Visualizing the Uptake of C60 to the Cytoplasm and Nucleus of


Feb 28, 2007 - part b regions of lighter contrast are marked on the surface of the crystals which suggest the presence of surface plasmons (p). The hi...
0 downloads 0 Views 713KB Size


Environ. Sci. Technol. 2007, 41, 3012-3017

Visualizing the Uptake of C60 to the Cytoplasm and Nucleus of Human Monocyte-Derived Macrophage Cells Using Energy-Filtered Transmission Electron Microscopy and Electron Tomography A L E X A N D R A E . P O R T E R , * ,† MHAIRI GASS,‡ KARIN MULLER,§ JEREMY N. SKEPPER,§ PAUL MIDGLEY,| AND MARK WELLAND† The Nanoscience Centre, University of Cambridge, 11 JJ Thompson Avenue, Cambridge CB3 OFF, U.K., UK SuperSTEM, Daresbury Laboratory, Daresbury, Cheshire WA4 4AD, U.K., Multi-imaging Centre, Department of Physiology, Development and Neuroscience, Anatomy Building, University of Cambridge, Downing Street, Cambridge CB2 3DY, U.K., and Department of Materials Science and Metallurgy, University of Cambridge, Pembroke Street, Cambridge CB2 3QZ, U.K.

Concerns have been raised over the release of C60 nanoparticles into the environment and the potential risk to human health. To address these concerns it is essential to understand the pathways by which nanoparticles enter the cell, where they migrate to, and to establish whether the particles are transformed or modified within the cell. Imaging the subcellular distribution of carbon-based nanoparticles is particularly challenging. It is difficult to achieve high spatial resolution with sufficient image contrast to enable the nanoparticles to be identified within the cell. We have exposed human monocyte-derived macrophages (HMMs) to C60 and used energy filtered transmission electron microscopy (EFTEM) to image the distribution of C60 aggregates within intracellular compartments. We demonstrate that images recorded using low-loss electrons provide a significant improvement in contrast between the cellular material and the C60 allowing a clear differentiation between C60 and unstained cellular compartments and also between ordered and disordered forms of aggregated C60. We confirm that C60 is taken up by HMMs in vitro and is sequestered at several sites within the cell. These sites include the cytoplasm, lysosomes, and most significantly the cell nuclei.

Introduction Carbon-based nanoparticles, such as Buckminster fullerene (C60), have attracted a great deal of scientific interest because of their unique chemical and physical properties. However, * Corresponding author e-mail: [email protected] † The Nanoscience Centre, University of Cambridge. ‡ UK SuperSTEM. § Department of Physiology, Development and Neuroscience, University of Cambridge. | Department of Materials Science and Metallurgy, University of Cambridge. 3012

