NO326968B1 - PAF-AH polypeptide fragment, isolated polynucleotide, DNA vector, host cell, process for producing a PAF-AH polypeptide fragment, variant or variant fragment of plasma PAF-AH, PAF-AH polypeptide fragment, variant or variant fragment, pharmaceutical composition, use of the PAF-AH fragment, variant or variant fragment for the preparation of a medicament for treating a mammal. - Google Patents

Description

Field of the invention

The present invention relates to PAF-AH polypeptide fragment, isolated polynucleotide, DNA vector, host cell, method for producing a PAF-AH polypeptide fragment, a variant or a variant fragment of plasma PAF-AH, PAF-AH polypeptide fragment, variant or variant fragment , pharmaceutical preparation, use of the PAF-AH fragment, variant or variant fragment for the manufacture of a medicament for the treatment of a mammal.

Background

Platelet-activating factor (PAF) is a biologically active phospholipid that is synthesized by various cell types. In vivo and at normal concentrations of 10"<10> to 10"<9> M, PAF activates target cells, such as platelets and neutrophils, by binding to specific G protein-coupled cell surface receptors [Venable et al., J. Lipid Res. , 34:691-701 (1993)]. PAF has the structure 1-O-alkyl-2-acetyl-sn-glycero-3-phosphocholine. For optimal biological activity, the sn-1 position of the PAF glycerol backbone must be in an ether bond with a fatty alcohol, and the sn-3 position must have a phosphocholine head group.

PAF functions in normal physiological processes (eg, inflammation, hemostasis, and parturition) and is implicated in pathological inflammatory responses (eg, asthma, anaphylaxis, septic shock, and arthritis) [Venable et al., supra, and Lindsberg et al., Ann. Neurol., 30:117-129

(1991)]. The likelihood of PAF involvement in pathological responses has prompted attempts to modulate the activity of PAF, and the main focus of these attempts has been the development of antagonists of PAF activity that interfere with binding of PAF to cell surface receptors. See e.g. Heuer et al., Clin. Exp. Allergy, 22:980-983 (1992).

Synthesis and secretion of PAF, as well as its degradation and secretion, appear to be tightly controlled. To the extent that pathological, inflammatory effects of PAF result from a failure of PAF regulatory mechanisms resulting in excessive production, insufficient production, or lack of degradation, an alternative method of modulating the activity of PAF would involve mimicking or enhancing the natural process of resolution of inflammation. Macrophages [Stafforini et al., J. Biol. Chem., 255(17):9682-9687 (1990)], hepatocytes and the human hepatoma cell line HepG2 [Satoh et al., J. Clin. Invest., 87:476-481 (1991), and Tarbet et al., J. Biol. Chem., 266(25): 16667-16673 (1991)] has been reported to release an enzymatic activity, PAF-acetyl hydrolase (PAF-AH), which inactivates PAF. In addition to inactivating PAF, PAF-AH also inactivates oxidatively fragmented phospholipids, such as products of the arachidonic acid cascade that mediate inflammation. See Stremler et al., J. Biol. Chem., 266(17):11095-11103 (1991). The inactivation of PAF by PAF-AH takes place primarily by hydrolysis of the PAF sn-2-acetyl group, and PAF-AH metabolizes oxidatively fragmented phospholipids by removing sn-2-acyl groups. Two types of PAF-AH have been identified: cytoplasmic forms found in many different cell types and tissues, such as endothelial cells and erythrocytes, and an extracellular form found in plasma and serum. Plasma PAF-AH does not hydrolyze intact phospholipids, with the exception of PAF, and this substrate specificity allows the enzyme to circulate in vivo in a fully active state without adverse effects. Plasma PAF-AH appears to account for all PAF degradation in human blood ex vivo [Stafforini et al., J. Biol. Chem., 262(9):4223-4230 (1987)].

Although the cytoplasmic and plasma forms of PAF-AH appear to have identical substrate specificity, plasma PAF-AH has biochemical characteristics that distinguish it from cytoplasmic PAF-AH and from other characterized lipases. Specifically, plasma PAF-AH is associated with lipoprotein particles, is inhibited by diisopropylfluorophosphate, is unaffected by calcium ions, is relatively insensitive to proteolysis, and has an apparent molecular weight of 43,000 daltons. See Stafforini et al., (1987), supra. The same article by Stafforini et al. describes a procedure for the partial purification of PAF-AH from human plasma, and the amino acid composition of the plasma material obtained using the procedure. Cytoplasmic PAF-AH has been purified from erythrocytes, as reported by Stafforini et al., J. Biol. Chem., 268(6):3857-3865 (1993), and 10 amino-terminal residues of cytoplasmic PAF-AH are also described in the article. Hattori et al., J. Biol. Chem., 268(25):18748-18753 (1993) describes purification of cytoplasmic PAF-AH from bovine brain. After submission of the parent application of the present patent application, the nucleotide sequence of bovine brain cytoplasmic PAF-AH was published by Hattori et al., J. Biol. Chem., 259(237):23150-23155 (1994). On January 5, 1995, 3 months after the filing date of the parent application of the present patent application, a nucleotide sequence for a lipoprotein-associated phospholipase A2 (Lp-PLA2) was published in Smithkline Beecham PLC Patent Cooperation Treaty (PCT), International Publication No. WO 95/00649 . The nucleotide sequence of Lp-PLA2 differs in one position compared to the nucleotide sequence of PAF-AH according to the present invention. The nucleotide difference (corresponding to position 1297 in SEQ ID NO: 7) results in an amino acid difference between the enzymes encoded by the polynucleotides. The amino acid at position 379 in SEQ ID NO. NO. 8 is valine, while the amino acid in the corresponding position in Lp-PLA2 is alanine. The nucleotide sequence of PAF-AH according to the present invention also includes 124 bases at the 5' end and 20 bases at the 3' end which are not present in the Lp-PLA2 sequence. 3 months later, on 10 April 1995, an Lp-PLA2 sequence was deposited in GenBank under accession no. U24577 which differs in 11 positions compared to the nucleotide sequence of PAF-AH according to the present invention. The nucleotide differences (corresponding to positions 79, 81, 84, 85, 86, 121, 122, 904, 905, 911, 983 and 1327 in SEQ ID NO: 7) result in four amino acid differences between the enzymes encoded by the polynucleotides. The amino acids in positions 249, 250, 274 and 389 in SEQ ID NO. NO. 8 are respectively lysine, aspartic acid, phenylalanine and leucine, while the respective amino acids in the corresponding positions in the GenBank sequence are isoleucine, arginine, leucine and serine.

The recombinant production of PAF-AH would enable the use of exogenous PAF-AH to mimic or enhance normal processes of resolution of inflammation in vivo. Administration of PAF-AH would provide a physiological advantage over administration of PAF receptor antagonists, because PAF-AH is a product normally found in plasma. Furthermore, because PAF receptor antagonists structurally related to PAF inhibit native PAF-AH activity, the desirable metabolism of PAF and of oxidatively fragmented phospholipids is prevented. Inhibition of PAF-AH activity with PAF receptor antagonists thus counteracts the competitive blocking of the PAF receptor by the antagonists. See Stremler et al. supra. In places with e.g. moreover, acute inflammation results in the release of oxidants in inactivating the native PAF-AH enzyme, which in turn results in elevated local levels of PAF and PAF-like compounds that would compete with any exogenously administered PAF receptor antagonist for binding to PAF - the receptor. In contrast, treatment with recombinant PAF-AH would enhance endogenous PAF-AH activity and compensate for any inactivated endogenous enzyme.

WO 95 09921 includes PAF-AH. WO 95 00649 relates to lipoprotein associated phospholipase A2, LP-PLA2 in purified form, and an isolated nucleic acid molecule encoding this enzyme.

There is thus a need in the art to identify and isolate polynucleotide sequences that code for human plasma PAF-AH, and to develop materials and methods applicable for the recombinant production of PAF-AH, and to produce reagents for the detection of PAF-AH in plasma .

Summary of the invention

The present invention describes new purified and isolated polynucleotides (that is, DNA and RNA, both sense and antisense strands) which code for human plasma PAF-AH or enzymatically active fragments thereof. Preferred DNA sequences include genomic and cDNA sequences, as well as wholly or partially chemically synthesized DNA sequences. The DNA sequences encoding PAF-AH, which are shown in SEQ ID NO. NO. 7, and DNA sequences that hybridize to the non-coding strand thereof under standard stringency conditions, or that would hybridize were it not for the redundancy of the genetic code, are encompassed by the present invention. Also included are biological copies (that is, copies of isolated DNA sequences formed in vivo or in vitro) of DNA sequences. Autonomously replicating recombinant constructs, such as plasmid and viral DNA vectors incorporating PAF-AH sequences, and particularly vectors in which DNA encoding PAF-AH is operably linked to an endogenous or exogenous expression control DNA sequence and a transcription terminator, are also provided.

It is further described that prokaryotic or eukaryotic host cells are stably transformed with DNA sequences as described herein in a manner that allows the desired PAF-AH to be expressed therein. Host cells expressing PAF-AH products can serve many different useful purposes. Such cells provide a valuable source of immunogens for the development of antibodies specifically immunoreactive with PAF-AH. Host cells according to the invention are remarkably useful in methods for the large-scale production of PAF-AH, in which the cells are grown in a suitable culture medium and the desired polypeptide products are isolated from the cells or from the medium in which the cells are grown, e.g. by immunoaffinity purification.

A non-immunological method described herein for purifying PAF-AH from plasma includes the following steps: (a) isolation of low-density lipoprotein particles; (b) dissolving the low density lipoprotein particles in a buffer comprising 10 mM CHAPS to form a first PAF-AH enzyme solution; (c) applying this first PAF-AH enzyme solution to a DEAE anion exchange column; (d) washing the DEAE anion exchange column using an appropriate pH 7.5 buffer comprising 1 mM CHAPS; (e) elution of PAF-AH enzyme from the DEAE anion exchange column in fractions using ca. pH 7.5 buffers comprising a gradient from 0 to 0.5 M NaCl; (f) pooling fractions eluted from the DEAE anion exchange column with PAF-AH enzymatic activity; (g) adjusting the pooled active fractions from the DEAE anion exchange column to 10 mM CHAPS to form a second PAF-AH enzyme solution; (h) applying the second PAF-AH enzyme solution to a blue dye ligand affinity column; (i) elution of PAF-AH enzyme from the blue dye ligand affinity column using a buffer comprising 10 mM CHAPS and a chaotropic salt; (j) applying the eluate from the blue dye ligand affinity column to a Cu ligand affinity column; (k) elution of PAF-AH enzyme from the Cu ligand affinity column using a buffer comprising 10 mM CHAPS and imidazole; (1) performing SDS-PAGE on the eluate from the Cu ligand affinity column; and (m) insulation of the approx. 44 kDa PAF-AH enzyme from the SDS-polyacrylamide gel. The buffer in step (b) is preferably 25 mM Tris-HCl, 10 mM CHAPS, pH 7.5; the buffer in step (d) is 25 mM Tris-HCl, 1 mM CHAPS; the column in step (h) is a blue Sepharose Fast Flow column; the buffer in step (i) is 25 mM Tris-HCl, 10 mM CHAPS, 0.5 M KSCN, pH 7.5; the column in step (j) is a Cu-chelating Sepharose column; and the buffer in step (k) is 25 mM Tris-HCl, 10 mM CHAPS, 0.5 M NaCl, 50 mM imidazole, at a pH in the range 7.5-8.0.

A method for purifying enzymatically active PAF-AH from E. coli producing PAF-AH includes the steps: (a) preparing a centrifugation supernatant from lysed E. coli producing PAF-AH enzyme; (b) applying the centrifugation supernatant to a blue dye ligand affinity column; (c) elution of PAF-AH enzyme from the blue dye ligand affinity column using a buffer comprising 10 mM CHAPS and a chaotropic salt; (d) applying the eluate from the blue dye ligand affinity column to a Cu ligand affinity column; and (e) elution of PAF-AH enzyme from the Cu ligand affinity column using a buffer comprising 10 mM CHAPS and imidazole. The column in step (b) is preferably a blue Sepharose Fast Flow column; the buffer in step (c) is 25 mM Tris-HCl, 10 mM CHAPS, 0.5 M KSCN, pH 7.5; the column in step (d) is a Cu-chelating Sepharose column; and the buffer in step (e) is 25 mM Tris-HCl, 10 mM CHAPS, 0.5 M NaCl, 100 mM imidazole, pH 7.5.

Another method for purifying enzymatically active PAF-AH from E. coli producing PAF-AH includes the steps: (a) preparing a centrifugation supernatant from lysed E. coli producing PAF-AH enzyme; (b) diluting the centrifugation supernatant in a low pH buffer comprising 10 mM CHAPS; (c) applying the diluted centrifugation supernatant to a cation exchange column equilibrated to about pH 7.5; (d) elution of PAF-AH enzyme from the cation exchange column using 1 M salt; (e) increasing the pH in the eluate from the cation exchange column and adjusting the salt concentration in the eluate to approx. 0.5 M salt; (f) applying the adjusted eluate from the cation exchange column to a blue dye ligand affinity column; (g) elution of PAF-AH enzyme from the blue dye ligand affinity column using a buffer comprising approx. 2 M to 3 M salt; and (h) dialyzing the eluate from the blue dye ligand affinity column using a buffer comprising approx. 0.1% Tween. The buffer in step (b) is preferably 25 mM MES, 10 mM CHAPS, 1 mM EDTA, pH 4.9; the column in step (c) is an S-Sepharose column equilibrated in 25 mM MES, 10 mM CHAPS, 1 mM EDTA, 50 mM NaCl, pH 5.5; PAF-AH is eluted in step (d) using 1 mM NaCl; The pH of the eluate in step (e) is adjusted to pH 7.5 using 2 M Tris base; the column in step (f) is a Sepharose column; the buffer in step (g) is 25 mM Tris, 10 mM CHAPS, 3 M NaCl, 1 mM EDTA, pH 7.5; and the buffer in step (h) is 25 mM Tris, 0.5 mM NaCl, 0.1% Tween 80, pH 7.5.

Still another method for purifying enzymatically active PAF-AH from E. coli includes the steps: (a) preparing an E. coli extract yielding dissolved PAF-AH supernatant after lysis in a buffer containing CHAPS; (b) dilution of the supernatant and application to an anion exchange column equilibrated at approx. pH 8.0; (c) elution of PAF-AH enzyme from the anion exchange column; (d) applying the adjusted eluate from the anion exchange column to a blue dye ligand affinity column; (e) elution of the blue dye ligand affinity column using a buffer comprising 3.0 M salt; (f) diluting the blue dye eluate in a suitable buffer for performing hydroxylapatite chromatography; (g) performing hydroxylapatite chromatography in which washing and elution are carried out using buffers (with or without CHAPS; (h) diluting the hydroxylapatite eluate to a suitable salt concentration for cation exchange chromatography; (i) applying the diluted hydroxylapatite eluate to a cation exchange column at a pH varying between about 6.0 and 7.0; (j) elution of PAF-AH from the cation exchange column with a suitable formulation buffer; (k) performing cation exchange chromatography at low temperature; and (1) formulation of PAF-AH in liq. or frozen form in the absence of CHAPS.

In step (a) above, the lysis buffer is preferably 25 mM Tris, 100 mM NaCl, 1 mM EDTA, 20 mM CHAPS, pH 8.0; in step (b), the dilution of the supernatant for anion exchange chromatography is 3-4 times in 25 mM Tris, 1 mM EDTA, 10 mM CHAPS, pH 8.0, and the column is a Q-Sepharose column equilibrated with 25 mM Tris, 1 mM EDTA, 50 mM NaCl, 10 mM CHAPS, pH 8.0; in step (c), the anion exchange column is eluted using 25 mM Tris, 1 mM EDTA, 350 mM NaCl, 10 mM CHAPS, pH 8.0; in step (d) the eluate from step (c) is applied directly to a blue dye affinity column; in step (e), the column is eluted with 3 M NaCl, 10 mM CHAPS, 25 mM Tris, pH 8.0 buffer; in step (f), dilution of the blue color eluate for hydroxylapatite chromatography is carried out by dilution in 10 mM sodium phosphate, 100 mM NaCl, 10 mM CHAPS, pH 6.2; in step (g), hydroxylapatite chromatography is performed using a hydroxylapatite column equilibrated with 10 mM sodium phosphate, 100 mM NaCl, 10 mM CHAPS, and the elution is performed using 50 mM sodium phosphate, 100 mM NaCl (with or without 10 mM CHAPS), pH 7 .5; in step (h), dilution of the hydroxylapatite elute for cation exchange chromatography is carried out by dilution in a buffer at a pH varying from approx. 6.0 to 7.0, and comprising sodium phosphate (with or without CHAPS); in step (i) an S-Sepharose column is equilibrated with 50 mM sodium phosphate (with or without 10 mM CHAPS), pH 6.8; in step (j), elution is performed with an appropriate formulation buffer, such as potassium phosphate, 50 mM, 12.5 mM aspartic acid, 125 mM NaCl, pH 7.5, containing 0.01% Tween 80; and in step (k) cation exchange chromatography is performed at 2-8 °C. Examples of suitable formulation buffers for use in step (1) that stabilize PAF-AH include 50 mM potassium phosphate, 12.5 mM aspartic acid, 125 mM NaCl, pH 7.4

(with and without the addition of Tween 80 and/or Pluronic F68)

or 25 mM potassium phosphate buffer containing (at least) 125 mM NaCl, 25 mM arginine and 0.1% Tween 80 (with or without Pluronic F68 at approx. 0.1 and 0.5%).

A further method for purifying enzymatically active rPAF-AH products from E. coli includes the steps: (a) preparing an E. coli extract that yields dissolved rPAF-AH product supernatant after lysis in a buffer containing Triton X-100; (b) diluting the supernatant and applying to an immobilized metal affinity yielder column equilibrated at ca. pH 8.0; (c) elution of rPAF-AH product from the immobilized metal affinity extractor column with a buffer comprising imidazole; (d) adjusting the salt concentration and applying the eluate from the immobilized metal affinity exchanger column to a hydrophobic interaction column (HIC No. 1); (e) elution of HIC No. 1 by decreasing the salt concentration and/or increasing the detergent concentration; (f) titrating the HIC No. 1 eluate to a pH of about 6.4; (g) applying the adjusted HIC No. 1 eluate to a cation exchange column (CEX No. 1) equilibrated at ca. pH 6.4; (h) eluting CEX No. 1 with concentrated sodium chloride; (i) adjusting the CEX No. 1 eluate with sodium chloride to a concentration of about 2.0M; (j) applying the adjusted CEX No. 1 eluate to a hydrophobic interaction column (HIC No. 2) equilibrated at approx. pH 8.0 and approx. 2.0 M sodium chloride; (k) elution of HIC No. 2 by decreasing the salt concentration and/or increasing the detergent concentration; (1) dilution of the HIC No. 2 eluate and adjustment to a pH of approx. 6.0; (m) applying the adjusted HIC No. 2 eluate to a cation exchange column (CEX No. 2) equilibrated at approx. pH 6.0; (n) elution of the rPAF-AH product from CEX No. 2 with a suitable formulation buffer.

In step (a) above, the lysis buffer is preferably 90 mM Tris, 0.125% Triton X-100, 0.6 M NaCl, pH 8.0, and the lysis is performed in a high-pressure homogenizer; in step (b), the supernatant is diluted in equilibration buffer (20 mM Tris, 0.5 M NaCl, 0.1% Triton X-100, pH 8.0), a zinc chelate column (Chelating Sepharose Fast Flow, Pharmacia, Uppsala, Sweden ) is loaded, equilibrated with equilibration buffer, the diluted supernatant is applied, and the column is washed with 20 mM Tris, 0.5 M NaCl, 4 M urea, 0.1% Triton X-100, pH 8.0, followed by washing with 20 mM Tris, 0.5 M NaCl, 0.02% Triton X-100, pH 8.0; in step (c) elution is performed with 20 mM Tris, 50 mM imidazole, 0.02% Triton X-100, pH 8.0; in step (d), the eluate is adjusted to 1 mM EDTA and 2 M NaCl, a phenyl-Sepharose 6 Fast Flow (Pharmacia) is equilibrated with equilibration buffer (2.0 M NaCl, 25 mM Tris, 0.02% Triton X-100, pH 8.0), the adjusted eluate from step (c) is applied at room temperature, the column is washed with equilibration buffer and washed with 25 mM NaPO 4 , 0.02% Triton X-100, pH 6.5, at a flow rate of 30 cm/h ; in step (e), elution is performed with 25 mM NaPO 4 , 3% Triton X-100, pH 6.5; in step (g), a Macro-Prep High S column (Bio-Rad Labs, Richmond, CA) is equilibrated with equilibration buffer (20 mM NaPO 4 , 0.02% Triton X-100, pH 6.4), the adjusted eluate from step (f) is applied, the column is washed with equilibration buffer and washed with 25 mM Tris, 0.02% Triton X-100, pH 8.0; in step (h) elution is performed with 25 mM Tris, 0.02% Triton X-100, 1.3 M NaCl, pH 8.0; in step (j), a Bakerbond Wide Pore Hi-Propyl C3 (Baker, Phillipsburg, NJ) is equilibrated with equilibration buffer (2.0 M NaCl, 25 mM Tris, 0.02% Triton X-100, pH 8.0), the adjusted eluates from step (i) are applied at room temperature, the column is washed with equilibration buffer and washed with 25 mM Tris, 0.02% Triton X-100, pH 8.0, at 30 cm/h; in step (k), elution is performed with 10 mM Tris, 3.0% Triton X-100, pH 8.0; in step (1) the dilution is carried out in equilibration buffer (20 mM succinate, 0.1% Pluronic F68, pH 6.0); in step (m) an SP Sepharose Fast Flow column (Pharmacia) is equilibrated with the equilibration buffer from step (1), the eluate from step (1) is applied, the column is washed with equilibration buffer; and in step (n) elution is performed with 50 mM NaPO 4 , 0.7 M NaCl, 0.1% Pluronic F68, 0.02% Tween 80, pH 7.5.

