Toxicity and regulatory perspectives of carbon nanotubes
In 2010, the global capacity for production of carbon nanotubes was conservatively estimated at 350 tonnes/year.1 With a substantial increase in manufacture, handling, use and disposal forecast, increasing human and environmental exposure is inevitable, and, as a result, carbon nanotubes (CNTs) in their many forms are beginning to increasingly come under toxicological scrutiny. The enormous number of permutations and combinations of CNTs being handled in the materials’ discovery, production and biomedical or industrial application phases has led to a daunting bottleneck in relation to ensuring human and environmental safety. After a brief discussion of public perceptions and risk, this chapter summarizes many of the current CNT toxicity studies and provides a breakdown of the various CNT parameters thought to influence their toxicological properties. Subsequent sections give perspectives on potential future biological applications of CNTs, and on the shifting trends in regulatory frameworks and knowledge gaps which need to be addressed in order to ensure the safe and sustainable utility of CNTs in a range of important applications.
20.1 Toxic effects of nanomaterials and nanoparticles: public perception and the necessary ‘risk-versus-reward’ debate
Current public commentary and debate on the risk governance of nanotechnology and nanomaterials have drawn polarized opinions, with some bodies calling for a ban or moratorium on the research, development and sale of nanomaterials and nanoscale products. Carbon nanotubes (CNTs) feature most prominently in this debate, both in the scientific literature and in the media. Due to their fibrous morphology, coupled with insolubility in the lungs, they have drawn natural comparisons with asbestos. Recent high impact findings provide evidence to support the hypothesis that an asbestos-like biological response is possible, and have strongly urged that inhalation during handling should be strictly minimized until further studies establish quantitative workplace exposure limits.2−4 These cautionary studies are balanced by many promising examples and applications where CNTs have already engendered considerable improvements to health, information technology, energy, environment and aerospace sectors. As is the case with many rapidly evolving new technologies, our understanding of human health and environmental effects is incomplete. There are several excellent articles which articulate this dilemma,5−7 and while the in-vitro and in-vivo toxicology and epidemiological studies steadily accumulate and serve to inform regulatory legislation, it appears that one of the biggest hurdles to be overcome is in occupational medicine, where, for each specific nano-object of interest, new and robust methodologies for measuring, detecting and tracking nanomaterials need to be developed. Only then can the complex relationship between long-term cumulative workplace and environmental exposure and acute/chronic molecular, cellular and pathogenic toxicological mechanisms be understood. Thankfully, based on our current awareness of issues such as air pollution and asbestos, those researchers and industries working with nanomaterials have largely taken a conservative approach to handling these materials, thereby inherently minimizing the risks.
From a manufacturing perspective, a dilemma of a different kind faces manufacturers of nano-enabled products. In an excellent article, Richman8 explained that in order for nanotechnology discoveries to progress fully to mature applications, more efficient high throughput characterization tools with multiple capabilities are required in order to assure customers of quality, performance and safety. This relationship between the required level of characterization and stage of commercialization is shown in Fig. 20.1 for a new technology introduced in a mature field (e.g. semiconductor industry) contrasted with new nanotechnology implementation, showing a much slower rate of decrease in metrology required for the nanomaterial-enabled product assurance. This relationship goes some way towards explaining the tremendous lag between early stage nanotechnologies and workable regulatory frameworks. This is particularly pertinent for CNTs, which are arguably the most inherently variable nanomaterials available.
20.1 Characterization level α as a function of known properties. Heavy characterization is necessary to establish existence, synthesis and properties. Production characterization is mostly concerned with batch-to-batch variation and maintaining advertised functions or properties. Ideally, characterization of developed materials for application is limited to quality control of the applied device. In a mature field (dotted line), this function drops off rapidly between stages of development. Generally, nanotechnology (dashed line) is not achieving the drop-off necessary for profitable deployment. The value of addressing this bottleneck is the impact that faster, facile and standardized characterization can make in developing nanotechnology.
For CNT-reinforced composites, studies also need to be done to ascertain downstream end-user exposure. For example, several manufacturers of composite golf shafts now incorporate small percentages of CNTs in the resin to improve the toughness, ‘feel’ and playability of these golf shafts. Rather than buy clubs off the shelf, it is commonplace in the industry for players to employ a club professional or club fitter to trim the tip of a graphite golf shaft in order to adjust its ‘flex profile’ and ‘kickpoint’ to match the player’s particular swing speed, tempo, etc. Are free CNTs released in the dust during this cutting process, or are the dust particles essentially the same as when CNTs are not incorporated? Should the club fitter be wearing more suitable personal protective equipment or be using a dust extraction unit? There are many of these questions of a pragmatic nature that still remain unanswered.
For a better understanding of CNTs’ toxicological response, a number of studies9−27 have explored CNT toxicology properties in comparison with other carbon-based and mineral materials. These studies found that single-walled carbon nanotubes (SWCNTs)9,10 and multi-walled carbon nanotubes (MWCNTs)11−14 interact with cells, and reduce cell activity in a dose-dependent and size-dependent manner. In comparison with silver (Ag) nano-sized particles, Shin et al.11 found that nano Ag materials increased apoptosis and cytokine expression (IL-6 and IL-8) of human bronchial epithelial cells (BEAS-2B) when applied over 100 μg/ml concentration, while MWCNTs (diameters: 4 ~ 6 and 10 ~ 15 nm) weakly increased apoptosis and did not significantly increase cytokine expression in concentration less than 200 μg/ml.
Carbon nanomaterials are most commonly used as control materials.10,12,13,15,21,28−30 These nanomaterials are all made of carbon atoms but with distinct geometries. They provide an opportunity to explore the comparative toxicity particularly associated with geometrical structure and physical properties.29 Thus, Walker et al.10 examined the effects of purified single-walled (BET average 56 m2/g; Fe content 0.27% by weight) and multi-walled (a specific surface area (BET) average 641 m2/g; Fe content 8.8% by weight) carbon nanotubes on human aortic endothelial cells by evaluating actin filament integrity and VE-cadherin distribution by fluorescence microscopy, membrane permeability by measuring the lactate dehydrogenase (LDH) release, proliferation/viability by WST-1 assay, and overall functionality by tubule formation assay. It was observed that marked actin filament and VE-cadherin disruption, cytotoxicity, and reduced tubule formation occurred consistently at 24 hrs post-exposure to the highest concentrations [50–150 μg/106 cells (1.5–4.5 μg/ml)] for both SWCNTs and MWCNTs tested in their studies. Researchers did not observe these effects with carbon black exposure and CNT exposure in lower concentrations [1–10 μg/106 cells (0.04–0.4 μg/ml)] or in any tested concentrations at 3 hrs post-exposure.
The cytotoxicity of highly-purified MWCNTs obtained from catalytic chemical vapour deposition (average diameter, 67 nm; surface area, 26 m2/g; carbon purity, 99.79 wt%; iron impurity, ca. 2000 ppm; fibre length not specified)22,23 was found to be higher than that of crocidolite (also known as blue asbestos, the most hazardous of the amphibole asbestos family). Thus, mouse macrophages were exposed to MWCNTs or to Union Internationale Contre le Cancer (UICC) crocidolite in order to evaluate the toxicity of these nano-size fibres. The cytotoxicity of MWCNTs was found to be higher than that of crocidolite. Several proteins were found to adsorb onto MWCNTs when MWCNT-exposed macrophages were gently analysed. One of these proteins was macrophage receptor with collagenous structure (MARCO). MARCO-transfected CHO-K1 cells associated with MWCNTs more rapidly than mock-transfected cells. These results indicate that MWCNTs probably trigger cytotoxic effects in phagocytotic cells by reacting with MARCO on the plasma membrane and rupturing the plasma membrane. Studies using SWCNTs (High Pressure CO derived (HiPco), 10% iron catalyst residue, diameter 0.8–1.2 nm, average length 800 nm and bundle size of 2.6 × 1014 m2) evaluated the inflammatory response of immortalized and primary human lung epithelial cells (A549 and NHBE) with a special focus on the mediating role of dipalmitoylphosphatidylcholine (DPPC, the main component of lung lining fluid) on particle toxicity. It was revealed that DPPC helped improve the degree of SWCNT dispersion in A549 medium and in turn, increased particle toxicity, It was concluded that effective dispersion of the CNTs in biologically relevant media are important for the accuracy of the toxicological data.22,23
Other studies have shown that with respect to certain metal oxide nanoparticles (TiO2, Al2O3), MWCNTs were more toxic to A549 human pneumocytes,24 while CuO and CuZnFe2O4 nanoparticles were found to cause far greater cytotoxicity and DNA damage than carbon nanoparticles and MWCNTs.25 Both CNTs and metal oxide nanoparticles were able to rapidly enter A549 cells, distributes in the cytoplasm and intracellular vesicles.
