CA2102129C - Growth hormone antagonists - Google Patents

Abstract

The present invention relates to growth inhibitory antagonists of bovine growth hormone obtained by mutation of the third alpha helix of that protein. These novel hormones may be administered exogenously to animals, or transgenic animals may be made that express the antagonist and thereby exhibited a reduced growth phenotype.

Description

'I~VO 92!19736 P~.'flUS92103532 C~~tO~GY~OI~ ~TPTISTS
This application is a continuation-in-part of USSN
07/693, 305, filed May 1, 1991, now pending, which is a continuation-in-part of PCT/US90/05874, filed October 12, 1990, which is a continuation-in-part of USSN 07/'419,51, filed October 12, 1989, all of which axe hereby incorporated by reference herein.
BA.~CGR~'~ ~P' TIt~DT
Field of_ the Invention This xn~srention relates to novel growth hormone, especially bovine growth hormone, muteins which inhibit the growth of animals or otherwise antagonize the effects of endogenous gr~wth hormone. These analogues znay be expressed in transgenic animals which thereby acquire a '°reduced growth"
phenotype .
Information Disclosure Statement ~o~rine growth hormone (GH) is a protein of 192 amino ~eids that is naturally synthesized in the anterior pittaitary.
Tl~e molecular weight of the mature protein is about' 22, 000 d~ltons, but it as init3:ally m~,d.e a~ .a pre-growth horcmone ~rith axa extra 26 a:~~.no acids ~n the amino t~rnninal. Tha.e leader (or signal pept~:de) is normally cleaved, during secretion of the hormone by bovine pituita~,r c~ll~Several forms of the mature protean have been f ound- in nat~.re . The N- teraninal can ~ra,ry (due to variation in the s~:~e of clea~rage during secretion) so that the matureprotein begins with e~-then NH2-Ala-Phe-Pxo or l~z -Phe-Pro. Additi~nal~.y, the amino arid at posit3.on 126 may be either le~aciaae or valise; agParently a.s a result of allel~.c ~r~.~iatior~ in the bovirae gaoPu~.atioa~.
Exogeno~zs administration'of bGH to cattle increases milk production, feed efficieaacy, = growth rate, and the lean-to-fat ratio, and decreases fattening time.

''V0 92/19735 P(_'fJUS921~3532 ~~..~'z~2~
bGH has been produced by recombinant DNA techniques, see e.g., Fraser, U.S. 4',443,539 (yeast); Buell, EP Appl..
203,395 (bacteria); Krivl, EP Appl. 193,515 (bacteria);, Kopchicl~, EP Appl. 161,640 (encapsulated mouse cells implanted, into animals); DeHoer, EP Appl. 75,444 (bacteria; gene modified to eliminate harmful secondary structure) and this has facilitated the production of analogues of bGH by site-specific mutagenesis. Thus, Aviv, GB 2,073,245 describes production of Met Pro (des A1a) bGH, Met Arg (des Ala) bGH, Met-Glu-Gly (des Ala) bGH, and des (Alas Phe2 - Pro3 -Ala4 ) bGH in E . r~ol i . Brews , et al., PTMAS (USA) 85:336?-?1 (1988) reported preparation of the bGH mutant IC112L, which extended the hydraplnobic face of the third alpha helix of bGH. The 96-133 fragment of this mutant was also prepared.
The biological activity of proteolytic fragments of bGH has als~ been ~tudied. Brews, et al., Biochemistry, 26:???4 (198?)~ Sw~.sIOC~,i, et al., Endocrinology, x:900 (1970) ; Faladi,ni; et ~.~,. , TIBS; X56 (r3ov. 19?9) . The fragment of bGH containing ~m~.no acids 96-133 is superior in growth pr~moti~ag assays to bGH 1-95 and bGH 151-191. Hara, et al., Biochemistry, 1~.'7:5~0 (19?8) ; S~nenberg, 'U.S. Pat:~nt Nos.

3;664,925 and 4r~5,520; then arad SOnenb~rg,-J. Biol. Chew., 250:2510-14 (19??), An o~tag~ptide der~.ved from the amino-te~;l of 1'.~H has ~e~en shown to have' hypoglycemic activity, see Ng, et al., Diabetes, 23:943-949 (19?4), but it has no ef'~eC~ on growth: Similar re~u7s ~e~,e observed with the fragment bGH (96-133). Graf; et al., Eur: J. Biochem., 64:333-340 (19°6); Hara, et al., Biochem., 1?:550-56 (1978).
Anal~gues of bGH have varied in growth-promoting.
activity, as have the known analogues of other growth hormones.
However, a growth hormone analogue having growth-inhibitory.
activity hay not previously been reported.

93~~ 9xt19736 PCTt~J59Zt~353~

A variety of transgenic animals have been produced.
Haarnmer, et al., Nature, 315:080-683 (1985) (rabbits, sheep and pigs). Certain of these animals have been caused to express a growth hormone, and increased growth of such transgenic animals has been reported. Palmiter, et al., Nature X00_:611 (1982) microinjeeted the male pronueleus of fertilized mouse eggs with a DNA fragment containing the promoter of the mouse metallothionein-I gene fused to the structural gene of rat growth hormone. Several of the transgenic mice developed from the genetically moda.fied zygote exhibited a growth rate substantially higher than that of control mice. (Zn effect, the gehetically modified mouse serves as a test environment for determining the effect of the hormone on animal growth).
Later, Palm~.ter, et al., Science, 2220809 (1983) demonstrated that a similar enhancement of growth could be obtained in tr~.nsgenic mice bearing an expressible human growth hormone gene. A lake effect is observed when human growth hormone releasing factor is'expre~sed in transgenic mice. Hammer, 3t a~.r nature, 315:413 (1985).
8~vine growth hormone i~as also been expressed in transgenic aninnals. ~cGrane, et al. J. viol. Chem., 2C3~:1144351 (1988),- ~opch~.ck, et al., brazil. J. Genetics, 12:3?--54 (1989): However; transg~nic animals characterized by an exogenous gene which confers a reduced growth phenotype were hitherto unkyaown.
y ~F ~1f The present invention,relates to proteins which are ~~stantially homologous with a vertebrate growth hormone brat h,~ve growth-inhibitory activit~r.
We h~.ve disc~vered that mutation of Gly~ 1 ~ in bGH to erg ( "G119~2" ) , P'r0 c "G11~P" ) , I°ys ( "G1~:9R" ~ , Trp ( "G119W"
) Or ~,eu ( "G11~L°° ) , or the homologous ~lyl a o in hGH to 1-erg or Trp, WO 92/I9736 P~'/1.1~92/03532 results in a mutein~r. (mutant protein or peptide fragment thereof) which has growth-inhibitory activity in vertebrates, especially mammals. This novel hormone may be administered to mammals (or other vertebrates), in particular humans and bovines, when growth inhibition is desirable.
2n one embodiment of the invention, the hormone is produced e~cogenously and administered to the subject. In view of the sixe of the hormone, it is preferably produced by expression in a suitable host of a gene coding for it. Such a gene is most readily prepared by site-specific mutagenesis of a bGH gene. However, the hormone may also be produced ,by other techniques, such as by condensation of fragments of native bGH
with a syhthetic peptide carrying the replacement amino acid, 7Cf a peptide fragment has the desired growth-inhibitory a~ti~rity, it may be prepared in toto by a Merrifield-type synthesis.
Tn a second embodiment of the invention, this gene is introduced ~.n.tr~ a prenatal foam of a ar~amma7. by known techniques, and the prenatal fe~rm is developed into a transgen~.c mammal which ea~pres~e~ a reduced growth phenotype .
Corzceivably, .a anamsnal c~uld be genetically rctodifi~~t after b1rth, i.e., negene therapy". -Thus, growth-~:nhibited ~namals may be produced either by adz~~.nastration of the gx~wtJh' ix~ibitory hormone of this inven~a.on in pharmaceutical fdr~n; or ~Y 9Eneta:c transformation of a prenatal or postnatal. foran of ' the anammal a The growth-inhibitory hormone, or tYae gene encodzng i~' i~ useful in the production Qf sanall animals f or use in research facilities where space is restricted, as pets for pet lovers with limited barters; and as livestock for farmers having small tracts. The hormone may also be useful in the treatment of human gigantism; and in research on gigantism and WO 9211976 ~ ~ ~ ~ P~ I'/ZJS92/~3532 -dwarfism, in the treatment of diabetes and its sequelae, in the control of cholesterol, and in the prevention and treatment of certain cancers.
Character~sta.cally, patients with poorly controlled diabetes have been found to have high levels of circulating growth hormone. See, e.g., Lundbaek, et al., Lancet, 2:13-83 (~.g70). It has been speculated that high levels of growth hormone may contribute to poor diabetic contral, as opposed, to being merely a caneequence thereof. Press, et al., New England L~fed. ; X10: 810-14 (3984) . d~ttempts have been made to inhibit ~.rowth hormone release by means of somatostatin analogues .
However, use of growth hormone antagonists has not been reported previously. Thus, a further aspect of the present invention is the use of the disclosed G~3 axxtagonists to improve diabetic control.
.bong the complications of diabetes are retinopathy, n.~phr~pathy a~ad ahg'z;opathy. Diabetic retinopathy is believed to ~,ri~e as a r~stxlt of the proliferation of microvascular endothelial yells in the retina. Human growth hormone is known to stimulate pr~liferation of microvascular endothelial cells.
See Itymasze~rskie et al:; Proc. Nat. Acad. ~Gi. ~JS1~ ,~:617-2~.
(1990 . The' growth hr~rtnone antagonists -of the present invention ~.y t~aerefore be useful in countering the adverse effects .of e7euated levels of endogenous growth hormone on microvascular tissues, such as the re~lna, In d~.abet~.CS, Or In ether individuals experiencing excessive growth hormone le°c~els .
Glomerulc~~clerosis occurs in a variety of glomerular d~a~ages, including diabet~.c n~~hropathy. The cause is unknovem, but mesangial cell p~'ol~.feration precedes or ~c~ompan~.es mesangial sclerosis. Thus, dysregulation of resident glomerular cells may be an imp~~tant issue in the development of glomerulosclsrosis (Doi, e~ al., Vim. J. Pathol.;
1.37:541, 1990) . _ VV~ 92!1973b P'(°f/~J892/03532 c c ~ _ 6 _ tz ~~ z~. ~
Transgenic mice which express bGH have been shown to have enlarged glomeruli which progressed to a state of.
glomerulosclerosis. Thus, GH has been implicated in the development of diabetic glomerulosclerosis (Doi, et al., 1990, and Bell, stn. J. Nled. ~ci. , 301:195, 1991) . ~.1 GH antagonist could alleviate the GH-dependent effect an the cells of the diabetic kidney, and thereby be useful in the prevention or treatment of glomerulosclerosis.
tnThile growth, hormones have not previously been implicated in hypercholesterolemia, in another embodiment, GH
aratagonaats are used to reduce serum cholesterol levels.
It has been suggested that long-activity somatostatin analogues may have ~ralue in the control of breast and prostate cancers. lNZanni; ~iotherapy, x:32-36 (1992). Manni hyp~hh~si~es that they could inhibit tumor growth by a number ~f mechana,smsg including inhibiting growth hormone secretion.
Growth. h.armaone is implicated because it is lactogenic and because it elevates I~F-1 levels; ale suggest that the growth ~orrc~cane antagonists of the present in~rention may be used in the treatment oaf cancers whose growth, is facilitated by ~xadogenous yro~vth horanor~e r~~ IGF'-1.' .
In general. these antagonis s are therapeutically or prdphyl~ct~:cally, us~f~ul ia~ c~unterihg the adverse effects of gr~~rhh horm~ne~, b~th endogenous hormones and hormones administered clinically.
In the caaarae o~ our raork, we have discovered a cor~elavion betv~reen the ability of mouse L cel3s to secrete the Protein and the protean having an effect (pos~.tive or negative) on growth rate in a transgenic an~.mal. The use of an L cell secretion assay to identify growth-modulating proteins is also a part of this inver~tion.

