NANO

LETTERS

Cellular Toxicity of Carbon-Based 2006

Vol. 6, No. 6
Nanomaterials 1121-1125

Arnaud Magrez,*,† Sandor Kasas,‡,§ Valérie Salicio,| Nathalie Pasquier,⊥

Jin Won Seo,† Marco Celio,| Stefan Catsicas,‡ Beat Schwaller,| and László Forró†

Institut de Physique de la Matière Complexe (IPMC), Ecole Polytechnique Fédérale de

Lausanne, 1015 Lausanne, Switzerland, Laboratoire de Neurobiologie Cellulaire,
Faculté des Sciences de la Vie, Ecole Polytechnique Fédérale de Lausanne,
1015 Lausanne, Switzerland, Institut de Biologie Cellulaire et de Morphologie,
UniVersité de Lausanne, rue du Bugnon 9, 1005 Lausanne, Switzerland, UniVersity of
Fribourg, DiVision of Histology, Department of Medicine, 14, Chemin du Musée,
CH-1705 Fribourg, Switzerland, and Laboratoire Cytopath,
AVenue Cardinal-Mermillod 36, 1227 Carouge, Switzerland
Published on May 20, 2006 on http://pubs.acs.org | doi: 10.1021/nl060162e

Received January 24, 2006; Revised Manuscript Received April 10, 2006
Downloaded by UNIV OF GEORGIA on June 29, 2009

ABSTRACT
The cellular toxicity of carbon-based nanomaterials was studied as a function of their aspect ratio and surface chemistry. These structures
were multiwalled carbon nanotubes, carbon nanofibers, and carbon nanoparticles. Their toxicity was tested in vitro on lung tumor cells. Our
work clearly indicated that these materials are toxic while the hazardous effect is size-dependent. Moreover, cytotoxicity is enhanced when
the surface of the particles is functionalized after an acid treatment.

The discovery of numerous nanomaterials has added a new
dimension to the rapid development of nanotechnology.
Consequently, the professional and public exposure to
nanomaterials is supposed to increase dramatically in the
coming years. Especially, carbon-based nanomaterials (CBNs)
are currently considered to be one of the key elements in
nanotechnology. Their potential applications range from
biomedicine through nanoelectronics to mechanical engineer-
ing. Thus, it is primordial to know the health hazards related
to their exposure. Here, we performed studies on cultured
cells exposed to three different types of CBNs, carbon
nanotubes, carbon fibers, and carbon nanoparticles. Severe
inhibition of cell proliferation and cell death have been
observed, which become more pronounced as the aspect ratio
of CBNs decreases and with the presence of chemically
active functional groups on the graphene surfaces. These
results indicate that more attention has to be paid to the health
concerns of CBNs before pushing their application forward.
* To whom correspondence should be addressed. E-mail:
arnaud.magrez@epfl.ch.
† Institut de Physique de la Matière Complexe (IPMC), Ecole Polytech-
nique Fédérale de Lausanne.
‡ Laboratoire de Neurobiologie Cellulaire, Faculté des Sciences de la
Vie, Ecole Polytechnique Fédérale de Lausanne.
§ Institut de Biologie Cellulaire et de Morphologie, Université de
Lausanne.
| University of Fribourg, Division of Histology, Department of Medicine.
⊥ Laboratoire Cytopath.
10.1021/nl060162e CCC: $33.50 © 2006 American Chemical Society
Published on Web 05/20/2006
Carbon-based nanomaterials are currently one of the most
attractive nanomaterials with their different forms, such as
fullerenes, single- and multiple-walled carbon nanotubes,
carbon nanoparticles, nanofibers, and so forth. Although
some of them are already in mass production, for carbon
nanotubes only in the last two years have novel methods
been able to considerably improve their synthesis and yield.
Recent studies demonstrated that CBNs also aggregate in
combustion streams of fuel gas and air commonly used in
our everyday life, indicating that we are already strongly
exposed to CBNs in the atmospheric environment both in-
and outdoor.1 Significant progress has been made in incor-
porating CBNs in potential applications. In particular,
biological applications that employ CBNs for DNA, proteins,
and drug delivery2,3 have attracted much attention. Unfor-
tunately, the information concerning the potential hazards
related to CBN exposure is rare and still under debate.4-10
In particular, the toxicity of water-soluble CNTs has been
discussed.11,12 The scientific community is mostly concerned
about the toxicity of carbon nanotubes because of their
structural resemblance to asbestos. Inhalation of asbestos
fibers is known to induce asbestosis (a progressive fibrotic
disease of the lung), lung cancer, and malignant mesothe-
lioma of the pleura.13 The role of asbestos in lung cancer is
still under debate and, unfortunately, experiments performed
on rats or guinea pigs14 are not conclusive because the
 Figure 1. SEM images of (a) MWCNTs, (b) CNFs, and (c) carbon black. The aspect ratio of these nanoparticles is about 80-90, 30-40,
and 1, respectively. The scale bars correspond to 2 µm.

