Preparation, Sensitivity, and Specificity of Limulus Lysate for Endotoxin Assay

Limulus amoebocyte lysate was prepared from a total of 180 crabs during 1971 and 1972 by using a slightly modified lysate preparation procedure. Marked variability of lysate potency was noted both years. In addition, lysate quality appeared diminished in 1972 as compared with 1971. Different lysate batches were evaluated for potency by using a variety of endotoxin preparations. Variations in batch potencies were observed, but little variation in reactivity among different endotoxin preparations was noted. Use of potent lysate batches allowed detection of endotoxin concentrations as low as 100 pg/ml. No endotoxinlike activity was observed from 11 different strains of yeast by use of the Limulus assay.     

Carcinoscorpius rotunda cauda amoebocyte lysate for detection of endotoxins--its preparation, stability, sensitivity and comparison with Limulus amoebocyte lysate

Carcinoscorpius amoebocyte lysate (CAL) was prepared from C. rotunda cauda by a modification of the method described by Mahalanabis et al. [Indian J Med Res, 70 (1979) 35]. Seasonal variation as well as batch variation was observed in the yield of haemolymph and the total lysate protein. In the presence of E. coli lipopolysaccharide (pure, free endotoxin) and E. coli and Salmonella cell suspensions (bound endotoxin), the CAL formed a gel after incubation at 37 degrees C. The gelling time varied from 10-90 min depending on the concentration of endotoxin used; higher concentrations formed gel more rapidly. The endotoxin detection capacity (sensitivity) of the lysate preparations was influenced by the season in which prepared, but not by the total protein content. Ten fold increase in the sensitivity was achieved by a purification step using chloroform. Although subsequent frozen storage with or without lyophilization did not alter the initial sensitivity, it was either decreased considerably or lost totally when the lysate was stored for 4 months at 4 degrees C or for 2 months at 30 degrees C. Under the same conditions, Limulus lysate was more stable. The lost sensitivity could not be regained by the incorporation of divalent cations (Ca2+ and Mg2+). The CAL preparations in general were able to detect as little as 10-100 pg of endotoxin or as few as 10(3) cells of E. coli or 10(4) cells of Salmonella and were comparable to LAL. CAL could be used successfully in lieu of Limulus amoebocyte lysate in the detection and assay of endotoxins.     

Sodium-Dependent Growth of Azotobacter chroococcum

Selected ion-monitoring gas chromatography-mass spectrometry was used for detection of beta-hydroxy fatty acids as an independent assay for the presence or absence of endotoxin in materials claimed to induce nonspecific activation of Limulus amoebocyte lysate. To this end, suspensions of gram-negative and -positive bacteria, one fungal species, cerebrospinal fluid, and hollow-fiber hemodialyzer rinses were assayed for endotoxin by gas chromatography-mass spectrometry and the Limulus amoebocyte lysate assay. Good qualitative agreement was shown for both methods when suspensions of test organisms were assayed. Two false-negative results were obtained by gas chromatography-mass spectrometry assays of cerebrospinal fluid and were shown to be a result of insufficient endotoxin in the cerebrospinal fluid specimens for detection by gas chromatography-mass spectrometry. Hemodialyzer rinses were Limulus assay positive; however, no beta-hydroxy fatty acids were detected by gas chromatography-mass spectrometry. These data were compared with data obtained from USP rabbit pyrogen tests of the rinse materials (nonpyrogenic) and chemical characterization of the Limulus assay-reactive rinses, which showed the rinses to be cellulosic in nature. It is suggested that beta-hydroxy fatty acids, as assayed by selected ion-monitoring gas chromatography-mass spectrometry, be used as chemical marker molecules for the presence or absence of endotoxin in materials reported to cause nonspecific activation of Limulus amoebocyte lysate.     

