The influence of engineered Fe2O3 nanoparticles and soluble (FeCl3) iron on the developmental toxicity caused by CO2-induced seawater acidification

lable at ScienceDirect

Environmental Pollution 158 (2010) 3490e3497

Contents lists avai

Environmental Pollution

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

The influence of engineered Fe2O3 nanoparticles and soluble (FeCl3) ironon the developmental toxicity caused by CO2-induced seawater acidification

E. Kadar a,*, F. Simmance a,b, O. Martin b, N. Voulvoulis b, S. Widdicombe a, S. Mitov c,J.R. Lead c, J.W. Readman a

a Plymouth Marine Laboratory, Prospect Place, The Hoe, Plymouth PL1 3DH, UKbCentre for Environmental Policy, Imperial College London, UKc School of Geography, Earth and Environmental Sciences, University of Birmingham, Birmingham B15 2TT, UK

Developmental toxicity of hypercapnia mediated by exposure to engin

a r t i c l e i n f o

Article history:Received 1 November 2009Received in revised form17 March 2010Accepted 19 March 2010

Keywords:IronpHCarbon capture and storageD-shell larvaeNanotoxicology

* Corresponding author.E-mail address: [email protected] (E. Kadar).URL: http://www.pml.ac.uk.

0269-7491/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.envpol.2010.03.025

a b s t r a c t

An embryodevelopment assay using a common test organism, the ediblemussel (Mytilus galloprovincialis),exposed to both Fe2O3 nanoparticles and soluble FeCl3 at 3 acidic pHs, has provided evidence for thefollowing: (1) CO2 enriched seawater adjusted to pH projections for carbon capture leakage scenarios (CCS)significantly impaired embryo development; (2) under natural pH conditions, no significant effect wasdetected following exposure of embryos to Fe, no matter if in nano- or soluble form; (3) at pH of naturalseawater nano-Fe particles aggregate into large, polydisperse and porous particles, with no biologicalimpact detected; (4) at pH 6 and 7, such aggregates may moderate the damage associated with CO2

enrichment as indicated by an increased prevalence of normal D-shell larvaewhen nano-Fewas present inthe seawater at pH 7, while soluble iron benefited embryo development at pH 6, and (5) the observedeffects of iron on pH-induced development toxicity were concentration dependent.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Pressure to reduce CO2 emissions in response to the threat toglobal warming has highlighted various mitigation options, someinvolving geological sequestration, i.e. injection of CO2 into under-ground reservoirs (Holloway, 2005). Such seemingly convenientmethods, however, have raised concerns over the risks of subsur-face storage leaks (Hawkins, 2004). Whether the ecological risks ofexcess CO2 emission by passive diffusion into global ocean surfacewaters would outweigh those of carbon capture leakage scenarios(CCS) in vulnerable areas are currently being pondered (Stern,2006). Regardless of the mechanisms of entry, an increase in CO2partial pressure will give rise to pH reduction that is predicted toinfluence metal speciation in water. Theoretical speciation modelscould to a certain extent, predict changes in metal speciation, buttheir accuracy is limited due to the complex and dynamic nature ofthe system. Further complications are expected regarding inter-actions of pH with emerging contaminants such as metaloxide nanoparticles, whose potential bioaccumulation and envi-ronmental impact are equally difficult to predict due to lack of

All rights reserved.

information to date (reviewed by Limbach et al., 2009; Klaineet al., 2008).

Chemicals at the nano-scale have unique characteristicscompared to their bulk analogues as a result of a higher specificsurface area (SSA) and surface energies, and a higher proportion ofatoms on their surface versus interior (Handy et al., 2008a). Thesame properties of nanoparticles that give rise to their novelapplications, however, may render unpredicted toxicity (Keenanet al., 2009). Structural and chemical properties such as size,shape, surface chemistry and composition of nanoparticles havebeen shown to affect their behaviour in the environment(Baalousha et al., 2008; Auffan et al., 2008; Navarro et al., 2008;Handy et al., 2008b; Hassellov et al., 2008). Therefore, a prerequi-site for suitable risk assessment criteria is the characterisation ofthe physical and chemical properties of nanoparticles in order tounderstand their fate and behaviourwithin environmentalmatrices(Alvarez et al., 2009; Hassellov et al., 2008).

Engineered iron oxide nanoparticles are widely used in themedical sector such as magnetic drug targeting systems (Berry andCurtis, 2003) or contrast agents in magnetic resonance imaging(Schulze et al., 1995), in the food/feed industry as additives andnutritional supplements for humans (Fidler et al., 2004) andaquaculture, and more recently in the waste water treatmentindustry in Europe (de Boer et al., 2009) and the USA (Shan et al.,

Table 1Characterisation measurements performed with nano-Fe2O3 and soluble FeCl3.

Particle characterisation Nano-Fe2O3 Soluble FeCl3

Z-average diameter (nm) (DLS) 3167 � 135 ndPolydispersity index 0.41 ndZeta potential (mV) �2.4 � 0.3 2.2 � 0.2Hydrodynamic diameter (nm) (NTA) 130 � 15 404 � 20Particle number (mL�1) (1.3 � 0.5) � 108 (0.5 � 0.1) � 108

Mean particle size (nm) (AMF) 40.1 � 1.3 239.4 � 15.4Crystallite size (nm) (XRD) 50.4 � 3.7 123.6 � 7.5Specific surface area (m2 g�1) (BET) 37.1 � 0.2 e

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e3497 3491

2009). Such usage entails production volumes that, if dischargedinto estuarine habitats, may synergistically interact with the effectsof CO2 emission-induced acidification.

