Lung Toxicity of 16 Fine Particles on Intratracheal Instillation in a Bioassay Model Using F344 Male Rats

Abstract

We have developed a bioassay model to estimate toxicity of fine particles in the lungs at an early stage after intratracheal instillation (Yokohira et al. 2005; Yokohira et al. 2007). The present experiment was conducted to improve the model by estimating appropriate doses based on dose-dependent toxicity of instilled quartz (4 mg to 0 mg) as a positive control and assessing the impact of powdered particles without suspension (Experiment 1). In addition, examination of the toxicity of a series of particles was performed with the developed bioassay (Experiments 2A, 2B, and 2C). The materials chosen were sixteen particles, including nanoparticles and diesel powder. Histopathological and immunohistochemical analysis of bromodeoxyuridine (BrdU) incorporation and inducible nitric oxide synthase (iNOS) were performed after exposure of the lungs.

A dose of 2 mg quartz suspended in 0.2 mL saline was suggested to be most appropriate for sensitive detection of acute and subchronic inflammatory changes. Although some materials, including nanoparticles, demonstrated toxicity that was too strong for sensitive assessment, the ranking order could be given as follows: CuO > quartz > neutralized Na₂PdCl₄ > NiO > hydrotalcite > MnO₂ > diesel > titanium dioxide (in Experiment 2B) > β-cyclodextrin > diesel standard > titanium dioxide (in Experiment 2A) > CaCO₃.

Introduction

Recently, the health risk of fine particles in the air has attracted much more attention with the development of nanotechnology. Nanoparticles clearly could be a source of air pollution, but there are very few established evaluation methods for detection of lung toxicity resulting from fine particles in our environment.

We have described an in vivo bioassay for detection of hazards owing to fine particles by intratracheal instillation (Yokohira et al. 2005) which can be used, at an early stage, for risk assessment of materials inhaled into human deep lung tissue. Sequential analysis of the effects of quartz (DQ-12, 4 mg/rat), a typical lung toxic agent (Bruch et al. 1993), revealed days 1 and 28 after instillation to be most appropriate for detection of acute and subacute inflammatory changes, respectively, with bromodeoxyuridine (BrdU) incorporation on day 1 and inducible nitric oxide synthase (iNOS) levels on day 28 as suitable end points. For further validation, toxicity of fine particles from five various materials (quartz, hydrotalcite, potassium octatitanate, palladium oxide, and carbon black) was examined in our in vivo bioassay, with a special focus on correlations between immunohistochemical and histopathological findings (Yokohira et al. 2007). Antibodies specific for BrdU provide a sensitive method for detecting DNA replication for DNA repair and cell proliferation in situ (Gratzner 1982), whereas iNOS is associated with development of lung damage, inflammation, granulomas, and fibrosis induced by inhalation of silica (Castranova et al. 2002).

The dose of 4 mg used by Mercer et al. (2003) is too large to allow assessment of slight changes. In fact, toxicity of the five materials in our previous experiment was found to be only high (quartz only) or low (other particles), with no intermediate toxicity. In many reports of assessing toxicity of fine particles using quartz as a positive control, the doses employed were under 4 mg (Ernst et al. 2002; Sayes et al. 2007; Zhang et al. 1998; Zhang et al. 2002).

In our previous experiments (Yokohira et al. 2005; Yokohira et al. 2007), test particles were suspended in saline, but this is associated with considerable agglutination. To prevent this situation as much as possible, we tested here the efficacy of a dry powder instillation method, using a dry powder insufflator without a vehicle (Bot et al. 2000; Suarez et al. 2001).

One of the purposes of the present experiment (Experiment 1) was thus to determine appropriate doses of instilled quartz as a positive control to detect toxicological changes sensitively for our bioassay as a screening model, and to assess the impact of powdered quartz particles not in suspension. The toxicity of a series of different particle types, including nanoparticles and diesel exhaust particles, was also tested in our new bioassay (Experiment 2A–C).

