Sage Journals: Discover world-class research

Abstract

Endotoxin contaminations of engineered nanomaterials can be responsible for observed biological responses, especially for misleading results in in vitro test systems, as well as in vivo studies. Therefore, endotoxin testing of nanomaterials is necessary to benchmark their influence on cells. Here, we tested the traditional Limulus amebocyte lysate gel clot assay for the detection of endotoxins in nanoparticle suspensions with a focus on possible interference of the particles with the test system. We systematically investigated the effects of nanomaterials made of, or covered by, the same material. Different types of bare or PEGylated silica nanoparticles, as well as iron oxide–silica core shell nanoparticles, were tested. Detailed inhibition/enhancement controls revealed enhanced activity in the Limulus coagulation cascade for all particles with bare silica surface. In comparison, PEGylation led to a lower degree of enhancement. These results indicate that the protein–particle interactions are the basis for the observed inhibition and enhancement effects. The enhancement activity of a particle type was positively related to the calculated particle surface area. For most silica particles tested, a dilution of the sample within the maximum valid dilution was sufficient to overcome non-valid enhancement, enabling semi-quantification of the endotoxin contamination.

Keywords

Coagulation endotoxin horseshoe crab lipopolysaccharide nanomaterials

Introduction

The development and production of nanomaterials and their applications are steadily increasing. There is a large diversity of nanomaterials available because they can be produced from virtually every material,¹ made in different shapes and sizes, and can be functionalized by additional surface coatings. Especially for medical applications, engineered nanomaterials have to be tested for toxicologic and immunologic effects on living cells and organisms.^2–4 In addition to thorough physicochemical characterization of the nanomaterials, the detection of contaminations with pyrogenic substances, especially bacterial endotoxins, is of major importance. Since at least the 1990s it has been known that endotoxin contamination of commercial particles can be responsible for observed biological responses.^5,6 Just recently, it was demonstrated^7–9 again that undetected endotoxin contaminations in particle suspensions can cause misleading results in in vitro test systems. Therefore, endotoxin contamination of nanomaterials is a critical factor that influences the general assessment of the safety of nano-objects, as well as the preclinical development and approval of the nanomaterials as pharmaceutical product or medical device. One of the most widely used endotoxin detection methods is the Limulus amebocyte lysate (LAL) gel clot assay. This in vitro test system is applied by the pharmaceutical industry for the evaluation of several medical products. It is prepared of the hemolymphe (blood) of the horseshoe crab Limulus polyphemus (Arthropoda), which contains LPS-sensitive hemocytes—the so-called amebocytes. Exposure of the amebocyte to Gram-negative bacteria or LPS induces a rapid degranulation of the cells and exocytosis of defense molecules of the innate immune system, resulting in an extracellular coagulation reaction.¹⁰ Even today, the blood of wild horseshoe crabs is used to produce the LAL, which contains all the components necessary to perform the coagulation cascade. The presence of endotoxin in a sample triggers the LPS-sensitive serine protease factor C, which initiates a serine protease mediated series of enzymatic reactions, leading to an amplification of the LPS stimulus. The final hydrolyzation of the clot table protein coagulogen to coagulin results in the formation of a firm gel clot by self-association of the coagulin monomers.¹¹ For a review of the molecular details of the coagulation cascade see Kawabata et al.;¹² for the history, assay methodology and other aspects of the LAL detection system, see Novitsky.¹³ It is known that several traditional cell- or enzyme-based assays can be affected by nanoparticles.^14–16 The LAL-based assays are no exception.^17,18 For LAL assays with optical read-out (turbidimetric, chromogenic), interference is mostly based on the inherent optical properties of the nano-objects. Additionally, other mechanisms, such as enzyme inhibition, as proposed by Schulze et al.,¹⁹ or substrate-masking, are potential factors leading to false results. The mechanisms behind the observed interference events are often unclear. As mentioned above, the under- and overestimation of the endotoxin content in a medical product can have a significant effect on its approval (in case of overestimation) and, in the worst case, in severe health impacts on the patients (in case of underestimation).²⁰ Therefore, the establishment of a reliable method for endotoxin determination and quantification in nanoparticle suspensions is urgently needed. Moreover, researchers have to be aware of the necessity of testing their nanoparticle suspensions for endotoxin contamination and to document in future publications the results of performed controls, as well as observed interferences.

Here, we tested the traditional LAL gel clot assay for the detection and semi-quantification of endotoxin in dispersions of silica nanoparticles synthesized for different biological and biomedical applications. In contrast to other studies that compared the effects of different kind of nanomaterials on the assay systems,^18,19 we systematically investigated the effects of nanomaterials made of the same outer material. Six different dispersions of bare or PEGylated silica nanoparticles, as well as of four dispersions of iron oxide–silica core shell nanoparticles, were tested. The gel clot assay—the only LAL-based assay which is not based on an automatic optical read out—was selected owing to the strong absorbance of the tested iron oxide–silica core shell particles at the typical read-out wavelength (405 nm) of the chromogenic LAL assays. In addition, the gel clot assay is considered to be less prone to interfering substances.²¹ According to the European Pharmacopoeia, for the bacterial endotoxin test, in case of ‘doubt and dispute’, the final decision is made based upon the LAL gel clot assay limit test.²² The focus of this study was on the detection of possible interferences, inhibition or enhancement effects of the particles with the assay test system.

