Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The lung is continuously exposed to environmental particles. The importance of the interaction between the lung and particles has been emphasized by recent studies implicating particulate air pollution in exacerbating respiratory disease and even in elevating mortality rates (1). Particulate matter (PM) < 2.5 µm (PM_2.5) can reach the deep lung and interact with the alveolar epithelium. However, little is known about the alveolar epithelium's ability to internalize inhaled particles. Therefore, knowledge of this cell's ability to internalize inhaled particles has become important.

Several studies have reported phagocytosis of particulate matter by alveolar epithelial cells (6). One study intratracheally instilled two different particles, carbon and thorium dioxide, into rats (6). The two particle types elicited very different responses from alveolar epithelial cells: carbon particles were not ingested by either type I epithelium or type II cells, whereas thorium dioxide particles were ingested by type II but not type I cells. After 19 to 26 h, thorium dioxide was observed in 52% of type II cells and the particles were also found associated with lamellar bodies. In other related studies, type I cells were found to ingest iron oxide particles (9) but nickel particles were not ingested by either alveolar cell type (11). However, it was noted by the authors in the latter study that the type II cells increased their volume densities by 2- to 3-fold after nickel particle exposure and that granular material was associated with extracellular lamellated structures. Together these limited data imply that there are differential capacities of the alveolar epithelial cells to internalize particles and that particle composition may play a role in the process.

The internalization of fibers has been demonstrated in both type I and type II cells. Studies by Suzuki (12) and Pinkerton and colleagues (13) located chrysotile fibers in type II as well as type I cells. They also noted that type II cells undergo substantial morphologic changes, including enlargement of lamellar bodies. Comparing two similar studies with rats inhaling chrysotile fibers, Brody and associates (14) found ingestion only by type I and not type II cells, whereas Pinkerton and coworkers (13) reported ingestion by both cell types. One factor that may explain this discrepancy is the experimental parameters used, specifically particle size. Brody's group (14) used only chrysotile fiber ( 10 µm), whereas Pinkerton's group used a mixture of < 5- and < 10-µm chrysotile particles. It is also possible that Pinkerton's design may have increased the number of particles getting to the deep lung by longer exposures (3- or 12-mo exposure times versus 1 h) and different exposure methods (whole-body exposure chambers versus nose-only exposure). These studies suggest that size and number of fibers contacting the deep lung epithelium may influence their apparent interaction with the alveolar epithelium.

Currently, the role of ultrafine particles as a toxicant, which may explain the epidemiologic associations of increases in PM_2.5 and health effects, is receiving considerable attention (15). It appears that these particles may pass more easily from lung to distant sites (18). How this occurs has not been clearly defined, and the capability of type II epithelial cells to phagocytize ultrafine particles is unknown. For these studies, we chose the human epithelial cell line A549, which originated from a human lung carcinoma (19). This cell line has proved useful, and numerous investigators have used it to study the interactions of human epithelial cells with inhaled foreign material (20- 24). Indeed, one study (25) has proposed the A549 cell line as a standardized model to study lung type II epithelium and drug metabolism. The first objective of the present study was to create an in vitro model composed primarily of polarized, alveolar type II cells. The second objective was to investigate the interaction of A549 cells with an ultrafine particle of titanium dioxide (TiO₂) (50 nm diameter) as a surrogate for ultrafine airborne particles. These particles are inert, nonpathogenic, and found in dusty workplaces such as industries involved in the crushing and grinding of rutile (26).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

A549 Cell Model

The A549 cell line (19) was obtained from American Tissue Type Culture (Rockville, MD). Cells (passage numbers 77-90) were maintained mycoplasma-free in continuous culture in 100-mm tissue culture dishes (Corning, Corning, NY), in F12K Kaighn's modification medium (F12K) (Gibco Laboratories, Grand Island, NY) with 10% heat-inactivated fetal bovine serum (Hyclone, Logan, UT), and 100 U/ml penicillin-100 µg/ml streptomycin at 37°C in 5% CO₂ in a humidified incubator.

