Toxicology and Applied Pharmacology182, 197–207 (2002)doi:10.1006/taap.2002.9430

Winfried Moller, Thomas Hofer, Axel Ziesenis, Erwin Karg, and Joachim Heyder

GSF–National Research Center for Environment and Health, Institute for Inhalation Biology, Robert Koch Allee 29, D-82131 Gauting/Munich, Germany

Received November 14, 2001; accepted April 10, 2002

(in-ron-.masslswith

terthe

icles.here-finef theircro-

, be-en innin-

rown;e

dto

ipherytra-finelungerin,e

lation

nsegestrigger92).emo-te ofe in-

intollen

Ultrafine Particles Cause Cytoskeletal Dysfunctions in Macro-phages. Moller, W., Hofer, T., Ziesenis, A., Karg, E., and Heyder,J. (2002). Toxicol. Appl. Pharmacol. 182, 197–207.

Essential cytoskeletal functions of macrophages are migration,phagocytosis of foreign materials, and intracellular transport anddigestion The influence of fine and ultrafine test particles (UFP),such as TiO2, elemental carbon, commercial carbon black, dieselexhaust particulate matter, and urban dust (UrbD), on cytoskel-eton-related functions of macrophages, such as phagocytosis,phagosome transport mechanisms, and mechanical cytoskeletalintegrity, were studied by flow cytometry and by cytomagnetom-etry. Additionally, necrosis and apoptosis caused by the test par-ticles was detected. The diameter of the test particles ranged from12 to 220 nm and the Brunauer–Emmet–Teller specific surfacearea ranged from 6 to 600 m2/g. Primary alveolar macrophagesfrom beagle dogs (BD-AM), obtained by bronchoalveolar lavage,were used as well as macrophages originating from the cell lineJ774A.1. For cytomagnetometry studies, spherical 1.8-�m ferro-magnetic particles served as probes for cytoskeletal functions andwere incubated together with the macrophages 24 h prior to UFPexposure. Macrophages were exposed in vitro with 10–320 �gUFP/ml/106 cells up to 24 h. In all experiments, J774A.1 macro-phages were more sensitive than BD-AM to UFP exposure. Cy-toskeletal dysfunctions evaluated by cytomagnetometry were animpaired phagosome transport and an increased cytoskeletal stiff-ness and occurred at concentrations of 100 �g UFP/ml/106 cellsand above, in both BD-AM and J774A.1. Only fine TiO2 did notshow any effect. Urban dust (standard reference material 1649a)and diesel exhaust particles (DEP, standard reference material1650) caused comparable cytoskeletal dysfunctions to elementalcarbon with high specific surface area. Cytoskeletal dysfunctionsinduced by DEP or UrbD could be reduced after washing theparticles by dichloromethane. UFP caused an impaired phagocy-tosis of 1-�m diameter fluorescent latex beads, inhibited cell pro-liferation, and decreased cell viability. All recorded cytotoxic pa-rameters showed only weak correlations with the specific surfacearea or the total number of UFP, which can result from thedifferent types of particles and different surface compositions.UFP cause cytoskeletal toxicity in vitro in macrophages, which cancause cellular dysfunctions, such as impaired proliferation, im-paired phagocytic activity, and retarded intracellular transportprocesses as well as increased cell stiffness and can result inimpaired defense ability in the lung. © 2002 Elsevier Science (USA)

Key Words: ultrafine particles; diesel exhaust; urban dust; cy-toskeleton; stiffness; viscoelasticity; apoptosis; phagosome trans-port; phagocytosis.

197

Epidemiological studies suggest increased health riskscreased morbidity and mortality) after exposure to envimental particles (Dockeryet al., 1993; Popeet al., 1995)Recent epidemiological studies suggest that, not only theof deposited urban particles (i.e., PM10 or PM2.5� mass of alparticles being smaller than 10 or 2.5�m, respectively) iassociated with increased morbidity and mortality, but,even higher significance, the number of particles (Peterset al.,1997). Because of the vanishing mass of ultrafine (diame�100 nm) compared to fine particles, they contribute little tomass, but greatly to the total number concentration of partMost urban particles result from combustion processes, tfore, the dominant fraction of the urban particles is ultracarbon. Those particles are less soluble and, because osmall size, may not only be phagocytized by alveolar maphages, but can also enter epithelial cells (Takenakaet al.,2000) and can penetrate into the circulation. In additioncause of their high specific surface area, they can catalyzchemical reactions, which may induce chronic inflammatiothe lung. Animal studies andin vitro cell studies have showthat high concentrations of ultrafine particles can induceflammatory processes and increased calcium transients (Bet al., 2000; Johnstonet al., 1996; Li et al., 1997, 1999Oberdoersteret al., 1992a; Stoneet al., 2000), can induce thproduction of oxidative reagents (H2O2, O2

�) (MacNee anDonaldson, 2000; Zhanget al., 1998), or can act as vehiclestransport toxic gasses and substances to the lung per(Johnstonet al., 2000; Niessner and Wilbring, 1989). In intracheal instillation experiments it could be shown that ultraparticles show higher accumulation in the interstitium of theand a slower clearance in comparison to fine particles (F1994; Ferinet al., 1992; Warheitet al., 1997). Because of thvanishing mass of ultrafine compared to fine particles, the reto specific toxicological cellular effects is unclear.

Alveolar macrophages (AM) play a key role in the defereaction within the lungs. They ingest foreign materials, dibacteria and viruses, and present antigens in order to tspecific (immunological) defense mechanisms (Brain, 19Macrophages are resident in each of the alveoli, and chtaxis directs the macrophages within minutes to the siparticle deposition. During phagocytosis the particles arcorporated into a membranous vesicle and are ingestedintracellular phagosomes, which fuse with lysosomes (Tjeetal., 2000). Phagolysosomes are acidic (pH� 5) and contai

Ultrafine Particles Cause Cytoske

tal Dysfunctions in Macrophages le

0041-008X/02 $35.00© 2002 Elsevier Science (USA)

All rights reserved.

enzymes and reactive substances (H2O2) (Mullins and Boni-facino, 2001; Nyberg et al., 1989). In this environment mostbacteria and fungi are digested. Badly digestible particles arerecovered in the lung for longer times (ICRP, 1994). Thecytoskeleton of AM is crucially involved in these defensereactions, including locomotion and cell migration, phagocy-tosis, intracellular transport, phagosome–lysosome fusion, andsignal transduction (Janmey, 1998; Stossel, 1993; Valerius etal., 1982). The cytoskeleton consists of three different filamen-tous structures. The microfilaments (actin) are primarily dis-tributed in the periphery of the cell and are involved in dy-namic processes of the cell, such as crawling and phagocytosis.The microtubuli originate in the microtubuli organizing centernear the nucleus and spread out to the periphery. They areinvolved in cell shape and intracellular vesicle transport. In-termediate filaments contribute to the static part of the cy-toskeleton. Every filamentous structure has its own family ofmotor proteins being responsible for the transport of moleculesand vesicles (Hirokawa, 1998). Those transport processes re-quire energy being provided by ATP (Nemoto et al., 1989;Qian, 2000).