Acidification of body fluids as a result of increasing external CO2known as hypercapnia, occurs rapidly, in a matter of hours byaltering the acidebase regulation (Michaelidis et al., 2005; Mileset al., 2007), growth, reproduction, feeding and ultimatelycausing mortality in many aquatic species (Kurihara et al., 2004;Kurihara and Shirayama, 2004; Raven et al., 2005). Most marineanimals are sensitive to reduced pH (Ormerod and Angel, 1996), butthose that rely on calcification processes for survival, includingechinoderms (Miles et al., 2007), corals (Seibel and Fabry, 2003),molluscs (Lindinger et al., 1984; Michaelidis et al., 2005), crusta-ceans (Spicer et al., 2007; DeFur and McMahon, 1984) and calcifiedalgal species (Riebesell et al., 2000) are the most vulnerable. Earlystages of the life cycle are suggested to be most sensitive toincreased CO2 concentrations, with potentially widespread impli-cations for population size, community structure and biodiversity(Widdicombe and Spicer, 2008). Acidic pH causes loss of spermmotility of Pacific oysters (Crassostrea gigas) (Dong et al., 2002),diminished reproductive success of sea urchins (Kurihara andShirayama, 2004), copepods (Kurihara et al., 2004), mussels(Kurihara et al., 2008) and barnacles (Findley et al., 2009).

Oxidised tertiary iron is thought to be biologically unavailable,unless at low pH and anoxic conditions (Sunda, 2001). However,little is known about Fe nanoparticle uptake pathways and fate inaquatic organisms (Kadar et al., 2010), and there is no informationon low pH mediated nanoparticle toxicology. A number of studiesinfer that exposure to nano-iron (oxides and/or zero valent iron)renders generation of free radicals. These reactive oxygen species(ROS) cause inflammation of epithelial tissue and alteration of anti-oxidant enzymatic activity (Kadar et al., 2010; Auffan et al., 2009;Keenan et al., 2009; Li et al., 2009; Fernaeus and Land, 2005; Zhouet al., 2003). Moreover, recent investigations comparing nZVI andiron oxide NPs conclude that nZVI produces most ROS and thus it ismost toxic (Auffan et al., 2009). In contrast, other experiments,showed no toxicity response to nano-iron in mice (Rohner et al.,2007), or in bovine sperm cells (Makhluf et al., 2006). Suchcontradictory results may originate from specific experimentalconditions (organism type, exposure media, duration of exposure,pH, etc.) that greatly influence toxicity of nanoparticles.

To test the hypothesis that nano-metal oxidesmay bemore toxicunder reduced pH conditions as compared to the soluble analogue,we have exposed embryos of Mytilus galloprovincialis obtainedby in vitro cross-fertilisation to both Fe2O3 nanoparticles (withdiameter <50 nm) and soluble FeCl3 under static conditions for upto 48 h. A 3-way crossed experimental design was used to assessthe percentage of normally developing larvae at a well distin-guished growth-stage in response to hypercapnia mediated bynano- versus soluble Fe exposure.

2. Materials and methods

2.1. Experimental model

Commercially available mussels of Mytilus spp. were purchased from RockShellfish, Cornwall (batch number: 28/6/T1) and were identified as Mytilusgalloprovincialis, using the PCR method adapted from Inoue et al. (1995). Briefly,Me15 and Me16 primers and GoTaq polymerase were used in the following cycle:a 5 min denaturation at 95 �C followed by 40 cycles of 95, 55, 72 �C for 1 min eachand a 5 min elongation at 72 �C.

2.2. Induction of spawning and in vitro fertilisation

Mussels were kept in aerated oceanic seawater (T ¼ 15 � 0.5 �C andsalinity ¼ 35.1 � 0.52 s.u.) until induction of spawning (3e5 h following theacquisition from the local fish market).

In vitro cross-fertilisation was performed using oocytes pooled from 5 mussels,and sperm extracted from a separate group of 5 males. Oocytes and sperm wereobtained by chemical induction of spawning (i.e. injection of 1.0 mL of 0.5% w/v KClinto the posterior of the adductor muscle) followed by provision of food supply(Phytofeast� administered at approximately 1.25 � 104 cells mL�1) combined withheat shock from 11 �C to 16 �C. Gametes were inspected for general health, asindicated by spherical shape for eggs and mobility for sperm, prior to fertilisation.Fertilisation was performed using sterile laboratory equipment in salinity- andtemperature-adjusted, 0.2 mm-filtered (Anachem�) seawater (T ¼ 15 � 0.5 �C andsalinity ¼ 35.1 � 0.52 s.u.). The number of eggs was adjusted to an optimum of30 eggs mL�1 suspension in filtered seawater (Sprung and Bayne, 1984). Spermcounts were carried out using a haemocytometer, and sperm was added to the eggsuspension to yield a 105e107 sperm mL�1 mixture. Constant temperature of15 � 1 �C was maintained during fertilisation and throughout larval development.Excess spermwas removed 1 h after fertilisation to prevent polyspermy (Sprung andBayne, 1984). Fertilisation success was determined using the formula: (numberof fertilised eggs)/(number of fertilised þ unfertilised eggs) � 100, and was above80%. The embryo suspension was concentrated using plastic sieves to 1000embryos mL�1, to yield approximately 100 embryos in each experimental vial.

2.3. Preparation of nanoparticle suspension

Commercially available nanoparticles (50 nm diameter ferric iron, SigmaeAldrich, CAS: 1309-37-1) were suspended in 0.2 mm-filtered seawater which wasthen sonicated for 30 min in a Lucas Dawe Ultrasonics bath (Type no. 6456-A1). Astock solution of 1.2 g L�1 was made in freshly filtered (0.2 mm pore size filters), pre-cooled seawater. Equimolar Fe3þ was provided from the soluble form of the metal(FeCl3), i.e. 96 mg L�1. Estimation of equimolar nano- versus soluble Fe3þ has beenpreviously described by Kadar et al. (2010). Briefly, the number of surface boundFe3þwas calculated using the average 150m2 g�1 surface toweight ratio of the Fe2O3

nanoparticles (SigmaeAldrich product description) and the covalent atomic radiusof Fe2O3. Then, a FeCl3 solution was made having the number of Fe3þ of the sameorder of magnitude, i.e. w1020. Calculations were performed under the assumptionthat all particles are spherical. WHAM/Model VI (Windermere Humic AqueousModel) (Tipping, 1994) was used to simulate the precipitation of amorphous Fe (III)hydroxide. The typical DOC (dissolved organic carbon) concentrations of the surfaceocean range between 60 and 80 mmol L�1 (Sharp, 2002). Assuming a 66% fulvic acidcontribution to the DOC (Thurman, 1985), simulations were performed with threedifferent Fe (III) concentrations (8 mg L�1, 0.8 mg L�1 and 0.08 mg L�1) at threedifferent pH values (6, 7 and 8.1) and with four different fulvic acid concentrations(0.48mg L�1, 0.55mg L�1, 0.63mg L�1 and 1mg L�1). Fe (III) was chosen as a solutioncomponent and FA as a colloidal phase, and Fe (III) (OH)3 was allowed to precipitatein the initial parameter settings. Comparing the concentrations of Fe (III) and ofprecipitate Fe (III) (OH)3, all simulations showed that nearly 100% of dissolvedFe precipitated as expected.

