WO2000056879A1 - TRANSFORMING GROWTH FACTOR (TFG) β SUPERFAMILY ANTAGONISTS - Google Patents

Abstract

Antagonists for biological processes induced by dimers or multimers are described that are easily derivable from the respective dimers or multimers by omission of at least one nonomer unit and/or at least one wrongly bound or folded nonomer unit. Particularly suitable antagonists of this kind are BMP antagonists that are a very efficient agent against undesired ossification due to trauma or operations such as hip replacement.

Description

Transforming Growth Factor (TFG) β Superfamily Antagonists

Technical Field

The present invention concerns antagonists against diseases induced by agonists needing for their activation more than one receptor-agonist interaction. In particular the present invention concerns antagonists for cytokines, preferably of the transforming growth factor (TGF) β superfamily.

Background Art

It is known that several biologically active molecules, below referred to as agonists, are dimers or multimers (trimers, tetramers, etc.) and need for their activation the simultaneous interaction with two or more receptor sites of the same or different receptors. A well known family of such agonists are the cytokines, in particular the transforming growth factor (TGF) β superfamily. Such factors are e.g. involved in heterotopic ossification (HO) .

Heterotopic ossification (HO) is a normal bone formation at ectopic sites like muscle and connective tissue, that can lead to a decreased range of motion, pain, or even total ankylosis predominantly of hip or elbow joints (see (6, 14)). In contrast, orthotopic ossifications is characterised by normal bone formation contiguous with the normal skeleton. HO may occur due to genetic disorders, but it is commonest following surgical trauma especially total hip replacement or head and neck trauma .

Although the mechanisms that cause HO after trauma are not entirely understood, one key player of all ossifications are certainly bone morphogenetic proteins (BMPs) . They are member of the TGF β superfamily and are able to induce heterotopic bone formation (33,34). Like TGF β, BMPs are synthesised as precursor molecule, which dimerise and get glycosylated (32). The majority of the BMP protein is cleaved at a basic peptide sequence, yielding the active, mature BMP molecule as a disulfide dimer of the carboxy-terminal quarter of the proprotein. Mature BMPs appear as a broad band around 30 kDa on SDS- PAGE, and reduce to monomers in the range of 16-20 kDa.

BMPs, or bone morphogenetic proteins are well described (1, 19, 26-29, 34) . Until now more than 30 different BMP-like proteins are known (for review see (20)) . Together they form the BMP family which comprises all BMPs, all OPs (osteogenic protein) , CDMPs (cartilage- derived morphogenetic protein) , GDFs (growth/differen- tiation factor) , Dpp (decapentaplegic) and Vg (vegetal) . The BMP family belongs to a larger family known as the transforming growth factor β (TGF β) superfamily, which includes besides TGF βs, also activins/inhibins, and Mόl- lerian inhibiting substances (35) . Analogous to other members of the TGF β superfamily the BMP receptors I and II are serine/threonine kinases . Upon BMP binding, BMP receptor I and II form a heteromeric-activated receptor complex, which initiates the signal transduction cascade (20) and leads to the activation of different genes in- volved in osteogenesis. The prerequest for the induction of osteogenesis by BMPs is the presence of cells with BMP receptors present in their cell membrane. Only these cells are able to receive BMP signals and respond to them. For the occurrence of the phenomena of heterotopic ossification, especially after total hip replacement, Friedenstein (7) and Owen (17,18) postulated the existence of two osteogenic precursor cells in the periar- ticular tissue of the hip. The inducible progenitor cells need BMP to develop into osteoblasts. They are located in the periarticular soft tissue and migrate and circulate in the blood stream. The second type called determined osteogenic progenitor cells originate in stromal parent cells of the bone marrow and develop into osteoblasts upon contact with non-resident tissue. Although they don't need BMPs for the induction of bone formation, they respond to BMPs with an accelerated differentiation (21) . After the induction of bone formation, the bone formation itself is influenced by BMPs, because all osteoblasts respond to BMP signals (8,22) by a stimulation of proliferation and differentiation.

Besides BMPs there are other effectors on bone formation, which were shown to promote the osteogenic activity of BMPs. Most of them are cytokines but also prostaglandins are known to enhance bone formation and to promote the activity of BMPs (16) . A decrease of prostaglandins in patients treated with non-steroidal anti-inflammatory drugs (NSAIDs) is the reason for a decrease in the number of patients developing HO after total hip replacement (6) . In 1975 Dahl (3) was the first to demonstrate that NSAIDs are useful for the prophylaxis against HO. He used indomethacin, but also ibuprofen, acetylsalicylic acid, and diclofenac have been investigated in the following years (10) and shown to reduce the occurrence of HO. The treatment has to start just before or immediately after operation and should be continued for at least 8 days (5) . In 1981 postoperative radiation in preventing

HO was introduced by Coventry and Scanlon (2) This treatment aims directly at the osteogenic precursors (6) and inhibits their differentiation to osteoblasts. If radiation therapy is used to reduce the occurrence of HO, it is sufficient to apply 1x7 Gy preoperatively or 17.5 Gy postoperatively if administered not later than 96 hours after hip replacement 2.

The state of the art treatments bear a lot of disadvantages. About 30% of patients treated with NSAIDs develop side effects like: pyrexia, allergic reaction, gastrointestinal discomfort, gastric ulceration or central nervous effects (10) and the osseointegration of the implant is retarded. The length and onset of the treatment with NSAIDs is still in dispute, but in animal studies it was shown that indomethacin could only prevent de- mineralized bone-induced heterotopic ossification if ad- ministered 6 hours before the implantation (4). Thus, HO induced through accidents, like head and neck traumas, can not be treated to date.

Compared to the treatment with NSAIDs radiotherapy is very cost intensive. Other disadvantages of radiotherapy are linked to the overall deterious effects of radiation, like the induction of transient oligosper- mia or infertility and initiation of secondary malignancies . Radiation can even be deterious for the osseointegration of alloplastic implants and reduce short- and long-term stability of the prosthesis (9,31) Also cancer patients which underwent radiation therapy should not be exposed to additional radiation and are therefore excluded from HO prophylaxis by radiation therapy.

