WO1997033996A2 - Human bikunin - Google Patents

Abstract

The instant invention provides for proteins, polypeptides, nucleic acid sequences, constructs, expression vectors, host cells, pharmaceutical compositions of, and methods for using human placental bikunin, serine protease inhibitor domains, and fragments thereof.

Description

Title of the Invention: Human Bikunin

Field of the Invention

The compositions of the invention relate to the field of proteins which inhibit serine protease activity. The invention also relates to the field of nucleic acid constructs, vectors and host cells for producing serine protease inhibiting proteins, pharmaceutical compositions containing the protein, and methods for their use.

Background of the Invention Problem Addressed

Blood loss is a serious complication of major surgeries such as open heart surgery and other complicated procedures. Cardiac surgery patients account for a significant proportion of transfused donor blood. Blood transfusion carries risks of disease transmission and adverse reactions. In addition, donor blood is expensive and demands often exceed supply. Pharmacological methods for reducing blood loss and the resultant need for transfusion have been described (reviewed by Scott et al., Ann. Thorac. Surg. 50: 843-851, 1990).

Protein Serine Protease Inhibitor

Aprotinin, a bovine serine protease inhibitor of the Kunitz family is the active substance in the medicament Trasylol®. Aprotinin (Trasylol®) has been reported as being effective in reducing perioperative blood loss (Royston et al., Lancet ϋ: 1289-1291, 1987; Dietrich et al., Thorac. Cardiovasc. Surg. 37: 92-98, 1989; Fraedrich et al., Thorac. Cardiovasc. Surg.37: 89-91, 1989); W. van

Oeveren et al. (1987), Ann Thorac. Surg.44, pp 640-645; Bistrup et al., (1988) Lancet 1, 366-367), but adverse effects, including hypotension and flushing (Bohrer et al., Anesthesia 45: 853-854, 1990) and allergic reactions (Dietrich et al., Supra) have been reported. Use of aprotinin in patients previously exposed to it is not recommended (Dietrich et al., Supra). Trasylol® has also been used for the treatment of hyperfibrinolytic hemorrhages and traumatic hemorrhagic shock.

Aprotinin is known to inhibit several serine proteases including trypsin, chymotrypsin, plasmin and kallikrein, and is used therapeutically in the treatment of acute pancreatitis, various states of shock syndrome, hyperfibrinolytic hemorrhage and myocardial infarction (Trapnell et al., (1974) Brit J. Surg. 61: 177; J. McMichan et al., (1982) Circulatory Shock 9: 107; Auer et al., (1979)Acta Neurochir. 49: 207; Sher (1977) Am J. Obstet. Gynecol. 129: 164; Schneider (1976) , Artzneim.-Firsch. 26: 1606). It is generally thought that Trasylol® reduces blood loss in vivo through inhibition of kallikrein and plasmin. It has been found that aprotinin (3-58, Argl5, Alal7, Ser42) exhibits improved plasma kallikrein inhibitory potency as compared to native aprotinin itself (WO 89/10374).

Problems With Aprotinin

Because aprotinin is of bovine origin, there is a finite risk of inducing anaphylaxis in human patients upon re-exposure to the drug. Thus, a human functional equivalent to aprotinin, by virtue of a lower risk of anaphylaxis, would be most useful and desirable to have.

Aprotinin is also nephrotoxic in rodents and dogs when administered repeatedly at high dose (Bayer, Trasylol®, Inhibitor of proteinase; Glasser et al., in "Verhandlungen der Deutschen Gesellschaft fur Innere Medizin, 78.

Kongress", Bergmann, Munchen, 1972 pp. 1612-1614). One hypothesis ascribes this effect to the accumulation of aprotinin in the negatively charged proximal tubules of the kidney, due to its high net positive charge (WO 93/14120). Accordingly, an object of the present invention is to identify human proteins with functional activity similar to aprotinin. It was also an object of the instant invention to identify human proteins, that would be less charged, yet exhibit the same, highly similar, or improved protease specificities as found for aprotinin, especially with respect to the potency of plasmin and kallikrein inhibition. Such inhibitors could then be used repeatedly as medicaments in human patients with reduced risk of adverse immune response and reduced nepruOtoxiάty.

Brief Summary of the Invention

The instant invention provides for a purified human serine protease inhibitor which can specifically inhibit kallikrein, that has been isolated from human placental tissue via affinity chromatography.

The instant invention provides a newly identified human protein herein called human placental bikunin that contains two serine protease inhibitor domains of the Kunitz class. In one particular embodiment, the instant invention embodies a protein having the amino acid sequence: ADRERSIHDF CLVSKWGRC RASMPRWWYN VTDGSCQLFV YGGCDGNSNN 50

YLTKEECLKK CATVTENATG DLATSRNAAD SSVPSAPRRQ DSEDHSSDMF 100

NYEEYCTANA VTGPCRASFP RWYFDVERNS CNNFIYGGCR GNKNSYRSEE 150

ACMLRCFRQQ ENPPLPLGSK VWLAGAVS 179 (SEQIDNO: 1)

In a prefered embodiment the instant invention provides for native human placental bikunin protein having the amino acid sequence:

ADRERSIHDF CLVSKWGRC RASMPRWWYN VTDGSCQLFV YGGCDGNSNN 50 YLTKEECLKK CATVTENATG DLATSRNAAD SSVPSAPRRQ DSEDHSSDMF 100 NYEEYCTANA VTGPCRASFP RWYFDVERNS CNNFIYGGCR GNKNSYRSEE 150 ACMLRCFRQQ ENPPLPLGSK 170

(SEQIDNO:52)

In one aspect, the biological activity of the protein of the instant invention is that it can bind to and substantially inhibit the biological activity of trypsin, human plasma and tissue kallikreins, human plasmin and Factor Xlla. In a preferred embodiment, the present invention provides for a native human placental bikunin protein, in glycosylated form. In a further embodiment the instant invention encompasses native human bikunin protein which has been formed such that it contains at least one cysteine-cysteine disulfide bond. In a preferred embodiment, the protein contains at least one intra-chain cysteine- cysteine disulfide bond formed between a pair of cysteines selected from the group consisting of CYS11-CYS61, CYS20-CYS44, CYS36-CYS57, CYS106- CYS156, CYS115-CYS139, and CYS131-CYS152, wherein the cysteines are numbered according to the amino acid sequence of native human placental bikunin. One of ordinary skill will recognize that the protein of the instant invention may fold into the proper three-dimensional conformation, such that the biological activity of native human bikunin is maintained, where none, one or more, or all of the native intra-chain cysteine-cysteine disulfide bonds are present. In a most preferred embodiment, the protein of the instant invention is properly folded and is formed with all of the proper native cysteine-cysteine disulfide bonds. Active protein of the instant invention can be obtained by purification from human tissue, such as placenta, or via synthetic protein chemistry techniques, as illustrated by the Examples below. It is also understood that the protein of the instant invention may be obtained using molecular biology techniques, where self-replicating vectors are capable of expressing the protein of the instant invention from transformed cells. Such protein can be made as non-secreted, or secreted forms from transformed cells. In order to facilitate secretion from transformed cells, to enhance the functional stability of the translated protein, or to aid folding of the bikunin protein, certain signal peptide sequences may be added to the NH2-terminal portion of the native human bikunin protein.

In one embodiment, the instant invention thus provides for the native human bikunin protein with at least a portion of the native signal peptide sequence intact. Thus one embodiment of the invention provides for native human bikunin with at least part of the signal peptide, having the amino acid sequence:

AGSFLAWLGSLLLSGVLA -1

ADRERSIHDFCLVSKWGRCRASMPRWWYNVTDGSCQLFVYGGCDGNSNN 50 YLTKEECLKKCATVTENATGDLATSRNAADSSVPSAPRRQDSEDHSSDMF 100 NYEEYCTANAVTGPCRASFPRWYFDVERNSCNNFIYGGCRGNKNSYRSEE 150 ACMLRCFRQQENPPLPLGSKVWLAGAVS 179 (SEQ ID NO: 2)

In a prefered embodiment the instant invention provides for a native human placental bikunin protein with part of the leader sequence intact, having the amino acid sequence of SEQ ID NO: 52 with an intact leader segment having the amino acid sequence:

MAQLCGL RRSRAFLALL GSLLLSGVLA - 1 (SEQ ID NO: 53)

In another embodiment, the instant invention provides for bikunin protein with part of the leader sequence intact, having the amino acid sequence of SEQ ID NO: 52 with the intact leader segment having the amino acid sequence:

MLR AEADGVSRLL GSLLLSGVLA -1 (SEQ ID NO: 54)

In a preferred numbering system used herein the amino acid numbered +1 is assigned to the NH2-terminus of the amino acid sequence for native human placental bikunin. One will readily recognize that functional protein fragments can be derived from native human placental bikunin, which will maintain at least part of the biological activity of native human placental bikunin, and act as serine protease inhibitors. In one embodiment, the protein of the instant invention comprises a fragment of native human placental bikunin, which contains at least one functional Kunitz-like domain, having the amino acid sequence of native human placental bikunin amino acids 7-159, hereinafter called "bikunin (7- 159)". Thus the instant invention embodies a protein having the amino acid sequence:

IHDFCLVSKWGRCRASMPRWWYNVTDGSCQLFVYGGCDGNSNN 50

YLTKEECLKKCATVTENATGDLATSRNAADSSVPSAPRRQDSEDHSSDMF 100

NYEEYCTANAVTGPCRASFPRWYFDVERNSCNNFIYGGCRGNKNSYRSEE 150 ACMLRCFRQ 159 (SEQ ID NO: 3)

where the amino acid numbering corresponds to that of the amino acid sequence of native human placental bikunin. Another functional variant of this embodiment can be the fragment of native human placental bikunin, which contains at least one functional Kunitz-like domain, having the amino acid sequence of native human placental bikunin amino acids 11-156, bikunin (11- 156)

CLVSKWGRCRASMPRWWYNVTDGSCQLFVYGGCDGNSNN 50

YLTKEECLKKCATVTENATGDLATSRNAADSSVPSAPRRQDSEDHSSDMF 100 NYEEYCTANAVTGPCRASFPRWYFDVERNSCNNFI YGGCRGNKNSYRSEE 150 ACMLRC 156

(SEQ ID NO: 50).

One can recognize that the individual Kunitz-like domains are also fragments of the native placental bikunin. In particular, the instant invention provides for a protein having the amino acid sequence of a first Kunitz-like domain consisting of the amino acid sequence of native human placental bikunin amino acids 7-64, hereinafter called "bikunin (7-64)". Thus in one embodiment the instant invention encompasses a protein which contains at least one Kunitz-like domain having the amino acid sequence: IHDFCLVSKWGRCRASMPRWWYNVTDGSCQLFVYGGCDGNSNN 50

YLTKEECLKKCATV 64

(SEQ ID NO: 4)

where the amino acid numbering corresponds to that of the amino acid sequence of native human placental bikunin. Another form of the protein of the instant invention can be a first Kunitz-like domain consisting of the amino acid sequence of native human placental bikunin amino acids 11-61, "bikunin (11- 61)" having the amino acid sequence:

CLVSKWGRCRASMPRWWYNVTDGSCQLFVYGGCDGNSNN 50

YLTKEECLKKC 61

(SEQ ID NO: 5)

The instant invention also provides for a protein having the amino acid sequence of a Kunitz-like domain consisting of the amino acid sequence of native human placental bikunin amino acids 102-159, hereinafter called "bikunin (102-159)". Thus one embodiment the instant invention encompasses a protein which contains at least one Kunitz-like domain having the amino acid sequence:

YEEYCTANAVTGPCRASFPRWYFDVERNSCNNFI YGGCRGNKNS YRSEE 150

ACMLRCFRQ 159 (SEQ ID NO: 6)

where the amino acid numbering corresponds to that of the amino acid sequence of native human placental bikunin. Another form of this domain can be a Kunitz-like domain consisting of the amino acid sequence of native human placental bikunin amino acids 106-156, "bikunin (106-156)" having the amino acid sequence:

CTANAVTGPCRASFPRWYFDVERNSCNNFIYGGCRGNKNSYRSEE 150 ACMLRC 156 (SEQIDNO:7)

Thus one of ordinary skill will recognize that fragments of the native human bikunin protein can be made which will retain at least some of the native protein biological activity. Such fragments can also be combined in different orientations or multiple combinations to provide for alternative proteins which retain some of, the same, or more biological activity of the native human bikunin protein.

One will readily recognize that biologically active protein of the instant invention may comprise one or more of the instant Kunitz-like domains in combination with additional Kunitz-like domains from other sources. Biologically active protein of the instant invention may comprise one or more of the instant Kunitz-like domains in combination with additional protein domains from other sources with a variety of biological activities. The biological activity of the protein of the instant invention can be combined with that of other known protein or proteins to provide for multifunctional fusion proteins having predictable biological activity. Thus, in one embodiment, the instant invention encompasses a protein which contains at least one amino acid sequence segment the same as, or functionally equivalent to the amino acid sequence of either SEQ ID NO: 5 or SEQ ID NO: 7.

An open reading frame which terminates at an early stop codon can still code for a functional protein. The instant invention encompasses such alternative termination, and in one embodiment provides for a protein of the amino acid sequence:

ADRERSIHDFCLVSKWGRCRASMPRWWYNVTDGSCQLFVYGGCDGNSNN 50

YLTKEECLKKCATVTENATGDLATSRNAADSSVPSAPRRQDS 92 (SEQIDNO:8)

In one embodiment, the instant invention provides for substantially purified, or recombinantly produced native human bikunin protein with an intact segment of the leader sequence, and at least a portion of the native transmembrane region intact. Thus one embodiment of the invention provides for native human bikunin, with an intact leader sequence, and with at least part of the tΛ.nsmembrane domain (underlined), having an amino acid sequence selected from: 1)EST •MLR AEADGVSRLL GSLLLSGVLA -1 2)PCR MACjLCGL RRSRAFLALL GSLLLSGVLA -1 3)λcDNA MAQLCGL RRSRAFLALL GSLLLSGVLA -1

1) ADRERSIHDF CLVSKWGRC RASMPRWWYN VTDGSCQLFV YGGCDGNSNN 50 2) ADRERSIHDF CLVSKWGRC RASMPRWWYN VTDGSCQLFV YGGCDGNSNN 50 3) ADRERSIHDF CLVSKWGRC RASMPRWWYN VTDGSCQLFV YGGCDGNSNN 50

1)YLTKEECLKK CATVTENATG DLATSRNAAD SSVPSAPRRQ DSEDHSSDMF 100 2)YLTKEECLKK CATVTENATG DLATSRNAAD SSVPSAPRRQ DSEDHSSDMF 100

3)YLTKEECLKK CATVTENATG DLATSRNAAD SSVPSAPRRQ DSEDHSSDMF 100

1)NYEEYCTANA VTGPCRASFP RWYFDVERNS CNNFIYGGCR GNKNSYRSEE 150 2)NYEEYCTANA VTGPCRASFP RWYFDVERNS CNNFIYGGCR GNKNSYRSEE 150 3)NYEEYCTANA VTGPCRASFP RWYFDVERNS CNNFIYGGCR GNKNSYRSEE 150

1)ACMLRCFRQQ ENPPLPLGSK VWLAGLFVM VLILFLGASM VYLIRVARRN 200

2)ACMLRCFRQQ ENPPLPLGSK VWLAGLFVM VLILFLGASM VYLIRVARRN 200 3)ACMLRCFRQQ ENPPLPLGSK VWLAGLFVM VLILFLGASM VYLIRVARRN 200

D QERALRTVWS SGDDKEQLVK NTYVL 225

2 ) QERALRTVWS FGD 213

3 ) QERALRTVWS SGDDKEQLVK NTYVL 225 where sequence 1) is EST derived consensus SEQ ID NO: 45, 2) is PCR clone SEQ ID NO:47, and 3) is lambda cDNA clone SEQ ID NO:49. In a preferred embodiment a protein of the instant invention comprises one of the amino acid sequence of SEQ ID NO: 45, 47 or 49 wherein the protein has been cleaved in the region between the end of the last Kunitz domain and the transmembrane region.

The instant invention also embodies the protein wherein the signal peptide is deleted. Thus the instant invention provides for a protein having the amino acid sequence of SEQ ID NO: 52 continuous with a transmembrane amino acid sequence: EST VWLAGLFVM VLILFLGASM VYLIRVARRN 200 EST QERALRTVWS SGDDKEQLVK NTYVL 225

(SEQ ID NO: 69) a transmembrane amino acid sequence:

PCR VWLAGLFVM VLILFLGASM VYLIRVARRN 200 PCR QERALRTVWS FGD 213

(SEQ ID NO: 68) or a transmembrane amino acid sequence: λcDNA VWLAGLFVM VLILFLGASM VYLIRVARRN 200 λcDNA QERALRTVWS SGDDKEQLVK NTYVL 225 (SEQ ID NO: 67) . The protein amino acid sequences of the instant invention clearly teach one of the art the appropriate nucleic acid sequences which can be used in molecular biology techniques to produce the proteins of the instant invention. Thus, one embodiment of the instant invention provides for a nucleic acid sequence which encodes for a human bikunin having the consensus DNA sequence of Figure 3 (SEQ ID NO: 9), which translates into the amino acid sequence for native human placental bikunin sequence of Figure 3 (SEQ ID NO: 10). In another embodiment, the instant invention provides for a consensus nucleic acid sequence of Figure 4C (SEQ ID NO: 51) which encodes for an amino acid sequence of Figure 4D (SEQ ID NO: 45).

In a preferred embodiment, the instant invention provides for a nucleic acid sequence which encodes for native human placental bikunin having the DNA sequence of Figure 4F (SEQ ID NO: 48) which encodes for the protein sequence of SEQ ID NO: 49. In an another embodiment, the instant invention provides for a nucleic acid sequence of Figure 4E (SEQ ID NO: 46) which encodes for a protein sequence of SEQ ID NO: 47.