VI~~ 92J1973~ 2 ~ fl 2 ~ 2 ~ PCT/~.JS92/03532 The appended claims are hereby incorporated by reference as a .further enumeration of the preferred embodiments. All patents and publications cited in this specification are incorporated by reference.
~ItI~F I~~~C,R~PTIi~~1 tala' T~ DRI1GS
Figure 1 Amino acid sequence of bGH (G119R) and nucleotide sequence of the gene encoding this analogue. The alpha helices are marled and the amino acids are numbered, with number 1 being the first amino acid of the mature protein. The boldfaced bases and amino acids are those mutageni~ed in the G119R mutant:
Figure 2 General strategy of oligonucleot2de directed mutagenesis. p$GH10 6 was used as the parental vector. It contains mouse metal7.othi~nein I traa~scriptional regulatory sec,~uences (E~lT- 3. ) fused to the bGH gene (BamHI j oined with EgI2I ) which c~ntai:ns f ive exons ( shaded boxes I--'T) and intron A. This fusion. gene wad incorporated. into pER~22 at the EcoRI
s~.te: The pER322 orygin of repla,cation (OR.I) , ampicillin resistant gene (Amp), as well as the bGH translation start (~1~CG) and stop (TAG) codona are ixadicated. 5' and 3' non-translat~d regions a.re shown in hatching. The .~cleotide s~~en~~; between restriction sites Tth111I and r~mal is shown.
Substihutaon mutabi~an~ are indicted: One silent mutation is also iyadi:cate~ ( * ) ~~ich created a uniqe.te Eam'HI site . The p~~itaon of the principal ammo acid residues mutated in our e~erlments ~~1~, 117; 119, 122') are indicated.
Flc~ure 3 is an idealised surface net (cylindrical plot) representation of most of the thud alpha helm of bovine growth hormone : Tl°~e surface net is produced bY prod action of the helix onto a coaxial cylindrical sheet of paper, cutting this paper parallel o the hel~:cal ~~cis and flattening it. The ~rolumes of the amino acids aye given ~.n parentheses . A dashed line indicates the cleft or depression formed by A1a122-G1y119-Asp115:

'V6~~ 92/ ~ 9736 ~'f,'T/~JS92103532 - 8 - , Fi"c~~:re 4. is a' '~p~ot of the secondary structure prediction (alpha-helix, beta-sheet, reverse turn, random coil).
for amino acids 208-127 of bovine growth hormone (a) wild-type (b) the mutant G119R and (c) the mutant A12~L. These plots.
were generated by the '"Micxo-Genie" program.
Figure 5 Scatchard plots of data from competitive binding experiments for wild type bGH and bGH-N!8 using mouse liver membrane preparations. The ordinate represents the ratio of bound to-free bGH and the abscissa the concentration of total bGH bound. Each point represents the mean of four experiments which were carried out in triplicate.
Figure 6 provides a growth rate comparison among control (non-traa~sgenic), G119R; Gil9L, G119R. and G119P mice, a:llustrating the growth-inhibitory effect of these mutants.
Firs, 7 presents an axial view of the third alpha helix (1t39-126) of bGHf showing its amphipathic tendencies.
Hydxoplaobic ~.no acid sectc~xs are sh~.ded by dots; hydrophilic amino acids ark indicated b~ o~en'sectors; the glycine sector, a neutral am3.no acid, by ~laxated lines . ~'he residue's numbers ~d hyc~xophi~.i~ity values (Hopp and Wood scale) are given.
T~i~re 8 pxes~nte ssde views of the third alpha helix of wild tae (left) and G119R mutant (right) bGHs projected on tie plane in which ;the side cha~.n of the ~rginine-119 of the mutant ~~.19R lies . The gl~rcine 219 residue found at the bottom of th.e cleft °is indicated by an arrow.
The ~riews were prepared b~ use of molecular modelling.
software (QUAN~°A and CF3An, Pol~rg~ne; Waltham, Massachusetts, USA) . .
Figux'e 9 compares serum glucose, urea/nitrogen, and triglyceride levels of control mice, transgenic bGH-I~8 (ElI7:L, 1P(.°f/iJS92/03532 W~ 92/19736 G1~.9R, A122D)-producing mice, and transgenic wtbGH-producing mice, of bath sexes.
Figrure _ 20 compares serum cholesterol for transgenic wtbGH-producing mice, control mice, and transgenic bGH-1~2a-producing mice, of both sexes.
Figure ~.1 plots GPDH activity against bGH-MS dosage in a competitive inhibition assay for the antagonism of the ability of GH (here, wild-type bGH) to promote the differentiation of preadipocytes (NTH 3T3-F9:42A cells).
Fio~are 1.2 compares of the effect of bGH and bGH-M8 on the differentiation ~f 3T3-F442A cells. At confluence, cells were incubated .with increasing concentrations of bGH or bGH-N18.
Cells were h~~°vested on day 8 for determination of GPDH
activity. The experiment ~a~s repeated twice with similar reeults. Hacl~ bar represents the mean value obtained from tripli~at~e agsay~The error bar represents the standard di~r~a i oh.
Fi,~~ure 3. sh~ws the re~.ati:onship between serum bGH
an~ilog con.~entratioxis ~:md the graw~h ratio of transge.~.c mice (TG)/n~ntr~nsgenic (lef~G). The ordinate represents bGH analog concen~ra~s~ns in ~exum. The ab~~iea~~ rep~esex~ts the growth ratio of T~/I3'f~ mike, ~'zc~u~°,~ 14 shows the relationship between serum hGH
analog concentrations and the growth ratio of transgenic mice . , (TG) /nontransgeni~ (NT~) . The ordinate represents bGH ar~a~:og don~ent~ations in sezum. The abscissa represents the growth ~~ti~ of ~'~/~fG mice.
mD~~ ~~ TPREF~'3 D $
The present invention relates to growth hormone antagonists, especially growth inhibitor, which are peptides ~r proteins having a similarity in sequence and secondary VhO 92A19736 ~ PC,°T/11~92A03S32 structure to a vertebrate growth hormone, including but not limited to mammalian growth horanones, especially human and bovine growth hormones. Preferably, the compound comprises an.
alpha helix having an amino acid sequence homology of at least about 50% with the third alpha helix of a vertebrate growth, hormone, especially bovine or human growth hormone. Other alpha helices of the native hormone may be omitted if this can be done without 1~ss of growth-inhibitory and/or other growth hormone antagonist activity. The use of the term "antagonist"
is in a functional sense and is not intended to limit the invention to compounds having a particular mechanism of action.
The overall percentage homology of bovine growth hormone with other mammalian growth hormones is highs porcine (9a%): ovine (99%). human (66%), and rat (87%). Insofar as the third alpha helix (amino acid sequence homologous to bGH
109-126) is conceded, the percentage homology is comparable to the ovegal:l f~.gurea porcine (94%), ovine (94%), human (66%), and rat (94%) .
The second~.x~r Structure of a polypeptide is a regular ar~ange~nent of a liaaear segment ~f the polypeptide c~~.n. The m~~t c~~nlyy enco~.~ered secondary stx~xctur~s are the beta-eh~ets and the alpha-he~:a,ces. See Schulz and Schimer, F in a. 1 s f ~~ in ruc re ~~ (Springer-Verlag: 1979).
The alpha. - helix 'is' stabilized by hydrogen bonding between ~aeptide. amide and carbonyl groups of residues separated by a single turn of the'helix. Secondary structure predictions are based sin observation of the frequency of occurrence of the aanin~ acid a.n a bets.-sheet, alpha~helix, etc. in a protein haring a lsno~ran three dimensional stx~zctuxe.
The three-dimensional structure of porcine growth hormone has been determined by X-ray diffraction and compared to that of other. growth hormones. ~bdel-Meguid, et al., Proc.
fat. Adad. Sci:, 84:6434 (1987). Like the other growth ~V~ 92119736 PG~'/US32/03532 hormones thus studied, it is a single domain protein arranged as a four helix bundle with the helices in an antipara11e1 relationship. Its four helixes are made up of residues 7-34, 75-87, 106-127 and 152-183. For X-ray. studies of bGH and hGH, see Hell, et al., J. Hiol. Chem., 260:8520-25 01985) and DeVos, et al., Science, 255:306-312 (1992). 'the three-dimensional structures of other growth hormones may be deduced by comparison of the sequences with due regard for the secondary structure tendencies of substituted amino acids.
Hovix~e growth hormone is 93% homologous at the amino acid sequenG~ level w~.th porcine growth hormone, and bGH's structure has been deduced by study of the two sequences and of the structure of porcine growth hormone. Its four alpha helixes have been reported to be assumed by amino acids 4-33, &6-80, 308-127 and 150-179. The third alpha helix of bGH is defined as amino acids 106-229. However, it will be noted that the ends of th~.a helix have a less marked alpha helical ~econdarr~r structure than does the central region, which is 109-126. a ect bounds of the third alpha helix may differ for ~them GH°s, dependiaag on the alpha hel~.cal tendencies of the "and" amino ada.ds. the conformat3,on as reasonably consistent wa.th the preda.ction~ ana.de , by Chen and Sonenberg, Bio,~a.s~ry, ~6:2'~.10 ~~:977) u~~.ng tae method of Chou and Fasman;
Hiochemi~try, 1:222 ~1~74) (mss 10-34, 66-87, 111-127, 186-x.91) .
The ~nino acid ses~xen~e of the growth hormones isolated from various ~rert~brate species are highly conserved.
In a. comparison of flounder growth ~.ornnone with other growth hormones; including bGI3, ~latahiki, et al., J. Biol. Chem., 264:312 (1989) a.dezatified five, con~er~red regions. Watahiki's conser~red region G~4 c~mprises the stretch T~KDLFEGI~EbHf~ of bovihe growth h~rm~ne, i . e. , r~esid~,tes 1~.3 to x29 . tfatahiki' s Figure 3 identifies residues conserved among the GHs and residues predicted to be important for the manifestation of growth-promoting activity.