Ossifications can also be a manifestation of inherited and acquired bone forming lesions. Here the only possible treatment is the surgical removal of the new formed bone. But this could prove to be very difficult and dangerous like in spinal hyperostosis or mye- lopathy caused by the ossification of the posterior lon- gitudinal ligament in the cervical spine (12) which leads to spinal cord compressions . Not operable is Spondylitis ankylosans (Bechterew-Strϋmpel-Marie-Krankheit) where ossification is linked to inflammation and possible auto immune response triggered by Klebsiella antigens (23). The main problem for all different ossifications is that even if the extra bone can be removed by surgery the trigger for ossifications still exists and recurrence will take place. Thus, it is very much desirable that a successful removal of the extra bone by surgery is accom- panied by an efficient inhibition of ossification at the operation site, whereby such inhibition should be almost free of undesired side effects. Operations to reduce extra bone can also be performed in the case of cranio-metaphyseal dysplasia. But in the inherited diseases myositis ossificans and fi- brodysplasia ossificans progressiva no treatment can be applied, due to the fact that any treatment tested so far led to severe progression of the disease (25,36).

Known antagonists are e.g. receptor mimics such as fetuin or neutralizing anti-TGF-beta antibodies that block osteogenesis (37) , IL-2 and I -6 variants carrying specific substitutions (38 and 39) , RANTES extended by addition of one amino acid at its N-terminus (40) , and soluble receptors, namely the interleukin-4-receptor (41) .

However, to find and/or generate such antago- nists is rather time consuming and often limited to good knowledge of the receptors and agonists as well as their interactions or interacting sites, respectively.

It is thus still very much desired to get products to inhibit unwanted biological effects, such as for example ossification, that are efficient, poor in or free of undesired side effects, and, preferably, also effective soon after administration, and that preferably are readily available.

Disclosure of the invention

The present invention concerns an antagonist for biological processes induced by a dimer or a multimer activating such processes due to interactions with more than one receptor site, which antagonist is characterised in that it interacts with at least one first receptor site needed to activate the biological process and in that it does not interact with at least one second recep- tor site needed for such activation, and whereby said interaction with said second receptor site does not take place due to at least one monomer unit of said dimer or multimer being missing or folded thus that a biological process activating interaction with said second receptor site is impossible. Such antagonists either lack at least one monomer unit or have one of their monomer units folded thus that no interaction is possible.

The terms "folded", "folding" etc. as they are used in the scope of the present invention comprise any conformation of dimers or multimers wherein at least one monomer unit is differently positioned relative to at least one further monomer unit if compared with the conformation of the respective agonist. Such folding can actually be due to a different folding, but also to a wrong binding of one of the monomer units, e.g. at binding at a wrong dimerisation site. Preferred antagonists are antagonists to cytokines, such as for example TGF β, interleukin-5, inter- leukin-6, interleukin-10, interleukin-12 , hepatocyte growth factor, platelet derived growth factor, and macro- phage-colony stimulating factor. In particular preferred are antagonists to members of the TGF-β superfamily, especially antagonists to members of the BMP superfamily.

Antagonists lacking at least one monomer unit or consisting of one monomer unit can be produced in that a host cell is transformed with a DNA sequence encoding the respective agonist and cultured under conditions allowing the expression of the monomer units building up said agonist, and in that the product of the expression is solubilized and treated under non-oxidising or reductive conditions. If only a part of a monomer unit is de- sired, a monomer unit produced as described above can be treated by well known methods to reduce the size of amino acid sequences. Alternatively it is of course also possible to produce such antagonists by expression of a DNA encoding such a modifies monomer unit in a suitable host cell.

Antagonist acting due to folding can be produced in that a host cell is transformed with a DNA se- quence encoding the respective agonist and cultured under conditions allowing the expression of the monomer units building up said agonist, and in that the product of the expression is solubilized and treated under oxidising conditions.

Antagonists of the present invention do also comprise such antagonists wherein the at least one lacking interaction with at least one receptor site does not take place due to at least one monomer unit of said dimer or multimer being folded thus that a biological process activating interaction with said second receptor site is impossible, due to an extension of the amino acid sequence of one of said monomer units at its N-terminal end by at least 5, preferably 10 to 30 amino acids, e.g. the sequence N-Met-Gly-Ser-Ser-His-His-His-His-His-His-Ser- Ser-Gly-Leu-Val-Pro-Arg-Gly-Ser-His-Met-C. Also such an antagonist is preferably an antagonist to members of the TGF-β superfamily , in particular an antagonist to members of the BMP superfamily. The present invention also concerns a DNA sequence encoding BMP that is extended at its N-terminus by a sequence as defined above.

An antagonist folded due to an extension at its N-terminus can be produced - in that a DNA sequence encoding the respective dimer or multimer is extended at its N-terminus by a sequence according to claim 12 either prior to its introduction into a suitable vector or by the introduction into a vector comprising such an extension, - in that a suitable host cell is transfected with said vector and cultured under conditions allowing the expression of the antagonist or monomer units of said antagonist, and

- in that the product of the expression of said DNA sequence is solubilised and treated under oxidising conditions. The antagonist of the present invention are very valuable effective substances in the treatment of a broad variety of dimer or multimer induced diseases . In particular antagonists to members of the BMP superfamily are very valuable agents against heterotopic ossification (HO) .