Sodium-Dependent Growth of Azotobacter chroococcum

Selected ion-monitoring gas chromatography-mass spectrometry was used for detection of beta-hydroxy fatty acids as an independent assay for the presence or absence of endotoxin in materials claimed to induce nonspecific activation of Limulus amoebocyte lysate. To this end, suspensions of gram-negative and -positive bacteria, one fungal species, cerebrospinal fluid, and hollow-fiber hemodialyzer rinses were assayed for endotoxin by gas chromatography-mass spectrometry and the Limulus amoebocyte lysate assay. Good qualitative agreement was shown for both methods when suspensions of test organisms were assayed. Two false-negative results were obtained by gas chromatography-mass spectrometry assays of cerebrospinal fluid and were shown to be a result of insufficient endotoxin in the cerebrospinal fluid specimens for detection by gas chromatography-mass spectrometry. Hemodialyzer rinses were Limulus assay positive; however, no beta-hydroxy fatty acids were detected by gas chromatography-mass spectrometry. These data were compared with data obtained from USP rabbit pyrogen tests of the rinse materials (nonpyrogenic) and chemical characterization of the Limulus assay-reactive rinses, which showed the rinses to be cellulosic in nature. It is suggested that beta-hydroxy fatty acids, as assayed by selected ion-monitoring gas chromatography-mass spectrometry, be used as chemical marker molecules for the presence or absence of endotoxin in materials reported to cause nonspecific activation of Limulus amoebocyte lysate.     

Limulus amoebocyte lysate assay for detection and quantitation of endotoxin in a small-volume parenteral product.

A Limulus amoebocyte lysate gel-clotting method for the determination of endotoxin in a small-volume parenteral product has been described. Sample dilution with 0.1 M potassium phosphate monobasic buffer (pH 8.0) effectively eliminated assay interference, whereas dilution with water did not. The threshold pyrogenic dose for Escherichia coli EC-2 and O127:B8 endotoxins was determined to be 1.0 ng of endotoxin per kg of body weight. Not more than 1.0 ng of endotoxin (the threshold pyrogenic dose) per the highest recommended human dose or the USP pyrogen test dose per kg of body weight, whichever dose is more stringent, is a logical limit for the quantity of bacterial endotoxin in small-volume parenteral products. Excellent correlation was attained when this criterion was used to compare the Limulus amoebocyte lysate assay with the USP pyrogen test.     

Limulus amoebocyte lysate assay for detection and quantitation of endotoxin in a small-volume parenteral product

A Limulus amoebocyte lysate gel-clotting method for the determination of endotoxin in a small-volume parenteral product has been described. Sample dilution with 0.1 M potassium phosphate monobasic buffer (pH 8.0) effectively eliminated assay interference, whereas dilution with water did not. The threshold pyrogenic dose for Escherichia coli EC-2 and O127:B8 endotoxins was determined to be 1.0 ng of endotoxin per kg of body weight. Not more than 1.0 ng of endotoxin (the threshold pyrogenic dose) per the highest recommended human dose or the USP pyrogen test dose per kg of body weight, whichever dose is more stringent, is a logical limit for the quantity of bacterial endotoxin in small-volume parenteral products. Excellent correlation was attained when this criterion was used to compare the Limulus amoebocyte lysate assay with the USP pyrogen test.     

Characterization of Limulus amoebocyte lysate-reactive material from hollow-fiber dialyzers.

Hollow-fiber hemodialyzers containing cellulose-based membranes have been shown to produce positive results with the Limulus amoebocyte lysate test. This study was undertaken to determine whether endotoxin was causing the reaction. Rinses from 45 parallel-plate and hollow-fiber dialyzers from eight different manufacturers were tested before and after treatment with cellulase, using three lysates and four Limulus amoebocyte lysate methods. In addition, four in vitro cellular methods--human leukocytic pyrogen, lymphocytic activating factor, peritoneal macrophage, and arginase release--were used to evaluate endotoxin activity. The substance causing the reaction was identified by chromatographic methods. Results indicate that the Limulus amoebocyte lysate reactive material is cellulose derived and not pyrogenic.     

Investigations on the Specificity of the Limulus Test for the Detection of Endotoxin

Lysates obtained from amoebocytes of Limulus polyphemus, the horseshoe crab, showed gel formation after the addition of bacterial endotoxin. In contrast to living gram-negative bacteria, viable gram-positive microorganisms did not cause gelation of lysate. Nevertheless, peptidoglycan isolated from the cell walls of various gram-positive organisms did induce the reaction. However, the activity of peptidoglycan was 1,000 to 400,000 times less than that of Escherichia coli lipopolysaccharide. After exposure to lysozyme, peptidoglycan no longer gelled amoebocyte lysate, therefore apparently excluding endotoxin contamination. Gelation of amoebocyte lysate by endotoxin or peptidoglycan was inhibited by different concentrations of sodium polystyrolsulfonate. Whereas these studies confirm the specificity of the Limulus test for bacterial endotoxins, they also indicate that other substances of bacterial origin should be investigated for their ability to gel amoebocyte lysate.     

