Distribution and fibrotic response following inhalation exposure to multi-walled carbon nanotubes

A number of studies have used inhalation exposure of experimental animals to assess the health risks of multi-walled carbon nanotubes (MWCNTs). Acute inhalation studies have generally studied the pulmonary and systemic responses without a significant recovery in the post-exposure period following exposure. For instance, Li et al. [1] found no significant inflammation and thickening of the alveolar septa in studies at 32 mg/m³ for periods varying from 8 to 24 days with sacrifice at the end of exposure. Mitchell et al. [2] did a range of inhalation exposures from 0.3 to 5 mg/m³ for 7 and 14 days with sacrifice the day after exposure termination. While nonmonotonic systemic immunosuppression after 14 days was observed, no significant histopathology was observed in these acute exposures. In contrast, an acute mouse inhalation study conducted in our laboratory at 10 mg/m³ for 2, 4, 8 or 12 days demonstrated dose-dependent pulmonary inflammation and the rapid development of a fibrotic response [3].

Ryan-Rasmussen et al. [4] conducted MWCNT aerosol exposures in control and ovalbumin-sensitized mice at 100 mg/m³ for 6 hours with recovery periods of 1 and 14 days after inhalation. Fibrotic responses were found in the tissues surrounding airways of ovalbumin-sensitized mice, and there was elevation in whole lung soluble collagen but no significant changes were observed in unsensitized mice exposed to MWCNT aerosol. Rats exposed to MWCNT aerosols up to 2.5 mg/m³ for 13 weeks showed multifocal granulomatous inflammation, diffuse histiocytic and neutrophilic inflammation, and intra-alveolar lipoproteinosis [5], while another 13 week inhalation study of rats to MWCNT reported persistent pulmonary inflammation and granulomas in rats exposed to 0.4 – 6 mg/m³ [6].

We have previously reported the distribution and fibrotic response following aspiration of MWCNTs with analysis of the post-exposure response up to 56 days in mice [7, 8]. In that study, there was an initial inflammatory phase observed in the lungs which lasted approximately 2 weeks. Administration of the MWCNTs by aspiration was conducted with the aid of a previously described dispersion solution [9], which mimicked the contents of the lungs liquid lining layer and resulted in absence of large agglomerates of MWCNTs in the lungs following aspiration. As a result, granulomatous lesions were only rarely observed, and there were no large, airspace agglomerates of MWCNTs in the lungs. The majority of the initial MWCNT lung burden was demonstrated to be in alveolar macrophages. The alveolar interstitium received approximately 8% of the MWCNT lung burden, which was generally in the form of singlets or small, micron diameter MWCNT structures containing only a few fibers. There was development of a fibrotic response in the alveolar interstitium that progressed throughout the 56 day post-exposure period and was observed to be directly correlated to the presence of MWCNTs in the alveolar interstitium.

In this study, we sought to compare the responses between aspiration and inhalation and to extend the analysis out to a post-exposure response of 336 days using a recently developed inhalation exposure system for whole body exposure [10]. This system is capable of delivering a well dispersed aerosol to optimize the alveolar level lung burden without deposition of large, airspace masses.

The use of a well dispersed aerosol establishes a uniform initial distribution of MWCNTs within the lungs. The number of fibers per MWCNT structure was measured in the lungs to address the issue of redistribution of MWCNTs during the ensuing post-exposure period. The number distribution of fibers/nanotubes per MWCNT structure at different times during the post-exposure period was determined by using a combination of morphometry, serial sections and enhanced darkfield microscopy to count the number of MWCNT structures in each size class. Although limited to classifying MWCNT structures containing one, two, three, four and greater than four fibers per structure, our prior studies [11, 12] indicate that the smaller MWCNT structures are the ones which are found within the alveolar interstitium and may play a critical role in the fibrotic response.