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 8, 2007

recent studies have raised concerns about their potential toxicity, with the release of C60 into the environment proposed to be a potential long-term risk to human health, with ecological implications (1-3). The toxicity of C60 is of particular toxic concern, because the individual C60 molecules have a diameter of only 0.7 nm and are therefore small enough to potentially enter the cell via passive diffusion through ion channels and to be able to diffuse through pores in the nuclear membrane. It is, therefore, essential to determine the degree of toxicity of C60 before its use becomes widespread. Imaging the cellular distribution of carbon-based nanoparticles is challenging because of the difficulty in achieving high spatial resolution with sufficient image contrast that enables the nanoparticles to be distinguished clearly from the organelles of the cell. One technique available for acquiring high spatial resolution carbon elemental maps (4-6) is energy-filtered transmission electron microscopy (EFTEM). As electrons pass through the material they scatter elastically and lose no energy (zero-loss electrons) or scatter inelastically. If the latter, the vast majority of the energy loss is imparted to weakly bound electrons generating low-loss excitations, especially “plasmons”. In addition, some energy is lost exciting electrons from the core levels into the vacuum known as core-loss excitations. With EFTEM, it is possible to acquire two-dimensional, elemental maps using electrons with an energy loss characteristic of a core level, interband transition, or plasmon resonance energy. Specifically, in the case of core-loss EFTEM carbon mapping, fine structure is seen at the onset of the K-shell carbon ionization edge (7), and this offers a means to map similar but different carbon species. In practice the high electron dose needed for this type of mapping will lead to rapid degradation of the C60 and amorphization of any fullerene crystals. Other techniques, such as scanning transmission electron microscopy (STEM) high annular angular dark-field (HAADF), Z-contrast, imaging (8), are available, but these are unlikely to succeed because of the lack of contrast between the composite components of interest. Therefore an alternative, relatively low-dose, technique is needed that can distinguish between the carbonaceous components in the sample. The low-loss region of the energy-loss spectrum contains information about the energy distribution of the electrons transmitted though a thin film in the energy range of ∼0-75 eV. The low-loss region of the electron energy loss spectrum (EELS) has two advantages for imaging carbonaceous species: (i) the scattering cross-sections are many orders of magnitude larger than for core-loss scattering, and thus acquisition times and beam damage can be minimized, and (ii) the (π + σ) volume plasmon excitation which dominates the low-loss spectrum is sensitive to the electronic structure of the carbon species (9). In the low-loss spectrum, fullerene clusters exhibit a bulk plasmon energy, arising from oscillations of the σ and π electrons, of around 28 eV, whereas amorphous carbon exhibits a plasmon energy of ∼23 eV (9). Graphitic-like carbonaceous materials, such as C60 and carbon nanotubes, also exhibit an interband, πfπ*, transition at 6 eV. This πfπ* transition results from the excitation of electrons in the π bond and is therefore very weak or missing in the amorphous material. We hypothesize that differences in plasmon energies between the carbon rich cellular compartments and the fullerenes and the occurrence of the πfπ* transition at 6 eV will enable us to differentiate between these very similar carbonaceous materials. Macrophages are sentinel cells, found in all tissues, which perform a wide range of functions associated with defense of the organism against microbial and foreign body invasion 10.1021/es062541f CCC: $37.00

 2007 American Chemical Society Published on Web 02/28/2007

(10, 11). Macrophages play an important role in clearing particles from the lung surface via energy-dependent phagocytosis and in orchestrating immune and inflammatory responses. Macrophages have been widely studied to address their role in particle-related lung disease (10, 11), but little is known about the fate of C60 inside macrophages. C60 molecules are hydrophobic and are only 0.7 nm in diameter. We hypothesize that the individual molecules may be able to enter the cell via passive diffusion across lipid bilayers as well as by fluid phase transport and by active processes. The first objective of this work is to monitor the exposure of human monocyte-derived macrophage (HMM) cells to C60 and evaluate the use of EFTEM techniques for imaging the distribution of C60 aggregates within intracellular compartments. The second objective will be use an energy-filter at both the zero-loss and plasmon energies, to study the distribution of C60 within the cell. Three-dimensional (3-D) electron tomography is a wellestablished technique in the biological sciences for determining the 3-D structure of organelles and the structure of macromolecular assemblies. During image acquisition, a specimen is rotated around a single axis and a micrograph is recorded at a range of tilts angles, typically every 1-2° for >100 degrees. A back-projection algorithm is then used to compute the 3-D reconstruction of the sample. In 2-D micrographs it is often difficult to determine whether the C60 is present on the surface of the resin section, as an artifact from sample preparation, or whether the C60 is distributed within the 3-D volume of the section. The final objective of this study will be to show that the EFTEM images are suitable for tomographic reconstruction and that the position of the C60 aggregates within the cell can be identified unambiguously.

