ilable at ScienceDirect

Toxicon 76 (2013) 187–196

Contents lists ava

Toxicon

journal homepage: www.elsevier .com/locate/ toxicon

The isolation and characterization of lipopolysaccharidesfrom Microcystis aeruginosa, a prominent toxic water bloomforming cyanobacteria

Lucie Bláhová a, Ond�rej Adamovský a, Luká�s Kubala b, Lenka �Svihálková�Sindlerová b, Radka Zounková a, Lud�ek Bláha a,*

aMasaryk University, Faculty of Science, Research Centre for Toxic Compounds in the Environment (RECETOX), Kamenice 753/5,Building A29, CZ62500 Brno, Czech Republicb Institute of Biophysics, v.v.i., Academy of Sciences of the Czech Republic, Department of Free Radical Pathophysiology,Kralovopolska 135, CZ62165 Brno, Czech Republic

a r t i c l e i n f o

Article history:Received 11 July 2013Received in revised form 30 September 2013Accepted 8 October 2013Available online 16 October 2013

Keywords:EndotoxinCyanobacteriaWater bloomLipopolysaccharideMicrocystis

* Corresponding author. Tel.: þ420 549493194, þE-mail address: blaha@recetox.muni.cz (L. BláhaURL: http://www.recetox.muni.cz

0041-0101/$ – see front matter � 2013 Elsevier Ltdhttp://dx.doi.org/10.1016/j.toxicon.2013.10.011

a b s t r a c t

Massive toxic blooms of cyanobacteria represent a major threat to water supplies world-wide, yet serious gaps exist in understanding their complex toxic effects, including the roleof lipopolysaccharides (LPS). The present comparative study focused on the levels andbiological activities of LPS isolated from Microcystis aeruginosa, which is one of the mostglobally distributed toxic species. Using hot phenol extraction, LPS was isolated from 3laboratory cultures and 11 natural water blooms. It formed 0.2–0.7% of the original drybiomass of the cyanobacteria, based on gravimetry. Additional analyses by commercialanti-LPS ELISA were correlated with gravimetry but showed concentrations that wereabout 7-times lower, which indicated either impurities in isolated LPS or the poor cross-reactivity of the antibodies used. LPS isolates from M. aeruginosa were potent pyrogensin the traditional Limulus amebocyte lysate (LAL)-test, but comparison with the PyroGenetest demonstrated the limited selectivity of LAL with several interferences. The determinedpyrogenicity (endotoxin units, EU) ranged from very low values in laboratory cultures (lessthan 0.003 up to 0.008–EU per 100 pg LPS) to higher values in complex bloom samples(0.01–0.078 EU per 100 pg of LPS), which suggested the role of bloom-associated bacteriain the overall effects. Potent pro-inflammatory effects of the studied LPS from both culturesand bloom samples were observed in a highly-relevant ex vivo human blood model bystudying reactive oxygen species production in phagocytes as well as increased pro-ductions of interleukin 8, IL-8, and tumor necrosis factor a, TNF-a. LPS from M. aeruginosaseem to modulate several pathways involved in the regulation of both innate immunityand specific responses. In comparison to the standard pathogenic bacterial LPS (WorldHealth Organization Escherichia coli O113:10 endotoxin; activity 1 EU per 100 pg), thestudied cyanobacterial samples had pyrogenicity potencies that were at least 12-timeslower. However, the health risks associated with LPS from M. aeruginosa should not beunderestimated, especially with respect to diverse biological effects observed ex vivo andin the case of massive blooms in drinking water reservoirs, where the estimated pyroge-nicity can reach up to 46,000 EU per mL of water.

� 2013 Elsevier Ltd. All rights reserved.

420 605510953 (mobile).).

. All rights reserved.

L. Bláhová et al. / Toxicon 76 (2013) 187–196188

1. Introduction

Massive cyanobacterial water blooms cause seriousenvironmental problems worldwide, and the production ofcyanobacterial toxins (cyanotoxins) is regarded as posingmajor risks to health (Codd et al., 2005; Funari and Testai,2008). While several known cyanobacterial toxins such asmicrocystins or cylindrospermopsin have been investi-gated in detail (Campos and Vasconcelos, 2010; Zeguraet al., 2011; Funari and Testai, 2008), there is a lack of in-formation on other potentially toxic components of cya-nobacteria such as lipopolysaccharides. These can reachvery high concentrations in water reservoirs during thelarge-scale growth of cyanobacterial blooms, and exposureto humans via drinking water or recreational activities maylead to adverse health outcomes (Anderson et al., 2002;Rapala et al., 2002).

Several different cyanobacterial species contribute tothe formation of dense cyanobacterial blooms in differentparts of the world. These include Cylindrospermopsis raci-borskii (mostly in tropical regions (Sinha et al., 2012)) andPlanktothrix sp. in different parts of the world, includingScandinavia (Bonilla et al., 2012). However, the mostglobally-distributed cyanobacteria is probably Microcystisaeruginosa, which is responsible for major toxic bloomproblems in Europe (Via-Ordorika et al., 2004), Asia (Zhanget al., 2012), North America (Wilson et al., 2005), but also inother regions.

Lipopolysaccharides are major components of the cellwall of all Gram-negative bacteria, including cyanobacteria(Buttke and Ingram, 1975). In pathogenic bacteria, LPS isknown to mediate immune system responses and to causefever, inflammation, shock or even death (Saluk-Jusczakand Wachowicz, 2005), and endotoxins are recognized asone of the important water quality parameters (Can et al.,2013). While the structure and biological significance ofLPS derived from heterotrophic bacteria has been studiedin the past, much less information exists for cyanobacteria(Stewart et al., 2006). LPS originating from cyanobacteriawas shown to be less toxic in comparison to prototypicaland pathogenic bacteria. For example, the endotoxin ac-tivity of LPS from a laboratory culture of M. aeruginosa PCC7806 was found to have more than fifty times lower py-rogenic activity (fever-causing affects) than Escherichia coliLPS (Bernardová et al., 2008). On the other hand, thetoxicity of cyanobacterial LPS is generally comparable tothat of other bacteria occurring in the aquatic environmentsuch as Pseudomonas fluorescens or Kluyvera intermedia(Bernardová et al., 2008; Rapala et al., 2002). Some labo-ratory studies indicated the pyrogenicity and biologicalactivities of cyanobacterial LPS in vivo (Weckesser et al.,1979; Dogo et al., 2011; Palikova et al., 2013). In addition,other documented biological effects of cyanobacterial LPSinclude the potentiation of heavy metal toxicity (Notchet al., 2011) and modulations of microcystin biotransfor-mation and toxicity (Best et al., 2001, 2002; Lindsay et al.,2006; Jaja-Chimedza et al., 2012).

The most widely used method for endotoxin detectionis the Limulus amebocyte lysate test (LAL test), which isbased on the precipitation of the haemolymph of Limuluspolyphemus after exposure to endotoxin (Lindsay et al.,

1989). Consequently, turbidity or color reaction can bedetermined. While the LAL test is known to be sensitiveand fast, its disadvantage can be its lower specificity. Analternative advanced method for the quantification ofendotoxin is based on isolated recombinant Factor C –

rFC, i.e. the PyroGene test, which uses only the first stepin the precipitating cascade of the LAL test (Ding and Ho,2001; Ding et al., 2001). In comparison to the wholehaemolymph test, the activation of rFC by the binding ofendotoxin is more specific avoiding the overestimation ofresults.

