Experimental Lung Research 28:641± 670, 2002 Copyright # 2002 Taylor & Francis 0190-2148/02 $12.00 + .00 DOI: 10.1080=01902140290108337
TIME COURSE OF QUARTZ AND TiO2 PARTICLE–INDUCED PULMONARY INFLAMMATION AND NEUTROPHIL APOPTOTIC RESPONSES IN RATS
DuPont Haskell Donna D. Zhang, Mark A. Hartsky, and David B. Warheit Laboratory for Health and Environmental Sciences, Newark, Delaware, USA
Apoptosis, or programmed cell death, has been reported to play an important role in the resolution of pulmonar y in¯ ammation. This study was undertaken to investigate the role of apoptosis in resolving particle-induced lung in¯ ammatory responses in exposed rats, using a dose-respons e=time course experimental design. Groups of rats were exposed via intratracheal instillation to 0, 0.5, 1, 5, 10, or 50 mg=kg body weight of quartz (i.e., crystalline silica) particles or to 0, 0.5, 1, 5, 10, 20, or 50 mg=kg of pigment-grade titanium dioxide (TiO2) particles and evaluated for lung in¯ ammation parameters and evidence of apoptosis of in¯ ammator y cells at 24, 48, 72, or 168 hours post exposure. At each post exposure evaluation period, bronchoalveola r lavage (BAL)-recovered cells from control and particle-exposed rats were assesse d for apoptosis using 4 different techniques. The results in silica-exposed rats demonstrated a signi® cant dose-relate d increase in in¯ ammation concomitant with apoptosis of pulmonary in¯ ammator y cells at 24 to 48 hours post exposure. At later postexposure time points, both the silica-induced in¯ ammatory responses and apoptotic levels of in¯ ammatory cells at higher doses (i.e., ¶5 mg=kg) were reduced but persisted through 1 week. TUNEL (TdTmediated dUTP nick end-labeling ) assay studies con® rmed that the vast majority of apoptotic cells were neutrophils. In contrast, titanium dioxide particle exposures produced transient pulmonar y in¯ ammation but only small measurabl e and nonsigni® cant apoptotic response s at higher exposure concentrations. These results suggest that the sustained lung in¯ ammatory response in rats exposed to ¶5 mg=kg silica may be related to the ineffectiveness of the normal apoptotic mechanisms associated with resolution of in¯ ammation. However, because quartz particles are known to be cytotoxic to alveolar macrophages and other lung cells, normal apoptotic mechanisms may have limited utility for resolving particle-induced in¯ ammation, particularly because silica may not be representativ e of other particle-types. Alternatively, it seems unlikely that apoptosis served to promote silica-induced lung in¯ ammatory responses because the initial increase of apoptosis in in¯ ammator y cells was subsequently correlated with a reduction of the pulmonar y in¯ ammator y response in silica-exposed rats. The ® ndings from this in vivo study demonstrate that the neutrophil, and not the alveolar macrophage, is the primar y in¯ ammator y cell-type that undergoes apoptosis in response to particles. Received 11 March 2002; accepted 17 May 2002. This study was supported by Dupont Titanium Technologies and the Advanced Fiber Systems SBUs of the Dupont Company. The authors wish to thank Dr. John W. Green for technical assistance with the statistical analyses of the data and Dr. Leonard Davis for contributing helpful scienti® c insights throughout this study. In addition, Ken Reed and Denise Janney are acknowledged for their expert technical assistance in these studies. Address correspondence to David B. Warheit, DuPont Haskell Laboratory, 1090 Elkton Road, Newark, DE 19714, USA. E-mail:
[email protected] m
641
642
D. D. Zhang et al. Furthermore , at doses causing similar degrees of in¯ ammation at 24 hours post exposure, the magnitude of apoptosis induced by silica is signi® cantly larger than that induced by TiO2, indicating that there are potency differences in lung in¯ ammation as well as apoptotic responses among different particle-types. Keywords apoptosis, lung in¯ ammation, neutrophil apoptosis, quartz particles, resolution of in¯ ammation, silica particles, titanium dioxide particles
Apoptosis, or programmed cell death, plays a key role in the steady-state regulation of tissues. Under normal conditions, nearly all cells undergo a controlled physiological demise via apoptosis [1]. Inappropriate or insuf® cient apoptotic responses are also considered to be involved in the pathogenesis of various diseases, including cancer [1, 2]. It is noteworthy that apoptosis has also been reported to play an important role in the resolution of pulmonary in¯ ammation [1, 2, 4]. In this scenario, following infection-related recruitment of in¯ ammatory cells into the lung, apoptotic leukocytes, which express speci® c signals in their plasma membranes, are recognized and phagocytized by macrophages, a process bene® cial to resolution of the in¯ ammatory response [1, 5, 6]. However, in the absence of an appropriate apoptotic response, or when the recognition and clearance mechanisms of macrophages are overwhelmed, recruited neutrophils may become necrotic, releasing toxic cellular contents, such as reactive oxygen species and proteases. As evidenced in chronic in¯ ammatory events, these toxic products liberated by recruited leukocytes are associated with tissue damage and persistent in¯ ammation, which ultimately lead to the development of pulmonary ® brosis or tumors [7, 8]. The role, if any, of apoptosis in particle-induced pulmonary in¯ ammation remains to be determined. On the one hand, a number of investigators have suggested that neutrophil apoptosis plays a pivotal role in the resolution and control of lung in¯ ammation and provides for lung tissue to return to normal status following infection [1, 3, 9± 12]. In contrast to this concept, Borges and colleagues [13] have suggested that activation of Fas ligand, a critical component of apoptosis, plays a central proin¯ ammatory role in the induction of pulmonary silicosis in mice; and Kitamura and coworkers [14] have concluded that endotoxin-induced lung injury in mice was associated with up-regulation of Fas in alveolar and in¯ ammatory cells. Because pulmonary in¯ ammatory cells are known to play a central role in directing particle-induced pulmonary in¯ ammation and ® brosis, it would be important to investigate the role of apoptosis in pulmonary in¯ ammatory cells following in vivo exposures to ® brogenic and non® brogenic dusts. In this regard, little is known about the occurrence of apoptosis in pulmonary in¯ ammatory cells following in vivo exposures to crystalline silica particles or pigment-grade titanium dioxide (TiO2) particles. Accordingly, we have postulated that, at high doses, TiO2 particle exposure results in pulmonary
Lung In¯ ammation and Apoptosis
643