At ~ 85% confluence, cells were washed twice with phosphate-buffered saline (PBS)/200 µM ethylenediaminetetraacetic acid and then incubated with 4 ml of 0.4% trypsin in PBS for 5 min. Detached cells were pelleted at 1,200 × g (Beckman GPR centrifuge; Beckman Instruments, Columbia, MD), resuspended in media, and counted with a hemacytometer. Cells were then plated on aclar discs at 0.5 × 10⁶ cells in 24-well plates (Corning).

Aclar Embedding Sheets

Aclar embedding sheets (Ted Pella, Inc., Redding, CA) were cut into circles (< 16 mm diameter) and notched on one edge to indicate orientation (27). They were sterilized in 70% alcohol for 0.5 h and air-dried before placement in sterile 24-well plates and the addition of A549 cells.

TiO₂ Particles

TiO₂ particles were generously provided by Dr. Gunter Oberdörster (University of Rochester, Rochester, NY). Particles were suspended in balanced salt solution (124 mM NaCl, 5.8 mM KCl, 10 mM dextrose, 20 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid, and 1.0 mM CaCl₂), titrated with NaOH to pH 7.4 as stock solutions of 2.5 mg/ml, and sonicated ~ 30 s immediately before use in these assays. Sonications were performed with a Sonifier Cell Disruptor W200P (Heat Systems-Ultrasonics, Inc., Plainview, NY) at a setting of 4. A final concentration of 40 µg/ ml of particles in serum-free F12K was used for most cell treatments. Other treatments used two lower doses (5 and 0.25 µg/ml) which were calculated to cover less than the total surface area of the cells. The stock solution was sonicated 3 min before the final low-dose concentrations were prepared in serum-free F12K approximately 1 h before use. This time allowed for settling of particle aggregates before aliquots were removed.

A549 Monolayer Fixation and Embedment

After treatment, A549 cells attached to aclar discs were fixed with 2.5% glutaraldehyde in 0.1 M potassium phosphate buffer for at least 1 h, washed with 0.1 M potassium phosphate buffer, and then postfixed in 1% OsO₄ in 0.085 M sodium cacodylate buffer for 0.5 h. Samples were washed with 0.085 M sodium cacodylate buffer, dehydrated in a graded ethanol series and propylene oxide, infiltrated with a mixture of araldite 502 (Ernest F. Fullam, Inc., Latham, NY) and propylene oxide (1:2) for 2 h, and followed by a mixture of araldite 502 and propylene oxide (1:1) overnight. Aclar discs were then cut into two to three rectangles, placed on the bottom of flat bed molds (with the cells toward the top of the molds) with 100% araldite 502, and incubated at 68°C for 48 h. Blocks were then removed from the oven, separated from the aclar, and re-embedded with fresh araldite 502 at 68°C for an additional 48 h. Sections of 30 nm from each group were made with an RMC MT6000 ultramicrotome (RMC, Inc., Tucson, AZ), and placed unstained onto uncoated 600-mesh copper grids for examination with a LEO EM902 energy-filtering transmission electron microscope (Leo Electron Microscopy, Inc., Thornwood, NY). Occasionally, 30-nm sections on 600-mesh grids were poststained with uranyl acetate and then examined with energy-filtering transmission electron microscopy (EFTEM). Sections of 60 nm from each group were made with an ultramicrotome, placed onto 200-mesh grids, stained with uranyl acetate and Reynolds lead citrate, and examined using conventional transmission electron microscopy (TEM) (Leo Electron Microscopy).

