Evaluation of the endotoxin binding efficiency of clay minerals using the Limulus Amebocyte lysate test: an in vitro study

Endotoxins are toxins that are kept in the cell wall of Gram-negative bacteria. In pure chemical form endotoxins are so called Lipopolysaccharides (LPS). LPS consist of a lipid part (lipid A) and a polysaccharide part with an inner core and O-specific side chains. They are characterised as amphiphilic (hydrophilic and lipophilic) molecules (Hodgson2006), which are heat resistant and very pH stable. Endotoxin activity is indicated as endotoxin units (EU). In general, around 10 EU are equivalent to 1 ng endotoxin.

Being part of bacteria, endotoxins are continuously released into the environment (Magalhaes et al.2007). In the gastrointestinal tract, endotoxins are potent immune stimulators (Mani et al.2012), which in healthy animals are removed from the gut via the reticuloendothelial system (Munford2005). If there is a failure in endotoxin deactivation, e.g. in cases of stress in animals, circulating endotoxin can lead to systemic inflammation, endotoxemia, shock and even death (Mani et al.2012; Rice et al.2003; Zweifach and Janoff1965).

A feed additive, which binds endotoxins in the gastrointestinal tract, especially in a state of overwhelming release of endotoxins and in a state of a suppressed immune system, could help to inhibit the negative impact in the animal. The idea of feeding various adsorbents (e.g. bentonite, organoclays, kaopectate, charcoal, kaolin, terra fullonica, smectite) to endotoxin or E. coli challenged animals for prevention of related diseases (especially endotoxemia), has been investigated regularly (Ditter et al.1983; Gardiner et al.1993; Song et al.2011; Trckova et al.2004). However, none of the investigated adsorbents could be clearly assigned to endotoxin binding.

Besides the in vivo approaches, the efficacy of different substances to remove endotoxins from blood (dialysis), protein solutions and pharmaceuticals has been tested extensively. Different methods (e.g. ion-exchange chromatography, ultrafiltration, sucrose gradient centrifugation) and materials (e.g. L-histidine, poly-L-lysine, polymyxin B) have been known since the early 80′s and were described by Tosa et al.(1987), Petsch and Anspach (2000) and later reviewed by Magalhaes et al.(2007).

Currently, the most sensitive test for detecting endotoxins is the kinetic chromogenic Limulus Amebocyte lysate (LAL) test. The principle of the LAL test, discovered by Bang (1956), is that a lysate prepared from the blood of Limulus polyphemus, the horseshoe crab, forms clots when endotoxins are present (Young et al.1972). Coagulation is caused by the activation of a number of enzymes located in the blood cells (amebocytes) of Limulus polyphemus, and this activation is initiated by endotoxins. In the chromogenic test, the coagulogen has been replaced by a chromogenic substrate. The last enzyme activated in the cascade splits the chromophore from the chromogenic substrate, producing a yellow colour reaction proportional to the amount of endotoxins in the system in a time dependant manner (Hurley1995). The biological reagent and the high sensitivity (0.005 endotoxin units/mL) of the test imply susceptibility for interferences (Dubczak2011). These interferences have to be overcome with sample preparation and dilution finding for each matrix.

Although testing of potential endotoxin binding materials, like bone char (a form of charcoal produced by heating bone in the presence of a limited amount of air), using the LAL test has been reported (Rezaee et al.2009), there is still not much data available on promising protocols for the kinetic chromogenic LAL assay for the evaluation of clays like bentonites and organoclays as potent endotoxin sorbents.

The objective of our study was to determine in vitro the potential of bentonites and organoclays for an application as feed additive to prevent animals from endotoxin triggered diseases. The LAL test was adapted to evaluate endotoxin binding in aqueous solutions and artificial intestinal fluid. Results of our study helped us to determine the binding efficiencies of the tested materials.