The naturally high variability and complexity of thechemical composition of LPS from various types of bacteriabring significant challenges with respect to quantificationof the endotoxin activity of various LPS by biochemicaltests. Among the classical and still widely usedmethods aredirect determinations of pyrogenic activity in rabbits orother laboratory animals in vivo (Hoffmann et al., 2005).Recently, alternative in vitro tests have been introducedemploying permanent cell lines (Hoffmann et al., 2005;Poole et al., 2003), isolated primary cells, or whole blood(Gaines Das et al., 2004; Hoffmann et al., 2005). In thesesystems, the in vitro activation of cells by the tested samplesis evaluated by determination of the release of various cy-tokines (TNFa, IL-1, IL-6 etc.) or the production of nitricoxide or reactive oxygen species (Mayer et al., 2011;Hoffmann et al., 2005; Matiasovic et al., 2011). Generally,in vitro biological tests overcome the limitations of the LALand PyroGene tests (which assess a single activationpathway) and better reflect the reaction of the biologicalsystem to pyrogenic agent. However, at the same time, thisis also a limitation, since the positive response of cellsin vitro can be induced also by a number of other com-pounds such as cell wall components muramyl dipeptideand peptidoglycan, DNA, and poly (I:C) (Ding and Ho, 2001;McKenzie et al., 2011). Therefore, in most cases, biologicaltests are not specific to a particular LPS structure. All themethods (toxicological assays and the LAL and PyroGenetests) provide important information on the pyrogenic ac-tivity of the studied sample, but they cannot provide validinformation on the “concentration” (or mass) of theendotoxin LPS.

Despite the complexity of LPS, some bioanalyticalmethods for the determination of LPS concentrations havebeen designed, such as Enzyme-Linked ImmunosorbentAssay (ELISA). This method uses immuno-enzymatic re-actions and structural recognition of the antigen, i.e. LPS. Incomparison to the effect-based methods of endotoxin an-alyses, ELISA should provide information on the concen-tration of LPS molecules. In addition, more selectivechemical methods based on gas chromatography–massspectrometry (GC–MS) have only recently been developedfor the quantification of cyanobacterial LPS. These methodstrack some important markers of bacterial endotoxin, suchas 3-deoxy-D-manno-oct-2-ulosonic acid – (“KDO”) and 3-hydroxy fatty acids (de Santana-Filho et al., 2012; Casa-buono et al., 2012; McKenzie et al., 2011). However, thecompositions of oligosaccharides and A-lipids of cyano-bacterial LPS were shown to differ from bacteria (Fujii et al.,2012; Snyder et al., 2009), and therefore selective markersof cyanobacterial LPS are not available.

L. Bláhová et al. / Toxicon 76 (2013) 187–196 189

The present comparative study investigated the levels(concentrations) and biological activities of LPS isolatedfrom the widespread cyanobacteria M. aeruginosa. Thestudy focused on LPS from both laboratory cyanobacterialcultures and natural water blooms dominated by M. aeru-ginosa and assessed various traditional as well as less-explored endpoints such as (i) endotoxin activity in tradi-tional LAL tests, (ii) the alternative determination of pyro-genicity using the PyroGene test, and (iii) the potency toactivate leukocytes in human whole blood determined asthe production of reactive oxygen species (ROS), inter-leukin 8 (IL-8), and tumor necrosis factor a (TNF-a). Inaddition, concentrations of LPS isolated from cyanobacteriawere determined by commercial ELISA. The outcomes andperformance of all the biochemical and biological methodsused are compared, and the health implications of theobserved results are discussed.

2. Materials and methods

2.1. Sample material

Three laboratory strains of M. aeruginosa were used inthe present study, including M. aeruginosa PCC 7806 pur-chased from the Pasteur Culture Collection of Cyanobac-teria (this strain produces the well-known peptide toxinsmicrocystins), M. aeruginosa A017 obtained from the Cya-nobacterial Culture Collection of the Czech Academy ofScience (microcystin non-producer), and M. aeruginosaUTEX 2667 from the Culture Collection of Algae at theUniversity of Texas in Austin (microcystin producer). All themicroorganisms were grown in a mixture of Zehnder me-dium (Schlosser, 1994), Bristol (modified Bold) medium

Table 1Characterization of the studied environmental samples: water blooms dominate

Samplecode

Locality Samplingmonth

Surfacearea/km2

Compositio

TB01 Slapy October 2008 12.7 MicrocystiWoronichinþ Aphanizo

TB02 Pí�s�tany August 2008 0.9 MicrocystiTB03 Hracholusky July 2008 3.6 Microcysti

MicrocystisMicrocystisAphanizom

TB04 Rathan August 2008 0.1 MicrocystiMicrocystis

TB05 B�rezová September 2008 0.3 MicrocystiTB06 Újezd July 2008 1.1 Microcysti

AphanizomTB07 Nové Mlýny II August 2008 9.2 MicrocystiTB08 Pí�s�tany September 2008 0.9 Microcysti

MicrocystisTB09 Vajgar July 2008 0.5 Microcysti

MicrocystisAnab. spiroaquae, A. ccircinalis 4

TB10 �Svihov August 2008 14.0 MicrocystiWoronichin

TB11 Plumlov August 2008 0.6 MicrocystiMicrocystis

(Stein, 1973), and distilled water in the ratio 1:1:2 (v/v/v)under continuous illumination (cool white fluorescenttubes, 3000 lux) at 22 �C � 2 �C and aerated with sterilizedair passed through a 0.22 mm filter (Bártová et al., 2011). Inaddition to laboratory cultures, water blooms dominatedby M. aeruginosa were collected from Czech Republic res-ervoirs using a 25 mm plankton net (Table 1). Identificationof the studied localities is shown in Fig. 1. Part of thebiomass was used for identification of phytoplankton spe-cies according to Komárek and Anagnostidis (2005).Dominance of cyanobacteria and M. aeruginosa (%) wascalculated based on the counts of phytoplankton cells in thesample. The harvested biomass from laboratory culturesand natural blooms was frozen at �18 �C and then freeze-dried (Heto Power Dry LL3000) prior to the extraction ofLPS.