Materials and methods

Particle synthesis

Silica particles were prepared by a modified Stöber process²³ or according to a modified method described by Hartlen et al.²⁴ through L-arginine-catalyzed hydrolysis of tetraethoxysilane (TEOS) in a biphasic water/cyclohexane system.²⁴ Iron oxide cores for the iron oxide–silica core shell particles were supplied by Nanogate AG (Quierschied-Göttelborn, Germany) and further coated with silica using a modified Stöber process. Several tested nanoparticles contained an integrated fluorescence label (^f) with the deep red fluorescence dye Atto647N-NHS, which was introduced into the silica matrix by coupling to a silane linker molecule and by addition of the modified Atto dye to the biphasic water/cyclohexane system. PEGylation was achieved by deposition of a mPEG750-modified silane linker onto the particle silica surface. All chemicals used for particle synthesis were purchased from Sigma-Aldrich (Schnelldorf, Germany). Atto647N-NHS ester was supplied by Atto-Tec (Siegen, Germany). Endorem, a formulation of iron oxide nanoparticles used for magnet resonance imaging (MRI), was obtained from Guerbet (Sulzbach, Germany) and used as a reference material. After synthesis, silica and iron oxide–silica core shell particles were purified by dialysis against endotoxin-free Milli-Q water, followed by sterile filtration in a sterile work bench using cellulose acetate membranes with 0.2-µm pore size. Particles were stored at 4–8℃ until further use.

Particle characterization

Aliquots of the sterile filtered nanoparticle suspensions were applied for particle characterization. Particle size and morphology were determined by scanning or transmission electron microscopy (SEM/TEM). SEM samples of undiluted nanoparticle suspensions were prepared by vacuum drying onto silicon wafers. The hydrodynamic particle diameter was recorded by dynamic light scattering (DLS; Dyna Pro Titan, Wyatt Technology Europe, Dernbach, Germany). Zeta-(ζ)-potential measurements of 1-ml samples were carried out in disposable cuvettes at 25℃ with a Zetasizer Nano (Malvern, Germany). UV and visible light spectroscopy (UV-vis) measurements in the wavelength range of 350–800 nm were performed with a Cary 5000 spectrophotometer (Varian, Darmstadt, Germany). The pH values of all tested samples were controlled and in the required pH range of the assay (pH 6–8). No adjustment of the pH was necessary.

Endotoxin tests

Endotoxin tests were performed with the LAL Gel Clot Assay Lonza (Pyrogent™ Plus N294-03; Lonza Walkersville, Walkersville, MD, USA). All materials used were sterile and labeled to be pyrogen-free by the manufacturers. Endotoxin detection was performed according to the protocol ‘Nanokon SOP 2.2.2: Detection and semi-Quantification of Endotoxin Contaminations in Nanoparticle Suspensions – Limulus amebocyte lysate (LAL) Gel Clot Assay’ published on the DaNa homepage.²⁵ In principle, the protocol is based on information given by the European Pharmacopoeia, Monograph 2.6.14. Bacterial Endotoxins,²² ISO 29701.^26,27 The labeled assay sensitivity of the LAL gel clot assay (0.03 EU/ml) was controlled according to the manufacturer’s instructions and confirmed for all assay lots used. Based on Neun et al.,²⁸ a standard particle concentration of 1 mg/ml was used as a starting point for a twofold dilution series down to the maximum valid dilution (MVD). The MVD was calculated under the application of an endotoxin limit of 0.5 EU/ml for medical devices. This endotoxin limit for medical devices is applied for nanoparticle formulations as long as no real dose is established. For a standard particle concentration of 1 mg/ml, the calculated MVD for the LAL gel clot assay with a sensitivity of 0.03 EU/ml was 16. Nanoparticle suspensions were diluted in endotoxin-free reagent water (LAL reagent water W50; Lonza Walkersville). In case of positive results, nanoparticle suspensions were tested for possible interference of (1→3)-β-D-glucans by the application of β-glucan blocker (Beta-G-Blocker Kit N190; Lonza Walkersville). To further validate the test results, inhibition/enhancement controls (IEC) were performed for all particles tested and according to the above-mentioned standard operating protocol and the manufacturer’s instructions. Endotoxin-free tested silica nanoparticle samples were spiked with the certified reference standard endotoxin Escherichia coli O55:B5 (Lonza Walkersville). Each nanoparticle sample and IEC was tested with two replicates. Measurements were repeated with independent dilution series.

Results

Several nanoparticle suspensions were tested for endotoxin contamination. The physicochemical properties of the tested nanoparticle dispersions are given in Table 1. An overview of the IEC results is given in Table 2.

Table 1.

Physicochemical properties of the tested particle suspensions.

Particle	Synthesis	Hydrodynamic diameter (nm), in water (DLS)	Primary particle size (nm) (TEM/SEM)	Zeta-potential (mV), 25℃, in water
Endorem	–	246	5	−35
Fe_xO_y@SiO₂-1	Stöber	16.6 ± 5.0	3–5	−44.7
Fe_xO_y@SiO₂-2^f	Stöber	22.1 ± 6.6	3–5	−42.9
Fe_xO_y@SiO₂-3^f	Stöber	88.9 ± 49	3–5	−43.2
Fe_xO_y@SiO₂-4^f	Stöber	71.1 ± 10	75	−41.4
Silica-1-25	Hartlen	20.8 ± 2.6	24.6 ± 3.4	−24.2
Silica-2	Hartlen	97.6 ± 39	42 and 108	−42.9
Silica-3-80	Hartlen	84.2 ± 6.7	80 ± 5	−42.3
Silica-4-130	Hartlen	141.4 ± 15	128.3 ± 11.0	−40.5
Silica-5^f	Stöber	104.0 ± 35	113 ± 14	−43.2
Silica-6 + PEG^f	Stöber	96.0 ± 34	113 ± 14	−41.2

Table 2.

IEC results for all tested particles.