Because it is generally agreed that CNT respiratory exposure is the route of highest concern, rodent studies looking at lung injury, inflammation and early signs of tumour formation are perhaps the best indicators of CNT-induced health risks.
There are various in vivo studies focused on assessing the potential impact of both unrefined and pristine CNTs on the lungs.15,21,30−33 Despite the first study showing no initial indication of lung toxicity,31 subsequent studies have found histological evidence of lung inflammation and granuloma formation.15,21,32,33 The initial study in 2001 of Huczko et al.31 investigated the effects of unrefined CNT on the pulmonary function of guinea pigs. The authors did not indicate whether the CNTs used were SWCNTs or MWCNTs. Guinea pigs were given a single intratracheal instillation (IT) with 25 mg of unrefined CNT in 0.5 mL of saline solution, while control animals received 25 mg of CNT-free soot. Small amounts of surfactant were used to help disperse the nanotubes. Pulmonary function was investigated via non-invasive procedures 4 weeks after instillation. The animals were sacrificed at this time for further study. Neither the non-invasive procedures, nor the final bronchoalveolar lavage examination (BAL) showed any difference between the exposed or control groups, and it was concluded that ‘working with soot containing CNT is unlikely to be associated with any health risk’.31 Subsequent detailed studies into CNT lung toxicity15,21 in 2004 were the first to present data raising concerns over potential CNT health hazards. Lam et al.21 investigated the pulmonary toxicity of three batches of SWCNTs (unrefined and purified samples from iron catalyzed HiPco synthesis and unrefined SWCNTs, containing nickel (25.99% by Wt) and yittrium (5.01% by Wt) synthesized via electrical arc discharge, dimensions not reported) and compared them to carbon black and quartz particles. Mice were again exposed via IT and then observed for either 7 or 90 days at which point they were sacrificed for histological examination. All three SWCNT products induced dose-dependent lung lesions, characterized by interstitial granulomas, regardless of the levels of metal impurities, although it was noted that the SWCNTs containing Ni produced a higher mortality rate. The purified CNTs were prepared by rigorous treatment (45-h reflux) with concentrated acids (2 M to 3 M nitric acid) to remove metal impurities. Although the procedure reduced the metal impurity content to 2% by weight, these harsh conditions can also modify the surface of the CNTs and introduce defects, which have not been investigated. Significantly, this study concluded that SWCNT were more toxic than carbon black, and CNT containing Ni was more toxic than quartz, the recognized positive inducer of lung toxicity.
Warheit et al.15 investigated the lung toxicity of SWCNTs in rats intratracheally instilled. The study exposed rats to unrefined SWCNTs (made by laser ablation, and containing 30–40% amorphous carbon and ~ 5% Co and ~ 5% Ni catalyst. The nominal size of these SWCNTs was 1.4 nm in diameter and more than 1 μm in length), quartz (as a positive control), carbonyl iron particles (as a negative control), graphite particles (with the same percentage Ni/Co as the SWCNT sample) and phosphate buffered saline solution. The rats were studied using BAL fluid markers, cell proliferation assays and histopathological examination at 24 h, 1 week, 1 month and 3 months after instillation. Initially, a mortality rate of approximately 15% was observed in rats exposed to 5 mg/kg SWCNT. However, this was later attributed to mechanical blockage of the upper airways causing asphyxiation, rather than toxicity of the SWCNT. Researchers found that exposure to SWCNT produced only transient inflammation, as assessed by cell proliferation and cytotoxicity indices. Histological examination of exposed animals identified a series of non-dose-dependent multifocal granulomas. The granulomas were non-uniform in distribution and not progressive after 1 month. The presence of the granulomas was considered inconsistent with the lack of lung toxicity as determined by other parameters and the authors concluded that more research was needed.
Following their preliminary study in 2001,31 Huczko et al.32 published a follow-up study in 2005. In this investigation, five different samples of MWCNT (unrefined MWCNTs produced by both CVD and arc-discharge, and commercially available CNTs from a number of suppliers) were intratracheally instilled into guinea pigs. The animals were studied using BAL markers, lung resistance tests and histopathological examination at 90 days. Unlike the preliminary study, significant evidence of pulmonary toxicity was observed. These researchers found alveolar macrophage infiltration in the BAL of most of the non-control animals. Abnormal lung resistance was observed in all animals exposed to CNT samples, while lung histology reported multiple lesions in all CNT-exposed animals. The authors concluded that the exposure time was critical for induction of lung pathology.
Muller et al.33 investigated the effect of intratracheally instilled MWCNTs on the pulmonary function of rats. Exposure to purified and milled (via oscillatory agate ball milling to reduce aggregation) MWCNTs (see Table 20.1) (CVD synthesis from ethylene over alumina support with Co and Fe catalyst, purified in NaOH) and monitored for up to 60 days. Asbestos and carbon black were used as controls. Rats were studied using BAL, inflammatory and fibrotic markers, biopersistance tests, and histopathological examination. Dose-dependent inflammation and granuloma formation was observed, but unlike previous studies,15 the inflammation persisted for the full 60 days. In all cases, both MWCNTs and milled MWCNTs were considered more inflammatory than carbon black, but less inflammatory than the asbestos fibers. In addition, the shorter milled MWCNTs were cleared faster than non-ground MWCNTs. The authors urged the introduction of appropriate safety measures for handling CNTs, while calling for more studies to accurately establish the toxicology of CNTs.
Characteristics of carbon nanotube (CNT) samples before and after agate milling33
|Length (μm)||5.9 ± 0.05||0.7 ± 0.07|
|Average inner diameter (nm)||5.2 ± 1.5||5.1 ± 2.1|
|Average outer diameter (nm)||9.7 ± 2.1||11.3 ± 3.9|
|Specific surface area (m2/g)||378 ± 20||307 ± 15|
|Oxidized forms (atomic %)||13.7 ± 0.7||13.1 ± 0.7|
|Carbon content (%)||97.8 ± 0.2||98.0 ± 0.2|
In the only study of its kind to date, Shvedova et al.34 investigated the pharyngeal aspiration of SWCNTs (synthesized via the HiPco process and purified to > 99%, Fe(CO)5 catalyst, purified by acid treatment, diameter 1–4 nm) in mice. The study found that SWCNT exposure (10, 20, 40 μg/animal) led to a dose-dependent increase in inflammatory markers at 1–3 days post exposure, granuloma formation and progressive interstitial fibrosis and alveolar wall thickening up to 60 days post exposure. Contrary studies utilizing intratracheal instillation for delivery,15,21,31−33 pharyngeal aspiration resulted in two distinct particle morphologies. The first particle morphology observed, compact SWCNT aggregates (> 500 nm in diameter), was associated with acute inflammation and granuloma formation at the particle deposition sites. In contrast, the second particle morphology, dispersed SWCNT structures, was associated with diffuse interstitial fibrosis and alveolar wall thickening in areas further away from SWCNT aggregate. The researchers again concluded that more extensive inhalation studies were required to confirm these initial observations.
All studies that have utilized intratracheal instillation have reported major difficulties dealing with the agglomerative nature of CNTs in aqueous solutions.15,21,32,33 Although this method carries the advantage of being able to deliver a specific dose to the lung epithelium, it does not resemble exposure by natural respiration. There is also evidence that it takes significant energy and agitation to release fine CNT particles into the air.35
In somewhat of a landmark study in 2008, Poland et al.2 used a well-respected peritoneal cavity model (often used as a surrogate assay for lung toxicity) to demonstrate length-dependent MWCNT pathogenicity, reporting the existence of granulomas on the surface of the diaphragm in cases where mice were exposed to high aspect ratio MWCNTs (‘NTlong1’ from Mitsui, 85 nm diameter, 24% of bundles longer than 15 μm, residual Cu, Ni and Co catalyst (all < 6%), and, ‘NTlong2’ from University of Manchester, 165 nm diameter, 84% of bundles longer than 15 μm, residual Fe (37%), Cu, Ni and Co catalyst (Cu, Ni and Co all < 6%)). These were compared to results where mice were exposed to low aspect ratio MWCNTs (‘NTtang1’ and ‘NTtang2’) consisting of CNTs arranged in low-aspect-ratio tangled aggregates (see Table 20.2). Motivated by this study, in the first published study to employ ‘the gold standard’3 of inhalation exposure, Ryman-Rasmussen et al.4 reported results from a study where mice were allowed to inhale (30 mg/m3) MWCNT for 6 hours, after which their lung tissues were harvested and studied. MWCNTs were found throughout the lungs and within macrophages, and most importantly, they were seen deposited in the sub-pleural wall close to the very sensitive mesothelial tissue. Also significant was the lack of fibrosis observed when a lower inhalation exposure (1 mg/m) was employed; suggesting again that the minimization of inhalation during handling is very wise until further long-term inhalation exposure data comes to hand.