Figure 1 shows micrographs from various imaging methods, which demonstrate the distribution of MWCNTs in the lungs 1 day after the termination of the 12 day inhalation exposure period. Figure 1A is a light micrograph of a Sirius Red stained section showing MWCNTs in a bronchiole, alveolar macrophage, and a case of a fiber midway in penetration between the alveolar airspace and alveolar epithelial surface. Granulomatous responses surrounding airspace MWCNT structures were not observed after inhalation of the well dispersed MWCNT aerosol. Figure 1B shows a MWCNT-loaded macrophage in a bronchiole of the lungs. An FESEM image of a MWCNT-loaded alveolar macrophage is shown in Figure 1C. No large, airspace clumps of MWCNTs or granulomatous lesions surrounding MWCNT masses were observed. Penetrations of the visceral pleura by MWCNT fibers (Figure 1D, arrow) were also observed in inhalation-exposed lungs similar to that previously reported for aspiration exposure [13] and acute inhalation exposure [3].

At 14 days and later times post-exposure, MWCNT-loaded alveolar macrophages, similar to Figure 1C and singlet or small clusters of MWCNTs in the alveolar tissue were the most prominent compartments in which MWCNTs were observed. The enhanced darkfield image of Figure 2 shows singlet MWCNTs and small clusters of MWCNTs within alveolar septa of the transition region between a terminal bronchiole and first alveolar duct bifurcation at 168 days after exposure to MWCNTs. As was typically observed at 14 days and later times, a cluster of MWCNTs are present within the ridge of the first alveolar duct bifurcation of Figure 2. Inhaled particles on the ridge of the first alveolar duct bifurcation are the principal site of deposition following inhalation exposures to particles [14] and asbestos fibers [15]. Typically in these inhalation studies, particles or fibers are densely focused on the ridge of the first alveolar duct bifurcation, are proportionately lower in subsequent alveolar duct bifurcations, and are only sparsely found in alveolar septa adjacent to alveolar duct bifurcation ridges. As shown in Figure 2, inhaled MWCNTs also produced a high concentration of fibers on the bifurcation ridge. In addition, singlet or small MWCNT structures were consistently observed in adjacent alveolar septa (small arrows) and throughout the more distal alveolar septa of the lungs. The occurrence of MWCNTs throughout alveolar tissues of the lungs is unlike that observed for other particles.

The lung burden of MWCNTs in airway and alveolar region components of the lungs at 1, 14, 84, 168 and 336 days post-exposure are given in Figure 3. Initially, MWCNTs in the airways accounted for 16 percent of the initial distribution with 84 percent in the alveolar region (including 1.2 percent of MWCNTs in the subpleural tissue region). At 336 days, 4.2 percent of the initial lung burden was found in the airways and 95.8 percent of the initial lung burden remained in the alveolar region (including 4.8 percent in the subpleural tissue region). Alveolar macrophages initially contained nearly 3 fold the burden present in the alveolar tissue and also had the greatest clearance, declining from 15.8 ug at 1 day post-exposure to 10.2 ug at 168 days post-exposure. Over the same period, the content of MWCNTs in the alveolar tissue increased from 5.8 ug at 1 day post-exposure to 9.5 ug at 168 days post-exposure. The airway region, after an initial rapid decline between 1 and 14 days post-exposure, contained a nearly constant, low level of burden between day 14 and 168 days post-exposure. Total lung burdens of MWCNTs were 28.1 ± 0.8, 25.4 ± 1.2, 20.7 ± 1.4, 21.0 ± 1.2, and 18.3 ± 1.1 ug (Mean ± S.E., N = 8) at 1, 14, 84, 168 and 336 days, respectively (Mean ± S.E., N = 8). A highly significant (P < 0.0001) decrease in MWCNT lung burden with post-exposure time was determined.

Examples of MWCNT structures containing 1, 2, 3, 4 and >4 fibers per MWCNT structure are shown in the enhanced darkfield image of Figure 4 (upper image) taken from a 1 day post-exposure lung. Counting of these cases was based on using serial sections to avoid bias of selecting larger fibers and to accurately count the number of fibers per structure and thus determine how the distribution of MWCNT fibers changed once deposited in the lungs. MWCNT fibers and most nanoparticles in general, are efficient at scattering light over a broad wavelength. When the scattered light is imaged by an enhanced darkfield microscope, the fibers appear bright white due to scattering of light, while tissue is dark and airspaces are black. This contrast significantly enhances the ability to detect isolated or rare nanoparticles. The right hand, FESEM image of Figure 4 is an example of a singlet MWCNT penetration of the alveolar epithelium at 168 days post-exposure. Partial shells of the MWCNT structure are visible at the airspace end (arrow).

