CA2190752A1 - Modified insulin-like growth factors - Google Patents

Abstract

Modified forms of insulin-like growth factor (IGF) are provided which demonstrate improved pharmacological and biological properties. These modifications include IGF muteins produced by substituting or adding a cysteine in the amino acid sequence of native IGF as well as such muteins attached to polyethylene glycol (PEG) at the free cysteine site. The present invention further provides methods of making such modified forms. The IGF-PEG conjugates can be formulated into pharmaceutical compositions and used for the therapeutic treatment of IGF associated conditions.

Description

WO 9513~003 1 ~ S' 't MODIFIED INSUL~ E GROWTE[ FACI~RS
FiPI~I of the I
This invention relates to the ' ~ of ~/ly~Li~kD~ and more ~Li~.ul~uly to the "~ of insulin-like growth factors and to methods of making and using such modified yulyy~lJLhh~
r ~ ~ of ~hP ~vP~
The insulin gene family, comprised of insulin, relaxin, insulin-like growth factors 1 ^Dnd 2, and possibly the beta subunit of 7S nerve growth factor, represents a group of structurally related puly~ Lid~D whose biological functions have diverged as reported in 10 Dull, et al., ~ 310:777-781 (1984).
T " I;IU; growth factors I and 2 aGF-l and IGF-2) are about 7-8 kilodalton proteins that are structurally related to each other and to insulin. IGF-1 and IGF-2 s_are about 70% amino acid identity with each other and about 30% amino acid identity with insulin. IGF-I and IGF-2 are believed to have related tertiary structures as reported in PCT
Application Publication No. WO 90100569, published on January 25, 1990. The structural similarity between IGF-I and IGF-2 permits both to bind to IGF receptors. Two IGF
receptors are known to exist. IGF-I and IGF-2 bind to the IGF type I receptor, while msulin binds with less affinity to this receptor. The type I receptor l r '' ~ binds IGF-I and is believed to transduce the mitogenic effects of IGF-I and IGF-2. IGF-2 binds 20 to the type I receptor with a 10-fold lower affinity than IGF-I. The second or type ~ IGF
receptor, r " ~y binds IGF-2. Receptor binding is beheved to be necessary for the biolodcal activities of IGF-I and IGF-2.
IGF-I and IGF-2 are mitogenic for a large number of cell types, including fibroblasts, I " ~U~D, endothelialcellsandosteoblasts (bone-formingcells). IGF-I and 25 IGF-2 also stimulate .~ of many cell types, e.g., synthesis and secretion of collagens by, ' ' IGF-I and IGF-2 exert their mitogenic and cell ;~
effects by binding to the specific IGF cell surface receptors. IGF-I also has been shown to inhibit protein catabolism in vivo, stimulate glucose uptake by cells and to promote survival of isolated neurons in culture. These properties have led to IGF-I being tested as a therapeutic agent for a variety of disease indications as reported in Froesch et al.,

2 1 ~ 0 7 5 2 F~~ S ~ ~c ~
;n r..~ ry and ~ nl 254-260 ~aay/June 1990) and Cotterill, Clinical r.~ Il" ~ vy 37~ 16 (1992). In addition lO specif~c cell surface receptors, there exist at least six distinct IGF binding proteins aGFBP-l through IGFBP-6) that circulate throughout the body. These proteins bind IGF-I and IGF-2 with high affinity. The birlding S of IGF-1 and IGF-2 to binding proteins reduces the action of these IGFs on cells by preventing their interaction with cell surface IGF receptors. IGF bmding proteins, IGFBP-3, also function to prolong the circulating half-lives of IGF-1 and IGF-2 in the blood stream. In the absence of IGF binding proteins, the half-life of IGF-1 in blood is less than 10 minutes. In contrast, when IGF-1 is bound to IGFBP-3, its half-life in blood is lengthened to about 8 hours. The circulating half-life of IGF-1 bound to the other smaller binding proteins is about 30 minutes as reported in Davis et al., I~Lf rn.l,.. .;...~l,~vy~ 123 469-475 (1989); Guler et al., Acta r-~- ' ' 121:753-758(1989); and T~ l ' et al., J. of r ~.- i l..~. 123:461-468 (1989). When IGF is bound to bindimg proteims, it is unable to bind to the IGF receptors and is therefore, no longer active in the body. Decreased affu ity to binding proteins allows more of the IGF
to be active im the body. Situations when this decreased affnity to brnding proteins may be useful include, for exarnple, cachexia, . u,;~, and peripheral i..,.,... ' -r the therapeutic utility of IGF can be modifled by the presence or absence of these IGF binding proteins, vhich may potentiate or inhibit the beneficial effects of MF. The levels of certain IGF binding proteins can vary greatly, depending upon the disease state. For example, IGFBP-1 levels are very high in diabetes patients, whereas they are nearly ~ ' ' '- in normal patients as reported in Brismar et al., J. of rn.l.,..; ~ 11:599-602 (1988); Suil~i et al., J. of t'l-rn~ y and Metabolism, 66:266-273 (1988); and Unterman et al., Biochem.
. Biophys. Res. Cornm., 163:882-887 (1989). IGFBP-3 levels are reduced in severely ill patients such as those tbat have undergone major surgery as repo~ted in Davies et al., L
r~ , 130 469-473 (1991); Davenport et al., J. t'lin r ~ MP~q~, 75:590-595 (1992). The reduced levels of IGFBP-3, and consequent shorter circulating half-life of IGF-l, may contribute to the cachexia (weight loss) seen in these patients.
Insuhn-like growth factor 1 aGF-l), also known as " C, has long been studied for its role in the growth of various tissues. Its role as a useful therapeutic agent for several disease conditions has been suggested. Ci" ~ 'y reduced levels of IGF-1 -Wo s5/32003 2 ~ ~ 0 7 5 2 ~ 5.c were found in 23 patients with varying extent and severity of burns as reported in Moller et al., Bums, 17(4):279-281 (1991). A marked rise in serum type m ,,.~ G' '', a marker of bone formation, occurred after one week of ~ ' ' ' " of '~, produced IGF-1 to patients with dwarfism otherwise non-responsive to growth hommone as reported in Laron et al., '1 ' T'-'- ' ' , 35: 145-150 (1991). The effects of the infusion of IGF-I in a child with Laron Dwarfism is described in Walker et al., Th,- New T~.n~
Tm~ of M~irinA 324(21):1483-1488 (1991). Increased weight g~un, nitrogen retention and muscle protein synthesis following treatment of diabetic rats with IGF-I or IGF-I
having a deletion of the first three amino acids ordinarily found in IGF-I (referred to as "(desl-3)IGF-l") were .~ ' in Tomas et al., r ~ J.. 276:547-554 (1991).
Growth restoration in msulin-deficient diabetic rats by ' of .- .1~, produced human IGF-I is reported in Scheiwiller et al., ~hlL, 323:169 (1986). IGF-I
and (desl-3)IGF-I enhance growth in rats after gut resection, as reported in Lemmey et al., J. Physiol.. 260 (F-' ' Metab. 23) E213-E219 (1991). A . ' of platelet-derived growth factor and insulin-like growth factors, including IGF-I, enhanced "'G~ ''1;"" in beagle dogs as reported by Lynch et al., J. ~
16:545-548 (1989). The synergistic effects of platelet-derived growth &ctor and IGF-1 in wound healing are reported in Lynch et al., P~oc. I~Tqt7 A~ l Sci.. 84:7696-7700 (1987).
'rhe effects of IGF-I and growth hormone on 1 ~ " I bone growth in vitro are set forth in Scheven and Hamilton, .A~`tq rn ~ 1"~ (cove~h~ 124:602-607 (1991). In vivo actions of IGF-I on bone formation and resorption in rats are shown in Spencer et al., ~, 12:21-26 (1991). The use of IGF-1 and IGF-2 for enhancing the survival of non-mitotic, cholinergic neuronal cells in a mammal is described rn U.S. Patent 5,093,317 to Lewis et aT. In addition, PCT Application Publication No. WO 92111865 published on July 23, 1992, describes the use of IGF-I for the treatment of caTdiac disorders.
Various ' ~ to the naturally occurring or wild-type IGF-I have been described. For example, the naturally occurring variant of IGF-I missing from the first three N-terminal amino acids, (desl-3)IGF-I, was discovered in cerebral spinal fluid and - in colostrum as reported in Sara, et al., Proc.l~Tqtl Ar-q~l .Cri 83:4904~907 (1986) and Francis et al., Rin.~Amirql Jmlr~ql 251:95-103 (1988). In vitro studies have shown that this variant is equipotent to IGF-I in binding to IGF cell surface receptorS and in stimulating cell X,, Thus, the fTrst three amino acids of IGF-1 appear to be Wo gs/32003 2 1 9 0 7 5 2 r~.,u~: c'oc. 10 ~
I for the binding of IGF-I to its specific cell surface receptors. (Desl-3)IGF-Iwas found to have greatly reduced affinity (100-fold less) for certain IGF binding proteins, specifically IGFBP-I and IGFBP-2, as reported in Forbes et al., Ri~h~ m Biovhys. Res.
Comm., 157:196-202 (1988); and Carlsson-Skwirut et al., r Bi~-h~ys. A~
1101:192-197 (1989). The binding of IGFBP-3 to (desl-3)IGF-I also is affected, being reduced by two to three fold. Animal stndies have shown that (desl-3)IGF-I, when given by continual ' infusion, is more potent than IGF-I m stimulating a mlmber of anabolic functions, such as growth, reported in Cascieli et al., J. E ' ' ~y, 123:373-381 (1988) and Gillespie et al., J. ~ Iu~ -y, 127:401-405 (1990). The enhanced properties of (desl-3)1GF-1 are believed to result from its reduced affinity for IGF binding proteins, reported in Carlsson-Skwirut et al., ~- - Bi~-h!ys. A~ 1101:192-197 (1989). The reduced affuuty of (desl-3)IGF-I for IGF binding proteiAs results in (desl-

