Gene Expression Studies Reveal That DNA Damage,Vascular Perturbation,and Inflammation Contribute to the Pathogenesis of Carbonyl Sulfide Neurotoxicity

Abstract

Carbonyl sulfide (COS) is an odorless gas that produces highly reproducible lesions in the central nervous system. In the present study, the time course for the development of the neurotoxicological lesions was defined and the gene expression changes occurring in the posterior colliculus upon exposure to COS were characterized. Fischer 344 rats were exposed to 0 or 500 ppm COS for one, two, three, four, five, eight, or ten days, six hours per day. On days 1 and 2, no morphological changes were detected; on day 3, 10/10 (100%) rats had necrosis in the posterior colliculi; and on day 4 and later, necrosis was observed in numerous areas of the brain. Important gene expression changes occurring in the posterior colliculi after one or two days of COS exposure that were predictive of the subsequent morphological findings included up-regulation of genes associated with DNA damage and G1/S checkpoint regulation (KLF4, BTG2, GADD45g), apoptosis (TGM2, GADD45g, RIPK3), and vascular mediators (ADAMTS, CTGF, CYR61, VEGFC). Proinflammatory mediators (CCL2, CEBPD) were up-regulated prior to increases in expression of the astrocytic marker GFAP and macrophage marker CSF2rb1. These gene expression findings were predictive of later CNS lesions caused by COS exposure and serve as a model for future investigations into the mechanisms of disease in the central nervous system.

Introduction

Carbonyl sulfide (COS) is a low-molecular-weight gas originating from a variety of natural and anthropogenic sources. The total worldwide production and release into the atmosphere is estimated to exceed 9500 tons annually. The man-made sources are diverse and include many industrial and combustive processes, such as viscous rayon production, thiocarbamate herbicide production, petroleum and rubber product manufacture, automobile exhaust, and cigarette smoke. Based on the high annual production, the relative lack of toxicity data and estimated persistence of the gas in the atmosphere for about two years, COS has been classified as a high priority Clean Air Act chemical by the United States Environmental Protection Agency (U.S. EPA 1990).

Previous studies by our group characterized the morphologic features of the highly specific and reproducible neurotoxicity that occurs upon repeated exposure to COS in rats. In these previous experiments, the posterior colliculi and parietal cortex were the most consistently affected areas, with necrosis in the posterior colliculus occurring in 100% of exposed animals. Other brain regions affected less consistently included the putamen, thalamus, anterior olivary nucleus, nucleus of the lateral lemniscus, red nucleus, hippocampus CA1 region, and retrosplenial cortex. The necrosis was typically bilaterally symmetrical. The posterior colliculus, anterior olivary nucleus, and the nucleus of the lateral lemniscus are all brain regions involved in the auditory system, indicating that the system is specifically targeted by COS exposure. Consistent with these histologic findings, functional abnormalities in the auditory system, reflected in altered brain stem auditory evoked responses (BAER), were present in exposed rats (Morgan et al. 2004; Sills et al.; Herr et al. 2007; Sills et al. 2004/2005).

Although interest in COS neurotoxicity has increased considerably in recent years, there is still little known about the mechanism of COS toxicity. The goals of the current study were two-fold: (1) to further characterize the morphological lesions so as to refine the time course of their development and develop a model for mechanistic studies, and (2) to then use this model to identify and characterize the transcriptional responses and cellular pathways whose activation or perturbation resulted in the development of the described morphological changes. The various pathways that were identified, including DNA damage, apoptosis, and inflammation, among others, provide insight into the mechanism of toxicity, which culminates in the characteristic neurodegeneration after exposure to COS.

Materials and Methods

Chemical

Carbonyl sulfide (CAS# 463-58-1) was purchased from Tex-La Gases (Houston, TX, USA). Carbonyl sulfide was procured as a liquid in gas cylinders equipped with valves configured to provide the vapor phase of COS. Chemical purity was determined to be >98.1% by gas chromatography/thermal conductivity analysis of the vapor phase over the liquid COS. The 1.9% residual was composed primarily of CO₂ with <0.6% H₂S. After dilution of the bulk gas to the desired COS exposure concentrations, the H₂S concentrations were reduced to insignificant levels. No H₂S could be detected in samples collected from the chambers during COS exposures.

