Eosinophilic Inclusions in Rat Clara Cells and the Effect of an Inhaled Corticosteroid

Abstract

Large eosinophilic cytoplasmic inclusions (ECIs) are occasionally seen in untreated rat Clara cells. Following inhalation exposure to a corticosteroid, the number of ECIs was increased. This is the first histopathological description of rat ECIs and attempted characterization by immunohistochemistry, in situ hybridization, and electron microscopy. ECIs were strongly positive for surfactant protein D (SP-D) and weakly positive for Clara cell specific protein (CCSP). Clara cell cytoplasm was positive for CCSP mRNA regardless of ECIs, but not within ECIs. Corticosteroid treatment and ECI presence did not affect the immunohistochemistry and in situ hybridization staining intensities. Electron microscopy revealed large intracytoplasmic granules with an irregular limiting membrane. The ECI number was microscopically quantified in rats from three-, six-, and twenty-four-month studies. The mean ECI counts in treated rats increased from three- to fifty-four-fold with a positive dose-related trend, when compared with vehicle controls. Although the mechanism is unclear, SP-D and to a lesser extent CCSP accumulate in the ECIs. As human bronchial epithelium does not appear to contain structures analogous to the ECI, it is suggested that the observation of an increased number of ECIs in the treated rats is not likely to be relevant for human clinical risk assessment.

Introduction

Clara cells are components of the nonciliated epithelium of respiratory conducting and transitional airways and are predominantly seen in distal bronchioles. The morphology and distribution of this cell varies between species (Plopper 1983). Rat Clara cells have a characteristic apical dome shape composed of protruding cytoplasm containing smooth endoplasmic reticulum (SER) and secretory granules. Compared to the rat, SER is less abundant in the Clara cells of dogs and primates, and electron-dense secretory granules are rare in dogs (Haschek, Witschi, and Nikula 2002). Using a standard H&E section, large eosinophilic cytoplasmic inclusions (ECIs) are occasionally seen at an apical location in a small proportion of Clara cells in normal rats, although little is known about these structures. In general terms, cytoplasmic inclusions may be seen in association with various conditions including increased or altered cytoplasmic organelles such as secretory granules (Dinsdale, Verschoyle, and Ingham 1984), rough endoplasmic reticulum (Ward et al. 2001), mitochondria (Spicer et al. 1990), or in pathological conditions such as virus infections (Percy and Barthold 2001).

Clara cells secrete a variety of proteins including Clara cell secretory protein (CCSP), surfactant proteins (SP-A, SP-B and SP-D), and tryptase. They are also a major site in the lung for cytochrome P450 enzymes (Singh and Katyal 2000). These constituents can be demonstrated within the cytoplasm by immunohistochemical techniques (Singh, Katyal, and Gottron 1985; Voigt et al. 1990; Crouch et al. 1991; Phelps and Floros 1991; Hackett, Shimizu, and Gitlin 1992). Increased synthesis, impaired secretion, or endocytosis of proteins are potential causes of ECIs.

Recently, we observed an increased number of ECIs in rats treated with an inhaled corticosteroid in toxicity studies ranging in duration from three months up to two years. Corticosteroids are known to up-regulate various molecules including CCSP, IL-1 receptor antagonist, IL-1R2, IκBα, β₂-adrenoceptor (Adcock 2003), surfactants, and P450 enzymes (Berg et al. 2002), which could be associated with an increased incidence of the ECIs in the corticosteroid treated rats. To our knowledge, this is the first histopathological description of the ECIs in the rat. We have characterized the ECIs in the rat lung and report the effect of a corticosteroid on their development.

Materials and Methods

Animals and Study Design

Wistar Han rats (Crl: WIBR) were obtained from Charles River (UK) Ltd. Serologic monitoring for microbial agents including common viruses was regularly performed by the supplier. All animals were kept in the barrier facility, and the air supply was filtered and not recirculated throughout the experiments.