3)IGF-I having a shorter circulating half-life than wild type IGF-I as reported in Cascieri et al., J. rA~ py. 123:373-381 (1988). Therefore, (desl-3)IGF-I must be - ' l by continual infusion or by multiple daily injections in order to effoct its enhanced potency.
PCT Application Publication No. WO 89105822 publishod on June 29, 1989, describes other ,~ of IGF-I. This application describes substituting the third amino acid from the N-terminal end of the naturally occurring IGF-I with glycime, glut~mime, leucine, arginine or Iysine to form IGF-I muteins. This reference however doos not teach replacing the third amino acid with cysteine.
The potential therapeutic usefulness of IGF-I and (desl-3) IGF-I is limited by their short circulating alf-lives to situations when the proteins can be r ' by continual infusion or by multiple daily injections to achieve thoir maximum therapoutic potontial. As am example, Woodall et al., T~ M~-h Res., 23: 581-584 (1991), reports that the same total dose of IG~-I r four times daily by ' injection was far superior in stimulating growth im mutant lit/lit mice (growth hormone deficient mice) than was the same total dose ~ ' ' once daily.
There are many cases in wmch it would be preferable to administer IGF-1 in a single onc~a-day injection or in a single injection given once every several days. For injeciable drugs, patient . , " is expected to be higher for drugs that can be r- ~ once a day rather than several times a day. In order for IGF-I to be ~ woss/320n3 2 l 9 3 7 52 r~llu ~r~
lly effective when given once a day or once every few days, methods must be developeo to increase its circulating half-life.
Increasing the molecular weight of a protein by covalently bonding am iner~ polymer chain such as ~ .h~k ..., glycol (PEG) to the protein is known to increase the circulating 5 half-life of the proteirl in the body. See, for example, Davis et al, F~ 1 polymp~
Polymeric r~ - and F' '~ for p~inmPrlir~ll Use. p. 441-451 (1980).
However, since multiple PEG molecules can bind to each protein molecule, and because there are typically a large number of sites on each protein molecule suitable for bindirlg to several PEG molecules using known methods, there has been little success in attaching PEG
to yield l- O reaction products. See Goodson et al, Biotechnologv, 8:343 (1990),and U.S. Patent 4,904,584. This lack of site attachment specificity can give rise to a number of problems, including loss of activity of the protein.
Thus, a need exists for prolonging the circulating half life of IGF without , its usefulness as a therapeutic agent. The present invention satisfies this need 15 by increasmg the molecular weight of the IGF. This is . ' ' ' by providing muteins of IGF suitable for thiol-specific attachment of PEG to IGF and PEG conjugates formed from such muteins.
Q of the InvP ~inn The mvention relates to various modifled forms of IGF. One type of modified IGF,20 referred to as mutems, is produced by replacmg cysteine residues for specific amino acids in the wild type molecule, or by inse~ting cysteine residues adjacent to or bet~veen amino acids in the wild type molecule. Cysteine residues which are not involved in disulfide bonds are considered to be "free". Cysteine residues can be substituted or inserted in regions of the IGF molecule that are exposed on the protein's surface. For example, the 25 non-native cysteime can be inserted or substituted within the first or last twenty amino acids of wild-type IGF-I. This non-native cysteine is expected to be free and to serve as the attachment site for the pol,~,LI~ , glycol (PEG) molecules to IGF, resulting in PEGylated molecules. In some cases, however, during refolding of the mutem or during reduction of the mutein before reaction with PEG, the non-native cysteine can become mvolved in a 30 disulfide bond and thereby free a native cysteine for PEGylation. Attachment of the PEG

wo gsl32003 2 1 9 0 7 5 2 F~l/u..: cr r A ~
molecule to a mutein creates a further modified for n of IGF, or IGF-PEG conjugate of defmed structure, where the PEG is attached to the IGF at a free cysteine residue.
Thus, the present invention is directed to a PO~ glycol (PEG) conjugate comprising PEG and a mutein of IGF, and IJ~ti~,ul~ly IGF-1, where the PEG is attached to the mutein at a non-native cysteine. PEG can be attached to the free cysteine through a thiol-specific activating group including, for example, maleimide, sulfhydryl, thiol, triflate, tresylate, aziridine, exirane and S-pyridyl. A suitable PEG can have a molecular weight of 5 kDa, 8.5 kDa, 10 kDa or 20 kDa. The PEG conjugate of tùe present invention can also contain a second protein to form a dumbbell. Methods of making the PEG
conjugates are also provided.
Moreover, PEG is known to increase the circulating half-life of a protein in thebody. Thus, IGF can be ~' ' to patients less frequently with equal or better .,rr~L;~ than in the past.
The present invention is further directed to muteins of IGF l.~u Liuul~iy those having IS a non-native cysteime in the N-terrninal or C-terrninal region of the mutein. The muteins can be produced by I ` methods.
Also provided in the present invention are I ' ' l~ comprising the IGF-PEG conjugate and methods of using the IGF-PEG conjugate to treat a patient havimg or potentially having an IGF associated condition.
~ P.~r.lrfinn of fhP Tnv,onri~n The present invention is directed to modified forms of insulin-like growth factors aGF) that provide beneficial proper~ies not exhibited by wild-type IGF. The modified forms of IGF include muteins of these growth factors, containing at least one free cysteime.
Conjugates containing the IGF muteins attached to pol~.LhJh,..~, glycol (PEG) are also 25 considered modified forms of IGF.
Terms used throughout the ~ and claims are defined as follows:
The term "IGF" refers to any pol~id~ that binds to the IGF type I Receptor, including, for ex~unple, IGF-l, IGF-2, (desl-3)1GF-1, (R3)1GF-1 which is a mutein having a non-native arginine at residue number 3, and insulin. This hormone family is described in Blundell and Humbel, ~a~L, 287:781-787 (1980). Due to this common receptor w0 95/32003 2 1 9 ~ 7 5 2 r~"u ~r ^
binding, the teachings of the present invention which are described with respect to IGF-1 are mtended to encompass IGF-2, des(1-3) IGF-l, (R3)IGF-1, and insulin also.
The term "wild type IGF" refers to the, "~ ' or naturally occurring IGF. This term is used , ' ~ ' 'y with "IGF," "naturally occurring IGF," or "native IGF." The S term "wild type IGF" also refers to native IGF to which a methionine residue has been added at the N-termimus.
The term "wild type IGF-1" refers to the I ' ~ ' or naturally-occurring 70 amino acid form of IGF-1. This term is used ' , "y with "IGF-1," "naturally occurring IGF-1," and "native IGF-1." The term "wild type IGF-1" also refers to native IGF-1 to which a methionine residue has been added at the N:
The term "~.u.. ~ " refers tO that which is not found im the native molecule.
The term "IGF-PEG conjugate" refers to an IGF molecule attached to a IJUI~
glycol molecule. This is also referred to as a "Peg conjugaten.
The term "N-terminal region" refers to ~ , the first twenty ammo acids IS from the N-termimus of IGF or an IGF mutein, and up to twelve amimo acids preceding the frst amino acid of the N-terminus of IGF.
The term "N; " refers to the frst amimo acid at the N-ter~ninal region im the sequence of wild-type IGF, for example, glycine in IGF-1.
The term ~C-terminal region" refers to ~I,UI~ / the last twenty amino acids 20 from the C-termimus of IGF or an IGF mutein and up to twelve amino acids following the last amino acid of the C-termimus of IGF.
The term "C-termimus" refers to the last amimo acid at the C-termmal region in the sequence of wild-type IGF, for example, alanine m IGF-I.
The term "mutein" refers to a modified form of IGF, which has been modified to 25 contain a non-native cysteine.
The term "retain biological activity" refers to having at least 10% of the mitogenic activity of wild type . ' IGF as measnred by the relative amount of 3H-thymidine , ~ into UMR106 rat . cells, in the absence of IGF binding protein-1, usimg the assay described herein. The muteins and the conjugates of the present invention 30 retain biûlogical activity.
The term "activating group" refers to a site on the PEG molecule which attaches to the muteim.