Previously, we have developed protocols to study cytoskel-eton-associated functions in macrophages in vivo and in vitro,such as phagosome transport, mechanical cytoskeletal integ-rity, and phagocytosis, using ferromagnetic microparticles invitro in cell cultures and in vivo after voluntary inhalation(Barth et al., 1994; Moller et al., 1997, 2000, 2001a). Ferro-magnetic microparticles resist intraphagosomal digestion formany weeks. In in vitro studies the magnetic microparticles(1.8-�m diameter) are incubated for 24 h together with culti-vated macrophages, after which more than 95% are ingested.The particles are magnetized and aligned in a short magneticfield pulse. All aligned particles produce a remanent magneticfield (rmf), which can be detected by a magnetic field sensor.Intracellular phagosome transport causes stochastic disorienta-tions of the particles, which results in a decay of the rmf of thecell probe (relaxation). The mechanical viscoelastic propertiesof the cytoskeleton can be investigated by twisting the micro-magnets in a weak magnetic field in a method called magnetictwisting cytometry (MTC). MTC has been used to investigatethe role of the different cytoskeletal structures in macrophagefunction after application of selectively acting cytoskeletaldrugs (Cytochalasin D and Nocodazole; Moller et al., 2000).Additionally, it has been shown that MTC can monitor cyto-toxicity of GaAs-particles in vivo in lung macrophages ofanimals (Aizawa et al., 1993; Okada et al., 1999).

In this study, we set out to test the hypothesis that environ-mental ultrafine particles (UFP) induce toxic reactions on thecytoskeleton of alveolar macrophages, using the MTC tech-nique. Cultivated macrophages, which have ingested magnetictracer particles for 24 h, were incubated with increasingamounts of fine and ultrafine (diameter � 100 nm) test parti-cles. Intracellular phagosome transport (relaxation), stiffness ofthe cytoskeleton, and phagocytosis were recorded after incu-

bation times up to 24 h. Fine and ultrafine test particles havingdifferent chemical compositions and impurities and a widerange of specific surface area (6–600 m2/g) were tested. Thecytoskeletal toxicity of the elemental carbon particles is com-pared to urban dust (UrbD) and to diesel exhaust particles(DEP). Additionally, the ability to phagocytize fluorescentlatex beads was tested after 24 h of incubation with UFP aswell as the ability of UFP to induce apoptosis and necrosis.This allows the correlation of the cytoskeletal dysfunctionswith mass, incubation time, and specific surface area of theparticles. Macrophages originating from permanent cell lines(J774A.1) were used, as well as primary alveolar macrophagesfrom beagle dogs.

MATERIALS AND METHODS

Target cells. Primary alveolar macrophages from beagle dogs (BD-AM)were harvested by bronchoalveolar lavage (BAL). Diluted epithelial liningfluid was obtained from the lungs of anesthetized dogs by segmental bron-choalveolar lavage of three lung lobes using a flexible fiber-optic broncho-scope. Each lung segment was lavaged with 3 � 20 ml sterile, body-warmphosphate-buffered saline (PBS). The BAL fluid was filtered through sterilegauze and centrifuged for 20 min at 400g to separate cells from the superna-tant. The cell pellet was resuspended in PBS. The total cell number wasdetermined in a hemocytometer, and cell viability was measured by trypan blueexclusion. Differential cell numbers were determined by microscopic exami-nation of 600 cells after May–Grunwald–Giemsa staining of air-dried cytospinpreparations. Cells were incubated at 37°C and 5% CO2 in RPMI 1640Medium (Sigma), supplemented with 5% fetal calf serum, 100 U/ml penicillin,100 �g/ml streptomycin, 2.5 �g/ml amphotericin, and 0.3 g/L-glutamine inNaHCO3 buffer.

J774A.1 macrophages originate from a BALB/c/NIH mouse (Ralph andNakoinz, 1975) and were obtained from the German Collection of Animal CellCultures (Tumorbank, DKFZ, Heidelberg, Germany). Cells were incubated inthe same medium as used for primary BD-AM. Cells grew with a doublingtime of �2 days and were divided every 4 days.

Magnetic particle binding assay. Macrophages (0.2 � 106) were incu-bated together with 10 �g of 1.8-�m spherical ferromagnetic microparticles inglass vials (12 mm od) for 24 h prior to adding drugs or ultrafine particles. Thisensures adherence of the macrophages and more than 95% of phagocytosis ofthe magnetic beads. Before the addition of UFP or drugs, nonadherent cells andfree particles were removed by medium exchange. Ferromagnetic micropar-ticles (beads) were prepared with narrow size variation and spherical shape(Moller et al., 1990, 2001b), which is important for data analysis usingmathematical models to estimate the viscous and elastic properties of thecytoskeleton. The particles were not further coated in order not to activatespecific cell surface receptors.

Ultrafine test particles and cytoskeletal drugs. Table 1 gives an overviewof the used test particles together with their physical properties. Because mosturban particles result from combustion processes, the dominant fraction of theurban particles is ultrafine carbon. Elemental ultrafine carbon (EC90) particlesare produced by an electrical spark generator (Roth et al., 1998) understandardized conditions with low impurities. The mobility diameter of thoseparticles was 90 nm and the high specific surface area of 600 m2/g indicatesaggregates having much smaller subunits. The specific surface area, as shownin Table 1, was determined based on N2 gas adsorption measurements accord-ing to the Brunauer–Emmet–Teller method (Brunauer et al., 1938; Gregg andSing, 1982). Alternatively, commercial ultrafine carbon black particles wereused (Printex 90, Printex G, Degussa, Frankfurt, Germany), which haveimpurities of �1% of transition metals. Diesel exhaust particles (standardreference material 1650) and urban dust (standard reference material 1649)

198 MOLLER ET AL.

were used to simulate urban particle exposure (NIST, 1998, 2000). Addition-ally, fine and ultrafine TiO2 particles were used as a second type of materialhaving different chemical composition and surface properties. The fine andultrafine test particles were incubated in vitro with the AMs in concentrationsof 10 to 320 �g/mL of medium/106 macrophages for 1 h up to 1 day. Theparticles were suspended within glass tubes in deionized water by subsequentsonication and then further diluted in medium. In order to investigate theinfluence of the organic components being adsorbed on DEP or UrbD, theparticles were first dispersed in DMSO and then further diluted in medium, inwhich the total concentration of DMSO was kept below 0.1% and which didnot have any influence on the parameters. Additionally, the DEP and UrbDparticles were twice washed in dichloromethane (DClM) and, after removal ofthe solvent, redispersed in medium. Extraction of organic components alsoyields an increase of the specific surface area (Dasenbrock et al., 1996). Thetotal extractable organic carbon is 4.6 � 0.04% in UrbD and 20.2 � 0.4% inDEP, respectively. Microscopic investigation showed that a large fraction ofthe carbon particles appeared as aggregates in the macrophages.

Cytoskeletal drugs were used in order to deliver reference data of cytotoxicreactions on the different filamentous subunits of the cytoskeleton. Cytocha-lasin D (1–5 �M) was used to disrupt microfilaments. It has previously beenshown that Cytochalasin D inhibits phagocytosis of bacteria and particles.Nocodazole (10 �g/ml) or Colchicine (10–50 �M) disrupt microtubuli.Colchicin retards, but does not completely inhibit, the process of phagocytosis.

