Endotoxins: A Review

Biomedical Significance and its Detection with the Limulus Amebocyte Lysate (LAL) Assays

Bacterial endotoxins or lipopolysaccharides (LPS)

Most people are surprised to learn that a normal human being hosts a mass of bacteria equivalent to about one half kilogram; the bulk is in the gut lumen and the remainder is mainly dispersed in the skin, the rhino-oral-pharyngeal cavities and the genital mucosa. Their weight almost equals that of our heart, and because they are smaller than our cells, their total number (about 10^14) exceeds that of our own cells.

A large fraction of the bacterial flora residing in our body belongs to a certain group of bacteria termed Gram-negative bacteria, which differ from other microorganisms by an unique framework of their cell wall (Bacteria are classified as Gram-positive or Gram-negative depending on whether or not, in a procedure devised by the Danish physician Hans Christian Joachim Gram, they retain a particular blue dye). These bacteria are ubiquitous, existing not only in the gut and on the surface of our own bodies, but also on the surface of the entire planet.

Endotoxins were first described separately by the Danish pathophysiologist Peter Ludwig Panum, the German bacteriologist Richard Pfeiffer and the Italian pathologist Eugenio Centanni more than 100 years ago. In his experiment, Pfeiffer noted that lysates of the heat-inactivated Gram-negative bacterium Vibrio cholerae caused shock and death in laboratory animals, and he called the heat-resistant toxic substance(s), then not yet characterized, "endotoxin" to distinguish it from the heat-labile exotoxin secreted by the bacterial cell. It is now known that endotoxins are integral components of the outer membrane of the cell wall of Gram-negative bacteria and essential for bacterial growth and viability. They are released when the bacteria disintegrate or multiply. Endotoxins derived from various bacterial families share a common framework that consists of a polysaccharide and a covalently bound lipid A, and are hence called lipopolysaccharides(LPS).

In enterobacterial smooth-type LPS, the polysaccharide component consists of two regions, namely the O-specific chain and the core oligosaccharide, which differ in their genetic determination, mechanisms of biosynthesis and chemical structure. A variety of non-enterobacterial smooth-type strains of pathogenic Gram-negative bacteria form LPS which lack the O-specific chain and these LPS are similar to those of enterobacterial roughmutants. Lipid A is composed of a b-1,6-linked D-glucosamine disaccharide that carries two phosphoryl groups and four hydroxyfatty acids, and has been fully synthesized. It is also established that lipid A is the most highly conserved region and represents the endotoxic principle of LPS.

1

2

Actions of Endotoxin

LPS are potent and toxic macromolecules. As a constituent of the outer membrane, they play an important role in the interaction of Gram-negative bacteria with higher organisms. The host's immune system recognizes invading bacteria according to their LPS structures and reacts by the formation of antibacterial antibodies directed against LPS. On the other hand, LPS may be released by bacteria that undergo multiplication or disintegration. Once released, LPS exhibit a broad spectrum of biologic (endotoxic) activities, such as pyrogenicity, hypotension, hypermetabolism, anorexia, tissue damage, disseminated intravascular coagulation and lethal shock. Furthermore, LPS are potent immunostimulators because of their strong activation of B ymphocytes, granulocytes and mononuclear cells. It is now known that LPS do not exert their manifold actions directly. Instead, they recruit and induce host cells, particularly macrophages, to produce and secrete a variety of mediators. These mediators consists of three groups:

Proteins, which include tumor necrosis factor (TNF), interleukin-1 (IL-1), IL-6 and IL-8
Oxygen free radicals, which include O2-, hydrogen peroxide (H2O2) and nitric oxide (NO)
Lipids, which include prostaglandin E2, thromboxane A2 and platelet-activating factor.

TNF is one of the prime LPS mediators made by macrophages. Depending on the dose, the exogenous administration of TNF mimics several responses attributed to LPS, including fever, cachexia, tissue injury, irreversible shock and even death. TNF acts synergistically with IL-1, gamma interferon, or both, to trigger a systemic inflammatory response and cause damage to the vascular endothelium. TNF also provokes the release of prostaglandins and other lipid mediators.

The adverse effects of these lipid mediators include increased vascular permeability and vasoactivity and the contraction of pulm-nary smooth muscle. On the other hand, TNF can also mobilize various defensive cells to sites of infection and destroy tumor cells. IL-1 exerts many of the same effects as TNF and induces the production of more IL-1 and TNF and vice versa. In addition to fever, IL-1 induces inflammatory, hematologic and metablic changes typical of host responses to infection, such as neutrophilia, hypoferremia, synthesis of hepatic acute phase proteins and glucocorticoids.