Modes for carrying out the invention

The present invention provides specific antagonists and a method for the production of antagonists for biological processes induced by dimers or multimers activating such processes due to multiple interactions with receptor sites located within a specific distance, usually on one cell. Said antagonists are characterised in that they interact with or bond to at least one site of at least one receptor needed to activate the corresponding agonist and leave at least one further reception site free, whereby at least one of the monomeric units is cleaved off or wrongly bound thus that an activating interaction with at least one receptor site is impossible due to a change in the localisation of at least one binding site relative to another binding site. The inventive antagonists for homodimers thus are monomers or parts of monomers or wrongly connected dimers. Also for heterodimers the inventive antagonists can be monomers or parts of monomers, whereby in the case of a preferred first receptor site the monomer unit bind- ing to said preferred first binding site is the preferred monomer. If the preferred receptor site is not known or if no great difference in the binding speed is present, a mixture of monomers might be used. Furthermore also heterodimers with a wrongly bound unit are suitable, since they provide both interacting sites but in a configuration allowing only one of the two possible (and for an agonist activity needed) interactions. The same applies for multimers with three or more subunits of the same or different kind. Antagonists to such multimers comprise monomers with the preferences discussed with regard to the heterodimers, as well as dimers or multimers provided that they lack at least one monomer unit or have at least one monomer unit wrongly bound so that at least one interaction needed for agonist activity lacks. Furthermore also for heterodimers and multimers parts of a monomer unit can be used as antagonist provided that they have a sufficiently fast and strong interaction with one of the first binding receptor sites.

It was surprisingly found that the antagonists of the present invention readily bind at least one receptor site and are not replaced by the agonist although the agonist has two or more binding sites so that a preference of the correctly binding agonist had been expected, the more so since the steric effect of at least a monomer or a binding part of a monomer is much reduced over a dimer or multimer.

In particular the present invention concerns antagonists to cytokines, such as for example TGF β, in- terleukin-5, interleukin-6 (multimer), interleukin-10 , interleukin-12 , hepatocyte growth factor, platelet derived growth factor, and macrophage-colony stimulating factor. A preferred group of antagonists are those to members of the TGF-β superfamily, and as ossification antagonists those to the BMP superfamily are much preferred. Members of the TGF-β superfamily, for example, with the cytokines being homodimers or heterodimers and the signal transduction being initiated by a cytokine me- diated dimerisation of two receptors upon cytokine binding are especially suitable for the generation of antagonists according to the present invention.

Especially the invention comprises the production and/or folding of antagonists for cytokines, preferably members of the TGF-β superfamily, especially members of the BMP-superfamily, as well as respective antagonists . The agonists suitable for deriving therefrom the antagonists of the present invention also include specific mutants of agonists needing for being active the simultaneous interaction with at least two receptor sites, in particular mutants of cytokines, and very much preferred specific mutants of bone morphogenetic proteins with improved refolding properties . From such agonists monomeric and dimeric antagonists lacking at least one "possibility" for simultaneous interaction easily deriv- able. The "possibility" for simultaneous interaction means the presence of a correctly placed binding site, the lack of such possibility either the lack of such a binding site (absence of at least one monomer unit) or a wrongly situated binding site. Such polypeptides are ob- tainable by a method that is also an object of this invention, and in general such polypeptides are obtained in high yields.

In accordance with the present invention and as specific embodiments thereof, mutant forms or native forms of recombinant bone morphogenetic proteins 2, 4 and 7 and other BMP-like proteins may be used to produce large quantities of BMP monomers homo- or heterodimers from bacteria that are folded into biologically active dimer or monomer molecules acting as BMP-antagonists. As already defined above, folded refers to the antagonists with at least one monomer unit wrongly folded, e.g. due to its binding to a wrong dimerisation site. In contrary, the term refolding refers to the conformation of the polypeptide associated with the natural biological activity and includes the dimerisation.

The molecules of the present invention also include DNA molecules comprising a nucleotide sequence encoding BMP-2,-4 and -7 except that the N-terminus was extended by some amino acids, preferably about 21 amino acids. The extension is based on the idea, that in nature all BMPs are synthesised first as pro-peptides and therefore the folding of the C-terminus, which finally forms the mature BMP, might be influenced by a N-terminal extension. As DNA source which codes for the 21 amino acid- extension any suitable vector can be used such as the commercially available vector pET-28-a from Novagen. The original mature DNA sequence for BMPs can be mutagenized by PCR and a Nde-restriction site can be created at the N-terminus. With such an Nde restriction site it is easy to place the mutagenized DNA for BMPs into e.g. the above mentioned vector. Such modifications are well within the level of ordinary skill in the art. The N-terminal extension of the protein produced e.g. in said specific vector is : N-Met-Gly-Ser-Ser-His-His-His-His-His-His-Ser-Ser- Gly-Leu-Val-Pro-Arg-Gly-Ser-His-Met-C. But other extensions may be used to influence the folding of the mature BMP and are also covered by this invention.

Nucleotide sequences encoding no N-terminal extension can also be cloned into a commercially available vector such as e.g. pET-23-a, from Novagen using the same Nde-restriction site as stated earlier. But any bac- terial expression vector may be used, as long as it is capable of directing the expression of a heterologous protein such as BMP in the bacteria chosen. The modification including the extension are well within the level of ordinary skill in the art. The bacterial expression plasmid may be transformed into a competent bacterial cell using known methods. Transformants are selected for growth on medium containing an appropriate drug, when drug resistance is used as the selective pressure. For the production of re- combinant BMP any bacterial species may be used.

The BMP expressed in such transformed bacteria cells is present in inclusion bodies, which are aggregates of precipitated BMP monomers and can be isolated from disrupted cells by centrifugation. The inclusion bodies can be solubilized by acidification with acetic acid or trifluoroacetic acid and reduced by a reducing agent such as β-mercaptoethanol or dithiotreitol . The pH for solubilization and denaturation proved to be preferably between 2 to about 4, but also basic conditions with pH above 10 can be used. The inclusion bodies can also be dissolved in chaotropic agents, well known in the art, like urea or guanidine hydrochloride . The concentration of chaotropic agents are normally in the range between 4 to 9 M.