One can easily recognize that certain allelic mutations, and conservative substitutions made in the nucleic acid sequence can be made which will still result in a protein amino acid sequence encompassed by the instant invention. One of skill in the art can recognize that certain natural allelic mutations of the protein of the instant invention, and conservative substitutions of amino acids in the protein of the instant invention will not significantly alter the biological activity of the protein, and are encompassed by the instant invention.

The instant invention also provides for pharmaceutical compositions containing human placental bikunin and fragments thereof that are useful for the reduction of perioperative blood loss in a patient undergoing surgery.

The present invention also provides methods for reducing perioperative blood loss in a patient undergoing surgery, wherein an effective amount of the disclosed human serine protease inhibitors of the present invention in a biologically compatible vehicle is administered to the patient.

The present invention also provides for variants of placental bikunin, and the specific Kunitz domains described above, that contain amino acid substitutions that alter the protease specificity. Preferred sites of substitution are indicated below as positions Xaa 1 through Xaa^2 in the amino acid sequence for native placental bikunin. Substitutions at Xaa through Xaa 16 are also preferred for variants of bikunin (7-64), while substitutions at Xaa¹^ through Xaa^ are preferred for variants of bikunin (102-159). Thus the present invention embodies protein having an amino acid sequence:

Ala Asp Arg Glu Arg Ser lie Xaa¹ Asp Phe 10

Cys Leu Val Ser Lys Val Xaa² Gly Xaa³ Cys 20 Xaa⁴ Xaa⁵ Xaa⁶ Xaa⁷ Xaa⁸ Xaa⁹ Trp Trp Tyr Asn 30

Val Thr Asp Gly Ser Cys Gin Leu Phe Xaa¹⁰ 40

Tyr Xa ¹¹ Gly Cys Xaa¹² Xaa¹³ Xaa¹⁴ Ser Asn Asn 50

Tyr Xaa¹⁵ Thr Lys Glu Glu Cys Leu Lys Lys 60

Cys Ala Thr Xaa¹⁶ Thr Glu Asn Ala Thr Gly 70 Asp Leu Ser Thr Ser Arg Asn Ala Ala Asp 80

Ser Ser Val Pro Ser Ala Pro Arg Arg Gin 90

Asp Ser Glu His Asp Ser Ser Asp Met Phe 100

Asn Tyr Xaa¹⁷ Glu Tyr Cys Thr Ala Asn Ala 110

Val Xaa¹⁸ Gly Xaa¹⁹ Cys Xaa²⁰ Xaa²¹ Xaa²² Xaa²³ Xaa^{24 120} Xaa²⁵ Trp Tyr Phe Asp Val Glu Arg Asn Ser 130

Cys Asn Asn Phe Xaa²⁶ Tyr Xaa²⁷ Gly Cys Xaa^{28 140}

Xaa²⁹ Xaa³⁰ Lys Asn Ser Tyr Xaa³¹ Ser Glu Glu 150

Ala Cys Met Leu Arg Cys Phe Arg Xaa³² Gin 160

Glu Asn Pro Pro Leu Pro Leu Gly Ser Lys 170 Val Val Val Leu Ala Gly Ala Val Ser 179 (SEQ ID NO: 11) .

where Xaa - Xaa 32 each independently represents a naturally occurring amino acid residue except Cys, with the proviso that at least one of the amino acid residues Xaa -Xaa^l is different from the corresponding amino acid residue of the native sequence.

In the present context, the term "naturally occurring amino acid residue" is intended to indicate any one of the 20 commonly occurring amino acids, i.e., Ala, Arg, Asn, Asp, Cys, Gin, Glu, Gly, His, He, Leu, Lys, Met, Phe, Pro, Ser, Thr, Trp, Tyr and Val.

By substituting one or more amino acids in one or more of the positions indicated above, it may be possible to change the inhibitor specificity profile of native placental bikunin or that of the individual Kunitz-like domains, bikunin(7-64) or bikunin (102-159) so that it preferentially inhibits other serine proteases such as, but not limited to, the enzymes of the complement cascade, TF/FVπa, FXa, thrombin, neutrophil elastase, cathepsin G or proteinase-3. Examples of preferred variants of placental bikunin include those wherein Xaa* is an amino acid residue selected from the group consisting of His, Glu, Pro; Ala, Val or Lys, in particular wherein Xaa is His or Pro; or wherein Xaa 2 is an amino acid residue selected from the group consisting of Val, Thr, Asp, Pro, Arg, Tyr, Glu, Ala, Lys, in particular wherein Xaa 2 is Val or Thr; or wherein Xaa3 is an amino acid residue selected from the group consisting of Arg, Pro, He, Leu, Thr, in particular wherein Xaa^ is Arg or Pro; or wherein Xaa⁴ is an amino acid residue selected from the group consisting of Arg, Lys and Ser, Gin, in particular wherein Xaa⁴ is Arg or Lys; or wherein Xaa^ is an amino acid residue selected from the group consisting of Ala, Gly, Asp, Thr, in particular wherein Xaa^ is Ala; or wherein Xaa⁶ is an amino acid residue selected from the group consisting of Ser, lie, Tyr, Asn, Leu, Val, Arg, Phe, in particular wherein Xaa⁶ is Ser or Arg ; or wherein Xaa⁷ is an amino acid residue selected from the group consisting of Met, Phe, lie, Glu, Leu, Thr and Val, in particular wherein Xaa⁷ is Met or lie; or wherein Xaa * is an amino acid residue selected from the group consisting of Pro, Lys, Thr, Gin, Asn, Leu, Ser or lie, in particular wherein Xaa * is Pro or lie; or wherein Xaa 9 is an amino acid residue selected from the group consisting of Arg, Lys or Leu, in particular wherein Xaa 9 is Arg: or wherein XaalO is an amino acid residue selected from the group consisting of Val, lie, Lys, Ala, Pro, Phe, Trp, Gin, Leu and Thr, in particular wherein Xaa 10 is Val; or wherein Xaa H is an amino acid residue selected from the group consisting of Gly, Ser and Thr, in particular wherein Xaa 1 is Gly; or wherein Xaa 2 is an amino acid residue selected from the group consisting of Asp, Arg, Glu, Leu, Gin, Gly, in particular wherein Xaal2 is Asp or Arg; or wherein Xaa 13 is an amino acid residue selected from the group consisting of Gly and Ala; or wherein Xaa ⁱ⁴ is an amino acid residue selected from the group consisting of Asn or Lys; or wherein Xaa 5 is an amino acid residue selected from the group consisting of Gly, Asp, Leu, Arg, Glu, Thr, Tyr, Val, and Lys, in particular wherein Xaa 15 is Leu or Lys; or wherein Xaal is an amino acid residue selected from the group consisting of Val, Gin, Asp, Gly, lie, Ala, Met, and Val, in particular wherein Xaa 16 is Val or Ala; or wherein Xaa ⁱ⁷ is an amino acid residue selected from the group consisting of His, Glu, Pro, Ala, Lys and Val, in particular wherein Xaa I⁷ is Glu or Pro; or wherein Xaa 18 is an amino acid residue selected from the group consisting of Val, Thr, Asp, Pro, Arg, Tyr, Glu, Ala or Lys, in particular wherein Xaa 18 is Thr; or wherein Xaa ⁹ is an amino acid residue selected from the group consisting of Arg, Pro, He, Leu or Thr, in particular wherein Xaa 9 is Pro; or wherein Xaa20 j_s an amino acid residue selected from the group consisting of Arg, Lys, Gin and Ser, in particular wherein Xaa 20 is Arg or Lys; or wherein Xaa21 is an amino acid residue selected from the group consisting of Ala, Asp, Thr or Gly ; in particular wherein Xaa 21 is Ala; or wherein Xaa22 i_s an amino acid residue selected from the group consisting of Ser, He, Tyr, Asn, Leu, Val, Arg or Phe, in particular wherein Xaa 2 i_s Ser or Arg ; or wherein Xaa23 is an amino acid residue selected from the group consisting of Met, Phe, lie, Glu, Leu, Thr and Val, in particular wherein Xaa 3 is Phe or He; or wherein Xaa2⁴ is an amino acid residue selected from the group consisting of Pro, Lys, Thr, Asn, Leu, Gin, Ser or He, in particular wherein Xaa2⁴ is Pro or He; or wherein Xaa 5 i_s an amino acid residue selected from the group consisting of Arg, Lys or Leu, in particular wherein Xaa 25 is Arg: or wherein Xaa²⁶ is an amino acid residue selected from the group consisting of Val, He, Lys, Leu, Ala, Pro, Phe, Gin, Trp and Thr, in particular wherein Xaa 26 is Val or He; or wherein Xaa2 is an amino acid residue selected from the group consisting of Gly, Ser and Thr, in particular wherein Xaa2 is Gly; or wherein Xaa 28 is an amino acid residue selected from the group consisting of Asp, Arg, Glu, Leu, Gly or Gin, in particular wherein Xaa28 is Arg; or wherein Xaa29 is an amino acid residue selected from the group consisting of Gly and Ala; or wherein Xaa30 is an amino acid residue selected from the group consisting of Asn or Lys; or wherein Xaa31 is an amino acid residue selected from the group consisting of Gly, Asp, Leu, Arg, Glu, Thr, Tyr, Val, and Lys, in particular wherein Xaa31 is Arg or Lys; or wherein Xaa³2 is an amino acid residue selected from the group consisting of Val, Gin, Asp, Gly, He, Ala, Met, and Thr, in particular wherein Xaa32 is Gin or Ala.

Description of the Drawings

The invention wUl be better understood from a consideration of the following detailed description and claims, taken in conjunction with the drawings, in which:

Figure 1 depicts the nucleotide sequence of EST R35464 (SEQ ID NO: 12) and the translation of this DNA sequence (SEQ ID NO: 13) which yielded an open reading frame with some sequence similarity to aprotinin. The translation product contains 5 of the 6 cysteines in the correct spacing that is characteristic for Kunitz-like inhibitor domains (indicated in bold). The position normally occupied by the remaining cysteine (at codon 38) contained instead a phenylalanine (indicated by an asterisk).

Figure 2 depicts the nucleotide sequence of EST R74593 (SEQ ID NO: 14), and the translation of this DNA sequence (SEQ ID NO: 15) which yielded an open reading frame with homology to the Kunitz class of serine protease inhibitor domains. The translation product contained 6 cysteines in the correct spacing that is characteristic for Kunitz-like inhibitor domains (indicated in bold). However, this reading frame sequence includes stop codons at codon 3 and 23.

Figure 3 depicts a deduced nucleic acid sequence of human placental bikunin (SEQ ID NO: 9) labeled "consensus" and matched with the translated protein amino acid sequence labeled "translated" (SEQ ID NO: 10). Also as comparison are shown the nucleic acid sequence for ESTs H94519 (SEQ ID NO: 16), N39798 (SEQ ID NO: 17), R74593 (SEQ ID NO: 14) and R35464 (SEQ ID NO: 12). The underlined nucleotides in the consensus sequence correspond to the site of PCR primers described in the Examples. Underlined amino acids in the translated consensus sequence are residues whose identity have been confirmed by amino acid sequencing of purified native human placental bikunin. Nucleotide and amino acid code are standard single letter code, "N" in the nucleic acid code indicates an unassigned nucleic acid, and "*" indicates a stop codon in the amino acid sequence.

Figure 4A depicts the original overlay of a series of ESTs with some nucleic acid sequence homology to ESTs encoding human placental bikunin, or portions thereof. Shown for reference are the relative positions of bikunin (7-64) and bikunin (102-159), labeled KID1 and KTD2 respectively.

Figure 4B depicts a subsequent more comprehensive EST overlay incorporating additional ESTs. Numbers on the upper X-axis refer to length in base pairs, starting at the first base from the most 5' EST sequence. The length of each bar is in proportion to the length in base pairs of the individual ESTs including gaps. The EST accession numbers are indicated to the right of their respective EST bars.

Figure 4C depicts the corresponding alignment of the oligonucleotide sequences of each of the overlapping ESTs shown schematicaUy in Figure 4B. The upper sequence (SEQ ID NO: 51) labeled bikunin represents the consensus oligonucleotide sequence derived from the overlapping nucleotides at each position. The numbers refer to base-pair position within the EST map. The oligonucleotides in EST R74593 that are bold underlined (at map positions 994 and 1005) are base insertions observed in R74593 that were consistently absent in each of the other overlapping ESTs.

Figure 4D depicts the amino acid translation of the consensus oligonucleotide sequence for bikunin depicted in Figure 4C (SEQ ID NO: 45). Figure 4E depicts the nucleotide sequence (SEQ ID NO: 46) and corresponding amino acid translation (SEQ ID NO: 47) of a placental bikunin encoding sequence that was derived from a human placental cDNA library by PCR-based amplification. Figure 4F depicts the nucleotide sequence (SEQ ID NO: 48) and corresponding amino acid translation (SEQ ID NO: 49) of a native human placental bikunin encoding clone that was isolated from a human placental lambda cDNA library by colony hybridization.

Figure 4G compares the alignment of the amino acid translated oligonucleotide sequences for placental bikunin obtained by EST overlay (SEQ ID NO: 45), PCR based cloning (SEQ ID NO: 47), and conventional lambda colony hybridization (SEQ ID NO: 49).

Figure 5 shows a graph of purification of human placental bikunin from placental tissue after Superdex 75 Gel-Filtration. The plot is an overlay of the protein elution profile as measured by OD 280 nm (solid line), activity of eluted protein in a trypsin inhibition assay (% inhibition shown by circles), and activity of eluted protein in a kallikrein inhibition assay (% inhibition shown by squares).

Figure 6 shows a graph which plots the purification of human placental bikunin from placental tissue using C18 Reverse-Phase Chromatography. The plot is an overlay of the protein elution profile as measured by OD 215 nm (solid line), activity of eluted protein in a trypsin inhibition assay (% inhibition shown by circles), and activity of eluted protein in a kallikrein inhibition assay (% inhibition shown by squares).

Figure 7 depicts a silver stained SDS-PAGE gel of highly purified placental bikunin (lane 2), and a series of molecular size marker proteins (lane 1) of the indicated sizes in kilodaltons. Migration was from top to bottom.

Figure 8 shows the amount of trypsin inhibitory activity present in the ceU-free fermentation broth from the growth of yeast strains SC101 (panel 8A) or WHL341 (panel 8B) that were stably transformed with a plasmid (pS604) that directs the expression of placental bikunin (102-159). Figure 9 shows both a silver stained SDS-PAGE (left panel) and a

Western blot with anti-placental bikunin (102-159) pAb (right panel) of cell-free fermentation broth from the growth of yeast strain SCIOI (recombinants 2.4 and 2.5) that was stably transformed with a plasmid directing the expression of either bovine aprotinin, or placental bikunin (102-159). Migration was from top to bottom.

Figure 10 is a photograph which shows a silver stained SDS-PAGE of highly purified placental bikunin (102-159) (lane 2) and a series of molecular size marker proteins (lane 1) of the indicated sizes in Kilodaltons. Migration was from top to bottom.

Figure 11 is a photograph which shows the results of Northern blots of mRNA from various human tissues that was hybridized to a 32p labeled cDNA probe encoding either placental bikunin (102-159) (panel HA) or encoding placental bikunin (1-213) (panel 11B). Migration was from top to bottom. The numbers to the right of each blot refer to the size in kHobases of the adjacent RNA markers. The organs from which mRNA was derived is described under each lane of the blot. Figure 12 depicts an immunoblot of placental derived placental bikunin with rabbit antiserum raised against either synthetic reduced placental bikunin (7-64) (panel A) or 102-159 (panel B). For each panel, contents were: molecular size markers (lanes 1); native placental bikunin isolated from human placenta (lanes 2); synthetic placental bikunin (7-64) (lanes 3) and synthetic placental bikunin (102-159) (lanes 4). Tridne 10-20% SDS-PAGE gels were blotted and developed with protein A-purified primary polydonal antibody (8 ug IgG in 20 ml 0.1% BSA/Tris-buffered saline (pH 7.5), foUowed by alkaline phosphatase-conjugated goat anti-rabbit secondary antibody. Migration was from top to bottom.

Figure 13 depicts a Coomassie Blue stained 10-20% Tridne SDS-PAGE gel of 3 micrograms of highly purified placental bikunin (1-170) derived from a baculovirus / Sf9 expression system (lane 2). Lane 1 contains molecular size markers. Migration was from top to bottom.

Figure 14 depicts a comparison of the effect of increasing concentrations of either Sf9-derived human placental bikunin (1-170) (fiHed circles), synthetic placental bikunin (102-159) (open drcles), or aprotinin (open squares) on the activated partial thromboplastin time of human plasma. Clotting was initiated with CaCl2. The concentration of proteins are plotted versus the -fold prolongation in clotting time. The uninhibited clotting time was 30.8 seconds.

Detailed Description of the Invention

The present invention encompasses a newly identified human protein herein called human placental bikunin that contains two serine protease inhibitor domains of the Kunitz class. The instant invention also encompasses pharmaceutical compositions containing placental bikunin and fragments thereof that are useful for the reduction of perioperative blood loss in a patient undergoing surgery, or with major trauma.