Studies were conducted separately for three months (twelve rats/sex/group), six months (twelve rats/sex/group), and two years (sixty animals/sex/group). Animals were exposed to the vehicle (lactose) or corticosteroid dry powder at three levels—low, intermediate, and high doses—by an inhalation system comprising a snout-only inhalation exposure chamber, restraining tubes, and a dust generator to produce an aerosol from powder. Animals were placed in the chamber and exposed for sixty minutes per day. The average mass median aerodynamic diameter of the corticosteroid dry powder ranged between 2.5–4.7 μm.

All dose levels in the studies were expressed as achieved doses estimated by the following calculation with an assumption of 100% deposition in the respiratory tract: D = (RMV × T × C)/BW.

D: dose (μg/kg/day);

RMV: 4.19 × BW(g)^0.66 (mL/min) as described by McMahon, Brain, and Lemott (1977);

T: duration of exposure/day (minutes);

C: aerosol concentration (μg/L); and

BW: group average body weight for study (g).

The estimated achieved doses of the corticosteroid at three levels were as follows: three-month study at 0, 4.1, 8.0, and 22.8 μg/kg/day in males and at 0, 4.7, 9.1, and 26.3 μg/kg/day in females; six-month study at 0, 3.0, 7.7, and 18.8 μg/kg/day in males and at 0, 3.5, 9.2, and 22.3 μg/kg/day in females; two-year study at 0, 0.9, 3.0, and 8.0 μg/kg/day in males and at 0, 1.1, 3.5, and 9.5 μg/kg/day in females. There were two vehicle control groups in the two-year study.

All studies complied with the Animals Act 1986 and the associated Codes of Practice for the Housing and Care of Animals used in Scientific Procedures and the Humane Killing of Animals under the Act.

Histology

Lung tissues were lightly inflated with 10% neutral buffered formalin with a syringe using steady pressure at necropsy and further fixed with the same fixative, processed to 4–5 μm sections and stained with haematoxylin and eosin (H&E). Selected sections were also stained with periodic acid Schiff (PAS). Longitudinal sections of the left lobe and right caudal lobe and transverse sections of the right cranial and middle lobes from each animal were examined under a light microscope by a pathologist and peer-reviewed by other pathologists. The same sections from four lung lobes were used for quantitation of the ECIs.

Immunohistochemistry (IHC)

IHC was carried out on formalin fixed and paraffin embedded sections. A standard indirect peroxidase method (Avidin/Biotin Complex/Horseradish Peroxidase–Diaminobenzidine) was used. Primary antibodies included Clara cell secretory protein (CCSP) (USB and Upstate, multiple suppliers due to antibody availability), cytochrome P450 enzymes (CYP2B1 and CYP2B2: RDI, RDI-CYP2B12abr), SP-A and SP-B (Santa Cruz Biotech), SP-D (Chemicon), mitochondrial inner membrane (MIM, Abcam), TOM20 (mitochondrial outer membrane marker, Santa Cruz Biotech), and mast cell tryptase (MCT, Dako Cytomation).

Antigen retrieval techniques included autoclave for ten minutes (SP-A and SP-B), trypsin for ten minutes at 37°C (CCSP), microwave for fifteen minutes (SP-D and TOM20), proteinase K (MIM), and target retrieval solution (Dako Cytomation) for twenty minutes at pH6 (MCT). Isotype control antibodies appropriate to each primary antibody were included in all IHC runs. Isotype controls included IgG (normal rabbit immunoglobulin fraction, Dako Cytomation), IgG2b (mouse monoclonal, Dako Cytomation), IgG2a (mouse monoclonal, Dako Cytomation), IgG1 (mouse monoclonal, Dako Cytomation), and IgG (goat F(ab)₂ fragment, Serotec). Vehicle control rat lung tissue was used as a positive control for all antibodies.

