The primary objective of this investigation was to evaluate if exposure to aerosolized MWCNT promotes the growth of DNA damaged cells and/or is a complete carcinogen. To accomplish this, mice were exposed to aerosolized MWCNT (5 mg/m3, 5 hours/day) for 15 days. Initial MWCNT lung burden in these mice was 31.2 ± 0.9 μg MWCNT/mouse . In order to evaluate the relationship of these MWCNT lung burdens to human MWCNT exposures, we compared the MWCNT lung burdens in these mice with potential human occupational exposures. OSHA has not yet established exposure limits for carbon nanotubes; however, MWCNT are regulated as respirable particulates not otherwise regulated (PNOR). The PNOR have an OSHA Permissible Exposure Limit (PEL) of 5 mg/m3. NIOSH recently published a Current Intelligence Bulletin with a Recommended Exposure Limit (REL) of 1 μg/m3 for carbon nanotubes which is 5000-fold lower than the OSHA PNOR PEL .
Assuming a mouse alveolar epithelium surface area of 0.05 m2, the 31.2 μg MWCNT lung burden would result in 624 μg MWCNT/m2 alveolar epithelium. Using the alveolar epithelial surface area of 102 m2 for human, the equivalent human lung burden would be 63.6 mg .
If the MWCNT mass median aerodynamic diameter (MMAD) =1.5 μm were used, minute ventilation of 20 L/minute for a person performing light work  and an alveolar deposition fraction of 30%  (for 240 work days per year), the equivalent lung burden in workers exposed at the previous draft REL for CNT of 7 μg/m3 would be achieved in approximately 13 years . This indicates that the mouse MWCNT lung burdens in this study approximate feasible human occupational exposures.
Inhalation of multi-walled carbon nanotubes (MWCNT) for 15 days following a single intraperitoneal injection of the known initiator MCA led to increased incidence and numbers of bronchiolo-alveolar adenomas and bronchiolo-alveolar adenocarcinomas in B6C3F1 male mice. The combined incidence of bronchiolo-alveolar adenomas and bronchiolo-alveolar carcinomas of 90.5% for the MCA + MWCNT group greatly exceeded that in groups of mice exposed to air or MCA and the NTP historical vehicle control range for B6C3F1 male mice . Additionally, the numbers of bronchiolo-alveolar adenomas or bronchiolo-alveolar adenocarcinomas were greatest in the MCA + MWCNT group compared to other groups. The data demonstrate that MWCNT may act as a carcinogen that promotes the growth of initiated lung cells, resulting in the development of lung adenocarcinoma.
The strong tumor promotion that was observed in the current study may have resulted from a combination of effects that have been observed following exposure to carbon nanotubes. MWCNT material was observed in the diaphragm and in the lungs, both within macrophages and the interstitium. MWCNT are internalized by macrophages following pulmonary exposure. In addition, MWCNT material has been observed in the interstitium. MWCNT exposure has been shown to induce fibrosis as early as seven days post-exposure [8, 10, 45]. In previous experiments, the post-exposure pulmonary distribution, pulmonary fibrotic response and transport of MWCNT to systemic organs were examined at various times post-exposure, from 1 to 336 days [46, 47]. Fibrillar collagen in the lungs was specifically stained and the quantity of fibrillar collagen in the alveolar region was measured by morphometry. These measurements of fibrillar collagen in the alveolar region of the lungs demonstrated a fibrotic response to inhaled MWCNT which was significantly above vehicle controls and progressively increased throughout the 336 days post-exposure study period . These studies have demonstrated that inhaled MWCNTs are deposited throughout the alveolar region of the lungs and are retained in the alveolar tissue. Additionally MWCNT were demonstrated in the visceral pleura, pleural space and parietal space . MWCNT that penetrate the visceral pleural induce pleural inflammation and cell proliferation in a manner similar to asbestos [8, 48, 49]. MWCNT have further been shown to penetrate the cytoplasmic membrane and nuclear envelope [50, 51].