The present invention also provides methods for reducing perioperative blood loss^" in a patient undergoing surgery or due to major trauma, wherein an effective amount of the disclosed human serine protease inhibitors of the present invention, in a biologically compatible vehicle, is administered to the patient. A preferred application for placental bikunin, isolated domains, and other variants is for the reduction of blood loss resulting from trauma or surgery that has the potential for loss of large volumes of blood. These methods and compositions reduce or eliminate the need for whole donor blood or blood products, thereby redudng the risk of infection and other adverse side effects , as well as the cost of surgery. The methods are thus useful in reducing blood loss in normal patients, i.e., those not suffering from inborn or other pre- operative defidendes in coagulation factors. The reduction in blood loss is seen as a reduction in blood loss during surgery , as reduced post surgical drainage or both. Preferred surgical applications indude but are not limited to use in thoradc and abdominal surgery, total and partial hip replacement surgeries and surgeries to treat a patient having an epithelial lesion of the eye. Preferred thoracic surgical procedures include but are not limited to aortocoronary bypass, exdsion of cardiac and aortic aneurysms, and surgery for esophageal varices, and coronary artery bypass surgery. Preferred abdominal surgeries indude but are not limited to liver transplants, radical prostatectomy, surgery for diverticulitis of colon, tumor debulking, surgery on the abdominal aorta and surgery for duodenal ulcers, and repair of liver or spleen trauma. Preferred use for the treatment of trauma include but are not limited to the use in stabilization of severely injured patients at acddent sites suffering from e.g., limb loss or major thoradc /abdominal wounds. In case of use for the reduction of blood loss resulting from surgery it is preferred to administer the placental bikunin, isolated domains, or other variant prior to and during surgery, whereas in case of use in trauma settings the placental bikunin variant, isolated domain or other variant is to be administered as rapidly as possible foUowing injury, and should be contained on emergency vehides traveling to the acddent sites.

Factor XQ (also known as Hageman Factor) is a serine protease that is found in the circulation in a zymogen form (80 kD) at approximately 29-40 μg/ml (see Pixley, et al. (1993) Meth. in Enz., 222, 51-64) and is activated by tissue and plasma kallikrein. Once activated, it participates in the intrinsic pathway of blood coagulation which is activated when blood or plasma contacts a "foreign" or anionic surface. Once activated, Factor Xlla can then deave and activate a number of other plasma proteases including Factor XI, prekallikreύv and CI of the complement system. Thus Factor XII may be involved in causing hypotensive reactions since activated kallikrein can cleave kininogen releasing bradykinin (see Colman, (1984) /. Clin. Invest., 73, 1249). Sepsis is a disease that results from bacterial infection due to bacterial endotoxin or lipopolysaccharide (LPS). Exposure of Factor XII to LPS results in the activation of Factor XQ. Patients with sepsis frequently have symptoms of intravascular coagulation which may also be due to activation of Factor XII by LPS. Septic shock can result from bacterial infection and is associated with fever, low systemic vascular resistance, and low arterial pressure. It is a common cause of death in intensive care units in the United States, where seventy five percent of the patients that die from septic shock have a persistent hypotension (see Parillo, et al. (1989) Ann Rev. Med._r 40, 469-485).

Adult respiratory distress syndrome is characterized by pulmonary edema, hypoxemia, and decreased pulmonary compliance. The pathogenesis of the disease is currently unknown although the proteolytic pathways of coagulation and fibrinolysis are believed to play a role (see Carvalho, et al. (1988) /. Lab Clin. Med., 112: 270-277).

The proteins of the instant invention are also a novel human Kunitz type inhibitor of kaUikrein, an activator of Factor XII. Thus another object of the current invention is to present a method for the prophylactic or therapeutic treatment of systemic inflammatory reactions such as septic shock, adult respiratory distress syndrome (ARDS), preeclampsia, multiple organ failure and disseminated intravascular coagulation (DIC). The therapeutic or prophylactic administration of the peptides of the instant invention would result in the modulation of these inflammatory conditions and be benefidal to the patient. Plasmin plays an important role in extracellular matrix degradation and the activation of matrix-metaUo protease (MMP) cascades. CoUectively these proteases mediate migration of and tissue invasion by both endothelial cells during angiogenesis/neovascularization, and cancer cells during metastasis. Neovascularization is essential to support tumor growth and metastasis is a process which mediates the spreading of tumors and which is assodated with extremely poor patient prognosis.

Several preclinical studies suggest that Kunitz like serine protease inhibitors with a protease specificity similar to aprotinin are useful as medicaments for cancer. For example, aprotinin reduced tumor growth and invasion, with increased tumor necrosis when administered to hamsters bearing a highly invasive fibrosarcoma or to mice bearing a similarly malignant mammary carcinoma (Latner et al., (1974), Br. J. Cancer 30: 60-67; Latner and Turner, (1976), Br. J. Cancer 33: 535-538). Furthermore, administration of 200,000 KIU of aprotinin i.p. to C57B1/6 Cr male mice on days 1 to 14 post-inoculation with Lewis lung carcinoma ceUs, reduced pulmonary metastases by 50% although had no effect on primary tumor mass (Giraldi et al., (1977) Eur. J. Cancer, 13: 1321-1323). Similarly, administration of 10,000 KIU i.p. on each of days 13-16 post-inoculation of C57BL/6J mice with Lewis tumor cells inhibited pulmonary metastases by 90% without affecting the primary tumor growth (Uetsuji et al., (1992), Jpn. J. Surg. 22: 429-442). In this same study, administration of plasmin or kallikrein with the same dosing schedule was argued to increase the number of pulmonary metastases. These results prompted the authors to suggest that perioperative administration of aprotinin to cancer patients may reduce the likelihood of metastases. Black and Steger (1976, Eur. J. Pharmacol., 38: 313-319) found that aprotinin inhibited the growth of the transplanted rodent Murphy-Strum lymphosarcoma in rats and suggested that the effect involved the inhibition of the kinin-forming enzyme system. Twice daily i.p. injection of female ddY mice with 10,000 KIU of aprotinin for 7 weeks to mice each bearing a single autochtonous squamous cell carcinoma resulting from 3-methylcholanthrene treatment reduced the growth rate of the primary tumors by 90%. In some animals tumor regression was observed. While all vehicle treated animals had died within the seven weeks, all of the aprotinin treatment group remained alive. Reduced tumor growth was associated with hyperkeratosis (Ohkoshi, Gann (1980), 71: 246-250).

Clinically, a surgically cured group of 26 patients who received aprotinin i.v. exhibited a 70% survival two years post surgery with no recurrence of tumors whereas a placebo group of 26 patients at the same time exhibited only a 38% survival with a significant rate of tumor recurrence (Freeman et al. Br. Soc. Gastroenterol. (1980) supplement A: 902). In a case study (Guthrie et al., Br. J. Clin. Pract (1981) 35: 330-332), administration of bromocriptine plus aprotinin to a patient with advanced cancer of the cervix caused remission. Aprotinin was administerd both as a 500,000 KIU i.p. bolus every eight hours concurrently with a continuous i.v. infusion of aprotinin at a rate of 200,000 KIU per 6 hr for a total of seven days once a month. Treatment was ended at the end of the fourth month due to the development of an allergic reaction to aprotinin. More recent evidence has further underscored a role of plasmin as a target for these effects of aprotinin on metastases.

The mechanism for these events could be related to the fact that aprotinin blocks the invasive potential of cancer cell lines (Liu G., et al., Int J. Cancer (1995), 60: 501-506). Furthermore, since the proteins of the instant invention are also potent inhibitors of plasmin and kaUikrien, they are contemplated for use as anti-cancer agents. For example they are contemplated for use in blocking primary tumor growth by restricting neovascularization, primary tumor invasion and in blocking metastasis through inhibition of tissue infiltration. The compounds may be administered locally to tumors or systemically. In a preferred mode of treatment, the protein would be administered perioperatively during tumor debulking to minimize the risk of metastasis. In such a regime, the blood sparing properties of the compound would be additionaUy advantageous in providing a clearer surgical field of view. Another preferred mode of administration would be as a combination therapy with either MMP inhibitors or chemotherapy. An additional preferred mode of administration would be as a locally administered gene therapy designed to achieve selective expression of placental bikunin within the tumor cells, or their associated stroma and vascular beds.

Preferred types of cancers targeted for therapy would be vasular- dependent solid tumors such as breast, colon, lung, prostate and ovarian carcinomas which exhibit a high metastatic potential, and those for which local delivery of a high concentration of the protein is feasible such as lung cancers through pulmonary delivery, colon cardnomas through hepatic delivery to liver metastasis, or skin cancers such as head and neck carcinomas or melanomas through subcutaneous delivery. Since the proteins of the present invention are of human origin they would be less likely to be assodated with aUergic or anaphylactic reactions of the kind observed by Guthrie et al., supra, upon reuse.

Additionally, the proteins of the present invention are contemplated for use in the reduction of thromboembolic complications assodated with activation of the intrinsic pathway of coagulation. This would indude prevention of pulmonary emboUsm in late stage cancer patients, a frequent cause of death (Donati MB., (1994), Haemostasis 24: 128-131).

Edema of the brain and spinal cord is a complication resulting from traumatic brain or spinal cord injury, stroke, cerebral ischemia, cerebral and sub-arachnoid hemhorrhage, surgery (induding open heart surgery), infectious diseases such as encephalitis and meningitis, granulomatous diseases such as Sarcoid and focal or diffuse carcinomas, and is a contributor to the high level of morbidity and death foUowing these events. Bradykinin is known to disrupt the blood brain barrier experimentaHy (Greenwood J., (1991), Neuroradiology, 33: 95-100; Whittle et al., (1992), Acta Neurochir., 115: 53-59), and infusion of bradykinin into the internal carotid artery induced brain edema in spontaneously hypertensive rats (SHR) subjected to common carotid artery occlusion (Kamiya, (1990), Nippon Oca Daigaku Zasshi. 57: 180-191). Elevated levels of bradykinin are found in extraceUular fluids foUowing trauma in a model involving traumatized rat spinal chord (Xu et al., (1991), J. Neurochem, 57: 975-980), and in plasma and tissue from rats with brain edema resulting from cerebral ischaemia (Kamiya et al., (1993), Stroke, 24: 571-575). Bradykinin is released from high molecular weight kininogen by serine proteases including kaUikrein (Coleman (1984) J. Clin Invest, 73: 1249), and the serine protease inhibitor aprotinin was found to block the magnitude of brain edema resulting from cerebralsd-emia in SHR rats (Kamiya, (1990), Nippon Ika Daigaku Zasshi. 57: 180- 191; Kamiya et al., (1993), Stroke, 24: 571-575) and rabbits subjected to a cold lesion of the brain (Unterberg et al., (1986), J. Neurosurgery, 64: 269-276).

These observations indicate that brain edema results from local proteolytic release of kinins such as bradykinin from high molecular weight kininogen, followed by bradykinin-induced increases in blood brain barrier permeabUity. Accordingly, placental bikunin and fragments thereof are contemplated as medicaments for the prevention of edema in patients at risk for this condition, particularly those of high risk of mortality or brain injury. This would include head and spinal trauma patients, polytrauma patients, patients undergoing surgery of the brain or spinal cord and their assodated vessels or other generalsurgeries including open-heart surgery, patients who have suffered from a stroke, cerebral or sub-arachnoid hemorrhage, infectious diseases of the brain, granulomatous disease of the brain or diffuse or focal carcinomas and tumors of the brain or any conditions such as multiple sclerosis involving breakdown of the blood brain barrier or patients suffering from any other inflammatory processes of the brain or spinal cord. Patients would receive an administration of placental bikunin either as an infusion or bolus injection, intravenously or intracraniaUy. Additional doses of placental bikunin could be administered intermittently over the following one to three weeks. Dose levels would be designed to attain circulating concentrations in excess of those required to neutralize elevations in plasma levels or bradykinin and other vasoactive peptides formed through the action of serine proteases, and suffident to reduce edema. Since the protein is of human origin, repeated administration in this course of therapy would not lead to development of an immune reaction to the protein. Placental bikunin and fragments thereof would be contemplated for monotherapy or prophylacsis as weU as for use in combination with other medicaments such as neurotherapeutics and neuroprotectants.

Recent evidence (Dela Cadena R. A. at al., (1995), FASEB J. 9: 446-452) has indicated that the contact activation pathway may contribute to the pathogenesis of arthritis and anemia, and that kallikrein inhibitors may be of therapeutic benefit. Accordingly, protease inhibitors of the present invention are contemplated according to their capadty to inhibit human kalUkrein, as medicaments for the treatment of arthritis and anemia in humans.

Treatment of male non-insulin diabetic (NIDDM) patients with aprotinin significantly improved total glucose uptake and decreased the metabolic clearance rate of insulin (Laurenti et al., (1996), Diabetic Medicine 13: 642-645). Accordingly, the human proteins of the present invention are contemplated for chronic use as medicaments for the treatment of NIDDM.

DaHy treatment of patients at risk of preterm delivery with urinary trypsin inhibitor for two weeks significantly reduced recurrent uterine contractions (Kanayama et al., (1996), Eur J. Obstet. Gynecol. & Reprod. Biol. 67: 133-138). Accordingly, the human proteins of the present invention are contemplated for use in the prevention of preterm delivery.

Aprotinin has been shown to stimulate differentiation of mouse myoblasts in culture (Wells and Strickland, Development, (1994), 120: 3639-3647)), a process that is inhibited by TGFb. TGFb exists as an inactive pro-polypeptide which is activated by limited proteolysis. The mechanism of aprotinin action has been proposed to involve inhibition of proteases which process pro-TGFb to the mature active form. TGFb has been shown to be up-regulated in various fibrotic lesions and has long thought to be a potential target for anti-fibrotic therapies. In a rat model of pulmonary fibrosis for example, TGF-b concentrations paralleled the extent of bleomydn-induced inflammation. Furthermore, plasmin levels in the alveolar macrophage coincided with mature TGF-b levels, and the addition of the plasmin inhibitor a-2-antiplasmin abrogated the post translational activation of pro-TGFb by the macrophage (Khal et al., (1996), Am. J. Respir. CeU Mol. Biol. 15: 252-259.) The data suggest that plasmin contributes to the formation of active TGFb by alveolar macrophage, and that this process plays a pathologic role in the bleomydn-induced lung inflammation. In light of these observations, placental bikunin and fragments thereof are contemplated as therapeutics for various fibrotic disorders, including pulmonary, hepatic, renal and dermal (scleroderma) fibrosis.

AerosUized aprotinin was shown to protect >50% of mice infected with lethal doses of either influenza virus or paramyxovirus (Ovcharenko and 2-hirnov, Antiviral Research, (1994), 23: 107-118). A suppression of the development of fatal hemorrhagic bronchopneumonia and a normalization of body weight gain were also noted with aerosUized aprotinin treatment. In light of these observations, placental bikunin and fragments thereof are contemplated as therapeutics for various respiratory related influenza-Uke diseases.

^' The human placental bikunin, isolated domains, and other variants of the invention are contemplated for use in the medical/therapeutic applications suggested for native aprotinin or aprotinin analogues with other inhibitory profiles, in particular those which necessitate usage of large doses. These would include diseases for which use of the human protein is indicated by virtue of its abUity to inhibit human serine proteases such as trypsin, plasmin, kaUikrein^", elastase, cathepsin G and proteinase-3, which indude and are not limited to: acute pancreatitis (pancreatic elastase and trypsin), inflammation, thrombocytopenia, preservation of platelet function, organ preservation, wound healing, various forms of shock, induding shock lung, endotoxin shock and post operative complications; disturbances of blood coagulation such as hyperfibrinolytic hemorrhage; acute and chronic inflammatory reactions, in particular for the therapy and prophylaxis of organ lesions, such as for example pancreatitis and radiation induced enteritis, complex-mediated inflammatory reactions such as immunovasculitis, glomerulonephritis and types of arthritis; collagenoses in particular rheumatoid arthritis; types of arthritis caused by metabolism-related deposits (for example gout); degeneration of the elastic constituents of the connective tissue parts of organs, such as in atherosderosis (serum elastase) or pulmonary emphysema (neutrophil elastase); adult respiratory distress syndrome, inflammatory bowel disease, and psoriasis. A major unexpected finding was that the synthetic peptides encoding bikunin (7-64), and bikunin (102-159), could properly fold into the correct three- dimensional conformation having active protease inhibitor bioactivity (Examples 2 and 1, respectively). Upon folding, each of these fragments of Bikunin underwent a reduction in mass of 6 mass units, consistent with the formation in each case, of three intrachain disulfide bonds between six cysteine residues of each fragment. Another surprising finding is that the synthetic peptides encoding bikunin (7-64), bikunin (102-159), and bikunin (1-170) are highly inhibitory of plasmin and both tissue and plasma kaUikrein (Example 4 , 3, and 10 respectively). Inhibition of plasmin and kaUikrein by Trasylol® is thought to be involved in the mechanism by which Trasylol® reduces blood loss during open heart surgery. Our unexpected findings of the specifidty of the Kunitz domains of the present invention make them suitable therapeutic agents for blood sparing during surgery or trauma where there is significant blood loss, or for any other condition where inhibition of plasmin and /or kaUikrein would be beneficial.

Furthermore, we showed in this disdosure (Example 10) that placental bikunin (1-170) is a potent inhibitor of factor XIa and a moderate inhibitor of factor Xa. Factor XIa plays an essential role in the intrinsic pathway of coagulation, serving to interconvert inactive factor IX into active factor IXa. Thus, Placental Bikunin inhibits two key enzymes of the intrinsic pathway, kaUikrein and factor XIa. Consistent with these observations, we also showed that placental bikunin (1-170) is a potent inhibitor of the activated partial thromboplastin time, which is. a measure of the speed of coagulation driven by the intrinsic pathway. On the other hand, we showed that Placental bikunin (1- 170) is an extremely weak inhibitor of the tissue factor Vila complex, suggesting that it is not important in the regulation of the extrinsic coagulation cascade. Based on these unexpected findings, placental bikunin is contemplated as a medicament for diseases in which activation of the intrinsic pathway of coagulation contributes significantly to the disease mechanism. Examples of such diseases would include post-traumatic shock and disseminated intravascular coagulation. A significant advantage of the Kunitz domains of the present invention-is that they are human proteins, and also less positively charged than Trasylol® (Example 1), thereby reducing the risk of kidney damage on administration of large doses of the proteins. Being of human origin, the protein of the instant invention can thus be administered to human patients with significantly reduced risk of undesired immunological reactions as compared to administration of siπύlar doses of Trasylol®. Furthermore, it was found that bikunin (102-159), bikunin (7-64), and bikunin (1-170) are significantly more potent inhibitors of plasma kalUkrein than Trasylol® in vitro (Example 3, 4 and 10). Thus bikunin and fragments thereof are expected to be more effective in vivo at lowering blood loss in patients.