It seems that when Gram-negative bacteria invade tissue and release moderate amounts of endotoxin there, this array of macrophage products can help eradicate the immediate infection by generating a localized and controlled immune response. The typical effects, such as mild fever, recruitment of both microbe-specific and less specialized immune components, usually serve to promote recovery and help to protect against other microbial assaults. But when an infection is severe and a large amount of endotoxin accumulates in circulating blood, acting on macrophages throughout the body and overstimulating the coagulation, complement and kinin systems, systemic release of potent mediators can produce lifethreatening shock; as the circulation fails, cells everywhere malfunction and die.

Endotoxins can enter the peripheral circulation through wounds or the portal circulation by absorption and/or translocation from the gut. Generally, however, lethal effects occur when bacteria themselves gain access to the blood. They multiply rapidly in that medium and, in the process, can liberate huge supplies of toxin to act on macrophages and other immune cells. The biological responses to LPS in the pathogenesis of Gram-negative sepsis are listed in Table 1.

3

Limulus and its blood coagulation system

Limulus is an ancient marine arthropod commonly known as the 'horseshoe crab', with a history stretching back at least 400 million years. There are only four extant species of horseshoe crab in the world, namely Limulus polyphemus, Tachypleus tridentatus, Tachypleus gigas and Carcinoscorpius rotundicauda, and all belong to the family Limulidae. Limulus blood has always held a certain fascination because of its blue color, which is due to the presence of a copper-containing hemocyanin that functions as the oxygen carrier. However, it is the amebocyte, the only type of circulating blood cell, that is of particular interest, since amebocytes contain all the components of the blood coagulation system of Limulus.

More than 40 years ago, the Danish-born pathologist Frederik B. Bang reported that a bacterial infection of Limulus polyphemus caused fatal intravascular clotting and this clotting could also be induced by a heat-stable derivative of the bacterium, subsequently shown to be a Gram-negative Vibrio species. In 1964, Levin and Bang demonstrated that extracts of the amebocytes, but not the cell-free hemolymph, would gel in the presence of endotoxin and the rate of gelation was directly related to the concentration of endotoxin.

It is now known that the coagulation of Limulus amebocyte lysate (LAL) is brought about by an enzymatic cascade, which consists of three proenzymes, factor C, factor B, and pro-clotting enzyme and one clottable protein, coagulogen. LPS activates factor C, which in turn activates factor B and proclotting enzyme, the latter cleaving coagulogen to yield an insoluble gel and releasing a soluble peptide C. The catalytic nature of each activated enzyme in the coagulation cascade serves in turn to amplify the next step, resulting in a high sensitivity of LAL to LPS.

In addition to the LPS-mediated cascade, another proenzyme, factor G, represents an alternative coagulation pathway. Factor G is activated by certain b-glucans and then directly activates the proclotting enzyme, which in turn causes the gelation of coagulogen. The mechanism of the LAL coagulation cascade is illustrated in Figure 1.

Figure 1 Mechanism of the coagulation cascade in horseshoe crab amebocyte lysate

4

Detection of endotoxin with the LAL assays

The most commonly used method for detecting the reaction of LAL and endotoxin is the gel-clot assay. Basically, equal volumes of sample and LAL (typically 0.1 ml each) are combined in a 10 x 75 mm glass tube. After an incubation period of 60 min at 37°C, the tubes are inverted 180°. A positive result is indicated by a clot that withstands the inversion. By titrating the LAL with serially diluted standard endotoxin, the minimum endotoxin concentration required to yield a positive clot can be determined. This minimum endotoxin concentration, or end point, is referred to as the LAL sensitivity. The gel-clot assay can be used as a purely qualitative limit test to rank samples as either positive or negative. However, by titrating positive samples one can obtain a semiquantitative measurement of the endotoxin concentration in unknowns by multiplying the last positive sample dilution by the LAL sensitivity.

The main drawback of the assay is the poor objectivity in judging the end point, thus giving 50% assay variation. The best available sensitivity of the assay is 0.03 EU/ml.

During the process of LAL clot formation, the reaction mixture becomes increasingly more turbid and contains a larger proportion of insoluble clotting protein. The turbidimetric assays measure the increase in turbidity as a function of the endotoxin concentration. Endotoxin concentrations in unknowns are determined by comparing the resultant turbidity with a standard curve. Since the occurrence of turbidity precedes gelation, these assays are more sensitive than the gel-clot assay. Depending on whether the turbidity is measured at the end of an incubation period or throughout the reaction, the turbidi-metric assays can be performed as an end-point assay or kinetic assay. The latter has the advantage of permitting the quantification of endotoxin over a greater concentration range. The sensitivity of the kinetic assay is about 0.01 - 0.001 EU/ml.