PAF-AH products can be obtained as isolates from natural cell sources, or they can be chemically synthesized, but they are preferably produced by recombinant procedures involving prokaryotic or eukaryotic host cells according to the invention. PAF-AH products where part or all of the amino acid sequence is shown in SEQ ID NO. NO. 8, is included. The invention particularly relates to fragments lacking up to the first 12 N-terminal amino acids of the mature human PAF-AH amino acid sequence shown in SEQ ID NO. NO. 8, especially those having Met46, Ala47 or Ala48 in SEQ ID NO. NO. 8 as the initial N-terminal amino acid. Fragments lacking up to 30 C-terminal amino acids of the amino acid sequence according to SEQ ID NO. NO. 8, are also included, especially those having Ile429 and Leu431 as the C-terminal residue. Also included are variants of PAF-AH that have an amino acid substitution in the sequence according to SEQ ID NO: NO. 8 selected from the group consisting of S 108 A, S 273 A, D 286 A, D 286 N, D 296 A, D 304 A, D 338 A, H 351 A, H 395 A, H 399 A, C 67 S, C 229 S, C 291 S, C 334 S, C 407 S, D 286 A, D 286 N and D 304 A. As stated above, the present invention provides polynucleotides (including DNA) encoding such fragments or variant fragments, as well as as methods for the recombinant production of such fragments or variants, by culturing host cells comprising such DNA. Preferred PAF-AH products include the prokaryotic polypeptide expression products from DNA encoding amino acid residues Met46-Asn441 in SEQ ID NO. NO. 8, designated rPH.2, and the prokaryotic polypeptide expression products from DNA encoding amino acid residues Met46-Ile429 in SEQ ID NO. NO. 8, designated rPH.9. Both the rPH.2 and rPH.9 products show less amino-terminal heterogeneity than e.g. the corresponding prokaryotic expression products from DNA encoding the complete mature sequence of PAF-AH with a preceding translation initiation codon. The rPH.9 product also exhibits higher carboxy-terminal homogeneity (consistency). The use of mammalian host cells is expected to provide such post-translational modifications (e.g. myristolation, glycosylation, truncation, lipidation and tyrosine, serine or threonine phosphorylation) as may be necessary to give recombinant expression products according to the invention optimal biological activity. PAF-AH products according to the invention can be unabridged polypeptides, fragments or variants. Variants may include PAF-AH analogs in which one or more of the specified (ie, encoded) amino acids have been split or replaced, or in which one or more unspecified amino acids have been added: (1) without loss of one or more of the enzymatic activities or immunological characteristics specific to PAF-AH; or (2) with specific loss of a particular biological activity of PAF-AH. Proteins or other molecules that bind to PAF-AH can be used to modulate its activity.

Antibodies (eg, monoclonal and polyclonal antibodies, single-chain antibodies, chimeric antibodies, CDR-grafted antibodies, and the like) and other binding proteins specific for PAF-AH are also described. Particularly illustrative binding proteins are the monoclonal antibodies produced by hybridomas 90G11D and 90F2D, which were deposited with the American Type Culture Collection (ATCC), 12301 Parklawn Drive, Rockville, MD 20852, on September 30

1994, and which were respectively given accession numbers HB 11724 and HB 11725. Illustrative binding proteins are also the monoclonal antibody produced by hybridoma 143A, which was deposited with ATCC on June 1, 1995 and given accession no. HB 11900. Proteins or other molecules (eg, lipids or small molecules) that specifically bind to PAF-AH can be identified using PAF-AH isolated from plasma, recombinant PAF-AH, PAF-AH variants, or cells that expresses such products. Binding proteins are in turn useful in preparations for immunization, as well as for purification of PAF-AH, and are useful for detection or quantification of PAF-AH in fluid samples and tissue samples using known immunological procedures. Antiidiotypic antibodies specific for PAF-AH-specific antibodies are also included.

The scientific value of the information provided by the descriptions of DNA and amino acid sequences is self-evident. As one series of examples, knowledge of the sequence of a PAF-AH cDNA enables isolation by DNA/DNA hybridization of genomic DNA sequences encoding PAF-AH, and specification of PAF-AH expression control regulatory sequences, such as as promoters, operators and the like. DNA/DNA hybridization procedures performed with DNA-17 (1994)], ulcerative colitis (Denizot et al., supra], ischemic stroke [Satoh et al., Stroke, 23:1090-1092 (1992)], ischemic brain injury [Lindsberg et al., Stroke, 21:1452-1457

(1990), and Lindsberg et al. (1991), supra], systemic lupus erythematosus [Matsuzaki et al., Clinica Chimica Acta, 210:139-144 (1992)], acute pancreatitis [Kald et al., Pancreas, 8(4):440-442 (1993 )], septicemia (Kald et al., supra), acute glomerulonephritis after streptococcal infection [Mezzano et al., J. Am. Soc. Nephrol., 4:235-242 (1993)], pulmonary edema after IL-2 therapy [Rabinovici et al., J. Clin. Invest., 89:1669-1673 (1992)], allergic inflammation [Watanabe et al., Br. J. Pharmacol., 111:123-130 (1994)], ischemic renal failure [Grino et al., Annals of Internal Medicine, 121(5):345-347 (1994); premature labor [Hoffman et al.. Am. J. Obstet. Gynecol., 162(2):525-528 (1990), and Maki et al., Proe. Nati. Acad. Sei., USA, 85:728-732

(1988)]; respiratory distress syndrome in adults [Rabinovici et al., J. Appl. Physiol., 74(4):1791-1802 (1993); Matsumoto et al., Clin. Exp. Pharmacol. Physiol., 19:509-515 (1992); and Rodriguez-Roisin et al., J. Clin. Invest., 93:188-194

(1994)]. The present invention also relates to the use of PAF-AH preparations for the treatment of human immunodeficiency virus (HIV) infection of the central nervous system. "Treatment" as used herein includes both prophylactic and therapeutic treatment.

Animal models for many of the preceding pathological conditions are described in the art. A mouse model for asthma and rhinitis is e.g. described in Example 16 herein; a rabbit model of arthritis is described by Zarco et al., supra; rat models of ischemic intestinal necrosis/necrotizing enterocolitis are described by Furukawa et al., Ped. Res., 34(2):237-241 (1993), and Caplan et al., supra; a rabbit model for stroke is described by Lindsberg et al.

(1990), supra; a mouse model for lupus is described by Matsuzaki et al., supra; a rat model of acute pancreatitis is described by Kald et al., supra; a rat model of pulmonary edema following IL-2 therapy is described by Rabinovici et al., supra; a rat model of allergic inflammation is described by Watanabe et al., supra; a canine model of renal allotransplantation is described by Watson et al., Transplantation, 56(4):1047-1049 (1993); and rat and guinea pig models of adult respiratory distress syndrome are respectively described by Rabinovici et al., supra, and Lellouch-Tubiana, Am. Fox. Breathe. Dis., 137:948-954 (1988).

Specifically disclosed are PAF-AH preparations for use in methods of treating a mammal susceptible to or suffering from PAF-mediated pathological conditions, comprising administering to the mammal PAF-AH in an amount sufficient to supplement endogenous PAF -AH activity and inactivate pathological amounts of PAF in the mammal.

Therapeutic/pharmaceutical compositions as described herein include PAF-AH products and a physiologically acceptable diluent or carrier, and may also include other agents with anti-inflammatory effects. Indicated dosage amounts will be sufficient to supplement endogenous PAF-AH activity and inactivate pathological amounts of PAF. For general dosage considerations, see Remington's Pharmaceutical Sciences, 18th ed., Mack Publishing Co., Easton, PA (1990). Doses will vary between approx. 0.1 and 1000 p,g PAF-AH/kg body weight. Therapeutic preparations according to the present invention can be administered in different ways, depending on the pathological condition to be treated. The administration can e.g. is carried out intra-venously, subcutaneously, orally, by means of suppositories and/or pulmonary.

For pathological conditions in the lungs, pulmonary administration of PAF-AH is particularly indicated. For pulmonary administration, many different delivery devices can be used, including e.g. nebulizers, metered dose inhalers and powder inhalers, which are standard in the art. Delivery of various proteins to the lungs and circulatory system by inhalation of aerosol formulations is described by Adjei et al., Pharm. Res., 7(6):565-569 (1990) (leuprolide acetate); Braquet et al., J. Cardio. Pharm., 13 (suppl. 5): pp. 143-146 (1989) (endothelin-1); Hubbard et al., Annals of Internal Medicine, IJI(3), 206-212

(1989) (α-antitrypsin); Smith et al., J. Clin. Invest., 84:1145-1146 (1989) (α-proteinase inhibitor); Debs et al., J. Immunol., 140:3482-3488 (1933) (recombinant gamma interferon and tumor necrosis factor alpha); PCT patent application no.

WO 94/20069, published September 15, 1994 (recombinant pegylated granulocyte colony-stimulating factor).

Brief description of the figures

A large number of other aspects and advantages of the present invention will become apparent upon consideration of the following detailed description of the invention with reference to the figures, in which: FIGURE 1 is a photograph of a PVDF membrane containing PAF-AH purified from human plasma; FIGURE 2 is a diagram showing the enzymatic activity of recombinant human plasma PAF-AH; FIGURE 3 is a schematic drawing depicting recombinant PAF-AH fragments and their catalytic activity; FIGURE 4 shows mass spectroscopy results for a recombinant PAF-AH product, rPH.2; FIGURE 5 shows mass spectroscopy results for a recombinant PAF-AH product, rPH.9; FIGURE 6 is a bar graph illustrating blockade of PAF-induced rat paw edema by topically administered recombinant PAF-AH of the invention; FIGURE 7 is a bar graph illustrating blockade of PAF-induced rat paw edema by intravenously administered PAF-AH; FIGURE 8 is a bar graph showing that PAF-AH blocks PAF-induced edema but not zymosan A-induced edema; FIGURES 9A and 9B show dose response results for PAF-AH anti-inflammatory activity in rat foot edema; FIGURES 10A and 10B show results indicating the in vivo efficacy of a single dose of PAF-AH over time. FIGURE 11 is a line graph showing the pharmacokinetics of PAF-AH in rat circulation; and FIGURE 12 is a bar graph showing the anti-inflammatory effects of PAF-AH compared to the lesser effects of PAF antagonists in rat paw edema; FIGURE 13 shows results indicating that PAF-AH neutralizes the apoptotic effects of conditioned medium from HIV-1 infected and activated monocytes.

Detailed description

The following examples illustrate the invention. Example shows a new method for purifying PAF-AH from human plasma. Example 2 describes amino acid microsequencing of the purified human plasma PAF-AH. The cloning of a truncated cDNA encoding human plasma PAF-AH is described in example 3. Identification of a putative splice variant of the human plasma PAF-AH gene is described in example 4. The cloning of genomic sequences encoding human plasma PAF-AH, is described in Example 5. Example 6 describes the cloning of dog, mouse, bovine, chicken, rodent and macaque cDNA homologous to the human plasma PAF-AH cDNA. Example 7 shows the results of an assay to demonstrate the enzymatic activity of recombinant PAF-AH transiently expressed in COS 7 cells. Example 8 describes expression of untruncated, truncated and chimeric human PAF-AH DNA in E. coli, S. cerevisiae and mammalian cells. Example 9 shows protocols for purification of recombinant PAF-AH from E. coli and assays confirming its enzymatic activity. Example 10 describes various recombinant PAF-AH products, including amino acid substitution analogs and amino and carboxy truncated products, and describes experiments showing that native PAF-AH isolated from plasma is glycosylated. Results from a Northern blot analysis for expression of human plasma PAF-AH-RNA in various tissues and cell lines are shown in Example 11, while results of in situ hybridization are shown in Example 12. Example 13 describes development of monoclonal and polyclonal antibodies specific for human plasma PAF-AH. Examples 14, 15, 16, 17, 18 and 19 show the therapeutic effect in vivo of administration of recombinant PAF-AH products according to the invention on acute inflammation, pleurisy, asthma, necrotizing enterocolitis, respiratory distress syndrome in adults and pancreatitis in animal models respectively . Example 20 describes the in vitro effect of recombinant PAF-AH product on neurotoxicity associated with HIV infection. Example 21 shows the results of immunoassays of serum from human patients exhibiting a failure in PAF-AH activity, and describes the identification of a genetic lesion in the patients which is clearly responsible for the failure.

Example 1

PAF-AH was purified from human plasma to provide material for amino acid sequencing.

A. Optimization of purification conditions

Low-density lipoprotein (LDL) particles were first precipitated from plasma with phosphotungstate and dissolved in 0.1% Tween 20 and subjected to chromatography on a DEAE column (Pharmacia, Uppsala, Sweden) according to the method of Stafforini et al. (1987), supra, but inconsistent elution of PAF-AH activity from the DEAE column required re-evaluation of the dissolution and subsequent purification conditions.

Tween 20, CHAPS (Pierce Chemical Co., Rockford,

IL) and octylglucoside were evaluated by centrifugation and gel filtration chromatography for their ability to solubilize LDL particles. CHAPS gave 25% higher recovery of dissolved activity than Tween 20 and 300% higher recovery than octylglucoside. LDL precipitates solubilized with 10 mM CHAPS were then fractionated on a DEAE-Sepharose Fast Flow column (an anion exchange column; Pharmacia) with buffer containing 1 mM CHAPS, to provide a large aliquot of partially purified PAF-AH (the "DEAE aliquot" ) for evaluation of additional columns.

The DEAE portion was used as starting material to test different chromatography columns for applicability in further purification of the PAF-AH activity. The columns tested included: Blue Sepharose Fast Flow (Pharmacia), a dye ligand affinity column; S-Sepharose Fast Flow (Pharmacia), a cation exchange column; Cu-chelating Sepharose (Pharmacia), a metal-ligand affinity column; Fractogel S (EM Separations, Gibbstown, NJ), a cation exchange column; and. Sephacryl-200 (Pharmacia), a gel filtration column. These chromatographic procedures all gave low, unsatisfactory purification levels when performed in 1 mM CHAPS. Subsequent gel filtration chromatography on Sephacryl S-200 in 1 mM CHAPS gave an enzymatically active fraction that eluted in a wide size range instead of the expected size of approx. 44 kDa. Taken together, these results indicated that the LDL proteins aggregated in solution.

Different LDL samples were therefore evaluated by means of analytical gel filtration chromatography for aggregation of the PAF-AH activity. Samples from the DEAE portion and of freshly dissolved LDL precipitate were analyzed on Superose 12 (Pharmacia) equilibrated in buffer with 1 mM CHAPS. Both samples eluted in a very broad molecular weight range, with most of the activity eluting above 150 kDa. When the samples were then analyzed on Superose 12 equilibrated with 10 mM CHAPS, the bulk of the activity eluted near 44 kDa, as expected, for PAF-AH activity. However, the samples contained some PAF-AH activity in the high molecular area corresponding to aggregates.

Other samples eluted PAF-AH activity exclusively in the ca. 44 kDa range when they were later tested by gel filtration. These samples were an LDL precipitate dissolved in 10 mM CHAPS in the presence of 0.5 M NaCl, and a freshly prepared DEAE aliquot adjusted to 10 mM CHAPS after elution from the DEAE column. These data indicate that at least 10 mM CHAPS is required to maintain non-aggregated PAF-AH. Increasing the CHAPS concentration from 1 mM to 10 mM after chromatography on DEAE, but before subsequent chromatographic steps, resulted in dramatic purification differences. The degree of PAF-AH purification on S-Sepharose Fast Flow was e.g. increased from 2 times to 10 times. PAF-AH activity was bound irreversibly to the Blue Sepharose Fast Flow column in 1 mM CHAPS, but the column gave the highest level of purification in 10 mM CHAPS. DEAE chromatography was not improved by prior addition of 10 mM

CHAPS.

Chromatography on Cu-chelating Sepharose following the Blue Sepharose Fast Flow column concentrated the PAF-AH activity 15-fold. It was also determined that PAF-AH activity could be recovered from a reduced SDS-polyacrylamide gel, as long as the samples were not boiled. The activity of the material eluting from the Cu-chelating Sepharose column when subjected to SDS-polyacrylamide gel electrophoresis coincided with a major protein band when the gel was silver stained.

B. PAF-AH cleaning protocol1

The new protocol used to purify PAF-AH for amino acid sequencing therefore included the following steps, which were performed at 4 °C. Human plasma was aliquoted into 900 ml aliquots into Nalgene bottles and adjusted to pH 8.6. LDL particles were then precipitated by the addition of 90 mL of 3.85% sodium phosphotungstate, followed by 23 mL of 2 M MgCl 2 . Plasma was then centrifuged for 15 minutes at 3600 g. Pellets were resuspended in 800 ml of 0.2% sodium citrate. LDL was reprecipitated by addition of 10 g NaCl and 24 ml 2 M MgCl 2 . LDL particles were pelleted by centrifugation for 15 minutes at 3600 g. This was repeated twice. The pellets were then frozen at -20 °C. LDL particles from 5 L of plasma were resuspended in 5 L of buffer A (25 mM Tris-HCl, 10 mM CHAPS, pH 7.5) and stirred overnight. Solubilized LDL particles were centrifuged at 3600 g for 1.5 h. Supernatants were combined and filtered with Whatman 113 filter paper to remove any remaining solid materials. Solubilized LDL supernatant was applied to a DEAE-Sepharose Fast Flow column (11 cm x 10 cm; 1 L resin volume; 80 ml/min) equilibrated in buffer B (25 mM Tris-HCl, 1 mM CHAPS, pH 7.5 ). The column was washed with buffer B until the absorbance returned to baseline. Protein was eluted with an 8 L, 0-0.5 M NaCl gradient, and 480 ml fractions were collected. This step was necessary to achieve binding to the Blue Sepharose Fast Flow column indicated below. Fractions were analyzed for acetyl hydrolase activity, essentially using the method described in Example 4.

Active fractions were pooled, and sufficient CHAPS was added to give a concentration of approx. 10 mM

CHAPS. The DEAE aliquot was applied overnight at a rate of 4 ml/min to a Blue Sepharose Fast Flow column (5 cm x 10 cm; 200 ml material volume) equilibrated in buffer A containing 0.5 M NaCl. The column was washed with the equilibration buffer at a rate of 16 ml/min until the absorbance returned to baseline. PAF-AH activity was eluted stepwise with buffer A containing 0.5 M KSCN (a chaotropic salt) at a rate of 16 ml/min and collected in 50 ml fractions. This step gave a purification higher than 1000 times. Active fractions were pooled, and the aliquot was adjusted to pH 8.0 with 1 M Tris-HCl, pH 8.0. The active portion from Blue Sepharose Fast Flow chromatography was applied to a Cu-chelating Sepharose column (2.5 cm x 2 cm; 10 ml material volume; 4 ml/min) equilibrated in buffer C [25 mM Tris-HCl, 10 mM CHAPS, 0.5 M NaCl, pH 8.0 (pH 7.5 also worked)], and the column was washed with 50 ml of buffer C. PAF-AH activity was eluted with 100 ml of 50 mM imidazole in buffer C, and collected in 10 ml fractions. Fractions containing PAF-AH activity were pooled and dialyzed against buffer A. In addition to providing a 15-fold concentration of PAF-AH activity, the Cu-chelating Sepharose column provided a small purification. The Cu-chelating Sepharose portion was reduced in 50 mM DTT for 15 min at 37°C and applied to a 0.75 mm, 7.5% polyacrylamide gel. Gel pieces were excised every 0.5 cm and placed in disposable microcentrifuge tubes containing 200 µl of 25 mM Tris-HCl, 10 mM CHAPS, 150 mM NaCl. The pieces were ground and incubated overnight at 4 °C. The supernatant from each gel piece was then analyzed for PAF-AH activity to determine which protein band on SDS-PAGE contained PAF-AH activity. PAF-AH activity was found in the band of approx. 44 kDa. Protein from a duplicate gel was electrotransferred to a PVDF membrane (Immobilon-P, Millipore) and stained with Coomassie Blue. A photograph of the PVDF membrane is shown in Figure 1.

