Description

Cross-Reference to Related Application

This is a continuation-in-part of application Serial No. 728,833 filed April 30, 1985, whose disclosures are incorporated by reference. Technical Field

The present invention relates to mitigation of the effects of gram-negative bacterially secreted lipopolysaccharides on an infected host animal, and particularly the mitigation provided by a particular glycoprotein found in acute phase serum that is substantially absent from normal serum and binds to lipopolysaccharides.

Background Art

The wide variety of microorganisms commonly found in the gastrointestinal tract, particularly the gram-negative, nonsporulating bacilli, have become increasingly important in clinical medicine. They are the principal organisms found in infections of the abdominal viscera, peritoneum, and urinary tract, as well as being frequent secondary invaders of the respiratory tract, burned or traumatized skin, and sites of decreased host resistance and instrumentation. Currently, they are the most frequent cause of life-threatening bacteremia. The gastrointestinal flora are exceedingly complex. The large intestine contains about 10 to 10 organisms per gram of contents. Of these, 90 to 95 percent are obligate anaerobes. Most common are the gram-negative bacilli, Bacteroides and Fusobacterium, gram-positive bacilli, including

Bifidobacterium, Eubacterium, and Corynebacterium species, and a wide variety of anaerobic streptococci. Other anaerobes include the gram-positive spore-forming rods of the clostridia species and gram-negative cocci, Veillonella. Enterococci are also present. The well-known aerobic gram-negative rods, which are members of the family Enterobacteriaceae, account for only 5 to 10 percent of the total flora. These include the most common, E. coli, as well as the

Klebsiella-Enterobacter group, Proteus, Providencia, Edwardsiella, Serratia, and under pathologic conditions. Salmonella and Shigella.

Drugs often used to treat gram-negative bacterial diseases include aminoglycosides such as gentamicin, tobramycin and amikacin as well as carbenicillin and ticarcillin. Penacillinase- resistant drugs such as ethicillin, oxacillin and nafcillin or members of the cephalsporin family such as cepthlothin, cefazolin and cephaprin are also used. Penicillin, clindamycin and chloramphenicol are often suggested for use when anaerobes are implicated.

However, drug therapy has several drawbacks. First, the above usually used drugs act against the infective agent, the bacterium,- and not against the problem-causing agents, that include the toxin known as bacterial endotoxin or lipopolysaccharide. Second, drugs may kill great numbers of otherwise useful gastrointestinal flora. Third, many pathogenic strains of bacteria have developed resistance for the usually used drugs as is evidenced by the use of penaσillinase-resistant drugs.

Thus, it would be beneficial if a therapy could be devised to mitigate the effects of the lipopolysaccharide endotoxin while the body's natural humoral and cellular protective systems combat the 5 infecting bacteria. "V The gram-negative bacteria of the

^ gastrointestinal tract produce disease by invasion of tissue and by release of a pharmacologically active lipopolysaccharide (LPS) from the cell wall, known as 10 endotoxin. Endotoxins from a wide variey of unrelated species behave quite similarly, regardless of the inherent pathogeniσity of the microorganism from which they are derived or their antigenic structure. 15 In the intact microorganism, endotoxins exist as complexes of lipid, polysaccharide, glycolipid and non-covalently-bound protein. The biologic activity seems to be a property of a lipid and carbohydrate portion. 20 The lipopolysaccharides of gram-negative bacteria may be roughly divided into three structural regions. The outer-most' region contains the chains of specific sugars that characterize the O-specific antigens and determine individual serotypes within a 25 species. The specific sugars are linked to a core polysaccharide that is of similar structure among related groups of bacteria. The core is in turn linked through 2-keto-3-deoxyoctonate disaccharides to the major lipid component termed lipid A. _t , 30 Evidence has now accumulated to indicate that the properties of endotoxins may be accounted for by this complex lipid substance.

Lipid A is a glucosamine disaccharide esterified with phosphoric and pyrophosphoric acid 35 and also contains ester- or amide-linked lauric,

palmitic, and myristic acids. Perhaps the most important finding in recent years is that the lipid A and core-polysacchride regions are immunogenic and can induce antibodies that cross-react among the gram-negative bacteria.

Animal studies reveal that antibodies prepared against these components of endotoxin protect against challenge from heterologous gram-negative bacteria. However, better protection is reportedly obtained by immunization with specific O-antigens that induce opsonizing antibodies. Upon entry into the bloodstream and initiation of endotoxemia, LPS and blood humoral and cellular elements interact. The work of several groups has shown that the blood-borne LPS partitions between the tissues and plasma lipoproteins, with specific binding to high density lipoprotein (HDL) . Freudenberg et al. (1980) Infee. Immun. 28;373; Mathison and Ulevitch (1979) J. Immunol. 123:2133; Munford et al. (1982) J. Clin. Invest. 70:877; and Ulevitch et al. (1981) J. Clini. Invest. 67:827.

Depending upon the method of introduction, about 10-50 percent of the initially administered LPS partitions to the plasma lipoproteins (HDL) , with the remainder going to the tissues. Clearance from the animal body appears to be via the tissues and into bile. Thus, if the partitioning between plasma lipoproteins and tissues could be adjusted to be less favorable to HDL, LPS could be more quickly cleared from the body of the infected animal.

In view of the above reports and the findings discussed hereinafter, an early example of improved clearance may have been reported by Filkins

(1976) Proc. Soc. Exptl. Biol. Med. 151:89. It was there reported that rats treated with whole rat blood

plasma and serum from either post-endotoxic or post-trauma donors manifested detoxifying potential in rats into which Salmonella enteritidis LPS had been injected. In contrast, normal rat blood and

5 phosphate-buffered saline controls exhibited no detoxification. A role for the reticuloendothelial system in elaboration of the blood anti-endotoxin system was postulated.

It has also been reported that acute phase

10 rabbit serum (APRS) modifies the interactions of LPS with HDL by retarding the _in vitro rate of binding of LPS to HDL, thereby modulating the endotoxic effect of the LPS. Tobias and Ulevitch (1983) J. Immunol. 131:1913-1916. Binding of LPS to components of

15 normal rabbit serum (NRS) has also been reported. Ulevitch and Johnston (1978) J. Clin. Invest. .62:1313; Ulevitch et al. (1979) J. Clin. Invest. 4:1516; Ulevitch et al. (1981) J. Clin. Invest. 62:827; Munford et al. (1981) J. Clin. Invest.

20 7fJ:877; and Freudenberg et al. (1980) Infect. Immun. 2J3:373.

In rabbits, the interaction of LPS with HDL can be accounted for by a two-step mechanism in which the LPS is first disaggregated by the action of serum

25 proteins. The disaggregated LPS thereafter binds with HDL to form the observed complex. It is believed that similar mechanisms apply in other animals, including man.

Mixtures of LPS with rabbit serum that are

30 permitted to react for 30 minutes at 37 degrees C provide an LPS complex with a density of less than about 1.2 grams per cubic centimeter (g/cc) . When NRS is used, the complex contains components of HDL including apolipoprotein Al (apo Al) [Ulevitch et al.

35 (1981) J. Clin. Invest. 67:827], while with APRS the

complex contains apo Al and also serum amyloid A apolipoprotein (apo SSA) [Tobias et al. (1982) J. Immunol. 128:1420] .

While complexes with densities of about 1.2 g/cc are ultimately formed by admixture of LPS with NRS and with APRS, the times for formation of similar amounts of those complexes differ. Thus, for

NRS, the formation of the 1.2 g/cc complex is about

90 percent complete within about 30 minutes, while in APRS, the complex is about 95 percent formed after a time period of about 6 hours. Tobias and Ulevitch (1983) J. Immunol. 131:1913.

In addition to the time courses of complex formation being different in NRS as compared to APRS, initial complexes formed in the two serum types also differ in density. Thus, in NRS, the density of the initially formed complex is 1.33 g/cc, while in APRS, the density of the intial complex is 1.3. An LPS-containing serum complex with a density of 1.3 g/cc was also reported when Balbc/Strong mice were injected with AgN0 ₃ or LPS. Tobias and Ulevitch (1983) J. Immunol. 131:1913.

Precipitated euglobulin fractions formed from mixtures of LPS and NRS or APRS were examined for their solubilities in saline. It was found that substantially all of the LPS in the NRS-formed precipitate dissolved (52 percent of recovered LPS and 52/53 of precipitate) leaving only about 1 percent of the recovered, precipitated LPS undissolved, while most of the LPS that precipitated from APRS (59 percent of recovered LPS and 59/64 of

• the precipitate) did not dissolve in saline.

Dissolution of the saline-insoluble precipitates followed by SDS-PAGE analysis indicated that the APRS-formed precipitated complex contained a

newly identified protein having an apparent relative molecular weight of about 60,000. That new protein was found to be a glycoprotein by staining with periodic acid-Schiff stain, and is referred to

5 hereinafter as gp60. Tobias and Ulevitch (1983) J. Immunol. 131:1913. ^ ts Data for the time-dependent shift of density and LPS precipitability from NRS and APRS showed similar time courses. In addition, SDS-PAGE analysis

10 of precipitates taken at various times after admixture of LPS with APRS showed the gpβO material as well as possibly two other proteins of molecular weights of about 57,000 and about 79,000 interact with LPS to modify LPS/HDL binding kinetics. Tobias

15 and Ulevitch (1983) J. Immunol. 131:1913.

It thus appeared that at least the before-described gp60 material was involved in mediating the binding of LPS to HDL. Subsequent work, discussed hereinafter, has however shown that

20 that gp60 material is not the. substance that retards binding between HDL and LPS.

It would be beneficial if a glycoprotein shown to be present in acute phase rabbit serum (but misidentified in the earlier reported work) could be

25 added to the normal or acute phase serum of other animals such as cattle, swine and man that suffer from gram-negative bacterial ^* infections to slow or retard the binding between LPS and HDL, and thereby provide for faster clearing of the LPS from the blood

30 and thus from the infected host. Brief Summary of the Invention

The present invention relates to mitigation of the effects on an infected animal host of lipopolysaccharides secreted by gram-negative

35 bacteria that infect the host animal.

One aspect of the invention constitutes a therapeutic composition for introduction into the bloodstream of an animal host that is susceptible to infection by lipopolysaccharide-secreting gram-negative bacteria. The composition comprises a purified glycoprotein that is dispersed in a liquid, physiologically tolerable diluent. The purified glycoprotein: (a) is a material that is present in the acute phase serum of the animal host, but is substantially absent from the normal serum of the host; i.e., it is present at a concentration of less than about 0.5 micrograms per milliliter of normal serum; (b) binds to a gram-negative bacterially-secreted lipopolysaccharide when the purified glycoprotein and lipopolysaccharide are admixed ^n vitro in normal serum ^' of the animal host; (c) retards the iτi vitro binding of the lipopolysaccharide to high density lipoprotein present in the normal serum of the host animal; and (d) is substantially homogeneous. The purified glycoprotein is often referred to herein as lipopolysaccharide binding protein (LBP) . The effective amount of the purified glycoprotein is an amount that is sufficient, when administered as a unit dose to the animal host, to provide an animal host serum level of the glycoprotein that is in excess of the level present a-t a time immediately prior to treatment of the animal and is at least an amount sufficient to retard binding in an ^n vitro determination of the lipopolysaccharide endotoxin to high density lipoprotein present in the normal serum of an animal of the same species as the animal host when the lipopolysacchride is present in the normal serum of that jln vitro determination at a concentration of 10 micrograms per milliliter.

For use in humans, the purified glycoprotein is preferably isolated from human serum and is an alpha ₂ ,beta,-glycoprotein having a molecular weight of about 60,000, a sedimentation coefficient of about 4 S, and contains about 25 weight percent carbohydrate. That carbohydrate includes galactose, mannose, glucosamine, sialic acid and fructose moieties and is free from galactosamine moieties. For animals other than humans, the glycoprotein is preferably a homolog of the above-described glycoprotein.

A method for treating an animal to mitigate the effects of an infection caused by lipopolysaccharide-secreting gram-negative bacteria constitutes another aspect of the present invention. In accordance with that method, a unit dose of the before-described therapeutic composition is introduced into the bloodstream of the animal to be treated. The unit dose preferably contains about 0.3 to about 5 milligrams per kilogram of treated animal body weight.