woss/32003 2! 90~52 ~ 5~ l^
, The term "I' lly acceptable carrier" refers to a ylly..;~lu~ y ' ', aqueous or non-aqueous solvent.
The term "free cysteine" refers to any cysteine residue not involved in an i l.- . -l. '-. disulfidebond.
The term "IGF associated condition" refers to an existing or potential adverse Ly ~;ulù~ l condition which results from an over-production or ' , ' of IGF, IGF binding protem or IGF receptor, , ~,, or inadequate binding of IGF to binding proteins or receptors and any disease in which IGF: ' aLeviates disease symptoms. An IGF associated condition also refers to a condition in ~hich - ' of IGF to a normal patient has a desired effect.
The term "patient" refers to any human or animal in need of treatment for an IGFassociated condition.
The IGF muteins oFthe present rnvention are produced by modifying wild-type IGF,particularly at the N-terminal or C-terminal region of the protein. Such . ~.~;ri. ~ can be l or additions of at least one cysteine residue. An IGF mutein can be produced by replacing a specific amimo acid with a cysteine, such as, for example, ~, one of the frrst or last four amino acids of IGF-1 with a cysteine residue. The amino acid sequence of wild type IGF-1 starting from the N-terminal end is: G P E T L
CGAELVDALQFVCGDRGFYFNKPTGYGSSSRRAPQTG
IVDECCFRSCDLRRLEMYCAPLKPAKSA~SEQIDNO. 1).
Other ' include, for example, adding at least one cysteine residue in Front of the first or after the last amino acid of IGF. Fûr example, a cysteine residue can be inserted in front of and adjacent to the ftrst amino acid of IGF. For muteins produced by E. cûli, the non-native cysteine can appear between Met and the frrst amino acid of IGF.
A free cysteine residue can also appear in a group of about twelve or less amino acids inserted before the first or after the last amimo acid of IGF to form a longer IGF mutein.
In IJ~u~uLuly useful " the non-native cysteine residues are located in regiûns of the IGF-I molecule exposed to the protein's surface. The N-terminal region, for example, is involved in the binding of the IGF to binding proteins, but is not involved in binding of IGF to cell surface IGF receptors. ~
An IGF-I mutein of the present invention is also referred to as "a cysteine mutein of IGF-I." The non-native cysteine residue can act as the attachment site for covalent wo 95/32003 2 ~ 9 ~ 7 5 2 r~ r - 10 linkage of the activating group on the pvl.~ l.jh..~ glycol. Altemavively, the non-native cysteine can become involved in disulfide bonding thereby freeing a native cysteine residue for thiol-specific attachmem to PEG. The newly created molecule comprising the cysteme mutem of IGF with the PEG attached is referred to as a "PEG conjugate of IGF~.
S The IGF muteins of the present invention can be prcpared by methods well known to one skilled in the art. Such metbods include mutagenic techniques for ~ ~; ormsertion of an ammo acid residue, as described, for example, in U.S. Patent 4,518,584, l ' herein by reference. The mutems produced by mutagenic techniques can then be expressed as ' products according to procedures known to those skilled in the10 art. The muteins can ~llt~ V~,ly be synthesized by Cu_llliv~ methods known in the art. The IGF muteims can also be prepared according to the methods atld techniques described in the examples set forth below.
The present invention also provides IGF-PEG conjugates and methods of making such conjugates by attaching the IGF muteins to pol~,lhjk,l.c glycol to mcrease the circulating half-life of the molecule in the body as well as decrease its afftnity to IGF
bmding prvteins.
In the present invention long chain polymer units of pvl~.hgl~llC glycol (PEG) are bonded to the mutein via a covalent attachment to the sulfhydryl group of a free cysteine residue on the IGF mutem. Various PEG polymers with different molecular weights, 5.0 kDa (PEG~ ), 8.5 kDa (PEGUoo), 10 kDa (PEGIoooo), amd 20 kDa (PEG200o~) can be used to make the IGF-PEG conjugates. In order to obtain selectivity of reaction and ~" reaction mixtures, it is useful to use ' ' ' polymer units that will react specifically with sulfhydryl or tniol grvups. The functional or reactive group attached to the long chain pvl~ glycol polymer is the activatmg grvup to which the IGF
mutein attaches at a free cysteme site. Useful activating groups include, for example, maleimide, sulfhydryl, thiol, triflate, tresylate, aziridme, exirane, or 5-pyridyl.
In another I t, PVIJ~hjh~IIC glycol (PEG) polymers containing two activating grvups can be used to create "dumbbell" type molecules containing two IGF
- muteins attached to one molecule of PEG at each end of the PEG molecule. For example, 30 PEG 1.;1 ' (a pUl~, ~k,llc glycol polymer containing a maleimide activating group on each end of the PEG molecule) can be used to create these "dumbbell" type molecules.
These dumbbell molecules can also comprise a single IGF mutein covalently attached via wos~/32003 21 9 07 52 r~ ,5,.~
PEG to a second protein or peptide of different structure. The second prooein or peptide can be, for example, a growth factor such as platelet-derived growth factor, or fibroblast growth factor.
One skilled in the art can readily deoermine the appropriate pH, of S protein, and ratio of protein:PEG necessary to produce a useful yield of either mono-pegylated IGF- I aGF-PEG), or dumbbell IGF- I aGF-PEG-IGF, IGF-PEG-PDGF, or IGF-PEG-FGF) using ~;u..~v...~u.~l methods known to one skilled in the alt for making these , . . . .
The invention present also includes l ' l ~ . The IGF muteins 1û and PEG conjugates can be in a 1' ~ 'ly-acceptable catrier to form the ~,h ".. . ~..';. Al , ~' of the present invention. The term ~l' "S~
acceptable carrier" as used herein means a non-toxic, generaUy inert vehicle for the active ingredient, wmch does not adversely affect the ingredient or the patient to whom the is _' ' Suitable vehicles or catriers can be found in standard ~' 'texts,forexample,inr 's~' 'Sri~n~ 16thed.,Mack Publishing Co., Easton, PA (1980), i . ' herein by reference. Such carriers include, for example, aqueous solutions such as L- ' buffers, phosphate buffers,Ringer's solution and ,uh~;vlv~ saline. In addition, the carrier can contitin other l' '~S,-acceptable excipients for modifying or v the pII, osmolarity, viscosity, clarity, color, sterility, stability, rate of ~' ' or odor of the r~
The l' ', , can be prepared by methods krlown in the art, imcluding, by way of am example, the simple mixing of reagents. Those skilled in the art will know that the choice of the ,ul~ ' carrier and the appropriate preparation of the, l depend on the intended use and mode of In one i ' ' t, it is envisioned that the carrier and the IGF muoeirl or conjugate constitutes a ,u~ ;vlvo;~lly- , ' ' ~lv....1~., ' ' The primary solvent in such a carrier can be either aqueous or non-aqueous in nature. In addition, the carrier can contilin ULh~ -acceptable excipients for modifyimg or ,, the pH, osmolarity, viscosity, clarity, color, sterility, stability, rate of ~" ' or odor of the ' ' Simil3rly, the catrier can contain still other I ' ' g l'~-acceptable excipients for modifying or l" the stability, raoe of ' ' release, or absorption of the IGF mutein or conjugate. Such excipients are those substances usually . .

woss/32003 2 1 93752 1~ s~
and customarily employed to formulate dosages for parenteral ~ ;.... in either ulut dose or multi-dose form.
Once the, ' I, . has been r 1~ it can be stored im sterile vials as a solution, suspension, gel, emulsion, soLid, or dehydrated or IyophiLized powder.
5 Such 1( 1 may be stored either in a ready to use form or requiring 'y prior to r' ' The preferred storage of such r ~ is at at least as low as 4rc and preferably at -70C. It is also preferred that such containing the IGF mutem or conjugate are stored and ' ~ at or near J~h~lo~h~l pH. It is presently believed that ' im a r~ at a high pH
(i.e. greater thfm 8) or at a low pH (i.e. Iess than 5) is l ' ' ' The manner of ' v the r ~ ' contaming the IGF mutem or conjugate for systemic delivery can be via ' , ~ 0~15~ oral, intranasal, or vaginal or rectal - r r " y . Preferably the manner of: ' of the COntSdilliiîg the IGF muteins or conjugates for local delivery is via ;.. l. - i;. .. l~., ' l, orimstiLlationorinhalationstotherespiratorytract. Inaddition it may be desirable to administer the IGF muteins or conjugates to specified portions of the alimentary can~ either by oral - ' of the IGF muteims or conjugates in an or device For oral: ' the the IGF muteins or conjugates are r ~ ~ The , ' ' IGF muteins or conjugates may be formulated with or without cceptable carriers . ~!~ used in the . , " _ of solid dosage forms. Preferably, the capsule is designed so that the active portion of the r ~ ' is released at that poimt m the gastro-intestinal tract when ~ ,/ is maximized and pre-systemic .l. v,~ ;.... is minimized. Additional excipients may be included to facilitate absorption of the IGF muteins or conjugates. Diluents, flavorings, low melting point waxes, vegetable oils, lubricants, suspending agents, tablet ~" v v agents, and bimders may also be employed.
Regardless of the manner of r ' ' ' ' " , the specific dose is calculated accordmg to the, . body weight of the patient. Other factors in !' ' ' ' ~ the appropriate dosage can imclude the disease or condition to be treated or prevented, route of: ' and the age, sex and medical condition of the pateint. In certain ' ' the dosage and -' is designed to create a preselected woss/32003 21 q07 52 r~ c--10 ~
range of the IGF muteins or conjugates in the patient's blood stream. It is believed that the of circulating .,. . ~ of the IGF muteins or conjugates of less than 0.01 ng per ml of plasma may not be an effective ~ , while the prolonged of circulating levels in excess of 100 ~g per ml may have S undesirable side effects. Further refmement of the ~ .nc necessary to determine the appropriate dosage for treatment involving each of the above mentioned r '- isroutinely made by those of ordinary skiU in the art and is ~ithin the ambit of tasks routinely performed by them without undue, l , especiaUy in light of the dosage ' ~ and assays disclosed herein. These dosages may be ascertamed through use 10 of the established assays for .' ~ l,, dosages utilized in ; with appropriatedù~_ ...,~u...._ data.
It should be noted that the IGF mutein and conjugate l ' ' ~
described herem may be used for veterinary as weU as humam d~ ;- .-- - and that the term "patient" should not be construed in a limiting manner. In the case of veterinary "1~ 'A, the dosage ranges should be the same as specified above.
The l' ~ f the present iAvention can be used to treat a patient having or potentiaUy having an IGF associated condition. Some of these conditions can include, for example, dwarfism, diabetes, periodontal disease and U~tW,UU.J~;~. The ': , . of the present invention can also be used to treat a condition in which ' of IGF to a normal patient has a desired effect; for example, using IGF-I to enh~mce growth of a patient ûf normal shture.
The foUowing examples are intended to illustrate the present imvention and are not intended to be limiting.
EXAMPL~ I
A. C of the IGF-I ge~e The IGF-1 gene was assembled in two shges. InitiaUy, the DNA sequence encoding IGF-1 was joined to DNA sequences encoding the secretory leader sequence of the E. coli OMP A protein (ompAL). This gene fusion was rnnctr~ t~ in order to determine whether IGF-I could be efficiently secreted from E. coli. A second construct, in which IGF-1 is expressed as an int~.ll ' protein in E. coli, was created by deleting DNA sequences wo9sl32003 2 1 ~0752 F~"~,~ s~6 1~
encoding the OmpA leader sequence and replacing them with DNA sequences that allow inrr~rf~ r expression of IGF-l.
B. r - of tbe OmpAL-IGF-I gene fusion The four synthetic oligonucleotides labeled OmpAlU:
5'GATCCGATCGTGGAGGATGATTAAATGAAAAAGACAGCTATCGCGATCGCA3' (SEQ ID NO. 2), OmpA2U: 5'GTGGCACTGGCTGGTTTCGCTACCGTA
GCGCAGGCCGCTCCGTGGCAGTGC3' ~SEQ ID NO. 3), OmpAlL: 5'CAGTGC
CACTGCGATCGCGATAG~ ATTTAATCATCCTCCACGATCG3 ' (SEQ
ID NO. 4) and OmpA2L: 5'GCACTGCCACGGAGCGGCCTGCGCTAC
GGTAGCGAAACCAGC3' (SEQ ID NO. 5), were annealed pairvise (lU + lL amd 2U
+ 2L) and the pairs ligated together. All four of these ~ , ' ' were synthesizedusing DNA ~ purchased from Applied Biosystems (Models 391 and 380A). The ligation mixture was then digested with the restriction enzyme Haem. The resulting BamE~/EIaem restriction fragment coding for a ~ l start signal and the first 21 amino acids of the ompA signal sequence was purified. This DNA fragment was mixed with BamElI + PstI-digested PUC18 DNA (. ~;~I r available from Boehringer Marmhein r -~ 's, T " . ', IN) and the two synthetic 'i~ ~ ' [IGF-I
(1-14) U + L] 5'CCGGTCCGGAGACTCTGTGCGGCGCAGAACTGGTTGAC
GCTCTGCA3' (SEQ ID NO. 6) and 5'GAGCGTCAACCAGTTCTGCGCCGC
ACAGAGTCTCCGGACCGG3' (SEQ ID NO. 7) were ligated together. The ligation mixture was used to tr.msform ~QIi strain ~M109 (. ~Iy available from New England Biolabs, Beverly, MA) and individual colonies isolated. These plasmids (OmpALlGF-lpUC18) have a i ' ' start signal followed by DNA sequences encoding the OmpA signal sequence amd the first 14 amino acids of IGF-I.
An M13 phage containing DNA sequences encoding amino acids 15 through 70 of IGF-I was created by ligating together the two ~ ,' y pairs of ~-"~, ' ' (tGFIU + IL and IGF2U + 2L) 5'GTTCGTATGCGGCGACCGTGGCTTC
TACTTCAACAAACCGACTGGCTACGGTTCCAGCTCTCGTCGTGCACCGCAG
ACTGGTATC3' (SEQ ID NO. 8) and ~ ~lC(JluAACGATACCAGTCTGCGGTGC
ACGACGAGAGCTGGAACCGTAGCCAGTCG~ l l l ~ l l ~AAGTAGAAGCCACG
GTCGCCGCATACGAACTGCA3' (SEQ ID NO. 9) and cloning the DNA fragment into woss/320o3 21 90752 P~llu~ ~c 10 Pstl + Hindm-digested M13 mplg DNA (commercially available from New England Biolabs, Beverly, MA). Double-stranded DNA was purified from a phage clone and the PstI/Hindm fragment encoding ammo acids 15-70 of the IGF-1 protein were isolated. This DNA fragment was ligated together with Pstl + Hindm-digested plasmid OrnpALlGF-lpUC18 DNA and used to transform E. coli strain 3M107 (. l~y available from GIBCOBRL, t~ h ..~l....v, MD). TheBamHI/HindmfragmentcontainingthelGF-I gene fused to the OmpAL sequence was isolated and cloned into the BamHI + Eindm generated site of plasmid pT3~-2 (described in PCT Application pubhication WO 91/08285 pub~ished on June 13, 1991). The completed plasmid containing the ompA~IGF-I gene fusion is called pT3XI-2 010C(TC3)ompALlGF-1.
C. I' of the ?' - ~1 IGF-l gene The BamHI/Hindm fragment contaming the OmpA~IGF-I gene fusion described above was purified from plasmid pT3XI-2010C(TC3)ompALlGF-I and digested with Hinf~.
The all 200 bp Hinff/Hindm DNA fragment was mixed with the almealed, , . y synthetic -'iv ' ' (MetIGFlU + lL) 5'GATCCGATCGT
GGAGGATGATTAAATGGCCG~l~CGGAG3' (SEQ ID NO. 10) and 5'AGT
CTCCGGACCGGCCATTTAATCATCCTCCACGATCG3' (SEQ ID NO. I l) arid ligated with BamHI + Hindm-digested plasmid pT3XI2 DNA, and used to transform E. coli JM107. The completed plasmid construct is called 010C(TC3)IGF-lpT3XI-2 and contains an extra alanine residue in between the initiator methionine and the beginning of the IGF-1 sequence. The BamHI/Hindm fragment containing the mutant IGF-1 gene was isolated and ligated into the BamE~ + Hindm generated site of plasmid pT5T (described in Nature, Vol. 343, No. 6256, pp. 341-346, 1990). The ligation mixture was used to trarlsform E.
coli BL21/DE3 described in US Patent 4,952,496 and the resulting individual colorlies were isolated. This construct was named 010C(TC3)IGF-lpT5T.
The extra alanine codon was removed by in vitro v In vitro was performed usmg a Muta-Gene kit purchased from Bio-Rad T ' ' (R' ' 1, CA). The v procedure followed was essentially that described in the i -~. . I ;., -that accompany the l~it. Plasmid 010C(TC3)IGF-lpT3XI-2 was digested with BamHI +Hindm and the 200 bp DNA fragment contair~ing the mutant IGF-1 gene was purified and cloned into the BamE~ and Hindm sites of plasmid M13 mpl9.

