|Home | About | Journals | Submit | Contact Us | Français|
The role of SPARC in the in vivo lung response to crocidolite asbestos was addressed by instillation of crocidolite asbestos in a series of wild type or SPARC -null mice. Animals were sacrificed at one week, one month, and three months post-instillation to assess the impact of SPARC on multiple stages in the development of fibrosis. RNA was harvested from 10 animals/time point, pooled, and used to probe a mouse array containing ~10,000 probes. Gene expression data was analyzed for fold-change, and for broader functional group alterations. As expected, the one-week time point displayed alterations in genes involved in immune recognition, energy utilization, and growth factor production. Later time points showed expression alterations for genes involved in protein degradation, Wnt receptor signaling, membrane protein activity, and transport. Molecules in the Wnt pathway have been implicated in bone growth, mediation of fibroblast activity, and have been directly linked to SPARC regulation.
As a member of the matricellular class of secreted proteins, SPARC (a secreted protein acidic and rich in cysteine) is involved in the regulation of extracellular matrix (ECM) — cellular interactions. SPARC has been localized to fibroblasts found in idiopathic pulmonary fibrosis . After embryogenesis, SPARC is most often expressed in tissues undergoing remodeling [1, 2]. SPARC is a calcium-binding glycoprotein that interacts with many ECM components including the fibrillar collagens (types I, II, III, and V) as well as type IV, thrombospondin1, vitronectin, and fibrinogen fragments D and E [3, 4]. SPARC has been shown to be a modulator of growth factors involved in fibrosis and has also been implicated in the development of cancers [5, 6, 7]. These characteristics may be linked to the activities of SPARC that include stimulating the TGF-β signaling system, serving as a target for and/or inducer of matrix metalloproteinases, and modulating integrin-linked kinase activity [5, 8, 9]. It has also been shown that SPARC interacts with the TGF-β-receptor complex to modulate phosphorylation of Smad-2, to affect the activity of JNK, and to increase the expression of c-jun . More recently, SPARC has been shown to interact with Stabilin-1, a scavenger receptor expressed on alternatively-activated tissue macrophages and sinusoidal endothelial cells . This interaction regulates the extracellular concentration of SPARC and thus potentially modulates ECM remodeling.
The 32 kDa SPARC protein is a member of a gene family with structural similarities in the arrangements of the protein modules. Other matricellular proteins in this family include SC1/Hevin, QR1, and Testican [12, 13]. The protein domains include a N-terminal acidic domain that contains a low-affinity calcium-binding domain and a transglutaminase cross-linking site. This domain has been shown to inhibit cell spreading. The central domain of the protein (aa 50 – 130) is similar to follistatin, and characteristically acts by inhibiting proliferation and abrogating focal adhesions; however, release of an internal peptide ([K]GHK) stimulates proliferation and angiogenesis. The C-terminal domain (E-C) contains a high affinity calcium binding E-F hand domain that inhibits cell spreading and proliferation, and is responsible for binding to cells and matrix [3, 12].
SPARC-null mice were developed independently by two groups [14, 15]. These mice are born phenotypically normal but develop early cataracts, kinked tails, osteopenia and increased adipose tissue. SPARC-null mice also demonstrate enhanced growth and metastasis of implanted tumors, accelerated closure of cutaneous wounds, collagen fibrils with smaller and more regular diameters and a diminished ability to encapsulate subcutaneously implanted foreign bodies (see Framson  for review). It is unknown how the lung will respond to asbestos exposures in the absence of SPARC. Our investigations are designed to discover the role SPARC plays in the development of lung fibrosis in response to asbestos exposure by examining gene expression in lungs derived from asbestos-exposed SPARC-null and wild-type mice. The specific analyses in this report target those genes or classes of genes in mouse lung tissue that change their expression during exposure to asbestos, with or without the presence of SPARC. This approach would logically describe a set of genes that can only be activated by the combination of SPARC activity and asbestos exposure.