in¯ ammation and its resolution is associated with apoptosis of in¯ ammatory cells and phagocytosis by macrophages. In contrast, we hypothesize that quartz particle exposure, which is known to produce a more persistent lung in¯ ammatory response in rats and humans, fosters an insuf® cient apoptotic response in pulmonary in¯ ammatory cells, and thus results in poor lung clearance of instilled particles and, as a consequence, the development of sustained, unresolved, chronic in¯ ammation. The current study was undertaken to evaluate the role of neutrophil apoptosis in either promoting or resolving particle-induced pulmonary in¯ ammation in exposed rats, using a dose-response=time course experimental design. MATERIALS AND METHODS General Experimental Design The experimental design for these studies was approved by our AAALAC-approved Haskel Animal Welfare Committee. The rats used in this study were pathogen-free male Crl:CD BR rats (approximately 250 g) obtained from Charles River Breeding Laboratory (Kingston, NY). All rats were quarantined for at least 1 week prior to exposures. Groups of rats were used to assess pulmonary in¯ ammation and apoptosis after intratracheal instillation of quartz (i.e., crystalline silica) or TiO2 particles. In the silica particle group, rats were exposed to 0 (saline), 0.5, 1, 5, 10, or 50 mg silica=kg body weight. The 10- and 50-mg=kg doses were considered to be ``overload’’ concentrations of cytotoxic particles and thus the focus for interpretation of results was on the 0.5, 1, and 5 mg=kg doses. In the TiO2 particle group, rats were exposed to 0 (saline), 0.5, 1, 5, 10, 20, or 50 mg TiO2 particles=kg body weight. Similarly, the 10-, 20-, and 50-mg=kg doses were considered to be ``overload’’ concentrations of low toxicity particles and the focus for interpretation of results was on the 0.5, 1, 5, and 10 mg=kg doses (note that the 10-mg=kg group was included herein because of the low in¯ ammogenic potency of TiO2 particulates). Postexposure evaluation periods were 24 hours (1 day), 48 hours (2 days), 72 hours (3 days), and 168 hours (7 days). Twelve experiments (5 TiO2 and 7 silica) were conducted and 1 to 2 rats=dose=time point was utilized for each experiment. Particles and Bronchoalveolar Lavage Quartz particles (henceforth referred to as crystalline silica particles) (size range ˆ 1± 4 mm) in the form of Berkeley Min-U-Sil 5 were purchased from Pennsylvania Glass and Sand Corp. (Pittsburgh, PA). Pigment-grade TiO2 particles (mean particle diameter ˆ 0.25 mm; rutile type) were obtained from the DuPont Co. (Wilmington, DE). All particles were autoclaved for
644
D. D. Zhang et al.
sterilization and the elimination of endotoxin contamination. Following particle exposure, rats were euthanized by intraperitoneal injections of sodium pentobarbital. Bronchoalveolar lavage (BAL) was performed as described previously [15]. Cell Differential Analyses Lavage ¯ uids were centrifuged and resuspended in phosphate-buffered saline (PBS); an aliquot was cytocentrifuged at 1500 rpm for 3 minutes on glass slides using a Shandon cytospin3 cytocentrifuge (Shandon, PA). The slides were then stained with Diff-Quik (Dade-Behring, Aguada, PR). For cell differential analyses, a minimum of 500 cells were counted under oil immersion (£1000). Two slides were made and counted for each sample. Cell Death Detection ELISA Cytosolic histone-associated DNA fragments in particle- and salineexposed BAL cells were detected using an enzyme-linked immunosorbent assay (ELISA) kit (Boehringer Mannheim, Germany) with some modi® cations. Lavaged ¯ uids were centrifuged and resuspended in PBS. Subsequently, cells were counted with a hemocytometer. An aliquot (equivalent to 7.5£105 cells) was transferred to an Eppendorf tube and centrifuged. The pellets were resuspended in 0.75 mL of lysis buffer. After incubation for 30 minutes at room temperature, the cell lysate was diluted 20 times (5£104 cells =mL) and centrifuged at 200£g for 10 minutes. Twenty microliters of the supernatant (cytoplasmic fraction, corresponding to a cell equivalent of 1£103 cells) was transferred into the streptavidin coated MTP (microtiter plate) for further analysis according to the manufacturer’s protocol. Duplicates of each sample were then analyzed. Apoptotic DNA Ladder BAL cells from rats exposed to varying concentrations of TiO2 and silica particles were recovered at 1 day post exposure. Genomic DNA of BAL cells was extracted using an apoptotic DNA ladder kit (Sigma, St. Louis, MO) modi® ed by adding an additional step, that is, RNase treatment and phenol= chloroform extraction; 2 mg of DNA per lane was electrophoresed in 2.0% agarose gels containing 0.5 mg=mL ethidium bromide. TUNEL (TdT-Mediated dUTP Nick End-Labeling) Assay Cytocentrifuge preparation of BAL cells were made as described above. DNA strand breaks in individual apoptotic cells were ¯ uorescein-labeled by an in situ cell death detection kit (Boehringer Mannheim). Fluorescein was
Lung In¯ ammation and Apoptosis
645
then converted to a color-based stain using an alkaline phosphatase± conjugated, anti-¯ uorescein antibody and BM purple AP substrate (Boehringer Mannheim). After color-staining, the slides were analyzed under light microscopy (£100). Several ® elds were scanned and recorded on a computer connected to the microscope. The number of positive- and negative-staining macrophages and neutrophils were counted and recorded. Statistics Experiments were conducted to investigate the role of apoptosis of in¯ ammatory cells in the resolution of particle-induced pulmonary in¯ ammation. Rats were exposed to saline or to 0.5, 1, 5, 10, 20, or 50 mg=kg silica or TiO2 particles and the lungs were evaluated at 1, 2, 3, or 7 days post exposure. These evaluations consisted of lung lavage recoveries of cells from the treated animals and included cell counts, percent neutrophils, and percent apoptosis. Each experiment was conducted with 1 rat per dose group per time course. Included in these analyses are the results from 5 experiments with TiO2 particles and 7 experiments with silica particles. In order to analyze these data, animals with the same exposure and time course were treated as though they came from a single experimental group. Although this may confound among-animal and among-experiment variability, the likely effect is to increase the background variability, so that some small effects may be missed, but any effect found statistically signi® cant by this analysis is truly signi® cant. Data from all time courses and all doses for a given particle-type (i.e., TiO2 or silica) were analyzed together using a 2-way analysis of variance (ANOVA) with interaction. Where the measured response was not normally distributed, a normalizing and variance stabilizing transformation was found. Normality was assessed by the Shapiro-Wilk test and variance homogeneity by Levene’s test. For silica and TiO2 counts, a log transform was used. For percent neutrophils (polymorphonuclear neutrophils [PMNs]) and apoptosis measurements, a log transform was not possible because of the presence of zero response values. A cube-root transform of PMNs for both particle-types was used, whereas for apoptosis, a 5th-root was used for TiO2 and a 6th-root for silica responses. No common-power transform could be found. Comparisons were made at each time point between each dose group and the control, as well as a linear contrast to test for dose-response trend. In addition, for each dose, the means among all time points were compared. The day-by-dose interaction term was statistically signi® cant only for silica cell counts, whereas the main effects of both day and dose were statistically signi® cant for all 6 models. Each contrast was evaluated by an unadjusted
