536
Chem. Res. Toxicol. 1997, 10, 536-545
Articles Levuglandin E2-Protein Adducts in Human Plasma and Vasculature Robert G. Salomon,*,† Ganesamoorthy Subbanagounder,† June O’Neil,‡ Kamaljit Kaur,† Mark A. Smith,§ Henry F. Hoff,‡ George Perry,§ and Vincent M. Monnier§ Department of Chemistry and Institute of Pathology, Case Western Reserve University, Cleveland, Ohio 44106, and Department of Cell Biology, Research Institute, Cleveland Clinic Foundation, Cleveland, Ohio 44195 Received September 9, 1996X
The prostaglandin endoperoxide PGH2 rearranges nonenzymatically to generate prostaglandins and secoprostanoic acid levulinaldehyde derivatives such as PGE2 and levuglandin (LG) E2, respectively. Direct detection of LGE2 in biological samples is complicated because it is rapidly sequestered by covalent adduction to endogenous nucleophiles including proteins, which produces LGE2-derived protein-bound pyrroles. Therefore, to detect LGE2-protein adducts in vivo, antibodies were raised against a covalent adduct of LGE2 with keyhole limpet hemocyanin (KLH). This antigen enabled the production of high-titer antibodies that exhibit minimal cross-specificity and are sensitive for detecting LGE2-derived pyrroles. Although pyrrole yields are low at LG/protein ratios found in vivo, an enzyme-linked immunosorbent assay with the LGE2-KLH antibodies detects LGE2-derived protein-bound pyrrole immunoreactivity in human plasma from specific patient populations. Furthermore, prominent immunocytochemical staining of human brain thin sections revealed the presence of LGE2-derived pyrrole immunoreactivity, especially in the meningeal vessels of some patients. This demonstration of LG-protein adducts in human plasma and vasculature provides the first evidence for the biological occurrence of levuglandins in vivo and further suggests that these antibodies might prove useful in diagnostic and mechanistic studies of various disease conditions.
Introduction In addition to serving as a pivotal intermediate in the enzymatic generation of prostaglandins (PG),1 thromboxanes, and prostacyclins, the prostaglandin endoperoxide PGH2 rearranges nonenzymatically under the conditions of its cyclooxygenase-promoted biosynthesis from arachidonic acid (AA). The rearrangement (Figure 1) generates PGs, e.g., PGE2 (1), and secoprostanoic acid levulinaldehyde derivatives, e.g., levuglandin (LG) E2 (2). In earlier studies we showed that LGE2 binds covalently with proteins (3) and that Paal-Knorr condensation (4) generates LG-derived protein-bound pyrroles 1 (5), making detection of LGE2 difficult. Further complicating this, we found that pyrroles generated by the reaction of LGE2 with simple amines are chemically sensitive compounds * Author to whom correspondence should be addressed [telephone, (216) 368-2592; fax, (216) 368-3006; e-mail,
[email protected]]. † Department of Chemistry. ‡ Department of Cell Biology. § Institute of Pathology. X Abstract published in Advance ACS Abstracts, March 15, 1997. 1 Abbreviations: AA, arachidonic acid; AS, atherosclerosis; BHT, butylated hydroxytoluene; BSA, bovine serum albumin; CEO, chicken egg ovalbumin; ELISA, enzyme-linked immunosorbent assay; FPLC, fast protein liquid chromatography; HNE, (E)-4-hydroxy-2-nonenal; HSA, human serum albumin; IgG, immunoglobin G; KLH, keyhole limpet hemocyanin; LDL, low-density lipoprotein; LG, levuglandin; LSC, liquid scintillation counting; MDA, malondialdehyde; NGS, normal goat serum; PBS, phosphate-buffered saline; PC, phosphatidylcholine; PG, prostaglandin; TB, Tris buffer; RD, renal disease.
S0893-228x(96)00157-9 CCC: $14.00
Figure 1. Reaction pathway leading to the formation of both PGE2 and LGE2-derived protein-bound pyrrole from PGH2.
which readily decompose to form insoluble, presumably polymeric, products. Therefore, we previously raised antibodies against a stable pyrazole isostere antigen 2 (Figure 2) and used them to detect LGE2-derived proteinbound pyrroles generated in vitro. Quite unexpectedly, we found no loss of immunoreactivity over several weeks for an adduct (LGE2-KLH) of LGE2 with keyhole limpet hemocyanin (KLH) (6). The apparent stability of such protein-bound LGE2-derived pyrroles may be the consequence of inhibition of pyrrole polymerization by the protein matrix. More important, whatever the chemical basis, this observation had two practical consequences: (i) LG-protein adducts that are generated in vivo should accumulate and (ii) immunological detection of such protein modifications might be feasible with antibodies raised directly against LG-protein adducts. © 1997 American Chemical Society
LGE2-Protein Adducts in Human Plasma
Figure 2. Molecular structures of a pyrazole isostere antigen 2 and the LGE2-KLH antigen.