development of the asbestos-induced disease takes longer

than the lifetime of these test animals.
In the case of asbestos, where a benign silicate mineral
becomes carcinogenic in its fibrous form, the size, aspect
ratio, and surface charges have proven to have a strong
influence on their toxicity.15 How these parameters affect
the biotoxicity of CBNs is totally unknown, although it is
generally expected that they play a significant role. To
explore the shape influence of CBNs on cell toxicity, we
exposed different cell types to CBNs in vitro with different
Published on May 20, 2006 on http://pubs.acs.org | doi: 10.1021/nl060162e

aspect ratios, namely, multiwalled carbon nanotubes

(MWCNTs, produced by chemical vapor deposition16), with
an average diameter of 20 nm and aspect ratios ranging from
Downloaded by UNIV OF GEORGIA on June 29, 2009

80 to 90 (Figure 1a); carbon nanofibers, (CNFs, obtained

from Pyrograf Products, Inc.) with a mean diameter of 150
nm and aspect ratios of 30-40 (Figure 1b); and finally
flakelike-shaped carbon nanoparticles (carbon black, obtained
after grinding graphite, used as electrodes for the arc
synthesis of carbon nanotubes) with aspect ratios of about 1
and size distributions ranging in the submicrometer range
(Figure 1c). All CBNs were suspended in a strongly diluted
gelatin solution to avoid aggregation.
The effect on the cell proliferation and cytotoxicity of the
CBNs was evaluated by the widely established MTT assay17
performed with three different human lung-tumor cell lines,
H596, H446, and Calu-1. The assay is based on the
accumulation of dark-blue formazan crystals inside living
cells (but not in dead cells) after their exposure to MTT (for
details, see the Supporting Information). For each cell type,
a linear relationship between the number of living cells and
the optical density was established, and this allowed us an
accurate quantification of cell numbers. At first we approved
the suitability of different lung cancer cell lines for the
cytotoxicity assay by treating them with MWCNTs at
concentrations ranging from 0.002 to 0.2 µg/mL for 4 days
(Figure 2). The number of viable cells showed a CBN
concentration-dependent decrease in all cell lines. Neverthe-
less, because H596 cells showed the highest sensitivity as
well as the best reproducibility, H596 cells were used for
the principal experiments. We also validated that the cell
growth in a standard medium and in a medium containing Figure 2. Representative growth curve for three different human
the same gelatin concentration as the CBN-treated cells was lung tumor cells grown in normal medium (control), medium
not significantly different, indicating that addition of gelatin containing gelatin, or medium containing 0.02 and 0.2 µg/mL of
MWCNTs (see the Supporting Information for details). In control
at this concentration does not affect the proliferation.
cells, the increase in optical density, which is proportional to the
An average growth curve of H596 cells in a standard number of viable cells, is due to cell division. In MWCNT-treated
medium is shown in Figure 3a (control). Although growth samples, the number of viable cells is lower at all time points.
curves of cells grown in a medium with the same gelatin
concentration but without exposure to CBNs were not cells in all CBN-treated samples was decreased. The toxic
different from the control conditions, the number of viable action on living cells already appeared within 24 h after
1122 Nano Lett., Vol. 6, No. 6, 2006
 Figure 3. (a) Representative growth curve for H596 cells grown in normal medium (control), medium containing gelatin (see the Supporting
Information for details), or 0.02 µg/mL of CBNs. MWCNTs, CNFs, and carbon black are dispersed in the same gelatin-containing medium.
(b) Dose-dependent toxicity of H596 cells exposed to CBNs for 2 days. In all cases, the toxicity order was carbon black > CNFs >
Published on May 20, 2006 on http://pubs.acs.org | doi: 10.1021/nl060162e