2.2. Extraction of LPS

Lipopolysaccharides were extracted using hot phenol–water extraction (Papageorgiou et al., 2004; Bernardováet al., 2008). A suspension of the lyophilized biomass(200 mg) in distilled water (50 ml) was mixed with 90%phenol (50 ml) and stirred at 68 �C for 20 min. After coolingto 4 �C, the mix was centrifuged (3026 g/30 min), and thesupernatant (aqueous phase with LPS) was separated. Thephenol layer was re-extracted once more with distilledwater (50 ml). Pooled supernatants (aqueous phase) werepurified by dialysis using cellulose membranes(33 � 21 mm, Sigma–Aldrich) against distilled water(1000 ml) for 48 h and then centrifuged and lyophilized.The semi-purified freeze-dried extract of LPS was resus-pended in 750 ml of 0.1 M Tris–HCl buffer (pH 7.4)

d by Microcystis aeruginosa.

n of cyanobacteria % Cyanobacteriain biomass

Dominant microcystins inbiomass (MC-LR, RR, YR)mg/g d.w.

s aeruginosa 90%,ia naegelianamenon sp. 5%

95 152

s aeruginosa 80% 95 1561s aeruginosa 80%,viridis 5%,flos-aquae 5%,enon sp. 10%

100 1506

s aeruginosa 85%,ichtyoblabe 10%

95 1337

s aeruginosa 64% 88 102s aeruginosa 75%,enon sp. 25%

100 1670

s aeruginosa 85% 85 1127s aeruginosa 45%,viridis 25 %

70 2147

s aeruginosa 49%,ichtyoblabe 5%,ides þ A. flos-rassa þ A.6%

100 306

s aeruginosa 45%,ia 10%

55 455

s aeruginosa 80%,ichtyoblabe 20%

100 780

Fig. 1. Map of the major reservoirs in the Czech Republic affected by the repeated occurrence of massive blooms dominated byMicrocystis aeruginosa. The elevenstudied localities are marked with full symbols; the size of the symbols corresponds to surface area.

L. Bláhová et al. / Toxicon 76 (2013) 187–196190

containing 25 mg ml�1 RibonucleaseA (Sigma–Aldrich). Thesolution was incubated for 16 h at 37 �C and then 750 ml of90% phenol in water was added. After 4 min incubation atroom temperature, the solution was centrifuged (18 407 g/15 min) and the aqueous phase separated and purified for48 h by dialysis (see above) and lyophilized. The freeze-dried powder constituting purified LPS was weighed forassessment of the content of LPS in the biomass and kept at�18 �C.

2.3. Pyrogenicity by LAL test

The chromogenic kinetic Limulus amebocyte lysate(LAL) test was performed according to the manufacturer’sinstructions (Associates of Cape Code, Inc., USA). Duplicatesof diluted samples of LPS in LAL-reagent water, negativecontrols and standards were preincubated at 37 �C in a 96-well microplate for 15 min. A chromogenic substrate wasthen added and the development of absorbance wasmeasured for 50 min in kinetic mode at 405 nm (BioTek-Power Wave spectrophotometer). The time at which theabsorbance exceeded the normalized value of 0.03 absor-bance units (threshold time) was recorded. Pyrogenicitywas calculated from the calibration of the standard endo-toxin from E. coli 0111:B4 (0.031–1.0 of Endotoxin Units/ml;EU/ml) tested against the reference standard endotoxin(RSE; prepared from E. coli 0113:H10:K). Activity wasexpressed as EU per mg of isolated LPS (EU/mg LPS) and EUper mg of the original biomass dry weight (EU/mg d.w.).

2.4. Pyrogenicity by the PyroGene rFC endotoxin system

In addition to the standard LAL test, the alternativePyroGene rFC endotoxin detection system (Cambrex BioScience, USA, 50-658U) was used to detect the endotoxicityof cyanobacterial LPS. The method uses a one-stepapproach incubating 100 ml of blank (or standard or

diluted sample) with 100 ml of the reaction solution (rFCenzyme solution, rFC assay buffer, and fluorogenic sub-strate in the ratio of 1:4:5). The incubation was performedin a 96-well microplate for 1 h at 37 �C according to themanufacturer’s instructions (PyroGene rFC EndotoxinDetection System, Cambrex Bio Science, USA, 50-658U).Fluorescence was measured (excitation 390 nm, emission440 nm) using a microplate reader (BMG-Labtech POLAR-star OPTIMA), and endotoxin concentrations calculatedaccording to the standard curve.

2.5. Effects on human whole blood

Heparinized (50 IU/mL) blood samples were obtainedfrom the cubital vein of healthy volunteers, who gave theirinformed consent. The luminol-enhanced chem-iluminescence (CL) of whole blood samples incubated withthe tested compounds was measured using an LM-01Tmicroplate luminometer (Immunotech, Czech Republic)according to previously optimized experiments (Kubalaet al., 2003; Matiasovic et al., 2011). The reaction mixture(total volume of 200 ml HBSS – Hank’s Balanced Salt Solu-tion) contained 10 ml of whole blood plus the tested LPSsamples (1 mg/ml), blanks or positive control (commercialLPS – 0.001 mg/ml; E. coli serotype 026:B6, Sigma–Aldrich).After incubation (2 h, 37 �C), 1 mM luminol (stock solutionof 10 mM luminol in 0.2 M borate buffer; Invitrogen, USA)was added and the CL emission expressed as relative lightunits (RLU) was recorded continuously for 90 min at 37 �C(duplicate analyses). The integral value of the CL reaction,which represents the total ROS production by bloodphagocytes, was evaluated, and final datawere recalculatedas a percentage of the untreated control (100%). For thedetermination of cytokine release, the whole blood wasincubated with the tested LPS samples (1 mg/ml) or controlLPS from E. coli (0.001 mg/ml) for 5 h (analysis of TNF-a) or24 h (analysis IL-8) at 37 �C. Then, the samples were

L. Bláhová et al. / Toxicon 76 (2013) 187–196 191

centrifuged at room temperature (800 g/5 min), and thesupernatant collected and frozen (�80 �C). Concentrationsof cytokines in the collected supernatant were determinedby commercial kits according to the manufacturers’ in-structions (IL-8 kit from Bender MedSystems, Vienna,Austria; TNF-a kit from DuoSet� R&D Systems, Minneap-olis, USA).

2.6. LPS content – the ELISA test

For additional quantification of LPS content in the iso-lated samples from cyanobacteria, a commercially availableELISA kit was used (LPS ELISA Kit, Novatein Biosciences,USA) according to the manufacturer’s instructions. Theassay is based on double-antibody sandwich ELISA. Theprimary antibody was coated on the surface of the micro-plate wells, which had been incubated with LPS from thestudied samples (or blanks or LPS standards). Afterwashing, which removed unbound substances, secondaryHRP (horseradish peroxidase)-labeled antibody was added.The third step included the addition of the chromogenicsubstrate. The reaction was stopped and the absorbance ofthe resulting yellow color determined at 450 nm using aBioTek-Power Wave spectrophotometer. The concentrationof LPS in the samples was calculated on the basis of thecalibration curves of LPS using the GraphPad Prism soft-ware (four-parameter logistic variable slope curve).

2.7. Microcystin analyses in LPS extracts

To check for the presence of known cyanotoxinsmicrocystins (MC), Liquid Chromatography ElectrosprayIonization Mass Spectrometry Analyses were performedusing an HPLC Agilent 1290 series instrument (AgilentTechnologies, Waldbronn, Germany) with a Phenomenexcolumn (LUNA C-18 endcapped (3 mm) 100 � 2 mm i.d.)

Table 2Content of LPS in biomass d.w. (according to weight of LPS extract or analysis by ctests.