Materials and Methods C60 (0.01 g) (99.9% pure, Aldrich, Dorset, U.K.) was dispersed in 10 mL of freshly distilled, filtered, and degassed tetrahydrofuran (THF) (1 g/L). The solution was left in a vial with a screw cap for 24 h, stirred at ambient temperature, and allowed to become saturated with soluble C60. For TEM observations of C60 in THF, lacy carbon 300 mesh copper grids (Agar Scientific, Dorset, U.K.) were immersed in stirred solutions of C60 in THF. Full characterization of the C60 in THF was carried out in a previous study, and full details of this characterization are available elsewhere (12). In summary, single and polycrystals of varying shapes and sizes were observed when C60 was dispersed in THF (12). Crystals of C60 had a hexagonal close-packed structure in THF solution (12). The diameters of the individual particles were 60-270 nm, and the diameters of the clusters were 420-1300 nm (12). The detailed method used for the in vitro cell culture study and electron microscopy has been published previously (12). In summary, mature human macrophages were obtained by in vitro culture of human monocytes isolated from human buffy coat residues (National Blood Transfusion Service, Brentwood, U.K.). Buffy coat residue was washed once with phosphate buffered saline (PBS), and the resulting cell sediment was mixed with an equal volume of fresh PBS. Thirty milliliters of diluted buffy coat residue was layered onto 15 mL of LymphoPrep (Axis-Shields, Oslo, Norway), and, after centrifugation at 20 °C for 30 min at 600g, the opaque interphase of mononuclear cells was removed and washed three times with PBS containing 4 mg/mL bovine serum albumin (BSA) to remove platelets. Monocytes were then enriched by an additional centrifugation step in a Percoll gradient. Mononuclear cells were resuspended in 4 mL of PBS and mixed with 8 mL of Percoll:Hanks’ Balanced Salt Solution (10× concentrate) (6:1, at pH 7.0). After centrifugation at 20 °C for 30 min at 450g, the monocytes were collected

from the top of the gradient, washed in PBS/BSA, and seeded in 24- and 48-well tissue culture plates at 1-2 × 106 cells/ well and 0.5-1 × 106 cells/well, respectively, using Mφ-SFM (Macrophage Serum-Free Medium, Invitrogen) unless otherwise stated. Adherent monocytes were cultured at 37 °C in humidified air/5% CO2 using Mφ-SFM for at least 6-7 days prior to experiments unless stated otherwise, renewing the culture medium twice a week. HMMs were treated with C60 for 24 h at concentrations of 5 µg/mL in cell culture medium. For EFTEM analysis, following exposure washed cell monolayers were fixed with 4% glutaraldehyde in PIPES buffer (0.1 M, pH 7.4) for 1 h at 4 °C. Then, cells were treated with graded solutions of ethanol (70, 95, and 100%) for 5 min in each solution. Samples were infiltrated under vacuum in LR white resin (Agar Scientific, U.K.) for 3 days. Samples were then cured in fresh LR white for 23 h at 60 °C. Seventy nanometer thick sections of both HMMs exposed and not exposed to C60 were cut onto distilled water with an ultramicrotome using a 35° wedge angle diamond knife. Sections were collected immediately on lacey and plain carbon 300 mesh copper grids (Agar Scientific, U.K.) and dried for an hour at 37 °C. Selected sections were stained with uranyl acetate and lead citrate for 5 min in each. Preliminary EFTEM of cells exposed and not exposed to C60 was employed to assess the distribution of particles within the HMMs. Subsequent tomographic reconstructions were performed to allow an improved visualization of the particles within the cells in three dimensions. EFTEM was performed on a Philips CM300 operated at 300 kV, using a 10 µm objective aperture to optimize spatial resolution (14). An operating voltage of 300 keV was used as the Philips CM300 TEM is well aligned for EFTEM at 300 keV. This is particularly the case for the low magnification search mode which is extremely useful for conducting these experiments. This voltage was also chosen to minimize beam damage to both the cell and the C60, where operating at high voltages reduces damage to the cell by beam heating and operating at low voltages reduces damage to the C60 via knock on damage. To optimize contrast zero-loss images of nonstained sections were taken using a 3 eV slit, and zero-loss images of stained sections were taken using a 20 eV slit. Low-loss energy filtered series were recorded from 0 to 34 eV using a 2 eV slit and 2 eV step size (15). Electron energy loss spectra (EELS) were extracted from the EFTEM series using IDL image processing software. Electron tomography was performed on 70 nm thick sections. The data sets were acquired over a tilt range of -70 to +70° using a step size of 2°. Low-loss energy filtered series, using the parameters described above, were recorded at each tilt angle. Each image was corrected for any shift relative to the reference image using a cross-correlation routine. Spatial and rotational alignment through the tilt series was corrected by sequential cross-correlation and series averaging. Threedimensional reconstruction was carried out using weighted back-projection (16) and a simultaneous iterative reconstruction technique (SIRT) (17). Reconstructions were performed using IDL software. The anisotropic nonlinear diffusion (AND) algorithm was employed to denoise and enhance the 3-D tomograms (18).