To date, several studies have compared toxicology properties of different CNT types: SWCNTs and MWCNTs. Most researchers36−40 concluded that the size (diameter) of carbon nanotubes is a key factor governing their antibacterial effects and that it is likely the main CNT-cytotoxicity mechanism is cell membrane damage by direct contact with CNTs.
Kang et al.36,37 studied the microbial cytotoxicity of four carbon-based nanomaterials [SWCNT ((Stanford Materials), purified (95% (w/w)), MWCNT ‘as-prepared’, CVD synthesized using iron catalyst (NanoTechLabs)], dry oxidized MWCNTs in air at 350 °C to remove amorphous carbon and ‘acid-treated’ MWCNTs [10 M HCl at 70 °C for 18 hrs, and finally, ‘functionalized’, via sonication in H2SO4/HNO3 (3:1 v/v)], aqueous phase of fullerene (C60) nanoparticles, and colloidal graphite in Gram negative and Gram positive bacteria. Thus, SWCNTs inactivated the highest percentage of cells in monocultures of Escherichia coli, Pseudomonas aeruginosa, Bacillus subtilis, and Staphylococcus epidermis, as well as in the diverse microbial communities of river water and wastewater effluent.37
Other results38,40 supported the finding that SWCNTs are more toxic than MWCNTs. De Nicola et al.38 explored the toxicity of different types of carbon nanotubes (see Table 20.3), differing in preparation (arc discharge versus catalysed chemical vapour deposition); size (10–50 versus 100–150 nm wide × 1–10 μm long); contaminants (amorphous C, graphite, fullerenes or iron) and morphological type (multi-walled, MW, or single-walled, SW) on human U937 cells and Cheng et al.39 studied nanoparticle effects on the cytoplasm, lysosomes, and nucleus of human monocyte-derived macrophage cells, and found that the most pronounced effects on nuclear and plasma membrane potential were slightly more pronounced with SWCNT cell exposure than MWCNT.
Description of the three types of carbon nanotube used in the present study38
Jia et al.29 tested cytotoxicity parameters of purified SWCNT (synthesized by arc discharge, ~ 90% purity (mainly amorphous C impurities, with traces of Fe, Y and Ni catalyst)), MWCNT (synthesized by CVD with diameters ranging from 10 to 20 nm, purity > 95% (< 3% amorphous C, ~ 0.6% Ni catalyst)), and fullerene (C60). The cytotoxicity increased by as high as 35% when the dosage of SWCNTs was increased by 11.30 μg/cm2. No significant toxicity was observed for C60 up to a dose of 226 μg/cm2. The cytotoxicity apparently follows a sequence order on a mass basis: SWCNT > MWCNT > quartz > C60. SWCNTs significantly impaired phagocytosis of AM at the low dose of 0.38 μg/cm2, whereas MWCNT and C60 induced injury only at the high dose of 3.06 μg/cm2.
Schrand et al.41 results confirmed a previous study indicating that carbon nanomaterials (nanodiamonds (ND), SWCNTs, MWCNTs and carbon black (CB)] displayed differential biocompatibility in neuronal and lung cell lines. The greatest biocompatibility was found after incubation with NDs and both cell types followed the trend: ND > CB > MWCNT > SWCNT.
As shown above, the major studies concluded that SWNTs have higher toxicity in comparison to MWCNTs. But most of studies did not provide details of CNT structure, physio-chemical properties, synthesis and purification methods. For example, it is well documented that nickel, which in combination with yttrium is used as a catalyst in the production of arc-discharged SWCNT nanotubes, is cytotoxic.42 Common SWCNT purification methods based on oxidation have the potential disadvantage of modifying the CNTs by introducing functional groups and defects which also can lead to CNT toxicity.43−46 These data confirm the necessity for reliable controlled synthetic methods and well-characterized carbon nanotubes, complex analyses of carbon nanotubes’ properties in relation to their toxicity.
Carbon nanotubes can vary in length and shape, consist of different metal catalysts, may have some defects or functional groups (covalent and non-covalent) on the walls. Also, the extent and location of the functionalization can vary from sample to sample. All these key parameters, which are summarized in Fig. 20.2, can contribute to CNT toxicity.
The length, shape and aspect ratio are some of the most important parameters determining CNT toxicity. However, it is difficult to properly correlate the impact of size dependency on CNT toxicity studies due to the lack of systematic investigation employing CNTs of known dimensions.
Takagi47 and co-workers reported that MWCNTs induced mesothelioma along with a positive control, crocidolite (blue asbestos), when administered intraperitoneally to p53 heterozygous mice that have been reported to be sensitive to asbestos. The average diameter of MWCNTs employed was about 100 nm, and 27.5% of the particles were longer than 5 micrometers (5–20 μm). The authors concluded that the genotoxic effect of MWCNTs is unclear, and current results suggested that intraperitoneal administration of MWCNTs possesses carcinogenic potential in p53 mice presumably depending on their size/shape and persistency in the organism.
Poland et al.2 show that exposing the mesothelial lining of the body cavity of mice, as a surrogate for the mesothelial lining of the chest cavity, to long MWCNTs resulted in asbestos-like, length-dependent, pathogenic behaviour. They explored MWCNTs in a different length (short MWCNTs forming tightly packed spherical agglomerates, a large proportion of which are in the respirable size range < 5 μm; bundles of intermediate-length MWCNTs, size range < 5 μm; dispersed bundles and singlets of long and intermediate-length MWCNTs, many in the range 10–20 μm and longer; regular bundles and ropes of MWCNTs with a fairly constant length and diameter in more than 20 μm in length). The accumulation of MWCNTs in the diaphragmatic mesothelium and the subsequent degree of granuloma lesion formation were significantly higher after injection of rigid MWCNTs longer than 20 μm compared with low-aspect ratio, tangled nanotube aggregates or a negative control carbon-containing compound that is not needle-shaped. This suggests increased risk associated with lung exposure to long and rigid CNTs. However, the differences in the source, preparation and purification of different commercial CNTs and therefore potential differences in physicochemical properties and contaminations were not properly taken into toxicological consideration.48 For example, the two short tangled samples were acid-treated and this may have resulted in modifications to the CNTs, such as the removal of impurities (amorphous carbon and metal catalyst). Also the long nanotubes contained a higher amount of Fe (37.3 (μg g−1) in comparison to SWCNTs with total Fe content of 7.9 μg g−1.48
Sato et al.45 used 220 nm and 825 nm-long MWCNTs (purity is about 80 wt% with impurities such as amorphous carbon, Fe, Mo, Cr and Al) samples for testing. The results indicated that the degree of inflammation around 825-CNTs was stronger than that around 220-CNTs since macrophages could envelop 220-CNTs more readily than 825-CNTs. However, no severe inflammatory response such as necrosis, degeneration or neutrophil infiltration in vivo was observed around both CNTs examined throughout the experimental period.49
Agharkar with co-workers50 hypothesized that short MWCNT treatment induces cytotoxicity in different cell phenotypes and functionalization of MWCNTs impacts the cytotoxic profile. The results demonstrated dose-dependent cytotoxicity of three different MWCNT types (carboxylated MWCNT, hydroxylated MWCNT and non-functionalized MWCNT) and it was found43 that the reason CNT toxicity increases may be due to the presence of carboxyl and hydroxyl functional groups – not the increase in nanotubes’ length by functionalization.
In another investigation,51 protein-binding mechanisms of pristine and functionalized MWCNTs (f-MWCNTs) were investigated by varying f-MWCNTs diameter, nanotube surface chemistry, and proteins using steady-state and time-resolved fluorescence, and circular dichroism spectroscopic techniques. The f-MWCNTs with a larger diameter (40 nm) generally exhibited stronger protein binding compared to those with a smaller diameter (10 nm), demonstrating that the curvature of nanoparticles plays a key role in determining the protein binding affinity.