The chromogenic assays use a synthetic chromogenic peptide as substrate for the clotting enzyme in place of coagulogen. The chromogenic substrate is hydrolyzed by the clotting enzyme, releasing the terminal chromogenic moiety with a yellow color. The chromogenic end-point assay is normally performed in two steps, in which LAL and sample are first incubated, and substrate is then added. The resultant absorbance is determined spectrophotometrically after the reaction is stopped with acetic acid. The chromogenic kinetic assay utilizes a single colyophilized LAL/substrate reagent, which is incubated with the sample and monitored spectrophotometrically for the appearance of hydrolyzed substrate. Endotoxin in unknowns is calculated from a standard curve. Both the end-point and kinetic assays have a sensitivity of 0.01 EU/ml. The main drawback of the assays is the interference from the yellow color of specimens.

Limulus ELISA is a general technique for quantifying endotoxin by combining LAL activation with ELISA measurement of LAL components degraded during LAL-LPS reaction. During the reaction of LAL and endotoxin, coagulogen is split by the clotting enzyme and thus loses its capacity to react with certain anti-bodies raised against the whole molecule or its split product. The enzyme-linked immunosorbent assays (ELISA) measure the quantitative consumption of coagulogen or generation of the immunoreactivity of the split coagulogen using murine monoclonal antibodies (MAb) against the intact or split coagulogen (also called peptide C).

In the coagulogen ELISA, LAL and LPS are first incubated in microplate, followed by the addition of the anti-coagulogen MAb. Aliquots of the mixture are then applied to the ELISA microplate precoated with diluted LAL (source of coagulogen).

The MAb bound to the plate is detected with a horseradish peroxidase-conjugated rabbit anti-mouse antibodies. The amount of the MAb bound to the plate is directly proportional to the concentration of endotoxin in specimens. In the peptide-C ELISA, LAL and LPS are first incubated in microplate and stopped with NaOH. Aliquots of the mixture are then applied to the ELISA microplate, allowing the split coagulogen bound to the plate and subsequent detection with a horseradish peroxidase-conjugated anti-peptide-C MAb. The amount of the split coagulogen bound to the plate is directly proportional to the concentration of endotoxin in specimens. The coagulogen ELISA has a sensitivity similar to the chromogenic assay, while the non-competitive peptide-C ELISA and the sandwich ELISA can detect endotoxin concentrations as low as 0.001 EU/ml. The advantages of the ELISAs are the minimal consumption of LAL (using less than 5% of the LAL in other LAL assays) and non-interference from any color of specimens.

5

Application of the Limulus Amebocyte Lysate (LAL) Assays

With the improvement of sensitivity and specificity of the LAL methodology over the past two decades, it is now possible to quantitate endotoxins in almost any type of sample - from water to air, from drugs to medical devices, and from white milk to red blood. The practical applications include three major areas:

Control of endotoxin in drugs (including nutritional formulae) for parenteral administration, in vaccines and in medical devices for surgical use.
Quantitation of endotoxin in clinical settings as diagnostic aid.
Hygienic control of air, water, dairy products, foods, raw materials, etc.

Since its introduction in the late 1960's, the most common application of the LAL assay has been its substitution for the rabbit pyrogen test to control endotoxin contamination of parenteral drugs and medical devices (e.g. surgical gloves, catheters, implants and tubings). The assay is finding its increasing use in biological products made by genetic engineering such as peptide hormones and plasma proteins used for substitution therapy.

The LAL assay has been found to correlate with the rabbit pyrogen test. It is simpler to perform and at least 10-fold more sensitive than the latter. The classical gel-clot method was allowed by the US FDA in 1977 and included in the U.S. and European Pharmacopoeias as the Bacterial Endotoxin Test in 1979 and 1987, respectively, while the turbidimetric and chromogenic methods remained optional until their inclusion in the USP (1995), the Australian Therapeutic Goods Act (1995), the Japanese Pharmacopoeia (JP XIII, 1996) and the European Pharmacopoeia (1998). Besides its use for testing final products (e.g. the entire range of medical devices), the assay is frequently applied for inprocess control, in order to minimize the risk of contamination of end-products.

