CA2083688A1

CA2083688A1 - Growth factor which inhibits the growth of cells overexpressing the human oncogene erbb-2

Info

Publication number: CA2083688A1
Application number: CA002083688A
Authority: CA
Inventors: Marc Lippman; Ruth Lupu
Original assignee: Individual
Current assignee: Georgetown University
Priority date: 1990-05-25
Filing date: 1991-05-22
Publication date: 1991-11-26
Also published as: AU7960591A; AU667026B2; EP0597845A4; WO1991018921A1; EP0597845A1; JPH05509079A

Abstract

A method for inhibiting the growth of adenocarcinoma cells in a human, which cells overexpress the oncogene erbB-2, which entails administering to said human an amount of a 30 Kd glycoprotein effective to inhibit the growth of said cells.

Description

~0 91/18921 ~ 3 PCT/I S9I/03~3 Descri~tlon of Cells overexeressinq the Human Oncogene er~B-2 Technical _ield The pr~sent invention relates to a growth factor which interacts with the hu~an oncogene er~B-2, and which inhibits the growth of cells overexpressing this oncogene.

Bac~qround Art Carcino~enesis is b~lieved to be a multi-step process of alteration of genes which are involved in the growth control of cells. A variety of proto-oncog~nes and oncogenes have been implicated in t~e actiYation of tumor cells as regulating factors. For example, oncogenic protein ~inases are b21ieved to induce cellular trans~orma~ion through either inappropriate or excessive prot~in phosphcrylation, resulting in the uncontroll~d growth Or malignant neoplasms. See HistopathQloqY, Wrba, F. et al, : 15, 71-7~ (l9a9)~

: one group of proto-oncogenes encodes cellular growt~ factors or their receptors. ThQ c-erbB-1 gene encodes the epidermal growth ~actor or their receptors. T~e c-sis gene encodes the ~-chain o~ the platelet-derived growth factor.
The c-fm5 qene encodes a ralated or identical molecule for th~ ~ecap~or o~ the granulocyte-ma~rophag~ colony stimulatinq fac~or. A fourth member of this group of proto-onco~enQs~ called neu was iden~i~ied in ethylni~rosourea-induced rat neuroblasto~as. The human counterpart or nau, called ~ER-2/neu or c-erbB-2, has been ~, ~, . . . . . .
, , -. ' i~;. W~ 921 ~ PCT/~S9]/0~3 sequenced and ~apped to th~ chromosomal locus 17q21. See Cancer Research, Schneider, P.M. et al, 49, 4968-4971 (September 15, 1989).

~he HE~-2/NEU O~ c-erb~-2 oncogene belongs to the erbB like oncoqene group, and is related to, but distinct fro~ the epidermal growth factor receptor ~EGFR). The oncogene has ~ en demonstrated to be implicated in a number of human adenocarcinomas leading to elev~ted levels of expression of the pl85 protein productO Eor example, the oncogene has been found to be amplified in breast, ovarian, gas~ric an~ even lung adenocarcinomas.

Furthermore, the amplification of the c-erbB-2 oncogene has been fcund in many cases to be a significant, if not the most significant, predictor of ~oth overall survival time and time to relapse in patients sufferin~ ~rom such forms of cancer.

Carcinoma of the breast and ovary account for approximately one-third of all cancers occurring in women ~: and together are responsible for approximately on~-fourth of -~ 20 cancer related deaths in ~emales. Signif~cantly, the c-srbB-2 oncagene has been found to ~e ampli~ied in 25 to 30% of human p~imary breast cancers. See Sc~nce., Slamon, D. e~ al, 244, 707-712 (May 12, 1989).

The c-er~B-2 oncogene is known to express a 18SKd : ~S trans~embran2 glycoprotein (pl85~bB-2). ~he expressed protein ha~ ~een suggested to be a growth ~actor receptor ~ due tD its homology with EGFR. However, known EGFR ligands, :~ such as E~F or ~GF~, do not bind to p185~rb8-2. At present, :;
~`

WO 91 /18921 ~.., ~ ~; .J ~ ! ~3 ~3 PCl /1_ S9 i /03443 no ligand is known which binds to this protein.

Thus, a need continues to exist .or a li~and for (pl85er~8-2). Such a ligand might be used to counteract the e~fects of c-erbB-2 onco~ene over expression in facilitatlng 5 carcinogenesis.

DiscLQsur~ o~ the In~ention Accordingly, it is an object of the present invention to provid~-a growth factor which interacts directly with the erbB-2 oncoqene.

It is also an object of the present in~ention to provide a m~thod for the isolation and purification of the abovedescribed growth factor.

It is also an object of the present invention to provide a method for inhibiti~g t~e growt~ of cells which overexpress the human oncogene erbB-2.

Accordingly, the above objects and others ar~ provided ~y a 30kDa TGF~-like glycoprotein.

BrLef~ s~ on o~ the Drawinqs Fig~re 1 illus~ra~es the isolation of the present 30 Xd growth factor. Portion A illustra~es th2 us~ of low a~inity heparin chromatography, while portion B illustrates t~e use of reversed-phase chromatography.

Figure 2 illus~rates the detection o~ phosphorylated .

;.

.

WO 91/18921 ~ PCT/~S91/03~a3 proteins in SK-~r-3 cells.

Figure 3 illustrates the detection of phosphorylated proteins in MDA-453 cells.

Figure 4 illustrates the phosphorylation of 5 pl85er~8-2 protein in intact CH0/DHFR and CH0/erbB-2 cells.

Figure 5 illustrates a pl85e~bB-2 reoeptor competition assay in SX~ 3 cells.

Figure 6 illus~rates the inhibition of pl85~rbB-2 crosslinking with 4D5 antibcsdy by gp30.

10 Best Mode f~?~carr~inc7 out the_~nvent~on The human c-er~3B-2 oncogene encodes a ~8~ Kd transmem}~ran~ glycoprotein havin~ prot~in lcinase activity.
Thi glycoprotQin, known as pl85~rbB-2, ~hows esctensive s1:ructural similarity with the pl70 epiders~al growth factor receptor (EGFR) and is therefore thought to be growth factor recep~or. Howaver, neither EGF nor TGF~, the normal ligands ~or the EGF~, interact directly with p~85erba-2. In ~act, no ligand ~or this glycoprotein has b~en described. It would be ext~emely de3irable to find a ligand ~or ~his 185 Kd glycoprotein inasmuch as er~B-2 oncogene is amplified in msny adenocarcinomas and is over oxpressed in nearly 30% of human ~roast cancer patients. Additionally, it is known that pl85rb~-2 iS necassary for the ~aintenance o~ the malignant p~enotype of cells transfor~ed ~y the-oncogene.

., . 25 In accor~ance with the present invention, it has 'Deen ' .

.. . ~
;,, ... ..

WO91/18921 ~ pCT/~S91/0~W3 surprisingly discovered that a 30 Kd growth factor which is secreted fxom ~he estrogen rec~ptor negative cell ~in~
~DA-231 is effective as a ligand for pl85erbB-2 glycoprotein. The 30 Kd glycoprotein of the present invention also exhibits TGFc-liXe activity. For example, the present 30 Kd glycoprotein binds to EGFR, is capable of phosphorylating EGFR as well as inducing NRX colony for~ation. This is quite surprising inas~uch as the present 30 Kd growth factor is distinct from the normal precursor for TG~ or mature TGF~ as shown by peptide mapping of the translated proteins.

Generally, the 30 Kd glycoprotein was i~munoprecipitated by an anti-TGF~ polyclonal antibody and exhibited TGF~ e biological activity as assayed by E~F
radioreceptor a5say and NRK and AlN4T cell colony formati~r.
:assays. The 30 Kd growth factor also stimulated autophosphorylation o~ the EGF receptor more ef~iciently t~an ~atu~e 6 Kd TFG~.

The 30 Xd glycoprotein was observed, unlike EGF and TGF~, to bind to heparin-sepharosa, and was purified to apparent homogeneity by heparin a~inity chromato~raphy and subsequ~nt reversed pha~e chromatography.

