Studies in humans.

As summarized elsewhere (7, 107), inhaled particles of different sizes exhibit different fractional depositions within the human respiratory tract. Although inhaled ultrafine particles (<100 nm) deposit in all regions, tracheobronchial deposition is highest for particles <10 nm in size, whereas alveolar deposition is highest for particles approximately 10–20 nm in size (7, 107). Particles <20 nm in size also efficiently deposit in the nasopharyngeal-laryngeal region. Human studies of potential adverse pulmonary effects resulting from exposure to engineered nanoparticles appear to be limited, although a number of investigations into pulmonary deposition patterns of inhaled nanoparticles in the healthy and diseased lung have been conducted (5, 24, 28, 93). Computational models predict increased deposition of inhaled nanoparticles in diseased or constricted airways (44), and, consistent with this prediction, obstructive lung disease and asthma have both been demonstrated to increase their pulmonary retention (5, 24). Nonetheless, Pietropaoli et al. (114) did not observe differences between healthy and asthmatic subjects in respiratory parameters assessed up to 45 h after a 2-h inhalation of ultrafine carbon particles (up to 25 μg/m²), nor was airway inflammation observed in either group (measured as exhaled nitric oxide). Moreover, the same study reported that exposure of healthy subjects to a higher concentration of ultrafine carbon particles (50 μg/m² for 2 h) resulted in decreased midexpiratory flow rate and carbon monoxide diffusing capacity 21 h after exposure, albeit still in the absence of airway inflammation (114). Thus nanoparticles may influence respiratory function and gas exchange without a concomitant induction of inflammation.

Several studies have also examined the potential for inhaled manufactured ultrafine particles (i.e., ^99mtechnetium-labeled carbon nanoparticles) to translocate from the lungs to the systemic circulation in humans. This is an important issue to consider as inhaled engineered nanoparticles may exert adverse cardiovascular effects, similar to the proposed mechanism for the nanoparticulate fraction of urban air pollution (15, 40). All but one of the studies reported to date indicate that inhaled ^99mtechnetium-labeled carbon nanoparticles are not detected outside of the lungs in appreciable quantities after inhalation (17, 91, 93, 100, 150, 151). However, as alluded to by Mills et al. (91), these findings do not indicate that other nanoparticles will behave in the same manner, nor do they rule out the possibility that nanoparticles may interact with and influence the vasculature. Moreover, the studies conducted to date have used a single inhalation exposure protocol, and it is possible that repeated exposures may result in greater pulmonary accumulation and translocation of significant quantities of nanoparticles to the circulation.

Studies in experimental animals.

Pulmonary effects resulting from airway administration of nanoparticles have been examined in a number of experimental animal studies, a summary of which is presented in Table 2. Although the primary outcomes of interest in the majority of these studies have been pulmonary inflammation and fibrosis, several have investigated distribution patterns within the lung and the potential translocation and systemic distribution of nanoparticles following pulmonary administration; these are summarized in Table 3. In addition to the endpoints listed in Tables 2 and ​and3,3, carcinogenic effects of inhaled nanoparticles (ultrafine particles) have, in some instances, been found to be more severe than those of larger size analogs. This is thought to result primarily from lung particle overload due to the inability of alveolar macrophages to recognize and/or clear particles of this size, leading to particle build up, chronic inflammation, fibrosis, and tumorigenesis. These effects are discussed in detail elsewhere (14, 101) and will not be covered here.