Results Imaging unstained sections of cells without energy filtering produces very little contrast. The contrast from cell organelles was significantly improved by imaging at electron losses between 0 and 50 eV with a 3 eV energy window. Optimum imaging from cell membranes and the nucleus was achieved at 45 eV (approximately twice the plasmon energy of the VOL. 41, NO. 8, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3013

FIGURE 1. Low-loss EELS spectra from C60 within a cell. (a) Crystal inside a vacuole. Spectrum taken from C60 crystals in Figure 3a. (b) Cytoplasm of a cell exposed to C60. Spectrum taken from the cytoplasm c in Figure 3a. (c) Disordered aggregate of C60. (d) Cell not exposed to C60. (e) Lacy carbon support film.

FIGURE 2. HMM cells exposed to C60. Zero-loss EFTEM images of (a) C60 fused with the plasma membrane. m - plasma membrane. (b) A hexagonal cluster of C60 adjacent to the nucleus (n) of the cell. Inset A shows fine particles distributed within the nucleus. nm nuclear membrane. cell). EELS spectra taken from the cell and C60 are shown in Figure 1. The following descriptions document observations that were made in analyzing several representative areas of cells from three different experiments. C60 was observed at several locations within cells: C60 was found aligned along the plasma membrane (Figure 2a), and disordered C60 was seen clustered within the nucleus 3014

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 8, 2007

FIGURE 3. (a) 0 eV, (b) 20 eV, (c) 26 eV, and (d) 26/20 eV images using a 2 eV slit of C60 within the nucleus. The images in Figure 2b,c have a similar contrast and dividing one by the other (Figure 2d) indicates that the loss behavior of the C60 and the surrounding cell medium is near-identical. nm - nuclear membrane, n - nucleus. (Figure 2b) and within lysosomes (Figure 2b). The lacey carbon support film was frequently imaged superimposed on the section. To confirm the subcellular localization of the C60 particles, sections were stained with lead citrate and uranyl acetate and analyzed using zero-loss EFTEM (12). Low-loss EFTEM series were taken to confirm the particles were C60. Low-loss EFTEM images taken at 20 and 26 eV with a 2 eV window provide a “reverse” contrast image between the C60 and the cell. Figure 3a-c compares 0, 20, and 26 eV loss images of C60 aggregates within the nucleus. This can, of course, be reversed electronically to provide an image with positive contrast to compare with heavy metal stained sections imaged by bright field TEM. Figure 4a,b shows similar 20 and 26 eV loss images of C60 within a lysosome. The images in Figure 3b,c have a similar contrast and dividing one by the other (Figure 3d) indicates that the loss behavior of the C60 and the surrounding cell medium is near-identical. This can be explained if the C60 has very little graphitic-like bonding and is amorphous in nature. In comparison the C60 in Figure 4a has a relatively higher intensity in the 26 eV image (Figure 4c) than in the 20 eV image (Figure 4b). This is consistent with a higher plasmon energy in these C60 aggregates. Dividing one by the other (Figure 4d) indicates that the loss behavior of the C60 is greater from the crystalline C60. Dividing the 6 eV image by the 4 eV image maps shows the variation of the πf π* transition in the cytoplasm and shows an increase in intensity at the plasma membrane which is probably due to an increased concentration of graphitic C60 at this site (Figure 4e). This is confirmed by the intensity profile across the plasma membrane (Figure 4f). Spectra extracted from an EFTEM image series of crystals of C60 in Figure 4 revealed two characteristic peaks at 6 and 26 eV, corresponding to the πf π* transition and the π+σ plasmon for crystalline C60, respectively, thus confirming that these aggregates were well ordered, probably crystalline, C60 (Figure 1) (9). Energy loss spectra from particles within the “hexagon” in Figure 2b revealed they were mainly disordered carbon (π+σ plasmon at 21 eV and the absence of the πf π* transition) with a small 6 eV shoulder, suggesting that some free C60 is remaining. The cytoplasm of the cell exposed to C60 (Figure 4) revealed two characteristic peaks at 6 and 22 eV, corresponding to the πf π* transition, and the π+σ plasmon for disordered C60, respectively. Comparison of spectra from the cytoplasm of a cell not exposed and exposed