LPS weighted (mg LPS/mg biomass d. w.)

LPS ELISA (mg LPS/mg biomass d.w.)

LALmg

Laboratory culturesLC01 (M. aerug.,

PCC 7806)3.8 0.5 4.

LC02 (M. aerug.,UTEX 2667)

1.9 0.3 8.

LC03 (M. aerug.mut., A017)

2.5 0.4 <3.

Toxic bloomsTB01 5.9 0.8 67.TB02 6.4 1.0 78.TB03 2.5 0.3 13.TB04 4.9 1.0 18.TB05 2.5 0.4 67.TB06 3.5 0.5 47.TB07 4.9 0.7 48.TB08 6.8 1.2 34.TB09 6.8 0.8 13.TB10 2.5 0.3 67.TB11 2.0 0.3 10.

a Concentration of LPS obtained by weighing used for calculations.

coupled with a guard column (SecureGuard C18; Phe-nomenex, Torrance, CA, USA). The mobile phase consistedof (A) 5 mM ammonium acetate in water, pH 4, and (B) amethanol–acetonitrile mixture (1:1) with 5 mM ammo-nium acetate. The binary pump gradient was non-linear (anincrease from 25% B at 0 min to 80% B at 2 min, then anincrease to 95% B at 10 min, then 95% B for following 4 minand 4 min column equilibration back to the initial condi-tions of 25% B); the flow rate was 0.25 ml/min. 5 mL of thesample were injected for the analyses. The mass spec-trometer was an AB Sciex Qtrap 5500 (AB Sciex, Concord,ON, Canada) with electrospray ionization (ESI). Ions weredetected in the positive mode. The ionization parameterswere as follows: capillary voltage, 5.5 kV; desolvationtemperature, 350 �C; Curtain gas 15 psi, Gas1 40 psi, Gas230 psi. The following m/z transitions of different MC vari-ants were monitored using the scheduled MRM mode(with corresponding values of declustering potential – DP(V), entrance potential – EP (V) and collision energy – CE(V)): MC-RR – m/z tran519.8 > 135.1 (DP 156, EP 10, CE 37)and 102.9 (DP 156, EP 10, CE 91), MC-YR – 1045.5 > 102.9(DP 241, EP 10, CE 129) and 212.9 (DP 241, EP 10, CE 69), MC-LR – 995.5>102.9 (DP 171, EP 10, CE 129) and 105.1 (DP 171,EP 10, CE 127). Quantification was based on external stan-dards and the detection limit of the method was 0.25 ng/mL for MC-RR, 1.0 ng/mL for MC-YR, and 1.0 ng/mL for MC–LR.

3. Results

LPS was isolated from eleven natural cyanobacterialwater bloom samples that were dominated by M. aerugi-nosa (45–90% dominance) and three differentM. aeruginosalaboratory strains (Table 2). Based onweighings of the finalisolates (gravimetry), the content of LPS was found to bebetween 2 and 7 mg LPS/mg biomass d.w., which

ommercial ELISA), and endotoxin activities assessed by LAL and PyroGene

(�104 EU/LPSa)

PyroGene (�104 EU/mg LPSa)

LAL (EU/mg d.w.)

PyroGene(EU/mg d.w.)

6 0.22 176 8

8 0.95 169 18

0 <0.1 <150 <4

9 8.97 4017 5303 14.28 5039 9196 1.53 336 382 2.13 901 1059 4.29 1673 1067 1.90 1644 654 18.96 2349 9205 1.01 2353 698 0.93 941 646 9.70 1674 2402 0.91 202 18

A

B

1 2 3 4 5 6 7

LPS weighted (ug/mg dw)

0.2

0.4

0.6

0.8

1.0

1.2

1.4

ELIS

A LP

S (u

g/m

g dw

)

Pearson's R = 0.9 2, p<0.0001, y = -0.019 + 0.15*x ; r2 = 0.85

-1000 0 1000 2000 3000 4000 5000 6000

LAL (EU / mg dw)

-200

0

200

400

600

800

1000

Pyro

Gen

e (E

U /

mg

dw)

Pearson's R = 0.80, p=0.001, y = -53.2 + 0.18*x ; r2 = 0. 64

Fig. 2. Correlations between concentrations of LPS (mg/mg d.w.) determinedby ELISA and gravimetry (A), and pyrogenic activities determined by LAL andPyroGene tests (B). Graphs also show linear regression, its 95% confidenceintervals and coefficient of determinance (r2).

Fig. 3. Effects of Microcystis aeruginosa LPS isolates (1 mg/ml) in humanwhole blood - induction of chemiluminescence caused by the release ofreactive oxygen species by phagocytes (A), induction of IL-8 (B), and in-duction of TNF-a (C). The results are means of five (chemiluminescence) orthree (TNF-a and IL-8) independent replicates � S.D.; positive control (graycolumn) was LPS from E. coli serotype 026:B6 (0.001 mg/ml). *NA – samplenot analyzed.

L. Bláhová et al. / Toxicon 76 (2013) 187–196192

corresponded to approximately 0.2–0.4% biomass d.w. forcyanobacterial cultures and 0.2–0.7% for complex blooms.In addition to the simple weighing of the isolates, thecontent of LPS in the extracts was determined by a com-plementary ELISA method (Table 2), which showed lowervalues ranging from 0.3 to 1.2 mg LPS/mg d.w. (approxi-mately 15% of the values determined by gravimetry).Nevertheless, the values were highly correlated (Fig. 2A,Pearson’s R ¼ 0.91; P < 0.001).

Endotoxin activities were determined by the traditionalLAL test and PyroGene test and expressed as EndotoxinUnits (EU) related to both LPS (determined by gravimetry)and biomass (Table 2). The LAL activity test measurementsof both cultures and natural water blooms ranged withinapproximately one order of magnitude – less than 150 EU/mg d.w. for sample LC03 (M. aeruginosa mut. A017 strain),up to 5040 EU/mg d.w. for TB02 sample (the water bloomfrom Pistany locality). The alternative PyroGene testshowed generally lower activities ranging within two or-ders of magnitude – less than 4 EU/mg d.w. for sample LC03up to more than 900 EU/mg d.w. for water bloom samples

TB07 and TB02 (localities Nove Mlyny II and Pistany). Theresults of both LAL and PyroGene assays were significantlycorrelated (Fig. 2B, Pearson’s R ¼ 0.81, p ¼ 0.0005). Values(from both the PyroGene and LAL tests) found for complexbloom samples were higher than those found for laboratorycultures (non-parametric Mann Whitney U test p ¼ 0.013and p ¼ 0.016 for PyroGene and LAL, respectively). Thisdifference should, however, be treated with caution whenconsidering the low number (N ¼ 3) of laboratory strains.

The evaluation of leukocyte activation in human wholeblood clearly revealed the significant potential of all sam-ples to activate leukocytes; in addition, increases in theproduction of ROS, IL-8 and TNF-a were observed.