Although it is recognized that the pathogenic mechanisms underlying animal models of lung fibrosis and human lung fibrosis are not necessarily the same, increased collagen deposition and structural changes to the lungs that can result in altered respiratory mechanics are common features of both. As such, for the purpose of this review, fibrosis in animal models is considered to be defined by increased collagen content and/or histopathological evidence of structural alterations to the lung that are consistent with fibrosis. With regard to inflammation and fibrosis, studies have focused primarily on the effects of carbon nanotubes, carbon black, fullerenes, silica, and metal-based nanoparticles including titanium dioxide, silver, and nickel. The reader is referred to recent reviews for more detailed descriptions of some of the available data for carbon nanotubes (34, 56, 76). In brief, single- and multi-walled carbon nanotubes have been shown to induce inflammatory and fibrotic responses in mice and rats following administration by intratracheal instillation or oropharyngeal aspiration. However, differences in the type and extent of injury induced by the two forms of nanotubes have been observed, and some reports indicate a lack of inflammation due to carbon nanotube instillation or inhalation (88, 92). Of note, when administered at equal doses (on a milligrams per kilogram body weight basis) to mice by intratracheal instillation, one type of single-walled carbon nanotube was found to be more toxic than silica quartz (77), which itself poses a serious occupational health hazard in chronic inhalation exposure scenarios. However, the authors of this study were unable to rule out the possibility of nickel contamination in the nanotube suspensions as contributing to the observed toxicity (77).

Although not the most physiologically relevant exposure routes, instillation and aspiration have dominated pulmonary toxicity studies of carbon nanotubes, likely due in part to their relative ease and cost efficiency compared with inhalation exposure protocols. However, concerns including the potential formation of nanotube aggregates and how responses to these aggregates might differ from responses to more dispersed species necessitate caution in interpreting the available data. Indeed, Mercer et al. (90) recently reported that the deposition pattern and inflammatory/fibrotic response to aspirated single-walled carbon nanotubes in mice depended on whether the nanotubes were dispersed before instillation. Specifically, dispersion was found to result in incorporation of the single-walled carbon nanotubes into the alveolar interstitium and an increase in collagen accumulation following aspiration administration (90), whereas the same authors had previously demonstrated that nondispersed single-walled carbon nanotubes caused a primarily granulomatous response associated with deposition of nanotube agglomerates (127). To date, very few inhalation exposure studies with carbon nanotubes have been reported (80, 92). No biochemical analyses were performed in the study by Li et al. (80), but the histopathological responses of the lung to instilled multi-walled carbon nanotubes (0.05 mg/mouse) were compared in a nonquantitative manner to those following inhalation of aerosols of multi-walled carbon nanotubes for up to 15 days over the course of a 24-day period (resulting in estimated lung burdens of 0.07, 0.14, or 0.21 mg/mouse). The authors observed considerable differences in nanotube deposition pattern and bronchial and alveolar responses between the two administration methods. In particular, aggregated clumps of nanotubes were evident following instillation and were associated with areas of bronchial and alveolar inflammation and destruction of alveoli, whereas the aerosol inhalation exposure resulted in much less severe inflammation and a general thickening of alveolar walls. Similarly, Mitchell and colleagues (92) did not observe an inflammatory response in the lungs of mice following 7 or 14 days of inhalation exposure to multi-walled carbon nanotubes, nor did they observe any evidence of fibrosis. Two important differences between exposure via aerosol inhalation and exposure via instillation or aspiration include the dose rate and dose distribution. Both of these factors are likely to significantly impact the nanoparticle agglomerate state in the lungs following exposure, the responses of the lung cells to the deposited material, and the rate of translocation (if it occurs at all). Thus the observed dissimilar findings in toxicity outcomes reported to date highlight the importance of considering the method of administration when assessing lung inflammatory and fibrotic responses to carbon nanotubes and likely to other nanoparticles as well. This is particularly relevant given that human exposures are likely to occur via inhalation of nanoparticle-containing aerosols and not by bolus instillation or aspiration of nanoparticle-containing solutions.