~ unicamycin treatment in vivo or N-~lyconase deqlycosylation in vitr~ revealed a precursor of 2~ Kd in : 25 contrast to th~ 18 Kd precursor for ~ature TGF~.
:Furt~er~ore, in vitro translation of total ~RN~ from MDA-M8-231 cells confirmed these observations. Biochemical characterization of the 30Xd TGF~-}ike protein was obtained by V8-prot~ase digestion of the de-glycosylated polypeptides :~ ` - ' . ' - . -~091/18921 ~ r ~ CT/~S91~0~3 - ~6-and translated products. P~ptide mappinq of the Va~diges~ed, immunoprecipita~2d material suggests an ~ln.e acid sequence distinct from TGF~. Hence, the 30Kd polypeptide, whil~ related to ~he EGF/TGF~ family, is encoded by a different gene and is not a post-~ranslation modification of mature TGF~.

Having obtained t~e present 30 ~d glycoprotein, in accordanca with another aspect of the present in~ention, the same i~ used to inhibit the growth of cells which overexpress ~e c-~rbB-2 oncoyene.

In accordance with the present invention, the pr~sent 30 Kd glycoprotein may be used, by itself, or in conjunction with other medicinal substances to inhibit the growth of any cells which overexpress the c~erb~-2 oncogene.

Generally, the present 30 Kd glycoprotein may be used advantageausly to inhibit the growth of adanocarcinoma cells, preferably those of breast, ovarian, gastric and lung tissue which overexpress the erbB-2 oncogene and EGFR.

., In order to furt~er exemplify t~e present invention, reference will now bc made to certain examples which are : provided solely for purposes of illustration and are not intended ~o be limitati~e.

~ER~A~S AN~ ME~DS FOR OBT~INI~G TRE 30 Kd GLYCO~ROTEIU

l ~L~ -Z5 Cell~ from th~ following sources were used: MDA-M3 231 :: .

:'~
WO91/18921 ~:,r ~ !~3 and NRX clone 49F fibroblasts were obtalned from the American type Culture Collection (Rockville, MD). HsS78T
cells, A43l cells, and H8 cells, a TGF~trans~ected MCF-7 br~ast cancer cell line, were available upon request from a variety of sources. Carcinogen~immortalized normal mammary epithelial cell subline 184AlN4 and its SV40-transfected derivative l84AlN4T, were also available on request.
Rat-FeSrV transfected cells w~re also provided upon reques~.
All cell lines were propagated in improved modi~ied Eagle ~ s 10 medium (IMEM, Gibco, Grand Island NY) supplemented with 10%
fetal bovine serum tFss~ Gibco~.

Condi~i~ed Media Pre~aration~ Collection and Concentration Conditioned media collections were carried using a well-known procedure. The media were concan~rated lO0-fold in an Amicon ultr~-filtration c811 (Y~5 membrane) (Amicon, Danvers, MA) . Once clarif ied and conc~ntrated, the media were stored at -20C while consecuti~e collections were ~ade during the following days. The concen~rated media were dialyzed using Spec~raphore 3 tubing (Spec~ral Medical Industries, : Los An~eles, CA) against lO0 volumes of O.l M acetic acid over a two day period at 40C. The material that precipitated during dialysis was removed by centrifugation at 4000 rpm ~or 30 min. at 4C; protease inhibitors were 2~ a~dQd. The clariSied sample was then lyophilized.

~ç~aboLîc ~abel}L~a-~ad I~muno~recipitation Cells were grown ~o 80% confluen~e in IMEM. C~ll :~ monolayers were washed three times with PBS and incubated ,. , '~
.

.. WO91/18921 PCT/~91/0~W3 for two hours in serum-free IMEM which lacked methionine and cysteine and was supple~ented with glu~a~ine (2.9 g/l) (Biofluids, Roc~ville, MD). This medium was then re~oved and replaced with serum-free IMEM witAout met~ionine and cysteine containing 2.5 mCi/ml ~35S~ cyst~ine and methionine : (~mersham, ~rli~gton H~iqhts, IL, 1175 Ci/mmole). A total of 2.5 ml of this m~dium was used for a 5 cm dish. The medium was harvested from the culture after 16 hrs at 37OC
and clarified by centrifugation. Cells were washed once with PBS, har~ested by scraping, and lysed in 1 ~l of RIPA
buffer (300 mM ~aCl, 100 ~M Tris-HCl, containing 2% Triton xlOO, 2~ Nadeoxycholat~, Q.2% SDS, 0.4S BSA and 2 mM PMSF).
Following an incu~ation o~ 30 minutes on ice, the lysate was clarified by centrifugation (30 mi~utes a~ 4000 rp~ and : 15 used i~mediately or was stored at -70C. ~35S]- labelled proteins released into the conditioned media by the dlfferent cell line~ wer~ immunoprQcipitated with lO g (specific or non specific) antibody partially puri~ied by 45% ammonium sulfat~ p~ecipitation. A~ter solubilization the immunoprecipitates were analyzed by 15% SDS-PAGE and subsequent fluorography. ~res~ained molecular weight markers (Biorad, Richmond, CA) ware run in parallel lanes.
unic~y~LT~çatment Tunicamycin ~Sigma, St. Louis, MO) was dissol~ed in 50 mM sodiu~ carbo~ate (pH 10.0) and filter-sterilized with a 0.~2 m ~ilter. Confluen~ monolay~rs of MDA-M3-231, MCF-7 and ~s578~ cells were grown in IME~ in the p~esence of 20 g/ml tunicamycin ~unless otherwise specified) for 4 hours prior to ~etabolic labelling. Metabolic labelling was ~hen performed as described above with continued tunicamycin .
' !~IA
WO 91/18921 ~ PCr/l ~i91 03~3 _g_ treatment.

Elastase Treatment The samples containing TGF~-like activity were incubated with 20 g porcine pancreatic elastase (Si~ma) dissolved in 50 mM ~l~cylglycine, pH 7.9, for 1 hour at 22C. The samples were then subjected to immunoprecipitation and SDS-PAGE analy~is.

PolYclonal Antibodies; Antiserum against human TGF2 was obtained by immu~ization of a rabbit on day O with 400 g of reco~binant TGF~ synthesized in E. coli, provided by ; Genentech Corp. The immunogen was first conjugated to keyhole limpet hemocyanin ~XLH) and wa~ emulsifled in complete Freund's adjuvant and was injected intradermally at multlple sites. Additional injeotions were given as follows: day 60, 175 g TGF~ and day5 90, 150~ 180, and 210, 100 g TGF~. The booster injections were giYen subc~taneously at multiple sites in incomplete Freund's adjuvant. The rabbit serum was assayed for antibody titer by ELISA at 10 to 14 days following each i~jection. The antiser~ collected at day 180, designated R399, was used ~or immunoprecipitation and radioimmunoassay.

Monocl~nal Antibodies; A monoclonal antibody against recombinant TGF~ w~s kind~y provided by Genetech Corp.

ZS Measurement of Anti TGFU An~ibod~ ~R1199)_~evels~by_~3~L~B
~icro-Elisa plates (Dynatech-Immunolon II, Dynatech - WO91/læ92l ~ `' PCT/~S91/03~3 --~o--Laboratories, I~c. Chantilly VA.) were coated for 16 hours at 4C with 500 -.g~l of recombinant TGF~ in 50 mM sodium -carbonate buffer (pH 9.6). Th~ samples to be assayed (antibody) were serially diluted 1:1,000-1:64,000 with O.lS
M NaCl, 0.05 M Tris-HCl (pH 7.4), 2 ~M EDT~, 5 mg/ml bovine serum albumin, 0 . 05% 5~een 20 (TBS-BSA-'rween) and were incubated in the wells for 2 hours at 37C. 'rhe plates were washed fi~a ti~es with PBS-Tween and then incubated for 1 hr at 37ac with horseradish peroxidase-conju~ated goat anti-rabbit i~munoglobulin in ~8S-8SA-Tween. The plates were then washed five ~i~es with P~S-Tween and incubated for 4 hrs at 22C with 100 l per WPll 0~ O- 1 mg/ml o-phenylenediamine, 0O012% H202 in o.l M Phosphate-citrate buf ~er ( p~ 5.0). ~h~ reaction was stopped by the addition of so l/well of 2.5 N H2SO4 and th~ absorbance was measl~red at 492 nm using a U~ 700 Microplate Reader (Dynatech Lab., Inc. Chantilly, VA).