The goal of all work known in the art on the refolding of proteins especially of TGF-like proteins is aimed on the recovery of biological active TGF-like proteins, with the same function as the TGF-like proteins in nature. In contrary, the goal of the present invention is to receive folded TGF-like proteins which antagonise the natural action of TGF-like proteins. Based on the knowl- edge that in order to transduce the signal of TGF-like proteins, homodimers or heterodimers have to bind two receptors to form a heteromeric-activated receptor complex, possible antagonist have to bind receptors and block the formation of heteromeric activated receptor complexes . As example to show the efficacy of the antagonists of the present invention BMPs were chosen, due to the fact that natural BMPs induce bone formation and BMP-antagonists should inhibit bone formation. Both actions can be tested in vivo, using demineralized bone ma- trix as carrier to determine the effect as antagonist and inactive collagenous bone matrix as carrier to determine the natural action of BMPs (13). As parameter for ossification the amount of calcium per mg implant can be used, as well as the extent of ossification judged on the basis of toluidin-blue stained or Goldner stained histo- sections . All these measurements and histo-staining are well within the ordinary skill in the art.

As one type of antagonists for natural actions of BMPs, BMP monomers were synthesised, which bind BMP-receptors without leading to the formation of activated receptor complexes . For the production of monomers the same procedures as described in US patent 5756308 or EP 0433225 Al can be used. Preferably the solubilization of BMP-monomer inclusion bodies is performed with urea without reducing agents. A preferred solubilization buffer contains: 6M urea, 20 mM Tris-HCl pH 7,9; 0,5 M NaCl; 5 mM imidazole. The solubilization could be shown to be complete after 4-6 days at 4°C under agitation at high protein concentrations (1-5 mg/ml) . The solubilized protein can be further purified using known chro a- tographic methods such as size exclusion chromatography, or exchange chromatography, or reverse phase high performance liquid chromatography. As last purification step gel filtration is preferably used to separate monomers from dimer or oligomers. The final gel filtration step is preferably performed in a buffer containing 6M urea and 25 mM Tris-HCl pH 8, but other buffers with chaotropic salts serve the same purpose.

This invention also comprises the formation of dimers, preferably BMP-dimers, which are folded to antagonise the natural action of BMPs. That BMP-dimers act- ing as antagonists can easily be obtained is due to the fact that in each mature BMP-monomer 6 cysteins are present, which represent potential dimerisation sites. In nature only specific cysteins are involved in dimerisation. Thus, if no reducing agent is used during the solubiliza- tion process, a substantial amount of monomers is oxidised to dimers . The oxidation is due to the presence of air or can be favoured by the use of oxidising agents, like glutathione . If choatropic or other denaturing agents are present during the dimerisation or if the pro- tein concentration is high, dimerisation can occur at unnatural sites and thus effect the 3 dimensional overall structure of the dimer. This can lead to products where one or both halves can bind to receptors, but the formation of a heteromeric activated receptor complex is hin- dered due to steric reasons caused by the change of the overall 3-dimensional structure of the dimer. Inhibiting monomers can also be produced by reducing bioactive BMP-dimers with for example β- mercaptoethanol or dithiotreitol . As source for bioactive BMP-dimers, refolded bacterial BMP, BMPs produced in mam- malian cell lines like CHO-cells or BHK cells, or natural BMPs isolated from animals like bovines, or other species can be used.

As already mentioned above, the present invention is not restricted to BMPs or BMP-like proteins, but covers also other cytokines which have to bind two receptors for the transduction of the signal like other members of the TGF-superfamily. In addition, possible mutants of cytokines and random polypeptides which bind cytokine receptors and antagonise the action of cytokines are also covered.

By virtue of the invention it is possible to obtain polypeptides for use in the prevention of cytokine actions requiring simultaneous interaction with at least two receptor sites, such as ossifications including het- erotopic ossifications and other diseases linked to ossifications .

Heterotopic ossification and ankylosis due to HO can be inhibited locally by applying antagonists of the present invention during the operation at the oper- ated hip or directly at the most likely affected knee or elbow-joints . In order to prevent heterotopic ossification induced by head and neck traumas or other traumatic incidences, the invention can be applied directly at knee, hip, and elbow joints, where ankylosis is most dra- matic and most likely to occur. The advantage of the antagonists of the present invention in comparison to NSAIDs and radiotherapy is, that they are more specific and therefore much reduced in, or even without side effects. Therefore a routine prophylaxis treatment with an- tagonists of the present invention is justified after all hip replacements. In addition, such antagonists can also be applied shortly after the heterotopic ossifications were induced. Thus patients which suffer from head and neck traumas or other accidental traumas could also undergo a routine HO prophylaxis treatment. With the arising availability and use of human recombinant BMPs for the initiation of ossification, antagonists of the present invention can also serve to restrict the ossification to a certain region or to stop ossification induced by human recombinant BMPs .

The present invention also opens a new field of treating diseases, linked to bone formation. For the treatment of inherited or induced bone forming lesions like heterotopic ossification, spinal hyperostosis, spon- dylitis ankylosans, cranio-metaphyseal dysplasia, myosi- tis ossificans, or fibrodysplasia ossificans progressiva antagonists of the present invention could be administered at the operation site in order to inhibit further ossification, or the recurrence of HO. In the case that operations can not be performed, the inventive antagonists can be applied by injections in the regions where ossification occurs, or applied systemically. In the case of fibrodysplasia ossificans progressiva, where any injection leads to new ossifications, an inventive antagonist is preferably administered systemically. If injections in such patients are mandatory, the inventive an- tagonists with or without carrier can be mixed with the medication in order to inhibit bone formation at the injection site.

Vascular invasion is a prerequest for bone formation. The very early and transient increase in BMP-4 mRNA during fracture healing (43, 44) and its decisive role in hematopoiesis (42) indicate a crucial role of BMP-4 especially for the blood supply during bone formation. Vascularisation is a critical point for the ossification but it is also important for cancer development and tumor growth, because fast growing tumors need a lot of nutrients. In the last years, the efficacy of inhibitors for angiogenesis in stopping the growth of tumors were shown. This is a further field of application of antagonists of the present invention such as BMP-4 antagonists. Thus BMP-antagonists, in particular BMP-4- antagonists, additionally represent a new class of tumor suppressors.