~O 92119?36 P~'1US9210~3532 Studying .&~Tatahiki~s GD~4 consensus region, several families of growth hormones may be discerned. 'The first family.
(I) domprises cGH, rGH, pGH oGH, bGH, and hGH. These begin with LKDLEEGI. They then continue with IQA (cGH, rGH, pGH),, ILA (oGH, bGH) or IQT (hGH). All members of family I then conclude GD4 with L1KR,ELED (except for rGH, LMQELED, and hGH, LMGRLED). The second family (II) comprises fGH, yGH, tGH and sGH. These have the consensus sequence LS (E/D) LK (M/T) G(L/I) (L/G/H/N) (K/L) LI (E/T/R/I) (A/G) (N/S) QD.
Four amino acids in GD4 are conserved among all of the grovath hormones noted by Watahiki: Leu 113, Leu 116, Gly 119, ~eu 123 and Asp 12 (numbering according to the bGH
sec~uerace) . ~f the amino acids nearest Gly 119 on the face of the thzrd alpha helix, Asp115 a.s strongly conserved (replaced r by Glu in the fish hormones); Leu 116 is invariant, Glu 118 is Cd~~erved a~aox~g the a~amma.ls and birds; but replaced by Met, Thr or Val in fish, Ile 120 ~.~ almost invariant (replaced by Leu in f~H); and Ala 122 as well conserved, especially in mammals and bir~.s (replaced by Thr a~n hGH and L~u or Lys in fish GHs ) . ( It sh~uld be understood that the present ~.nvention is not limited to mutants in ~rhieh these conse~°a~at~.ons a~~e maintained ~
zt his been shown that a recombinant molecule conta~:na.n~ a hGH- (1'-13~) fragment linl~~d to a human placental lactog~n- (~:~41-1.91). fragment retained full hGH immunological activity ~d bindi~.g affinity to 'GH rece~tore isolated from rabbit liver. R.us~eJ.~:, et al. , J. Hiol: Chem. , 256 : 296-300 (193). By ~xsin~ the homolog-scanning mutagenesis technique, gene fragments ref homologous ' hor~none~ -~..e. , human placental lactogen or human pro~actin - wex°e systematically substituted through~~zt the h~H gene, thus producing various chimeric ho~nones. Cunningham, et al., Science, 23:1330-36 (1989). A
o,~mpari~on of the binding affinities o~ those mutants GHs and wild-type hGH to a. cloned liver hGH receptor led to the conclusion that there were three discontinuous polypeptide VN~ 92/19736 PCT/~JS92i03S32 determinants in hGH involved in receptor binding. They were located at the 1~1H2 terminus, C00H terminus, and within a loop between amino acid residues 54 and 74. These putative binding domains were further analyzed by an alanine-scanning mutacgenesis technique in which alanine residues were systematically substituted throughout those regions. Amino acid residues at positions 10, 58, 64, 68, x.'72, 174, 175 and 176 of hGH were shown to be important for Gi3 receptor binding.
However, none of the mutant GHs were reported to inhibit growth. Cunningham, et al., Science, 244:108x-85 (3.989).
The present invention is not limited to the mutation of the third alpha helix of bovine or human growth. hormone.
Rather, it encompasses the mutation of the third alpha helix of any mat~rcnalian or other vertebrate growth hormone, including, but not limited to; the growth Norm~nes whose sequences are given in t~atahi~i (1.989): flounder, yellowtail, tuna, salmon, chieken, rato p~rcihe; ovine, bovine and human growth hormones.
~xpreseion of Cants o~ ether growth hormones is facilitated b~ the availability of genes encoding the latter. See, e.g., G~eddel; l~3atttr~, 281::544-683 (1979) (hGH) .
The concept ~f a polypeptide ~rhich is substantially hom~log~us to bovine growth hormone is deemad to include (but is n,~t l,~:mited tc~) any polypeptide which differs from bovine or hen growth ho~none by (a) a substitutibn at an amino acid cc~rrespondin~ to amino acids x.15 ~~ x.19 of bovine growth hora~none, (b) a substitution at an amino acid corresponding to an amino acid of bo~ri:~e or human growth hormone which is not consezvred among the vertebrate growth hormones, especially the replacement of that amino acid by one found at the site in a different growth hormone, and/or (c) truncation of amino acids g5 and/or 1:34-13~:: (Conserved amino acids are identified in Watahiki, et a.1., 19~~.) Thus, all non-bovine vertebrate gro~rth hax~mones are "substantially homologous" with bovine and/or human growth hormone. Preferably, the palypeptide is at least about SOo homologous, more preferably at least 80%

'VV~J 92!1973b PC.°TlUS92103532 .~

homologous, with bovine or human growth hormone in the subsequence substantially corresponding to the third alpha helix (approximately, residues 106-129) of. bGH, and more.
preferably over the entire ~.ength of the polypeptide (ignoring extraneous non-bGH-related fusions to the amino- or carboxy-terminal).
The compound is Considered to be growth-inhibitory if t2~e growth of test animals of at least one vertebrate species which are treated with the compound (or which have been genetically engineered to express it themselves) is significantly (at a 0.95 confidence level) slower than the growth of control animals (the term "significant" being used in its statistical sense). Preferably, it is growth-inhibitory in a plurality of species, or at least in humans and/or bovines.
Growth hormones have considerable interspecies cross-reaCtivity. Gil3., et al., Biotechnology, 3:63 (1985) reported y that recombiaiant chicken and. bovine growth hormones accelerate growth in juvenile pacific salmon.
It i~ k~aov~ra that Certain fragments of growth h~rmones also have gr~wth-pr~moting activity, and it is expected that the gro~rth-ihhibibory peptides (the reran is used here,~after to Exclude proteins) of the present invexation need not be as large ~~ b~T~d Preferab7:y, the peptides are at least 11 amino acids 1~ng (three turns o~ an alpha he3.ix) and more preferably at last 50 amino acids ~:ong. ~h~se peptides may retain the growth inhibiting ~Ction of; e.g:, bGH (G119R), yet lack other, undesirable biological a~ti~riti~~ of the native size mutant.
They a.y also hive more desirable pharmacokinetic ch~.r~cteraatics The growth inhibitory peptides of the present invention may also be larger than bGH, provided that the additional amino acids do not ~'es~.lt in the Compound being unable to reduce the growth rate of a vertebrate.

~V~ 92/19736 P~CT/~JS921035~2 ~~.~JN:~.~~

While the mechanism of action of applicant°s growth inhibitory peptides is not known, it is believed that they function as antagonists to wild-type growth hormones endogenously produced by the target animal. We have shown tk~at, e.g., bGH (G119R) and bGH (G119R, E117L, A122D), both competitively inhibit the binding of wild type bGH to liver membrane preparations. Thus, it is believed that the compound has a net result of inhibiting growth because its growth-pr~moting acta.vity is substantially less than that of wild type growth hormones (and perhaps is negligible) yet it can displace from growth hormone receptor (GHR) sites the endogenous native growth hormone (whose stimulation of growth would have been more pronounced). However, applicants are not bound by this theory.
De~Tos, et al., Science, 255:306 (1992) examined the comply of hG~i and the extrace~.lule.r domain of its receptor (hGI~) bar X-ra~,diffraction. The f~.rst receptor-binding region of hGH is concave and 1s formed mainly by residues an exposed faces of he~:xx 4, but also by exposed residues Of helix 1 and residues in the region connecting helices ~. and 2, The second receptor-binding region d~r~prises the exposed sides o~~helices 1 and 3 end is. relatively Elate The role of~the helix 3 is shov~n, best ~.n DeVos ° Fic~. 5; there is a s~.gnificant decrease in salvent acces~ibil'~ay a.r~~and hGH E1~:9 upon complex formation.
The c~mplex had' the form hGT3 (hG~Il3.) Z ; that is, the receptor dim~ra.~es to interact w~.th. hGH. It is possa.ble that our GH
antagonists interfere with tha.s da.mer~.~ata.on.
Preferably, the connpounds of the present invention have an ED50 which i.s less than about lC~ t~:mes the E~50 of wild t~rp~ UGH in an 'assay of the ability o~ the compound to displace r~.diolabeled ~r~ld tie bGH from a li~rer membrane preparation made as descra:bed below. More pref~rablyp the compounds Yaave an ED50 at least comparable to that of wild type bGH. N~ost preferably, the compounds have a higher of f inity f or ~gro~wth ~VU 92/19736 P~'lLJS92/0353Z
_ _ .., hormone receptors than does the growth hormone native to the animal receiving ~ the 'compound. For purification and characterization of a human growth hormone receptor, see Leung,.
et al., Nature, 330:537-43 (1987).
A GH mutein may be considered an antagonist, even if it lacks growth-inhibitory activity, if it antagonizes another GH-med:~ated activity, e.g., its diabetogenic, glomerulosclerotic, hypercholesterolemic, or tumorigenic activities.
the preferred growth-inhibitory peptides are characterized by ~. ~nod~.fication of the surface topography of the third alpha helix. It will be seen from Figure 3 that in the third alpha helix of "wild-type" bovine growth hormone, there ~a a surface cleft or depression beginning, at the A~partate-115, deepening at the Glycine-119, and ending with the Alanine-12~. ~1I1 of the mutants prepared so far, both t.hase which retain the gild-type cdrowth-promoting activity and those which do not, are consistent with the theory'that growth-promatin~ .activity rec~,uires the presence of this cleft or de~reseion end that, if the center of ~h~.s cleft is "filled in"
by substitution of amino aids with bulkier side ch~-ras, the mut~iaa inh~.bits the growth of the ~~ab~ ect o i~;atat~.ons' which, ~ub~tanti~.lly destabilize the alpha-helix are undeai~abl~ sine they may result in the loss of all grQ~th-related activity. ~1'e have obs~r~red such loss in the c~.se ~f several xnutatir~ns which were expected to disrupt the ~I~~a helix.
For a discussion o~ alpha helix farmers and breakers, .
set Claou and F~sman; su ra. GZu,.Ala anal Leu are the preferred al~aha helix fo~ners' while Pro- anc3 Gl~r are characterized as strong helix breakers. Substitutions which introduce strong alpha helix lareakers are less desirable, but may be tolerated in a particu~.ar case, such as the ea~d of the helix. 'the 'WfJ 92/19730 ~CTli1S92/03532 2~.~ ~ ~.Z~

secondary structures of our analogues have been predicted using the "Micro Genie'° computer program, which uses the algorithm of Garnier, et al., J. Biol. Chem., 120:97-12A (1978).
With respect to amino acid 119, glycine is both the smallest amino acid residue and the one least favorable to alpha-helix formation. Thus, it is believed that any other amino acid ma.y be substituted for it without destabilising the alpha helix, while at the same time filling in the aforementioned cleft. A11 of the 6118 bGI~ substitutions tested resulted in a "small animal" phenotype. These substitutions wire arginine (a large, positively charged ~), proline (a cxelie aliphatic AA), lysine (a large, positively charged .~A), tryptophan (a large aromatic AA) and leucine (a large, nonpolar; aliphatic AFa). In hGH, the homologous glycine is at pos~.tion 3.19. Subati~ution of arginine or tryptophan resulted in an antagonist, ho~rever, hGH G120~ retained gro'c~ath-promoting activ~ay. Corxseguent~.y, it is presently believed tk~at this glycine; which is conserved in X11 vertebrate GHs, may be replaced by away amino acid other than alanine (the second smallest amino. acid),'and mope preferably by any amino acid which i~ at least as large as proline (the smallest replacement ami~ao acid known t0 re8~tlt in a "sm~.ll°' animal phen~ty.~~) . The del.eti~n of G$"'' 9 ~e also Y.nov~n to result ~n 'a °'small" animal phenotype Nlod~.f icati~n - of p~~i~ion x.15 is suggested by our ~°tleft'n thet~ry. The' aspartate at position. 115 may be replaced b~ a b~lk~.er amira~ acid wh~.ch does not destroy the alpha helix.
Preferably, tfie replacement amino acid has a sire greater than that of ~lut~tate. The aani~.o acids histidine, methiox~ine, is~leucine; leucix~e, lysine, a.rginine, phenylalanine, tyrosine and trypt~phan are substantially larger than glutamate. Of these; H~.s, Met, Leu and Trp are more preferred because they combine the advantages of bulk with a reasonably sarong alpha.helical propensity. Note, however, that the wild-type Glu is the stron.Best alpha-helix former of all of the annino acids .