Ra~ munoassaY (RIA) TGF~RIA; The presence of peptides i~unclogically related to TGF~ was determined using a RIA kit with a polyclonal anti-rat ~GF~ and rat t2sI]TGFQ (aiotope~ Inc., Seattle, WA). This antibody does not cross-react with human EGF. Aliquots o~ conditioned media were reduced wit~ 40 mm dithiothreitol and denaturQd by immersion for 1 minute in a ~5 boiling water bath. Assays were done in duplicate according to the manu~oturer's protocol and each collection of conditioned media was assayed at least twic~.

- Solid P~ase RI~; 96 well microtiter plates were coated with anti-TGF~ antibody (R399 or monoclonal antibody) for 2 .

. :
: : --~W~91/18921 ~ PCT/~S91/0~3 --11~

hours at 3~C. The wells were ~hen filled with lO0 l of the column fraction to be assayed for TGF~ activi~y. A standard curve was constructed using 0.075 to 15 ng unlabelled TGF~.
After the 2 hours incubation 5x104 cpm of [l25I] TFG~ or 2x105 cpm o4 metabolically labelled antigen was added per well. The plates were incubated further for 16 hours at 4C. The wells were then washed and counted using a ga~ma counter (Model B5002, Packard Instru~ants Co~, Sterling, VA). The E~F RIA was per~ormed with an anti-EGF antibody (Oncogene Science clone 144-8, ~anhasset, NY). A standard cur~e was constructed using human EGF (HEGF, receptor grade, Gollaborative Research, Waltham, MA).

T~e puri~ied 30 kDa ~GF~-like protein was subjec~d to diqestion with N-glyconase. Sa~ples equivalent to iO0 ng were incubated with 50 1 ot 0.2 M sodium phosphate (pH 8.6~, 1.25% NP40 and 2-6 g N-Glycanase (Genzym2 Corp., Boston, MA) were subsequently added to each sample and incubated at 37C
for 16 hours. 50 l of 3-fold concentrated loading buffer was added ~e~ore e~ectrophoretic analysis, per~ormed as outlined above. The gel was silver stained.

EGF ~adiorece~ptor Assay A431 membranes wer~ prepared according to t~e method of Xlmball and Warner~ A431 cells were disrupted under nitrogen and the nuclei and organelles pelle~ed by low speed cen~ri~Ugation. The ~embranes were then pelleted by cent~ifugation at 35,000 rpm for l hour and re~uspended in -20 mM HEPES b~ffer, p~ 7.4. Membranes (2.5 gt~l~ were -~.
:

, W091~18921 ~ 3 PCT/~S91/0~3 plated into ~6 well plates and allowed to dry overnight at 3,~C before use. Standard binding competition studies we~e performed using tl25I]EGF (ICN, Costa M~sa, California, specific activity- lOO ci/g, about 50,000 CPM/well). A
standard curve was constructed with 0.075-lOng of unlabelled hEGF (rece~tor grade, Collaborative Res~arch). Tha di~ferent fractions to be analyzed were lyophili ed and reconstituted in PBs (0.5 ~1/500 ml conditioned media).
After incubation of the labelled EGF and 10 1 of the ~a~ples for 2 hours at 37~ in binding bu~fer (IMEM containing 50 mM
HEPES a~d 0.1% BSA pH 7.7), the w~lls were washed, cut from the plate and counted. EGF comp~Eing activity, was computed using a Hewlett Pac~ard RIA Progr~m.

Anchoraae-indePendent Growth Assay lS Soft a~ar cloning assay~ were carried out using a l ~i botto~ layer of IMEM contalning 0.6% Bacto agar (Di~co, Detroit, MI), lOS FBS, and 2 mM glu~amine in 35~m tissue ~ishes (Costar, Ca~ridge, MA). A O.8 ~l top layer of IMEM
containing the test samples, 0.3~% agar, 10% FBS, and 3 X
104 NRX cell~ was added after solidi~ication of the bottom layer. Each sample was plated in triplicate. All samples were s~erilized by filtration using a 0.22 m Millex CU
millipora ~ilter before platiny. Platas were incubated in a humidi~i~d, 5~ C02 atmosphere at 37C and were counted aPter 25 12 days incubat~on with a Bausch and Lo~b S~em Cell Colony Count~r (Artex Sy~tem~ Corp, Farmingdale, NY).

Ancho~aae-deDen~ent Growth Assay CQ11~ w2re grown in IMEM containing 5% FCS. Upon .

. ~ -- . " -: .

.. :. ~. ~ - .;
.
~. . . . , . . ~

~091/1~9~ PCTt~S91/0~3 confluence cells were detached using trypsin-versene (Bi~fl~ids, Rockville MD) and ~a~sed at 1:20 to 1:50 dilutions. Cells were seeded in l2~well plates at - 4,000-lO,000 cells/well, depending on the cell type (MDA-MB-231~8,000 cells/well in serum free IMEM). After 24 hours the media was changed and the cells were treated with : EGF, TGF~ or ~G~a-like protein weEe harvested at l, 2 and 4 days using t~ypsin-versen~. The cells were coun~ed using a coulter counter.

Heparin Affinity Chromato~raphy Media conditioned by MDA-M~-231 cells were clarified by centrifugation for 20 minutes at 2,000 rp~ at 4C. The supernatant was collected and stored at -70C. After allowin~ the heparin-sepharose (Pharmacia, Piscataway, NJ) to expand in PBS, 2 ml of gel was loaded on an Econo column ~Biorad, ~ichmond; CA) and washad with abou~ lOO bead volumes of PBS. Conditioned media were run through the beads by gravity (flow rat~ 20 to 50 ~l/hr). The gel was : then wa~hed with 5 volumes of PBS and ~luted stepwise with an increasing qradient o~ NaCl in lO mM Tris-~Cl, p~ 7.0 (elution ~uffer). Gradient steps of 0.4 M, l.l M, 2.0 M and 3.0 M NaCl were used in the elution bu~er until the 280nm ; abqorption during each step returned to baselin~ (usually 3 to 5 column bed volumes). The eluate was d~salted on G-25 columns (Pharmaci , Piscataway, NJ) and filter-s~eri~ized : hePore use in the dif~erent ~ioa~says. Poaled fraction~
containin~ active materials were also de~alted on PDlO
columns tPharmacia, Piscataway, NJ) before running t~rough HPLC and FP~C.

.

, .
. ~ .

. , .
, WO 91/1~921 ~ `S'r~ 3 PCr/~S91/03443 Lyophilized conditioned medium was dissolved in 1 M
acetic acid to a final concentration of a}~out 25 mg/3IIl total protein. Insoluble ~naterial was removed by centrifugation at 10, 000 rpm for 15 minutes. The sample was then loaded o;~to a Sephadex G-100 cs:~lumn (XX 16, Pharmac:ia, Piscataway, NJ~, was e~uilibrated and wa~ subjected to elution with 1 M
acetic acid at 4 C with an upward ~low of 30 ml/hr . 100 ng of protein was processed from 4 ml of 100-fold concentrated medium. F~actions containing 3 ml of elu~te were lyophilized and resuspended in 300 1 PBS for assay and served as a source for further purification.

Reversed-phase Hiqh Pressure Liquid Chromatoqraphy (~PLC~

~teeD Ace~Q~i~ril~_Ç~adient; Step acetonitrile gradiant and all ot~er HPLC step~ were carried out at roo~
temperat~re after equilibration o~ the C3-Reversed phase column with o.oS% TFA (Trifluoroacetic acid) in water ~H~LC-grad~). The sampleq were loaded and fractions eluted with a linear gradient (0-45% acetonitrile in 0.05~ TFA) at a ~low rate of 1 ml/min over a 30 minute period, Absorbance was monitored at 280nm. One ml ~ractions were collected and lyophilized bQfore analysis for EGF receptor-competing activity.

~hallQw Ac~toni~ile ~radlen~ pool of active ~ractions ~o~ th~ pr~vious ~P~C step was rechomatographed over th~ saMe colu~n. Elution was performed with a 0-13~
ace~onitril~ gradient in 0.05S TF~ over a 5 minute period followed by a linear 18-45~ acetonitrile gradient in 0.05%

, . , . ~ , .
.; ~ ~, ~ , . -.
:

W091/18921 ~3'~ PCT/~S91/0~3 TFA over a 30 minute period. The flow rate was 1.0 ml/min and 1 ml fractions were collected. H~man TGF~-likP factor was eluted at a '0-32% acetonitrile concentration as a single peak detectable by RRA.