2.4. Multimethod approach for nanoparticle characterisation

Particle sizes (hydrodynamic diameters), polydispersity index and zeta potentialwere measured on a Zetasizer Nano ZS ZEN3600 (Malvern Instruments Ltd.,Malvern, UK) operating with a HeeNe laser at a wavelength of 633 nm using backscattered light. Measurements were carried out on the stock solutions with particleconcentrations of 10 mg L�1 (Fe2O3) and 8 mg L�1 (FeCl3), after sonication of thestabilized solutions for 30 min. Results were the means of triplicate runs and in eachrun ten measurements were made (Table 1).

Particle sizes and particle numbers were also measured using a NanoSight LM10system with a laser output of 30 mW at 650 nm. Mean square displacements ofsingle particles were determined by tracking the scattered light followed by analysisby the NTA (Nanoparticle Tracking Analysis) software. Measurements were carriedout at particle concentrations of 0.1 mg L�1 (Fe2O3) and 0.08 mg L�1 (FeCl3) andresults were the means of triplicate runs (Table 1).

The adsorption method was used to prepare samples for AFM (Atomic ForceMicroscopy) analysis, which preferentially investigates small particles, which arestrongly sorbed to mica (Lead et al., 2005). In this method, mica sheets were cleavedon both sides, and then immersed vertically into the sample solution (10 mg L�1 and

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e34973492

8 mg L�1 for Fe2O3 and FeCl3, respectively) for 2 h. Following adsorption, the micasheets were withdrawn from the solution and gently rinsed by immersion indeionised water to remove non-adsorbed sample. All AFM images were obtainedusing an XE-100 AFM (Park Systems). All scans were performed in air, at roomtemperature and AFM height measurements were recorded. Images were acquiredin a true non-contact mode and recorded in topography mode with a pixel size of256 � 256 and a scan rate of 0.5e1.0 Hz (Fig. 1A, B).

Fig. 1. Multimethod characterisation of engineered nano-Fe2O3 versus soluble FeCl3 particmicroscopy (TEM in subfigures C and D) and X-ray diffraction (XRD in (E)); A) partially repreparticles loosely aggregated, while in B) FeCl3 particles show less coverage and larger sizeparticles loosely bound and with large size range while D) shows FeCl3 particles amassedshowing distinct crystallite morphology with smaller crystallites (w50 nm) in tetragonal phsalt.

Particles in both iron stock solutions were observed by Transmission ElectronMicroscopy (TEM). Drops were partially air-dried onto formvar-coated grids fora short time before removal of excess sample and a rinse with water to removeexcess, non-sorbed salt. Grids were examined with a JEOL EXII with Soft ImagingSystem iTEM software and Megaview III camera (Fig. 1C, D).

X-ray diffraction (XRD) was performed using a Bruker AXS D8 Autosampler. EVAsoftware programwas used for the assignment of reflections and analysis of the XRD

les using atomic force microscopy (AFM in subfigures A and B), transmission electronsentative AFM image obtained with nano-Fe2O3 showing high coverage of polydisperse(size under assumption of spherical shape); C) TEM micrograph showing nano-Fe2O3

into large aggregates; E) evolution of XRD patterns of both Fe2O3 and FeCl3 particlesase for nanoparticles and larger crystallites (w124 nm) in monoclinic phase for the Fe

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e3497 3493

patterns. The Fe2O3 and FeCl3 crystallite sizes were calculated using the Scherrerequation (Table 1).

BET (Brunauer, Emmett and Teller) surface areas (m2 g�1) were determinedusing a Coulter� SA3100� series surface area and pore size analyser. Samples wereoutgassed for 10 h at 200 �C prior to analysis (Table 1).

2.5. Acidification of natural seawater

Seawater was collected from station L4, approximately 8 miles south ofPlymouth, UK (50�150N, 04�130W, water depth w55 m) and shipped to supply theseawater acidification system at PML described in detail by Widdicombe andNeedham (2007). This system uses 100% CO2 gas bubbled through seawater con-tained within large (450 L) reservoir tanks with pre-set pH. The water used in thisstudy was taken from this system by removing aliquots of acidified seawater fromreservoir tanks set at either pH 7 or pH 6. Soluble CO2 levels in this water weremeasured using the Licor LI-6262 CO2 analyser and total carbon dioxide (tCO2) wascalculated using the System Program described in Pierrot et al. (2006). The averagevalues of physico-chemical parameters of the seawater used in various treatmentsare summarized in Table 2. There was no significant change in the pH of testsolutions following the addition of Fe. There was a small (DpH 0.02e0.04) increasethat was systemic in all treatments.

2.6. Experimental design

One hundred and twenty-six 12-mL gastight sample containers (Labco�) wereused to expose embryos to the metals in static conditions, using three replicates ofeach treatment (metal form � metal concentration � pH; 2 � 3 � 3) for each timeinterval (T0 and T48 h). Each vial contained approximately 100 embryos (within 2 hof fertilisation) and 11mL of test solution yielding the nominal Fe2O3 concentrationsof 0, 100, 1000, and 10,000 mg L�1, and 0, 8, 80 and 800 mg L�1 of FeCl3, respectively.Vials were assessed for numbers of normal or abnormal embryos after 48 h, whenthe distinguishable D-shell stage was unanimously reached (at 15 �C). Counts andlight-microscopic photographs of typical normal and abnormal larvae were madeon fixed specimens (see below). An embryo was considered malformed if it hasnot reached D-shell by 48 h or when some developmental defect was observed(i.e. concave, malformed or damaged shell, protruding mantle, etc.).