Materials and Methods

Chemicals

Quartz dust (DQ-12) was obtained from Douche Montan Technologie (GmbH, Germany) with a particle diameter less than 7 μm. Titanium dioxide (rutile form, lot TCG4139) was from Wako Pure Chemical Industries, Ltd. (Osaka, Japan) with a particle diameter less than 5 μm. Hydrotalcite (trade name Kyoward 500, PL-1686), with an average diameter of 7.81 ± 1.52 μm, was obtained from Kyowa Chemical Industry Co., Ltd. (Kagawa, Japan). β-cyclodextrin (CAS7585–39–9) was from Ensuiko Sugar Refining Co., Ltd. (Tokyo, Japan). Potassium tetrachloropalladate (II) (K₂PdCl₄: CAS10025–98–6) and sodium tetrachloropalladate (II) (Na₂PdCl₄: CAS13820–53–6) were from Kanto Chemical Co., Inc. (Tokyo, Japan), and CuO (CAS1317–38–0) with a particle diameter less than 5 μm was obtained from Sigma-Aldrich (St. Louis, MO, USA). CuO (CuO nano: CAS1317–38–0) with a particle diameter of 33 nm was from Sigma-Aldrich (St. Louis, MO, USA). MnO₂ (CAS1313–13–9) with a particle diameter of 10μm was from Sigma-Aldrich (St. Louis, MO, USA). NiO (CAS1313–99–1) with a particle diameter less than 10 μm was from Sigma-Aldrich (St. Louis, MO, USA). NiO (NiO nano: CAS1313–99–1) with a particle diameter of 10–20 nm was from Sigma-Aldrich (St. Louis, MO, USA). Diesel powder standard (Certified Reference Material, CRM No. 8: Vehicle exhaust particulates) and diesel powder (exhaust particles of a diesel engine) were special gifts from the National Institute for Environmental Studies (Tsukuba, Japan). Palladium (II) propionate (C₆H₁₀O₄Pd: CAS3386–65–0) was obtained from Sigma-Aldrich (St. Louis, MO, USA). CaCO₃ (CAS471–34–1) with a particle diameter less than 10 μm was from Sigma-Aldrich (St. Louis, MO, USA). K₂PdCl₄ and Na₂PdCl₄ were dissolved in distilled water. Other particles or chemicals were dissolved in saline (Otsuka Isotonic sodium chloride solution from Otsuka Pharmaceutical Factory, Inc., Tokushima, Japan).

Animal Treatment

Male F344/DuCrlCrj rats (eight weeks of age), purchased from Japan Charles River (Atsugi, Japan), were maintained in the Kagawa University Animal Facility according to the Regulations for Animal Experiments of Kagawa University. The protocols of the experiments were approved by the Animal Care and Use Committee. The animals were housed in polycarbonate cages with white wood chips for bedding; and they were given free access to drinking water and a basal diet, CE-2 (CLEA Japan Inc., Tokyo, Japan); and kept under controlled conditions of humidity (60 ± 10%), lighting (twelve-hour light/dark cycle), and temperature (24 ± 2°C). The experiments were started after a two-week acclimation period.

Experiment 1

Sixty 10-week-old male F344 rats were randomly separated into six groups of ten rats each. Groups 1–4 were exposed by intratracheal instillation to 4 mg, 2 mg, 1 mg, and 0 mg (control) quartz (DQ-12) suspended in saline (0.2 mL) using a specially designed aerosolizer (Penn Century, PA, USA), and subgroups of five rats each were sacrificed on days 1 and 28 thereafter. Groups 5 and 6 were exposed to 4 mg and 0 mg (air-only control) quartz powder without suspension in any vehicle using another type of aerolizer, the DP-4 Insufflator (dry powder insufflator) (Penn Century, PA, USA) and sacrificed on the same days.

Experiment 2A

A total of 108 rats were randomly separated into nine groups of twelve animals each. Groups 1–4 were exposed by intratracheal instillation of 2 mg/rat quartz, titanium dioxide, hydrotalcite, and β-cyclodextrin suspended in 0.2 mL saline, and Groups 5 and 6 were exposed by intratracheal instillation of 2 mg/rat K₂PdCl₄ and Na₂PdCl₄ suspended in 0.2 mL distilled water. Distilled water was employed as a vehicle for the reason that there was a difference only in the bases, Na or K, between K₂PdCl₄ and Na₂PdCl₄. The remaining groups, Groups 7–9, were maintained as controls (saline, distilled water, and untreated). Subgroups of six rats were sacrificed on days 1 and 28.