Inhalation Exposure

The COS vapor, at reduced pressure, was supplied to pumps that controlled the vapor flow necessary to achieve the desired concentration. The COS vapor was mixed with conditioned air (HEPA filtered, charcoal scrubbed, temperature and humidity controlled) and delivered to the Hazleton 2000 exposure chambers at approximately 400 L min⁻¹. Carbonyl sulfide in each chamber was sampled at ninety-second intervals and measured using Orbital Scientific Model Diamond 20 Fourier transform infrared (FTIR) spectrophotometers. Exhaust from the exposure chambers was passed through activated charcoal scrubbers (Safe-mod model CA-500) connected in series. The scrubbed exhaust was sampled about every fifteen minutes and analyzed for COS. The concentrations of COS in the exposure chambers were independently verified before and during the animal exposures. Chamber samples were analyzed by gas chromatography (GC) with photo-ionization and GC-thermal conductivity detectors.

Animals

Male and female Fischer 344 rats (Charles River Laboratories, Raleigh, NC, USA), six or seven weeks old on arrival, were held for ten to fourteen days to provide time to assess their overall health and to acclimate them to the exposure facility. During the holding period, rats were weighed and randomized to treatment groups. Animals were placed in exposure chambers without food six hours per day for two days before chemical exposure for acclimation to the exposure conditions. Animals were eight or nine weeks old at the start of the exposures. Animals were individually housed in Hazleton 2000 inhalation exposure chambers for the duration of the study. The chambers were contained in a humidity and temperature-controlled, HEPA-filtered, mass air displacement room in facilities accredited by the American Association for Accreditation of Laboratory Animal Care. Animal rooms were maintained with a light–dark cycle of twelve hours (light from 7:00 AM to 7:00 PM). Sentinel animals, housed in the animal facility as part of an ongoing surveillance program for parasitic, bacterial, and viral infections, were specific pathogen-free throughout the study. Feed (NIH-07) was removed during the six-hour exposures and for six hours per day on nonexposure days. Water (chlorinated, City of Durham) was provided ad libitum by an automatic watering system during the nonexposure as well as the exposure periods. Individual body weights were recorded for all animals on the day before the first exposure. Rats were observed for clinical signs of toxicity twice daily (early morning and late afternoon). Individual animal clinical observations were documented for all rats at the time of weighing.

This study was conducted under federal guidelines for the use and care of laboratory animals and was approved by the NIEHS Animal Care and Use Committee.

Development of a Model for Gene Expression Studies

Before the gene expression studies were conducted, a time course study was performed to determine when the initial lesions of COS neurotoxicity occurred.

Ten male F344 rats per group were exposed by inhalation to 500 ppm of COS, six hours per day for one, two, three, or four consecutive days. Control animals, also housed within exposure chambers, breathed filtered, conditioned air. After the last inhalation exposure, rats were anesthetized with intraperitoneal sodium pentobarbital and perfused via left ventricular cannulation with McDowell-Trump fixative (McDowell and Trump 1976). After further immersion fixation with the same fixative, brains were removed and sectioned at the following levels: (1) frontoparietal cortex immediately anterior to the optic chiasm; (2) parietal cortex through the hippocampus and infundibulum; (3) mesencephalon through the anterior colliculi; (4) mesencephalon through the posterior colliculi; (5) mid-cerebellum; and (6) posterior medulla through the obex.

Tissues were processed and embedded in paraffin following standard protocols, and paraffin sections were stained with hematoxylin and eosin using standard procedures. All available areas of brain were examined, but the neuropathological findings for twenty-one specific brain regions were recorded based on the findings from previous studies (Morgan et al. 2004; Sills et al. 2004).

Neuronal Degeneration and GFAP Assessment

For degeneration and glial fibrillary acidic protein (GFAP) assessment, five male F344 rats per group were exposed to COS, six hours per day for one, three, five, eight, and ten days. Following euthanasia via an overdose of pentobarbital, rats were perfused through the left ventricle with a solution of 0.8% NaCl and 0.8% sucrose in cacodylate buffer (one minute), followed by a fixative solution of 4% paraformaldehyde and 4% sucrose in the same buffer (eight to ten minutes). Brains were then immersion-fixed in situ for two to four days in the same fixative. After removal, brains were cryoprotected in a glycerol-based solution for six to eight hours. Whole brains were embedded collectively in gelatin blocks (fifteen brains/ block) with parallel alignment of their rostrocaudal axes for coronal sectioning (Multibrain Technology, NeuroScience Associates, Knoxville, TN, USA). Gelatin blocks were hardened in a net 4% formaldehyde solution for forty-eight hours, then frozen in a slurry of dry ice in 2-methyl butane. With an American Optical 860 sliding microtome, 40 μm composite freeze-cut sections were cut throughout the brain. A serial set comprising every eighth section was stained with the amino cupric silver method specific for neuronal degeneration (de Olmos et al. 1994) and by GFAP immunohistochemistry for astrocytes (Fix et al. 1996).