646
D. D. Zhang et al.
2-sided t test due to the signi® cant ANOVA F tests for both day and dose and, in one case, for the day-by-dose interaction. Statistical signi® cance was de® ned to occur at the P < .05 level. RESULTS Pulmonary In ammation and Apoptosis in Silica-Exposed Rats The viabilities of lavage-recovered cells, as measured by Trypan blue exclusion, from all sham- or particle-exposed rats were always greater than 95%. In general, the percentages of neutrophils were used as an index of pulmonary in¯ ammation. Silica-Induced BAL Fluid Cell Counts …Figure 1A† At each time point, there was a signi® cant increasing trend in the doseresponse. On day 1, every treatment mean was signi® cantly greater than the control. On days 2, 3, and 7, all treatment means except the 2 smallest were signi® cantly greater than the control mean. For the 0.5-, 1.0- and 5.0-mg=kg dose groups, the mean on each of days 2, 3, and 7 was signi® cantly smaller than on day 1. There were no statistically signi® cant differences among days 2, 3, and 7. For the 10-mg=kg group, the mean on day 2 was signi® cantly lower than on day 1. Otherwise, there were no statistically signi® cant differences among days. For the 50-mg=kg group, there were no statistically signi® cant differences among days. These results suggest maximum recovery occurred by day 2 for the 3 lowest dose groups, but for the highest 2 dose groups, there had been little recovery even after 7 days.
" FIGURE 1 (A) Bronchoalveolar lavage ¯ uid (BALF) cell numbers recovered from rats 1, 2, 3, and 7 days following silica instillation exposures. Values given are expressed as mean cell numbers (§SD). Groups of rats were intratracheally instilled with saline, 0.5, 1, 5, 10, or 50 mg=kg of crystalline silica particles. At each time point, there was a signi® cant increasing trend in the dose-response. On day 1, every treatment mean was signi® cantly greater than the control. On days 2, 3, and 7, all treatment means, except the two smallest, were signi® cantly greater than the control mean (*:P < .05). (B) Bronchoalveolar lavage ¯ uid (BALF) cell numbers recovered from rats 1, 2, 3, and 7 days following titanium dioxide particle instillation exposures. Values given are expressed as mean cell numbers (§SD). Groups of rats were intratracheally instilled with saline, 0.5, 1, 5, 10, 20, or 50 mg=kg of titanium oxide particles. At each time point except day 3, there was a signi® cant increasing trend in the dose-response. On day 1, all but the 2 lowest-dose groups had signi® cantly larger means than the control. On days 2 and 3, only the 50- and 10mg=kg exposure groups had means signi® cantly larger than the control, with the mean of the 20-mg=kg exposure group marginally larger (*:P ˆ.09) on day 2. On day 7, only the high-dose group mean was signi® cantly higher than the control. These results suggest signi® cant recovery in all dose groups by day 7 and in all but the 2 highest dose groups by day 2 (P < .05).
647
FIGURE 1 Continued.
648
FIGURE 1 Continued.
Lung In¯ ammation and Apoptosis
649
Silica-Induced PMNs …Figure 2A† At each time point, there was a signi® cant increasing trend in the doseresponse. On each day, every treatment mean was signi® cantly greater than the control. For the 0.5-mg=kg dose group, the day 7 mean was signi® cantly smaller than the day 1 mean. Otherwise, there were no signi® cant differences among daily means. For the 1.0-mg=kg dose group, the mean on each of days 2, 3, and 7 was signi® cantly smaller than on day 1. There were no statistically signi® cant differences among days 2, 3, and 7. For the 5.0- and 10.0-mg=kg dose groups, the day 3 mean was signi® cantly smaller than the day 1 mean. Otherwise, there were no signi® cant differences among daily means. For the 50-mg=kg group, there were no statistically signi® cant differences among days. There appear to be recoveries in all dose groups by day 3, with subsequent increase at day 7 in the 3 intermediate doses, but these observations were not statistically signi® cant. Apoptosis in Silica-Exposed Lung Cells …Figure 3A† At each time point, there was a signi® cant increasing trend in the doseresponse. On days 1, 2, and 7, every treatment mean was signi® cantly greater than the control. On day 3, all but the 2 lowest dose group means were signi® cantly greater than the control mean. For the 0.5- and 1.0-mg=kg groups, the day 2 and day 1 means were statistically indistinguishable, as were the day 3 and day 7 means. All other daily means were signi® cantly different, with the means responses on later days smaller than on earlier days. The same is true of the 5.0-mg=kg group, except that the day 7 mean was indistinguishable from days 2 and 3. At 10 mg=kg, the means on days 3 and 7 were signi® cantly smaller than on day 1, and the day 7 mean was signi® cantly smaller than the day 2 mean. The days 3 to 2 difference was mildly signi® cant (P ˆ.08). For the high-dose group, there were no signi® cant differences among the daily means. These results suggest a signi® cant recovery by day 3, which persisted through day 7. " FIGURE 2 (A) Percentages of neutrophils in BAL ¯ uids of rats 1, 2, 3, and 7 days after silica exposures. Values given are expressed as mean percentages of neutrophils (§SD). Groups of rats were intratracheally instilled with saline, 0.5, 1, 5, 10, or 50 mg=kg of crystalline silica. At each time point, there was a signi® cant increasing trend in the dose-response. On each day, every treatment mean was signi® cantly greater than the control. (*:P < .05). (B) Percentages of neutrophils in BAL ¯ uids of rats 1, 2, 3, and 7 days after TiO2 particle exposures. Values given are expressed as mean percentages of neutrophils (§SD). Groups of rats were intratracheally instilled with saline, 0.5, 1, 5, 10, or 50 mg=kg of titanium oxide particles. At each time point, there was a signi® cant increasing trend in the dose-response. On days 1 and 2, all but the 2 lowest dose groups had signi® cantly greater mean response than the control. By day 3, the 5.0-mg=kg group likewise was no greater than the control. There was no further change by day 7. These results suggest signi® cant recovery by day 3 for exposures less than 10 mg=kg (*:P < .05).