Particle	IEC results at the MVD^a	Measured assay sensitivity (EU/ml)^a	IEC acceptance criteria met?^b
Endorem	Labeled sensitivity	0.03	Yes
Fe_xO_y@SiO₂-1	Enhancement	<0.000875	No
Fe_xO_y@SiO₂-2^f	Enhancement	<0.007	No
Fe_xO_y@SiO₂-3^f	Enhancement	<0.007	No
Fe_xO_y@SiO₂-4^f	Enhancement	0.007	No
Silica-1-25	Labeled sensitivity	0.03	Yes
Silica-2	Enhancement	0.015	Yes
Silica-3-80	Enhancement	0.015	Yes
Silica-4-130	Enhancement	0.007	No
Silica-5^f	Enhancement	0.015	Yes
Silica-6 + PEG^f	Enhancement	0.015	Yes
No particle control (water)	Labeled sensitivity	0.03	Yes

Measured at MVD of the sample, corresponding to a particle concentration of 62.5 µg/ml in ultra-pure water (MilliporeQ).

The labeled assay sensitivity λ is 0.03 EU/ml and was confirmed by us. IEC acceptance criteria are met if the measured assay sensitivity is between 2 λ and ½ λ.

Silica particles: bare and PEGylated

For the majority of silica particles no endotoxin contamination (>0.03 EU/ml) was detected for all particle concentrations tested (1 mg/ml to 62.5 µg/ml particles). Only one of six silica nanoparticle samples contained low amounts of endotoxin. For all particles, IEC were performed at the MVD, as well as at least one other particle concentration. Generally, all silica particles showed coagulation enhancement activity. If tested at the MVD, which corresponds to a concentration of 62.5 µg/ml particles, only Silica-1-25 did not change the sensitivity of the assay. In contrast to the other silica nanoparticle suspensions, Silica-1-25 showed a zeta-potential of −24.2 mV instead of −40 to −45 mV. Four silica particle suspensions (Silica-2, Silica–3–80, Silica-5^f and Silica-6-PEG^f) showed slight enhancement activity at a concentration of 62.5 µg/ml particles, which was within the limits of acceptance (0.5–2λ) given by the European Pharmacopoeia.²² Silica-4-130 (see Figure 1), a suspension of monodisperse silica particles with a primary particle size of around 130 nm, showed enhancement beyond the limits of acceptance if tested in concentrations within the valid dilution range. Concentration-dependent IEC were performed for Silica-3-80, showing a slight assay enhancement (sensitivity to 0.015 EU/ml) at all tested particle concentrations (500 µg/ml, 250 µg/ml, 125 µg/ml and 62.5 µg/ml), but, interestingly, there was no visible difference in the degree of enhancement. Concentration-dependent IEC were also performed for Silica-2, a mixture of medium (42 nm size, 72%) and large (108 nm size, 28%) nanoparticles. Here, a clear concentration dependency of the degree of enhancement was obtained, decreasing by reduction of the applied particle concentration. The results are given in Table 3.

Figure 1.

SEM image of monodisperse silica particles silica-4-130 with a primary particle size of 128.3 ± 11 nm. The spherical particles were obtained by a modified method described by Hartlen et al.²⁴ through L-arginine-catalyzed hydrolysis of TEOS in a biphasic water/cyclohexane system.

Table 3.

Concentration-dependent IEC for Silica-2.

Particle concentration (µg/ml)	IEC endotoxin value (EU/ml)
Particle concentration (µg/ml)	0.03	0.015	0.007	0.0035	No spike
1000	+	+	+	+	−
500	+	+	+	+	−
250	+	+	+	−	−
125	+	+	−	−	−
62.5 (MVD)	+	+	−	−	−
Assay sensitivity enhancement factor	1	2	4	8

+: positive reaction, clot formation; −: negative reaction, no clot formation.

The labeled assay sensitivity (λ) is 0.03 EU/ml. The limit of acceptance for assay enhancement is 0.015 EU/ml (enhancement factor 2).

To investigate the influence of surface modification on the assay results and enhancement, silica particles modified with PEG were included and compared with particles without PEGylation. At a high particle concentration (500 µg/ml SiO₂), the results obtained with Silica-5^f and Silica-6 + PEG^f diverged, and the detected assay enhancement was of different intensity. PEGylated particles (Silica-6 + PEG^f) showed a lower degree of enhancement, which was within the acceptance range, whereas the enhancement by the non-PEGylated particles (Silica-5^f) was beyond the acceptance range of the assay. The results indicate that PEGylation reduces the enhancement activity of silica nanoparticles.

Iron oxide–silica core shell particles (Fe_xO_y@SiO₂)

The tested iron oxide–silica core shell particles Fe_xO_y@SiO₂-1, Fe_xO_y@SiO₂-2^f and Fe_xO_y@SiO₂-3^f originated from three different synthesis batches following the same basic synthesis protocol. In case of Fe_xO_y@SiO₂-2^f and Fe_xO_y@SiO₂-3^f, the protocol was slightly modified to introduce a fluorescence dye within the particle matrix. All three particle-suspensions showed similar behavior in the LAL gel clot assay. There was no difference observed due to absence or presence of fluorescence dye within the matrix. At high concentrations (1 mg/ml and 500 µg/ml particles) total inhibition of the assay occurred, indicated by a lack of gel clot formation. In addition, particle precipitation was observed. At medium concentrations (250 µg/ml and 125 µg/ml particles) no visible particle agglomeration occurred. Instead, the reaction was positive, resulting in the formation of a stable and homogenously light brown-colored gel clot. At the MVD (62.5 µg/ml particles), no clot formation was observed (see Table 4); however, the IEC for all three particle types at particle concentrations tested to be free of endotoxin showed strong enhancement beyond the limits of acceptance (0.5–2λ), given by the European Pharmacopoeia.²² Therefore, no valid endotoxin value could be determined for all these particle suspensions. For the unlabeled particle suspension, Fe_xO_y@SiO₂-1, concentration-dependent IEC were performed. The tested concentrations in part exceeded the MVD, ranging to 15.625 µg/ml particles (see Table 5). The results clearly show that the activity of enhancement is dependent on the particle concentration. At the lowest tested particle concentration (15.625 µg/ml) the assay sensitivity reached the labeled assay sensitivity of 0.03 EU/ml. For comparison, larger iron oxide–silica core shell particles Fe_xO_y@SiO₂-4^f with a primary particle size of 75 nm were tested. Similar to all other tested particles with a silica surface, the particles showed enhancement of the LAL assay when diluted within the MVD. In contrast to the small iron oxide–silica core shell particles with a primary particle size of 3–5 nm, no full inhibition of the assay could be observed at a high concentration of 500 µg/ml Fe_xO_y@SiO₂-4^f particles.