Studies on Limulus amoebocyte lysate. III. Purification of an endotoxin-binding protein from Limulus amoebocyte membranes

A protein that has been isolated from Limulus polyphemus amoebocyte membranes binds endotoxin. The protein was purified by two independent methods, organic solvent extraction and affinity chromatography, both followed by gel filtration. Immunologic studies confirm that the protein is a component of amoebocyte membranes. Although without enzymatic activity, the binding protein enhances Limulus lysate gelation. As a membrane-associated endotoxin binding "protein," it may be involved in Limulus lysate coagulation, which is initiated by minute amounts of Gram-negative bacterial endotoxin. The protein has an apparent molecular weight of 80,000.     

Detection of pyrogens in intravenous IgG preparations.

Five different intravenous IgG (i.v. IgG) preparations were assessed for their capacity to modify the pyrogenic response to bacterial lipopolysaccharide (LPS) of rabbits under the conditions of a pharmacopoeal test. Four of the five preparations were found to mitigate the reaction rendering the result "non-pyrogenic" with an LPS dose proved pyrogenic when administered in saline or in albumin. Bacterial LPS was found readily detectable by a simple Limulus amoebocyte lysate (LAL) gelation test. Four of six brands of i.v. IgG were found reactive in the test under conditions adjusted to detect the FDA limit. The reaction obtained upon addition of standard LPS to the negative preparations supported the validity of the assay. The LAL reactivity of two of the reactive preparations was inhibited by laminarin, a compound known to inhibit Limulus lysate gelation by beta-D-glucan, but not by Polymyxin B. Specific detection of bacterial endotoxins in i.v. IgG solutions requires inhibition of the beta-D-glucan pathway of the Limulus lysate coagulation. Using an appropriate inhibitor, the LAL gelation test is suitable to detect a potential endotoxin contamination in i.v. IgG which might have not been unravelled by the in vivo test for pyrogens.     

Measurements of lipopolysaccharide (endotoxin) in meningococcal protein and polysaccharide preparations for vaccine usage

Lipopolysaccharide (LPS, i.e. endotoxin) present in meningococcal outer-membrane protein and polysaccharide preparations made for vaccine use was quantitated by a silver-stain method following SDS-PAGE. The reactivities of LPS in the preparations were also measured by rabbit pyrogenicity and Limulus amoebocyte lysate (LAL) assay. Although rabbit pyrogenicity and LAL assay are more sensitive than the silver stain method, the latter provided an actual amount of LPS present in the protein or in the polysaccharide. For a meningococcal protein preparation, rabbit pyrogenicity showed about one-tenth, and even less by LAL assay, of the actual amount of LPS. This is because protein-bound LPS in meningococcal protein preparations is about 10-fold less active in causing fever in rabbits, and 20- to 40-fold less active in the gelation of LAL than the same amount of a purified free LPS which is generally used as a reference in quantitating LPS in these two assays. As for the small amount of LPS present in a meningococcal polysaccharide preparation, similar LPS content was obtained when measured by the three methods suggesting that the LPS is not bound to the polysaccharide in contrast to that in the proteins mentioned above. The purified meningococcal LPS was pyrogenic in rabbits at 1 ng/kg.     

Comparison of several control standard endotoxins to the National Reference Standard Endotoxin--an HIMA collaborative study.

A collaborative study, initiated under the auspices of the Health Industry Manufacturers Association (HIMA), was designed to establish the relationship of Escherichia coli O55:B5 endotoxin (the control standard endotoxin of HIMA and the Food and Drug Administration's Office of Medical Devices) to the U.S. National Reference Standard Endotoxin and to two internationally used control standard endotoxins. By using two Limulus amoebocyte lysate test systems, it was established that the E. coli O55:B5 endotoxin lot originally used by HIMA and the Office of Medical Devices to establish Limulus amoebocyte lysate release test criteria for pyrogen testing of medical devices contains approximately 4.5 endotoxin units (EU) per ng. Thus, the 1.0-ng/kg endotoxin dose limit currently established for medical devices is approximately the same as the 5.0-EU/kg endotoxin limit (on an activity basis) established by several other Food and Drug Administration agencies for human and animal parenteral drugs and biological products.     

