Introduction

Respirable quartz has been classified as a human group I carcinogen (IARC 1997), and the development of a chronic inflammation and silicosis is known as a pathological event following quartz exposure (Castranova 2000; Rimal et al. 2005). A human study in coal miners relates pulmonary inflammatory response to the amount of crystalline silica within the inhaled coal mine dust (Kuempel et al. 2003). In various animal studies, the effects of quartz on lung inflammation and damage have been shown (Albrecht et al. 2004; Porter et al. 2004). Generation of reactive oxygen species (ROS) and the activation of redox-sensitive transcription factors like nuclear factor kappa B (NFκB) have been demonstrated in inflammatory cells including alveolar macrophages (AM) as well as in lung epithelial cells of rats (Duffin et al. 2001; Albrecht et al. 2004, 2005) and mice (Hubbard et al. 2002). In vivo as well as in vitro exposure to quartz induces the transcription of inflammatory genes and the subsequent release of cytokines and chemokines from AM (Rao et al. 2004). Furthermore, the relation between silica-induced ROS generation and the activation of the transcription factors like NFκB and activator protein-1 (AP-1) has been demonstrated (Kang et al. 2000; Gwinn and Vallayathan 2006). In vitro exposure of RAW 264.7 cells to crystalline silica induces tumour necrosis factor-α (TNF-α) and interleukin-6 (IL-6) release as well as a reduction of the phagocytic capacity (Balduzzi et al. 2004). Also an in vivo exposure of monkeys to DQ12 resulted in a reduced rate of phagocytosis of fluorescent polystyrene beads in cells obtained by bronchoalveolar lavage (BAL) (Hildemann et al. 1992). These data provide a broad knowledge about the consequences of quartz exposure in which AM seem to be a key player in the clearance as well as the initial inflammatory response.

In the lungs, AM functions are governed by a sensitive equilibrium and the imbalance may cause the development of chronic diffuse lung diseases (Fels and Cohn 1986; Keane and Strieter 2002; Rubins 2003; Fujiwara and Kobayashi 2005). Such an imbalance might also be the initial step in the development of chronic inflammation after quartz exposure. However, the relevance and the mechanism of particle uptake in this process still remains to be elucidated.

The process of phagocytosis requires the recruitment of actin filaments for the internalization process (Allen and Underhill 1999). This actin recruitment by polymerisation/depolymerisation exhibits the critical step during the phagocytic process. Cytochalasin-D (Cyt-D), known to block actin polymerisation by binding to the barbed (plus) end of the actin filament, is described as a potent inhibitor of phagocytosis (Wakatsuki et al. 2001; Tjelle et al. 2000). During internalization, AM generate ROS (O −2 and H2O2) under activation of NADPH oxidase (Park 2003). This ROS generation may also act as a second messenger, which can trigger several signalling pathways including the activation of NFκB resulting in the production of pro-inflammatory cytokines such as TNF-α (Iles and Forman 2002). Also an influence of ROS generation on particle uptake in epithelial cells is discussed (Churg 1996).

Although, an uptake of quartz particles by a classical phagocytosis mechanism is indicated by several hints, the mechanism of quartz uptake by macrophages as well as the importance of the uptake process itself or the involvement of actin in the inflammatory response to particles is still to be elucidated. Therefore, the aim of the present study was to investigate the role of actin as a mediator of particle uptake as well as particle-induced oxidative stress generation and TNF-α release. Consequently, we studied the dose- and time-dependent uptake of quartz particles in macrophages using NR8383 cells, a macrophage cell line, in comparison to primary AM and interstitial macrophages (IM). Furthermore, we focused on the role of actin in particle uptake and the subsequent inflammatory response by measurement of ROS generation and TNF-α release.

Materials and methods

Cell culture

NR8383 cells, an immortalized rat alveolar macrophage cell line, were cultured in Kaihn´s modified medium (F12-K Nutrient Mixture, Gibco, Eggenstein, Germany) containing 15% fetal calf serum (FCS), 1% Penicillin/Streptomycin and 1% Glutamine (Sigma-Aldrich, Seelze, Germany) by 37°C, 5% CO2. For experiments, 1.1×105 cells/cm2 were seeded and kept for 3 days in culture before treatment.