Measurements of the fiber number per MWCNT structure at 1, 14 and 168 days after the termination of inhalation exposure are given in Figure 5. The results given in Figure 5 demonstrate that clearance depends on the number of fibers in the MWCNT structure. Large agglomerates (>4 fibers/ MWCNT structure) account for the majority of MWCNT clearance from the lungs. From day 1 to 168 days, 28.3 percent of the initial lung burden was cleared from the lungs by removal of large agglomerate (>4 fibers/MWCNT structure). The initial lung burden present as singlet MWCNTs was essentially unchanged over the 168 day post-exposure period. Intermediate percentages of initial lung burden decreasing in the lungs in the cases of 2, 3 and 4 fibers per MWCNT structure. Total MWCNT fiber number of the initial lung burden was 1,321 million based on an initial lung burden of 28.1 ug/lung and a conversion of 47 million MWCNT fibers per ug [16].

MWCNT-induced cytotoxicity in the lung was assessed by determining BAL fluid LDH activities (Figure 6A) and albumin while pulmonary inflammation was assessed by determining BAL polymorphonuclear leukocytes (PMNs) (Figure 6B). The time course of albumin in first BAL fluid followed the same pattern as that of LDH activity (data not shown). MWCNT-exposed mice had significantly higher BAL fluid LDH activities, albumin and BAL PMN levels in comparison to the corresponding air-exposed mice at all post-exposure times. For mice exposed to MWCNTs, BAL fluid LDH activities and BAL PMNs decreased significantly with post-exposure time, but statistically significantly different from controls out to 168 days. At 336 days post-exposure BAL fluid LDH activities, albumin and BAL PMN levels still elevated above controls but no longer statistically different.

Figure 7 shows representative light micrographs of Sirius Red stained sections from a clean-air control lung and a lung at 168 days after MWCNT inhalation. As illustrated in Figure 7, areas of the lung where MWCNT fibers were observed in the interstitial space were found to develop dense bands of fibrillar collagen (arrows). In our prior study using aspiration exposure to MWCNTs, sites with concentrations of interstitial MWCNTs were also found to be foci for a fibrotic response [7, 8].

Morphometric assessment of the fibrillar collagen response, based on Sirius Red binding to fibrillar collagen, at 1, 14, 84, 168 and 336 day post-exposure is given in Figure 8. The average thickness of fibrillar collagen for the clean-air controls was 0.16 ± 0.02 um. In a prior morphometric study, where high resolution transmission electron microscopy imaging was used to identify and quantify fibrillar collagen, at a resolution of individual fibrils, a value of 0.15 ± 0.02 um was obtained in comparable body weight of CD-1 mice [17]. Morphometric measurements of that study were shown to be comparable, albeit approximately 30 percent lower, than to those obtained by biochemical measurements. Biochemical measurement of whole lung or lobe fibrillar collagen also include the connective tissue (fibrillar collagen and elastin) of the airway, blood vessels and their associated bronchovascular cuff. Biochemical measurements significantly over-estimate collagen in the alveolar region and are relatively insensitive to changes in the alveolar region as these non-alveolar areas are rich in fibrillar collagen and contain 68% of the total tissue mass of the lungs and approximately 70% of total lung connective tissue [18].

A trend toward increased fibrillar collagen within alveolar septa is apparent at 14 days post-exposure, which is consistent with our earlier observations of a rapid fibrotic response in acute inhalation exposure to MWCNTs [3]. The increase in fibrillar collagen was significantly elevated over control at 84, 168 and 336 days post-exposure and increases by 62, 56 and 70% above 1 day post-exposure, respectively. Despite the increased thickness of the alveolar fibrillar collagen, mean linear intercept length, a measure of the average size of airspaces in the alveolar region of the lungs was 29.5 ± 0.5, 29.6 ± 0.6, 29.1 ± 0.5, 29.7 ± 0.4, 29.5 ± .0.2 and 29.1 ± .0.2 and 29.5 ± .2 microns for clean-air control, 1 day, 14 day, 84 day and 168 day post-exposure MWCNT-exposed groups. Mean linear intercept length were not significantly different between groups.