The central questions in this study center concern 1) the types and magnitude of gene expression changes in SPARC-null mouse lungs versus wild-type mouse lungs and 2) the types and magnitude of changes seen in SPARC-null mouse lung following the stress induced by asbestos exposure. SPARC-null and wild type mice were exposed to saline or crocidolite asbestos for periods of one week, one month, or three months. Animals were euthanized and lungs were removed and submitted for RNA expression analysis. Microarray analysis was performed in a standard test versus reference RNA co-hybridization on MWG 10K Group A mouse oligo-based arrays. The use of reference RNA allowed the comparison of values from multiple time points and exposure conditions. Ten mice were pooled from each treatment group for each data point. Pooling tends to obscure subtle gene expression changes as the variance of ten biological specimens (mice) generates a high baseline of transcriptional noise. Typically only those expression changes that are of a large magnitude are seen in pooled data. In order to focus primarily on genes that reproducibly were altered, we performed two sequential threshold or filter steps. Three replicate arrays were analyzed for each experiment and mean and standard deviation was established for each gene. Genes achieving a significant (p<0.01) difference in expression compared to the saline control experiments were further analyzed. An arbitrary cutoff of two-fold up or down regulation (ratio of <0.50 or > 2) was used to further filter the data to delineate genes that were highly activated or suppressed by asbestos exposure. A general description of the numbers of genes meeting these criteria is delineated in Table 1. The largest sets of gene expression alterations occurred early (one week) in both wild type and SPARC-null mice, reflecting an acute response, but significantly more changes occurred in SPARC-null mice than the wild type mice at this time point (112 genes up or down in SPARC-null vs. 62 in wild-type mice). Smaller sets of gene expression changes take place at the one month and three month time points (Table 1). The resulting sets of gene transcripts for each time point and condition are listed in tables tables22--4.4. The datasets are from one-week exposures of crocidolite vs. saline in both wild-type and SPARC-null mice (Table 2), one-month exposures of crocidolite vs. saline in both wild type and SPARC-null mice (Table 3), and three-month exposures of crocidolite vs. saline in both wild type and SPARC-null mice (Table 4). Table 5 presents transcripts that were altered in expression level in SPARC-null mice versus wild type mice without the addition of asbestos. This represents the steady state gene expression differences within the lung and a potential list of genes related to constitutive SPARC function. Genes with expression altered by at least two-fold (up or down) with a significance of p<0.01 are listed. Notable genes in the list related to fibrosis, bone formation, lung function, or SPARC mechanisms are Gdf5, Sftpc, Fgf13, Saa3, Zmat3, and Wnt3.
Several transcripts were alternatively measured by real-time PCR for validation of the array. The values for both analyses are presented in Table 6.
GoMiner analysis combined with a cutoff of greater than 2-fold expression changes was used to derive a set of functional (gene ontology) groups that were altered. These groups were examined at each of the timepoints. These groups are described in the text below and full tables of the results will be uploaded to the NCBI-GEO database.
A comparison of transcriptional changes in mouse lung following various times of exposure to asbestos was studied in this report. The experimental design involved pooling lung RNA from a group of 10 mice for each condition/time point. This tended to compromise the smaller gene expression changes due to the variability inherent in mouse lung gene expression from mouse to mouse. However, this had the advantage of providing a biological filter for the data, only allowing significant and profound changes to survive the analysis. In addition, a fold-change cutoff of two-fold was applied to further filter the data. Asbestos exposures of one week, one month, and three months provided for examination of the process of the mouse lung response to asbestos at critical time intervals. Early response may reflect significant gene expression activity involved in recognition, immune response, and signal transduction or in the case of the knockout mice, a larger number of pathways dysregulated by the absence of SPARC. The largest number of transcription changes occurred at this time point as shown in Table 2. Later time points may reflect the physical processes of fibrosis and resolution. In separate reports [16, M. Smart et al., manuscript submitted], the level of fibrosis as assessed by hydroxyproline levels was studied in these mice. Fibrosis was significant by one month in both wild type and SPARC-null animals, however, at three months after asbestos instillation fibrosis continued to increase in wild type mice but had been reduced to control levels in SPARC-null mice, demonstrating a resolution of fibrosis in the absence of SPARC expression. This finding is similar to that of Strandjord and colleagues  who found that Sparc-null mice had reduced collagen accumulation compared to wild-type mice after bleomycin exposure. Distinct gene sets were altered at this later time point and, interestingly, more genes were down-regulated than up-regulated by three months after exposure. This may reflect an attempt by the lung to modulate or dampen the response to stress. It may well be more successful in SPARC-null than wild-type mice due to the absence of positive regulators of fibrosis. SPARC normally plays a key role in TGF-β signaling to c-jun/AP-1/junk kinase and a corresponding induction of collagen synthesis. In this data set, the transforming growth factor beta 2 gene and the transforming growth factor beta regulated gene 4 are both up regulated at one week in SPARC-null mice.
