TOXICOLOGICAL SCIENCES 92(1), 5–22 (2006)
doi:10.1093/toxsci/kfj130
Advance Access publication February 16, 2006
REVIEW
Carbon Nanotubes: A Review of Their Properties in Relation to
Pulmonary Toxicology and Workplace Safety
Ken Donaldson,*,1 Robert Aitken,† Lang Tran,† Vicki Stone,‡ Rodger Duffin,* Gavin Forrest,§ and Andrew Alexander§
*MRC/University of Edinburgh Centre for Inflammation Research, ELEGI Colt Laboratory, Queen’s Medical Research Institute, Edinburgh EH16 4TJ,
United Kingdom; †Institute of Occupational Medicine, Riccarton, Edinburgh, United Kingdom EH14 4AP; ‡Napier University, School of Life Sciences,
Edinburgh, United Kingdom EH10 5DT; and §School of Chemistry, University of Edinburgh, Edinburgh, United Kingdom EH9 3JJ
Received November 22, 2005; accepted February 6, 2006
INTRODUCTION TO CARBON NANOTUBES
Carbon nanotubes (CNT) are an important new class of technological materials that have numerous novel and useful properties. The forecast increase in manufacture makes it likely that
increasing human exposure will occur, and as a result, CNT are
beginning to come under toxicological scrutiny. This review seeks
to set out the toxicological paradigms applicable to the toxicity of
inhaled CNT, building on the toxicological database on nanoparticles (NP) and fibers. Relevant workplace regulation regarding exposure is also considered in the light of our knowledge of
CNT. CNT could have features of both NP and conventional
fibers, and so the current paradigm for fiber toxicology, which is
based on mineral fibers and synthetic vitreous fibers, is discussed.
The NP toxicology paradigm is also discussed in relation to CNT.
The available peer-reviewed literature suggests that CNT may
have unusual toxicity properties. In particular, CNT seem to have
a special ability to stimulate mesenchymal cell growth and to
cause granuloma formation and fibrogenesis. In several studies,
CNT have more adverse effects than the same mass of NP carbon
and quartz, the latter a commonly used benchmark of particle
toxicity. There is, however, no definitive inhalation study available
that would avoid the potential for artifactual effects due to large
mats and aggregates forming during instillation exposure procedures. Studies also show that CNT may exhibit some of their
effects through oxidative stress and inflammation. CNT represent
a group of particles that are growing in production and use, and
therefore, research into their toxicology and safe use is warranted.
Key Words: carbon nanotubes; graphite; inflammation; oxidative
stress; nanoparticles; fibre; toxicology; asbestos; lungs; workplace.
1
To whom correspondence should be addressed at MRC/University of
Edinburgh Centre for Inflammation Research, ELEGI Colt Laboratory, Queen’s
Medical Research Institute, 47 Little France Crescent, Edinburgh EH16 4TJ,
United Kingdom. Fax: 0131 242 6582. E-mail: [email protected].
Nanotechnology is a global cross-disciplinary undertaking
that has extraordinary potential to change our lives by
improving existing products and enabling new ones. Like most
new technologies, the emphasis on benefits has been offset by
considerable debate about the uses and safety of nanotechnologies (Brumfiel, 2003; Dreher, 2004; Giles, 2003; Hood, 2004;
Maynard and Kuempel, 2005; The Royal Society and the Royal
Academy of Engineering, 2004; Seaton and Donaldson, 2005).
Occupational medicine is replete with examples of respirable
dust particle exposures causing disease, and so matters of
safety have been focused largely on nanoparticles (NP), which
represent only one aspect of nanotechnology. NP, generally
defined as particles less than 100 nm in any dimension and
previously called ultrafine particles by inhalation toxicologists,
are a concern because they may enter the body through the
lungs, skin, or gut depending on the type of exposure. Several
recent papers have highlighted this area of toxicology, its
novelty, the gaps in research, and possible testing strategies for
NP and nanotubes (Donaldson et al., 2004a; Oberdorster et al.,
2005a,b; Seaton and Donaldson, 2005).
There is sufficient ‘‘ultrafine particle’’ toxicology literature
to consider that the large and diverse group of manufactured
NP arising from nanotechnology comprises an inhalation hazard of unknown potential. Most of the concern has flowed from
this perception, enhanced by limited studies indicating that
they may also be taken up through the skin or gut. The notion,
for which some limited support exists, that NP may gain access
to the blood poses a relatively new particle toxicological
problem, i.e., particle effects at sites other than the lungs.
This paper examines the toxicology issues relating to carbon
nanotubes (CNT), one of the major new NP that are set to
be produced in bulk for many diverse purposes. CNT are discussed here exclusively as an inhalation hazard. Global revenues
from CNT in 2006 are estimated at ~ $230 million with a
growth rate of ~ 170% (data from BCC Research, Norwalk, CT,
The Author 2006. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved.
For Permissions, please email: [email protected]
6
DONALDSON ET AL.
FIG. 1. Molecular representations of SWCNT (top left) and MWCNT (top right) with typical transmission electron micrographs below.
http://www.bccresearch.com/). This provides potential for
workplace and even eventual general exposure, if there is
attrition of materials that contain them. Concern has been
raised over the safety of CNT because they have three properties that are clearly associated with pathogenicity in particles.
1. They are NP and so could have more toxicity than larger
sized particles.
2. They are fiber shaped and so might behave like asbestos
and other pathogenic fibers which have toxicity associated with their needle-like shape.
3. They are essentially graphitic and so are expected to be
biologically biopersistent.
In this paper we describe the synthesis, structure, and
composition of CNT and put this in the context of the known
toxicology of NP and pathogenic fibers such as asbestos to
draw conclusions regarding the toxicology issues relating to
these materials. We also discuss the exposure measurement
issue in relation to particles and fiber and highlight the
difficulties that this presents for nanotubes.
SYNTHESIS OF CNT
There are three methods commonly used to synthesize CNT:
arc discharge, chemical vapor deposition (CVD), and laser
ablation (Fig. 2). The common feature of these methods is
addition of energy to a carbon source to produce fragments
(groups or single C atoms) that can recombine to generate
CNT. The energy source may be electricity from an arc
discharge, heat from a furnace (~ 900C) for CVD, or the
high-intensity light from a laser (laser ablation).
The mechanisms of CNT formation are widely debated
(Cassell et al., 1999; Sinnott et al., 1999). Where metal
catalysts are used, it is thought that carbon is deposited on
a catalyst NP, from which the C network finally extrudes to
form a graphite tube (Vinciguerra et al., 2003) (Fig. 3). This
Arc discharge (Popov, 2004)
Graphite electrodes
Electric discharge
Pure electrodes
multi-walled tubes
Metal-doped electrodes
single-walled tubes
High quantity of non-CNT carbon impurities
Chemical vapour deposition (CVD) (Cheung
et al., 2002)
TYPES OF CNT
Comprised entirely of carbon, the structure of pure CNT can
be visualized as a single sheet of graphite rolled to form
a seamless cylinder. There are two classes of CNT: singlewalled carbon nanotubes (SWCNT) and multiwalled carbon
nanotubes (MWCNT; Fig. 1). MWCNT are larger and consist
of many single-walled tubes stacked one inside the other. CNT
are distinct from carbon fibers, which are not single molecules
but strands of layered graphite sheets. CNT are reported to be
physically strong and stiff and, e.g., SWCNT can be as much as
10 times as strong as steel and 1.2 times as stiff as diamond
(Walters et al., 1999; Yu et al., 2000).
Carbon
+
Energy
Gaseous carbon source (e.g., methane,
carbon monoxide, or acetylene)
High temperature (ca. 900 ºC)
Metal-catalyzed decomposition of carbon
source on solid support (e.g., alumina Al2O3)
Laser ablation (Dai, 2002b)
Graphite target
High-intensity laser pulses in high temperature
inert atmosphere (ca. 1200ºC)
Pure electrodes
multi-walled tubes
single-walled tubes
Metal-doped electrodes
FIG. 2. Common methods used to synthesize CNT.
TOXICOLOGY OF CARBON NANOTUBES
7
FIG. 3. Proposed growth mechanism of CNT (left) and images of growing nanotubes captured using TEM (right); scale bar ¼ 10 nm (Dai, 2002) [permission
for TEM granted by Accounts of Chemical Research].
growth has been observed using in situ electron microscopy
(Helveg et al., 2004) (Fig. 3). The size of the metal particle
determines the diameter of the tube; SWCNT grow from metal
particles a few nanometers in diameter, while larger particles
tend to produce wider MWCNT.
Commercially available CNT are often synthesized by CVD
as this process is easily scaled up for industrial production.
The annual global production of single-walled tubes was 9 tons
in 2004 (Cientifica, 2004). Typically, a major producer of
SWCNT produces ~ 450 kg/day (Borowiak-Palen et al., 2002;
Bronikowski et al., 2001; Islam et al., 2003). The high-pressure
carbon monoxide method, which uses iron as the catalyst and
carbon monoxide as the carbon source (Bronikowski et al.,
2001), is commonly employed. MWCNT are also in commercial production, and one company (see web site http://www.
fibrils.com/) has been making plastic composites incorporating
multiwalled nanotubes since 1983. The annual global production capacity for MWCNT was estimated at 100 tons in
2004 (Cientifica, 2004).
After synthesis, CNT are usually purified to remove
amorphous carbon (‘‘soot’’), free residual catalyst (metal from
which nanotubes failed to grow), and any support material.
This is typically achieved by washing or ultrasonicating with
dilute acid. Highly purified tubes may have additionally
undergone some form of oxidation. Removal of some supports
like silica and alumina necessitates a stronger treatment
process (e.g., with concentrated acid; Borowiak-Palen et al.,
2002; Chiang et al., 2001). Since purification also destroys
CNT, the removal of impurities must be balanced against the
introduction of defects into tubes.
COMPOSITION AND REACTIVITY OF CNT
The chemistry of pure CNT is surprisingly uninteresting.
They are significantly unreactive; e.g., SWCNT must be heated
to 500C before they burn in air (Zhang et al., 2002). However,
there are points in the structure of CNT which are more reactive
than others, such as defects due to missing carbon atoms and
the more strained curved-end caps (Lin et al., 2003). Smaller
nanotubes are more ‘‘strained’’ because they deviate further
from the ideal planar structure of graphite. Purified nanotubes
are likely to contain additional defects in the form of
carboxylic acid (-COOH) residues.
A sample of CNT invariably contains a variety of residual
impurities. There are three classes of impurities that may
remain from the synthesis process: metals, organics, and
support material.
Co, Fe, Ni, and Mo are by far the most commonly used
metals in CNT synthesis, and mixtures of these are also widely
employed; Mo is often used as an additive to promote SWCNT
growth (Kitiyanan et al., 2000). Postproduction processing
removes the majority of metal catalyst; however, Carbon
Nanotechnologies Incorporated (CNI, Houston, TX) purified
grade tubes, e.g., may still contain up to 15% residual metal by
mass (Brelsford, personal communication, CNI). Residual metal
is usually encapsulated in a layer of carbon, either amorphous
soot or layers of graphite; metal NP that fail to grow tubes often
end up encapsulated in this way. Additionally, it is observed that
metal NP often become incorporated into the ends of nanotubes;
in this case, it would appear that the metal is inaccessible to the
outside environment (Dai, 2002). Most residual metal is nanoparticulate, although aggregates may be formed. Any metal NP
not coated in carbon will be coated with a metal oxide layer.
Residual organics can be divided into two subclasses:
various forms of bulk (amorphous or microstructured) carbon
and residual organic molecules. Bulk carbon structures formed
include amorphous soot particles and a variety of structures
built up from graphite sheets of different sizes; e.g., sheets can
linearly stack to make carbon fibers or they can nest in spheres.
Carbon fibers may consist of planar or ‘‘paper cup’’–shaped
graphite sheets stacked together (Ertl et al., 1997).
CNT synthesis involves the use of a material to support the
catalyst or growth region. Aluminates and silicates (e.g.,
zeolites) are common supports but require strong acids or
bases to remove them later. Magnesium oxide is increasingly
used as a CNT growth support as it can be dissolved in milder
acids. A high surface area is a desirable property for these
8
DONALDSON ET AL.
supports and helps prevent coalescence of metal NP during synthesis; for this reason, nanopowders are often used as supports.
solubility. Functionalization with different groups is likely to
result in different toxicity (Sayes et al., 2005) since the particle
surface is important in interacting with biological systems.