MWCNTs.
Downloaded by UNIV OF GEORGIA on June 29, 2009

Figure 4. Cytopathological analyses of H596 cells (a) A typical control image of H596 cells is depicted; cells were stained with hematoxyline-
eosine; the nuclei appear purple and patchy; the cytoplasm is weekly stained (pink). Clusters of cells are characterized by close cell/cell
contacts, and individual cells are polygonal-shaped. (b) H596 cells after 1 day treatment with 0.02 µg/mL of MWCNT. Cells have lost their
mutual attachments, retracted their cytoplasm (arrows) such that the pink color appears stronger, and the nuclei are smaller and more
condensed (picnotic) also evidenced by the stronger purple staining.

exposure, and differences between control and CBN-treated
become even more pronounced after culturing the cells for
5 days. Analysis of the dose-dependent toxicity after 2 days
of CBN exposure (Figure 3b) revealed that the number of
viable cells decreases as a function of the exposed CBN dose
for all CBNs tested. Moreover, a clear CBN morphological
dependence on the cell toxicity was observed: Contrary to
our expectations, carbon black particles exhibited the highest
cytotoxicity evidenced by the lowest number of viable cells
at all concentrations and time points tested (Figure 3 and
also by the large amount of cell debris in the culture medium
(not shown)). In particular, at low CBNs concentrations
(0.002 and 0.02 µg/mL) the number of viable cells decreased
in the following sequence: carbon black > CNFs > carbon
nanotubes. Thus, filaments were less toxic than particles in
our experiments. At a high CBNs concentration of 0.2 µg/
mL, differences diminished especially for CNFs and carbon
black. However, MWCNTs always appeared to be less toxic
Nano Lett., Vol. 6, No. 6, 2006
than the other two materials. Thus, surprisingly, filaments
were less toxic than particles in our experiments.
Light microscopic examination of the cells exposed to the
different CBNs revealed several morphological alterations
compared to the control cells grown in standard medium
(Figure 4). Already after 1 day of exposure, a fraction of
cells were only loosely attached to the culture dishes or even
completely detached. In the remaining cells cytoplasm
retraction with eosinophilia and shrunken (picnotic) nuclei
were observed, which are typical for irreversible cell injuries
and cell death. It has to be pointed out that no specific
differences could be observed at the light microscopic level
with respect to the cell morphology between cells exposed
to the different CBNs. Thus, differences in the number of
viable cells determined from the MTT assays do merely
reflect the sensitivity of cells to different CBNs, while the
morphological alterations obtained from light microscopy
indicate a common final cell death pathway for all CBNs.
1123