2.7. Metal analysis

Water samples (10 mL) were taken immediately following addition of the metal,and after 48 h. The samples were filtered through 0.2 mm (Anotop�) filters. Filterswere digested in Aqua Regia overnight and the diluted digests (1:3) together withthe acidified water samples (pH 2) were analysed for total Fe using a Varian 725-ESICP-OES instrument (detection limit was in the region of 5e10 mg L�1). Standardsolutions ranging between 0 and 2000 mg L�1 were made up in 0.2 mm-filtered andacidified, natural seawater.

2.8. Embryo preparation for scoring by light microscopy

Larvaewere fixed in 5% buffered formaldehyde andwere observed using anMCANikon (Japan) inverted light microscope fitted with a JVC Widefield digital camera.

2.9. Statistical analysis

Permutation-based Analysis of Variance was used to detect statistically signifi-cant differences between treatments (Control, exposed to soluble Fe and exposed tonano-Fe). This procedure is formally equivalent to a standard ANOVA but the flex-ibility and robustness of the permutation approach ameliorates the necessity forvariables to fulfill standard assumption, such as normality (Somerfield et al., 2008).Calculations were made using PRIMER 6 (Clarke and Gorely, 2006) andPERMANOVA þ b18 software.

3. Results

3.1. Nanoparticle characterisation

Zeta potential measurements of Fe2O3 and FeCl3 particles inseawater suggested a high propensity for aggregation, which was

Table 2Physico-chemical parameters in experimental vials measured prior to addition ofnano- and soluble Fe.

Treatment DO (%) pH Salinity CO2 (mM kg�1)

Control 85.5 � 11 8.1 � 0.4 35.1 � 0.52 357.8Acidic 82.5 � 10 6.9 � 0.5 35.1 � 0.52 6616.3Highly acidic 86.5 � 15 6.0 � 0.05 35.1 � 0.52 54,201.2

confirmed by the hydrodynamic diameter measurements (Table 1),as expected from iron nanoparticle behaviour (Baalousha et al.,2008). The Z-average diameters from DLS (Dynamic Light Scat-tering) measurements indicate the formation of large aggregates inwater and the high polydispersity indexes suggest that both largerand smaller aggregate sizes are present. In the case of FeCl3 inseawater the intensity of the signal was very low and no Z-averagediameter is reported. Particle sizes determined fromNTA techniqueare lower compared to the values obtained from DLS technique(Table 1), as previously found (Domingos et al., 2009), most likelydue to the bias towards large aggregates in DLS measurements.Individual particle trajectories are tracked in NTA and the effect ofaggregation is less important than in DLS. Particle numberconcentration was higher for the Fe2O3 sample compared to theFeCl3 sample shown by both NTA (Table 1) and AFM (Fig. 1A, B), andin agreement with the observed low intensity DLS signal for thelatter sample. It seems that the observed particles are due to thebackground seawater in the case of FeCl3 as the particle number isvery close to the limit of detection. Increasing the FeCl3 measure-ment concentration to 0.8 mg L�1 led to the presence of large(strongly scattering) aggregates between 300 and 400 nm andno consistent signal was observed leading to large standarddeviations.

Semi-quantitative AFM analysis showed the presence of parti-cles with approximate sizes of 40 nm in the case of Fe2O3 and240 nm for FeCl3 (Table 1 and Fig. 1A, B). Large particles are notexpected to stick to the mica support due to low diffusivities andpossibly low adhesion to the mica.

Transmission electron micrographs of air-dried particles in bothmedia confirmed distinct aggregation patterns: loose aggregates ofvarying sizes for nano-Fe2O3, and less polydisperse, more compactaggregates in the salt solution (Fig. 1C, D).

The BET (Brunauer, Emmett and Teller) method was used fordetermining Fe2O3 specific surface area (37.1 m2 g�1) (Table 1).

The Fe2O3 and FeCl3 crystallite sizes were calculated using theScherrer equation. As expected and in agreement with the Braggpeak broadening of the XRD patterns (Fig. 1E), the Fe2O3 nano-powder showed smaller crystallite size (w50 nm) compared to theFeCl3 salt (w124 nm) (Table 1). XRD analysis confirmed monocliniccrystal system (base-centered lattice) in case of FeCl3 and tetrag-onal crystal system (primitive lattice) in case of Fe2O3.

3.2. Morphological changes induced by acidic pH and nano- versussoluble Fe in M. galloprovincialis larvae

Acidic seawater (pH 6) caused the most severe impact onembryogenesis; there were practically no normal D-shape larvae(a typical healthy D-shell stage larva is shown in Fig. 2A) 48 hfollowing exposure; most larvae were held back in trochophorestage (equivalent to a 24 h development stage in controls undersame day degrees shown in Fig. 2D) or presenting malformationswith various severities (Fig. 2). Frequently the shells that normallyhave a perfect D-shape (Fig. 2A) were concave (Fig. 2C). Protrusionof the mantle (Fig. 2B) indicated that the larva was dead at the timeof the fixation, and also accounted for abnormality. The normalD-shape larvae developed in the presence of nano-Fe were swim-ming actively in spite of the presence of aggregated particles(Fig. 2F).

3.3. Developmental effects of acidic seawater in interactionwith nano-Fe versus soluble Fe

Exposure of Mytilus galloprovincialis embryos to increasingconcentrations of CO2 equivalent to predictions for differentscenarios of carbon capture leakage (pH 6 and 7) resulted in

Fig. 2. Light-microscopic photographs of the typical early development stage of M. galloprovincialis at 48 h day degree at 15 �C in A) controls (pH 8.1), normal D-shell larvae;B) abnormal D-shell with protruding mantle at pH 7; C) abnormal D-shell with concave hinge at pH 7; D) delayed embryo that shows normal trochophore morphology at pH 7;E) abnormal trochophore with shrunk diameter at pH 6; and F) abnormal trochophore unable to swim at pH 6; note large aggregates of nanoparticles (asterisk).