wo gs/32003 2 1 9 0 7 5 2 l'~ u~,S C :^
Uracll-containing single-str~mded template DNA was pre~oared follu . . illg ,UI~J.~""LiOII
of the phage in E. coli strain CJ236 (supplieA with Muta-Gene Kit purchased from Bio-Rad T ' - ,Richmûnd,CA). The,-li, ' ' usedfor , hadthesequence:
5'GATGATTAAATGGGTCCGGAGACT3' (SEQII) NO. 12). The ~ reaction 5 producf was used to transform E. coli strain JM109 and individual plaques picked.
Double-stranded replicative form DNA from imdividualphages was isolated, digested with BamHI + Hindm and the 200 bp fragment containing the IGF-1 gene purified. The purified DNA was cloned into the BamHI + Hindm generated site of plasmid pTST and used to transform E. coli strain BI~1/DE3. One bacterial cûlony with the correct plasmid was named 010(TC3)mutlGF-lpT5T. Several isolates were sequenced, and all were correct.

C of IGF-1 Muteins Several muteins of IGF-1 were cn"cfn~rf~A Three of the muteins replaced each of the first three amino acids of IGF-I with a cysteine residue. These muteins are referred to as Cl, C2, and C3, I~i.,U~i~.,ly. A fûurth muteim introduced a cysteine residue between the N-terminal methiûnine residue and the first amino acid of IGF-I. This mutein is referred to as -lC.
The -lC, C1, C2 and C3 muteins of IGF-1 were made using the 1,~ chain reaction (PCR) technique as described below. The starting plasmid used for the Ie, was 010(TC3)mutIGF-lpT5T, which is described in Example 1. This plasmid contains DNA sequences encûding an irlitiator methionine followed by the sequence of the natural humam IGF-1 protein. Mut mt IGF-1 DNA sequences were amplified frvm tms gene using a 5' ~ iv ' ' that contained the desired mutation and a 3' ~i~, ' ' of correct sequence. The 5' ~ ' were designed sû that they , ' the first methionine of the gene as part of am Nde I restriction enzyme cleavage site (CATATG). Each ~ ~ 'i,, ' ' contained the desired mutation followed by lS to 21 nucleotides that were a perfffl match to the IGF-1 gene sequences im plasmid 010(TC3)mutlGF-lpTST. The 3' ~'iv ' ' was 25 nucleotides long and was designed to encode the last 6 codons of IGF-1 amd to cûntain the Hind m site that follows the stop codon.

wo 95/32003 2 l 9 0 7 5 2 r~l~L~ ~ , ~
In addition, a wild type clone was made in the same malmer using the Ndel site of pTST (described above) as the first I ' This wild type clone is designated 85p-11.
The, 'i~, ' ' used to construct the -1 C, C1, C2 and C3 muteins and the 85p-11 wild type clone are set forth in Table 1.

-r . _ ~
, ' J
-' J
X ~
V ~ O .
~, m E~ 3 . m P~ m m ~
x O ~ ~
a w0 95/32003 2 19 0 7 5 2 r "u.
Polymerase chain reaction (PCR) was performed in 100ul reactions containing 20 mM Tris pH 8.8, 10 mM KCI, 6 mM (NH4)2S04, 1.5 mM MgC12, 0.1% Triton X-100 using 20 pmole of each oligo and ~ 'y Ing of plasmid 010(TC3)mutIGF-lpT5T
as template DNA. 0.5ul (1.25 units) of Pfu pol~ (Stratagene, San Diego, CA) was 5 added after the first .l ~ ,.. step with the tubes held at 65C. The reactions were oYerlayed with 2 drops of mineral oil at that time. The reactions were cycled 30 times for I mim. at 95C, I min. at 65C, and I min. at 72C in an Ericomp Twinblock'Y thermal cycler (Ericomp, San Diego, CA). After the last cycle the reactions were held at 720C for 10 minutes.
After PCR, 80ul of the reactions were phenol extracted one time then ethanol l ' The r ~ DNA was . ' ' in 80ul of TE buffer (10 mM Tris-HCI pH 8.0, I nlM EDTA) and 20ul was digested with Nde I and Hind m and ~,I~L.I I ' ' on a 1.5 % agarose gel. The amplified DNA bands that ran at 210bp were eluted using NA45 paper (Schleicher and Schuell, Keene, NH) 15 according to the r 'S ' ' '' The eluted DNA was ~ in 20ul of TE buffer and 2ul was ligated to gel-purifled Nde I and Hind m digested plasmid pTST in a volume of 20ul. Plasmid pT5T is described in Example 1. The ligation reactions were used to transform ~.~QU stram BL21/DE3 and colonies selected on LB agar plates containing 50ug/ml of ampicillm. Milu plasmid DNA preps were made from several 20 colonies from the j r " plates. The DNAs were digested with Eco RV and Hmd m to determine which j r ' contained IGF-I DNA inserts. Plasmids containing IGF-I DNA inserts were sequenced to verify that the inserts were correctly mutated (the entire IGF gene was sequenced for each mutein).
P~ - ' ' y growth studies were performed by growing a - ~ ; V~; i r 25 for each mutein in Luria Broth + 12ug/ml t~ a~ " to an ~, OD60o of about 1Ø Isopropyl-beta-D-i'-~ y ' aPTG) was added to I mM to induce expression of T7 ~,ol~ and the subsequent ~ ;.... and translation of the IGF
muteins. ~A~ '!~ 0.1 OD unit of cells were Iysed in SDS sample buffer by boilmg for two mimutes and ~ .' ' on a 16% P~IJ~ SDS gel. The gel was 0 stained with Coomassie blue. IGF-1 protein bands of the expected size, which is / 7-8 kDa, were observed in the lanes loaded with induced cells for each mutein as well as for the wild-type control.