Isolation of primary cells

AM as well as IM were isolated from 12-week-old female Wistar rats (mean body weight: 260 g, Janvier, Le Genest St. Isle, France). The animals were housed and maintained in an accredited on-site testing facility, according to the guidelines of the Society for Laboratory Animal Science (GV-SOLAS). Rats were sacrificed by deep anaesthetization with Na-pentobarbital (50 mg/kg body weight) and subsequent exsanguination via the abdominal aorta. To obtain AM, lungs were cannulated via the trachea in situ and BAL was performed as described previously (Knaapen et al. 2002). BAL macrophages obtained from perfused lungs were centrifuged (4°C, 800 g, 10 min), washed with PBS and resuspended in Kaihn´s modified culture medium containing FCS (15%) as well as 1% Penicillin/Streptomycin and 1% Glutamine. IM were isolated by lung digestion according to the method described by Richards et al. (1987) with minor modifications (Knaapen et al. 2002). Briefly, after trypsination, lungs (20 ml trypsin per lung, 30 min, 37°C water bath) were chopped for 5 min and incubated with 5 ml FCS containing 500 μl DNase (4 mg/ml, Sigma-Aldrich, Seelze, Germany). After adding 20 ml PBS, shaking and filtration through gauze and a 30-μm nylon filter, cells were separated by a Percoll gradient centrifugation (20 min, 800 g, 10°C). Then, isolated cells were washed with PBS and resuspended in DMEM/5% FCS containing DNase (4 mg/ml). Incubation for 60 min at 37°C and 5% CO2 allows IM as well as neutrophils and fibroblasts to settle down, whereas epithelial cells stay in suspension. The adherent fraction was harvested by scraping, and IM were obtained by centrifugation (4°C, 800 g, 5 min), washed with PBS and resuspended in culture medium. For treatment, wherefore only passage one was used, primary cells were seeded (1.1×105 cells) and cultured overnight in Kaihn´s modified medium (15% FCS, 1% Penicillin/Streptomycin, 1% Glutamine) at 37°C and 5% CO2. Purification of 98 and 89% for AM and IM, respectively, was determined by analyzing May-Grünwald-Giemsa (MGG)-stained cytospin preparations. In addition, ED-1 immunohistochemical staining as a macrophage specific marker was performed to confirm the results of the purification analysis.

Particles and cell treatment

All cell types (AM, IM and NR8383 cells) were equally treated with DQ12 (Dörentruper quartz, Batch 6, mean diameter 0.96 μm) at concentrations of 0, 10, 20, 40 μg/cm2 for 1, 2, 4 or 24 h using a freshly prepared DQ12 suspension (5 mg/ml). Analysis using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay revealed that these treatment conditions were not toxic (data not shown). In order to prepare quartz suspensions, DQ12 was baked at 220°C for 14 h to destroy endotoxins, and subsequently sonicated for 5 min (water bath sonicator, Sonorex TK 52, Schaltech, Mörfelden-Walldorf, Germany) either in medium, or for measurement of respiratory burst in, HBSS+/+ (Mg2+/Ca2+). For inhibition of actin-recruitment, cells were pre-incubated with Cyt-D (1.5 μg/ml, Sigma-Aldrich, Seelze, Germany) or DMSO (Dimethyl sulphoxide, 0.75%) as vehicle control at non-toxic doses for 30 min at 37°C and 5% CO2. Treatments for uptake studies as well as analyses of TNF-α release were performed in medium.

For measurement of ROS generation, the medium was removed to avoid scavenging effects. Therefore, macrophages were washed once with HBSS+/+, supernatants were centrifuged (10 min, 25°C, 900 g), and the resuspended cell pellet (non-adherent fraction) was added to the adherent cell-layer. Following 30 min recovery (5% CO2, 37°C), cells were pre-incubated with freshly prepared dihydrorhodamine (DHR)-solution (4.3 μM, DHR 123, Molecular Probes, Leiden, Netherlands) for 30 min for flow cytometric analysis of ROS generation. Pretreatment with Cyt-D as well as quartz treatment was performed as described before. For electron spin resonance (ESR) analysis of ROS generation, the quartz preparation (DQ12 suspended in HBSS+/+, 10 or 40 μg/cm2) and the spin trap DMPO (5,5-Dimethyl-1-pyrroline N-oxide, 0.11 M, Sigma-Aldrich, Seelze, Germany) were added simultaneously after pre-incubation with the actin-inhibitor Cyt-D (1.5 μg/ml) or the vehicle control DMSO and incubated for 1 or 3 h at 37°C, 5% CO2.