Clinical application of the LAL Assay has been sought since the early 1970's. Its utility for the detection of endotoxin has been well established in four body fluids: Cerebrospinal fluid, urine, cervical and urethral exudates, and ocular exudates. Retrospective reviews of data indicate that the LAL assay is a reliable indicator of the presence of Gram-negative bacteria in cerebrospinal fluid, with its overall diagnostic sensitivity and specificity approaching 100%. Similar sensitivities and specificities have been obtained when the assay is applied to the detection of clinically significant Gram-negative bacteriuria (>105 bacteria per ml urine) and the presumptive diagnosis of gonorrhea; for both applications, the specimens need to be properly diluted to exclude false-positive reactions due to the presence of normal commensal bacteria. In contrast to bacteriological methods, the LAL assay does not give false negative results due to antibiotic treatment.

The detection of endotoxin in blood has been impeded by the presence of factors in plasma that interfere with the LAL-LPS reaction and confound the reading of the assay. To overcome the interference, various methods of plasma pretreatment have been described.

These include dilution plus heating, chloroform extraction, alkalinization with NaOH-KCl, acid +/- oxidation with perchloric, trifluoroacetic and nitric acids, and the use of detergents. Although none of the methods mentioned are ideal, comparative studies favor the choice of dilution plus heating or oxidation/precipitation with perchloric acid. A recent metaanalysis of 45 studies revealed a relatively low concordance of endotoxinaemia with Gram-negative bacteria, implying that these two phenomna may not necessarily be linked.

In another study including 483 consecutive febrile patients and using the chromogenic LAL assay, it was concluded that the detection of endotoxin is a more accurate predictor of sepsis than the traditional bacterial culture test. Due to the rapid progress of bacterial infection, multiple sampling and simultaneous assay of other parameters such as cytokines released during sepsis are recommended in order to obtain more reliable diagnosis of Gramnegative infections.

Body fluids in which the LAL assay is suitable for detection of endotoxin:
Plasma, urine, cerebrospinal fluid, urethral exudates, cervical secretions, synovial fluid, corneal ulcer material, amniotic fluid, middle ear fluid, saliva, sputum, plaque from teeth, tear fluid.

Another important application of the LAL assay is to control the bioburden (the index for microbial contamination) of food and dairy products. Since Gram-negative bacteria (the source of endotoxin) are ubiquitous in our environment, the level of endotoxin can be used to reflect the general hygienic condition of foods and raw material. We have found that the bacterial growth in refrigerated minced meat is directly proportional to the increase of endotoxin level. Since endotoxins are heat-stable, their level in sterilized foods reflect the microbial contamination before sterilization. Bacterial culture and endotoxin determination are complementary, both giving relevant information about the hygienic status of the material tested.

Some areas and products where endotoxin determination is suitable for hygienic control:

Infant formulae, dairy products, meat, fish and eggs. Evaluation of new techniques for food storage and shelf life; bottled water and drinking water in general; dust and air in factories; evaluation of air conditioning and humidifiers.

Since the LAL assay is based on a series of enzymatic reactions activated by environmentally ubiquitous endotoxins, it is very susceptible to both chemical interferences and operational contamination. Expertise and experience with the assay are essential to obtain relevant and reliable quantitation of endotoxin.

A number of modifications of the LAL assay have been reported in the scientific literature and each of them has its own advantages and disadvantages. However, with an experienced hand and proper selection of published methods, it is possible to determine the endotoxin concentration in almost any sample.

Most interferences with the LAL assay can be eliminated by simple dilution with endotoxin-free water. The higher sensitivity a method has, the more effective it can eliminate interference by dilution. When an interference (by e.g. protease inhibitors) cannot be overcome by the simple dilution procedure, the combination of dilution with heating of samples at 75°C usually gives a better result, as these interfering factors are heat-labile.

Due to the diversity of interfering factors, assay procedures for each kind of sample should be evaluated. A pass-fail test may be used for samples usually free of endotoxin. However, a quantitative assay is advantageous in practice.

References

Zhang, G-H:
Immunochemical detection of reaction between endotoxin and the Limulus amebocyte lysate and its potential applications.
Ph.D. Thesis, University of Copenhagen, 1995.

Friberger, P:
Endotoxin determination with Limulus lysate in clinical diagnosis and hygienic control.
In Gupta A (ed.): Immunology of insects and other arthropodes. pp. 449-479, 1991.
CRC Press, Boca Raton, U.S.A.

Mikkelsen, T:
The Secret in the Blue Blood.
Science Press, Science
Press Foreign Language Book No. 134, Beijing, People´s Republic of China, 1988.

Neither this publication nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage and retrieval system, without permission in writing from the CEO of Scan Dia Labs Aps.