, 385-390 (1988)1 by adding 225 μl of extraction buffer composed of 500 mM KC1, 10 mM KH2PO4, 1.5 mM EDTA and 5 mM Na-molybdate. The pH was adjusted to 7.4 with IM KOH. Beta-mercaptoethanol was added freshly before use at a concentration of 0.01 %. The buffer was at 4° C. After 90 min, the extracts were centrifuged in 1 ml Eppendorf tubes for 5 min at 12,000 x g and 4°C. The supematants were coUected and 100 μl of extract used for receptor determination. ER and PR levels were measured using EIA kits manufactured by Abbott Laboratories, North Chicago, IL, USA. The results (Fig. 15) showed that 6 h after stimulation with MM, the ER protein levels of MCF-7 ceUs were slightly and after 48 h strongly reduced in comparison to controls. PR levels were similar to controls at both time points when ceUs were stimulated with MM.

As a control for normal ceU function of MCF-7 ceUs, ER and PR levels were also measured in ceUs treated with estradiol. Estradiol (E2) is known to inhibit the expression of both ER mRNA and ER protein and to stimulate PR expression of MCF-7 ceUs. Furthermore, PR expression is dependent on ER activation by estradiol [Ree et al, Endocrinology 124, 2577-2583, (1989)]. The control experiments showed that estradiol reduced ER protein expression and stimulated PR protein expression. These results with MCF-7 cells are consistent with the data obtained by Ree et al., supra, and Read et al. [Molec. Endoσinol. 3, 295-304 (1989)].

F) Tyrosine phosphorylation of cell membrane proteins of hormone- dependent cells: The effects of purified MM (fraction "m" of Fig. 6) on tyrosine phosphorylation of MCF-7 membrane proteins were tested.

MCF-7 ceUs were seeded into 24-weU Falcon culture plates (2 cm ² /weU) at a density of 150,000 ceUs/weU in ceU culture medium containing 5 % fetal calf serum (FCS). After 24 hours, the medium was exchanged with medium containing 1 % FCS. The following day, the ceUs were incubated for two hours with serum-free medium before stimulation. CeUs were treated with MM concentrations of 20 and 200 U/ml. An unstimulated control was included. After 30 min at 37°C in a ceU culture incubator, the media were removed and the ceUs extracted by heating for 5 min in an oven at 100°C with 100 μl/weU of reducing sample buffer for SDS polyacrylamide gel electrophoresis containing 4 % SDS and 2.5 % mercaptoethanol. 60 μl of each extract and an aUquot of EGFR reference sample were loaded and resolved on mini Tris-Tricine gels (6 %). The proteins were transferred to a PVDF membrane (Bio-Rad Laboratories, Hercules, CA) by Western blotting using 10 mM CAPS buffer, pH 11.0 with 10 % methanol and subsequently probed with an antibody raised against tyrosine phosphate. The detection of tyrosine phosphorylated proteins was performed with a second, alkaline phosphatase- labeled antibody obtained from Upstate Biotechnology Incorporated (UBI, Lake Placid, NY, product No. 17-105) and used as recommended by the manufacturer. The results showed that MM stimulated the tyrosine-specific phosphorylation of a membrane protein of MCF-7 ceUs with an apparent molecular mass of approximately 180-190 kDa when compared with molecular weight standards and with a phosphorylated EGF receptor reference from EGF-stimulated A431 (Fig. 16). The identity of the phosporylated protein(s) was not determined.

G) Elevated expression of cell surface EGF receptor and erbB2 levels after stimulation of hormone-dependent cells with MM: 12 x IO ⁶ MCF-7 ceUs were grown on culture plastic dishes with 10 cm diameter in serum- and phenol-free ceU culture medium for 1 day and stimulated with 30 U/ml or 100 U/ml of highly purified MM (fraction "m" of Fig. 6) for 24 hours. The surface receptor numbers were determined with flow cytometry using a mouse monoclonal antibody directed against the ceU surface domain of the human EGF receptor (Ab-1, Oncogene Science, Manhasset, NY) or the cell surface domain of erbB2 and a second, fiuorescein isothiocyanate (FITC)- labeled goat anti-mouse antibody (Becton Dickinson, San Jose, CA) was