The results reported by Becker et al.52 measured the length-dependence of DNA-wrapped SWCNTs. An assay was used to look for reductions in metabolic activity of the cell population upon exposure to the DNA-wrapped SWCNTs. Their results suggested that shorter tubes [shorter than (189 ± 17) nm] may be more toxic to cells than longer SWCNTs.
On the other hand, Simon-Deckers et al.24 studied the response of A549 human pneumocytes to carbon nanotubes and found that neither the CNT length nor the presence of metal catalysts impurities influenced cytotoxicity.
Journeay with co-workers53 demonstrated that water-soluble rosette nanotube structures represent low pulmonary toxicity, likely due to their biologically inspired design, and their self-assembled architecture. The researches suggested that the novel nanostructures with biological design may negate toxicity concerns for biomedical applications of nanotubes. But this class of self-assembling rosette nanotubes was found to be naturally water-soluble and free of metal content upon synthesis. All these factors may have an influence on CNT toxicity.
MWCNTs range from 10 to 100 nm in diameter, they have a strong tendency to bundle together in ‘ropes’ due to van der Waals and electrostatic forces.54 Bundles typically contain many tens of nanotubes and can be considerably longer and wider than the nanotubes from which they are formed. This is a very important factor modifying CNT toxicity.55
Aggregates might have a much larger aerodynamic diameter (diameter of a particle of unit density) than singlet CNTs 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.55
The in-vitro cytotoxic experimentation was carried out with MSTO-211H cells using four samples of SWCNTs.56 Non-cytotoxic polyoxyethylene sorbitan monooleate was found to well disperse CNTs. In the present study, the cytotoxic effects of well-dispersed SWCNTs were compared with that of conventionally purified rope-like agglomerated SWCNTs and asbestos as a reference. While suspended CNT-bundles were less cytotoxic than asbestos, rope-like agglomerates induced more pronounced cytotoxic effects than asbestos fibres at the same concentrations.
On the other hand, Soto et al.57 explored cytotoxic effect of different aggregated nanomaterials, including Ag, TiO2, Fe2O3, Al2O3, ZrO2, Si3N4, naturally occurring mineral chrysotile asbestos and carbonaceous nanoparticulate materials such as multiwall carbon nanotube aggregates and black carbon aggregates. Cytotoxicological assays of these nanomaterials were performed utilizing a murine alveolar macrophage cell line and human macrophage and epithelial lung cell lines as comparators. The nanoparticulate materials exhibited varying degrees of cytotoxicity for all cell lines and the general trends were similar for both the murine and human macrophage cell lines. Authors concluded that the particle morphology or aggregate morphology is not correlated with the cytotoxicity response for either the murine or the human cell line exposure since a variety of morphologies within the nanorange exhibit equivalent or similar cytotoxicities.
The non-dependence of CNT toxicity on aggregation level was found by Muller and co-workers.33 In their study, MWCNTs and ground CNTs were administered intratracheally (0.5, 2 or 5 mg) to Sprague Dawley rat’s lung. CNTs and ground CNTs were still present in the lung after 60 days (80% and 40% of the lowest dose) and both induced inflammatory and fibrotic reactions. At 2 months, pulmonary lesions induced by CNTs were characterized by the formation of collagen-rich granulomas protruding in the bronchial lumen, in association with alveolitis in the surrounding tissues. These lesions were caused by the accumulation of large CNT agglomerates in the airways. Milled CNTs were better dispersed in the lung parenchyma and also induced inflammatory and fibrotic responses.
As-produced CNTs are insoluble in all organic solvents and aqueous solutions.58 Since many applications of CNTs (biomedical or otherwise) require their dispersion in a variety of solvents (for example, organic solvents for polymer interactions and aqueous solvents for drug delivery), there have been many investigations into improving CNT solubility. This section discusses some of the successful and popular methods, namely: sonication, stabilization with surfactant and covalent functionalization.59
Carbon nanotubes can be solubilized by a series of methods,60 including their functionalization by the aryl diazonium process, use of elemental metals, simple inorganics, acids, esters, aldehydes, amines, aromatics, macrocycles, thiols, biomolecules, polymers, and using such techniques as pulsed streamer discharge, microwave treatment, cryogenic crushing and γ-irradiation.
Lacerda et al.61 show the importance of ensuring sufficient chemical functionalization to achieve a stable dispersion of individual f-MWCNTs in physiological media and conditions and revealed that urinary excretion rates are higher when f-MWCNTs are individualized rather than aggregated in the bloodstream. The study addresses the short-term impact (first 24 h) of intravenous administration of various types of multi-walled nanotubes on the physiology of healthy mice. Non-functionalized, purified MWCNTs and different types of water-dispersible, functionalized MWCNTs were tail-vein injected. Histological examination of tissues (kidney, liver, spleen and lung) harvested 24 h post-administration indicated that organ accumulation depended on the degree of ammonium (–NH3+) functionalization at the f-MWCNT surface. The results concluded that the higher the degree of functionalization of MWCNT-NH3+, the less their accumulation in biological tissues.
A novel biomimetic amphiphilic polymer,62 cholesterol-end-capped poly(2-methacryloyloxyethyl phosphorylcholine) (CPMPC), was used as surfactant to achieve water soluble and biocompatible carbon nanotubes. Crude CNTs were facilely dispersed in aqueous media by ultrasonication with the help of CPMPC, with cholesterol-end-capped PEG and lecithin as controls. While CPEG-coated CNTs showed obviously detrimental effects on the growth of HUVEC304 cells, CPMPC-coated CNTs did not exhibit obvious toxicity. The biocompatible CPMPC-coated CNTs represent an excellent nano-object for potential biomedical applications.
Deng et al.63 used water soluble multi-walled carbon nanotubes functionalized by taurine (S-MWCNTs) as a model to investigate the possible toxicity of CNT to mouse spleen. The toxicity of various doses of S-MWCNTs was examined by carbon clearance measurement, oxidative stress assay, histopathologic and electron-microscopic examination. Compared with the control group, phagocytic activity of RES, activity of reduced glutathione, superoxide dismutase and malondialdehyde in splenic homogenate did not change significantly in 2 months. The histopathologic examination showed no observable sign of damage in spleen; however, the accumulated S-MWCNTs gradually transferred from the red pulp to the white pulp over the exposure time and it is possible to initiate the adaptive immune response of spleen.
The importance of solubilization was confirmed by Dumortier et al.64 The authors address the impact of functionalized carbon nanotubes (f-CNTs) on cells of the immune system. Two types of f-CNTs were prepared, following the 1,3-dipolar cycloaddition reaction (f-CNT 1) and the oxidation/amidation treatment (f-CNT 3), respectively. It was found that both types of f-CNTs are taken up by B and T lymphocytes as well as macrophages in vitro, without affecting cell viability. It was discovered that f-CNT 1, which is highly water soluble, did not influence the functional activity of immunoregulatory cells. f-CNT 3, which instead possesses reduced solubility and forms mainly stable water suspensions, preserved lymphocytes’ functionality while provoking secretion of proinflammatory cytokines by macrophages.
New potential applications are being investigated by introducing defects and modifying the surface chemistry of CNTs.65 Nevertheless, since CNTs are considered as almost inert substrates, they are previously subjected to oxidizing treatments and, depending on the conditions and the oxidizing agents used, is possible to obtain modifications on the surface chemistry and in the structural integrity at distinct levels.
Acid functionalization is a technique used to increase the solubility, dispersion and purification of many different types of materials including nanotubes.44 The previous investigations43−45,66,67 show that acid functionalization can significantly increase CNTs’ toxic effects, particularly if they have been modified to express functionally reactive chemical groups on their surface.
To explore the effect of CNTs’ surface chemistry on their toxicity, Magrez et al.43 decorated the surface of MWCNTs (produced by chemical vapour deposition, with an average diameter of 20 nm and aspect ratios ranging from 80 to 90) by a chemical modification of the outer layer after acid treatment. This procedure resulted in adding carbonyl (C = O), carboxyl (COOH), and/or hydroxyl (OH) groups onto the nanotube surfaces. The toxicities of the functionalized MWCNTs were tested in vitro on lung tumour cells. The results show that toxicity increases significantly when carbonyl, carboxyl, and/or hydroxyl groups are present on their surface. In the same study, a comparison of the toxicities of carbon filaments revealed that MWCNTs exhibited less toxicity. However, the exact mechanisms that lead to cell death are still unclear, and can induce cell death either after contact with cell membranes or cellular internalization, which highlights the need to carry out further studies on carcinogenicity and on the handling of MWCNTs.