A549 Cell Pellet Fixation and Embedment

To compare cell morphology with aclar-plated cells, freshly harvested cells were fixed just before plating (0 h control). Cells grown on 100-mm petri dish were harvested as described, then 1 × 10⁶ cells were pelleted in a microcentrifuge at 735 × g for 1 min and cells were then resuspended in 2.5% glutaraldehyde in 0.1 M potassium phosphate buffer. Fixation, dehydration, and infiltration procedures were as described earlier. After infiltration, the pellet was embedded in 100% araldite 502 and incubated at 68°C for 48 h. Sections of 30 and 60 nm were taken as described for A549 monolayers in the preceding section, and then examined with EFTEM or conventional TEM, respectively.

Cell Morphology Assay

Initial trials were performed to determine the culturing time required for A549 cells to obtain cuboidal, epithelial-like characteristics. Distinguishing morphologic characteristics, as determined by examination with EFTEM and conventional TEM, included lamellar body structures (a verifying feature) as well as a cuboidal shape and apical surface projections. Only a minimal number of attenuated, epithelial-like cells (characteristic of type I alveolar epithelial cells) were observed in these optimal growing conditions in any given field at low magnification within the ultrathin section.

Cells (0.5 × 10⁶) were plated onto aclar discs in a 24-well plate and incubated for 0 or 24 h (± 1 h), or 2, 3, 4, 7, 9, or 11 d. At the times indicated, the cells were fixed and processed for ultrastructural examination. After an initial experiment, two subsequent trials focused solely on Days 2, 3, and 4 to determine optimal culture times.

Particle Exposure Assay

A549 cells (5 × 10⁵ cells) were plated onto aclar discs in 24-well plates and cultured for 4 d. Cells were then washed twice with serum-free F12K before treatment with 40 µg/ml (final concentration) of TiO₂ particles for varying exposure times (1, 2, 3, 4, and 24 h). For the subsequent two experiments, the exposure times used were 3, 6, and 24 h. The cells were then processed and examined by EFTEM to identify Ti and O by electron energy loss spectroscopy (EELS) and electron spectroscopic imaging (ESI). Two control experiments were performed. Cells (0.5 × 10⁶) were plated onto aclar discs in 24-well plates for 4 d (0 h control) and 5 d (corresponding to particle exposure for 24 h) and then fixed with 2.5% glutaraldehyde. Microscopy analysis was used to verify internalization and distribution of TiO₂ particles within the alveolar type II cells. These exposure time points (6 and 24 h) yielded substantial internalization and were used for the final assay experiments.

In a separate experiment and before exposure to particles (6 and 24 h), some wells received 2 µg/ml of cytochalasin D (cyto D) (Sigma, St. Louis, MO) in 0.1% dimethyl sulfoxide for 30 min. Three controls were employed as follows: (1) A group of A549 cells was fixed after a 4-d incubation and before particle exposure. (2) A group was incubated with serum-free F12K without TiO₂ particles or cyto D. After a 24-h exposure, they were gently washed twice with serum-free F12K and then fixed. (3) A final group tested the viability of A549 cells that had been incubated with cyto D for 30 min before particle exposure and maintained throughout the 24-h exposure period. Viability was determined with trypan blue exclusion. After 30 min of incubation with cyto D, wells were gently washed twice with serum-free F12K. They were then treated with 40 µg/ml TiO₂ particles for 6 and 24 h. At the end of the exposures the cells were washed twice with serum-free F12K, fixed, and then processed for EFTEM or conventional TEM examination. Each experiment was repeated three times.

Another set of experiments examined the uptake of ultrafine particles at reduced particle exposure levels. A549 cells were treated with 0.25 or 5 µg/ml TiO₂ particles for 24 h. At the end of the exposures, cells were washed twice with serum-free F12K, fixed, and then processed for EFTEM examination.