appUed to stain the ceU-bound mouse anti EGF receptor or anti erbB2 antibodies. The ceUs were scraped from the dishes, washed in serum-free medium and counted using a micro ceU counter (Sysmex F-300, TOA Medical Electronics, Kobe, Japan). To 10^ ceUs in 80 μl serum-free medium, 20 μl first antibody solution was added and incubated for 45 min on ice. Then the ceUs were washed 3 times in cold serum-free medium, resuspended in 100 μl and incubated with 4 μl FITC-labeled goat anti mouse antibody for 45 min on ice in the dark. Afterwards, the ceUs were washed twice with serum- free medium, resuspended in 250 μl serum-free medium and analyzed at room temperature in the flow cytometer (FACScan, Becton Dickinson, San Jose, CA). Histograms of the FITC-fluorescence of the individual suspensions were acquired and data were expressed as the mean fluorescence channel numbers.

At least 90% of aU ceUs were staining with the fluorescent antibody, as was confirmed by determining the number of ceUs with increased fluorescence in a gated dot plot on the flow cytometer. Bound fluorescence was calculated by subtracting the fluorescence of ceUs incubated only with the second (FITC-labeled) antibody from the fluorescence obtained from samples incubated with the first antibodies (the anti receptor antibodies) and the FITC-labeled antibodies.

The results of the experiment are shown in Figures 18 and 19. The values represent the means of 3 determinations. In hormone-dependent MCF-7 ceUs stimulated by MM, the EGF receptor levels were approximately 3-fold higher after 24 hours than the receptor level of control ceUs (Fig. 18). ErbB-2 levels were found to be elevated by 35 % after 24 h of stimulation by 100 U/ml of MM (Fig. 19).

Example β

Determination of MM activity during isolation.

Determination of MM activity is based on the induction of fast morphological changes of the estrogen receptor-containing cell Une T47-D. T47-D ceUs were grown in serum and phenol red-free ceU culture medium supplemented with transferrin, low levels of insulin and 1 mg/ ml bovine serum albumin (Albumax I from Gibco). The seeding density was between

25,000 and 35,000 ceUs per cm ² . Two days after seeding, the cells were used for MM activity testing. T47-D ceUs reacted to MM similar to ZR-75-1 ceUs (compare Fig. 11a) but were sUghtly more sensitive and easier to handle because their attachment to the substrate is tighter. MM containing solutions were added in appropriate amounts to weUs containing the test ceUs. The culture plates were then immediately returned to the ceU culture incubator to maintain the normal 37°C temperature for ceU cultures. After 10 to 15 minutes, and a second time after 30-45 minutes, the ceUs were inspected by phase contrast microscopy at a magnification of 200 x for induced membrane ruffling and lameUipodia formation. Five levels of activation were distinguished: a) no activation (-), b) weak and transitory, but unequivocaUy detectable formation of membrane ruffles in many ceU patches (+), c) lameUipodia formation in many ceU patches and membrane ruffles in all cell patches (++), d) formation of lammeUpodia in aU and of enlarged lameUipodia in many ceU patches (+++) and e) intensive membrane ruffling and formation of large lameUipodia in aU cell patches (++++). This gradation of test results corresponded roughly to the ceU activation pattern of the 5 last steps in a 1 to 2 dilution of MM containing media or chromatographic fractions. The dilution with an assigned one plus (+) activity was defined as containing 1 unit MM per ml.

The results were subjected to some variabiHty, depending mostly on the actual conditions of the ceU cultures (ceU density, age of cultures). UsuaUy, differences between measurements of MM activity at other time points and other cultures of T47-D ceUs were not greater than 50% to 200%.

Example 7

Antibody to MM

MM is antigenic in npnhuman mammals, so that antibody is produced by the animals after they have been inoculated with the MM. Two doses of 5 μg MM are sufficient to induce production of antibody in mice. Polyclonal antisera can be isolated from the blood of the inoculated animals.