The amount of serine protease inhibitor administered should be sufficient to provide a supra normal plasma level. For the prophylactic reduction of bleeding during and foUowing coronary aortic by-pass surgery (C ABG), the proteins of the instant invention may be used in place of Trasylol® whUe taking into account the differences in potency. The use of Trasylol® is outlined in the Physidans Desk Reference, (1995), listing for Trasylol® supplement A. Briefly, with the patient in a supine position, the loading dose of placental bikunin, isolated domain or other variant is given slowly over about 20 to 30 minutes, after induction of anesthesia but prior to sternotomy. In general, a total dose of between about 2x10⁶ KIU (kalUkrein inhibitory units) and 8 X10⁶ KIU will be used, depending on such factors as patient weight and the length of the surgery. Preferred loading doses are those that contain a total of 1 to 2 miUion kaUikrein inhibitory units (KIU). When the loading dose is complete, it is followed by the constant infusion dose, which is continued until surgery is complete and the patient leaves the operating room. Preferred constant infusion doses are in the range of about 250,000 to 500,000 KIU per hour. The pump prime dose is added to the priming fluid of the cardiopulmonary bypass circuit, by replacement of an aUquot of the priming fluid prior to the institution of the cardiopulmonary bypass. Preferred pump prime doses are those that contain a total of about one to two miUion KIU. The proteins of the instant invention are employed in pharmaceutical compositions formulated in the manner known to the art. Such compositions contain active ingredient(s) plus one or more pharmaceuticaUy acceptable carriers, dUuents, fUlers, binders, and other excipients, depending on the administration mode and dosage form contemplated. Examples of therapeutically inert inorganic or organic carriers known to those skiUed in the art include, but are not limited to, lactose, corn starch or derivatives thereof, talc, vegetable oUs, waxes, fats, polyols such as polyethylene glycol, water, saccharose, alcohols, glycerin and the Uke. Various preservatives, emulsifiers, dispersants, flavorants, wetting agents, antioxidants, sweeteners, colorants, stabUizers, salts, buffers and the like can also be added, as required to assist in the stabUization of the formulation or to assist in increasing bioavaUabiUty of the active ingredient(s) or to yield a formulation of acceptable flavor or odor in the case of oral dosing. The inhibitor employed in such compositions may be in the form of the original compound itself, or optionally, in the form of a pharmaceutically acceptable salt. The proteins of the instant invention can be adminstered alone, or in various combinations, and in combination with other therapeutic compositions. The compositions so formulated are selected as needed for administration of the inhibitor by any suitable mode known to those skiUed in the art.

Parenteral administration modes indude intravenous (i.v.), subcutaneous (s.c. ), intraperitoneal (i.p.), and intramuscular (i.m. ) routes. Intravenous administration can be used to obtain acute regulation of peak plasma concentrations of the drug as might be needed. Alternatively, the drug can be administered at a desired rate continuously by i.v. catheter. Suitable vehicles indude sterile, non-pyrogenic aqueous diluents, such as sterile water for injection, sterile-buffered solutions or sterile saline. The resulting composition is administered to the patient prior to and/or during surgery by intravenous injection or infusion.

Improved half-Ufe and targeting of the drug to phagosomes such as neutrophUs and macrophage involved in inflammation may be aided by entrapment of the drug in liposomes. It should be possible to improve the selectivity of liposomal targeting by incorporating into the outside of the liposomes ligands that bind to macromolecules specific to target organs /tissues such as the GI tract and lungs. Alternatively, i.m. or s.c. deposit injection with or without encapsulation of the drug into degradable microspheres (e.g., comprising poly-DL-lactide-co-glycoUde) or protective formulations containing collagen can be used to obtain prolonged sustained drug release. For improved convenience of the dosage form it is possible to use an i.p. implanted reservoir and septum such as the percuseai system. Improved convenience and patient compUance may also be achieved by use of either injector pens (e.g., the Novo Pin or Q-pen) or needle-free jet injectors (e.g., from Bioject, Mediject or Becton Dickinson). Precisely controUed release can also be achieved using implantable pumps with delivery to the desired site via a cannula. Examples include the subcutaneously implanted osmotic pumps available from ALZA such as the ALZET osmotic pump.

Nasal delivery may be achieved by incorporating the drug into bioadhesive particulate carriers (<200 mm) such as those comprising cellulose, polyacrylate or polycarbopbil, in conjunction with suitable absorption enhancers such as phosphoUpids or acylcarnitines. CommerdaUy available systems indude those developed by Dan Biosys and Sdos Nova.

Pulmonary delivery represents a nonparenteral mode of administration of the drug to the circulation. The lower airway epithelia are highly permeable to a wide range of proteins of molecular sizes up to about 20 kDa. Micron-sized dry powders containing the medicament in a suitable carrier such as mannitol, sucrose or lactose may be deUvered to the distal alveolar surface using dry powder inhalers such as those of Inhale™, Dura™, Fisons (Spinhaler™), and Glaxo (Rotahaler™), or Astra (Turbohaler™) propellant based metered dose inhalers. Solution formulations with or without liposomes may be deUvered using ultrasonic nebulizers.

Oral delivery may be achieved by incorporating the drug into tablets, coated tablets, dragέes, hard and soft gelatin capsules, solutions, emulsions, suspensions or enteric coated capsules designed to release the drug into the colon where digestive protease activity is low. Examples of the latter indude the OROS-CT/Osmet™ system of ALZA, and the PULSINCAP™ system of Scherer Drug Delivery Systems. Other systems use azo-crosslinked polymers that are degraded by colon-specific bacterial azoreductases, or pH sensitive polyacrylate polymers that are activated by the rise in pH in the colon. The above systems may be used in conjunction with a wide range of avaUable absorption enhancers. Rectal delivery may be achieved by incorporating the drug into suppositories.

In its preferred medicinal application, for reduction of perioperative blood loss, the preferred mode of administration of the placental bikunin variants of the present invention is parenterally, preferably by i.v. route through a central line.

The amount of the pharmaceutical composition to be employed wUl depend on the redpient and the condition being treated. The requisite amount may be determined without undue experimentation by protocols known to those skiUed in the art. Alternatively, the requisite amount may be calculated, based on a determination of the amount of target protease such as plasmin or kaUikrein which must be inhibited in order to treat the condition. As the active materials contemplated in this invention are deemed to be nontoxic, treatment preferably involves administration of an excess of the optimally required amount of active agent.

Additionally, placental bikunin, isolated domains or other variants may be used to isolate natural substances such as its cognate proteases from human material using affinity based separation methods, as weU as to eUdt antibodies to the protease that can be further used to explore the tissue distribution and useful functions of Placental bikunin.

Searching Human Sequence Data The existence of a distinct human protein homologous in function to aprotinin, was deduced following a unique analysis of sequence entries to the expressed-sequence-tag data-base (hereafter termed dbEST) at the NCBI (National Center for Biological Information, Maryland). Using the TBlastN algorithm (BLAST, or Basic Local Alignment Search Tool uses the method of Altschul et a., (1990) J. Mol Biol 215: 403-410, to search for similarities between a query sequence and aU the sequences in a data-base, protein or nudeic add in any combination), the data-base was examined for nudeotide sequences bearing homology to the sequence of bovine pre-pro-aprotinin, Trasylol®. This search of numerous dones was selectively narrowed to two particular clones which could possibly encode for a deduced amino acid sequence that would correspond to a human protein homologous in function to aprotinin. The selected nucleic add sequences were R35464 (SEQ ID NO: 12) and R74593 (SEQ ID NO: 14) that were generated from a human placental nudeic add library. The translated protein sequence in the longest open reading frame for R35464 (SEQ ID NO: 13) was missing one of the 6 cysteines that are critical for formation of the Kunitz-domain covalent structure, meaning that the nucleic add sequence of R35464 could not yield a functional inhibitor. Similarly, the longest translated open reading frame from clone R74593 (SEQ ID NO: 15) contained a stop codon 5' to the region encoding the Kunitz Uke sequence, meaning that this sequence, could not be translated to yield a functional secreted Kunitz domain. The significance of these sequences alone was undear. It was possible that they represented a) the products of pseudogenes, b) regions of untranslated mRNA, or c) the products of viable mRNA which had been sequenced incorrectly.

Discovery of Human Bikunin To spedfically isolate and determine the actual human sequence, cDNA primers were designed to be capable of hybridizing to sequences located 5' and 3' to the segment of cDNA encoding our proposed Kunitz like sequences found within R35464 and R74593. The primers used to ampUfy a fragment encoding the Kunitz Uke sequence of R74593 were CGAAGCTTCATCTCCGAAGCTCCAGACG (the 3'primer with a HindHI site; SEQ ID NO:33) and AGGATCTAGACAATAATTACCTGACCAAGGA (the 5'primer with an Xbal site; SEQ ID NO:34).

These primers were used to amplify by PCR (30 cycles) a 500 base pair product from a human placental cDNA library from Clontech (MATCHMAKER, Cat #HL4003AB, Clontech Laboratories, Palo Alto, CA), which was subcloned into Bluescript-SK+ and sequenced with the T3 primer with a Sequenase™ kit version 2.0. Surprisingly, the sequence of the fragment obtained using our primers was different from the sequence Usted in the dbEST data base for done R74593. In particular, our new sequence contained an additional guanosine base inserted 3' to the putative stop codon, but 5' to the segment encoding the Kunitz-like sequence (Figure 3). The insertion of an additional G shifted the stop codon out of the reading frame for the Kunitz-like domain (G at base pair 114 of the corrected sequence for R74593; Figure 3).

Subsequent query of the dbEST for sequences homologous to the Kunitz- like peptide sequence of R74593 yielded H94519 derived from human retina library and N39798. These sequences contained a Kunitz-Uke sequence that was almost identical to the Kunitz-Uke domain encoded in R35464 except that it contained all six of the characteristic cysteines. Overlay of each of the nucleotide sequences with that of R74593 (corrected by the insertion of G at b,p, 114) and R35464 was used to obtain a consensus nucleotide sequence for a partial human placental bikunin (SEQ ID NO: 9; Figure 3). The translated consensus sequence yielded an open reading frame extending from residue -18 to +179 (Figure 3; full translation SEQ ID NO: 10) that contained two complete Kunitz-like domain sequences, within the region of amino acid residues 17-64 and 102-159 respectively.

Further efforts attempted to obtain additional 5' sequence by querying dbEST with the sequence of R35464. Possible matches from such searches, that possessed additional 5' sequence were then in turn used to re-query the dbEST. In such an iterative fashion, a series of overlapping 5' sequences were identified which included clones H16866, T66058, R34808, R87894, N40851 and N39876 (Figure 4). AUgnment of some of these sequences suggested the presence of a 5' ATG which might serve as a start site for synthesis of the consensus translated protein sequence. From this selected information, it was now possible to selectively screen for, and determine the nudeic acid and polypeptide sequences of a human protein with homologous function to aprotinin.

Re-interrogation of the dbEST revealed a number of new EST entries shown schematically in Figure 4B. Overlap with these additional ESTs allowed us to construct a much longer consensus oUgonucleotide sequence (Figure 4C) that extended both 5' and 3' beyond the original oligonudeotide sequence depicted in Figure 3. In fact, the new sequence of total length 1.6 kUobases extended all the way to the 3' poly-A taU. The increased number of overlapping ESTs at each base-pair position along the sequence improved the level of confidence in certain regions such as the sequence overlapping with the 3' end of EST R74593 (Figure 3). Several overlapping ESTs in this region corroborated two critical base deletions relative to R74593 (located as bold underUned in Figure 4C, map positions 994 and 1005). Translation of the new consensus sequence (Figure 4D) in the bikunin encoding frame yielded a form of placental bikunin that was larger (248 amino acids) than the mature sequence (179 amino adds) encoded from the original consensus (SEQ ID NO: 1), and was terminated by an in-frame stop codon within the oligonudeotide consensus. The size increase was due to a frame shift in the 3' coding region resulting from removal of the two base insertions unique to EST R74593. The frame shift moved the stop codon of the original consensus (Figure 3) out of frame enabling read through into a new frame encoding the additional amino acid sequence. The new translation product (Figure 4D) was identical to the original protein consensus sequence (SEQ ID NO: 1) between residues +1 to +175 (encoding the Kunitz domains), but contained a new C-terminal extension exhibiting a putative 24 residue long transmembrane domain (underlined in Figure 4D) foUowed by a short 31 residue cytoplasmic domain. The predse sequence around the initiator methionine and signal peptide was somewhat tentative due to considerable heterogeneity amongst the overlapping ESTs in this region.

Analysis of the protein sequence by Geneworks™, highlighted asparagine residues at positions 30 and 67 as consensus sites for putative N- linked glycosylation. Asparagine 30 was not observed during N-terminal sequendng of the fuU length protein isolated from human placenta, consistent with it being glycosylated.

Cloning of Human Bikunin The existence of a human mRNA corresponding to the putative human bikunin nudeotide sequence inferred from the analysis of Figure 3, was confirmed as foUows. The nudeic acid primer hybridizing 5' to the Kunitz - encoding cDNA sequence of R35464 (b.p. 3-27 of consensus nudeotide sequence in Figure 3): GGTCTAGAGGCCGGGTCGTTTCTCGCCTGGCTGGGA

(a 5' primer derived from R35464 sequence with an Xbal site; SEQ ID NO: 35), and the nudeic acid primer hybridizing 3' to the Kunitz encoding sequence of R74593 (b.p. 680-700 of consensus nudeotide sequence in Figure 3), was used to PCR ampUfy, from a Clontech human placental Ubrary, a fragment of the size (ca. 670 b.p) expected from a cDNA consensus nudeotide sequence encoding the placental bikunin sequence of Figure 3 (Shown schematically in Figure 4A). Using a 5' primer hybridizing to a sequence in R87894 that is 126 b.p 5' to the putative ATG start site discussed above, (shown schematicaUy in Figure 4A at b.p. 110) plus the same 3' primer to R74593 as used above, it was possible to amplify a fragment from a Clontech human placental Ubrary of the expected size (approximately 872 b.p) predicted by EST overlay (Shown schematically in Figure 4).

Sequencing of the 872 b.p. fragment showed it to contain nudeotide segment corresponding to b.p. 110 to 218 of EST R87894 at its 5' end and b.p. 310 to 542 of the consensus sequence for placental bikunin inferred from the EST overlay analysis (of Figure 3), at its 3' end. This 3' nucleotide sequence contained aU of the Kunitz-like domain encoded by placental bikunin (102-159). To obtain a cDNA encoding the entire extracellular region of the protein, the following 5' PCR primer: CACCTGATCGCGAGACCCC (SEQ ID NO: 36) designed to hybridize to a sequence within EST R34808 was used with the same 3' primer to EST 74593 to amplify (30 cycles) an approximately 780 base-pair cDNA product from the human placental cDNA library. This product was gel purified, and doned into the TA vector (Invitrogen) for DNA sequencing by the dideoxy method (Sanger F., et al., (1977) Proc. Natl. Acad. Sci (USA), 74: 5463- 5467) with the following primers:

Vector Specific: GATTTAGGTGACACTATAG ( SP6 ) ( SEQ ID NO : 37 )

TAATACGACTCACTATAGGG (T7) (SEQ ID NO: 38)

GeneSpedfi T ACCTGACCAAGGAGGAGTGC (SEQ ID NO: 39) AATCCGCTGCATTCCTGCTGGTG (SEQ ID NO: 40)

CAGTCACTGGGCCTTGCCGT (SEQ ID NO: 41)

The resulting cDNA sequence is depirted in Figure 4E together with its translation product. At the nudeotide level, the sequence exhibited only minor differences from the consensus EST sequence (Figure 4D). Translation of the sequence yielded a coding sequence containing an in-frame initiator ATG site, signal peptide and mature placental bikunin sequence and transmembrane domain. The translated sequence of the PCR product was missing the last 12 amino add residues from the cytoplasmic domain as a consequence of the choice of selection of the 3' primer for PCR amplification. This choice of 3' PCR primer (designed based on the sequence of R74593) was also responsible for the introduction of an artif actual S to F mutation at amino add position 211 of the translated PCR-derived sequence. The signal peptide deduced from translation of the PCR fragment was somewhat different to that of the EST consensus. To obtain a full length placental bikunin cDNA, the PCR derived product

(Figure 4E) was gel purified and used to isolate a non-PCR based full length clone representing the bikunin sequence. The PCR derived cDNA sequence was labeled with ³²P-CTP by High Prime (Boehringer Mannheim) and used to probe a placental cDNA Library (Stratagene, Unizap™ λ Ubrary) using colony hybridization techniques. Approximately 2 X 10⁶ phage plaques underwent 3 rounds of screening and plaque purification. Two clones were deemed full length (-1.5 kUobases) as determined by restriction enzyme analysis and based on comparison with the size of the EST consensus sequence (see above). Sequendng of one of these done by the dideoxy method yielded the oligonucleotide sequence depicted in Figure 4F. The translation product from this sequence yielded a protein with inframe initiator methionine, signal peptide and mature placental bikunin sequence. The mature placental bikunin sequence was identical to the sequence of the mature protein derived by translation of the EST consensus although the signal peptide sequence lengths and sequences differed. Unlike the PCR derived product, the cDNA derived by colony hybridization contained the entire ectodomain, transmembrane domain, cytoplasmic domain and in-frame stop codon. In fact, the clone extended aU the way to the poly-A tail. The initiator methionine was followed by a hydrophobic signal peptide which was identical to the signal peptide encoded in the PCR derived clone. Subsequently we expressed and purified a soluble fragment of placental bikunin, bikunin (1-170), from Sf9 cells (Example 9), and found it to be a functional protease inhibitor (Example 10). Furthermore, we isolated from human placenta a soluble fragment of placental bikunin which was also an active protease inhibitor (Example 7). Both the natural protein and the form of the protein expressed in Sf9 ceUs are probably glycosylated at the asparagine residue at position 30 based on the recoveries of PTH-amino acids during N- terminal sequencing (Examples 7 and 9).