FIGURE 5. Zero-loss tomographic reconstruction of a cluster of C60 particles penetrating into the nucleus of the cell: (a) Voltex projection and (b) orthoslices at 20 nm intervals in height though the cell. Showing the nucleus, the cytoplasm, and the C60. FIGURE 4. (a) 0 eV, (b) 20 eV, (c) 26 eV, (d) 26/20 eV, and (e) 6/4 eV images using a 2 eV slit of crystals of C60 within a lysosome. Clusters in the 26 eV image have a higher intensity than the cell medium in the 26 eV image than in the 20 eV image, and this increase in intensity is confirmed in the ratio image (26 eV/20 eV) in part d. In part b regions of lighter contrast are marked on the surface of the crystals which suggest the presence of surface plasmons (p). The histogram (Figure 2f) across the plasma membrane in part e shows an increase in intensity at the plasma membrane which is indicative of a greater density of C60 at this site. m - plasma membrane, c cytoplasm. to C60 showed two main differences: first a shift in the π+σ plasmon energy from 22 eV for the exposed cytoplasm to 20 eV for the unexposed, and second the presence of the πf π* transition only in the cytoplasm of cells exposed to C60. The presence of the πf π* transition arises from excitation of the π-bonded electrons in the graphitic C60, suggesting that C60 in molecular form must be present within the cytoplasm. The fact that the C60 has not aggregated implies that individual C60 is bound to the cytoplasm. The π+σ plasmon for C60 usually occurs at around 26 eV, the apparent shift to 22 eV results from the contribution from the cytoplasm (as the beam transmits through the sample in the region of C60 it will interact with both the cytoplasm and the C60), and also a contribution resulting from the surface plasmon which, assuming a Drude model, occurs at an energy of Ep/x2 as the C60 has a high surface area. Finally, as a control, loss spectra of the lacy amorphous carbon support showed one broad peak at 23 eV, as expected. During all acquisitions the C60 eventually degraded into disordered carbon after approximately 5 min exposure to the electron beam, with loss of the πf π* transition and shift of the plasmon energy to 22 eV. For this reason experiments were conducted with fast acquisition times and minimum exposure to the electron beam. It was not possible, using 2-D imaging, to be completely certain whether these particles were simply lying on the surface of the section or have actually entered the cell. For