SIZE OF CNT
CNT vary significantly in length and diameter depending on
the synthetic procedure. Lengths are generally dependent on
synthesis time but are typically tens of microns, although significantly shorter and longer nanotubes have been made (Motta
et al., 2005; Puretzky et al., 2002). Cleaned and processed CNT
are usually shorter than as-produced nanotubes due to the
destructive conditions used in purification (Liu et al., 1998).
The diameters of SWCNT are controled by the sizes of the
metal NP from which they are grown, which vary between
about 0.7 and 3 nm (Jorio et al., 2001). MWCNT generally
range from 10 to 200 nm in diameter (Hou et al., 2003). More
important, however, is the strong tendency of both SWCNT and
MWCNT to bundle together in ropes as a consequence of
attractive van der Waals forces analogous to forces that bind
sheets of graphite (Thess et al., 1996). Bundles typically
contain many tens of nanotubes and can be considerably longer
and wider than the nanotubes from which they are formed. This
could have important toxicological consequences.
Much research has focused on creating homogeneous dispersions of CNT in various solvents. CNT are slightly dispersed
in some organic solvents such as dichlorobenzene (95 mg/l),
chloroform (31 mg/l), and carbon disulphide (2.6 mg/l; Bahr
et al., 2001). More complete dispersion is generally achieved by
prolonged sonication, but dispersions of CNT tend to aggregate
over time. CNT are not readily dispersible in water without the
aid of surfactant molecules to exfoliate individual or small bundles of nanotubes from larger aggregates, enclosing them in a
hydrophilic coating. Common surfactants used for this purpose
are sodium dodecylbenzenesulfonate, Triton X-100, sodium
dodecyl sulfate, and Cetyl Trimethyl Ammonium Bromide
(Islam et al., 2003; Moore et al., 2003; O’Connell et al., 2002).
As a consequence of their dimensions, CNT have very high
surface areas. The available surface area is dependent on the
length, diameter, and degree of bundling. Theoretically, discrete
SWCNT have surface areas of ~ 1300 m2/g, whereas MWCNT
generally have surface areas of a few hundred m2/g (Peigney
et al., 2001). As a result of bundling, the surface area of most
samples of SWCNT is dramatically lowered to ~ 300 m2/g, or
less, which is still a very high value (Ye et al., 1999b).
Functionalization of CNT Surfaces
Much research on CNT focuses on modifying these asproduced tubes by the addition of different chemical groups,
leading to a significant change in many of their properties
(Banerjee et al., 2003). Many companies sell these ‘‘derivatized’’ or ‘‘functionalized’’ nanotubes including fluorine
groups as a starting point for various other chemical modifications. Functionalization with polymer groups is used to enhance
PARTICLE TOXICOLOGY OF RELEVANCE TO CNT
Nanoparticles
Compact particles have their harmful effects as a consequence of two factors that act together to contribute to their
potential to cause harm: their surface area and the reactivity or
intrinsic toxicity of that surface (Donaldson and Tran, 2002;
Tran et al., 2000). The smaller the particles are, the more
surface area they have per unit mass; therefore, any intrinsic
toxicity of the particle surface will be emphasized (Duffin
et al., 2002). Therefore, as particles become generally smaller,
their likelihood of causing harm to the lung increases. Thus,
particles with a relatively inert surface can exert their effects on
cells by having a large surface area. Alternatively, particles
with a highly reactive surface can affect cells without
necessarily having a large surface area. NP are currently
available in a variety of compositions that range from very
simple—almost pure carbon or titanium dioxide (TiO2)—to
very complex structures, where surface modifications or coatings are applied. The nanotechnology industry is also developing new NP for use in a variety of purposes (The Royal
Society and the Royal Academy of Engineering, 2004), and the
toxicology of these materials is largely unknown.
NP in the lungs. We are already exposed to large numbers
of ambient NP in environmental air pollution (Maynard and
Howard, 1999). In this regard, combustion-derived NP have
been the focus of much research as a likely driver of adverse
health effects, such as exacerbations of airways disease as well
as deaths and hospitalization from cardiovascular disease
(Donaldson et al., 2005). There is a large existing database in
the lung particle toxicology literature that shows NP of various
sorts to have extra toxicity (Donaldson et al., 2002) by which
we mean that the same material in the form of NP is more toxic
than in the form of larger, still respirable particles. The
mechanism of lung injury caused by combustion-derived NP
is predominantly via oxidative stress, produced at the particle
surface that leads to cell injury and lung inflammation. The
oxidative stress is produced as a consequence of the properties
of the combustion-derived NP, principally the large surface
area, metals, and organic molecules (reviewed in Donaldson
et al., 2004). Oxidative stress–responsive transcription factors
such as nuclear factor kB (NF-kB) and Activator Protein-1 then
translate this oxidative stress into proinflammatory proteins,
a sequence of events described for a number of pathogenic
particle types (Donaldson et al., 2004a; Schins and Donaldson,
2000). Inflammation can then impact a number of pathological
processes such as airways disease, cardiovascular disease,
fibrosis, or cancer (Mauderly et al., 1994).
TOXICOLOGY OF CARBON NANOTUBES
Earlier carcinogenesis studies with ultrafine carbon black
and TiO2 were complicated by rat lung overload, a nonspecific
response of the rat lung to high levels of even low-toxicity
particles that probably does not operate in humans (Lee et al.,
1986, 1985a,b; Mauderly et al., 1990). However, it has recently
been shown that low (1000 mg/m3) short-term (7 h) exposures
to NP carbon black cause mild inflammatory effects in rat lungs
(Gilmour et al., 2004b). Proinflammatory effects of NP have
been described in a number of in vitro models (Brown et al.,
2004; Gilmour et al., 2004a), as has their ability to generate
reactive oxygen species and oxidative stress (Stone et al., 1998;
Wilson et al., 2002). Additionally NP have been seen to inhibit
phagocytosis (Renwick et al., 2001) and increase macrophage
sensitivity to chemotactic complement components (Renwick
et al., 2004).
Very small particles and structures could have a range of
other effects that are not seen with conventional particles. For
instance, they may not be detected by the normal phagocytic
defenses allowing them to gain access to the blood or the
nervous system. Very small particles are smaller than some
molecules and could act like haptens to modify protein
structures: either altering their function or rendering them
antigenic, raising the potential for autoimmune effects (Borm
and Kreyling, 2004).
NP and extrapulmonary responses. Of special concern is
the apparent ability of NP to redistribute from their site of
deposition. Thus, following inhalation exposure, NP have been
reported to travel via the nasal nerves to the brain (Kreuter
et al., 2002; Oberdorster et al., 2004), as has been described for
poliovirus (Bodian and Howe, 1941), and to gain access to the
blood and other organs (Kreyling et al., 2004). Thus, adverse
effects of NP are likely to occur in highly different scenarios,
dependent on particle type and portal of entry. For NP that are
made and handled in bulk, there is potential for lung exposure.
For some NP, such as those in sunblock creams, dermal
exposure is already occurring, and the range of different NP
in cosmetics is likely to increase (Pinnell et al., 2000; Tan et al.,
1996). NP in food are reported to cross into the gut lymphatics
and redistribute to other organs more readily than larger
particles (Hillery et al., 1994; Jani et al., 1990). Furthermore,
a huge class of NP is designed to be introduced directly into the
body for diagnostic and therapeutic reasons (Duncan, 2003).
However, it is worth noting that the chemistry and properties of
these different classes of NP are very diverse, and it is likely
that while some are likely to be toxic, others such as those
screened extensively for drug delivery may be relatively low in
their toxic potential.
Fibers
The vast majority of information on fiber toxicology is
derived from asbestos, but the synthetic vitreous fibers (SVF)
have contributed greatly to our current understanding of the
roles of length and biopersistence in mediating pathogenicity.
9
Consequently, the fiber toxicology paradigm discussed here is
based on mineral fibers and SVF (Bernstein et al., 2005); other
paradigms, as yet unknown, could be involved in pathogenicity
of other fibers types such as CNT. Inhalation of pathogenic
fibers is associated with fibrosis of the lung, i.e., asbestosis,
lung cancer, and mesothelioma (i.e., cancer of the mesothelium
lining the pleural and peritoneal cavities). Despite the decline
in use of asbestos in many countries, the number of deaths
attributed to mesothelioma, a disease often attributed solely to
asbestos exposure, continues to rise, and in the United
Kingdom it is predicted to peak between 2011 and 2015 at
somewhere between 1950 and 2450 deaths/year. The proposed
mechanisms of lung disease caused by fibers are numerous and
include oxidative stress, inflammation, and both direct and
indirect genotoxicity (Bernstein et al., 2005; Kane, 1996).
Fiber deposition. If they are sufficiently thin, even very long
fibers can readily penetrate to the distal lung, traveling into the
lung by aligning their longitudinal axis with the airstream. For
such particles, deposition is dictated by aerodynamic diameter,
which, for fibers with an actual diameter that is very small, is
not unduly influenced by length (Kennedy and Kelly, 1993).
There is substantial penetration of long thin fibers (for a
qualification of the term ‘‘long’’ see below) beyond the ciliated
airways. Brody et al. (1984) described the deposition of long
chrysotile fibers at and around the alveolar ducts, where the net
flow of air becomes zero, and so the potential for fibers settling
out is great (Lippmann, 1993). The ability of long and thin fibers
to reach the gas-exchange region of the lungs is important
because a failure to clear these fibers during ongoing exposure
leads to a build up of dose. The deposition of CNT in the lung is
expected to depend on the form that the CNT take when they are
aerosolized. In an aggregated form, the CNT are likely to be
deposited in the lung like a particle with a similar aerodynamic
diameter. However, if CNT can be aerosolized as single fibers,
their deposition will be conventional, i.e., based on aerodynamic
diameter and the possibility of aligning with the airstream.
Fiber length and inflammation. Length is an important
feature of fibers that modulates their pathogenicity. In a seminal
study (Davis et al., 1986), rats were exposed to both long
amosite and the same amosite shortened by milling. Following
inhalation exposure to equal airborne mass concentrations,
there was substantially more fibrosis and cancer with the long
fibers. Following ip instillation, mesothelioma was caused by
the long amosite but not the short amosite. The same group
(Donaldson et al., 1989) instilled the same fiber samples into
the mouse peritoneal cavity and reported intense inflammation
with the long amosite and virtually no inflammation with the
short amosite. The role of fiber length was further emphasized
when the length of glass fibers was found to be critical in NFkB activation (Ye et al., 1999a) in macrophages, leading to the
production of TNF-a. Using different size classes, the activation of this transcription factor, vital in the initiation of
inflammation, was far greater with the longest fraction.
10
DONALDSON ET AL.
Unpurified nanotube sample
Residual material
Nanotubes
Single-walled
1–3 nm diameter
Often form
bundles
Multi-walled
Metals
10–100 nm diameter
Typically Co, Fe,
Ni, Mo
Widely varying lengths, typically tens of
microns
Nanoparticulate
Carbonencapsulated
Oxides
Organics
Support
Amorphous carbon
Typically fine
alumina,
magnesium
oxide, or silica.
Microstructured
carbon (carbon
blacks, onions,
fibers, etc.)
FIG. 4. Possible components in a sample of unpurified CNT.
Fiber length, clearance, and accumulated dose. Long
fibers can be defined as fibers that significantly exceed the
size of macrophages and are usually taken to be 10- to 20-mm
long. Such fibers present problems to alveolar macrophages,
which will have difficulty in effectively phagocytosing and
‘‘clearing’’ them to the mucociliary escalator. Long fibers are,
therefore, more slowly cleared from the deep lung than shorter
fibers, as demonstrated in several studies (e.g., Coin et al.,
1994; Searl et al., 1999). Therefore, with ongoing exposure, the
dose of long fibers will accumulate more than for the more
rapidly clearing short fibers. Clearly, however, dose is not the
only explanation for the greater pathogenicity of long fibers. In
studies using the mouse peritoneal cavity (Donaldson et al.,
1989) and in vitro (Donaldson et al., 1992; Ye et al., 1999a),
where failed clearance is not the explanation, long fibers have
much more pronounced proinflammatory effects than short
fibers. There is, therefore, a ‘‘double hit’’ caused by long
fibers—firstly, they are more likely to be retained, so building
up the dose available to make contact with cells; secondly, they
are more stimulatory and biologically active than short fibers
when they make contact with cells.