L. Bláhová et al. / Toxicon 76 (2013) 187–196 193

However, the effects of two samples (TB10 – Svihov andTB11 – Plumlov) were lower compared to other samples(Fig. 3). In contrast, endotoxin samples isolated from lab-oratory cultures LC01 and LC02 and water bloom sampleTB07 (Nove Mlyny II) revealed the highest potential toactivate leukocytes. The comparison of endotoxin samplesisolated from M. aeruginosa (1 mg/ml tested) with positivecontrol (i.e. highly purified commercial LPS from E. colitested at 0.001 mg/ml) revealed similar levels of stimulation(Fig. 3A–C). All three parameters of blood leukocyte acti-vation were significantly correlated (Pearson’s R,p < 0.005), but no clear correlation was found between theparameters of leukocyte activation and endotoxin activitiesdetermined by LAL and PyroGene (Pearson’s R, P > 0.05).

M. aeruginosa is a well-known producer of microcystin,and the original content of microcystins in the studiedbiomasses varied from 100 to more than 2000 mg/g d.w.(Table 1). On the other hand, none of the studied LPS ex-tracts contained any of the investigated microcystin vari-ants (MC-LR, YR or MC-RR) above their limits of detection(1 ng MC-LR, -YR or 0.25 ng MC-RR per 1 mg of LPS extract).

4. Discussion

Despite the fact that previous studies have investigatedcyanobacterial LPS (Mohamed and Al Shehri, 2007;Bernardová et al., 2008; Wiegand and Pflugmacher,2005), broader characterization of the bioactivities andrisks associated with LPS of the globally-distributed bloom-forming species M. aeruginosa is missing.

The amount of LPS inM. aeruginosa isolates (0.2–0.7% ofthe total cyanobacterial biomass in different samples) cor-responded well to the previously reported content in abroader range of cyanobacterial and bacterial strains andsamples (Bernardová et al., 2008; Jaja-Chimedza et al.,2012). It is technically difficult to confirm the purity of LPSisolates obtained with the hot phenol–water extractionmethod (Papageorgiou et al., 2004), because no quantita-tive chemical method for analyzing the whole of cyano-bacterial LPS is currently available. Existing methods suchas GS-MS (Fujii et al., 2012; Snyder et al., 2009; de Santana-Filho et al., 2012) usually address certain parts of LPS fromselected microorganisms. Using ELISA as the analyticalmethod of reference, the purity of the isolated LPS was nothigh, because ELISA-derived concentrations correspondedto about 15% of the gravimetry, while the remaining 85%could have been various impurities that did not react withcommercial antibodies. Nevertheless, these impurities(most likely fragments or degradation products of LPScreated during the isolation process) could contribute tothe biological activities discussed below. However, otherexplanations of the difference between ELISA and gravim-etry should be considered. Most importantly, the manu-facturer of the commercial ELISA mentions only the use ofE. coli LPS during ELISA development but does not provideany detailed information about the actual epitope, i.e. thetarget domain within the LPS molecule. There was noresponse from the manufacturer although specificallyasked by the authors of the present study. It is thereforedifficult to discuss or hypothesize on the cross-reactivity ofthe ELISA with LPS from different microorganisms,

although some structures of LPS such as the core domainare known to be conserved among species (Martin et al.,1989). Because similarities between cyanobacterial LPSand prototypical Gram-negative bacteria such as E. coli areunknown, the lower cross-reactivity of ELISA antibodiescould significantly contribute to the observed differencesbetween gravimetry and ELISA. This should be carefullyconsidered if ELISA is planned for screening purposes,because it may actually provide false negative results orunderestimate real LPS concentrations.

The present study also addressed another general aspectof LPS analysis by comparing the widely used LAL test withthe PyroGene test. The PyroGene test is expected to bemore selective, as it employs recombinant protein Cmodulating only the first steps of the complex clottingcascade, which is actually monitored in the whole hae-molymph LAL test (McKenzie et al., 2011). The presentstudy of M. aeruginosa LPS showed significant pyrogenicityin both the LAL test and PyroGene test, and the effects werewell correlated. However, the values detected with thePyroGenemethodwere about 13-times lower on average incomparison to the LAL test (the difference ranged from 3 to34 times). This seems to correspond well to the insufficientsensitivity of, and number of interferences in the LAL test,as discussed previously by e.g. Anderson et al., 2002 andDungan, 2011. Various constituents of bacterial cell wallssuch as peptidoglycans, polysaccharides and glucans havebeen shown to cause false positive results in LAL tests (Dingand Ho, 2001; McKenzie et al., 2011). The observed differ-ence between the LAL and PyroGene methods can thus alsoreflect possible impurities, which could affect complexcoagulation in the LAL test but not in the more selectivePyroGene test. From this perspective, use of the PyroGenemethod for testing complex materials such as water orisolates (as shown in this study) has the advantage ofproviding a more selective response to LPS.

The actual pyrogenicity of studied M. aeruginosa LPS(endotoxin units – EU) was highly variable among differentsamples (ranging over an order of magnitude), with lowervalues observed in laboratory cultures in comparison to thebloom samples. A similar difference was reported also in aprevious study that compared LPS from various cyano-bacterial species (Bernardová et al., 2008). Such differencescould be attributed to the contribution of LPS from othercyanobacterial species (Bernardová et al., 2008) or het-erotrophic bacteria associated with the complex naturalcyanobacterial biomass (Rapala et al., 2002). This is alsosupported by other recent studies that documented highinflammatory reactions to bacterial LPS originating fromaquatic samples (Ohkouchi et al., 2012). In addition, variousLPS-like chemicals could also contribute to the overall py-rogenicity of isolates determined by the LAL or PyroGenetests (Conrad et al., 2006; Bedick et al., 2001). However, thecompositions of bacterial communities associated withcyanobacterial blooms have only been partially character-ized (Eiler et al., 2006; Salomon et al., 2003), and the po-tential role of these bacteria in creating the overall healthrisks associated with cyanobacterial blooms remainselusive. In spite of the apparent variability in EU values,there was no clear correspondence with the biologicalcomposition, % of cyanobacteria in the sample, dominance

L. Bláhová et al. / Toxicon 76 (2013) 187–196194

of M. aeruginosa or to the production of microcystin inoriginal samples (compare Tables 1 and 2). The determinedpyrogenicity (EU values per mg of LPS) allows direct com-parisons to different samples as well as to prototypicalmicroorganisms such as E. coli. According to the WorldHealth Organization (WHO), standard E. coli O113:10endotoxin has an activity of 1 EUper 100 pg of endotoxin. Incomparison (considering the standard LAL test), the max-ima observed in the present study were 0.009 EU per100 pg for the laboratory culture sample M. aeruginosaUTEX 2667 and 0.078 EU per 100 pg for the complex bloomsample TB02. Taken together, M. aeruginosa LPS were atleast 12- to 100-times less potent than WHO standardO113:10 endotoxin (and this phenomenon was even morepronounced when considering the results of the PyroGenetest or ELISA).