Table 4.

Results of the endotoxin tests for Fe_xO_y@SiO₂-2.

Negative control (No particles)	Particle concentration (µg/ml)
Negative control (No particles)	1000	500	250	125	62.5 (MVD)	Positive control 0.06 EU/ml
−	−^a	−^a	+	+	−	+
−	−^a	−^a	+	+	−	+

Assay inhibition is visible by a lack of clot formation at high particle concentrations. Precipitation of particle aggregates occurred, forming a dark brown bottom layer covered by a clear liquid upper layer.

Table 5.

Concentration-dependent IEC for Fe_xO_y@SiO₂-1.

Particle concentration (µg/ml)	IEC endotoxin value (EU/ml)
Particle concentration (µg/ml)	0.03	0.015	0.007	0.0035	0.00175	0.000875
62.5 (MVD)	+	+	+	+	+	+
50	+	+	+	+	−	ND
40	+	+	+	−	−	ND
31.25	+	+	−	−	−	−
15.625	+	−	−	ND	ND	ND

ND: not determined; +: positive reaction, clot formation; −: negative reaction, no clot formation.

The labeled assay sensitivity (λ) is 0.03 EU/ml; the limit of acceptance for assay enhancement is 0.015 EU/ml.

Endorem (Fe_xO_y)

Endorem, a formulation based on super paramagnetic iron oxide nanoparticles with a primary particle size of 5 nm, was tested as a reference to the applied iron oxide–silica core shell nanoparticles with a primary particle size of around 3 nm. Endorem (AMI 25) has been approved as contrast agent for the detection of liver metastases by MRI and is applied via intravenous injection. As expected, no endotoxin contamination was detected in diluted and undiluted Endorem. Endorem showed no interference with the LAL gel clot assay if applied in the lowest tested particle concentration (62.5 µg/ml particles). Even in undiluted state, at a particle concentration of 11.2 mg/ml spiked with endotoxin, no full inhibition of the LAL gel clot assay was detected. A stable and homogenously colored brown gel clot was formed, without any sign of visible particle precipitation.

Summary

Detailed inhibition/enhancement screening revealed that in the presence of endotoxins, all tested particles with a silica surface showed enhancement activity in the Limulus coagulation cascade. The results of the control experiments in the absence of endotoxins show that the interaction of the silica particles with the LAL coagulation system alone is not sufficient to lead to a gel clot within 60 min of incubation. Therefore, the presence of endotoxins is necessary to sufficiently trigger the coagulation cascade and to result in a stable gel clot. The enhancement effect of silica and iron oxide–silica core shell particles was shown to be clearly dependent on the particle concentration. For Fe_xO_y@SiO₂-1, it was demonstrated that an increase in particle concentration, which is directly coupled with an increase of the available surface area (essentially the sum of all the particles’ surfaces), is positively related to the enhancement activity. In contrast, no direct relation between the particle size and the enhancement effect could be found. On the basis of mass concentration, small particles are expected to show a higher enhancement activity than larger particles owing to the larger specific surface area. Instead, the particles with the largest sphere size, silica-4-130, caused significant enhancement, whereas silica particles with a diameter of 25 nm did not lead to enhancement if both were tested at the same particle concentration. Comparison of PEGylated and non-PEGylated silica particles showed differences in enhancement activity if tested at high concentrations or, rather, at a low dilution of the suspension. PEGylation resulted in a decrease of the enhancement activity.

At high particle concentrations, iron oxide–silica core shell particles with a small primary particle size of around 3–5 nm led to inhibition of the assay, visible by a lack of clotting. This was not observed using larger iron oxide–silica core shell particles if tested at similar mass concentration. In contrast, dextran-coated iron oxide nanoparticles, Endorem, which have a primary particle size smaller than the tested iron oxide–silica core shell particles, did not lead to a full inhibition of the assay, even if the particle concentration was elevated by a factor of 10 compared with the iron oxide–silica core shell particles.

Reliability of the test results

For most tested silica particles a dilution of the sample within the MVD was sufficient to overcome non-valid enhancement, enabling semi-quantification of the endotoxin contamination. In contrast, for one of the tested silica particle suspensions and all tested iron oxide–silica core shell nanoparticles, enhancement at the MVD was still beyond these limits. For these nanoparticle suspensions, other endotoxin detection methods must be evaluated.