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e34973494

a severe decrease in the number of normal D-shape larvae ascompared to average 80% in control conditions in FSW at pH 8.1(Fig. 3A). At pH 6 an average 40% of the larvae were in trochophorestage 48 h following fertilisation, as opposed to controls where only10% were delayed (Fig. 3B).

Curiously, for embryos in natural seawater, i.e. pH 8.1, thepercentage of normal D-shape larvae after 48 h from fertilisationwas not significantly affected by any of either the nano-Fe or thesoluble Fe concentrations (Fig. 4). In contrast, at pH 7 in the presenceof nano-Fe, the prevalence of normal D-shell larvae increased sig-nificantly, while at pH 6 soluble Fe had the same effect (Fig. 5A, B).

3.4. Water chemistry

Data from the total Fe analysis of the filtrate, i.e. particles withdiameter <200 nm suggests that regardless of the pH this frac-tion was relatively stable over 48 h at the low and mid concen-trations of added nano-Fe (Fig. 6, left side). Furthermore, at pH 6this fraction increased while the retentate fraction diminished by48 h, indicating dissolution or disaggregation of larger nano-particles or colloidal stability of smaller aggregates or individualnanoparticles. However, when the higher concentration(10 mg L�1) was added, total Fe in the filtrate decreased in a pHdependent manner for both forms of the metal, by 48 h followingaddition.

4. Discussion

This study showed that relatively small pH changes over shorttime (48 h) may cause serious disruption of development inMytilusgalloprovincialis consisting of, not only over 50% mortality and/ordamaged larvae, but also a delay in development, i.e. larvae weresuspended in the trochophore stage. The delay may be a naturaldefence mechanism to survive temporary poor environmentalconditions. Similar pH-induced delays in embryogenesis were

recently observed in several other marine species (DeFur andMcMahon, 1984; Lindinger et al., 1984; Riebesell et al., 2000;Dong et al., 2002; Seibel and Fabry, 2003; Kurihara andShirayama, 2004; Kurihara et al., 2004, 2008; Michaelidis et al.,2005; Raven et al., 2005; Dupont et al., 2008; Findley et al., 2009)and potentially could have large scale ecological consequences. It isnot yet known whether such delays are reversible or would affectthe general health of the larvae at a later stage in development aswell. However, the prolonged residence of larvae in the planktonitself would have indirect systemic effects on communities.

The mechanisms of how external pH affects embryo develop-ment are not clear. However, in adult Mytilus reduced pH is knownto induce a reduction of haemolymph pH (Michaelidis et al., 2005)rendering the organism vulnerable to slight changes in seawaterpH. Further longer term exposures combined with recovery studiesare needed to predict accurately the extent of the impact and itsputative reversibility.

Our results provide evidence for the lack of atypical toxicresponses to either form of the iron at natural pH conditions, whichare consistent with previous studies on doseeresponse to bothsoluble (Valko et al., 2005) and nano-iron exposures (reviewed byLewinski et al., 2008) at physiological pH. Similarly, little or notoxicity was reported on in vitro cell viability assays in the mouse(Gupta and Wells, 2004; Hussain et al., 2005; Cheng et al., 2005)over relatively short exposure duration. Lack of severe toxicity ofnano- versus soluble Fe was previously inferred from our in vitroexposure study on excised Mytilus sp. gills (Kadar et al., 2010) andwas associated with the aggregation and sedimentation of Fenanoparticles, and the precipitation of amorphous Fe (III) hydrox-ides that are known to form in alkaline aqueous solutions (Turnerand Hunter, 2001) when the metal is present as Fe (III). Indeed,our speciation simulations showed that nearly 100% of dissolved Feprecipitated, as expected. The lack of nano-specific toxicity in thisassay, together with our previous studies may suggest that undernatural seawater pH and ionic strength conditions, engineered Fe

0

10

20

30

40

50

60

70

pH 6 pH 7 pH 8.1

slle

hs

Dd

ep

ole

ve

dyll

am

ro

nf

o%

A

B

0

10

20

30

40

50

pH 6 pH 7 pH 8.1

oy

rb

me

de

yal

ed

fo

%s

Fig. 3. Embryo scores expressed as A) % of normal D-shells of total larvae in the watercolumn following 48 h of exposure to three concentrations of CO2 as equivalent to pH6, 7 and 8.1, respectively; and as B) % of trochophores, i.e. delayed embryos following48 h of exposure to three concentrations of CO2 as equivalent to pH 6, 7 and 8.1.

0

10

20

30

40

50

60

70

Control nano-Fe soluble-Fe

low mid high

ea

vr

al

ll

eh

s-

Dl

am

ro

n%

Fig. 4. Normal embryo scores expressed as % of total larvae in the water columnfollowing 48 h of exposure to three concentrations of nano-Fe2O3, i.e. low (100 mg L�1),mid (1000 mg L�1) and high (10,000 mg L�1) concentrations, and soluble FeCl3, i.e. 8, 80and 800 mg L�1 under natural ocean pH (i.e. 8.1). Vertical bars represent averagevalues � SEM, N ¼ 3.

Control Low Mid High

pH 6

0

2

4

6

8

10

12

14

pH 7

0

0.5

1

1.5

2

2.5

3

nano-Fe soluble-Fee

av

ra

ll

le

hs

-D

la

mr

on

fo

%

Control Low Mid High

A

B

Fig. 5. Embryo scores at acidic pHs: A) pH 7 and B) pH 6 following 48 h of exposure tothree concentrations of nano-Fe2O3, i.e. 100, 1000 and 10,000 mg L�1 (referred as low,mid and high), and soluble FeCl3, i.e. 8, 80 and 800 mg L�1 (referred as low, mid andhigh). Points represent average values � SEM, N ¼ 3.

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e3497 3495

nanoparticles aggregate into larger particles with properties anal-ogous to those of naturally present amorphous Fe. This was alsoconfirmed by our multimethod approach for characterisations ofthe aggregates.