This hypothesis was confirmed by Tong et al.,44 where acid functionalization (AF) enhanced the cardiopulmonary toxicity of SWCNTs. Mice were exposed by oropharyngeal aspiration to 10 or 40 μg of saline-suspended SWCNTs, and acid-functionalized SWCNTs (AF–SWCNTs, synthesized using 1:1 HNO3:H2SO4, 20 psi, ~ 150 °C, 3 min in a microwave digester). The pulmonary inflammatory responses and cardiac effects were assessed by bronchoalveolar lavage and isolated cardiac perfusion respectively, and compared to saline or LPS-instilled animals. Additional mice were assessed for histological changes in lung and heart. Instillation of 40 μg of AF–SWCNTs increased the percentage of pulmonary neutrophils. Isolated perfused hearts from mice exposed to 40 μg of AF–SWCNTs had significantly lower cardiac functional recovery, greater infarct size, and higher coronary flow rate than other particle-exposed animals and controls, and also exhibited signs of focal cardiac myofiber degeneration. Therefore, it was concluded that the intrapulmonary instillation of AF–SWCNTs enhanced cardiac ischemia/reperfusion injury and caused myocardial degeneration in mice. There was no significant ischemia/reperfusion injury in mice exposed to non-functionalized SWCNTs. However, correlation of the results between the toxicity and the functionalization of SWCNTs may not be straightforward. For example, it is possible that not only a chemical modification of the surface is taking place, but also a combination of the chemical and structural changes (especially in a microwave digester), leading to more defects in the SWCNTs. In addition to the cytotoxicity of SWCNTs, the studies by Sato et al.67 revealed that surface-modified carbon nanofibres prepared by sonication using a sulfuric acid–nitric acid mixture exhibited a significant cytotoxic factor that affected cell activation. Therefore, toxicological data related to the nature of the surface modification on SWCNTs needs further investigation.
On the other hand, functionalization or coating of carbon nanotubes with biomolecules68−76 such as nucleotide acids, proteins, and artificial polymers has emerged showing these agents improving the biocompatibility of carbon nanotubes. Thus, Dutta and co-workers70 have investigated whether re-suspension methods that alter protein adsorption profiles would modulate nanomaterial toxicity and the characterization of proteins adsorbed to the surface of two distinct classes of nanomaterials (SWCNTs (the measured BET was 274.1 m2/g; the contaminations of silica, yttrium and nickel were 0.12, 2.9 and 17.29% by weight respectively); and 10-nm amorphous silica) provides insight into pathway-specific effects. Pre-coating SWCNTs with a nonionic surfactant (Pluronic F127) inhibited albumin adsorption and anti-inflammatory properties. Albumin-coated SWCNTs reduced LPS-mediated Cox-2 induction under serum-free conditions. The profile of proteins adsorbed onto amorphous silica particles (50–1000 nm) was qualitatively different, relative to SWCNTs, and pre-coating amorphous silica with Pluronic F127 dramatically reduced the adsorption of serum proteins and toxicity. The importance of using Pluronic F127 in the CNTs’ coating has been confirmed by Bardi et al.72 They also show that MWCNTs coated with Pluronic F127 surfactant can be injected in the mouse’s cerebral cortex without causing degeneration of the neurons surrounding the site of injection. This means that the use of a surfactant to disperse SWCNTs affects protein binding affinities and toxicity. Therefore, while surfactants may increase dispersion and prevent any clumping of materials, they can also prevent formation of desirable protein–CNT complexes that would govern the uptake and disposition, which in turn could be responsible for CNT toxicity. In addition, some surfactants treated CNT would provide platform to be physisorbed by lipids and proteins in the surrounding environment. Therefore, it needs to be understood that these surface modifications will play a key role in governing uptake and disposition, which in turn can be important determinants of toxicity.
Two-photon confocal microscopy studies,71 Raman microscopy and lactate dehydrogenase (LDH) assay, were employed to study the effect of CNTs on rat glioma cell line (C6 cells) and rat alveolar macrophage continuous cell line (AM 11). Two-photon confocal microscopy reveals that unmodified CNT is toxic to the C6 cell line but, when modified by DNA, it becomes relatively non-toxic. Both LDH assay and Raman microscopy show that CNTs dispersed in serum are not toxic to the AM11 and C6 cell lines. They have illustrated that the controversy in CNT toxicity will continue, even within the same laboratory, if the changes in the surface properties of CNTs are not considered and well understood before these materials are tested for toxicological properties.
Sayes et al.73 performed in vitro cytotoxicity screens on cultured human dermal fibroblasts (HDF). The SWCNTs (synthesized using the HiPCo method, about 400 nm long) were subjected to functionalization techniques to synthesize SWCNT-phenyl-SO3H and SWCNT-phenyl-SO3Na (six samples with carbon/-phenyl-SO3X ratios of 18, 41, and 80), SWCNT-phenyl-(COOH)2 (one sample with carbon/-phenyl-(COOH)2 ratio of 23), and underivatized SWCNT stabilized in 1% Pluronic F108. Authors have found that as the degree of side-walled functionalization increases, the SWCNTs sample becomes less cytotoxic. Further, side-walled functionalized SWCNT samples are substantially less cytotoxic than surfactant stabilized SWCNTs and direct contact between cellular membranes and water-dispersible SWCNTs has been observed by atomic absorption micrographs.
Bianco et al.74 reported that functionalized CNTs are able to cross the cell membrane. This assay confirmed the results of the WST-1 test75 in that almost no decrease in viability was observed after 24 hr exposure of NR8383 cells to CNTs at a concentration of 100 μg/ml. This is in accordance with other published results that demonstrate no cytotoxic effect of intact CNTs on cell membrane integrity by assessing the LDH levels of the supernatants of macrophage cultures.33
Carrero-Sánchez et al.69 compared the toxicological effects between pure MWCNTs and N-doped multiwalled carbon (CNx) nanotubes. They found that when MWCNTs were injected into the mice’s trachea, the mice could die by dyspnea depending on the MWCNTs’ doses. However, CNx nanotubes did not cause the death of any mouse. Extremely high concentrations of CNx nanotubes administrated directly into the mice’s trachea only induced granulomatous inflammatory responses. Therefore CNx nanotubes are less harmful than MWCNTs or SWCNTs and may have the potential to be used in bioapplications.
Polymer-functionalized carbon nanotubes76 hold great promise for their use in environmental and biomedical applications. In this work, polyethyleneimine (PEI) was covalently bonded to acid-treated MWCNTs through amide bond formation. Neutral and negatively charged MWCNTs are nontoxic to both cell lines at a concentration up to 100 μg/mL, whereas positively charged MWCNTs are toxic to FRO cells at 10 μg/mL. The results of this study demonstrate that PEI-modified MWCNTs can be chemically modified to alter their surface charges and cytotoxicity, thereby significantly improving the biocompatibility of the materials for a variety of biomedical applications.
Ideal CNTs are made of one (SWCNT) or more (MWCNT) graphene sheets with hexagonal display of sp2 hybridized carbon atoms. However, CNTs are not the perfect structures they once were thought to be.65 Several properties, which were studied for ideal CNTs, depending mainly on their diameter and chirality, are also strongly affected by the presence of defects such as pentagons, heptagons, vacancies or dopant species.77
For most carbon materials, including carbon nanotubes, carbon atoms are arranged in aromatic rings which build up lamellae (graphene sheets) of various sizes and stacking heights. These aromatic layers are usually not structurally perfect and several types of defects may be present.78 Carbon atoms in these defects as well as those present on the edges of the graphene sheets are much more reactive than the atoms in the interior of the graphene sheets. The common structural defects existing in CNTs are atomic vacancies and topologic defects, of which the former corresponds to the deficiency of carbon atoms in a CNT, and the latter is associated with the network topology deviated from the hexagon rings.79
High-resolution transmission electron microscopy (TEM) images indicate80 that acid treatment induce defects on the sidewalls of the nanotubes. The inner walls can be damaged without affecting the outer walls, while the inner walls are opened along with the outer ones by heating in air. Also amorphous carbon was found inside the nanotubes after oxidation.80
To examine how structural properties may modulate the toxicity of CNTs, Fenoglio et al.46,81 and Muller et al.81 explored the physicochemical determinants of these toxic responses with progressively and selectively modified CNTs: ground multi-walled CNTs modified by heating at 600 °C (loss of oxygenated carbon functionalities and reduction of oxidized metals) or at 2400 °C (annealing of structural defects and elimination of metals) and by grinding the material that had been heated at 2400 °C before (introduction of structural defects in a metal-deprived framework). The CNTs were administered intratracheally (2 mg/rat) to Wistar rats to evaluate the short-term response (3 days) in bronchoalveolar lavage fluid (LDH, proteins, cellular infiltration, IL-1β, and TNF-α). The long-term (60 days) lung response was assessed biochemically by measuring the lung hydroxyproline content and histologically. In vitro experiments were also performed on rat lung epithelial cells to assess the genotoxic potential of the modified CNTs with the cytokinesis block micronucleus assay. The results show that the acute pulmonary toxicity and the genotoxicity of these CNTs were reduced upon heating but restored upon grinding, indicating that the intrinsic toxicity of CNTs is mainly mediated by the presence of defective sites in their carbon framework. Thus, defects may be one of the major factors governing the toxic potential of CNTs. The exact method of formation of defects is not well understood. However, there are certain spectroscopic evidence which confirms the creation of defects in CNTs under strong acid treatment, microwave digestion and sonication. This again highlights the fact that experimental procedures and characterization techniques for toxicological studies are highly necessary and such studies are probably ongoing in many laboratories in the world.