Experiment 2B

A total of 106 rats were randomly separated into nine groups (Group 1–8: twelve rats each, Group 9: ten rats). Groups 1–7 were exposed by intratracheal instillation of 2 mg/rat quartz, titanium dioxide, CuO, CuO nano, MnO₂, NiO, and NiO nano suspended in 0.2 mL saline, and the remaining Groups 8 and 9 were maintained as controls (saline and untreated). Subgroups of six rats were sacrificed on days 1 and 28.

Experiment 2C

A total of ninety-six rats were randomly separated into eight groups of twelve rats each. Groups 1–4 and 6 were exposed by intratracheal instillation of 2 mg/rat quartz, diesel standard powder, diesel powder, and C₆H₁₀O₄Pd, CaCO₃ suspended in 0.2 mL saline, and Group 5 was exposed to neutralized Na₂PdCl₄ (final pH was 6.5) in 20 mM phosphate buffer and 5N NaCl. Groups 7 and 8 were maintained as vehicle controls (saline and vehicle of Group 5). Subgroups of six rats were sacrificed on days 1 and 28.

All rats in Experiments 2A, 2B, and 2C received an intraperitoneal injection of BrdU (100 mg/5 mL saline/kg/rat) (Tokyo Kasei Kogyo Co., Ltd., Tokyo, Japan) at one hour before sacrifice. At autopsy, the lungs, liver, and kidneys were removed. The lungs were weighed, including trachea and heart, first, rinsed in 10% neutral buffered formalin after excision, and then infused with 10% neutral buffered formalin. The weights of the lungs were calculated by subtraction of the weight of the remaining trachea and heart from that of lungs including trachea and heart. Six hours later, the fixative solution was changed to 100% ether to preserve antigens. Other organs were immersed in 10% neutral buffered formalin for a week. Slices of organs were routinely processed and embedded in paraffin for histopathological examination of hematoxylin and eosin (H & E)-stained sections (Experiments 1, 2A, 2B, and 2C) and immunohistochemistry (Experiments 2A, 2B, and 2C). Histopathological specimens were two slices of the left lobe, and one slice each of the other lobes. In Experiment 1, histopathological examination of H & E-stained sections was conducted without immunohistochemistry.

Histopathological Analysis (Experiments 1, 2A, 2B, and 2C)

Each lung lobe was examined histopathologically for neutrophil infiltration in the walls and spaces of the alveoli, pulmonary edema, pulmonary fibrosis, macrophage infiltration in the alveoli, restructuring of walls, and granulation. A granulomatous inflammatory change was classified as a granuloma when small collections of epithelioid cells were observed surrounded by a rim of lymphocytes and at least several alveolar spaces were filled and obstructed (Figure 1A). Severity for each parameter was divided into: 0, no change; 1, weak; 2, moderate; and 3, severe. The largest total occupancy of granulation per slide was taken as a standard for “3, severe.” Other levels of severity of granulation were determined with reference to this standard. These scores were used for histopathological assessment.

Immunohistochemical Analysis (Experiments 2A, 2B, and 2C)

Lungs were immunostained for BrdU and iNOS by the avidinbiotin complex (ABC) method, and all staining processes from deparaffinization to counterstaining with hematoxylin were performed automatically using the Ventana Discovery staining system (Ventana Medical Systems, AZ, USA). Antimouse BrdU monoclonal antibody (code No. M0744 purchased from DAKO, Glostrup, Denmark) and antirabbit inducible nitric oxide synthase (iNOS) polyclonal antibody, NOS2 (N-20) (sc-650 purchased from Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used at 1:100 dilution.

Image Analysis (Experiments 2A, 2B, and 2C)

Each section stained immunohistochemically was examined with the assistance of an image analyzer (IPAP; Image Processor for Analytical Pathology, Sumika Technoservice Co., Hyogo, Japan) (Watanabe et al. 1994). More than twenty microscope images (400X) from all lung lobes of the rats were assessed for numbers and areas of immunohistochemically positive cells. Numbers and areas were indicated as follows: %No (labeling indices) = (number of positive cells/total number of lung cells) × 100; %area = (total area of positive cells/total area of lung) × 100. In this experiment, data for area as well as number were used to indicate positive parameters. Since an image analyzer was used for counting positive cells, adjacent positive cells were sometimes counted as only 1. Therefore, %area was also calculated to indicate positive lesions for each parameter. At least 4000 cells per group were counted to generate BrdU and NOS labeling indices.