Composition and Biopersistence
Biopersistence effects arise when fibers do not undergo
chemical dissolution in the lung tissue, weaken, break, or
dissolve away and therefore persist in the lungs. Asbestos fibers
are crystalline silicates with a number of other elements present
in the crystal structure that cause them to be identifiable by
their elemental composition into the amphiboles and chrysotile. Among the SVF there is also considerable variation in
chemical composition, but all are basically amorphous and
noncrystalline. Typically, these are made from melted rock or
slag or by melting together a mixture of components of which
the most common is SiO with varying amounts of other oxides,
such as BaO2, TiO2, and CaO, to produce fibers with different
properties (WHO IARC I.L.F., 2002). There are some crystalline man-made fibers such as SiC.
These differences in structure provide for differences in the
ability of fibers to resist chemical attack in the lungs. When
fibers are treated in vitro with salt solutions that are typical of
those found in biological systems, the fibers differ in their
‘‘durability’’ (Eastes et al., 2000), i.e., some types of fibers
undergo dissolution and either break up into smaller fragments
or dissolve entirely. Other types of fibers, like amphibole
asbestos, resist dissolution. In the lung, these differences are
reflected in a tendency for long fibers of nonbiopersistent
material to be cleared, as they dissolve or break up into shorter
fibers that can be removed (Hesterberg et al., 1998; Searl et al.,
1999). Thus, the harmful doses of fibers in the lungs are long
biopersistent fibers (Miller et al., 1999). This has lead to the
paradigm shown in Figure 5.
There was evidence in the earliest studies that chrysotile
fibers disappear much faster from the lungs of asbestos-exposed
workers, and differential losses of chrysotile from the lungs of
Canadian chrysotile miners, with resulting dramatic increases
in the tremolite (amphibole) chrysotile ratio, were well documented (McDonald, 1998). Recent studies have shown chrysotile to be relatively nonbiopersistent in rat lungs (Bernstein
et al., 2003a,b). Based on extensive studies with SVF, biopersistence is now seen as the most important factor dictating
pathogenicity of a fiber, and industries are manipulating the
composition of their products to obtain less biopersistent fibers.
Composition and Pathogenicity
Proinflammatory effects may arise from fibers as a result of
their reactive surface or the release of biologically active ions
such as transition metals. The central hypothesis of the
proinflammatory effects of fibers implies an important role
for oxidative stress. The oxidative stress/NF-jB pathway has
great potential importance for the production of fiber-mediated
inflammation. Transition metal–mediated oxidative stress appears to be important in the adverse effects of asbestos fibers,
advanced in the early 1980’s (Ghio et al., 1998; Hardy and
Aust, 1995; Weitzman and Graceffa, 1984). However, although
TOXICOLOGY OF CARBON NANOTUBES
11
FIG. 5. Paradigm for the role of long fibers and biopersistence in the pathogenic effects of fibers.
bioavailable transition metals may play a role in the genesis of
inflammation and DNA damage by crocidolite, which is the
most extensively tested fiber, a study of a range of pathogenic
and nonpathogenic fibers found that bioavailable iron does not
represent a generic pathway for the adverse health effects of
fibers (Fisher et al., 1998). It is possible that oxidative stress at
fiber surfaces might be important in leading to inflammation by
pathways other than via transition metals. These pathways,
although postulated (Fubini, 1996), are not well understood but
include the reduction of oxygen directly by electrons to form
hydrogen peroxide and its further reduction to hydroxyl radical
or hydrogen abstraction from an organic molecule by a free
radical to produce an organic radical (Fubini, 1996).
The relevance of the above to nanotubes relates to the fact
that nanotubes are being made increasingly longer, and so they
definitively satisfy the requirement for length in the above
paradigm when they become greater than ~ 15 lm. Likewise,
being essentially graphitic, they will not be soluble in a neutral
or mild acid pH and so are potentially biopersistent. Furthermore, their tendency to be contaminated with metals may
contribute to oxidative-induced inflammation and toxicity.
TOXICOLOGICAL CONSIDERATIONS OF NANOTUBES
Based on the above description of CNT synthesis size and
composition, coupled with the known toxicology of NP and
fibers, we can address the toxicological paradigms that could
be relevant to CNT.
Length—The Fiber Issue
If CNT are longer than about 20 lm then, analogous to
mineral and SVF fibers, if inhaled we can expect them to cause
the same types of pathology that long biopersistent fibers cause
in rats and humans if the fiber number is sufficient. These
pathologies are fibrosis, cancer, pleural changes, and mesothelioma. This current paradigm of fiber pathogenicity does not
discriminate between different compositions of long biopersistent fibers, except where composition can determine biopersistence. There are, however, two cases where long biopersistent
fiber types—erionite (Wagner et al., 1985) and SiC (Davis
et al., 1996)—showed special proclivity to cause mesothelioma
for reasons that were not easily explained, since they were not
especially long or more biopersistent than amphiboles. The
chemical basis of the extra pathogenicity of these two fibers has
not been elucidated. This suggests that some fiber types may
possess surface reactivity that imparts added pathogenicity; we
do not know if CNT have this potential. Carbon particles would
not normally be anticipated to have specially active surfaces;
however, due to their small size, NP carbon black, composed of
degenerated graphitic crystallites, is able to generate more
oxidative stress in vitro than fine carbon black (Wilson et al.,
2002). The role of nanodimensions in determining CNT
toxicity is yet to be investigated.
SWCNT are thin, but MWCNT are not much thinner than
chrysotile asbestos fibrils, which are 20 nm in diameter (Roggli
and Coin, 2003). Chrysotile fibers are considered to be
nonbiopersistent in human lungs (Bernstein et al., 2003a,b),
and so there is no experience of a workforce being potentially
exposed to a biopersistent fiber of this degree of thinness.
The strategy for assessing the pathogenicity of ‘‘long CNT’’
would, therefore, identify end points recognized from classical
fiber pathogenicity, such as durability, biopersistence, persistent inflammogenicity, and proliferation (Donaldson and Tran,
2004; Warheit et al., 1995). Chronic studies should emphasize
carcinogenic and mesothelioma end points in the lungs, and
positive controls would include amphibole asbestos. Pleural
mesothelioma is a key end point to study but is a rare tumor
12
DONALDSON ET AL.
even in animal experiments; however, hamsters have been
shown to be a species especially susceptible to mesothelioma
(Mcconnell et al., 1999) and so may have some utility. Since
some types of CNT appear to be flexible and have the ability to
bend or curl up to form a more compact particle (Maynard
et al., 2004; Shvedova et al., 2005), the rigidity of CNT may be
a factor in driving toxicity, since less rigid particles would be
more easily cleared by phagocytic cells.
Surface Area—The NP Issue
If CNT are appreciably shorter than 15–20 lm, then we may
expect them to be adequately phagocytosed by macrophages
and to act like NP in that their dominant adverse effect is likely
by virtue of their small size and large surface area. However, it
is important to also bear in mind that very small nanotubes may
be essentially ‘‘stealth particles’’ that avoid the phagocytic
defenses (Borm and Kreyling, 2004). It should be noted that the
surface area has been used as a metric for describing asbestos
fiber toxicity (Lippmann, 1988). The overall surface area of
CNT is, of course, much greater than that of a NP of similar
diameter. Therefore, on a per-particle basis, we would expect
even a ‘‘short’’ nanotube of 20 3 2000 nm to deliver 100 times
more surface area than a spherical 20-nm diameter NP. Equal
masses of the same 20-nm NP and the 20 3 2000-nm nanotubes would be composed of about a hundred times more compact NP, and so on a mass basis, the surface area dose would
be roughly similar. It seems likely that because of the difficulty
of detecting nanotubes by phase-contrast optical microscopy
(PCOM), the current standard technique for fibers, they may
be regulated on the basis of mass. Given the argument relating
to surface area, CNT number alone may not be a suitable
metric, and a surface area metric might be more appropriate
taking into account the extra surface area of short nanotubes
compared to NP.
The mechanism of the direct local adverse effects of NP in
the lungs will likely depend on the nature of the NP surface,
and the ability to cause oxidative stress and inflammation
appear to be important. The NP that are most analogous to CNT
are NP of carbon black (as opposed to diesel soot, welding
fumes, etc.). Nanoparticulate carbon black (NPCB) is composed of disordered graphite sheets and so differs from the
continuous graphitic sheet nature of the nanotube surface. As
described previously, NPCB cause oxidative stress in cells
(Stone et al., 1998) and cell-free systems (Wilson et al., 2002)
and oxidative stress–mediated cell signaling (Beck-Speier
et al., 2005; Tamaoki et al., 2004) and cause inflammation
after short-term inhalation (Gilmour et al., 2004b) and instillation exposure (Renwick et al., 2004). These effects are
highly likely driven through surface area effects as carbon
black has the same proinflammatory surface properties as
a range of low-toxicity, low-solubility (LTLS) particles that
have their acute inflammatory effects through surface area
alone (Duffin et al., 2002). Soluble leachates of NPCB had no
cellular or proinflammatory effects, again suggesting that no
metals or soluble organics are involved (Brown et al., 2000).
Therefore, we can anticipate that the CNT ‘‘skeleton’’ will
have proinflammatory potential that is driven by their surface
area, as found for other graphitic NP.
The toxicology strategy for ‘‘short CNT’’ could, therefore,
include comparison with NPCB and other LTLS particles for
their ability to cause inflammation. Studies should benchmark
CNT against the same surface area dose of LTLS to determine
whether there is extra toxicity. Oxidative stress potential and
proinflammatory signaling should be assessed.
Several in vivo instillation studies (Lam et al., 2004; Muller
et al., 2005; Shvedova et al., 2005; Warheit et al., 2004) have
compared NPCB with CNT and found the CNT to be more
pathogenic, especially with regard to granuloma formation. It
is not yet clear whether this is a result of chemical difference in
the graphitic surface or physical size of the aggregates
produced in the studies, a factor suggested by the authors of
one study (Warheit et al., 2004).
Metals
Depending on the mode of synthesis, CNT may contain
a number of toxic metals that can be viewed as contaminants in
that they are not required for the function of the nanotubes.
These metals include Co, Fe, Ni, and Mo, all of which have
documented toxic effects. Transition metals such as Fe can be
singled out as metals that are important in the toxicity of a range
of pathogenic dusts through their ability to redox-cycle and
cause oxidative stress (Donaldson et al., 1996; Ghio et al., 1999;
Kennedy et al., 1989). The potential role of these soluble metals
and matrix-bound metals should be a key strategy in determining the overall toxicity of the nanotubes and their mechanism(s).
The strategy for determining the role of metals should include
testing of soluble fractions to determine the role of metals. This
can be coupled with the use of leached and chelator-treated
nanotubes, depleted of metals, to determine the effect of metal
depletion on nanotube toxicity. The role of translocation and
redistribution of metals should also be considered.
Organics
The argument for a potential role for organics in the toxicity
of nanotubes comes from PM10 where there is a large organic
component in the NP fraction. Many of these organics are
associated with diesel soot and other combustion NP and have
the ability to generate free radicals and cause oxidative stress
(Kendall et al., 2001). The organics in PM10 (the mass of
airborne particulate material per unit volume collected by
a sampling convention with 50% efficiency for particles with
an aerodynamic diameter of 10 lm) are complex and include
combustion-derived molecules and also unburnt fuel and lube
oil (Kendall et al., 2001; Squadrito et al., 2001). The latter will,
of course, be absent in the combustion mix used for nanotube
synthesis, but airborne organics could be present along with
TOXICOLOGY OF CARBON NANOTUBES
contaminating organics on the reactants. If organics are produced in appreciable quantities, they can interact with enzyme
systems in the lungs to result in oxidative stress (Hiura et al.,
2000) via the production of superoxide (Kumagai et al., 1997),
etc. Oxidative stress may then activate oxidative stress–
responsive signaling pathways in cells, such as mitogenactivated protein kinase and NF-kB (Bonvallot et al., 2001),
and deplete antioxidant defenses (Abe et al., 2000), leading to
release of proinflammatory cytokines (Abe et al., 2000).
Strategies for assessing the role of organics would include
solvent extraction and testing of the extracts and the organicsdepleted CNT.
Aggregation
As described earlier, although MWCNT range from 10 to
100 nm in diameter, they have a strong tendency to bundle together in ‘‘ropes’’ due to van der Waals forces. Bundles typically
contain many tens of nanotubes and can be considerably longer
and wider than the nanotubes from which they are formed. This
would be a very important factor modifying toxicity.
Deposition. Aggregates might have a much larger aerodynamic diameter than singlet CNT and so could be less
respirable and could be large enough to deposit with a different
anatomic pattern in the lungs compared to the singlet CNT.