As shown in table 1 below, approx. 200 (ug PAF-AH purified 2 x 10<6> fold from 5 l human plasma. Comparison with this is a 3 x 10<*> fold purification of PAF-AH activity, described by Stafforini et al., (1987) , supra. ;In summary, the following steps were unique and critical for the successful purification of plasma PAF-AH for micro-sequencing: (1) resolution and chromatography in 10 mM CHAPS, (2) chromatography on a blue ligand affinity column, such as Blue Sepharose Fast Flow, (3) chromatography on a Cu-ligand affinity column, such as Cu-chelating Sepharose, and (4) elution of PAF-AH from SDS-PAGE. ;Example 2 ;For amino acid sequencing, the approximate 44 kDa- protein band from the PAF-AH-containing PVDF membrane described in Example 1 excised and sequenced using an Applied Biosystems 473A protein sequencer N-terminal sequence analysis of the approximately 44 kDa protein band corresponding to PAF-AH activity indicated that the band contained two major sequences and two minor sequences The relationship between the two ho wood sequences were 1:1, and it was therefore difficult to interpret the sequence data. ;To distinguish the sequences of the two major proteins that had been separated on the SDS gel, a duplicate PVDF membrane containing the approximately 44 kDa band was split in the middle, so that separate sequencing of the upper part and the lower part of the membrane. ;The N-terminal sequence obtained for the lower part of the membrane was: ;A search of protein databases revealed that this sequence was a fragment of human serum albumin. The upper part of the same PVDF membrane was also sequenced, and the N-terminal amino acid sequence was: ;This sequence did not correspond to any protein in the searched databases and was different from the N-terminal amino acid sequence: ;which was reported for erythrocyte cytoplasmic PAF- AH by Stafforini et al. (1993), supra. The new sequence (SEQ ID NO: 2) was used for cDNA cloning of human plasma PAF-AH, as described below in Example 3. ;Example 3 ;An unabridged clone encoding human plasma PAF-AH, was isolated from a macrophage cDNA library. A. Construction of a macrophage cDNA library; Poly A<+>RNA was harvested from monocyte-derived macrophages from peripheral blood. Double-stranded butt-ended cDNA was generated using the Invitrogen Copy Kit (San Diego, CA), and BstXI adapters were ligated to the cDNA before insertion into the mammalian expression vector pRc/CMV (Invitrogen). The resulting plasmids were introduced into E. coli strain XL-1 Blue by electroporation. Transformed bacteria were spread at a density of approx. 3000 colonies per agarose plate, on a total of 978 plates. Plasmid DNA prepared separately from each plate was kept in individual aliquots and was also combined into larger aliquots each representing 300,000 clones. B. Library screening by PCR; The macrophage library was screened by polymerase chain reaction using a degenerate antisense oligonucleotide PCR primer based on the novel N-terminal amino acid sequence described in Example 2. The sequence of the primer is shown below in IUPAC nomenclature, and where " I" is an inosine. ;The codon selection tables of Wada et al., Nuc. Acids Res., 19S:1981-1986 (1991), was used to select nucleotides in the third position of each codon in the primer. The primer was used in combination with a primer specific for either the SP6 or T7 promoter sequences, both of which flank the pRc/CMV cloning site, to screen the macrophage library portions from 300,000 clones. All PCR reactions contained 100 ng of template cDNA, 1 µg of each primer, 0.125 mM of each dNTP, 10 mM Tris-HCl, pH 8.4, 50 mM MgCl 2 , and 2.5 units of Taq polymerase. An initial denaturation step at 94°C for 4 minutes was followed by 30 amplification cycles of 1 minute at 94°C, 1 minute at 60°C and 2 minutes at 72°C. The resulting PCR product was cloned into pBluescript SK" (Stratagene, La Jolla, CA), and its nucleotide sequence was determined by the dideoxy chain termination method. The PCR product contained the sequence predicted by the new peptide sequence and corresponds to nucleotides 1-331 of SEQ .ID NO 7. ;The PCR primers shown below, specific for the cloned PCR fragment described above, were then designed for identification of an unabridged clone. ;PCR reactions using the primers were performed as described above for first screening the cDNA portions of 300,000 clones, and then the appropriate subset of the smaller portions of 3,000 clones.Three portions of ;3,000 clones that produced a PCR product of the expected size were then used to transform bacteria. ;C.Library Screening by Hybridization ;DNA from the transformed bacteria was then screened by hybridization using the original cloned PCR fragment as a probe.Colony were blotted onto nitrocellulose and prehybridized and hybridized in 50% formamide, 0.75 M sodium chloride, 0.075 M sodium citrate, 0.05 M sodium phosphate, pH 6.5, 1% polyvinylpyrrolidine, 1% Ficoll, 1% bovine serum albumin, and 50 ng /ml sonicated salmon sperm DNA. The hybridization probe was labeled by random hexamer priming. After overnight hybridization at 42°C, blots were washed extensively in 0.03 M sodium chloride, 3 mM sodium citrate, 0.1% SDS, at 42°C. The nucleotide sequence of 10 hybridization clones was determined. One of the clones, clone sAH 406-3, contained the sequence predicted by the original peptide sequence for the PAF-AH activity purified from human plasma. The DNA and the deduced amino acid sequences of human plasma PAF-AH are shown in SEQ ID NO:2, respectively. NO. 7 and 8. Clone sAH 406-3 contains a 1.52 kb insert with an open reading frame encoding a predicted protein of 441 amino acids. At the amino terminus lies a relatively hydrophobic segment with 41 residues in front of the N-terminal amino acid (isoleucine in position 42 in SEQ ID NO: 8) identified by means of protein microsequencing. The encoded protein can thus either have a long signal sequence or a signal sequence plus a further peptide which is cleaved off to give the mature functional enzyme. The presence of a signal sequence is a characteristic feature of secreted proteins. The protein encoded by clone sAH 406-3 also includes the consensus GxSxG motif (amino acids 271-275 of SEQ ID NO: 8), which is believed to contain the active site serine of all known mammalian lipases, microbial lipases and serine proteases. See Chapus et al., Biochimie, 70:1223-1224 (1988), and Brenner, Nature, 334:528-530 (1988). ;Table 2 below is a comparison between the amino acid composition of the human plasma PAF-AH of the present invention, as predicted from SEQ ID NO. NO. 8, and the amino acid composition of the allegedly purified material described by Stafforini et al. (1987), supra. The amino acid composition of the mature form of the human plasma PAF-AH of the present invention and the amino acid composition of the previously purified material claimed to be the human plasma PAF-AH are clearly different from each other. ;When the nucleotide of Hattori et al., supra, and the deduced amino acid sequences of bovine brain cytoplasmic PAF-AH were compared with the amino acid sequences of the human plasma PAF-AH of the present invention, no significant structural similarity was observed in the sequences . ;Example 4 ;A putative splice variant of the human PAF-AH gene was detected when PCR was performed on macrophage and stimulated ;PBMC cDNA using primers that hybridized to the 5'-untranslated region (nucleotides 31-52 of SEQ .ID NO.7) and the region spanning the translation termination codon at the 3<1> end of the PAF-AH cDNA (nucleotides 1456-1487 in SEQ ID NO.7). The PCR reactions gave two bands on a gel, one band corresponding to the expected size of the PAF-AH cDNA according to Example 3, and the other which was approx. 100 bp shorter. Sequencing of both bands revealed that the larger band was the PAF-AH cDNA of Example 3, while the shorter band lacked exon 2 (Example 5 below) of the PAF-AH sequence encoding the putative signal and propeptide sequences of plasma PAF -Ah. The predicted catalytic triad and all cysteines were present in the shorter clone, and therefore the biochemical activity of the protein encoded by the clone probably corresponds to the activity of the plasma enzyme. ;To begin to assess the biological relevance of the PAF-AH splice variant predicted to encode a cytoplasmically active enzyme, the relative frequency of ;the two forms in blood monocyte-derived macrophages was analyzed using RNase protection. Neither of the two messages was present in freshly isolated monocytes, but both were found on day 2 of in vitro differentiation of the monocytes into macrophages, and persisted during 6 days of culture. The amount of the two messages was approximately equivalent throughout the differentiation period. In contrast, similar analyzes of neural tissue revealed that only the truncated message predicted to encode the truncated extracellular form of PAF-AH is expressed. ;Example 5 ;Genomic human plasma PAF-AH sequences were also isolated. The structure of the PAF-AH gene was determined by isolating lambda and P1 phage clones containing human genomic DNA by DNA hybridization under high stringency conditions. Fragments of the phage clones were subcloned and sequenced using primers designed to anneal at regular intervals through the cDNA clone sAH 406-3. New sequencing primers designed to anneal to the intron regions flanking the exons were also used to sequence back through the exon-intron boundaries to confirm the sequences. Exon/intron boundaries were defined as the points where the genomic sequences and cDNA sequences diverged. These analyzed revealed that the human PAF-AH gene consists of 12 exons. ;Exons 1, 2, 3, 4, 5, 6 and part of 7 were isolated from male fetal placenta library constructed in lambda FIX (Stratagene). Phage plaques were blotted onto nitrocellulose and prehybridized and hybridized in 50% formamide, 0.75 M sodium chloride, 75 mM sodium citrate, 50 mM sodium phosphate (pH 6.5), 1% polyvinylpyrrolidine, 1% Ficoll, 1% bovine serum albumin, and 50 ng /ml sonicated salmon sperm DNA. The hybridization probe used to identify a phage clone containing exons 2-6 and part of 7 consisted of the entire cDNA clone sAH-406-3. A clone containing exon 1 was identified using a fragment derived from the 5' end of the cDNA clone (nucleotides 1-312 of SEQ ID NO: 7). Both probes were labeled with <32>P using random hexamer priming. After overnight hybridization at 42°C, blots were washed extensively in 30 mM sodium chloride, 3 mM sodium citrate, 0.1% SDS, at 42°C. The DNA sequences of exons 1, 2, 3, 4, 5, and 6, together with partially surrounding intron sequences, are shown in SEQ ID NO:2, respectively. NO. 9, 10, 11, 12, 13 and 14. ;The rest of exon 7, as well as exons 8, 9, 10, 11 and 12, were subcloned from a P1 clone isolated from human P1 genomic library. P1 phage plaques were blotted onto nitrocellulose and prehybridized and hybridized in 0.75 M sodium chloride, 50 mM sodium phosphate (pH 7.4), 5 mM EDTA, 1% polyvinylpyrrolidine, 1% Ficoll, 1% bovine serum albumin, 0.5% SDS and 0.1 mg/ml total human DNA. The hybridization probe, labeled with <32>P using random hexamer priming, consisted of a 2.6 kb EcoRI fragment of genomic DNA derived from the 3' end of a lambda clone isolated above. This fragment contained exon 6 and the part of exon 7 that was present on the phage clone. After overnight hybridization at 65°C, the blots were washed as described above. The DNA sequences of exons 7, 8, 9, 10, 11 and 12, together with partially surrounding intron sequences, are shown in SEQ ID NO: 1, respectively. NO. 15, 16, 17, 18, 19 and 20. ;Example 6 ;Untruncated plasma PAF-AH cDNA clones were isolated from mouse spleen, dog spleen, bovine spleen and chicken spleen cDNA libraries, and a partial rodent clone was isolated from a rat thymus cDNA library. The clones were identified by low-stringency hybridization to the human cDNA (hybridization conditions were the same as described for exons 1-6 in Example 5 above, except that 20% formamide was used instead of 50% formamide). A 1 kb HindIII fragment of the human PAF-AH-sAH 406-3 cDNA clone (nucleotides 309-1322 of SEQ ID NO:7) was used as a probe. A partial monkey clone was also isolated from macaque brain cDNA by PCR using primers based on nucleotides 285-303 and 851-867 of SEQ ID NO. NO. 7. The nucleotide sequences and the deduced amino acid sequences of mouse, dog, bovine, chicken, rat and macaque cDNA clones are shown in SEQ ID NO: NO. 21," 22, 23, 24, 25 and 26. ;A comparison between the deduced amino acid sequences of the cDNA clones and the human cDNA clone results in the amino acid percent identity values shown in Table 3 below. ;Approximately 38% of the residues are complete conserved in all the sequences. The most divergent regions are at the amino-terminal end (containing the signal sequence) and the carboxyl-terminal end shown in Example 10 as non-critical for enzymatic activity. The Gly-Xaa-Ser-Xaa-Gly motif (SEQ. ID NO 27) found in neutral lipases and other esterases was conserved in bovine, dog, mouse, rat and chicken PAF-AH The central serine in this motif serves as the active site nucleophile for these enzymes . The predicted aspartate and histidine components in the active site (Example 10A) were also conserved. The human plasma PAF-AH of the present invention therefore appears to utilize a catalytic triad and may adopt the α/β hydrolase conformation of the neutral lipases, even if it does not show any other sequence homology with the lipases. ;Human plasma PAF-AH is also expected to contain a region that mediates its specific interaction with the low-density and high-density lipoprotein particles in plasma. Interaction with these particles can be mediated by the N-terminal half of the molecule, which contains large stretches of amino acids that are highly conserved among species, but do not contain the enzyme's catalytic triad. ;Example 7 ;To determine whether human plasma PAF-AH cDNA clone sAH 406-3 (Example 3) encodes a protein with PAF-AH activity, the pRc/CMV expression construct was transiently expressed in COS 7 cells . Three days after transfection using a DEAE-dextran method, COS cell medium was assayed for PAF-AH activity. Cells were seeded at a density of 300,000 cells per 60 mm tissue culture dish. The following day, the cells were incubated for 2 hours in DMEM containing 0.5 mg/ml DEAE-dextran, 0.1 mM chloroquine and 5-10 µg plasmid DNA. The cells were then treated with 10% DMSO in phosphate-buffered saline for 1 min, washed with medium, and incubated in DMEM containing 10% fetal calf serum previously treated with diisopropylfluorophosphate (DFP) to inactivate endogenous bovine serum PAF-AH. After 3 days of incubation, media from transfected cells were assayed for PAF-AH activity. Assays were performed in the presence and absence of either 10 mM EDTA or 1 mM DFP to determine whether the recombinant enzyme was calcium independent, and was inhibited by the serine esterase inhibitor DFP, as previously described for plasma PAF-AH by Stafforini et al. (1987), supra. Negative controls included cells transfected with pRc/CMV that either lacked an insert or contained the sAH 406-3 insert in the opposite orientation. ;PAF-AH activity in transfectant supernatants was determined using the method of Stafforini et al. (1990), supra, with the following modifications. Briefly, PAF-AH activity was determined by measuring the hydrolysis of 3H-acetate from [acetyl-3H]PAF (New England Nuclear, Boston, MA). The aqueous, free <3>H-acetate was separated from labeled substrate by reverse-phase column chromatography over octadecyl silica gel cartridges (Baker Research Products, Phillipsburg, PA). Assays were performed using 10 µl of transfectant supernatant in 0.1 M HEPES buffer, pH 7.2, in a 50 µl reaction volume. A total of 50 pmol substrate was used per reaction with a 1:5 ratio of labeled to unlabeled PAF. The reaction mixtures were incubated for 30 minutes at 37°C, and the reactions were stopped by the addition of 40 µl of 10 M acetic acid. The solution was then washed through the octadecyl silica gel cartridges, which were then rinsed with 0.1 M sodium acetate. The aqueous eluate from each sample was collected and counted in a liquid scintillation counter for 1 minute. Enzyme activity was expressed in counts per minute. As shown in Figure 2, media from cells transfected with sAH 406-3 contained PAF-AH activity at levels fourfold higher than background. This activity was unaffected by the presence of EDTA, but was abolished by 1 mM DFP. These observations show that clone sAH 406-3 encodes an activity consistent with the human plasma enzyme PAF-AH. Example 8 Unabridged and variously truncated human plasma PAF-AH DNA and a chimeric mouse-human PAF-AH DNA were expressed in E. coli and yeast, and stably expressed in mammalian cells using recombinant methods. A. Expression in E . coli ; PCR was used to generate a protein-coding fragment of human plasma PAF-AH cDNA from clone sAH 406-3 that could be readily subcloned into an E. coli expression vector. The subcloned segment started at the 5<1> end of the human gene with the codon encoding Ile42 (SEQ ID NO: 8), the N-terminal residue of the enzyme purified from human plasma. The remainder of the gene, through the native termination codon, was included in the construct. The 5'-sense PCR primer used was: and contained an XbaI cloning site as well as a translation initiation codon (underlined). The 3'-antisense primer used was: ; and included the termination codon for sAH 406-3, and contained an EcoRV cloning site. PCR reactions were performed essentially as described in Example 3. The resulting PCR product was digested with XbaI and EcoRV and subcloned into a pBR322 vector containing the Trp promoter [deBoer et al., PNAS, 80:21-25;(1983 )] immediately upstream of the cloning site. E. coli strain XL-1 Blue was transformed with the expression construct and grown in L nutrient solution containing 100 g/ml carbenicillin. Transformants from overnight cultures were pelleted and resuspended in lysis buffer containing 50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 10 mM CHAPS, 1 mM EDTA, 100 µg/ml lysozyme, and 0.05 trypsin inhibitory units (TIU)/ml aprotinin. After incubation for 1 hour on ice and sonication for 2 minutes, the lysates were assayed for PAF-AH activity using the method described in Example 4. E. coli transformed with the expression construct (designated trp AH) produced a product with PAF-AH activity . See Table 6 of Example 9. ;Constructs that included three additional promoters, the tacJJ promoter (deBoer, supra), the arabinose (ara)B promoter from Salmonella typhimurium [Horwitz et al., Gene, 14:309-319 (1981) ] and the bacteriophage T7 promoter, were also used to drive expression of human PAF-AH sequences in E. coli. Constructs comprising the Trp promoter (pUC trp AH), the tacIJ promoter (pUC tac AH) and the araB promoter (pUC ara AH) were assembled into plasmid pUC19 (New England Biolabs, MA), while the construct comprising the T7 promoter (pET AH), was assembled into plasmid pET15B (Novagen, Madison, WI). A construct containing a hybrid promoter, pHAB/PH consisting of the araB promoter fused to the ribosome binding sites of the T7 promoter region, was also assembled into pET15B. All E. coli constructs produced PAF-AH activity within a range of 20 to 50 U/ml/OD600 . This activity corresponded to a total recombinant protein mass of > 1% of the total amount of cell protein. ;Several E. coli expression constructs producing PAF-AH with extended amino termini were also evaluated. The N-terminus of native plasma PAF-AH was identified as Ile42 by amino acid sequencing (Example 2). However, the sequence immediately upstream of Ile42 is not consistent with amino acids found at signal sequence cleavage sites [that is, the "-3-1 rule" is not followed because lysine is not found in position -1; see von Heijne, Nuc. Acids Res., 14:4683-4690 (1986)]. Presumably a more classical signal sequence (M^A^ or M^P^) is recognized by the cellular secretion system, followed by endo-proteolytic cleavage. The entire coding sequence for PAF-AH, starting at the initial methionine (nucleotides 162-1487 of SEQ ID NO:7), was designed for expression in E. coli using the trp promoter. As shown in Table 4, this construct produced active PAF-AH, but the expression was approx. one fiftieth of the level of the original construct, starting at Ile42. Another expression construct, starting at Val18 (nucleotides 213-1487 of SEQ ID NO:7), produced active PAF-AH at ca. one third of the level of the original construction. These results suggest that amino-terminal end extensions are not critical or necessary for activity of recombinant PAF-AH produced in E. coli. Truncated recombinant human PAF-AH products were also produced in E. coli using a low copy number plasmid and a promoter inducible by addition of arabinose to the culture. Such an N-terminal, truncated PAF-AH product is the recombinant expression product from DNA encoding amino acid residues Met46-Asn441 of the polypeptide encoded by the unabridged PAF-AH cDNA (SEQ ID NO: 8), and is designated rpH.2. The plasmid used for production of rpH.2 in bacterial cells was pBAR2/PH.2, a pBR322-based plasmid containing (1) nucleotides 297-1487 of SEQ ID NO. NO. 7 encoding human PAF-AH starting with the methionine codon at position 46, (2) the araB-C promoters and the araC gene from the arabinose operon of Salmonella typhimurium, (3) a transcription termination sequence from bacteriophage T7, and (4) a replication start region from bacteriophage etc. ;pBAR2/PH.2 specifically included the following segments of DNA: (1) from the broken AatII site at position 1994 to the EcoRI site at nucleotide 6274, vector sequence containing an origin of replication region and genes encoding resistance to either ampicillin or tetracycline-derived from the bacterial plasmid pBR322; (2) from the EcoRI site at position 6274 to the XbaI site at position 131, DNA from the Salmonella typhimurium arabinose operon (Genbank accession numbers M11045, M11046, M11047, J01797); (3) from the XbaI site at position 131 to the NcoI site at position 170, DNA containing a ribosome binding site from pET-21b (Novagen, Madison, WI); ;(4) from the NcoI site at position 170 to the XhoI site at position 1363, human PAF-AH cDNA sequence; and (5) from the XhoI site at position 1363 to the broken AatII site at position 1993, a DNA fragment from pET-21B (Novagen) containing a transcription termination sequence from bacteriophage T7 and a replication start region from bacteriophage fl. Another PAF-AH product, designated rPH.9, is the recombinant expression product from DNA encoding amino acid residues Met46-Ile429 of the polypeptide encoded by the unabridged PAF-AH cDNA (SEQ ID NO: 8). DNA encoding rpH.9 was inserted into the same vector used for the production of rPH.2 in bacterial cells. This plasmid was designated pBAR2/PH.9 and specifically included the following segments of DNA: (1) from the broken AatII site at position 1958 to the EcoRI site at nucleotide 623 9 of the vector sequence containing an origin of replication and genes encoding resistance to either ampicillin or tetracycline derived from the bacterial plasmid pBR322; (2) from the EcoRI site at position 6239 to the XbaI site at position 131, DNA from the Salmonella typhimurium arabinose operon (Genbank accession numbers M11045, M11046, M11047, J01797); (3) from the XbaI site at position 131 to the NcoI site at position 170, DNA containing a ribosome binding site from pET-21b (Novagen, Madison, WI); (4) from the NcoI site at position 170 to the XhoI site at position 1328, human PAF-AH cDNA sequence; (5) from the XhoI site at position 1328 to the broken AatII site at position 1958, a DNA fragment from pET-21B (Novagen, Madison, WI) containing a transcription termination sequence from bacteriophage T7 and a replication start region from bacteriophage fl. ;Expression of PAF-AH products in pBAR2/PH.2 and pBAR2/PH.9 is under the control of the araB promoter, which is strongly repressed in the presence of glucose and in the absence of arabinose, but acts as a strong promoter when L -arabinose is added to cultures without glucose. Selection for cells containing plasmids can be achieved by adding either ampicillin (or related antibiotics) or tetracycline to the culture medium. Many different E. coli strains can be used as a host for recombinant expression of PAF-AH products, including, but not limited to, strains that are proteotrophic for arabinose metabolism, such as W3110, DH5, BL21, C600, JM101 and their derivatives, strains containing mutations that reduce proteolysis, such as CAG62 9, KY1429, and strains defective in their ability to degrade arabinose, such as SB7219 and MC1061. The advantage of using a strain that cannot degrade arabinose is that the inducer (arabinose) for the production of PAF-AH is not used up in the medium during the induction period, resulting in higher levels of PAF-AH compared to that obtained with strains that are able to metabolize arabinose. Any suitable medium and culture conditions can be used to express active PAF-AH products in different E. coli strains. For example, either rich medium formulations, such as LB, EDM2 95 (an M9-based minimal medium supplemented with yeast extract and acid hydrolyzed casein), or "defined" medium, such as A675, an A-based minimal medium adjusted to pH 6, allow 75 with glycerol as carbon source and supplemented with trace elements and vitamins, significant production of rPAF-AH products. Tetracycline is included in the medium to maintain selection of the plasmid.

The plasmid pBAR2/PH.2 was transformed into E. coli strain MC1061 (ATCC 53338) which contains a deletion of the arabinose operon, and thereby cannot metabolize arabinose. MC1061 is also a leucine auxotroph and was grown by a batch process using a defined medium containing casamino acids that complement the 1euc mutation.

E. coli MC1061 cells transformed with pBAR2/PH.2 were grown at 30°C in batch medium containing 2 g/l glucose. Glucose has two purposes, being a carbon source for cell growth and a repressor of the arabinose promoter. When the batch glucose levels were depleted (< 50 mg/l), an infusion of nutrient (containing 300 g/l glucose) was initiated. The nutrient supply was increased linearly for 16 hours to a degree that limited formation of acid by-product. At this point, the nutrient supply was changed to medium containing glycerol instead of glucose. At the same time, 500 g/l L-arabinose was added to a final concentration of 5 g/l. The glycerol feed was maintained at a constant feed rate for 22 hours. The cells were harvested using hollow fiber filtration to concentrate the suspension approx. 10 times. Cell paste was stored at -70 °C. A final cell mass of approx. 80 g/l (OD600 = 50 - 60) with a PAF-AH activity of 65-70 U/OD/ml, which represents approx. 10

% of the total cell protein. The final cultivation volume of approx. 75 1 contained 50-60 g of PAF-AH.