In terms of classes of genes determined by GoMiner analysis (NCBI-GEO database) that would have been expected to be altered based on the known function of SPARC versus the observed data, there was a reasonable correlation. The largest number of genes, and thus the largest number of gene categories, was altered at the one-week time point. In the mouse lung, many processes can be expected to be taking place, such as immune recognition, inflammation, attempts at asbestos removal, vascularization, and cell-cell communication. We saw significant expression changes in genes involved in circulation, growth factor activity, response to bacterium and response to other organisms (possibly similar to genes involved in asbestos response), energy use, and microtubule mobilization.
At one month post asbestos instillation, we might expect that SPARC-null mice would have a decrease in activity in pathways related to fibrosis, although the SPARC-null mice had similar levels of fibrosis at one month in studies employing hydroxyproline as a measure of fibrosis [A. Smartt et al., manuscript submitted]. We saw significant changes in GoMiner categories titled cell soma, regulation of actin filament polymerization, Golgi-associated vesicles, and the Wnt receptor signaling pathway. The Wnt receptor signaling pathway has been implicated in many related processes, including mesothelioma formation, tissue regeneration, and fibrosis [18, 19, 20].
After three months exposure to instilled asbestos, there were clear differences in the levels of fibrosis in SPARC-null versus wild type mice. We might expect several categories of genes to be altered as the fibrosis resolved in SPARC-null mice by 3 months. Interestingly, we found significant numbers of genes altered in the GoMiner categories of lyase activity, energy production, membrane activity, and transport.
An intriguing result is that each time point reflects changes in gene expression for a unique set of genes. Very few genes are altered at all three time points, or even at two time points. This result suggests that each timepoint is distinct in terms of cellular processes taking place. At all time points, in SPARC-null mice more genes are down regulated than up-regulated. This may be due to the absence of TGF-β signaling, as well as the absence of a robust fibrotic process.
Finally, we examined the set of genes that were distinctly expressed when SPARC-null mice were compared to wild type mice without the exposure to asbestos. Several genes in Table 5 had known roles in SPARC or fibrosis-related processes. These included Gdf5, Sftpc, Fgf13, Saa3, Zmat3, and Wnt3. Gdf5 encodes growth differentiation factor 5, a cartilage morphogenic factor and regulator of apoptosis . Sftpa1 encodes surfactant associated protein-C, an important innate host defense protein . Fgf13 is a fibroblast growth factor , while Saa3 is a key protein involved in the early inflammatory response . Zmat3 is involved in p53-mediated regulation of apoptosis , and Wnt3 is a key molecule in the fibrotic process .
In summary, differential gene expression analysis was performed at various times following asbestos exposure in SPARC-null or wild type mice to assess the role of SPARC in the fibrotic process. Tens of genes (61) were identified that vary in their expression at least 2 fold at a significance level of <0.01 between SPARC-null and wild-type mice without asbestos exposure, and additional transcripts were identified that may play a role in SPARC modulation of the fibrotic response to asbestos or other stressors in the lung. This report is a companion to another report [16, A. Smartt et al., manuscript submitted] in which fibrosis was directly quantified over this time frame. Thus the biological endpoint of fibrosis and its severity can be correlated with the changes in gene transcription. Following validation of key transcripts by alternative methodology, an informative model of SPARC-mediated modulation of fibrosis can be described. In order to explain the remodeling or reduction in fibrosis present in SPARC-null mice, it is important to have a broad picture of the molecular changes that are or are not taking place in this tissue. The finding of fibrosis resolution is rare in the literature, and considering the absence of effective therapies for fibrosis in humans exposed to asbestos these findings can be very useful in understanding the molecular mechanisms taking place in these mice, and potentially can be exploited for therapeutic advantage or increased diagnostic advantage.