Cell preparation and flow cytometric measurement of particle uptake and intracellular ROS generation

In order to measure the uptake of quartz particles, a flow cytometric approach developed by Stringer et al. (Stringer et al. 1995) was used. Therefore, treated cells were scraped on ice, centrifuged (900 g, 5 min, 4°C) and washed twice with 1 ml ice cold HBSS-/- (Mg2+/Ca2+). Finally, macrophages were resuspended in 400 μl HBSS-/- and stored on ice until measurement. The uptake of particles was determined by flow cytometry using FACSCalibur (fluorescence-activated cell sorter, Becton Dickinson, Heidelberg, Germany) measuring side-scatter angle (SSC) and forward-scatter angle (FSC) of 10,000 counts. The SSC is directly related to the cell granularity and was used as a measurement of particle uptake (Palecanda and Kobzik 2000), whereas FSC correlates to the cell size and was used as a cofactor. Data were analyzed with the Cell Quest 3.3.Software (Becton Dickinson, Heidelberg, Germany) as follows. In bivariant FSC-SSC-histograms of SSC and FSC, cell debris and free particles were excluded by an electronical gate containing cells of all sizes and granularities. In univariant histograms of SSC or FSC, the median of cell granularity or cell size was determined. From these values, the mean of cell size and granularity of untreated controls as well as quartz-treated cells were calculated to analyze the cell size as well as the uptake which is presented as percentage of control. For the measurement of intracellular ROS generation by the oxidation of DHR, additionally to FSC and SSC the DHR fluorescence was analyzed by the fluorescence (530 nm)-detector of the flow cytometer (Imrich et al. 1999). After the preparation of macrophages as described previously, the simultaneous measurement of cell size, particle uptake and intracellular ROS generation was performed counting 10,000 cells with the FACSCalibur. Afterwards, the median of FSC, SSC and DHR fluorescence were determined and analyzed as described before. All flow cytometric experiments were done in triplicates and repeated in three to five independent experiments as noted in the results.

Electron spin resonance (ESR) measurement of extracellular ROS generation

DQ12-induced formation of hydroxyl radicals was measured extracellularly by ESR using DMPO as spin trapping agent (Shi et al. 2003). Therefore, 1 or 3 h after addition of DMPO, radical formation was measured in the cell supernatant using the MiniScope MS100 Spectrometer (Magnettech, Berlin, Germany) with the following instrumental setting: room temperature, microwave frequency = 9.39 GHz, magnetic field = 3,360 G, sweep width = 100G, scan time = 30 s, number of scans = 3, modulation amplitude = 1.8 G, receiver gain = 1,000. For quantification, the amplitudes of all four measured ESR signals were determined and summarized to total amplitude. The mean of total amplitude obtained from five independently measured spectra was analyzed and expressed as mean (ESR-signal intensity) in arbitrary units.

Cytospin preparation

In order to prepare cytospin slides, cells were scraped, centrifuged (900 g, 5 min, 4°C), washed four times with ice cold PBS and resuspended in PBS containing 2% BSA to increase the cell attachment in a homogenous distribution on the glass slide. For preparation of the cytospin slides 1×106 cells were used (5 min, 300 rpm, Thermo shandon, Dreieich, Germany).

Investigation of particle uptake using light and fluorescence microscopy

In a second independent method, particle uptake was determined by counting the number of cells with incorporated particles using May-Gruenwald-Giemsa (MGG)-stained cytospins (Merck, Darmstadt, Germany). Samples were microscopically examined at visual and ultraviolet (555 nm) light on an Olympus BX60 fluorescence microscope (Hamburg, Germany, original magnification ×1,000). Light microscopy was used to determine the percentage of phagocytic positive cells of 100 cells/cytospin of three independent experiments.

Immunohistochemistry (IHC) of actin

Cells on dried cytospin preparations were fixed (3.7% paraformaldehyde, PFA/ PBS; pH 7.4, 10 min at room temperature) and permeabilised (0.01% Triton X-100/PBS, 3 min at room temperature). After blocking in 1% BSA/PBS for 15 min at room temperature, fixed cells were incubated with AlexaFluor488®-coupled phalloidin (0.16 μM, 5units/ml, Molecular Probes, Leiden, Netherlands; Waterman-Storer et al. 2000) overnight at 4°C in order to stain the F-actin-component of the cytoskeleton. After washing with PBS, cytospins were covered with Immuno-Fluore Mounting Medium (MP Biomedicals, Aurora, USA) and analyzed with a fluorescence microscope (Olympus BX60, Hamburg, Germany).