Clearance and cell activation. Macrophages may very
well be able to phagocytose a large aggregate of nanotubes
as a single entity and clear it when the same amount of singlet
CNT would be more difficult to handle. This would be evident
as a different oxidative burst or gene expression response to
a large aggregate than to singlet particles as a result of the
surface area or volume of the aggregate (Duffin et al., 2002).
All deposited particles enter into the mucus or epithelial
lining fluid, and so strategically, the role of these fluids is
important and should be addressed in toxicity studies. The
epithelial lining fluid of the centriacinar region contains
surfactant and protein that could promote dispersion of CNT,
including exfoliation or breakup of larger CNT aggregates.
Strategies for investigating the role of aggregation in in vitro
studies and studies where CNT are presented to rat lungs should
mimic the degree of aggregation seen in CNT in workplace air.
Preliminary hygiene studies suggest that there is a considerable
degree of aggregation of CNT in the air in workplace situations
(Maynard et al., 2004). However, experimental studies should
fully explore the role of aggregation status where possible. In
this regard, the use of dispersants is key and this is discussed
below for published studies.
Pathogenic Processes
The dominant mechanism for particle toxicity is oxidative
stress, which plays a role in the adverse effects of all pathogenic
particles and fibers so far studied (Donaldson and Tran, 2002).
Both NP and fibers fit this paradigm, and it is likely that CNT
13
will also cause oxidative stress, and there are some data to this
effect. Oxidative stress triggers inflammation via oxidative
stress–responsive transcription factors, and inflammation is
central to the pathogenic effects of other particles (Donaldson
and Tran, 2002). Oxidative stress from particles and inflammatory cells also causes the guanine adduct 8-hydroxy deoxyguanosine (8-OH-dG) that may be important in carcinogenesis
by particles (Schins, 2002; Schins et al., 2002). Chronic
inflammation and oxidative stress caused by ongoing exposure
lead to a number of particle-specific effects. Fibers such as
asbestos cause fibrosis and cancer (Mossman and Churg, 1998)
that could be due to the direct effects of fibers on cells or as
a result of oxidative stress from the fibers or the inflammatory
response causing 8-OH-dG adducts leading to mutation (Kane,
1996). Acute increases in ambient air pollution particles (PM10)
are linked to increases in exacerbations of airways disease and
deaths and hospitalizations for cardiovascular disease (Pope
and Dockery, 1999) that are considered to be driven by
combustion-derived NP (Donaldson et al., 2005; Lighty et al.,
2000; Neumann, 2001; Pope, 1995; Pope, III et al., 1999;
Schwartz, 2000). In chronic exposure studies, increased incidence of lung cancer is associated with living in an area with
increased PM10 (Pope, III et al., 2002), an effect that is likely
driven by combustion-derived NP (Dybdahl et al., 2004). These
same mechanisms are likely to be important in the pathogenic
effects of CNT and are the end points that should be studied.
The proclivity of asbestos to cause pleural pathology should be
borne in mind in any strategy for assessing nanotube toxicology. The sensitivity of mesothelial cells to asbestos fibers is
well known (Lechner et al., 1985), and so the consequences of
nanotube-mediated effects on mesothelial cells is warranted.
Translocation
Some NP are characterized by their ability to translocate,
from their site of deposition in the lungs to the blood (Nemmar
et al., 2001, 2002) and the brain (Oberdorster et al., 2004).
Chrysotile fibrils, which are in the order of thickness of
MWCNT (20 nm), have been detected in the blood after gut
ingestion (Weinzweig and Richards, 1983). Also, fibers have
been reported to translocate to the pleural space where their
presence is linked to the development of mesothelioma
(Gelzleichter et al., 1999; Viallat et al., 1986). There are no
data on this property for CNT, but this is clearly a potentially
important process that should be included in any CNT research
strategy, especially translocation to the pleura.
Figure 6 summarizes the likely toxicopathological pathways
that could be important in the adverse effects of CNT.
PUBLISHED STUDIES ON THE EFFECTS OF SWCNT
AND MWCNT IN TOXICOLOGICAL MODELS
Several existing studies address the potential pulmonary and
cellular toxicity of nanotubes, and yet, none have addressed
14
DONALDSON ET AL.
FIG. 6. Important characteristics of nanotubes and their potential effects on processes that could impact on adverse effects.
whether the particle or fiber paradigm is dominant. Of the
in vivo studies, none has used inhalation, and all use instillation
where the high dose and dose rate raise questions about
physiological relevance; no study has addressed the role of
length, in the sense that none has compared long (greater than
~ 20 lm) to short (< 10 lm) nanotubes. Although there are
several in vitro studies in various models and end points, no
biopersistence study has been carried out. Lam et al. (2004)
demonstrated that a single intratracheal instillation in mice
with three different types of SWCNT resulted in dosedependent ‘‘epithelioid granulomas’’ and some evidence of
interstitial inflammation. No size distribution was given for the
CNT. Quartz and carbon NP at equal mass dose were used as
controls, and the authors concluded that, on an equal mass
basis, SWCNT in the lungs were far more toxic than carbon
black and even quartz. The relative toxicity compared to quartz
was evident from histological images where the nanotubes
caused extensive granulomas and fibrous lesions, whereas the
controls did not. Based on histological assessment, inflammation was variable with the different samples of nanotubes.
However, bronchoalveolar lavage would have provided a more
accurate quantitation of inflammation. The different samples of
nanotubes had different metal contents, e.g., Fe and Ni, but this
did not result in differences in their ability to produce granulomas, suggesting that metals could not explain this effect.
Warheit et al., evaluated the acute lung toxicity of intratracheally instilled SWCNT that were 1.4 nm wide 3 > 1 lm
long and contained appreciable amounts of Ni, Co, and
amorphous carbon. In their study, the authors showed that
exposure to SWCNT produced a transient inflammation and
a non–dose-dependent accumulation of multifocal granulomas.
The toxicological significance of these granulomas was,
however, questioned as their formation was possibly due to
the instillation of a bolus of agglomerated nanotubes (Warheit
et al., 2004).
Muller et al., (2005) exposed rats intratracheally to whole or
ground MWCNT of 0.7 or 5.9 lm length, respectively, and
characterized their biological activity. These workers also
measured biopersistence at two time points. The longer,
unground nanotubes were more biopersistent than the short
nanotubes at 60 days. Although this is in keeping with the
greater biopersistence of long fibers seen in studies with
asbestos and other mineral fibers, it should be noted that these
long nanotubes were much shorter than those mineral fibers
defined as long, which approach 20 lm (Donaldson and Tran,
2004). The CNT were highly fibrogenic and inflammogenic,
being roughly equivalent to a chrysotile asbestos control (no
size distribution given), and more active than NPCB (Muller
et al., 2005). Ground nanotubes were on the whole more active
than unground nanotubes, and this may have been a result of
greater dispersion of the dose within the lungs, or grinding may
release metals previously trapped within CNT structure,
allowing them to be biologically active. An in vitro study
generally supported the in vivo study with ground being more
potent than unground in stimulating TNF-a production.
Shvedova et al. (2005) studied mice exposed to SWCNT of
99.7% weight elemental carbon and 0.23% weight iron. The
primary nanotubes were 1–4 nm in diameter, but, as delivered
by pharyngeal aspiration, two distinct particle morphologies
were observed: compact aggregates and dispersed structures.
These two morphologies were linked to two distinct responses
with dense SWCNT aggregates being associated with foci of
granulomatous inflammation, including discrete granulomas
with hypertrophic epithelial cells. In areas that did not have
TOXICOLOGY OF CARBON NANOTUBES
these aggregates, diffuse interstitial fibrosis dominated, assumed to have been caused by exposure to the dispersed
SWCNT. Control mice received NPCB or quartz, but these
treatments did not cause thickening of alveolar walls or induce
the formation of granulomas.
All the four studies had a similar design with quartz and
NPCB as controls. All showed a similar florid fibrotic response
to nanotubes that was absent with NPCB and quartz. Both CNT
and carbon black are predominantly forms of graphite. This
graphite is a rolled seamless sheet in the case of nanotubes plus
various metals, amorphous carbon, and contaminants. In the
case of carbon black, the graphite has a degenerated graphitic
crystallite structure with little in the way of contaminants
(Donaldson et al., 2005). The extended form of the graphite in
nanotubes could, therefore, be important in encouraging
fibroblast proliferation or activation of local cells to release
stimulants of fibroblast proliferation and growth; the role of the
contaminants of the nanotubes remains undetermined. None of
these studies looked at translocation to the pleura or to
extrapulmonary tissues.
No inhalation studies with nanotubes are yet available, and
the extent of the necessary investment in such an undertaking
cannot be overemphasized. A large and, therefore, expensive
quantity of CNT is needed, a representative CNT type must be
chosen, and the target concentration and the aggregations status
and the relevance to the actual workplace cloud needs also to be
considered. The inflammatory and fibrogenic effects seen so far
may be a result of large nanotube complexes being formed
during the instillation or pharyngeal aspiration procedure and
that would not be present in a respirable cloud. Such
aggregates, while pathogenic after instillation, might never
reach the periphery of the lung under normal inhalation
exposure because of their large aerodynamic size. However,
large ‘‘feathery’’ aggregates may have a large physical diameter, but, since their density is low, they may have a small
aerodynamic diameter and be respirable.
In in vitro studies, CNT have been tested in a range of different
cell types to assess their potential toxicity. Treatments of human
keratinocytes, mimicking potential dermal exposure, have shown
that both SWCNT and MWCNT are capable of localizing within
and causing cellular toxicity (Monteiro-Riviere et al., 2005;
Shvedova et al., 2003). Human T lymphocytes, a possible target
cell for nanotubes that gain access to lymphoid tissue following
interstitialization, exposed to specifically oxidized MWCNTwere
found to be killed in a time- and dose-dependent manner, with
higher concentrations inducing a marked decrease in cellular
viability via an apoptotic mechanism (Bottini et al., 2006).
Inhibition of cell function and the induction of apoptosis in vitro
were reported in kidney cells treated with SWCNT (Cui et al.,
2005). In a study with alveolar macrophages, SWCNT were more
cytotoxic than MWCNT at equal mass dose using an assay that
measures mitochondrial metabolism through the production of
a colored formazan molecule (the 3-[4,5-dimethylthiazol-2-yl]2,5-diphenyltetrazolium (1)bromide assay). However, this assay
15
is potentially complicated in macrophages, as they release reactive oxygen species during phagocytosis that could potentially
interfere with the oxidative process that generates the formazan
product in the assay (Jia et al., 2005). Manna et al. (2005) demonstrated a clear dose-dependent activation of NF-kB and oxidative
stress in human keratinocytes along with IkB depletion and
MAPK phosphorylation. The ability to cause oxidative stress and
inflammation is in clear concordance with other proinflammatory
particles such as asbestos, PM10, and quartz that activate these
pathways (Donaldson and Tran, 2002) and suggests that SWCNT
may have some similarities to other pathogenic particles.
The extended repetitive structure of long asbestos fibers has
been implicated in causing aggregation of the epidermal
growth factor receptor (EGF-R) and initiating signaling
cascades (Pache et al., 1998). The EGF-R is implicated in
the induction of inflammation and fibrosis. The ordered
graphitic structure of a long CNT might also have some ability
to cause receptor aggregation. The idea that the graphitic
structure could be important is supported by the finding that the
more the surface of SWCNT is derivatized/functionalized, the
lesser is its in vitro toxicity (Sayes et al., 2005).
Dispersion of CNT in Toxicology Studies
As discussed above, dispersion of CNT is problematic, and
a number of different surfactants have been used for this
purpose. Because of the importance of the particle surface in
toxicology, the use of surfactants needs careful consideration
and control. A number of different treatments have been used
to disperse CNT in toxicological studies. In in vivo studies,
Tween 80 has been used (Muller et al., 2005; Warheit et al.,
2004) as well as sonication in phosphate-buffered saline
(Shvedova et al., 2005) and shearing in a glass homogenizing
tube plus sonication in heat-inactivated mouse serum (Lam
et al., 2004). In in vitro studies, water or medium have been
used as the dispersant in some studies (Bottini et al., 2006; Cui
et al., 2005; Monteiro-Riviere et al., 2005), while one study
used a Dounce homogenizer, sonication, and vortexing of
particles in culture medium (Jia et al., 2005). One study
dissolved the CNT first in dimethylformamide, and control
treatments received an equivalent volume of dimethylformamide (Manna et al., 2005). Sayes et al. (2005) used ‘‘waterdispersible SWCNT’’ in their studies of functionalized CNT.