High-level production of rPAF-AH products can be achieved when pBAR2/PH.2 or -PH.9 is expressed by strain SB7219 or MC1061. Other strains with failure of arabinose degradation are suitable for the production of high-density cells. Preferably, the cells are cultured under the following conditions. Exponentially growing SB7219;pBAR2/PH.2 and SB7219;pBAR2/PH.9 strains were inoculated into fermenters containing batch medium containing 2 g/l glucose. When the glucose is consumed, a glycerol solution containing trace elements, vitamins, magnesium and ammonium salt is added to the tanks to maintain healthy exponential growth. The tanks are maintained at 30 °C, supplied with air to supply oxygen and agitated to maintain the level of dissolved oxygen higher than approx. 15% saturation. When the cell density in the culture is higher than 110 g/l (wet cell mass), a constant feed rate is introduced, and a bolus addition of L-arabinose is added to the culture (approx. 0.5% final amount). Product formation is observed for 16-22 hours. 40-50 g/l (dry cell weight) is typically obtained from the cultures. The cells are harvested by centrifugation, stored at -70 °C and rPAF-AH product is purified for analysis. Specific product ivities higher than 150 units/ml/OD600 are routinely achieved.

B. Expression in yeast cells

Recombinant human PAF-AH was also expressed in Saccharomyces cerevisiae. The yeast ADH2 promoter was used to drive rPAF-AH expression and produced 7 U/ml/OD600 (Table 5 below).

C. Expression of PAF-AH in mammalian cells

1. Expression of human PAF-AH cDNA constructs

Plasmids engineered for expression of PAF-AH, with the exception of pSFN/PAFAH.1, utilize a strong viral promoter from cytomegalovirus, a polyadenylation site from the bovine growth hormone gene, and the SV40 origin of replication to allow high copy number replication of the plasmid in COS cells. Plasmids were electroporated into cells.

A first set of plasmids was constructed in which the 5'-flanking sequence (pDCl/PAFAH.1) or both the 5'- and 3<1->flanking sequences (pDCl/PAFAH.2) of the human PAF-AH cDNA was replaced with flanking sequences from other genes known to be expressed at high levels in mammalian cells. Transfection of these plasmids into COS, CHO or 293 cells led to the production of PAF-AH at approximately the same level (0.01 units/ml or 2-4 times higher than background) as the level reported for clone sAH 406- 3 in example 7 after short-term transfection of COS cells. Another plasmid was constructed that included a Friend splenic focus-forming virus promoter in place of the cytomegalovirus promoter. The human PAF-AH cDNA was inserted into plasmid pmH-neo [Hahn et al., Gene, 127:267 (1993)] under the control of the Friend spleen focus-forming virus promoter. Transfection of the myeloma cell line NSO with the plasmid designated pSFN/PAFAH.1 and screening of several hundred clones resulted in the isolation of two transfectants (4B11 and 1C11) that represented 0.15-0.5 units/ml PAF-AH activity. . If a specific activity of 5000 units/mg is assumed, the productivity of these two NSO transfectants corresponds to approx. 0.1 mg/l.

2. Expression of mouse-human chimeric PAF-AH gene constructs

A construct (pRc/MS9) containing the cDNA encoding mouse PAF-AH in the mammalian expression vector pRc/CMV resulted in the production of secreted PAF-AH at the level of 5-10 units/ml (1000-fold higher than background) after transfection in COS- cells. Assuming that the specific activity of mouse PAF-AH is approximately the same as the activity of the human enzyme, the mouse cDNA is therefore expressed at a 500-1000 times higher level than the human PAF-AH cDNA.

To examine the difference between the expression levels of human and mouse PAF-AH in COS cells, two mouse-human chimeric genes were constructed and tested for expression in COS cells. The first of these constructs, pRc/PH.MHCl, contains the coding sequence for the N-terminal 97 amino acids of the mouse PAF-AH polypeptide (SEQ ID NO: 21) fused to the C-terminal 343 amino acids of human PAF- AH in the expression vector pRc/CMV (Invitrogen, San Diego, CA). The second chimeric gene, in plasmid pRc/PH.MHC2, contains the coding sequence for the N-terminal 40 amino acids of the mouse PAF-AH polypeptide fused to the C-terminal 400 residues of human PAF-AH in pRc/CMV. Transfection of COS cells with pRc/PH.MHCl led to the accumulation of 1-2 units/ml PAF-AH activity in the medium. Conditioned medium derived from cells transfected with pRc/PH.MHC2 was found to contain only 0.01 units/ml of PAF-AH activity. From these experiments, it appears that the difference in expression level between mouse and human PAF-AH genes is at least partially due to the polypeptide segment between residues 40 and 97, or the corresponding RNA or DNA segment encoding this region of PAF-AH -the protein.

3. Recoding of the first 290 bp of the PAF-AH coding sequence

One hypothesis for the low level of human PAF-AH synthesized in transfected mammalian cells is that the codons used by the native gene are suboptimal for efficient expression. However, it is unlikely that codon usage can explain the 500-1000-fold difference in expression levels between mouse genes and human genes, because optimization of codons generally only has a 10-fold effect on expression at most. A second hypothesis to explain the difference between mouse and human PAF-AH expression levels is that the human PAF-AH mRNA in the 5' coding region forms a secondary structure that either leads to relatively rapid degradation of the mRNA or causes inefficient translation initiation or extension.

To test these hypotheses, a synthetic fragment was constructed that encodes the authentic human PAF-AH protein from the amino terminus to residue 96, but in which most of the codons have been replaced ("recoded") with a codon of a different sequence, but which codes for the same amino acid. Changing the second codon from GTG to GTA resulted in the creation of an Asp718 site that was at one end of the synthetic fragment and is present in mouse cDNA. The other end of the fragment contained the BamHI site, which is normally found at codon 97 in the human gene. It approx. The 290 bp Asp718/BamHI fragment was derived from a PCR fragment generated using the doubly asymmetric PCR method for the construction of synthetic genes described by Sandhu et al., Biotechniques, 12:14-16 (1992). The synthetic Asp718/BamHI fragment was ligated with DNA fragments coding for the rest of the human PAF-AH molecule starting with nucleotide 453 in SEQ ID NO. NO. 7, so that a sequence encoding authentic human PAF-AH enzyme was inserted into the mammalian expression vector pRc/CMV (Invitrogen, San Diego) to form plasmid pRc/HPH.4. The complete sequence of the recoded gene is shown in SEQ ID NO. NO. 30. The 5' flanking sequence adjacent to the human PAF-AH coding sequence in pRc/HPH.4 is from the sequence of a mouse cDNA encoding PAF-AH in pRc/MS9 (nucleotides 1-116 in SEQ ID NO 21).

To test expression of human PAF-AH from pRc/HPH.4, COS cells were transiently transfected with pRc/HPH.4 (reencoded human gene), pRc/MS9 (mouse PAF-AH), or pRc/PH.MHCl ( mouse-human hybrid 1). The conditioned medium from the transfected cells was tested for PAF-AH activity and found to contain 5.7 units/ml (mousegene), 0.9 units/ml (mouse-human hybrid 1) or 2.6 units/ml ( recoded human gene). Thus, the strategy of recoding the first 290 bp of the human PAF-AH coding sequence was successful in increasing expression levels of human PAF-AH from a few nanograms/ml to approx. 0.5 micrograms/ml in a transient COS cell transfection. The recoded PAF-AH gene from pRc/HPH.4 will be inserted into a mammalian expression vector containing the dihydrofolate reductase (DHFR) gene, and DHFR-negative Chinese hamster ovary cells will be transfected with the vector. The transfected cells will be subjected to methotrexate selection to obtain clones that produce high levels of human PAF-AH due to gene amplification.

Example 9

Recombinant human plasma PAF-AH (start at Ile42) expressed in E. coli was purified to a single Coomassie-stained SDS-PAGE band using various methods and assayed for activities exhibited by the native PAF-AH enzyme.

A. Purification of recombinant PAF-AH

The first purification procedure used is similar to that described in Example 1 for native PAF-AH. The following steps were performed at 4 °C. Pellets from 50 ml of PAF-AH producing E. coli (transformed with expression construct trp AH) were lysed as described in Example 8. Solids were removed by centrifugation at 10,000 g for 20 minutes. The supernatant was applied at 0.8 ml/min to a Blue Sepharose Fast Flow column (2.5 cm x 4 cm; 20 ml material volume) equilibrated in buffer D (25 mM Tris-HCl, 10 mM CHAPS, 0.5 M NaCl, pH 7.5). The column was washed with 100 ml of buffer D and eluted with 100 ml of buffer A containing 0.5 M KSCN at 3.2 ml/min. A 15 ml active fraction was applied to a 1 ml Cu-chelating Sepharose column equilibrated in buffer D. The column was washed with 5 ml buffer D, followed by elution with 5 ml buffer D containing 100 mM imidazole, with gravity flow. Fractions containing PAF-AH activity were analyzed by SDS-PAGE.

The results from the purification are shown in table 6, in which one unit corresponds to moles of PAF hydrolysis per hour. The purified product obtained at 4 °C appeared on SDS-PAGE as a single intense band below the 43 kDa marker, with some diffuse staining directly above and below the band. The recombinant material is significantly purer and exhibits higher specific activity compared to PAF-AH preparations from plasma, as described in Example 1.

When the same purification protocol was performed at ambient temperature, in addition to the band below the 43 kDa marker, a group of bands below the 29 kDa marker was observed that correlated with PAF-AH activity in analyzed gel pieces. These lower molecular weight bands may be proteolytic fragments of PAF-AH that retain enzymatic activity.

A different purification procedure was also performed at ambient temperature. Pellets (100 g) of PAF-AH-producing E. coli (transformed with the expression construct pUC trp AH) were resuspended in 200 ml lysis buffer, 25 mM Tris, 20 mM CHAPS, 50 mM NaCl, 1 mM EDTA, 50 g/ ml benz-amidine, pH 7.5) and lysed by passing three times through a microfluidizer at 1054.5 kg/cm<2>. Solids were removed by centrifugation at 14,300 x g for 1 hour. The supernatant was diluted 10-fold in dilution buffer [25 mM MES (2-[N-morpholino]ethanesulfonic acid), 10 mM CHAPS, 1 mM EDTA, pH 4.9] and applied at 25 ml/min to an S-Sepharose Fast Flow- column (200 ml) (a cation exchange column) equilibrated in buffer E (25 mM MES 10 mM CHAPS, 1 mM EDTA, 50 mM NaCl, pH 5.5). The column was washed with 1 L of buffer E, eluted with 1 M NaCl, and the eluate was collected in 50 ml fractions and adjusted to pH 7.5 with 0.5 ml of 2 M Tris base. Fractions containing PAF-AH activity were pooled and adjusted to 0.5 M NaCl. The S portion was applied at 1 ml/min to a Blue Sepharose Fast Flow column (2.5 cm x 4 cm; 20 ml) equilibrated in buffer F (25 mM Tris, 10 mM CHAPS, 0.5 M NaCl, 1 mM EDTA, pH 7.5). The column was washed with 100 ml of buffer F and eluted with 100 ml of buffer F containing 3 M NaCl at 4 ml/minute. The Blue Sepharose Fast Flow chromatography step was then repeated to reduce endotoxin levels in the sample. Fractions containing PAF-AH activity were pooled and dialyzed against buffer G (25 mM Tris, pH 7.5, 0.5 M NaCl, 0.1% Tween 80, 1 mM EDTA).

The results of the purification are shown in table 7, in which one unit equals moles of PAF hydrolysis per hour.

The purified product obtained appeared by SDS-PAGE as a single intense band below the 43 kDa marker, with some diffuse staining directly above and below the band. The recombinant material is significantly purer and exhibits higher specific activity compared to PAF-AH preparations from plasma, as described in Example 1.

A further purification procedure according to the invention involves cell lysis, clarification and first column step. Cells are diluted 1:1 in lysis buffer (25 mM Tris, pH 7.5,

150 mM NaCl, 1% Tween 80, 2 mM EDTA). Lysis is performed in a cooled microfluidizer at 1054.5-1406 kg/cm<2> with three passes of the material to give >99% cell degradation. The lysate is diluted 1:20 in dilution buffer (25 mM Tris, pH 8.5, 1 mM EDTA) and applied to a column packed with Q-Sepharose Big Bead chromatography medium (Pharmacia) and equilibrated in 25 mM Tris, pH 8.5, 1 mM EDTA, 0.015% Tween 80. The eluate is diluted 1:10 in 25 mM 'MES, pH 5.5, 1.2 M ammonium sulfate, 1 mM EDTA, and applied to butyl-Sepharose chromatography medium (Pharmacia) equilibrated in the same buffer. PAF-AH activity is eluted in 25 mM MES, pH 5.5, 0.1% Tween 80, 1 mM EDTA.

A further method according to the invention for purifying enzymatically active PAF-AH from E. coli includes the steps: (a) preparing an E. coli extract which gives dissolved PAF-AH supernatant after lysis in a buffer containing CHAPS; (b) dilution of the supernatant and application to an anion exchange column equilibrated at approx. pH 8.0; (c) elution of PAF-AH enzyme from the anion exchange column;

(d) applying the adjusted eluate from the anion exchange column to a blue dye ligand affinity column; (e) elution of the blue dye ligand affinity column using a buffer comprising 3.0 M salt; (f) diluting the blue dye eluate in a suitable buffer for performing hydroxylapatite chromatograph i; (g) performing hydroxylapatite chromatography in which washing and elution are carried out using buffers (with or without CHAPS; (h) diluting the hydroxylapatite eluate to a suitable salt concentration for cation exchange chromatography; (i) applying the diluted hydroxylapatite eluate to a cation exchange column at a pH that varies

between approx. 6.0 and 7.0; (j) elution of PAF-AH from the cation exchange column with a suitable formulation buffer; (k) performing cold state cation exchange chromatography; and (1) formulation of PAF-AH in liquid or frozen form in the absence of CHAPS.

In step (a) above, the lysis buffer is preferably 25 mM Tris, 100 mM NaCl, 1 mM EDTA, 20 mM CHAPS, pH 8.0; in step (b), the dilution of the supernatant for anion exchange chromatography is 3-4 times in 25 mM Tris, 1 mM EDTA, 10 mM CHAPS, pH 8.0, and the column is a Q-Sepharose column equilibrated with 25 mM Tris, 1 mM EDTA, 50 mM NaCl, 10 mM CHAPS, pH 8.0; in step (c), the anion exchange column is eluted using 25 mM Tris, 1 mM EDTA, 350 mM NaCl, 10 mM CHAPS, pH 8.0; in step (d) the eluate from step (c) is applied directly to a blue dye affinity column; in step (e), the column is eluted with 3 M NaCl, 10 mM CHAPS, 25 mM Tris, pH 8.0 buffer; in step (f), dilution of the blue color eluate for hydroxylapatite chromatography is carried out by dilution in 10 mM sodium phosphate, 100 mM NaCl, 10 mM CHAPS, pH 6.2; in step (g), hydroxylapatite chromatography is performed using a hydroxylapatite column equilibrated with 10 mM sodium phosphate, 100 mM NaCl, 10 mM CHAPS, and the elution is performed using 50 mM sodium phosphate, 100 mM NaCl (with or without 10 mM CHAPS), pH 7 .5; in step (h), dilution of the hydroxylapatite elute for cation exchange chromatography is carried out by dilution in a buffer at a pH varying from approx. 6.0 to 7.0, and comprising sodium phosphate (with or without CHAPS); in step (i) an S-Sepharose column is equilibrated with 50 mM sodium phosphate (with or without 10 mM CHAPS), pH 6.8; in step (j), elution is performed with an appropriate formulation buffer, such as potassium phosphate, 50 mM, 12.5 mM aspartic acid, 125 mM NaCl, pH 7.5, containing 0.01% Tween 80; and in step (k) cation exchange chromatography is performed at 2-8 °C. Examples of suitable formulation buffers for use in step (1) that stabilize PAF-AH include 50 mM potassium phosphate, 12.5 mM aspartic acid, 125 mM NaCl, pH 7.4 (approximately, with and without the addition of Tween 80 and/or Pluronic F68) or 25 mM potassium phosphate buffer containing

(minimum) 125 mM NaCl, 25 mM arginine and 0.1% Tween 80 (with or without Pluronic F68 at approximately 0.1 and 0.5%).

B. Activity of recombinant PAF-AH

The most notable property of PAF acetylhydrolase is its marked specificity for substrates with a short residue in the sn-2 position of the substrate. This strict specificity distinguishes PAF acetyl hydrolase from other forms of PLA2. Thus, to determine whether recombinant PAF-AH degrades phospholipids with long-chain fatty acids in the sn-2 position, hydrolysis of l-palmitoyl-2-arachidonyl-sn-glycero-3-phosphocholine (arachidonoylPC) was investigated, because this is the preferred substrate for a well-characterized form of PLA2. As predicted from previous studies with native PAF-AH, this phospholipid was not hydrolyzed when incubated with recombinant PAF-AH. In additional experiments, arachidonoylPC was included in a standard PAF hydrolysis assay at concentrations ranging from 0 to 125 M to determine whether it inhibited the hydrolysis of PAF by recombinant PAF-AH. There was no inhibition of PAF hydrolysis, even at the highest concentration of PAF-AH, which was 5 times higher than the concentration of PAF. Recombinant PAF-AH thus exhibits the same substrate selectivity as the native enzyme; long-chain substrates are not recognized. Recombinant PAF-AH enzyme also rapidly degraded an oxidized phospholipid (glutaroylPC) that had undergone oxidative cleavage of the sn-2 fatty acid. Native plasma PAF-AH has several other properties that distinguish it from other phospholipases, including calcium independence and resistance to compounds that modify sulfhydryl groups or degrade disulfides.

Both the native and recombinant plasma PAF-AH enzymes are sensitive to DFP, indicating that et

serine forms part of their active sites. An unusual feature of the native plasma PAF acetylhydrolase is that it is tightly associated with circulating lipoproteins, and its catalytic efficiency is influenced by the lipoprotein environment. When recombinant PAF-AH according to the invention was incubated with human plasma (pretreated with DFP to abolish the endogenous enzyme activity), it associated with low- and high-density

tetslipoproteins in the same way as the native activity. This result is significant because there is substantial evidence that modification of low-density lipoproteins is essential for the cholesterol deposition observed in atheromas, and that oxidation of lipids is an initiating factor in this process. PAF-AH protects low-density lipoproteins from modification under oxidizing conditions in vitro, and may have such a role in vivo. Administration of PAF-AH is thus indicated for suppression of the oxidation of lipoproteins in atherosclerotic plaques, as well as for removing inflammation.

All these results confirm that the cDNA clone sAH 406-3 encodes a protein with the same activities as the human plasma PAF acetylhydrolase.

Example 10

Various other recombinant PAF-AH products were expressed in E. coli. The products included PAF-AH analogs with single amino acid mutations, and PAF-AH fragments.

A. PAF-AH- amino acid substitution products

PAF-AH is a lipase because it hydrolyzes the phospholipid PAF. Although no obvious overall similarity exists between PAF-AH and other characterized lipases, conserved residues have been found in comparisons between structurally characterized lipases. A serine has been identified as a member of the active site. The serine, together with an aspartate residue and a histidine residue, form a catalytic triad that represents the lipase's active site. The three residues are not contiguous in the primary protein sequence, but structural investigations have shown that the three residues are contiguous in three-dimensional space. Comparisons between structures of mammalian lipases suggest that the aspartate residue is generally located 24 amino acids C-terminal to the active seteserine. In addition to this, the histidine is usually 109-111 amino acids C-terminal to the active seteserine.

By site-directed mutagenesis and PCR, individual codons in the human PAF-AH coding sequence were modified to encode alanine residues and were expressed in E. coli. As shown in table 8, where e.g. the abbreviation "S108A" indicates that the serine residue at position 108 was changed to alanine, point mutations of Ser273, A<sp>296 or His351 completely destroy PAF-AH activity. The distances between active site residues are similar for PAF-AH (Ser to Asp, 23 amino acids;

Sees His, 78 amino acids) and other lipases. These experiments show that Ser273, As<p>296 and His351 are critical residues for activity, and are therefore likely candidates for catalytic triad residues. Cysteines are often critical to the functional integrity of proteins due to their capacity to form disulfide bonds. The plasma PAF-AH enzyme contains five cysteines. To determine whether any of the five are critical for enzyme activity, each cysteine was mutated individually to serine, and the resulting mutants were expressed in E. coli. Preliminary activity results using partially purified preparations of these recombinantly produced mutants are shown below in the second column of Table 8, while results using more purified preparations are shown below in the third column of Table 8. The data show that all the cysteine mutants had essentially equivalent activity, so neither cysteine appears to be required for PAF-AH activity. Other point mutations also had little or no effect on PAF-AH catalytic activity. In Table 8, "++++" represents we 11 type-PAF-AH activity of approx. 40-60 U/ml/OD600, "+++" represents approx. 20-40 U/ml/OD600 activity, "++" represents approx. 10-20 U/ml/OD600 activity, " + " represents 1-10 U/ml/OD600 activity, and "-" represents < 1 U/ml/OD600 activity.

B. PAF-AH fragment products

C-terminal deletions were prepared by cleaving the 3' end of the PAF-AH coding sequence with exon nuclease III for different times, and then ligating the shortened coding sequence to plasmid DNA coding stop codons in all three reading frames. Ten different deletion constructs were characterized by DNA sequence analysis, protein expression and PAF-AH activity. Removal of 21-30 C-terminal amino acids greatly reduced catalytic activity, and removal of 52 residues completely abolished activity. See Figure 3.

Similar deletions were made at the amino-terminal end of PAF-AH. Fusions of PAF-AH with E. coli thio-redoxin at the N-terminus were prepared to facilitate consistent high-level expression of PAF-AH activity [LaVallie et al., Bio/technology, 11:187-193 (1993) ]. Removal of 19 amino acids from the naturally processed N-terminus (Ile42) reduced the activity by 99%, while removal of 26 amino acids completely destroyed enzymatic activity in the fusion protein. See figure 3. Deletion of 12 amino acids seems to increase enzyme activity approx. four times.

In subsequent purifications of PAF-AH from fresh human plasma by a method similar to that described in Example 1 (Microcon 30 filter from Amicon was used to concentrate Blue Sepharose eluate instead of a Cu column), two N- terminals in addition to Ile42, Ser35 and Lys55. The heterogeneity may be the natural state of the enzyme in plasma, or it may arise during purification.