Magnetic twisting cytometry. For in vitro measurement of relaxation andmechanical cell properties, a magnetic twisting cytometry device was used(Valberg and Butler, 1987). The glass vials (12-mm diameter) containingmacrophages with ingested magnetic particles were positioned in a secondgradiometer array of flux-gate sensors (Forster GmbH, Reutlingen, Germany;Fig. 1). The particles in the cells were aligned parallel to the direction of thesensors by a 200-mT, 10-�s magnetic field pulse. Ferromagnetic particles (10�g) induced a cell field of �1 nT in the sensor array. The probe was rotatedat 6 Hz and the signal of the fluxgate array was amplified and phase sensitivedetected. Particle twisting was performed in a magnetic twisting field (1–2.5mT) perpendicular to the direction of detection (parallel to the axis of rotation).

Stochastic phagosome motion (relaxation). The motion of vesicles andphagosomes happens permanently within living cells and reflects part of theintracellular transport system. After particle alignment in a magnetic fieldpulse, a remanent magnetic field is recorded. Due to the random motion of themagnetic phagosomes, the rmf decays (Fig. 2A). This decay is called relax-ation and happens within �10 min, after which the rmf vanishes, and allmagnetic dipoles are randomly oriented. The phagosomes are coupled to thecytoskeletal filaments by motor proteins and the hydrolization of ATP providesthe energy to move the phagosomes along cytoskeletal filaments. This trans-

latory motion implies random rotational kicks to the phagosomes, resulting ina loss of alignment of the magnetic dipoles and a decay of the rmf. Relaxationis described as a rotational Brownian motion process (Nemoto, 1982; Nemotoand Moller, 2000; Nemoto et al., 1989), which implies an exponential decay ina Newtonian viscosity. Relaxation measurements in living cells show twophases, with an initially fast phase followed by a slow exponential phase. Adouble exponential decay was fitted to the experimental data by a nonlinearregression algorithm (Solver, Microsoft Excel 8.0) and yields a good approx-imation of all relaxation measurements (Moller et al., 2000). From this fit, tworobust relaxation parameters were analyzed, being independent of any model(Fig. 2A). This is the normalized rmf after 1 min, b1 � B(1 min)/B0, whichcharacterizes the initial fast phase of decay, and rmf after 5 min, b5 � B(5min)/B0, which is characteristic for the decay in the slow phase.

Magnetic phagosome twisting. After particle alignment by pulse magne-tization, a weak magnetic field was applied to twist the particles for investi-gation of cytoplasmic rheology and mechanical integrity (Moller et al., 1997;Valberg and Feldman, 1987). Particle rotation in a rheological fluid followseither Newton’s law in cases of viscosity, where the applied shear stress � isproportional to the shear rate (d�/dt), or Hoke’s law in cases of elasticity,where the applied shear stress is proportional to the strain (rotation angle �).The constant of proportionality is either the viscosity � or the elasticitymodulus �, respectively. The shear stress to a magnetic dipole having aremanent magnetization, M, in a magnetic twisting field, BTW , is described by:

� �M � BTW

�(1)

where � is the rotational shape-factor (� � 6 for spheres). The strain wasestimated from the measurement of the cell field B(t) according to �(t) �arccos B(t)/B 0. First the particles were magnetized and aligned parallel to thefield sensors (Fig. 2B). After 20 s relaxation the twisting field BTW was appliedfor 10-s duration and viscoelastic recoil was recorded for 3 min. Cell stiffnesswas estimated as the ratio between mean stress to strain after a constanttwisting duration of 10 s:

Stiffness ���

��(2)

This analysis of stiffness does not view for specific viscous or elastic proper-ties. Therefore, this parameter provides an integral description of the cytoskel-etal mechanics.

Measurement of phagocytosis capacity and cell proliferation. Macro-phages were prepared in suspension in nontreated Geiner petri dishes (GreinerGmbH, Frickenhausen, Germany) together with different types and amounts of

FIG. 1. Schematic view of the magnetic twisting device (MTC) to mea-sure relaxation and twisting of ferromagnetic microparticles ingested by mac-rophages. The macrophages containing 1.8-�m-diameter ferromagnetic parti-cles are prepared in 12-mm-diameter glass vials. After pulse magnetization theremanent magnetic cell field is detected by an array of four magnetic fluxgatesensors (Forster device). The cell probe is rotated at 6 Hz and phase-sensitivesignal amplification achieves a good signal-to-noise ratio.

TABLE 1Fine and Ultrafine Test Particles Used in the Study Together

with Physical Properties

Particle type (material)Diameter

(nm)Specific surface

area (m2/g) Source

fTiO2 220 6 DegussaufTiO2 20 48 DegussaufP-G (carbon black, Printex-G) 51 30 DegussaufP90 (carbon black, Printex90) 12 300 DegussaufEC90 (elemental carbon) 90 600 GSFDiesel exhaust particles (DEP) 120 108 NISTa

Urban dust (UrbD) ? ? NISTb

Note. f, fine; uf, ultrafine.a Standard reference material 1650a (NIST, 2000).b Standard reference material 1649a (NIST, 1998).

199ULTRAFINE PARTICLES CAUSE CYTOSKELETAL TOXICITY

UFP and incubated for 24 h. Fluorescent 1.0-�m diameter latex beads (10 �g;yellow-green, excitation, 505 nm; emission, 515 nm, carboxylate modified;Molecular Probes, Leiden, The Netherlands) were added to 1 ml cell suspen-sion (1 � 106 cells/ml) and further incubated for up to 24 h. This correspondsto 18 particles per macrophage. After 1.5, 4, 8, and 24 h, the probe was mixedand 100 �l of the cell suspension was removed, dispersed in 600 �l PBSbuffer, and analyzed by flow cytometry (Epics XL Flowcytometer, BeckmanCoulter Inc., Fullerton, CA). From the plot of forward scatter against sidescatter, macrophages were extracted (gated) and their mean fluorescenceintensity per cell was analyzed. Cells having only one particle phagocytizedappeared as a single peak in the fluorescence spectrum, which was used forcalibration in order to predict the mean number of phagocytized particles percell. After calibrating the flow of the flow cytometer, the mean cell concen-tration was estimated. This value was comparable to the measurements with acell counter (CASY, Scharfe GmbH, Reutlingen, Germany). Using theJ774A.1 cell line, the measurement of cell concentration allowed the predictionof impairment of cell proliferation after UFP exposure. The total yield ofphagocytosis was estimated by normalizing to a maximum of 18 particles permacrophage and adjusting for the cell concentration. For every type of particleexposure as well as for control probes the mean phagocytic yield was calcu-lated from the four different time points of analysis.