Discussion

The present study shows that nanoparticles with silica surface can interfere with the LAL gel clot assay. Assay interference was detectable as either enhancement or inhibition. The degree of enhancement was variable between nanoparticle samples made of the same material. As mentioned before, assay interference by nanoparticles is a common problem to diverse assay formats.^14–16 However, to date, only a few studies have investigated the effects of nanomaterials on the endotoxin detection systems, and showed that several organic and inorganic nanoparticle suspensions interfere with various LAL formats.^9,17,18 In this context a systematic study of nanoparticles made of the same material is missing. It is important to note that most tested conventional substances also show some degree of interference with LAL assays,²⁰ for example by masking the presence of endotoxin, alteration of the endotoxin aggregate structure or coordination divalent cations (Ca²⁺, Mg²⁺). Owing to their anionic and amphiphilic character, endotoxins have the potential to bind to virtually every organic and inorganic material. In addition to accumulation on the surface of nanoparticles especially designed for endotoxin binding,^29–31 endotoxins were found to bind to several non-functionalized surfaces, for example aluminum silicates,³² polyethylene, titanium and silica.^5,6,33 In case of the here-tested nanoparticle samples, it would be expected that a tight binding of the LPS to the silica surface would reduce the detected endotoxin values. But the presence of a LPS spike within the nanoparticle suspension does not prevent recognition by the factor C. This might be based on the fact that endotoxin binding to certain molecules is dependent on the ionic strength of the surrounding medium. Through contact of the nanoparticle suspension with the LAL, the ionic strength of the surrounding medium can be changed significantly. The nanoparticles become coated with proteins once they enter a biological fluid and various proteins are introduced that compete for the available particle surface. Both the ionic strength and the presence of proteins has the potential to affect binding of the LPS to the silica particle surface. In addition, it cannot be excluded that LPS is bound to the silica particle surface in a way that makes the LPS more accessible for the LPS-sensitive factor C.

Another explanation for the observed assay enhancement activity goes into the mechanism of coagulation. The human coagulation and immune system is very different from the ancient system of the horseshoe crab. The coagulation cascade of the horseshoe crab is one of the best studied invertebrate coagulation systems, and several defense factors and mechanisms have been identified within the last decades.³⁴ Just a few years ago it was found that factor C, next to its role as pattern recognition receptor for LPS on amebocytes³⁵ and initiator of the LPS-triggered coagulation cascade, acts as an LPS-responsive C3 convertase.³⁶ Factor C contains a sushi domain found in human complement factors.³⁷ Despite the differences, it might be possible that observations made on the human coagulation and complement system can be used to interpret the data obtained here. Studies on the protein corona of different sized silica particles in human plasma showed a high percentage of bound coagulation factors and members of the complement system.³⁸ As these factors contain highly conserved structural motifs, it is likely that factors of the Limulus coagulation cascade, which possess similar structural motifs, also bind to the silica surface of the nanoparticles.

At least since the beginning of the twentieth century it has been known that exposure of human blood to foreign surfaces can result in blood coagulation.³⁹ These observations and further studies led to the discovery of the contact activation pathway, also known as the intrinsic pathway of the human coagulation system. Blood in contact with negatively-charged surfaces, for example amorphous silica, quartz and kaolin, triggers the intrinsic pathway by activation of factor XII, the Hageman factor,⁴⁰ which can also be activated by the presence of endotoxins.⁴¹ This contact activation by silica is used to shorten coagulation times in blood collection tubes.⁴² In the LAL gel clot assay, shorter coagulation times would cause clotting of samples that would normally not result in clot formation by the end of the assay incubation time. Experimental proof of the above-mentioned assumptions must be made in future experiments, especially measurement and comparison of the coagulation times of the LAL in the presence of the different nanoparticle suspensions. The pro-coagulant properties of several silica nanoparticles on human whole blood and plasma were demonstrated a few years ago.⁴³ It was shown that coagulation enhancement relies on an immediate activation of factor XII of the intrinsic coagulation pathway. Even if a contact activation mechanism similar to the human coagulation system has not yet been reported in Limulus, it is obvious that the observed enhancement is based on a close interaction of proteins and the nanoparticle surface. Maitz et al.⁴⁴ also reported assay enhancement in the LAL gel clot assay when endotoxin detection was performed on bulk materials and biomedical devices directly immersed in the LAL reagent. The detected endotoxin values were dependent on the surface properties. As mentioned above, our results show that an increase in particle concentration, which is directly coupled with an increase of the available surface area, is positively related to the enhancement activity (see Figure 2). The divergent results obtained with nanoparticle suspensions of different particle sizes might be explained by differences in the physicochemical characteristics of the suspensions. Based on a mass-to-mass basis, it was expected that smaller particles lead to enhanced coagulation activity in comparison with large particles. In this study, there was no clear size dependency, but a tendency for smaller particles to lead to a higher enhancement activity. The results obtained for Silica-1-25 might be explained by a difference in surface charge. In comparison with the other tested silica particle suspensions, Silica-1-25 showed a shift in the zeta potential from around −40 mV to around −20 mV. The zeta potential displays the amount of certain surface groups, for example −OH groups and the acidity of the surface, which, in turn, influence the protein adsorption.⁴⁵

Figure 2.

Sensitivity of the LAL gel clot assay in dependance on the applied concentration of the iron oxide-silica core shell particles Fe_xO_y@SiO₂-1 and calculated surface area respectively. Labeled sensitivity of the assay is 0.03 EU/ml. Elevation of the particle concentration leads to an increase in surface area available for protein–particle interactions, which becomes noticeable in enhancement of the assay sensitivity.

The assumption that protein–particle interactions are the driving force for the assay enhancement is strengthened by the results obtained from the comparison of silica particles with and without PEGylation. As expected, the PEGylated particles differ in enhancement intensity compared with the non-PEGylated particles. This can be explained by the protein-repelling characteristics of PEGylated materials. PEGylation of titanium surfaces, known to initiate the intrinsic pathway, was shown to enhance blood compatibility by reducing protein–surface interaction and contact activation, resulting in prolonged coagulation times of human blood in contrast to non-modified titanium surfaces.⁴⁶ Similar observations were made by Tavano et al.⁴³ by comparison of PEGylated and non-PEGylated ORMOSIL nanoparticles.