The purified material described above was also analyzed for glycosylation. Purified native PAF-AH was incubated in the presence or absence of N-glycanase, an enzyme that removes N-linked carbohydrates from glycoproteins. The treated PAF-AH samples were electrophoresed through a 12% SDS-polyacrylamide gel and then visualized by Western blotting using rabbit polyclonal antisera. Protein that was not treated with N-glycanase migrated as a diffuse band of 45-50 kDa, while the protein treated with the glycanase migrated as a dense band of approx. 44 kDa, showing that native PAF-AH is glycosylated.

N-terminal heterogeneity was also observed in purified preparations of recombinant PAF-AH (Ile42-N-terminal). These preparations were a mixture of polypeptides with N-termini starting at Ala47, Ile42 or the artificial initiation site Met.x adjacent to Ile42.

1. Initial comparison of PAF-AH fragments

with PAF-AH

Considering the observed heterogeneity of recombinantly produced PAF-AH, other recombinant products were prepared and tested for homogeneity after recombinant expression and purification. The preparation of the recombinant expression products of pBAR2/PH.2 and pBAR2/PH.9 in E. coli strain MC1061 was analyzed at different times during the production phase of cell fermentation. Partially purified samples of the recombinant PH.2 and PH.9 from cells collected at time points varying between 5 and 22 hours after induction of protein expression were analyzed by matrix-assisted laser desorption ionization mass spectrometry (MALDI-MS).

When the PH.2 expression vector was used, two peaks were observed in the spectrum for the partially purified protein at a mass value expected for rPAF-AH protein. Two peaks were observed at all time points, and higher heterogeneity was observed at time points when fermentation is stressed, as indicated by an accumulation of acetate and/or a depletion of oxygen in the medium. The accuracy of the MALDI-MS technique in this mass range was approx. ± 0.3%, which is approximately the mass of one amino acid. The higher mass peak observed was consistent with the presence of the expected unabridged translation product for the PH.2 vector, minus the translation initiating methionine which is expected to be removed after translation. The lower mass peak was approx. 1200 atomic masses less.

When the PH.9 expression vector was used, a single peak dominated the spectrum for the partially purified protein at a mass value expected for rPAF-AH protein. This single peak was observed at all time points, and no increase in heterogeneity was observed at different time points. The observed mass of this protein was consistent with the presence of the expected unabridged translation product for the PH.9 vector minus the initial methionine.

2. Purification of PAF-AH fragments Recombinantly expressed rPH.2 preparations (expression product of DNA encoding Met46-Asn441) and rPH9 preparations (expression product of DNA encoding Met46-Ile429) were purified for further comparison with purified r PAF-AH (expression product of DNA encoding Ile42-Asn441). rPH.9 was produced using E. coli strain SB7219 and generally purified according to the zinc chelate purification procedure described above, while rPH.2 was produced by E. coli strain MC1061 and purified as described below. The transformed cells were lysed by diluting the cell paste with lysis buffer (100 mM succinate, 100 mM NaCl, 20 mM CHAPS., pH 6.0). The slurry was mixed and lysed using high pressure digestion. The lysed cells were centrifuged, and the supernatant containing rPH.2 was retained. The clarified supernatant was diluted five times in 25 mM sodium phosphate buffer containing 1 mM EDTA, 10 mM CHAPS, pH 7.0. The diluted supernatant was then applied to the Q-Sepharose column. The column was first washed with 3 column volumes of 25 mM sodium phosphate buffer containing 1 mM EDTA, 50 mM NaCl, 10 mM CHAPS, pH 7.0 (wash 1), then washed with 10 column volumes of 25 mM Tris buffer containing 1 mM EDTA, 10 mM CHAPS , pH 8.0 (wash 2), and with 10 column volumes of 25 mM Tris buffer containing 1 mM EDTA, 100 mM NaCl, 10 mM CHAPS, pH 8.0 (wash 3). The elution was

performed with 25 mM Tris buffer containing 1 mM EDTA, 350 mM NaCl, 10 mM CHAPS, pH 8.0. The Q-Sepharose eluate was diluted three times in 25 mM Tris, 1 mM EDTA, 10 mM CHAPS, pH 8.0, and then applied to a Blue Sepharose column. The column was first washed with 10 column volumes of 25 mM Tris, 1 mM EDTA, 10 mM CHAPS, pH 8.0. The column was then washed with 3 column volumes of 25 mM Tris, 0.5 M NaCl, 10 mM CHAPS, pH 8.0. Elution was performed with 25 mM Tris, 3.0 M NaCl, 10 mM CHAPS, pH 8.0. The Blue Sepharose eluate was diluted five times in 10 mM sodium phosphate, 10 mM CHAPS, pH 6.2, and then applied to the chromatography column. The column was washed with 10 column volumes of 10 mM sodium phosphate, 100 mM NaCl, 0.1% Pluronic F68, pH 6.2. rPH.2 was eluted with 120 mM sodium phosphate, 100 mM NaCl, 0.1% Pluronic F68, pH 7.5. The hydroxyapatite telate was diluted sixfold with 10 mM sodium phosphate, 0.1% Pluronic F68, pH 6.8. The diluted hydroxyapatite teluate was adjusted to pH 6.8 using 0.5 N succinic acid and then applied to an SP-Sepharose column. The SP-Sepharose column was washed with 10 column volumes of 50 mM sodium phosphate, 0.1% Pluronic F68, pH 6.8, and eluted with 50 mM sodium phosphate, 125 mM NaCl, 0.1% Pluronic F68, pH 7.5. The eluted rPH.2 was formulated by diluting to a final concentration of 4 mg/ml in 50 mM sodium phosphate, 125 mM NaCl, 0.15% Pluronic F68, pH 7.5, and Tween 80 was

added to a final concentration of 0.02% Tween 80. The formulated product was then filtered through a 0.2 membrane and stored before use.

3. Comparison of PAF-AH fragments with PAF-AH

by sequencing

The purified rPH.2 and rPH.9 preparations were compared to purified rPAF-AH preparations by N-terminal sequencing using an Applied Biosystems model 473A protein sequencer (Applied Biosystems, Foster City, CA) and by of C-terminal sequencing using a Hewlett-Packard model G1009A C-terminal protein sequencer. The rPH2 preparation had less N-terminal heterogeneity compared to rPAF-AH. The N-terminal analysis of the rPH.9 preparation was similar to the analysis of rPH.2, but less C-terminal heterogeneity was observed for the rPH.9 preparation compared to rPH.2.

The purified rPH.2 preparation contained a major sequence where the N-terminus was Ala47 (approx. 86-89%) and a minor sequence where the N-terminus was Ala48 (approx. 11-14%), the ratio between the two N -terminals was fairly constant under different fermentation conditions. The purified rPH.9 preparation also contained a major sequence where the N-terminus was Ala47 (about 83-90%) and a minor sequence where the N-terminus was Ala4B (10-17%). In contrast, attempts in bacteria to produce the polypeptide starting with Ile42 (rPAF-AH) resulted in a variable mixture of polypeptides with N-termini starting with Ala47 (20-53%), Ile42 (8-10%) or with the artificial initial Met^-methionine (37-72%) adjacent to Ile42. For rPH.2 and rPH.9, the initial methionine is effectively removed by an amino-terminal peptidase after bacterial synthesis of the polypeptide, leaving the alanine at position 47 (or the alanine at position 48) as the N-terminal residue.

C-terminal sequencing was performed on one lot of rPH.2, which was observed to have a major sequence in which the C-terminus was HOOC-Asn-Tyr (about 80%), which was consistent with the predicted H00C-Asn441-Tyr440- C-terminus of the translation product, while approx. 20% were HOOC-Leu. After the rPH.2 preparation was fractionated by SDS-PAGE, further sequencing of the primary and secondary bands yielded a C-terminal sequence of HOOC-Leu-Met from a lower secondary band (AHL, described below in section B .5), which is consistent with a product that is 10 amino acids shorter than the unabridged translation product, as well as low levels of HOOC-His. Further peptide mapping has shown that additional C termini are present in some portions of PH.2 protein. The C-terminus of rPH.9 was primarily HOOC-Ile-His (about 78-91%, depending on the lot) by direct sequencing, which is consistent with the predicted HOOC-Ile429-His428-C-terminus of the translation product. There appears to be some background ("noise") in this technique, so low levels of other sequences cannot be ruled out.

4. Comparison of PAF-AH fragments with PAF-AH

using MALDI-MS

MALDI-MS was performed on purified rPH.2 and rPH.9 preparations. The rPH.2 spectrum exhibited two peaks in the spectrum at a mass value expected for the rPAF-AH product (see Figure 4) similar to the pattern observed with the partially purified protein in section B.l above. The lower molecular weight secondary peak was typically present at ca. 20-30% of the total quantity. The rPH.9 spectrum showed a dominant peak at a mass consistent with that expected for the truncated translation product of the rPH.9 vector, minus the translation initiating methionine (see Figure 5). A shoulder peak with a slightly lower molecular weight was also observed for rPH.9, which represented approx. 5% of the total.

5. Comparison of PAF-AH fragments with PAF-AH

using SDS-PAGE

Sodium dodecyl sulfate-polyacrylamide electrophoresis (SDS-PAGE) was performed on purified rPAF-AH, rPH.2, and rPH.9 preparations. A complex pattern of bands was observed for rPH.2 around the electrophoretic migration range expected for the rPAF-AH product, based on protein molecular weight standards. In one or some gels two dominant bands were observed with easily observable secondary bands above and below the primary band. These upper secondary, middle primary and lower secondary bands were respectively designated AHUf AHM and AHL. All of these bands reacted with an anti-rPAF-AH monoclonal antibody on Western blot, and have thus been identified as rPAF-AH-related products. The upper secondary band, AH,, increased in intensity over time with storage of the protein, and presumably represents a modified form of the rPAF-AH product. SDS-PAGE of the rPAF-AH preparation is similar to that of rPH.2. There are two major bands migrating close to the expected molecular weight for rPAF-AH, as well as a minor band above and a shadow below the major bands. In contrast, rPH.9 displayed a single dominant band on SDS-PAGE with no apparent cleavage. Weak bands with a slightly lower molecular weight and in an expected dimer position were also observed. No AH„-like band was observed.

The composition of the purified rPH.2 and rPH.9 preparations was also analyzed on 2D gels (isoelectric focusing (IEF) in urea, followed by SDS-PAGE in the second dimension). For rPH9, the 2D gels showed 5 main spots separated in the IEF direction. The charge heterogeneity appeared to be constant between batches of rpH9. In contrast, the 2D gel pattern of rPH.2 was more complicated because it contained approx. 15 spots separated in the IEF and SDS-PAGE dimensions.

6. Comparison of activity of PAF-AH fragments with PAF-AH

Purified rPH.2 and rPH.9 have enzymatic activity indistinguishable from that of endogenous PAF-AH purified from serum, and rPH.2 and rPH.9 bind to lipoprotein in a similar manner to purified endogenous PAF-AH.

Example 11

A preliminary analysis of the expression pattern of human plasma PAF-AH mRNA in human tissues was performed by Northern blot hybridization.

RNA was prepared from human cerebral cortex, heart, kidney, placenta, thymus, and tonsil using RNA Stat 60 (Tel-Test "B", Friendswood, TX). RNA was also prepared from the human hematopoietic progenitor-like cell line THP-1 (ATCC TIB 202), which was also induced to differentiate into a macrophage-like phenotype using the phorbol ester phorbol myristyl acetate (PMA). Tissue RNA and RNA prepared from the premyelocytic THP-1 cell line before and 1-3 days after induction were electrophoresed through a 1.2% agarose-formaldehyde gel and then transferred to a nitrocellulose membrane. The untruncated human plasma PAF-AH cDNA, sAH 406-3, was labeled by random priming and hybridized to the membrane under conditions identical to those described in Example 3 for library screening. Initial results indicate that the PAF-AH probe hybridized to a 1.8 kb band in thymus, tonsil and to a lesser extent placental RNA.

PAF is synthesized in the brain under normal physiological and pathophysiological conditions. Given the molecule's known proinflammatory and potential neurotoxic properties, a mechanism for the localization of PAF synthesis or for its rapid catabolism would be expected to be critical for nervous tissue health. The presence of PAF acetylhydrolase in nervous tissue is consistent with it playing such a protective role. It is interesting that both a bovine heterotrimeric, intracellular PAF-AH [the cloning of which is described by Hattori et al., J. Biol. Chem., 269(37):23150-23155 (1994)] and PAF-AH according to the invention have been identified in the brain. To determine whether the two enzymes are expressed in similar or different regions of the brain, the human homologue of the bovine brain intracellular PAF-AH cDNA was cloned, and its mRNA expression pattern in the brain was compared by Northern blotting with mRNA- the expression pattern for PAF-AH according to the invention, using substantially the same methods as described in the previous section. The areas of the brain examined using Northern blotting were the cerebellum, medulla, spinal cord, putamen, amygdala, caudate nucleus, thalamus and the occipital, frontal and temporal lobes of the cerebral cortex. Indications for both enzymes were detected in all these tissues, although the heterotrimeric intracellular form was more abundant than the secreted form. Northern blot analysis of additional tissues further revealed that the heterotrimeric, intracellular form is expressed in many different tissues and cells, including thymus, prostate, testis, ovary, small intestine, colon, peripheral blood leukocytes, macrophages, brain, liver, skeletal muscle, kidney, pancreas and adrenal gland. This ubiquitous expression suggests that the heterotrimeric, intracellular PAF-AH has a general housekeeping function in cells.

The expression of PAF-AH-RNA in monocytes isolated from human blood and during their spontaneous differentiation into macrophages in culture was also investigated. Little or no RNA was detected in healthy monocytes, but expression was induced and maintained during differentiation into macrophages. There was a concomitant accumulation of PAF-AH activity in the culture medium of the differentiating cells. Expression of the human plasma PAF-AH transcript was also observed in THP-1 cell RNA at 1 day, but not 3 days, after induction. THP-1 cells did not express mRNA for PAF-AH in the basal state.

Example 12

PAF-AH expression in human and mouse tissues was examined by in situ hybridization.

Human tissues were obtained from the National Disease Research Interchange and the Cooperative Human Tissue Network. Normal mouse brain and spinal cord and EAE stage-3 mouse spinal cords were harvested from S/JLJ mice. Normal S/JLJ mouse embryos were harvested from 11 to 18 days post fertilization.

The tissue sections were placed in Tissue Tek II cryoforms (Miles Laboratories, Inc., Naperville, IL) with a small amount of OCT compound (Miles, Inc., Elkhart, IN). They were centered in the cryoform, the cryoform was filled with OCT compound and then placed in a container of 2-methyl-butane [C2H5CH(CH3)2, Aldrich Chemical Company, Inc., Milwaukee, WI], and the container was placed in liquid nitrogen. Once the tissue and OCT compound in the cryoform were frozen, the blots were stored at -80°C until sectioned. The tissue blocks were cut at a thickness of 6 µm and mounted on Vectabond (Veetor Laboratories, Inc., Burlingame, CA) coated glass slides and stored at -70 °C and placed at 50 °C for approximately 5 min for heating and removal of condensation, and then fixed in 4% paraformaldehyde for 20 min at 4°C, dehydrated (70%, 95%, 100% ethanol) for 1 min at 4°C in each step and then air dried for 30 min at room temperature. Sections were denatured for 2 min at 70°C in 70% formamide/2 x SSC, rinsed twice in 2 x SSC, dehydrated, and then air-dried for 30 min. The tissues were hybridized in situ with radiolabeled, single-stranded mRNA generated from DNA derived from an internal 1 kb HindiII fragment of the PAF-AH gene (nucleotides 308-1323 of SEQ ID NO:7) by in vitro RNA transcription incorporation of 35S-UTP (Amersham) or from DNA derived from the heterotrimeric, intracellular PAF-AH cDNA identified by Hattori et al. The probes were used in different lengths from 2 50 to 500 bp. Hybridization was performed overnight (12-16 hours) at 50 °C; the <35>S-labeled riboprobes (6 x 10<5> cpm/section), tRNA (0.5 (µg/section) and diethyl pyrocarbonate (depc)-treated water were added to the hybridization buffer to bring it to a final concentration of 50 % formamide, 0.3 M NaCl, 20 mM Tris, pH 7.5, 10% dextran sulfate, 1x Denhardt's solution, 100 mM dithiothreitol (DTT), and 5 mM EDTA.After hybridization, the sections were washed for 1 h at room temperature for 4 x SSC/10 mM DTT, then for 40 min at 60 °C in 50% formamide/1 x SSC/10 mM DTT, for 30 min at room temperature in 2 x SSC and for 30 min at room temperature in 0.1 x SSC. Sections were dehydrated, air-dried for 2 hours, coated with Kodak NTB2 photographic emulsion, air-dried for 2 hours, developed (after storage at 4°C in complete darkness) and counterstained with hematoxylin/eosin.

A. The brains

Cerebellum. In both mouse brains and human brains, a strong signal was observed in the Purkinje cell layer of the cerebellum, in basket cells and individual neuronal cell bodies in the dentate nucleus (one of the four deep nuclei in the cerebellum). Message for the heterotrimeric, intracellular PAF-AH was also observed in these cell types. Plasma PAF-AH signal was also observed on individual cells in the granular and molecular layers of the gray matter.

Hippocampus. In the human hippocampal sections, individual cells throughout the section, which appeared to be neuronal cell bodies, showed a strong signal. These were identified as polymorphic cell bodies and granule cells. Message for the heterotrimeric, intracellular PAF-AH was also observed in the hippocampus.

Brain stem. In sections from both human and mouse brainstem, there was a strong signal on individual cells in the gray matter.

Cortex. In human cortical sections taken from the cerebral, occipital and temporal cortices and in sections from whole mouse brains, individual cells throughout the cortex showed a strong signal. These cells were identified as pyramidal, stellate and polymorphic cell bodies. There does not seem to be differentiation in the expression pattern in the different layers of the cortex. These in situ hybridization results differ from the cerebral cortex results obtained by Northern blotting. The difference probably results from the higher sensitivity of in situ hybridization compared to Northern blotting. As in the cerebellum and hippocampus, a similar pattern of expression of the heterotrimeric intracellular was observed

PAF-AH.

Pituitary gland. A somewhat weak signal was observed on scattered individual cells in the pars distalis on the human tissue section.

B. Human colon

Both normal colons and colons with Crohn's disease showed signal in the lymphatic aggregates present in the mucosa in the sections, the signal level being slightly higher in the section from the patient with Crohn's disease. The colon with Crohn's disease also had a strong signal in the lamina propria. A high signal level was likewise observed in a disease-affected appendix section, while the normal appendix showed a lower, but still detectable, signal. The sections from the patient with ulcerative colitis showed no clear signal in either the lymphatic aggregates or in the lamina propria.

C. Human tonsil and thymus

A strong signal was observed on scattered groups of individual cells in the germinal centers of the tonsil and in the thymus.

D. Human lymph node

Strong signal was observed on the lymph node section taken from a normal donor, while a somewhat weak signal was observed in the lymph nodes on the section from a donor with septic shock.

E. Human small intestine

Both normal small intestine and small intestine with Crohn's disease had a weak signal in Peyer's patches and lamina propria in the sections, with the signal on the affected tissue being slightly higher.

F. Human spleen and lung

No signal was observed on any of the spleen tissues (normal section and section with splenic abscess) or lung tissue (normal section and section with emphysema).

G. Mouse spinal cord

In both the normal and EAE stage 3 spinal cords, there was a strong signal in the gray matter of the spinal cord, the expression being slightly higher in the EAE stage 3 spinal cord. In the EAE stage-3 spinal cord, cells in the white matter and the perivascular cuffs showed presumably infiltrating macrophages and/or leukocytes, a signal that was absent in the normal spinal cord.

F. Mouse embryos

In 11-day-old embryos, there was clear central nervous system signal in the fourth ventricle that remained constant throughout the embryonic time course as it developed into the cerebellum and brainstem. As the embryos matured, signal became evident in the central nervous system in the spinal cord (day 12), primary cortex and ganglion Gasseri (day 14), and pituitary (day 16). Signal was observed in the peripheral nervous system (starting on day 14 or 15) on nerves exiting the spinal cord, and on day 17 a strong signal appeared around the embryo's whiskers. Expression was also observed in the liver and lung on day 14, the intestine (starting on day 15) and in the posterior part of the mouth/pharynx (starting on day 16). On day 18, the expression pattern had differentiated into signal in the cortex, hindbrain (cerebellum and brainstem), nerves that exit from the lumbar region to the spinal cord, the posterior part of the mouth/pharynx, the liver, the kidney and possibly a weak signal in the lung and intestine .

G. Summary

PAF-AH mRNA expression in tonsil, thymus, lymph node, Peyer's patches, appendix, and colonic lymphatic aggregates is consistent with the conclusions that the presumably predominant in vivo source of PAF-AH is macrophages, because these tissues are all populated with tissue macrophages that serve such as phagocytic and antigen processing cells.

Expression of PAF-AH in inflamed tissues would be consistent with the hypothesis that one of the roles of monocyte-derived macrophages is to reduce inflammation. PAF-AH would be expected to inactivate PAF and the proinflammatory phospholipids and thus downregulate the inflammatory cascade of events initiated by these mediators.

PAF has been detected in brain tissue and is secreted by rat cerebellum granule cells in culture. Experiments in vitro and in vivo have shown that PAF binds a specific receptor in nervous tissue and induces functional and phenotypic changes, such as calcium mobilization, up-regulation of transcription-activating genes and differentiation of the neural precursor cell line PC12. These observations suggested a physiological role for PAF in the brain, and, consistent with this, recent experiments using hippocampal tissue cultures and PAF analogs and antagonists have implicated PAF as an important retrograde messenger in hippocampal long-term potentiation. In addition to its pathological effect in inflammation, PAF therefore appears to participate in routine neuronal signaling processes. Expression of the extracellular PAF-AH in the brain may serve to regulate the duration and magnitude of PAF-mediated signaling.

Example 13

Monoclonal antibodies specific for recombinant human plasma PAF-AH were raised using E. coli-produced PAF-AH as an immunogen.