~i'~ 92J19736 P(.'TlL3S32/03532 The D115A mutant of bGH is not a GH antagonist, but Alanine is smaller than Aspartic Acid, .~so this is not probative of the value of replacing Asp115 with a bulkier amino acid.
zt is possible that G119A might lead to a t~small°' phenotype if coupled with other mutations, e.g., at 115 and 122.
zt is Qossi.ble to systematically screen for the effect of all possible am~.no acid substitutions at positions 115 and 119. ('here are 202 -1 or 399 combinatorial ~eassibilitie~.) DNA which encodes bGH and is degenerate at these positions, so as to there encode ..all possible amino acids, or only those with acceptable alpha-helical pro~ensa.ta.es, is prepared, a.g., by a 'dirty bottleo~ synthesis.
Phage are pregared.; as taught by Ladner, et al., pC~/~S89j03731, W090/02809, which display the mutant bGHs as a deznaira of a dhimeri~ coat protein. The phage are incubated with ~ chromatographic support bearing a growth hormone receptor. (For tl~e techniques ~f isolata.ng growth hormone receptors, see Leung, et al., Mature 330:537 (1987) and Spe~a.ce~, et a~.. , J. Biol : Chem. , 263 :7862 (3.988) ) . l~Tative bG~I is ;also incubated with the support, before, during...-~r after th:~; phage a:ncLabation. Bound phag~ are recovered, amplified and ein~d. to determine the sequence of the mutant bGH (usually by': sequencing the'c~rresponding g~he a.n the phage genom~).
These z~utants haV~ dem~nstrated the ability to compete with wild tie bGH for ~ growth h~~one receptor. Their ability to a~~agor~i~e GH acti~rity in ~ri°vo i~ then confirmed by, a . g . , administering'' them directly to an animal or by preparing a suitab~.e transgenic animal:, ~r by the ~n vitro assay described ix~' Ele 7. This approach ~.y be extended, if desired, to other ami,rao acid posit~:ons in the third alpha helix. Amino aCyds which are particularly preferred for screening are the six an~,ino aca.d~ spatially nearest bGH ~ s Glyl~.~ , that is , ~a122, I~eu123, zlea20, Leu116, ~sp1.1.5 and G1u118. zt should be noted that Bass, et al., Proteins: Structure, Function and wc~ 9zim~~6 p~re~sgz/Q3s~z Genetics, 8x309-314 (1990), prepared "hormone phage" which express and display hGH-geneTII fusion proteins and which were bound by anti-hGH monoclonal antibodies. Moreover, it was possible to separate phage bearing inTt-hGH from phage bearing the low affinity hGH mutant R64A by means of affinity chromatography (using the extracellular domain of the hGH
receptor bound to nonporous oxirane beads).
besides the mutations at position 119, which is deemed necessary to impart the desired growth-inhibitory acti~rity, additional mutations are possible which; will leave the growth-inhibitory activity or other antagonist activity intact: I'hese mutations may take the form of single or multiple substitutions, deletions, or insertions, in nonessential regions of the polypeptide.
For example, it is possible to alter another amino acid in the alpha helix provided that the substitution does not destroy the alpha ~e~i~. preferably; such alterations replace an amino acid with one of eimi3:ar size ~.nd polarity. It may be advantageous to modify amino acid$ f~,an)c~:ng the primary mutation site 119 in oxder to increase the alpha-helical propensities of the s~queh.ce, paxticul.a.rly ~.f the mutation at 119 is one expected to destabilize the helix<

the following table may be helpful in identifying candidate mutar~ts s ~ V~7:ume A3.pkia f anctstxoms ) Helicitv Gly(G) 60.1 ~.93 Ala (.~) ~ 8 ~ 6 1. 45 Ser(S) 89.0 0.79 s (c) 108.5- 0: 77 Asp (D) 1.11: 2 0 .98 Thr (T) 1.1.6 1 0 . 82 Asp (N) 11.7 . 7 0 . 73 Px~o (p) 122 . 7 0 . 59 G1u(H) 18.4 1.53 .

val(v) 140.0 1.14 Gln (Q) 3.43 . 9 1.17 His(H) 18.2 1.24 .

i~b'O l2/1973~6 ~ PC'TA~JS32/~3532 ' ..

Met{M) 162.9 1.20 Ile(T) 166.7 1.00 Leu(L) 166.? 1.34 Lys(K) 168.6 1. d7 Arg(R) 173.4 0.79 Phe {~') 1~9 . 9 1..12 Tyr(Y) 193.6 0.61 Trp{W) 227. 1.14 Several of the cited references provide guidance as to where and where not the polypeptide will tolerate mutagenesis. Watahiki, et al. (19H9) compared the sequences of flounder, yellowtail, tuna, salmon, chicken, rat, porcine, ovine, bovine and human growth hormones. He identified five consez~red domains which he labeled GDl-GnS. Mutations in these conse~red domains are more' likely to affect activity; GD4 corresponds to the third alpha helix of bGH. In mutating a known GH antag~na.~t with the desire to retain inha.bitory activity, mutations outside the c~nserved domains are more pedant. Howe~rer, mutations in these conserved regions, if caref~x~.ly claosez~; may be to~.erated; for example, the mutation E11.?L does not modify the activity of either wild-type bGH or a bGH G119R mutant. Note that not only substitutions, but also ihserta:ons and deletions, are suggested by the example of the cogne;te harm~nes.
~bdel-Nleguid~ et al. (197) determined the 3ri-st~ucture of reco~niaia~~~t anethionyl porcine growth hormone, and suggestedthat it revealed the wg~n~~al three-dimensional fold"
of the growth horanones a The 3I7-structure can be used to id~ntif~r interior and surface residues; generally speaking, proteir~~ mutated at surface residues (other than the receptor binding site) are moxe likely to remain functional. However, Creigh~On ~.nd Chothia., Platuxe, 339'14 (199) diSCUS9 the toleration of mutations at-buried residues. The structure may , al9o be'used to determine flexible surface "loops°°; proteins arm more tolerant of e~eletions and insert~.ons in such regions.
~unningham; et al. (;1989) used homolog-scanning mutagenesis to a.d~ntify the epitopes of hGH for its cloned ., -.,.i t.. . Y ,' .x c a.:, o 7 . , ~'J , .
n ..>. . . .. ,.., . r .. n., . . ., . t .. .. ~W.v:. n. , , ~nriw:..~..m. ,. ......, s .'.:.tf....~.. ~....r . . . ,. . . .. .......~..A:
W . In..pl..;, . .'f~~.'~. . .. r.... . .. , ,.. .... .. . ..

ego ~zn9~36 ~mus9z~n3~~z ~1~~~.Z9 -~~-liver receptor. Only variant hormones having mutations in regions C (54-'74) , x'.(164-19 t7) , and, to a lesser extent, A(11-33) exhibited reduced binding affinity. Cunningham and in7ells, Science, 24~s1081 (1989) used a related technique, alanine-scanning mutagenesis, to further study these regions. Note, however, that binding to the receptor utilized by Cunningham is not necessarily critical to the growth-promoting or growth-inhibitory activity of the mutant.
k'or example; it seems likely that major amino- and carboxy-terminal truncations can be made without adverse effects on growth-inhibitory activity, since the 96-133 fxagment of bGH (Ttll2L) is understood to retain bioactivity.
Truncations may be generated by gene modification or by e~copeptidase treatment .
7n terms of the kinds of substitutions which may be ode, one may 1~ok first to analyses of the frequencies of amino acid ~Iaangea between homologous proteins of different organisms, such as those presented in Table ~.-2 of Schulz and Schimer; a ra ~d figure 3 - 9 of Creighton, s_ u~ara . Based on such analyses, we define conservative substitutions as exchanges within the groups set f~rth below: ,d I sma.l~: ala.phatic, nonpolar or slightly polar re~i:~lues -Ala, Ser, Thr (Fro, Gly) 1CI negata.°vely, charged residues and their amides Asn g Gnu Gln 2II positively charged residues - ~iis Arg Lys ZV large aliphatic nonpolar residues l~Iet Leu Ile Val (Cys) large aromat~.c residues -Fhe ~.'1rr Trp Three residues are parenthesised because of their special ro~.es in protein architecture. Gly is the only residue without a side chain and therefore imparts flexibility to the chain. Fro has an unusual geometry which tightly constrains 1~V0 92/19736 P~'/~JS92/~3532 _ ~2 _ the chain. Cys can participate in,disulfide bonds which hold proteins into a particular folding; the four cysteines of bGH
are ?highly conserved. Mote that' Schulz and Schimer would merge.
I and II above. Note also that Tyr, because of its hydrogen bonding potential, has some kinship with Ser, Thr, etc.
Within the growth hormone family itself, we see a wide variety of substitutions of other amino acids for the residues of bGH. k'pr example, among the vertebrate GHs set forth in Watahiki, et al., (1989), Pro appears 6 times in bGH.
The first Pro is not substituted, but is absent from sGH. The second is replaced by Leu in fGH; Thr in yGH and tGH. The third is replaced by Leu in fGH, Ile in yGH and tGH, Val in sGF~: ~'he fourth Pro is conserved. The fifth Pro is replaced by Phe in fGH; Ser in yGH. The sixth Pro is replaced by Phe in fGH; yGH and tGH; and Leu in sGH. Overall, Pro is replaced 4 times by Phi, 3~ tunes by Leu, 2 times each by Thr and Ile, and ~nce each by dal and Sir. When this azxalysis is extended to all ,no acids of' bGH; . vae obtain the following tallies:
( 1~ ; 2 conser~red) - >
Thr 14 . Ser fly.) , Asp (~) , Ile (4) . Glu (4) .
Glv ~, 4~ . Val ( 3 ) , Gln ( 3 ) D Leu ( 3 ) . Lys ( ~ ) .
Pale ( 1. ) v ASn ( 1 ) j , J S
~p ( 1.0; ~ Conserved) - >
~~y~(6)'a .Ash t4) . Val (~) D ~ 3) . Lys (2) , ~'g (2); Gly (2), Thr (1), Ser (1), Phe (1). ~.a (1a Glu (13; 2 conserved)->
,~~ f14) . Lys (1~) a Gln ~4) . Ala (3) . Pro (2) .
T~°l~' ( 2 ) , Asn 1 , Ser ( ~l. ) , Val ( 1 ) , Met ( ~,. ) , Arg (1) D G1Y (1) Phe (13; 3 conserved)->
Tvr t7). Leu (~), Asn (6), Ser (5). Gln (3). Ile (~) D Gly (2) I Has (1) D Thr (1) , Val (1) ~l0 92119?35 P~l~J~92/~3~32 ~~~~~~z~

fly (lo; 1 conserved) ->
Arg . ( 9 ) r Cs"7.u ( 8 ) r ~P ('1 ) r 'dal ( ~ ) , Pro ( 4 ) . Sex' Asn ( 3 ) . Phe ( 3 ) . Asp ( ~ ) , Hi s ( 1 ) , Thr (1) r Tyr (~.) r Aa.a 1 Hia (3; 1 conserved) ->
~,~~(1) ~ Asn (1) , Tyr (1) , Asp (1) I1e (~') ->

Gl.n (7) , .Asn (5) , laeu Phe (4) . Val (4) (4) . , .

Al.:a(2) , Ser (1) . ~g (~.) ~~s ( ;

conserved) ->

Ser (11) . Ar 7 , Gly (~4) Gln (2) ~eu (2) , r , Ann (3.) ~eu (~?;
~:1 Conserved) -a her (11) ; Va7: t91 . Asn bet Gln (7) , ('7) , ('!) .