With such a wide spectrum of different pretreatments of CNT
in different assays, care needs to be taken in relating such
toxicity data to the actual toxicity of CNT in the lungs. In the
lungs, all particles encounter lung-lining fluid, which is likely
to bind to the particles and can modify the toxicity of some
particles (Brown et al., 1998; Fisher et al., 1998). No recommendations can presently be made as to the best dispersant
to use since they have not been systematically compared as to
effectiveness in causing dispersion or their toxicity.
Dispersion in air for inhalation studies is also potentially
problematic in view of the well-documented tendency of the
16
DONALDSON ET AL.
nanotubes to aggregate together. However, if the workplace
cloud is similarly aggregated, then the use of aggregated CNT
in animal inhalation experiments may be warranted as the most
relevant exposure.
REGULATION OF EXPOSURE IN WORKPLACES
Particles
Regulation intended to limit workplace exposure to chemicals, including particles, has been in place for many years in
most industrialized countries. The required approach is highly
similar in most cases. For example, in the United Kingdom, the
legislation dealing with the control of exposure to harmful
chemicals is the Control of Substances Hazardous to Health
Regulations 2002. These requirements and associated guidance
provide a step-by-step approach for employers to assess risks,
implement measures needed to control exposure, assess the
efficacy of these measures, and to establish good and safe
working practices. In the United States, the Occupational
Safety and Health Administration is responsible for developing
and enforcing workplace safety and health regulations based on
the same principles. The regulations specify control limits for
workplace air expressed in terms of mass concentration per unit
volume. Where it is appropriate to carry out assessment of
exposure, the current best practice for airborne particles is to
use a personal sampling device to collect a sample of the most
appropriate, biologically relevant fraction (HSE, 2000) of the
aerosol. It is common practice to use samplers conforming to
the inhalable or respirable convention (see below), and samples
collected via a personal sampling device are subsequently
assessed either gravimetrically or via chemical analysis to
determine the mass and hence average concentration over the
defined period. These samplers provide an estimate of timeweighted mass concentration, from which personal exposure
may be derived.
Currently, biological relevance is determined as a function of
particle size, with the appropriate metric being aerodynamic
diameter (Jones et al., 1988).
The two biologically relevant size fractions currently used
are the inhalable (< 10-lm aerodynamic diameter) and
respirable fractions (< 4.5-lm aerodynamic diameter). These
are considered to represent the fraction of airborne material
that can enter the human respiratory tract (inhalable) and can
penetrate beyond the ciliated airways (respirable). Definitions
for these fractions are well established and have international
agreement (ACGIH, 1993; CEN, 1993; ISO, 1993). Mostly,
occupational exposure limit values are defined either in terms
of inhalable or respirable, although some substances have
values for both. These well-established conventions represent
size fractions which are much larger than those normally
considered relevant to NP. One recent development, which
reflects increasing concerns about the potential risks of NP, is
the publication for comment, of a draft-recommended exposure
limit for TiO2 by NIOSH (2005). They recommend exposure
limits of 1.5 mg/m3 for fine TiO2 and 0.1 mg/m3 for ultrafine
TiO2. ‘‘Fine’’ is defined in this document as equivalent to the
respirable fraction, and ‘‘ultrafine’’ is defined as the fraction of
respirable particles with primary particle diameter < 100 nm.
Determination of the proportion of ultrafine particles in the
sample is based on transmission electron microscopy (TEM).
Potentially, this approach may be used as a basis for and
assessing compliance with standards for other NP aerosols.
Other than their use in conventions, there is little requirement to collect information about particle size of workplace
aerosols. With the exception of fibers (see below), there is no
requirement to collect information about particle number for
workplace aerosols nor is there any requirement to collect
information about particle surface area. In the case of NP,
the surface area is likely to be the most relevant metric
(Oberdorster et al., 2005a).
Fibers
In contrast to other workplace aerosols, where mass is the
metric which is used to determine regulatory compliance, fibers
are regulated in terms of particle number. The two main groups
of fibers that have been of interest for regulation in the
workplace are various types of asbestos fibers and various
types of SVF. Many countries have established personal
exposure limits for airborne fibers in workplace atmospheres
expressed in terms of fiber number concentration. The method
typically used to determine these concentrations is the socalled membrane filter method. Various specifications for the
membrane filter method have been published, and while there
are differences in detail between these specifications, the broad
thrust of each method is the same.
Publication of the World Health Organization (WHO)
method for determination of airborne fiber number concentrations (WHO, 1997) was an important attempt to harmonize
these various methods. This WHO method is expected to be the
basis of current and new regulation in most countries.
A key issue in the method is the definition of a fiber. WHO
methods define a fiber as an object with length greater than
5 lm, a width less than 3 lm, and a length to width to ratio
(aspect ratio) greater than 3:1. The method is based on PCOM
determination of the airborne fiber number concentrations,
collected on a membrane filter by means of sampling pump.
The analysis relies on manual fiber counting of a relatively
small number of fibers, and the method is widely recognized as
one of the least precise analytical techniques used in the
occupational environment, being subject to a number of
systematic and random errors as well as individual counter
bias including limit of detection issues. The minimum visible
width according to the WHO method is 0.13–0.15 3 106 m.
However, in practice the smallest visible fibers will be about
0.2–0.25 lm wide. Since some fibers will fall below the limit
17
TOXICOLOGY OF CARBON NANOTUBES
of visibility, the count represents only a certain proportion of
the total number of fibers present. Thus, the count represents
only an index of the numerical concentration of fibers and is
not an absolute measure of the number of fibers present. Fibers
falling within a graticule within the eyepiece of the microscope
are counted, and various counting rules are specified by the
WHO for fibers partially out of the field of view, for fibers
attached to particles, and for aggregates. Neither standard error
of the mean (SEM) nor TEM are routinely used for fiber
counting but can be used where more information, particularly
on fiber type, is required, as recommended by the U.K. Health
and Safety Executive (HSE, 1998).
CNT will only be visible by this method if they have
sufficient contrast and are greater than 0.2–0.25 lm wide; in
fact, this means that singlet CNT will not be countable by this
methodology. As described earlier, CNT tend to aggregate into
ropes, and these may be visible but they may not be greater than
5 lm in length, especially if, as has been observed for some
CNT, they are not rigid and so they curl up (Maynard et al.,
2004; Shvedova et al., 2005). It is clear that conventional fiber
counting, as it applies to mineral fibers, is not currently
practical for the determination of airborne CNT.
Implications for Regulation of CNT in Workplaces
Current regulations for the control of fibers in the workplace
do provide a framework by which exposure to CNT in the
workplace may be regulated. However, the use of these
regulations in controlling CNT exposure presents several
challenges. At this stage we cannot be sure how these materials
will present in the workplace. It will depend on the specific
material, the process of this manufacture or use, type of fiber,
i.e., single walled or multiwalled, and the scenario.
From what we know about MWCNT, we might expect that
these would have diameter from 20 to greater than 200 nm and
that lengths could vary from less than 1000 nm to more than
106 nm. They may be straight and partly rigid, or they may be
bent or curled and partly flexible. They could appear as single
particles or entangled as clumps or ropes. As single particles,
those which were greater than 5 lm in length would certainly
comply with the WHO definition of a fiber, having an aspect
ratio greater than 3:1. However, those with diameters less than
200 nm would not be easily detected by PCOM due to the limit
of optical detection. It is well recognized that fibers such as
asbestos have a distribution of diameters. For this reason, while
the count obtained in a PCOM analysis would also exclude
fibers less than 200 nm, it is considered that the count would
produce an acceptable index of the fibers which are present. It is
clearly the case, however, that MWCNT will have a highly
different distribution and may well be designed to have a very
narrow distribution of diameters. So, e.g., a sample of MWCNT
having a diameter of 30 nm would not be detected by PCOM.
MWCNT may also appear as clumps or ropes. In this case, if
the individual fibers were not ‘‘easily distinguishable,’’ the rope
would count as a single fiber provided that its characteristics
lay within the definition of the fiber. If the rope did not lie
within that definition, it would not be counted. This implies that
in health terms fibers within the rope will continue to behave as
a single entity as they enter the respiratory tract, which may or
may not be the case.
It is not clear whether SWCNT can appear as single entities
in workplace exposures. Exposures in scenarios for the production and use of SWCNT (and MWCNT) have not been well
investigated. In the one study of SWCNT production that has
been reported (Maynard et al., 2004), there is evidence to
suggest that it is possible for single SWCNT entities to enter the
workplace air. Much current effort and development relating to
SWCNT is geared towards producing such material. Certainly,
based on the dimensions of SWCNT we would not expect them
to be detected by PCOM; therefore, while their physical
dimensions are consistent with fibers, they would not be
counted as such by routine approaches. Due to the uniformity
of their dimensions, a count of fibers greater than 200 nm
would not provide a reliable index. In the Maynard study, much
of the SWCNT were found as clumps of materials. Although
such entities would clearly be seen by PCOM, they would not
generally be expected to comply with the definition of a fiber
and would not be counted as such.
At least some forms of CNT are expected to have a structure
which would require them to be classified as fibers. However, it
is clear that many of them would not be detected by current
optical methods. For SEM (and to some extent TEM) methods
that have been used, these are not used routinely and are not
expected as part of current regulatory processes. Use of SEM or
TEM methods would be necessary to adequately assess levels
of CNT present in workplaces. However, as these methods do
not currently benefit from the rigor and quality approaches used
in PCOM methods, much work would need to be undertaken
(e.g., to establish sample collection and preparation methods,
graticule field selection, magnification limits, counting rules,
and proficiency testing schemes) to develop appropriate
approaches by which routine assessment using these methods
would be possible.
CONCLUSION
CNT are an important class of new materials that have
numerous properties which make them useful to technology
and industry. The predicted increase in manufacture and use
makes human exposure likely, and so CNT are beginning to
come under toxicological scrutiny in order to assess the hazard
they present. This review has sought to set out the toxicological
paradigms that can be used to investigate the toxicity of
CNT, building on the toxicological database on NP and fibers.
We also describe the current workplace regulation regarding exposure to particles and fibers and make suggestions
about potential problems associated with nanotube exposure
18
DONALDSON ET AL.
measurement. We conclude that nanotubes could have features
of both NP and conventional fibers. According to the current
paradigm for fiber toxicology, based on mineral fibers and SVF,
they would pose a problem if they were longer than about 20
lm, a length so far not evaluated in any of the toxicological
studies. If CNT behave like NP, then the existing paradigm for
LTLS surfaces such as graphite is that the surface area dose
likely drives toxicity/inflammation. Such peer-reviewed literature as is currently available suggests that CNT may have
toxic effects beyond those anticipated for their mass exposure.
For example, they have more adverse effects than the same
mass of NP carbon and quartz, a commonly used yardstick of
harmful particles. They seem to have special ability to
stimulate mesenchymal cell growth. However, it should be
noted that there is, as yet, no definitive inhalation study
available that would avoid the potential for artifactual effects
due to large mats and aggregates forming during instillation
exposure procedures. CNT may have their effects through
oxidative stress and inflammation (Fig. 6), and this is borne out
by some published toxicology studies. Additionally, the studies
so far suggest that they may have an unexpected ability to cause
granuloma formation and fibrogenesis. The NP toxicology
paradigm also emphasizes the potential for NP to translocate
from their portal of entry to other tissues, and this should
remain a potential target in CNT research.
The increasing use of CNT in industry means that the safety
of those who are working with nanotubes requires consideration of the workplace regulation in the light of the peculiar
problems presented by monitoring such small materials and the
uncertainty of the nature, mechanism, and exposure response
for adverse effects. Until better information becomes available,
CNT should be considered in the same way as other biopersistent fibers in workplace risk assessments implying similar
control and assessment approaches. However, difficulties in
applying PCOM methods means that assessment based on
electron microscopy should be strongly considered. Clearly,
more research is needed to gain an insight into the mechanism
of these adverse effects and on the best ways to measure CNT
in the air in order to protect those that are exposed to this new
material.
ACKNOWLEDGMENTS
K.D. acknowledges the support of the Colt Foundation. K.D. has carried out
paid consultancy for companies in the field of nanotechnology, none of which
use or make CNT. None of the other authors has a conflict of interest.