Transmission electron microscopy (TEM)

In order to analyze the intracellular distribution of ingested quartz particles, TEM images were examined. Therefore, monolayer cultures grown in 35 mm culture dishes to 85% confluence were treated with 10 μg/cm2 DQ12 for 24 h. After incubation, monolayers were washed three times and fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer (1 h, 4°C). Post fixation (2% OsO4/0.1 M sodium cacodylate buffer, 1 h, 4°C), cells were en bloc stained with 1.5% uranylacetate dihydrate and phosphotungstic acid, dehydrated in an ethanol series and embedded in epoxy resin (Epon, Serva, Heidelberg, Germany). Blocks of monolayer samples were cut in ultra thin sections (50 nm, Ultracut E, Leica, Bensheim, Germany), placed on 150 mesh grids and stained with uranylacetate and lead citrate (Knaapen et al. 2002) before examination by TEM (STEM CM12, Philips, Eindhoven, Netherlands).

TNF-α Elisa

After 24 h of treatment, cell-free supernatants of NR8383 cells were collected and frozen at −80°C until use. All samples were prepared in triplicate, four times. Supernatants were analyzed using a commercial TNF-α kit (R&D Systems, Wiesbaden, Germany) according to the manufacturer’s manual. Calculated data are presented as mean ± SEM (standard error of the mean) normalized to the control.

Statistical analyses

The median of cell granularity or cell size were chosen as outcome variable from the FACS measurements. Data are presented as percentage of control and mean with standard error (±SEM) are given. Analysis of covariance was used for the statistical analysis of the flow cytometric experiments. The median of FSC was used as a covariate in the statistical analyses of SSC, in order to adjust for the influence of the cell size on cellular functionality (Haugen et al. 1999). The data obtained from ESR experiments were presented as mean of ESR-signal intensity ± SEM. DHR results are expressed as mean ± SEM of the median of fluorescence. Comparison between groups was done using ANOVA post hoc testing (LSD) unless mentioned otherwise (SPSS, version 10) and judged significantly with an error probability P < 0.05.

Results

DQ12 uptake in NR8383 cells in comparison to primary macrophages

Quartz-treatment of NR8383 macrophages for 1, 2, 4 and 24 h resulted in a dose- as well as a time-dependent DQ12 uptake as determined by flow cytometry (Fig. 1). Independent analyses of MGG-stained cytospin preparations using fluorescence and light microscopy showed 4 h after DQ12 treatment an increase of brightness with increasing quartz dose (Fig. 2 a–d) due to the auto-fluorescence of DQ12 at an excitation of 555 nm. Visual light images from these cytospin slides were used to quantify cells with DQ12 uptake (Fig. 2 e–h) showing a similar dose-dependency as observed by the flow cytometric approach (Fig. 2i).

Fig. 1
figure 1

DQ12 uptake in NR8383 cells treated with 10, 20 or 40 μg/cm2 quartz for 1, 2, 4 and 24 h measured by flow cytometry as increase of the median of side scatter angle (SSC). Data are presented as percentage of control (mean ± SEM, n = 3). Significant time- (# P < 0.05, ## P < 0.01, ### P < 0.001) and dose-dependent (**P < 0.01, ***P < 0.001) increase in granularity was determined by analysis of covariance

Full size image
Fig. 2
figure 2

Representative fluorescence images (a–d) and visible light images (e–h) of MGG-stained cytospin preparations of control (a, e) and quartz treated (10 μg/cm2, B, F; 20 μg/cm2, C, G; 40 μg/cm2, D, H) NR8383 cells for analysis of quartz uptake after 4 h (n = 3, original magnification ×1,000). Fluorescence images demonstrate an enhancement of brightness with increasing DQ12 dose. i Quantification of phagocytic positive cells resulted in a significant (*P < 0.05, **P < 0.01; ANOVA, LSD) uptake of DQ12 in NR8383 cells treated with 10, 20 or 40 μg/cm2 quartz for 4 h. Data are presented as percentage of DQ12-positive cells of 100 total cells (mean ± SEM, n = 3)

Full size image

Additionally, the internalization of quartz particles was shown in ultra-thin sections of DQ12-treated NR8383 cells (24 h, 10 μg/cm2) analyzed by TEM (Fig. 3). Electron microscopic images demonstrate a cytosolic location (a, b) of the quartz particles (white arrows) surrounded by a membranous structure (phagosome, black arrowheads, c).