A further aspect of the present invention contemplates a method of assaying an animal body sample for the presence of a lipopolysaccharide endotoxin secreted by gram-negative bacteria. In this method, an aliquot of an animal body sample is provided, and is admixed with an unmasking reagent to to unmask any endotoxin present in the sample aliquot. The aliquot containing unmasked endotoxin is admixed in an aqueous medium with a purified glycoprotein as described hereinbefore. The admixture so formed is maintained for a predetermined time period sufficient for the purified glycoprotein to react with lipopolysaccharide that may be present in the aliquot and form a complex. The presence of a

complex formed between the glycoprotein and lipopolysaccharide is then determined, the presence of such a complex indicating that the lipopolysaccharide was present in the body sample aliquot.

The above assay method can be carried out using techniques analogous to those of receptor-ligand assays such as antibody-antigen assays wherein the purified glycoprotein is treated as the receptor and the lipopolysaccharide is the ligand. The assay method can also be utilized in a centrifugal density gradient assay where the presence of a formed complex can be determined by its density relative to the densitites of the admixed protein or glycoprotein.

Still another aspect of the invention consists essentially of a synthetic polypeptide corresponding in sequence to all or a portion of the amino-terminal 39 residues of the rabbit (lapine) glycoprotein whose amino residue sequence, from left to right and in the direction from amino-terminus to carboxy-terminus, is:

10

Thr-Asn-Pro-Gly-Leu-Ile-Thr-Arg-Ile-Thr-

11 20

Asp-Lys-Gly-Leu-Glu-Tyr-Ala-Ala-Arg-Glu-

21 30

Gly-Leu-Leu-Ala-Leu-Gln-Arg-Lys-Leu-XXX-

21 39

Gly-Val-Thr-Leu-Pro-Asp-Phe-Asp-Gly-

wherein residue number 30 (XXX) is indeterminate, and believed to be asparagine (Asn) .

The polypeptide of this invention can contain a sequence of about 6 to about 39 residues corresponding to the above sequence, and more preferably contains about 10 to about 25 residues. The present invention provides several benefits and advantages. One benefit is that the invention provides a treating composition that helps mitigate the effects of gram-negative bacterially secreted lipopolysaccharides on a treated animal.

Another benefit is that the above-mitigation can be effected by a composition that acts upon the . endotoxin rather than on the infecting bacteria so that useful gastrointestinal flora need not be destroyed during therapy.

A salient advantage of the present invention is that the therapeutic composition is preferably free from drugs usually used to treat gram-negative infections. Thus, the problems often associated with such drug treatments such as drug-resistance of the infecting gram-negative pathogenic 'bacteria can be obviated.

Still further benefits and advantages of the present invention will be apparent from the description that follows.

Brief Description of the Drawings

In the drawings forming a portion of this disclosure:

Figure 1 contains two graphs that illustrate the observation and quantitation of the 1.3

grams/cubic centimeter (g/cc) complex (C1.3) formed between the lipopolysaccharide endotoxin (LPS) secreted by Salmonella minnesota Re595 and human serum components in which the serum was collected before and after acute phase induction.

The graph of Figure la is a plot of individual cesium chloride (CsCl) density gradients showing the ability of normal (B) and acute (A) human sera, respectively, to form C1.3. The ordinate of that graph is in counts per minute (CPM) times

10 -3 observed from 3H-LPS in each fraction collected from the gradient. The abscissa lists the fraction numbers in which the counts were observed.

Figure lb shows the time dependence of (i) the ability of human serum to form C1.3 after acute phase was induced by etiocholanolone (A) and (ii) the C-reactive protein (CRP) concentration in human serum after similar acute phase induction ( ■) . The left-hand ordinate is in units of percent... H-LPS present as C1.3, while the right-hand ordinate shows the concentration of CRP in nanograms per milliliter (ng/ml) . The abscissa is in units of hours after acute phase induction.

Figure 2 is a graph showing plots of the kinetics of Re595 LPS-HDL complex formation in normal human serum that was collected immediately prior to injection of etiocholanolone (A) and in acute phase serum that was collected 32 hours after the etiocholanolone injection (D). The ordinate is in units of log(100-percent H-LPS present as C1.3) , while the abscissa is time in minutes after admixture of the -Η-LPS with either serum. Arrows indicate the first order one-half times (t, ,~ ) in minutes (min) for the respective plots.

Figure 3 is a photograph of a polyacrylamide gel eleσtrophoresis analysis carried out in the presence of sodium dodecyl sulfate (SDS-PAGE) as described by Laemmli (1970) Nature (London) 222:680. A 5 percent stacking gel was used in conjunction with a 10 percent separating gel. Protein-containing bands were visualized with Coomassie blue. Apparent relative molecular weight markers are shown in the right-hand lane. Those markers were phosphorylase B (94,000), bovine serum albumin (67,000), ovalbumin (40,000), and soybean trypsin inhibitor (30,000). The position of the gp60 material described herein and in Tobias and Ulevitch (1983) J. Immun. 131:1913 is indicated on the left by the designation "gp60- ^π . Protein preparations fractionated by the

SDS-PAGE analysis of this figure were: Lane A-rabbit euglobulin precipitated from APRS with LPS; Lane B-euglobulin precipitated from normal human serum (NHS) ; Lane C-euglobulin precipitated from NHS with LPS; Lane D-euglobulin precipitated from acute phase human serum (APHS) ; Lane E-euglobulin precipitated from APHS with LPS; and Lane F-molecular weight markers.

Figure 4 is a graph showing the .in vivo clearance from rabbits of H-LPS premcubated with delipoproteinated NRS ( ■ ) and APRS (A). Each point in the graphs is the averaged value from 4 to 6 normal rabbits. A catheter was placed in a femoral artery of each rabbit from which blood samples could be taken at desired times.

30 Micrograms (ug) of ³ H-LPS in 3 milliliters (ml) NRS or APRS that additionally contained 20 millimolar (mM) ethylenediaminetetraacetiσ acid (EDTA) . The resultant admixtures were maintained for a time

period of 10 to 30 minutes at room temperature to provide the preincubation. The preincubated admixtures were then injected into the rabbits (time zero) . Blood samples were taken at the times indicated and were allowed to clot. Sera were then collected and assayed by liquid scintillation. The ordinate of the graph is in units of percent of the initial counts per minute (CPM) detected in the blood, while the abscissa is in minutes after the injection of the radiolabeled LPS.

Figure 5 illustrates examples of the NRS reconstitution assay for LBP activity in which mixtures of APRS with NRS (O) or reconstitution of NRS with LBP (t) are shown. Details are given hereinafter.

Figure 6 illustrates CsCl density gradient analyses for C1.3 formation in delipoproteinated APRS (A, upper panel) pr delipoproteinated NRS (B, lower panel) reconstituted with lipoproteins from NRS (solid squares) or APRS (open squares) .

Figure 7 illustrates ion exchange chromatography on Bio-Rex 70 resin of APRS (A, upper panel) or NRS (B, lower panel) . Fractions pooled are denoted by the horizontal lines and letters with the graph. Solid lines, absorbance at 280 nanometers (nm) ; broken lines, conductivity.

Figure 8 illustrates CsCl density gradient analyses of LBP activity in pools from Bio-Rex 70 chromatography of APRS (left set of panels) or NRS

(right set of panels) . Pools tested for activity are identified in Figure 7. Not all gradients yielded exactly the same numbers of fractions. Thus, the x-axis, "relative fraction position" is used to standardize display of the gradients. A value of

zero (0) represents the bottom of the gradient and 1 the top of the gradient.

Figure 9 illustrates ion exchange chromatography on Mono-Q resin of Pools C from • Bio-Rex 70 chromatography of APRS (A, upper panel) or NRS (B, lower panel). Solid lines, absorbance at 280 nm; broken lines, molarity of ammonium sulfate. One milliliter (ml) fractions were taken using a 1 milliliter per minute (ml/min) flow rate. Figure 10 shows CsCl density gradient analyses of LBP activity in fractions eluting at 20 minutes from Mono-Q chromatography as shown in Figure 9. Panel A, fractions collected from APRS; Panel B, fractions collected from NRS. Figure 11 illustrates SDS-PAGE analyses of chromatography fractions. Lanes 1, 2 and 3 are pools A, B and C from Bio-Rex 70 chromatography of APRS, respectively. Lanes 4, 5 and 6 are pools A, B and C from Bio-Rex 70 chromatography of NRS, respectively. Lane 8 is from the 20 minute fraction from Mono-Q chromatography of pool C, APRS. Lanes 7 and 9 are molecular weight markers having the following apparent molecular masses in kilodaltons (kD) : 94 kD, 67 kD, 43 kD, 30 kD. Figure 12 shows further SDS-PAGE analysis of the 58 kD (lane 1) and 60.5 kD (lane 3) proteins separated electrophoretically from the mixture (lane 2) obtained after Mono-Q chromatography. Details are given hereinafter. Figure 13 illustrates CsCl density gradient analyses for LBP activity of immunopreσipitate supernates obtained from mixtures of APRS or NRS and rat anti-LBP antiserum. APRS, no antiserum, 0; APRS, 1.7% (v/v) antiserum, ; APRS, 4.6% (v/v) antiserum, ; APRS, 14% (v/v) antiserum, ; NRS, 14% (v/v) antiserum, •.

Figure 14 shows SDS-PAGE analyses of immunoprecipitates obtained from APRS or NRS and anti-LBP antisera. Lane 1, molecular weight (94 kD, 62 kD, 43 kD, 30 kD) markers; lane 2, NRS plus 14% (v/v) pre-immune serum; lane 3, NRS plus 14% (v/v) antiserum; lane 4, APRS plus 14% (v/v) pre-immune serum; lane 5, APRS plus 14% (v/v) antiserum; lane 6, APRS plus 4.6% (v/v) antiserum; lane 7, APRS plus 1.7% (v/v) antiserum; lane 8, APRS plus 0.6% (v/v) antiserum; lane 9, purified LBP.

Figure 15 shows SDS-PAGE analyses of and 3 quantitation of H-LPS immunoprecipitates obtained from mixtures of APRS or NRS and anti-LBP antisera.

₃ Reaction mixtures and H-LPS precipitation are shown below SDS-PAGE gel lanes. Lane 6 contained CRP.

Figure 16 illustrates SDS-PAGE analysis of

125 I-ASD-LPS reaction mixtures. Lanes 1-3 contained samples of the reaction mixtures in which samples were applied directly to the gels. Lanes 4-7 contained immunoprecipitates of the reaction mixtures precipitated with 14% (v/v) anti-LBP antiserum. Each pair of lanes, a and b, represents the Coomaassie Blue stained gel and the autoradiographic print, respectively. Lane 1, ¹²⁵ I-ASD-LPS; lane 2, ^ώJ I-ASD-LPS pre-photolysed and then admixted with immunopurified anti-LPS; lane 3, ⁵ I-ASD-LPS photolysed with immunopurfied anti-LPS; lane 4, pre-photolysed ¹²⁵ I-ASD-LPS with NRS; lane 5, ¹²⁵ I-ASD-LPS photolysed with NRS; lane 6 pre-photolysed ¹²⁵ I-ASD-LPS with APRS; lane 7, ¹²⁵ I-ASD-LPS admixed with APRS; and lane 8, molecular weight markers, 94 kD, 43 kD, 30 kD. Detailed Description of the Invention I. INTRODUCTION The present invention contemplates a therapeutic composition for treating an animal host

that is susceptible of gram-negative bacterial infection, and methods related thereto.

The therapeutic compositions and associated methods of this invention relate to the finding ^" that a glycoprotein present in acute phase serum, but substantially absent from normal serum retards the binding of lipopolysaccharide endotoxins to the high density lipoprotein present in blood serum.

During the first few days following an insult to an animal, a vast number of systemic and metabolic changes occur that are referred to as the acute phase response. Kushner (1982) Ann. N. Y. Acad. Sci. 389:39. Insults leading to an acute phase response include tissue-injuring infection, surgical or other trauma, drug-related effects, burns, tissue infarction and various idiopathic inflammatory states.