Based on the above observations, it seems that full length placental bikunin has the capadty to exist as a transmembrane protein on the surface of cells as well as a soluble protein. Other transmembrane proteins that contain Kunitz domains are known to undergo proteolytic processing to yield mixtures of soluble and membrane associated forms. These include two forms of the Amyloid Precursor Protein termed APP751 (Esch F., et al., (1990) Sdence, 248: 1122-1124) and APP 770 (Wang R., et al., (1991), J. Biol Chem, 266: 16960-16964).

Contact activation is a process which is activated by exposure of damaged vascular surfaces to components of the coagulation cascade. Angiogenesis is a process that involves local activation of plasmin at endotheUal surfaces. The specifidty of placental bikunin and its putative capadty to anchor to ceU surfaces, suggest that the physiologic functions of transmembranous placental bikunin may indude regulation of contact activation and angiogenesis. The amino add sequences for placental bikunin (7-64), bikunin (102-159), and fuU length placental bikunin (Figure 4F) were searched against the PIR (Vers.46.0) and PatchX (Vers. 46.0) protein databases as weU as the GeneSeq (Vers. 20.0) protein database of patented sequences using the Genetics Computer Group program FastA. Using the Genetics Computer Group program TFastA (Pearson and Lipman, 1988, Proc. Natl. Acad. Sci. USA 85: 2444-2448), these same protein sequences were searched versus the six-frame translations of the GenBank (Vers. 92.0 with updates to 1/26/96) and EMBL (modified Vers. 45.0) nucleotide databases as weU as the GeneSeq (Vers. 20.0) nudeotide database of patented sequences. The EST and STS subsets of GenBank and EMBL were not included in this set of searches. The best matches resulting from these searches contained sequences which were only about 50% identical over their full length to the 58-amino acid protein sequence derived from our analysis of dones R74593 and R35464.

Isolation of Human Bikunin

As mentioned above, synthetic peptides corresponding to bikunin (7-64) and bikunin (102-159) as determined from the translated consensus sequence for bikunin (Figure 3), could be refolded (Examples 2 and 1, respectively) to yield active kalUkrein inhibitor protein (Example 4 and 3, respectively). We exploited this unexpected property to devise a purification scheme to isolate native placental bikunin from human tissue. Using a purification scheme which employed kallikrein-sepharose affinity chromatography as a first step, highly purified native potent kallikrein inhibitor was isolated. The isolated native human bikunin had an identical N- terminus (sequenced for 50 amino add residues) as the sequence predicted by the translation of the consensus nudeic add sequence (Figure 3) amino add residues +1 to +50 (Example 7). This confirmed for the first time the existence of a novel native kalUkrein inhibitor isolated from human placenta.

Known Kunitz-Uke domains are listed below. Residues believed to be making contact with target proteases are highUghted as of spedal interest (bold /underlined). These particular residues are named positions Xaal'l⁶ f_or specific reference as shown by label Xaa below:

1 1 111 1 1

1 2 3 456789 0 1 234 5 6

1) IHDFCLVSKVV GRCRASMPRW WYNVTDGSCQ LFVYGGCDON SNNY TKEEC LKKCATV 2) YEEYCTA AVT GPCRASPPRW YFDVERNSCN NFIYGGCRGW KNSYRSEEAC MLRCFRQ

3) -HSFCAFKADD GPCKAIMKRF FFNIFTRQCE EFIYOGCBQW Q RFBSLEEC KKMCTRD

4) -PDFCFLEEDP GICROTITRY FY NQTKQCE RFKYOGCLCM MNNFETLEEC KNICEDQ

5) -PSWCLTPADR GLCRAHEHRF YYNSVIGKCR PFKYSGCOGW EN FTSKQEC LRACKKG

6) -AEICLLPLDY GPCRALLLRY YYRYRTQSCR QF YCGCBGN ANNFYTWEAC DDACWRI 7) -PSFCYSPKDB G CSAKVTRY YFNPRYRTCD AFTYTQCOOM DNNFVSREDC KRACAKA

8) -KAVCSQEAMT GPCRAVMPRT TFDLSKGKCV RFITGGCOGN RN FBSEDYC MAVCKAM

9) RPDFC EPPYT GPCKARIIRY FYNAKAGLCQ TFVYOGCR-U RNNFKSAEDC iMRTCGGA

10) CQLGYSA GPCMQMTSRY FYNGTSMACE TFQYOGC-*» GNNFVTEKEC LQTC

11) VAACN PIVR GPCRAFIQIiW AFDAVKGKCV LFPYGGCQf-W GMKFYSEKEC REYCGVP 12) -BVCCSEQAET GPCRAMI3RW YFDVTEGKCA PFFYGGCOGK RNNFDTEEYC MAVCGSA

13) CK PKDB GTCRDPILKW YYDPNTKSCA RFWYGGCOOH EMKFOSQKEC EKVC

14) -PNVCAFPMEK GPCQTϊMTR FFNFETGECE FAYGGCOCM SN F RKEKC EKFCKFT

Where sequence number 1) is Bikunin (7-64) (SEQ ID NO: 4); sequence 2) is Bikunin (102-159) (SEQ ID NO: 6); sequence 3) is Tissue factor pathway inhibitor precursor 1 (SEQ ID NO: 18); sequence 4) is Tissue fartor pathway inhibitor precursor 1 (SEQ ID NO: 19); sequence 5) is Tissue factor pathway inhibitor precursor (SEQ ID NO: 20); sequence 6) is Tissue factor pathway inhibitor precursor 2 (SEQ ID NO: 21); sequence 7) is Tissue factor pathway inhibitor precursor 2 (SEQ ID NO: 22); sequence 8) is Amyloid precursor protein homologue (SEQ ID NO: 23); sequence 9) is Aprotinin (SEQ ID NO: 24); sequence 10) is Inter-α-trypsin inhibitor precursor (SEQ ID NOs: 25); sequence 11) is Inter-α-trypsin inhibitor precursor (SEQ ID NOs: 26); sequence 12) is Amyloid precursor protein (SEQ ID NO: 27); sequence 13) is Collagen α-3(VI) precursor (SEQ ID NO: 28); and squence 14) is HK3-B9 (SEQ ID NO: 29).

It can be seen that Placental Bikunin (7-64) and (102-159) each have the same number (six) and spadng of cysteine residues as is found in members of the Kunitz dass of serine protease inhibitors. The precise bonding of cysteine residues to form the three intrachain disulfide bonds is known and invarient for aU previously known Kunitz family members (Laskowski, M et al., 1980, Ann. Rev. Biochem. 49:593-626). Based on this known bonding pattern and the fact that the folding of Placental Bikunin (7-64) and (102-159) into active protease inhibitors is accompanied by a mass reduction consistent with the formation of three intrachain disulfide bonds (Examples 2 and 1), it is highly probable that the disulfide bonding within the Kunitz domains of Placental Bikunin occur between cysteine residues: Cll and C61; C20 and C44; C36 and C57; C106 and C156; C115 and C139; C131 and C152. Furthermore, this pattern of disulfide bonding is highly probable in larger forms of Placental Bikunin containing both Kunitz domains since such forms of the protein are also active serine protease inhibitors and because N-terminal sequencing (Example 7) of native Placental Bikunin for 50 cycles yielded a sequence that was silent at positions where the cysteine residues were expected. The placental bikunin, isolated domains or other variants of the present invention may be produced by standard soUd phase peptide synthesis using either t-Boc chemistry as described by Merrifield R.B. and Barany G., in: The peptides, Analysis, Synthesis, Biology, 2, Gross E. et al., Eds. Academic Press (1980) Chapter 1; or using F-moc chemistry as described by Carpino L. A., and Han G.Y., (1970) J. Amer Chem Soc, 92, 5748-5749, and Ulustrated in Example 2. Alternatively, expression of a DNA encoding the placental bikunin variant may be used to produce recombinant placental bikunin variants.

The invention also relates to DNA constructs that encode the Placental bikunin protein variants of the present invention. These constructs may be prepared by synthetic methods such as those described in Beaucage S.L. and Caruthers M.H., (1981) Tetrahedron Lett, 22, ppl859-1862; Matteucd M.D and Caruthers M.H., (1981), J. Am. Chem. Soc. 103, p 3185; or from genomic or cDNA which may have been obtained by screening genomic or cDNA libraries with cDNA probes designed to hybridize with placental bikunin encoding DNA sequence. Genomic or cDNA sequence can be modified at one or more sites to obtain cDNA encoding any of the amino add substitutions or deletions described in this disdosure.

The instant invention also relates to expression vectors containing the DNA constructs encoding the placental bikunin, isolated domains or other variants of the present invention that can be used for the production of recombinant placental bikunin variants. The cDNA should be connected to a suitable promoter sequence which shows transcriptional activity in the host cell of choice, possess a suitable terminator and a poly-adenylation signal. The cDNA encoding the placental bikunin variant can be fused to a 5' signal peptide that wUl result in the protein encoded by the cDNA to undergo secretion. The signal peptide can be one that is recognized by the host organism. In the case of a mammaUan host ceU, the signal peptide can also be the natural signal peptide present in fuU length placental bikunin. The procedures used to prepare such vectors for expression of placental bikunin variants are weU known in the art and are for example described in Sambrook et al., Molecular Cloning: A laboratory Manual, Cold Spring Harbor, New York, (1989). The instant invention also relates to transformed cells containing the

DNA constructs encoding the placental bikunin, isolated domains or other variants of the present invention that can be used for the production of recombinant placental bikunin variants. A variety of combinations of expression vector and host organism exist which can be used for the production of the placental bikunin variants. Suitable host ceUs indude baculovirus infected Sf9 insert cells, mammaUan ceUs such as BHK, CHO, Hela and C-127, bacteria such as E. coU, and yeasts such as Saccharomyces cervisiae. Methods for the use of mammaUan, insect and microbial expressions systems needed to achieve expression of placental bikunin are weU known in the art and are described, for example, in Ausubel F.M et al., Current Protocols in Molecular Biology, John Wiley & Sons (1995), Chapter 16. For fragments of placental bikunin containing a single Kunitz inhibitor domain such as bikunin (7-64) and (102-159), yeast and E. coli expression systems are preferable, with yeast systems being most preferred. TypicaUy, yeast expression would be carried out as described in US patent 5,164,482 for aprotinin variants and adapted in Example 5 of the present spedfication for placental bikunin (102-159). E.coU expression could be carried out using the methods described in US patent 5,032,573. Use of mammaUan and yeast systems are most preferred for the expression of larger placental bikunin variants containing both inhibitor domains such as the variant bikunin (7-159).

DNA encoding variants of placental bikunin that possess amino acid substitution of the natural amino sequence can be prepared for expression of recombinant protein using the methods of Kunkel T. A., (1985) Proc. Natl. Acad. Sd USA 82: 488-492. Briefly, the DNA to be mutagenized is cloned into a single stranded bacteriophage vector such as M13. An oligonucleotide spanning the region to be dianged and encoding the substitution is hybridized to the single stranded DNA and made double stranded by standard molecular biology techniques. This DNA is then transformed into an appropriate bacterial host and verified by dideoxynudeotide sequencing. The correct DNA is then cloned into the expression plasmid. Alternatively, the target DNA may be mutagenized by standard PCR techniques, sequenced, and inserted into the appropriate expression plasmid. The following particular examples are offered by way of iUustration, and not Umitation, of certain aspects and preferred embodiments of the instant invention.

Example 1

Preparation of synthetic placental bikunin (102-159)

Materials and methods/Reagents used. The fluorogenic substrate Tos- Gly-Pro-Lys-AMC was purchased from Bachem BioScience Inc (King of Prussia, PA). PNGB, Pro-Phe-Arg-AMC, Ala-Ala-Pro-Met-AMC, bovine trypsin (type HI), human plasma kalUkrein, and human plasmin were from Sigma (St. Louis, MO).

Recombinant aprotinin (Trasylol®) was from Bayer AG (Wuppertal, Germany). Pre-loaded Gin Wang resin was from Novabiochem (La Jolla, C A). Thioanisole, ethanedithiol and t-butyl methyl ether was from Aldrich (MUwaukee, WI).

Quantification of functional placental bikunin (7-64) and (102-159)

The amount of trypsin inhibitory activity present in the refolded sample at various stages of purification was measured using GPK-AMC as a substrate. Bovine trypsin (200 pmoles) was incubated for 5 min at 37%C with bikunin (7- 64) or (102-159), from various stages of purification, in buffer A (50 mM Hepes, pH 7.5, 0.1 M NaCl, 2 mM CaCl2 and 0.01% triton X-100). GPK-AMC was added (20 μM final) and the amount of coumarin produced was determined by measuring the fluorescence (ex = 370 nm, em = 432 nm) on a Perkin-Elmer LS- 50B fluorimeter over a 2 min. period. For samples being tested the % inhibition for each was calculated according to equation 1; where R_Q is the rate of fluorescence increase in the presence of inhibitor and R\ is the rate determined in the absence of added sample. One unit of activity for the inhibitor is defined as the amount needed to achieve 50% inhibition in the assay using the conditions as described.

% inhibition = 100 x [1 - R₀/Rl. (1)

Synthesis. Placental bikunin (102-159) was synthesized on an Applied Biosystems model 420A peptide synthesizer using NMP-HBTU Fmoc chemistry. The peptide was synthesized on pre loaded Gin resin with an 8-fold excess of amino acid for each coupling. Cleavage and deprotection was performed in 84.6% trifluoroacetic acid (TFA), 4.4% thioanisole, 2.2% ethanedithiol, 4.4% Uquified phenol, and 4.4% H2O for 2 hours at room temperature. The crude peptide was predpitated, centrifuged and washed twice in t-butyl methyl ether. The peptide was purified on a Dynamax 60A C18 reverse-phase HPLC column using a TFA/acetonitrUe gradient. The final preparation (61.0 mg) yielded the correct amino acid composition and molecular mass by Electrospray mass spectroscopy (MH+ =6836.1; calcd = 6835.5) for the predicted sequence:

YEEYCTANAV TGPCRASFPR WYFDVERNSC NNFIYGGCRG NKNSYRSEEA CMLRCFRQ ( SEQ ID NO : 6 )

Purification. Refolding of placental bikunin (102-159) was performed according to the method of Tarn et al., (J. Am. Chem. Soc. 1991, 113: 6657-62). A portion of the purified peptide (15.2 mg) was dissolved in 4.0 ml of 0.1 M Tris, pH 6.0, and 8 M urea. Oxidation of the disulfides was accomplished by dropwise addition of a solution containing 23% DMSO, and 0.1 M Tris, pH 6.0 to obtain a final concentration of 0.5 mg/ml peptide in 20% DMSO, 0.1 M Tris, pH 6.0, and 1 M urea. The solution was aUowed to stir for 24 hr at 25°C after which it was diluted 1:10 in buffer containing 50 mM Tris, pH 8.0, and 0.1 M NaCl. The material was purified using a kallikrein affinity column made by covalently attaching 30 mg of bovine pancreatic kaUikrein (Bayer AG) to 3.5 mis of CNBr activated Sepharose (Pharmada) according to the manufacturers instructions. The refolded material was loaded onto the affinity column at a flow rate of 1 ml/min and washed with 50 mM Tris, pH 8.0, and 0.1 M NaCl until absorbance at 280 nm of the wash could no longer be detected. The column was eluted with 3 volumes each of 0.2 M acetic add, pH 4.0 and 1.7. Active fractions were pooled (see below) and the pH of the solution adjusted to 2.5. The material was directly appUed to a Vydac C18 reverse-phase column (5 micron, 0.46 x 25 cm) which had been equilibrated in 22.5% acetonitrile in 0.1% TFA. Separation was achieved using a Unear gradient of 22.5 to 40% acetonitrile in 0.1% TFA at 1.0 ml/min over 40 min. Active fractions were pooled, lyophUized, redissolved in 0.1% TFA, and stored at -20°C until needed.

Results. Synthetic placental bikunin (102-159) was refolded using 20% DMSO as the oxidizing agent as described above, and purified by a 2-step purification protocol as shown below, to yield an active trypsin inhibitor (Table 1 below). Table 1

Purification table for the isolation of synthetic placental bikunin (102-159)

^aProtein determined by AAA.

^Protein determined by OD280 nm using the extinction coeffident determined for the purified protein (1.7 x 10⁴ Lmol"! cm'l). ^cOne Unit is defined as the amount of material required to inhibit 50% of trypsin activity in a standard assay.

Chromatography of the crude refolded material over an immobilized bovine pancreatic kaUikrein column selectively isolated 6.0% of the protein and 97% of the trypsin inhibitory activity present. Subsequent chromatography using C18 reverse-phase yielded a further purification of 2-fold, with an overall recovery of 74%. On RPHPLC, the reduced and refolded placental bikunin (102-159), exhibited elution times of 26.3 and 20.1 minutes, respectively. Mass spectroscopy analysis of the purified material revealed a molecular mass of 6829.8; a loss of 6 mass units from the starting material. This demonstrates the complete formation of the 3 disulfides predicted from the peptide sequence.

