A selective bi-site immunoenzymatic procedure for human Lp[a] lipoprotein quantification using monoclonal antibodies against apo[a] and apoB

A selective bi-site ELISA assay procedure for quantification of Lp[a] lipoprotein in human plasma based on linkage of apo[a] to apoB is described. The lipoproteins referred to as apo[a]:B were captured by a mixture of two anti-apo[a] monoclonal antibodies (K07, K09) and were revealed by a mixture of six anti-apoB monoclonal antibodies coupled to peroxidase. Since apo[a] and plasminogen have striking similarities in protein structure, the selective binding of Lp[a]:B in our assay depended upon the marked difference in affinity of the K07 and K09 mixture for Lp[a]:B (Kd = 0.32 x 10(-10) M) versus plasminogen (Kd = 0.47 x 10(-7)M). The high sensitivity (the Lp[a]:B working range 0.06-0.40 micrograms/ml) and the use of anti-apoB as antibody tracer added to the selectivity of the assay. The expression of K07 and K09 epitopes determined by competitive inhibition method and the reactivity of Lp[a]:B particles measured by bi-site ELISA were similar on individual lipoproteins, independent to their plasma levels. The assay is precise, and intra- and interassay coefficients of variation were 4.7% and 9.6%, respectively. It yields quantitative Lp[a]:B values that correlate highly with Lp[a] levels obtained by electroimmunoassay with polyclonal antibody (r = 0.73) or with Lp[a] levels measured by the other bi-site ELISA using only K07 and K09 antibodies (r = 0.96). However, upon analyzing each individual plasma with an arbitrary Lp[a]-cut off of 15 mg/dl, evidence of the qualitative aspect of the lipoprotein was obtained. The group with Lp[a] less than 15 mg/dl had higher frequency of subjects (65%) with the ratio Lp[a]/Lp[a]:B above 1.5.(ABSTRACT TRUNCATED AT 250 WORDS)     

Enzyme-linked immunoassay for Lp[a]

Based on our findings that rabbit antisera raised against human Lp[a] or apo[a] have the potential to cross-react with plasminogen, and in some cases have nearly equal affinities for plasminogen and Lp[a], we have developed an assay for plasma Lp[a] based on a "sandwich" ELISA that is insensitive to the presence of plasminogen. This was accomplished through the use of anti-apo[a] as a capture antibody and quantitation of the bound Lp[a], i.e., the apoB-100-apo[a] complex, with an anti-apoB antibody. Although apo[a] is heterogeneous in size, all Lp[a] particles tested, either in pure form or contained in whole plasma, gave parallel dose-response curves and were immunologically equivalent. However, when purified Lp[a] particles with different apo[a] isoforms were studied, those having larger isoforms were, on a weight basis, less reactive than those having a smaller size. Nearly equivalent reactivity was observed when protein concentration was expressed on a molar basis. The distribution of Lp[a] in a population of 84 subjects was skewed with one-third of the individuals having less than 1 mg/dl Lp[a] protein. All subjects tested had measurable concentrations of Lp[a] with a lower limit of detection of 0.030 mg/dl Lp[a] protein. The mean level was 3.2 mg/dl with a range of 0.045 to 13.3 mg/dl. These studies demonstrate the successful development of an ELISA for Lp[a] protein that is insensitive to the presence of plasminogen; that heterogeneity of Lp[a] and apo[a] are an important source of variation in the assay; and the need for an appropriate Lp[a] standard in order to minimize this variation.     

Hereditary and dietary effects on apolipoprotein[a] isoforms and Lp[a] in baboons

Baboons possess Lp[a] that is similar to human Lp[a], including the presence of the unique protein, apo[a]. Baboon apo[a] occurred in at least nine isoforms distinguishable by size. Isoforms were resolved by 3-12% polyacrylamide gradient gel electrophoretic separation of serum proteins, and were detected with baboon apo[a]-specific antibodies. Thirty one different apo[a] isoform phenotypes were detected in a population of 165 unrelated baboons. Identical isoform phenotypes were observed in different samples from individual baboons, and isoform phenotypes were unaffected by changes in diet. In one experiment, 16 baboons were fed a series of five diets differing in amounts of cholesterol and saturated or unsaturated fats. There was no significant effect of diet on serum Lp[a] levels. In another group of baboons (n = 70) controlled for age and dietary history, enrichment of the diet with cholesterol and saturated fat caused a small, but significant (P less than 0.005), increase (means = 0.6 mg/dl) in serum Lp[a] concentration. Analysis of two large sire families suggested that apo[a] isoform patterns and serum Lp[a] concentrations were inherited. Putative parental alleles responsible for specific isoform bands appeared to segregate randomly. Heritability (h2) of serum Lp[a] concentration was estimated to be 0.95 +/- 0.04. We conclude that apo[a] isoform phenotypes and serum Lp[a] concentrations are inherited, and that Lp[a] concentrations are only slightly influenced by diet.     