Fig. 3
figure 3

Transmission electron microscopic images of NR8383 cells treated with 10 μg/cm2 DQ12 for 24 h (a original magnification ×25,000; b, c original magnification ×53,000). Internalized particles (indicated by white arrows) are localized within membranous structures (black arrowheads)

Full size image

In order to examine whether the NR8383 cell line governs a good in vitro model, quartz uptake was comparatively analyzed in primary AM and IM using flow cytometry. After a 4-h treatment with 10, 20 or 40 μg/cm2 DQ12, all types of macrophages (NR8383, AM, IM) showed a significant dose-dependent uptake rate (Fig. 4). Although no significant differences in uptake were observed between the three macrophage populations, the uptake rate of IM tended to be lower compared to AM and NR8383 cells. To evaluate whether cell size and consequential lower cell surface-particle-interaction area might be responsible for the observed lower uptake rate, cell sizes of NR8383, AM as well as IM were examined by flow cytometry and microscopy. No differences neither in the microscopic images of MGG-stained cytospin preparations (5a–c) nor in the FSC values (5d) between IM and AM were observed, while NR8383 cells are about 40% larger than primary macrophages (Fig. 5a, d).

Fig. 4
figure 4

DQ12 uptake in the NR8383 cell line, alveolar macrophages (AM) and interstitial macrophages (IM) treated with 10, 20 or 40 μg/cm2 quartz for 4 h measured by flow cytometry as increase of median of side scatter angle (SSC). Uptake is presented as percentage of control (mean ± SEM, n = 3). Significant dose-dependent increase (*P < 0.05, **P < 0.01, ***P < 0.001) in granularity was determined by analysis of covariance

Full size image
Fig. 5
figure 5

Representative images (original magnification ×1,000, bar ≡ 200 μm) of MGG-stained cytospin preparations of NR8383 cells (a), AM (b) and IM (c) verifying the size distribution obtained by flow cytometry (d). Median of forward scatter angle (FSC, mean ± SEM, n = 3) measured by flow cytometry of NR8383 cells, AM and IM as an indicator of the cell size. Significant (**P < 0.01, ***P < 0.001) size differences were observed between NR8383 cells and primary macrophages but not between AM and IM

Full size image

Involvement of actin in the phagocytosis of quartz particles

The relevance of actin polymerisation for the quartz uptake was examined using Cyt-D pretreatment at low (10 μg/cm2) as well as higher (40 μg/cm2) DQ12 dose to enclose possible dose-dependent differences (Xiao et al. 2003). Results of the flow cytometric measurement are summarized in Fig. 6. The particle uptake could be significantly reduced by inhibition of actin polymerization at 10 as well as at 40 μg/cm2 after 1 h (b) and 4 h (c) treatment. Although inhibition of the actin cytoskeleton leads to a potent reduction of uptake, the internalisation could not be completely blocked at any concentration or time point. DMSO, the Cyt-D solvent, was found not to influence particle uptake at both concentrations after 4 h DQ12 exposure and at the low concentration after 1 h treatment. However, a transient yet significant influence of DMSO on DQ12 uptake could be observed following 1 h at the high dose exposure.

Fig. 6
figure 6

a Representative univariant gated histogram of SSC of DQ12 treated NR8383 without (upper panel) or after inhibition of actin polymerisation with Cytochalasin-D (Cyt-D, lower panel). Uptake is detectable as an increase of the number of cells with higher granularity (side scatter angle, SSC) of controls (0 μg/cm2) compared to 10 and 40 μg/cm2 quartz treatment (1 h). After actin inhibition, the particle-induced increase in cell granularity is clearly reduced. b, cDQ12 uptake in NR8383 cells after treatment with 10 or 40 μg/cm2 quartz for 1 h (b) or 4 h (c) measured by flow cytometry as increase of the median of side scatter angle (SSC) with or without pre-incubation with Cytochalasin-D (Cyt-D) or DMSO. Uptake is presented as percentage of control (mean ± SEM, n = 5). Pre-treatment with Cyt-D caused significant inhibition (*P < 0.05, **P < 0.01, ***P < 0.001, analysis of covariance) of cell granularity

Full size image

Immunocytochemical investigation supports the participation of actin in the phagocytic process of quartz particles (Fig. 7). Particle binding and ingestion is demonstrated in the phase contrast images of control (a) versus the treated cells (c, 40 μg/cm2 DQ12, 15 min). Fluorescence images of the same cells (e, control; g, DQ12) show an increased intensity as well as morphological changes of the F-actin framework during particle internalisation. The uniform distribution of F-actin over the cell (e, control) changes after DQ12 treatment, as shown by condensation in the extending pseudopodia and around the particle binding sites. In comparison, phase contrast images obtained from the Cyt-D pretreated cells (d) do not show particle internalisation. Both fluorescence images of Cyt-D pretreated cells (f, without; h with DQ12) show disruption and clustering of the actin cytoskeleton. In addition, DMSO (vehicle control) pretreatment had no effect on particle uptake or actin structure (images not presented).