~.g (4) , flu (~) , Phe (3) ~'r (3) ~~.y (1) , , . Pro (1) I~~:S (1) s ~~t ( ~, a' :.ale Ser (3) Leu (1~ Asn ~'7) , .
~1:~
(4) ;
r'hr (3) :

( ~1~1 1 )';

~n (~) ~~

Ile' (5? , leis (4) . Aao ~lx~ (3) C,~lu (2) _I,'3) . . , Ser (1) Pro (6; canserved)>

~Yae (~) : Lei: (~)' , ~h~ Ile (2) ~'a~. t1) (2~ . , her ~

) Gl.n ( 1 conserved) -1~;

L~u (13) . .Ax3 (6) . lG~s Ser (~) Glu t4) .
(~) , , llis (~.) . G1Y (1) . ~ (~) Pro (1) ~

1~~ 92/19736 PCT/US92/~3x32 ..
_ 2~ _ Arg ( 9.3 ; 1 conserved) .. j hvg (19.) , Ser (9) , Thr (7) , lle (3) , Glu (2) , Gl~r ( 2 ) . Asn ( 2 ) . H~a. s ( 2 ) . Vas ( 1, ) , Gln ( ~. ) . Asp (~.) , Ala (l) Ser (13; 3 conserved)->
Ala (~) ~ Asn (~) , Gln (~) , Leu (4) . ~~.v (3) . Glu (~) . AgP (2) . Thr (2) . Arg (1) ~ Va1 (2) Thr ( 12 ; ~. conserved) - >
gar lle~) . Ala (13) . Val (7) , Tyr (5) . Phe (4) .
=l a ( ~ ) . Iv~et ( 3 ) , Leu ( 3 ) , Pro ( 2 ) . Asn ( 2 ) , G
Val (6) ->
Ala (4), Ser (~). ~1e (3). Thr (2). G1n (6), Gly (2) . l~f~~ (2) ; Leu (~) , L~,ts (1) ~r (6) °~
~eu: ( 5 ) , Pro ( 4 ) . Gln ( ~ ) t. Phe ( ~ ) . Glu ( 1 ) , Ser (1) Note t that the above figures are not- normalized to adjust for the relative fre~a~ncies of occurrence of the ~rioug ~t~.no adids . ~Te ~ux~lher note that in our own mu~.agenesas ~~gerixnents; changing Lys 112 to heu or Lys ~.~.4 to T(I~), Glu to G3y (~126G) br Leu (M4), or Ala to Thr (A~.22T) did xaot alter acti°~ity; while changing T~ys, G~.u or veu to Pro abolished activ-it~:
The present invention ~.s not limited t~ a.n~r particu3ar method of producarag the desixed GH antagonists.
Pref~rab~:y, these antagona.sts ark, produced bg~ first altering a gene encoding a vertebrate GPI (e.g., bGH rrr hGH) having the "r~ati~re" third alpha. helix by site~°specific Hnutagenesis, and ~~eh cloning and expressing the altered gene in asuitable W~ 92/19736 1'C.°T/1JS92103532 _ 25 _ host. Molecular biology techniques are described in, e.g., Sambrook, et al., Molecular Cloning: A Laboratory Manual (Cold Spring ~iarbor Lab Press; 2nd ed., 1989). The gene may be of genomic origin, it may be cDNA prepared from bGH messenger RNA, it may be synthetic, or it may be a combination thereof. For the amino acid sequence of bGH and for the cDNA sequence of the bGH gene, see Miller, et al., J, Hiol. Chem., 255:7521-24 (1980). 1"'or the genomic bGH sequence, see lrloychick, et al., N'tRGlel.C Acids Res. , 10:7197-?210 (1982) . The cDNA sequence for hGH a.s given by Chang, et al., Gene, 55:189 (1987) and DeNoto, et al., Nucleic Acid Res. 9:3719 (1981), and the: genomic hGH
sequence is in. Robbins, et al., Cell, 29:623 (1982).
The host may be any convenient organism, including a bacterial, yeast, or mammalian cell. The gene is operably linked to a promoter functional in the host. A constitutive proanoter mould a~ti.vate gene expression in a general manner, i.e., in many tissue end at all times during development. A
~egulatable promoter may be activated in a tissue or cell specific mannerp at precise time during development, or in response to chaxages in the environment. A constitutive pr~moter is usually employed when larger amounts of gene product (usually ~gotein) is requa.red or when the g~n~ product as required in anan~ cells of many tissues.' A regulatable Promoter is uta.l3zed when one gene product is required an a soall number ~~ cells of a particular tissue or at a given time during d~~rel~pmen~
°I'he e~pressifln system may be engineered so that the antag~nist ~.s secreted into the culture medium, or the host dells may be grown to a high cell density and f.hen lysed to release the c~mpound.
~ne method suitable for the purification of bGH
(G119R) and the like is described' in Leung, et al., Endocrinolog~rp ~:~.9:1489-1.496 (1986) . Hssentially, this procedure involves purification by (a) ammonium sulfate dV0 9Z/19736 ~ PC3'flLJS92/~3532 precipitation, (b) fractionation on DEAE-cellulose (or any equivalent ion-exchange cohuann), and (c) gel filtration (e. g., on a Sephadex G-25 and/or Sephacryl S-200 column). Other procedures applicable to purification of growth hormone-related compounds are set forth in Reichert, Jr., "Purification of Anterior Pituitary Hormones: Bovine, Rat and Rabbit," Meth.
Enzyrnol., 37:360 et seq. (Academic Press, ~1'.Y.:1975).
Polyclonal or monoclonal antibodies which specifically recognize the protein of interest may also be used in the purification process.
The purified antagonist may then be combined with compatible, nontoxic pharmaceutical excipients and administered to an animal, e.g. to treat a condition characterized by an excessive growth rate. (The term "animal"
is intended to include humans.) zn the case of administration to nonhuman animals, it may be preferable to incorporate the dxvg into the animal's feed, possibly ~.n a prepared combination c~f drug and, nutritional material ready for use by the farmer.
The antagonist may be administered orally or parenterally (~.r~cluding intravenously, subcutaneously arid intramuscularly) to hang in any suitable pharmaceutical dosage form. ~n the case. of treatment of xetinopathy, i~ may be adm~,a~ristered da~r~ct~:y to the eye by means of a conventional ocular phacw.tical form. An effee~ive -dosage and treatment prat~col ziciay be d.et~rm~:n~d by:conve~tional means, starting with a low d~se in laboratory anim~al~ and ~&~en increasing the dosage while monitoring the effects; and systematically varying the d~~ag~ re~~.men as wel'1. The trial dosages would be chosen after consi~3eration of 'the clinical literature with respect to administration of gr~wth hormones, and of somatostatin (a ~r~wth hormone release inhibitor).
In another embodiment; the gene is introduced into a host cell which is developed into genetically transformed cells of a transgenic animal. Linearized I~IvTTA, bearing the growth hormone antagonist gene may be introduced into a gamete, or :., , ..<.S'n . a.. . a9 .; . ..,u. a.
.. . . ',r . .. .> a .
...'~...-.a~. .n ..,stw.. ~...~ . u.l.f.. . . ..... ... f...'... , . w,... ..v .... . , .._.... ..:.5~~.......,..,. v. .. . , WO 92/19736 PCT/US92f03532 ~~.~>~~~