In Situ Hybridization (ISH)

Formalin fixed and paraffin embedded sections were used for ISH. To detect target mRNA, a cocktail of four 30-mer cDNA oligonucleotide probes was designed in-house based on either the rat CCSP or SP-D mRNA sequence as shown below:

CCSP

Probe 1: 5′ gagctgcagcagatggacagcatgagcaca 3′

Probe 2: 5′ ttctggtctcctgtgggagggtatccacca 3′

Probe 3: 5′ aaggagggagaggggaatgacagggggctt 3′

Probe 4: 5′ gagcctaggaggagggcctcaaggacttga 3′

SP-D

Probe 1: 5′-gagaaagtgaagcatggcagattgcttcct-3′

Probe 2: 5′-cagataaccaggcgctgctctccacaagcc-3′

Probe 3: 5′-tattcggatggtggcagcatagaggtctga-3′

Probe 4: 5′-tctctcctttgggcccaactggacctctag-3′

After rehydration, sections were pretreated in proteinase K (0.5 mg/ml for fifty minutes at 37°C) and subsequently placed in a non-formamide containing prehybridization buffer (4X Denhardt’s solution, 4X standard saline citrate (SSC), 83 μg/ ml salmon testes DNA, 5 mg/ml sodium dodecyl sulfate, 4 mg/ml sodium pyrophosphate, 0.01M tris HCl) for one hour at 42°C. After washing in graded SSC, bound probe (optimal concentration 0.1 ng/μl) was localized with alkaline phosphatase-linked anti-digoxigenin antibody (Boehringer, 1:500 dilution) for one hour. The final detection step was carried out overnight using nitroblue tetrazolium (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP). Sections were counterstained with haematoxylin. After dehydration and mounting, slides were examined and a positive signal was demonstrated by the presence of dark purple or black diformazan cytoplasmic reaction product. Control procedures included hybridization with inappropriate probes of similar length and G:C ratio (Calcitonin, 3 x oligo-probe cocktail, MWG Biotech AG, Ebersberg, Germany) and hybridization in the absence of labeled probes. Vehicle control rat lung tissue was used as a positive control for both probe sets.

Electron Microscopy (EM)

Lung tissues fixed in 10% neutral buffered formalin from the three-month study (two controls and two high-dose animals) were processed for EM. They were further fixed in 4% formaldehyde/1% glutaraldehyde, postfixed in 1% Millonig’s buffered osmium tetroxide, and then processed into Spurr’s resin. Ultrathin sections 60–90nm were prepared, stained with uranyl acetate and lead citrate, and examined using a Hitachi H7500 transmission electron microscope.

Quantitation of the ECIs

All ECIs were counted manually by light microscopy in four lung lobes per animal from the three studies examined (three-, six-, and twenty-four-month duration exposure) by the same pathologist. In the study for twenty-four months, only animals surviving to the termination of the study (thirty-seven to forty-eight males and thirty to forty-three females per group) were used for quantitation. Total ECI numbers of four lung lobes from each animal were recorded for statistical analysis.

Statistical Analysis

All three treatment groups were compared against the vehicle control, using a one-sided upper Dunnett’s t-test, to determine which dose levels showed a significant increase. Six separate regression analyses for each study and for each sex were run. In addition, all the data from three studies were combined into one analysis of variance (ANOVA), so that the factors dose, exposure duration, and age were all included. Two identical analyses were run for male and female rats separately. The response variable, ECI counts was cube-root transformed, as this was found to be the transformation that produced a residual distribution closest to normal. The dose variable was log-transformed (to base 10, after adding an offset of 0.1 to avoid taking the log of zero), thus allowing a more linear fit to the data. In addition, the squared log-dose (log-dose²) effect was also included to allow for a departure from a linear relationship. Therefore, the factors in both ANOVA models were log-dose, log-dose², exposure duration, and age.

