Subchronic Pulmonary Pathology, Iron Overload, and Transcriptional Activity after Libby Amphibole Exposure in Rat Models of Cardiovascular Disease

Background: Surface-available iron (Fe) is proposed to contribute to asbestos-induced toxicity through the production of reactive oxygen species. Objective: Our goal was to evaluate the hypothesis that rat models of cardiovascular disease with coexistent Fe overload would be increasingly sensitive to Libby amphibole (LA)-induced subchronic lung injury. Methods: Male healthy Wistar Kyoto (WKY), spontaneously hypertensive (SH), and SH heart failure (SHHF) rats were intratracheally instilled with 0.0, 0.25, or 1.0 mg LA (with saline as the vehicle). We examined bronchoalveolar lavage fluid (BALF) and histological lung sections after 1 week, 1 month, or 3 months for pulmonary biomarkers and pathology. SHHF rats were also assessed at 6 months for pathological changes. Results: All animals developed concentration- and time-dependent interstitial fibrosis. Time-dependent Fe accumulation occurred in LA-laden macrophages in all strains but was exacerbated in SHHF rats. LA-exposed SHHF rats developed atypical hyperplastic lesions of bronchiolar epithelial cell origin at 3 and 6 months. Strain-related baseline differences existed in gene expression at 3 months, with persistent LA effects in WKY but not SH or SHHF rats. LA exposure altered genes for a number of pathways, including inflammation, immune regulation, and cell-cycle control. Cell-cycle control genes were inhibited after LA exposure in SH and SHHF but not WKY rats, whereas tumor suppressor genes were induced only in WKY rats. The inflammatory gene expression also was apparent only in WKY rats. Conclusion: These data show that in Fe-overload conditions, progressive Fe accumulation occurs in fiber-laden macrophages within LA-induced lesions. Fe overload does not appear to contribute to chronic inflammation, and its role in hyperplastic lesion development requires further examination.


Libby Amphibole
The Libby amphibole (LA) asbestos sample used in this study was collected from the Rainy Creek Complex near Libby, Montana in 2007 by the United States Geological Survey and was processed to produce inhalable material by Meeker et al., as indicated earlier (Meeker et al., 2003). The sample was further size fractionated by water elutriation as described previously (Webber et al., 2008) in order to isolate a rat respirable fraction (PM 2.5 ) using a settling velocity of 3.4 x 10 -4 cm s -1 . LA fiber size distribution and surface chemical properties have been described in recent publications (Shannahan et al. 2011a;2011b). The mean fiber length of LA was shorter (~5 µm) relative to other asbestos materials used in many studies.

Intratracheal Instillation of Libby Amphibole
The concentrations of LA selected for intratracheal instillation although high were, in general, comparable to instillation studies using other fiber types (Adamson andBakowska, 2001, Hirano et al., 1988). Doses were chosen to assure a response in the lung upon instillation allowing for a comparative analysis between strains. Theoretically, a rat will deposit 0.07 mg of fibers during 6 hour inhalation at 10 mg/m 3 based on the assumption that minute volume is 200 ml and the deposition fraction to pulmonary region is 0.10. Intratracheal instillation ensured the delivery of exact concentrations of LA into the lung and allowed us to control for likely strain-related deposition differences due to their variation in breathing parameters (Shannahan et al., 2010). In our previous study approximately 30% of WKY presented with non-pathogenic cardiac hypertrophy 5 (Shannahan et al., 2010). Therefore WKY group size was increased (n=12) to eliminate the data from those with hypertrophic hearts (heart weight > 1.3g; normal heart weight ~1.1g) and still maintain appropriate group sizes for statistical comparisons.

Cell Differential and Bronchoalveolar Lavage Fluid (BALF) Analysis
Aliquots of BALF were taken for total cell counts (Coulter Inc., Miami, FL, USA), cell differentials, and analyses of lung injury markers. Cell differentials were conducted on Cytospin preparations (Shandon, Pittsburgh, PA), and slides were stained with LeukoStat