Carbon black, fullerenes, silica, and metal-based nanoparticles have also been studied for their ability to induce inflammatory and fibrotic responses in the lungs of experimental animals following delivery via instillation, aspiration, and/or inhalation (Table 2). Increased lung inflammation resulting from exposure to nano-sized particles compared with that resulting from an equivalent mass of micron-sized particles has been demonstrated in some studies (16, 45, 49, 54, 82, 118, 160), whereas others have found this not to be the case (9, 120, 135, 149). Potential factors in the increased inflammatory profile observed for nanoscale materials in some studies include their size, increased number, and higher surface area per unit mass compared with that of larger particles of the same material (15, 98, 104). Titanium dioxide is a good example of how both the size and form of a nanoparticle can influence its pulmonary toxicity, as a nanoscale anatase form of titanium dioxide was found to induce greater lung inflammatory responses than those resulting from a nanoscale rutile form and from a micron-sized anatase form following intratracheal administration in rats (148). The increased ratio of surface area to mass for nanoparticles means that a greater percentage of the atoms or molecules of a given particle are present on the surface of the particle, thereby providing an increased number of potential reactive groups at the particle surface that may influence toxicity. Although this appears to be a useful metric for assessing the toxic potential of some nanoparticles, there is consensus among experts in the field that no single dose metric (i.e., particle number, size, surface area, or other) has emerged to be useful for assessment of the reactivity and potential toxicity of nanoparticles in general (89, 144). Rather, it is likely that the most appropriate means of expressing dose-related toxicity for nanoparticles of interest will continue to be determined on an individual basis.

Although the majority of studies have examined the pulmonary effects of nanoparticles administered alone, some have incorporated nanoparticle exposure as a potential modifying factor in established models of airway inflammation (Table 2). Carbon black nanoparticles of 14-nm size, but not of 56-nm size, were found to exacerbate endotoxin-induced airway inflammation in mice when given concomitantly by intratracheal administration (62). Similarly, both 14- and 56-nm carbon black nanoparticles promoted allergen-induced inflammation in mice that was associated with increased levels of eotaxin, TARC, MIP-1α, GM-CSF, and IL-2, -5, -6, -10, and -13 protein in lung tissue (63, 64), although the exacerbating effect was more prominent for the smaller sized particles. Enhancement of allergic airway inflammation in mice by the intranasal administration of carbon black nanoparticles at the time of allergen sensitization (32) or by inhalation 24 h before allergen challenge (3) has also been reported. Finally, Shvedova and coworkers (126) recently demonstrated that intratracheal administration of single-walled carbon nanotubes to mice increased the inflammatory response to a subsequent infection with Listeria monocytogenes and that this effect was associated with decreased bacterial clearance and increased airway levels of several acute phase cytokines and chemokines. Although only two of the six studies described above contained sufficient nanoparticle characterization data (3, 126), the cumulative data from all six studies suggest a detrimental effect of nanoparticle exposure on airway inflammation induced by other agents.

As discussed earlier, studies in humans generally indicate that inhaled manufactured ultrafine particles (i.e., ^99mtechnetium-labeled carbon nanoparticles) do not translocate from the lungs to the systemic circulation (17, 91, 93, 150, 151). A number of studies in experimental animals have also addressed the possibility of nanoparticle translocation from the lungs to the circulation and extrapulmonary tissues (Table 3), some of which indicate that certain nanoparticles may have the capacity to do so. For example, Semmler et al. (123) observed detectable albeit very low levels of insoluble iridium nanoparticles in the liver, spleen, brain, and kidney of rats following a single inhalation exposure, whereas Ji et al. (65) reported that the content of silver in the liver of male rats increased in a concentration-dependent manner following inhalation of silver nanoparticles for 5 days/wk for 4 wk. Detection of nanoparticles in the liver following pulmonary delivery may not reflect a direct translocation from the lungs into the bloodstream, however, as the possibility of mucociliary transport and subsequent swallowing of the nanoparticles, leading to gastrointestinal absorption and detectable liver accumulation, must be considered. Also of significant interest is the potential translocation of inhaled nanoparticles to the brain via the olfactory nerve, as has been demonstrated in rats following exposure to inhaled nanoparticles of elemental carbon (36 nm), manganese oxide (30 nm), and silver (12–15 nm) (39, 65, 105). Whether such translocation to the brain has neurological or other consequences remains to be determined.