Electrophoretic Elution of Radiolabelled Protein from Gels After fluorography of an SDS-PAGE, bands of interest were excised and the protein eluted by electrophoresis into a di~lysis tubing over 16 hrs at 120 volts. The contents of the dialysis bag were cooled at 4 C and then precipitated by lo the addition o~ trichloroac~tic acid to a final concentration of 20~. The precipitates were pelleted by centrifugation, washed twice with ethyl ether, and resuspended in loading buffer.

Di~estion P~ocedure for Pu~lfied E~uted.Proteins Electroeluted proteins were dissolved at approximately 0.~ mg/ml in loading buffer which contained 0.125 M Tris-HCl (pH 6.8), 0.5~ SDS, 10% Glycerol and 0.001~ Bromophenol Blue. The samples were then heated at 100C for 5 minutes.
Proteolytic digestion were carried out at 37C for 30 minUtes by the addition of StaDhylococcus ~ure~s Protease V8 ~Sigma, St. Louis, M0) to a final concentration of 25 g/ml according to methods. P-mercaptoethanol and SDS were subsequently added to final concentrations of 20~ and 2%, respectivQly. Proteolysis was stopped by boiling for 2 min.
Th~ sample~ were then injec~ed on a C18 Reversed Phase HPLC
column.

~- Ph~sphoryl~h~n_~ 5he EGF ~eceDtor . . . - , :

, .

W091/1X921 ~ PCT/~S9l/~3 Subconfluent A431 c~lls were cultured i~ IMEM or 10-12 hours. The cells ~r~ trcated with 10-30 nM TGFQ, EGF or TFG~-like growth factor for 30 minutes at 37C. Cells were lyse~ in 20 mM Trls-HCl (pH 7.4), 150 ~M NaCl, 1% NP40, 1 S EDTA, 2 mM PMSF, 42 mM leupeptin and immunoprecipitated as described a~ove using monoclonal antibody 225 directed against the E~;F receptor Oncogene Science, Manhasset, NY)~
The i~munoprecipitates were washed three ti~nes with RIPA
bu~fer and resuspended in 40 1 TNE ~O.OlM Tris-HCl, pH 7.5, 0.15 M NaCl, 1 mM EDTA). Five Ci o~ [~-32P]ATP was added to the immunoprecipitates and the total ATP concentration was adjusted to lS mM (final) in a volu~e of 60 1. The reaction mixture was incubated for 5 minutes on ice before addition of 20 l of 3x sample bu~fer. The samples were boiled for 5 minutes and analyzed by denaturing 7 . 5~ SDS-PAGE.

RNA Extraction Total cellular RNA was extracted from ce~ls by ` homogeni2ing in quanidine isothiocyanate followed ~y ; c~ntrifugation over a cesium chloride cushion. Poly (A)+
~RNA was eluted in 10 ~M ~ris after passing total cellular RNA over an oligo (DT) cellulose column (Pharmacia, ~iscataway, NJ) equilibrated with 10 mM TrisoO.5 M NaCl pH
8Ø A~ter precipitation in ethanol ~66~ vol/vol~ and 0.1 M
ac~tic acid, both total and poly(A)~ selected RNA were resusp~nded in lO m Tris-l m~ EDTA buffer and separated on 1% agaros~, 6% formaldehyde gels. Electrophoresis was carried out at 20 volt~ over ~4-16 hours in: 5 mm NaAs 1 mM
EDTA, 20 mM 3-~N-morpholino~ propane sulfonic aeid pH 7.0 ~` ~MOPS-Sigma). The gels were stained with ethidium bro~ide 2.Og/ml to allow inspection of th~ quality and quantity o~
:; .

::' - ~ , -. . :, .
.. - , WO91/189~ PCT/~S91/0~3 RNA (). In vitro translation assays were perfor~ed using Weat Germ kit accor~lng to tha manufacturer's instructions (Promega).

Identification of a TGF~-like Poly~eptide in MDA MB 231 Human 3reast Cancer Cells To determine whether the 30 Kd TGF~-like pro~ein was recognized by antibodies developed against mature 6 kDa TFG~, MDA-MB-231 cells were metabolically labelled With ~35S] methlonine and [35S~ cysteine. M~tabolically labelled condikioned ~edia from MDA-MB-231, TGF~-transfected MCF-7 (H8), and HS578T cells were te~ted by solid phase RIA for i~unoreactivity with a polyclonal antibody (R399) and a monoclonal antibody raised against reco~binant 6 XDa TGF~.
Metabolically labelled TGF~-like mat~rial fro~ MDGA-mb-~31 cells reacted only with the polyclonal antibody. In contrast, the two anti~odies cro~s-reacted with meta~olically labelled material derived from H8 cells and no immune-reaction was noted with prei~mune serum (normal rabbit seru~ NRS) or metabolically labelled conditioned media ~rom H~5~Bt breast carcinosarco~a cells (Fig. l), which do not produce TGF~ M~N~. Thus only the monoclonal an~i~ody is able to distinguish bstween the di~ferent species of TGF~. Specificity of the assay was demonstrated ~ing a co~petition RIA with unla~elled recombinant TGF~.

Labelled material from MDG-MG-231, H8, ~nd ~at-FeSrY
c~lls was i~munoprecipitated with the an~i-TGF3 polyclonal an~body. D~ection of.an immunoreactive specie~ of app~oximately 30 kDa size veri~ied the secretion of a high molecular weight TGF~-like polypeptide in M~A-`~B-231 cells.

, W091/1892~ ,3~ PCT/~91/0 ~8 cells, which overexpress classical TGF~, yielded a 6 kDa product. The expected 18 kDa precursor of the classical 6 kDa TGF~ was precipitated fro~ Rat-FeSrV, which are ~nown to secrete the "normal" precursor. The intensi~y of the bands diminished when the immunoprecipitation was performed in the presencP of excess unlabelled TGF~. No specific bands were immunoprecipitated by preimmune rabbit serum.

The apparent heterogeneity in size o~ the larger TFG~
spç~ies and the potential ~or N-linked glycosylation of the TFG~ precursor at ~sn 25 led to the consid~ration of whether the high molecular weight TGF~-like polypeptide secreted from MDA-NB-231 cell~ was a glycosylated ~orm of TGF~. When MDA-MB-231 cells were incubated with tunicamycin, an i~i~itor o~ co-translatio~al N-linked glyco~ylation, and the media was immunoprecipitated with the anti-TGFc polyclonal antibody a species of 22 kDa substituted tha~
: , previously observed of 30 kDa. Additional cleavage of the 22 kDa polypeptide with elastase yi~lded an apparent ll kDa product different ~rom the matu~a 6 kDa TGFa that was ob~erved in Rat-FeSrV labelled ~edia. The ll kDa product h~d a higher immunoreactivity with the R399 antibody than the 30 kDa and the 22 kDa polypeptides. Shor~er exposure of th~ gel showed clearly a p~ecipitated ~and ~ear the ll kDa molecular weight. Tunicamycin treatment did not significantly affect the l2vels of se~reted TGF~ ac~ivity as determined by both RIA ~nd EGF receptor binding assays.

When the purified 30 ~Da polypeptide was treated with Nglycona~e a 22 kDa product was detected ~y silver staining.

WO91/18921 ~ `3 ~ PCT/~S91/0~3 The absence of cleavage of the purified 30 kDa polypeptide 2f~er O-glyconase treatment suggests that no O-glycosylatle~
occurs in this system.