Comparison of several control standard endotoxins to the National Reference Standard Endotoxin--an HIMA collaborative study

A collaborative study, initiated under the auspices of the Health Industry Manufacturers Association (HIMA), was designed to establish the relationship of Escherichia coli O55:B5 endotoxin (the control standard endotoxin of HIMA and the Food and Drug Administration's Office of Medical Devices) to the U.S. National Reference Standard Endotoxin and to two internationally used control standard endotoxins. By using two Limulus amoebocyte lysate test systems, it was established that the E. coli O55:B5 endotoxin lot originally used by HIMA and the Office of Medical Devices to establish Limulus amoebocyte lysate release test criteria for pyrogen testing of medical devices contains approximately 4.5 endotoxin units (EU) per ng. Thus, the 1.0-ng/kg endotoxin dose limit currently established for medical devices is approximately the same as the 5.0-EU/kg endotoxin limit (on an activity basis) established by several other Food and Drug Administration agencies for human and animal parenteral drugs and biological products.     

A novel endotoxin-specific assay by turbidimetry with Limulus amoebocyte lysate containing beta-glucan

The gelation of standard Limulus amoebocyte lysate (LAL) is triggered by the addition of a small amount of beta-glucan (1-1000 ng/ml plasma), but in the presence of an excessive amount of beta-glucan (1 mg/ml plasma), the gelation becomes insensitive to beta-glucan. Utilizing this property, a method to determine quantitatively the amount of endotoxin circulating in humans was developed. When a modified LAL, or LAL-ES, which contains an excessive amount of CM-curdlan as beta-glucan, was used for the assay, a linear relation in the logarithmic scales was obtained between the gelation time measured by the turbidimetry (min) and the concentration of endotoxin. This relation was not affected by a considerable amount of beta-glucan (100 ng/ml). The sensitivity of the endotoxin assay was estimated to be as low as 3 pg/ml. The following aspects of the method were found by clinical application to normal and febrile subjects. (1) Using both LAL and LAL-ES, it was possible to distinguish the effect of endotoxin from that of beta-glucan in plasma, i.e., bacterial sepsis from fungal sepsis. (2) The amount of circulating endotoxin determined by the present method showed good correlation to those obtained by chromogenic assay using modified LAL devoid of Factor G which could be activated by beta-glucan.     

Selective assay for endotoxin using poly(ε-lysine)-immobilized Cellufine and Limulus amoebocyte lysate (LAL)

We developed a selective endotoxin (lipopolysaccharide; LPS) assay using poly(ε-lysine)-immobilized cellulose beads (PL-Cellufine) and Limulus amoebocyte lysate (LAL). First, LPS was selectively adsorbed on the beads in a solution containing various LAL-inhibiting or LAL-enhancing compounds (e.g., amino acids, enzymes) and the LPS adsorbed on the beads was separated from the compounds by centrifugation. Second, the LPS adsorbed on the beads directly reacted with the LAL reagent, and the LPS concentration was determined by a turbidimetric time assay. The accuracy of the adsorption method with PL-Cellufine was high compared with that of a common solution method. Apparent recovery of LPS from compound solution was 88-120%.     

Comparison of the Limulus amoebocyte lysate test with plate counts and chemical analyses for assessment of the quality of lean fish.

The quality of lean fish was assessed simply and rapidly with Limulus amoebocyte lysate. The endotoxin levels agreed with aerobic plate counts and chemical indices of spoilage. Correlation between level of endotoxin and level of total volatile bases was found to be highly significant (r = 0.8579; P less than 0.001).     

Modulation of the biological activity of bacterial endotoxin by incorporation into liposomes

In an attempt to define the mechanism by which endotoxin induces its biological activity, we studied the effect of the incorporation of lipopolysaccharide and lipid A into phospholipid vesicles (liposomes) on the stimulation of the macrophage cell-line RAW 264.7 and on the coagulation of Limulus amoebocyte lysate. The incorporation of Salmonella minnesota smooth-and rough (Re) lipopolysaccharide or primarily monophosphoryl lipid A into multilamellar and small unilamellar vesicles consisting of phosphatidylcholine, phosphatidylserine and cholesterol (molar ratio 4:1:4) reduced the interleukin 1 inducing potency of these substances about 1000-fold. When corrected for the actual uptake of radiolabeled free and liposome-incorporated lipopolysaccharide by the cells, this difference amounted to 100- to 1000-fold. In addition, liposome-associated Re-lipopolysaccharide was about 1000-fold less potent in stimulating the Fc-receptor mediated uptake of IgG-coated sheep erythrocytes by the cells. The ability of lipopolysaccharide and lipid A to coagulate the Limulus amoebocyte lysate appeared to be at least 100-fold decreased upon incorporation into phospholipid vesicles. Control experiments demonstrated that liposomes prepared without lipopolysaccharide did not reduce the studied activities of free lipopolysaccharide. In conclusion, the incorporation of lipopolysaccharide into the liposomal membrane probably prevents the interaction of the hydrophobic portion of the lipid A component of lipopolysaccharide with the plasma-membrane structures involved in the activation of macrophages and with the proteins of the Limulus amoebocyte lysate. This indicates that the direct interaction of the lipid A moiety of lipopolysaccharide with the macrophage plasma-membrane is required to optimally trigger the studied responses.     