For the present study, antibodies were raised in rabbits against LGE2-KLH. After purification on protein A or LGE2-BSA affinity columns, these antibodies provide a sensitive analytical method that is specific for LG-derived pyrroles 1. We found that pyrrole formation by the reaction of LGE2 with human serum albumin (HSA) varies nonlinearly over a wide range of LG/protein ratios. Moreover, we also present the first evidence for the biological occurrence of such LG-modified proteins in human plasma and vasculature.
Experimental Procedures Cautionary Notice: Seroclear is an ozone-depleting reagent and appropriate handling/disposal procedures are necessary. Sodium azide is highly toxic if internalized and can generate the highly toxic gas, hydrazoic acid, in the presence of acids. General Methods. Liquid scintillation counting (LSC) was done with a Beckman LS 5801 counter. Samples were prepared with 5 mL of xylene-based scintillation fluid (Fisher). Centrifugation was done with a Sorvall centrifuge with an H-1000B rotor or with a Beckman AccuSpin FR centrifuge with an AH-4 rotor. Absorbance values of enzyme-linked immunosorbent assays (ELISAs) were measured on a Bio-Rad microplate reader using dual wavelengths (405 nm to read the plate and 650 nm as a reference). TLC was performed on glass plates precoated with silica gel (Kieselgel 60 F254, E. Merck, Darmstadt, Germany); Rf values are quoted for 0.25 mm thick plates. The plates were visualized by viewing the developed plates under shortwavelength UV light or by heating the plates after spraying with vanillin-sulfuric acid. Flash column chromatography was performed on 230-400 mesh silica gel supplied by E. Merck. Materials. Spectrapor membrane tubing (Mr cutoff 14 000 No. 2) for standard dialysis was obtained from Fisher Scientific Co. (Chicago, IL). Low-binding Ultrafree-CL PLGC regenerated cellulose 2 mL ultrafiltration units (NMWL ) 10 000) were obtained from Millipore Corp. (Bedford, MA). The following commercially available materials were used as received: chicken egg ovalbumin (CEO; grade V, 99%), bovine serum albumin (BSA; fraction V, 96-99%), HSA (fraction V), disodium pnitrophenyl phosphate, AA, PGF2R, and PGB2 were from Sigma Chemical Co. (St. Louis, MO); PGE2 and PGD2 were from Cayman Chemical, Ann Arbor, MI; KLH was from ICN Biochemicals (Irvine, CA); goat anti-rabbit IgG-alkaline phosphatase was from Boehringer-Mannheim (Indianapolis, IN); rabbit peroxidase anti-peroxidase was from Sternberger Monoclonals Inc. (Baltimore, MD). Phosphate-buffered saline (PBS) was prepared from a pH 7.4 stock solution containing 0.2 M NaH2PO4/Na2HPO4, 3.0 M NaCl, and 0.02% NaN3 (w/w). This solution was diluted 20× as needed. Tris buffer (TB), containing 1 mM NaCl and 0.6 mM CaCl2, was prepared from 5 mM trizma (Sigma); LGE2 (7), 3-[6-carboxy-2(Z)hexenyl]-1-neopentyl-4-[3hydroxy-1(E)octenyl]-2-methylpyrrole (3a) (6), 3-[6-carboxy-2(Z)hexenyl]-1-(6-hydroxyhexyl)-4-[3-hydroxy-1(E)octenyl]-2-methylpyrrole (3b) (6), (Z)-7-hydroxy-5-heptenoic acid (4) (6), (E)-1chloro-1-octen-3-ol (5) (6), 6-(2-pentylpyrrol-1-yl)hexanoic acid (6a) (6), 6-[2-formyl-5-(hydroxymethyl)pyrrol-1-yl]hexanoic acid (6b) (6), 2-[8-acetyl-12-tert-butyldimethylsiloxy-9-[1(S),2-isopropylidenedioxyethyl]]-5(Z),10(E)-heptadecadienoic acid (7) (7),