The liver is particularly affected during the acute phase response, and causes a rise in concentration of a large number of plasma proteins that have been grouped together as acute phase plasma proteins. Proteins whose concentrations rise by as much as 25 percent have been included in the group of acute phase plasma proteins.

The best studied acute phase proteins rise in concentration still more. Among those proteins are ceruloplasmin and complement component C3 whose concentrations increase by about 50 percent; alpha- _^ -acid glycoprotein, alpha-_-antitrypsin, alpha,-antichymotrypsin, fibrinogen and haptoglobin whose concentrations increase about two to about four fold; and C-reactive protein (CRP) and serum amyloid A protein (SAA) whose concentrations usually increase several hundred times.

Serum obtained from an animal in an acute phase response is referred to as acute phase serum (APS) .

Animals that are free from infection, or tissue injury as described before are referred to as normal animals. Serum from such a normal animal is referred to as normal serum (NS) . II. THE GLYCOPROTEIN AND ITS USES A. The Glycoprotein (LBP) The compositions and methods of the present invention utilize a purified glycoprotein that is often referred to herein as lipopolysaccharide binding protein (LBP) .

The term "purified glycoprotein" is used herein to mean that the glycoprotein (LBP) moves as a single band in SDS-PAGE analysis. Preferably, the purified glycoprotein contains no more than about 30 weight percent of other proteinaceous material that is stainable by Coomassie blue, more preferably no more than about 20 weight percent of such other material, and most preferably no more than about 10 weight percent of such other material. The foregoing percentages are based on the total weight of proteinaceous material present in the single band. The glycoprotein is present in acute phase serum (APS) of animals. In humans, the glycoprotein is present- in an amount of about 5 to about 10 micrograms per milliliter (ug/ml) of APS.

The glycoprotein is substantially absent from normal serum (NS) of such animals. By

"substantially absent", it is meant that less than about 0.5 ug/ml of the glycoprotein is present in such sera.

Since the concentration of the useful glycoprotein rises many-fold from substantial absence

in NS to an identifiable presence in APS, the useful glycoprotein can be classed as an acute phase protein, as discussed hereinbefore.

The useful purified glycoprotein can be further identified by its binding to LPS secreted by gram-negative bacteria when the purified glycoprotein and LPS are admixed jji vitro in a normal animal serum (e.g., a non-acute phase serum) such as that of an animal host to be treated, as is discussed hereinafter. The binding of the glycoprotein to LPS can be assayed in several ways as is also discussed hereinafter. However, the centrifugation density gradient method described hereinafter and in Tobias and Ulevitch (1983) J. Immunol. 131:1913, whose disclosures are incorporated by reference, is preferred.

The purified glycoprotein, also retards the in vitro binding of LPS to high density lipoprotein that is present in a normal animal serum. Again, the method of determining the binding rate retardation caused by the glycoprotein can be assessed in several manners. However, the centrifugation density gradient technique that assesses rates of density shifts from 1.33 or 1.30 to less than 1.2 g/cc described in Tobias and Ulevitch (1983) J. Immunol. 133 :1913 is preferred.

The purified glycoprotein, while being pure relative to other proteins and glycoproteins generally, is also substantially free from other acute phase proteins such as those mentioned hereinbefore. Particularly absent are CRP, SAA, murine serum glycoprotein gp70 and their homologs obtained from other animal species.

The purified glycoprotein has a molecular weight of about 55,000 to about 70,000. Exemplary of

such materials are the newly identified lapine glycoprotein having an apparent relative molecular weight of about 60,000 and the human glycoprotein having an apparent relative molecular weight of about 59,500, both of which are discussed further hereinafter. Apparent relative molecular weights (masses) , "M " are hereinafter referred to as "molecular weights" or "molecular masses".

It is believed that the useful human glycoprotein is the material first reported by

Iwasaki and Schmid (1970) J. Biol. Chem. 245:1814 and later noted by Schwick and Haupt, Chapter 3, "Human Plasma Proteins of Unknown Function", in The Plasma Proteins IV, Academic Press, Inc. (1984) page 196, both of which disclosures are hereby incorporated by reference. Amino acid and carbohydrate analyses of the glycoprotein are provided in Table II while further physical properties and solubilities in various media are found in Table I of the Iwasaki and Schmid paper.

Among the characteristic data reported by both sets of authors are that the protein is an alpha^^eta^-glycoprotein that was isolated from Cohn Fraction VI. The glycoprotein has a sedimentation coefficient of 4 Svedbergs (S) . The glycoprotein was denominated 4SGP by Schwick and Haupt and that designation is used herein.

Both groups of authors also reported that 4SGP is a single-chain glycoprotein having about 25 weight percent carbohydrate. The carbohydrate includes galactose, mannose, glucosamine, sialic acid and fructose moieties, but is free of galactosamine moieties.

4SGP is further reported to have an Arg (Arginine) residue at its amino-terminus and a Ser

(Serine) residue at its carboxy-terminus, and to be electrophoretically homogeneous at pH values of 8.6 and 5.0. Its isoelectric point is reported to be at pH 4.0. 4SGP was reportedly isolated from a pool of normal serum. It was reported by Iwasaki and Schmid to be present at an estimated amount of 20 milligrams per 100 liters of human serum. That amount corresponds to about 0.2 micrograms per milliliter of normal serum, and is thus substantially absent from normal serum using the before-mentioned definition.

Several studies have reported homologies among the acute phase serum proteins across many animal species. See, for example Baltz et al. (1982) Ann. N.Y. Acad. Sci. 389:49 as to phylogenetiσ aspects of CRP and serum amyloid P component (SAP) .

It is believed from the concentrations of the glycoprotein in human, and rabbit sera; the formation of complexes having a density of 1.3 g/cc in APS of humans, rabbits and mice that respond to

LPS; the observed binding retardation between LPS and HDL shown in rabbits and humans; and the before-mentioned acute phase serum protein homologies, that a protein homologous to 4SGP is present in all animals that are susceptible or respond to LPS and can mount an acute phase response.

B. Therapeutic Composition

One aspect of the present invention is a therapeutic composition for introduction into an animal host such as man, cattle, swine, poultry like chickens, sheep, rabbits, goats and the like that are susceptible (respond) to LPS secreted by gram-negative bacteria. The composition comprises an effective amount of the before-described purified

glycoprotein dispersed in a liquid, physiologically tolerable diluent.

Exemplary of such diluents are normal saline, phosphate-buffered saline, distilled or deionized water. Ringer's injection, lactated Ringer's injection and the like.

The purified glycoprotein is dispersed in an effective amount in that composition. As used herein, an effective amount is an amount that is sufficient, when administered as a unit dose to the animal host, to provide an animal host serum level of the glycoprotein that is in excess of the level present at a time immediately prior to treatment of the animal, and is at least an amount sufficient to retard binding, in an ji vitro determination, of the lipopolysaccharide endotoxin to high density lipoprotein (HDL) present in the normal serum of an animal of the same species as the animal host when the lipopolysaccharide is present in the normal serum of that ϋ vitro determination at 10 micrograms per milliliter. Preferably, the serum level of the glycoprotein is less than about 50 micrograms per milliliter.

Typical unit doses contain about 0.3 to about 5 milligrams of purified glycoprotein per kilogram of treated animal body weight. More preferably,' the unit dose contains about 1 to about 3 milligrams per kilogram.

Infections of gram-negative LPS-secreting bacteria that an above composition is useful in treating have already been discussed. However, treatment of infections caused by the genus Salmonella, and particularly by Salmonella minnesota and Salmonella typhimurium, is contemplated.

It is preferred that the purified glycoprotein be the principal treating agent of the composition. It is further preferred that the treating composition be free from drugs such as those discussed hereinbefore that are usually used to treat gram-negative bacterial infections (anti-bacterial agents) , and that such drugs not be used in a treatment regimen that includes a therapeutic composition of this invention. C. Treatment Method

Another aspect of the present invention constitutes a method of treating an animal such as those already mentioned to mitigate the effects of an infection caused by lipopolysaccharide-secreting gram-negative bacteria. This method comprises introducing a unit dose of the before-described composition into the bloodstream of an animal to be treated.

The composition can be introduced into the animal in a number of ways well known in the art. Exemplary introductions include injections given intramuscularly, intraperitoneally or intraparenterally, infusion as by drip bottle via a catheter, and the like. As was the case with the therapeutic composition, it is preferred that treatments of gram-negative bacterial infections using a composition of this invention be free of the before-mentioned anti-bacterial agents drugs that are usually used in therapies against gram-negative bacteria. Thus, the detriments of such therapies as discussed before can be avoided, the effects of the lipopolysaccharide endotoxin can be mitigated and the body's natural cellular and humoral defensive mechanisms can be utilized to fight the infection.

Of course, drugs that stimulate the body's natural cellular and humoral defenses can be beneficially included in the therapeutic composition or as a part of the treatment method regimen. D. Assay Methods

The purified glycoprotein described before is also useful in assay methods for the presence of lipopolysaccharide endotoxin secreted by gram-negative bacteria in a liquid animal body sample. Exemplary liquid animal body .samples include blood, serum, plasma, abdominal exudate, saliva, urine, cerebrospinal fluid, tears and joint fluid. Serum and plasma are preferred liquid animal body samples. In accordance with this method, an aliquot of a liquid animal body sample is provided, and is admixed with an endotoxin unmasking reagent to unmask any endotoxin present in the sample aliquot and form a liquid aliquot containing unmasked endotoxin. Unmasking reagents and techniques for their use are described in U.S. Patent No. 4,276,050 to Firca and Rudbach whose teachings are incorporated by reference. Briefly, exemplary unmasking reagents include aqueous solutions containing 2 percent Tween-80 [polyoxyethylene (20) sorbitan monooleate] , 2 percent dextransulfate,.3 percent sodium chloride, 2 percent ammonium thiocyanate, and most preferably 0.002 molar benzamidine and its biologically compatible acid addition salts. The unmasking reagent is preferably admixed with about an equal volume of body sample aliquot, and the composition is agitated gently.

The unmasked endotoxin-σontaining body sample aliquot is thereafter admixed with a before-described glycoprotein to form an admixture.

The admixture so formed is maintained for a predetermined time period sufficient for the purified glycoprotein to react and form a complex with lipopolysaccharide endotoxin present in the body sample; an exemplary time period being about 10 minutes. As is well known in the art, the maintenance or incubation time is a function, inter alia, of the amounts of both materials that are present in the admixture, e.g., LPS and glycoprotein (LBP) , with lower amounts typically requiring longer maintenance times.

The presence of the complex formed between the admixed, purified glycoprotein and lipopolysaccharide endotoxin is determined. The admixed, purified glycoprotein preferably includes a covalently-linked label that provides a means for indicating the formation of the complex, and preferably the amount of complex formed.

Exemplary of such covalently-linked labels ι o- ^* " _ι are radioisotopes such as H and -'I whose methods of covalent linkage to proteins are well known.

The discussion hereinafter in Section III relates primarily to formation of LPS/glycoprotein complexes in which the LPS bears a radiolabel ( ³ H) and in which the complex formation was assessed by measuring radioactivity in various fractions taken following CsCl density gradient centrifugation.

Similar assays can be performed using modifications of the above method in which the admixed glycoprotein contains the radiolabel.

Enzyme labels and their substrates can also be used. Exemplary enzymes and substrates include horseradish peroxidase normally used with hydrogen peroxide and an oxidative dye precursor such as

o-phenylenediamine and alkaline phosphatase that is typically used with £-nitrophenyl phosphate. Methods for covalently linking enzymes to proteins are also well known in the art. Substantially any assay method similar to those receptor-ligand assays used in immunological tests between antibody and antigen is also useful herein. Particularly preferred receptor-ligand assays are solid phase assays. Thus, in one method, a known amount of LPS as ligand is affixed to a solid matrix as a solid phase support. The liquid body sample aliquot and a known, excess amount of the purified, labeled glycoprotein over that of any amount of LPS expected in the sample are admixed to form a liquid phase admixture, and the liquid phase admixture is maintained as described before for a time sufficient for the purified, labeled glycoprotein to react and complex with LPS present in the body sample. An unmasking agent is preferably admixed with the body sample prior to admixture of the body sample and purified glycoprotein.