Immunogenicity and toxicity of lipopolysaccharide from Francisella tularensis LVS

Abstract Lipopolysaccharide (LPS) from the live vaccine strain of Francisella tularensis ( F . tularensis LVS) was isolated and purified. The LPS did not stimulate lymphocytes from previously tularaemia-vaccinated individuals or lymphocytes from nonprimed individuals. However, serum antibodies from tularaemia vaccines reacted with the LPS whereas virtually no reactivity was found with antibodies from individuals not exposed to F. tularensis LVS. Antibodies of immunoglobulin class M displayed the antibody reactivity predominantly. The LPS failed to induce the mononuclear cell-derived cytokine interleukin-1 and only low levels of tumour necrosis factor were detected. Furthermore, no LPS endotoxin properties were found in galactosamine-treated mice or in the Limulus amoebocyte lysate assay. From these results it can be concluded that F. tularensis LVS possesses a lipopolysaccharide-like molecule, which does not exhibit properties of a classical endotoxin.     

Statistical procedure for evaluating the sensitivity of Limulus amoebocyte lysate by using a reference lysate.

A study was designed to estimate variability of the Limulus amoebocyte lysate test by comparing a reference lysate against itself. Three technicians performed parallel tests, i.e., titrated side by side, the contents of two vials of reference lysate on 4 different days using 24 vials of the United States reference lysate and 12 vials of the United States reference endotoxin. Each parallel test was replicated three times. From the sensitivity endpoints, ratios were calculated for each parallel test. These ratios were converted to the logarithm for estimating variability among technicians and among vials of endotoxin. By using the overall variability of log ratios, a statistical procedure was developed to evaluate the sensitivity of each lot of licensed lysate submitted to the Bureau of Biologics for release.     

Structural requirements of lipid A species in activation of clotting enzymes from the horseshoe crab, and the human complement cascade

The structure/activity relationship of lipid A, a bioactive center of endotoxic lipopolysaccharides, in the activation of the clotting enzyme cascade of a horseshoe crab amoebocyte lysate (Limulus activity) and the complement system in human serum, was examined using synthetic lipids A and related compounds. Regarding Limulus activity, a newly developed colorimetric method, which utilizes a mixture of recombined clotting factors and a chromogenic substance, was much more sensitive for detecting changes in the chemical structure of test compounds than the conventional gelation method using the amoebocyte whole lysate. (beta 1-6)-D-Glucosamine disaccharide bisphosphates, which had neither 3-hydroxyacyl nor 3-acyloxyacyl groups, and acylglucosamine phosphates, which in structure correspond or are analogous to the non-reducing or reducing moieties of lipids A and biosynthetic disaccharide lipid A precursors showed only negligible activity in the colorimetric tests, but they exhibited a distinct though much weaker gelation activity than the parent disaccharide molecules. The assay results obtained by the colorimetric Limulus test correlate better with the pyrogenicity of the test synthetic compounds than those given by the gelation method, although the dependence of pyrogenicity on chemical structure is greater. The presence of 3-hydroxyacyl groups on the bisphosphorylated (beta 1-6)-D-glucosamine disaccharide backbone is the prerequisite for effective activation of the clotting enzyme cascade of horseshoe crab amoebocyte lysate, while the presence of an adequate number (one or two) of 3-acyloxyacyl groups on the disaccharide bisphosphate backbone is needed for full pyrogenicity. Complement activation, on the other hand, showed structural requirements quite different from those for the colorimetric Limulus activity and the pyrogenicity. The disaccharide compounds that had only non-hydroxylated acyl groups, acylated glucosamine phosphates that had the structure of the non-reducing portion of lipids A and biosynthetic disaccharide precursors, which were scarcely active in the colorimetric Limulus test, caused complement activation comparable to or sometimes stronger than that of the parent disaccharide molecules. Acylglucosamine phosphates, corresponding in structure to the reducing moiety of disaccharide compounds, however, showed little activity.