Lung Histopathology
The left lung from each animal was trimmed, embedded in paraffin, sectioned to a thickness of approximately 3 microns (transverse), and stained with hematoxylin and eosin (H&E), Perls' Prussian blue, or Masson's trichrome. The lung lesions were evaluated with particular attention to the lesion location (e.g., bronchi, terminal bronchioles, alveolar duct, alveoli, interstiteum, centriacinar regions, pleura). The morphological evaluation took into consideration the characteristics of each 6 inflammatory component, e.g., polymorphonuclear cells, macrophages, fibrosis, microgranulomas, as well as changes in the alveolar epithelium. Lesions were also evaluated for Perls' Prussian blue staining for ferric Fe content and localization and Masson's Trichrome for collagen in order to understand modulation of pulmonary Fe content and progression of fibrosis. Histopathological changes were scored using semiquantitative grading at five levels (0 = normal; 1 = minimal; 2 = mild; 3 = moderate; 4 = severe) taking into consideration the degree of severity and the type of lesion (Shackelford et al., 2002, Nyska et al., 2005. In order to verify pathological changes and possible chronic progression of lesion development in SHHF after LA exposure the study was repeated in 36 additional SHHF rats examining lung pathology 3 months and 6 months post-exposure. SHHF rats were instilled with saline containing 0, 0.25, or 1.0 mg of LA and necropsies were performed 3 or 6 months later for detailed pathological evaluation.

Immunohistochemistry
The cellular origin of hyperplasic changes in SHHF rats was examined immunohistochemically. Five-micron thick lung sections were cut from paraffin embedded lung tissue from rats exposed to 0.0, 0.25 and 1.0 mg LA after 3 months and immunohistochemically stained for several epithelial cell markers. The hydrated lung sections were first incubated with 3% hydrogen peroxide to block endogenous peroxidase. Sections were incubated with primary antibodies specific for selected proteins. Biotinylated secondary antibodies were used with peroxidase-conjugated streptavidin to detect cellular localization of proteins of interest. Positive control slides 7 using known reactive tissues and negative control slides in which the primary antibody was omitted were also prepared. A board certified pathologist then evaluated the slides.

Gene Array
There were 6 samples per group and gene expression in each sample was assayed on separate chips. Biotin-labeled cRNA was produced from 15 µg total RNA using an This software also provided summary reports by which array QA metrics were evaluated including average background, average signal, and 3'/5' expression ratios for spike-in controls, β-actin, and GAPDH.

Functional Analysis of Gene Array Dataset
All DEGs (FDR with significance at < 0.05) were filtered to meet the criteria of intensity greater than or equal to 30 and had a fold change greater than or equal to 1.5 or less than or equal to -1.5. Significant KEGG pathway were identified and among those the pathways involved in inflammation, and cell cycle control were analyzed further at gene expression level. Inflammation genes were assembled from KEGG pathways: 4060, 8 4062, 4070, 4310, 4510, 4514, 4520, 4620, 4621, and 4670. Cell cycle control genes were assembled from KEGG pathways: 4115, and 5200. For each of the functional lists, genes that were on the pathway-specific list and on the master list were submitted to Eisen's Cluster (Eisen et al., 1998) for hierarchical clustering. The genes were median centered with average linkage. The resulting cluster was displayed using Treeview (Eisen et al., 1998). Two other functional lists, growth and fibrosis, were obtained from NetAffx (http://www.affymetrix.com/estore/analysis/index.affx). Genes that were on both the growth and fibrosis lists were clustered with Eisen's Cluster and displayed with Treeview.

Libby Amphibole Characterization
Fiber characterization parameters were recently published (Shannahan et al. 2011b

Pulmonary Inflammation, and Injury as Determined using BALF
Baseline levels (saline control) of BALF protein and albumin were significantly higher in SH and SHHF when compared to WKY (SHHF>SH>WKY) (Supplemental Material Figure 1), as evident in our previous study (Shannahan et al. 2010). Only SHHF exposed to 1 mg LA demonstrated elevated BALF protein at 3 months, however, because of the variable and high levels at baseline, it was difficult to ascertain if this increase was related to LA exposure. The recovery of alveolar macrophages in BALF was unchanged 10 after LA exposure in all strains despite increased histological cellularity and presence of fiber-laden macrophages. As seen at earlier time points, neutrophils were significantly increased at baseline in SHHF compared to SH and WKY while LA-induced increases in neutrophils noted 1 day through 1 month (Shannahan et al., 2011a) time points, were largely reversed by the 3 months in all strains. Baseline levels (saline control) of BALF ferritin and transferrin were significantly higher in SHHF when compared to WKY (SHHF>SH>WKY) (Supplemental Material Figure 1) suggesting an underlying state of CVD associated Fe-overload. However, at 3 months LA-induced increases noted during earlier time points (Shannahan et al. 2011b) were largely reversed. Figure Legends Supplemental Material, Figure 1 Supplemental Material, Figure 1 Legend: Alterations in BALF protein, albumin, macrophages, neutrophils, ferritin, and transferrin in WKY, SH and SHHF rats at 3months following intratracheal instillation of saline (control), 0.25 mg LA, or 1 mg LA.