The use of clay minerals as a feed additive for farm animals to prevent the effects of different mycotoxins (e.g. aflatoxin and ergotamine), bacteria and other toxic compounds is widespread and has been discussed for years (Slamova et al.2011; Tateo and Summa2007; Trckova et al.2004). Therefore, the hypothesis is that binding endotoxins in the gut lumen will reduce the number of endotoxins entering the organism. As animal studies are expensive, and the effect of the additive has to be assured, the in vitro screening of sorbents is important for estimating their adsorptive potential. Although, results of in vitro studies may not be reflected in in vivo studies, conclusions regarding the mode of action under controlled conditions can be gained (Ganner and Schatzmayr2012). Approaches for screening binding materials in vitro and collecting information on the underlying functional mechanisms, are of great interest.

The role of endotoxins in inflammatory processes induced by Gram-negative bacteria is well known (Rietschel et al.1994). Already small amounts of endotoxin (< 1 ng/mL; approximately 10 EU/mL) (Anspach and Hilbeck1995; Mani et al.2012) given intravenously are sufficient to activate the immune system and trigger symptoms (Gorbet and Sefton2005). Even far higher dosages are expected to be present in the environment of farm animals (Ratzinger2010; Vogelzang et al.1998).

PMB is a well investigated polypeptide antibiotic, which deactivates LPS via disrupting the molecular structure (Pristovsek and Kidric1999). A two way mechanism explains the reaction. Firstly, the positively charged PMB adheres to LPS lamella because of electrostatic interaction and secondly, the acyl chain is inserted in the LPS lamella. Due to this stoichiometric binding effect, PMB was chosen as positive control in our experiments. The limitation of in vivo use of PMB is its toxicity (Danner et al.1989) and therefore, alternatives are necessary.

In our experiments, non-treated and organophilised bentonites were tested to get an idea of the endotoxin binding behaviour of sorbents in different media, and whether the LAL assay can be used to quantify endotoxin activity in different media. It has to be kept in mind that the use of organoclays as feed additives is not permitted in the European Union, but since organoclays are known to bind a broad spectrum of substances they were included in this study. In contrast, Bentonite-montmorillonite is authorized as a technological feed additive according to European Parliament regulation No 1831/2003 and can therefore, be used as a feed additive.

In older literature, fluorescein isothiocyanate (FITC)-LPS was used for fluorescence measurements (Nolan et al.1975) or the gel tube test was used to obtain quantitative results to measure the endotoxin binding in vitro (Ditter et al.1983). Gorbet and Sefton (2005) stated that many studies on materials incubated in endotoxin solutions were never reported. This fact points out the necessity of defined approaches for testing potential endotoxin sorbents.

(Our investigations were based on the set-up of testing aflatoxin binding by Lemke et al. (2001). Our first approach was a single sorption experiment with a low amount (about 10 EU/mL) of endotoxin. Differences in adsorption efficiencies of organophilic and non-treated clays were revealed. The reason that organoclays bind endotoxin to a high percentage in a clear buffers solution may be explained by the hydrophobic forces between LPS and treated clays (Hou and Zaniewski1991). Another explanation could be that, low endotoxin values in aqueous solution may not activate the binding sites of untreated bentonite.

However, far higher dosages than 10 EU/mL are expected in the environment. Thus, the next step in our experiments was, to test the influence of endotoxin concentrations (2,450 – 51,700 EU/mL in LRW solution) on the adsorption efficiency of sorbents. Organoclay 1 revealed on average a comparable binding efficiency to PMB (P = 0.385). The same behaviour was expected for Organoclay 2, however, binding efficiencies were significantly (P = 0.013) decreased compared to PMB. This could be either due to the differences of the original clay (before treatment) itself or to the treatment with different surfactants. In contrast to the swelling bentonites, which partially disperse themselves spontaneously in water owing to the swelling caused by penetration of water between the unit layers, the three-dimensional structure of attapulgite prohibits any internal swelling action, which reduces access to adsorbed molecules.

(Further calculations with isothermal models were conducted. Data of PMB, Organoclay 1 and Bentonite 1 were fitted to linear and Freundlich isotherm according to the method explained by Giles et al. (1974). PMB and Organoclay 1 revealed comparable results and suggest a Langmuir isotherm, whereas Bentonite 1 indicated a linear isotherm. No saturation levels were reached for tested sorbents, although high concentrations of endotoxin were used.