650
FIGURE 2 Continued.
651
FIGURE 2 Continued.
652
D. D. Zhang et al.
It should be noted, however, that there was a signi® cant decrease in the control mean on day 7, which casts some doubt on the other conclusions. Pulmonary In ammation and Apoptosis in TiO2-Exposed Rats TiO 2-Induced BAL Fluid Cell Counts …Figure 1B† At each time point except day 3, there was a signi® cant increasing trend in the dose-response. On day 1, all but the 2 lowest dose groups had signi® cantly larger means than the control. On days 2 and 3, only the 50- and 10mg=kg exposure groups had means signi® cantly larger than the control, with the 20-mg=kg mean marginally larger (P ˆ.09) on day 2. On day 7, only the high-dose group mean was signi® cantly higher than the control. These results suggest signi® cant recovery in all dose groups by day 7 and in all but the 2 highest dose groups by day 2. For ® xed dose, the only signi® cant comparisons were at 1 mg=kg (day 7 < day 1); at 5.0 mg=kg (day 7 < day 1); 10 mg=kg (days 2, 3, 7 < day 1); 20 mg=kg (days 3, 7 < day 1). It may be that there was too little damage at the low dose to show a recovery and too much damage at the high dose, whereas at the intermediate doses, there was signi® cant recovery by at least day 7. TiO 2-Induced PMNs …Figure 2B† At each time point, there was a signi® cant increasing trend in the doseresponse. On days 1 and 2, all but the 2 lowest dose groups had signi® cantly greater mean responses than the control. By day 3, the 5.0-mg=kg group likewise was no greater than the control. There was no further change by day 7. These results suggest signi® cant recovery by day 3 for exposures less than 10 mg=kg. For ® xed dose, the signi® cant comparisons were for 0.5 mg=kg (none); 1.0 mg=kg (days 3, 7 < day 1); 5.0 mg=kg (days 3, 7 < days 1, 2); 10 mg=kg (days 2, 3, 7 < day 1; day 7 < day 2, marginally day 3 < day 2, P ˆ.08); 20 mg=kg (day 7 < day 1); 50 mg=kg (none). As with cell counts, it may be that there was too little damage at the low dose to show a recovery and too much damage at the high dose, whereas at the intermediate doses, there was signi® cant recovery by at least day 7, and earlier for lower exposures. " FIGURE 3 (A) Apoptotic values of BAL ¯ uid in¯ ammatory cells in silica-exposed rats. Cytosolic histoneassociated DNA fragments in BAL cells recovered from silica-exposed rats. At each time point, there was a signi® cant increasing trend in the dose-response. On days 1, 2, and 7, every treatment mean was signi® cantly greater than the control. On day 3, all but the 2 lowest dose group means were signi® cantly less than the control mean (*:P < .05). (B) Apoptotic values of BAL ¯ uid in¯ ammatory cells in TiO2 exposed rats. Cytosolic histone-associated DNA fragments in BAL cells recovered from TiO2 -exposed rats. There were signi® cant increasing dose-response trends only on days 1 and 2 and there was no day on which any dose group mean was signi® cantly different from the control mean.
653
FIGURE 3 Continued.
654
FIGURE 3 Continued.
Lung In¯ ammation and Apoptosis
655
Apoptosis in TiO2-Exposed Lung Cells …Figure 3B† There were signi® cant increasing dose-response trends only on days 1 and 2 and there was no day on which any dose group mean was signi® cantly different from the control mean. For ® xed dose, there were no signi® cant differences among daily means. These results suggest no signi® cant effect on apoptosis resulting from exposure to TiO2 at the levels included in these experiments. A time course comparison of neutrophilic in¯ ammation and apoptosis in silica- or TiO2-exposed rats is presented in Figures 4 and 5, respectively. For silica-exposed rats, the apoptotic trend correlated with the degree of pulmonary in¯ ammation, both during high (i.e., 24 and 48 hours post exposure) and lower levels (at 72 and 168 hours post exposure). However, apoptosis of in¯ ammatory cells was not associated with resolution of the in¯ ammatory response in silica-exposed rats, as evidenced by a sustained pulmonary in¯ ammatory response. For the TiO2-exposed rats, a transient in¯ ammatory response was measured and this was resolved by 3 days post exposure. However, this effect was not correlated with an increase in apoptosis of BAL cells. Apoptotic DNA Ladder To con® rm the results obtained from the cell death ELISA assay, DNA agarose gels were utilized to detect the presence of apoptotic DNA ladders, because internucleosomal DNA fragmentation is a biochemical hallmark of apoptosis. As illustrated in Figure 6A, silica exposure resulted in a DNA ladder formation in BAL cells. A DNA ladder was evident in the 0.5-mg=kg silica-treated sample, and the intensity of the DNA ladders were enhanced at higher silica doses (i.e., 1-, 5-, 10-mg=kg silica-treated samples). In TiO2exposed samples (Figure 6B), DNA ladders were visible only at extremely high doses, that is, 5 or 10 mg=kg. These results are in accordance with and con® rm the data obtained from the cell death ELISA assay (Figure 3A, B), demonstrating a robust apoptotic response in BAL cells from silica-exposed rats and a underwhelming apoptotic response in BAL cells recovered from TiO2-exposed rats. Morphology of BAL Cells In order to identify the primary cell type(s) undergoing apoptosis in BAL cells following in vivo particle exposure, BAL cells were stained with Diff-Quik to identify morphological changes following particle exposure. Diff-Quik staining in silica-exposed (Figure 7A) BAL ¯ uid± recovered cells revealed that some neutrophils demonstrated a condensed morphology
656
FIGURE 4 Time course comparison of silica- versus TiO2 -induced lung in¯ ammatory responses. Time course of neutrophil in¯ ux in BALF of rats instilled with 1 or 5 mg=kg crystalline silica or TiO2 particles. All values given are mean percentages of PMN in¯ ux. This line graph demonstrates that the differences in rat lung responses to a cytotoxic versus low toxicity dust; that is, silica instillation induced a persistent in¯ ammatory response whereas titanium dioxide exposures produced a transient pulmonary response.
657
FIGURE 5 Comparison of silica- versus TiO2-induced lung in¯ ammatory cell apoptotic responses. Time course of apoptotic responses in in¯ ammatory cells recovered from the lungs of rats instilled with 1 or 5 mg=kg crystalline silica or TiO2 particles. All values given are mean values of apoptotic units. For silicaexposed rats, the apoptotic trend correlated with the degree of pulmonary in¯ ammation, both during high (i.e., 24 and 48 hours post exposure) and lower levels (at 72 and 168 hours post exposure). However, apoptosis of in¯ ammatory cells was not associated with resolution of the in¯ ammatory response in silica-exposed rats, as evidenced by a sustained pulmonary in¯ ammatory response. For TiO2 -exposed rats, a transient in¯ ammatory response was measured, which resolved by 3 days post exposure. However, this effect was not correlated with an increase in apoptosis of BAL cells.