Results

Histology

ECIs were large homogeneous round globules stained with eosin in the cytoplasm of bronchiolar epithelial cells (Figure 1A). The ECIs exceeded normal size of epithelial cells, and were also stained with PAS. In both vehicle control and treated animals, the ECIs were more commonly present at the level of distal terminal bronchioles than proximal airways where increased ciliated epithelial cells were present. The ECIs were more frequently seen in treated rats (Figure 1B) when compared with vehicle controls. However, there were no morphological differences between ECIs located in vehicle controls and treated animals, and there were no other morphological differences in the bronchiolar epithelial cells of vehicle controls and treated animals. No degeneration or necrosis was evident in ECI-bearing cells. There was no increased incidence of adaptive or proliferative lesions in the lung from any of the studies. There was no morphological evidence of known respiratory viral infections.

IHC and ISH

Positive control sections (containing rat Clara cells and type II pneumocytes) were satisfactorily stained using all primary antibodies, and with the CCSP mRNA probe set. No signal was obtained using the SP-D mRNA probe set in lung sections. Isotype controls for IHC were satisfactory for all primary antibodies (Figure 1E). Negative controls for ISH were also satisfactory for the CCSP mRNA.

IHC and ISH results in the ECIs are summarized in Table 1. Clara cells were identified as CCSP positive cells in the bronchiolar epithelium with characteristically strong cytoplasmic staining. In both vehicle control and treated animals, CCSP positive cells were the predominant type at the level of the bronchiolar epithelium with increased CCSP negative ciliated epithelial cells towards proximal airways. In the terminal bronchioles, the majority of the epithelial cells were CCSP positive. There was no morphological difference in the population of the CCSP-positive cells at the same level of bronchioles between vehicle controls and treated animals. Clara cells were positively stained for SP-D in the ECIs and cytoplasm (Figure 1C). CCSP was weakly positive in the ECIs with greater positive signal in the cytoplasm of Clara cells (Figure 1D). All observed ECIs were within positively stained Clara cells. The intensities of IHC signal were not demonstrably different between vehicle control and treated animals for any of the positively staining antibodies in either Clara cell cytoplasm or ECIs. This was also the case for CCSP mRNA in the cytoplasm (Figure 1F), for which no clear difference between vehicle control and treated Clara cells with or without ECIs could be demonstrated.

EM

The ECI was seen within nonciliated epithelial cells in the bronchiole and recognized as a vesicle containing moderately electron-dense amorphous material enclosed by an irregular limiting membrane (Figure 2A). Nonciliated cells had small round granules and rod-shaped granules in the cytoplasm regardless of the ECI. Desmosomes were also demonstrated between adjacent cells (Fig 2B). The ECI measured up to 17 μm in diameter. There were no morphological changes of the smooth endoplasmic reticulum (SER), rough endoplasmic reticulum (RER), and mitochondria in the ECI-bearing cells from vehicle control and treated animals. No structural formation including granular, filamentous, or lattice-work patterns, which are characteristic of viral inclusions, was detected in the ECIs or otherwise within the Clara cells from any animals.

Quantitation of the ECIs

All treated groups showed a statistically significant increase in ECI counts over the vehicle controls, for both sexes in all studies (p < .05). The linear regression of ECI counts against dose was statistically significant (p < .0001) from all six analyses, for each study, and for each sex. Dose effect on ECI counts was statistically significant by ANOVA (p < .0001). It was observed that the ECI counts tend to increase with age in both sexes, though this was only statistically significant in the males (p = .0143). The effect of exposure duration was deemed to be nonsignificant in both sexes. Therefore, the ECI counts did not progress with increased duration of treatment, but with age of animals.

The group means of the ECI counts (Figure 3) in treated animals increased between three-fold and fifty-four-fold when compared with vehicle controls. The highest group mean of the ECI count was seen in the male high-dose group from the three-month study, where there was an increase in mean from 0.7 (vehicle control group) to 37.8 (high-dose group). The highest individual ECI counts observed were 90 in a male from the high-dose group. There were increased ECI counts in vehicle controls from the two-year study when compared with vehicle controls from the three- or six-month studies.