microinjected into the pronuclei of fertilized eggs, into the cytoplasm, into the. nuclei of two-cell embryos, into individual cells of a blastocyst, or into the blastocoel cavity. (Some of these targets may be reached by electroporation instead of microinjection.) Alternatively, a retrovirus bearing the gene may be constructed and used to infect preimplantation embryos or tissue culture cells (e.g., embryonic stem cells) which may be aggregated with such embryos. In either case, the genetically modified zygote, after a brief in vi ro cultivation, is implanted into a foster mother and carried to term. For '~~ene therapy" post partum, see Clime, et al., Nature, 284:422-425 (1980); Williamson, Nature, 298:416-18 (~.g~2). Again, the gene is operably linked to a promoter functional in the host, and the promoter may be constitutive or regulatable. Preferably, expression is regulated so abnormal ionic or fetal development is avoided.
The invention is further illustrated, without limitation; by the following examples.
1e 1: Ge~eralb~~n ~f fi~tatgons C~nferr~.ng tlae ~teduced ~P3~e~~tgrpe TE1ZI~..LS .~ vl~T~iODS
the pla~mid, pBGH-lOdelta6, was derived from pBGH-10 and contains the eomplete codang region of bGH and intron A.
Bovine growth h~rrnone introns B, C and D are absent (Figure 1).
This p3asmid eneode~ "wild type°° bGH, and its expression is controlled by a 100 base pair segment of the mouse ~etall~athione~:n I t~anscriptional regulatory sequence.
Plasmids pBGH-lt~delta6-G1 s s R and pBGH-lOdelta6-E11' ~,, ~a.zzD here derived from pBGH-lOdelta6 and were generated by segment-directed gnuaage~esis using complementary oligonucleotides to replace the I~NA between the Tth111I site (found near the 3 ~ end of Exon IV) and the Xma. I site (located 'VV~ 92/19736 PGT/US92/03532 _ 2$ _ near the S' end of Exon 'il). The other mutations described herein were generated similarly.
The complementary oligonucleotides used for pBGHIO
delta 6-G:2~gR were:
a'GTGTCTATGAGAAGCTGAAGGACCTGGAGGAAAGGATCCTGGCCTGATGCGGGAGCTGGA
AGATGGCACCCC 3'; 73-MER) and (5'CCGGGGGGTGCCATCTTCCAGCTCCCGCAT
CAGGGCCAGGATCCTTTCCTCCAGGTCCTTCAGCTTCTCATAGAGA 3'; 76-MER).
The complementary oligonucleotides used for pBGHlQdelta6-E11?L~
Ga x a R d Ai 2 z y~ veers (5'GTGTCTATGAGAAGCTGAAGGACCTGCTGGAFaAGGATCCTGGACCTGATGCGGGAGCTG
GAAGATGGCACCCC 3'; 73-mer) and 5' CC~',GGGGGTGCCATCTTCCAGCTCCCGC
ATE,A.GGTCCAGGATCCTTTCCAGCAGGTCCTTCAGCTTCTCATAGACA 76-mer).
These oligonucleotides hybridi~~e as follows;
G~l9I~
GT FTC TAT GAG A~1G CI'G AAG GAC CTG GAG GAPa AGG ATC CTG GCC
A~ CAG ATA CTC TTC GAC TTC CTG GAC C;TC CTT TCC TAG GAC CGG
Arg 'Tel Tyr Glu Lys Leu hys Asp Leu Glu Glu Arg Tle Leu A1a CTS ATG ~G.AG CTG GAT GGC ACC CC
GAC TAC GCC C~'G GAC CTT CTA CCG ~G GCC
Iaeu I~et Az.~g Gha ~~u .
~7.2:~~0 G1~~R.~ F~..2~D
GT GTR ~,'AT GAG AAG CTG ~t,G GAC CTG C3~'G GAA AGG ATC CTG GAC
ACA CAG'ATA CTC TTC G~.C TTC CTG GAC'GTC CTT TCC TAG GAC CGG
Arg Va1 r Glu hy~ Leu Lys Asp Leu Leu Glu Arg I1a Leu Asp C~'~ ~,TG' CGG Gt'aG CTG GAA GAT GGC ' ACC CC
GAC TAC GCC CTC GAC CTT CTA CCG 'fGG GGG GCC
Leu. filet ~g Glu L~u Thes~ ol~:gonucleotides encode D1VA changes which result a.n the substitutions of arginine for glycine at position 12.9 a.n pBGH-l.Cdelta6-G119R; and leucine for glutamate at 1~~ 9219736 P~.°T1~S92!~3532 ~.~~~~~~ - 29 position 117, arginine for glycine at position 119 and aspartate for alanine at position 122 in pBGH-lOdelta6-E11'L, G$~9R, and AlaaD: These amino acids were chosen because they have hydrophilic (arginine and aspartic acid) or hydrophobic (leucine) character [See Hopp and Woods, PNAS (USA), 78:3824-28 (1981)], positively (arginine) or negatively (aspartic acid) charged side chains LSee Kaiaer and Kezdy, Science acid) 223:249-55 (1984)]; and high cx-helical-forming potential (See Claau and Fasman; Ann. Rev. Biochem., 47:251-76 (1978)]
furtheringgeneration of an idealized amphiphilic a-helix (See 7Nlargalit, et al., J. Immunol., 138:2223-29 (1987);; Brems, et alo; Biochemistr~:26:7774-78 (1987); Kaiser and Kezdy, supra_;
Cherl; et al. , P1L~AS (U$A) ; 87:50f1-65 (July 19903 . Tn addition, these oligonucleotide duplexes encode a silent base-pair change designed to create a unique BamHI restriction site which s~.mplified gcreeni~g px'oc~edures. The oligonucleotides were annea7.ed and aubcloned between the Tthllll and Kanal sites using standard pracedures (l~Iani.atis ~t al:. Molecular Cloning (Cold Spring Harbar: (1982)) ~tant plasrcdid DNA's were identified b~: ~i~eatian with Bami~l restriction site which simplified screening pracedures. the oligonucleotides were annealed and su~cloned between the Tthlll~ end ~na.T sates using standard procedures IMahiatis et ~1.; Molecular Cloning (Cs~~r Spring Narborr 1982~)a mutant plasmid DrTA's Were identified by diges~aLdn with ~a~il.
The nucleotide segu~hce of the mut~.ted bovine growth hormone target geg~ons ,were determined by using the dideoxy chain-termination matched with modified T7 DNA pcalymerase (Sequenase, United States Biochemical; Sanger et al.. PNAS
(USA), 74:5463-67 (1977)): Oli~ohucleotide primers for manual DNA: sequeneing mere synthesized using the I~u~ont Coder #3~0.~DNA
synthesizer and purified by denat~arir~g polyac~ylamide gel electrophoresis, passive eluti~n and c~ncentratian by ethanol precipitation. The oligon~:eleotide primers used for the direct sequencing analysis of the two mutants was the following: I~mer (5'AAATTTGTCATAGGTCTG 3!). Briefly; 1-~~ag of doublestranded !WO 92/9736 ~'C°T/EJS92/~3532 plasmid DNA was denatured in the presence of 0.2N Na~H, and 10-~0 pmoles of oligonucleotide 'primer was allowed to anneal (65°C, a min. followed by 30 min. slow cool) to the denatured.
template. A two-step polymerization was performed by using the modified T7 DNA polymerase which extends the oligonucleoti,de-primed chain in the presence of dNTP's and deoxyadenosine triotriphosphate (~~.000 Ci/mmole, Amersham) followed by transfer of equal aliquots into each of four specific dideoxynucleotide mixes which randomly terminate chain elongation. Following addition of a formamide termination buffer to each reaction, the samples were incubated,at ~0°C for 2 min. and the ,DNA sequence was determined after size fractionation of the four sets of fragments by 10~
polyacrylamide/~M urea electrophoresis end autoradiography.
1e 2: reesi~n ~ ~.l.ian Cells ~.n. Culture using the in ~ritro mutagenesis protocols described above, two mutant bGH genes were generated initially: one con~rerts glycinell9 to arginine ("~219R") and the second convgrts~ glutamat~lx' to leucixae, glycine219 to arginine, and alanir~e~ 2 2 to aspar~ate (E1~.TL, G1~.9I~, A~:22D) .
~, ~h~ ~la~ana.~~ encoding theta mutations- as well as wild type bGH ANA (pBGHlCdelta) were transiently introduced into dultured m~use L cells; wha.ch were svxbsequently analyzed for b~H exgres~ion. Fall~wing "western analysis", protein bands of approximately 22,000 daltons wire observed for wild type bGH
end bG~ derived from the two mutant genes.
Mouse L cells were maintained in Dfi~N! (Gibco) plus 10% calf serum and ~5 ~,/m1 gentamicin (Gibco). In this study, a , m~dification of a previously described transfection procedure vas employed (~capata et al . , Nucleic kids Res. , 1.2 : 5'07-5717 (1984)). Briefly, 2~,g of plasmid DNA was added to l.~ ml of DI~lEM c~ntaining 0.2mg DEAF-dextran. This solution was added to approximately 106 cells in a 35-n~ tissue culture plate which '~O 92f 19736 CPC T/US921d~3532 had been washed previously with 2.0m1 of DMEM. Following incubation of the cells for 1 hour at 37°C, the DNA-DEAE-dextran solution was removed and the cells "shocked" for 90 seconds with 2.0m1 of 10% DMSO in Hepes buffered saline, at room temperature. Subsequently, the "shock" solution was removed and cells washed with 2.0m1 DIEM. Media containing ~.0%
Nu-Serum .(Collaborative Research) plus 50~g/ml gentamicin were changed daily, Culture fluids were stored at -20°C. For bGH
binding assays, transfected cells were incubated in DMEM minus serum for ~.6 hours, after which the culture fluids were removed and frozen at --20°C.
Sodium dodecyl sulfate (SDS) PAGE analyses of secreted bGH have been described (Kopchick et al.. DNA, 4:23-31 (1985); Kelder ~t al., Gene, 76:75-80 (1589). In this study, we used a polycloa~al anti-bGH serum for "western°° analysis.
1e ~ m ~r~~r~ne R.~cepttOr Ba.ndi.n~ Bltud3..e~
Culture fluids lacking serum were collected from ells tramsf~c ed by pBGH-1~delta6 (wild type bGH) and the m~~s,nt bGH genes: Following lyophilization of the culture media and bGH c~ncentration determinations, cgmpetitive mernbran~ ~inda:rag studies were c~.rried out- as previously described (Smith & Talamants, .Js B301. Chem., 26222213-19 (1987)). Diver membrane prepa.rati:ons from C57BDj6JxSJD hybrid mice ~f either sex'(f0-120 days ald? were homogenized with a B~a.nkman Polytron in 4 volumes (w/v) of 0.3M sucrose, 1~mM
EDTA, SOmM Hepes, (7. lrnM TPCK and llmM PI~ISF at pH $ . 0 . The' above step and all the follo~iing protocols were carried out at 4°C.
The; homogenate was centrifuged at 20; 000xg for 30 nain. and the supernatant was centrifuged at 7.~C,OC~~xg for 1 hour. The pellets were washed once with lOmi~I Hepes, pH 8~0 and ~ecentrifuged. These pellets were xesuspend~ed in lOmM Hepes, ~H 8.0; to a protein concentration of. approximately 50mg/ml.
The membranes were aliqucated, frozen on dry ice, and stored at -20°C. Membrane protein concentrations were determined by the 1~V~ 92>19736 ~Cf/US92A03532 Lowry protein assay (Lowry et al., J. Biol. Chem., 193:265-275 (191)).
Competitive hinding assays were performed using the followa.ng protocol. Microsomal membranes corresponding to.
three mgs. protein were incubated with 30,000 cpm/tube lzsl bGH
(Cambridge lMedical Diagnostics) and unlabeled bGH ranging from 0.3m1 assay buffer (20mM Hepes, lOmM CaCl2 0.1% DSA, and 0.05%
NaN3 pH 8.0). All assays were performed in triplicate. After overnight incubation at room temperature, membrane bound hormone was separated from free hormone by the addition of 1 ml of ice cold assay buffer followed by centrifugation at 10,000xg for 20 min: Membrane pellets were then assayed for radioactivity. Specifically bound radioactivity was determined by subtraction from the value produced by incubation of membranes ~ri~h 5~Cg unlabeled bGH (Smith and Talamants, 1987) .
Hffective doses which resulted in 50% displacement C~~~O) of lzsl-bGH fr~m the membrane preparations were determizaed. Mutant bGH encoded by pHGH-lOdelta6-G1~9R. and ~BGHl~delte. 6-El z ~ I,,, Gl i a gz~ Ai 2 2 D revealed an HD50 value eirn~~.lar to wild type bGH.
~' 1~ ~4: gean~.c ~ ~x~uct~.c~n silent study 1~ 'series ,of tran~genic mouse dines which contain wild type and mutant IbGH genes were produced by standard mieroinjecta:on techniques (~dGrane et al.. 1988). DNA
eactraction from mouse tails, dot blots, and serum determinations ;was as described (~IcGrane et al. . 1988) .
~,e Renee contain the ~ranscriptional regulatory seqbzenc~s- of the mouse metallothionein I promoter which has been shown to be aetive in liver tissue as well as other tissues of the transgen~~ mouse (Pahniter et al.. Nature, 300:611-6~.5 (19'82) ) . Offspring generated by the znicroinjection procedure were assayed for bGFi DNA by 'slot blot hybridization '1y~ 92119736 ~'(.'T/USg2/03532 _ 33 -analysis. Mouse lines were generated which contain approximately one copy of the recombinant bGH I7TdA sequences derived from pBGH-lOdelta6, (wild type), pBGH-14de1ta6-G119R, and pBGHlOdelta6-E~ 1 ~ L, G11 a R, A12 2 D, Serum from transgenic animals were assayed for bGH levels by the Inlestern technique.
All mice which expressed the wild type bGH transgene in serum also possessed a corresponding enhanced growth rate. Mice which expressed mutant bGH (G11 s R or Ea 1 ~ L, 6119 R, p,1 2 2;p) in serum were dramatically and significantly smaller. After eight weeks' growth, the growth ratio for wild type bGH tranagenic mice relative to control littermates was ~..5 while the ratio for the two bGH mutant mice to control littermates was -0.6.
In the case of the triple mutant, we generated ~.0 founder mice that express the mutated bGH gene. The growth ratio between the transgenic and nontransgenic littermates ranged from 0.58 to 1.00. The degree o~ suppression of growth was directly related to the serum levels of the mutated bGH. Three founders haws been bred that pass the trait to offspring; ~50% of these ~ffspring are positive for the gene and possess the corresponding small phenotype.
It has been demonstrated that many activities of GH
are meda.ated t~rou,gh a family of peptides known as in~lin-like growth factors (IGF)~ in particular ~GF-1, which is believed to be produced primarily in the liver following GH binding to its recept~r(e). (Bee ~ruesch, et al., Ann. Rev. Physiol., ,~~7~e~,~36? (19s5) ; Zapts et al., T3~rm. Res., 24:12.-130 (~.9s6) ) .
1GF-l has been sh~wn to decrease GH production in the pituitary b~ a classical negati~cre feedback mechanism. (Leung, et al., endocrinology, 119:189-96 (1986)). One hypothesis to explain the growth suppression in pBGHlO~5-M8 transgenic mice is that ' b~H-M8 is aetive as an in vivo antagonist to mouse GH (mGH), tYa~~eby suppressing mouse IGF-1 prdductic~n. Tf this is true, .hen ane would e~sect not only a reduction in serum mouse IGF-1 levels in ~aGH M8 transgenic mice but also an increase in mGH
production in the pituitary. We have found that the 2GF-1 leirels in the serum of the "~mal1" transgenic mice are- ~50%
s. .. . . , v . ...d ~..~ , ... . .. . . ... ...
. .".....,... . .... . ...., ._.., .,...".........., .. ';',a:S r . ., . . ..
... ..... ,.... ..... . , ':.e ~ .. ..