Chem. Res. Toxicol., Vol. 10, No. 5, 1997 537 2-[8-acetyl-12-tert-butyldimethylsiloxy-5,6-ditritio-9-[1(S),2-isopropylidenedioxyethyl]]-5(Z),10(E)-heptadecadienoic acid (7-t2) with specific activity ) 35.4 Ci/mmol (3), a BSA-pyrazole isostere conjugate (8), LGE2-KLH (6), MDA-LDL (9), and HNE-LDL that was not NaBH4-reduced (9) were prepared as described previously. Low Level Tritiated Samples. To allow quantitative radiochemical analysis, samples of lightly tritium tagged LGE2 precursor, 2-[8-acetyl-12-tert-butyldimethylsiloxy-9-[1(S),2-isopropylidenedioxyethyl]]-5(Z),10(E)-heptadecadienoic acid, were prepared by adding small amounts of 7-t2 to 7. These mixtures are referred to as tritium-labeled 7. The specific activities of these samples were determined by LSC of an aliquot. LGE2 prepared from this precursor is referred to as tritium-labeled LGE2 and is presumed to have the same specific activity. LGE2-KLH Antigen. A PBS solution containing 28.5 mM tritium-labeled (45 µCi/mmol) LGE2 (12.48 µmol) and 2.3 µM KLH (6.7 mg, 3.34 µmol of lysyl residues) was incubated at room temperature for 1 h followed by dialysis against PBS (200 mL for 8 h and then 200 mL for 16 h). The fraction of LG bound (17%) was determined by quantitative radiochemical analysis. The adduct solution was diluted to provide 0.72 mM KLH-bound LGE2. LGE2-HSA Standard. Tritium-labeled LGE2 (specific activity ) 167.6 µCi/mmol, 1.6 mg, 4.54 µmol) was mixed with HSA (13.2 mg, 0.2 mmol) in PBS (1 mL) and incubated at 37 °C for 4 h. The solution was then dialyzed (MW cutoff 14 000) against PBS (2 × 500 mL) for 48 h. The volume was then adjusted to 7 mL with PBS. The total radioactivity in the sample, 399.1 nCi, was determined by LSC of aliquots. This corresponds to 2.38 µmol of LGE2 bound to HSA. The final concentration of HSA-bound LGE2 was 0.34 mM or 11.9 mol of bound LGE2/mol of HSA. Adducts with Various HSA/LGE2 Ratios. HSA-LGE2, adducts containing 0.04-12.8 mol of LGE2/mol of HSA, were prepared. Thus, samples containing HSA (6 mg) in pH 7.4 PBS (2 mL) were incubated at 37 °C for 18 h with various amounts (3.5-720 µg) of tritium-labeled (3.93 µCi/µmol) LGE2 and then dialyzed against PBS (2 × 500 mL, 16 h each) at 22 °C. The amount of LG bound was determined by quantitative radiochemical analysis, and LG-derived pyrrole immunoreactivity was determined by ELISA (vide infra). LGE2-BSA Coating Agent. Tritium-labeled LGE2 (specific activity ) 1.14 µCi/mmol, 16 mg, 46.8 µmol) was incubated at 37 °C for 4 h with BSA (66 mg) in PBS (2 mL). The resulting mixture was dialyzed (MW cutoff 14 000) against pH 7.4 TB (2 × 125 mL) for 48 h. The volume was then adjusted to 5 mL. The total radioactivity in the sample, 1.20 µCi, was determined by LSC of an aliquot (100 µL). This corresponds to 10.4 µmol of LGE2 bound to BSA. The final concentration of LGE2 was thus 2.08 mM bound to BSA (13 mg/mL). Lipoproteins. LDL was isolated from human plasma by sequential ultracentrifugation as a 1.019 < d < 1.063 g/mL fraction as described previously (10). The LDL was dialyzed against 0.15 M, pH 8.5, NaCl containing 0.5 mM Na2EDTA, filter-sterilized, and then stored at 4 °C. Human Plasma. Blood was collected in 7 mL vacutubes (purple top) which contained EDTA (10.5 mg). Cells were removed by centrifugation at 2500 rpm (1300g) for 20-30 min. After transfer of the supernatant serum to a plastic vial, butylated hydroxytoluene (BHT) in absolute ethanol was added to give a final concentration of 1 mg/mL, and the plasma was then stored at 3 °C for no more than a few days. For longer term storage, samples were kept at -80 °C. Immunization. The immunogen, an LGE2-KLH adduct (1.12 mg) containing 0.32 µmol of LGE2/mg of KLH, in PBS (500 µL) at pH 7.4 was emulsified with Freund’s complete adjuvant (500 µL). Each of three Pasturella free, New Zealand white rabbits (Hazelton) was inoculated intradermally into several sites on the back (125 µL) and rear leg (125 µL). Booster injections of LGE2-KLH in Freund’s incomplete adjuvant were given every 21 days. Antibody titer was monitored 10 days after each inoculation by ELISA as described below.