Mice #1342 were injected on day 0, day 19 and day 40 with recombinant PAF-AH. The mouse was first injected with the immunogen in PBS, 4 days later the mouse was euthanized, and the spleen was removed sterilely and placed in 10 ml of serum-free RPMI 1640. A single-cell suspension was formed by rubbing the spleen between the mat-ground ends of two microscope slides immersed in serum-free RPMI 1640. , supplemented with 2 mM L-glutamine, 1 mM sodium pyruvate, 100 units/ml penicillin and 100 µg/ml streptomycin (RPMI) (Gibco, Canada). The cell suspension was filtered through 70 mesh Nitex cell filter (Becton Dickinson, Parsippany, New Jersey) and washed twice by centrifugation at 200 g for 5 minutes and resuspending the pellet in 20 ml of serum-free RPMI. Thymocytes taken from three normal Balb/c mice were prepared in a similar manner. NS-1 myeloma cells, maintained in log phase in RPMI with 10% fetal bovine serum (FBS) (Hyclone Laboratories, Inc., Logan, Utah) for 3 days before fusion, were centrifuged at 200 g for 5 min, and the pellet was washed twice as described in the previous section.

1 x 10<8> spleen cells were combined with 2.0 x 10<7> NS-1 cells, centrifuged, and the supernatant was aspirated. The cell pellet was removed by tapping the tube, and 1 ml of 37 °C PEG 1500 (50% in 75 mM Hepes, pH 8.0) (Boehringer Mannheim) was added with stirring over 1 min, followed by the addition of 7 ml of serum-free RPMI during 7 minutes. An additional 8 ml of RPMI was added, and the cells were centrifuged at 200 g for 10 min. After discarding the supernatant, the pellet was resuspended in 200 ml RPMI containing 15% FBS, 10 (IM sodium hypoxanthine, 0.4 µM aminopterin, 16

\ iM thymidine (HAT) (Gibco), 25 units/ml IL-6 (Boehringer Mannheim) and 1.5 x 10<6> thymocytes/ml, and spread on 10 Corning 96-well flat-bottom tissue culture plates (Corning, Corning New York).

On days 2, 4 and 6 after fusion, 100 µl medium was removed from the wells of the fusion plates and replaced with fresh medium. On day 8, the fusion was screened by ELISA testing for the presence of mouse IgG binding to recombinant PAF-AH. Immulon 4 plates (Dynatech, Cambridge, MA) were coated for 2 h at 37°C with 100 ng/well recombinant PAF-AH diluted in 25 mM Tris, pH 7.5. The superimposed solution was aspirated, and 200 µl/well blocking solution [0.5% fish skin gelatin (Sigma) diluted in CMF-PBS] was added and incubated for 30 minutes at 37°C. The plates were washed three times with PBS with 0.05% Tween 20 (PBST), and 50 µl of culture supernatant was added. After incubation at 37°C for 30 min and washing as indicated above, 50 µl of horseradish peroxidase-conjugated goat anti-mouse -IgG(fc)

(Jackson ImmunoResearch, West Grove, Pennsylvania) diluted 1:3500 in PBST added. Plates were incubated as above, washed four times with PBST, and 100 µl substrate consisting of 1 mg/ml o-phenylenediamine (Sigma) and 0.1 (µl/ml 30% H 2 O in 100 mM citrate, pH 4.5, was added. The color reaction was stopped after 5 minutes by the addition of 50 µl of 15% H 2 SO 4 . A490 was read on a plate reader (Dynatech).

Selected fusion wells were cloned twice by dilution in 96-well plates and visual counting of the number of colonies/well after 5 days. Cloned hybridomas were 90D1E, 90E3A, 90E6C, 90G11D (ATCC HB 11724) and 90F2D (ATCC HB 11725).

The monoclonal antibodies produced by hybridomas were isotyped using the Isostrip system (Boehringer Mannheim, Indianapolis, IN). The results showed that the monoclonal antibodies produced by fusion 90 hybridomas were all IgG^

All of the monoclonal antibodies produced by fusion 90 hybridomas performed well in ELISA assays, but were unable to bind PAF-AH on Western blots. In order to generate antibodies which could recognize PAF-AH by Western, mice No. 1958 were immunized with recombinant enzyme. Hybridomas were generated as described for fusion 90, but were screened by Western blotting instead of ELISA to identify Western competent clones.

For Western analyses, recombinant PAF-AH was mixed with an equal volume of sample buffer containing 125 mM Tris, pH 6.8, 4% SDS, 100 mM dithiothreitol, and 0.05% bromophenol blue, and boiled for 5 min before application to a 12 % SDS-polyacrylamide gel (Novex). After electrophoresis at 40 mAmp, proteins were electrotransferred onto a polyvinylidene fluoride membrane (Pierce) for 1 h at 125 V in 192 mM glycine, 25 mM Tris base, 20% methanol, and 0.01% SDS. The membrane was incubated in 20 mM Tris, 100 mM NaCl (TBS) containing 5% bovine serum albumin (BSA, Sigma) overnight at 4 °C. The blot was incubated for 1 hour at room temperature with rabbit polyclonal antisera diluted 1:8000 in TBS containing 5% BSA, and then washed with TBS and incubated with alkaline phosphatase-conjugated goat anti-mouse IgG in TBS containing 5% BSA for 1 hour at room temperature . The blot was again washed with TBS and then incubated with 0.02% 5-bromo-4-chloro-3-indolyl phosphate and 0.03% nitro blue tetrazolium in 100 mM Tris-HCl, pH 9.5, 100 mM NaCl and 5 mM MgCl2. The reaction was stopped by repeated rinsing with water.

Selected fusion wells, in which the supernatants were positive in Western analyses, were processed as described above. Hybridoma 143A reacted with PAF-AH in Western blots and was cloned (ATCC HBZ 11900).

Polyclonal antisera specific for human plasma PAF-AH were produced in rabbits by three monthly immunizations with 100 µg of purified recombinant enzyme in Freund's adjuvant.

Example 14

Experimental studies were performed to evaluate the in vivo therapeutic effects of recombinant PAF-AH according to the present invention on acute inflammation, using a rat foot edema model [Henriques et al., Br. J. Pharmacol., 106:579-582 (1992)]. The results of these studies showed that rPAF-AH blocks PAF-AH-induced edema. Parallel studies were conducted to compare the efficacy of PAF-AH with two commercially available PAF antagonists.

A. Preparation of PAF-AH

E. coli transformed with the PAF-AH expression vector puc trp AH was lysed in a microfluidizer, solids were centrifuged off, and the cell supernatants were applied to an S-Sepharose column (Pharmacia). The column was washed thoroughly with buffer consisting of 50 mM NaCl, 10 mM CHAPS, 25 mM MES and 1 mM EDTA, pH 5.5. PAF-AH was eluted by increasing the NaCl concentration in the buffer to 1 M. Affinity chromatography using a Blue Sepharose column (Pharmacia) was then used as an additional purification step. Before applying PAF-AH preparations to the Blue Sepharose column, the sample was diluted 1:2 to reduce the NaCl concentration to 0.5 M, and the pH was adjusted to 7.5. After thoroughly washing the Blue Sepharose column with buffer consisting of 0.5 M NaCl, 25 mM Tris, 10 mM CHAPS and 1 mM EDTA, pH 7.5, PAF-AH was eluted by increasing the NaCl concentration to 3.0 M.

The purity of PAF-AH isolated in this way was generally 95%, as determined by SDS-PAGE, with activity in the range of 5000-10,000 U/ml. Additional quality controls performed on each PAF-AH preparation included determination of endotoxin levels and hemolytic activity in freshly prepared rat erythrocytes. A buffer containing 25 mM Tris, 10 mM CHAPS, 0.5 M NaCl, pH 7.5, functioned both as a storage medium for the enzyme and as a carrier during administration. Doses used in the experiments were based on enzyme activity analyzes carried out immediately before the experiments.

B. Induction of edema

6-8 week old female Long Evans rats (Charles River Wilmington, MA) weighing 180-200 g were used in all experiments. Before experimental manipulations, the animals were anesthetized with a mixture of the anesthetic

agent Ketaset (Fort Dodge Laboratories, Fort Dodge, IA), Rompun (Miles, Shawnee Mission, KS) and Ace Promazine (Aveco, Fort Dodge, IA), administered subcutaneously in an amount of approx. 2.5 mg Ketaset, 1.6 mg Rompun, 0.2 mg Ace Promazine per animals per dose. Edema was induced in the foot by administration of either PAF or zymosan as follows. PAF (Sigma no. P-1402) was freshly prepared for each experiment from a 19.1 mM stock solution stored in chloroform/methanol (9:1) at -20 °C. The required volumes were dried under N 2 , diluted 1:1000 in a buffer containing 150 mM NaCl, 10 mM Tris, pH 7.5, and 0.25% BSA, and sonicated for 5 minutes. Animals received 50 µl of PAF (final dose 0.96 nmol) subcutaneously between the hind paw pads, and edema was assessed at 1 hour, and again at 2 hours in some experiments. Zymosan A (Sigma No. A-8800) was freshly prepared for each experiment as a suspension of 10 mg/ml in PBS Animals received 50 l of zymosan (final dose 500 µg) subcutaneously between the hind paw pads, and edema was assessed after 2 hours.

Edema was quantified by measuring the foot volume immediately before administration of PAF or zymosan, and at an indicated time point after administration of PAF or zymosan. Edema is expressed as the increase in foot volume in ml. Volume change measurements were performed on anesthetized animals using a plethysmometer (UGO Basile, model no. 7150) which measures the changed band volume of the submerged foot. To ensure that foot immersion was comparable from one time point to the next, the rear foot was marked with waterproof ink where the hairline meets the heel. Repeated measurements of the same foot using this technique showed that the accuracy was within 5%.

C. Administration methods and dosages for PAF-AH

PAF-AH was injected locally between the footpads or systemically by IV injection into the tail vein. For local administration, the rats received 100 1 PAF-AH (4000-6000 U/ml) subcutaneously between the right hind paw pads. The left foot served as a control by administration of 100 µl vehicle (buffered saline). For systemic administration of PAF-AH, rats received the indicated units of PAF-AH in 300 µl vehicle IV in the tail vein. Control animals received the corresponding volume of vehicle intravenously in the tail vein.

D. Local administration of PAF-AH

Rats (N = 4) were injected with 100 µL of PAF-AH (4000-6000 U/ml) subcutaneously between the right footpads. The left was injected with 100 L of vehicle (buffered saline). Four other rats were injected with vehicle only. All rats were immediately given PAF via subcutaneous foot injection, and foot volumes were assessed 1 hour after injection. Figure 6, in which edema is expressed as the mean increase in foot volume (ml) ± SEM for each treatment group, illustrates that PAF-induced foot edema is blocked by local administration of PAF-AH. The group that received local PAF-AH treatment before administration of PAF showed reduced inflammation compared to the control group. An increase in foot volume of 0.08 ml ± 0.08 (SEM) was observed in the PAF-AH group compared to 0.63 0.14 (SEM) for the vehicle-treated control rats. The increase in foot volume was a direct result of PAF injection because the animals injected with vehicle only in the foot did not show any increase in foot volume.

E. Intravenous administration of PAF-AH

Rats (N = 4 per group) were pretreated IV with either PAF-AH (2000 U in 300 L of vehicle) or vehicle alone 15 minutes prior to administration of PAF. Edema was assessed 1 and 2 hours after administration of PAF. Figure 7, in which edema is expressed as the mean increase in volume (ml) ± SEM for each treatment group, illustrates that IV administration of PAF-AH blocked PAF-induced foot edema at 1 and 2 hours after administration. The group receiving 2000 U PAF-AH given IV showed a reduction in inflammation within 2 hours. The mean volume increase for the PAF-AH-treated group at 2 hours was 0.10 ml ± 0.08 (SEM) compared to 0.56 ± 0.11 vehicle-treated controls.

F. Comparison of PAF-AH protection in edema induced by PAF or zymosan

Rats (N = 4 per group) were pretreated IV with either PAF-AH (2000 U in 300 µl vehicle) or vehicle alone. 15 minutes after the pretreatment, the groups received either PAF or zymosan A, and foot volume was assessed after 1 and 2 hours, respectively. As shown in Figure 8, in which edema is expressed as mean increase in volume (ml) ± SEM for each treatment group, systemic administration of PAF-AH (2000 U) was effective in reducing PAF-induced foot edema but did not block zymosan-induced edema. A mean increase in volume of 0.08 ± 0.02 was observed in the PAF-AH-treated group compared to 0.49 ± 0.03 for the control group.

G. Titration of effective dose for PAF-AH protection

In two separate experiments, groups of rats (N = 3-4 per group) were pretreated IV with either serial dilutions of PAF-AH or vehicle control in a volume of 300 µl 15 min prior to PAF administration. Both feet were administered PAF (as described above), and edema was assessed after 1 hour. Figure 9, in which edema is expressed as mean increase in volume (ml) ± SEM for each treatment group, illustrates the increase in protection against PAF-induced edema in rats injected with increasing doses of PAF-AH. In the experiments, the ID50 for PAF-AH given IV was found to be between 40 and 80 U per rat.

H. Efficacy in vitro of PAF-AH as a function of time after administration

In two separate experiments, groups of rats (N = 3-4 per group) were pretreated IV with either PAF-AH (2000 U in 300 µl vehicle) or vehicle alone. After administration, groups received PAF at time points ranging from 15 minutes to 47 hours after administration of PAF-AH. Edema was then assessed 1 hour after PAF administration. As shown in Figure 10, in which edema is expressed as mean increase in volume (ml) ± SEM for each treatment group, administration of 2000 U PAF-AH protected the rats against PAF-induced edema for at least 24 hours.

I. Pharmacokinetics of PAF-AH

Four rats received 2000 U of PAF-AH by intravenous injection in a volume of 300 µl. Plasma was collected at various time points and stored at 4°C, and plasma concentrations of PAF-AH were determined by ELISA using a dual mAb capture assay. Briefly, monoclonal antibody 90G11D (Example 13) was diluted in 50 mM carbonate buffer, pH 9.6, at 100 ng/ml and immobilized on Immulon 4 ELISA plates overnight at 4°C. After extensive washing with PBS containing 0.05% Tween 20, the plates were blocked for 1 hour at room temperature with 0.5% fish skin gelatin (Sigma) diluted in PBS. Two equal serum samples diluted in PBS with 15 mM CHAPS were added to the washed ELISA plate and incubated for 1 hour at room temperature. After washing, a biotin conjugate of monoclonal antibody 90F2D (Example 13) was added to the wells at a concentration of 5 µg/ml diluted in PBS and incubated for 1 hour at room temperature. After washing, 50 µl of a 1:1000 dilution of ExtraAvidin (Sigma) was added to the wells and incubated for 1 hour at room temperature. After washing, the wells were developed using OPD as substrate and quantified. Enzyme activity was then calculated from a standard curve. Figure 11, in which data points represent mean ± SEM, shows that after 1 hour plasma enzyme levels approached the predicted concentration based on a 5-6 ml plasma volume of 180-200 g rats, mean = 374 U/ml ± 58.2. After 1 h, plasma levels declined steadily, reaching a mean plasma concentration of 19.3 U/ml ±3.4 at 24 h, which is still significantly higher than endogenous rat PAF-AH levels, which have been found to be ca. 4 U/ml using enzymatic analyses.

J. Efficacy of PAF-AH in relation to PAF antagonists

Groups of rats (N = 4 per group) were pretreated with one of three potential anti-inflammatory agents: the PAF antagonist CV3988 (Biomol No. L-103) administered IP (2 mg in 200 µl EtOH), the PAF antagonist Alprazolam (Sigma No. A-8800) administered IP (2 mg in 200 µl EtOH) or PAF-AH (2000 U) administered IV. Control rats were injected IV with 300 µl vehicle. The PAF antagonists were administered IP because they were dissolved in ethanol. Rats injected with either CV3988 or Alprazolam were given PAF 30 min after administration of the PAF antagonist to allow the PAF antagonist to enter the circulation, while rats treated with PAF-AH and vehicle were injected 15 min after enzyme administration. Rats injected with PAF-AH showed a reduction of PAF-induced edema beyond that achieved with the established PAF antagonists CV3 988 and Alprazolam. See Figure 12, in which edema is expressed as mean increase in volume (ml) ± SEM for each treatment group.

In summary, rPAF-AH provides effective blockade of PAF-mediated edema in vivo. Administration of PAF-AH products can be either local or systemic by intra-venous injection. In dosing studies, IV injections in the range of 160-2000 U/rat were found to dramatically reduce PAF-mediated inflammation, while the ID50 dose appears to be in the range of 40-80 U/rat. Calculations based on the plasma volume for rats of 180-200 g predict that a plasma concentration in the range of 25-40 U/ml should block PAF-induced edema. These predictions are supported by preliminary pharmacokinetic studies. A dose of 2000 U PAF-AH was found to effectively block PAF-mediated edema for at least 24 hours. 24 hours after administration of PAF-AH, plasma concentrations of the enzyme were found to be approx. 25 U/ml. PAF-AH was found to block PAF-induced edema more effectively than the two known PAF antagonists tested.

Collectively, these results show that PAF-AH effectively blocks PAF-induced inflammation, and may be of therapeutic value in diseases where PAF is the primary mediator.

Example 15

Recombinant PAF-AH according to the present invention was tested in a second in vivo model, PAF-induced pleurisy. PAF has previously been shown to induce vascular leakage when introduced into the alveolus [Henriques et al., supra]. Female rats

(Charles River, 180-200 g) were injected in the tail vein with 200 µl 1% Evans blue dye in 0.9% saline with 300 µl recombinant PAF-AH (1500 umol/ml/h, prepared as described in Example 14) or with an equivalent volume of control buffer. 15 min later, the rats received an injection of 100 µl PAF (2.0 nmol) into the alveolar sac. 1 hour after PAF administration, the rats were sacrificed, and the pleural fluid was collected by flushing the pleura with 3 ml of heparinized, phosphate-buffered saline. The degree of vascular leakage was determined by the amount of Evans blue dye in the alveoli, which was quantified by absorbance at 620 nm. Rats pretreated with PAF-AH were found to have much less vascular leakage than control animals (representing more than 80% reduction of inflammation).

The foregoing results support the treatment of individuals suffering from pleurisy with recombinant PAF-AH enzyme according to the present invention.

Example 16

Recombinant PAF-AH enzyme of the present invention was also tested for efficacy in a model of antigen-induced eosinophil recruitment. Accumulation of eosinophils in the airways is a characteristic feature of late-phase immune responses that occur in asthma, rhinitis and eczema. Balb/c mice (Charles River) were sensitized by two intraperitoneal injections consisting of 1 µg ovalbumin (OVA) in 4 mg aluminum hydroxide (Imject alum, Pierce Laboratories, Rockford, IL) given 2 weeks apart. 14 days after the second immunization, the sensitized mice were either challenged with OVA in the form of an aerosol or saline as a control.

Before the challenge, the mice were randomly placed into four groups of four mice/group. The mice in groups 1 and 3

were pretreated with 140 µl control buffer consisting of 25 mM Tris, 0.5 M NaCl, 1 mM EDTA and 0.1% Tween 80 given by intravenous injection. The mice in groups 2 and 4 were pretreated with 750 units of PAF-AH (activity 5500 units/ml given in 140 µl of PAF-AH buffer). 30 minutes after administration of PAF-AH or buffer, the mice in groups 1 and 2

exposed to PBS in aerosol form, as described below, while the mice in groups 3 and 4 were exposed to OVA in aerosol form. 24 hours later, the mice were treated a second time with either 140 µl buffer (groups 1 and 3) or 750 units of PAF-AH in 140 µl buffer (groups 2 and 4) given by intravenous injection.

Eosinophil infiltration into the trachea was induced in the sensitized mice by exposing the animals to OVA in aerosol form. Sensitized mice were placed in 50 mL conical centrifuge tubes (Corning) and forced to inhale aerosolized OVA (50 mg/mL) dissolved in 0.9% saline for 20 min using a nebulizer (model 646, DeVilbiss Corp., Somerset , PA). Control mice were treated in a similar manner, with the exception that 0.9% saline was used in the nebulizer. 48 hours after exposure to OVA or saline in aerosol form, the mice were euthanized, and the tracheas were excised. The trachea from each group was embedded in OCT and stored in wood

-70 °C until sections were cut.

To evaluate eosinophil infiltration in the trachea

tissue sections from the four groups of mice were stained with either Luna solution and hematoxylin-eosin solution or peroxidase. Twelve 6 µm thick sections were cut from each group of mice and numbered accordingly. Odd-numbered sections were stained with Luna stain as follows. The sections were fixed in formal alcohol for 5 minutes at room temperature, rinsed three times in tap water for 2 minutes at room temperature and then rinsed twice in dH2O for 1 minute at room temperature. Tissue sections were stained with Luna stain for 5 min at room temperature (Luna stain consists of 90 ml Weigert's iron hematoxylin and 10 ml 1% Biebrich Scarlet). Stained sections were dipped in 1% acid

alcohol six times, rinsed in tap water for 1 minute at room temperature, dipped in 0.5% lithium carbonate solution five times and rinsed under running tap water for 2 minutes at room temperature. The sections were dehydrated with 70%-95%-100% ethanol, 1 minute each time, at room temperature, then cleared with xylene for 1 minute, twice, at room temperature and mounted in Cytoseal 60.

For the peroxidase staining, even numbered sections were fixed in acetone at 4°C for 10 minutes and allowed to air dry. 200 1 DAB solution was added to each section and left for 5 minutes at room temperature. The sections were rinsed in tap water for 5 minutes at room temperature, and 2 drops of 1% osmic acid were applied to each section for 3-5 seconds. Sections were rinsed in tap water for 5 minutes at room temperature and stained with Mayer's hematoxylin at 25°C at room temperature. Sections were then rinsed in tap water for 5 min and dehydrated for 1 min with 70%-95%-100% ethanol at room temperature. Sections were cleared twice with xylene for 1 minute at room temperature and mounted in Cytoseal 60.

The number of eosinophils in the tracheal submucosal tissue was evaluated. Tracheae from mice from groups 1 and 2 were found to have very few eosinophils scattered throughout the submucosal tissue. As expected, tracheas from mice in group 3, which were pretreated with buffer and exposed to nebulized OVA, were found to have large numbers of eosinophils throughout the submucosal tissue. In contrast, the tracheas of mice in group 4, which were pretreated with PAF-AH and exposed to nebulized OVA, were found to have very few eosinophils in the submucosal tissue compared to that observed in the two control groups, groups 1 and 2 .

Therapeutic treatment with PAF-AH of individuals exhibiting a late-phase immune response involving the accumulation of eosinophils in the airways, such as occurs in asthma and rhinitis, is thus demonstrated.