658
D. D. Zhang et al.
FIGURE 6 Agarose gel electrophoresis of genomic DNA of BAL cells from rats instilled with silica (A) or TiO2 (B). Lane 1: DNA marker; lanes 3± 7: DNA of BAL cells recovered from rats instilled with saline, 0.5, 1, 5, 10, or 50 mg=kg silica or TiO2 at 1 day post exposure. This is representative of 3 experiments. Continued.
(arrows). However, this characteristic could not be fully attributed to the effects of particles because artifacts in cell morphology may have been associated with the preparation of cells (i.e., cytocentrifugation). In contrast to neutrophils, no abnormal morphology was detected in BAL ¯ uid± recovered alveolar macrophages following either particletype instillation exposure. Figure 7B depicts an alveolar macrophage with 2 phagocytized neutrophils contained within its cytoplasm. TUNEL Assay Instead of using subjective cell morphological characteristics to quantify apoptosis of pulmonary in¯ ammatory cells, a TUNEL assay method was
Lung In¯ ammation and Apoptosis
659
FIGURE 6 Continued.
utilized. As shown in Figure 8A, there were no positive staining cells recovered from saline-exposed rats. However, silica exposure induced positive apoptotic staining in neutrophils, with the percentages of positively stained neutrophils increasing at higher doses (Figure 8B, C, Table 1). On rare occasion, positive-staining macrophages with enclosed apoptotic bodies (i.e., derived from neutrophils) were observed (Figure 8C ). In contrast, TiO2 induced a very low percentage of apoptotic staining in neutrophils at extremely high doses and no staining was observed in macrophages (Figure 8D, E). It was interesting to note that some macrophage surfaces were covered by TiO2 particles in TiO2-exposed BAL cells. However, the presence of adsorbed or internalized TiO2 particles were identi® ed and thus differentiated from a positive apoptotic TUNEL stain based on color criteria. A summary of TUNEL assay results at 24 hours post exposure of BAL cells to silica or TiO2 at equal in¯ ammatory doses is presented in Table 1. The equivalent in¯ ammatory doses are de® ned as doses that induce similar percentages of neutrophils in BAL cells (e.g., 1 mg=kg silica versus 5 mg=kg TiO2, and 5 mg=kg silica versus 10 mg=kg TiO2). Doses of silica or TiO2 particles that induced similar degrees of in¯ ammation (as de® ned by numbers of PMNs) resulted in different apoptotic responses (see Table 1). In addition, the TUNEL assay data provided evidence that the majority of apoptotic cells were neutrophils.
660
D. D. Zhang et al.
FIGURE 7 Diff-Quik staining of BAL cells recovered from rats instilled with 1 mg=kg silica at 1 day post exposure (A). Several neutrophils with more condensed morphology are depicted (arrows). (B) An alveolar macrophage, recovered from a rat instilled with 1 mg=kg silica at 1 day post exposure, is observed containing 2 phagocytized neutrophils (arrows). (Magni® cation ˆ 100£.)
Lung In¯ ammation and Apoptosis
661
FIGURE 8 TUNEL labeling of BAL cells recovered from rats instilled with saline (A), 1 (B), and 5 (C) mg=kg silica, 5 mg=kg TiO2 (D), and 10 mg=kg TiO2 (E) at 1 day post exposure. N ˆ neutrophils; M ˆ alveolar macrophages. Note that the many neutrophils from silica-exposed rats were apoptotic and the lack of TUNEL staining in neutrophils and macrophages recovered from the lungs of TiO2-exposed rats. It was interesting to note that some macrophages were covered by TiO2 particles in TiO2-exposed BAL cells. However, the presence of adsorbed or internalized TiO2 particles were identi® ed and thus differentiate d from a positive apoptotic stain based on color criteria. In addition, a silica-exposed macrophage containing apoptotic neutrophils is observed in C. (Magni® cation ˆ 100£.) Continued.
DISCUSSION This study was designed to assess the time course of in¯ ammation and apoptosis in pulmonary in¯ ammatory cells following in vivo particle exposures to crystalline silica and pigment-grade TiO2. Our results demonstrate that instilled silica particles were potent in producing pulmonary in¯ ammation and apoptosis in BAL-recovered cells. It is interesting to note that although the pulmonary in¯ ammatory response to silica was sustained through 7 days post exposure, the maximum response of apoptotic cells was measured at 24 and 48 hours post exposure, and was reduced but sustained and greater than control levels thereafter. Apoptosis has been reported to play an important role in the resolution of in¯ ammation [1]. Because high-dose silica inhalation exposure produces
662
FIGURE 8 Continued.
D. D. Zhang et al.
Lung In¯ ammation and Apoptosis
FIGURE 8 Continued.
663
664
D. D. Zhang et al.
TABLE 1 Results of the TUNEL Assay in Particle-Exposed BAL Cells
Saline control Si (1 mg=kg) TiO2 (5 mg=kg) Si (5 mg=kg) TiO2 (10 mg=kg)
N
N‡
M
3 42 214 137 72
0 12 2 231 6
38 34 244 59 28
M‡ 0 0 0 2* 0
% of N ‡
% of M ‡
0 22.2 0.9 62.8 7.7
0 0 0 3.4 0
Note. The area of cell stained and the intensity of staining were the two criteria used to score positive cells to distinguish them from background staining. N ˆ Number of negative-staining neutrophils. N ‡ ˆ Number of positive-staining neutrophils. M ˆ Number of negative-staining macrophages. M ‡ ˆ Number of positive-staining macrophages. % of N ‡ ˆ percentage of positive-staining neutrophils. % of M ‡ ˆ percentage of positive-staining macrophages. *Staining of these two cells are in cytoplasmic regions and thus most likely that the staining was derived from phagocytized apoptotic bodies, see Figure 8C.