Fig. 7
figure 7

Representative phase contrast (upper panel) and fluorescence (lower panel) images (n = 4, original magnification ×1,000) of 488-phalloidin-stained cytospin preparations were shown of non-particle treated (a, b, e, f) and DQ12 treated (40 μg/cm2, 15 min; c, d, g, h) NR8383 cells. Cells without cyt-D preincubation demonstrate the morphological changes of actin cytoskeleton during the phagocytosis process of DQ12 (c, g) compared to control cells (a, e). In contrast, pre-incubation with Cyt-D (b, d, f, h) resulted in a complete clustering of actin

Full size image

The involvement of actin in the TNF-α release

In order to investigate the impact of particle uptake and related actin cytoskeleton recruitment on the release of inflammatory mediators, TNF-α release of NR8383 cells was measured. Long term (up to 24 h) pretreatment with Cyt-D or particles did not show cytotoxic effects measured by the MTT assay (data not shown). Treatment with DQ12 resulted in a dose-dependent TNF-α increase in the cell supernatant (Fig. 8) which becomes significant at a dose of 20 μg/cm2. Pretreatment with Cyt-D significantly blocks TNF-α release at both doses indicating that uptake or related actin cytoskeleton recruitment are necessary for the release of the cytokine TNF-α.

Fig. 8
figure 8

TNF-α release expressed as percentage of control (mean ± SEM, n = 4) of cells without or after pre-treated with Cyt-D or DMSO followed by quartz exposure (10, 20 or 40 μg/cm2, 24 h). Dose-dependent increases of TNF-α in the cell supernatant after DQ12 treatment (#P < 0.05, ##P < 0.01, ###P < 0.001, ANOVA, LSD) is significantly impaired by actin inhibition using Cyt-D (*P < 0.05, **P < 0.01, ANOVA, LSD)

Full size image

The involvement of actin in the generation of oxidative stress

In order to investigate the relationship between particle internalisation, related actin recruitment and the generation of extra- as well as intracellular oxidative stress response, the ROS generation of macrophages was analyzed using two different methods.

The extracellular ROS generation was measured 1 and 3 h (Fig. 9a, b) after DQ12 exposure by ESR using DMPO as a spin trap. Analysis of the ESR-signal intensity showed at both time points, a DQ12-related increase in extracellular ROS generation reaching significance at the high dose (40 μg/cm2) independent of vehicle control (DMSO). Cyt-D pretreatment resulted in a complete inhibition of the oxidative response at both doses and time points. As such these data were indicative of the necessity of actin in both quartz particle uptake and DQ12-induced ROS generation.

Fig. 9
figure 9

a, b Extracellular generation of reactive oxygen species (ROS) measured in cells treated with 10 or 40 μg/cm2 quartz for 1 h (a) or 3 h (b) by ESR using DMPO as a spintrap. Data are presented as mean ± SEM (n = 5). DQ12-treatment (40 μg/cm2) significantly increases the radical concentration in cell supernatant independent of DMSO pretreatment (##P < 0.01, ###P < 0.001; ANOVA, LSD). After actin inhibition using Cyt-D, quartz treatment did not result in an increased ROS generation. c Correlation between particle uptake and intracellular ROS generation measured as increase of granularity and Dihydrorhodamine (DHR) oxidation by flow cytometry, respectively. Treatment of NR8383 cells with DQ12 (10 μg/cm2, black; 40 μg/cm2, grey) without (square) or with pre-treatment with DMSO (circle) or Cyt-D (triangle) shows a linear correlation between uptake and ROS generation [regression coefficients: 0.985 (10 μg/cm2) and 0.996 (40 μg/cm2)]. Data are presented as mean ± SEM (n = 4)

Full size image

One advantage of the flow cytometric approach consists in the simultaneous measurement of different parameters. Therefore, intracellular ROS generation combined with particle uptake was analyzed by simultaneous measurement of SSC and oxidized DHR fluorescence (FL1). Auto-fluorescence of DQ12 particles has no effect on the Rhodamine fluorescence and in turn DHR did not influence the SSC signal (data not shown). Analysis of DHR fluorescence, 1 h after DQ12 treatment showed a dose-dependent increase in intracellular ROS generation and this could be blocked by actin inhibition. A strong correlation was observed between quartz uptake and ROS generation with regression coefficients of 0.985 and 0.996 for quartz concentrations of 10 and 40 μg/cm2, respectively (Fig. 9c). Taken together with the results as obtained using ESR measurements, these flow cytometry data show that actin inhibition results in a reduction of particle uptake as well as a complete inhibition of ROS generation, indicative of the involvement of actin in particle uptake as well as ROS generation.