1~ wo 95/32003 2 19 0 7 5 2 r~l,u~ c -- ^
Muteins in which cysteine was substituted for the amino acids at residues 11, 12, 15, 16, 55, 64, 65, 67 and 69 of the wild type IGF-l were also prepared using PCR. The PCR template was either the full-length plasmid from clone 85p-11 or the NdeI-Hindm fragment from clone 85p-1 I containing the wild type IGF-I gene.
S For muteims Cll, C12, C15 and C16, 5' ~" ~ ' ' were synthesized that ' the furst methionine as part of the NdeI site (CATATG) and each contained the mutation desired followed by 21-24 nucleotides that were a perfect match to clone 85p-11. These, 'i, ' ' are set forth im Table 2. For the 3' end of the gene, a 25-mer ~ aGF(262p)25) that nnatched the last 6 codons of IGF-I plus the Hindm site followmg the stop codon was used. IGF(262p)25 is set forth in Table 1.

W09V3~003 21~0~2 r~ o A
~`
.
~' . _ ô~ 6 ~ ~ o~ ~ .
t 8 : c~
S '~
" ~

V'~

wo 95t32003 2 1 9 0 7 5 2 r~ l,u~s loG~ ^
For muteins CSS, C64, C65, C67 and C69, 3' oli~,. ' ' that I ~ the Hindm site (AAGCTT), the stop codon (TAA) and containing the mutation desired followed by 21-27 nucleotides that were a perfect match to clone 85p-11 were synthesized.
For the S' end of the gene, a 28-mer, li, ' ' (IGF(85p)28) that matched the first 5 . 7 codons of IGF-I plus the initiation codon for methionine ill~,ul~ ' into an NdeI site was used. These ~ ' ~ ' ' are set forth in Table 3.

W~9~32003 21 90752 r ~u~ ~ ~

~", ~ . `' ~ , .
. .
& ~ ~
YY
.
u ~ . ~ u ~
y ~ ~ ~J l ~ -U ~ ~ ~
, ~ o ~ WO 95/32003 2 1 9 7 5 2 ~ ~, ., ~ ~ c Polymerase chain reaction (PCR) was performed in 100ul reactions containing 20 mM Tris p~ 8.8, 10 mM KCI, 6 mM (NH4)2SO4, 1.5 mM MgC12, 0.1% Triton X-100 usmg 20 pmole of each oligo, rr ' ' ~1 lng of template DNA, 200uM each of dATP, dCTP, dGTP, TTP, 20pmole of each oligo primer, and lul (2.5Tl) of Pfu polymeraseS (Stratagene, San Diego, CA). The reactions vere cycled 30 times for I mirl. at 95C, 1 min. at 65C and I min. at 72C in a GeneAmp PCR System 9600 (Perkin Elmer Cetus).
After the last cycle the reactions were held at 72C for 10 min.
After PCR, the reaction mixtures were purified by passing through ~'' S1 100 columns (Clontech Lab. Inc., Palo Alto, CA). Purified PCR fragments were digested with NdeI and lIindm and the--210bp bands eluted from 1.5% agarose gel using NA45 paper (Schleicher and SchueD, Keene, NEI) according to the r ~S ~ The double cut and gel-purif~ed DNA fragments were ligated to similarly cut and gel purifled pTST plasmid at 15C for 18 hrs. The resulting ligation mixtures were used to transform E. Coli strain DHSa and plated on LB + ampiciDin (SOuglml) plates. Plasmid DNAs were prepared from several colonies using the Qiagen plasmid kit and the entire IGF-I gene sequenced with the Taq DyeDeoxy Terminator Cycle Sequencmg Kit (Applied Biu~y ) A correct construct was selected for each mutein (clones Cl 1-1, C12-3, C15-1, C16-4, C55-1, C64-1, C65-1, C67-2, C69-1) and j r ~ mto E. Coli strain BL21/DE3 for expression.
r y expression studies of the muteins were performed by growing two t, j r ' for each mutem in LB + t~LI,.~,' (12uglml) to an lj r OD600 0f about 0.6-0.8. IPTG was added to a final . of ImM and the ceDs aDowed to grow for an additional 2 hrs. The IPTG induces the expression of T7 polymerase and the subsequent and translation o~ the IGF muteins.
~, '~, 0.1 OD unit of the cells (both uninduced and IPIG-induced) were Iysed in SDS sample buffer containing B , ' ~ ' and ~1~, , ' ' on 15 % SDS-PAGE.
The gel was stained by Coomassie blue and the IPTG-inducible IGF mutein bands of the expected size were observed in lanes loaded with induced cells for each mutein.

w095/32003 21 907 52 P~

r Refolding, and I'; ~ of MuteiLfs Although the following is written for the C3 mutein, the satne procedure applies to other muteins l , ' ' by the instant invention. The only difference is in the star~utg S cells used.
E coli cells expressing the IGF-1 C3 mutein were suspended in Buffer A (50 rnM
Tris, pH 7.5, 20 mM NaCI and l r~tM " ' ' ' (DTI~ at a . of ~fO ml Buffer A to 10 g cell paste, and disrupted at 1800 psi using a French pressure ceil (SIM
Inc., Urbana IL). The suspension was centrifuged at 20,000 X g for 30 min, 0 and a1iquots of the pellet and supernatant analyzed by SDS-PAGE. A major band to the IGF-1 C3 mutein was present in the pellet, but not the C~ f~ f The pellet was suspended in Buffer A at a . of 40 ml Buffer A to 10 g cell paste, and re-centrifuged at 20,000 X g for 30 ~un. Tilis wash procedure was repeafed 2 times. The final pellet containing the IGF-I C3 mutein was suspended in 6 M guanidine, 50 rn~I Tris, pE 7.5, 6 mM DTT at a: of 25 ml to 10 g cells using a ground glass I ~ . The suspension was incubated at room i r ' for 15 r un. The '._;1 protein was removed by ~ ;r..L,.-:;.... at 20,000 X g for 30 mirt. The funal of the C3 mutein was 1.0 mg/ml. SDS-PAGE analysis of the pellet and ~ ~ following the procedure of Example 2 showed that IGF-I C3 mutein was 20 present in the supernatant only.
The denatured and reduced IGF-I mutein was subjected to the following three-steprefolding procedure:
1) Oxidized ~ , the mixed-disulfide producing reagent (GSSG), was added to the supernatant to a fmal of 25 rnM, and incubated at room i , for 15 min.
2) The solution was then diluted 10 fold graduaUy with 50 mM TAs, pH 9.7, amd l' ~ and cysteine were added to funal of lrnM and S rltM, I~,~l;.~ly. Filtal . of protein was 100ug/ml.
3) The refolding mixture was incubated overnight at 40C, and then centrifitged at 20,000 x g for 15 min. SDS-PAGE analysis of the pellet and ~ Wo gs/32003 2 ~ 9 0 7 5 2 PCTIUS95/06540 showed that the supematant was composed of relatively I O IGF-I
C3 muteim.
Aliquots (50ul) of the supernatant were diluted to 200ul with Buffer B (0.05%
TFA), injected onto a reverse phase column (RP~, I x 250mm, Synchrom, Lafayette, IN), amd eluted with 80 % ~ 1~, 0.042 % TFA (Buffer C) using a linear gradient (increase of 2% Buffer C/min) at a flow rate of 0.25 ml/min.
A smgle major peak ~ , refolded IGF-I C3 mutein eluted at 26.5 mm.
Refolded IGF-1 C2 mutein eluted at 26.0 min. The retention times of refolded IGF-1 C3 and IGF-I C2 muteins shifted to 32.2 min and 31.7 min, I~~ iv~l~, after beimg completely reduced and denatured in 5 M guanidine, 50 mM Tris p:EI 7.5, 100 mM DTT.
These results imdicate that both the C3 and C2 muteins refold to a smgle major species under the conditions described. N-terminal sequence analysis of IGF-I C3 mutein eluting at 26.6 min gave the sequence: M G P C T L C (SEQ ID NO. 29) con}lrming that a cysteine residue has been substituted for glutamic acid im the 3 position of the N-terminal sequence of natural human IGF-I . An extra methionine residue is present at the N-terminus of the ' proteim expressed by E coli. N-terminal sequence analysis of IGF-I C2 muteim eluting at 26.0 min gave the sequence: G C E T L C (SEQ ID NO. 30) conflrming that a cysteime residue has been substituted for proline in the 2 position of the N-temlinal sequence of natural human IGF-l.
The refold . containing the C55, C64, C65, C67 or C69 muteins were analyzed by diluting 50 ul aliquots of the . to 200ul with Buffer B (0.05 % TFA), injecting the diluted , onto a reverse phase column (RP-4, I x 250mm, Synchrom, Lafayette, IN), and eluting with 100% ~1 ', 0.042% TFA (Buffer D) usimg a linear gradient (increase of 2% Buffer D/min) at a flow rate of 0.25 ml/mm. Two major distinct ,.~ I peaks (1:1 ratio of each), Pld & Pkll, eluted at 20.5 and 21.5 min, l~D~ . This pattem closely resembles the pattem observed for wild type ("WT") IGF-I (See, Meng, et al., J. Cl i Vol. 443, pp. 183-192 (1988)), specifically ~ ' herein by reference. The earlier eluting peak observed in the WT
refold has been shown to be an isomeric form of IGF-I with S-S ~ C6-C47, C48-C52, and C18-C61; whereas the later eluting peak is correctly refolded with S-S
,, C6-C48, C47-C52, C18-C61 (See, Raschdorf et al., Ri~m~Ai~