Apart from endotoxin activities determined usingbiochemical LAL and PyroGene systems, the present studyalso assessed the biological effects of isolated endotoxins inwhole human blood. The activation of a primarily non-specific immune reaction response ex vivo is a highlyrelevant and sensitive system containing all componentsimportant for the reaction of an organism to endotoxinexposure in vivo. Data showed that all endotoxin isolatesfromM. aeruginosa possess a potential to induce a complexinflammatory reaction corresponding to previously re-ported in vivo observations (Dogo et al., 2011; Palikovaet al., 2013). Interestingly, the levels of the biological ef-fects did not correlate with pyrogenicity in LAL or Pyro-Gene, which also corresponds with previously reportedobservations of other authors (the release of cytokine inmononuclear cells or in whole blood assay; Brandenburget al., 2009), and indicate different processes beyond bothtypes of endpoints. On the other hand, Peters et al. (2012)observed a correlation between the LAL test and the pro-duction of IL-8 in a specific cell line, HEK293, where thecells were transfected by the specific LPS-responding cellreceptors TLR4/MD/CD14. The induction of ROS productionduring the so-called “oxidative burst” of phagocytes (inwhole blood, predominantly presented by neutrophilicgranulocytes) was evaluated as a sensitive marker of im-mune reaction to pathogens or pathogen-associatedstructures (Kubala et al., 2003; Via�cková et al., 2011). Theinduction of ROS production mediated by the studiedendotoxin samples has significant implications for thein vivo situation, since the induction of ROS at the site ofexposure leads to damage of surrounding tissue, thederegulation of physiological cell functions, and the pro-motion of local inflammation (Nathan and Cunningham-Bussel, 2013; Via�cková et al., 2011). Interestingly, ROSproduction elicited by microglia cells in response to M.aeruginosa endotoxin was previously reported by Mayeret al. (2011). The analysis of ROS production was com-plemented by studying the release of two crucial regulatorycytokines IL-8 and TNF-a. In vivo, IL-8 contributes to theinduction of the chemotactic migration of leukocytes to thesite of exposure to the inflammatory agent (Pichert et al.,2012). TNF-a is among the first cytokines released uponexposure of an organism to pathogenic structures andpromotes an inflammatory reaction both locally and sys-tematically through the activation of a wide range of both

leukocytes and other non-immune cells. Determination ofthe potency to induce TNF-a in whole blood is a vitalmethodological approach during the testing for pyrogencontamination, and TNF-a release is also known to belinked to the release of other cytokines such as IL-6 and Il-1b (Hoffmann et al., 2005; Gaines Das et al., 2004). Signif-icant correlation among all three biological endpointssuggests the coincident activation of various mechanismsand functions of blood leukocytes by the tested M. aerugi-nosa endotoxin samples. Overall, the potential of M. aeru-ginosa LPS isolates to induce inflammatory reactions inanimals and humans has been revealed and its risks shouldbe evaluated in detail.

Although the pyrogenicity of cyanobacterial LPS mightseem lower in comparison to highly toxic endotoxin such asE. coli O113:10, the environmental effects of toxic bloomsmight increase associated risks. Exposure to highly toxic E.coli is generally occasional; however, cyanobacteria M.aeruginosa often form blooms inwater reservoirs, includingdrinking water supplies, with densities ranging from 105 to107 cells ml�1. A simple calculation of the “worst-casescenario” was performed for dense cyanobacterial scum(up to 50mg of dry biomass per ml) and using results of themore conservative PyroGene test. The estimated pyroge-nicity of bloom samples ranged from 900 EU/mL (sampleTB11) up to 46,000 EU/mL (samples TB02, TB07). Interest-ingly, these values also reflect biological endpoints, as TB11and TB07 were, in the whole blood tests, samples with verylow and very high potencies, respectively. The calculatedvalues are about an order of magnitude lower than thepreviously reported maximum values from Finnish densewater blooms (up to 380,000 EU/ml, Rapala et al., 2002) orvalues estimated in a previous study on LPS from differentcyanobacteria (Bernardová et al., 2008). However, the dif-ferences could be attributed to our calculation method, inwhich results of the conservative PyroGene test were used.

5. Conclusions

The present study demonstrated that a combination ofseveral techniques is needed to fully characterize both theamount and biological potential of LPS isolated from haz-ardous material such as cyanobacterial bloom. The resultsof the ELISA method indicated that around 15% of cyano-bacterial isolates corresponded to LPS; however, the dif-ference could also relate to the poorly characterized cross-reactivity of antibodies used in the commercial ELISA. ThePyroGene test (in comparison to the LAL test) can be rec-ommended as a more appropriate method for screeningpyrogenicity in complex materials, since it is less affectedby interference and false positives. M. aeruginosa, which isone of the major bloom-forming cyanobacteria worldwide,was found to produce significant amounts of bioactive LPSin both laboratory cultures and natural bloom samples. Inspite of the existence of lower pyrogenicity in comparisonto that of standard E. coli LPS, the present study revealedcomplex biological potencies of cyanobacterial endotoxinsin human blood. In summary, attention should be paid tothe monitoring and removal of both well-known cyano-toxins and endotoxins from drinking water supplies andrecreational water bodies affected by toxic blooms. The

L. Bláhová et al. / Toxicon 76 (2013) 187–196 195

present results indicate existing health risks associatedwith M. aeruginosa LPS.

Ethical statement

Authors declare that they are not aware of any ethicalissues related to the present manuscript, which is beingsubmitted to Toxicon. Blood samples used for testingex vivo effects were obtained from healthy volunteers, whogave their informed consent.

Acknowledgments

The research was supported by Czech National ScienceFoundation grant No. GACR13-27695P. The infrastructurewas supported by the project CETOCOEN (no. CZ.1.05/2.1.00/01.0001) from the European Regional DevelopmentFund. Authors would like to express great thanks to Ms. IdaSou�cková Ol�sová for helping with the map preparation.

Conflict of interest statement

The authors declare that there are no conflicts ofinterest

References

Anderson, W.B., Slawson, R.M., Mayfield, C.I., 2002. A review of drinkingwater-associated endotoxin, including potential routes of humanexposure. Can. J. Microbiol. 48, 567–587.

Bártová, K., Hilscherová, K., Babica, P., Mar�sálek, B., Bláha, L., 2011. Effectsof microcystin and complex cyanobacterial samples on the growthand oxidative stress parameters in green alga Pseudokirchneriellasubcapitata and comparison with the model oxidative Stressor-Herbicide Paraquat. Environ. Toxicol. 26 (6), 641–648.

Bedick, J.C., Shnyra, A., Stanley, D.W., Pardy, R., 2001. Innate immunereactions stimulated by a lipopolysaccharide-like component ofthe alga Prototheca (strain 289). Naturwissenschaften 88, 482–485.

Bernardová, K., Babica, P., Mar�sálek, B., Bláha, L., 2008. Isolation andendotoxin activities of lipopolysaccharides from cyanobacterial cul-tures and complex water blooms and comparison with effects ofheterotrophic bacteria and green alga. J. Appl. Toxicol. 28, 72–77.

Best, J.H., Eddy, F.B., Codd, G.A., 2001. Effects of purified microcystin-LRand cell extracts of Microcystis strains PCC 7813 and CYA 43 on car-diac function in brown trout (Salmo trutta) alevins. Fish Physiol. Bio-chem. 24, 171–178.