Figure 5. Effect of filament decoration on cell toxicity in H596
cells. The growth curves obtained from chemically decorated
MWCNTs and CNFs are denoted De-MWCNTs and De-CNFs,
respectively. The filament concentration to which all samples were
Published on May 20, 2006 on http://pubs.acs.org | doi: 10.1021/nl060162e
exposed to was 0.02 µg/mL. In both cases, the number of viable
cells is lower in the decorated samples, indicative of increased
toxicity.
Downloaded by UNIV OF GEORGIA on June 29, 2009
A possible explanation for the observed aspect-ratio
dependence of the CBN toxicity is the presence of dangling
bonds, which are highly reactive sites. In general, they are
present in carbon black with a high density, whereas in
carbon nanotubes they preferentially occur at lattice defects
or end caps.
To explore the effect of surface chemical properties on
the toxicity, we performed another set of experiments in
which we modified the surface chemistry of the filaments.
The surface of MWCNTs and CNFs has been decorated
according to the method reported by H. Hiura et al.18 that
involved a chemical modification of the outer layer of the
carbon nanotube after acid treatment. This procedure results
in adding carbonyl (CdO), carboxyl (COOH), and/or hy-
droxyl (OH) groups onto the nanotube and nanofiber
surfaces. H596 cells were grown in a gelatin-containing
medium containing 0.02 µg/mL of dispersed CBNs. The
MTT assay was carried out between one and 4 days after
exposure to CBNs. Cells grown in a plain gelatin-containing
medium served as reference. The chemical decoration effect
on cytotoxicity is displayed in Figure 5. When the number
of viable cells is compared after the treatment with CBNs,
it becomes evident that the toxicity increases with the
chemical surface treatment. This is significant in the case of
MWCNTs and moderate for CNFs. The latter observation
might be somewhat obscured by the fact that a relatively
high toxicity already occurred with unmodified CNFs.
Nevertheless, these results clearly demonstrate that graft-
ing additional putatively “toxic” chemical groups on the
surface of MWCNTs reduces the number of viable cells
significantly.
In conclusion, our experiments demonstrate that CBNs
generally lead to proliferation inhibition and cell death.
Although carbon nanotubes are less toxic than carbon fibers
1124
and nanoparticles, the toxicity of carbon nanotubes increases
significantly when carbonyl (CdO), carboxyl (COOH), and/
or hydroxyl (OH) groups are present on their surface. The
exact mechanisms that lead to cell death are still unclear,
but CBNs can induce cell death either after contact with cell
membranes or after their internalization. One has to consider
that the toxicity of the investigated CBNs could be specific
to the processing methods applied. Moreover, it has to be
emphasized that this study does not address the carcinoge-
nicity of CBNs, that is, the potential to transform a normal
cell to a tumor cell, which requires a detailed follow-up
investigation on that specific topic. In the last five years,
the question about potential toxicity of carbon nanotubes was
raised steadily. Although our study shows a lower toxicity
of carbon nanotubes compared to carbon black or filaments,
precautions in their manipulation need to be taken. In