In this study, next to enhancement we also observed assay inhibition. Inhibition was caused by small-sized nanoparticles (∼ 3–5 nm), which have an extremely large surface-to-volume ratio, a strong curvature and high number of atoms on the particle surface. At a comparable mass concentration, the available surface area for interaction with proteins is much larger for small particles than for large particles (see Figure 3).¹ The ratio of the available surface to the amount of proteins present in the LAL is different for small and large particles. There are two possible explanations for the observed inhibition. First, on the one hand, the small particles might destabilize the gel clot by binding to coagulin monomers to a high extent, impairing the formation of a firm gel clot. On the other hand, low particle concentrations might lead to a stabilization of the gel clot, resulting in enhancement. Second, concerning only the available surface area for protein binding, enhanced adsorption of a significant amount of coagulation factors or other proteins involved in coagulation could reduce the coagulation rate up to a rate that does not lead to the formation of a firm gel clot within the assay incubation time. An excess of surface area over the amount of proteins available for binding can lead to binding of proteins that have lower binding affinity to the particles,⁴⁷ and are normally not able to compete with high-affinity proteins. Coagulogen is a possible candidate to bind to the particle surface. Coagulogen has functional, but not structural, similarity to vertebrate fibrinogen, and the coagulin monomer is a functional analogue to the fibrin monomer.^11,48,49 In single protein studies, as well as in studies on human blood samples, it has been shown that fibrinogen can adsorb to silica particles,^50,51 and that its secondary structure changes when adsorbed onto small silica particles with a radius < 30 nm.⁵² In the 1980s Vroman et al.^53,54 observed that on clot-promoting surfaces (glass or oxidized metals) fibrinogen is initially deposited from plasma, but is displaced by high molecular mass kininogen and other molecules within a few minutes—an effect called the ‘Vroman effect’. The increased surface area of the small iron oxide silica particles might reduce the competition with high affinity molecules and enable the adsorption of a significant amount of coagulogen/coagulin molecules onto the particle surface, which, in turn, would lead to a reduction of free coagulogen/coagulin molecules within the lysate. In addition, a change in secondary structure would decrease the amount of intact coagulogen molecules. Both effects could lead to a significant reduction of the coagulation rate and clot forming ability by the small iron oxide–silica core shell particles, but not by the large iron oxide–silica core shell particles with a radius >30 nm. In our study, small dextran-coated iron oxide nanoparticles (5 nm in size) in a 10-fold higher mass concentration than the iron oxide–silica nanoparticles did not lead to a full inhibition of the assay. This might be due to the presence of 10% low molecular dextran within the formulation, in part covering the nanoparticle surface, or simply in altered surface characteristics due to the different material. In preliminary studies, high concentrations of citrate–stabilized iron oxide nanoparticles of comparable particle size also showed inhibition of the LAL gel clot assay. But the possibility remains that this effect is based on another mechanism, the anti-coagulation activity of citrate, which is known to affect LAL detection systems if applied in certain concentrations. Owing to the complexity of the LAL, as well as the multiple factors that influence protein–particle interaction, thus far only assumptions can be made about the mechanisms of the observed assay interference. Further information is needed, especially on the protein composition of the commercial LAL reagent. As the LAL reagent is a biological product, fluctuations in the composition of different batches might also lead to different results. Lot-to-lot variability of the results must be investigated in future studies.

Figure 3.

Calculated particle surface area as a function of primary particle size. A decrease in particle size is connected to an exponential increase in particle surface area. Squares correspond to the size of the tested particle samples.

Conclusions

The results, once again, show that every nanoparticle type and individual batch has to be carefully tested for inhibition and enhancement of a biological assay. Results obtained with one particle type cannot be necessarily applied to a second particle type made of the same material. Even minor changes of the physicochemical properties of the nanoparticle suspension can render the assay results invalid. The size, composition, zeta potential, surface curvature, accessible surface area, and aggregation behavior of the nanoparticles, as well as the protein–surface interaction seem to be crucial factors that influence the outcome of the LAL gel clot assay. As biological assays consist of complex media, it is difficult to draw conclusions from the results to the underlying mechanisms. For nanoparticle suspensions with strong inherent optical properties, for example light absorption and scattering, as well as strong interference with the LAL gel clot assay, other alternative methods without optical read out must be found to obtain reliable results concerning endotoxin contamination.

Footnotes

Acknowledgements

We thank Thomas Jung, Nanogate AG, Germany, for providing iron oxide nanoparticles. We also thank Aljosha Rakim Jochem, Sarah Schmidt, Anika Schwarz and Anika Weber for assistance in particle synthesis and physicochemical characterization; Marcus Koch for SEM/TEM investigation; and Henrike Peuschel, Jennifer Atchison and Julia Callender (all INM) for proofreading the manuscript. We thank Eduard Arzt for continuing support of our work at INM.

Funding

This work was funded by the Federal Ministry of Education and Research (BMBF), in frame of the project ‘NanoKon - Systematic evaluation of health effects of nanoscale contrast agents’ [grant number 03X0100C].

Conflict of interest

The authors do not have any conflicts of interest to declare.

References

Krug

Wick

. Nanotoxicology: an interdisciplinary challenge. Angew Chem Int Ed Engl 2011; 50: 1260–1278.

Dobrovolskaia

McNeil

. Immunological properties of engineered nanomaterials. Nat Nanotechnol 2007; 2: 469–478.

Dobrovolskaia

Aggarwal

Hall

McNeil

. Preclinical studies to understand nanoparticle interaction with the immune system and its potential effects on nanoparticle biodistribution. MolPharm 2008; 5: 487–495.

Zolnik

Gonzalez-Fernandez

Sadrieh

Dobrovolskaia

. Minireview: nanoparticles and the immune system. Endocrinology 2009; 151: 458–465.

Ragab

Van De Motter

Lavish

Goldberg

Ninomiya

Carlin

. Measurement and removal of adherent endotoxin from titanium particles and implant surfaces. J Orthop Res 1999; 17: 803–809.