The maintained liquid phase admixture is then admixed with the solid phase to form a solid/liquid phase admixture. That admixture is maintained for a further time period sufficient for a further complex to form between the LPS of the solid phase support and the excess glycoprotein that did not bind with LPS present in the body sample. Separation of the phases and determination of the amount of solid phase-bound, labeled glycoprotein provides a measure of the amount of LPS in the aliquot and therefore in the body sample.

It is to be understood that the glycoprotein can also be used as the solid phase-affixed portion

and a known amount of labeled LPS can be added to the body sample aliquot.

It is often convenient to augment the amount of LPS originally present in the body sample. This can conveniently be done by culturing the body sample in a culture medium that promotes growth of gram-negative bacteria that produce the LPS and are present in the sample. After a suitable growth period such as one or two days, the culture medium is concentrated to provide a concentration of solids having a molecular weight greater than about 10,000 of at least about one milligram per milliliter. The aliquot can then be provided from the concentrated culture medium. The amount of LPS present in the liquid body sample can also frequently be augmented by concentrating the sample or its aliquot prior to use in the method. Convenient methods for performing such concentrations include air drying and lyophilization followed by redissolution in a smaller amount of solvent than the original volume, and ultrafiltration as described in Section V F. Ultrafiltration removes many unwanted salts and low molecular weight species. Useful solid matrices are well known in the art. Such materials include the cross-linked dextran available under the trademark SEPHADEX from Pharmacia Fine Chemicals (Piscataway, NJ) , agarose, beads of glass, polyvinyl chloride, polystyrene, cross-linked acrylamide, nitrocellulose or nylon-based webs such as sheets or strips, or the wells of a microtiter plate such as those made from polystyrene or polyvinyl chloride.

Latex particles useul in agglutination-type assays are also useful solid matrices. Such

materials are supplied by the Japan Synthetic Rubber Company of Tokyo, Japan, and are described as carboxy-functional particles dispersed in an anionic soap. Typical lots of such particles have an average diameter of 0.308 microns, and may have an average carboxy-functional group distribution of about 15 to about 30 square Angstroms per carboxy group.

Prior to use, the particles are reacted with a diamine such as l,3-diamino-2-propanol to form a plurality of amide bonds with the particle carboxy groups while maintaining free amine groups. The free amines are thereafter reacted with a dialdehyde such as glutaraldehyde and the glycoprotein to form Schiff base reaction products. The Schiff base reaction products are thereafter reduced with a water-soluble reductant such as sodium borohydride to provide a useful solid support.

E. LBP Polypeptides and Antibodies

The discussion that follows in Section. IV describes work with the lapine (rabbit) homolog of the human glycoprotein (LBP) discussed earlier herein and in the following Section III. The studies described in Section IV particularly describe the amino-terminal thirty-nine residues of lapine LBP. A synthetic polypeptide that consists essentially of about 6 to about 39 amino acid residues, and more preferably about 10 to about 25 amino acid residues, corresponding to all or a. portion of those 39 amino-terminal residues of rabbit (lapine) LBP constitutes another aspect of the present invention. The complete amino-terminal lapine LBP 39-residue sequence, from left to right and in the direction from amino-terminus to carboxy-terminus, is shown below.

Thr-Asn-Pro-Gly-Leu-Ile-Thr-Arg-I le-Thr- Asp-Lys-Gly-Leu-Glu-Tyr-Ala-Ala-Arg-Glu- Gly-Leu-Leu-Ala-Leu-Gln-Arg-Lys-Leu-XXX- Gly-Val-Thr-Leu-Pro-Asp-Phe-Asp-Gly

wherein the residue denominated XXX is indeterminate, and believed to be an asparagine (Asn) residue. A synthetic polypeptide of this invention corresponding in sequence to a portion of the amino-terminal 39 residues of lapine LBP that includes position 30 with its indeterminate residue (XXX) includes an Asn residue at that position.

Exemplary synthetic polypeptides of the first group include those having the sequences shown below, from left to right and in the direction from amino-terminus to carboxy-terminus,

(1-6) Thr-Asn-Pro-Gly-Ile-Thr; (7-13) Thr-Arg-Ile-Thr-Asp-Lys-Gly; (1-30) Thr-Asn-Pro-Gly-Leu-Ile-Thr-Arg-Ile-Thr-

Asp-Lys-Gly-Leu-Glu-Tyr-Ala-Ala-Arg-Glu-Gly-Leu-Leu-Ala- Leu-Gln-Arg-Lys-Leu; and

(1-39) as already shown, wherein "XXX" is asparagine (Asn) .

Exemplary more preferred synthetic polypeptides having about 10 to about 25 residues that correspond in sequence to a portion of the amino-terminal 39 residue sequence of the lapine LBP molecule, from left to right and in the direction from amino-terminus to carboxy-terminus, are shown below.

(1-13) Thr-Asn-Pro-Gly-Leu-Ile-Thr-Arg-Ile- Thr-Asp-Lys-Gly;

(8-21) Arg-Ile-Thr-Asp-Lys-Gly-Leu-Glu-Tyr-Ala- Ala-Arg-Glu-Gly;

(8-29) Arg-Ile-Thr-Asp-Lys-Gly-Leu-Glu-Tyr-Ala- Ala-Arg-Glu-Gly-Leu-Leu-Ala-Gln-Arg-Lys-Leu; and (26-39) Gln-Arg-Lys-Leu-Asn-Gly-Val-Thr-Leu-Pro-

Asp-Phe-Asp-Gly.

The parenthesized numerals before the two sets of polypeptide sequences above refer to the numbered positions from the amino-terminus of LBP to which those polypeptides correspond.

The full names for individual amino acid residues are sometimes used herein as are the well-known three-letter abbreviations. Single-letter abbreviations (code) are also utilized. The Table of Correspondence, below, provides the full name as well as the three-letter and single-letter abbreviations for each amino acid residue named herein [See, for example, L._Stryer, Biochemistry, 2nd ed., W. H. Freeman and Company, San Francisco, (1981) , page

16.]. The amino acid residues utilized herein are in the natural, L, form unless otherwise stated.

Table of Correspondence

Three-letter One-letter Amino acid abbreviation symbol

Alanine Ala A

Arginine Arg R

Asparagine Asn N

Aspartic acid Asp D Asparagine or aspartic acid Asx B

Cysteine Cys C

Glutamine Gin Q

Glutamic acid Glu E

Glutamine or glutamic acid Glx Z Glyσine Gly G

Histidine His H

Isoleucine He I

Leucine Leu L

Lysine Lys K Methionine Met M

Phenylalanine Phe F

Proline Pro P

Serine Ser S

Threonine Thr T Tryptophan Trp W

Tyrosine Tyr Y

Valine Val V

The term "corresponds" in its various grammatical forms is used herein and in the claims in relation to polypeptide sequences to mean the polypeptide sequence described containing only conservative substitutions in particular amino acid residues along the polypeptide sequence. In addition, one polypeptide "corresponds to another if antibodies raised to one polypeptide immunoreact with the other polypeptide.

The term "conservative substitution" as used above is meant to denote that one amino acid residue has been replaced by another, biologically similar residue. Examples of conservative substitutions include the substitution of one hydrophobic residue such as He, Val, Leu or Met for another, or the substitution of one polar residue for another such as between Arg and Lys, between Glu and Asp or between Gin and Asn, and the like.

In some instances, the replacement of an ionic residue by an oppositely charged ionic residue such as Asp by Lys has been termed conservative in the art in that those ionic groups are thought to

merely provide solubility assistance. In general, however, since the replacements discussed herein are on relatively short synthetic polypeptide antigens, as compared to a whole protein, replacement of an ionic residue by another ionic residue of opposite charge is considered herein to be "radical replacement", as are replacements between nonionic and ionic residues, and bulky residues such as Phe, Tyr or Trp and less bulky residues such as Gly, He and Val.

The terms "nonionic" and "ionic" residues are used herein in their usual sense to designate those amino acid residues that normally either bear no charge or normally bear a charge, respectively, at physiological pH values. Exemplary nonionic residues include Thr and Gin, while exemplary ionic residues include Arg and Asp.

A synthetic polypeptide of the present invention can be prepared by several solid and liquid phase techniques as are well known in the art. Preferably, however, the solid phase so-called "Merrified" method is utilized. Exemplary syntheses are discussed in U.S. Patents No. 4,545,931 and No. 4,544,500, whose disclosures are incorporated herein by reference.

A polypeptide of the present invention is preferably linked to an immunogenic carrier such as a protein as a conjugate for use in production of antibodies and antibody preparations. Immunogenic carriers are well known in the art and include keyhole limpet hemocyanin (KLH) , edestin, curcubin, human serum albumin, tetanus toxoid, sheep erythrocytes, polyamino acids such as poly(D-lysine D-glutamic acid) and the like.

Methods of linking the polypeptide to the immunogenic carrier to form the conjugate are also well known. Exemplary techniques include use of glutaraldehyde, a water-soluble carbodiimide, and those described in United States Patents No. 4,544,500 and No. 4,545,931.

The polypeptide-carrier conjugate is dissolved or dispersed in an aqueous composition of a physiologically tolerable diluent when used to induce the production of antibodies. Suitable physiologically tolerable diluents are well known in the art and include phosphate-bu fered saline (PBS) and 0.9 normal saline, and preferably also include an adjuvant such as complete Freund's adjuvant or incomplete Freund's adjuvant.

An effective amount of a conjugate- containing composition is introduced into a host animal such as a goat, rabbit, mouse, rat, horse or the like to induce the production (secretion) of antibodies to the polypeptide. -Effective amounts of immunogens useful for inducing antibody secretions in host animals are well known in the art. Methods of introduction into the host animal are also well known and are typically carried out by parenteral administration as by injection. A plurality of such introductions is normally utilized so that the host is hyperimmunized to the immunogenic polypeptide-containing conjugate. For example, weekly introductions over a one-to-two-month time period can be utilized until a desired anti-polypeptide antibody titer is achieved.

The antibodies so induced are thereafter recovered from the host animal. The recovered antibodies can be utilized as a preparation in the host serum as recovered, or can be in substantially

pure form; i.e., substantially free from host serum proteins, polypeptides and cellular debris. The latter antibody preparation can be conveniently prepared by passage of the recovered serum over ^" an affinity column as prepared from Sepharose 4B

(Pharmacia Fine Chemicals, Piscataway NJ) linked to the polypeptide of the conjugate, as is known.

The recovered preparation of antibodies immunoreacts with a synthetic polypeptide of the invention, such as the polypeptide of the conjugate, as well as with denatured rabbit LBP. Exemplary denatured LBP is that material that has been treated with 2-mercaptoethanol in SDS-PAGE analysis, and typically has a protein structure that is relatively more open or expanded than is that of native protein. In more preferred practice, the antibodies also immunoreact with native, non-denatured rabbit LBP as is present in APRS. Most preferably, the antibodies immunoreact with human LBP in denatured and/or non-denatured forms.

An antibody preparation of this invention prepared from a polypeptide as described above can be in dry form as obtained by lyophilization. However, the antibodies are normally used and supplied in an aqueous liquid composition in serum or a suitable buffer such as PBS.

The antibodies and polypeptides described herein are useful in assay methods for the determination of the presence and amount of rabbit and human lipopolysaccharide binding protein (LBP) . The polypeptides are particularly useful in these assays for blocking studies as in connection with the Western blot-type assays. Similar blocking can also be carried out in solid phase assays such as the ELISA-type studies that are also described hereinafter.

The antibodies are particularly useful in assays because of their unique specificity for im unoreacting with LBP. For example, a liquid body sample as described before can be admixed and contacted with the antibodies affixed to a solid matrix as a solid support to form a solid/liquid phase admixture. After passage of a predetermined maintenance time for the contact, the phases are separated to remove any materials that did not immunoreact. The non-immunoreacted antibodies are thereafter visualized or otherwise assayed as with a label linked to a polypeptide of this invention.