538 Chem. Res. Toxicol., Vol. 10, No. 5, 1997 LGE2-KLH Antibody Purification: Protein A Column. The crude LGE2-KLH antibody serum from the day 94 bleeding of rabbit 3 contained 37.6 mg/mL of protein, as determined by absorbance at 280 nm [A280 ) 1.35 for 1.0 mg/mL (11)]. Seroclear (1.4 mL) was added to this crude antibody serum (1.4 mL), and the mixture was vortexed for 60 s. Then it was centrifuged for 10 min at 3000 rpm (1900g), and the upper delipidated layer (aqueous phase) was removed and added to an equal volume of binding buffer. This solution was eluted through a protein A-Superose (Pharmacia) column monitoring the eluant at 280 nm. The eluate was collected into fractions (1 mL) containing 1.0 M Tris buffer (pH 8.8; 160 µL). The fractions containing IgG proteins were pooled and dialyzed against pH 7.4 PBS (0.02% NaN3) for 24 h at 5 °C. The resulting solution (7.5 mL) contained 0.935 mg/mL of purified IgGs, as determined by absorbance at 280 nm. This corresponds to 13% of the IgG in the crude serum. LGE2-KLH Antibody Purification: Affinity Column. A PBS solution of LGE2-BSA (0.2 mg of LGE2 bound to 8 mg of BSA) was used as the affinity ligand that was coupled to cyanogen bromide-activated Sepharose 4B (Pharmacia, Piscataway, NJ) using the procedure recommended by the supplier. The remaining active sites on Sepharose were blocked with glycine buffer (0.2 M, pH 8.0). The LGE2-BSA-Sepharose was placed in a column and rinsed with pH 8.3 sodium bicarbonate buffer (0.1 M, 5 mL) containing NaCl (0.5 M), followed by acetate buffer (10 mL, 0.1 M, pH 4.0) containing NaCl (0.5 M). Crude rabbit serum (1.9 mL) was loaded onto the column and was allowed to equilibrate at 5 °C for 16 h. Unbound serum was then eluted with PBS (40 mL). Specifically retained antibodies were eluted with pH 2.2 glycine-HCl buffer (0.2 M containing 0.5 M NaCl, 30 × 0.8 mL) and collected in tubes containing pH 10 Tris buffer (1.0 M, 200 µL). The column was eluted with PBS (20 mL) and further eluted with pH 12 triethylamine buffer (0.1 M, 20 × 1 mL), and absorbance at 280 nm was measured for the fractions. The glycine-HCl eluate and the triethylamine eluate were dialyzed separately against pH 7.4 PBS (2 × 2 L) at 4 °C for 24 h. The dialyzed antibody solutions were concentrated by gel filtration through a membrane (MW cutoff 5000, 2 mL) using an ultracentrifuge at 5800 rpm for 1 h. The concentrations of antibodies for the glycine-HCl eluted fraction (687 µg/mL, 1.2 mL) and the triethylamine eluted fraction (579 µg/mL, 1.2 mL) were calculated from their absorbances at 280 nm. A 10% BSA solution (100 µL) was added to each purified antibody fraction, which was then stored at 4 °C. The antibody obtained by triethylamine elution was used for immunostaining studies. Enzyme-Linked Immunosorbent Assay (ELISA). For all ELISAs, unless otherwise noted, duplicates of each sample were run on the same plate. (a) Antibody Titers. For determining LGE2-KLH antibody levels in rabbit serum, the BSA conjugate of a stable pyrazole isostere (8) of levuglandin-derived protein-bound pyrroles 4 was used as coating agent. The pyrazole isostere hapten was conjugated with BSA in a 6.6:1 molar ratio. The isostere-BSA conjugate (100 µL of a solution containing 4.4 mg/mL in pH 7.4 PBS) was added to each well of a sterilized Baxter ELISA plate. The plate was then incubated at 37 °C for 1 h in a moist chamber. After the coating solution was discarded, each well was washed with PBS (3 × 300 µL), then filled with 1.0% CEO in PBS (300 µL), and incubated at 37 °C for 1 h to block remaining active sites on the plastic phase. Each well was washed with 0.1% CEO in PBS (300 µL) and then 100 µL of rabbit serum from each bleeding, diluted 1:100000 with 0.2% CEO in PBS, or 0.2% CEO in PBS without serum for a blank, was dispensed into the sample wells. Normal rabbit (not injected with antigen) serum diluted as above was employed as a negative response control. The plate was covered and incubated at room temperature for 1 h with shaking. After the supernatants were discarded and a wash with 0.1% CEO (3 × 300 µL), 100 µL of goat anti-rabbit IgG-alkaline phosphatase diluted 1:1000 with 1.0% CEO was added to each well and the plate was again incubated with shaking at room temperature
Salomon et al.
Figure 3. Serum anti LGE2-pyrrole antibody titer in New Zeland white rabbits immunized with LGE2-KLH using a pyrazole isostere-BSA conjugate as coating agent. for 1 h. After the supernatant was discarded, the wells were washed with 0.1% CEO (3 × 300 µL). Enzyme-linked antibody bound to the well was revealed by dispensing into each well 100 µL of a solution containing disodium p-nitrophenyl phosphate [10 mg in 11 mL of 50 mM aqueous glycine containing MgCl2 (1 mM) and sufficient 6 M NaOH to raise the pH to 9.6]. The plate was allowed to develop at room temperature, usually