Purification of the TGF~-like PolypeDtide TG~-like, material was isolated from s~rum free conditi~ned ~dia of MDA MB 231 cells. L~vels of T~Fu-like polypeptide were quantified by three inclependent assays:
capacity to indu~a anchorage-indapendent growth of NRK
fibroblasts in soft agar, ability to compete with (l2sI]EGF
for EGF receptor binding on A431 human carcino~a cell membranes and cross-reactivity with polyclonal antibodies raised against mature TGF~. EGF receptor binding activity and TGF~ i~unoreactivity were detected ~siny a R~A kit provided by Biotops. To determine th~ approximate molecular weight o~ the MDA-MB-231 deri~ed TGFQ-l~X~ polypeptide, ~ ~i o~ 100-~old concentrated, dialyzed conditioned mediu~ was chromatographed by g~l filtration usin~ Sephadex G-lO0.
Elution was performed with l.0 M acetic acid and ~ractions were characterized ~or protein content. TGF~-like activi~i~s were elut~d fram the column in a sin~le broad peak. Maximal activity was observed at an apparent molecular weight of 30 kDa and was separat~d from the bulk of contaminating proteins prasant in the bad volume. All t~e ~ractlons demonstrating TGF~ immunoreactivity also :~ 25 contained EGF receptor binding activity. The ralative a~ounts o~ rQceptor binding a~tivity and immunoreactivity present in these fractions, however appe~red to di~er.
Further analysis of the TGF~-~ike polypeptide from MD~-MB-231 cells was carried out using heparin-sepharose affini~y chromatography. Heparin-sepharos~ afPinity ,, . -. .
.~. . .
' : .

~091/18921 ~ PCT/~S91/0~3 chromatography was perf ormed on uAconcentrated conditioned medla from MDA-MB-231 cells. In all experiments, less t~an 20~ of the TGF~ activity loaded onto the column was recovered in the unabsorbed fractions. A sharp pea~ of EGF
receptor binding activity was eluted by heparin-sephar~s~
chromatography at a concentration of 0.4-0.5 ~ NaCl. This activity represented one major 30 k~a molecular weight protein, which retained 70%-80~ of the load activity.

The TGF~-like polypeptidP was further purif~ed by reversed phase chromatography ~HPLC) in two steps. A pool of fractions containing EGF receptor-competing activity from heparin-sepharose chromatography Was reconstituted in 0.05%
TFA in water and then chromatographed on a ~ondapak C3 column. A steep acetonitrile gradient (0-100%) was used in thiS step. TGFQ-like polypeptide elutes as a sharp peak in 30~ acetonitrile and is separated from the bulk ~f the contaminating proteins. The capacity of the individual fractions to compet~ for EGF receptor binding and to stimulate the growth ~f NRK cells in ~oft a~ar was determined. A pool of the active fractions ~indicated with an horizontal bar) was rechromatographed on the same column.
~rac~ions w~re eluted with a 0-20% acetonitrile gradient in O . 05% TFA for 5 minutes followed by a linear 20-40~
acetonitrile gradient. The TGF~-like polypeptid~ activity Was eluted at 25-30% acetonitrile and sffectively separa~ed ~rom oth~r contaminant proteins.

In order to achieve a complete separa~ion of TGF~-like polypeptide from those i~pUritie detected by sLlver staining (data no~ shown) we used size exclusion chromatography under acidic conditions. The ~ctive :, ~ - , . . .
: ' . - ' . W09~ll892l PCT/~Sgl/0~3 fractions for EGF receptor-competing activity were pooled and analyzed hy S~-PAGE. One single polypeptlde band was observed after silver staining.

A summary of the steps leading to the isolation and purifica~ion of T~F~olike polypeptide is presented in Table 1. A 27% recovery of activity and approximate 5400 fold purirication was achieved.

.

,~

' :

.. ~ ' ' ': , wo 91/18921 PCr/~S91/i)3~3 _ ~ _~

O ~ ~ ~ ~ ~

_~ C ~ _~ C
c~ o e ~ _~
_ ~ C~

, _ " :
. _ __, 1. C

~ ~ IJ ~ E ~ :~ c 3 e u c o-0 ~ :r e :
e c~ ~ ul 3 ~ c e ~ E . O c G

u ~ J C O ~ O, ~ ~ c~ ~ 0 0 ~ u ~
~. ~ c~ ~e~ ~c, c ~ O
.~ O O U C~O ~ ,~, _ o O U -- -- ~ U
U~ O--I U _ ~ L C
O 3 ~ E ~
E~ ~ o C C ~ O ~ ~
' U 3 C _ ~ ~ ~ C U C C
L 1: S' C U~ _ O ~ U 3 ~
O '3 --~ ~- I ~ ~ 1~.C~l O .CI --I
_ ~ _ L~ ~ ~
.. ~ O ~ 0 Ul L~
~.1 ~ " O X O
~U ~ _ I ~ E~ 0 0 v~
Q ~ ~ 5 ~ ~ I I I t~ y ~q C~ U~

, ~2r~
WO91/18921 PCT/~'S91/0~3 Bio1Oaical Characterization of the TGF~liXe ~aterial The EGF receptor binding aotivity of the 30 ~Da TGF~-like prot~in was compared with that of EGF in a radioreceptor assay. Both growth factors competed With [l25I]EGF or receptor sites on A431 membranes. The specific EGF-competing activity of the purified TGF~olike polypeptida was found to ~e 1-1.5 x lO6 units/~g; l.lng o~
TGF~-like polypeptide was r~quired to inhibit EGF binding by 50%. TGF~-like polypeptide was as e~fective as EGF in EGF
receptor bindin~. Furthermore, the purified 30 kDa TGF~-like polypeptide stimulated the growth of serum NRX
fibrobl~sts ~nd induced colsny formation of t~ese cells in so~ agar. The bioactivity of the purified T~F~-like polypeptide was also tested by anchorage-depend~nt growth assays of the carcinogen- immortalized huma~ ~ammary epithelial cell~ 184AlN4 and anchorage-independent growth assays of 184AlN4-derived cells partially transfor~d by SV~O T antig~n, 184AlN4T. Dose respo~se curve~ of TGF~-like polypeptide on these cells were similar to those observed with EGF and TGFU. The biological activity o~ the purified 30 kDa TGF~-like factor wa3 ~urther assessed ~y examinin~
its ability to induce autophosphorylation o~ the EGF
recep~or. A431 cells, which oversxpress th~ EGF receptor, w~re incubatQd with Yarious concentrations ~f EGF, TCFU or 25 TGF~-like growth factor~ Each of thQ three peptides s$milarly st$mulated phosphorylat$on of th~ EGF receptor.

Pep~ide Mappin~

In order to determine the degree of homology between the no~el 30 kDa TGF~-like growth factor and ~aturs TGF~, ., ~ .
' ' ' .

WO 91/18921 PCT/~S91/03~3 peptide lsLapping was performed using th~ method of Clevand.
Immur.opr~c~ L _tation of metabolically labelled conditi~ned media from MDA MB-231, H8, and ~at-FeSrV cells was carried out with the R3g9 anti-TCF~ polyclonal antibody.
Precipitate~ were analyzed on SDS-PAGE and the specific bands were electroeluted (~0 Kd from ~DA-MB-231 cells, 6 kDa from H8 cells, and 18 kDa from l:he Rat-F~SrV cells)~ These proteins were subjected to enzynatic treatment with N-glyoona~a and elastac~. The precipitated bands sizes are summarized in Table 2. The products were then su~jected to a peptide~~igestion using 25 g/ml V8-protease. After complete digestion, the samples were analyzed by Cl8-reversed phase chromatography. Three major peptide peaXs eluted at different acetonitrile concentrations by reversed phase chromatography. However, the concentrations at which thos2 peptides isolated from MDA-MB-231 cells eluted (16%, : 18.7~, and 21.~%) were di~ferent fro~ the peptides isolated from H8 and FeSrV cells (24~, 29~, and 32.6%). T~e peptide elution pattern of the TGF~ (6 Kd) d~rived from H8 cells and Rat-FeSrV cells was essential identical. Th~ same results were obtained with 40 g of V8 protease, indicating that concentration oi~ the enzyme wa~ not responsible for t~se differential peptide cleava~e. Moreover, in ~itro translation o~ mRNA d2ri~ed from MDA-MB-231 ~ells and X8 c~lls wa~ dona, the resulting polypeptide had the same pep~ mappi~g profile than th~ purified 30 kDa fac~or arter tr~tment with N-glyconase and elastase. These results p~ovide evidence that a precursor different than the l'normal" TGF~ precursor is translated from the mRNA of MDA-MB-231 cells. Moreover, the above results indicate that : th~ MDA-MB-231 derived TGF~-like polypeptide shares Yery raw, i~ any, common peptide sequences ~ith mature ~GF~.