Discussion

To investigate the mechanisms and biological consequences of uptake of quartz in alveolar macrophages, we have analyzed particle uptake using flow cytometry by measurement of the increase in cellular granularity. The necessity of actin recruitment for the process of quartz particle uptake was demonstrated, thereby identifying their uptake in the rat alveolar macrophage cell line NR8383 as classical phagocytosis. Furthermore, we showed that particle uptake or its related cytoskeleton recruitment is involved in the intra- as well as extracellular ROS generation and early cytokine release. Specifically, a causal relationship was observed between actin recruitment inhibition and the subsequent decline in particle uptake and the reduction in ROS generation as well as the release of TNF-α.

In a first approach, a dose- and time-dependent uptake of quartz particles in NR8383 cells could be demonstrated by flow cytometry. Furthermore, uptake of quartz particles by macrophages could be identified as a fast and continuing process as indicated by a two-fold increase in granularity already after 1 h reaching no saturation within the 24 h under our experimental conditions.

In a second approach, cells with internalized quartz particles were determined by light and fluorescence microscopy. Thereby, a good correlation was found between the uptake results measured by flow cytometry and microscopy. These data indicate the flow cytometric measurement as a valid method to determine particle uptake, which is in line with the study of Stringer et al. (1995) investigating the uptake of alpha quartz, diesel exhaust and titanium dioxide in hamster AM.

Also a good correlation was found between the uptake rate in NR8383 cells and primary BAL-AM. Both cell types showed after 4 h of quartz particle treatment a similar increase in the granularity, and respectively the internalization rate. These results govern the cell line as a suitable model to investigate the mechanism of quartz uptake.

Compared to NR8383 cells and AM, rat IM showed a lower quartz uptake which is in line with the data of Zetterberg et al. (1998) describing a lower number of attached and ingested particles per IM compared to AM. The authors discuss this effect as a consequence of a reduced cell surface area of the smaller IM. However, our results demonstrate that although IM and AM have similar cell sizes, they differ in their uptake behaviour (IM < AM). Furthermore, while NR8383 cells and AM significantly differ in their cell size, they were found to have similar uptake rates. Therefore, other reasonable parameters might be responsible for the lower uptake in IM, e.g. variations in differentiation status or different functionalities. In this context, a lower phagocytosis capacity of yeast (S. cerevisiae) has been described for IM compared to AM (Franke-Ullmann et al. 1996). The authors also demonstrated after LPS exposure, a higher TNF-α release by AM than by IM, whereas IM produced more IL-6 and IL-1, indicative of functional differences. This might be due to the localisation of AM within the alveolar spaces of the lung resulting in an exposure to high oxygen concentrations and pollutants. Their direct reaction to foreign invaders requires a high phagocytic capacity and the release of first defence-cytokines, like TNF-α. In contrast, IM are located in close contact with other cells and their function is, in addition to the first immune response, an immune regulation as evidenced by their high release of immune regulatory cytokines like IL-6 and IL-2 (Brain 1992; Iles and Forman 2002).

In order to investigate the mechanism of quartz uptake, we demonstrated the involvement of actin in this process by reduction of quartz internalisation using Cyt-D. Thus, the quartz uptake could be described as classical phagocytosis. However, the increase in granularity was not completely reduced, which might be due to particle adherence at the cell surface, a process which is not blocked by Cyt-D also described by Stringer et al. (1996). Additionally, TEM analysis also provides evidence for phagocytosis as quartz particle internalization mechanism, since DQ12 particles were intracellularly found in membranous structures (i.e. phagosomes). The influence of the Cyt-D vehicle, DMSO, on the DQ12 uptake at the high dose after 1 h of treatment might be related to physico-chemical changes of the particle surface, i.e. DMSO adsorption onto the surface or modification of the surface radical structure resulting in changes of its hydrophobic/hydrophilic characteristics. Since DMSO has no influence on the uptake after 4 h of treatment, these transient modifications might be abolished due to interaction with the medium. However, the DMSO effect was less pronounced compared to the inhibitory effects by Cyt-D.