It is clear that the catalytically synthesized carbon nanotubes typically contain 3–30 wt% metal as produced, whereas most commercially ‘purified’ products contain lower but significant amounts of metal residue.82 These metal impurities can be from the catalyst used and/or those introduced inadvertently during the purification and functionalization steps. At present, thermal gravimetric analysis (TGA) and X-ray fluorescence (XRF) analysis are used for quantitative and qualitative analysis of multiple elements in CNTs respectively. However, each technique has its own limitations. Although inductively coupled plasma mass spectrometry (ICP-MS) is widely used for geological, environmental and biological samples to be digested and prepared into solutions, the chemical inertness of CNTs does not help the solubilization of CNTs. It is evident that in order to understand the toxicological investigations, such analytical techniques capable of effectively quantifying metallic impurities in CNTs are needed. For example, iron has been hypothesized to contribute to the toxicity of Fe-containing carbon nanotubes.13,16,82−86 Recently, Ge and co-workers reported a practically useful neutron activation analysis (NAA) technique and inductively coupled plasma mass spectrometry (ICPMS) analytical method for quantitative determination of metallic impurities in CNTs.86 Guo et al.82 confirmed the hypothesis that iron-catalyzed free-radical generation can contribute to oxidative stress and toxicity upon exposure to ambient particulate, amphibole asbestos fibres and SWCNTs. Generally, the iron phase both in the hollow cores and/or at the tips of as-synthesized carbon nanotubes consist of γ-iron, β-iron, Fe3C and Fe1–xS.87 The Fe1–xS phase decomposes completely around 1500 °C while the iron carbide phase decomposes in the temperature range of 1500–2400 °C. Thus, a thermal treatment method was developed to reduce metal impurities from carbon nanotubes. But, on the other hand, authors82 concluded that iron bioavailability varies greatly from sample to sample and cannot be predicted from total iron content as only the small amount of total Fe (1–7% wt) is responsible for CNT toxicity. Iron bioavailability is not fully suppressed by vendor ‘purification’ and is sensitive to partial oxidation, mechanical stress, sample age, and intentional chelation.
The hypothesis that SWCNT toxicity may be dependent upon the metal (particularly iron) content has been confirmed by Murray et al.85 The toxic effect of SWCNTs may be explained via the metal’s ability to interact with the skin, initiate oxidative stress, and induce redox-sensitive transcription factors thereby affecting/leading to inflammation. To test this hypothesis, the effects of SWCNTs were assessed both in vitro and in vivo using EpiDerm FT engineered skin, murine epidermal cells (JB6 P +), and immune-competent hairless SKH-1 mice. No significant changes in AP-1 activation were detected when partially purified SWCNTs (0.23% iron) were introduced to the cells. The results indicated that topical exposure to unpurified SWCNTs induced free radical generation, oxidative stress, and inflammation, thus causing dermal toxicity.
In another work,75 NR8383 and human A549 lung cells were incubated with commercial SWCNTs and MWCNTs, carbon black and quartz as reference particles as well as an acid-treated SWCNT preparation (SWCNT a.t.) with reduced metal catalyst content. The authors did not observe any acute toxicity on cell viability (WST-1, PI-staining) upon incubation with all CNT products. None of the CNTs induced the inflammatory mediators NO, TNF-α and IL-8. A rising tendency of TNF-α release from LPS-primed cells due to CNT treatment could be observed. However, they detected a dose- and time-dependent increase of intracellular reactive oxygen species and a decrease of the mitochondrial membrane potential with the commercial CNTs in both cell types after particle treatment whereas incubation with the purified CNTs (SWCNT a.t.) had no effect. Thus, the authors concluded that metal traces associated with the commercial nanotubes may be responsible for the toxicological effects.
On the other hand, contradictory results were shown by Lam et al.21 in the toxicological study of nanotubes produced under different conditions containing different heavy metals (see Table 20.4). Samples of ‘raw’ and ‘purified’ nanotubes both contained iron, while a third nanotube product contained nickel and yttrium. All animals treated with 0.1 mg per mouse of nickel-yttrium containing nanotubes showed no overt clinical signs. But 5 of 9 mice treated with 0.5 mg died: 2/4 within the 7-day group and 3/5 in the 90-day group. The iron-containing nanotubes (both raw and purified) did not cause deaths in the mice.
Metal content of test samples and experimental design of the intratracheal instillation study in mice21
*Metals with concentrations < 0.01% by weight are not shown.
The Panessa-Warren et al.88 investigation examined ‘as-prepared’ and acid-cleaned carbon nanoparticle physicochemical characteristics and whether these characteristics changed following 2.5–7 yr exposure to pH neutral saline or fresh water. To determine if these aqueous aged nanotubes were cytotoxic, these nanotubes were incubated with human epithelial monolayers and analyzed for cell viability (vital staining) and ultrastructural nanoparticle binding/localization. The presence of Ni and Y catalyst was less damaging to cells than CNT lattice surface oxidation. Extended fresh water storage of oxidized CNTs did not reduce surface reactive groups, nor lessen cell membrane destruction or cell death. However, storing oxidized CNTs in saline or natural organic matter (NOM) significantly reduced CNT-induced cell membrane damage and increased cell survival to control levels.
Impurity atoms are always introduced during any growth method of a solid89 and/or during the purification step of CNTs. For a reproducible growth method, it is necessary to know the source of impurities and to be able to control the type, concentration and depth distribution of impurities.
One very important point regarding toxicity is the variation of impurities of the different CNT preparations.90 Most used fractions contain high amounts of metals. Another very important contamination is amorphous carbon, which exhibits comparable biological effects as carbon black. Impurities such as O, Si, P, S, Cl, Ar, Ti, Cr and Fe also were found.89 Also, the presence of these amorphous carbon impurities may favour the covalent functionalization much more readily than the stable CNTs.
Different purification methods yield different CNT characteristics and may be suitable for the production of different types of CNTs.91 The purification methods can be categorized into two main groups, namely chemical and physical purifications. The chemical methods, such as oxidation by heating, acids and oxidizing agents, alkali treatment and annealing in inert gases separate the synthesis products based on their reactivity which normally introduce unavoidable defects along the tubes and the pentagonal structure at the tube ends, causing remarkable damages to the structure and morphology of the CNTs. However, with proper control of the reaction conditions, purification of CNTs through the removal of metal catalyst particles may result in higher purity as well as the tips opening.91
On the other hand, the physical methods, such as ultrasonication, filtration, centrifugation, and size-exclusive chromatography, separate the impurities based on their size. These processes are relatively mild and do not cause severe damage to the tubes, but they are normally more complex and less effective.92
Koyama et al.93 clarified the effect of impurities within such tubes through systemic studies of immunological responses in mice by monitoring and examining changes in peripheral T-cell subset and peripheral cytokine levels and histology. Contaminated and thermally treated at 1800 °C and 2800 °C MWNTs (see Table 20.5) were subcutaneously implanted in mice. The implanted tubes with impurities clearly induced immunological toxicity and localized alopecia, whereas extremely pure implanted tubes showed good biocompatibility. Those studies suggest that such high-temperature thermal treatment is an effective way to improve the biocompatibility of carbon nanotubes.
Basic characteristic of as-grown and thermally treated multi-walled carbon nanotubes at 1800 and 2800 °C, respectively92
aR refers to the intensity of D band over the intensity of G band.
bWe have determined the dissolved amount of iron by refluxing 5 g of nanotubes in hydrochloric acid (0.6 N) for 25 h.
cWe have measured acetone-soluble components.
dWe have determined the oxidation temperatures via the derivation of TGA curve.