Detection of apoptosis and necrosis. The ability of induction of apoptosisand necrosis in macrophages after incubation with ultrafine particles wasdetected by using the Vybrant apoptosis kit (V-13241, Molecular Probes). Theassay consists of the Annexin V (AnV) antibody, which has a high affinity tothe phosphatidylserine (PS) ligands and which is coupled to the fluophoreAlexa Fluor 488. In viable cells, the PS ligands directly to the cytoplasm (vanEngeland et al., 1998; Vermes et al., 2000). When the apoptotic cell deathprogram is initiated, these ligands are expressed on the extracellular cellmembrane. Additionally, the cells are incubated with probidium iodide (PI). PIlinks to the nucleus of necrotic cells. Using both dyes in combination withtwo-wavelength flow cytometry, the fraction of viable (AnV�PI�), apoptotic(AnV), necrotic (PI), and apoptotic and necrotic cells (AnVPI) wasanalyzed, respectively. Cells being positive for PI (PI) were classified asnecrotic, cells having neither AnV nor PI were classified as viable. Cellsbeing positive for AnV, but negative for PI are expressed as apoptotic(AnVPI�). Macrophages were incubated together with ultrafine particles for24 h in nontreated 30-mm-diameter Greiner petri dishes. Cell suspension (100

�l) was removed and prepared with the fluorescent dyes for analysis by theflow cytometer. Displaying forward scatter against side scatter allows forextracted (gate) AMs. For those cells the fluorescence intensity of AnV wasplotted versus PI and the different cell populations were analyzed in a four-quadrant plot. At least 3000 gated AM were analyzed in every measurement.

Additionally, adherent cell probes were analyzed by fluorescence micros-copy, using an inverted microscope (Axiovert 25, Zeiss GmbH, Jena, Ger-many) together with a mercury-arc lamp, two filter systems (allowing for thedetection of Alexa Fluor 488 and PI dye, respectively), and a digital camerasystem (CoolSnap-pro, Mediacybernetics Inc., Silver Spring, MD). Digitalimages were recorded and processed using the Optimas software package(Mediacybernetics Inc.). For every probe a transmission image was recorded aswell as two fluorescence images using filter systems for either AnV or PI,respectively. The transmission image was used to analyze the projected area ofevery cell as regions of interest (ROI). These ROIs were used for analysis ofthe mean fluorescence intensity of either AnV or PI. All cells of the image werethen classified according to their AnV and PI intensity.

Data analysis. Every set of UFP (drug) measurements was performed onat least five separate probes together with a separate set of control measure-ments. All parameters estimated under the set of drugs were normalized by theset of control measurements. A significant deviation of the normalized param-eters from unity denotes an influence of the UFP (drug). Using a two-tailedStudent’s t test, deviations from unity were analyzed for their level of statisticalsignificance. An influence by the appropriate drug was accepted when the levelof significance was p � 0.05. Pearson’s correlation analysis was performedusing WINSTAT (version 2000.1, Fitch Software, Cambridge, MA).

RESULTS

Influence of UFP on Phagosome Motion and on MagneticParticle Twisting in J774A.1 Macrophages

The results of stochastic intracellular phagosome transportand of cytoskeletal mechanical integrity after 24 h of incuba-tion with fine and ultrafine particles is shown in Fig. 3 forJ774A.1 macrophages. None of the tested particles signifi-

FIG. 2. Measurement of the decay of the remanent magnetic cell field (relaxation, A) resulting from stochastic intracellular phagosome transport processesand of cytoskeletal stiffness (B) by MTC. Relaxation is characterized by the relative decay after 5 min [b5 � B(5 min)/B(t � 0)]. Cytoskeletal stiffness isestimated as the ratio between stress to strain after a 10-s step twist in BTW � 1.2 mT magnetic field, perpendicular to the initial direction of dipole alignment.

200 MOLLER ET AL.

cantly influenced relaxation and stiffness in concentrations of32 �g particles/ml/106 cells and below (data not shown). FineTiO2 particles caused no modulation of phagosome transportability and cytoskeletal stiffness at concentrations up to 320�g/ml in contrast to ultrafine TiO2, for which 320 �g/mlcaused a retarded relaxation (increase of b5) as well as asignificant stiffening of the cytoskeleton. From the group ofcarbon particles, P90 caused no modulation of either phago-some transport or cytoskeletal stiffness in concentrations up to320 �g/ml. Only EC90 particles caused a dose-dependentretardation of relaxation as well as an increase of stiffness.Both DEP and UrbD caused a retardation of relaxation and astiffening of the cytoskeleton of J774A.1 macrophages. Figure3B shows the influence of the different types of preparation ofUrbD and DEP on cytomagnetometric measurement. In bothUrbD and DEP there is no significant influence of a predisper-sion in DMSO prior to dilution in medium, but probes beingpredispersed in DMSO show stronger impairment of relaxation

and increased stiffness by tendency. Washing the particlestwice with DClM prior to dispersion in medium causes areduced impairment of relaxation and less increase in cytoskel-etal stiffness. In all data measured with J774A.1 macrophages,there is no significant correlation between either b5 or stiffnesswith the specific surface area of the particles or with the totalnumber of UFP.

Influence of UFP on Phagosome Motion and on MagneticParticle Twisting in Beagle Dog Alveolar Macrophages

The results of stochastic phagosome motions (relative relax-ation after 5 min, b5) and of cytoskeletal integrity (stiffness) ofbeagle dog alveolar macrophages after 24 h incubation withtest particles are shown in Fig. 4. Test particles composed ofTiO2 do not modulate relaxation and stiffness in concentrationsof 32 �g/ml/106 cells and below. Concentrations of 100 �g/mlof TiO2 and higher caused a moderate retardation of relaxation

FIG. 3. Stochastic phagosome motion (relative decay after 5 min, b5) and mechanical integrity (stiffness) of J774A.1 macrophages after 24 h incubation withdifferent types (A) of fine and ultrafine particles. (B) Influence of predispersion of DEP or UrbD in DMSO or stripping off organic components by washing twicein dichloromethane on relaxation and stiffness (data represent mean values of UFP-exposure probes normalized to control probes � SD; N � 5; *p � 0.05; **p �0.01). f, fine; uf, ultrafine.

201ULTRAFINE PARTICLES CAUSE CYTOSKELETAL TOXICITY

(increase of b5), both with fine and with ultrafine TiO2. Onlyultrafine, but not fine TiO2 showed a concentration-dependentincrease of cytoskeletal stiffness. Incubation of BD-AM withparticles composed of carbon showed a concentration-depen-dent acceleration of relaxation (decrease of b5), which isstrongest using EC90 particles. Microscopic examination ofthe cell probes after MTC measurement showed a loss of celladherence (but no complete detachment), which was strongestafter incubation with 320 �g/ml of EC90 particles. In this case,the requirements of sufficient attachment of the macrophagesto the substrate for MTC data recordings are not fulfilled andtherefore these data should not be interpreted as valid MTCresults. Acceleration of relaxation is stronger for EC90 com-pared to P90 particles. P90 particles did not induce any mod-ulation and EC90 caused a moderate increase in cytoskeletalstiffness. Because of the loss of cell adherence, stiffness after320 �g/ml EC90 incubation shows a large variation and shouldnot be interpreted as an effect. Diesel exhaust particles cause,like the other carbonaceous particles, an acceleration of relax-ation, but, in contrast to the other carbonaceous particles, adecrease in cytoskeletal stiffness. Incubation of BD-AM withall types of FP and UFP shows an acceleration of relaxationwith increasing specific surface area (coefficient of correlation0.72, p � 0.001). For example, relaxation (b5) is acceleratedby 25% using particles with a specific surface area of 600 m2/g(EC90). There is no significant correlation between cytoskel-etal stiffness and specific surface area.