WO95/32003 21 90752 r~ c~
F ~ r S~ u~ uvy~ Vol. 16, pp.3-8 (1988), specifically I ' herein by reference).
An ~,~ ' peak of varying size elutulg at 21.5 - 23.0 min is also present in the refûld , SDS-PAGE analysis of this material shows that it contains 5 misfolded monomer and multimer forms of IGF-I. RP4 analysis of the refold contaming the C11, C12, C15 or C16 muteins showed the presence of several peaks eluting from 2û.5 - 21.5 min, as well as signific~mt (50 to 75 % of the total) amounts of apparently mis-folded material elutmg at 21.5 - 23.0 min. The retention time of the refolded muteins shifted to 27 min after being completely reduced and denatured im 5 M guanidine, 50 n~M
Tris pH 7.5, 100 mM DTT. Tbis ~' that the refolded forms "collapse" to a single ~uul~ tiJc after reduction of disulfides. The RP4 peaks, therefore, represent distinct forms of IGF muteins with different disulflde bonds.
l~XA~E 4 Isolation of Refolded IGF-I Mutein lS Refold mixtures (435 ml) prepared from 20g of E. coli paste containing either the refolded C3 or C2 mutein of Example 3 were . ' to lOOml, acidified to pH 5.5 with 5M HCI, dialyzed agamst 20 mM sodium acetate, pH 5.5, and loaded onto an S-Sepharose (PharmacialLKB, Piscataway, N" column (1.6 X 15 cm) previously ~
with the sodium acetate buffer described above. The bound protein was eluted with a 300 20 ml linear gradient from 0 to 0.5M NaCI. Three ml fractions were collected. A smgle major protem peak eluted at 0.2-0.3M NaCI. Aliquots (100 ul and 25 ul) of each peak were analyzed separately by reverse phase (RP-4 1 X 250 mm) and gel filtration -" ' ~ A ',Y (Superdex 75 3.2 X 300 mm, Pharmacia/LKB, Piscataway, NJ). Gel filtration effectively separates monomers from dimeric and multimeric forms of IGF present 25 m the refold . Fractions contaming ~ , monomeric (determined from gel filtration and RP-4 analysis) C3 mutein were pooled, ~ ' to 2.0 mg protein/ml, and 2.0-5.0 ml aliquots were loaded onto a Sephacryl S-100 (Pharmacia/LKB, Piscataway, N" gel filtration column (2.6 X 100 cm) previously ur, "' ' with 20 mM
sodium acetate, pH 5.5, 250 mM NaCI. The proteim was eluted at 2.0 ml/min, and aliquots 30 (lOul) of each fraction were analyzed by RP-4 reverse phase ' . ~,, ' y and SDS-PAGE followmg the procedure of Example 2. Fractions containing a sirlgle refolded , ~ wO 9sl32003 2 1 9 0 7 5 2 . ~
species of IGF-1 C3 or IGF-1 C2 mutein monomer of 95 % or more purity were pooled and r ' to 250 ug/ml. This material was assayed for bioactivity and reacted with an 8.5 kDa p~ glycol as described below.
For the Cll, C12, C15, C16, C55, C64, C65, C67 and C69 muteins, the refold . was dialyzed imto 20mM Tris, pH 7.4,: l ' 10-lSX and loaded onto a Superdex 75 gel filtration column previously r~ ' ' with the same buffer. The monomers were then pooled and loaded onto an RP-4 column (RP-4, 2.1 x 250mm, Synchrom), and eluted with 100% ~ nifnl~, 0.042% TFA ~3uffer B) using a lmear gradient (increase of 2 % Bufer B/min) at a flow rate of 1.0 ml/min. The C12, C55, C64, C65, C67 & C69 muteins refolded into 2 distinct RP-4 monomer peaks (1:1 ratio of each, PkI & Pk~) closely resembling the pattern observed for WT-IGF-I (See Meng at al., cited above). The C12 & C15 monomer fractions from Superdex 75 also contained significant amounts of apparently mis-folded material eluting after PII (21.5 - 23 0 min). The Cll, C15 and C16 monomer fractions contained multiple (3-6) peaks when analyzed by RP-4.
The monomer peaks were coDected separately and assayed for bioactivity and subjected to mass analysis (see below).
Mass analysis of the monomer peaks was performed using an API m obtained from Sciex, Toronto, Canada. Mass analysis was performed on both the refolded (disulfide bouD Dba) ~I-d ledu d b d bll bled IlWllOmeD. ~ odowiD Inbse~ we~e obbDed:

WO95132003 2 1 90752 r~ 5~

.
~ ~5 Ql pl 7ns wn6 wn3 ~Dsn U~
Cll p2 wn- usn S C12 pl mn n~s n~s U7~ ~03 n2 p2 n~s U7~
ns pl 77~6 n77 n7s Wl u~n ns p2 nu ~33 cls p3 n7s W7 0 a6pl 7736 ns7 nss U l ~z n6 p2 ns6 0~2 csspl 7n~ n~s n~7 U33 L32 c5s p2 n~7 U32 C6 pl ml ng2 nu L7~ n7s 5 C ~2 n~ n7~
a~s pl 7757 nn n7~ W2 ~59 (Ys p n7s n62 C67pl m3 W3~ W32 ~ 117 C67p W31 ~119 20 aispl nn3 Wl~ wls ~m2 ~mn c6s p2 W15 ~ml ~ W095/32003 21 ~0752 r ~ c- ^
These masses of the reduc~ed muteins match, within VA~ erwr, the expected masses for a pol~lidv with the indicated cysteine mutation. The masses of the refolded monomers match the expected mass of ~oly~,~liJv~ having 3 ' ' disulfide bonds and a single mixed disulfide of either cysteine - glutathione (cyS-s-s-r' ~.) or5 cysteine - cysteine (Cys-S-S-Cys). The mixed disulfides form during refoldimg and remain intact because there is no other cysteine residue present in the molecule available to form an ' ' disuhf~de.
A scale up ~, .. ;1;. ~;.... for the C69 mutein was also perfornled. The refold mixture fwm 8 gm of washed imclusion bodies (WIBS) was . ' ~ 10X to 400 ml, dialyzed against 20mM sodium acetate, pH 5.5. 200 ul aliquots was loaded onto am Sephacryl S-100 (Pharmacia/LKB, Piscataway, N~) gel filtration column (10 X 80 cm) previously ~ . ' with 20 mM sodium acetate, pH 5.5 and 250 mM NaCI. The fractions were eluted at 25 ml per minute, and aliquots (50 ul) of each fraction were analyzed by SDS/PAGE. Fractions contanting monomers were pooled.
To separate the two isomeric forms of C69, 200 ml of the S-100 monomer pool was diluted with 800 ml of 1.1 M ammonium sulfate-20mM sodium acetate, pH 5.5 (Buffer A) and loaded onto an Octyl Sephawse (Pharmacia/IXB, Piscataway, NJ) column (2.5 X 20 cm) previously ~, ' ' ' with Buffer A. The bound proteim was eluted with a 750 ml linear gradient from Buffer A to 50% Buffer B (50% ethamol-20 mM sodium acetate, pH
5.5). 12 ml fractions were collected. Two major protein peaks eluted at 25% and 32% of Buffer B. Aliquots (50 ul) of each pealA were analyzed by reverse phase (RP-4, I X 250 mm). Fractions containing ~ , correctly refolded (determined from RP-4 analysis) C69, eluting at--32-38 ~ Buffer B were pooled. Reverse phase analysis showed the correctly refolded C69 pool was 95 % or more 1 - . ~L,. A ~ This material was assayed for bioactivity (see below).
EXAMPLE S
Pegylation of Mtlteins The C3, C2 and C69 muteins were covalently joined to an 8.5 kDa ~ul~vLlljlv,.v glycol (8.5 kDa PEG) or an 20 kDa pV~ ..rlv.lv glycol (20 kDa PEG) having a maleimide 30 activatimg group in a two step process:

wo 95/32003 2 1 9 0 7 5 2 P~ . C : ~
1) The purifled IGF-l muteins were partially reduced with DTT in a 15 ml reaction mixture containing 2.3 mg (296 nmoles) IGF-1 muteim, 170 ug DTT (1110 nmoles) im 14 mM sodium acetate, 33 mM sodium phosphate, pH 7Ø The final of protem was 10 ug/ml, and the molar ratio of DTT:proteim was 3.75:1. For reaction with the C69 S mutein, 91ug DTT (592 nmoles) was used and the molar ratio of DTT:protein was 2:1.
The reaction mixture was incubated at room i . ci for 3 hours (5 hours for reaction with the C69 mutein) and terminated by the addition of 1.0 ml of lM sodium acetate, pH
5.5. The reaction mixture was dialyzed at 4C overnight against 20 mM sodium acetate pH5.5.
2) The par~ially reduced IGF-I mutein was reacted with either the 8.5 kDa PEG orthe 20 kDa PEG in a 20 ml reaction mixture containing 2.3 mg (296 nmoles) of proteim, 9.985 mg (1174 mmoles) 8.5 kDa PEG im 15 mM sodium acetate, 26 mM sooium phosphate, pH 7Ø The final: of proteim was 112 mg/l. The molar ratio of 8.5 kDa PEG:protein was 4:1; for reaction with the C69 mutein the molar ratio of 20kDa PEG:protein was 4:1. The reaction mixture was incubated at room ~ for 3 hours, and terminated by placing at 40C or freezing at -20C. SDS-PAGE analysis of the reaction mixture following the procedure of Example 2 showed that ~ , 50% of the partially reduced PEGylated C2 and C3 mutein was conver~ed to a mono-PEGylated species. The C3 and C2 20kDa-PEG conjugates migrated at a relative molecular weight of a~ 'S, 60kDa on SDS PAGE; The C3 and C2 8.5kDa-Peg conjugates migrated at a relative molecular weight of about 23 kDa on SDS PAGE. A~l, 'S, 20 % of thepartially reduced PEGylated C69 mutein was converted to a mono-PEGylated speciesmigrating at a relative molecular weight of 67 kDa on SDS PAGE.
Wild type IGF-l subjected to the same partial reduction conditions and PEGylation procedures did not become PEGylated.
EXAI~LE 6 n - of Pegyla~d Muteins The pegylated C2 or C3 mutein reaction mixtures (containing ~ S~ 100-200 mg protein) were dialyzed extensively at 4C agamst 5 mM CitliC acid, pH 2.6. The pegylated mutein was separated from the unPEGylated muteim usmg an S-Sepharose (Pharmacia/LKB, Piscataway, NJ) cation exchange column (2.5 X 25 cm) previously wo s~/32oo3 2 1 9 0 7 5 2 ~ 0 eq.li1ih~rA with 5 mM citric acid buffer, pH 2.6. The bound protein was eluted with a 2000 ml linear Oradient from 0 to 1 M NaCl. 25 ml fractions were collected. Pegylated C2 or C3 muteins eluted at 0.25-0.4M NaCI and the h~ oy' ' protein eluted at 0.8-0.9 M NaCI.
S Fractions containing the pegylated C2 or C3 muteins were pooled, . ' and rulLh~ I y .i~cl by Sephacryl S-200 gel filtration ' . O , ' .~ . 15 ml of the .fractions containing ~.yyl, ~, 20 mg of total protein was loaded onto a Sephacryl S200 (PharmacialLKB, Piscataway, NJ) column (2.6 X 100 cm) previously , ' ' with 20 mM sodium acetate, pH 5.5 containing 250 mM NaCI. The protem was eluted at 2.0 ml per min. The bulk of the material eluted with an apparent MW of 200 kDa.
The C69-PEG was separated from the " ,.~ ' C69 muteim by Q-Sepharose anion exchange ~ L. , . y . 100 ml of the reaction mixture contaming 11 mg of total protein was loaded onto a Q-Sepharose anion exchange column column (Pharmacia/LE~B, Piscataway, NJ) column a.6 X 100 cm) previously , "' ' With 20mM Tris, pH 9.0 (buffer A). The bound protein was eluted with a linear gradient of 20mM tris Ph 9.0, IM
NaCI (buffer B) at 5.0 mUmm. 10 ml fraction were collected. The C69-PEG eluted at 50mM NaCI, well separated from the unreacted monomer which eluted at 100mM NaCI.C69-PEG was pooled, ' to 13 ml amd loaded onto an S-200 gel filtration column a.6 x 100cm) previously , with 20mM sodium acetate, ph 5.5, 250 mM NaCI.
BA~E 7 Bioassay of Pegylaled Muteins r human metIGF-I (rIGF-l) (13achem California, Torrance, CA), various I, ,,.~' ' and PEGylated muteins were tested for their relative mitogenic activity and affinity for .~ ' msulin-lLI~e growth factor binding protem I ("IGF-BPI"), whichis described in PCT Application publication WO 89/09792, published on October 19, 1989.
A. Relative Mitogenic Activity The relative mitogenic (growth s~ ' , ` activities of the C3 and C2 muteins and pegylated C3 and C2 muteins were compared to that of wild type IGF-I by measuring the relative amount of ~H ~ . ihl~,v~ ' ' into rat U~'~,V~olw~d cells when varying amounts of the proteins were present under serum free conditions. The rat o t .,- --wo gs/32003 2 1 9 0 7 5 2 r~l~u~ s.~ - ~
cells (the UMR106 cell line; obtained from American Type Culture Collection, Accession No. CRL-1661, Rockville, Maryland) were plated at 5-6 X 104 cells in 0.5 ml of Ham's F12 Media (M~ ' ' Herndon, VA) containing 7% fetal bovime serum, 100 U/ml penicillin and 100 ~g/ml ~ r' y~ ' and 2 mM L-glutamime per well im 48-well tissue S culture ptates (Costar, Cambridge, MA). After incubating for 72 hours at 37C when the cells were confluent, the cells were washed twice with phosphate buffered saline (PBS) md prc- ' ` ' in serum-free Ham's F12 medium containmg 100 U/ml penicillin amd 100 mg/ml ~ tu...y~ amd 2 mM L-glutamime for 24 hours. After the pre 0.5 ml of F12 serum-free medium containing serial dilutions (1.0 - 1,000 ng/ml) of dGF-1, C3 10 and C2 muteins, and pegylated C3 amd C2 muteims were incubated separately for an additional 20-24 hours. Each well was then pulse labeled with 0.5 uCi of 3II ~
(NEN Research Products, DuPont Co., Boston, MA) for 4 hours, then washed with cold PBS three times and . ' 3H-thymidine was , . ` ` ' with cold 7 %
l.;~,t.lu~ , acid a.T. Baker Inc., Phillipsburg, Nl). After rinsing with 95% ethanol, 15 cells were solubilized with 0.3 M NaOH and aliquots removed for ~ ;counting.
3H-thymidine was quantitated by liquid ' counting. All assays were performed im triplicate.
The C3 and C2 muteims and pegylated C3 and C2 muteins were found to stimulate the same maximal level of 3EI lt.~ " . into DNA as ' IGF-1.
The potencies of the C3 and C2 muteins and the pegylated C3 and C2 muteins were about 3 to 4 fold lower than ' IGF-I. The ED50 (dose required for half maximal activity) of IGF-1 was 5-10 ng/ml compared with 30-40 ng/ml for ~ . ~,y-C3 amd C2 muteims, and the pegylated C3 amd C2 muteins.
These CA~ ` ' that the mitogenic activity of IGF-1 has been substantiaUy retained by the U amd C2 mutems and the pegylated U and C2 muteins. All four molecules capable of simulating cells to divide, as measured by 3H-thymidine into DNA. All four molecules are capable of stunulating cells to divide to the same maximal extent.
Using the assay described above, the relative mitogenic activity of of the E~P-4 peaks (described above) of the C11, 12, C15, C16, CSS, C64, C65, C67 and C69 IGF muteins and of C69-PEG was also ,~ ' The results of the latter assay is set forth in Table ~ W09s/32003 21qO752 P~ Sc APPPROXn~IATE ED~o OF IGF-I MUT~ I~S ~ C69-PEG (NG/ ML) PEAK I PEAIC II PEAK m PEAK IV
(20.5 min) (21.5 min) (22.0 min) (22.5 min) Cl 1 - 200 ~ 200 ~ 200 ND

SC15 ~ 150 ~70 ~5 ~300 C16 ~20 ~8 ND ND
CSS ~40 5 - 7 ~40 ND
C64 ~ 90 ~ 22 ND ND
C65 ~ 45 ~ 45 ND ND
10 C67 ~25 5 - 6 ND ND
C69 .20 - 25 5 - 6 ND ND

~WT IGF-I 20 4-6 WT IGF-I ED~o ~4-6 ng/ml ND - NOT DETERMINED

wo9~/32003 2 1 90752 1~ S~
The mitogenic activity of Peaks ~ of the C12, C16, C55, C67 and C69 mutein monomers was not ~ -- ly different from the mitogenic activity of colrectly refolded WT rIGF-I (Peak II). The ED~o of wild type rIGF-I was 4-6 ng/ml compared with 5-8 ng/ml for these unPEGylated mutein monomers. Table 5 shows that Peaks I of the various S muteins had lower (5-30 fold) bioactivity than correctly refolded WT rIGF-I si~ilar to the bioactivity of WT rIGF-I peak I. The C69-PEG conjugate, synthesized from peak II of C69, had the same bioactivity as C69 peak II and correctly refolded WT rIGF-I.
B. Relative AffiDiq for IGF-BP1 The relative affinities of the C3, C2 and C69 muteins and the PEGylated foTms ofthose mutems for IGF binding protein-l (IGF-BP1) were compared to that of the wild type IGF-I by measunng the ability of IGF-BPI to inhibit the mitogenic activities of the protems on rat ~- cells. The rat U~IR106 ceDs were plated at 5-6 X 104 cells in 0.5 ml of Ham's F12 containing 7% fetal bovine serum, 100 U/ml peniciDin and 100 mg/ml , y~ih~ and 2 mM ~glutamine per well m 48-well tissue culture plates.
After mcubating for 72 hours at 3rC when the cells became confluent, the cells were washed twice with PBS and prc ' ' m serum-free Ham's F12 medium containing 100U/ml penicillin, 100 mg/ml ,tl~ , amd 2 mM ~glutamme for 24 hours. After the - 0.5 ml of F12 serum-free medium containing either 50 ng/ml or 200ng/ml of rIGF-I, C3 or C2 mutein, or pegylated C3 or C2 mutein were incubated separately with varying amounts of IGF-BPI (100 nglml - I X 104 ng/ml) for an additional 20-24 hours.
Each well was then pulse labeled with 0.5 uCi of 3H-thymidine (NEN Research Products, Du Pont Co., Boston, MA) for 4 hours, then washed with cold PBS three times and ~, ' 3H-thymidine was ~ , ' with cold 7% i ' acid a.T. Baker Inc., r ~ NJ). 3H-thymidine was quamtitated by liquid ~ -- counting. AU
assays were performed in triplicate.
The results of this experiment indicated that the affu~ities of the,, ~' ' C3 mutein and the pegylated C3 mutein for IGFBPI were greatly reduced. At a molar ratio of 20:1 (IGFBPl:rIGF-I), the mitogenic activity of rIGF-I (50 ng/ml) was reduced 80%;
however, the mitogenic activities of the same . of the ~ ' ' C3 mutein and pegylated C3 mutein were reduced 35% and 0%, I~ ~liv.l.~. Similarly, when 200 ng/ml of the proteins were incubated with a 20 fold molar excess of IGF-BPI, 70% of the ~ W095/32003 2 19~752 rV.,IJ~J~06~10 mitogenic activity of rIGF-I was inhibited, whereas none of the mitogenic activity of the pegylaoed C3 mutein was inhibited. The affmities of of both the ~ ' C2 muoein and the pegylated C2 mutein were identical to that of wild type IGF-1. The affinity of both the I, v.r' ' C69 and C69-PEG for IGF-BP1 were not v ~ different from that of WT rIGF-I
These data indicate that the pegylated C3 muoem has vreatly reduced affmity for IGFBP1 when compared to IGF-1. Thus the mitogenic activity of the pegylaoed C3 muoein wi~l not be inhibited by IGF binding prooeins under conditions where the mitogenic activity of IGF-1 will be inhibited. However, the affinity of pegylated C2 and C69 muoeins for IGFBP1 is the same as the affinity of wild type IGF-1. Thus the mitogenic activity of pegylaoed C2 and C69 muoeins will be inhibioed by IGF binding proteins under the same conditions where the mitogenic activity of IGF-1 will be inhibioed.
E~fA~LE 8 Animal Tests Animal studies were performed to compare the pl-~-.. ,ov; properties of the muoeins and PEG conjugates of the present invention to the l' '~ properties of wild type IGF-1.
Animals Male Spravue Dawley rats with pituitary glands surgicaUy removed (h~l~vlJhr.,~Lu.. ~vd or Hypox rats) and weighing 110-121 grams were o~tained from Charles River Co. The rats were maintained in cages with lighting controlled over a 12 h-lighV12 h-dark cycle.
The animals had continuous access to waoer and food. Five animals were housed per cage. The weights of the rats were monitored daily and only rats with weight gains of less than 2 grams per week during the 2-3 weeks after arrival were considered to be succesfully ~J~,ul~h~ l and used for the ~ r B. M~:thods In r, I, animals (10 Hypox rats per group) were mjecoed every third day (ETD) ' 'y (sc) with ~T rIGF-I (160 mg, 320mg), 1 . v.v ' C2 (320mg), Woss/32003 21 901752 r~ 'o~
I . ,,y' ' C3 (320mg), pegylated C2 O(C2-PEG, 320mg) or Pegylated C3 (C3-pF.G,r~r,n ) dissolved in 0.2 ml of binding buffer (0.1 M HEPES-0.05 M NaH2PO4). A
separate group of 10 animals received 0.2ml vehicle. Injections were given between 0700 hours and 0800 hours and body weights were recorded daily between 1600 h and 1?00 h.
5 rI'he weights of rats on the day after the last imjection were taken as the fmal weight.
F, ' ~, animals (9 Hypox rats per group) were imjected every third day sub-'S, with WT rIGF-I (320mg, single injection daily, S~; 320mg ETD; 640mg ErID), or C3-PEG (320mg ETD, 640mg ErD, 960mg ETD).dissolved in 0.2 ml of binding buffer (0.1 M HEPES-0.05 M NaH2PO~). A separate group of 9 animals received 0.2ml vehicle. Injections were given between 0700 h and 0800 h amd body weight were recorded daily between 1600 h and 1700 h. r~he weights of rats on the day after the last injection were taken as the fmal weight.
In F, m, animals (10 Hypox rats) were injected every third day sub-lS, with C3-PEG (160mg lV, 320mg ETD), dissolved in 0.2 ml vehicle (0.1 M
HEPES-0.05 M NaH2PO~). A separate group of 10 animals received 0.2ml vehicle.
Injections were given betwoen 0700 h and 0800 h amd body weight were recorded daily betwoen 1600-1700 h. The weights of rats on the day after the last injection were taken as the fmal weight.
At the end of F, I & II, rats were r, ' .~ ' ' with CO2 and weighed. In 20 F . m, the tibia were removed and the epiphyseal width measured.
C. R~sults r ~ Rats treated with sc injections with either 160mg or 320mg of WT
IGF-I ErD showed no significant weight gain compared with animals injected with vehicle (Table 6). Similarly, animals injected ErID with 320 ng un-PEGylated C2 IGF-I or un-25 PEGylated C3 mutein did not show signif~cant weight gain. Animals injected ErID with320 mg C2-PEG and C3-PEG gained 4.42 i:0-74 g and 5.45 i 0.98 g, ~ .,ly, which was ! ,, .r. '',Sl greater than the weight gain of animals injected with wild type IGF-I (p < 0.01) . The weight gain of animals injected with PEGylated C2 or PEGylated C3 was ,, ~ 'S greater than the weight gain of animals injected with the un-PEGylated C2 or 30 C3 mutein (p~0.05). The PEGylated proteins clearly showed effficacy; however, the ~ WO 95J32003 2 1 ~ 0 7 5 2 P~ C ~o identical dose of WT IGF-I showed no efficacy. S~ ;ly, the addition of PEG
improves the biological potency of the molecule.
TABA~ 6 THE EFl~ECT OF IGF~~ [~ANS (U~AA;~YA~ATAFD ~ PEGYI~ATAED) ON TA~E
GA~OWTHA OF AA YA~A ~Ai~DA~ 1 ~J~IZED ~ATS
? ''lT.TJrT~T.T~ A1)0SE AL~ 2UA~NI- Y MEAN WT GA~I P VALUE
ug/day (g) vs Vehicle ETD -1.28 + 0.95 WT IGF-I 160 ETAT) 0.23 + 0.87 WT IGF-I 320 AT~D 0.59 + 0.79 10 C2 320 ETD -0.52 + 0.67 C2-PEG 320 ETD 4.42 + 0.74 WT 320 0.01 C2 320 0.01 C3 320 <0.05 C3 320 ~D I .75 i 0.90 C3-PEG 320 ETAr) 5.45 + 0.98 WT 320 <0.01 C2 320 <0.05 C3 320 < .05 These results ' that the PEGylated muteins exhibit enhanced ~ over WT IGF-I and the un-PEGylated IGF muteins.
r lATA A~ats treated with sc injections of WT IGF-I 320mg SATD~ 320mg ETD
and 640mg ETAr) gained 4.02g + 0.46g, 0.81g + 0.81g and 1.41g + 0.52g, ~Dy~Li~vl~
(Table 7). However, animals given 160mg, 320mg, 640mg 960mg of C3-PEG ETAI) gained 5.22g + 0.46g, 5.50g + 0.52g, 8.69g + 0.67g, and 10.43g + 0.77g, ~ (Table 20 6). All doses of C3-PEG ETD stimulated ~ more weight gain than both WT