In conformance to the uptake data following actin inhibition obtained by flow cytometry as well as imaging of particle internalisation by TEM, the phagocytic process could also be shown by immunohistochemistry in the form of morphological changes of the cytoskeleton. Fluorescence images of actin-stained cytospin preparations showed F-actin condensation at the particle binding side as well as actin recruitment to the pseudopodia. This actin-based engulfment of particles and the formation of actin foci building up the phagosome have been described as the uptake mechanism of classical phagocytosis (Allen and Aderem 1996; Kwitatkowska and Sobota 1999). The increase in fluorescence intensity after treatment with DQ12 is in line with the data published by Brown et al. (2004) showing increased F-actin fluorescence in J774 cells after treatment with ambient particulate matter (i.e. PM10). This involvement of actin in the uptake of mineral particles has been discussed already for a considerable time. Churg (1996) refers to differences in the uptake rate of large and small particles as a result of different uptake mechanisms assuming uptake via endocytosis in coated vesicles for small particles and phagocytosis for larger particles. In more detail, Kobzik (1995) demonstrated the involvement of the scavenger receptor in the uptake of silica particles, already indicative of classical phagocytosis. Our data confirm this hypothesis using three independent methods (FACS, TEM, IHC), by demonstrating uptake of quartz particles by macrophages as an actin-mediated and membrane-bound process and therefore as classical phagocytosis.

Uptake rate, and even more, uptake mechanism are discussed to have a strong impact on the generation of ROS and the release of cytokines by AM. Zirconia or alumina treatment of blood-derived monocytes has been demonstrated to result in a reduced phagocytosis capacity of fluorescent latex beads and associated decrease in oxidative burst (Nkamgueu et al. 2000). Using various different particles, including quartz, residual oil fly ashes (ROFA), and TiO2, Imrich et al. (1999) showed a correlation between the number of internalized particles and ROS measured as enhanced DHR and hydro-ethidine fluorescence. In our current study, we investigated the relevance of uptake for ROS generation and TNF-α release in macrophages with a special focus on the involvement of actin. Flow cytometric analysis of intracellular ROS generation as well as measurement of extracellular radical generation by ESR showed conformance with results. Both methods demonstrated following quartz treatment a dose-dependent generation of ROS by macrophages in relation to the polymerisation of the cytoskeleton. Moreover, simultaneous measurement of uptake and DHR fluorescence showed a good correlation between increasing uptake rate and intracellular oxidative stress response, which assumes the necessity of either uptake itself or an intact actin framework for ROS production. Formation of ROS, specifically in the form of superoxide anion by phagocytes is a complex process, which is up to now, not completely understood. A variety of studies indicate the involvement of phagocytic receptors in ROS generation. For instance, neutrophilic oxidative burst following phagocytosis of complement opsonised yeast could be completely reduced by Cyt-B treatment (Granfeldt and Dahlgren 2001). In leukocytes, a network of diverse receptors (CR3, FcγIIR, FcγIIIR) is described which interact with the cytoskeleton resulting in the activation of NADPH oxidase and O −2 -production (Babior 1999). Our results demonstrating the involvement of the actin cytoskeleton in ROS generation are in line with the literature describing an influence of the actin cytoskeleton on the oxidative burst (Goldschmidt-Clermont and Moldovan 1999). An involvement of actin in ROS generation is also shown in endothelial cells by mechanotransduction (Ali et al. 2004).

Since ROS generation is also discussed to be involved in downstream events like the release of proinflammatory cytokines, exemplarily, TNF-α release was examined under consideration of the involvement of the cytoskeleton. In our study, DQ12 was found to cause a dose-dependent TNF-α release that could be completely blocked by inhibition of the cytoskeleton with Cyt-D. This indicates that the actin cytoskeleton may also be involved in the release of pro-inflammatory cytokines such as TNF-α. Actin requirement for TNF-α release is already shown in LPS-treated adherent human monocytes, whereas in the non-adherent fraction Cyt-D did not affect the secretion of TNF-α or IL-8 (Rosengart et al. 2002). On the other hand, TNF-α is also described to induce the actin cytoskeleton reorganization (Wojciak-Stothhard et al. 1998).

In conclusion, our study demonstrates the involvement of the actin framework in the phagocytic process of quartz particles. Our studies also suggest that actin may mediate DQ12-induced generation of ROS and TNF-α release. However, further mechanistic studies are required to demonstrate the active involvement of the cytoskeleton in phagocytic signalling pathways leading to these proinflammatory responses. Therefore, our current investigations focus on the involvement of phagocytotic receptors and their downstream signalling cascades which might be important events leading to particle-induced inflammation.