Modulation of Cell Proliferation and of PhagocytosisCapacity by UFP

The ability of macrophages to phagocytize 1 �m carboxy-late-modified fluorescent microspheres is shown in Fig. 5 andin Table 2. Because J774A.1 macrophages are a proliferating

cell line, having a mean doubling time of �2 days, the numberof cells at the moment of sampling has to be taken into account.Table 2 shows that UFP cause a significant inhibition of cellproliferation. ufTiO2 inhibits proliferation more (50% of con-trol) than fTiO2 (90% of control). The pure carbon blackprobes cause a small inhibition of cell proliferation (80–90%of control), for which the effect seems to be stronger in EC90compared to P90. UrbD has the strongest effect on cell prolif-eration (50–60% of control) and does not correlate with thedifferent types of preparation. DEP reduces cell proliferation to75–80% of control, for which the DClM-washed DEP seem tobe less restrictive. Table 2 also shows the influence of 24 hUFP exposure on mean cell volume. If all UFP added to theassay are equally distributed to all cells, a 2% increase of cellvolume should occur, which corresponds to an increase indiameter from 13.1 to 13.3 �m. The data in Table 2 indicate afurther swelling of the cells, which may indicate osmotic stresswithin the cells. Increasing cell diameter correlates with re-tarded relaxation and increased stiffness.

Phagocytosis of 1-�m latex beads by J774A.1 macrophagesis significantly impaired (65% of control) after 24 h incubationwith both fine and ultrafine TiO2 particles. The carbonaceousparticles also cause a significant impairment of phagocytosiscapacity (�60% of control), for which UrbD allows only 50%phagocytosis compared to control probes. In J774A.1 macro-phages impairment of phagocytosis does not correlate with animpairment of cell proliferation. Both fine and ultrafine TiO2

appear not to reduce the ability to phagocytize 1-�m latexbeads by BD-AM. BD-AM show an impaired phagocytosiscapacity after incubation with ultrafine carbon particles. Incu-bation with DEP allows only 20% of phagocytosis compared tountreated BD-AM.

FIG. 5. Phagocytosis (percentage of control) of 1 �m fluorescent carbox-ylate-modified latex microspheres by BD-AM and by J774A.1 macrophagesafter 24 h incubation with fine and ultrafine test particles (100 �g/ml/106 cells.Data represent means � SD; N � 4; *p � 0.05; **p � 0.01). f, fine; uf,ultrafine.

FIG. 4. Stochastic phagosome motion (relative decay after 5 min, b5) andmechanical integrity (stiffness) of BD-AM after 24 h incubation with differenttypes of fine and ultrafine particles (data represent mean values of UFPexposure probes normalized to control probes � SD; N � 5; *p � 0.05; **p �0.01). f, fine; uf, ultrafine.

202 MOLLER ET AL.

Apoptosis and Necrosis Caused by UFP

Flow cytometric analysis of apoptosis and necrosis after 24 hincubation of macrophages with 100 �g UFP/ml/106 cells isshown in Fig. 6. Viability in control cells is 89% in J774A.1macrophages and 94% in BD-AM. In the control probes ofboth cell strains, most nonviable cells are necrotic (PI) andthe fraction of apoptotic cells is vanishing. All particles causea decrease in cell viability, which does not remain under 75%for the desired particle concentration of 100 mg/ml/106 cellsand is more pronounced in J774A.1 macrophages compared toBD-AM. Ultrafine TiO2 particles enhance the fraction of apo-ptotic cells compared to fine TiO2. Cell viability is lower withEC particles compared to the Printex particles in both cellstrains. The fraction of apoptotic cells is increased in all

exposure experiments with carbon particles and is strongestusing laboratory-grade EC particles in both cell strains. InBD-AM there is no difference between the two commercialcarbon particles (P-G and P90), which only differ in theirspecific surface area. Cell viability after DEP or UrbD expo-sure is comparable to commercial carbon particles (Printex).Both for TiO2 and for the carbon particles, cell viability seemsto decrease with increasing specific surface area.

Measurements of cell viability by fluorescence microscopyis shown in Fig. 7. In this case, the cells adhere to the bottomof the glass bottles. The fraction of viable cells is higher forcontrol cells and for all UFP types compared to the flowcytometric analysis. The correlation between cell viability andthe specific particle types is comparable to the flow cytometric

TABLE 2Effect of 100 �g UFP/ml/106 J774A.1 Macrophages after 24–48 h of Incubation

Exposure PreparationCell diameter

(�m)

Proliferation (% control)Phagocytic yield

(% control)Viability

(%)CASY FACS

Control Medium 13.1 100 100 100 95fTiO2 Medium 13.7 90 78 62 � 5 94ufTiO2 Medium 14.7 58 53 66 � 7 85ufEC90 Medium 14.1 90 81 61 � 25 84ufP90 Medium 14.1 89 88 55 � 16 92ufP90 DMSO/medium 13.9 100 97 88 � 7 91UrbD Medium 14.4 55 55 72 � 22 93UrbD DCIM/medium 14.7 51 46 62 � 17 90UrbD DMSO/medium 13.7 56 56 57 � 7 92DEP Medium 15.4 78 75 50 � 18 92DEP DCIM/medium 14.3 78 88 48 � 10 94

Note. UFP were suspended in medium, suspended in DMSO and further diluted in medium, or washed twice in dichloromethane (DCIM) and afterwardsuspended in medium.

FIG. 6. Measurement of apoptosis and necrosis in J774A.1 macrophagesand in BD-AM by flow cytometry after 24 h of incubation with different typesof fine and ultrafine particles in concentrations of 100 �g/ml/106 cells. f, fine;uf, ultrafine.

FIG. 7. Measurement of apoptosis and necrosis in BD-AM by fluores-cence microscopy after 24 h of incubation with different types of fine andultrafine particles in concentrations of 100 or 320 �g UFP/ml/106 AM. F. fine;uf, ultrafine.

203ULTRAFINE PARTICLES CAUSE CYTOSKELETAL TOXICITY

analysis. Additionally, cells were incubated with 320 �g UFP/ml/106 cells, which shows a dose-dependent decrease of cellviability being caused by an increase of necrotic cells. In thisassay DEP appear more toxic than the other carbon particles.