The isoelectric points of the purified, refolded synthetic placental bikunin (102-159) was determined using a Multiphor II Electrophoresis System (Pharmada) run according to the manufacturers suggestions, together with pi standards, using a precast Ampholine® PAGplate (pH 3.5 to 9.5) and focused for 1.5 hrs. After staining, the migration distance from the cathodic edge of the gel to the different protein bands was measured. The pi of each unknown was determined by using a standard curve generated by a plot of the migration distance of standards versus the corresponding pi's. With this technique, the pi of placental bikunin (102-159) was determined to be 8.3, in agreement with the value predicted from the amino acid sequence. This is lower than the value of 10.5 established for the pi of aprotinin. (Tenstad et al., 1994, Acta Physiol. Scand. 152: 33-50). Example 2

Preparation of synthetic placental bikunin (7-64)

Placental bikunin (7-64) was synthesized, refolded and purified essentially as described for placental bikunin (102-159) but with the following modifications: during refolding, the synthetic peptide was stirred for 30 hr as a solution in 20% DMSO at 25°C; purification by C18 RP-HPLC was achieved with a linear gradient of 25 to 45% acetonitrile in 0.1% TFA over 40 min (lml/min). Active fractions from the first C18 run were reapplied to the column and fractionated with a linear gradient (60 min, 1 ml/min) of 20 to 40% acetorύtrUe in 0.1% TFA.

Results. The final purified reduced peptide exhibited an MH+ =

6563, consistent with the sequence:

IHDFCLVSKV VGRCRASMPR WWYNVTDGSC QLFVYGGCDG NSNNYLTKEE CLKKCATV (SEQ ID NO: 4)

The refolding and purification yielded a functional Kunitz domain that was active as an inhibitor of trypsin (Table 2 below).

Table 2A

Purification table for the isolation of synthetic placental bikunin (7-64)

TABLE __A

The purified refolded protein exhibited an MH+ = 6558, i.e. 5±1 mass units less than for the reduced peptide. This demonstrates that refolding caused the formation of at least one appropriate disulfide bond.

The pi of placental bikunin (7-64) was determined using the methods employed to determine the pi of placental bikunin (102-159). Placental bikunin (7-64) exhibited a pi that was much higher than the predicted value (pi = 7.9). Refolded placental bikunin (7-64) migrated to the cathodic edge of the gel (pH 9.5) and an accurate pi could not be determined under these conditions.

Continued Preparation of synthetic placental bikunin (7-64) Because the synthetic placental bikunin (7-64) may not have undergone complete deprotection prior to purification and refolding, refolding was repeated using protein which was certain to be completely deprotected. Placental bikunin (7- 64) was synthesized, refolded and purified essentially as described for placental bikunin (102-159) but with the foUowing modifications: during refolding, the synthetic peptide (0.27 mg/ml) was stirred for 30 hr as a solution in 20% DMSO at 25 C; purification by C18 RP-HPLC was achieved with a linear gradient of 22.5 to 50% acetonitrile in 0.1% TFA over 40 min (1 ml/min).

Results. The final purified reduced peptide exhibited an MH+ = 6567.5 , consistent with the sequence:

IHDFCLVSKV VGRCRASMPRW WYNVTDGSC QLFVYGGCDG NSNNYLTKEE CLKKCATV (SEQ ID NO: 4)

The refolding and purification yielded a functional Kunitz domain that was as active as an inhibitor of trypsin (Table 2B below).

Table 2B

Purification table for the isolation of synthetic placental bikunin (7-64)

The purified refolded protein exhibited an MH+ = 6561.2, i.e. 6.3 mass units less than for the reduced peptide. This demonstrates that refolding caused the formation of the expected three disulfide bonds.

The pi of refolded placental bikunin (7-64) was determined using the methods employed to determine the pi of placental bikunin (102-159). Refolded placental bikunin (7-64) exhibited a pi of 8.85, slightly higher than the predicted value (pi = 7.9). Example 3

In vitro specificity of functional placental bikunin fragment (102-159) Proteases . Bovine trypsin, human plasmin, and bovine pancreatic kaUikrein quantitation was carried out by active site titration using p- nitrophenyl p'-guanidinobenzoate HCI as previously described (Chase,T., and Shaw, E., (1970) Methods Enzmol., 19: 20-27). Human kaUikrein was quantitated by active site titration using bovine aprotinin as a standard and PFR-AMC as a substrate assuming a 1:1 complex formation. The K_m for GPK- AMC with trypsin and plasmin under the conditions used for each enzyme was 29 μM and 726 μM, respectively; the K_m for PFR-AMC with human plasma kaUikrein and bovine pancreatic kaUikrein was 457 μM and 81.5 μM, respectively; the K_m for AAPR-AMC with elastase was 1600 μM. Human tissue kaUikrein (Bayer, Germany) quantification was carried out by active site titration using p'nitrophenyl p'-guanidinobenzoate HCI as previously described (Chase, T., and Shaw, E., (1970) Methods Enzmol. 19: 20-27).

Inhibition Kinetics: The inhibition of trypsin by placental bikunin (102- 159) or aprotinin was measured by the incubation of 50 pM trypsin with placental bikunin (102-159) (0-2 nM) or aprotinin (0-3 nM) in buffer A in a total volume of 1.0 ml. After 5 min. at 37°C, 15 μl of 2 mM GPK-AMC was added and the change in fluorescence (as above) was monitored. The inhibition of human plasmin by placental bikunin (102-159) and aprotinin was determined with plasmin (50 pM) and placental bikunin (102-159) (0-10 nM) or aprotinin (0- 4 nM) in buffer containing 50 mM Tris-HCl (pH 7.5), 0.1 M NaCl, and 0.02% triton x-100. After 5 min. incubation at 37°C, 25 μl of 20 mM GPK-AMC was added and the change in fluorescence monitored. The inhibition of human plasma kallikrein by placental bikunin (102-159) or aprotinin was determined using kallikrein (2.5 nM) and placental bikunin (102-159) (0-3 nM) or aprotinin (0-45 nM) in 50 mM Tris-HCl (pH 8.0), 50 mM NaCl, and 0.02% triton x-100. After 5 min. at 37°C 15 μl of 20 mM PFR-AMC was added and the change in fluorescence monitored. The inhibition of bovine pancreatic kaUikrein by placental bikunin (102-159) and aprotinin was determined in a similar manner with kalUkrein (92 pM), placental bikunin (102-159) (0-1.6 nM) and aprotinin (0- 14 pM) and a final substrate concentration of 100 μM. The apparent inhibition constant. K_j* was determined using the nonlinear regression data analysis program Enzfitter software (Biosoft, Cambridge, UK): The kinetic data from each experiment were analyzed in terms of the equation for a tight binding inhibitor:

Vi/V₀ = 1 - (E₀ + Io + Kj* - [(E₀ + Io + K_i*)2 - 4 E₀I₀]l/2)/2E₀ (2)

where Vj/V₀ is the fractional enzyme activity (inhibited vs. uninhibited rate), and E₀ and ΪQ are the total concentrations of enzyme and inhibitor, respectively. Ki values were obtained by correcting for the effect of substrate according to the equation:

Ki = KiV(l + [So]/K_m) (3)

(Boudier, C, and Bieth, J. G., (1989) Biochim Biophys Acta., 995: 36-41)

For the inhibition of human neutrophil elastase by placental bikunin (102-159) and aprotinin, elastase (19 nM) was incubated with placental bikunin (102-159) (150 nM) or aprotinin (0-7.5 μM) in buffer containing 0.1 M Tris-HCl (pH 8.0), and 0.05% triton X-100. After 5 min at 37%C, AAPM-AMC (500 μM or 1000 μM) was added and the fluorescence measured over a two-minute period. Ki values were determined from Dixon plots of the form 1/V versus [I] performed at two different substrate concentrations (Dixon et al., 1979).

The inhibition of human tissue kaUikrein by aprotinin, placental bikunin fragment (7-64) or placental bikunin fragment (102-159) was measured by the incubation of 0.35 nM human tissue kalUkrein with placental bikunin (7-64) (0- 40 nM) or placental bikunin (102-159) (0-2.5 nM), or aprotinin (0-0.5 nM) in a 1 ml reaction volume containing 50 mM Tris-HCl buffer pH 9.0, 50 mM NaCl, and 0.1% triton x-100. After 5 min. at 37°C, 5 ul of 2 mM PFR-AMC was added achieving 10 uM final and the change in fluorescence monitored. The Km for PFR-AMC with human tissue kalUkrein under the conditions employed was 5.7 uM. The inhibition of human factor Xa (American Diagnostica, Inc, Greenwich, CT) by synthetic placental bikunin (102-159), recombinant placental bikunin, and aprotinin was measured by the incubation of 0.87 nM human factor Xa with increasing amounts of inhibitor in buffer containing 20 mM Tris (pH 7.5), 0.1 M NaCl, and 0.1% BSA. After 5 min. at 37°C, 30 ul of 20 mM LGR-AMC (Sigma) was added and the change in fluorescence monitored. The inhibition of human urokinase (Sigma) by Kunitz inhibitors was measured by the incubation of urokinase (2.7 ng) with inhibitor in a total volume of 1 ml buffer containing 50 mM Tris-HCl (pH 8.0), 50 mM NaCl, and 0.1% Triton x-100. After 5 min. at 37°C, 35 ul of 20 mM GGR-AMC (Sigma) was added and the change in fluorescence monitored. The inhibition of Factor XIa (from Enzyme Research Labs, Southbend, IN) was measured by incubating FXIa (0.1 nM) with either 0 to 800 nM placental bikunin (7-64), 0 to 140 nM placental bikunin (102-159) or 0 to 40 uM aprotinin in buffer containing 50 mM Hepes pH 7.5, 100 mM NaCl, 2 mM CaC12, 0.01% triton x-100, and 1% BSA in a total volume of 1 ml. After 5 min at 37 C, 10 ul of 40 mM Boc-Glu(OBzl)-Ala-Arg-AMC (Bachem Biosdences, King of Prussia, PA) was added and the change in fluorescence monitored.

Results: A direct comparison of the inhibition profUes of placental • bikunin (102-159) and aprotinin was made by measuring their inhibition constants with various proteases under identical conditions. The Kj values are listed in Table 3 below.

Table 3

Ki values for the inhibition of various proteases by bikunin (102-159)

Placental bikunin (102-159) and aprotinin inhibit bovine trypsin and human plasmin to a comparable extent under the conditions employed. Aprotinin inhibited elastase with a Ki of 8.5 μM. Placental bikunin (102-159) inhibited elastase with a Ki of 323nM. The K[ value for the placental bikunin (102-159) inhibition of bovine pancreatic kallikrein was 20-fold higher than that of aprotinin inhibition. In contrast, placental bikunin (102-159) is a more potent inhibitor of human plasma kalUkrein than aprotinin and binds with a 56-fold higher affinity.

Because placental bikunin (102-159) is greater than 50 times more potent than Trasylol® as an inhibitor of kaUikrein, smaUer amounts of human placental bikunin, or fragments thereof (i.e. placental bikunin (102-159)) are needed than Trasylol® in order to maintain the effective patient doses of inhibitor in KIU. This reduces the cost per dose of the drug and reduces the likeUhood of adverse nephrotoxic effects upon re-exposure of the medicament to patients. Furthermore, the protein is human derived, and thus much less immunogenic in man than aprotinin which is derived from cows. This results in significant reductions in the risk of incurring adverse immunologic events upon re-exposure of the medicament to patients.

Example 4 In vitro specificity of functional placental bikunin fragment (7-64)

In vitro spedfidty of functional human placental bikunin (7-64) was determined using the materials and methods as described in the Examples above.

Results: The table below shows the efficacy of placental bikunin (7-

64) as an inhibitor of various serine proteases in vitro. Data is shown compared against data obtained for screening inhibition using either placental bikunin (102-159), or aprotinin (Trasylol®).

Table 4 A

Ki values for the inhibition of various proteases by bikunin(7-64)

The results show that the amino acid sequence encoding placental bikunin (7- 64) can be refolded to obtain an active serine protease inhibitor that is effective against at least four trypsin-like serine proteases.

Table 4B below also shows the efficacy of refolded placental bikunin (7-64) as an inhibitor of various serine proteases in vitro. Refolded placental bikunin (7-64) was prepared from protein that was certain to be completely deprotected prior to purification and refolding. Data is shown compared against data obtained for screening inhibition using either placental bikunin (102-159), or aprotinin (Trasylol®).

Table 4B

Ki values for the inhibition of various proteases by refolded bikunin (7-64)

Suprisingly, placental bikunin (7-64) was more potent than aprotinin at inhibiting human plasma kallikrein, and at least similar in efficacy as a plasmin inhibitor. These data show that placental bikunin (7-64) is at least as effective as aprotinin, using in vitro assays, and that one would expect better or simUar potency in vivo.

Example 5

Expression of placental bikunin variant (102-159) in yeast

The DNA sequence encoding placental bikunin 102-159 (SEQ ID NO: 6) was generated using synthetic oUgonudeotides. The final DNA product consisted (5' to 3') of 15 nucleotides from the yeast α-mating factor propeptide sequence fused to the in-frame cDNA sequence encoding placental bikunin (102-159), foUowed by an in-frame stop codon. Upon cloning into a yeast expression vector pS604, the cDNA would direct the expression of a fusion protein comprising an N-terminal yeast α-mating factor propeptide fused to the 58 amino acid sequence of placental bikunin (102-159). Processing of this fusion protein at a KEX-2 deavage site at the junction between the α-mating factor and Kunitz domain was designed to Uberate the Kunitz domain at its native N- terminus.

A 5' sense oligonudeotide of the foUowing sequence and containing a HindHI site for cloning was synthesized:

GAA GGG GTA AGC TTG GAT AAA AGA TAT GAA GAA TAC TGC ACC GCC AAC GCA GTC ACT GGG CCT TGC CGT GCA TCC TTC CCA CGC TGG TAC TTT GAC GTG GAG AGG ( SEQ ID NO: 42 )

A 3' antisense oligonucleotide of the foUowing sequence and containing both a BamHI site for doning and a stop codon was synthesized:

CGC GGA TCC CTA CTG GCG GAA GCA GCG GAG CAT GCA GGC CTC CTC AGA GCG GTA GCT GTT CTT ATT GCC CCG GCA GCC TCC ATA GAT GAA GTT ATT GCA GGA GTT CCT CTC CAC GTC AAA GTA CCA GCG (SEQ ID NO: 43)

The oiigonudeotides were dissolved in 10 mM Tris buffer pH 8.0 containing 1 mM EDTA, and 12 ug of each oligo were added combined and brought to 0.25M NaCl. To hybridize, the oiigonudeotides were denatured by boUing for 5 minutes and aUowed to cool from 65 °C to room temp over 2 hrs. Overlaps were extended using the Klenow fragment and digested with Hindπi and BamHI. The resulting digested double stranded fragment was cloned into pUC19 and sequence confirmed. A clone containing the fragment of the correct sequence was digested with BamHI/Hindiπ to Uberate the bikunin containing fragment with the foUowing + strand sequence:

GAA GGG GTA AGC TTG GAT AAA AGA TAT GAA GAA TAC TGC ACC GCC AAC GCA GTC ACT GGG CCT TGC CGT GCA TCC TTC CCA CGC TGG TAC TTT GAC GTG GAG AGG AAC TCC TGC AAT AAC TTC ATC TAT GGA GGC TGC CGG GGC AAT AAG AAC AGC TAC CGC TCT GAG GAG GCC TGC ATG CTC CGC TGC TTC CGC CAG TAG GGA TCC (SEQ ID. : 44)

which was then gel purified and gated into BamHI/Hindiπ cut pS604. The ligation mixture was extracted into phenol /chloroform and purified over a S- 200 minispin column. The Ugation product was directed transformed into yeast strains SC101 and WHL341 and plated on ura selection plates. Twelve colonies from each strain were re-streaked on ura drop out plates. A single colony was inoculated into 2 ml of ura DO media and grown over night at 30°C. Cells were peUeted for 2 minutes at 14000x g and the supematants evaluated for their content of placental bikunin (102-159).

Detection of expression of placental bikunin (102-159) in transformed yeast Firstly, the supematants (50 ul per assay) were evaluated for their capacity to inhibit the in vitro activity of trypsin using the assay methods as described in Example 1 (1 ml assay volume). An un-used media only sample as weU as a yeast clone expressing an inactive variant of aprotinin served as negative controls. A yeast clone expressing natural aprotinin served as a positive control and is shown for comparison.

The second method to quantify placental bikunin (102-159) expression exploited use of polyclonal antibodies (pAbs) against the synthetic peptide to monitor the accumulation of the recombinant peptide using Western blots. These studies were performed only with recombinants derived from strain SC101, since these produced greater inhibitory activity than recombinants derived from strain WHL341.