Solid phase assays, such as enzyme-linked im unosorbant assays (ELISA) , radio-labeled immunosorbant assays (RIA) or flurochrome-linked immunosorbant assays (FIA) are particularly contemplated.

Thus, the amount of LBP present in a liquid body sample is assayed in an embodiment of this invention. Here, a solid phase matrix such as the sides and bottom of a polystyrene or polyvinylchloride microtiter plate is provided. Antibodies of this invention are affixed to the solid matrix as by physical binding to form a solid phase support, as is known.

A predetermined amount of a liquid body sample such as plasma or serum to be assayed for LBP is admixed with the solid phase support to form a solid/liquid admixture. Exemplary predetermined amounts typically are about 25 to about 150 microliters neat, or more preferably present at a known dilution in an aqueous medium such as PBS that contains a total volume of about 25 to about 150 microliters.

That solid/liquid admixture is maintained for a period of time sufficient for LBP present in the sample to immunoreact with the solid phase-affixed antibody to form a solid phase-bound immunoreactant, and a liquid phase depleted of LBP. Exemplary maintenance (incubation) times typically range from about 5 minutes to about 6 hours, with the temperature of that maintenance typically being from about 4°C to about 40°C, with room temperature (about 20°C) being exemplary.

The solid and liquid phases are then separated as by rinsing to remove any materials from the sample that were not bound to the solid support. The solid phase containing the bound immunoreactant is retained for further use in the assay. The amount of solid phase-bound immunoreactant formed is then determined, and thereby determines the amount of LBP present in the assayed sample. That amount can be determined in a number of well known manners.

For example, where the amount of antibodies affixed to the solid matrix is known, an aqueous composition containing a polypeptide of this invention operably linked to a label (labeling means) can be immunoreacted with the unreacted affixed antibodies to form a second solid/liquid phase admixture.

The second solid/liquid phase admixture is maintained for a second time period sufficient for the labeled polypeptide to immunoreact with the previously unreacted solid phase-bound antibodies and form a second solid phase-bound immunoreactant. Maintenance times and temperatures useful for this maintenance step are similar to those described before.

The solid and liquid phases are again separated to remove any labeled polypeptide not present in the second solid phase-bound immunoreactant, and the amount of labeled polypeptides bound is determined. That determination is conveniently accomplished by use of a label such as an enzyme, flurochrome dye or a radiolabel operably linked to the second antibodies.

To obtain the most accurate results in such assays, it is preferred that non-specific binding sites on the solid supports be blocked after the solid support is prepared. Such non-specific site blockage can be achieved by known techniques such as by admixture of an aqueous composition of a protein free from immunoreaction in the assay such as bovine serum albumin (BSA) with the solid support prior to admixture of the liquid body sample. The admixture so formed is typically maintained for the time period and at the temperature described before. The solid phase having its non-specific binding sites blocked and the liquid phase are then separated as by rinsing, and the liquid body sample is admixed.

The discussion above has included a label operably linked to a polypeptide. The term "operably linked" is used herein to mean that the label molecules are linked or bonded to the polypeptide molecules so that antibody binding of the polypeptide is not substantially impaired nor is the action of the label substantially impaired. Thus, polypeptides containing operably linked label molecules bind to their antibodies and the linked label molecules operate to indicate the presence of the bound polypeptide in the immunoreactant.

For an ELISA, typically used enzymes operably linked to a polypeptide as a label include

horseradish peroxidase, alkaline phosphatase and the like. Each of those enzymes is used with a color-forming reagent or reagents (substrate) such as hydrogen peroxide and -phenylenediamine; and £-nitrophenyl phosphate, respectively.

Similar assays can also be carried out using a fluorochrome dye operably linked to a polypeptide as a label to signal the presence of polypeptides bound in an immunoreaction product The fluorochrome dye is typically linked by means of an isothiocyanate group to form the conjugate. Exemplary fluorochrome dyes include fluorescein isothiocyanate (FITC) , rhodamine B isothiocyanate (RITC) and tetramethylrhodamine isothiocyanate (TRITC) . In another technique, biotin operably linked to an a polypeptide is utilized as a label to signal the presence of the immunoreactant in conjunction with avidin that is itself linked to a signalling means such as horseradish peroxidase. A radioactive element such as H or - ^*63 i as utilized herein can also be operably linked to the polypeptide to form the label. In this instance, the radioactive decay of the element serves to quantify the assay of bound polypeptide, and thereby bound LBP.

III. LPS/HDL BINDING RETARDATION

As noted earlier, it was initially believed that a previously identified gp60 material isolated from acute phase rabbit serum (APRS) was responsible for the retardation of binding of rabbit HDL by exogenously supplied LPS. Tobias and Ulevitch (1983) J. Immunol. 131:1913. However, antibodies raised to the isolated gp60 when admixed with both NRS and APRS provided gp60-containing immunoreaσtants from both types of serum. The amounts of gp60 isolated from

both serum types using those antibodies were similar. In addition, the isolated gp60, when reconstituted with NRS, provided a serum in which no retardation of LPS binding to HDL was observed." Thus, it became apparent that the originally identified and isolated rabbit gp60 was not the material whose presence retards binding of LPS with HDL in rabbit sera. However, the original results that indicated the presence in APS of a material (i) that is substantially absent from NS (ii) that binds to gram-negative bacterially secreted lipopolysaccharide when both are admixed _iιι vitro in normal serum, and (iii) retards the jj vitro binding of LPS to HDL in serum is still thought to be correct, as is discussed hereinbelow.

Thus, further studies with the lapine system identified another protein of molecular weight of about 60,000 that is also glycosylated according to the periodic acid-Schiff stain. Preliminary data indicated that this glycoprotein is present in APRS in an amount of about 5-10 ug/ml, while current results indicate that it is present at about 30-35 ug/ml. This glycoprotein is substantially absent in NRS; i.e., the presence if any of this newly identified glycoprotein is in an amount of less than about one-twentieth of the amount present in APRS and is thus less than about 0.5 ug/ml. The newly found glycoprotein binds .iii vitro to LPS secreted by gram-negative bacteria such as Salmonella minnesota Re595 when the glycoprotein and LPS are admixed in

NRS, and it also retards the .in vitro binding of LPS to HDL in NRS.

The further results discussed below relate to work in the human system that parallels work in the lapine system. The human homolog to the newly

identified lapine glycoprotein appears to have a slightly higher molecular weight, but is functionally equivalent in its presence and substantial absence in acute phase serum and normal serum, respectively, its binding to LPS .in vitro in normal human serum, and in its retardation of binding of LPS to serum HDL in in vitro determinations.

Etiocholanolone is a naturally occurring steroid metabolite experimentally useful for inducing local inflammatory reactions and fever in man. These responses typically begin within 8-20 hours after injection and last 2-6 hours [McAdam et al. (1978) J. Clin. Invest. £1:390; Wolff (1967) Ann. Intern. Med. 6:1268] . Etiocholanolone also induces typical plasma acute phase reactant responses, for example, CRP and serum amyloid A (SAA) , within 24-48 hours after injection [McAdam et al. (1978) J. Clin. Invest. £1:390] .

Samples of. human serum collected at various times before and after eitocholanolone injection were surveyed for their ability to form an Re595-LPS protein complex of density 1.3 g/cc (C1.3) by admixing LPS with serum and maintaining the LPS serum admixture for a time period of 10 minutes prior to equlibrium density gradient centrifugation with

CsCl. This reaction time was chosen from preliminary kinetic studies as a reaction time that would permit most of the Re595-LPS to complex with HDL in normal serum but trap C1.3 before its LPS transfers to HDL. Results of this survey are shown in Figure 1.

Figure la shows a CsCl density gradient using sera collected before and 32 hours after etiocholanolone injection; the appearance of a form of Re595-LPS in the bottom third of the gradient when APHS is used is evident. The density of this form of

Re595-LPS was found to be 1.30g/σc by measurement of refractive index.

The amount of Re595-LPS at d=1.3g/cc as a function of time before or after etiocholanolone injection is shown in Figure lb. Figure lb also shows the CRP concentrations of the same samples surveyed for C1.3. The ability of serum to form C1.3 follows a time course similar to the acute phase CRP response that etiocholanolone induces [McAdam et al. (1978) J. Clin. Invest. 61:390].

The kinetics of Re595-LPS-HDL complexation of sera taken either before (NHS) or 32 hours after (APHS) etiocholanolone injection was also studied. To obtain these data Re595-LPS serum mixtures were sampled at various times after mixing and examined using CsCl gradients.

The formation of Re595-LPS-HDL complexes is plotted as for a first order reaction in Figure 2. Complexation of Re595-LPS with HDL in NHS has a one-half time (t ^) of about 7 minutes, whereas in APHS the reaction has a one-half time of about 52 minutes. This difference is virtually identical to that seen in rabbit serum where the one-half times are 2-4 minutes and 40-80 minutes for NRS and APRS, respectively [Tobias and Ulevitch (1983) J. Immunol. 13_1:1913] .

The solubility properties of R595-LPS are different in NHS and APHS. The euglobulin precipitate dialysis procedure [Tobias and Ulevitch (1983) J. Immunol. .121:1913] results in Re595-LPS distributed in the dialyzed supernate, washes, and final precipitate as shown in Table I, below.

TABLE I

₃ Percent Recovery of H-Re595 LPS from

Euglobulin Precipitation

LPS Fraction NHS (St.Dev.) ² APHS (St.Dev.) ²

Not precipitated 91.3 (4.7) 39.5 (6.5)

Recovered in washes 2.5 (0.25) 1.5 (0.1) In precipitate 6.3 (4.4) 59.0 (7)

Data are presented as a percentage of recovered LPS from two determinations. Overall recovery of reactant LPS was 52 percent.

2 St. Dev. = Standard deviation.

Using NHS or APHS taken from the same volunteer before and after etiocholanolone injection, about 52 percent of the input LPS was recovered. Of this recovered LPS only 6.3 percent was recovered in the final precipitate when NHS was used, while 59 percent was recovered in the final pricipitate when APHS was used. These results are thus similar to those discussed before from earlier work in the lapine system.

When the final euglobulin precipitates prepared from NHS and APHS in the absence as well as in the presence of Re595-LPS were examined by SDS-PAGE (Figure 3) , the precipitate from the LPS-APHS reaction mixture (lane E) contained a unique protein of apparent molecular weight 59,500 not found in any of the other precipitates.

Also run on the gel shown in Figure 3 is a lane (A) containing rabbit LPS-APRS euglobulin precipitate with "gp60" marked. gp60 is the

previously identified and isolated glycoprotein that precipitates from acute phase rabbit serum only in the presence of C1.3.

While the human protein appears to be slightly larger than rabbit gp60, this comparison may not be valid since the molecular weights of glycoproteins do not dependably correlate with their mobility in SDS-PAGE [Bretscher (1971) Nature New Biol. 231:229] . The other difference noted between the lapine and human acute phase sera is that APRS C1.3-forming activity is stable at 4°C for months, while APHS C1.3-forming activity is not stable even when sera are stored at -20°C for more than several weeks. It is concluded from these results, that as with the rabbit, a component of the human acute phase response interacts with Re595-LPS to reduce the rate of binding of LPS to HDL. The identity of the acute phase reaσtant in the human system is believed to be the before-discussed alpha ₂ ,beta- _] _-glycαprotein.

It is further believed that a homolog of the human protein such as the newly identified glycoprotein in the lapine system carries out a similar function in other animals. Addition of rabbit CRP to normal rabbit serum to levels characteristic of acute phase serum does not reconstitute the observed phenomena. Additionally, reconstitution of ultracentrifugally delipoproteinated NRS with HDL from APRS does not reconstitute the observed phenomena. These observations argue that the prototypical acute phase reaσtants CRP and SAAS are not involved [Galanos et al. (1969) Eur. J. Biochem. 9:245]. The remaining known acute phase reactants do not undergo large concentration changes between the normal and acute

state and are therefore, unlikely to be able to cause the qualitative differences between normal and acute phase serum observed.