REFERENCES
Abe, S., Takizawa, H., Sugawara, I., and Kudoh, S. (2000). Diesel exhaust
(DE)-induced cytokine expression in human bronchial epithelial cells: A
study with a new cell exposure system to freshly generated DE in vitro. Am.
J. Respir. Cell Mol. Biol. 22, 296–303.
ACGIH (1993). Threshold limit values of chemical substances and physical
agents. American Conference of Governmental Occupational Hygienists.
Cincinatti, Ohio.
Bahr, J. L., Mickelson, E. T., Bronikowski, M. J., Smalley, R. E., and Tour, J. M.
(2001). Dissolution of small diameter single-wall carbon nanotubes in
organic solvents? Chem. Commun. 2, 193–194.
Banerjee, S., Kahn, M. G., and Wong, S. S. (2003). Rational chemical strategies
for carbon nanotube functionalization. Chemistry 9, 1898–1908.
Beck-Speier, I., Dayal, N., Karg, E., Maier, K. L., Schumann, G., Schulz, H.,
Semmler, M., Takenaka, S., Stettmaier, K., Bors, W., et al. (2005). Oxidative
stress and lipid mediators induced in alveolar macrophages by ultrafine
particles. Free Radic. Biol. Med. 38, 1080–1092.
Bernstein, D., Castranova, V., Donaldson, K., Fubini, B., Hadley, J., Hesterberg,
T., Kane, A., Lai, D., Mcconnell, E. E., Muhle, H., et al. (2005). Testing
of fibrous particles: Short-term assays and strategies. Inhal. Toxicol. 17,
497–537.
Bernstein, D. M., Chevalier, J., and Smith, P. (2003a). Comparison of Calidria
chrysotile asbestos to pure tremolite: Inhalation biopersistence and histopathology following short-term exposure. Inhal. Toxicol. 15, 1387–1419.
Bernstein, D. M., Rogers, R., and Smith, P. (2003b). The biopersistence of
Canadian chrysotile asbestos following inhalation. Inhal. Toxicol. 15, 1247–
1274.
Bodian, D., and Howe, H. A. (1941). Experimental studies on intraneural
spread of poliomyelitis virus. Bull. Johns Hopkins Hosp. LXVIII,
248–267.
Bonvallot, V., Baeza-Squiban, A., Baulig, A., Brulant, S., Boland, S., Muzeau,
F., Barouki, R., and Marano, F. (2001). Organic compounds from diesel
exhaust particles elicit a proinflammatory response in human airway
epithelial cells and induce cytochrome p450 1A1 expression. Am. J. Respir.
Cell Mol. Biol. 25, 515–521.
Borm, P. J., and Kreyling, W. (2004). Toxicological hazards of inhaled nanoparticles—Potential implications for drug delivery. J. Nanosci. Nanotechnol.
4, 521–531.
Borowiak-Palen, E., Pichler, T., Liu, X., Knupfer, M., Graff, A., Jost, O.,
Pompe, W., Kalenczuk, R. J., and Fink, J. (2002). Reduced diameter
distribution of single-wall carbon nanotubes by selective oxidation. Chem.
Phys. Lett. 363, 567–572.
Bottini, M., Bruckner, S., Nika, K., Bottini, N., Bellucci, S., Magrini, A.,
Bergamaschi, A., and Mustelin, T. (2006). Multi-walled carbon nanotubes
induce T lymphocyte apoptosis. Toxicol. Lett. 160, 121–126.
Brody, A. R., Warheit, D. B., Chang, L. Y., Roe, M. W., George, G., and Hill,
L. H. (1984). Initial deposition pattern of inhaled minerals and consequent
pathogenic events at the alveolar level. Ann. N.Y. Acad. Sci. 428, 108–120.
Bronikowski, M. J., Willis, P. A., Colbert, D. T., Smith, K. A., and Smalley,
R. E. (2001). Gas-phase production of carbon single-walled nanotubes from
carbon monoxide via the HiPco process: A parametric study. J. Vac. Sci.
Technol. A Vac. Surf. Films 19, 1800–1805.
Brown, D. M., Donaldson, K., Borm, P. J., Schins, R. P., Dehnhardt, M.,
Gilmour, P., Jimenez, L. A., and Stone, V. (2004). Calcium and ROSmediated activation of transcription factors and TNF-alpha cytokine gene
expression in macrophages exposed to ultrafine particles. Am. J. Physiol.
Lung Cell. Mol. Physiol. 286, L344–L353.
Brown, D. M., Roberts, N. K., and Donaldson, K. (1998). Effect of coating with
lung lining fluid on the ability of fibres to produce a respiratory burst in rat
alveolar macrophages. Toxicol. In Vitro 12, 15–24.
Brown, D. M., Stone, V., Findlay, P., MacNee, W., and Donaldson, K. (2000).
Increased inflammation and intracellular calcium caused by ultrafine carbon
black is independent of transition metals or other soluble components.
Occup. Environ. Med. 57, 685–691.
Brumfiel, G. (2003). Nanotechnology: A little knowledge. Nature 424,
246–248.
TOXICOLOGY OF CARBON NANOTUBES
Cassell, A. M., Raymakers, J. A., Kong, J., and Dai, H. J. (1999). Large scale
CVD synthesis of single-walled carbon nanotubes. J. Phys. Chem. B 103,
6484–6492.
CEN (1993). Workplace atmospheres: size fraction definitions for measurement of airborne particles in the workplace. CEN standard EN481 (European
Committee for Standardisation) Brussels, Belgium: CEN.
Chiang, I. W., Brinson, B. E., Smalley, R. E., Margrave, J. L., and Hauge, R. H.
(2001). Purification and characterization of single-wall carbon nanotubes.
J. Phys. Chem. B 105, 1157–1161.
Cientifica (2004). Nanotubes Report 2004. Cientifica.
Coin, P. G., Roggli, V. L., and Brody, A. R. (1994). Persistence of long, thin
chrysotile asbestos fibers in the lungs of rats [review]. Environ. Health
Perspect. 102(Suppl. 5), 197–199.
Cui, D., Tian, F., Ozkan, C. S., Wang, M., and Gao, H. (2005). Effect of single
wall carbon nanotubes on human HEK293 cells. Toxicol. Lett. 155, 73–85.
Dai, H. (2002). Carbon nanotubes: Synthesis, integration, and properties. Acc.
Chem. Res. 35, 1035–1044.
Davis, J. G., Addison, J., Bolton, R. E., Donaldson, K., Jones, A. D., and
Smith, T. (1986). The pathogenicity of long versus short fiber samples of
amosite asbestos administered to rats by inhalation and intraperitoneal
injection. Br. J. Exp. Pathol. 67, 415–430.
Davis, J. M., Brown, D. M., Cullen, R. T., Donaldson, K., Jones, A., Miller, B.
G., Mcintosh, C., and Searl, A. (1996). A comparison of methods for
determining and predicting the pathogenicity of mineral fibres. Inhal.
Toxicol. 8, 747–770.
Donaldson, K., Beswick, P. H., and Gilmour, P. S. (1996). Free radical activity
associated with the surface of particles: A unifying factor in determining
biological activity? Toxicol. Lett. 88, 293–298.
Donaldson, K., Brown, G. M., Brown, D. M., Bolton, R. E., and Davis, J. G.
(1989). Inflammation generating potential of long and short fiber amosite
asbestos samples. Br. J. Ind. Med. 46, 271–276.
Donaldson, K., Jimenez, L. A., Rahman, I., Faux, S. P., MacNee, W.,
Gilmour, P. S., Borm, P. J. A., Schins, R. P. F., Shi, T., and Stone, V.
(2004a). Respiratory health effects of ambient air pollution particles:
Role of reactive species. In Oxygen/Nitrogen Radicals: Lung Injury and
Disease (V. Vallyathan, X. Shi, and V. Castranova, Eds.). Lung Biology
in Health and Disease (C. Lenfant, Exec. Ed.), Vol. 187. Marcel Dekker,
New York.
Donaldson, K., Li, X. Y., Dogra, S., Miller, B. G., and Brown, G. M.
(1992). Asbestos-stimulated tumor-necrosis-factor release from alveolar
macrophages depends on fiber length and opsonization. J. Pathol. 168,
243–248.
Donaldson, K., Stone, V., Tran, C. L., Kreyling, W., and Borm, P. J. (2004b).
Nanotoxicology. Occup. Environ. Med. 61, 727–728.
Donaldson, K., and Tran, C. L. (2002). Inflammation caused by particles and
fibres. Inhal. Toxicol. 14, 5–27.
Donaldson, K., and Tran, C. L. (2004). An introduction to the short-term
toxicology of respirable industrial fibres. Mutat. Res. 553, 5–9.
19
in the acute pulmonary inflammatory response to particles. Ann Occup. Hyg
46(Suppl. 1), 242–245.
Duncan, R. (2003). The dawning era of polymer therapeutics. Nat. Rev. Drug
Discov. 2, 347–360.
Dybdahl, M., Risom, L., Bornholdt, J., Autrup, H., Loft, S., and Wallin, H.
(2004). Inflammatory and genotoxic effects of diesel particles in vitro and
in vivo. Mutat. Res. 562, 119–131.
Eastes, W., Potter, R. M., and Hadley, J. G. (2000). Estimation of dissolution
rate from in vivo studies of synthetic vitreous fibers. Inhal. Toxicol. 12, 1037–
1054.
Ertl, G., Knozinger, H., and Weitkamp, J. (1997). Handbook of Heterogeneous
Catalysis. Wiley-VCH, Weinheim, Germany.
Fisher, C. E., Brown, D. M., Shaw, J., Beswick, P. H., and Donaldson, K.
(1998). Respirable fibres: Surfactant coated fibres release more Fe3þ than
native fibres at both pH 4.5 and 7.2. Ann. Occup. Hyg. 42, 337–345.
Fubini, B. (1996). Use of physico-chemical and cell-free assays to evaluate the
potential carcinogenicity of fibres. In Mechanisms of Fibre Carcinogenesis
(A. B. Kane, P. Boffetta, R. Saracci, and J. D. Wilbourn, Eds.), IARC
Scientific Publication No 140, pp. 35–54. IARC.
Gelzleichter, T. R., Bermudez, E., Mangum, J. B., Wong, B. A., Janszen, D. B.,
Moss, O. R., and Everitt, J. I. (1999). Comparison of pulmonary and pleural
responses of rats and hamsters to inhaled refractory ceramic fibers. Toxicol.
Sci. 49, 93–101.
Ghio, A. J., Stonehuerner, J., Dailey, L. A., and Carter, J. D. (1999). Metals
associated with both the water-soluble and insoluble fractions of an
ambient air pollution particle catalyze an oxidative stress. Inhal. Toxicol.
11, 37–49.
Ghio, A. J., Taylor, D. E., Stonehuerner, J. G., Piantadosi, C. A., and Crumbliss,
A. L. (1998). The release of iron from different asbestos structures by
hydrogen peroxide with concomitant O-2 generation. Biometals 11, 41–47.
Giles, J. (2003). Nanotechnology: What is there to fear from something so
small? Nature 426, 750.
Gilmour, M. I., O’Connor, S., Dick, C. A., Miller, C. A., and Linak, W. P.
(2004a). Differential pulmonary inflammation and in vitro cytotoxicity of
size-fractionated fly ash particles from pulverized coal combustion. J. Air
Waste Manag. Assoc. 54, 286–295.
Gilmour, P. S., Ziesenis, A., Morrison, E. R., Vickers, M. A., Drost, E. M.,
Ford, I., Karg, E., Mossa, C., Schroeppel, A., Ferron, G. A., et al. (2004b).
Pulmonary and systemic effects of short-term inhalation exposure to ultrafine
carbon black particles. Toxicol. Appl. Pharmacol. 195, 35–44.
Hardy, J. A., and Aust, A. E. (1995). Iron in asbestos chemistry and
carcinogenicity. Chem. Rev. 95, 97–118.
Health and Safety Executive (HSE) (1998). Fibres in Air. Guidance on the
Discrimination between Fibre Types in Samples of Airborne Dust on Filters
Using Microscopy, MDHS 87. Health and Safety Executive.
Health and Safety Executive (HSE) (2000). General Methods for Sampling and
Gravimetric Analysis of Respirable and Total Inhalable Dust, MDHS 14/3.
HSE Books, Sudbury, United Kingdom.
Donaldson, K., Tran, C. L., and MacNee, W. (2002). Deposition and effects
of fine and ultrafine particles in the respiratory tract. Eur. Respir. Monogr. 7,
77–92.