Examination by ESI and EELS

Various fields (morphologic ultrastructure of numerous A549 cell cytosols viewed at high magnification by EFTEM) were mapped for Ti and O by ESI at ×20,000. The EFTEM took place with the following parameters constant: 80 kV; beam current at 60 µA; 20-eV spectrometer slit resolution; 60-µm objective aperture; 0.8-mm spectrometer aperture; in image mode. Background noise levels for the system used in these experiments were established by a modified, previously described method (28). A range of three images (428, 438, and 448 eV for Ti; and 514, 504, and 494 eV for O) was used to determine background noise. The LEO CEM 902 elemental analysis program (29) was used to collect images at 418 and 438 eV and a peak image from Ti at 458 eV. The same was done for O, collecting images at 494 and 514 eV and a peak image at 534 eV. Ti and O were discriminated on the basis of the background noise determination and their distributions were represented as a binary map of the element.

Serial spectra were collected from representative fields of the same sections that had been mapped for Ti and O distribution. The range of the spectra started at 400 eV and ended at 575 eV. TiO₂ was identified by EELS, with the Ti L₂ and L₃ edges at 456 and 462 eV, respectively. The O K edge was at 532 eV. Parameter settings for EFTEM were as follows: 80 kV; beam current at 60 µA; 1-eV slit resolution; and 0.1-mm spectrometer aperture, performed in spectrum mode.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Morphology of Cultured A549 Cells

The first objective of this study was to create an in vitro model of deep lung epithelium that would be composed primarily of type II cells. Type II cells have distinguishing ultrastructural features, including cuboidal-like shape, extensive cytoplasm with rough endoplasmic reticulum, tight junctions, and lamellar bodies. To an extent, many of these characteristics were observed in cells fixed as pellets (Figure 1A). However, they also had free surfaces with cytoplasmic extensions on all sides, and polarization, if present, could not be determined. To generate polarized cells with known orientation, A549 cells were cultured on aclar substrates. After 4 d in culture, distinct polarization was observed with a single, unattached surface (Figure 1C). Other ultrastructural characteristics common to type II cells were also observed, including complex Golgi apparatus (not shown), tight junctions, lamellar bodies, and extensive cytoplasmic extensions. An optimal growth time of 4 d was selected at which time a monolayer primarily of epithelial-like type II cells had been established with cytoplasm having distinguishing features of alveolar type II cells. Cells plated for only 1 and 2 d (not shown) had a preponderance of attenuated cytoplasm and few organelles, and resembled epithelial-like type I cells. Lamellar bodies, a definitive marker used in this study, were difficult to find in these type-I-like cells. At Day 3 in culture, fields of coexisting type I and type II cells were observed in ultrathin sections. Cells grown for 5, 7, 9, and 11 d no longer formed consistent monolayers but instead formed multilayers (not shown).

View larger version (162K): [in this window] [in a new window]   Figure 1.   Comparative morphology of A549 epithelial cells removed from culture plates with trypsin and pelleted, or of cells cultured 4 d on aclar film. (A) Representative conventional TEM of A549 cells after culture on plastic dishes for 4 d and subsequent removal by trypsin. (B) ESI at 250 eV of cultured A549 cell with loosely packed lamellar bodies (LB). (C) Conventional TEM low magnification of A549 cells grown in culture on aclar film. Note extensive cytoplasm in these cells, characteristic of type II cells. Asterisks indicate free cell surface; ooo indicates cell surface attached to aclar film. In C: Nu, nucleus; arrows identify clusters of lamellar bodies. Bar: 1.7 µm in A; 0.5 µm in B; 2.5 µm in C.

Ultrafine Particle Exposures

Initial experiments sought to determine whether A549 cells would take up ultrafine TiO₂ particles after 1, 2, 3, 4, or 24 h of exposure. Examination by EFTEM indicated minimal internalization at the earliest time points (1 and 2 h, not shown). Therefore, in subsequent experiments, A549 cells were exposed to particles for 3, 6, and 24 h. At each sampling time (3, 6, and 24 h), phagocytized TiO₂ particles could be easily observed (Figure 2). One type of phagosome, a membrane-bound vacuole containing TiO₂ particles, was observed most often in the early exposures. Occasionally, particles were also found associated with lamellar bodies. Membrane-bound vacuoles contained mostly large aggregates of TiO₂ particles, but in some instances as few as two or three particles were observed in a single internalized cluster. Conventional TEM illustrated one A549 cell in the act of engulfing aggregates of TiO₂ particles after 6 h of exposure (Figure 2A). This observation was made most easily after 6 and 24 h of exposure. We also found evidence of intracellular movement of particle aggregates when membrane-bound clusters were located centrally within the A549 cytoplasm (Figure 2B). External TiO₂ was quite abundant along the apical surface of the plasma membrane at 3, 6, and 24 h (Figure 3).