Other studies21,39 demonstrated that nanotubes themselves have a toxic effect on the cells, but not the impurities. Thus, Cheng et al.39 exposed human macrophage cells to unpurified MWCNTs and found that a decrease in cell viability due to mainly necrosis was correlated with uptake of MWCNTs. Cells treated with purified MWCNTs and the main contaminant Fe2O3 itself yielded toxicity only from the nanotubes and not from the Fe2O3. Researchers observed that unpurified MWCNTs entered the cell both actively and passively, frequently inserting through the plasma membrane into the cytoplasm and the nucleus. These suggest that MWCNTs may cause incomplete phagocytosis or mechanically pierce through the plasma membrane and result in oxidative stress and cell death.
On the other hand, Miyawaki et al.94 investigated in vitro and in vivo toxicities of as-grown single-walled carbon nanohorns (SWNHs), a tubular nanocarbon containing no metal impurity. The SWNHs were found to be a non-irritant and a nondermal sensitizer through skin primary and conjunctival irritation tests and skin sensitization test. Negative mutagenic and clastogenic potentials suggest that SWNHs are not carcinogenic. The acute peroral toxicity of SWNHs was found to be quite low – the lethal dosage for rats was more than 2000 mg/kg of body weight. Intratracheal instillation tests revealed that SWNHs rarely damaged rat lung tissue for a 90-day test period, although black pigmentation due to accumulated nanohorns was observed. While further toxicological assessments, including chronic (repeated dose), reproductive, and developmental toxicity studies are still required, the present results strongly suggest that as-grown SWNHs have low acute toxicities.
Carbon-based biomaterials have been employed with great success for decades. Most notably, pyrolytic carbon has been used in biomedical implants and coatings, particularly in mechanical heart valve prostheses.95 Pyrolytic carbons have excellent biocompatibility properties,96,97 with good adherence of endothelial cells and minimal adherence and activation of platelets demonstrated.98 A more recent development in carbon biomaterials is diamond-like carbon (DLC). This dense metastable form of amorphous carbon has several properties which makes it desirable for biomedical applications. Most notable are its high hardness, low coefficient of friction, chemical inertness and good corrosion and wear resistance.99−101 DLC coatings for orthopedic and cardiovascular applications have also performed very well in terms of biocompatibility.99,102−105
The potential biomedical utility of CNTs appears to carry momentum in three key areas: medical imaging and therapeutic delivery;106,107 regenerative medicine;108 and nanocomposite biomaterials with improved mechanical and/or biological performance.109−114 Many of these applications are described in more detail in Chapter 22 of this book, and so will only be discussed briefly here.
CNTs and many other nanoscale objects are attractive as delivery vehicles and imaging agents due to their inherent chemical stability, ability to be internalized by cells, tunable surface functionality and unique electrical and spectroscopic properties.106,107,115 The many challenges of engendering physiological solubility and therapeutic activity via surface modification,58,106,107,116 and imaging via labeling with various radionuclides61,117−120 are being addressed, but still yet-to-come are studies which unequivocally demonstrate better therapeutic efficacy and safety of CNT-based technologies over established alternatives.106
With the ageing population placing considerable pressure on the future need for therapies and devices based on regenerative rather than replacement modalities, CNTs are also finding their way into polymeric or ceramic composite scaffolds being developed for tissue engineering.108 Clever incorporation of CNTs into scaffolds for bone regeneration can have multiple advantages. Porous 3-D structures are required to induce and accommodate the integration of new, properly organized and vascularized tissue, ideally by mimicking the native bone extracellular matrix. The mechanical properties of grafting materials are important factors in determining their suitability for use in orthopaedic therapies. When a graft is placed on injured bone site, it should be able to withstand both compressive and bending loads without failing, and should do so for the time required for complete growth of the new bone tissue. Most porous scaffold materials do not have adequate stiffness, strength or impact properties, therefore CNTs are being investigated as efficient high aspect ratio particulate reinforcement.109,110 Moreover, the surface of CNTs can be modified to enhance calcium phosphate biomineralization,111 and the volume fraction and organization of CNTs can be altered to perturb the scaffold surface energy and texture,112,113,121 which have also been shown to lead to enhanced tissue regeneration. In a similar manner, the fatigue life acrylic bone cement (used essentially as a grout for metallic hip and knee implants and bone), has been improved through the reinforcement of MWCNTs.114 Naturally, one critical question which still needs answering is; when a biodegradable scaffold incorporating CNTs slowly breaks down, what happens to the CNTs over time, and do the risks of having remnant CNTs outweigh the advantages? Composite scaffold studies involving tracking of radiolabelled CNTs need to be performed in order to establish a better understanding of chronic biodistribution and biological interactions.
The unique electrical properties of CNTs have naturally attracted the attention of neural tissue engineers. Since the first reports of neuronal cell attachment and growth on MWCNTs in 2000,122 many other workers have reported improved neural generation activity when grown on polymer–SWCNT composites123 and MWCNT substrates prepared with or without attachment of growth factors such as neurotrophin.124,125 CNT coatings on traditional metal wire electrodes have been shown to promote both enhanced recoding and stimulating characteristics.126 Elegant layer-by-layer assembly of alternating SWCNT–polyelectrolyte layers has yielded exceptionally strong membranes which also demonstrated enhanced guided neurite outgrowth, with or without electrical stimulation.127−129 Another layer-by-layer CNT composite was shown to successfully enhance the differentiation of stem cells to neuronal cells when compared to a conducting polymer culture substrate.130
As nanotechnology discovery and innovation march rapidly ahead, there is a critical nexus which exists today, where a very difficult balancing act is taking place. Generation of new toxicological scientific data and methodologies, rate-limited by finite R&D funding for ‘risk-related’ nanomaterials research, is balanced and influenced more by political agenda. For example, the inertia of retrospectively adapting inadequate regulatory frameworks for nanomaterials, and also the influence of the media, public perceptions and recent memories of ‘unintended consequences’ are inherited from previous rapid technological advances.131
Regulators and policy decision-makers are trying hard to arrive at more workable frameworks. In a collaborative paper summarizing conclusions from a 2008 NATO workshop on nanomaterial regulation, a multidisciplinary working group reviewed the documents listed in Table 20.6 and compared and contrasted their content under the categories of: (1) science and research aspects; (2) legal and regulatory aspects; (3) social engagements and partnerships; and (4) leadership and governance.132 Clearly, such in-depth international information sharing is going to be required to more efficiently address the existing bottleneck.
Elements of nanomaterial regulation frameworks discussed in each document at NATO workshop on nanomaterial regulation in 2008132.
Criteria are numbered 1 to 4 under each category; for each document and criterion, = document discussed the criterion, = document mentioned the criterion, blank = document did not address the criterion; adapted from Linkov and Satterstrom 2008.
Choi et al.133 estimated that if all nanomaterials required long-term in vivo testing (comprehensive precautionary approach), in the US alone it would cost $1.18bn and take up to 53 years if all existing nanomaterials were to be thoroughly tested. They strongly supported a more tiered risk assessment strategy, similar to the EU’s new REACH legislation134 for regulating chemicals. In such a system, initial screening comprising simple and inexpensive tests is used to prioritize substances for further more complex and expensive tests, with increasing degrees of selectivity for adverse effect.133,134 Implicit here for suppliers of nanomaterials will be the need to provide more specific information and recommendation of risk management measures (not just classification and labelling) in the way of materials safety data sheets (MSDS). Information including physicochemical information, handling and storage, exposure controls/personal protection and toxicological/ecotoxicological information will be required. Clearly for this to happen most effectively, risk assessors need to be engaged in the nanomaterial research, development and manufacturing processes, so that lowest risk options or formulations can be selected without compromising ultimate product performance,132 and so that researchers and workers handling the nanomaterials can be properly protected from the outset.
1. Attain reliable measurements of human and environmental exposure levels135 to inform realistic and meaningful dosing and threshold bands for other toxicology studies. Development of associated metrology.
2. Assess genotoxic potential and improve methods for predicting potential long-term effects.136
3. Improve or develop robust, chemically stable and sensitive fluorescent or radionuclide labelling methodologies which do not affect nanoparticle physicochemical or biological properties,117,137 thereby giving true biodistribution and toxicokinetic data.
4. Improved fundamental understanding of nanoparticle–biomolecule interactions (e.g. single and competitive protein interactions) and nanoparticle–cell interactions (e.g. nanoparticle interactions with cell-signalling pathways).