DISCUSSION

Influence of UFP on Phagosome Motion and on MagneticParticle Twisting

MTC studies investigate the transport of micron-sizedphagosomes from two different views. Relaxation directlymonitors the coupling dynamics between phagosomes and thecytoskeleton while phagosome twisting views the mechanicalintegrity of the cytoskeletal filaments, which are linked to thephagosomes. Phagosome transport requires an intact cytoskel-eton and intact phagosomes including motor proteins and en-ergy (ATP) (Nemoto et al., 1989; Wang and Ingber, 1994).Additionally, intracellular calcium plays an important role incytoskeletal functions and intracellular signaling (Glogauer etal., 1998; Hartwig and Janmey, 1989). Cytotoxic reactions caninvolve the energy metabolism of the cell, intracellular calciumtransients (Stone et al., 2000), the dynamics of cytoskeletalfilaments, and the behavior of phagosomes (Aizawa et al.,1993; Okada et al., 1999). Intracellular calcium transientsarising after exposure to carbon black happened in conjunctionwith the release of oxidative metabolites (Grundler et al., 2000;Stone et al., 2000). In a previous study it was shown thatdestroying microfilaments by cytochalasin D resulted in aretarded relaxation and an increased stiffness (Moller et al.,2000). Colchicine, which disrupts microtubuli, resulted in anaccelerated relaxation and in a moderately increased stiffness.Figure 8 summarizes the MTC measurements of J774A.1 mac-rophages after 24 h of incubation with 320 �g/ml UFP incomparison to the influence of cytoskeletal drugs. The studiesusing high concentrations of ufTiO2 and EC90 particles indi-cate cytoskeletal toxicity, being primarily related to microfila-ment dysfunctions. Accelerated relaxation in BD-AM afterexposure to P90 particles indicates microtubuli-related dys-functions.

Modulations in cytoskeletal functions evaluated by MTC areseen at particle concentrations of 100 �g/ml/106 cells and forparticles having a specific surface area of 48 m2/g (ufTiO2) andabove. This is a high concentration compared to ambientconditions (20–50 �g/m3) (Pitz et al., 2001), but, during long-term inhalation and because of the chemical stability of carbon,high amounts may accumulate in the lungs. For example,cigarette smoke contains primarily tar particles, which can beseen in high concentrations in AMs obtained by BAL fromchronic cigarette smokers. As mentioned more than 100 �gUFP distributed over one million macrophages corresponds toa volume increase of about 2%, which is far below overloadconditions (Oberdoerster et al., 1992b). Although the totalamount of ambient particulate matter decreased in the eastern

part of Germany after reunification, the number of ultrafineparticles increased (Ebelt et al., 2001; Pitz et al., 2001), whichmay explain increasing incidences of respiratory effects (bron-chial hyperresponsiveness, bronchitis, and asthma) (Heinrich etal., 1999; Richter et al., 2000).

Modulation of Phagocytosis Capacity by UFP

Phagocytosis is an essential cytoskeleton-dependent defenseprocess of macrophages. Assessment of phagocytic capacityafter 24 h of incubation with 100 �g UFP/ml/106 cells showssignificant differences between the two cell types. Phagocyto-sis is much more strongly inhibited in J774A.1 macrophagescompared to BD-AM. Additionally, there is a material-specificdifference in phagocytic capacity, but no correlation with thespecific surface area of the particles. Only J774A.1 macro-phages showed reduced phagocytotic capacity of EC90 com-pared to P90 particles, which correlates with the higher specificsurface area. It is surprising that there is no correlation betweenphagocytosis and the cytomagnetometric data (relaxation andcytoskeletal stiffness), because both parameters represent cy-toskeletal functions. One reason for this missing correlationmay be the different surface compositions of the differentparticle types, which may influence differently signal transduc-tion pathways and cell surface receptors that are involved in thephagocytic process.

It is surprising that 100 �g/ml TiO2 particles does notmodulate phagocytosis in BD-AM, because other authorscould show that only 1 �g/ml/106 AM of aggregates of ultra-fine carbon could modulate phagocytosis (Lundborg et al.,1999). The influence of different types of fine and ultrafinecarbon and TiO2 particles on phagocytosis was also investi-gated by Renwick et al. (2001), using J774.2 macrophages.

FIG. 8. Modulation of relaxation and cytoskeletal stiffness after 24 hincubation with 320 �g/ml fine and ultrafine particles in J774A.1 macrophagesin comparison to cytoskeletal drugs. Cytochalasin D (4 �M) disrupts micro-filaments and Colchicine (10 mM) disrupts microtubuli (data represent meanvalues of UFP-exposure probes or cytoskeletal drugs normalized to controlprobes � SD; N � 5; *p � 0.05; **p � 0.01). CyD, Cytochalasin; Col,Colchicine.

204 MOLLER ET AL.

Particle concentrations of 0.39 �g/mm2 and above significantlyinhibited phagocytosis of fluorescent indicator beads, whichcorresponds to about 100 �g UFP/ml/106 cells in the presentstudy. They determined that the inhibition of phagocytosis wasstronger for ultrafine compared to fine particles. These resultswere not verified in our studies using TiO2 particles.

Diesel exhaust material showed the strongest inhibitory effecton phagocytic ability, both in J774A.1 and in BD-AM, and thedifferent steps of preparation (dilution or removing of organiccomponents) did not significantly influence the results. Urban dustinhibited phagocytic capacity less than DEP, but cell proliferationwas reduced to 55% of control, and the different steps of surfacepreparation also did not influence the results. This indicates thatcytotoxicity caused by organic substances being adsorbed on theparticle surface are of secondary relevance for the induction ofcytoskeletal dysfunctions and that the mechanisms behind thesedysfunctions are still unclear.

Apoptosis and Necrosis Caused by UFP

The decrease in cell viability after 24 h incubation with 100�g/ml/106 cells is only marginal and does not go below 70%for all types of used particles, although loss of cell viabilitycannot explain the effects obtained on cytoskeletal dysfunc-tion. Cell viability evaluated by flow cytometry is higher forBD-AM compared to J774A.1 macrophages, for control probesas well as for the particle exposure probes. This shows thatprimary cells are more resistive to incubation handling andexposure conditions. Nonetheless, both cell types show thesame type of response to UFP, with a decrease of cell viabilityin relation to increasing specific surface area. Ultrafine TiO2

are comparable to P90 in the decrease of cell viability but havea much smaller specific surface area. This demostrates anadditional material or particle surface effect (hydrophobic orhydrophilic), which may cause differences between TiO2 andcarbon particles in their ability to form aggregates. Micro-scopic images showed more and bigger aggregates of carbonparticles compared to TiO2 particles. Most TiO2 particles ap-peared as single particles and therefore may follow differentcytotoxic routes in the cells. We could not discriminate be-tween particles not within phagolysosomes from those free inthe cytoplasm of the macrophages. The dominant fraction ofnonviable cells is necrotic, which may result from the longincubation time of 24 h, when apoptotic cells move into thenecrotic fraction and/or become phagocytized by viable cells.

Fluorescence microscopy evaluated higher fractions of via-ble cells under comparable incubation conditions. In theseassays the cells were adherent and this may significantly in-fluence the behavior and resistance of the cells. Additionally,there was no further preparation step for microscopic comparedto flow cytometric analysis, which also may improve cellviability.

DEP and UrbD showed no additional loss in cell viabilitycompared to carbon particles at concentrations of 100 �g/ml.

Only at higher concentrations did the fraction of viable cellsdecrease dramatically. This may result from the fraction of cyto-toxic substances being adsorbed on the surface of those particles.