, - .

9~ ~;J"~ P~T/~S91/0~3 In order ~o characterize the cellular effects of the present 30 Xd glycoprotein, various experimen~s were conducted. The following Examples are provided solely for the purpose of illustra~ing the present inventi~n and are not intended to be limitative.

Exam~le 1 In order to characterize t~e cellular ef~ects o~ the present 30 Xd glycoprote~n ligand, its induc~ion of tyrosine phosphorylation was assessed in t~e hu~an breast cancer lines MDA-468 and SK-Br~3. Notably, MDA-46 8 cells ~ave amplification ~nd over expression of th~ EGFR gsne and do not expresi erbB-2 xeceptor-like protein. SK-Br-3 cells have a~plification and ove~ expression of the erhB-2 gene as well as ~elatively elevated levels o~ EGFR. T~e 30 Xd liqand, 15 TGF~ and EGF were found to induce tyrosine phosphorylation in both cell lines an EGFR blocking antibody abolished the phosphorylation induced by the three growth ~actors in ~DA-468 cells. This antibody did not, however, completely block the phoiphorylation induced by the present 30 Kd ligand and SK-Br-3 cells. ~owever, it did block the phosphorylation induced by ~F~.

From th~ abov~ result, it appears t~at tyrosine phosphorylation of a protein is di~ferent from EGFR occurs in sX-ar-3 cells treated with 30 Xd fac~or. No phosphorylation was obserYed in untreated SK-Br-3 cell~, and cells ~reated with the anti-EGF~ antibody alone.
: ~ .
.~ Exam~

.

:: :

S~?D ~ 3 ~
W091/189~1 PCT/~S91/0~3 In human mammary carcinoma cell line MDA-453, which over expresses er~B-2, but which has undetecta~le levels of the EGF receptor protein or mRNA, the 30 Kd ligand -~as observed to induce a sisnificant increase in ~yrosine S phosphorylation in a dose dependent manner at concentrations ranging fro~ 1.25 mg/ml to SO mg/~l. By contrast, EGF and TG~U were unable to induce tyrosine phosphorylation in the 185 kDa range, at a concentration of 25 mg/ml. No phosphorylation was observed in untrea~ed cells. Hence, o from the above, a direct interac~ion between the 30 Rd ligand and the 185 K~ glycopro~ein appears to occur.

Exam~le 3 In order to determine t~e effects of the present 30 Kd ligand on the proliferation in colony ~ormation of breast carcinoma cell lines, the followinq experiment wa~
conducted.

CQ11S were treated with the present 30 Kd growth ~actor, EGF, TGF~ and anti-QrbB-2 antibody in order to inhibit t~a proliferation o~ S~-Br-3 c~lls.

It was observed that the ~nti-erbB-2 antibody inhibited the proliferation of the SK-Br-3 and MDA~453 calls by 60-70%
bu~ did not inhibit the proliferation of MDA-468 cells.
Surprisingly, by exposing SK-Br-3, MDA-453 and ~DA-468 cells to the 30 Kd ligand protein of the present invention, a 60 70% inhibition of cell gro~th was observed ~or all cell lines.

Inhibition o~ growth by the 30 Kd ligand pro~ein was ' ~ ~

-:

J r \ (_~ ~
WO91/18921 PCT/~S91/0~3 reversed by an EGFR blocking antibody in ~DA-468 cells, but not in SX-~r-3 or ~A-4~3 cells. This is an indica~ion that the effects of Ihe 30 Xd protein on SK-Br~3 and MDA-453 cells are not mediated through EGFR.

By contrast, the present 30 ~d glycoprotein exhibited no effect on MCF-7 cells, which have nor~al levels of EGFR
and ~rbB-2. Additionally, EGF and TGF~ inhibi~ed the anchorag dependant growth of MDA-468 cells and SX-Br-3 cells, but not that of MDA-453 or MCF 7 oells. EGF-induced anchorage dependant growth inhibition of SK~Br-3 and MDA-468 cells was reversed by an anti-EGFR blocking anti~ody. In the presence of the 30 Kd glycoprotein, the growth inhibition of SR-Br-3, ~DA-453 and MDA-468 cell~ was neaxly complete.

The growth inh~bitory property of the pres~nt 3 0 Kd i ligand appears to be similar to that described ~or EGF on ; size whichever express EGFR such as A431 cells and MDA-46 . cells.

; Exampl9_4 Further, t~e growth of CHO/erb~-2 trans~ected cells was inhibi~ed by 70-80~ after treatment with th~ present 30 Kd . glycopro~Qin. No e~ct was obser~ed on thQ CH0/~HFR
control trans~ectants and the parenteral CHO line. TGF~ at the qame molar concentration did not exhibit any effect on th~ proliferat$on of any of the three lines. Tyrosin~
; phosphorylat~on and cel.l proliferation of thQ CHOJDHFR cells and the parenteral C~0 cell line is not ~ ected after tr~at~ente by t~e present 30 Xd ligand or TGF~.

-~ , . . .

... .
, s ~ 3~3 WO91/18~2~ PCT/~S91/0~w3 -2~-~xam~le 5: Cell Growth Inhibitlon by gp30 .Sk-Br-3, MDA 453, MDA 468 and MCF-7 cells were plated in 24 well plates in I~EM (Bio~luids) supplemented with 5%
FCS. Parental CHO cells, and CHO cells trans~ected With the 5 DHFR gene or the erbB-2 gene were plated in ~4 well plates (Costar) in ~-~EM (Biofluids) supplemented by 10% dialyzed FCS, 0.75 mg/ml G418 and Methotrexate (M~X) 50 n~ for the CH0 parental and CH0-DHFR CELLS for 259 nM ~or the C~0-erbB-2. After 24 hours media was removed and replaced with control serum free media (SFM) containing fibronectin, transferrin, hepes, glutamine, trace elements, and BSA, or SFM with the addition of 2.0 ng/ml gp30, lO ng/ml recombinant TGF~ (Genetech), or with 2.5 MgJ~l 4DS spacific anti-pl85~rb8 2 monoclonal antibody. Cells were grown in lS 90% confluence of control and counted. Each Group was assayed in triplicate. Resul~s are shown as growth relative to con'crol. The experiments were performed thre~ times and the results were reproducible. The results are shown ~elow in Table 2.
,~ ~
. _ _ _ _ _ _ _ SK-Br-3 _MDA153 CHO/crbB 2 C~O~HFR MDA~ M~
t ~30 3l 2~ 20 ~ la l~
4DS~nd~ 32 34 22 98 l~ ~
MF~ 73 9l 89 9S 7~ IOS
Con~l~ ~y __ _87 9l 87 ~ _9~ _ ~

In order to further d~scribe the various aspects of t~e present invention, reference will now be made to the ~igures of th~ present specification.

~,; F~qure 1: lsolatlon of qp30 . :
, . .
- . . , .~.......... . .
~ . ~ ~ , . . .
. , W091/18921 ~r~ PCT/~'S91/0~3 Part A illustrates the use of low affinity heparin chroma~ography. In particular, affinity chromatography of conditioned media ~rom MDA~231 cells was performed on a heparin sepharose column. Fractions were analy~ed for EGF
receptor binding ac ivity of A431 cell membranes. Ali~uots from the input media and ~rom the fractions containing activity were analyzed by a 15% SDS~PAGE, followed by silver staining. Lane 1 shows unconc~ntrated conditioned media.
Lane 2 r~p~sents the actiYe fraction.
Part B illustrates the use of reversed-phase chromatography. Notably, the EGF/TFGA active fractions obtained after heparin-sepharose chromatography were chromatographed twice on a ~Bondapak C3 column in O.O~S TFA.
Samples were eluted with a steep gradient of ac~tonitrile.
Fractions ~hat showed EGF r~ceptor binding activity were ~hen rechromatographed and e}uted wi~b a shallow acetonitrile gradient. EGF comp~eting activity waC
constantly eluted at a 25 30T acetonitrile gradient. The resulting fraction was analyzed on a 15% SDS-PAGE followed by silver staining. Sizes are shown in kilodaltons.