VV~ 92/19736 P~f/US92/03532 - 3~ -those of normal non-transgenic mice while mice containing wild type bGH (large mice) have approximately 2x the IGF-1 levels of non-transgenic mice. Results from immunoblot analysis of whole .
pituitary glands taken from bGH-MS .transgenic mice, bGH
transgenic mice, and their nontransgenic littermates suggest .
that the pituitary glands in those growth-suppressed mice contain higher levels of mGH relative to their nontransgenic littermates. In contrast, mGH levels in bGH transgenic mice were largely depressed because mouse serum IGF1 levels were increased up to twice as much, as levels in serum of their nontransgenic littermates. Palmiter, et al., Science, 2~2:809-1~ (1983) . Together, these results indicate that the altered bG~I molecules are acting as an antagonist to endogenous mouse GH. Thus, it is the first example to our knowledge of an a~n vivo growth hormone antagonist and the first examgle of uncou~aling of growth-promoting and receptor-binding activities of GHs .
1e 5: Gore~ng of other te3.~ of b~GH a~ hGH
~y simil~;r procedures; muteins of bGH and hGH with al~.er~.~.iox~s a~n the th~.rd al~aha helix have been prepared and tested for secretion in L cells, and, in selected cask their effect on the growth of transganic mice, with the following reat~ltg .
The mutants are de~~r~.bed by giving the original amino acid, gas positi~xa in the amin~ acid sequence of bGH, and the g~placement amino acid, with the amino acids set forth according to the internationally accepted single letter code.
~eOrge, et al., Protein Seq. Data Anal., 1.:27-39 (198'7).
A fir9t set of mutated bGH:genes, when expressed in transgenic mice, resulted in animals with a growth ratio sianilar to that of mice which express wild type bGH (i.a., -1.59 - 1.72). y~Te hive referred to these analogs as "full functional agonises" (Table I).

W~ 9~/~973G P(.'T/US92/~3532 ~~.~~~~9 _ ~5 -A second set of mutated bGH genes, when expressed in transgenic mice, resulted in mice with a growth ratio smaller than those animals which express wild type bGH (i.e., between 1,29 -1.35). We refer to these bGH analogs as "partial.
functional agonists" and have listed them in Table II.
A third set of mutated bGH genes, when expressed in transgenic mice, resulted in animals with a growth ratio sam3lar to nontransgenic mice (i,e., - 1.0). isle refer to these analogs as "non-functional agonists" (Table III), A fourth set of mutated bGH genes, when expressed in transgenic mice, resulted in mice with a growth ratio of between 0.5'~ and 1.0 (Table I~T). The growth ratio of the mice was negatively correlated with the serum level of the bgh ans:l.og, i.e., a~ the serum level of the bgh analog increased, the growth ratio of' the animals decreased. Thin correlation is sh~r~an graphically in Figure 13.
Also, these analogs, when expressed to NIH-3T3-preadipoeytes, did not result in stir~tulata.on of preadipocytes diffe~e~.t~:ation; however, native GH will pro~y,~a~e this differentiation (Fig. 12). In faGt~ these analogs will antagon~.ze the abil:ity ' of wild ty~~ GH t~ promote preadipocyte .differez~tiatiora (Fig. Zl) : ~Te have referred to these s:nalogs a,s "functional a?r~ta~oni~~s" (Table IV) .
hake also generated transgenic mice which express either ~nrild ty~ae h~~I, hGH G120A., hGH G120R and hGH GZ20'W (Table V) : Mice which express hGH G~.2~A s~aow a growth enhanced ~her~oty~e similar to mice which express wild type IaGH (Table ~)~ We call this hGH analog a a'functional agonist." In contrast, substitutican. of R or W f~r G at position 120 in hGH, and subsequent expression ~.r~ transgenic mice, results in arximals with a growth ratio between 0.?3 and 0.96 (Table V);
aid whose level of serum hGH analogy is negatively correlated W~ 92/19736 PC°'~'/Ug92/03532 N~~~
_ 36 _ with the growth phenotype; i.e., as the serum levels of these hGH 120 analogs increase, the growth ratios decrease. This correlation is shown in Figure 14. Therefore, like the bGH.
analogs which act as "functional antagonist," we termed these hGH 120 analogs as "functional antagonist.°' It is important to-note that the glycine residue in bGH at position 119 is the homologue of the glycine residue in hGH at position 120. They are both located in the central portion of the third a-helix.
A subset of bGH analogs is presented in Table VI in which we have evaluated their ability to be secreted following transfection of tkxe mutated DI~7A into mouse L cells. Transgenic animals have not been generated which contain these mutated DNAs.
The mutant K112L, R3.14~nT shows the effect of expanding the hydrophobic face of the helix. This mutant affects animal growth much a~ does wild ype growth hormone.
Th.e mutations R114P, H118P and L121P (and various combina.ti~ns thereof) apparently destroy the alpha helix (Prol~.ne is-a strong alpha he~.ix breaker.) The growth-related viol~gica7: aot~.vity is abolished. The mutation H~~6G is a special ca~ea glycin~'is a helix breakex, but position 126 is at bye ehd of the helix so the noranal biological activity is r~t~a.ned: ~T~.th G119P, howe'~erone strong helix breaker was substututed for ~n even stronger ones the alpha helix was aPparea~tly presexved:
The 'third alpha helix of wild type growth hormone diverges from a perfeet amphiphilic alpha helix at three poszti~n~: F~.rst, at 117, Glu is a hydrophilic amino acid in the hgrdroph~biC dace: ~SecOnd,.at 119, Gay is a neutral amino acid in the h~rdrophilic face: Finally, at 122, Ala is a hydrophoba.c am~.no acid in the hydrophilic face. The mutations E117L, G119~t ~.nd A~.22D, separately or in combination, increase !y~ 92f19736 P("TfLJS92f03532 ..
the amphiphilic character of the helix. G119R additionally increases the alpha-helical tendencies of the sequence.
Our initial hypothesis was that the growth-inhibitory activity of the mutants G119R and E117L/G119R/A122D was associated with the increased amphipathicity of the third alpha helix. We have since developed evidence that the amphipathicity of the third alpha helix is largely irrelevant to that activity, (1) The single E117L, like wt bGH, produced large aa~imal s (2) fi~utant G119P produced the small animal phenotype even though proline is as .hydrophilic as glycine.
(3) Mutant G119L produced the small animal phenotype wen though leucine is hydrophobic and therefore disrupts the hydrophilic face of the helix.
(4) ~utara.t E111L/G1~.9W/R~.25L produced the small ana.mal phenotype even though all three mutations disrupt the hydrophilic face of the helix, (S) the single A122D produces a mutein which has no effect on growth.
Thus., ~.n one emb~diment, the gresent invention relates to ~rautatioh~ of.the third alpha helix which result in ~rowth~~nhibitory activity yet reduce or leave unchanged the aniphiph.ilic character of the helix.
diagonal growth: hox-mone antagonists may be idehtified by'sy~tematically varying the codon corresponding to G~.19 in bGH, so as to eacpres~ the ~.~ other mutants ' having a single- amino acid change at, this p~sitican. This is readily ac~complish~d b~ synthesizing oligonucle~tid~s differing from hose set forth in Example 1 as ccadon 11.9 so as to encode the desired alternative amino acid. Similarly, one may alter the hoanologous glycine reside in the third alpha helix of other GHs, ~.g~e the G12° of hGH. Hy similar means, variations of VV~ 92119736 PC'~'/B.JS92/~3532 the codons corresponding to other amino acids of the third alpha helix of a GHlare investigated.
le ~6 , ~,t~.chol.esterolem~.c acti~.ty of Gr~wth ~~~one .~intagon~.sts Procedures for Clinical Chemistry Tests:
Blood samples were obtained from mouse tails. The samples were allo~red to clot at room temperature for 5 minutes and were then centrifuged and the serum was collected and frozen at -2a°C unt~.l analysis. Total Cholesterol (TC), Triglyceride (TR.), Glucose (GL), and Hlood Urea Nitrogen (BUN) Were analyzed on a Kodak Ektachem DT 6a Analyser using dry, multilayered, self-contained elements specific for each test.
~1. of serum was pipetted on individual slides specific for each test ~.nd were analyzed using co7.orimetric measurement by reflectance spectrophotometry methods and compared to dai~.y ~~,ala~ty Coxatrol reference samples.
Results:
There ~.s no significant difference in blood glucose, seraun urea/na.trogen and ~~rum triglycerir~e levels betwn bGHi~i~
tr~sge~ic miss an~i them n~ntrans~enic littermates . ~zowever, total serum ch~le~te~ol lev~l.~ in bGH-118 transgenic mice are sigax~.fa:cantly decreased (P<0, 053 as compared to their nontran~geniic littermates and bGH transgenic mice.
~.e ~: 1.n '~a.tr~ ~~.o~eay fir Gr hormone Antogon.:i..st l~cti~ay Studies of growth hor~nnne have shown that it promotes the forrn~t~.on of adipese from preadip~se 3T3 cells. iKurikawa, et al., Cell 29.789 (192). Glycerophosphate dehydrogenase (GPDI3) has been used as a differentiation marker for this G~h.nduced adipose conversion. ~rl~.~e and Green, J. Biol . Chem. , ~54:~73-75 (1979): TTixon and Green, Endocrinology, 114:527 ~JV~ 92/19736 P~'/L1S92/~3532 2~.~~~.~9 (1984); Pairault and Green, Proc. Nat. Acad. Sci. (USA), ?6:5138 (19?9) .
We have adapted this assay to. determine whether a bGH
mutant acts as a GH antagonist. Both bGH and bGH-M8 bind to receptors on these preadipocytes with a Kd value of 1(?mM. When exposed to native secxuence bovine growth hormone (30 pM) and cultured for seven days, the preadipocytes differentiate and GPDH activity is stimulated. If the bGH mutant is added to culture medium containing wild-type bGH, there is a dosedependent reduction in GPDH activity and, therefore, presumably, a.n adipose conversion (Figure 11).
This as~a~r is a convenient screening tool for identifying p~tential GH antagonists.
1e ~:
Mice transgenic for the wild type bGH gene are known t~ develop progressive se~rere glomeru.losclerosis and increased glomerular size. Doi, et a~.. , lam: J. Path. , 137: 541-52 (199); Resce; et ~1.; hab. Invest., 65: 601-5 (1991); Doi, et al ~ : ~- J. Path. ; 131: 398-40~ (19f39) ; see also St~ewart, et al:, Ez~docrinolagy, 1~0: 405-41.4 (192). This is not merely a function Of body ize, as bGH-Ml.l mice (i.e., hl~3.P, E126G
mutant). whose x~ut~nt bGH does ncat enhance growth, also exl~a.bit ~lumeru~.oscle~osis . In bGH-M8 (G119I2) mace, however, which had reduced ~eruzn IGF-1, body size, a.nd glomerular size relative to nontra~sgenic mice, glomerulosclerosis was absent.
Sumr~a.ry of growth ratio comparisons between transgenic mice express~.ng bGH analogs and their non-transgenic littermates at 6 to 8 weeks of age.