A comparison of the fibrillar collagen response between our earlier study [11] using aspiration delivery of MWCNTs to the present inhalation exposure is given in Figure 9. The earlier study, Figure 6 in Mercer et al. [11], used the same morphometric methods to determine the alveolar interstitial collagen fibers response to MWCNT, expressed as the average thickness of connective tissue in the alveolar interstitium, from 1 to 56 days after bolus aspiration.

The time course of response of bolus aspiration and inhalation exposure of MWCNTs to the lungs is similar in spite of the differences in the rate of exposure, bolus versus 12 day exposure period. The magnitudes of alveolar interstitial fibrosis for aspiration vs. inhalation of MWCNTs appear similar in Figure 9. However, it should be noted that lung burden from the aspiration study was approximately 3 fold greater than with inhalation, i.e., 80 and 28 ug, respectively. Therefore, on an equivalent lung burden basis, inhaled MWCNTs are 3 fold more fibrogenic than MWCNTs delivered by bulk aspiration. Although some may find this surprising, the potency difference may reflect greater dispersion of the dry aerosol vs. MWCNTs in a biologically compatible fluid, resulting in delivery of a greater number of small MWCNT structures per equal mass.

The comparable fibrotic responses between bolus aspiration and inhalation of MWCNT in spite of the significant differences in lung burden appear to be in large part to differences in effective delivery of the lung burden to the alveolar interstitium. Although relatively well dispersed on administration, the bolus nature of the aspiration is likely to produce a greater degree of MWCNT agglomeration on the lungs surface as higher concentrations of MWCNTs on the alveolar surface produced by the short duration of exposure are more likely to agglomerate.

Studies of inhaled particle clearance from the lungs are usually based on a rapid phase in which airway clearance occurs, usually within 24 hours after the termination of exposure, and a second slower phase in which particles deposited in the deep alveolar region of the lung are cleared via macrophages [19, 20]. Clearance from the alveolar region is slower and occurs as particle-laden macrophages move up the mucociliary escalator or via the lymphatic to the tracheobronchial lymph nodes [21] where the particles not retained by the lymph nodes are released into the venous system at the junction of the jugular and subclavian veins. Insoluble particles, which are not removed by these processes but remain in the airspaces are frequently walled off by foreign body responses mediated by an epithelioid macrophage reaction, forming granulomas [22]. Fibers are also cleared from the intrapleural space via ducts in the parietal pleura which drain into the lymphatics.

However, Donaldson et al. [23, 24] proposed that long fibers are retained in the stomata which leads to inflammation, fibrosis and potentially mesothelioma. As reviewed by Donaldson et al. [25], this pathway may be particular important as CNT seem to have a special ability to stimulate mesenchymal cell growth and to cause granuloma formation and fibrogenesis. Indeed, Takagi et al. [26] reported that intraperitoneal injection of MWCNT (3 ug/mouse) led to mesothelioma of the abdominal wall. Murphy et al. [24] reported persistent inflammation and fibrosis of the parietal pleural surface 24 weeks after intrapleural injection of long (>15 um) but not short (<4 um) MWCNT (5 ug/mouse). Recently, Schinwald et al. [27] determined the threshold fiber length which would result in pleural inflammation 1 week after intrapleural injection of silver nanofibers of well defined lengths. Results showed a distinct length threshold, with fibers longer than 4 um leading to a pathogenic response. The MWCNT fibers aerosolized in the current study have a mean length of 4.3 um [16]. Given that the mean length of MWCNT fibers in the present study is at the threshold for pathogenic response, it is possible that some effects, such as reduced clearance from the lungs may be due to these mechanisms. As described in the results approximately 1.2 percent of the lung burden was found to translocated to the subpleural tissue and/or visceral pleura in singlet form. This corresponds to 1.2 percent of the 28.1 micrograms of lung burden 1 day post-exposure (see Results) being translocated to the pleural compartment, or approximately 0.35 micrograms, which is an order of magnitude below the burdens at which the aforementioned studies found length-dependent pathologic effects. NIOSH is currently conducting a follow-up exposure to inhaled MWCNT to determine if fiber burden in pleural tissue results in inflammation, lesions and/or transformation of mesothelial cells.