a chronic lung in¯ ammatory response in rats that is never resolved [15], it is conceivable that silica-induced in¯ ammation and apoptotic effects have overwhelmed the capacity of macrophages to ingest apoptotic neutrophils, and, as a consequence, this could lead to a persistent pulmonary in¯ ammatory response. Alternatively, the apoptotic process may be insuf® cient and thus unable to compensate for the cytotoxic (i.e., necrotic) effects of silica particles on lung cells. It was also clearly demonstrated that although silica instillation produced both an in¯ ammatory and apoptotic response in the lungs of exposed rats, TiO2 particle exposure resulted in transient pulmonary in¯ ammation only at high doses, and this was associated with a weak neutrophil apoptotic response. At lower TiO2 doses (0.5 and 1 mg=kg), there was little in¯ ammation and essentially no detectable apoptosis in BAL cells. Conceivably, the limited number of apoptotic neutrophils were quickly phagocytized and degraded by macrophages, because this is a very rapid and ef® cient process [9]. Although the development of pulmonary ® brosis clearly is a complex process, the ® nding of in¯ ammation and apoptosis in the lung cells exposed to silica, a cytotoxic and ® brogenic dust, and the relatively weak levels of apoptotic effects associated with exposures to TiO2, a low toxicity particle, suggests that the mechanisms related to in¯ ammation control may be speci® c for each particle-type. The role of apoptosis in the resolution of pulmonary in¯ ammation remains controversial. Several investigators have postulated that following an in¯ ux of in¯ ammatory cells into the lung, apoptosis of neutrophils and consequent phagocytosis of and clearance of these in¯ ammatory cells by macrophages represents an important prerequisite step for the resolution and control of in¯ ammation following exposures to bacterial and particulate
Lung In¯ ammation and Apoptosis
665
pulmonary insults. This concept seems logical and is consistent with the development and resolution of an acute and transient in¯ ammatory response. Moreover, a failure to resolve even a minor pulmonary infection would likely result in the development of a chronic in¯ ammatory and consequent ® brotic response. Alternatively, Borges and colleagues [13] have recently reported that apoptotic mechanisms serve to trigger the development of silicosis in exposed mice. In their study, normal mice were intratracheally instilled with a high single dose (i.e., no dose-response assessments were made) of silica particles and subsequently developed severe pulmonary in¯ ammation and ® brosis. In contrast to the wild-type mice, similarly exposed, Fas ligand± de® cient, lymphoproliferative disease mutant (gld) mice did not develop silicosis. The authors concluded that Fas ligand plays an important proin¯ ammatory role in the development of silica-related pulmonary ® brosis [13]. How then can one reconcile the seemingly paradoxical role of apoptosis in regulating both proin¯ ammatory and anti-in¯ ammatory effects (e.g., apoptosis as a modulator of the resolution of in¯ ammation)? Haslett has suggested that the outcome of in¯ ammation can be regarded as a ``battle’’ between mechanisms that would cause injury and amplify in¯ ammation versus those that tend to protect tissues and promote resolution. In this regard, it is unknown why some stimuli such as Streptococcus induce a limited pulmonary in¯ ammatory response that resolves quickly without injury, whereas other agents such as silica or Staphylococcus elicit in¯ ammatory responses that are sustained and promote ® brotic responses [11]. In our experience in comparing the pulmonary effects of ® brogenic (crystalline silica) versus non® brogenic (TiO2) dust exposures in the lungs of rats, we have noted the differences in the degree of injury elicited by the particle-induced in¯ ammatory response. In our TiO2 studies, particle-overload inhalation exposures in rats to 250 mg=m3 for 4 weeks produced a signi® cant neutrophilic lung in¯ ammatory response (¹50% BAL ¯ uid PMNs), with little evidence of pulmonary cellular injury or necrosis (BAL ¯ uid lactate dehydrogenase [LDH] levels ˆ 150% when compared to sham-exposed controls) and no evidence of pulmonary ® brosis at 6 months post exposure [16]. In contrast, 3-day inhalation exposures to 100 mg=m3 crystalline silica particles in rats also produced a sustained neutrophilic in¯ ammation (¹40% to 60% BAL ¯ uid PMNs) concomitant with signi® cant evidence of lung cytotoxicity and necrosis (e.g., BAL ¯ uid LDH levels ˆ 1200% versus sham-exposed control values at 1 month post exposure) and the development of ® brosis at 3 months post 3-day exposure [15]. Despite the similar numbers of in¯ ammatory neutrophils recovered from the lungs of silicaor TiO2-exposed rats, there clearly was a difference in the numbers of necrotic cells and, as a consequence, the intensity of the in¯ ammatory response. Moreover, in another study, silica inhalation exposure produced signi® cant
666
D. D. Zhang et al.
pulmonary toxicity in PMN-depleted rats, indicating that silica particles are also cytotoxic to other lung cell-types [17]. Although neither apoptosis or necrosis endpoints were directly measured in these studies, it seems clear that inhalation of high concentrations of crystalline silica particles results in signi® cant cytotoxicity and likely tips the balance to necrosis versus apoptosis. This could have signi® cant implications for the inability of apoptosis to promote the resolution of pulmonary in¯ ammation in silica-exposed rats. Cox and coworkers [12] have reported that macrophage engulfment of apoptotic neutrophils contributes to the resolution of acute pulmonary in¯ ammation in vivo. The investigators exposed rats via intratracheal instillation to lipopolysaccharide (LPS) and carried out BAL studies at several time periods post exposure. LPS induced apoptotic effects in pulmonary neutrophils and the maximal effect for both neutrophil in¯ ux into the lung and apoptosis was measured at 24 hours post exposure, indicating a transient in¯ ammatory response. This time course for the resolution of in¯ ammation was consistent with the maximal response of macrophages phagocytizing neutrophils. It was concluded that the acute in¯ ammatory response produced by LPS instillation was resolved following apoptosis of neutrophils and phagocytosis of the apoptotic neutrophils by macrophages. In contrast to these results, Kitamura and colleagues reported that LPSinduced lung injury in mice was associated with upregulation of Fas in alveolar and in¯ ammatory cells [14]. Are species differences responsible for these discrepancies or are the experimental designs of these 2 studies suf® ciently different for addressing the question of the role of apoptosis in the regulation of pulmonary in¯ ammation? Although the development of pulmonary ® brosis clearly is a complex process, the ® nding of a pulmonary in¯ ammatory response that was not resolved by 7 days post instillation exposure of silica particles suggests that the clearance process may have been overwhelmed by excessive apoptotic response of neutrophils. This chronic in¯ ammatory response in the rat may contribute to the development of silica-induced pulmonary ® brosis. In support of our results, Hagimoto and colleagues [7, 8] have reported that excessive apoptosis induced by activation of Fas=FasL pathway results in pulmonary in¯ ammation and ® brosis in mice. Clearly, macrophage phagocytosis of apoptotic in¯ ammatory cells plays an important role in the resolution of in¯ ammation. And it seems reasonable to conclude that, in the absence of neutrophil apoptosis, the majority of leukocytes would undergo necrosis (as opposed to apoptosis) and release their cellular contents at sites of in¯ ammation. On the other hand, excessive apoptosis may overwhelm clearance processes and cause secondary necrosis of apoptotic bodies. Therefore, resolution of in¯ ammation is likely to be a tightly regulated process that occurs through the balance of apoptosis of neutrophils and removal of the apoptotic neutrophils by macrophages. Any impairment of this regulated
Lung In¯ ammation and Apoptosis
667
balance may result in persistent in¯ ammatory responses that could be injurious to the host tissue. The ® nding that the neutrophil is the key in¯ ammatory cell-type undergoing apoptosis following in vivo exposure to silica particles seems to be at variance with the conclusions of Iyer et al. [18], Sarih et al. [19] and Bingisser et al. [20]. These investigators reported that silica induced apoptosis in exposed macrophages in vitro; moreover, it was suggested that apoptosis of macrophages plays a key role in silica-induced pulmonary ® brosis. The discrepancy for the ® ndings reported previously and our results presented herein might be accounted for by the fact that all of the ``macrophage’’ studies were carried out in vitro. Iyer and coworkers [18] have postulated that the ® brotic potential of a particulate depends upon its ability to cause apoptosis in alveolar macrophages. These investigators exposed human alveolar macrophages to crystalline silica (133 mg=mL), amorphous silica (80 mg=mL), and TiO2 (60 mg=mL) and assessed the cells for apoptosis using morphological analysis, DNA fragmentation, and the cell death ELISA assay. Treatment of cells with crystalline silica produced an apoptotic effect, whereas exposure to amorphous silica or TiO2 particles produced no signi® cant apoptotic potential. Additional studies were conducted on the mechanism of silicainduced apoptosis in human alveolar macrophages and the authors concluded that the mechanisms related to the development of ® brosis may be associated with activation of a scavenger receptor. For the in vivo studies reported herein, the predominant BAL cell-type that underwent apoptosis following crystalline silica exposure was the neutrophil. Furthermore, we and others have demonstrated that, under in vivo silica-exposure conditions, macrophages participate in the in¯ ammatory process through a clearance mechanism, that is, through phagocytosis of apoptotic neutrophils. The discrepancy between in vivo and in vitro ® ndings may be accounted for, in part, by the studies of Boudreau and colleagues [21], who have reported that apoptosis of cells in vitro is strongly in¯ uenced by cell culture conditions. In this study, we measured apoptosis of pulmonary in¯ ammatory cells using 4 different techniques, namely, (1) standard morphological criteria; (2) a cell death detection ELISA assay; (3) development of an apoptotic DNA ladder; and (4) a TUNEL assay. Unlike necrotic cells, apoptotic cells remain intact, while retaining their cytoplasmic granules and maintain plasma membrane integrity. Other changes that are characteristic of apoptotic cells are a condensation of nuclear chromatin into dense, crescentshaped aggregated structures concomitant with the nucleus becoming more conspicuous [1, 22]. These are evident in light microscopy analyses of cytocentrifuge cellular preparations. The cell death detection ELISA assay quanti® es histone-complexed DNA fragments (mono- and oligonucleosomes) derived from the cytoplasm of cells following the induction of apoptosis. The assay can detect internucleosomal degradation of genomic DNA during
668
D. D. Zhang et al.
apoptosis. In contrast, the ELISA is speci® c for nucleosomes containing single- or double-stranded DNA. Moreover, a biochemical hallmark of apoptosis is the degradation of the genomic DNA, an irreversible event that commits the cells to die and occurs prior to changes in plasma membrane permeability. When DNA is extracted from apoptotic cells and separated by electrophoresis on an agarose gel, these DNA fragments reveal a distinctive ladder pattern consisting of multimers of the 180 to 200 base pairs of DNA associated with nucleosomes. In the fourth method, the TUNEL assay represents cleavage of genomic DNA during apoptosis and may yield double-stranded, low-molecular-weight DNA fragments (mono- and oligonucleosomes) as well as single-strand breaks (``nicks’’) in high-molecular-weight DNA. The TUNEL assay uses an enzyme (terminal deoxynucleotidyl transferase, TdT) to add biotinylated nucleotides to the strand breaks found in the DNA of apoptotic cells. They can then be detected by a ¯ uorochromelabeled streptavidin conjugate. Thus, TUNEL-positive nuclei demonstrate evidence of oligonucleosomal DNA strand breaks [23]. As discussed above, 4 different methods have been utilized in this study to assess measures of apoptosis in pulmonary in¯ ammatory cells recovered by BAL following exposures to quartz or TiO2 particles. Ideally, apoptosis would be quanti® ed as the percentage or proportion of neutrophils that were deemed to be of apoptotic as opposed to nonapoptotic morphology. The TUNEL assay, as well as morphological criteria, have the potential to provide this information, whereas measurements of DNA laddering and the cell death ELISA are only semiquantitative. However, it seems apparent that comparisons of quantitative results from morphological versus TUNEL criteria of apoptotic cells may yield differences, as evidenced by the small number of neutrophils that were identi® ed as apoptotic by the morphological criteria, which have been usually regarded as a gold standard for assessing apoptosis [9]. In contrast, the assessment of apoptosis using the TUNEL method appeared to identify a rather high number of TUNELpositive cells, with ¹60% counted in the quartz-induced in¯ ammatory model (see Table 1). Indeed, it seems highly unlikely that morphological and immunocytochemical estimations of apoptosis can yield such quantitatively disparate results. In attempting to reconcile these differences, it is conceivable that the TUNEL assay may be detecting DNA strand breaks within the neutrophils; as this has been described for silica-induced cell damage using a comet assay [24]. DNA strand breaks can be repaired and do not necessarily lead to the inevitable cell death by apoptosis or by necrosis. Alternatively, the morphological assessments of apoptosis may be somewhat distorted by the nature of the cytocentrifuge processing of lavaged cells, wherein the slides containing the pulmonary cells are spun at a high rate of speed and, as a consequence, ® ne morphological cellular features may be masked or altered (and thus underestimated). In either case, the