Curiously, low-pH associated stress was reduced in the pres-ence of iron, with detection of subtle but significant nano-formrelated disparity. In the pH range of natural estuarine waters (pH6e8) iron oxides aggregate into increasingly larger porous aggre-gates (Handy et al., 2008b; Baalousha et al., 2008; this study). It ispossible that in more acidic waters (pH < 7) the nano-Fe aggre-gates with highly porous surface act as buffers scavenging thetoxic compounds that cause toxicity at acidic pH (Kurihara et al.,2008). Possibly, the iron oxide buffers the pH change in theimmediate cell vicinity (or even within the cell, which deservesfurther investigations).

In summary, our results indicate that a conventional biomarkerof developmental toxicity, the Mytilus embryo assay is a suitabletool not only for the rapid assessment of harmful effects of CO2-induced seawater acidification, but also to provide evidence forputative mitigating effect of iron nanoparticles. The observedeffects on the early life stages of Mytilus galloprovincialis togetherwith several other species (Lindinger et al., 1984; DeFur andMcMahon, 1984; Riebesell et al., 2000; Dong et al., 2002; Seibeland Fabry, 2003; Kurihara et al., 2004; Michaelidis et al., 2005;

pH6

1

10

100

1000

10000

100 µg/L 1000 µg/L 10000 µg/L

100 µg/L 1000 µg/L 10000 µg/L

100 µg/L 1000 µg/L 10000 µg/L

pH7

1

10

100

1000

10000

pH8

1

10

100

1000

10000

0

20

40

60

80

100

120

140

80 µg/L 800 µg/L 8000 µg/L

80 µg/L 800 µg/L 8000 µg/L

80 µg/L 800 µg/L 8000 µg/L

T0 T48

0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

140

Lg

m(

eF

fo

no

it

ar

tn

ec

no

cl

ani

mo

N-1

)0

n

m0

2<

sel

ci

tr

ap

sa

Added soluble-Fe concentrationAdded nano-Fe concentration

Fig. 6. Concentration of total Fe (mg L�1) in the form of particles with diameter <200 nm in suspension in the experimental vials right after (T0) and 48 h following addition (T48) oflogarithmically increasing concentrations both of nanoparticles (100, 1000, 10,000 mg L�1 in the left panel) and of soluble FeCl3 (80, 800, 8000 mg L�1 in the right panel) at 3 distinctseawater pHs. Vertical bars represent average � SEM (N ¼ 3).

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e34973496

Findley et al., 2009) give rise to concerns over acidification affectingthe health and ecological productivity of marine organisms, bothnatural and farmed, which would have serious ecological andeconomic implications. Given the important ecological role ofmarine bivalves in littoral ecosystems (Kurihara et al., 2008), as wellas their importance as a source of seafood, it is likely that CO2-induced reduction in seawater pH could potentially affect thestructure, productivity, function and diversity of aquatic ecosys-tems, and may even influence human nutrition. The demandforecast for increased seafood consumption is approximately 8%through 2025 (FAO and the International Food Policy ResearchInstitute), but availability of supply under the changing environ-mental conditions could be impacted.

The research described in this paper affords a multi-stress ap-proach using a simple, conventional and rapid ecotoxicological assay.The technique is shown to be a valuable tool in nano-toxicologicalstudies. The novelty of the approach lies in moving beyond assess-ments of a single stressor, addressing simultaneous assessment ofexposure to an emerging contaminant combined with additionalstress associated with global climate change. Very strong

interactions are indicated. This parallel approach overcomes limi-tations of single assessments that alone are insufficient to meet theneeds of evidence based policies on regulation of the commercialuse of nanoparticles.

Acknowledgments

We thank Amanda Beesley and Christine Pascoe for assistancewith the embryo assays. Paul Somerfield is acknowledged for hisconstructive advice in statistical data analysis. John Bignell andTim Bean from CEFAS Weymouth are greatly acknowledged for thePCR identification of the commercial test species. NERC fundingfor the Facility for Environmental Nanoparticle Analysis andCharacterisation (FENAC) is acknowledged.

References

Alvarez, P.J., Colvin, V., Lead, J.R., Stone, V., 2009. Research priorities to advance eco-responsible nanotechnology. ACS Nano 3, 1616e1619.

Auffan, M., Achouak, W., Rose, J., Roncato, M.A., Chaneac, C., Waite, D.T., Masion, A.,Woicik, J.C., Wiesner, M.R., Bottero, J.Y., 2008. Relation between the redox state

E. Kadar et al. / Environmental Pollution 158 (2010) 3490e3497 3497

of iron-based nanoparticles and their cytotoxicity toward Escherichia coli.Environmental Science & Technology 42, 6730e6735.

Auffan, M., Rose, J., Bottero, J., Lowry, G.V., Jolivet, J.P., Wiesner, M.R., 2009. Towardsa definition of inorganic nanoparticles from an environmental, health andsafety perspective. Nature Nanotechnology 4, 634e641.

Baalousha, M., Manciulea, A., Cumberland, S., Kendall, K., Lead, J.R., 2008. Aggrega-tion and surface properties of iron oxide nanoparticles: influence of pH and nat-ural organic matter. Environmental Toxicology and Chemistry 27, 1875e1882.

Berry, C.C., Curtis, A.S.G., 2003. Functionalisation of magnetic nanoparticles forapplications in biomedicine. Journal of Physics D: Applied Physiology 36,198e206.

de Boer, C.V., Braun, J., Klaas, N., 2009. Injection of nano-scale iron for the in-situremediation of contaminated groundwater. In: Fourth International Conferenceon the Environmental Effects of Nanoparticles and Nanomaterials, University ofVienna, 6e9 September 2009.

Cheng, F., Su, C., Yang, Y., Yeh, Y., Tsai, C., Wu, C., Wu, M., Shieh, D., 2005. Charac-terization of aqueous dispersions of Fe3O4 nanoparticles and their biomedicalapplications. Biomaterials 26, 729e738.

Clarke, K.R., Gorely, R.N., 2006. PRIMER v6: User Manual/Tutorial. PRIMER-E Ltd.,Plymouth.