Best, J.H., Pflugmacher, S., Wiegand, C., Eddy, F.B., Metcalf, J.S., Codd, G.A.,2002. Effects of enteric bacterial and cyanobacterial lipopolysaccha-rides, and of microcystin-LR, on glutathione S-transferase activities inzebra fish (Danio rerio). Aquat. Toxicol. 60, 223–231.

Bonilla, S., Aubriot, L., Soares, M.C.S., Gonzalez-Piana, M., Fabre, A.,Huszar, V.L.M., Lurling, M., Antoniades, D., Padisak, J., Kruk, C., 2012.What drives the distribution of the bloom-forming cyanobacteriaPlanktothrix agardhii and Cylindrospermopsis raciborskii? FEMSMicrobiol. Ecol. 79, 594–607.

Brandenburg, K., Howe, J., Gutsman, T., Garidel, P., 2009. The expression ofendotoxic activity in the Limulus test as compared to cytokine pro-duction in immune cells. Curr. Med. Chem. 16, 2653–2660.

Buttke, T.M., Ingram, L.O., 1975. Comparison of lipopolysaccharides fromAgmenellum quadruplicatum to Escherichia coli and Salmonella typhi-murium by using thin-layer chromatography. J. Bacteriol. 124, 1566–1573.

Campos, A., Vasconcelos, V., 2010. Molecular mechanisms of microcystintoxicity in animal cells. Int. J. Mol. Sci. 11, 268–287.

Can, Z., Wenjun, L., Wen, S., Minglu, Z., Lingjia, Q., Cuiping, L., Fang, T.,2013. Endotoxin contamination and control in surface water sourcesand a drinking water treatment plant in Beijing, China. Water Res. 47(11), 3591–3599.

Casabuono, A.C., van der Ploeg, C.A., Roge, A.D., 2012. Characterization oflipid A profiles from Shigella flexneri variant X lipopolysaccharide.Rapid Commun. Mass Spectrom. 26 (17), 2011–2020.

Codd, G.A., Morrison, L.F., Metcalf, J.S., 2005. Cyanobacterial toxins: riskmanagement for health protection. Toxicol. Appl. Pharmacol. 203 (3),264–272.

Conrad, M.L., Pardy, R.L., Wainwright, N., Child, A., Armstrong, P.B.,2006. Response of the blood clotting system of the Americanhorseshoe crab, Limulus polyphemus, to a novel form of lipopoly-saccharide from a green alga. Comp. Biochem. Physiol. A Mol.Integr. Physiol. 144, 423–428.

de Santana-Filho, A.P., Noleto, G.R., Gorin, P.A.J., de Souza, L.M.,Iacomini, M., Sassaki, G.L., 2012. GC-MS detection and quantificationof lipopolysaccharides in polysaccharides through 3-O-acetyl fattyacid methyl esters. Carbohydr. Polym. 87 (4), 2730–2734.

Ding, J.L., Navas, M.A.A., Ho, B., 2001. Two forms of factor-C from theamebocytes carcinoscorpius-rotundicauda – purification and char-acterization. Biochim. Biophys. Acta 1202 (1), 149–156.

Ding, J.L., Ho, B., 2001. A new era in pyrogen testing. Trends Biotechnol. 19,277–281.

Dogo, C.R., Bruni, F.M., Elias, F., Rangel, M., Pantoja, P.A., Sant’Anna, C.L.,Lima, L., Lopes-Ferreira, M., de Carvalho, L.R., 2011. Inflammatory ef-fects of the toxic cyanobacterium Geitlerinema amphibium. Toxicon 58(6–7), 464–470.

Dungan, R.S., 2011. Airborne endotoxin from indoor and outdoor envi-ronments: effect of sample dilution on the kinetic Limulus amebocytelysate (LAL) assay. J. Occup. Environ. Hyg. 8 (3), 147–153.

Eiler, A., Olsson, J.A., Bertilsson, S., 2006. Diurnal variations in the auto-and heterotrophic activity of cyanobacterial phycospheres (Gloeo-trichia echinulata) and the identity of attached bacteria. FreshwaterBiol. 51, 298–311.

Fujii, M., Sato, Y., Ito, H., Masago, Y., Omura, T., 2012. Monosaccharidecomposition of the outer membrane lipopolysaccharide and O-chainfrom the freshwater cyanobacterium Microcystis aeruginosa NIES-87.J. Appl. Microbiol. 113, 896–903.

Funari, E., Testai, E., 2008. Human health risk assessment related tocyanotoxins exposure. Crit. Rev. Toxicol. 38 (2), 97–125.

Gaines Das, R.E., Brugger, P., Patel, M., Mistry, Y., Poole, S., 2004. Monocyteactivation test for pro-inflammatory and pyrogenic contaminants ofparenteral drugs: test design and data analysis. J. Immunol. Methods288, 165–177.

Hoffmann, S., Peterbauer, A., Schindler, S., Fennrich, S., Poole, S.,Mistry, Y., Montag-Lessing, T., Spreitzer, I., Loschner, B., vanAalderen, M., Bos, R., Gommer, M., Nibbeling, R., Werner-Felmayer, G., Loitzl, P., Jungi, T., Brcic, M., Brugger, P., Frey, E.,Bowe, G., Casado, J., Coecke, S., de Lange, J., Mogster, B., Naess, L.M.,Aaberge, I.S., Wendel, A., Hartung, T., 2005. International validationof novel pyrogen tests based on human monocytoid cells. J.Immunol. Methods 298, 161–173.

Jaja-Chimedza, A., Gantar, M., Mayer, G.D., Gibbs, P.D., Berry, J.P., 2012.Effects of cyanobacterial lipopolysaccharides from microcystis onglutathione-based detoxification pathways in the zebrafish (Daniorerio) embryo. Toxins (Basel) 4 (6), 390–404.

Komárek, J., Anagnostidis, K., 2005. Cyanoprokaryota. 2. Teil: oscillator-iales. In: Büdel, B., Gärtner, G., Krienitz, L., Schagerl, M. (Eds.), Süs-wasserflora von Mitteleuropa, 19/2. Elsevier, München, p. 759.

Kubala, L., Ruzickova, J., Nickova, K., Sandula, J., Ciz, M., Lojek, A., 2003.The effect of (1 / 3)-beta-D-glucans carboxymethylglucan andschizophyllan on human leukocytes in vitro. Carbohydr. Res. 338,2835–2840.

Lindsay, J., Metcalf, J.S., Codd, G.A., 2006. Protection against the toxicity ofmicrocystin-LR and cylindrospermopsin in Artemia salina andDaphnia spp. by pre-treatment with cyanobacterial lipopolysaccha-ride (LPS). Toxicon 48, 995–1001.

Lindsay, G.K., Roslansky, P.F., Novitsky, T., 1989. Single step, chromogenicLimulus amoebocyte lysate assay for endotoxin. J. Clin. Microbiol. 27,947–951.

Martin, C., Codd, G.A., Siegelman, H.W., Weckesser, J., 1989. Lipopolysac-charides and polysaccharides of the cell envelope of toxic Microcystisaeruginosa strains. Arch. Microbiol. 152, 90–94.