Still another field of application of an inhibitor for vascularization such as a BMP-4-antagonist is in the first non invasive treatment for hemagiomas and vascular malformations. Severe formes of these diseaeses are very difficult to treat and can even be deadly. Especially relapses after surgical removal of hemangiomas or after embolization are very frequent and can be avoided by the use of an inhibitor for vascularisation.

Antagonists of the present invention can be applied alone, i.e. as single inhibitor, or in combination with other antagonists of the present invention or known antagonists, as a mixture of inhibitors. The application can be performed locally in an aqueous solution in combination with a carrier like collagen or entrapped in a biodegradable material, for example polylactide-co- glycolide acids microspheres as described for the application of BMP dimers (30) . The composition of the later one is suitable to determine the release of antagonists of the present invention to the target sites. Systemic application can be achieved by a formulation as liquid, pill, tablet, lozenges for enteral administration, or in liquid form for parenteral administration.

Examples Example 1 : Solubilization and folding of BMPs in E.coli

The expression of the monomers was performed according the recommendations of Novagen from which the E.coli strain (BL21) and the different pET-vectors were bought. The cells in which the BMP monomers were ex- pressed, were harvested by centrifugation at 5000xg and frozen at -80°C.

6,2 grams of frozen cell pellet were thawed and resuspended in 40 ml Buffer 1 (20mM Tris-HCl pH 7,9; 0,5M NaCl; 5mM imidazole) . Cells were disrupted by 2 passes through a French press. The lysate was diluted to 50 ml and centrifuged at 15000xg for 15 minutes. The pellet formed by inclusion bodies was resuspended in buffer 1 with the help of a dounce homogenisator and centrifuged again at 15000xg for 15 minutes. The supernatant was discarded and the pellet homogenised with a dounce homogenisator, in 50 ml of Bufferl+6M Urea. The volume of the final ho ogenate was adjusted to 50 ml and rotated on a turning wheel for 6 days at 4°C. The resuspended inclusion bodies were centrifuged for 20 minutes at lOOOOxg, the pellet discarded and the supernatant filtered through a sterile filter from Nalgene with 0.8 μm pores size.

If a BMP mutant with the 21 additional amino acids N-Met-Gly-Ser-Ser-His-His-His-His-His-His-Ser-Ser- Gly-Leu-Val-Pro-Arg-Gly-Ser-His-Met-C was produced, the protein was purified by a nickel affinity column. This is possible because the 21 amino acid N-terminal extension contains a stretch of 6 histidines. As affinity resin the Ni-NTA Superflow™ (Quiagen) was used. The column size was l,6cmxl0cm (20 ml). 75- 100 mg of protein in 30-50 ml Bufferl+6M urea were loaded onto the column at 2 ml/min. After a wash step with 50 ml of [8M urea; 0 , 1M NaH2P04; lOmM Tris -HCl pH 6.3] and another one with 75 ml [8M urea; 0.1M NaH2P04; 10 mM Tris-HCl pH 4.5] the protein was eluted with 75 ml [20mM Tris-HCl pH 7 , 9 ; 0.5M NaCl; 1M imidazole] .

Before the purification could be continued, the protein solution (30 ml) was dialysed with a size ex- elusion of 10000 MW against 2x1000 ml TU (25 mM Tris pH 8.0; 6 M urea] over night. After the dialysation the protein solution was centrifuged at 5000xg for 15 minutes and sterile filtered with a filter of 0.45 um pore size. Alternatively the protein solution was concentrated with an Ultrafree® Biomax-IOK from Millipore and the buffer exchanged by 0.5M arginine; lOmM histidine pH 6.3. The protein precipitated in this buffer and was centrifuged at 15000xg for 10 minutes. The supernatant was discarded and the pellet dissolved in TU. With a further centrifu- gation step at 15000xg for 10 minutes insoluble material was pelleted and discarded. Alternatively the protein solution was concentrated with a 15 ml Ultrafree® Biomax- IOK from Millipore and the buffer exchanged by several passes of lOx the volume by TU. The protein solution in TU was centrifuged at 15000xg for 10 minutes and the supernatant was used for further purification steps . All the mentioned approaches finished with the protein being dissolved in TU. Now it could be applied to a Heparin column at 2 ml/min (5ml HiTrap® from Pharmacia Biotech) in batches of 20 mg, washed at 5ml/min with 25 ml TU, and eluted by a step gradient with TU+2M NaCl at the same speed. Alternatively the Heparin column was made of Heparin-Sepharose CL-6B (Pharmacia Biotech) . In this case all steps were performed at 2 ml/min. The dimension of the column was 2.6cm xl5 cm (80ml).

For gel filtration 2ml of the protein eluted from the Heparin columns were loaded with 2-15 mg on either a Hiload® Superdex® 75 or 200 prep grade column from Pharmacia Biotech (1,6 cm x 60 cm; 124 ml) . On this columns the oligomer, dimer, and monomer of all different BMPs or mutagenized BMPs could be separated. 1 ml frac- tions were collected and the column was run with TU.

Fractions, which only contained dimers or monomers were pooled. The decision on which fractions to pool was made based on the result of SDS-PAGE under non reducing conditions. If the protein content of the pools was below 0.4 mg/ml it was concentrated with a 15 ml Ultrafree® Biomax- IOK from Millipore to about 0.5 to 2 mg/ml. Protein con- centrations were determined by "Coomassie® Plus Protein Assay Reagent" from Pierce.

Example 2 : Determination of ossification and it ' s inhibition

The extent and potency to induce or inhibit ossification were determined in vivo in rats. Inactivated bone collagen matrix (IBCM) and demineralized bone (DB) were produced as described in (13) and used for carrier. These methods are within the normal scope of the skill of the art . The protein solution was added to the carrier in TU or in 5mM HCl . In the latter case buffer was exchanged with 5mM HCl by 3x1:10 concentration steps with a 15 ml Ultrafree® Biomax-IOK from Millipore.