~,~ ~zm ~73s we rms~zro~s~z Table 1. Transgenic mice which express the following bGH
analogs exhbited phenotypes similar to transgenic mice which express wild type bGH (we have termed these analogs "full functional agonists")*
bGH Analogs n Mean Growth Ratio SD

WT-bGH____ ? 1.61 0.14 bGH-111A 2 1.72 ____ bGH-K112L 12 1.?0 0.19 bG~f-~11~&'W 12 1.70 ' 0.19 bGH-L116A 6 1.?1 0.16 bGH-E117L 18 1.68 0.18 bGFi-A122T 10 1.6? 0.16 bGH-R125L 3 1.61 0.18 bGH-E126G 4 1.59 0.14 * There is no correlation between serum levels of these bGH
analogs and the growth phenotypes. These mutated bGH genes are e~garegsed in mouse L cells and the secretion pattern is similar to the wild type bGH~
T~&ale ~I. Transgenic mice which express the ~ollw~ring bGH
analogs eibit~d p~.en~t~~s smaller than transgenic mice which egress wild hype bGH, however, large than non-transgenic mice (we have termed these ana~.ogs "parta:al functional agonists)*
bGH n dean Growth Ratio SD
t~!°-bGH ? 1.61 0.14 D115s~3. 3 1. 35 0 .15 L123I 3 1.29 0.13 There is no correlation between serum levels of these bGH
analogs and the grov~th phenotypes. These mutated bGH genes are expressed in and secreted ~by mouse L cells with the pattern ~~D 92IR~736 ~CTlUS92~'~353 ~~.~~~.~9 - 4~ -similar to wild type b~H.
H'ote that for the purposes of Tables x-vz, the characterization pf a mutein as °'functional~' or "non-functional" is in the context of its effect on growth.

!fi'O 92/19736 PCT/IJS92/~3532 ~1 ~~~ . , _ 42 _ Table III. Transgenic mice which express the following bGH
analogs exhibited phenotypes similar to their non-transgenic.
littermates (we have termed these analogs as '°non-functional agonists")'"
bGH Analogs n Mean Growth Ratio sD
~~.z4P, EZasP ~ l . ax. o . a9 L121P,E226G ~.1 0.94 0.06 13a22D 3 0 . 9 0 0 .11 '" There is n~ correlation between levels of bGI-i analogs in serum and the growth phenotypes. These mutated bGH genes are expressed :in and secreted by mouse L cells with the ea~ceptions of bG~I-K1~.4P; El~.sP and bGH-L121P, E1.26G which are not secreted by mouse L cells.

'VV~ 92/19736 F~.'I'/~JS92/~3532 ~3 -Table TAT. Transgenic mice whisk express the following bGH
analogs exhibited phenotypes smaller than non-transgenic littermates (we hare teraned these analogs as "functional antagonists" )'' bGg Serum bGH Growth Ratia Analogs Animal Sex (ug/ml) (Two Month) #

E1~.7L, G11,9R, A12~D 6 F 3.4 0.58 M 3.3 : 0.69 32 F 3.7 0.57 51 F 5.1 0.63 55 M 2 . ~. 0 . 85 65 ~ 0.6 Z.0 ~7 F 0.6 0.87 70 F ~.3 0.70 F 2.6 0.70 89 F 1.8 0.85 G119R 2~ M 0.5 0.93 2g* M 0.9 0.88 49'' ~ 6.0 w'~.60 53 M 1.5 . 0.85 94 F 0.~ 0.98 138 F 3.0 0.74 G119P ~ F 2.0 0.81 G119K 110 M 0.5 0.8~

12 M 0.4 . 0.95 18 F 4.0 0.78 26 F 5.0 0.59 G129L~ 23 F . &.5 0.81 27 M 0.5 1.0 W~ 92/I9736 ~C.°I'1US92103532 - 44.-G119W 16 . M ~.0 0.64 6119~ 1~ M 0.5 0.96 15 M 0.~ 0.90 22 N! 6 . 0 0 . ?5 23 M 9.5 0.90 The level o~ mouse growth suppression is correlated. with serum levels of analogs (see Fig. 13). These mutated bGH genes ire expressed in and secreted mouse L cells. The secretion pattern is similar to wild tie 1~GH.

~~ 92119736 P~'/~JS92/~3532 Table V. Summary of transgenic mice which express hGH genes encoding single amino acid substitutions at position 120* (hGH-G120A is a "full-functional agonist". hGH-G120R and hGH-G120Tn~
serve as "functional antagonists") hGH animal # Sex Serum hGH Growth Ratio Analogs (ug~ml) (Two Months) WT-hGH n=7 1..62 ~ 0.15 G120A g5 M 3.9 1.48 6 F 21.5 1.76 G120R 20 F ?8.5 0.79 4g F 1.5 0.96 68 M 3.4 0.73 73 F 0.8 0.93 F ~ o ~ 0 . 93 G120~1 1~ M 5.5 0.82 39 M ' 2 . 7 0 . 7?

56 F 2.0 0.83 l~GT3 Gly 1~.9 ass et~uivalentto hGH Gly 120.
in a position Th~xefore, we xe~er .o hGH Gl~r x.20 consa.stently with the liaer~tur~.

* * Th.e lw~l of growth suppress~.on is correlated with serum .levels of hGI3 analogs (See F~:g.
3.4) 'W~ 92/19736 P'CT/US92l03S32 Table V'I. Summary of mutated UGH genes expressed in mouse L
cell without transgenic mice dais.
UGH Analogs L-Cell Secretion Wild fiype bGH +
K114P _ E11~P

E~.l?,G119R ~ +

r'rll~i ,~'.22D +

~I109D,Y110D,L116R +

E11.1L, G11~W +

L121R;1~d124K +

E111L; G1.19Tn1, L121R-, lvl124K +

D115V +

~115G +

'6TZ09D; Y~.~.1D, L3:16R;'L121R, hl124K -E~.1~.L, G11,9W, R125L +

E17L1L, G11.9W, L121R,1~124K +

VlO~D,Y110D,L11,6K,R125L +

Table vII '1~ ' ~ni.ce mutarats Ma. ~rezla~s KZZ~w) r~~o cxi~~~;E~l~P) X11 (&121PoE126G) T?~4 ( E 117I.~ ) 1~I6 (Orl~.9R) (22~) N.t'7 (E~.~.'lL, G119R) (E11?L,A3:22D) ~a (~z~.~L~:~y~R;~i2an)

Claims (17)

The embodiments of the invention in which an exclusive property or privilege is claimed are defined as follows:

1. Use of a growth hormone antagonist peptide or protein of at least 50 amino acids, comprising an alpha helix which is at least 50% but not entirely identical to the third alpha helix of bovine growth hormone (bGH) or human growth hormone (hGH), wherein at least one of the differences from said third alpha helix of bGH or hGH is at a residue corresponding to bGH119 or hGH120, said peptide or protein being capable of antagonizing at least the biological activity of a human or animal growth hormone in the manufacture of a composition for the prevention or treatment of a condition of said human or animal, which is characterized by the level of said biological activity being excessive and wherein the condition is diabetes.

2. Use of a recombinant DNA molecule capable of expressing in a suitable host cell, a growth hormone antagonist peptide or protein of at least 50 amino acids, comprising an alpha helix which is at least 50% but not entirely identical to the third alpha helix of bovine growth hormone (bGH) or human growth hormone (hGH), wherein at least one of the differences from said third alpha helix of bGH or hGH is at a residue corresponding to bGH119 or hGH120, said peptide or protein being capable of antagonizing at least the biological activity of a human or animal growth hormone, in the manufacture of a composition for the prevention or treatment of a condition of said human or animal, which is characterized by the level of said biological activity being excessive and wherein the condition is diabetes.

3. Use according to claim 1 or 2 wherein the biological activity causes damage to microvascular tissues such as retinal endothelial cells.

4. Use according to claim 1 or 2 wherein the biological activity contributes to the development of glomerulosclerosis.

5. Use of a growth hormone antagonist peptide or protein of at least 50 amino acids, comprising an alpha helix which is at least 50% but not entirely identical to the third alpha helix of bovine growth hormone (bGH) or human growth hormone (hGH), wherein at least one of the differences from said third alpha helix of bGH or hGH is at a residue corresponding to bGH119 or hGH120, said peptide or protein being capable of antagonizing at least the biological activity of a human or animal growth hormone in the manufacture of a composition for the prevention or treatment of a condition of said human or animal, which is characterized by the level of said biological activity being excessive and wherein the condition is excessive levels of serum cholesterol and the peptide or protein has a hypocholesterolemic effect.

6. Use of a recombinant DNA molecule capable of expressing in a suitable host cell, a growth hormone antagonist peptide or protein of at least 50 amino acids, comprising an alpha helix which is at least 50% but not entirely identical to the third alpha helix of bovine growth hormone (bGH) or human growth hormone (hGH), wherein at least one of the differences from said third alpha helix of bGH or hGH is at a residue corresponding to bGH119 or hGH120, said peptide or protein being capable of antagonizing at least the biological activity of a human or animal growth hormone, in the manufacture of a composition for the prevention or treatment of a condition of said human or animal, which is characterized by the level of said biological activity being excessive and wherein the condition is excessive levels of serum cholesterol and the peptide or protein has a hypocholesterolemic effect.

7. Use of a growth hormone antagonist peptide or protein of at least 50 amino acids, comprising an alpha helix which is at least 50% but not entirely identical to the third alpha helix of bovine growth hormone (bGH) or human growth hormone (hGH), wherein at least one of the differences from said third alpha helix of bGH or hGH is at a residue corresponding to bGH119 or hGH120, said peptide or protein being capable of antagonizing at least the biological.

activity of a human or animal growth hormone in the manufacture of a composition for the prevention or treatment of a condition of said human or animal, which is characterized by the level of said biological activity being excessive and wherein the biological activity is tumorigenic.

8. Use of a recombinant DNA molecule capable of expressing in a suitable host cell, a growth hormone antagonist peptide or protein of at least 50 amino acids, comprising an alpha helix which is at least 50% but not entirely identical to the third alpha helix of bovine growth hormone (bGH) or human growth hormone (hGH), wherein at least one of the differences from said third alpha helix of bGH or hGH is at a residue corresponding to bGH119 or hGH120, said peptide or protein being capable of antagonizing at least the biological activity of a human or animal growth hormone, in the manufacture of a composition for the prevention or treatment of a condition of said human or animal, which is characterized by the level of said biological activity being excessive and wherein the biological activity is tumorigenic.

9. The use according to any one of claims 1 - 8, wherein said peptide or protein comprises amino acids corresponding to amino acids 96 - 133 of bGH or hGH.

10. The use according to any one of claims 1 - 8, wherein said peptide or protein comprises alpha helices corresponding to helices 1 - 4 of bGH, hGH or porcine growth hormone (pGH).

11. The use according to any one of claims 1 - 8, wherein.
said alpha helix is at least 80% identical to the third alpha helix of bGH or hGH.

12. The use according to any one of claims 1 - 8, wherein said peptide or protein is at least 50% identical to a vertebrate growth hormone.

13. The use according to any one of claims 1 - 8, wherein said peptide or protein is at least 50% identical to a mammalian growth hormone.

14. The use according to any one of claims 1 - 8, wherein said peptide or protein is at least 50% identical to a hGH or bGH.

15. The use according to any one of claims 1 - 8, wherein said peptide or protein is at least 80% identical to a vertebrate growth hormone.

16. The use according to any one of claims 1 - 8, wherein said peptide or protein is at least 80% identical to a mammalian growth hormone.

17. The use according to any one of claims 1 - 8, wherein said peptide or protein is at least 80% identical to a hGH or bGH.