Helveg, S., Lopez-Cartes, C., Sehested, J., Hansen, P. L., Clausen, B. S.,
Rostrup-Nielsen, J. R., Abild-Pedersen, F., and Norskov, J. K. (2004).
Atomic-scale imaging of carbon nanofibre growth. Nature 427, 426–429.
Donaldson, K., Tran, L., Jimenez, L., Duffin, R., Newby, D. E., Mills, N.,
MacNee, W., and Stone, V. (2005). Combustion-derived nanoparticles: A
review of their toxicology following inhalation exposure. Part. Fibre Toxicol.
2, 10.
Hesterberg, T. W., Chase, G., Axten, C., Miller, W. C., Musselman, R. P.,
Kamstrup, O., Hadley, J., Morscheidt, C., Bernstein, D. M., and Thevenaz, P.
(1998). Biopersistence of synthetic vitreous fibers and amosite asbestos in
the rat lung following inhalation. Toxicol. Appl. Pharmacol. 151, 262–275.
Dreher, K. L. (2004). Health and environmental impact of nanotechnology:
Toxicological assessment of manufactured nanoparticles. Toxicol. Sci. 77,
3–5.
Hillery, A. M., Jani, P. U., and Florence, A. T. (1994). Comparative, quantitative
study of lymphoid and non-lymphoid uptake of 60 nm polystyrene particles.
J. Drug Target. 2, 151–156.
Duffin, R., Clouter, A., Brown, D., Tran, C. L., MacNee, W., Stone, V., and
Donaldson, K. (2002). The importance of surface area and specific reactivity
Hiura, T. S., Li, N., Kaplan, R., Horwitz, M., Seagrave, J. C., and Nel, A. E.
(2000). The role of a mitochondrial pathway in the induction of apoptosis
20
DONALDSON ET AL.
by chemicals extracted from diesel exhaust particles. J. Immunol. 165,
2703–2711.
Hood, E. (2004). Nanotechnology: Looking as we leap. Environ. Health
Perspect. 112, A740–A749.
Hou, P. X., Xu, S. T., Ying, Z., Yang, Q. H., Liu, C., and Cheng, H. M. (2003).
Hydrogen adsorption/desorption behavior of multi-walled carbon nanotubes
with different diameters. Carbon 41, 2471–2476.
Islam, M. F., Rojas, E., Bergey, D. M., Johnson, A. T., and Yodh, A. G. (2003).
High weight fraction surfactant solubilization of single-wall carbon nanotubes in water. Nano Lett. 3, 269–273.
ISO (1993). Air quality: particle size fraction definitions for health-related
sampling. CD7708 International Standards Organisation Geneva.
Lee, K. P., Henry, N. W., III, Trochimowicz, H. J., and Reinhardt, C. F. (1986).
Pulmonary response to impaired lung clearance in rats following excessive
TiO2 dust deposition. Environ. Res. 41, 144–167.
Lee, K. P., Trochimowicz, H. J., and Reinhardt, C. F. (1985a). Pulmonary
response of rats exposed to titanium dioxide (TiO2) by inhalation for two
years. Toxicol. Appl. Pharmacol. 79, 179–192.
Lee, K. P., Trochimowicz, H. J., and Reinhardt, C. F. (1985b). Transmigration
of titanium dioxide (TiO2) particles in rats after inhalation exposure. Exp.
Mol. Pathol. 42, 331–343.
Lighty, J. S., Veranth, J. M., and Sarofim, A. F. (2000). Combustion aerosols:
Factors governing their size and composition and implications to human
health. J. Air Waste Manag. Assoc. 50, 1565–1618.
Lin, T., Bajpai, V., Ji, T., and Dai, L. M. (2003). Chemistry of carbon nanotubes.
Aust. J. Chem. 56, 635–651.
Jani, P., Halbert, G. W., Langridge, J., and Florence, A. T. (1990). Nanoparticle
uptake by the rat gastrointestinal mucosa: Quantitation and particle size
dependency. J. Pharm. Pharmacol. 42, 821–826.
Lippmann, M. (1988). Asbestos exposure indices. Environ. Res. 46, 86–106.
Jia, G., Wang, H., Yan, L., Wang, X., Pei, R., Yan, T., Zhao, Y., and Guo, X.
(2005). Cytotoxicity of carbon nanomaterials: Single-wall nanotube, multiwall nanotube, and fullerene. Environ. Sci. Technol. 39, 1378–1383.
Lippmann, M. (1993). Biophysical factors affecting fiber toxicity. In Fiber
Toxicology (D. B. Warheit, Ed.), pp. 259–303. Academic Press, San Diego,
CA.
Jones, A. D., McMillan, C., Johnston, A. M., Mcintosh, C., Cowie, H.,
Parker, I., Donaldson, K., and Bolton, R. E. (1988). Animal Studies to Investigate the Deposition and Clearance of Inhaled Mineral Dusts. Final Report
on CEC Contract 7248/33/026 (IOM Report TM/88/05). Institute of Occupational Medicine, Edinburgh.
Liu, J., Rinzler, A. G., Dai, H. J., Hafner, J. H., Bradley, R. K., Boul, P. J.,
Lu, A., Iverson, T., Shelimov, K., Huffman, C. B., et al. (1998). Fullerene
pipes. Science 280, 1253–1256.
Jorio, A., Saito, R., Hafner, J. H., Lieber, C. M., Hunter, M., McClure, T.,
Dresselhaus, G., and Dresselhaus, M. S. (2001). Structural (n, m) determination of isolated single-wall carbon nanotubes by resonant Raman
scattering. Phys. Rev. Lett. 86, 1118–1121.
Kane, A. B. (1996). Mechanisms of mineral fibre carcinogenesis. In
Mechanisms of Fibre Carcinogenesis (A. B. Kane, P. Boffetta, R. Saracci,
and J. D. Wilbourn, Eds.), IARC Lyon, pp. 11–34.
Kendall, M., Hamilton, R. S., Watt, J., and Williams, I. D. (2001). Characteristics of selected speciated organic compounds associated with particulate
matter in London. Atmos. Environ. 35, 2483–2495.
Kennedy, G. L., and Kelly, D. P. (1993). Introduction to fiber toxicology. In Fiber
Toxicology (D. B. Warheit, Ed.), pp. 15–42. Academic Press, San Diego, CA.
Kennedy, T. P., Dodson, R., Rao, N. V., Ky, H., Hopkins, C., Baser, M., Tolley,
E., and Hoidal, J. R. (1989). Dusts causing pneumoconiosis generate OH and
produce hemolysis by acting as fenton catalysts. Arch. Biochem. Biophys.
269, 359–364.
Kitiyanan, B., Alvarez, W. E., Harwell, J. H., and Resasco, D. E. (2000).
Controlled production of single-wall carbon nanotubes by catalytic decomposition of CO on bimetallic Co-Mo catalysts. Chem. Phys. Lett. 317,
497–503.
Kreuter, J., Shamenkov, D., Petrov, V., Ramge, P., Cychutek, K., KochBrandt, C., and Alyautdin, R. (2002). Apolipoprotein-mediated transport of
nanoparticle-bound drugs across the blood-brain barrier. J. Drug Target. 10,
317–325.
Kreyling, W., Semmler, M., and Moller, W. (2004). Dosimetry and toxicology
of ultrafine particles. J. Aerosol Med. 17, 140–152.
Kumagai, Y., Arimoto, T., Shinyashiki, M., Shimojo, N., Nakai, Y.,
Yoshikawa, T., and Sagai, M. (1997). Generation of reactive oxygen species
during interaction of diesel exhaust particle components with NADPHcytochrome P450 reductase and involvement of the bioactivation in the
DNA damage. Free Radic. Biol. Med. 22, 479–487.
Lam, C. W., James, J. T., McCluskey, R., and Hunter, R. L. (2004). Pulmonary
toxicity of single-wall carbon nanotubes in mice 7 and 90 days after
intratracheal instillation. Toxicol. Sci. 77, 126–134.
Lechner, J. F., Tokiwa, T., LaVeck, M., Benedict, W. F., Banks-Schlegel, S.,
Yeager, H., Jr., Banerjee, A., and Harris, C. C. (1985). Asbestos-associated
chromosomal changes in human mesothelial cells. Proc. Natl. Acad. Sci.
U.S.A. 82, 3884–3888.
Manna, S. K., Sarkar, S., Barr, J., Wise, K., Barrera, E. V., Jejelowo, O., RiceFicht, A. C., and Ramesh, G. T. (2005). Single-walled carbon nanotube
induces oxidative stress and activates nuclear transcription factor-kappaB in
human keratinocytes. Nano. Lett. 5, 1676–1684.
Mauderly, J. L., Cheng, Y. S., and Snipes, M. B. (1990). Particle overload in
toxicological studies—Friend or foe. J. Aerosol Med. 3 (Suppl. 1), S169–S187.
Mauderly, J. L., Snipes, M. B., Barr, E. B., Belinsky, S. A., Bond, J. A., Brooks,
A. L., Chang, I. Y., Cheng, Y. S., Gillett, N. A., Griffith, W. C., et al. (1994).
Pulmonary toxicity of inhaled diesel exhaust and carbon black in chronically
exposed rats. Part I: Neoplastic and nonneoplastic lung lesions. Res. Rep.
Health Eff. Inst. 68(Pt. 1), 1–75.
Maynard, A. D., Baron, P. A., Foley, M., Shvedova, A. A., Kisin, E. R., and
Castranova, V. (2004). Exposure to carbon nanotube material: Aerosol
release during the handling of unrefined single-walled carbon nanotube
material. J. Toxicol. Environ. Health A 67, 87–107.
Maynard, A. D., and Kuempel, E. (2005). Airborne nanostructured particles
and occupational health. J. Nanopart. Res. 7, 587–614.
Maynard, R. L., and Howard, B. (Eds.). (1999). Particulate Matter: Properties
and Effects upon Health. Bios Scientific Publishers, Oxford.
Mcconnell, E. E., Axten, C., Hesterberg, T. W., Chevalier, J., Miiller, W. C.,
Everitt, J., Oberdorster, G., Chase, G. R., Thevenaz, P., and Kotin, P. (1999).
Studies on the inhalation toxicology of two fiberglasses and amosite asbestos
in the Syrian golden hamster. Part II. Results of chronic exposure. Inhal.
Toxicol. 11, 785–835.
McDonald, J. C. (1998). Mineral fibre persistence and carcinogenicity. Ind.
Health 36, 372–375.
Miller, B. G., Jones, A. D., Searl, A., Buchanan, D., Cullen, R. T., Soutar, C. A.,
Davis, J. M., and Donaldson, K. (1999). Influence of characteristics of
inhaled fibres on development of tumours in the rat lung. Ann. Occup. Hyg.
43, 167–179.
Monteiro-Riviere, N. A., Nemanich, R. J., Inman, A. O., Wang, Y. Y., and
Riviere, J. E. (2005). Multi-walled carbon nanotube interactions with human
epidermal keratinocytes. Toxicol. Lett. 155, 377–384.
Moore, V. C., Strano, M. S., Haroz, E. H., Hauge, R. H., Smalley, R. E.,
Schmidt, J., and Talmon, Y. (2003). Individually suspended single-walled
carbon nanotubes in various surfactants. Nano Lett. 3, 1379–1382.
Mossman, B. T., and Churg, A. (1998). Mechanisms in the pathogenesis of
asbestosis and silicosis. Am. J. Respir. Crit. Care Med. 157(5 Pt. 1), 1666–
1680.
TOXICOLOGY OF CARBON NANOTUBES
Motta, M., Li, Y. L., Kinloch, I., and Windle, A. (2005). Mechanical properties
of continuously spun fibers of carbon nanotubes. Nano Lett. 5, 1529–1533.
Muller, J., Huaux, F., Moreau, N., Misson, P., Heilier, J. F., Delos, M.,
Arras, M., Fonseca, A., Nagy, J. B., and Lison, D. (2005). Respiratory toxicity of multi-wall carbon nanotubes. Toxicol. Appl. Pharmacol. 207, 221–231.
Nemmar, A., Hoet, P. H., Vanquickenborne, B., Dinsdale, D., Thomeer, M.,
Hoylaerts, M. F., Vanbilloen, H., Mortelmans, L., and Nemery, B. (2002).
Passage of inhaled particles into the blood circulation in humans. Circulation
105, 411–414.