particular, in applications where carbon nanotubes are
injected into human body for drug delivery,19 for example,
as contrast agent carrying entities for MRI,20 the toxicity issue
must be carefully addressed.
Acknowledgment. The work in Lausanne was supported
by the Swiss National Science Foundation and its NCCR
“Nanoscale Science”. The research of B.S. is supported by
the Swiss National Science Foundation (grant no. 3100A0-
100400/1).
Supporting Information Available: Experimental meth-
ods as well as additional time and dose dependence plots of
CBNs toxicity on H596 cells. This material is available free
of charge via the Internet at http://pubs.acs.org.
References
(1) Murr, L. E.; Bang, J. J.; Esquivel, E. V.; Guerrero, P. A.; Lopez, A.
J. Nanopart. Res. 2004, 6, 241-251.
(2) Wong Shi Kam, N.; O’Connell, M.; Wisdom, J. A.; Dai, H. J. Proc.
Natl. Acad. Sci. U.S.A. 2005, 102, 11600.
(3) Wu, W.; Wieckoski, S.; Pastorin, G.; Benincasa, M.; Klumpp, C.;
Briand, J. P.; Gennaro, R.; Prato, M.; Bianco, A. Angew. Chem., Int.
Ed. 2004, 43, 5242-5246.
(4) Sayes, C. M.; Fortner, J. D.; Guo, W.; Lyon, D.; Boyd, A. M.;
Ausman, K. D.; Tao, Y. J.; Sitharaman, B.; Wilson, L. J.; Hughes,
J. B.; West, J. L.; Colvin, V. L. Nano Lett. 2004, 4, 1881-1887.
(5) Oberdorster, E. EnViron. Health Perspect. 2004, 112, 1058-1062.
(6) Shvedova, A. A.; Castranova, V.; Kisin, E. R.; Schwegler-Berry, D.;
Murray, A. R.; Gandelsman, V. Z.; Maynard, A.; Baron, P. J. Toxicol.
EnViron. Health 2003, 66, 1909-1926.
(7) Lam, C. W.; James, J. T.; McCluskey, R.; Hunter, R. L. Toxicol.
Sci. 2004, 77, 126-134.
(8) Warheit, D. B.; Laurence, B. R.; Reed, K. L.; Roach, D. H.; Reynolds,
G. A. M.; Webb, T. R. Toxicol. Sci. 2004, 77, 117-125.
(9) Ding, L.; Stilwell, J.; Zhang, T.; Elboudwarej, O.; Jiang, H.; Selegue,
J. P.; Cooke, P. A.; Gray, J. W.; Chen, F. F. Nano Lett. 2005, 5,
2448-2464.
(10) Hurt, R. H.; Monthioux, M.; Kane, A. Carbon 2006, 44, 1027-
1120.
(11) Singh, R.; Pantarotto, D.; Lacerda, L.; Pastorin, G.; Klumpp, C.; Prato,
M.; Bianco, A.; Kostarelos, K. Proc. Natl. Acad. Sci. U.S.A. 2006,
103, 3357-3362.
(12) Chen, X.; Chong Tam, U.; Cziapinski, J. L.; Lee, G. S.; Rabuka, D.;
Zettl, A.; Bertozzi, C. R. J. Am. Chem. Soc., in press, 2006.
(13) LaDou, J. EnViron. Health Perspect. 2004, 112, 285-290.
(14) Huczko, A.; Lange, H.; Calko, E.; Grubek-Jaworska, H.; Droszcz,
P. Fullerenes, Nanotubes, Carbon Nanostruct. 2001, 9, 253.
(15) Rogers, R. A.; Antonini, J. M.; Brismar, H.; Lai, J.; Hesterberg, T.
W.; Oldmixon, E. H.; Thevenaz, P.; Brain, J. D. EnViron. Health
Perspect. 1999, 107, 367-375.
Nano Lett., Vol. 6, No. 6, 2006
 (16) Magrez, A.; Seo, J. W.; Miko, C.; Hernadi, K.; Forró, L. J. Phys.
Chem. B 2005, 109, 10087-10091.
(17) Mosmann, T. J. Immunol. Methods 1983, 65, 55-63.
(18) Hiura, H.; Ebbesen, T. W.; Tanigaki, K. AdV. Mater. 1995, 7, 275-
277.
(19) Colvin, V. L. Nat. Biotechnol. 2003, 21, 1166-1170.
Published on May 20, 2006 on http://pubs.acs.org | doi: 10.1021/nl060162e
Downloaded by UNIV OF GEORGIA on June 29, 2009
Nano Lett., Vol. 6, No. 6, 2006
(20) Sitharaman, B.; Kissell, K. R.; Hartman, K. B.; Tran, L. A.; Baikalov,
A.; Rusakova, I.; Sun, Y.; Khant, H. A.; Ludtke, S. J.; Chiu, W.;
Laus, S.; Toth, E.; Helm, L.; Merbach, A. E.; Wilson, L. J. Chem.
Commun. 2005, 31, 3915-3917.
NL060162E
1125