Our studies indicate that the lung burden of MWCNTs is cleared from the alveolar region using these routes. However, the clearance is far from complete at extended times. At 336 days, 65 percent remained in the lungs (Figure 3). This material was most frequently either in smaller MWCNT structures in the alveolar interstitium or larger MWCNT structures associated with macrophages and a smaller but significant component due to MWCNTs in airways at 336 days post-exposure following inhalation (Figure 3). The slow clearance from the airways after inhalation exposure is one of the few differences between our prior aspiration studies [11, 28], where airways were largely cleared of MWCNTs by 14 days, and the inhalation studies of this paper. The significantly greater MWCNT retardation in inhalation-exposed lungs is likely due to differences in how the initial MWCNTs in the airspaces of the alveolar region are processed. In aspiration studies there was a much higher initial proportion of large MWCNT structures in the airspaces of alveoli (not directly contained by macrophages). These large airspace MWCNT structures were either cleared up the mucocilliary escalator or enveloped into granulomas by the foreign body response of epithelioid macrophages. In inhalation, the initial percentage of airspace MWCNTs was lower, since the MWCNT structures were much smaller. These smaller MWCNT structures were rapidly incorporated into the alveolar interstitium where clearance is low.

The rapid decline of lung burden for larger MWCNT structures (Figure 5), the relatively unchanged lung burden for singlet MWCNT structures (Figure 5), the increase in percentage of lung burden distributed to the alveolar tissue compartments at 168 days (Figure 3), and the microscopic observations that fibrotic regions were associated with singlets and smaller MWCNT structures (Figure 7) suggests that there is a disassociation of larger MWCNT structures into small or singlet MWCNT structures. In a parallel study of mice from this exposure, we have recently determined that a significant extrapulmonary transport of singlet MWCNTs (~8%) occurs from these lungs [29]. The absence of any significant decline in singlet MWCNTs in the lungs coupled with the demonstration of extrapulmonary transport of singlet MWCNTs indicates that a source of single MWCNTs exist in the lungs. Only disassociation of the MWCNT structures containing multiple fibers into smaller fiber structures has the capacity to simultaneously supply 8% of initial lung burden to the lymph nodes and maintain the burden of singlet MWCNTs in the lungs at slightly increased levels.

A long debated issue in pulmonary toxicology is the relevance of bolus exposure to particles by intratracheal instillation or pharyngeal aspiration to pulmonary responses resulting from inhalation exposure. We have previously reported that at nearly equivalent lung burdens of MWCNTs, pulmonary inflammation after pharyngeal aspiration and whole body inhalation was similar at 1 day post-exposure [3]. The current study supports this conclusion in that the number of PMNs harvested 1 day after a 12 day inhalation to MWCNTs was not significantly different than the PMN response 1 day after pharyngeal aspiration at a similar lung burden [12]. In the present study, we extend this comparison to a later post-exposure time. Specifically, at 28 days post-exposure, comparison of PMNs after pharyngeal aspiration [12] to those in this study shows that at a similar lung burden MWCNT inhalation induced approximately 4-fold greater pulmonary inflammation in comparison to mice exposed by pharyngeal aspiration. This discrepancy is because pulmonary inflammation after pharyngeal aspiration decreased at a greater rate than that following inhalation exposure to MWCNT. The basis for this difference may be related to two factors.

First, although the MWCNTs used in the pharyngeal aspiration study were well dispersed, they still had some larger structures as evidenced by the presence of granulomas in MWCNT-exposed mice [12]. Granulomas were not observed in the lungs of mice exposed by inhalation exposure in the present study as discussed above. MWCNT structures in the inhalation exposed lung were smaller, and thus more particles per mass were present in the lungs. In addition, as a percentage of lung burden, the MWCNT structures were decreasing in size, maintaining a relatively constant number of MWCNT singlets present in the lung (Figure 5). This disassociation process likely contributes to pulmonary inflammation, resulting in a slowly decreasing but more persistent inflammatory response in the lung.