View larger version (50K): [in this window] [in a new window]   Figure 2.   Conventional TEM of A549 cells cultured 4 d on aclar and followed by 6 h exposure to 40 µg/ml TiO2 particles. Arrows indicate ultrafine particles engulfed by cells. (A) Note the o below a lamellar body containing TiO2 particles. (B) Asterisks identify membrane-bound aggregate of particles. Nu, nucleus. Bars: 1.1 µm.

View larger version (60K): [in this window] [in a new window]   Figure 3.   ESI at 250 eV and low magnification of a group of A549 cells cultured on aclar and after 24 h exposure to 40 µg/ml of TiO2 particles. Asterisks indicate side approximating the aclar disk. Bar: 2.5 µm.

Unlike the occasional association of TiO₂ particles with lamellar bodies found in the 6-h-exposure group, particles were more easily observed in lamellar bodies after 24 h, suggesting that this placement occurs subsequent to vacuole formation. The lamellar bodies had TiO₂ particles not only associated but also enmeshed among the membranes within this membrane-bound vacuole. TiO₂ particles were associated with both loosely and tightly packed lamellar bodies (Figures 4A and 4B). Even after 24 h of exposure there were large amounts of externally associated TiO₂ particles on the apical side of the cell. Cells in these figures were not washed before fixation; therefore, particles may or may not be bound by the cells. A549 cells also contained large aggregates of TiO₂ particles in membrane-bound vacuoles (Figure 4B). Interestingly, A549 cells exposed for the longest time (24 h) had TiO₂ particles around the nuclear region as well as occasional particles found near the basal end of the cell.

View larger version (69K): [in this window] [in a new window]   Figure 4.   High-magnification ESI image at 250 eV of an A549 cell after 24 h exposure to 40 µg/ml of TiO2 particles. Typically, particles were associated with both loosely (A) and tightly packed (B) lamellar bodies in aclar-cultured A549 cells. Bars: 4.4 µm.

Examination by ESI and EELS

To show that the electron-dense particles described were those added to the cell culture, elemental mapping and EELS were performed to localize Ti and O. Intracellular particles (Figure 5A) in loosely packed lamellar bodies were assessed by ESI for O (Figure 5B) and Ti (Figure 5C). The elemental map was made using a 20-eV spectrometer window and CEM software Rel. 1.31 with subtraction of the element on the basis of a determination of background noise. Figure 5D is an EELS spectrum of another membrane-bound aggregate of intracellular particles showing both Ti and O signals. These observations confirm that the particles observed were ingested ultrafine TiO₂ particles that had been incorporated into membrane-bound vacuoles and lamellar bodies.

View larger version (146K): [in this window] [in a new window]   Figure 5.   ESI at 250 eV of an aggregate of ultrafine TiO2 particles in a loosely packed lamellar body of an A549 cell after a 24-h exposure (A). Overlay of O distribution (B) and Ti distribution (C ) determined with elemental mapping. EELS spectroscopy spectrum (D) indicating the O and Ti peaks for a different aggregate of particles. Bar: 200 nm.