High-level lipoprotein [a] expression in transgenic mice: evidence for oxidized phospholipids in lipoprotein [a] but not in low density lipoproteins

Efforts to elucidate the role of lipoprotein [a] (Lp[a]) in atherogenesis have been hampered by the lack of an animal model with high plasma Lp[a] levels. We produced two lines of transgenic mice expressing apolipoprotein [a] (apo[a]) in the liver and crossed them with mice expressing human apolipoprotein B-100 (apoB-100), generating two lines of Lp[a] mice. One had Lp[a] levels of approximately 700 mg/dl, well above the 30 mg/dl threshold associated with increased risk of atherosclerosis in humans; the other had levels of approximately 35 mg/dl. Most of the LDL in mice with high-level apo[a] expression was covalently bound to apo[a], but most of the LDL in the low-expressing line was free. Using an enzyme-linked sandwich assay with monoclonal antibody EO6, we found high levels of oxidized phospholipids in Lp[a] from high-expressing mice but not in LDL from low-expressing mice or in LDL from human apoB-100 transgenic mice (P <0.00001), even though all mice had similar plasma levels of human apoB-100. The increase in oxidized lipids specific to Lp[a] in high-level apo[a]-expressing mice suggests a mechanism by which increased circulating levels of Lp[a] could contribute to atherogenesis.     

Quantification of apo[a] and apoB in human atherosclerotic lesions.

Lipoprotein[a] or Lp[a] is a cholesterol-rich plasma lipoprotein that is associated with increased risk for cardiovascular disease. To better understand this association we determined the amount of apo[a] and apoB as possible estimates for Lp[a] and low density lipoprotein (LDL) accumulation in atherosclerotic lesions and in plasma, from patients undergoing vascular surgery, using specific radioimmunoassays for apolipoprotein[a] and apolipoprotein B. Apo[a] and apoB were operationally divided into a loosely bound fraction obtained by extracting minced samples of plaque with phosphate-buffered saline (PBS), and a tightly bound fraction obtained by extracting the residual tissue with 6 M guanidine-HCl (GuHCl). We found that 83% of all apo[a] but only 32% of all apoB in lesions was in the tightly bound fraction. When normalized for corresponding plasma levels, apo[a] accumulation in plaques was more than twice that of apoB. All fractions of tissue apo[a], loosely bound, tightly bound, and total, correlated significantly with plasma apo[a]. However, no significant correlations were found between any of the tissue fractions and plasma apoB. If all apo[a] and apoB had been associated with intact Lp[a] or LDL particles, the calculated mass of tightly bound Lp[a] would actually have exceeded that of tightly bound LDL in five cases with plasma Lp[a] levels above 5 mg apo[a] protein/dl. When PBS and GuHCl extracts of lesions were subjected to one-dimensional electrophoresis, the major band stained for lipid and immunoblotted positively for apo[a] and apoB, suggesting the presence of some intact Lp[a] in these extracts. These results suggest that Lp[a] accumulates preferentially to LDL in plaques, and that plaque apo[a] is directly associated with plasma apo[a] levels and is in a form that is less easily removable than most of the apoB. This preferential accumulation of apo[a] as a tightly bound fraction in lesions, could be responsible for the independent association of Lp[a] with cardiovascular disease in humans.     

Lipoprotein [a] is cleared from the plasma primarily by the liver in a process mediated by apolipoprotein [a

The cellular and molecular mechanisms responsible for lipoprotein [a] (Lp[a]) catabolism are unknown. We examined the plasma clearance of Lp[a] and LDL in mice using lipoproteins isolated from human plasma coupled to radiolabeled tyramine cellobiose. Lipoproteins were injected into wild-type, LDL receptor-deficient (Ldlr-/-), and apolipoprotein E-deficient (Apoe-/-) mice. The fractional catabolic rate of LDL was greatly slowed in Ldlr-/- mice and greatly accelerated in Apoe-/- mice compared with wild-type mice. In contrast, the plasma clearance of Lp[a] in Ldlr-/- mice was similar to that in wild-type mice and was only slightly accelerated in Apoe-/- mice. Hepatic uptake of Lp[a] in wild-type mice was 34.6% of the injected dose over a 24 h period. The kidney accounted for only a small fraction of tissue uptake (1.3%). To test whether apolipoprotein [a] (apo[a]) mediates the clearance of Lp[a] from plasma, we coinjected excess apo[a] with labeled Lp[a]. Apo[a] acted as a potent inhibitor of Lp[a] plasma clearance. Asialofetuin, a ligand of the asialoglycoprotein receptor, did not inhibit Lp[a] clearance. In summary, the liver is the major organ accounting for the clearance of Lp[a] in mice, with the LDL receptor and apolipoprotein E having no major roles. Our studies indicate that apo[a] is the primary ligand that mediates Lp[a] uptake and plasma clearance.     