As we described in the results, despite a nearly 3-fold difference in total lung burden between aspiration and inhalation, the alveolar fibrotic response in the post-exposure period was comparable in both its progression over time and magnitude of response. This appears to be due to the comparable disposition of the lung burden to the alveolar interstitial space by the two different routes of exposure. Of note, alveolar airspace granulomas were present in bolus aspiration exposed lungs but were absent in inhalation exposed lungs, yet both bolus aspiration and inhalation exposed lungs have comparable reactions in the alveolar wall. This occurred in spite of the fact that collagen was contained in granulomas. However, granulomatous collagen was specifically excluded from the fibrillar collagen responses given in Figures 7 and 8. The observation that the two routes of administration yield similar fibrotic response in the alveolar wall at similar doses of lung burden to the alveolar wall suggests that the granulomas with encased MWCNT structures do not significantly alter the alveolar interstitial fibrotic response to MWCNTs.

There are a number of possible causes for the bolus aspiration having a much lower final incorporation into the alveolar interstitium. Of these, the bolus or rapid administration and the uniformity of the initial dispersion on the alveolar surface are likely factors. In the bolus administration all of the MWCNTs are distributed on the surface at one time, greatly facilitating nanotube-nanotube interactions. In the inhalation conducted over 12 day less than 10% of the total burden are distributed on the surface on each day of exposure. The greater effective distance between the MWCNT structures during the inhalation exposures may reduce the likelihood of interaction and agglomeration of the MWCNTs before the MWCNTs translocate from the alveolar surface. Although the bolus method was suspended in a dispersion medium and well separated by sonication at the time of administration, the aerosol formation of material produced by the aspiration maneuver [30] essentially separates the MWCNT solution into a respirable fluid aerosol which contains larger MWCNT structures than the inhalation exposure where the CMAD was 420 nm. These larger MWCNT structures deposited by aspiration are not as rapidly incorporated into the alveolar interstitium as the smaller MWCNT structures.

Results coupling the measurement of MWCNT lung burden with the distribution of MWCNTs in different lung regions (Figure 3) and the size dependence of clearance of MWCNT structures (Figure 5) reveals that clearance from the lungs is neither a simple one or two component process. Lung clearance exhibits multiple distinct and interacting components, i.e., rapid clearance from the airways to a low steady-state level, slow clearance from the alveolar region of the lung by the decline in burden of alveolar macrophages, and an increase in the MWCNT burden incorporated into alveolar tissue (Figure 3). As shown in Figure 5, there were compositional changes of MWCNTs within these compartments as well. The lung burden of singlet MWCNTs and smaller MWCNT structures, which are predominately within the alveolar interstitial space, increased over the 168 day post-exposure period examined (Figure 5). There was also significant, but low level content of MWCNTs remaining in the airways due to the presence of macrophages loaded with MWCNTs. The results clearly demonstrate that the redistribution and clearance of MWCNTs from the lungs is not simply the decline of all MWCNT fibers from a single uniform compartment.

The inhalation exposure in the present study resulted in a lung burden of 28 ug/lung in mice. This resulted in a substantial inflammatory and injury response (Figure 6), which decreased with time post-exposure, and a progressive fibrotic response (Figure 8). Recently our laboratory conducted an inhalation study using the same MWCNT sample which resulted in lung burdens of 0.31, 3.1 and 31.0 ug/mouse [31]. As expected, dose-dependent pulmonary responses were seen. At the lowest lung burden, pulmonary responses were minimal with only small changes in induction of selected inflammatory genes. A full manuscript describing these results is in preparation.

Inhalation exposure to MWCNT produced a fibrotic response that was found to develop and persist out to 336 days after exposure, which is significantly longer that examined in prior bolus aspiration studies. This study also found large differences in the clearance rates from the lungs depending on the size of the MWCNT structures. Clearance of larger MWCNT structures containing greater than 4 fibers from the lung predominately occurs by reduction in the lung burden of MWCNTs within alveolar macrophages. Smaller MWCNT structures in general and singlet MWCNT in particular are formed from dissociation of larger structures and are retained principally within the alveolar septa and sub-pleural tissue. Singlet MWCNTs and smaller MWCNT structures retained within the alveolar septa produce a progressive fibrotic reaction in the lungs.