CONCLUSIONS

Cytotoxicity caused by high concentrations of UFP is related tocytoskeletal dysfunctions, such as retarded intracellular transportprocesses, increased cell stiffness, impaired phagocytic ability,and impaired cell proliferation. The impaired phagocytosis offluorescent test particles caused by UFP exposure does not cor-relate with retarded relaxation and increased cell stiffness. Thereis only a weak or no correlation of the cytotoxic parameters withspecific surface area or with the total number of UFP. UFP causea surface area-dependent increase in macrophage apoptosis andnecrosis. Cytoskeletal cytotoxicity reported in this study has mul-tiple points of action in the cells and can result in impairedmacrophage migration, impaired phagocytosis, and impaired de-fense ability. All of these dysfunctions can modulate the inductionof chronic inflammatory processes in the lung, which have beendescribed in animal studies.

ACKNOWLEDGMENTS

We thank Dr. Ben Fabry for MTC device setup and helpful discussions. Thiswork was supported by the CEC under Contract FIGD-CT-2000–00053.

REFERENCES

Aizawa, Y., Takata, T., Karube, H., Tatsumi, H., Inokuchi, N., Kotani, M., andChiyotani, K. (1993). Magnetometric evaluation of the effects of galliumarsenide on the clearance and relaxation of iron particles. Ind. Health 31,143–153.

Barth, W., Moller, W., Pohlit, W., Stahlhofen, W., and Wiegand, J. (1994).Magnetopneumographic estimation of particle phagocytosis in the humanlungs. J. Aerosol Sci. 25, S491–S492.

Brain, J. D. (1992). Mechanisms, measurement, and significance of lungmacrophage function. Environ. Health Perspect. 97, 5–10.

Brown, D. M., Stone, V., Findlay, P., MacNee, W., and Donaldson, K. (2000).Increased inflammation and intracellular calcium caused by ultrafine carbonblack is independent of transition metals or other soluble components.Occup. Environ. Med. 57, 685–691.

Brunauer, S., Emmett, P. H., and Teller, E. (1938). Adsorption of gases inmultimolecular layers. J. Am. Chem. Soc. 60, 309–319.

Dasenbrock, C., Peters, L., Creutzenberg, O., and Heinrich, U. (1996). Thecarcinogenic potency of carbon particles with and without PAH after re-peated intratracheal administration in the rat. Toxicol. Lett. 88, 1–3.

Dockery, D. W., Pope III, C. A., Xu, X., Spengler, J. D., Ware, J. H., Fay,M. E., Ferris, B. G., and Speizer, F. E. (1993). An association between airpollution and mortality in six U.S. cities. N. Engl. J. Med. 329, 1753–1759.

Ebelt, S., Brauer, M., Cyrys, J., Tuch, T., Kreyling, W. G., Wichmann, H.-E.,and Heinrich, J. (2001). Air quality in postunification Erfurt, East Germany:Associating changes in pollutant concentrations with changes in emissions.Environ. Health Perspect. 109, 325–333.

Ferin, J. (1994). Pulmonary retention and clearance of particles. Toxicol. Lett.72, 121–125.

Ferin, J., Oberdoerster, G., and Penney, D. P. (1992). Pulmonary retention ofultrafine and fine particles in rats. Am. J. Respir. Cell Mol. Biol. 6, 535–542.

205ULTRAFINE PARTICLES CAUSE CYTOSKELETAL TOXICITY

Glogauer, M., Arora, P., Chou, D., Janmey, P. A., Downey, G. P., andMcCulloch, C. A. (1998). The role of actin-binding protein 280 in integrin-dependent mechanoprotection. J. Biol. Chem. 273, 1689–1698.

Gregg, S. J., and Sing, K. S. W. (1982). Adsorption, Surface Area, andPorosity. Academic Press, London, New York.

Grundler, W., Dirscherl, P., Beck-Speier, I., Beisker, W., Stampfl, A., Zim-mermann, I., and Maier, K. (2000). Simultaneous recording of calciumtransients and reactive oxygen intermediates of human polymorphonucleargranulocytes in response to formyl-Met-Leu-Phe and the environmentalagent sulfite. Cytometry 40, 219–229.

Hartwig, J. H., and Janmey, P. A. (1989). Stimulation of a calcium-dependentactin nucleation activity by phorbol 12-myristate 13-acetate in rabbit mac-rophage cytoskeletons. Biochim. Biophys. Acta 1, 64–71.

Heinrich, J., Hoelscher, B., Jacob, B., Wjst, M., and Wichmann, H. E. (1999).Trends in allergies among children in a region of former East Germanybetween 1992–1993 and 1995–1996. Eur. J. Med. Res. 4, 107–113.

Hirokawa, N. (1998). Kinesin and dynein superfamily proteins and the mech-anism of organelle transport. Science 279, 519–526.

ICRP (1994). Human Respiratory Tract Model for Radiological Protection. Areport of a Task Group of the International Commission on RadiologicalProtection. Ann. ICRP 24, 1–482.

Janmey, P. A. (1998). The cytoskeleton and cell signaling: Component local-ization and mechanical coupling. Physiol. Rev. 78, 763–781.

Johnston, C. J., Finkelstein, J. N., Gelein, R., Baggs, R., and Oberdoerster, G.(1996). Characterization of the early pulmonary inflammatory response associ-ated with PTFE fume exposure. Toxicol. Appl. Pharmacol. 140, 154–163.

Johnston, C. J., Finkelstein, J. N., Mercer, P., Corson, N., Gelein, R., andOberdoerster, G. (2000). Pulmonary effects induced by ultrafine PTFEparticles. Toxicol. Appl. Pharmacol. 168, 208–215.

Li, X. Y., Brown, D., Smith, S., MacNee, W., and Donaldson, K. (1999).Short-term inflammatory responses following intratracheal instillation offine and ultrafine carbon black in rats. Inhal. Toxicol. 11, 709–731.

Li, X. Y., Gilmour, P. S., Donaldson, K., and MacNee, W. (1997). In vivo andin vitro proinflammatory effects of particulate air pollution (PM10). Envi-ron. Health Perspect. 105(Suppl. 5), 1279–1283.

Lundborg, M., Johansson, A., Lastbom, L., and Camner, P. (1999). Ingestedaggregates of ultrafine carbon particles and interferon-gamma impair ratalveolar macrophage function. Environ. Res. 81, 309–315.

MacNee, W., and Donaldson, K. (2000). How can ultrafine particles beresponsible for increased mortality? Monaldi Arch. Chest Dis. 55, 135–139.

Moller, W., Kreyling, W. G., Kohlhaufl, M., Haussinger, K., and Heyder, J.(2001a). Macrophage functions measured by magnetic microparticles invivo and in vitro. J. Magn. Magn. Mater. 225, 218–225.

Moller, W., Nemoto, I., Matsuzaki, T., Hofer, T., and Heyder, J. (2000).Magnetic phagosome motion in J774A.1 macrophages: Influence of cy-toskeletal drugs. Biophys. J. 79, 720–730.

Moller, W., Roth, C., and Stahlhofen, W. (1990). Improved spinning topaerosol-generator for the production of high concentrated ferrimagneticaerosols. J. Aerosol Sci. 21, S657–S660.

Moller, W., Scheuch, G., Sommerer, K., and Heyder, J. (2001b). Preparationof spherical monodisperse ferrimagnetic iron-oxide microparticles between1 and 5 �m diameter. J. Magn. Magn. Mater. 225, 8–16.

Moller, W., Takenaka, S., Rust, M., Stahlhofen, W., and Heyder, J. (1997).Probing mechanical properties of living cells by magnetopneumography. J.Aerosol Med. 10, 173–186.