Previous studies have shown that the mode of presentation of LPS; i.e., as a purified aggregated isolate or as an HDL complex, can significantly modify its endotoxic properties. An example of such presentation differences can be seen from Figure 4 wherein it can be seen that LPS preincubated with delipoproteinated APRS is more rapidly cleared from a rabbit than is LPS preincubated with delipoproteinated NRS. Thus, the acute phase response appears to incude a means for dealing with lipopolysaccharide endotoxins, and that means appears to be the human alpha ₂ ,beta,-glycoprotein discussed and described herein, and its homologs in other animals. IV. LAPINE LBP

A. NRS Reconstitution Assay Two examples of the NRS reconstitution assay for lipopolysaccharide binding protein (LBP) activity in the lapine (rabbit) system are shown in Figure 5. In control studies, no systematic dependence on the assay results were observed when smaller total assay volumes; i.e., 0.5 or 0.25 ml rather than the standrard 1 ml, were used. Addition of 0.05 percent CHAPS [an N-alkyl sulfobetaine derivative of a bile acid amide reported to have an empirical. formula of

^C 32 ^H 58 ^N 2 ^S0 7' ^tnat - ³ available from CALBIOCHEM, San Diego, CA] or 0.5 M urea to NRS did not qualitatively block reconstitution although the quantitative effects were not studied.

Reproducibility of the assay was found to be + 20 percent.

-45- B. Purification of LBP

Since LPS added to serum forms a complex with HDL, we determined whether an initial separation of the lipoprotein from APRS would be a useful first purification step. Mixing lipoproteins prepared from NRS or APRS with delipoproteinated NRS or APRS provided reconstituted sera with alternate sources of lipoproteins. As shown in Figure 6, C1.3 formed only when delipoproteinated APRS was used and was independent of the source of the lipoproteins used to form the reconstituted serum. Therefore, we turned to whole APRS as the starting material for LBP isolation.

Chromatography of APRS on Bio-Rex 70 said by its distributor to be a weakly acidic cation exchanger containing carboxylic acid exchange groups on a maσroreticular acrylic polymer lattice that is available from Bio-Rad, Richmond, CA proved to be a very effective initial step in purification of LBP. The absorbance profile of APRS and NRS passed over the column as well as the fractions pooled for analysis are shown in Figure 7.

When 400 ml APRS were passed over a 50 ml bed of Bio-Rex 70, the ability to form C1.3 was largely removed. This can be seen in Figure 8, where CsCl density gradient studies of the ability of the various pools to reconstitute NRS are shown. LBP activity was eluted only at salt concentrations above 300 millimolar (mM) NaCl, with the largest amount eluting in the 1 molar (M) NaCl wash; i.e., pool C, Figure 8. Washing the column with 3 M NaCl did not elute more LBP activity.

Assay results for the three pools eluted from the Bio-Rex 70 column are shown for a typical preparation of LBP in Table II, below. Overall, some 32 percent of the LBP activity was recovered for an increase in specific activity of 927 fold.

TABLE II Purification of LBP

LBP Total Total ^* Total

Volume Activity Prot. Sp.Act Activity Protein

(ml) ¹ (U/ml) ² (rag/ml) ³ (ϋ/mg) ⁴ U) ⁵ (__3__ ⁶ _

Starting Material

APRS 400 2.8 0.70 0.04 1120 28000

Bio-Rex 70 Pools

A 6 13.9 0.65 21.4 83 3.93

B 6 14.3 0.45 ^' 31.8 86 2.73

C 6 32.9 0.54 60.9 197 3.21

Overall for Bio-Rex: 37.1 366 9.87

Overall purification factor = 927

HPLC of Bio-Rex Pools

A 2 4.6 0.12 38.3 9.2 0.24

B 2 15.1 0.20 75.5 30.2 0.40

C 2 40.6 0.44 92.3 81.2 0.88

Overall for HPLC: 79.3 120.6 1.52

Overall purification = 1982 fold

Purfication of pool C through HPLC = 2307 fold

^Sample volume in milliliters (ml) .

² LBP activity in units per milliliter (U/ml) of the sample.

3 Protein (Prot.) in the sample in milligrams per milliliter (mg/ml) .

Specific activity (Sp.Act.) of LBP in the sample in units per milligram (U/mg) of protein present. Total activity of LBP in the sample in units (U) .

Total protein in the sample in milligrams (mg) .

When NRS rather than APRS was chroraatographed on the same column, an almost identical protein elution profile was obtained, as shown in Figure 7B, but none of the pooled fractions had any significant LBP activity (Figure 8) . SDS-PAGE gels of the "C" pools from APRS and NRS are shown in Figure 11, lanes 3 and 6, respectively. Further purification of LBP was accomplished with HPLC using a Mono-Q column, said to be a strong anion exchanger, specially designed for rapid, high resolution chromatography of proteins that is available from Pharmacia Fine Chemicals, Piscataway, NJ. When an aliquot of Pool C from Bio-Rex 70 chromatography of NRS or APRS was run on the Mono-Q column and eluted with a gradient of ammonium sulphate, the absorbance profile of the eluate was as shown in Figure 9. The profile obtained with Bio-Rex 70 Pool C derived from APRS shows a peak eluting near 20 minutes not seen in the profile obtained with Pool C derived from NRS.

Fractions containing the unique APRS-derived peak as well as the analogous fractions from the NRS material were assayed for LBP activity. As shown in Figure 10, the unique protein peak from APRS showed good LBP acitivity. All other fractions tested had no LBP activity.

An SDS-PAGE gel analysis of the LBP containing peak is shown in Figure 11, lane 8. From the mobilities of the two bands in lane 8, relative

to the standards in lane 9, the apparent relative masses (M ) (molecular weights) of the two proteins are 60,500 and 58,000 daltons (D) . Judging by staining intensity, the 60.5 kilodalton (kD) band usually comprises 90 percent of the mixture.

Final resolution of the two proteins present in the active pool from Mono-Q chromatography was accomplished by SDS-PAGE, slicing the two bands apart after staining and recovering the proteins by electroelution. The separation of the two bands is shown in Figure 12.

Amino acid sequence data, described below, suggests that both bands have very similar primary structures, arguing that both bands may be LBP. Both bands stain with periodic acid-Schiff reagent, thus they are both glycoproteins.

The above separation procedure, while carried out using lapine LBP is also effective for purification of human LBP from APHS. C. Amino Acid Sequence Data

Partial amino acid sequence data were obtained for two preparations of LBP, the mixture of 60.5 kD and 58 kD proteins obtained from Mono-Q chromatography and the 58 kD protein purified by SDS-PAGE. Initially, material collected from Mono-Q chromatography was sequenced. Since this material consists of 80-90 percent of the 60.5 kD protein, these data reflect the sequence of the major component recovered from the column. The amino acid residue sequence of the first

39 amino acids from the amino-terminus of lapine LBP were determined to be as shown below. With the exception of positions 1,36,38, and 39, all positions were determined in duplicate for two different preparations. Positions 1,36,38, and 39 were

identifiable in only one run of the sequenator. Position 30 did not yield an identifiable residue and may represent a site of glycosylation, most probably an asparagine (Asn) residue. Sequence data for the 58 kD minor protein was obtained for 36 residues.

The sequence of the 58 kD protein agreed completely with that of the mixture of proteins, even to the indeterminate residue 30.

1 10

Thr-Asn-Pro-Gly-Leu-Ile-Thr-Arg-Ile-Thr-

11 20

Asp-Lys-Gly-Leu-Glu-Tyr-Ala-Ala-Arg-Glu-

21 30

Gly-Leu-Leu-Ala-Leu-Gln-Arg-Lys-Leu-XXX-

21 39

Gly-Val-Thr-Leu-Pro-Asp-Phe-Asp-Gly

The 39 amino acid sequence was used to search for homologous sequences in the National Bioraedical Research Foundation protein sequence database using the Wordsearch program from the University of Wisconsin Genetics Computer Group. To distinguish between random and non-random matches found by the computer search, the Wordsearch program was submitted to a randomized sequence having the same composition as the peptide shown above. Those matches found with the LBP sequence were eliminated

from consideration whose "quality scores" were not better than the matches found with the randomized sequence.

This procedure resulted in two matches ^" to portions of sequences of previously reported proteins (as shown below) i.e., human influenza virus b hemagglutinin precursor (INFLUENZA) [Krysal et al., Proc. Natl. Acad. Sci. USA 80:4527 (1983)] and baker's yeast glyceraldehyde 3-ρhosphate dehydrogenase (BAKER'S YEAST). Holland et al., J. Biol. Chem. 258:5291 (1983). (The hyphen in the second LBP sequence represents the indeterminate residue at position 30 from the amino-terminus.)

5 LITRITDKGLEYAAREGLLALQRKLXGVTLP 35 LBP

459 LAVLLSNEGIINSEDEHLLALERKLKKMLGP 489 INFLUENZA

8 RITDKGLEYAAREGL-LALQRKLXGV 32 LBP

2 RIAINGFGRIGRLVLRLALQRKDIEV 27 BAKER'S YEAST

The Wordsearch program was also used to look for homology between the sequences of LBP and rabbit CRP [Wang et al., J. Biol. Chem. iZ ^:1 3610 (1982)], human serum amyloid P [Mantzouranis et al., J. Biol. Chem. 260:7752 (1985)], human serum amyloid a [Sipe et al.. Biochemistry 24:2931 (1985)], Syrian hamster female acute phase protein [Dowton et al.. Science 228:1206 (1985)], human alpha-1-antichymotrypsin precursor [Chandra et al. , Biochem 22:5055 (1983)], human alpha-1 acid glycoprotein [Dente et al..

Nucleic Acid Res. 1 :3941 (1985)], and the major acute phase alpha-1 glycoprotein of the rat. Cole et al., FEBS Lett. 182:57 (1985). No significant homology was found with any of these acute phase reactants.

D. Depletion of APRS with Rat

Polyclonal Anti-LBP

When examined by radial immunodiffusion, rat antisera induced by introduction of whole, substantially purified lapine LBP were reactive with APRS and LBP-containing fractions of APRS isolated by Bio-Rex 70 and Mono-Q columns, but not with NRS or NRS fractions corresponding to the fractions isolated from APRS. The antisera were then tested for their ability to immunoprecipitate LBP and simultaneously remove LBP activity from APRS. In these studies, both NRS and APRS were admixed with immune and non-immune rat sera. After precipitation, the supernates were collected to determine their ability to form C1.3 (Figure 13) , indicating the presence or absence of LBP activity, and the precipitates were saved for analysis by SDS-PAGE (Figure 14) ..

When the supernates from the immunoprecipitates were assayed for their ability to form C1.3, the amount of Cl.3 observed was inversely proportional to the amount of immune serum added to the APRS (Figure 13) . Rat sera, immune or not, did not inhibit the binding of ³ H-LPS to HDL in NRS. When the constituent proteins of the immunoprecipitate were visualized by SDS-PAGE, only the admixtures of immune rat serum reacted with APRS yielded significant immunoprecipitates with bands corresponding to LBP (Figure 14, lanes 5-7). The intensity of the LBP band in the immunoprecipitates correlated positively with the amount of immune serum admixed (Figure 14, lanes 5-8), and correlated negatively with the ability of the residual supernate to form Cl.3 (see Figure 13).

Bands other than LBP observed in the immunoprecipitates are attributable to rat albumin

and rat heavy and light immunoglobulin chains. The combination of NRS with immune rat serum did yield very weak bands ^' corresponding to rat immunoglobulin chains and to LBP (Figure 14, lane 3). Non-immiine rat serum reacted with either NRS or APRS yielded only bands corresponding to albumin (Figure 14, lanes 2 and 4) .

Thus, immunoprecipitation of LBP and APRS decreased the ability of APRS to form Cl.3 in a dose dependent manner.