25 mg of carrier in an eppendorf®-tube was mixed with 120μl antagonist probe in 5mM HCl or TU con- taining 0.5 or 1 mg chondroitin 6-sulfate sodium from

FLUKA. In the control probe no antagonist was added. The loading took place at RT. After lh, 300μl of collagen (Collagen R, 2mg/ml in 0.1% acetic acid from Serva) was added and mixed with the implant material . The proteins were precipitated by the addition of 1.2 ml 100% EtOH

(pre-cooled -80°C) and the material transferred to -80°C. After 1 h in the freezer the implant material was centrifuged at 15000xg 15 min at 4°C, the supernatant discarded and replaced by 80% EtOH (-20°C) . After another centrifu- gation for 5 min at 15000xg the entire supernatant was removed and pellets were formed using a 1 ml syringe. The implants were dried in a hood over night .

At the next morning rats were anaesthetised and the probes implanted either subcutaneously or intra- muscularly at bilateral sites over the thorax. After 23- 28 days the rats were killed by CO2 and the probes removed. The explant was freed from adherent tissue and cut in half. One half was used for histo-chemistry, the other half was weighed and homogenised in 1,5 ml of cold 3mM sodium-bicarbonate buffered saline, pH 9. The homogenate was centrifuged at 7500xg for 15 minutes. In the super- natant the alkaline phosphatase activity could be determined by the department of clinical chemistry of the Uni- versitatsSpital Zurich. The pellet was resuspended in 1 ml 5mM Tris-HCl pH 7.2, stirred for lh, and centrifuged for 15 min at 15000xg. This wash procedure was repeated 3 times. To the final pellet 1ml of 0.5M HCl was added and stirred over night. After another centrifugation (15000xg; 15 minutes) the supernatant was given to the department of clinical chemistry of the UniversitatsSpi- tal Zurich for the determination of the calcium concen- tration by atomic absorption spectrophotometry. The values for calcium concentration and the weight of the half of the implant determined the calcium content of the implant expressed in [mgCa/gr Implant] (see Example 3).

Example 3 : Test of inventive agonists

In the first set of experiments the 21 amino acid extended BMP2(=B2mat, BMP4 (=B4mat) , or BMP-7 (=B7mat) were expressed in E.coli and purified by a nickel- affinity column and heparin column as indicated earlier. In this preparation monomers and dimers were present. As carrier IBCM was used and up to 50μg BMP applied. After 21 days in the rat, none of the implants showed an increase in calcium.

In the next experiment the monomers and dimers of B4mat were separated by a gel filtration step. 30μg dimers were loaded on IBCM. After 21 day in the rat, one out of 3 probes showed an increase in calcium by the factor of 5. The experiment was repeated in 4 animals with 60μg B4mat dimer. After 26 days only in one animal no difference between the calcium content of the control and protein loaded was seen. In 3 animals the calcium content in comparison to the control implant was increased by the factor 1.2, 2.6, and 18.5. From these ex- periments we concluded that the monomer present in mono- mer-dimer mixtures can inhibit the action of BMP-dimers .

In order to test the hypothesis that BMP- monomers can inhibit heterotopic ossification, we used demineralized bone as carrier. If DB is implanted at ec- topic sites it leads to heterotopic bone formation. This animal model is one of the rare models for heterotopic ossification (15) and was also used to show the efficiency of NSAIDs and radiotherapy for the prevention of HO. In 4 experiments with rats of different age, 3 independent monomer-preparations (normal B4; B4mat (=21 amino acids extended B4) the hypothesis could be confirmed:

Results :

mg Ca/g implant with antagonist/ control [%] independent

BMP-monomers Days No mean SD site preparation

12μg B4mat 26 3 14 4.9 i .m. +

50μg B4 23 3 67 32 s .c . +

60μg B4 25 3 31 14 s .c.

50μg B4mat 27 4 72 7.5 s .c . +

13 48 30

The results on calcium content were confirmed by the Goldner and toluidin blue stained histosections . Without antagonists the ossification of the implant was advanced in comparison to implants with BMP-monomers. The results clearly show, that the monomer from BMP-4 with and without N-terminal extension inhibit or retard the ossification. In comparison DiCesare and co-workers (4) showed that after 28 days indomethacin inhibited the de- mineralized bone-induced heterotopic ossification expressed in calcium content to 21% of the control value, but only if the treatment was started at least 6 h before implantation. After implantation indomethacin was administered on daily basis.

In our treatment with BMP-monomers the treatment started at the time point of implantation and the monomers were only applied once. Thus, an administration with a slow release device could be a more efficient prophylactics for heterotopic ossification.

While there are shown and described presently preferred embodiments of the invention, it is to be dis- tinctly understood that the invention is not limited thereto but may be otherwise variously embodied and practiced within the scope of the following claims .

Bibliography

1. Celeste, A. J. , J. A. Iannazzi, R. C. Taylor, R. M. Hewick, V. Rosen, E. A. Wang, and J. M. Wozney. 1990. Identification of transforming growth factor beta-family members present in bone inductive protein purification from bovine bone. Proc . Natl . Acad. Sci . USA. 87:9843-9847.

2. Coventry, M. B., and P. W. Scanlon. 1981. The use of radiation to discourage ectopic bone: a nine- year study in surgery about the hip. J^" Bone Joint Surg. 63:201-208.

3. Dahl, H. K. 1975. Klinske Observasjoner. In : Symposium on Arthrose, Ed M. S. D. Blindern, Norway: 31 - 46.

4. DiCesare, P. E., M. E. Nimni, L. Peng, M. Yazdi, and D. T. Cheung. 1991. Effects of indomethacin on demineralized bone-induced heterotopic ossification in the rat. J^" Orthop Res . 9:855-861. 5. Dorn, U., C. Grethen, H. Effenberger, H.

Berka, T. Ramsauer, and T. Drekonja. 1998. Indomethacin for prevention of heterotopic ossification after hip ar- throplasty. Acta Orthop Scand. 69:107-110.