Example 17

A PAF-AH product of the present invention was also tested in two different rat models for the treatment of necrotizing enterocolitis (NEC), an acute hemorrhagic intestinal necrosis that occurs in low birth weight infants and causes significant morbidity and mortality. Previous experiments have shown that treatment with glucocorticoids reduces the incidence of NEC in animals and premature infants, and the activity of glucocorticoids has been suggested to occur via an increase in the activity of plasma PAF-AH.

A. Activity in rats with NEC induced by challenge with PAF

1. Prevention of NEC

A recombinant PAF-AH product, rPH.2 (25,500 units in 0.3 ml, groups 2 and 4), or vehicle/buffer alone (25 mM Tris, 0.5 M NaCl, 1 mM EDTA and 0.1 % Tween 80)

(groups 1 and 3) were administered into the tail veins of female Vistar rats (n = 3) weighing 180-220 g. Either BSA (0.25%)-saline (groups 1 and 2) or PAF (0.2 ug/ 100 g) suspended in BSA-saline (groups 3 and 4) was injected into the abdominal aorta at the level of the superior mesenteric artery 15 minutes after injection of rPH.2 or vehicle, as previously described by Furukawa et al. [J. Pediatrician Res., 34:237-241 (1993)]. After 2 hours, the small intestine was removed from the Treitzke ligament to the cecum, thoroughly washed with cold saline and superficially examined. From microscopic examination, samples were taken from the upper, middle and lower parts of the small intestine. The tissues were fixed in buffered formalin, and the samples were processed for microscopic examination by staining with hematoxylin and eosin. The experiment was repeated three times.

Superficial findings indicated a normal-appearing intestine in groups treated with the BSA-saline vehicle. rPH.2 injected in the absence of PAF likewise had no effect on the superficial findings. In contrast, injection of PAF into the descending aorta resulted in rapid, severe discoloration and bleeding in the surface of the intestinal serous membrane. A similar hemorrhage was observed when a section of the small intestine was examined on the mucosal side, and the intestine appeared to be quite necrotic. When rpH.2 was injected via the tail vein 15 minutes before the administration of PAF into the aorta, the intestine appeared normal.

On microscopic examination, the intestine from groups 1, 2 and 4 showed a normal intestinal tuft architecture and a normal population of cells in the lamina propria. In contrast, the group treated with PAF alone showed full-thickness necrosis and hemorrhage throughout the mucosa.

The plasma PAF-AH activities were also determined in the rats used in the experiment described above. PAF-AH activity was determined as follows. Before injection into the tail vein, blood samples were taken. Blood samples were then taken from the vena cava just before the injection of PAF and at the time of sacrifice. Approximately 50 ml of blood was collected in heparin-treated capillary tubes. Plasma was isolated after centrifugation (980 x g for 5 minutes). The enzyme was assayed as previously described by Yasuda and Johnston, Endocrinology, 130:708-716

(1992).

The mean plasma PAF-AH activity in all rats before injection was found to be 75.5 ± 2.5 units (1 unit is equal to 1 nmol x min"<1> x ml"<1> plasma). The mean plasma PAF-AH activities 15 minutes after injection of the vehicle were 75.2 ± 2.6 units for group 1 and 76.7 ± 3.5 units for group 3. After 15 minutes, plasma PAF-AH- the activity of the animals injected with 25,500 units of rPH.2 2,249 ± 341 units for group 2 and 2,494 ±

623 units for group 4. The activity of groups 2 and 4 remained elevated (1855 ± 257 units) until the time of sacrifice (2 1/4 hours after injection of rPH.2) (group 2 = 1771 ± 308; group 4 = 1939 ± 478) . These results indicate that plasma PAF-AH activity in the rats injected with vehicle alone (groups 1 and 3) did not change during the course of the experiment. All animals receiving the PAF injection alone developed NEC, while all rats injected with rPH.2 followed by PAF injection were completely protected.

2. Dose dependence for prevention of NEC

To determine whether protection against NEC in rats was dose-dependent, animals were treated with increasing doses of rPH.2 15 minutes before PAF administration. First, rPH.2, varying from 25.5 to 25,500 units, was injected into the rats' tail veins. PAF (0.4 µg in 0.2 ml BSA-saline) was then injected into the abdominal aorta 15 minutes after the administration of rPH.2. The small intestine was removed and examined for NEC development 2 hours after PAF administration. Plasma PAF-AH activity was determined before the exogenous administration of the enzyme and 15 minutes and 2 1/4 hours after rPH.2 administration. The results are the mean value of 2-5 animals in each group.

Superficial findings indicated that all rats receiving less than 2000 units of the enzyme developed NEC. Plasma PAF-AH activity in animals receiving the lowest protective amount of enzyme (2040 units) was 363 units per ml of plasma after 15 minutes, representing a 5-fold increase over baseline levels. When rPH.2 was administered in an amount less than 1020 total units, the resulting plasma enzyme activity averaged approx. 160 or lower, and all animals developed NEC.

3. Duration of NEC Protection

To determine the length of time the exogenous PAF-AH product provided protection against the development of NEC, rats were injected once with a fixed amount of the enzyme via the tail vein and then challenged with PAF at various time points. rPH.2 (8,500 units in 0.3 ml) or vehicle alone was administered into the rats' tail veins, and PAF (0.36 µg in 0.2 ml BSA-saline) was injected into the abdominal aorta at various times after the enzyme administration. The small intestines were removed 2 h after PAF injection for superficial and histological examinations to evaluate for NEC development. Plasma PAF-AH activities were determined at different time points after enzyme administration and 2 h after PAF administration. The mean ± standard error of enzyme activity was determined for each group.

The results indicated that none of the rats developed NEC within the first 8 hours after injection of rPH.2, but 100% of the animals challenged with PAF 24 and 48 hours after injection of the enzyme developed NEC.

4. Reversal of NEC

To determine whether administration of PAF-AH product was able to reverse the development of NEC induced by PAF injection, 25,500 units of enzyme were administered via injection into the vena cava 2 minutes after PAF administration (0.4 g). None of the animals developed NEC. However, when rPH.2 was administered in this manner 15 minutes after the PAF injection, all animals developed NEC, which is consistent with the rapid time course of NEC development, as induced by the administration of PAF, previously reported by Furukawa [supra].

The sum of these observations indicates that a relatively small (5-fold) increase in plasma PAF-AH activity is able to prevent NEC. These observations, combined with previous reports that plasma PAF-AH activity in rabbit fetuses [Maki et al., Proe. Nati. Acad. Pollock.

(USA), 85:728-732 (1988)] and premature individuals [Caplan et al., J. Pediatr. 116:908-964 (1990)] has been shown to be relatively low, suggesting that prophylactic administration of human recombinant PAF-AH products to low birth weight neonates may be useful for the treatment of NEC.

B. Activity in a neonatal model of NEC

The efficacy of a PAF-AH product, rPH.2, was evaluated as follows in a NEC model in which neonatal rats are stressed by prescription feeding and asphyxia, two common risk factors for the disease in humans. In this model, approx. 70-80% of animals superficial and microscopic intestinal damage similar to neonatal NEC on the 3rd day after birth. Newborn rats were obtained from pregnant Sprague-Dawley rats (Harlan Sprague-Dawley, Indianapolis, IN) that were anesthetized with CO 2 and delivered by cesarean section. Newborn animals were collected, dried and kept in a neonatal incubator throughout the experiment.

Separate groups of animals were first used to determine the dosage and absorption characteristics of rPH.2. Normal newborn rat pups were given one of three different enteral or intraperitoneal doses of rPH.2 (3 X, 15 X, or 75 X) at time 0, and blood was collected 1 hour, 6 hours, or 24 hours later for determination of plasma- PAF-AH activity. PAF-AH activity was measured using a substrate incubation assay [Gray et al., Nature, 374:549

(1995) ] and an ELISA using an anti-human rPAF-AH monoclonal antibody for each sample (90F2D and 90G11D, described in Example 13). For selected samples, immunohistochemical analysis was performed using two different monoclonal antibodies raised against human rPAF-AH (90F2D and 90G11D, described in Example 13). Immunohistochemistry was performed by standard techniques using a 1:100 dilution of the antibody and overnight incubations.

After enteral dosing of rPH.2 to normal newborn rats, there was no measurable plasma PAF-AH activity at any time, either using the substratin incubation assay or the ELISA technique. With intraperitoneal administration of rPH.2, significant circulating PAF-AH activity was measurable using both methods 1 hour after dosing, and this activity peaked at 6 hours. Higher doses of rPH.2 (from 3 to 75 X, 10-250 U) resulted in higher plasma PAF-AH activity. Immunohistochemical analysis revealed the presence of rPAF-AH product in the epithelial cells of the intestinal mucosa after enteral administration. The reactivity accumulated mostly in the intestinal tufts, and minimal staining was present in the crypt cells. There was more staining in the ileum than the jejunum, and some rPAF-AH product was immunochemically identified in parts of the colon. There was no detectable staining in any control samples or in samples from animals dosed intraperitoneally. Thus, enteral administration of rPAF-AH product resulted in local accumulation of the enzyme in the mucosal epithelium without any measurable systemic absorption, while in contrast intraperitoneal administration of rPAF-AH product resulted in high levels of circulating enzyme but no local accumulation in mucosa .

In the NEC model, NEC was induced in newborn rats in accordance with Caplan et al., Pediatr. Pathol., 14:1017-1028 (1994). Briefly, animals were fed feed for newborn pups reconstituted from powder (Esbiliac, Borden Inc.) every 3 hours via a casing tube. The feeding volume first started at 0.1 ml/portion and increased, as tolerated, to 0.4 ml/portion on the 4th day. In all animals, asphyxia was induced twice daily by inhalation of 100% nitrogen for 50 seconds in a closed plastic chamber, followed by exposure to cold (4°C) for 10 minutes. Bowel and bladder function was stimulated by gentle manipulation after each feeding. The animals were kept for 96 hours, or until they showed signs of distress. Sick animals had swollen abdomens, bleeding faeces, shortness of breath, cyanosis and lethargy and were euthanized by decapitation. After sacrifice, the intestine from each rat was examined superficially for signs of necrosis and then fixed with formalin for later histological analysis. The specimens were embedded in paraffin, sectioned with a microtome, stained with hematoxylin and eosin and examined by two observers. Intestinal damage was scored as 1+ for swelling or separation of epithelial cells, 2+ for dissolution of epithelial cells to the middle intestinal tuft level, 3+ for necrosis of entire intestinal tufts and 4+ for transmural necrosis.

To assess the efficacy of rPH.2, three different groups of rats were treated with the compound via enteral administration, intraperitoneal administration, or both. The rPH.2 preparation contained 0.8 mg/ml protein and with a PAF-AH activity of approx. 4000 units/mg, with an endotoxin/protein ratio < 0.5 EU/mg. Enterally dosed animals were given 25 X (80 U) rPH.2 added at each feeding (every 3 hours) via orogastric tube. Intraperitoneally dosed animals were given 75 X by intraperitoneal injection twice daily. Control animals received appropriate volumes of buffer (20 mM NaPO 4 , pH 7.4) without rPH.2 and were examined at the same time as each experimental group. Mortality and signs of NEC were evaluated for each treatment group, and differences were analyzed statistically using Fischer's exact test. A p-value

< 0.05 was considered significant. The results are shown in table 9 below.

The data represent cumulative results from four different experiments for intraperitoneal dosing, four experiments for enteral dosing, and three experiments for intraperitoneal + enteral dosing.

Enteral dosing of rPH.2 significantly reduced the incidence of both NEC and death compared to control animals. Results from four different enteral dosing experiments showed that pretreatment with rPH.2 reduced NEC from 19/26 (control) to 6/26 (p < 0.001). Intestinal damage was variable among treated and control animals, but was in most cases characterized by middle intestinal tuft necrosis in some segments, total intestinal tuft necrosis in other areas, random areas of transmural necrosis and remaining parts with normal intestinal histology. The worst degree of NEC in treated and control animals with intestinal damage was approximately the same (mean score 2.8 for control animals versus 2.4 for rPH.2-treated rats, p > 0.05).

Intraperitoneal dosing with rPH.2 had no significant effect on NEC or death in this model. The onset of symptoms was similar for this group and the control group (40 ± 5 hours for control rats versus 36 ± 7 hours for rPH.2-treated rats), and the degree of NEC in both groups was similar (mean score 2.6 for control animals compared to 2.5 for rPH.2-treated rats).

Additional experiments were performed in which rats were dosed both enterally and intraperitoneally with rPH.2 at the same doses as the single treatment groups (25 X of rPH.2 at each feeding every 3 hours plus 75 X by intraperitoneal injection twice daily. The results are shown above in Table 9. Although there were no significant differences between treatment and control groups with respect to death, rPH.2 treatment significantly reduced the incidence of NEC (10/17 for control animals compared to 3/14 for rPH.2-treated rats, p = 0.04).It should be noted that 6 of the 7 animals that died in the rPH.2-treated group had positive blood cultures for E. coli, determined shortly before death.

These results provide further support for PAF-AH products having a protective role in a neonatal model of non-PAF-induced NEC. Enteral treatment with rPAF-AH product prevents NEC, while intraperitoneal treatment with these doses had no detectable effect. These findings suggest that supplementation with PAF-AH product for artificially fed, premature neonates at risk of NEC may reduce the incidence of this disease.

Example 18

The efficacy of PAF-AH product in a guinea pig model of acute respiratory distress syndrome (ARDS) was investigated.

Platelet-activating factor (PAF) injected intra-venously in guinea pigs produces a profound lung inflammation reminiscent of early ARDS in humans. Within minutes of intravenous administration of PAF, the lung parenchyma becomes congested with blood with constricted bronchi and bronchioles [Lellouch-Tubiana et al., supra]. Platelets and polymorphonuclear neutrophils begin to migrate outward, and cellular aggregates can be easily identified along the pulmonary arterioles [Lellouch-Tubiana, Br. J. Exp. Path., 6"6:345-355 (1985)]. PAF infusion also damages bronchial epithelial cells that dissociate from the airway walls and accumulate in the airway cavities. This damage to the airway epithelial cells is consistent with hyaline membrane formation that occurs in humans during the development of ARDS. The outward migration of neutrophils and platelets is rapidly followed by diapedesis of these cells into the alveolar septa and alveolar cavities of the lung. Cellular infiltrates triggered by PAF are accompanied by significant vascular leakage leading to airway edema [Kirsch , Exp. Lung. Res., 18:447-459 (1992)]. Evidence for edema is further supported by in vitro studies, where PAF induces a dose-dependent (10-1000 ng/ml) release of 125I-labeled fibrinogen in perfused guinea pig lungs [Basran, Br. J. Pharmacol., 77:437 (1982)].

Based on the above observations, an ARDS model was developed in guinea pigs. A cannula is placed in

the jugular vein of anesthetized Hartly male guinea pigs (approx. 350-

400 g), and PAF diluted in 500 µl phosphate-buffered saline with 0.25% bovine serum albumin as a vehicle (PBS-BSA) is infused over 15 minutes at a total dose ranging from 100 to 400 ng/kg. At different intervals after PAF infusion, the animals are euthanized and lung tissue is collected. In guinea pigs infused with PAF, dose-dependent lung damage and inflammation are clearly visible after 15 minutes and continue to be present for 60 minutes. Neutrophils and red blood cells are present in the alveolar cavities of PAF-treated guinea pigs but absent in control or sham-infused animals. Evidence of epithelial cell damage is also evident and resembles hyaline membrane formation in human ARDS patients. Protein determinations performed in bronchoalveolar lavage (BAL) samples taken from guinea pigs infused with PAF show a dramatic accumulation of protein in the inflamed lung, which is clear evidence of vascular leakage.

rPH.2 was found to completely protect against PAF-mediated lung injury in the guinea pig model of ARDS. Groups of guinea pigs were pretreated with either rPH.2 (2000 units in 500 µl) or only 500 µl of the PAF-AH buffer. 15 minutes later, these guinea pigs were infused with 400 ng/kg PAF in a volume of 500 µl, infused over 15 minutes. A simulation group of guinea pigs was additionally infused with 500 µl of PBS-BSA. At the completion of the PAF infusion, the animals were sacrificed, and BAL fluid was collected by lavage of the lungs 2x with 10 ml saline containing 2 µg/ml heparin to prevent coagulation. To determine protein concentration in BAL, the samples were diluted 1:10 in saline, and OD

280 was determined. BAL fluid from the simulation group of guinea pigs was found to have a protein concentration of 2.10 ± 1.3 mg/ml. In sharp contrast, BAL fluid from animals infused with PAF was found to have a protein concentration of 12.55 ± 1.65 mg/ml. In guinea pigs pretreated with rPH.2, BAL fluid was found to have a protein concentration of 1.13 ± 0.25 mg/ml, which is comparable to the sham controls and shows that PAF-AH product completely blocks pulmonary edema in response to PAF.

Example 19

The efficacy of a PAF-AH product, rPH.2, was evaluated in two different models of acute pancreatitis.

A. Activity in a rat pancreatitis model

Male Wistar rats (200-250 g) were purchased from Charles River Laboratories (Wilmington, MA). They were kept in a climate-controlled room at 23 ± 2 °C with a 12-hour light-dark cycle and fed with standard laboratory food with water ad libitum. The animals were randomly placed in either control or experimental groups. The rats were anesthetized with 50 mg/kg pentobarbital sodium intraperitoneally, and a polyvinyl catheter (size V3, Biolab Products, Lake Havasu, AZ) was placed in the jugular vein. The catheter was guided under the skin and out into the dorsal hal region, and the animals were allowed to recover from anesthesia. The rats were given free access to water, but were fasted overnight. The experiment was performed the next day on conscious animals. Meanwhile, the catheter was kept open by constant infusion of saline (0.2 ml/h). On the day of the experiment, animals were injected intravenously with rPH.2 or vehicle control, followed by an infusion of either (1) 5 µg/kg per hour caerulein for 3.5 hours, or (2) 10 µg/kg per hour caerulein for 5 h (Research Plus, Bayonne, NJ). Immediately after completion of the infusion, the animals were anesthetized with pentobarbital sodium, the abdomens were opened, and 5 ml of blood was drawn from the inferior vena cava for subsequent analyses. The animals were then sacrificed by exsanguination . Serum amylase, serum lipase, and serum bilirubin were measured, and the pancreas was removed. Pieces of pancreas were either fixed in a 4% phosphate-buffered formaldehyde solution for histological examination or immediately deep-frozen at -80 °C for measurements of myeloperoxidase activity. Additional pieces of pancreas were examined for content of water and amylase and trypsin, as described below. Myeloperoxidase activity, a measure of neutrophil sequestration, was determined in pancreas and lung as described below. Pulmonary, vascular per mobility was also determined as described below. Statistical analysis of the data was achieved using the unpaired Student<1>'s t-test. The data reported represent mean values + SEM of at least three different experiments. Differences in the results were considered significant when p < 0.05.

1. Water content in the pancreas

Pieces of pancreas were dried and weighed (wet weight) and were then dried for 34 h at 120 °C and weighed again (dry weight). Pancreatic water content was calculated as the difference between wet weight and dry weight and expressed as a percentage of pancreatic wet weight. An increase in water content in the pancreas was considered to indicate the development of edema.

2. Serum and pancreatic amylase

Amylase activity in serum was measured using 4, 6-et<y>lidene(G7)-p-nitrophen<y>l (GJ -α^-maltoplaside (ET-G7PNP) (Sigma Chemical Co., St. Louis, MO) as substrate according to Pierre et al., Clin. Chem., 22:1219 (1976). Amylase activity in pancreatic tissue homogenized in 10 mM phosphate buffer, pH 7.4, was measured using the same method.

3. Trypsin activity in the pancreas

Trypsin activity was measured fluorimetrically using Boc-Gin-Ala-Arg-MCA as substrate. Briefly, 200 µl of the sample and 2.7 ml of 50 mM Tris buffer (pH 8.0) containing 150 mM NaCl, 1 mM CaCl2 and 0.1% bovine serum albumin were mixed in a cuvette. 100 µl of substrate was added to the sample after preincubation for 20 seconds to start the reaction. The fluorescence reading was performed (excitation 380 nm, emission 440 nm) and expressed as a slope curve. To allow pooling of data from different experiments, trypsin activity in the fractions was expressed as a percentage of total trypsin activity.

4. Histology and morphometry

For light microscopy, complete, arbitrary cross-sections of the head, body and tail of the pancreas were fixed in 10% neutral phosphate-buffered formalin. Paraffin-embedded 5 µm sections were stained with hematoxylin-eosin (H&E) and examined blindly by an experienced morphologist. Acinar cell injury/necrosis was defined as either (a) presence of empty acinar cells or (b) vacuolization and swelling of acinar cells and destruction of the histoarchitecture of all or part of the acinar cells, both of which had to be associated with an inflammatory reaction. The extent of acinar cell damage/necrosis and the total area occupied by acinar tissue were quantified morphometrically using a computerized planimetric image analysis video unit (model CCD-72, Dage-MTl, Michigan city, IN) with NIH-1200 image analysis software. Ten randomly selected microscopic fields (125x) were examined for each tissue sample. The degree of acinar cell damage/necrosis was expressed as the percentage of total acinar tissue that was occupied by areas satisfying the criteria for damage/necrosis.