DeFur, P.L., McMahon, B.R., 1984. Physiological compensation to short term airexposure in red rock crabs Cancer productus from littoral and sublittoralhabitats: acidebase balance. Physiological Zoology 57, 151e160.

Domingos, R.F., Baalousha, M.A., Ju-Nam, Y., Reid, M.M., Tufenkji, N., Lead, J.R.,Leppard, G.G., Wilkinson, K.J., 2009. Characterizing manufactured nanoparticlesin the environment: multimethod determination of particle sizes. Environ-mental Science & Technology 43, 7277e7284.

Dong, Q.X., Eudeline, B., Allen, S.K., Tiersch, T.R., 2002. Factors affecting spermmotility of tetraploid pacific oysters. Journal of Shellfish Research 21,719e723.

Dupont, S., Havenhand, J., Thorndyke, M., 2008. CO2-driven acidification radicallyaffects larval survival and development in marine organisms. ComparativeBiochemistry and Physiology e Part A: Molecular & Integrative Physiology150, S170.

Fernaeus, S., Land, T., 2005. Increased iron-induced oxidative stress and toxicity inscrapie-infected neuroblastoma cells. Neuroscience Letters 382, 217e220.

Fidler, M.C., Walczyk, T., Davidsson, L., Zeder, C., Sakaguchi, N., Juneja, L.R.,Hurrell, R.F., 2004. A micronised, dispersible ferric pyrophosphate withhigh relative bioavailability in man. British Journal of Nutrition 91, 107e112.

Findley, H., Kendall, M., Spicer, J., Widdicombe, S., 2009. Future high CO2 in theintertidal may compromise adult barnacle Semibalanus balanoides survival andembryonic development rate. Marine Ecology Progress Series 389, 193e202.

Gupta, A.K., Wells, S., 2004. Surface-modified superparamagnetic nanoparticles fordrug delivery: preparation, characterization, and cytotoxicity studies. IEEETransactions on Nanobioscience 3 (489), 66e73.

Handy, R.D., Henry, B.T., Scown, M.T., Johnston, D.B., Tyler, R.C., 2008a. Manufac-tured nanoparticles: their uptake and effects on fishda mechanistic analysis.Ecotoxicology 17, 396e409.

Handy, R.D., von der Kammer, F., Lead, R.J., Hassellov, M., Owen, R., Crane, M., 2008b.The ecotoxicology and chemistry of manufactured nanoparticles. Ecotoxicology17, 287e314.

Hassellov, M., Readman, W.J., Ranville, F.J., Tiede, K., 2008. Nanoparticle analysisand characterization methodologies in environmental risk assessment ofengineered nanoparticles. Ecotoxicology 17, 344e361.

Hawkins, D.G., 2004. No exit: thinking about leakage from geologic carbon storagesites. Energy 29, 1571e1578.

Holloway, S., 2005. Underground sequestration of carbon dioxide e a viablegreenhouse gas mitigation option. Energy 30, 2318e2333.

Hussain, S.M., Hess, K.L., Gearhart, J.M., Geiss, K.T., Schlager, J.J., 2005. In vitrotoxicity of nanoparticles in BRL 3A rat liver cells. Toxicology in Vitro 19,975e983.

Inoue, K., Waite, J.H., Matsuoka, M., Odo, S., Harayama, S., 1995. Interspecific vari-ations in adhesive protein sequences of Mytilus edulis, M. galloprovincialis, andM. trossulus. Biological Bulletin 189, 370e375.

Kadar, E., Lowe, D., Sole, M., Fisher, A.S., Jha, A.N., Readman, J.W., Hutchinson, T., 2010.The uptake and biological responses to nano-Fe versus soluble FeCl3 in excisedmussel gills. Analytical and Bioanalytical Chemistry 396, 657e666.

Keenan, C., Goth-Goldstein, R., Lucas, D., Sedlak, L.D., 2009. Oxidative stress inducedby zero-valent iron nanoparticles and Fe (II) in human bronchial epithelial cells.Environmental Science & Technology 43, 4555e4560.

Klaine, S.J., Alvarez, P.J.J., Batley, G.E., Fernandes, T.F., Handy, R.D., Lyon, D.,Mahendra, S., McLaughlin, M.J., Lead, J.R., 2008. Nanomaterials in the envi-ronment: fate, behaviour, bioavailability and effects. Environmental Toxicologyand Chemistry 27, 1825e1851.

Kurihara, H., Shirayama, Y., 2004. Effects of increased atmospheric CO2 on seaurchin early development. Marine Ecology Progress Series 274, 161e169.

Kurihara, H., Shimode, S., Shirayama, Y., 2004. Effects of raised CO2 concentration onthe egg production rate and early development of two marine copepods(Acartia steueri and A. erythraea). Marine Pollution Bulletin 49, 721e727.

Kurihara, H., Asai, T., Kato, S., Ishimatsu, A., 2008. Effects of elevated pCO2 on earlydevelopment in the mussel Mytilus galloprovincialis. Aquatic Biology 4,225e233.

Lead, J.R., Muirhead, D., Gibson, C.T., 2005. Characterization of freshwater naturalaquatic colloids by atomic force microscopy (AFM). Environmental Science &Technology 39, 6928e6934, 6921.

Lewinski, N., Colvin, V., Drezek, R., 2008. Cytotoxicity of nanoparticles. Small 4,26e49.

Li, H., Zhou, Q., Wu, Y., Fu, J., Wang, T., Jiang, G., 2009. Effects of waterborne nano-ironon medaka (Oryzias latipes): antioxidant enzymatic activity, lipid peroxidationand histopathology. Ecotoxicology and Environmental Safety 72, 684e692.

Limbach, K.L., Grass, R.N., Strak, W., 2009. Physico-chemical differences betweenparticle- and molecule-derived toxicity: can we make inherently safe nano-particles? Chimia 63, 38e43.

Lindinger, M.I., Lawren, D.J., McDonald, D.G., 1984. Acidebase balance in the seamussel Mytilus edulis. Effects of environmental hypercapnia on intra andextracellular acidebase balance. Marine Biology Letters 5, 371e381.