Discussion

Our current studies revealed the following new findings that have not been described before: (1) a type of large eosinophilic cytoplasmic inclusion (ECI) exceeding normal size of cytoplasm was identified in rat Clara cells by light microscopy; (2) SP-D and to a lesser extent CCSP was detected within the ECI; and (3) increased ECI numbers with a positive dose relationship were induced by an inhaled corticosteroid.

At the ultrastructural level, a large cytoplasmic granule within the Clara cell, similar to the ECI, has previously been described as a Type C granule containing CCSP in the normal rat (Wasano and Hirakawa 1992). Within more common small secretory granules, SP-D was demonstrated in the rat Clara cells by immunoelectron microscopy (Crouch et al. 1992). Since we found both SP-D and CCSP consistently positive in the all ECIs and Clara cell cytoplasm, these proteins may coexist in the same ECI or cytoplasm of Clara cells.

SP–D is increasingly recognized as an important regulator of innate immune function in the lung, modulating Th2 responses and as a biomarker in asthma (Haczku, Vass, and Kierstein 2004). However, SP-D is not only demonstrated in the Clara cell and type II cells of the lungs but also in many other nonrespiratory organs (Stahlman et al. 2002) including rat stomach (Fisher and Mason 1995). We examined non-respiratory organs including stomach from current studies (unpublished data), but no ECIs were observed in these organs.

Although the mitochondria of Clara cells can be very large in some species, this is not a recognized feature of rat Clara cells (Plopper, Hill, and Mariassy 1980). In nonrespiratory organs large mitochondria have been seen in a variety of induced conditions (Wakabayashi 2002) including cortisone treatment in the rat liver (Kimberg, Loud, and Wiener 1968). Our electron microscopic observation revealed ECIs up to 17 μm, larger than giant mitochondria, and no morphological evidence of altered mitochondria in affected Clara cells. In addition, immunohistochemical staining of ECIs for mitochondrial inner or outer membrane was negative for ECIs, suggesting these inclusions were not derived from mitochondria.

Eosinophilic globules are often seen in the olfactory and respiratory epithelium of rat nasal cavities (Monticello, Morgan, and Uraih 1990). This finding can be increased by xenobiotic inhalation, but there are no widely accepted mechanisms. Our studies were carried out by snout-only inhalation, and no increased eosinophilic globules were observed in the nose of treated rats when compared with controls. ECIs were never seen in the ciliated respiratory epithelium of the respiratory tract and were different in morphology and distribution from the eosinophilic globules commonly seen in the nose.

Globule leukocytes contain eosinophilic granules and are present in the upper respiratory system and to a lesser extent in the lung of rats (Kent 1966). Contrary to ECIs, these cells are reduced upon corticosteroid treatment (Le Roy Ladurie and Fournier 1986; Tam et al. 1988). In addition, ECIs are often singular and occupy the entire cytoplasm protruding into the bronchiolar lumen, whilst the granules of globule leukocytes are generally numerous and smaller in size. Desmosomes, indicative of an epithelial cell, were identified in the ECI-bearing cells. These data illustrate that the ECIs were demonstrably different to the cytoplasmic granules of globule leukocytes.

Our ultrastructural examination did not detect any structures characteristic of viral infection in the rat lung. The morphological appearance of the ECI as a membrane bounded vesicle containing amorphous substance without perturbation of surrounding cellular cytoplasmic machinery is not characteristic of a viral inclusion. Regular serologic monitoring did not detect any common viral infections. Although this does not categorically rule out viral infections, the weight of evidence does not support a viral etiology.