There are a number of important factors that must be considered when assessing data purporting to support direct translocation of a nanoparticle from the lungs to the systemic circulation and secondary organs. These include solubility, potential leaching of a radiolabel, inflammation or injury induced by the nanoparticle, and others. Results of rat studies with inhaled insoluble iridium and gold nanoparticles and soluble cadmium oxide nanoparticles offer some insight into some of these factors. Data from studies of inhaled iridium and gold nanoparticles, in which pulmonary inflammation was not induced, indicate that translocation of these insoluble particles from the lower respiratory tract to secondary organs accounts for only a minute fraction of the administered dose (73, 123, 134); in the case of iridium, the vast majority was found to be cleared via the thoracic airways to the larynx, gastrointestinal tract, and eventually the feces (73, 123). For soluble cadmium oxide, lung injury was reported to be the underlying cause of translocation and accumulation in the liver following inhalation as this phenomenon only occurred at a high dose at which injury was induced and not at a low dose in the absence of injury (135). Thus translocation to the systemic circulation can be low even for a relatively soluble nanoparticle, indicating that other mechanisms such as the affinity of binding to cell membranes or proteins may be important. These and other studies reinforce the necessity to consider a variety of factors that can influence the deposition, retention, clearance, and translocation of nanoparticles within and from the lung.

Studies in cell culture.

Studies performed to assess the in vitro toxicity of nanoparticles can be used as part of a screening strategy to identify potentially hazardous substances and to elucidate underlying mechanisms of toxicities observed in vivo. A considerable amount of information from in vitro (i.e., cell culture) analyses of the pulmonary toxicity of nanoparticles has been published (Table 4). Related to the in vivo reports of inflammation and fibrosis induced by carbon nanotubes discussed earlier, in vitro assessments indicate that nanotubes may have the ability to induce oxidative stress and cytotoxicity in pulmonary cells of animal and human origin, although it should be noted that not all studies are in agreement on this matter. Indeed, metal contaminants in nanotube preparations have been found to be causative of some of the adverse outcomes observed in certain in vitro investigations (68, 116), whereas anti-inflammatory properties of nanotubes have been reported in others (37). In addition, an important consideration in studies of nanotube cytotoxicity is the means by which cell injury or death is quantified, as nanotubes have been found to interact with some commonly used dye-based assays and to lead to false-positive indications of cytotoxicity (31, 153). Nonetheless, the cumulative data suggest that carbon nanotubes may have the capacity to induce oxidative stress and cytotoxicity in lung cells.

Studies have also been performed to assess the potential toxic effects of nano-sized metals and metal oxides in a variety of lung cell types in vitro (Table 4). In general, these materials also appear to have the capacity to induce oxidative stress and cytotoxicity, although, not surprisingly, there are apparent differences in the extent of these effects among the different nanoparticles studied. Furthermore, the concentrations required to induce such effects are often relatively high (i.e., in the micrograms per milliliter range) wherein potential effects of agglomeration should be, but are not always, considered. As with any in vitro toxicity study, the range of concentrations tested in nanoparticle toxicity studies should be broad so as to allow for assessment of potentially different responses (and underlying mechanisms) at the low and high ends of the concentration-response curve. Moreover, similar to the interference of cytotoxicity assays by carbon nanotubes, nano-sized metals and metal oxides may interfere with cytokine assays commonly used to assess oxidative stress and inflammation (139), thus necessitating caution in interpreting data derived using such assays. Titanium dioxide is a commercially important nanoparticle that has been shown to induce oxidative stress in, among others, A549 and BEAS-2B lung cell lines (51, 84, 94). Of note, the form of nanoscale titanium dioxide appears to be an important determinant of its cytotoxicity, as the anatase form has been found to be more potent than the rutile form in murine (RAW 267.9) and human (A549) lung cell lines (121, 129). These in vitro data documenting differences in cytotoxicity due to different forms of nanoscale titanium dioxide are generally consistent with findings in recent in vivo pulmonary toxicity studies in rats (148).

We are grateful to Drs. Michael Fessler and Steve Kleeberger for helpful comments during preparation of this manuscript.