Statistics

Percentage data from immunohistochemistry were analyzed by the Tukey-Kramer post-hoc test. P values less than .05 were considered significant.

Comparative Analysis

To assess toxicity to lungs for each fine particle material and to screen them for hazard in a relatively simple way, toxicity points for each particle or chemical were calculated and placed in order. The process to obtain toxicity points was as follows:

1. Histopathological change: average points for the seven parameters were summed for each test particle group, and ratios to the respective control value were calculated. For BrdU and iNOS, total percentages averaged from area and number were corrected for control group values.

2. Points for each marker (histopathology, BrdU, and iNOS) were the converted to a ten-point scale, with 10 as the value for quartz.

3. Total points added together for histopathological change, BrdU, and iNOS were used to rank the materials in order of toxicity (Table 1).

Results

Experiment 1

After intratracheal instillation, no symptoms of dyspnea were observed. The general condition of the rats in all groups demonstrated no remarkable change during the experimental period.

As macroscopic findings, white spots were seen on the surfaces of the lungs in some treated groups on day 28. The numbers of spots in rats treated with 1 or 0 mg of quartz were far fewer than in those treated with 2 or 4 mg quartz. The spots in lungs treated with 4 mg quartz powder in 3 mL of air were fewer than in lungs treated with 4 mg quartz suspended in 0.2 mL saline. The liver and kidneys demonstrated no remarkable changes (data not shown).

Histopathological changes of lungs are illustrated in Figure 1. Main findings were neutrophil infiltration in the walls and spaces of the alveoli, pulmonary edema, pulmonary fibrosis, histiocyte infiltration in the alveoli, restructuring of alveolar walls, and granuloma. Scoring indices for histopathological change are summarized in Table 1. In the groups of quartz suspended in saline, granulomas were observed at 2 mg and 4 mg, with only mild inflammation at 1 mg. The 4 mg quartz dry powder also caused mild inflammation. The liver and kidneys did not demonstrate significant histopathological alteration in any group.

Experiment 2A

Following intratracheal instillation, all rats instilled K₂PdCl₄ or Na₂PdCl₄ suffered from dyspnea and died. Autopsies were performed, and the lungs, liver, and kidneys were removed for histopathological examination (Figure 2). Severe hemorrhage, congestion, and edema in the lungs appeared to be the causes of death. The general condition of the rats in all other groups demonstrated no remarkable alteration during the experimental period.

As macroscopic findings, white spots were observed on the surfaces of the lungs after treatment with quartz or titanium dioxide (data not shown).

Histopathological changes of lungs on day 28 are illustrated in Figure 3 (quartz, titanium dioxide, hydrotalcite, and β-cyclodextrin), and scoring indices are summarized in Table 3. The liver and kidneys did not demonstrate significant histopathological alteration in any group.

Numbers and areas of BrdU-positive cells were the highest overall in lungs of rats in the quartz-treated group on days 1 and 28 (Table 4). iNOS-positive cells were also most frequent in the quartz-treated group. Values were also elevated with hydrotalcite, β-cyclodextrin, and titanium dioxide, in decreasing order.

Experiment 2B

Fourteen rats instilled with CuO nano rapidly died, and only three rats survived day 1 following intratracheal instillation. These surviving rats were sacrificed after intraperitoneal injection of BrdU on day 1. At 3 days after instillation of NiO nano, all six rats scheduled for sacrifice on day 28 had died. It was not possible to detect the cause of death. The general condition of the rats in other all groups demonstrated no remarkable change during the experimental period.

As lung macroscopic findings, white spots were observed on the surfaces of the lungs after treatment with quartz and titanium dioxide, but not in any of the other groups.

Histopathological changes of lungs are illustrated in Figure 4 (no findings shown for MnO₂ since only mild inflammation was evident), with scoring indices in Table 3. On day 1, the scoring indices with CuO nano and CuO were higher than with quartz, employed as the positive control. In particular, the lungs treated with CuO nano exhibited severe edema. On day 28, lesions in the lungs treated with CuO were nearly equal to the damage caused by quartz. The NiO group demonstrated only mild inflammation and titanium dioxide, and MnO₂ groups also featured only mild change compared with the saline control group. The liver and kidneys did not demonstrate significant histopathological alteration.