Cyto D Treatments

To begin assessing the role of microfilaments in the uptake of TiO₂ particles, cells were preincubated in cyto D for 0.5 h before particles were added, and the cells were examined with ESI after 6 and 24 h. After 6 and 24 h, cells had internalized aggregates of TiO₂ particles in both membrane-bound vacuoles and lamellar bodies. It was immediately evident that cells treated with cyto D appeared to have ingested most of the large aggregates of TiO₂ particles within membrane-bound vacuoles (Figure 6); whereas without cyto D, the aggregates were more often found enmeshed in lamellar bodies. The free surfaces of the cells were still associated with external particles even after two washes. The plasma membrane of the cyto D-treated A549 cells had markedly fewer cytoplasmic extensions than did the untreated cells (Figure 6A) and also appeared to have fewer aggregates of particles enmeshed in the lamellar membranes. In addition, particles within the lamellar bodies of cyto D-treated cells (Figure 6B) were most often observed around the periphery of these structures. Finally, cyto D-treated cells appeared qualitatively to have fewer lamellar bodies (not shown). A549 cells maintained high viability after being exposed to 2 µg/ml of cyto D for 0.5 h and serum-free media for an additional 6 or 24 h. Cell viability for cyto D-treated cells after 6.5 and 24.5 h (0.5-h cyto D pretreatment and 6- or 24-h particle exposure), determined by trypan blue, was 91.3 and 88.4%, respectively, (n = 3).

View larger version (69K): [in this window] [in a new window]   Figure 6.   ESI at 260 and 264 eV (A and B, respectively) of cyto D-pretreated A549 cells subsequently treated with 40 µg/ml TiO2 particles for 24 h. Note in A the smooth plasma membrane and the large aggregation of membrane-bound particles; arrows in A point to membrane-bound aggregates of ultrafine particles; Nu, nucleus. Asterisks indicate the cell surface approximating the aclar disk; X in B indicates particles engulfed with atypical projections relative to the untreated cells. A549 cells, 30-nm sections, poststained with uranyl acetate. Bar: 1.1 µm in A; 0.6 µm in B.

Uptake of Ultrafine Particles at Reduced Exposure Levels

These experiments examined A549 cell particle uptake at lower exposure concentrations (0.25 and 5 µg/ml) for 24 h. For cells exposed to the lowest concentration (0.25 µg/ml), extensive examination of multiple fields by EFTEM revealed no particles associated with the free surface of the plasma membrane, or particles engulfed or internalized in the cytosol (Figure 7A). At 5 µg/ml, aggregates of particles were easily observed both associated with plasma membrane and internalized into the cytosol (Figures 7B and 7C). Compared with the 40-µg/ml treatment, fewer individual particles and particle "clusters" were found associated with the free plasma membrane at the 5-µg/ml treatment.

View larger version (141K): [in this window] [in a new window]   Figure 7.   Images at 250 eV of A549 cells after 24 h exposure to low-dose concentrations of TiO2 particles. (A) ESI of A549 cells at low magnification after treatment with 0.25 µg/ml of TiO2 particles. Note the particle-free apical surface of plasma membrane as opposed to the cluttered free surface of the cells treated with 40 µg/ml of TiO2 particles. Asterisks approximate the aclar disks. (B) ESI at high magnification of A549 cells treated with 5 µg/ml TiO2 particles showing the endocytosis of particles. (C ) ESI of enmeshed TiO2 particles in a lamellar body (LB) from cell treated as in B; Nu, nucleus. Bars: 2.5 µm in A, 0.4 µm in B and C.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

We were successful in determining culture conditions that led to an approximate monolayer of epithelial type II-like cells displaying cellular polarity. A549 cells grown on aclar discs generated cells with a typical epithelial cell polarity, establishing two compositionally discrete but continuous plasma membrane regions. The cultured cells demonstrated one free surface with microvilli similar to the surface found in the lumen of the alveolus (apical side) and an aclar interface lacking microvilli. In contrast, cells trypsonized from standard tissue culture plasticware yielded high-quality morphology under the electron microscope but also yielded cells that had free plasma membrane with cytoplasmic projections evident on all surfaces. Another disadvantage of the cell-pellet method is that it is no longer possible to identify regions of the epithelial plasma membrane that may at one time have corresponded to the apical surface. By growing A549 cells on the aclar film we were able to expose cells to particles on an "appropriate" surface before fixation and electron microscopic examination.