Cho

Shanbhag

Hong

C-Y

Baran

Goldring

. The role of adsorbed endotoxin in particle-induced stimulation of cytokine release. J Orthop Res 2002; 20: 705–713.

Vallhov

Qin

Johansson

Ahlborg

Muhammed

Scheynius

. The Importance of an endotoxin-free environment during the production of nanoparticles used in medical applications. Nano Lett 2006; 6: 1682–1686.

Shi

Yadav

Wang

. Endotoxin promotes adverse effects of amorphous silica nanoparticles on lung epithelial cells in vitro. J Toxicol Environ Health A 2010; 73: 748–756.

Smulders

Kaiser

J-P

Zuin

Van Landuyt

Golanski

Vanoirbeek

. Contamination of nanoparticles by endotoxin: evaluation of different test methods. Part Fibre Toxicol 2012; 9: 41–41.

10.

Muta

Iwanaga

. The role of hemolymph coagulation in innate immunity. Curr Opin Immunol 1996; 8: 41–47.

11.

Solum

. The coagulation of Limulus polyphemus hemocytes – A comparison of the clotted and non-clotted forms of the molecule. Thromb Res 1973; 2: 55–70.

12.

Kawabata Si Muta

. Sadaaki Iwanaga: discovery of the lipopolysaccharide- and -1,3-D-glucan-mediated proteolytic cascade and unique proteins in invertebrate immunity. J Biochem 2010; 147: 611–618.

13.

Novitsky

. Limulus amebocyte lysate (LAL) detection of endotoxin in human blood. Innate Immun 1994; 1: 253–263.

14.

Wörle-Knirsch

Pulskamp

Krug

. Oops they did it again! Carbon nanotubes hoax scientists in viability assays. Nano Lett 2006; 6: 1261–1268.

15.

Monteiro-Riviere

Inman

Zhang

. Limitations and relative utility of screening assays to assess engineered nanoparticle toxicity in a human cell line. Toxicol Appl Pharmacol 2009; 234: 222–235.

16.

Oostingh

Casals

Italiani

Colognato

Stritzinger

Ponti

. Problems and challenges in the development and validation of human cell-based assays to determine nanoparticle-induced immunomodulatory effects. Part Fibre Toxicol 2011; 8: 8–8.

17.

Dobrovolskaia

Germolec

Weaver

. Evaluation of nanoparticle immunotoxicity. Nat Nanotechnol 2009; 4: 411–414.

18.

Dobrovolskaia

Neun

Clogston

Ding

Ljubimova

McNeil

. Ambiguities in applying traditional Limulus Amebocyte Lysate tests to quantify endotoxin in nanoparticle formulations. Nanomedicine 2010; 5: 555–562.

19.

Schulze

Kroll

Lehr

C-M

Schäfer

Becker

Schnekenburger

. Not ready to use – overcoming pitfalls when dispersing nanoparticles in physiological media. Nanotoxicology 2008; 2: 51–61.

20.

Dawson

. Interference with the LAL test and how to address it. LAL update, Associates of Cape Cod, Inc., Woods Hole, MA, USA. , http://www.acciusa.com/pdfs/newsletter/LALUpdateVol22_No3rev001.pdf (2005, accessed 17 September 2012).

21.

Fields M. Testing blood samples for endotoxin. LAL update, Associates of Cape Cod, Inc., Woods Hole, MA, USA, http://www.acciusa.com/pdfs/newsletter/LALUpdateVol23_No2.pdf (2006, accessed 17 September 2012).

22.

European Pharmacopoeia. European pharmacopoeia 5.0, 2.6.14. Bacterial endotoxins. France: EDQM Council of Europe, 2005. http://www.edqm.eu/en/european-pharmacopoeia-7th-edition-1401.html.

23.

Stoeber

Fink

Bohn

. Controlled growth of monodisperse silica spheres in the micron size range. J Coll Int Sci 1968; 26: 62–69.

24.

Hartlen

Athanasopoulos

APT

Kitaev

. Facile Preparation of highly monodisperse small silica spheres (15 to >200 nm) suitable for colloidal templating and formation of ordered arrays. Langmuir 2008; 24: 1714–1720.

25.

Kucki M. Detection and semi-quantification of endotoxin contaminations in nanoparticle suspensions - Limulus amebocyte lysate (LAL) gel clot assay. DaNa project, http://www.nanopartikel.info/files/content/dana/Dokumente/Projekte/SOPs/NanokonSOP222_geschützt.pdf (2012, accessed 3 August 2012).

26.

ISO 29701:2010. Nanotechnologies - Endotoxin test on nanomaterial samples for in vitro systems - Limulus amebocyte lysate (LAL) test.

27.

Dobrovolskaia MA and Neun B. NCL_Method_STE-1.3, detection and quantification of Gram negative bacterial endotoxin contamination in nanoparticle formulations by gel-clot LAL assay. Nanotechnology Characterization Laboratory, National Cancer Institute-Frederick, MD, USA, http://ncl.cancer.gov/NCL_Method_STE-1.3.pdf (2011, accessed 6 January 2012).

28.

Neun

Dobrovolskaia

Detection and quantitative evaluation of endotoxin contamination in nanoparticle formulations by LAL-Based assays. In: McNeil

(eds). Characterization of nanoparticles intended for drug delivery, methods in molecular biology, Series Volume 697, Springer Protocols. New York: Humana Press, 2011, pp. 121–130. DOI: 10.1007/978-1-60327-198-1_12.

29.

Darkow

Groth

Albrecht

Lützow

Paul

. Functionalized nanoparticles for endotoxin binding in aqueous solutions. Biomaterials 1999; 20: 1277–1283.

30.