Fiqure 2: ~etection of Phos~ho~Ylated Proteins Ln-~K-ar-3 Cells SK-~3r-3 cells were grown in 90~ con~luence in 24-2311 platss (Costar). Cells were treated at 30C with I~E~
~lanes 1 and 2), I~EM containing 25 nb/ml recombinant TGF~
(Genetech, CA) (lanes 3 and 4), and I~EM contain~ng 5 ng/ml of gp30 (laneB S and 6~, all o~ these in tAe pres nce (lanes 1, 4, 5) and khe absenc~ ~lanes 2, 3, 6) Of an anti~EGF
r~ceptor blocking an~ibody (G~netech, CA). After 20 ~inutes the ~edi~ was removed and cell5 were lysed in 100 ~1 of :

, .

~r ~ $~
~091~18921 ~ pcT/~ssl/o~3 ~30 sample buffer containing 1% SDS, O.l~ ~mercaptoethanol, 0015 M Tr~s~ h 6.8), lO~ ~lycerol, 0.02 ~
bromophenol blue, lMm EDTA, 2 Mm Pmsf and 42 Mm leupeptin.
After 5 minutes at 95C, 50 ~ of protein were loaded in a 7.5% SDS-PAGE. Proteins were then transferred to nitr~cellulose membrane for immunoblotting (Hoefer Scientific Instruments, California) by electrophoresis in a ~odified method of Towbin et al, using a electrophoretic transfer unit (Hoefer, TE 22). Electrophoretic trans~er was carried out at room tempsrature for one hour at 125 Ma in a buffer containing 25 Mm glycine, 129 Mm Tris (Ph 8.3) and 20% methanol. Followlng transfer, the filter was blocked with 5~ BSA in Tris-Buffered Saline containing 0.5% Tween 20. An antiphosphotryosine antibody (Amersham) was reactPd with ~he immobilized proteins in 5% BSA ~Sigma RIA Grade).
Immunecomplexes were detected by a ~oat anti-mou~e antibody conjugated to al~aline phosphata~e. Blots were then incubated with a color dev~lopment substrate solution containing N~T and BCIP (Promega).

Figure 3: Detection ~f PhQs~horYlated Proteins i n MDA-453 Cells ~`
MDA-45e cell~ were grown to 90~ confluence in 24-2311 plates (Costar) and treated at 37C with IME~ ~lane l), IMEM
containing 25 ng/ml of recombinant TGF~ (Genetech, CA) (lane 25 lO)~ or I~EM containing 1.25-40 ng/ml of gp30 ~lanes 2-9).
After 20 ~inute~ media was remo~ed` and c~115 were lysed in 100 ~1 of sample buffer as described in Figure 2. After 5 minutes at s50C, 50 ~g of protein was loaded in a 7.5~
SDS~AGE. Proteins were then transferred to nitrocellulose ~embrane for immunoblotting with an antiphosphotryosine : .
. .
, ~ ~, i, . .. .,.-:.,-.

WO 91/1~92I PCT/~S91/03~3 antlbody (Amersham) as described in Figure 2.

Fiqure 40 P~o sphorYlation o~ P185 Protein in IntactCHo/DHFR, and C~O/erb~-2 Cells Cells were grown to 90% confluence in 24-2311 plates (Costar) in A~EM (Biofluids~ supplemented with lOS dialyzed FCS, 0.75 mg/ml G418, and Methotrexate (~rX) at concentra~ions o~ 50nM (CHO parental and C~O-DHFR) OR 250 Nm (CHO-er~B-2~. CXO-DHFR (Fig. 4A) and CHO-erbB-2 ~Fig. 4B) cells, were treated at 37C with control media supplemented with 20 M~ Hepes (Ph 7.4) ~A and B lanes 1 and 4), With lO
ng/ml of recombinant TGFa (Genetech, CA) (A and B lanes 2 and 5), and control media supplemented ~.0 ng/ml of gp30 ~A
and B lanes 3 and 6). After 20 minutes, media was removed and cells were lysed in 100 gl of sampla buffer (as d~cr~bed in Figure 2). An anti-phosphotyrosine antioody (A
and a lane~ 1 to 3) [Amersham) and an anti-erbB-2 antibody (A and B lanes 4 to 6) (NEN), were reacted with tha ` im~obil~zed proteins in 5~ BSA (Sigma RIA Grade~. -Im~unocomplexes were detected as ~e~cribed for Fiqure 2.

Cell~

SK-Br-3 calls were plated in 24 wç~l plates in IMEM
(E~iofluid~) supplemented with 5% FC5. Ai~er a wash with binding bu~er tD~/Fl2 pH 7 . 4, containing 1 mg/ml ssA, lO
- 25 M~ hepes and 20 ~ glutamine) cells were incubat~d for ~0 minutes at 37C with binding buffer. The EGF~R were satu~a~ed with 30 nM EGF for 2 hours at 4 ~c. plas binding *l studi~s were t~en perfo~ed for 3 hours at 40c w1th 1 nM

.
.

~'091/18~21 ~ ~ PCT/~S91/~W3 iodinated 4D5 in the presence of various c~ncentrations of unlabeled gp30 ~or 4D5. After the incubation, cells w~_~
~ashed 3 times with binding buffer and then solubilized with 1% SDS. No specific binding was det~rmined with excess (100 : S nM) of unlabeled antibody. Each ~roup was assayed in triplicate. The experiments were p~rformed f ive times and th~ resul s were reproducible.

Fiqure 6: Inhibition_o~ Pl85-cross-linkinq with 4D5 The binding assays were performed as described in Figure 5. Binding was performed with iodinated 4D5 ~1 n~) alone (lane 1), in the presence o~ 100 nFM unlabeled 4D5 (lane 2) and in t~e presence of 2nM gp30 (lane 3). 100 nM
EG.F were used as a control (lane 4). Cell~ were then i~ treated with a cros5-linking agent EGS for 45 minutes at 4C, then quenched ~y adding 0.1 ml of 20 Mm NH4Cl. The solubilized c811s wer~ immunoprecipitated with a polyclonal antibody to the C-ter~inal domain o~ erbB 22 ~Genetech, CA).
The precipitate~ were analyzed on a 5% SDS-PAGE.

The 30 Xd gly~oprotein cf th~ present inven~ion may thus be used advantageously to in~ibit the growth o~ ~arious : types of ad~noca~cinoma cells which overexpres~ the erb3-2 oncogen~ and EGFR. Preferably, the present 30 Kd glycoprotein i~ used in inhibit the growth of ~denocarcinoma call~ of breast, ovarian, gastric and lun~ tissue which . overexpress ~he erbB-2 oncogene and EGFR.

In ucing the present 30 Xd ~lycoprotein to inhibit the growth o~ the above malignant cells in a ma~mal, pre~erably :

.

.- -, - - -:

r ~s~
~091/18921 PCT/~91~0~3 a human, relatlvely low concentrations of th~ glycoprotein ma-~ ~G ~s~d. For example, an aqueous solutio~ having a concentration of about 1-50 ng/ml may be convenie~tly ad~inistered to a patie~t such that a total of from about -l-lO,000 ng of glycoprotein are administered per day. It is preferred, howevar, if about l-l,000 ng are administered per day.

~ he present invention thus relates to the use of the present 30 Xd TGF~-like glycoprotein in direct interactions With EGFR and pl85erb8-2o Hence, in another aspect, the present invention provides conjugates of the ~0 Kd glycoprotein ligand with either EGFR or pl85~8-2. In still another aspect, the present invention provides diagnostic and therapeutic methods using t~ese conjugates. Further, lS the present invent~on provides a diagnostic test ~it using the present con~ugates.

In another aspect, the present invention relates to the preparation of monoclonal antibodies of gp3O, and the use of these monoclonal antibodies to detect the presence of gp30 in patient sera.
As gp30 is known to be produced by MDA-MB-231 breast cancer cells, and is also lik~ly to be produced by other adenocarcino~a cancer cells, the presen~ invention also provides a method for detecting gp30 in patient sera.
Generally, in accordance with the present invention, t~e mer~ d~tectio~ of eit~er pl8s or gp30 is a ~asis for concluding that the detected pro~ein is being overexpressed.
This conclusion, in turn, leads to a poor patient prognosis necessitating the use of more aggressive treatment of the tumor.