Our in vitro model is suitable for particle localization in epithelial cells because the aclar substrate allowed cells to polarize. Aclar appears to be superior to Snapwell polycarbonate filters (25) because mechanical shearing is avoided in making ultrathin 30-nm sections. In this study aclar was actually removed before sectioning by the double-embedment process, eliminating shearing problems. With the polycarbonate there was evidence of shearing that may be acceptable for some studies but would limit detailed ultrastructural analysis. An important feature of the aclar model is that particle placement is not altered due to shearing during the sectioning process.

Observing cellular morphology of A549 cells in culture over an 11-d period allowed us to select an optimal time for particle exposure assays. At culture times shorter than 4 d the cells were morphologically a mixture of type I and type II cells. Type I cells resemble squamous-like epithelial cells with limited cytoplasm and cytoplasmic organelles. Type II cells have a polygonal-to-cuboidal shape and more discreet organelles, including characteristic lamellar bodies, a more elaborate and well-developed Golgi complex, and extensive rough endoplasmic reticulum. With culture times substantially greater than 4 d, the polygonal-to-cuboidal shape prevailed; however, cells piled on top of each other, forming layers of variable thicknesses. Culturing for 4 d yielded an approximate monolayer of cells with characteristics of type II epithelium.

For the purposes of this study, particle treatment times of 6 and 24 h were selected. Initial observations indicated that before 6 h, ingestion of particles by the cells was not substantial and it was difficult to image cells demonstrating particle uptake. However, we were able to observe limited ingestion of particles as early as 1 h after treatment, indicating that internalization could be relatively rapid and was a continuous process. After the longer exposure periods we could more easily observe TiO₂ particle ingestion. In addition, movement within the cytoplasm toward the basal membrane was indicated by the perinuclear localization of the particle clusters not seen with shorter exposure times.

The uptake of particles by A549 cells was apparently limited to aggregates of ultrafine particles. We observed plasma membrane projections surrounding and engulfing these aggregates of particles before internalization. Additionally, there were only limited observations of single, membrane-associated particles, suggesting potential uptake of individual ultrafine particles. This observation is consistent with the widely held hypothesis that individual ultrafine particles are not targets of phagocytosis. However, at this time we cannot rule out the possibility that a vast majority of particles had aggregated before settling on the A549 cells. Interestingly, TiO₂ particles were never observed moving through the tight junctions between cells. Instead, aggregates of TiO₂ particles entered cells by ingestion and were later found in membrane-bound vacuoles and enmeshed in lamellar bodies. Association of particles with lamellar bodies has been noted in other studies (6, 9). No particles were observed on the aclar or basal side of the cell, indicating that the particles were not exocytized from the cell during the incubation period. The mechanism or sequence of events that leads to relocalization of particles to the basolateral membrane is not known.

The identity and composition of the electron-dense particles, assumed to be TiO₂, was determined by EFTEM. Both ESI and EELS established that the electron-dense particles in various fields at ×20,000 did contain Ti and O elements. Ti's major L2 and L3 edges were 456 and 462 eV, respectively, and O's K edge was 532 eV. Other fields and other time points were analyzed by elemental mapping and/or EELS and the particles were also found to contain both Ti and O elements. The power of this technique is that unequivocal identification of particles is possible, whereas in vivo studies in the past have relied on the visualization of the instilled particles' shape to establish identity (30). For example, the granular material reported associated with the lamellated structures in the lumen from type II cells exposed to nickel particles was most likely composed of the experimentally instilled metal, but this was not determined (11).