Crocidolite asbestos was obtained from the Research Triangle Institute (Research Triangle Park, NC). Samples were freshly prepared in sterile phosphate-buffered saline (PBS, pH 7.4) and sonicated before instillation.
All animal protocols were approved by the University of Montana Institutional Animal Care and Use Committee. Mice were exposed to asbestos according to methods previously described . Briefly, pathogen-free 6-8 week old female and male C57Bl/6 WT mice and SPARC-null mice on a C57BL/6 background (Sparctm1Hwe)  were instilled intratracheally with 100 mg of crocidolite asbestos in 30ml sterile PBS under anesthesia (n=8-10 for each condition/time point). Control mice received only sterile PBS. Mice were euthanized 1 week, 1 month, or 3 months after instillation. The left lung of the mice was removed for RNA isolation.
Lung tissue samples were homogenized in 1ml of TRIZOL and RNA isolated following the manufacturer’s protocol (Invitrogen, Carlsbad CA). The resulting RNA was purified using the RNeasy kit (Qiagen, Valencia, CA) and subsequently treated with DNAse (Qiagen).
For microarray analysis, reverse transcription and cDNA labeling were performed using 8 μg of total RNA according to the Superscript Plus Direct cDNA Labeling Kit (Invitrogen). Labeling efficiency was determined using a NanoDrop Spectrophotometer (NanoDrop Technologies, Wilmington, DE). Labeled cDNA was hybridized overnight at 42°C to in-house arrays of oligonucleotides spotted on epoxysilane coated microscope slides (Erie Scientific, Thermo Fisher Scientific, Waltham, MA). The oligonucleotide set was produced by MWG Biotech, Inc. (Cork, Ireland), and is composed of 50-base oligos specific to 9853 unique mouse genes. Almost all genes have a known function or clearly defined protein domains. Hybridized arrays were scanned and analyzed using an Axon GenePix 4000B microarray scanner (Molecular Devices, Sunnyvale, CA) and Axon GenePix Pro 5.1 software (Molecular Devices). Normalization and flagging was performed using an R script courtesy of Dr. Terry Speed University of California-Berkeley. Replicate analyses and comparisons were performed using a combination of custom PHP/MySQL programs and standard Excel spreadsheets. Further analyses were performed using GoMiner (NCBI), and PathwayStudio (Ariadne Genomics, Rockville, MD).
For quantitative RT-PCR analysis, a starting amount of 500ng total RNA was reverse transcribed using the Transcriptor First Strand cDNA Synthesis Kit (Roche, Indianapolis, IN) for each time point and condition. Labeled probes and primer sets for each gene were designed using the Universal ProbeLibrary system (Roche) and are listed in Table 7. Real-time PCR reactions were performed on a Bio-Rad iQ5 instrument (Hercules, CA). The PCR program utilized was: One cycle of 95°C for 10 min. followed by 40 cycles of 95° C for 15 sec. and 60°C for 1 min. Resulting data were normalized to a β-actin control using the ΔΔCt method.
The authors acknowledge E. Helene Sage (Bennaroya Research Institute at Virginia Mason) for the kind gift of the SPARC-null mice. Support towards the publication of this manuscript came from the Montana NSF EPSCoR grant: EPS-03464558 and the State of Montana MBRCT grant: Agreement #07-04 (2004-2007) (MP and EAP); and the CDC through a subproject under grant number CCR822092 (EP). This publication was also made possible by both subproject (MP and EAP) and core facility (MP) support from Grant Number P20RR017670 and core facility support (MP) from Grant Number P20RR015583, both from the National Center for Research Resources (NCRR), a component of the National Institutes of Health (NIH). The views expressed in this publication do not necessarily reflect the official policies of the Department of Health and Human Services; nor does mention by trade names, commercial practices, or organizations imply endorsement by the U.S. Government.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.