6. Eulert, J. , D. Knelles, and T. Barthel . 1997. Heterotope Ossifikation. Der Unfallchirurg.

100:667-674.

7. Friedenstein, A. Y. 1968. Induction of bone tissue by transitional epithelium. Clin Orthop. 59:21-37. 8. Hiraki, Y. , H. Inoue, C. Shigeno, H.

Bentz, D. M. Rosen, A. Asada, and F. Suzuki. 1991. Bone morphogenetic proteins (BMP-2 and BMP-3) promote growth and expression of the differented phenotype of rabbit chondrocytes and osteoblastic MC3T3-E1 cells in vitro. J. Bone Mineral Res . 6:63-74.

9. Kantorowitz, D. A., G. J. Miller, J. A. Ferrara, G. S. Ibbott, R. Fisher, and C. R. Ahrens . 1990. Preoperative vs postoperative irradiation in the prophylaxis of heterotopic bone formation in rats. Int J Radiat Oncol Biol Phys . 19:1431-1438.

10. Knelles, D. , T. Barthel, A. Karrer, U. Kraus, J. Eulert, and 0. Kolbl. 1997. Prevention of heterotopic ossification after total hip replacement. J^" Bone Joint Surg. 79:596-602.

11. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacte- riophage. Nature . 227:680-685.

12. Matsunaga, S., T. Sakou, E. Taketomi, M. Yamaguchi, and T. Okano . 1994. The natural course of mye- lopathy caused by ossification of the posterior longitudinal ligament in the cervical spine. Clin Orthop . 305:168-177.

13. Muthukumaran, Ν. , S. Ma, and A. H. Reddi . 1988. Dose-dependence of the threshold for optimal bone induction by collagenous bone matrix and osteogenin- enriched fraction. Collagen Rel Res . 8:433-441. 14. Νilsson, 0. S. 1998. Heterotopic ossification. Acta Orthop . Scand. 69:103-106.

15. O'Connor, J. P. 1998. Animal models of heterotopic ossification. Clin Orthop . 346:71-80.

16. Ono, I., M. Inoue, and Y. Kuboki . 1996. Promotion of osteogenetic activity of recombinant human bone morphogenetic protein by prostaglandin El. Bone . 19:581-588.

17. Owen, M. 1963. Cell population kinetics of an osteogenic tissue I. J. Cell Biol . 19:19-32. 18. Owen, M. 1980. The origin of bone cells in the postnatal organism. Arthr Rheum . 23:1073-1080.

19. Reddi, A. H. 1994. Bone and cartilage differentiation. Curr Opin Genet Dev. 4:737-744.

20. Reddi, H. 1997. Bone morphogenetic pro- teins : an unconventional approach to isolation of first mammalian morphogens . Cytokine & Growth Factor Reviews . 8:11-20. 21. Rickard, D. J., T. A. Sullivan, B. J. Shenker, P. S. Leboy, and I. Kazhdan. 1994. Induction of rapid osteoblast differentiation in rat bone marrow stro- mal cell cultures by dexamethasone and BMP-2. Dev. Biol . 161:218-228.

22. Sampath, T. K. , C. Maliakal, P. V. Hauschka, W. K. Jones, H. Sasak, R. F. Tucker, K. H. White, J. E. Coughlin, M. M. Tucker, R. H. Pang, and e. al . 1992. Recombinant human recombinant osteogenic pro- tein-1 (OP-1) induces new bone formation in vivo with a specific activity comparable with natural bovine osteogenic protein and stimulates osteoblast proliferation and differentiation in vitro. J Biol Chem. 267:20352-20362.

23. Schmidt, K. L. 1991. New pathogenetic as- pects of ankylosis spondylitis. Z Rheumatol . 50:65-73.

24. Seegenschmiedt, M. H. , L. Keilholz, P. Martus, A. Goldmann, R. Wolfel, F. Henning, and R. Sauer. 1997. Prevention of heterotopic ossification about the hip: final results of two randomized trials in 410 pa- tients using either preoperative or postoperative radiotherapy. Int . J. Radiation Oncology Biol . Phys . 39:161- 171.

25. Shafritz, A. B., E. M. Shore, F. H. Gannon, M. A. Zasloff, R. Traub, M. Murnke, and F. S. Kap- Ian. 1996. Overexpression of an osteogenic morphogen in fibrodysplasia ossificans progressiva. N Engl J Med. 335:555-561.

26. Urist, M. R. 1965. Bone: Formation by autoinduction. Science. 150:447-454. 27. Urist, M. R. , R. J. DeLange, and G. A. M.

Finerman. 1967. Bone cell differentiation and growth factors. Science . 220:680-686.

28. Urist, M. R. , B. F. Silverman, K. Buer- ing, F. L. Dubuc, and J. M. Rosenberg. 1967. The bone in- ductive principle. Clin . Oerthop. Rel . Res . 53:243-283.

29. Wang, E. A., V. Rosen, P. Cordes, and et.al. 1988. Purification and characterization of other destinct bone-inducing factors. Proc Natl Acad Sci USA . 85:9484-9488.

30. Weber, F. E., K. W. Gratz, and H. F. Sailer. 1998. Production of recombinant human bone mor- phogenetic protein-4 in protein-free medium and its application via biodegradable poly (lactide-co-glycolide) microspheres . Journal of Cranio-Maxillofacial Surgery. 26:207-208.

31. Wise, M. W. , I. D. Robertson, P. F. La- chiewicz, D. E. Thrall, and M. Metcalf. 1990. The effect of radiation therapy on the fixation strength of an ex- perimantal porous-coated implant in dogs. Clin Orthop. 261:276-280.

32. Wozney, J. M. 1990. Bone morphogenetic proteins. Progress in Growth Factor Research . 1:267-280.

33. Wozney, J. M. 1993. Bone Morthogenetic Proteins and Their Gene Expression. In Cellular and Molecular Biology of Bone. N. M. , editor^Λeditors . Academic Press, New York. 131-167. 34. Wozney, J. M. , V. Rosen, A. J. Celeste,

L. M. Mitsock, M. J. Whitters, R. W. Kriz, R. M. Hewick, and E. A. Wang. 1988. Novel Regulators of Bone Formation: Molecular Clones and Activities. Science . 242:1528-1534.