Surprisingly, phagocytosis of ultrafine particle aggregates was not inhibited by preincubation with cyto D. TiO₂ particles were ingested after 6 and 24 h and were localized to lamellar bodies as well as membrane-bound vacuoles (Figure 6). Indeed, the cyto D-treated A549 cells appeared to have more aggregates of TiO₂ in membrane-bound vacuoles than did the untreated group, although this observation was not quantified. However, it was apparent that the cyto D exposure did have a substantial effect on the treated cells inasmuch as the plasma membrane of these cells appeared smooth and was only punctated by projections encapsulating particles. In our protocol, cells were treated with 2 µg/ml cyto D for 30 min and the media were removed before addition of TiO₂ particles. Dose-response experiments with cyto D and the A549 cell line have shown that at least a 0.5 µg/ml concentration is required to inhibit ingestion of bacteria (21, 22). In our protocol it was important to remove the cyto D before adding TiO₂ particles because of our extended exposure times of 6 and 24 h. Extended exposure of high levels of cyto D can result in substantial retraction of cell spreading and in apparent cell death with cultured endothelial cells (31). Although morphometric analysis was not performed, we also observed substantially fewer lamellar bodies present in cyto D-treated cells when compared with untreated controls. With and without cyto D treatment, cells at 6 and 24 h had substantial accumulation of particles within the lamellar bodies. These results indicate that there may be some subtle effects of cyto D on the synthesis of lamellar bodies. Additional studies are planned using reduced levels of cyto D throughout the particle exposure separately and in combination with other inhibitors, such as colchicine.

The interaction between ultrafine particles and A549 cells at lower particle-to-cell ratios was also studied after 24 h of exposure. Our objective was to allow a single layer of TiO₂ particles to cover the surface of the cultured cells, thereby reducing particle clustering. A concentration of 5 µg/ ml was calculated to cover just under the total cellular surface area. A 0.25-µg/ml dose concentration was approximately 20 times less than the total cell surface area. After 24 h of exposure at 5 µg/ml, we observed a reduced overall amount of extracellular particles on the free surface of the cells (compare Figures 3 and 7). Unexpectedly, a majority of particles were still associated as aggregates, despite increased sonication time and the use of a settling period. Ingestion continued to involve aggregates of ultrafine particles, not individual particles. Churg and colleagues (32) recently reported that ultrafine TiO₂ particles were recognized and phagocytized by the upper airway epithelium but only as aggregates. Our observations also tend to support the hypothesis that individual ultrafine particles are rarely phagocytized by type II epithelium. Internalized, aggregated particles were found enmeshed within the lamellar bodies and were also found in perinuclear regions. Upon reducing the particle concentration to 0.25 µg/ml, we were unable to observe TiO₂ particles either internalized or associated with the plasma membrane (observing the same number of sectioned electron microscopy blocks as with the 5-µg/ml dose). It is possible that clusters did not form at this low level of particles and therefore the particles did not become associated with the cell monolayer.

In conclusion, we have established an in vitro model to study endocytosis by epithelial type II cells. A549 cells did internalize ultrafine (50-nm diameter) TiO₂ particles but apparently only as aggregates. Phagocytosis of particles occurred as early as 1 h after exposure and continued for the entire 24-h exposure period. After 24 h of exposure, considerably more internalization of particles was noted than at the 6-h time point, with a majority of particles closer to the free surface as opposed to the cell side approximating the aclar disc. TiO₂ particles were observed throughout the A549 cytoplasm. Even after the cells were washed twice, TiO₂ particles were still associated extracellularly with the free surface of the cells. Once internalized, large aggregates of TiO₂ particles were found predominantly in membrane-bound vacuoles or enmeshed in loosely or tightly bound lamellar bodies. In some rare instances, small numbers of ultrafine particles (two or three) were seen within a single membrane-bound vacuole. At lower particle concentrations there were fewer particles associated with the apical cell surface and these were still in aggregates. Pretreatment of cells with cyto D did not inhibit internalization of TiO₂ particles. There may be a subtle effect on the association of particle-containing vacuoles with the lamellar bodies inasmuch as very few cyto D-treated cells had TiO₂ particles enmeshed in the lamellar bodies as compared with the untreated cells.