this reason we performed 3-D electron tomography on selected sections to determine the 3-D distributions of the particles. First we performed zero-loss BF tomography of a stained section to confirm penetration of an aggregate of C60 into the nucleus and second used low-loss EFTEM tomography to assess the potential of combining techniques to improve the contrast between the C60 and cellular structures. Figure 5 shows a zero-loss BF tomography SIRT reconstruction of a cluster of particles piercing the nucleus. Reconstructions were visualized using voltex projections using Amira 3D visualization software (Mercury Computer Systems Inc., Me´rignac Cedex, France). A voltex projection of C60 piercing through the nuclear membrane is shown in Figure 5a. Figure 5b shows orthoslices through the reconstructed volume show that the cell is devoid of any C60 in the top slice, whereas the aggregate is clearly visible in the middle and bottom slices. The nuclear membrane appears clearer in the orthoslices than in previous zero-loss images. The origin of the median-scale “speckle” contrast is unknown. These projections confirm inclusion of the C60 into the nucleus of the cell. Low-loss EFTEM tomography was performed to improve characterization of 3-D distributions of C60 within the cell. This technique requires an acquisition time of several hours, during which time the C60 degrades under the electron beam. For this reason tomographic reconstructions were performed at 0 and 21 eV, as this energy will provide maximum contrast between the cell and the C60 (mapping the π+σ plasmon of the cell exposed to C60 and π+σ plasmon of C60 which has degraded under the electron beam). Figure 6 shows orthoslices through the 0 eV (Figure 6b,c) and 21 eV (Figure 6e,f) reconstructed volume also confirmed that C60 was present within the nucleus of the cell. Comparison of tomographic reconstructions from 0 and 21 eV series showed little change in the contrast of the C60 cluster (slightly clearer morphology in the 21 eV image), but the 21 eV tomogram had improved SNR. VOL. 41, NO. 8, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3015

FIGURE 6. Tomographic reconstruction of a cluster of C60 particles within the nucleus of the cell. (a-c) 0 eV reconstruction: (a) Voltex reconstruction of 0 eV reconstruction and (b and c) orthoslices at 20 and 40 nm, through the 0 eV reconstructed volume. (d-f) 21 eV reconstruction: (d) Voltex reconstruction of 21 eV reconstruction and (e and f) orthoslices at 20 and 40 nm though the reconstructed volume. nm - nuclear membrane.

Discussion The first objective of our study was to use images recorded using zero-loss and plasmon energies to reveal structures within the cell and to improve the image contrast between the cell and the C60 in unstained specimens; conventional TEM images of unstained human macrophage exposed to C60 show very little contrast. To achieve contrast, traditional biological electron microscopy employs the use of a heavy metal salt to improve contrast between cellular components. Although these stains improve contrast from cellular components, they may also obscure differentiation between fine aggregates of C60 and cellular components. We have successfully demonstrated that by using a 3 eV energy-window at specific energy losses, fine detail can be observed within an unstained cell section as well as differentiating between the cell and the C60 in both disordered and crystalline forms (18, 19). Furthermore, low-loss spectra allowed us to resolve C60 within the cytoplasm of the cell which could not be visualized in the EFTEM image. These results confirm our hypothesis that low-loss EFTEM techniques provide a significant improvement in contrast achievable between the cell and the C60. One objective of our study was to combine low-loss EFTEM and tomography to identify C60 within the cell. In a two-dimensional energy-filtered image series we found significant differences in the plasmon energy between crystalline C60, disordered C60, the cell, and the amorphous carbon support film. The macrophage is a carbonaceous material with other elements such as nitrogen and phosphorus and exhibits a plasmon maximum of 20 eV, lower than that of amorphous carbon. This difference in plasmon energies between the cell and fullerene enables the ratio of energy-filtered images recorded at 20 eV and 26 eV to be used to differentiate carbonaceous materials, as has been demonstrated for carbon nanotubes and nylon by Gass et al. 3016