35. Yamashita, H. , P. Ten Dijke, C.-H. Heldin, and K. Miyazono. 1996. Bone morphogenetic protein receptors. Bone . 19:569-574.

36. Zasloff, M. A., D. M. Rocke, L. J. Crof- ford, G. V. Hahn, and F. S. Kaplan. 1998. Treatment of patients who have fibrodysplasia ossificans progressiva with isotretionin. Clin Orthop. 346:121-129.

37. Demetriou, M. et al . 1996. J. Biol . Chem. Vol. 271, Issue 22, 12755 -61.

38. Xu, D. et al. 1995. Eur. Cytokine Netw. Vol. 6, Issue 4, 237-44. 39. Sporeno, E. et al . 1996. Blood, Vol. 87,

Issue 11, p. 4510-9. 40. Wells, T.N. et al . 1996. J^". Leukoc . Biol . , Vol. 59, Issue 1, 53-60.

41. Gessner, A. et al . 1994. Infect . Immun . , Vol. 62, Issue 10, 4112-7. 42. Huber, T.L., and L. I. Zon. 1998. Tran- scriptional regulation of blood formation during Xenopus development. Sem Immun . 10:103-109.

43. Kitazawa, R. , S. Kitazawa, and S. Maeda. 1998. Expression of bone morphogenetic proteins (BMPs) in fractured mouse bone tissue: In situ hybridisation with polymerase chain reaction (per) -derived antisense DNA probe. Acra Histochem Cytochem, 31:231-236.

44. Nakase, T. , S. Nomura, H. Yoshikawa, J. Hashimoto, S. Hirota, Y. Kitamura, S. Oikawa, K. Ono, and K. Takaoka. 1994. Transient and localized expression of bone morphogenetic protein 4 messenger RNA during fracture healing. J Bone Miner Res . 9:651-659.

Claims

Claims

1. An antagonist for biological processes induced by a dimer or a multimer activating such processes due to interactions with more than one receptor site, characterised in that said antagonist interacts with at least one first receptor site needed to activate the biological process and does not interact with at least one second receptor site needed for such activation, and whereby said interaction with said second receptor site does not take place due to at least one monomer unit of said dimer or multimer being missing or folded thus that a biological process activating interaction with said second receptor site is impossible.

2. The antagonist of claim 1 that lacks at least one monomer unit.

3. The antagonist of claim 1 or 2 wherein one of its monomer units is folded.

4. The antagonist of anyone of claims 1 to 3 which is an antagonist to cytokines, such as for example

TGF β, interleukin-5, interleukin-6, interleukin-10, in- terleukin-12 , hepatocyte growth factor, platelet derived growth factor, and macrophage-colony stimulating factor.

5. The antagonist of anyone of claims 1 to 3 which is an antagonist to members of the TGF-β superfamily, in particular an antagonist to members of the BMP superfamily.

6. A method for the production of the antagonist of claim 2, characterised in that a host cell is transformed with a DNA sequence encoding the respective agonist and cultured under conditions allowing the expression of the monomer units building up said agonist, and in that the product of the expression is solubilized and treated under non-oxidising or reductive conditions.

7. A method for the production of the antagonist of claim 3, characterised in that a host cell is transformed with a DNA sequence encoding the respective agonist and cultured under conditions allowing the expression of the monomer units building up said agonist, and in that the product of the expression is solubilized and treated under oxidising conditions.

8. An antagonist for biological processes induced by a dimer or a multimer activating such processes due to interactions with more than one receptor site, characterised in that said antagonist interacts with at least one first receptor site needed to activate the bio- logical process and does not interact with at least one second receptor site needed for such activation, and whereby said interaction with said second receptor site does not take place due to at least one monomer unit of said dimer or multimer being folded thus that a biologi- cal process activating interaction with said second receptor site is impossible, due to an extension of the amino acid sequence of one of said monomer units at its N-terminal end by at least 5 amino acids.

9. The antagonist of claim 8 wherein the ex- tension has a length of 10 to 30 amino acids.

10. The antagonist of claim 8 or 9 wherein the extension has the sequence

N-Met-Gly-Ser-Ser-His-His-His-His-His-His- Ser-Ser-Gly-Leu-Val-Pro-Arg-Gly-Ser-His-Met-C .

11. The antagonist of anyone of claims 8 to

10 which is an antagonist to members of the TGF-β super- family , in particular an antagonist to members of the BMP superfamily.

12. A DNA encoding BMP that is extended at its N-terminus by a sequence of at least 5, preferably 10 to 30 amino acids, in particular the sequence

N-Met-Gly-Ser-Ser-His-His-His-His-His-His- Ser-Ser-Gly-Leu-Val-Pro-Arg-Gly-Ser-His-Met-C.

13. A method for the production of an antago- nist of anyone of claims 8 to 11, characterised

- in that a DNA sequence encoding the respective dimer or multimer is extended at its N-terminus by a sequence according to claim 12 either prior to its introduction into a suitable vector or by the introduction into a vector comprising such an extension,

- in that a suitable host cell is transfected with said vector and cultured under conditions allowing the expression of the antagonist or monomer units of said antagonist, and

- in that the product of the expression of said DNA sequence is solubilised and treated under oxi- dising conditions.

14. The antagonist of anyone of claims 1 to 5 and 8 to 11 as effective substance in the treatment of diseases .

15. Use of an antagonist of anyone of claims 4, 5 or 11 for the preparation of a pharmaceutical against heterotopic ossification (HO) .

16. A pharmaceutical composition comprising an antagonist according to anyone of claims 1 to 5 and 8 to 11.

17. A pharmaceutical composition comprising an antagonist according to anyone of claims 4, 5 or 11 for the treatment of heterotopic ossification (HO) .