With the predicted increase in manufacture, use and disposal of CNTs in coming years, increased exposure of humans and the environment is inevitable. Currently, manufacturers and regulators face a dilemma due to rate-limiting finite R&D funding for ‘risk-related’ nanomaterials research, and also by the very high levels of metrology required to deliver CNT-related product assurance. These hurdles are balanced by the many very promising existing and potential future industrial and biomedical applications of CNTs and CNT composites.
Early results from toxicity studies comparing various forms of CNTs to other nanomaterials and asbestos do suggest that, for now, an extra conservative approach should be taken when handling CNTs of all forms to minimize exposure. There are also early indications that SWCNTs can display a higher level of toxicity than MWCNTs and that, in general, CNTs of higher aspect ratio are associated with increased health risks after lung exposure. However, differences in the source, preparation and purification of different commercial CNTs, and therefore potential differences in physicochemistry and contamination, are typically not fully taken into toxicological consideration. Improved CNT controls are required where at all possible, so that the number of variables known to affect CNT toxicity (aspect ratio, metal catalyst or other impurities, surface chemistry and defects, level of aggregation and hierarchical morphology and solubilization in physiological media) are minimized. As more long-term inhalation data linking measured exposure and chronic biological interactions comes to hand, a clearer picture will emerge. In order to do this successfully, more robust, chemically-stable and sensitive fluorescent or radionuclide-based labelling methodologies (which do not affect CNTs’ physicochemical or biological properties) will be required. New methods are also required to assess genotoxic potential and to be able to predict long-term biological effects. Understanding specific molecular interactions between nanomaterials and biomolecules and cells will be fundamentally critical.
Finally, large scale and multi-level cooperation between academia, industry and government sectors, media, NPOs and the community will be essential to avoid repeating past mistakes and ensuring that the tremendous utility of these materials is conservatively, responsibly and sustainably developed. The composites industry is one of the highest volume adopters of CNTs and therefore business and technical leaders in this sector will need to ensure they play their role in this evidence-based process.
1. Mueller, N.C., Som, C., Nowack, B. Nanotechnology 2009: Fabrication, Particles, Characterization, MEMS, Electronics and Photonics – Technical Proceedings of the 2009 NSTI-Nanotechnology Conference and Expo, NSTI-Nanotech 2009. 2009;Vol. 1:159–162. [Houston, TX].
4. Ryman-Rasmussen, J.P., Cesta, M.F., Brody, A.R., Shipley-Phillips, J.K., Everitt, J.I., Tewksbury, E.W., Moss, O.R., Wong, B.A., Dodd, D.E., Andersen, M.E., Bonner, J.C. Nature Nanotechnology. 2009; 4:747–751.
11. Shin, D.C., Jang, J.Y., Yang, J.Y., Kim, S.H., Choi, I.H., Lim, Y.W., Kim, C.S., Lee, G.T. Technical Proceedings of the 2008 NSTI Nanotechnology Conference and Trade Show, NSTI-Nanotech, Nanotechnology 2008. 2008; Vol. 2:142–145. [Quebec City, QC].
16. Shvedova, A.A., Castranova, V., Kisin, E.R., Schwegler-Berry, D., Murray, A.R., Gandelsman, V.Z., Maynard, A., Baron, P. Journal of Toxicology and Environmental Health – Part A. 2003; 66:1909–1926.
18. Pacurari, M., Yin, X.J., Zhao, J., Ding, M., Leonard, S.S., Schwegler-Berry, D., Ducatman, B.S., Sbarra, D., Hoover, M.D., Castranova, V., Vallyathan, V. Environmental Health Perspectives. 2008; 116:1211–1217.
32. Huczko, A., Lange, H., Bystrzejewski, M., Baranowski, P., Grubek-Jaworska, H., Nejman, P., Przybyàowski, T., Czumiàska, K., Glapiàski, J., Walton, D.R.M., Kroto, H.W. Fullerenes Nanotubes and Carbon Nanostructures. 2005; 13:141–145.
34. Shvedova, A.A., Kisin, E.R., Mercer, R., Murray, A.R., Johnson, V.J., Potapovich, A., Tyurina, T., Gorelik, Y.Y., Arepalli, O., Schwegler-Berry, S., Hubbs, D., Antonini, A.F., Evans, J., Ku, D.E., Ramsey, B.K., Maynard, D., Kagan, A., Castranova, V.E., Baron, V.P. American Journal of Physiology – Lung Cellular and Molecular Physiology. 2005; 289:L698–L708.
38. De Nicola, M., Gattia, D.M., Bellucci, S., De Bellis, G., Micciulla, F., Pastore, R., Tiberia, A., Cerella, C., D’Alessio, M., Antisari, M.V., Marazzi, R., Traversa, E., Magrini, A., Bergamaschi, A., Ghibelli, L. Journal of Physics Condensed Matter. 2007; 19:1–7.
45. Sato, Y., Yokoyama, A., Shibata, K.I., Akimoto, Y., Ogino, S.I., Nodasaka, Y., Kohgo, T., Tamura, K., Akasaka, T., Uo, M., Motomiya, K., Jeyadevan, B., Ishiguro, M., Hatakeyama, R., Watari, F., Tohji, K. Molecular BioSystems. 2005; 1:176–182.
50. Agharkar, V.A., Bhushan, A., Lai, J.C.K., Daniels, C.K. Technical Proceedings of the 2008 NSTI Nanotechnology Conference and Trade Show, NSTI-Nanotech, Nanotechnology 2008. 2008;Vol. 2:122–125. [Quebec City, QC].
61. Lacerda, L., Soundararajan, A., Singh, R., Pastorin, G., Al-Jamal, K.T., Turton, J., Frederik, P., Herrero, M.A., Li, S., Bao, A., Emfietzoglou, D., Mather, S., Phillips, W.T., Prato, M., Bianco, A., Goins, B., Kostarelos, K. Advanced Materials. 2008; 20:225–230.
67. Sato, Y., Shibata, K.I., Kataoka, H., Ogino, S.I., Bunshi, F., Yokoyama, A., Tamura, K., Akasaka, T., Uo, M., Motomiya, K., Jeyadevan, B., Hatakeyama, R., Watari, F., Tohji, K. Molecular BioSystems. 2005; 1:142–145.
71. Senanayake, V., Juurlink, B.H., Zhang, C., Zhan, E., Wilson, L.D., Kwon, J., Yang, J., Lim, Z.L., Brunet, S.M.K., Schatte, G., Maley, J.M., Hoffmeyer, R.E., Sammynaiken, R. Journal of Biomedical Nanotechnology. 2008; 4:515–523.
81. Muller, J., Huaux, F., Fonseca, A., Nagy, J.B., Moreau, N., Delos, M., Raymundo-Piàtero, E., Bàguin, F., Kirsch-Volders, M., Fenoglio, I., Fubini, B., Lison, D. Chemical Research in Toxicology. 2008; 21:1698–1705.
84. Kagan, V.E., Tyurina, Y.Y., Tyurin, V.A., Konduru, N.V., Potapovich, A.I., Osipov, A.N., Kisin, E.R., Schwegler-Berry, D., Mercer, R., Castranova, V., Shvedova, A.A. Toxicology Letters. 2006; 165:88–100.
89. Usov, I.O., Wang, Y.Q., Li, Q., Zhu, Y.T., DePaula, R.F., Arendt, P.N. Nuclear Instruments and Methods in Physics Research Section B: Beam Interactions with Materials and Atoms. 2007; 261:574–577.
115. Kostarelos, K., Lacerda, L., Pastorin, G., Wu, W., Wieckowski, S., Luangsivilay, J., Godefroy, S., Pantarotto, D., Briand, J.P., Muller, S., Prato, M., Bianco, A. Nature Nanotechnology. 2007; 2:108–113.
118. Singh, R., Pantarotto, D., Lacerda, L., Pastorin, G., Klumpp, C., Prato, M., Bianco, A., Kostarelos, K. Proceedings of the National Academy of Sciences of the United States of America. 2006; 103:3357–3362.
132. Linkov, I., Steevens, J., Adlakha-Hutcheon, G., Bennett, E., Chappell, M., Colvin, V., Davis, J.M., Davis, T., Elder, A., Foss Hansen, S., Hakkinen, P.B., Hussain, S.M., Karkan, D., Korenstein, R., Lynch, I., Metcalfe, C., Ramadan, A.B., Satterstrom, F.K. Journal of Nanoparticle Research. 2009; 11:513–527.