Polymorphic forms of human apolipoprotein[a]: inheritance and relationship of their molecular weights to plasma levels of lipoprotein[a]

The plasma concentration of human lipoprotein[a], Lp[a], is highly correlated with coronary artery disease. The protein moiety of Lp[a], apoLp[a], consists of two apoproteins, apo[a] and apoB-100, linked by one or more disulfide bonds(s). Apo[a], the protein unique to Lp[a], exists in polymorphic forms that exhibit different apparent molecular weights (Mr). Polyacrylamide gel electrophoresis in sodium dodecyl sulfate followed by immunoblotting was used to separate and visualize these different forms and to determine the polymorphic pattern of apo[a] in the plasma samples of 692 individuals. A total of 11 different polymorph bands ranging in Mr from 419 kD to 838 kD could be resolved, but only 1 or 2 bands were present per individual. The polymorphic band pattern for an individual was assigned to 1 of the 66 different phenotype designations representing the total number of possible single- and double-band combinations of the 11 detectable bands. All 11 of the possible single-band phenotypes but only 32 of the 55 possible double-band phenotypes were represented. There were 412 plasma samples (59.5%) that contained a single band, 274 (39.6%) contained two bands, and only 6 (0.9%) had no detectable apo[a] band. A highly significant inverse correlation was found between the Mr of the band(s) present and the plasma apoLp[a] concentration (r = -0.461; rho = 0.0001). The correlation was better between apoLp[a] and single-band (r = -0.495; rho = 0.0001) than double-band (r = -0.382; rho = 0.0001) phenotypes. Of the 274 individuals exhibiting double-band phenotypes, the lower Mr band was more intense in 141 (51.4%), the two bands were equally intense in 85 (31.0%), while the higher Mr band was more intense in 48 (17.5%). Based upon the hypothesis that apo[a] polymorphism is controlled by different alleles at a single locus, the frequency of the 11 alleles determined from the observe phenotypes (low Mr----high Mr) was: band 1) 419 kD, 0.00875; band 2) 489 kD, 0.00510; band 3) 536 kD, 0.0555; band 4) 553 kD, 0.0758; band 5) 613 kD, 0.135; band 6) 680 kD, 0.0824; band 7) 705 kD, 0.104; band 8) 742 kD, 0.151; band 9) 760 kD, 0.246; band 10) 796 kD, 0.128; band 11) 838 kD, 0.00802. The observed distribution of phenotypes in the population was compared by chi-square analysis to that predicted on the basis of simple Mendelian inheritance, and the hypothesis was rejected (chi 2 = 921.7; rho less than 0.001). Significantly, the singleband phenotypes are over-represented in the population compared to that predicted.(ABSTRACT TRUNCATED AT 400 WORDS)     

Characterization of the basis of lipoprotein [a] lysine-binding heterogeneity

Although elevated plasma concentrations of lipoprotein [a] (Lp[a]) are considered to be a risk factor for atherosclerosis, the mechanisms by which Lp[a] mediates its pathogenic effects have not been conclusively determined. The apolipoprotein [a] (apo[a]) component of Lp[a] confers unique structural properties to this lipoprotein, including the ability to bind to lysine residues in biological substrates. It has been shown, however, that only a fraction of plasma Lp[a] (Lp[a]-Lys(+)) binds to lysine-Sepharose in vitro. The nature of the non-lysine-binding Lp[a] fraction in plasma (Lp[a]-Lys(-)) is currently unknown. In the present study, the Lp[a]-Lys(+) fraction was determined in the plasma of six unrelated individuals; the Lp[a]-Lys(+) fraction in these plasma samples ranged from approximately 37 to approximately 48%. Interestingly, purification of the Lp[a] by density gradient ultracentrifugation followed by gel filtration and ion-exchange chromatography resulted in progressive increases in the Lp[a]-Lys(+) fraction. Addition of either purified low density lipoprotein (LDL) or fibronectin to the purified Lp[a] at a 1:1 molar ratio reduced the Lp[a]-Lys(+) fraction (maximal decrease of 34 and 20%, respectively) whereas addition of both fibronectin and LDL to the purified Lp[a] resulted in a further decrease (45% maximally) in this fraction. Similar results were obtained by using a recombinant expression system for apo[a]: addition of a 4-fold molar excess of either LDL or fibronectin to conditioned medium containing metabolically labeled recombinant apo[a] reduced the Lys(+) fraction by 49 and 23%, respectively.Taken together, our data suggest that the lysine-binding heterogeneity of plasma Lp[a] is not primarily an intrinsic property of the lipoprotein, but rather results in large part from its ability to noncovalently associate with abundant plasma components such as LDL and fibronectin. These interactions appear to mask the lysine-binding site in apo[a] kringle IV type 10, which mediates the interaction of Lp[a] with lysine-Sepharose. The contribution of these interactions to the function of Lp[a] in vivo remains to be investigated.     