Clara cells are capable of proliferation, and in the event of airway injury can act as epithelial stem cells (Otto 2002). Acute single exposure to insecticide has been shown to increase both the number of Clara cells and CCSP synthesis (Elia, Aoki, and Maldonado 2000). It is therefore possible that ECI counts might be higher if a treatment resulted in an increased ratio of Clara cell population to ciliated cells. In the mouse the number of Clara cells can be significantly increased by corticosteroid inhalation without concurrent increased secretory granules (Roth et al. 2007). Investigation of relative Clara cell population was not carried out in our studies. Such an analysis was not considered necessary due to the magnitude of increase in ECIs in treated animals when compared to vehicle controls (from three- to fifty-four-fold). As more than half of bronchial epithelial cells are Clara cells in untreated control rodents (Plopper, Mariassy, and Hill 1980; unpublished observations), this fold change can not be explained by increased Clara cell population alone. Even if the entire bronchiolar epithelium consisted of Clara cells, a less than two-fold increase could be achieved. Increased Clara cell population above this level would require cell proliferation with significant morphological changes, and no such effects were evident by microscopic observation. We did not see any morphological evidence of increased cell injury such as degeneration, necrosis, associated inflammation, or proliferative lesions in the treated animals. In addition, there was no alteration in the CCSP immunostaining between vehicle controls and treated animals that would indicate a major shift in cell population (e.g., greater distribution of CCSP-positive cells). Development of the ECI was not a progressive change, since the ECI counts did not increase significantly with an increasing duration of exposure to corticosteroid. Interestingly, ECIs were also increased in the vehicle controls animals by length of the studies in particular from the two-year study. As air-inhalation and noninhaled controls were not available, the effects of lactose vehicle, the inhalation procedure, and age of rats on the development of the ECI could not be determined. Additional work will be necessary to fully characterize if these factors have effects on the ECIs.

Why only a small number of Clara cells develop ECIs is unknown. Clara cells may have different subsets, based on profiles of surfactant proteins (Kasper et al. 1995) and enzymes (Keith et al. 1987; Plopper et al. 1993). This is reflected in the variable response of these cells to pulmonary toxicants (Plopper et al. 1992). In our studies it was not clear whether Clara cell heterogeneity was associated with ECI development based on the IHC and ISH profiles of ECI bearing cells compared to noninclusion bearing Clara cells.

Glucocorticoids can enhance synthesis of surfactant proteins (Fisher et al. 1991; Deterding et al. 1994), CCSP, and P450 (Berg et al. 2002), where CCAT/enhancer binding proteins (C/EBPs) are possibly one of mediators (Ramji and Foka 2002). C/EBPs are expressed in the pulmonary epithelium (Cassel and Nord 2003) and appear to increase expression of lung epithelial proteins including SP-A (Rosenberg et al. 2002), SP-D (He and Crouch 2002), CCSP (Nord et al. 1998), and cytochrome P450 enzyme CYP2B1 (Luc et al. 1996). Expression of the C/EBPs was not examined in our studies. However, we did not see any positive immunostaining for SP-A and P450 within the ECIs, nor any increase in cytoplasmic SP-A or P450 after the treatment. The effect of corticosteroids on SP-A might depend on different levels of exposure (Iannuzzi, Ertsey, and Ballard 1993). Age of animals might also be important, since the corticosteroid effect on SP-D was seen in fetal lung, but not in adult lung (Ogasawara et al. 1992).

No positive ISH signal for SP-D mRNA was detected despite using a cocktail of four individual probe sets based on the rat mRNA sequence. In mice an unrelated ISH study of surfactant protein mRNAs also failed to detect SP-D mRNA, despite clear signals for the other three proteins examined (Pilling et al. 1999). However, as other investigators have successfully demonstrated SP-D mRNA in the Clara cells in rats (Yano et al. 2000; Kasper et al. 2002) and mice (Wong et al. 1996), the inability to produce a positive ISH signal for SP-D mRNA throughout the lung tissue including type II pneumocytes is likely due to a technical failure rather than representing evidence of a lack of SP-D mRNA in the rat Clara cell.