To produce the pAb, two 6-8 week old New Zealand White female rabbits (Hazelton Research Labs, Denver, Pa) were immunized on day zero with 250 ug of purified reduced synthetic placental bikunin (102-159), in Complete Freund's adjuvant, followed by boosts on days 14, 35 and 56 and 77 each with 125 ug of the same antigen in Incomplete Freund's adjuvant. Antiserum used in the present studies was coUected after the third boost by estabUshed procedures. Polydonal antibodies were purified from the antiserum over protein A. Colonies 2.4 and 2.5 from transformation of yeast SC101 (Figure 8) as weU as an aprotinin control were grown overnight in 50 ml of ura DO media at 30°C CeUs were peUeted and the supernatant concentrated 100-fold using a Centriprep 3 (Amicon, Beverly, MA) concentrator. Samples of each (30 μl) were subjected to SDS-PAGE on 10-20% tricine buffered gels (Novex, San Diego, CA) using the manufacturers procedures. Duplicate gels were either developed with a sUver stain kit (Integrated Separation Systems, Nantick, MA) or transferred to nitroceUulose and developed with the purified polyclonal antibody eUdted to synthetic bikunin (102-159). AlkaUne-phosphatase conjugated goat anti-rabbit antibody was used as the secondary antibody according to the manufacturer's directions (Kirkegaard and Perry, Gaithersburg, MD). Purification of placental bikunin (102-159) from a transformed strain ofSClOl Fermentation broth from a IL culture of SClOl strain 2.4 was harvested by centrifugation (4,000 g x 30 min.) then appUed to a 1.0 ml column of anhydrochymotrypsin-sepharose (Takara Biochemical Inc., CA), that was previously equUibrated with 50 mM Hepes buffer pH 7.5 containing 0. IM NaCl, 2 mM CaC-2 and 0.01% (v/v) triton X-100. The column was washed with the same buffer but containing 1.0 M NaCl until the A280nm declined to zero, whereupon the column was eluted with 0.1M formic add pH 2.5. Eluted fractions were pooled and applied to a C18 column (Vydac, 5um, 4.6 x 250 mm) previously equiUbrated with 0.1% TFA, and eluted with a 50 min. linear gradient of 20 to 80% acetonitrile in 0.1% TFA. Fractions containing placental bikunin (102-159) were pooled and re-chromatographed on C18 employing elution with a linear 22.5 to 50% acetonitrile gradient in 0.1% TFA.

Results. Figure 8 shows the percent trypsin activity inhibited by twelve colonies derived from the transformation of each of strains SClOl and WHL341. The results show that aU twelve colonies of yeast strain SClOl transformed with the trypsin inhibitor placental bikunin (102-159) had the abiUty to produce a substantial amount of trypsin inhibitory activity compared to the negative controls both of which showed no abiUty to inhibit trypsin. The activity is therefore related to the expression of a specific inhibitor in the placental bikunin variant (102-159) transformed cells. The yeast WHL341 samples contained minimal trypsin inhibitory activity. This may be correlated to the slow growth observed with this strain under the conditions employed. Figure 9 shows the SDS-PAGE and western analysis of the yeast SClOl supematants. Silver stained SDS-PAGE of supematants derived from recombinant yeasts 2.4 and 2.5 expressing placental bikunin (102-159) as well as from the yeast expressing aprotinin yielded a protein band running at approximated 6 kDa, corresponding to the size expected for each recombinant Kunitz inhibitor domain. Western analysis showed that the 6 kDa bands expressed by stains 2.4. and 2.5 reacted with the pAb eUdted to placental bikunin (102-159). The same 6 kDa band in the aprotinin control did not react with the same antibody, demonstrating the specificity of the antibody for the placental bikunin variant (102-159). The final preparation of placental bikunin C-terminal domain was highly pure by sUver-stained SDS-PAGE (Figure 10). The overall recovery of broth- derived trypsin inhibitory activity in the final preparation was 31%. N-terminal sequendng of the purified inhibitor indicated that 40% of the protein is correctly processed to yield the correct N-terminus for placental bikunin (102- 159) whUe about 60 % of the material contained a portion of the yeast α-mating factor. The purified material comprised an active serine protease inhibitor exhibiting an apparent Ki of 0.35 nM for the in vitro inhibition of plasma kaUikrein.

In condusion, the accumulation both of a protease inhibitor activity and a protein immunochemicaUy related to synthetic bikunin (102-159) in fermentation broth as weU as the isolation of placental bikunin (102-159) from one of the transformed lines provided proof of expression of placental bikunin in the recombinant yeast strains described herein, showing for the first time the utility of yeasts for the production of placental bikunin fragments.

Additional constructs were prepared in an effort to augment the expression level of the Kunitz domain contained within placental bikunin 102-159, as weU as to increase the yield of protein with the correct N-terminus. We hypothesized that the N-terπύnal residues of placental bikunin 102-159 (YEEY-) may have presented a deavage site that is only poorly recognized by the yeast KEX-2 protease that enzymicaUy removes the yeast a-factor pro-region. Therefore, we prepared yeast expression constructs for the production of placental bikunin 103-159 (N-terminus of EEY...), 101-159 (N-terminus of NYEEY...) and 98-159 (DMFNYEEY..) in order to modify the P' subsites surrounding the KEX-2 cleavage site. To attempt to augment the levels of recombinant protein expression, we also used the yeast preferred codons rather than mammaUan preferred codons in preparing some of the constructs described below. The constructs were essentiaUy prepared as described above for placental bikunin 102-159 (defined as construct #1) but with the foUowing modifications:

Construct #2 placental bikunin 103-159, yeast codon usage

A 5' sense oUgonudeotide

GAAGGGGTAA GCTTGGATAA AAGAGAAGAA TACTGTACTG CTAATGCTGT TACTGGTCCA TGTAGAGCTT CTTTTCC AAG ATGGTACTTT GATGTTGAAA GA (SEQ ID NO: 55)

and 3' antisense oligonucleotide ACTGGATCCT C ATTGGCGA A AACATCTC A A CATACAGGCT TCTTC AGATC TGTAAGAATT TTTATTACCT CTACAACCAC CGTAAATAAA ATTATTACAA GAATTTCTTT CAACATCAAA GTACCATCT (SEQ ID NO: 56)

were manipulated as described for the production of an expression construct (construct #1 above) for the expression of placental bikunin 102-159

Construct #3 placental bikunin 101-159, yeast codon usage

A 5' sense oligonucleotide

GAAGGGGTAA GCTTGGATAA AAGAAATTAC GAAGAATACT GTACTGCTAA TGCTGTTACT GGTCCATGTA GAGCTTCTTT TCC AAGATGG TACTTTGATG TTGAAAGA (SEQ ID NO: 57)

and the same 3' antisense oUgonucleotide as used for construct #2, were manipulated as described for the production of an expression construct (construct #1 above) for the expression of placental bikunin 102-159.

Construct #4 placental bikunin 98-159, yeast codon usage A 5' sense oligonudeotide

GAAGGGGTAA GCTTGGATAA AAGAGATATG TTTAATTACG AAGAATACTG TACTGCTAAT GCTGTTACTG GTCCATGTAG AGCTTCTTTT CCAAGATGGT ACTTTGATGT TGAAAGA (SEQ ID NO: 58)

and the same 3' antisense oUgonucleotide as used for construct #2, were manipulated as described for the production of an expression construct (construct #1 above).

Yeast strain SClOl (MATα, ura 3-52, sue 2) was transformed with the plasmids containing each of the above cDNAs, and proteins were expressed using the methods that were described above for the production of placental bikunin 102- 159 with human codon usage. Approximately 250 ml of each yeast culture was harvested, and the supernatant from centrifugation (15 min x 3000 RPM) separately subjected to purification over 1 ml columns of kaUikrein-sepharose as described above. The relative amount of trypsin inhibitory activity in the applysate, the amount of purified protein recovered and the N-terminal sequence of the purified protein were determined and are listed below in Table 7. Table 7

Relative production levels of different proteins containing the C-terminal Kunitz domain of placental bikunin

The results show that placental bikunin fragments of different lengths that contain the C-terminal Kunitz domain show wide variation in capacity to express functional secreted protein. Constructs expressing fragments 101-159 and 103-159 yielded Uttle or low enzymic activity in the supematants prior to purification, and N-terminal sequencing of 0.05 ml aliquots of each purified fraction yielded undetectable amounts of inhibitor. On the other hand expression either of placental bikunin 102-159 or 98-159 yielded significant amounts of protease activity prior to purification. N-terminal sequencing however showed that the purified protein recovered from expression of 102-159 was once again largely incorrectly processed, exhibiting an N-terminus consistent with processing of the majority of the pre-protein at a site within the yeast α-mating factor pro-sequence. The purified protein recovered from expression of placental bikunin 98-159 however was processed entirely at the correct site to yield the correct N-terminus. Furthermore, nearly twice as much protein was recovered as compared to the recovery of placental bikunin 102- 159. Placental bikunin 98-159 thus represents a preferred fragment length for _the production of the C-terminal Kunitz domain of placental bikunin by the α- mating factor pre-pro sequence/ KEX-2 processing system of S. cerevisiae, Example 6

Alternative procedure for yeast expression

The 58 amino acid peptide derived from the R74593 translation product can also be PCR ampUfied from either the R87894-R74593 PCR product cloned into the TA vector™ (Invitrogen, San Diego, CA) after DNA sequencing or from human placental cDNA. The amplified DNA product wiU consist of 19 nudeotides from the yeast α-mating factor leader sequence mated to the R74593 sequence which codes for the YEEY--CFRQ (58 residues) so as to make the translation product in frame, constructing an α -mating factor/Kunitz domain fusion protein. The protein sequence also contains a kex 2 cleavage which wiU Uberate the Kunitz domain at its native N-terminus.

The 5' sense oligonucleotide which contains a Hindm site for cloning wUl contain the foUowing sequence: GCCAAGCTTG GATAAAAGAT ATGAAGAAT ACTGCACCGC CAACGCA ( SEQ ID NO : 30 )

The 3' antisense oUgonucleotide contains a BamHI site for doning as well as a stop codon and is of the foUowing sequence: GGGGATCCTC ACTGCTGGCG GAAGCAGCGG AGCAT ( SEQ ID NO : 31 )

The full 206 nudeotide cDNA sequence to be cloned into the yeast expression vector is of the foUowing sequence:

CCAAGCTTGG ATAAAAGATA TGAAGAATAC TGCACCGCCA ACGCAGTCAC TGGGCCTTGC CGTGCATCCT TCCCACGCTG GTACTTTGAC GTGGAGAGGA ACTCCTGCAA TAACTTCATC TATGGAGGCT GCCGGGGCAA TAAGAACAGC TACCGCTCTG AGGAGGCCTG CATGCTCCGC TGCTTCCGCC AGCAGTGAGG ATCCCC (SEQ ID NO: 32)

After PCR amplification, this DNA wUl be digested with HindHI, BamHI and doned into the yeast expression vector pMT15 (see US patent 5,164,482, incorporated by reference in the entirety) also digested with HindlH and BamHI. The resulting plasmid vector is used to transform yeast strain SC 106 using the methods described in US patent 5,164,482. The URA 3+ yeast transformants are isolated and cultivated under inducing conditions. The yield of recombinant Placental bikunin variants is determined according to the amount of trypsin inhibitory activity that accumulated in the culture supematants- over time using the in vitro assay method described above. Fermentation broths are centrifuged at 9000 rpm for 30 minutes. The supernatant is then fUtered through a 0.4 then a 0.2 μm filter, dUuted to a conductivity of 7.5 ms, and adjusted to pH 3 with citric add. The sample is then batch absorbed onto 200 ml of S-sepharose fast flow (Pharmada) in 50 mM sodium citrate pH 3 and stirred for 60 min. The gel is subsequently washed sequentially with 2 L of each of: 50 mM sodium citrate pH 3.0; 50 mM Tris-HCL pH 9.0; 20 mM HEPES pH 6.0. The washed gel is transferred into a suitable column and eluted with a linear gradient of 0 to 1 M sodium chloride in 20 mM HEPES pH 6.0. Eluted fractions containing in vitro trypsin inhibitory activity are then pooled and further purified either by a) chromatography over a column of immobilized anhydrotrypsin (essentiaUy as described in Example 2); b) by chromatography over a column of immobiUzed bovine kaUikrein; or c) a combination of conventional chromatographic steps induding gel filtration and /or anion-exchange chromatography.

Example 7

Isolation and characterization of native human placental bikunin from placenta

Bikunin protein was purified to apparent homogeniety from whole frozen placenta (Analytical Biological Services, Inc, WUmington, DE). The placenta (740 gm) was thawed to room temperature and cut into 0.5 to 1.0 cm pieces, placed on ice and washed with 600 ml PBS buffer. The wash was decanted and 240 ml of placenta pieces placed into a Waring blender. After adding 300 ml of buffer consisting of 0.1 M Tris (pH 8.0), and 0.1 M NaCl , the mixture was blended on high speed for 2 min, decanted into 750.0 ml centrifuge tubes, and placed on ice. This procedure was repeated until aU material was processed. The combined slurry was centrifuged at 4500 x g for 60 minutes at 4°C. The supernatant was filtered through cheese doth and the placental bikunin purified using a kaUikrein affinity column made by covalently attaching 70 mg of bovine pancreatic kalUkrein (Bayer AG) to 5.0 mis of CNBr activated Sepharose (Pharmada) according to manufacturers instruction. The material was loaded onto the affinity column at a flow rate of 2.0 ml/min and washed with 0.1 M Tris (pH 8.0), 0.1 M NaCl until absorbance at 280 nm of the wash could no longer be detected. The column was further washed with 0.1 M Tris (pH 8.0), 0.5 M NaCl and then eluted with 3 volumes of 0.2 M acetic acid, pH 4.0. Fractions containing kallikrein and trypsin inhibitory (see below) activity were pooled, frozen, and lyophUized. Placental bikunin was further purified by gel-filtration chromatography using a Superdex 75 10/30 (Pharmada) column attached to a Beckman System Gold HPLC system. Briefly, the column was equUibrated in 0.1 M Tris, 0.15 M NaCl, and 0.1% Triton X-100 at a flow rate of 0.5 ml/min. The lyophUized sample was reconstituted in 1.0 ml of 0.1 M Tris, pH 8.0 and injected onto the gel-filtration column in 200 μl aliquots. Fractions were collected (0.5 ml) and assayed for trypsin and kaUikrein inhibitory activity. Active fractions were pooled, and the pH of the solution adjusted to 2.5 by addition of TFA. The material was directly applied to a Vydac C18 reverse-phase column (5 micron, 0.46 x 25 cm) which had been' equiUbrated in 20% acetonitrile in 0.1 %TFA. Separation was achieved using a linear gradient of 20 to 80% acetonitrile in 0.1% TFA at 1.0 ml/min over 50 minutes after an initial 20 minute wash at 20% acetonitrile in 0.1% TFA. Fractions (1ml) were coUected and assayed for trypsin and kalUkrein inhibitory activity. Fractions containing inhibitory activity were concentrated using a speed-vac concentrator (Savant) and subjected to N-terminal sequence analysis.

Functional assays for Placental Bikunin:

Identification of functional placental bikunin was achieved by measuring its abUity to inhibit bovine trypsin and human plasma kaUikrein. Trypsin inhibitory activity was performed in assay buffer (50 mM Hepes, pH 7.5, 0.1 M NaCl, 2.0 mM CaC12, 0.1% Triton x-100) at room temperature in a 96-well microtiter plate (Perkin Elmer) using Gly-Pro-Lys-Aminomethylcoumarin as a substrate. The amount of coumarin produced by trypsin was determined by measuring the fluorescence (ex = 370 nm, em = 432 nm) on a Perkin-Elmer LS- 50B fluorimeter equipped with a plate reader. Trypsin (23 μg in 100 μl buffer) was mixed with 20 μl of the sample to be tested and incubated for 10 minutes at 25°C The reaction was started by the addition of 50 μl of the substrate GPK- AMC (33 μM final) in assay buffer. The fluorescence intensity was measured and the % inhibition for each fraction was determined by:

% inhibition = 100 x [1- Fo/Fl]

where Fo is the fluorescence of the unknown and FI is the fluorescence of the trypsin only control. KaUikrein inhibitory activity of the fractions was similarly measured using 7.0 nM kaUikrein in assay buffer (50 M Tris, pH 8.0, 50 mM NaCl, 0.1% triton x-100) and 66.0 μM Pro-Phe-Arg-AMC as a substrate. Determination of the in vitro specificity of placental bikunin

The In vitro specificity of native human placental bikunin was determined using the materials and methods as described in the preceding examples above. Placental bikunin was quantified by active site titration against a known concentration of trypsin using GPK-AMC as a substrate to monitor the fraction of unbound trypsin.

Protein Sequencing The 1 ml fraction (C18-29 Delaria) was reduced to 300 ml in volume, on a

Speed Vac, to reduce the amount of organic solvent. The sample was then loaded onto a Hewlett-Packard miniature biphasic reaction column, and washed with 1 ml of 2% trifluoroacetic acid. The sample was sequenced on a Hewlett-Packard Model G1005A protein sequendng system using Edman degradation. Version 3.0 sequencing methods and all reagents were supplied by Hewlett-Packard. Sequence was confirmed for 50 cydes.

Results. Placental Bikunin was purified to apparent homogeniety by sequential kalUkrein affinity, gel-filtration, and reverse-phase chromatography (see purification table below):

Table 5

Purification table for native Placental Bikunin (1-179)

TABLE S

Step Vol (ml) OD 280 OD 280 Units³ Units/OD 280 ( ml) (U)

Placenta 1800.0 41.7 75,060 3,000,000 40.0

Supernatant kallikrein 3)7) \W 536 Ϊ6JLW 4 §δ

Affinity pH 4.0 kaUikrein ϊδ 045 4.56 TOT. 2^630

Affinity pH 1.7

5uperdex 75 15.0 0.0085 0.13 3,191 24546 ^aOne Unit is defined as that amount which inhibits 50% of trypsin activity in a standard assay.