Lung In¯ ammation and Apoptosis
669
evidence presented herein is convincing that the results of the 4 analytical techniques demonstrate consistent apoptotic results in in¯ ammatory cells recovered from the lungs of quartz- or TiO2-exposed rats. In summary, we have demonstrated that instillation of varying concentrations of crystalline silica in rats produces pulmonary in¯ ammation and apoptosis in neutrophils. Although both the pulmonary in¯ ammation and apoptosis were sustained through 7 days post exposure, the observed effects were maximal at 24 to 48 hours, and decreased at the 3- and 7-day post exposure time periods. In contrast, exposure to pigment-grade TiO2 particles produced transient pulmonary in¯ ammation at high doses and was associated with a low level of apoptosis. This represents a signi® cant difference in the pulmonary in¯ ammatory fate between a ® brogenic dust (e.g., silicaÐ sustained) versus a low toxicity particle (e.g., TiO2 Ð transient). In addition, similar to the studies reported by Cox and colleagues [12], who exposed rats in vivo to LPS, we report here that the major cell-type undergoing apoptosis is the neutrophil, and not the alveolar macrophage. This could represent a difference in the cellular effects reported in in vitro versus in vivo apoptosis studies. Studies are ongoing to further investigate the relationship between silica-induced pulmonary in¯ ammation and apoptosis, and for the roles of macrophages, neutrophils, and epithelial cells following in vivo exposures to silica particles. REFERENCES 1. Rossi AG, Haslett C: In¯ ammation, cell injury, and apoptosis. In: Said SI, ed. Proin¯ ammatory and Antiin¯ ammatory Peptides. New York: Marcel Dekker; 1998: 9± 24. 2. Leigh, J, Wang J, Bonin A, Peters M, Ruan X: Silica-induced apoptosis in alveolar and granulomatous cells in vivo. Environ Health Perspect. 1997;105 (Suppl 5):1241± 1245. 3. Savill JS, Haslett C: Fate of neutrophils. In: Hellewell PG, Williams TJ, eds. Immunopharmacology of Neutrophils. London: Academic Press, 1994:295± 314. 4. Anderson GP: Resolution of chronic in¯ ammation by therapeutic induction of apoptosis. TIPS. 1996;17:438± 442. 5. Newman SL, Henson JE, Henson PM: Phagocytosis of senescent neutrophils by human monocytederived macrophages and rabbit in¯ ammatory macrophages. J Exp Med. 1982;156:430± 442. 6. Sanui H, Yoshia SI, Nomoto K, Ohhara R, Adachi Y: Peritoneal macrophages which phagocytose autologous polymorphonuclear leukocytes in guinea pigs. I. Induction by irritants and microorganisms and inhibition by colchicine. Br J Exp. Pathol. 1982;63:278± 284. 7. Hagimoto N, Kuwano K, Miyazaki H, Kunitake R, Fujita M, Kawasaki M, Kaneko Y, Hara N: Induction of apoptosis and pulmonary ® brosis in mice in response to ligation of Fas antigen. Am J Respir Cell Mol Biol. 1997;17:272± 278. 8. Hagimoto N, Kuwano K, Nomoto Y, Kunitake R, Hara N: Apoptosis and expression of Fas=Fas ligand mRNA in bleomycin-induced pulmonary ® brosis in mice. Am J Respir Cell Mol Biol. 1997;16:91± 101. 9. Savill JS, Wyllie AH, Henson JE, Walport MJ, Henson PM, Haslett C: Macrophage phagocytosis of aging neutrophils in in¯ ammation: programmed cell death in the neutrophil leads to its recognition by macrophages. J Clin Invest. 1989;83:865± 875. 10. Savill J, Fadok V: Corpse clearance de® nes the meaning of cell death. Nature. 2000;407:784± 787. 11. Haslett C: Granulocyte apoptosis and its role in the resolution and control of lung in¯ ammation. Am J Respir Crit Care Med. 1999;160:S5± S11.
670
D. D. Zhang et al.
12. Cox, G, Crossley J, Xing Z: Macrophage engulfment of apoptotic neutrophils contributes to the resolution of acute pulmonary in¯ ammation in vivo. Am J Respir Cell Mol Biol 1995;12:232± 237. 13. Borges VM, Falcao H, Leite-Junior JH, Alvim L, Teixeira GP, Russo M, Nobrega AF, Lopes MF, Rocco PM, Davidson WF, Linden R, Yagita H, Zin WA, DosReis GA: Fas ligand triggers pulmonary silicosis. J Exp Med. 2001;194:155± 163. 14. Kitamura Y, Hashimoto S, Mizuta N, Kobayashi A, Kooguchi K, Fujiwara I, Nakajima H: Fas=FasLdependent apoptosis of alveolar cells after lipopolysaccharide-induced lung injury in mice. Am J Crit Care Med. 2001;163:762± 769. 15. Warheit DB, Carakostas MC, Hartsky MA, Hansen JF: Development of a short-term inhalation bioassay to assess pulmonary toxicity of inhaled particles: comparisons of pulmonary responses to carbonyl iron and silica. Toxicol Appl Pharmacol. 1991;107:350± 368. 16. Warheit DB, Hansen JF, Yuen IS, Kelly DP, Snajdr SI, Hartsky MA: Inhalation of high concentrations of low toxicity dusts in rats results in impaired pulmonary clearance mechanisms and persistent in¯ ammation. Toxicol Appl Pharmacol. 1997;145:10± 22. 17. Gavett SH, Carakostas MC, Belcher LA, Warheit DB: Effect of circulating neutrophil depletion on lung injury induced by inhaled silica particles. J Leuk Biol. 1992;51:455± 461. 18. Iyer R, Hamilton RF, Li L, Holian A: Silica-induced apoptosis mediated via scavenger receptor in human alveolar macrophages. Toxicol Appl Pharmacol. 1996;141:84± 92. 19. Sarih M, Souvannavong V, Brown SC, Adam A: Silica induces apoptosis in macrophages and the release of interleukin-1a and interleukin1-b. J Leuk Biol. 1993;54:407± 413. 20. Bingisser R, Stey C, Weller M, Groscurth, Russi PE, Frei K: Apoptosis in human alveolar macrophages is induced by endotoxin and is modulated by cytokines. Am J Respir Mol Biol. 1996;15:64± 70. 21. Boudreau N, Sympson CJ, Werb Z, Bissell MJ: Suppression of ICE and apoptosis in mammary epithelial cells by extracellular matrix. Science. 1995;267:891± 893. 22. Renvoize C, Bop;a A. Pallardy M, Breard J: Apoptosis: identi® cation of dying cells. Cell Biol Toxicol. 1998;14:111± 120. 23. Seki Y, Kai H, Kai M, Muraishi A, Adachi K, Imaizumi T: Mycocardial DNA strand breaks are detected in biopsy tissue from patients with dilated cardiomyopathy. Clin Carciol. 1998;21:591± 596. 24. Zhang A, Shen H-M, Zhang Q-F, Ong C-N: Critical role of GSH in silica-induced oxidative stress, cytotoxicity, and genotoxicity in alveolar macrophages. Am J Physiol Lung Cell Mol Physiol. 1999; 277:L743± L748.