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 8, 2007

(6). In 3-D, differentiation between carbonaceous species was more difficult. The fullerene crystals are beam sensitive and quickly degrade to disordered carbon clusters after relatively short exposures, reducing the difference in plasmon energies from 8 eV to 3 eV. This shift to lower plasmon energies for the fullerenes makes it more difficult to obtain ratio images that differentiate materials. Despite this it is still possible to distinguish the more intense signal from the cell attributed to its peak in plasmon energy, not yet reached by the disordered carbon aggregates. This observation perhaps explains why the SNR was lower in the 21 eV loss than at zero-loss in our tomographic reconstructions. Furthermore, we have successfully used 3-D electron tomography to determine distributions of the C60 within the cell and were able to show a marked increase in the clarity of information gained from imaging in 3-D. These results reiterate the potential of 3-D electron tomography for assessing the 3-D distribution of carbon based nanoparticles within cells (12). Detection of C60 highly concentrated at the plasma membrane and in the cytoplasm suggests that it may be able to diffuse through cell membranes to sites in the cytosol where it is not compartmentalized, and therefore the cytoplasmic constituents of cells may not be protected from C60 entry by the lipid bilayer (20). Diffusion across membranes may be increased if the C60 generates free radicals close to the membrane where lipid peroxidation damaging the membrane and increasing the permeability of the lipid bilayer (20, 21). Molecular C60 does exist in the cytoplasm, and it may also enter the nucleus by diffusion through the cytoplasm and the nuclear pores, eventually recrystalizing in the nucleus. Nuclear C60 may generate free radicals leading to DNA adducts being formed, with subsequent potential for mutagenesis and carcinogenesis (22), events which are already associated with a number of free radical-generating pathogenic particles (23, 24). In conclusion, low-loss EFTEM techniques provide a significant improvement in contrast achievable between the organelles of the cell and C60. We were able to differentiate between crystalline and disordered forms of C60, and lowloss spectra allowed us to resolve C60 within the cytoplasm of the cell which could not otherwise be visualized in an EFTEM image. We confirm that C60 is taken up by human monocyte-derived macrophage cells and is distributed within the cytoplasm, in lysosomes, aligned along the plasma membrane and most significantly within the nucleus. There is a need to understand the distribution of nanoparticles at the whole body level as well as within cells so that the risks nanoparticles pose can be adequately assessed, and we demonstrate here that ultrastructural studies can make a considerable contribution to that aim.

Acknowledgments Financial support was provided by the IRC in Nanotechnology, Cambridge, U.K., Isaac Newton Trust, EPSRC, FEI Company, Royal Academy of Engineering, and the Leverhulme Trust for a Senior Research Fellowship. The Multiimaging Centre was established with funding from the Welcome Trust. We thank Dr. Steffi Friedichs and Prof. Ken Donaldson for informative discussions.

Literature Cited (1) Oberdorster, E. Manufactured nanomaterials (fullerenes, C60) induce oxidative stress in the brain of juvenile largemouth bass. Environ. Health Perspect. 2004, 112 (10), 1058-62. (2) Sayes, C. M.; Goin, A. M.; Ausman, K. D.; Mendez, J.; West, J.; Colvin, V. L. Nano-C(60) cytotoxicity is due to lipid peroxidation. Biomaterials 2005, 26 (36), 7587-95. (3) Sayes, C.; Fortner, J.; Lyon, D.; Boyd, A.; Ausman, K.; Tao, Y.; Sitharaman, B.; Wilson, L.; West, J.; Colvin, V. L. The differential cytotoxicity of water soluble fullerenes. Nano Lett. 2004, 4, 1881.