Numbers and areas of both BrdU- and iNOS-positive cells in lungs of rats in the CuO- or CuO nano-treated group were much higher than with quartz on day 1 (Table 3). But on day 28, both BrdU- and iNOS-positive cells were seen to be only slightly increased as compared with the saline control group.

Experiment 2C

At three hours after intratracheal instillation of C₆H₁₀O₄Pd, all twelve rats had died. The general condition of the rats in other all groups demonstrated no remarkable change during the experimental period.

As lung macroscopic findings, in addition to white spots with quartz, particle coloring was evident with diesel standard powder and diesel powder (not illustrated). Histopathological changes are illustrated in Figure 5, and scoring indices are summarized in Table 2. On day 1, the scoring indices for C₆H₁₀O₄Pd and neutralized Na₂PdCl₄ were much higher than for quartz. In particular, the lungs treated with C₆H₁₀O₄Pd exhibited severe hemorrhage. On day 28, high scores were apparent for lungs treated with quartz and neutralized Na₂PdCl₄. The liver and kidneys did not demonstrate significant histopathological alteration.

Numbers and areas of BrdU-positive cells with neutralized Na₂PdCl₄ were higher than with quartz on day 1 (Table 3), but on day 28, values were equal for both. BrdU- or iNOS-positive cells demonstrated only a slight increase with diesel standard powder and diesel powder.

Toxicity classification (Experiments 2A, 2B, and 2C)

Results of the evaluation of particle toxicities by “total toxicity point” are listed in Table 4. Summarized results for fine particle toxicity are detailed in Table 5. Only CuO caused greater toxicity than quartz, and the other particles had relatively minor effects.

Discussion

From the results of Experiment 1, 2 mg quartz suspended in 0.2 mL saline was suggested to be the most appropriate dose for sensitive detection of acute and subchronic inflammatory changes. The dose of 4 mg does not induce any novel extra alterations, and 1 mg is too weak to cause change in histopathological parameters on Day 28. Although the dry powder instillation method may be suitable for experiments needing dry powder formulations, for example to increase stability in vaccine studies (LiCalsi et al. 1999), effects may be weakened. There are also problems with exposure of researchers, since there is the possibility of aspiration and difficulty in maintaining a precise dose, because control is impossible with loss in expired air after intratracheal instillation.

The various particles employed in the present study—quartz, titanium dioxide, hydrotalcite and β-cyclodextrin, K₂PdCl₄, Na₂PdCl₄, CuO, CuO nano, MnO₂, NiO, NiO nano, diesel standard powder, diesel powder, and C₆H₁₀O₄Pd, CaCO₃— were chosen for their likely human exposure. Hydrotalcite is used as an antacid, and β-cyclodextrin as a food additive and in the pharmaceutical field to improve dissolution, chemical stability, and bioavailability (Hirayama and Uekama 1999; Monnaert et al. 2004). CaCO₃ is chalk (Zhang and Huang 2005), and in industrial plants, workers may breathe in K₂PdCl₄ (Omrani et al. 2000), Na₂PdCl₄ (Maiti et al. 2001) and C₆H₁₀O₄Pd (these three particles are used as catalysers), CuO, CuO nanoparticle, MnO₂, NiO, or NiO nanoparticles. Diesel powder is a problem air pollutant capable of inducing inflammation (Folkmann et al. 2007).

All rats instilled with K₂PdCl₄ or Na₂PdCl₄ solution died with severe hemorrhage and edema of the lungs following intratracheal instillation in Experiment 2A (Figure 2). The pH values of K₂PdCl₄ or Na₂PdCl₄ solutions were strongly acid at 3.2, which presumably was a major contributor to lung damage. The lungs instilled with neutralized Na₂PdCl₄ also demonstrated severe inflammatory changes on day 1. The lungs treated with C₆H₁₀O₄Pd indicated almost the same findings, severe hemorrhage and edema, as with K₂PdCl₄ and Na₂PdCl₄. However, the pH value of C₆H₁₀O₄Pd solution was 4.9, which was not sufficiently acidic to lead to death. Schmid et al. (2007) reported that PdSO₄ exerts significant effects on human epithelial lung cells, and we have confirmed PdO toxicity in rats (Yokohira et al. 2007). From the available results, there is a possibility that not only strong acids but also palladium itself have toxic potential to the lung.