w09~32003 21 907 52 r~ m~5,~
IGF-I doses given ETD. Animals injected with either 640mg or 960mg of C3-PEG ETDgained ~ , more weight than those dven 320mg WT IGF-I SID. C3-PEG doses of 160mg and 320mg ETD stimulated greater weight gam than 320mg of WT IGF-I SID;however, these differences did not reach statistical 1~ EFFECI OF C3-PEGylated IGF-I ON TF~ GR(~WTH OF
~Y~Ul'~D~l~D RATS
MOLE~ i Dasr. ~u~r MliAN r ~ALUE
14ddqr ~rr ~N ns Vdsicle Ell~ -0.77 i 0.38 0Silll IaF-I 320 SID 4.02 + sNT 320 ETD0.01 0,46 w~r 640 r.TD o.ol Wl IaF-I 320 ElD 0.81 i 0.81 WT IGF-I 640 EID 1.41 i 0.52 c3-PrSG 160 EID 5.22 i W8 320 ETD 0.01 0.46 Wl 640 EID 0.01 C3-PEa 320 El D 5 50 i Wl' 320 Ell~ 0.01 0.52 s~ 640 ETD 0.01 5C3 PEa 640 Era 8.69 + Wl 320 SID <0.01 0.67 C3PK~s 160 ETD C0.01 C3PECs 320 El D ~0.01 c3-Pr,a 960 r~D 10.43 + Wl 320SID <0.01 0.77 c3Pra 160 ElD ~0.01 C3PEa320ElD C0.01 F ' ~ ~ ' that C3-PEG r' ' ' ' SC ETD exhibits gS.~ater potency than WT IGF-I: ' ' sc ETD. All doses of C3-PEG stimulated greater mean weight gain tham animals given 320mg WT IGF-I SID. The e~hanced 20 1 ' ~ of C3-PEG make it more potent than WT IGF-I in the animal model described.

wo 9513z003 2 1 9 0 7 5 2 ~ 5.'06'l^
r~ - m: Rats treated with sc injections of C3-PEG 160 mg an~ 320 mg gained 8.3g + 0.7g and 9.0g + 0.6g, ~ (Table 4). Vehicle gained 4.2g + 0.3g.
The weight gained induced by U-PEG was statistically greater than anunals given vehicle.
Similarly, the tibial epiphyseal widths of rats receiving C3-PEG were statistically greater than rats receiving vehicle (Table 8).
.

'1~ EE7FECT OF C3-PEGylafed IGE7-l ON T~ GROWT~ Ol~
~lYl ~rr~ ~l~ RATS (WElGElT GAIN) MOLECUI~E WSE ~ N Wl' GAIN MEAN T1131A WIDTH P VALUE
y (d (~ ) V8 0 Ve~iclo ETD 4.2 + 03 0.136 + 0.0~4 C3-PEG 160 ETD 8.3 + 0.46 0.159 + o oo~ Vebicle ~0.01 C3-PEG 320 El'D 9.0 + 0.~ 0.151 _ 0.004 Vebicle <0.01 WO 95132003 , ., ~ ,C l f~

TEIE l~FFECT OF C3-PEGylat~d IGF-I ON T~IE GROWT~ OF
H Yl J~ 1 J~II~D RATS ~IIBIA ls~ I .. ~ ~AL W~) DOSE l,~U~ MEAN TIBIAL P VAllllE
t~ ~n.~ ugld-y WIIYIH ( S Vchicle ErD 0.136 i 0.008 C3-PEG 160 ETD 0.180 ~ 0.01 Vchicle <0.01 C3-PE1~ 320 ETD 0.167iO.06 Vehicle <0.01 F, m ~ that C3-PEG stimulates not only weight gain, but also bone growth in EIYPOX rats. This indicates that C3-PEG may be a useful r 10 for the induction of bone for~nation.
Although this invention has been described with respect to specific ' ' , it is not intended to be limited thereto. Various ' which will be apparent to thoseskilled in the art are deemed to fall within the spirit and scope of the present invention.

Claims (21)

Claims:

1. A polyethylene glycol (PEG) conjugate comprising PEG and a mutein of IGF, said PEG is attached to said mutein at a free cysteine.

2. The conjugate of claim 1, wherein said PEG is attached to the free cysteine through an activating group selected from the group consisting of maleimide, sulfhydryl, thiol, triflate, tresylate, aziridine, exirane, and 5-pyridyl.

3. The conjugate of claim 1, wherein the IGF is IGF-1.

4. The conjugate of claim 1, wherein the PEG has a molecular weight selected from the group consisting of 5 kDa, 8.5 kDa, 10 kDa, and 20 kDa.

5. The conjugate of claim 4, whereim the PEG has a molecular weight of 8.5 kDa.

6. The conjugate of claim 1, further comprising a second polypeptide attached to said PEG.

7. The conjugate of claim 6, whereim the second polypeptide is a mutein of IGF.

8. A mutem of IGF having a non-native cysteine.

9. The mutein of claims 8, wherein said non-native cysteine is in the N-terminalregion of the mutein.

10. The mutein of claim 8, wherein said muteim is a recombinant product.

11. The muteim of claim 8, wherein said mutem is expressed by E. coli.

12. The mutein of claim 8, whereim said non-native cysteine is m the C-terminal region of the mutein.

13. The mutein of claim 12, wherem said mutem is the C69 mutein.

14. A method of making the conjugate of claim 1, comprising attaching PEG to a free cysteine of an IGF mutein, said mutein having a non-native cysteine in the N-terminal region.

15. The method of claim 14, wherein said PEG is attached to the free cysteine through an activating group selected from the group consisting of maleimide, sulfhydryl, thiol, triflate, tresylate, aziridirle, exirane, and 5-pyridyl.

16. The method of claim 14, wherein the activating group is maleimide.

17. The method of claim 14, wherein the PEG is attached to an IGF mutein and another polypeptide.

18. The method of claim 17, wherein the other polypeptide is an IGF mutein.

41

20. A pharmaceutical composition comprising the conjugate of claim 1 in a pharmaceutically acceptable carrier.

21. A method for treating an IGF associated condition comprising administering the pharmaceutical composition of claim 20 to a patient.