.~ ~

W091/t~921 P~T/~S91/0~W3 . -34-In more detail, the present invention specifically first contemplateS the use of con~sate, ~' the 30 Kd glycoprotein and EGFR, and of the 30 Kd glycoprotein and pl85e~b8-2 in detecting the presence of adenocarcinoma cells which overexpress either EGFR or erbB-2 oncogene.
Preferably, the adenocarcinoma cells dete ted are of breast, ovarian, gastric and lung tissue.
Generally, the pr~sent conjugates may be used advantageously in a biochemical detPction method in which the 30 Kd glycoprotein ligand is bound to a surface and put into contact with aqueous solution containing a tumor portion containing cells whi~h are suspected o~
overexpressing either E~FR or erbB-2 oncogene. This is conveniently done as either EGFR or pl85 may be found on the cell surfaces. If such calls are present, eit~er t~e EGFR
or pl85erb~-2 will become bound to the ligand. Iherea~ter, the aqueous solution is separated ~ro~ t~e bound antiligand material, a~d the antiligand material may ~e conveniently detected with a known detection mean5 associatad therewith.
For example, an ampli~ied enzymelinked immunoa~say may be used. ~he sur~ace to which the li~and i8 bound is treated with one or more agents ~or limiting the amoun~ of non-speci~ic binding. Such agents reduce the "noise"
arising due to non-specific binding when int~rpreting the as ay.

In acco~dance with the abov~ procedure, a diagnostic t2st kit may be constructed in a variety of ways.
.

For example, a test kit may be constructed to contain a : vessel con~aining a test liquid having a surface to which gp30 llgand is bound. This is preferably a multi-well test ~' , ~`

: :

, - W091/1~921 PCT/~S9~/0~3 3~-plate. Also contained is at least one other vessel containing reagent sclut~ he agQnt for limitlng non-specific binding may be incorporatPd wil:hin a solution of the kit or may have been used to treat the surface of the first vessel before it is supplied.

Then, a portion of the tumor or a tumor sample may be worked up into an aqueous solution and put into contact wit~
: the bound gp30.

In order to conveniently detect the overexpression of EGFR or erbB-2 oncogene in a human patient it is advantageous to use the well-known sandwich assay tech~ique.

For example, one assay method and test kit which may be used in accordance with the present invention are described in U.S. Patent 4,668,639 which is incorporated herein in the entirety.

Henc~, th~ presant invention contemplates and is specificalay ~i~ected to any diagnos~ic o~ therapeutic .. method ~ar the-~detect~on of adenocarcinoma cells whic~
`!
overexpre~S EGFR or erbB-2 oncogene, whic~ method uses the formation o~ a conjugate between the 30 Xd glycoprotein of the present inYention and either EGFR or pl85~r~-2.

AQ noted above, the present invention also provides an assay and a test kit for the detection of gp30 using monoclonal antibodies to gp30.
.~
It is noted that although either polyclonal or monoclonal antibodies can be used for this purpose, it is :`:
~ .
:

. . .

W091/18921 ~ PCT/~S91/03 preferred that monoclonal antibodies be used.

In such an assay, the monoclonal antibodies to gp~0 are preferably bound to the microtiter or ~ulti-well plate and exposed to patient sera suspected of containing gp30.
Upon de~ecting the presence of gp30 by a conventional de~ecting means, a conclusion o~ poor prognosis would be mad~ necessitating the use o~ more aggressive treatment for the tumor. IDportantly, however, the presence oS the 30 Rd glycoprotein (gp30) in patient sera can ~e detected utilizing either monoclonal or polyclonal antibodies in virtually any type of immunoassay. This includes both single-site or twosite or "sandwich" assays of the non-competitive types, as well as in traditional competiti~e binding assays.

With the above assay, a test kit is also provided.
Generally, the kit contains a first container containing an antibody having specificity for gp 30 and a seoond ~ontainer containinq a s~cond antibody having specificity for gp30 and : being labelled with a reporter molecule capabl~ giving a detec~ahle si~nal. The first antibody i9 immobilized on a solid surface.

; The above assay and test kit for the detection of gp30 may b~, re3pectively, conducted and constructed by analogy in accordance with U.S. Patent 4,92l,790, which is in~orporated herein in the entirety.

The ~0 Kd glycoprotein of the present invention is wellcharacterized by:

' ~ " , : ' .

3$

.W09~/18921 P~Ti~S9t/0~3 1) being a heparin-bInding protein;

2 ) being capable of binding to EGF reoeptor;

3) exhibiting cross-reactivity to antibodies to TGF~;

4~ being capable of cleavage by elastase; and 5) being capable of stimulating transforming activity in normal rat kidney (NRK) cells.

T~ polyclonal or monoclonal antibo~ies produced against gp30 may be produced in accordance with well-known techniques. For example, see current P~otocols in ~olecular 10 3ioloa~, edited by F.M. Ausubel et al (Wiley 1987), in particular Chapter 11 on Immunology. Also, the i~munoassays used in the assays and diagnostic test.kit~ of the pr-sent invention are well ~nown to the artisan as evidenced by the above treatise, and by thQ methods disclosed in U.S. Patent 15 4,g21790 which patent has been specifically incorpurated herein in the entirety.
., .
As descri~ed above, the diagnostic aspects of the pressnt invention relate to the use of methods and test Xits for the detection of either pl85, EGFR or gp30. T~e . 20 de~ection of any one of these proteins may form the basis :~ ~or a poor prognosis necessitating the use of aggressive treatment of one or more adenocarcinomas.

~e present invention also relates to gp30, itself, and .~ conjugates of gp30-EGF~ and/or gp30-pl85e~b3~2.
.~ .
: 25 :''' ~,''`

~.
~; :

,:

~ WO91/1~92~ s~ ~ ~ PCT/~591/0~3 The therapeutic aspects of the present inv~ntion relate ~O ~.le ~ls2 of gp30 to inhibit the growth of adenocarcinom~
cells which overexpress EGFR and/or erbB-2 oncogene.

Generally, the amount of gp30 to be administered as a th~rapeutic agent will bê determinêd on a case by case basis by the attending physician. As a guideline, the extent of the adenocarcinoma, body weight and age o~ th~ patient are considered while up to about 10,000 ng per day may be used, generally not ~ore than 1,000 ng per day of gp30 is administered~ It is preferred, however, if ~rom about 5-500 ng per day are used. Notably, however, the above amounts may vary on a case-by-case basis.

While the present 30 Xd glyccoprotein may be administered by itself, as a therapeutic agent, it ~ay be administered in comb~nation wit~ one or more other therapeuti~ aqents. For example, the 30 Kd qlycoprotein ~ay administered with any chemotherapeutic substance, growth inhibitor or immune~timulating substance. The present invention specifioally contemplates such combinations.
.
: 20 Havinq now descri~ed the invention, it will now be apparent to one of ordinary skill in the art that many ch~nge and mod~fications can be made to the above em~odiments without departing ~rom t~e scope and spirit of the present invention.

;

,

Claims

1. A method of inhibiting the growth of adenocarcinoma cells in a human, which cells overexpress the oncogene erbB-2 or EGFR, which comprises administering to said human an amount of a 30 Kd glycoprotein effective to inhibit the growth of said cells.

2. The method of Claim 1, wherein said adenocarcinoma cells are adenocarcinoma cells of breast, ovarian, gastric or lung tissue.

3. The method of Claim 1, which comprises administering about 1-10,000 ng Of 30 Kd glycoprotein per day to said human.

4. A 30 Kd glycoprotein obtained from MDA-MB-231 human breast cancer cells.

5. A conjugate of a 30 Kd glycoprotein obtained from MDA-MB-231 human breast cancer cells and a 185 Kd glycoprotein expressed by human erbB-2 oncogene.

6. A conjugate of a 30 Kd glycoprotein obtained from MDA-MB-231 human breast cancer cells and human EGFR.

7. A method of detecting a 185 Kd glycoprotein expressed by human erbB-2 oncogene or EGFR in human sera, which comprises:

a) contacting a solution of a tumor portion from a patient suspected of containing said 185 Kd glycoprotein or EGFR with a bound 30 Kd glycoprotein obtained from MDA-MB-231 human breast cancer cells, thereby forming a conjugate of said 185 Kd glycoprotein or EGFR and said bound 30 Kd glycoprotein, and b) detecting or attempting to detect said formed conjugate by detecting means.