Monoclonal antibody to MM is produced by conventional Kδhler-MUstein processes: immunization of a suitable animal species by injection with MM, recovery of antibody-producing ceUs (i.e., spleen cells) sensitized to MM,

immortalization of the antibody-producing ceUs by fusion with a compatible myeloma ceU line, and isolation of the monoclonal antibody from a selected immortalized ceU line thus established. Preferably, female Balb/c mice are injected with MM in Titremax (CytRx, 150 Technology Parkway, Technology Park Atlanta, Norcross, Georgia 30092), administered by i.p. injection. After two weeks, this injection is repeated. One week later, blood samples are collected, and tested, e.g., by western blot, for presence of antibody. The mice which exhibit the highest antibody levels are sacrificed, and their spleen ceUs are immortalized by fusion with mouse (Balb/c) myeloma ceUs, e.g., using PEG 4000, and distributed into 24 X 24 weUs. The hybridoma lines thus produced are screened and selected for the production of antibody.

Example 8

Cloning and expression of MM

MM cDNA may be isolated from a random-primed cDNA library created using poly(A) ⁺ RNA from MDA-MB-231 ceUs. A vector, e.g., Lambda ZAP II vector, is utilized to form the MM-fusion protein, and the expressed protein is then screened using MM antibody. Alternatively, the cDNA library may be screened using MM amino acid sequence information, e.g., by colony hybridization techniques, for example expressing the Ubrary in an expression system, preferably E. coU, lysing the colonies, e.g., on nitroceUulose filters, denaturing their DNA in situ and fixing it on the filter, hybridizing with labeled, preferably radiolabeled, oligonucleotide probes of at least 24 base pairs having cDNA base sequences corresponding to aU or a portion of the amino acid sequence of MM, identifying hybridized colonies, and retrieving the corresponding vectors from the Ubrary, using chromosome walking techniques if necessary to isolate and characterize the entire cDNA. Once the cDNA has been isolated, it may be expressed in a suitable expression system, e.g., a prokaryotic system such as E. coU. The MM may be isolated from the culture medium of the expression system, e.g., using the procedures outlined above.

Trample 9

Use o heparin as MM inhibitor

Heparin is useful in the treatment or control of breast cancer due to its inhibitory effect on MM activity. Heparin for this use is preferably administered parenteraUy, most preferably by means of an implantable pump permitting long term, continuous intravenous infusion. Although heparin is relatively nontoxic, a trial dose of 1000 units should precede usual therapeutic dosages to confirm that the patient wiU not have an aUergic reaction. Heparin preparations should be used which are designed for long- term administration, e.g., have low anticoagulant and angiogenic activity.

fi* τ_I.1fi 10

Therapeutic use of monoclonal antibody to MM

Monoclonal antibodies to MM may be used therapeuticaUy to block MM activity, thereby controlling or reducing the metastasis of breast cancer.

Suitable dosages and forms of administration will be apparent to one skiUed in the art.

Example 11

Diagnostic use of polvclonal or monoclonal antibody to MM

Assay kits utilizing polyclonal or monoclonal antibodies (hereinafter collectively referred to as antibodies) to MM may be comprised of the following components: (i) antibody, preferably lyophilised; (ii) labeled (preferably radiolabeled) MM or fragment thereof; and (iii) MM standard containing a known amount of MM. If radiolabeled MM is used, it is preferably ¹²⁵ I- labeled MM, labeled to about 50-100 μCi/μg.

The antibody is dissolved and incubated together with the labeled MM or fragment and either the sample to be assayed or the MM standard. Incubation takes place preferably at cool temperatures, e.g., about 4°C, and

lasts for at least two hours, preferably 4-6 hours. The pH of the incubating mixture is preferably kept in the range of from 5 to 8, more preferably at 7 or 8, preferably with the aid of a buffering agent such as a citrate or tris buffer. After incubation, the fraction of labeled MM or fragment is separated from the unbound fragment, e.g., by the use of charcoal such as dextran-coated charcoal. The unbound fraction adsorbs to the charcoal and may then be separated by filtration or by centrifugation. The amount of radioactivity in one fraction is then measured by standard techniques, e.g., by Uquid scintillation counting after the addition of a secondary solute. The proportion of labeled MM or fraction bound to the antibody is inversely proportional to the amount of MM in the unknown plasma sample. For quantitative analysis, a caUbration curve may be prepared by analyzing solutions of MM of known concentration.