Asbestos and MWCNT also induce inflammation, fibrosis, cell proliferation and cellular atypia in the lung [8, 48, 49]. Cell proliferation and inflammation are important events in the promotion of cancer [16, 17, 52–55]. Indeed in the current study, inhaled MWCNT induced dramatic hyperplasia and a moderate increase in adenomas however, the increase in adenomas was not statistically significant. In addition, MWCNT exposure did not result in an increased number of adenocarcinomas. The significant hyperplastic response that was observed after exposure to MWCNT material without prior initiation indicates that the material was a tumor promoter. The dramatic increase in adenomas and adenocarcinomas after MCA initiation followed by MWCNT-exposure demonstrate that inhaled MWCNT material is a strong tumor promoter. Strong tumor promoters increase the growth of chemically initiated as well as spontaneously initiated cells [56, 57]. Although the data of the current investigation do not indicate that inhaled MWCNT material act as tumor initiators, the data demonstrate the strongest promotion response observed in the lung using occupationally relevant material [15, 52, 58–60]. The data further indicate that MWCNT may initiate lung responses similar to the carcinogenic fiber asbestos [4, 61–63].
The dimensions of the nanotubes as well as their surface properties are important in the inflammatory response. Pulmonary exposure to SWCNT and MWCNT causes inflammation and fibrosis; however, the inflammatory response following MWCNT exposure is more pronounced than the response observed following SWCNT exposure [8, 64, 65]. The degree of inflammation resulting from asbestos and MWCNT is determined by the diameter and the length [4, 66, 67]. Carbon nanotubes of approximately 50 nm in diameter cause more inflammation than nanotubes of less than 20 nm or greater than 150 nm . In addition, the rigid MWCNT of 40–50 nm diameter and at least 4 microns long were the most inflammatory [67, 68]. Although these studies indicate that the diameter and length of carbon nanotubes may alter MWCNT-induced carcinogenicity further investigations are required to fully characterize the role of the dimensions as well as the physical properties in the carcinogenic response.
Classical multistep carcinogenesis models involve initiation, promotion and progression. Exposure to a genotoxic agent initiates a population of genetically altered cells which expand in number through the action of promoters and undergo additional genetic changes during the progression process [16, 17, 69]. Several studies suggest both genotoxicity and promotion from the classical carcinogenic high aspect ratio particle, asbestos. Oxidant generation from inflammation has been shown to damage the DNA and can initiate cancer . Asbestos and carbon nanotubes have been shown to induce disruption of the cell division apparatus and errors in chromosome number (aneuploidy) in vitro[12, 14, 70]. The long, thin asbestos fibers of less than 0.25 μm diameter and at least 5 microns in length are the most genotoxic [71, 72]. The mutagenicity of asbestos fibers is correlated with the potency as a carcinogen . Evidence from epidemiological studies has demonstrated that asbestos can act as a tumor promoter at low doses as well as a tumor initiator at longer and/or higher exposure levels . In several human epidemiology studies, smoking exposure and asbestos interact in a more than additive fashion in causing lung cancer [75–79]. There are multiple mutagens in cigarette smoke that have the potential to initiate cancer . Humans are also potentially exposed to many other mutagens that could initiate cancer such as radon, polychlorinated biphenyls, hexavalent chromium, naphthalene and benzo-a-pyrene in diesel exhaust [81–85]. Thus, it is plausible to suggest MWCNT could potentially act as promoters in individuals who smoke or are exposed to other initiators.
Previous studies to examine rodent exposure to asbestos by inhalation or pharyngeal aspiration have shown that asbestos is carcinogenic in the rat lung by this route but only weakly positive in the mouse [86–88]. The data demonstrating that asbestos induces mitotic spindle disruption and aneuploidy would suggest that asbestos would be a strong tumor promoter; however, asbestos has not been administered in an initiation/promotion protocol in a mouse model. Although lung cancer has not been observed in either rats or mice following the intraperitoneal injection of asbestos or carbon nanotubes, mesothelioma has been reported. Abdominal or scrotal injection of mice with asbestos or long thin MWCNT of at least 3.9 micron in length and 50 nm in diameter caused mesotheliomas in p53 +/- transgenic mice and Fischer rats [20, 21]. Recent investigations demonstrated that intraperitoneal injection of as little as 3 μg of MWCNT in genetically modified mice (p53+/-) induced mesothelioma . By contrast, an IP exposure of Wistar rats to short MWCNT of < 1 micron in length resulted in mesothelioma in 5/150 MWCNT-exposed animals but those findings were not statistically significant due to a high peritoneal mesothelioma rate in the control group . The high background rate of peritoneal mesotheliomas (1/26) is unusual for the Wistar rat [89–91]. A subsequent study demonstrated that high exposures (1 and 10 mg/rat) of thin, rigid MWCNT by intraperitoneal injection caused mesotheliomas (54). When the diameter of the nanotubes was considered, MWCNT of 50 nm in diameter were more carcinogenic than nanotubes of less than 20 nm or greater than 150 nm . These findings suggest that the diameter and length are critical in the carcinogenic response to MWCNT, a finding that is similar to classical studies of asbestos fiber carcinogenicity .