Makhluf, D.B.S., Qasem, R., Rubinstein, S., Gedanken, A., Breitbart, H., 2006. Loadingmagnetic nanoparticles into sperm cells does not affect their functionality.Langmuir 22, 9480e9482.

Michaelidis, B., Ouzounis, C., Paleras, A., Pörtner, H.O., 2005. Effects of long-termmoderate hypercapnia on acidebase balance and growth rate in marinemussels Mytilus galloprovincialis. Marine Ecology Progress Series 293, 109e118.

Miles, H., Widdicombe, S., Spicer, J.I., Hall-Spencer, J., 2007. Effects of anthropogenicseawater acidification on acidebase balance in the sea urchin Psammechinusmiliaris. Marine Pollution Bulletin 54, 89e96.

Navarro, E., Piccapietra, F., Wagner, B., Marconi, F., Kaegi, R., Odzak, N., Sigg, L.,Behra, R., 2008. Toxicity of silver nanoparticles to Chlamydomonas reinhardtii.Environmental Science & Technology 42, 8959e8964.

Ormerod, B., Angel, M.V., 1996. Ocean storage of carbon dioxide. In: Ormerod, B.,Angel, M.V. (Eds.), Workshop 2: Environmental Impact. International EnergyAgency, Greenhouse Gas R&D Program, Cheltenham, pp. 9e40.

Pierrot, D.E., Lewis, D.E., Wallace, D.W.R., 2006. CO2SYS Dos Program Developedfor CO2 System Calculations. ORNL/CDIAC-105. Carbon Dioxide InformationAnalysis Center, Oak Ridge National Laboratory, US Department of Energy, OakRidge, Tennessee.

Raven, J., Caldeira, K., Elderfield, H., Hoegh-Guldberg, O., Liss, P., Riebesell, U.,Shepherd, J., Turley, C., Watson, A., 2005. Ocean Acidification Due to IncreasingAtmospheric Carbon Dioxide. The Royal Society Policy Document 12/05. TheClyvedon Press Ltd., Cardiff, UK.

Riebesell, U., Zondervan, I., Rost, B., Tortell, P.D., Zeebe, R.E., Morel, F.M.M., 2000.Reduced calcification of marine plankton in response to increased atmosphericCO2. Nature 407, 364e367.

Rohner, F., Ernst, F.O., Arnold, M., Hilbe, M., Biebinger, R., Ehrensperger, F.,Pratsinis, S.E., Langhans, W., Hurrell, R.F., Zimmermann, M.B., 2007. Synthesis,characterization, and bioavailability in rats of ferric phosphate nanoparticles.Journal of Nutrition 137, 614e619.

Schulze, E., Ferrucci, J.T., Poss, K., Lapointe, L., Bogdanova, A., Weissleder, R., 1995.Cellular uptake and trafficking of prototypical magnetic iron-oxide label invitro. Investigative Radiology 30, 604e610.

Seibel, B.A., Fabry, V.J., 2003. Marine biotic response to elevated carbon dioxide.Advances in Applied Biodiversity Science 4, 59e67.

Shan, G.B., Surampalli, R.Y., Tyagi, R.D., Zhang, T.C., 2009. Nanomaterials forenvironmental burden reduction, waste treatment, and nonpoint sourcepollution control: a review. Frontiers of Environmental Science & Engineeringin China 3, 249e264.

Sharp, J.H., 2002. Analytical methods for total DOM pools. In: Hansell, D.A.,Carlson, C.A. (Eds.), Biogeochemistry of Marine Dissolved Organic Matter.Academic Press, San Diego, USA.

Somerfield, P.J., Jaap, W.C., Clarke, K.R., Callahan, M., Hackett, K., Porter, J., Lybolt, M.,Tsokos, C., Yanev, G., 2008. Changes in coral reef communities among theFlorida Keys, 1996e2003. Coral Reefs 27, 951e965.

Spicer, J.I., Raffo, A., Widdicombe, S., 2007. Influence of CO2-related seawateracidification on extracellular acidebase balance in the velvet swimming crabNecora puber. Marine Biology 151, 1117e1125.

Sprung, M., Bayne, B.L., 1984. Some practical aspects of fertilizing the eggs of themussel Mytilus edulis L. Journal du Conseil International pour L’Exploration dela Mer 41, 125e128.

Stern, N., 2006. Review on the Economics of Climate Change. http://www.hm-treasury.gov.uk/sternreview_index.htm.

Sunda, W.G., 2001. Bioavailability and bioaccumulation of iron in the sea. In:Turner, D.R., Hunter, K.A. (Eds.), The Biogeochemistry of Iron in Seawater.IUPAC Series on Analytical and Physical Chemistry. John Wiley & Sons Ltd.

Tipping, E., 1994. WHAMC e a chemical equilibrium model and computer code forwaters, sediments, and soils incorporating a discrete site/electrostatic model ofion-binding by humic substances. Computers & Geosciences 20, 973e1023.

Thurman, E.M., 1985. Organic Geochemistry of Natural Waters. Niyhoff/Junk, Boston.Turner, D.R., Hunter, K.A., 2001. The Biogeochemistry of Iron in Seawater. In:

Series on Analytical and Physical Chemistry of Environmental Systems. WileyBlackwell, London.

Valko, M., Morris, H., Cronin, M.T.D., 2005. Metals, toxicity and oxidative stress.Current Medicinal Chemistry 12, 1161e1208.

Widdicombe, S., Needham, H.R., 2007. Impact of CO2-induced seawater acidificationon the burrowing activity of Nereis virens and sediment nutrient flux. MarineEcology Progress Series 341, 111e122.

Widdicombe, S., Spicer, J.I., 2008. Predicting the impact of ocean acidificationon benthic biodiversity: what can animal physiology tell us? Journal ofExperimental Marine Biology and Ecology 366, 187e197.

Zhou, Y.M., Zhong, C.Y., Kennedy, I.M., Pinkerton, K.E., 2003. Pulmonary responses ofacute exposure to ultrafine iron particles in healthy adult rats. EnvironmentalToxicology 18, 227e235.