We did not observe any greater intensity of the IHC or ISH signals in treated animals compared to vehicle controls, or in the ECI-bearing Clara cells compared to normal Clara cells. There are a number of possible explanations for the increase in ECIs in treated rats, and up-regulation of SP-D and CCSP is one possibility. SP-D and CCSP can accumulate without up-regulation of these proteins, if cellular secretion was reduced, or if extracellular proteins were endocytosed by Clara cells. SP-D is known to be bound and internalized in Type II pneumocytes (Herbein, Savov, and Wright 2000) and alveolar macrophages (Kuan et al. 1994). Clara cells appear to have different endocytotic pathways to type II pneumocytes (Voorhout et al. 1992). It was not clear whether ECIs are related to Golgi-derived secretory vesicles or other types of vesicles including cell surface-derived endocytotic vesicles or lysosomes, as discussed elsewhere (Wasano and Hirakawa 1992). If endocytosis by the Clara cell is enhanced by corticosteroid, as is the case for other types of cells (Piemonti et al. 1999; Franchimont 2004), corticosteroid-induced ECIs might result from increased endocytosis of SP-D/CCSP by Clara cells. Although this mechanism is possible, we do not know why Clara cells would fail to degrade internalized SP-D/CCSP. Corticosteroid itself may also be endocytosed by Clara cells and be present in the ECI. In some species such as the rabbit, CCSP can bind lipophilic molecules including steroids. It has been reported that rat CCSP does not bind cortisol (Singh et al. 1990) but that it does bind to progesterone and exogenous molecules such as PCB and its metabolite (Lund et al. 1985). Species difference in binding capability of CCSP may explain why ECIs were only seen in rats. No information is available on the binding properties of the particular corticosteroid studied here, for CCSP and SP-D. If the corticosteroid were to bind CCSP forming a complex, this might be poorly metabolized and secreted from the cytoplasm. Therefore, the complex may accumulate in Clara cells, leading to the development of ECIs. In normal rats endogenous lipophilic molecules including retinoic acids and steroid hormones (Lopez de Haro, Alvarez, and Nieto 1988) can contribute to such a complex, while inhaled exogenous corticosteroid may exacerbate the development of ECIs in treated animals. A similar mechanism might also apply to SP-D, since some lipids are known to bind SP-D (Hawgood and Poulain 2001).

Inhalation toxicity studies were conducted in rats, mice, and dogs with the same corticosteroid (unpublished data). In untreated control dogs, small eosinophilic granules were often seen in the basal region of the bronchiolar epithelium, but the granules were not positive for CCSP and SP-D by IHC. They did not increase in treated dogs following corticosteroid exposure. ECIs seen in the rat often exceeded the normal size of the Clara cell, and such ECIs were not observed in any control or treated dogs and mice. Both rat and mouse have greater SP-D expression in the distal bronchioles when compared to human tissue (Mason, Greene, and Voelker 1998), and the Clara cell population in the bronchioles are higher in mice than rats (Plopper 1983). Regardless of Clara cell population and SP-D expression, for reasons unknown, ECIs appear to be a rat-specific phenomenon. As human bronchial epithelium does not appear to contain structures analogous to ECIs (A. Nicholson, professor of pulmonary pathology, Royal Brompton Hospital, London, personal communication, 2006), it is suggested that the observation of an increased number of ECIs in the lungs of corticosteroid treated rats may not be relevant for human clinical risk assessment. In the corticosteroid-treated rats, there was no morphological evidence of adverse response associated with the ECI formation. The two-year study (the approximate lifespan of rats) did not reveal any increased adaptive proliferation or neoplastic lesions. In addition, the ECI was not associated with body weight change or altered survival rate.

In conclusion, ECIs were seen in both vehicle controls and rats treated by inhalation with a corticosteroid. This feature appears to be rat-specific and is therefore considered unlikely to be relevant in terms of human risk assessment. There was no morphological evidence of degeneration or necrosis in the ECI-bearing Clara cells in either vehicle controls or treated rats. The corticosteroid induced a minor, dose-related increase in the very low background incidence of ECIs, but with no evidence of progression despite continued treatment up to two years. Accumulation of SP-D and to a lesser extent CCSP was responsible for the development of ECIs. The mechanism of the ECI development in the Clara cells is yet to be elucidated. Since the ECIs were seen in only a small population of rat Clara cells, careful or quantitative microscopic assessment may be required to identify whether the number of ECIs are altered.