Matiasovic, J., Stepanova, H., Volf, J., Kubala, L., Ovesna, P., Rychlik, I.,Faldyna, M., 12, 2011. Influence of the lipopolysaccharide structure ofSalmonella enterica serovar Enteritidis on interactions with pig neu-trophils. Vet. Microbiol. 150 (1–2), 167–172.

Mayer, A.M.S., Clifford, J.A., Aldulescu, M., Frenkel, J.A., Holland, M.A.,Hall, M.L., Glaser, K.B., Berry, J., 2011. Cyanobacterial Microcystisaeruginosa lipopolysaccharide elicits release of superoxide anion,thromboxane B2, cytokines, chemokines and matrixmetalloproteinase-9 by rat microglia. Toxicol. Sci. 121, 63–72.

L. Bláhová et al. / Toxicon 76 (2013) 187–196196

McKenzie, J.H., Alwis, K.U., Sordillo, J.E., Kalluri, K.S., Milton, D.K., 2011.Evaluation of lot-to-lot repeatability and effect of assay media choicein the recombinant factor C assay. J. Environ. Monit. 13, 1739–1745.

Mohamed, Z.A., Al Shehri, A.M., 2007. Cyanobacteria and their toxins intreated-water storage reservoirs in Abha city, Saudi Arabia. Toxicon50, 75–84.

Nathan, C., Cunningham-Bussel, A., 2013. Beyond oxidative stress: animmunologist’s guide to reactive oxygen species. Nat. Rev. Immunol.13 (5), 349–361.

Notch, E.G., Miniutti, D.M., Berry, J.P., Mayer, G.D., 2011. Cyanobacterial LPSpotentiates cadmium toxicity in zebrafish (Danio rerio) embryos.Environ. Toxicol. 26 (5), 498–505.

Ohkouchi, Y., Tajima, S., Nomura, M., Itoh, S., 2012. Comparison of in-flammatory responses in human cells caused by lipopolysaccharidesfrom Escherichia coli and from indigenous bacteria in aquatic envi-ronment. J. Environ. Sci. Health Part A Toxic/Hazardous Subst. Envi-ron. Eng. 47 (13), 1966–1974.

Palikova, M., Ondrackova, P., Mares, J., Adamovsky, O., Pikula, J.,Kohoutek, J., Navratil, S., Blaha, L., Kopp, R., 2013. In vivo effects ofmicrocystins and complex cyanobacterial biomass on rats (Rattusnorvegicus var. alba): changes in immunological and haematologicalparameters. Toxicon 73, 1–8.

Papageorgiou, J., Linke, T.A., Kapralos, C., Nicholson, B.C., Steffensen, D.A.,2004. Extraction of cyanobacterial endotoxin. Environ. Toxicol. 19,82–87.

Peters, M., Fritz, P., Bufe, A., 2012. A bioassay for determination of lipo-polysaccharide in environmental samples. Innate Immun. 18 (5), 694–699.

Pichert, A., Schlorke, D., Franz, S., Arnhold, J., 2012. Functional aspects ofthe interaction between interleukin-8 and sulfated glycosaminogly-cans. Biomatter 2 (3), 142–148.

Poole, S., Mistry, Y., Ball, C., Gaines Das, R.E., Opie, L.P., Tucker, G., Patel, M.,2003. A rapid done-plateT in vitro test for pyrogens. J. Immunol.Methods 274 (1–2), 209–220.

Rapala, J., Lahti, K., Rasanen, L.A., Esala, A.L., Niemela, S.I., Sivonen, K.,2002. Endotoxins associated with cyanobacteria and their removalduring drinking water treatment. Water Res. 36, 2627–2635.

Salomon, P.S., Janson, S., Graneli, E., 2003. Molecular identification ofbacteria associated with filaments of Nodularia spumigena and theireffect on the cyanobacterial growth. Harmful Algae 2, 261–272.

Saluk-Jusczak, J., Wachowicz, B., 2005. The proinflammatory activity oflipopolysaccharide. Postepy Biochem. 51, 280–287.

Schlosser, U.G., 1994. Sag – Sammlung-Von-Algenkulturen at theUniversity-of-Gottingen – catalog of strains. Bot. Acta 107 (3), 113–186.

Sinha, R., Pearson, L.A., Davis, T.W., Burford, M.A., Orr, P.T., Neilan, B.A.,2012. Increased incidence of Cylindrospermopsis raciborskii intemperate zones – is climate change responsible? Water Res. 46,1408–1419.

Snyder, D.S., Brahamsha, B., Azadi, P., Palenik, B., 2009. Structure ofcompositionally simple lipopolysaccharide from marine Synecho-coccus. J. Bacteriol. 191, 5499–5509.

Stein, J., 1973. Hanbook of Phycological Methods. Culture Methods andGrowth Measurements. Cambridge University Press, Cambridge,p. 477.

Stewart, I., Schluter, P.J., Shaw, R.G., 2006. Cyanobacterial lipopolysac-charides and human healthdA review. Environ. Health A Glob. AccessSci. Source 5, 7.

Via-Ordorika, L., Fastner, J., Kurmayer, R., Hisbergues, M., Dittmann, E.,Komarek, J., Erhard, M., Chorus, I., 2004. Distribution of microcystin-producing and non-microcystin-producing Microcystis sp. in Euro-pean freshwater bodies: detection of microcystins and microcystingenes in individual colonies. Syst. Appl. Microbiol. 27, 592–602.

Via�cková, D., Pekarová, M., Crhák, T., Búcsaiová, M., Matia�sovic, J.,Lojek, A., Kubala, L., 2011. Redox-sensitive regulation of macrophage-inducible nitric oxide synthase expression in vitro does not correlatewith the failure of apocynin to prevent lung inflammation induced byendotoxin. Immunobiology 216 (4), 457–465.

Weckesser, J., Drews, G., Mayer, H., 1979. Lipopolysaccharides of photo-synthetic prokaryotes. Annu. Rev. Microbiol. 33, 215–239.

Wiegand, C., Pflugmacher, S., 2005. Ecotoxicological effects of selectedcyanobacterial secondary metabolites a short review. Toxicol. Appl.Pharmacol. 203, 201–218.

Wilson, A.E., Sarnelle, O., Neilan, B.A., Salmon, T.P., Gehringer, M.M.,Hay, M.E., 2005. Genetic variation of the bloom-forming cyanobac-terium Microcystis aeruginosa within and among lakes: implicationsfor harmful algal blooms. Appl. Environ. Microbiol. 71, 6126–6133.

Zegura, B., Straser, A., Filipic, M., 2011. Genotoxicity and potential carci-nogenicity of cyanobacterial toxins – a review. Mutat. Res. 727 (1–2),16–41.

Zhang, P., Zhai, C., Chen, R., Liu, C., Xue, Y., Jiang, J., 2012. The dynamics ofthe water bloom-forming Microcystis aeruginosa and its relationshipwith biotic and abiotic factors in Lake Taihu, China. Ecol. Eng. 47, 274–277.