The majority of the kalUkrein and trypsin inhibitory activity eluted from the kalUkrein affinity column in the pH 4.0 elution. Subsequent gel-filtration chromatography (Figure 5) yielded a peak of kaUikrein and trypsin inhibitory activity with a molecular weight range of 10 to 40 kDa as judged by a standard curve generated by running molecular weight standards under identical conditions. Reverse-phase C18 chromatography (Figure 6) yielded 4 peaks of inhibitory activity with the most potent eluting at approximately 30 % acetonitrile. The activity assodated with the first peak to elute from C18 (fraction 29) exhibited an amino acid sequence starting with amino add 1 of the predicted amino acid sequence of placental bikunin (ADRER...; SEQ ID NO: 1), and was identical to the predicted sequence for 50 cycles of sequencing (underlined amino adds in Figure 3). Cysteine residues within this sequence stretch were silent as expected for sequencing of oxidized protein. The cysteine residues at amino add positions 11 and 20 of mature placental bikunin were later identified from sequencing of the S-pyridylethylated protein whereupon FTH-pyridylethyl-cysteine was recovered at cydes 11 and 20.

Interestingly, the asparagine at amino add residue number 30 of the sequence (Figure 3) was sUent showing that this site is likely to be glycosylated. Fraction 29 yielded one major sequence corresponding to that of placental bikunin starting at residue #1 (27 pmol at cyde 1) plus a minor sequence (2 pmol) also derived from placental bikunin starting at residue 6 (SIHD...). This shows that the final preparation sequenced in fraction 29 is highly pure, and most likely responsible for the protease inhibitory activity associated with this fraction (Figure 6).

Accordingly, the final preparation of placental bikunin from C18 chromatography was highly pure based on a silver-stained SDS-PAGE analysis (Figure 7), where the protein migrated with an apparent Mr of 24 kDa on a 10 to 20 % acrylamide tridne gel (Novex, San Diego, CA) calibrated with the foUowing molecular weight markers: insulin (2.9 kDa); bovine trypsin inhibitor (5.8 kDa); lysozyme (14.7 kDa); β-lactaglobulin (18.4 kDa); carbonic anhydrase (29 kDa); and ovalbumin (43 kDa). The above size of placental bikunin on SDS- PAGE is consistent with that predicted from the fuU length coding sequence (Figure 4F). As expected based on the N-terminal sequencing results described above, the purified protein reacted with an antibody elicited to placental bikunin (7-64) to yield a band with the same Mr (Figure 12A) as observed for the purified preparation detected on gels by silver stain (Figure 7). However, when the same preparation was reacted with an antibody elicited to synthetic placental bikunin (102-159), a band corresponding to the full length protein was not observed. Rather, a fragment that co-migrated with synthetic bikunin (102-159) of approximately 6 kDa was observed. The simplest interpretation of these results is that the purified preparation had undergone degradation subsequent to purification to yield an N-terminal fragment comprising the N-terminal domain and a C-terminal fragment comprising the C- terminal domain. Assuming that the fragment reactive against antiserum to placental bikunin (7-64) is devoid of the C-terminal end of the full length protein, the size (24 kDa) would suggest a high state of glycosylation.

Table 6. below shows the potency of in vitro inhibition of various serine proteases by placental bikunin. Data are compared with that obtained with aprotinin (Trasylol®).

Table 6

Ki values for the inhibition of various proteases by placental bikunin

The results show that placental bikunin isolated from a natural source (human placenta) is a potent inhibitor of trypsin-like serine proteases.

Example 8

Expression pattern of placental bikunin amongst different human organs and tissues A multiple tissue northern was purchased from Clontech which contained 2 μg of polyA+ RNA from human heart, brain, placenta, lung, Uver, skeletal musde, kidney, and pancreas. Two different cDNA probes were used: 1) a gel purified cDNA encoding placental bikunin (102-159); 2) the 780 base pair PCR-derived cDNA (Figure 4E) liberated from a TA done by digestion with EcoRI and gel purified. Each probe was labeled using ^V-dCTP and a random priming labeUng kit from Boehringer Mannheim Biochemicals (Indiana), then used to hybridize to the multiple tissue northern according to the manufacturers specifications. Autoradiographs were generated using Biomax film with an 18 hr exposure time, and developed using a Umax Scanner and scanned using Adobe Photoshop.

Results. The pattern of tissue expression observed using a placental bikunin (102-159) probe (Figure 11 A) or a larger probe containing both Kunitz domains of placental bikunin (Figure 11B) was essentiaUy the same as might be expected. The placental bikunin mRNA was most abundant in pancreas and placenta. Significant levels were also observed in lung, brain and kidney, whUe lower levels were observed in heart and liver, and the mRNA was undetectable in skeletal muscle. The transcript size was 1.95 kilobases in all cases, in close agreement with the predicted size of placental bikunin deduced both from EST overlay and cloning of full length cDNA described in preceding sections.

The broad tissue distribution of the mRNA shows that placental bikunin is broadly expressed. Since the protein also contains a leader sequence it would have ample exposure to the human immune system, requiring that it become recognized as a self protein. Additional evidence for a broad tissue distribution of placental bikunin mRNA expression was derived from the fact that some of the EST entries with homology to placental bikunin (Figure 4B) were derived from human adult and infant brain, and human retina, breast, ovary, olfactory epitheUum, and placenta. It is concluded therefore that administration of the native human protein to human patients would be unUkely to elicit an immune response.

Interestingly, the expression pattern of placental bikunin is somewhat reminiscent of that for bovine aprotinin which is found in high levels in bovine lung and pancreas. To further eluddate the expression pattern of placental bikunin, RT-PCR of total RNA from the foUowing human cells was determined: un-stimulated human umbiUcal vein endotheUal cells (HUVECs), HK-2 (line derived from kidney proximal tubule), TF-1 (erythroleukemia Une) and phorbolester (PMA)-stimulated human peripheral blood leukocytes. The probes used:

CACCTGATCGCGAGACCCC (sense; SEQ ID NO: 59); CTGGCGGAAGCAGCGGAGC ATGC (antisense; SEQ ID NO: 60),

were designed to ampUfy a 600 b.p placental bikunin encoding cDNA fragment. Comparisons were normaUzed by inclusion of actin primers to amplify an 800 b.p. actin fragment. Whereas the 800 b.p fragment identified on agarose gels with ethidium bromide was of equal intensity in all lanes, the 600 b.p. placental bikunin fragment was absent from the HUVECs but present in significant amounts in each of the other cell lines. We condude that placental bikunin is not expressed in at least some endotheUal cells but is expressed in some leukocyte populations. Example 9

Purification and properties of Placental Bikunin (1-170) highly purified from a

Baculovirus / Sf9 expression system A large fragment of Placental bikunin containing both Kunitz domains

(Placental Bikunin 1-170) was expressed in Sf9 ceUs as follows. Placental bikunin cDNA obtained by PCR (Figure 4E) and contained within a TA vector (see previous Examples) was liberated by digestion with HindHI and Xbal yielding a fragment flanked by a 5' Xbal site and 3' HindHI site. This fragment was gel purified and then cloned into the M13mpl9 vector (New England Biolabs, Beverly, MA). In vitro' mutagenesis (Kunkel T.A., (1985) Proc. Natl. Acad. Sd. USA, 82: 488-492) was used to generate a Pstl site 3' to the Xbal site at the 5' end, but 5' to the sequence encoding the ATG start site, natural placental bikunin signal peptide and mature placental bikunin coding sequence. The oligonucleotide used for the mutagenesis had the sequence:

5' CGC GTC TCG GCT GAC CTG GCC CTG CAG ATG GCG C AC GTG TGC GGG 3' (SEQ ID NO: 61)

A stop codon (TAG) and Bgiπ / Xmal site was simUarly engineered at the 3' end of the cDNA using the oUgonucleotide:

5' CTG CCC CTT GGC TCA AAG TAG GAA GAT CTT CCC CCC GGG GGG GTG GTT CTG GCG GGG CTG 3' (SEQ ID NO: 62).

The stop codon was in frame with the sequence encoding placental bikunin and caused termination immediately following the Lysine at amino add residue 170, thus encoding a truncated placental bikunin fragment devoid of the putative transmembrane domain. The product from digestion with Pstl and Bgiπ was isolated and cloned into the BacPacδ vector for expression of Placental bikunin fragment (1-170) which contains both Kunitz domains but which is truncated immediately N-terminal to the putative transmembrane segment.

The expression of Bikunin by Sf-9 insect ceUs was optimal at a multiplicity of infection of 1 to 1 when the medium was harvested at 72 h post infection. After harvesting, the baculovirus ceU culture supernatant (2L) was adjusted to pH 8.0 by the addition of Tris-HCl. Bikunin was purified by chromatography using a 5 ml bovine pancreatic kaUikrein affinity column as previously described in Example 7 for the purification of native placental bikunin from placenta. Eluted material was adjusted to pH 2.5 with TFA and subjected to chromatography on a C18 reverse - phase column (1.0 x 25 cm) equUibrated in 10% acetonitrile in 0.1% TFA at a flow rate of 1 ml/min. The bikunin was eluted with a linear gradient of 10 to 80% acetonitrile in 0.1% TFA over 40 min. Active fractions were pooled, lyophUized, redissolved in 50 mM Hepes (pH 7.5), 0.1 M NaCl, 2 mM CaC12, and 0.1% triton x-100, and stored at -20°C until needed. The concentration of recombinant bikunin was determined by amino acid analysis.

Results. Recombinant bikunin was purified from baculovirus ceU culture supernatant using a 2-step purification protocol as shown below, to yield an active trypsin inhibitor (Table 8 below).

Table 8

Purification of recombinant bikunin from transformed culture supernatant

TABLE 8

Chromatography of the crude material over an immobUized bovine pancreatic kaUikrein affinity column selectively isolated 0.013 % of the protein and 0.67 % of the trypsin inhibitory activity present. The majority of the trypsin inhibitory activity present in the starting supernatant did not bind to the immobUized kaUikrein and is not related to bikunin (results not shown). Subsequent chromatography using C18 reverse-phase yielded a further purification of 5-fold, with a recovery of 0.2%. The final preparation was highly pure by SDS-PAGE (Figure 13), exhibiting an Mr of 21.3 kDa, and reacted on immunoblots to rabbit anti-placental bikunin 102-159 (not shown). N-terminal sequencing (26 cycles) yielded the expected sequence for mature placental bikunin (Figure 4F) starting at residue +1(ADRER....) , showing that the signal peptide was correctly processed in Sf9 cells.

Purified placental bikunin from Sf9 cells (100 pmol) was pyridylethyl- alkylated, CNBr digested and then sequenced without resolution of the resulting fragments. Sequencing for 20 cycles yielded the following N-terminii: Sequence Amount Placental bikunin residue #

LRCFrQQENPP-PLG 21 pmol 154 - 168 (SEQIDNO:63)

ADRERSIHDFCLVSKWGRC 20 pmol 1 - 20 (SEQIDNO:64) FNYeEYCTANAVTGPCRASF 16 pmol 100 - 119 (SEQIDNO:65)

Pr--Y-V-dGS-Q-F-Y-G 6 pmol 25 - 43 (SEQIDNO:66)

Thus N-terminii corresponding to each of the expected four fragments were recovered. This confirms that the Sf9 expressed protein contained the entire ectodomain sequence of placental bikunin (1-170). N-terminal sequencing (50 cycles) of an additional sample of undigested Placental Bikunin (1-170) resulted in an amino add sequence which at cycle 30 was devoid of any PTH-amino add (PTH-asparagine was expected). A simUar result was obtained upon sequendng of the natural protein from human placenta (Example 7) and is consistent with this residue being glycosylated as predicted from the amino add sequence surrounding this asparagine residue. Furthermore, the cysteine residues within this region were also sUent consistent with their partidpation in disulfide bonding.

Example 10 Inhibition specificity of purified placental bikunin derived from Sf9 cells.

The in vitro specificity of recombinant bikunin was determined using the materials and methods as described in Examples 3, 4 and 7. In addition, the inhibition of human tissue kalUkrein by bikunin was measured by the incubation of 0.35 nM human tissue kaUikrein recombinant bikunin in buffer containing 50 mM Tris (pH 9.0), 50 mM NaCl, and 0.01% triton x-100. After 5 min. at 37°C, 5 μl of 2 mM PFR-AMC was added and the change in fluorescence monitored.

Inhibition of tissue plasminogen activator (tPA) was also determined as follows: tPA (single chain form from human melanoma cell culture from Sigma Chemical Co, St Louis, MO) was pre-incubated with inhibitor for 2 hr at room temperature in 20 mM Tris buffer pH 7.2 containing 150 mM NaCl, and 0.02% sodium azide. Reactions were subsequently initiated by transfer to a reaction system comprising the foUowing initial component concentrations: tPA (7.5 nM), inhibitor 0 to 6.6 μM, DDe-Lpro-Larg-pNitroaniUne (ImM) in 28 mM Tris buffer pH 8.5 containing 0.004 % (v/v) triton x-100 and 0.005% (v/v) sodium azide. Formation of p-NitroaniUne was determined from the A405nm measured foUowing incubation at 37 C for 2hr.

The table below show the efficacy of recombinant bikunin as an inhibitor oi various serine proteases in vitro. Data is shown compared against data obtained for screening inhibition using either recombinant bikunin, or aprotinin.

Table 9

Comparisons of Ki values for the inhibition of various proteases by recombinant placental bikunin (1-170) or aprotinin

The results show that recombinant bikunin can be expressed in insect cells to yield an active protease inhibitor that is effective against at least five different serine protease inhibitors. Recombinant bikunin was more potent than aprotinin against human plasma kalUkrein, trypsin and plasmin. Surprisingly, the recombinant bikunin was more potent that the syntheticaUy derived bikunin fragments (7-64) and (102-159) against aU enzymes tested. These data show that recombinant bikunin is more effective than aprotinin, using in vitro assays, and that one would expect better in vivo potency.

Besides measuring the potencies against spedfic proteases, the capacity of placental bikunin (1-170) to prolong the activated partial thromboplastin time (APTT) was evaluated and compared with the activity associated with aprotinin. Inhibitor was diluted in 20 mM Tris buffer pH 7.2 containing 150 mM NaCl and 0.02% sodium azide and added (0.1 ml) to a cuvette contained within an MLA Electra^ 800 Automatic Coagulation Timer coagulometer (Medical Laboratory Automation, Inc., PleasantvUle, N.Y.). The instrument was set to APTT mode with a 300 sec. activation time and the duplicate mode. Following addition of 0.1 ml of plasma (Specialty Assayed Reference Plasma lot 1-6-5185, Helena Laboratories, Beaumont, TX), the APTT reagent (Automated APTT-lot 102345, from Organon Teknika Corp., Durhan, NC) and 25 mM CaC12 were automatically dispensed to initiate clotting, and the clotting time was monitored automatically. The results (Figure 14) showed that a doubling of the clotting time required approximately 2 μM final aprotinin, but only 0.3 μM Sf9 derived placental bikunin. These data show that placental bikunin is an effective anticoagulant, and usefuU as a medicament for diseases involving pathologic activation of the intrinsic pathway of coagulation.

Although certain embodiments of the invention have been described in detaU for the purpose of Ulustration, it wiU be readily apparent to those sldlled in the art that the methods and formulations described herein may be modified without departing from the spirit and scope of the invention. Accordingly, the invention is not limited except as by the appended daims.

Claims

WE CLAIM:

1. A substantially purified protein, having serine protease inhibitory activity, selected from the group of proteins consisting of materials each of which comprises one of the following amino acid sequences, the amino acids of said sequences being numbered in accordance with the amino add sequence of native human placental bikunin shown in figure 4F in which the N-terminal residue generated by removal of signal peptide is designated as residue 1:

2. A protein as in claim 1, wherein said protein is glycosylated, or contains at least one intra-chain cysteine-cysteine disulfide bond, or is both glycosylated and contains at least one intra-chain cysteine-cysteine disulfide bond.

3. A pharmaceutical composition for inhibiting serine protease activity, comprising a protein of claim 1 or claim 2 plus a pharmaceuticall y acceptable carrier.

4. An isolated nucleic acid sequence which encodes for a protein of claim 1.

5. A self-replicating protein expression vector containing a nucleic acid sequence which encodes for and is capable of expressing a protein of claim or claim 2.

6. A method for inhibiting serine protease activity comprising contacting serine protease with an effective amount of at least one protein of claim 1 or claim 2.

7. A method for treating a condition of brain edema, spinal cord edema, multiple sderosis, ischemia, perioperative blood loss, sepsis, septic shock, fibrosis, disease associated with pathologic blood coagulation or clotting, polytrauma, stroke, cerebral or subarachnoid hemorrhage, inflamation of the brain, inflamation of the spinal cord, cerebral infection, cerebral

granulomatosis, spinal infection, spinal granulomatosis, open heart surgery, gastric cancer, cervical cancer, or prevention of metastasis comprising

administering to a subject having such a condition an effective amount of the protein of claim 1 or claim 2.

8. The method of Claim 7 wherein said condition is brain edema, spinal cord edema, multiple sclerosis, ischemia, perioperative blood loss, sepsis, septic shock, fibrosis, disease assodated with pathologic blood coagulation or clotting, stroke, cerebral or subarachnoid hemorrhage, inflamation of the brain, inflamation of the spinal cord, cerebral infection, cerebral granulomatosis, spinal infection, spinal granulomatosis, or open heart surgery.

9. The method of Claim 7 wherein said condition is gastric cancer, cervical cancer, or prevention of metastasis.

10. A method for the preparation of a medicament for the treatment of brain edema, spinal cord edema, multiple sclerosis, ischemia, perioperative blood loss, sepsis, septic shock, fibrosis, disease associated with pathologic blood coagulation or dotting, stroke, cerebral or subarachnoid hemorrhage, inflamation of the brain, inflamation of the spinal cord, cerebral infection, cerebral granulomatosis, spinal infection, spinal granulomatosis, open heart surgery, gastric cancer, cervical cancer, or prevention of metastasis.

11. A method for preparing a protien of claim 1 or claim 2 using

recombinant DNA technology.