Mullins, C., and Bonifacino, J. S. (2001). The molecular machinery forlysosome biogenesis. Bioessays 23, 333–343.

Nemoto, I. (1982). A model of magnetization and relaxation of ferrimagneticparticles in the lung. IEEE Trans. Biomed. Eng. 29, 745–752.

Nemoto, I., and Moller, W. (2000). A viscoelastic model of phagosome motion

within cells based on cytomagnetometric measurements. IEEE Trans.Biomed. Eng. 47, 170–182.

Nemoto, I., Ogura, K., and Toyotama, H. (1989). Estimation of the energy ofcytoplasmic movements by magnetometry: Effects of temperature and in-tracellular concentration of ATP. IEEE Trans. Biomed. Eng. 36, 598–607.

Niessner, R., and Wilbring, P. (1989). Ultrafine particles as trace catchers forpolycyclic aromatic hydrocarbons: The photoelectric aerosol sensor as a toolfor in situ sorption and desorption studies. Anal. Chem. 61, 708–714.

NIST (1998). Standard Reference Material 1649a, Urban Dust, Certificate ofAnalysis, pp. 1–4. National Institute of Standards and Technology (NIST),Gaithersburg, MD.

NIST (2000). Standard Reference Material. 1650a, Diesel Particulate Matter,Certificate of Analysis, pp. 1–7. National Institute of Standards and Tech-nology (NIST), Gaithersburg, MD.

Nyberg, K., Johansson, A., and Camner, P. (1989). Intraphagosomal pH inalveolar macrophages studied with fluorescein-labeled amorphous silicaparticles. Exp. Lung Res. 15, 49–62.

Oberdoerster, G., Ferin, J., Gelein, R., Soderholm, S. C., and Finkelstein, J.(1992a). Role of the alveolar macrophage in lung injury: Studies withultrafine particles. Environ. Health Perspect. 97, 193–199.

Oberdoerster, G., Ferin, J., and Morrow, P. E. (1992b). Volumetric loading ofalveolar macrophages (AM): A possible basis for diminished AM-mediatedparticle clearance. Exp. Lung Res. 18, 87–104.

Okada, M., Karube, H., Niitsuya, M., Aizawa, Y., Okayasu, I., and Kotani, M.(1999). In vitro toxicity of gallium arsenide in alveolar macrophages eval-uated by magnetometry, cytochemistry and morphology. Tohoku J. Exp.Med. 189, 267–281.

Peters, A., Wichmann, H. E., Tuch, T., Heinrich, J., and Heyder, J. (1997).Respiratory effects are associated with the number of ultrafine particles.Am. J. Respir. Crit. Care Med. 155, 1376–1383.

Pitz, M., Kreyling, W. G., Holscher, B., Cyrys, J., Wichmann, H. E., andHeinrich, J. (2001). Change of the ambient particle size distribution in EastGermany between 1993 and 1999. Atmos. Environ. 35, 4357–4366.

Pope, C. A., Thun, M. J., Namboodiri, M. M., Dockery, D. W., Evans, J. S.,Speizer, F. E., and Heath, C. W. (1995). Particulate air pollution as apredictor of mortality in a prospective study of U.S. adults. Am. J. Respir.Crit. Care Med. 151, 669–674.

Qian, H. (2000). A simple theory of motor protein kinetics and energetics.Biophys. Chem. 83, 35–43.

Ralph, P., and Nakoinz, I. (1975). Phagocytosis and cytolysis by a macrophagetumour and its cloned cell line. Nature 257, 393–394.

Renwick, L. C., Donaldson, K., and Clouter, A. (2001). Impairment of alveolarmacrophage phagocytosis by ultrafine particles. Toxicol. Appl. Pharmacol.172, 119–127.

Richter, K., Heinrich, J., Jorres, R. A., Magnussen, H., and Wichmann, H. E.(2000). Trends in bronchial hyperresponsiveness, respiratory symptoms andlung function among adults: West and East Germany. INGA Study Group.Indoor Factors and Genetics in Asthma. Respir. Med. 94, 668–677.

Roth, C., Karg, E., and Heyder, J. (1998). Do inhaled ultrafine particles causeacute health effects in rats? I. Particle production. J. Aerosol Sci. 29,S679–S680.

Stone, V., Tuinman, M., Vamvakopoulos, J. E., Shaw, J., Brown, D., Petterson,S., Faux, S. P., Borm, P., MacNee, W., Michaelangeli, F., and Donaldson,K. (2000). Increased calcium influx in a monocytic cell line on exposure toultrafine carbon black. Eur. Respir. J. 15, 297–303.

Stossel, T. P. (1993). On the crawling of animal cells. Science 260, 1086–1094.

Takenaka, S., Karg, E., Moller, W., Roth, C., Ziesenis, A., Heinzmann, U.,Schramel, P., and Heyder, J. (2000). A morphologic study on the fate of

206 MOLLER ET AL.

ultrafine silver particles: Distribution pattern of phagocytized metallic silverin vitro and in vivo. Inhal. Toxicol. 12, 291–299.

Tjelle, T. E., Lovdal, T., and Berg, T. (2000). Phagosome dynamics andfunction. Bioessays 22, 255–263.

Valberg, P. A., and Butler, J. P. (1987). Magnetic particle motions withinliving cells: Physical theory and techniques. Biophys. J. 52, 537–550.

Valberg, P. A., and Feldman, H. A. (1987). Magnetic particle motions withinliving cells: Measurement of cytoplasmic viscosity and motile activity.Biophys. J. 52, 551–561.

Valerius, N. H., Stendahl, O. I., Hartwig, J. H., and Stossel, T. P. (1982).Distribution of actin-binding protein and myosin in neutrophils duringchemotaxis and phagocytosis. Adv. Exp. Med. Biol. 141, 19–28.

van Engeland, M., Nieland, L. J., Ramaekers, F. C., Schutte, B., and Reutel-ingsperger, C. P. (1998). Annexin V-affinity assay: A review on an apopto-

sis detection system based on phosphatidylserine exposure. Cytometry 31,1–9.

Vermes, I., Haanen, C., and Reutelingsperger, C. (2000). Flow cytometry ofapoptotic cell death. J. Immunol. Methods 243, 1–2.

Wang, N., and Ingber, D. E. (1994). Control of cytoskeletal mechanics byextracellular matrix, cell shape, and mechanical tension. Biophys. J. 66,2181–2189.

Warheit, D. B., Hansen, J. F., Yuen, I. S., Kelly, D. P., Snajdr, S. I., andHartsky, M. A. (1997). Inhalation of high concentrations of low toxicitydusts in rats results in impaired pulmonary clearance mechanisms andpersistent inflammation. Toxicol. Appl. Pharmacol. 145, 10–22.

Zhang, Q., Kusaka, Y., Sato, K., Nakakuki, K., Kohyama, N., and Donaldson,K. (1998). Differences in the extent of inflammation caused by intratrachealexposure to three ultrafine metals: Role of free radicals. J. Toxicol. Environ.Health 53, 423–438.

207ULTRAFINE PARTICLES CAUSE CYTOSKELETAL TOXICITY