Example 12

Animal models

Nude mice inoculated with human mammary tumor cells are a preferred animal model for study of MM action and effects of blocking MM action by antagonists. For example, nude mice are inoculated with hormone- dependent mammary tumor ceUs, e.g., MCF-7 ceUs, and the developing tumors are treated with MM (infusion) in the absence and presence of MM blockers such as heparin, neutraUzing MAbs, and MM receptor blockers, to evaluate the actions of MM in vivo on responder ceUs. Alternatively, nude mice are inoculated with hormone-independent mammary tumor ceUs, e.g., MDA-MB-231 ceUs, and the developing tumors are treated with heparin, blocking MAbs for MM, and MM receptor blockers to evaluate the role of MM in producer cell proliferation in vivo.

Table I.

Human tumor and normal mammary ceUs producing or responding to a factor (mammamoduUn, MM) inducing membrane ruffling and cell motiUty.

cells and expression of secretion of reaction to reaction to cell lines estrogen cell activating CM from purified receptors factor ( ) MDA-M&-231 MM( ² )

MDA-MB-231 ATCC HTB 26 +

HBL-100 ATCC HTB 124 - + - -

Hs578T ( ³ ) ATCC HTB 126 - + - -

Hs578Bst ( ⁴ ) ATCC HTB 125 - (+) - -

SK-BR-2 III ATCC HTB 29 - + - -

MDA-MB-453 ATCC HTB 131 - - - -

BT-20 ATCCHTB 19 - - - -

MCF-7 ATCC HTB 22 + . + +

ZR-75-1 ATCC CRL 1500 + - + +

ZR-75-30 ATCC CRL 1504 + - + +

T-47D ATCC HTB 133 + - + +

MDA-MB-361 ATCC HTB 27 + - + +

BT-474 ( ⁵ ) ATCC HTB 20 (+) - (+) +

normal epithelial cells/ „ (+) _β orga-noids normal stro a cells ( ⁶ ) + " -

0) Results are based on the observation of morphological changes after stimulation of the mammamoduUn responder cells ZR-75-1 and T-47D by conditioned media (CM) of the ceUs listed. Fresh serum-free medium was added to confluent ceU cultures of each ceU (grown serum-free) and coUected 3 days later for the testing. The tests were performed as described in Example 6. "+" stands for unequivocaUy detectable ceU activation which is inhibited by 100 μg/ml heparin; (+) stands for marginal ceU activation.

( ² ) highly purified factor (fraction "m" of Fig. 6 ).

( ³ ) tumor and ( ) myoepithelial ceU lines of the same tumor specimen.

( ⁵ ) BT-474 cells express very low levels of ER and PR.

(6) from dysplastic mammary tissues.

Hs578Bst, MDA-MB-453 and BT-20 are estrogen receptor-negative cells that have low growth rates in ceU culture and produce only very low or undetectable amounts of mammamodulin activity.

SEQUENCE LISTING

( 1 ) GENERAL INFORMATION :

( i) APPLICANT :

(A) NAME: F.HOFFMANN-LA ROCHE AG

(B) STREET: Grenzacherstrasse 124

(C) CITY: Basle

(D) STATE: BS

(E) COUNTRY: Switzerland

(F) POSTAL CODE (ZIP) : CH-4002

(G) TELEPHONE: 061 - 688 42 56 (H) TELEFAX: 061 - 688 13 95

(I) TELEX: 962292/965542 hlr ch

(ii) TITLE OF INVENTION: NOVEL ORGANIC COMPOUNDS AND USES THEREFOR (iii) NUMBER OF SEQUENCES: 3

(iv) COMPUTER READABLE FORM:

(A) MEDIUM TYPE: Floppy disk

(B) COMPUTER: Apple Macintosh

(C) OPERATING SYSTEM: System 7.1 (Macintosh)

(D) SOFTWARE: Word 5.0

(2) INFORMATION FOR SEQ ID NO: 1:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 8 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: peptide

(v) FRAGMENT TYPE: internal

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 1:

Leu Val Leu Arg Xaa Xaa Glu Thr

1 5

(2) INFORMATION FOR SEQ ID NO: 2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 7 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: peptide

(v) FRAGMENT TYPE: internal

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 2:

Ser Glu Leu Arg lie Asn Lys

1 5

(2) INFORMATION FOR SEQ ID NO: 3:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 9 amino acids

(B) TYPE: amino acid (D) TOPOLOGY: linear

(ii) MOLECULE TYPE: peptide

(v) FRAGMENT TYPE: internal

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 3:

Xaa Leu Xaa Asn Pro Xaa Xaa Tyr Leu

1 5