(4) Weyland, M.; Midgley, P. A. Extending energy filtered transmission electron microscopy (EFTEM) into three dimensions using electron tomography. Microsc. Microanal. 2003, 9, 542. (5) Midgley, P. A.; Weyland, M. 3D electron microscopy in the physical sciences: the development of Z-contrast and EFTEM tomography. Ultramicroscopy 2003, 96, 413. (6) Gass, M. H.; Koziol, K. K.; Windle, A.; Midgley, P. FourDimensional Spectral Tomography of Carbonaceous Nanocomposites. Nano Lett. 2006, 6 (3), 376-379. (7) Papworth, A. J; Kiely, C. J; Burden, A. P.; Silva, S. R. P.; Amaratunga, G. A. J. Electron-energy-loss spectroscopy characterization of the sp2 bonding fraction within carbon thin films. Phys. Rev. B 2000, 62, 12628. (8) Midgley, P. A.; Weyland, M.; Thomas, J. M.; Johnson, B. F. G. Z-Contrast tomography: a technique in three-dimensional nanostructural analysis based on Rutherford scattering. Chem. Commun. 2001, 10, 907. (9) Sto¨ckli, T.; Bonard, J. M.; Chaˆtelain, A. Plasmon excitations in graphitic carbon spheres measured by EELS. Phys. Rev. B 2000, 61, 5751-5759. (10) Brown, D. M.; Donaldson, K.; Borm, P. J.; Schins, R. P.; Dehnhardt, M.; Gilmour, P. Calcium and ROS-mediated activation of transcription factors and TNF-alpha cytokine gene expression in macrophages exposed to ultrafine particles. Am. J. Physiol. Lung Cell Mol. Physiol. 2004, 286, L344-L353. (11) Donaldson, K.; Tran, L.; Jimenez, L.; Duffin, R.; Newby, D. E.; Mills, N. Combustion-derived nanoparticles: A review of their toxicology following inhalation exposure 1. Part. Fibre Toxicol. 2005, 2, 10. (12) Porter, A. E.; Muller, K.; Skepper, J.; Midgley, P.; Welland, M. Uptake of C60 by human monocyte macrophages, its localization and implications for toxicity: studied by high resolution electron microscopy and electron tomography. Acta Biomater. 2006, 2 (4), 409-19. (13) Denholm, E. M.; Wolber, F. M. A simple method for the purification of human peripheral blood monocytes: a substitute for Sepracell-MN. J. Immunol. Methods 1991, 144, 247. (14) Krivanek, O. L.; Kundmann, M. K.; Kimotok, K. Spatial resolution in EFTEM maps. J. Microsc. 1995, 180, 277-287.

(15) Thomas, P. J.; Midgley, P. A. 3D electron microscopy in the physical sciences: the development of Z-contrast and EFTEM tomography. Ultramicroscopy 2001, 88, 179-186. (16) Kak, A. C.; Slaney, M. Principles of Computed Tomography Imaging; IEEE: NY, 1988. (17) Gilbert, P. Iterative methods for the three dimensional reconstruction of an object from projections. J. Theor. Biol. 1972, 36, 105-17. (18) Fernandez, J. J.; Li, S. J. An improved algorithm for anisotropic nonlinear diffusion for denoising cryotomograms. Struct. Biol. 2003, 144, 152-161. (19) Yao, N.; Klein, C. F.; Behal, S. K.; Disko, M. M.; Sherwood, R. D.; Creegan, K. M.; Cox, D. M. Transmission electron diffraction of the ordering transformation in crystalline C60. Phys. Rev. B 1990, 45 (19), 3366-399. (20) Geiser, M.; Rothen-Rutlshauser, B.; Kapp, N.; Schurch, S.; Kreyling, W.; Schulz, H. Ultrafine particles cross cellular membranes by non-phagocytic mechanisms in lungs and in cultured cells. Environ. Health Perspect. 2005, 113, 1555-1560. (21) Ali, S. S.; Hardt, J. I.; Quick, K. L.; Kim-Han, J. S.; Erlanger, B. F.; Huang, T. T. A biologically effective fullerene (C60) derivative with superoxide dismutase mimetic properties. Free Radical Biol. Med. 2004, 37, 1191-1202. (22) Marnett, L. J. Oxyradicals and DNA damage. Carcinogenesis 2000, 21, 361-370. (23) Carlson, H. L.; Nygren, J.; Moller, L. Genotoxicity of airborne particulate matter: the role of cell-particle interaction and of substances with adduct-forming and oxidizing capacity. Mutat. Res. 2004, 565, 1-10. (24) Schins, R. P., Knaapen, A. M., Cakmak, G. D.; Shi, T.; Weishaupt, C.; Borm, P. J. Oxidant induced DNA damage by quartz in alveolar epithelial cells. Mutat. Res. 2002, 517, 77-86.

Received for review October 23, 2006. Revised manuscript received January 19, 2007. Accepted January 26, 2007. ES062541F

VOL. 41, NO. 8, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

3017