A limitation of the current study is that suitable non-carcinogenic particle controls do not exist in this model. We considered using the short multi-walled carbon nanotubes investigated by Muller et al., as a potential negative control . However, since the interpretation of the Muller et al. study is affected by the unusual high background rate of peritoneal mesothelioma in the control group, this particle cannot be considered a confirmed negative control nor could we identify any carbon nanotube as a confirmed negative particle control for a carcinogenicity study. The identification of a suitable negative control nanotube will require further carcinogenicity studies that have yet to be published.
Malignant mesothelioma in humans has three major histologic patterns: epithelial, sarcomatous (sarcomatoid), and biphasic [93, 94]. Using standard histopathology, the major differential diagnoses for malignant mesothelioma include broncho-alveolar adenocarcinoma of the lung, metastatic carcinoma and metastatic sarcoma . The diagnosis of malignant mesothelioma in humans can be supported by staining for proteins commonly expressed in mesotheliomas, including calretinin, cytokeratins, mesothelin, WT-1 and podoplanin (D2-40) . Podoplanin is among the markers most consistently expressed in human malignant sarcomatous mesotheliomas [34, 95]. The malignant serosal tumors seen in the mice in our study consistently expressed podoplanin. However, the staining for cytokeratins was negative to equivocal, a finding that is also sometimes seen in sarcomatous mesothelioma in humans [35, 36]. One review noted that only 13% of human sarcomatous mesotheliomas were positive for cytokeratin 5/6 and none were positive for seven other epithelial markers . However, in one study using a cocktail of mouse anti-human monoclonal antibodies, 93% of the cases of human sarcomatous mesotheliomas demonstrated cytokeratin expression . However, there are protein sequence differences between human and mouse cytokeratins. In addition, even with blocking steps, indirect immunohistochemistry using mouse antibodies on mouse tissues results in some degree of binding of the secondary anti-mouse IgG antibody with endogenous IgG located in the mouse tissue. It is for this reason that we used both a mouse monoclonal antibody and a rabbit anti-pancytokeratin antibody to stain for cytokeratins in this study. However, the negative to equivocal staining of the serosal tumors for cytokeratins in this study should be interpreted with an understanding that mesotheliomas are very rare in the mouse and that techniques for identifying mesothelioma markers in mice are not as advanced as they are for identifying those markers in human tissue. The negative staining does not mean that there are no cytokeratins in the serosal tumors of our study, only that no cytokeratins could be identified with the antibodies used in this study.
In humans, podoplanin staining in the absence of cytokeratin staining can be seen in several different sarcomas as well as in malignant mesothelioma [34, 97]. The malignant serosal tumors seen in this study were morphologically consistent with malignant sarcomatous mesotheliomas with five of the six tumors involving multiple peritoneal serosal surfaces. The remaining malignant serosal tumor was limited to the male urogenital tract, a common site for mesothelioma in rats but not in control mice . In humans, a diagnosis of cytokeratin negative sarcomatous mesothelioma is usually made by excluding other potential diagnoses . Given the rarity of mesotheliomas in the mouse, we could not exclude other diagnoses with absolute certainty . Thus, the characteristics of these tumors are consistent with, but not diagnostic of, mesothelioma. The principal differential diagnosis is pleural sarcoma.
These tumors are considered similar to the serosal tumors diagnosed as pleural sarcomas or malignant mesenchymal neoplasms in the classical asbestos studies in rats conducted by Stanton and co-workers who noted their comparability to human mesotheliomas .