5. Measurement of activity of pancreatic and lung myeloperoxidase (MPO)

Neutrophil sequestration in the pancreas and lung was evaluated by measuring tissue myeloperoxidase activity. Tissue samples taken at the time of sacrifice were stored at -70 °C until the time of analysis. Samples (50 mg) were thawed and homogenized in 1 ml of 20 mM phosphate buffer (pH 7.4) and centrifuged (10,000 x g, 10 minutes, 4 °C). The resulting pellet was resuspended in 50 mM phosphate buffer (pH 6.0) containing 0.5% hexadecyltrimethylammonium bromide (Sigma, St. Louis, MO) and subjected to four freeze-thaw cycles. The suspension was then further disintegrated by sonication for 40 seconds and centrifuged (10,000 x g, 5 minutes, 4°C). A reaction mixture consisting of the extracted enzyme,

1.6 mM tetramethylbenzidine (Sigma Chemical Co., St. Louis, MO), 80 mM sodium phosphate buffer (pH 5.4), and 0.3 mM hydrogen peroxide were incubated at 37°C for 110 seconds, and absorbance was measured at 655 nm in a CobasBio autoanalyzer. This absorbance was then corrected for the dry weight of the tissue sample fraction. 6. Measurement of pulmonary artery permeability

Obstruction of the common bile duct also typically results in severe pancreatitis-associated lung injury that can be quantified by pulmonary vascular permeability and histological examination. 2 hours before the animals were sacrificed, they were given an intravenous bolus injection of 5 mg/kg fluorescein isothiocyanate-albumin (FITC-albumin, Sigma Chemical Co., St. Louis, MO). Pulmonary microvascular permeability was evaluated by quantifying the leakage of FITC-albumin from the vascular space into the bronchoalveolar area. Briefly, immediately after sacrifice, the right bronchus was blocked using a clamp, and the trachea was exposed. The right lung was then lavaged using a cannula inserted into the trachea. Three washes with saline (60 ml washout) were pooled, and the FITC fluorescence in serum and washout was measured at excitation 494 nm and emission 520 nm. The fluorescence ratio between lavage fluid and blood was calculated and taken as a measure of microvascular permeability in the lung. The lung was also stained with H&E and examined histologically. 7. Effect of administration of caerulein and rPH. 2

Infusion of caerulein alone in an amount of

5 ug/kg/h for 3.5 hours resulted in a typical mild secretagogue-induced pancreatitis in the rat, which was characterized by hyperamylasemia, pancreatic edema as measured by pancreatic water content, and histological changes, including marked acinar cell vacuolization and pancreatic edema. Infusion of saline in control animals produced none of these biochemical or histological changes. Administration of rPH.2

intravenously at doses of 5, 10, or 20 mg/kg 30 minutes before the start of the caerulein infusion did not significantly alter the magnitude of the changes in pancreatic edema (water content) and histology induced by infusion of caerulein alone. Administration of rPH.2 also had no effect on caerulein-induced activation of pancreatic trypsinogen or other amylase content.

Infusion of a higher dose of caerulein,

10 ug/kg/h for 5 hours to rats resulted in a more severe pancreatitis, characterized, relative to controls, by a more pronounced increase in serum amylase activity and pancreatic edema, a marked increase in pancreatic MPO activity, and a significant increase in trypsinogen activation and amylase activity in the pancreas. Pancreatic histology indicated not only pancreatic edema and acinar cell vacuolization, but also some scattered necrosis and a few infiltrating cells.

Administration of rPH.2 (5 or 10 mg/kg intravenously) 30 minutes before the start of caerulein infusion

(10 ug/kg/h) reduced the extent of many of the pancreatic changes induced by infusion of caerulein alone. The results are shown in table 10 below. rPH.2 treatment at a dose of 5 mg/kg resulted in reduction of serum amylase activity (from 10,984 ± 1,412 to 6,763 ± 1,256). The higher dose of 10 mg/kg of rPH.2 produced no further improvement in hyperamylasemia. Treatment with either 5 or 10 mg/kg rPH.2 also produced some reduction in caerulein-induced development of pancreatic edema, as measured by water content (90.61 ± 0.27 for caerulein alone, vs. 88.21 ± 0.61 for caerulein + 5 mg/kg rPH.2 The 5 mg/kg dose of rPH.2 produced a significant reduction in pancreatic MPO activity (2.92 + 0.32-fold increase over controls for caerulein alone vs. 1.19 + 0.21 for caerulein with rPH.2, p < 0.05). Higher doses of rPH.2 produced no further improvement in MPO activity. Neither dose of rPH.2 produced any significant change in the degree of trypsinogen activation or amylase content in pancreas Pancreatic histology indicated some improvement of microscopic necrosis and infiltration after pretreatment with rPH.2.

Pancreas-associated lung damage has been observed both clinically and in several models of pancreatitis. Infusion of caerulein at 5 ug/kg/h for 3.5 hours, which produced a mild form of pancreatitis, produced no significant damage to the lungs. However, infusion of caerulein at 10 ug/kg/hr for 5 h, which resulted in more severe pancreatitis, also produced lung injury quantified by increased pulmonary vascular permeability (0.31 ± 0.04 to 0.79 ± 0.09), lung MPO -activity (indicating neutrophil sequestration) and neutrophil infiltration on histological examination.

Administration of rPH.2 at a dose of 5 mg/kg

30 min before infusion of caerulein significantly reduced the increase of lung MPO activity induced by infusion of caerulein alone (3.55 ± 0.93 for caerulein alone vs. 1.51 0.26 for caerulein with rPH.2). rPH.2 treatment significantly reduced the severity of microscopic changes observed in lung tissue after caerulein infusion. The caerulein-induced increase in pulmonary artery permeability was reduced by treatment with rPH.2, although not statistically significant. The higher dose of 10 mg/kg of rPH.2 was not more effective than the lower dose in reducing the severity of caerulein-induced lung injury.

B. Activity in a model of opossum pancreatitis

Healthy, randomly captured American opossums ( Didelphis virginiana ) of both sexes (2.0 kg to 4.0 kg) were obtained from Scott-Haas and housed in climate-controlled rooms at 23 ± 2 °C with a light-dark cycle in 12 hours, and fed with a standard laboratory chow with water ad libitum. After fasting overnight, the animals were anesthetized with 50 mg/kg sodium pentobarbital i.p.

(Veterinary Laboratories Inc., Lenexa, KS). A coeliotomy was performed through a midline incision under sterile conditions, and the common bile-pancreatic duct was ligated in all animals to induce acute necrotizing pancreatitis. The gallbladder duct was also ligated to prevent the gallbladder from serving as a bile reservoir. The animals were randomly placed in either control or experimental groups. Starting 2 days after ligation of the pancreatic duct, the experimental group received 5 mg/kg body weight per day of rPH.2 (in a 4 mg/ml solution) intravenously via the tail vein, while the control group only received an intravenous injection of the same volume of placebo vehicle. After 1 and 2 days of treatment (on days 3 and 4 after ligation of the pancreatic duct), the animals were euthanized with an overdose of sodium pentobarbital. Blood samples were taken from the heart for measurements of serum amylase, serum lipase and serum bilirubin, and the pancreas was removed. Pieces of pancreas were either fixed in a 4% phosphate-buffered formaldehyde solution for histological examination, or immediately deep-frozen at -80 °C for measurements of myeloperoxidase activity. Additional pieces of pancreas were examined for water content and pancreatic amylase, as described above in section A of this example. Myeloperoxidase activity, a measure of neutrophil sequestration, was determined in the pancreas as described above. Pulmonary artery permeability was also determined as described above.

The results reported represent means ± standard error (SEM) of multiple determinations in three or more separate experiments. The significance of the changes was evaluated using Student's t-test when the data only consisted of two groups, or by analysis of variance (ANOVA) when comparing three or more groups. If the ANOVA indicated a significant difference, the data were analyzed using Tukey's method as a post hoc test for the difference between groups. A p-value < 0.05 was considered to indicate a significant difference.

The results are shown in Table 11. Obstruction of the common bile-pancreatic duct resulted in severe

necrotizing pancreatitis, characterized by hyperamylasemia, hyperlipasemia and extensive necrosis of the pancreas. Obstruction of the common bile duct was also associated with a marked increase in serum bilirubin levels. Intravenous administration of rPH.2 (5 mg/kg/day) starting on day 2 after ligation of the pancreatic duct reduced the magnitude of many of the pancreatic changes induced by duct obstruction and placebo treatment alone. One day of rPH.2 treatment reduced serum amylase levels compared to placebo-treated animals, although the difference was not statistically significant, and 2 days of rPH.2 treatment (on day 4 after pancreatic duct ligation) significantly reduced serum amylase levels compared to placebo. One or 2 days of rPH.2 treatment reduced serum lipase levels relative to controls, although the difference was not statistically significant. Two days of rPH.2 treatment reduced pancreatic amylase content relative to controls, although 1 day of treatment resulted in an increase in pancreatic amylase. Treatment with rPH.2 was not observed to affect serum bilirubin levels, pancreatic myeloperoxidase activity or pancreatic water content.

The main characteristic histological changes induced by obstruction of the biliary-pancreatic duct included marked necrosis, infiltration of inflammatory cells, acinar cell vacuolization, and marked distension of the acinar cavities. Morphometric examination of the pancreas for acinar cell damage showed that rPH.2 had an extensive protective effect on the pancreas after 1 and 2 days of treatment with rPH.2. After 1 day of treatment with rPH.2, the acinar cell damage was reduced to approx. 23% of total acinar cell tissue compared to 48% damage for the placebo-treated animals. This reduction in acinar cell damage was even more pronounced after 2 days of treatment, at which time the rPH.2 treatment resulted in approx. 35% damage of the total acinus cell tissue compared to approx. 60% damage for the placebo-treated animals.

Pulmonary blood vessel permeability, quantified by FITC injection, showed a highly significant difference after treatment 1 and 2 days with rPH.2 compared to the placebo group. Histological examination of the lung showed severe lung damage in all placebo-treated animals. Lung damage was characterized by an extensive inflammatory response with interstitial and intraalveolar infiltration of mainly macrophages, lymphocytes and neutrophils, and by a diffuse, but marked, interstitial edema and thickening of the alveolar membranes. Administration of rPH.2 resulted in a marked reduction of inflammatory cell infiltration and a reduction of interstitial edema at all time points.

In summary, these results show that administration of rPH.2 intravenously at a dose of 5 mg/kg/day, starting 48 hours after ligation of the pancreatic duct, resulted in significant reduction of the increase in blood levels of amylase and lipase and acinar cell damage, as quantified by morphometric analysis of H&E-stained sections, and a significant reduction in the severity of pancreatitis-induced lung injury. Administration of rPAF-AH product in this clinically relevant model of pancreatitis showed beneficial effects in terms of reducing the severity of pancreatitis.

Example 20

A study was conducted to evaluate the effect of a PAF-AH product, rpH.2, on neurotoxicity associated with HIV infection. Human immunodeficiency virus type l (HIV-l) infection of the central nervous system results in neuronal loss by apoptosis. HIV-1-infected monocytes activated by many different antigenic stimuli, including contact with nerve cells, secrete high levels of neurotoxic, proinflammatory cytokines, including PAF. The effect of rPH.2 on the neurotoxicity of conditioned media from HIV-infected and activated monocytes was determined.

Monocytes were infected with HIV and activated as follows. Monocytes were obtained from peripheral bone marrow cells (PBMC) from HIV- and hepatitis B-serum-negative donors after leukopheresis and purified (>98%) by countercurrent centrifugal washing, as described by Genis et al., J. Exp. Med., 176: 1703-1718 (1992). Cells were grown as adherent monolayers (1 x 10<4> cells/ml in T-75 culture flasks) in DMEM (Sigma, St. Louis, MO) with recombinant human macrophage colony-stimulating factor (MSCF) (Genetics Institute, Inc ., Cambridge, MA). Under these conditions, monocytes differentiated into macrophages. After culture for 7-10 days, the macrophages were exposed to HIV-lmA (accession no. M60472) at a multiplicity of infection (MOI) of 0.01 infectious virus particles/target cell. Under these conditions, 20-50% of the monocytes were infected 7 days after HIV-1 inoculation, as determined by immunofluorescence and in situ hybridization techniques [Kalter et al., J. Immunol., 146:298-306

(1991)]. All cultures were supplied with fresh medium every 2 to 3 days. 5-7 days after HIV-1 infection, and below the peak for reverse transcriptase activity (10<7> cpm/ml), determined in accordance with Kalter et al., supra, cultures of HIV-1-infected monocytes and parallel cultures of uninfected monocytes stimulated with LPS (10 ng/ml) or vehicle for 30 min at 37 °C, and then snap-frozen at -80 °C until used in the neurotoxicity assay.

Cultures of human cerebral cortical neuronal cells were established as follows. Human fetal brain tissue was obtained from the telencephalon of second trimester (13-16 weeks gestation) human fetal brain tissue, according to a modified procedure of Banker and Cowan, Brain Res., 126:397-425 (1977). Briefly, brain tissue was collected, washed in 30 ml of cold Hank's BSS (containing Ca<2+> and Mg<2+> + 25 mM HEPES and 5 x gentamicin), separated from adherent meninges and blood, and cut into 2 mm <3> bits. The tissue was pressed through a 230 µM Nitex bag and gently triturated through a flame polished Pasteur pipette 10-15 times. The tissue was centrifuged at 550 rpm for 5 min at 4 °C, and the pellet was resuspended in 5-10 ml MEM-hip (D-glucose, 5 g/l; L-glutamine, 2 mM; HEPES, 10 mM; Na- pyruvate, 1 mM; KC1, 20 mM) containing NI components (insulin, 5 mg/l; transferrin,

5 mg/l; selenite, 5 ug/l, progesterone, 20 nM; putreskin,

10 MU), as well as 10% fetal calf serum (FCS), PSN antibiotic mixture (penicillin, 50 mg/l; streptomycin, 50 mg/l; neomycin, 100 mg/l) and fungizone (2.5 mg/l ). The number and viability of cells were determined by dilution with Hank's BSS with 0.4% trypan blue (1:1, vol/vol) and counting with a hemocytometer. The cells were gently triturated five times with a 10 ml pipette and spread at a density of 10s cells/12 mm slide precoated with poly-L-lysine (70K-150K MW, Sigma, St. Louis, MO), placed in 24-well culture dishes. 1 ml of medium was pipetted into each culture well. Cells were grown for 10-28 days at 37°C in a humidified atmosphere of 5% CO 2 /95% air, and with replacement of medium every 3 days. Under these conditions, cultures were > 60-70% homogeneous with respect to neutrons, with 20-30% astrocytes, < 1% microglia and approx. 10% macrophage and microglia staining. After culture for 14-28 days, neuronal cultures express sufficient levels of N-methyl-D-aspartate (NMDA) or non-NMDA receptors to die after toxic doses of NMDA or alpha-amino-3-hydroxy-5-methyl-4 -isoxazole-propionic acid (AMPA).

The neurotoxicity assay was performed as follows. The test samples which were (a) conditioned medium from LPS-stimulated HIV-1 infected monocytes, (b) control medium, (c) conditioned medium with added rPH.2 to 51 µg/ml or (d) conditioned medium with added vehicle for rPH .2, was added to the neuronal cell cultures at a concentration of 1:10 vol/vol for 24 h. Neurotoxicity was measured by identification of apoptotic nuclei in situ on neuron coverslips fixed in 4% paraformaldehyde using a commercial kit (Apop Tag; ONCOR, Gaithersburg, MD) that uses terminal deoxynucleotidyl transferase (TdT) to bind digoxigenin-dUPT to free 3' -OH ends of newly cleaved DNA (TUNEL staining). Digitized images of TUNEL-stained neurons in > 15 arbitrarily selected microscopic areas were analyzed for the number of TUNEL-stained nuclei/number of total neurons per 50 x field, using computerized morphometry (MCID, Imaging Research, St. Catherine, Ontario, Canada). Data were expressed as % neurons positive for TUNEL staining ± SEM, and are shown in Figure 13. Tests of statistical significance between controls and experimental treatments were determined using ANOVA or paired t-tests, with significance p < 0.05 . Quantification of these cultures confirmed that conditioned media from HIV-infected and activated monocytes induced neuronal cell death in nearly 25% of the total population of cortical neurons, and rPH.2 was able to reduce this toxicity to less than 5% of the total neurons. rPH.2 itself was not neurotoxic, because 50 ug/ml rPH.2 had no effect on nerve cell death compared to cultures treated with control medium. These results clearly show that a major component of the induced neurotoxicity by application of conditioned medium from activated HIV-1 infected monocytes must be due to PAF, because neurotoxicity can be almost completely counteracted by co-incubation with PAF-AH product, the enzyme responsible for metabolism of PAF in the central nervous system. These findings suggest potential therapeutic interventions in the treatment of the CNS neurological disease associated with HIV-1 infection.

Example 21

Almost 4% of the Japanese population have low or undetectable levels of PAF-AH activity in their plasma. This deficiency has been correlated with severe respiratory symptoms in asthmatic children [Miwa et al., J. Clin. Invest., 82:1983-1991 (1988)] who appear to have inherited the deficiency in an autosomal recessive manner.

To determine whether the deficiency arises from an inactive but present enzyme, or from an inability to synthesize PAF-AH, plasma from many patients with deficient PAF-AH activity was analyzed for both PAF-AH activity (using the method described in Example 10 for transfectants) and for the presence of PAF-AH using the monoclonal antibodies 90G11D and 90F2D

(Example 13) in a sandwich ELISA as follows. Immulon 4 flat bottom plates (Dynatech, Chantilly, VA) were coated with 100 ng/well of monoclonal antibody 90G11D and stored overnight. Plates were blocked for 1 h at room temperature with 0.5% fish skin gelatin (Sigma), diluted in CMF-PBS, and then washed three times. Patient plasma was diluted in PBS containing 15 mM CHAPS and added to each well of the plates (50 µl/well). The plates were incubated for 1 hour at room temperature and washed four times. 50 µl of 5 µg/ml monoclonal antibody 90F2D, which was biotinylated by standard methods and diluted in PBST, was added to each well, and the plates were incubated for 1 hour at room temperature and then washed three times. 50 µl of ExtraAvidin (Sigma) diluted 1/1000 in CMF-PBST was then added to each well, and the plates were incubated for 1 h at room temperature before being developed.

A direct correlation between PAF-AH activity and enzyme levels was observed. An absence of activity in a patient's serum was reflected by an absence of detectable enzyme. Plasma samples with half the normal activity also contained half the normal levels of PAF-AH. These observations suggested that the lack of PAF-AH activity was caused by an inability to synthesize the enzyme, or was caused by an inactive enzyme that was not recognized by the monoclonal antibodies.

Further experiments revealed that the deficiency was caused by a genetic lesion in the human plasma PAF-AH gene. Genomic DNA from PAF-AH-deficient individuals was isolated and used as template for PCR reactions with PAF-AH gene-specific primers. Each of the coding sequence exons was first amplified and sequenced from one individual. A single nucleotide change within exon 9 was observed (a G to T at position 996 of SEQ ID NO:7). The nucleotide change results in an amino acid substitution of valine with phenylalanine at position 279 of the PAF-AH sequence (V279F). Exon 9 was amplified from genomic DNA of an additional 11 PAF-AH-deficient individuals found to have the same point mutation.

To test whether this mutation destroyed the enzyme, an E. coli expression construct containing the mutation was prepared using methods similar to that described in Example 10. When introduced into E. coli, the expression construct produced no PAF-AH activity, while a control construct lacking the mutation was fully active. This amino acid substitution probably results in a structural modification which causes the observed lack of activity and lack of immunoreactivity with the PAF-AH antibodies of the present invention.

PAF-AH-specific antibodies according to the invention can thus be used in diagnostic methods to detect abnormal levels of PAF-AH in serum (normal levels are approx. 1-5 U/ml), and to follow the progression of treatment of pathological conditions with PAF-AH. Furthermore, identification of a genetic lesion in the PAF-AH gene allows genetic screening for the PAF-AH deficiency exhibited by the Japanese patients. The mutation causes the addition of a restriction endonuclease site (Maell) and thus allows the simple method of restriction fragment length polymorphism (RFLP) analysis to differentiate between active and mutant alleles. See Lewin,

pp. 136-141 in Genes V, Oxford University Press, New York, New York (1994).

Screening of genomic DNA from 12 PAF-AH-deficient patients was performed by digestion of DNA with Maell, Southern blotting and hybridization with an exon 9 probe (nucleotides 1-3 96 of SEQ ID NO: 17). All patients were found to have RFLP consistent with the mutant allele.

Although the present invention has been described in terms of specific embodiments, it is understood that variations and modifications will be apparent to those skilled in the art. The present invention is thus only limited by the accompanying claims.

Claims (14)

1. PAF-AH polypeptide fragment, characterized in that it is selected from the group consisting of polypeptides having Met46 in SEQ ID NO. NO. 8 as the initial N-terminal amino acid. 2. PAF-AH polypeptide fragment according to claim 1, characterized in that up to 30 C-terminal amino acids of the amino acid sequence with SEQ.ID are missing. NO. 8. 3. PAF-AH polypeptide fragment according to claim 1, characterized in that its C-terminal residue is a residue in SEQ ID NO. NO. 8 selected from the group consisting of: (a) Ile429, (b) Leu431, and (c) Asn441. 4. Variant of the PAF-AH polypeptide fragment according to claim 1, characterized in that it has an amino acid replacement in the sequence with SEQ.ID. NO. 8 selected from the group consisting of: (a) S 108 A, (b) S 273 A, (c) D 286 A, (d) D 286 N, (e) D 296 A, (f) D 304 A, ( g) D 338 A, (h) H 351 A, (i) H 395 A, (j) H 399 A, (k) C 67 S, (1) C 229 S, (m) C 291 S, (n ) C 334 S, and (o) C 407 S. 5. Isolated polynucleotide, characterized in that it codes for any PAF-AH polypeptide fragment according to claims 1-3, any variant fragment according to claim 4. 6. Isolated polynucleotide, characterized in that it codes for a human PAF-AH fragment or a variant fragment having Met46 in SEQ ID NO. NO. 8 as the N-terminal residue and Ile429 or Asn441 as the C-terminal residue. 7. Polynucleotide according to any one of claims 5 or 6, characterized in that it is a DNA. 8. DNA vector, characterized in that it comprises a DNA according to claim 7. 9. Host cell, characterized in that it is stably transformed or transfected with a DNA according to claim 8 in a way that allows expression in the host cell of a PAF-AH polypeptide fragment, a variant or a variant fragment. 10. Method for producing a PAF-AH polypeptide fragment, a variant or a variant fragment of plasma PAF-AH, characterized by cultivation of a host cell according to claim 9 in a suitable nutrient preparation and isolation of the PAF-AH fragment, variant or variant fragment from the cell or the growth medium. 11. PAF-AH polypeptide fragment, variant or variant fragment, characterized in that it is produced by the method according to claim 10. 12. Pharmaceutical preparation, characterized in that it comprises the PAF-AH fragment, variant or variant fragment according to any one of claims 1 or 11, and a pharmaceutically acceptable diluent, an adjuvant or a carrier. 13. Use of PAF-AH fragment, variant or variant fragment according to any one of claims or 11 for the manufacture of a medicament for the treatment of a mammal susceptible to or suffering from a PAF-mediated pathological condition, wherein the medicament is administered in an amount sufficient to supplement PAF-AH activity and inactivate pathological effects of PAF in the mammal. 14. Use according to claim 13 where the pathological condition is pleurisy, asthma, rhinitis, necrotizing enterocolitis, acute respiratory distress syndrome, acute pancreatitis or neurological disease associated with HIV infection.