E. Interaction of LPS with LPB Two types of studies were performed to determine whether LPS and LBP interact directly; immunoprecipitation of H-LPS as Cl.3 in APRS by anti-LBP sera, and delivery of 125I to LBP by photolysis of ¹²⁵ I-ASD-LPS ⁵ [LPS coupled to suϊfosuccinimidyl-2-(p_-azido sulicylamido)-l,3'- dithiopropionate that was radio-iodinated after coupling] as Cl.3 in APRS. In the immunoprecipitation studies, H-LPS and APRS (or NRS) were admixed, allowed to react and form a complex at 37 degrees C for 10 minutes, and the admixture was then cooled to zero degrees C before rat anti-LBP was added. The H-LPS content of a portion of the immunoprecipitate was determined by liquid scintillation counting while the remainder of the precipitate was taken for SDS-PAGE. Since LPS may bind non-speσifically to immune precipitates [Ginsberg et al., J. Immunol. 120:317 (1978)], immunoprecipitation of rabbit C-reactive protein

(CRP) from APRS by goat anti-CRP in the presence of ³ H-LPS was used as a control experiment.

The SDS-PAGE analysis of the immunoprecipitates is shown in Figure 15 together with the data for ³ H-LPS recovered with the

immunoprecipitate. The results show a clear positive correlation between LBS precipitation and H-LPS precipitation.

However, in preliminary studies,

125I-ASD-LPS was shown to cosediment with underivatized LPS in a CsCl gradient, form Cl.3 with

APRS, and bind to HDL in APRS more slowly than in NRS,

Further evidence for interaction of LPS with LBP was obtained through the use of 125I-ASD-LPS. APRS (or NRS) was admixed and allowed to react with ¹²⁵ I-ASD-LPS for 5 minutes at 37 degrees C in the dark to form a complex, then chilled to stop further transfer to HDL, and photolysed. Anti-LBP antiserum was then added to collect LBP for SDS-PAGE for autoradiographic analysis. The results as well as the results of control studies, are presented in Figure 16.

Lane 1 of Figure 16 shows that ¹²⁵ I-ASD-LPS photolysed in 20 mM EDTA, 150 mM NaCl, pH 7.4 does not yield a Coomassie blue stainable band (la) and the ¹²⁵ ι runs with the dye front (lb) . Lanes 2 and 3 show that ¹²⁵ I-ASD-LPS mixed with immunopurified anti-LPS labels immunoglubulin heavy chains if photolysed after admixing (lane 3) , but only very weakly if photolysed also before admixing (lane 2). For lanes 1-3, aliquots of the reaction mixtures were applied directly to the gels. For lanes 4-7, the reaction mixtures were immunoprecipitated with anti-LPS before application to the gel.

Lanes 4 and 5 show that ¹²⁵ I-ASD-LPS, whether photolysed only after admixture with NRS (lane 5) or also photolysed before with NRS did not label any material immunoprecipitable with anti-LBP; i.e., lanes 4b and 5b are clear. Lanes 6 and 7 show

that ¹²⁵ I-ASD-LPS admixed with APRS labels LBP strongly if photolysed after admixing with APRS (lane 7) , but labels LBP only weakly if also photolysed before admixing with APRS (lane 6) .V. V. MATERIALS AND METHODS

Part I - Related to Sections II and III A. Lipopolysaccharide Purification LPS, either biosynthetically tritiated or unlabeled, was purified from Salmonella minnesota Re595 as described by Galanos et al. Euro. J.

Biochem. (1969) ^{:245 and} modified by Ulevitch et al. (1981) J. Clin. Invest. £2 ^:827 « Briefly, 50 grams (g) of dried bacteria, prepared as described hereinafter, were admixed with 200 milliliters (ml) of extraction mixture [a monophasic solution containing aqueous phenol (90 g dry phenol + 11.0 ml H ₂ 0) , chloroform and petroleum ether (b.p. 40-60 degrees C) in a volume ratio of 2:5:8, respectively] . The admixture was then stirred for 5 minutes producing a fine suspension of whole bacteria. The suspension was then centrifuged (5000 rpm for 15 minutes in a Sorvall GSA rotor) to form a bacterial pellet and an LPS-containing supernatant. The pellet and supernatant were separated and the supernatant was filtered to remove any remaining bacteria or cellular debris. The bacterial pellet was subjected to the same extraction procedure one or two more times with each resulting LPS-containing supernatant being admixed with the first. Petroleum ether and chloroform were then removed from the pooled supernatants by rotary evaporation at 30-40 degrees C. To the resulting LPS-containing aqueous phenol solution was slowly admixed water until the LPS precipitated. The admixture was subsequently centrifuged (3000 rpm for

10 minutes in a Sorvall GSA rotor) so as to form a LPS pellet and supernatant.

The LPS pellet was separated from the supernatant and washed two or three times with about 5 ml of 80 percent phenol (w/v in H.O) by suspension and centrifugation. The LPS pellet was then washed three times with 5 ml of ether by suspension and centrifugation to remove any remaining phenol, and dried jj vacuo. The dried Re595 was then admixed with sufficient 20 mM EDTA (pH 7.5) to dissolve the LPS upon sonic oscillation in a model W-375 sonicator from Heat Systems-Ultrasonics, Inc., Plainview, NY. The LPS solution was dialyzed against 3 liters of sterile water for 72 hours, with a change of the dialysis bath every 12 hours, and then lyophilized. Fresh stock solutions of 5 mg/ml LPS were prepared by addition of the LPS to the appropriate buffer, followed by sonic oscillaiton at 25 degrees C. B. Bacteria Production

The LPS-producing bacterial strain Salmonella minnesota Re595, obtained from New England Enzyme Center, Tufts University, Boston, MA, was cultured in a growth medium containing Ardamine Z at 22.5 g/1, NZ amine NAK at 11.0 g/1, K ₂ P0 ₄ at 16.6 mg/1, K0 ₄ at 4.0 g/1 and Cerelose at 66.6 mg/1, all also obtained from New England Enzyme Center. Growth medium, typically one liter in a three liter flask with baffle plates, was inoculated with 20 ml of a confluent (plateau) S_. minnesota Re595 culture and inoculated at 37 degrees C with vigorous agitation for about 8 hours. Bacteria containing LPS were subsequently harvested by centrifuging the cultures to form a bacterial pellet and supernatant. The bacterial pellet was separated from the supernatant

and washed three times by resuspension in deionized water and repelleting by centrifugation. The final washed pellet was resuspended in about 50 ml of deionized water and lyophlized, typically yielding about 4.5 g dry weight bacteria per liter of culture.

To prepare S_. minnesota Re595 containing LPS

3 biosynthetically labeled with tritium ( H) , about

62.5 mCi of H-sodium acetate (Amersham, Santa Ana,

CA) were admixed with one liter of the growth medium just prior to inoculation. The bacteria were then cultured and harvested as previously described,

₃ typically yielding about 7-8 mCi H per 4.5 g dry weight bacteria.

C. Complex Formation Etiocholanolone was used to induce an acute phase response [McAdam et al. (1978) J. Clin. Invest. £1^390] in three human volunteers by intramuscular injection of 0.3 milligrams etiocholanolone per kilogram of volunteer body weight. Serum was collected at various times from 24 hours previous to etiocholanolone injection, to 120 hours after injection. Observation and quantitation of Re595-LPS complex formation in sera using CsCl gradients was as described in Tobias and Ulevitch (1983) J. Immunol. 131:1913.

Preparation of euglobulin precipitate from Re595-LPS serum reaction mixtures was accomplished as described by Tobias and Ulevitch (1983) J. Immunol. 131:1913. Briefly, the reaction was allowed to proceed for 2 minutes at 37°C, followed by stopping the reaction by rapid chilling in an ice bath. The chilled reaction mixture was dialysed overnight at 4°C versus 2.5 millimolar (mM) HEPES, 15 mM NaCl, at pH 7.65. The resulting precipitated euglobulin fraction was collected by centrifugation, washed

twice with lOmM HEPES, 140 mM NaCl, at pH 7.4, and was then suspended in distilled water for ease of handling. Re595-LPS in the supernate, washes, and

3 euglobulin precipitate was quantitated by H measurement.

D. Electrophoresis and CRP Analysis Polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate (SDS-PAGE) was performed using the recipes of Laemmli [Laemmli (1970) Nature (London) 222:680]. Measurement of serum C-reactive protein (CRP) levels were performed in the clinical laboratory of Scripps Clinic and Research Foundation following usual procedures.

E. Purification of the Acute Phase Reactant from Rabbit Serum

Acute phase rabbit serum was prepared as described in Tobias and Ulevitch (1983) J. Immunol. 131:1913. The newly identified gp60 material was purified from APRS as follows. The primary purification step was ion-exchange chromatography on Bio-Rex 70 (Bio-Rad Laboratories, Richmond, CA) . A 25 ml column of resin was equilibrated with 0.05 molar (M) sodium phosphate, 2 mM EDTA, pH 7.3 buffer (Pi/EDTA buffer) in the cold (about 4°C) . Thereafter, 200 ml of APRS was run through the column. The column was then washed first with Pi/EDTA buffer, and then with 0.22 M NaCl, 0.05 M sodium phosphate, 2 mM EDTA as a pH 7.3 buffer until the ultraviolet absorption of the eluate at 280 nanometers was less than 0.1.

A linear salt gradient of 30 ml each of 0.22 M and 0.5 M NaCl in Pi/EDTA buffer was then started followed by 50 ml of 1 M NaCl in Pi/EDTA. The desired glycoprotein-LPS binding (complex forming) activity primarily eluted at the end of the gradient with 1 M NaCl.

The active fractions were further purified by gel filtration using G-150 Sephadex (Pharmacia Fine Chemicals, Piscataway, NJ) in 5 mM sodium phosphate at pH 7.3 to remove low molecular weight contaminants. The activity eluted from the gel filtration column was very close to the elution position of BSA. Thus, the activity-bearing glycoprotein had an apparent relative molecular weight of about 60,000. The acute phase glycoprotein reactant bound to DE-52 cellulose (Whatman, Inc., Clifton, NJ) equilibrated with 5 mM sodium phosphate pH 8.3 and eluted in active form with 1 M NaCl. Salt gradient elution from the column provided a further purification step.

The presence of desired glycoprotein during the before-described purifications (Section D) was monitored by assaying the collected fractions for the ability of their components to form a complex with LPS. The presence of a complex-forming activity was ascertained by a centrifugal density gradient assay. F. Sucrose Density Gradient

Ultracentrifugation

The activity was recovered from a sucrose density gradient prepared with 5-20 percent sucrose and centrifugation for 2 hours at 45,000 RPM using a TV 865 rotor (DuPont Co., Instrument Products Biomedical Div., Newtown, CT) at an average value of 4 Svedbergs (S) . Each gradient so prepared was divided into 8 fractions. Each fraction so obtained was assayed for its ability to form a 1.3 g/cc complex using a CsCl density gradient method analogous to that described hereinafter using ^J H-LPS and APRS. These assays serve to confirm the roughly 60,000 molecular weight of the activity-containing glycoprotein.

G. Analysis of LPS Binding to

Newly Identified Glycoprotein

LPS binding to the newly identified glycoprotein in rabbit serum was carried out as ^" follows.

An aliquot of an animal body sample to be assayed for LPS binding activity was first concentrated, as necessary, to provide a composition containing about 1 mg/ml of solids having a molecular weight greater than about 10,000 using an Amicon ultrafiltration apparatus with a YM 10 membrane

(Amicon Corp., Scientific Systems Div., Danvers,

MA) . About 10 to about 250 microliters (ul) of the concentrated body sample were admixed with 0.5 ml of NRS containing 20 mM EDTA. At time zero, 5 ug of H-LPS was admixed with the above admixture. Ten minutes thereafter, t=10, 4.3 ml of normal saline containing 1.8 g of

CsCl previously maintained at 0°C was admixed with the H-LPS/body sample aliquot admixture.

The resulting admixture was centrifuged for a time period of 16 hours at 45,000 RPM using a TV865 rotor. The resulting density gradient was fractionated and the counts in fractions of varying densities were determined as is shown in Figure 1.

H. HDL Deficient Serum

High density lipoprotein (HDL) deficient

(depleted) serum, from either normal or acute phase rabbits or humans, was prepared by methods well known in the art. Briefly, the density of 35 ml of serum was adjusted to about 1.24 g/cirr by admixing 13.2 g of KBr. The serum/salt admixture was then centrifuged to gradient equilibrium in a high gravitational field, i.e., about 113,000 x gravity [48-60 hours at 40,000 rpm in a 60 ti rotor (Beckman

Instruments, Palo Alto, CA] . HDL, being less dense than the other components of the admixture, concentrates as a stable band with a yellowish color at the top of the gradient. The HDL band was separated from the gradient and discarded. The volume of the remaining serum/salt admixture was then adjusted to 35 ml and its density adjusted to 1.24

₃ g/cm by admixture of an appropriate amount of KBr.