Influence of short term dietary cholesterol and fat on human plasma Lp[a] and LDL levels.

The relationship between plasma levels of Lp[a] and LDL was examined using dietary regimens. In 81 normolipidemic male outpatients, dietary cholesterol was increased by consuming six eggs per day from a mean (SD) level of 311 (162) to 1430 (198) mg per day. Mean (SD) LDL-cholesterol levels increased from 102 (26) mg/dl to 120 (33) mg/dl (P less than 0.001), while mean (SD) Lp[a] levels were 5.5 (6.1) mg/dl on the basal diet and 5.6 (6.4) mg/dl on the cholesterol-rich diet. No significant correlation was observed between increases in either LDL-cholesterol or apolipoprotein B to Lp[a], nor was there any relationship between individual baseline levels of Lp[a] and dietary-induced changes of Lp[a]. Fourteen of the 81 participants were reexamined under strict nutritional control. Four diets with 40% of calories as fat, but differing in the type of fat and the amount of cholesterol, were administered sequentially to all subjects. As expected, mean (SD) LDL-cholesterol and apolipoprotein B levels were highest on the saturated fat, high cholesterol diet (112 (32) mg/dl and 79 (22) mg/dl) and lowest on the polyunsaturated fat, low cholesterol diet (77 (27) mg/dl and 53 (18) mg/dl). In contrast, mean Lp[a] levels did not significantly change among the four diets (range 4.2-4.9 mg/dl). No correlation of Lp[a] responses with changes in plasma lipids, apolipoproteins, or lipoproteins was observed on any diet. These data suggest that determinants of plasma Lp[a] levels are distinctly different from the determinants of plasma LDL levels in normolipidemic males.     

Characterization of five mouse monoclonal antibodies to apolipoprotein[a] from human Lp[a]: evidence for weak plasminogen reactivity

We describe the development of five murine monoclonal antibodies (14A12, 39A1, 53A9, 73A7, and 128A6) specific to human apolipoprotein[a] (Mr approximately 570,000), and their characterization by a number of procedures including cotitration, competition and inhibition enzyme-linked immunosorbent assays (ELISA), immunoblotting of native lipoproteins and of SDS-solubilized apolipoproteins electrophoresed in polyacrylamide gels, and dot immunobinding assays. The patterns of immunoreactivity of these antibodies were similar. Each reacted in ELISA assays and upon electroimmunoblotting with purified apo[a], with apo[a] liberated by reduction of Lp[a], and with delipidated Lp[a] solubilized in SDS, but by contrast, they reacted with native Lp[a] to a significant degree only upon electroimmunoblotting. No reactivity was seen with LDL-apoB-100 or with other apolipoproteins. The cross-reactivity of these antibodies with the homologous protein, plasminogen, was examined by comparison of the amount of plasminogen or apo[a] required for 50% inhibition of antibody binding to apo[a], and by an ELISA assay. The inhibition assay showed reactivity with plasminogen to be 37- to 50-fold lower than with apo[a], while dot immunobinding showed the lower limit of detection of plasminogen and of apo[a] to be approximately 320 and 31 micrograms, respectively. In an ELISA sandwich assay based on monoclonal antibodies LHLP-1, 14A12, and 53A9, the lower limit of Lp[a] detection (approximately 1 ng/ml protein) was about 100-fold less than that of plasminogen. Chemical modification of apo[a] revealed a significant contribution of arginine residues to the epitopes of 14A12, 39A1, and 53A9. Modification of cysteine residues with iodoacetamide was without effect, thereby distinguishing these antibodies from LHLP-1. Each antibody reacted with the six major size forms of apo[a] (Mr approximately 450,000-750,000) in immunoblots of human sera electrophoresed in SDS-polyacrylamide gels. Marked heterogeneity in apo[a] phenotype was detected and both single and double band phenotypes were observed in a randomized study. Cotitration and competition binding studies showed varying degrees of interaction between all five epitopes, with the exception of 128A6 which appeared to be independent of 39A1 and 53A9 (and vice versa). These data suggest that our five monoclonal antibodies recognize epitopes on apolipoprotein[a] that are exposed and accessible on the native Lp[a] particle. We conclude that our monoclonal antibodies recognize a specific region of apo[a], and that this region undergoes a conformational change upon adsorption of Lp[a] to plastic thereby diminishing epitope recognition.(ABSTRACT TRUNCATED AT 400 WORDS)