With CuO and NiO, diameters of the particles were in the micrometer and nanometer ranges. With the same volume of material, the nanoparticles of both CuO and NiO caused much more severe acute toxicity. In general, nanoparticles are likely to exert a greater impact because of the area and character of the surfaces (Powers et al. 2006). Furthermore, effects may be exerted in secondary sites. For example, an oral administration study in mice indicated titanium dioxide to be retained in the liver, spleen, kidneys, and lung tissues after uptake through the gastrointestinal tract (Wang et al. 2007). However, no remarkable histopathological changes were seen in the liver and kidneys after treatment with nanoparticles of CuO or NiO in the present study. Also, NiO nanoparticles induced only mild inflammatory change in the lungs on day 1. Toya et al. (1997) similarly found limited histopathological effects on the lungs with 13.0 mg NiO powder/kg body weight. Further examination is necessary to clarify the causes of death observed here.

To establish a bioassay model for detection of lung toxicity resulting from fine particles for screening purposes, both positive and negative controls are needed. Quartz is known as a typical lung toxic agent (Bruch et al. 1993) and provides a reliable positive control. Titanium dioxide was classified as belonging to Group 2B (possibly carcinogenic to humans) in a recent International Agency for Research on Cancer (IARC) publication (IARC Monograph 2006), but Hext et al. (2005) concluded from an overview of the epidemiology studies and toxicology studies in mice, rats, and hamsters that TiO₂ dust should not exert carcinogeniceffects on the human lung. On the other hand, titanium dioxide has carcinogenic potential (Lee et al. 1985) and causes formation of DNA adducts (Gallagher et al. 1994). With earlier assessment of acute or subacute findings, severity of toxicity differed with instilled particles: quartz > 80:20 anatase:rutile, ultrafine TiO₂ > fine-sized rutile TiO₂ = rutile TiO₂ ultrafine (Warheit, Hoke et al. 2007; Warheit, Webb et al. 2007). It should be considered within an allowance to employ rutile titanium dioxide as a negative control in our experimental purposes. In both Experiments 2A and 2B, titanium dioxide was employed as a negative control. But in Experiments 2A and 2B, total toxicity points were 3.0 and 8.9, respectively. A total point value of 8.9 seems excessively high as a negative control. However, from the total points in these experiments, at least, three grades of toxicity level (severe, moderate, and mild) could roughly be determined. It is expected that further examinations to obtain more particle toxicity data should allow more sensitive assessment of the toxicity of particles in the future.

Our bioassay does suffer from a weakness in that even low doses of some particles may lead to death, for example withn K₂PdCl₄ (Experiment 2A), Na₂PdCl₄ (Experiment 2A), CuO nano (Experiment 2B), NiO nano (Experiment 2B), and C₆H₁₀O₄Pd (Experiment 2C). Further dose range studies are necessary. Intratracheal instillation has been proposed as the most reliable route for assessing the pulmonary toxicity of particles in rodents (Warheit et al. 2005), although there are biologically different responses to inhalation and instillation (Osier and Oberdorster 1997). Our experiments indicated the order of toxicity by intratracheal instillation to be as follows: CuO > quartz > neutralized Na₂PdCl₄ > NiO > hydrotalcite > MnO₂ > diesel > titanium dioxide (in Experiment 2B) > β-cyclodextrin > diesel standard > titanium dioxide (in Experiment 2A) > CaCO₃ (Table 5).

Another limitation with this bioassay is that toxicity of particles can be detected only at an early stage because the experimental period is limited to 28 days. In this experiment, a dose response study was performed for quartz, not for risk assessment, but for improvement of this bioassay and was not performed with other test particles. For risk assessment, extra experiments including dose response studies of each test particle with hazard characterization should be employed. That said, the bioassay was originally designed to be used for hazard identification at an early stage with a ranking order toxicity of various particles at a single concentration. The approach adopted is clearly also useful for detection of acute and subacute pulmonary particle characteristics with a histopathological scoring system and markers like BrdU and iNOS for screening purposes.