The centrifugation and separation procedure was repeated. The resulting serum/salt admixture was then dialyzed against 0.9 percent saline to substantially remove KBr from the admixture. Finally, the admixture was adjusted to its original volume (35 ml) by admixture of an appropriate amount of 150 mM NaCl.

I. LPS Clearance Kinetics

The effects of acute phase reactants on the kinetics of LPS clearance from mammalian blood were studied in rabbits using biosynthetically labelled LPS ( H-LPS) . One hundred fifty micrograms of

³ H-LPS were admixed with 15 ml of either (1) HDL deficient APRS; or (2) HDL deficient NRS. The admixtures were incubated for 10-30 minutes at 20 degrees C to allow binding of ³ H-LPS to any acute phase reactants in the admixtures.

Three groups of 5 rabbits each were catheterized and the described admixtures were injected into the femoral vein and artery. Three milliliters of one of the above described admixtures were injected into the femoral vein of all rabbits in one group. Blood samples were taken from the femoral artery catheter at the time intervals shown in Figure 4. The serum was separated from each blood sample and the amount of H-LPS present in each serum sample was determined by liquid scintillation.

Part II - Related to SECTION IV

Materials:

₃ Biosynthetically tritiated LPS ( H-LPS) and unlabelled LPS were isolated from Salmonella minnesota Re595 as described previously. (Tobias et al.. Infect. Immun. 50.:73 (1985); Galanos et al. Eur. J. Biochem. 9:245 (1969). Rabbit blood was collected either by bleeding from the median ear artery or by heart puncture, allowed to clot at 37 degrees C for 2-6 hours and at zero degrees C overnight, centrifuged to remove clot fragments and cells, and the serum was stored frozen without preservative. Acute phase rabbit serum (APRS) was collected 24 hours after induction of an acute phase response by subcutaneous injection of 1 ml of 3% (W/V) silver nitrate in distilled water. Serum collected from non-induced rabbits was tested for "normality" before being used as normal rabbit serum (NRS) . The initial assay used was immunodiffusion versus antiserum to rabbit c-reactive protein (CRP) . Sera assaying negative for CRP were further assayed as described below to ensure a sufficiently rapid rate of binding of LPS to HDL. These precautions were instituted after observing that more than 50% of a batch of newly acquired rabbits had readily detectable acute phase reactants in their sera.

Polyclonal rat antisera to whole, substantially purified lapine LBP, as obtained from the two-column (Bio-Rex 70 and Mono-Q) purification procedure described herein, were raised in Lewis rats by intraperitoneal injection of each rat with 25 micrograms (ug) LBP in complete Freund's adjuvant, with 25 ug LBP in incomplete Freund's adjuvant at 3 weeks, and with 10 ug LBP in buffer at 6 weeks. Animals were bled by heart puncture under

Innovar-Vets anesthesia and serum collected as described above. Immunoprecipitation studies using these sera were performed by incubating rabbit serum together with varying volumes of antiserum for at least 3 hours at 37 degrees or 4 hours at 4 degrees. Precipitates were collected by centrifugation and washed twice with 50 mM phosphate buffer, 150 millimolar (mM) NaCl, 0.1% Tween-20 [polyoxyethylene (20) sorbitan monolaurate] , pH 7.4. Unfractionated lipoproteins and deliproproteinated sera were prepared by ultracentrifugation. To 35 milliliters (ml) of serum were added 13.23 grams (g) of KBr, after which the serum was spun at 40,000 RPM in a 60 Ti (Beσkman Instruments, Fullerton, CA) rotor for 36-60 hours at 4 degrees. After fractionation, protein assay, and cholesterol assay (CALBIOCHEM, La Jolla, CA) , the lipoproteins and serum proteins were separately pooled and dialysed extensively against 10 mM HEPES, 150 mM NaCl, pH 7.4. Finally the delipoproteinated sera and the lipoproteins were brought to 75 percent and 25 percent of the original serum volumes, respectively, by dilution or concentration as required. Delipoproteinated sera and lipoproteins were recombined in a 3:1 ratio, respectively, to prepare lipoprotein reconstituted sera.

LPS was coupled to sulfosuccinimidyl-2-- (£-azido salicylamido)-l,3'-dithiopropionate (Pierce) as described [Wollenweber et al. J. Biol. Chem. 260:5068 (1985)], and the resulting derivative

I C

(ASD-LPS) was radiolabelled with ^J - ^£ - ^, i using chloramine T [Ulevitch, Immunochemistry 15:157 (1978)] to yield ¹²⁵ I-ASD-LPS. The product had a specific activity of 7.1x10-' counts per milligram (cpm/mg) LPS from which the incorporation of ¹² 5ι

into LPS is calculated to be approximately 0.3 moles percent. Preliminary data indicate that

125 I-ASD—LPS co-sediments with LPS in CsCl gradients. 125I-ASD-LPS is quantitatively taken up by HDL and NRS and APRS, and ASD-LPS has the same mitogenicity as LPS when assayed with murine splenic B cells. Photolysis of ¹²⁵ I-ASD-LPS was accomplished using a Rayonet photochemical reactor (Southern N.E. Ultraviolet Co., Middletown, CT) equipped with General Electric F8T5.BLB lamps with a peak output at 370 nanometers (nm) . Reaction mixtures were exposed for 10 minutes on ice. METHODS:

A. General Sodium dodecyl sulfate polycrylamide gel electrophoresis (SDS-PAGE) with staining by Coomassie blue or periodic acid-Schiff reagent was performed by published procedures. Tobias et al. Immunol. 128:1420 (1982). All samples were reduced with 2-mercaρtoethanol prior to SDS-PAGE concentration of LPS, and its derivatives were determined using the ketodeoxyoctanoate assay [Cyubin et al.. Nature (Lond.) 186:155 (I960)] with LPS as standard. Protein concentrations were determined by either the Folin [Lowrey et al., J. Biol. Chem. 193:265 (1951)] or BCA (Pierce Chem. Co., Rockford, IL) reagents using bovine serum albumin as standard. All reactions of LPS or LPS derivatives with APRS or NRS were carried out at 10 milligrams per milliliter (mg/ml) unless otherwise noted.

B. Kinetics of LPS binding to HDL

The kinetics of LPS binding to HDL in serum were observed and quantitated by CsCl isopycnic density gradient ultracentrifugation. To 8 ml of rabbit serum were added 0.4 ml of 0.4 M EDTA at pH

7.4, and the mixture was warmed to 37 degrees C in a water bath. At time zero, 0.4 ml of 200 micrograms per milliliter (ug/ml) ³ H-LPS in 0.02 M EDTA, pH 7.4, were added. At suitable times, 1.0 ml aliquots of the reaction mixture ere removed and added to 3.8 ml of ice cold 2.81 M CsCl, 0.15 M NaCl. These samples were then spun to equilibrium for 16 hours at 45,000 RPM in a TV-865 rotor (DuPont Sorvall, Wilmington, DE.) at 0-4 degrees C. Following centrifugation, the gradients were fractionated, the refractive index measured if the density profile of

3 the gradient was to be determined, and the H-LPS in each fraction determined. The efficiency of measuring 3A-Re 595 LPS was found to be independent of the amount of CsCl in each vial. After graphing the H-LPS profile for each gradient, the amount of radioactivity in the body of the gradient; i.e., not bound to the HDL which floats at the gradient, was calculated as a,percentage of the radioactivity recovered in the entire gradient. A logarithmic of thi-s percentage as a function of the time of removal of the aliquot from the LPS serum reaction mixture yielded the half time for the binding of LPS to HDL. C. Reconstitution Assay for LBP Activity The basic method used during development of the purification procedure for LBP was a reconstitution assay in which fractions of acute phase serum were assayed for their ability to reconstitute "acute phase behavior" in NRS. The screening assay used was to mix a sample of the material to be tested with 1.0 ml of NRS at 37 degrees C for 30 minutes. LPS and EDTA were then added to the concentrations given above, and the LPS-HDL binding reaction was allowed to proceed for ten minutes at 37 degrees before addition of CsCl and

centrifugation. The 10 minute reaction time was chosen as a compromise between the times required for

LPS to bind to HDL in NRS and APRS. Observation of a

₃ peak of H-LPS at a density of 1.30 g/cm signalled the presence of LBP activity; i.e., complex 1.3 (Cl.3) formed.

To quantitate LBP activity in purified fractions of APRS, the reconstitution assay just described was performed with a series of different amounts of the sample being assayed up to a maximum of 200 ul per ml NRS. After centrifugation and quantitation of the LPS in Cl.3, a plot of percent LPS as Cl.3 versus sample volume was made. One LBP unit is defined as that amount of LBP activity that causes 50% of the recovered LPS to be recovered as Cl.3 in the above procedure. LBP activity in APRS was assayed similarly, except that the final volume of the NRS-APRS mixture was held constant and the final plot was then (percent LPS as Cl.3) versus (percent APRS) . On ocassion, to conserve materials, the total assay volume was reduced to 0.5 or 0.25 ml with proportional reduction of all components. D. Purification of LPB Two chromatographic procedures comprise the purification procedure for LBP. Serum was first fractionated using Bio-Rex 70 resin (Bio-Rad, Richmond, CA) • Fifty milliliters of resin were equilibrated with 41 mM NaCl in 50 mM phosphate buffer, pH 7.3, containing 2 mM EDTA (phosphate/EDTA) . Four hundred milliliters of APRS, containing 5 mM EDTA, was run over the column at approximately 65 ml/hour. The column was then washed with column equilibration buffer overnight (about 18 hours) or until the absorbance at 280 nm of the eluate was less than 0.2 absorbance units (AU) .

Washing was continued with 220 mM NaCl in phosphate/EDTA again until the absorbance was below 0.2, followed by a linear gradient formed from 60 ml each of 220 mM and 500 mM NaCl in phosphate/EDTA. Finally, the column was washed with 1 M NaCl in phosphate/EDTA. Pools of fractions to be assayed for LBP activity were dialysed against 5 mM HEPES, pH 7.3, concentrated to 6 ml using PTGC (millipore, Bedford, MA) or YM10 (Amicon, Danvers, MA) membranes in an Amicon ultrafiltration cell, and any precipitate formed was removed by centrifugation.

The second chromatographic step used high performance liquid chromatography (HPLC) (Perkin-Elmer) with a Mono-Q column (Pharmacia, Piscataway, NJ) as the adsorbent using the instructions provided with the pre-packed column. Unless otherwise noted, the flow rate was 1 ml per minute. The column was equilibrated with 20 mM diethanolamine buffer, pH 8.3. Injection of the sample was followed immediately by a 15 ml gradient of zero to 50 mM ammonium sulfate in 20 mM diethanolamine, pH 8.3. The gradient was then steepened, going in 15 minutes from 50 mM to 333 mM ammonium sulfate in the same buffer. Finally the column was washed for 5 minutes at a flow rate of 2 ml/min with 333 mM ammonium sulfate, again in 20 mM diethanolamine, pH 8.3. Fractions were collected and, on the basis of the absorbance profile of the column eluate, pooled into three pools, dialysed against 5 mM HEPES buffer, pH 7.4, and concentrated to 2 ml.

Final separation of the last two components of the LBP-containing fractions from Mono-Q chromatography was accomplished by SDS-PAGE using a 10 percent acrylamide gel. Following electrophoresis

and light staining with Coomassie Blue the protein bands were cut apart and recovered from the gel by eleσtroelution. Hunkapiller et al. Meths. Enzymol. 91:227 (1983). Amino acid sequencing was carried out by the

Protein Structure Core Laboratory of Scripps Research Foundation according to published procedures. Hewick et al., J. Biol. Chem. 16.."7990 (1981).

The present invention has been described with respect to preferred embodiments. It will be clear to those skilled in the art that modifications and/or variations of the disclosed subject matter can be made without departing from the scope of the invention set forth herein.