Nemmar, A., Vanbilloen, H., Hoylaerts, M. F., Hoet, P. H., Verbruggen, A., and
Nemery, B. (2001). Passage of intratracheally instilled ultrafine particles
from the lung into the systemic circulation in hamster. Am. J. Respir. Crit.
Care Med. 164, 1665–1668.
Neumann, H. G. (2001). Health risk of combustion products: Toxicological
considerations. Chemosphere 42, 473–479.
NIOSH (2005). Evaluation of health hazard and recommendations for
occupational exposure to titanium dioxide. Draft NIOSH Current Intelligence Bulletin. Federal Register 70, 250.
Oberdorster, G., Maynard, A., Donaldson, K., Castranova, V., Fitzpatrick, J.,
Ausman, K., Carter, J., Karn, B., Kreyling, W., Lai, D., et al. (2005a).
Principles for characterizing the potential human health effects from
exposure to nanomaterials: Elements of a screening strategy. Part. Fibre
Toxicol. 2, 8.
Oberdorster, G., Oberdorster, E., and Oberdorster, J. (2005b). Nanotoxicology:
An emerging discipline evolving from studies of ultrafine particles. Environ
Health Perspect. 113, 823–839.
Oberdorster, G., Sharp, Z., Elder, A. P., Gelein, R., Kreyling, W., and Cox, C.
(2004). Translocation of inhaled ultrafine particles to the brain. Inhal.
Toxicol. 16, 437–445.
O’Connell, M. J., Bachilo, S. M., Huffman, C. B., Moore, V. C., Strano, M. S.,
Haroz, E. H., Rialon, K. L., Boul, P. J., Noon, W. H., Kittrell, C., et al. (2002).
Band gap fluorescence from individual single-walled carbon nanotubes.
Science 297, 593–596.
Pache, J. C., Janssen, Y. M., Walsh, E. S., Quinlan, T. R., Zanella, C. L., Low,
R. B., Taatjes, D. J., and Mossman, B. T. (1998). Increased epidermal growth
factor-receptor protein in a human mesothelial cell line in response to long
asbestos fibers. Am. J. Pathol. 152, 333–340.
21
Renwick, L. C., Brown, D., Clouter, A., and Donaldson, K. (2004). Increased
inflammation and altered macrophage chemotactic responses caused by two
ultrafine particle types. Occup. Environ. Med. 61, 442–447.
Renwick, L. C., Donaldson, K., and Clouter, A. (2001). Impairment of alveolar
macrophage phagocytosis by ultrafine particles. Toxicol. Appl. Pharmacol.
172, 119–127.
Roggli, V. L., and Coin, P. G. (2003). Mineralogy of asbestos. In Pathology of
Asbestos-related Diseases, 2nd ed. (V. L. Roggli, T. D. Oury, and T. A.
Sporn, Eds.), pp. 1–16. Springer, New York.
The Royal Society and the Royal Academy of Engineering (2004). Nanoscience and Nanotechnologies: Opportunities and Uncertainties. London:
The Royal Society.
Sayes, C. M., Liang, F., Hudson, J. L., Mendez, J., Guo, W., Beach, J. M.,
Moore, V. C., Doyle, C. D., West, J. L., Billups, W. E., et al. (2005).
Functionalization density dependence of single-walled carbon nanotubes
cytotoxicity in vitro. Toxicol. Lett. 161, 135–142.
Schins, R. P. (2002). Mechanisms of genotoxicity of particles and fibers. Inhal.
Toxicol. 14, 57–78.
Schins, R. P. F., and Donaldson, K. (2000). Nuclear factor kappa B activation by
particles and fibres. Inhal. Toxicol. 12(Suppl. 3), 317–326.
Schins, R. P., Knaapen, A. M., Cakmak, G. D., Shi, T., Weishaupt, C., and
Borm, P. J. (2002). Oxidant-induced DNA damage by quartz in alveolar
epithelial cells. Mutat. Res. 517, 77–86.
Schwartz, J. (2000). Daily deaths are associated with combustion particles
rather than SO(2) in Philadelphia. Occup. Environ. Med. 57, 692–697.
Searl, A., Buchanan, D., Cullen, R. T., Jones, A. D., Miller, B. G., and Soutar,
C. A. (1999). Biopersistence and durability of nine mineral fibre types in rat
lungs over 12 months. Ann. Occup. Hyg. 43, 143–153.
Seaton, A., and Donaldson, K. (2005). Nanoscience, nanotoxicology, and the
need to think small. Lancet 365, 923–924.
Shvedova, A. A., Castranova, V., Kisin, E. R., Schwegler-Berry, D., Murray,
A. R., Gandelsman, V. Z., Maynard, A., and Baron, P. (2003). Exposure to
carbon nanotube material: Assessment of nanotube cytotoxicity using human
keratinocyte cells. J. Toxicol. Environ. Health A 66, 1909–1926.
Peigney, A., Laurent, C., Flahaut, E., Bacsa, R. R., and Rousset, A. (2001).
Specific surface area of carbon nanotubes and bundles of carbon nanotubes.
Carbon 39, 507–514.
Shvedova, A. A., Kisin, E. R., Mercer, R., Murray, A. R., Johnson, V. J.,
Potapovich, A. I., Tyurina, Y. Y., Gorelik, O., Arepalli, S., SchweglerBerry, D., et al. (2005). Unusual inflammatory and fibrogenic pulmonary
responses to single-walled carbon nanotubes in mice. Am. J. Physiol. Lung
Cell. Mol. Physiol. 289, L698–L708.
Pinnell, S. R., Fairhurst, D., Gillies, R., Mitchnick, M. A., and Kollias, N.
(2000). Microfine zinc oxide is a superior sunscreen ingredient to microfine
titanium dioxide. Dermatol. Surg. 26, 309–314.
Sinnott, S. B., Andrews, R., Qian, D., Rao, A. M., Mao, Z., Dickey, E. C., and
Derbyshire, F. (1999). Model of carbon nanotube growth through chemical
vapor deposition. Chem. Phys. Lett. 315, 25–30.
Pope, C. A. (1995). Combustion source paniculate air pollution and human
health causal associations or confounding? Paniculate matter: health and
regulatory issues (VIP.49) Proceedings, International Specialty Conference
hosted by Air and Waste Management Association, 4–6 April, 1995,
Pittsburgh, PA, pp. 60–77.
Squadrito, G. L., Cueto, R., Dellinger, B., and Pryor, W. A. (2001). Quinoid
redox cycling as a mechanism for sustained free radical generation by
inhaled airborne particulate matter. Free Radic. Biol. Med. 31, 1132–
1138.
Pope, C. A., III, Burnett, R. T., Thun, M. J., Calle, E. E., Krewski, D., Ito, K.,
and Thurston, G. D. (2002). Lung cancer, cardiopulmonary mortality, and
long-term exposure to fine particulate air pollution. JAMA 287, 1132–1141.
Pope, C. A., and Dockery, D. W. (1999). Epidemiology of particle effects. In
Air Pollution and Health (S. T. Holgate, J. M. Samet, H. S. Koren, and R. L.
Maynard, Eds.), pp. 673–705. Academic Press, San Diego, CA.
Pope, C. A., III, Hill, R. W., and Villegas, G. M. (1999). Particulate air pollution
and daily mortality on Utah’s Wasatch Front. Environ. Health Perspect. 107,
567–573.
Puretzky, A. A., Schittenhelm, H., Fan, X. D., Lance, M. J., Allard, L. F., and
Geohegan, D. B. (2002). Investigations of single-wall carbon nanotube
growth by time-restricted laser vaporization. Phys. Rev. B 65, 245425-1 to
245425-9.
Stone, V., Shaw, J., Brown, D. M., MacNee, W., Faux, S. P., and Donaldson,
K. (1998). The role of oxidative stress in the prolonged inhibitory effect
of ultrafine carbon black on epithelial cell function. Toxicol. In Vitro 12,
649–659.
Tamaoki, J., Isono, K., Takeyama, K., Tagaya, E., Nakata, J., and Nagai, A.
(2004). Ultrafine carbon black particles stimulate proliferation of human
airway epithelium via EGF receptor-mediated signaling pathway. Am. J.
Physiol. Lung Cell. Mol. Physiol. 287, L1127–L1133.
Tan, M. H., Commens, C. A., Burnett, L., and Snitch, P. J. (1996). A pilot study
on the percutaneous absorption of microfine titanium dioxide from
sunscreens. Australas. J. Dermatol. 37, 185–187.
Thess, A., Lee, R., Nikolaev, P., Dai, H. J., Petit, P., Robert, J., Xu, C. H., Lee,
Y. H., Kim, S. G., Rinzler, A. G., et al. (1996). Crystalline ropes of metallic
carbon nanotubes. Science 273, 483–487.
22
DONALDSON ET AL.
Tran, C. L., Buchanan, D., Cullen, R. T., Searl, A., Jones, A. D., and
Donaldson, K. (2000). Inhalation of poorly soluble particles. II. Influence
of particle surface area on inflammation and clearance. Inhal. Toxicol. 12,
1113–1126.
Viallat, J. R., Raybuad, F., Passarel, M., and Boutin, C. (1986). Pleural
migration of chrysotile fibers after intratracheal injection in rats. Arch.
Environ. Health 41, 282–286.
Vinciguerra, V., Buonocore, F., Panzera, G., and Occhipinti, L. (2003). Growth
mechanisms in chemical vapour deposited carbon nanotubes. Nanotechnology 14, 655–660.
Wagner, J. C., Skidmore, J. W., Hill, R. J., and Griffiths, D. M. (1985). Erionite
exposure and mesotheliomas in rats. Br. J. Cancer 51, 727–730.
Weitzman, S. A., and Graceffa, P. (1984). Asbestos catalyzes hydroxyl and
superoxide radical generation from hydrogen peroxide. Arch. Biochem.
Biophys. 228, 373–376.
Wilson, M. R., Lightbody, J. H., Donaldson, K., Sales, J., and Stone, V. (2002).
Interactions between ultrafine particles and transition metals in vivo and
in vitro. Toxicol. Appl. Pharmacol. 184, 172–179.
World Health Organization (WHO) (1997). Determination of Airborne Fibre
Number Concentrations: A Recommended Method by Phase Contrast
Optical Microscopy. World Health Organization, Geneva.
WHO IARC (2002). IARC monographs on the evaluation of carcinogenic
risks to humans. Man-Made Vitreous Fibres, Vol. 81. Lyon: IARC.
Walters, D. A., Ericson, L. M., Casavant, M. J., Liu, J., Colbert, D. T., Smith,
K. A., and Smalley, R. E. (1999). Elastic strain of freely suspended singlewall carbon nanotube ropes. Appl. Phys. Lett. 74, 3803–3805.
Ye, J., Shi, X., Jones, W., Rojanasakul, Y., Cheng, N., Schwegler-Berry, D.,
Baron, P., Deye, G. J., Li, C., and Castranova, V. (1999a). Critical role of
glass fiber length in TNF-alpha production and transcription factor activation
in macrophages. Am. J. Physiol. 276(3 Pt. 1), L426–L434.
Warheit, D. B., Driscoll, K. E., Oberdoerster, G., Walker, C., Kuschner, M., and
Hesterberg, T. W. (1995). Contemporary issues in fiber toxicology. Fundam.
Appl. Toxicol. 25, 171–183.
Ye, Y., Ahn, C. C., Witham, C., Fultz, B., Liu, J., Rinzler, A. G., Colbert, D.,
Smith, K. A., and Smalley, R. E. (1999b). Hydrogen adsorption and cohesive
energy of single-walled carbon nanotubes. Appl. Phys. Lett. 74, 2307–2309.
Warheit, D. B., Laurence, B. R., Reed, K. L., Roach, D. H., Reynolds, G. A.,
and Webb, T. R. (2004). Comparative pulmonary toxicity assessment of
single-wall carbon nanotubes in rats. Toxicol. Sci. 77, 117–125.
Yu, M. F., Files, B. S., Arepalli, S., and Ruoff, R. S. (2000). Tensile loading of
ropes of single wall carbon nanotubes and their mechanical properties. Phys.
Rev. Lett. 84, 5552–5555.
Weinzweig, M., and Richards, R. J. (1983). Quantitative assessment of
chrysotile fibrils in the bloodstream of rats which have ingested the mineral
under different dietary conditions. Environ. Res. 31, 245–255.
Zhang, M. F., Yudasaka, M., Koshio, A., and Iijima, S. (2002). Thermogravimetric analysis of single-wall carbon nanotubes ultrasonicated in monochlorobenzene. Chem. Phys. Lett. 364, 420–426.