Bromberg

Chang

Alvarez-Lorenzo

Magariños

Concheiro

Hatton

. Binding of functionalized paramagnetic nanoparticles to bacterial lipopolysaccharides and DNA. Langmuir 2010; 26: 8829–8835.

31.

Piazza

Colombo

Zanoni

Granucci

Tortora

Gioannini

. Uniform lipopolysaccharide (LPS)-loaded magnetic nanoparticles for the investigation of LPS-TLR4 signaling. Angew Chem 2011; 123: 648–652.

32.

Gardiner

Anderson

McCaigue

Erwin

Halliday

Rowlands

. Adsorbents as antiendotoxin agents in experimental colitis. Gut 1993; 34: 51–55.

33.

Xing

Pabst

Hasty

Smith

. Accumulation of LPS by polyethylene particles decreases bone attachment to implants. J Orthop Res 2006; 24: 959–966.

34.

Iwanaga

. The molecular basis of innate immunity in the horseshoe crab. Curr Opin Immunol 2002; 14: 87–95.

35.

Ariki

. A serine protease zymogen functions as a pattern-recognition receptor for lipopolysaccharides. Proc Natl Acad Sci USA 2004; 101: 953–958.

36.

Ariki

Takahara

Shibata

Fukuoka

Ozaki

Endo

. Factor C acts as a lipopolysaccharide-responsive C3 convertase in horseshoe crab complement activation. J Immunol 2008; 181: 7994–8001.

37.

Iwanaga

Miyata

Tokunaga

Muta

. Molecular mechanism of hemolymph clotting system in Limulus. Thromb Res 1992; 68: 1–32.

38.

Tenzer

Docter

Rosfa

Wlodarski

Kuharev

Rekik

. Nanoparticle size is a critical physicochemical determinant of the human blood plasma corona:a comprehensive quantitative proteomic analysis. ACS Nano 2011; 5: 7155–7167.

39.

Lozner

Taylor

FHL

. The effect of foreign surfaces on blood coagulation. J Clin Invest 1942; 21: 241–245.

40.

Ratnoff

Rosenblum

. Role of Hageman factor in the initiation of clotting by glass. Am J Med 1958; 25: 160–168.

41.

Morrison

Cochrane

. Direct evidence for Hageman factor (factor XII) activation by bacterial lipopolysaccharides (endotoxins). J Exp Med 1974; 140: 797–811.

42.

Vogler EA and Graper JC. Partikel zur Verbesserung der Gerinnung in einem Blutentnahmeröhrchen durch Aktivieren zweier Reaktionswege, Blutentnahmeanordnung und Verfahren zur Herstellung der Partikel. DE patent EP 0 628 816 B1 DE 694 14 915 T2, Germany, 1994.

43.

Tavano r, Segat

Reddi

Kos

Rojnik

Kocbek

. Procoagulant properties of bare and highly PEGylated vinyl-modified silica nanoparticles. Nanomedicine 2010; 5: 881–896.

44.

Maitz

Teichmann

Sperling

Werner

. Surface endotoxin contamination and hemocompatibility evaluation of materials. J Biomed Mater B Appl Biomater 2008; 90B: 18–25.

45.

Patwardhan

Emami

Berry

Jones

Naik

Deschaume

. Chemistry of aqueous silica nanoparticle surfaces and the mechanism of selective peptide adsorption. J Am Chem Soc 2012; 134: 6244–6256.

46.

Hansson

Tosatti

Isaksson

Wetterö

Textor

Lindahl

. Whole blood coagulation on protein adsorption-resistant PEG and peptide functionalised PEG-coated titanium surfaces. Biomaterials 2005; 26: 861–872.

47.

Cedervall

Lynch

Lindman

Berggard

Thulin

Nilsson

. From the Cover: Understanding the nanoparticle-protein corona using methods to quantify exchange rates and affinities of proteins for nanoparticles. Proc Natl Acad Sci USA 2007; 104: 2050–2055.

48.

Solum

Mürer

. Early studies on the coagulogen (clotTable protein) of Limulus polyphemus (horseshoe crab). Thromb Haemost 2007; 98: 129–130.

49.

Nakamura

Takagi

Iwanaga

Niwa

Takahashi

. Amino acid sequence studies on the fragments produced from horseshoe crab coagulogen during gel formation: homologies with primate fibrinopeptide B. Biochem Biophys Res Commun 1976; 72: 902–908.

50.

Deng

Mortimer

Schiller

Musumeci

Martin

Minchin

. Differential plasma protein binding to metal oxide nanoparticles. Nanotechnology 2009; 20: 455101–455101.

51.

Kendall

Ding

Kendall

. Particle and nanoparticle interactions with fibrinogen: the importance of aggregation in nanotoxicology. Nanotoxicology 2011; 5: 55–65.

52.

Roach

Farrar

Perry

. Surface tailoring for controlled protein adsorption: effect of topography at the nanometer scale and chemistry. J Am Chem Soc 2006; 128: 3939–3945.

53.

Vroman

Adams

Fischer

Munoz

. Interaction of high molecular weight kininogen, factor xii, and fibrinogen in plasma at interfaces. Blood 1980; 55: 156–159.

54.

Vroman

. The life of an artificial device in contact with blood: initial events and their effect on its final state. Bull NY Acad Med 1988; 64: 352–357.

Interference of silica nanoparticles with the traditional Limulus amebocyte lysate gel clot assay

Abstract

Keywords

Introduction

Materials and methods

Particle synthesis

Particle characterization

Endotoxin tests

Results

Silica particles: bare and PEGylated

Iron oxide–silica core shell particles (FexOy@SiO2)

Endorem (FexOy)

Summary

Reliability of the test results

Discussion

Conclusions

Footnotes

Acknowledgements

Funding

Conflict of interest

References

Iron oxide–silica core shell particles (Fe_xO_y@SiO₂)

Endorem (Fe_xO_y)