In anticipation that the material might be used as feed additive, we implemented binding experiments in artificial intestinal fluid. Samples prepared with the lower concentrated endotoxin solution showed a decreased binding efficiency of sorbents compared to higher concentrated LPS test solutions. This observation can have two reasons: Firstly, toxins bind to sites with higher affinity, followed by binding to lower affinity sites, when toxin concentration increases (Ferraz et al.2004). Secondly, increased toxin concentrations change the surface of the sorbent and activate further binding sites (Giles et al.1974) or start binding previously bound molecules (Grant et al.1998).

When compared to binding results in LRW solutions with 23,660-30,000 EU/mL (around 72%), Organoclay 1 showed a significant decrease in binding efficiency in AIF (20%; P < 0.005). In comparison to these results, binding efficiencies in LRW solution (22%) of Bentonite 1 are comparable to results in AIF solution (26%). Thus, complex media can have an influence on binding capacity of clay minerals, in our case Organoclay 1 showed a decrease of binding efficiency of over 50%. This is a very important observation, since the binding agent should act in the gastrointestinal tract, where conditions are rather complex.

These results lead to the conclusion that the adsorption of endotoxins to organoclays is mostly unspecific which limits its practical application. In addition to that organoclays do not meet regulatory requirements for feed additives.

Overall, there are only a few references on in vitro endotoxin binding studies which were done using the kinetic chromogenic LAL test, and most of these studies were done with other sorbents than clay minerals. Anspach and Hilbeck (1995) investigated various affinity sorbents (histidine, histamine and PMB) and recovered a binding capacity of 0.48 mg endotoxin/g wet sorbent in a working solution with 70,000 EU/mL (7 μg/mL). In a master’s thesis from Xu Ying (2006) the E. coli and endotoxin adsorption capacity of montmorillonite, zeolite and carbon nanotube with and without cetylpyridinium chloride (quaternary ammonium compound) treatment was described. Maximum binding capacity was reached by treated carbon nanotubes (with 357 EU/mg sorbent ≈ 35.7 ng/mg) and lowest was with zeolite (76.8 EU/mg ≈ 7.68 ng/mg). All tests were performed in 10 EU/mL working solution. Whereas, Perianayagam and Jaber (2008) tested Sevelamer Hydrochloride (a nonadsorbable, cross-linked polymer) in a 100 EU/mL solution and found a concentration dependent effect with best results (96%) at 50 mg/mL Sevelamer Hydrochloride. In a study testing bone char, Rezaee et al. (2009) revealed a binding capacity of 2.9 ng Endotoxin/g sorbent in a working solution with 80 EU/mL. With regard to a different experimental set up, to the studies mentioned above, in aqueous solution we could calculate the following absorbent capacities for the highest endotoxin concentration (50,000 EU/mL): 5.59 mg/g (Organoclay 1) > 3.97 mg/g (PMB) > 2.58.mg/g (Organoclay 2) > 1.55 mg/g (Bentonite 1) > 1.23 mg/g (Bentonite 2). Even higher maximum binding capacities of sorbents may be expected as saturation was not reached in our studies. In general we could achieve higher binding efficiencies in our experiments compared to other studies.

In conclusion, the LAL assay could be successfully used to test supernatants of endotoxin binding studies in aqueous solution and artificial intestinal fluid. The studies we evaluated (single concentration, endotoxin concentrations on adsorption efficiency in LRW solutions and sorption experiments in AIF) provide a good tool for the screening of sorbents for their potential endotoxin binding efficiency. Our results show that incubation media used for the endotoxin binding assay have an influence on the binding capacity of clay minerals. Binding mechanisms of treated and non-treated clays are different. Organoclay 1 (0.1%) showed good binding efficiencies in LRW solutions, whereas Bentonite 1 (0.1%) obtained a lower, but constant binding efficiency of high concentrated LPS solutions in LRW and AIF. In the 50,000 EU/ml aqueous buffer solution best adsorbent capacity was revealed for Organoclay 1 (5.59 mg/g), which showed a comparable binding efficiency to PMB (3.97 mg/g). Even higher binding capacities may be approached if saturation would be reached. Clay minerals, non-treated and treated, showed a promising endotoxin binding capacity in vitro. Further investigations are necessary to better understand the binding mechanism and to prove the effect of the used clays for in vivo trials.