RNA Isolation

There were three animals per group at each time point. Immediately after the cessation of COS exposure on the appropriate day of sacrifice (day 1 or day 2) the rats were anesthetized with intraperitoneal sodium pentobarbital. The posterior colliculi were surgically exposed and removed using a pair of corneal scissors and fine forceps. The samples were immediately placed in RNase-free plastic vials and snap-frozen in liquid nitrogen. Three control animals (nonexposed) were processed in the same manner.

Total RNAs from frozen brain tissue (posterior colliculus) were isolated with an RNeasy tissue kit (Qiagen, Valencia, CA, USA) according to the manufacturer’s protocols. The quality and integrity of the RNA was verified by 260/280 nm absorbance ratio, formaldehyde agarose gel electrophoresis, and an Agilent Bioanalyzer (Agilent Technologies, Palo Alto, CA, USA). Since the posterior colliculi were removed en masse for analysis, the total RNA sample contained a mixture of RNA of all the different cell types present in the tissue, including neurons, glia, endothelial cells, pericytes, and potentially others.

Reverse Transcription

To remove genomic DNA, RNA samples were incubated with 1 U of RNase-free deoxyribonuclease I (DNase I; Invitrogen Corp., Carlsbad, CA, USA) per microgram of RNA for fifteen minutes at room temperature. DNase I was then inactivated by the addition of 2.5 mM EDTA (pH 8.0) and heated at 65°C for ten minutes. Reverse transcription of RNA (1 μg) using MuLV reverse transcriptase (Applied Biosystems, Foster City, CA, USA) was carried out according to the manufacturer’s instructions using oligo-deoxythymidine primers (Invitrogen). As a negative control, a sample containing RNA but no reverse transcriptase was also included. A 1:10 dilution with DNase-free water of this cDNA was used for real-time polymerase chain reaction (PCR) analysis in ninety-six-well plates.

Microarray Analysis

Gene expression analysis was conducted using Agilent Rat Oligo Microarrays containing approximately 22,000 probes (Agilent Technologies, Palo Alto, CA, USA). Total RNA was amplified using the Agilent Low RNA Input Fluorescent Linear Amplification Kit protocol. Starting with 500 ng of total RNA, Cy3 or Cy5 labeled cRNA was produced according to the manufacturer’s protocol. For each two-color comparison, 750 ng of each Cy3 and Cy5 labeled cRNAs were mixed and fragmented using the Agilent In Situ Hybridization Kit protocol. Hybridizations were performed for seventeen hours in a rotating hybridization oven using the Agilent 60-mer oligo microarray processing protocol. Slides were washed and then scanned with an Agilent Scanner. Two arrays were used for each comparison allowing for dye reversals. Data were obtained using the Agilent Feature Extraction software, version 8.1, using defaults for all parameters and loaded into the Rosetta Resolver system, version 6.0 (Rosetta Biosoftware, Kirkland, WA). The Resolver system combines ratio profiles to create ratio experiments using an error-weighted average, as described in Weng et al. (2006). P values are generated and propagated throughout the system and represent the probability that a given gene is significantly expressed. Data were expressed as a fold change in the expression level compared to the control samples and an associated p value. Only those transcripts with a positive or negative ≥1.3-fold change and a p value ≤.01 were selected for further analysis.

The microarray data discussed in this publication have been deposited in NCBI’s Gene Expression Omnibus (Edgar et al. 2002) and are accessible through GEO Series accession number GSE12018 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE12018).

Real-time PCR Analysis

Quantitative gene expression levels were detected using real-time PCR with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA) and Taq-Man MGB probes (FAM dye labeled). Primers and probes for all genes analyzed were purchased from Applied Biosystems Assays-on-Demand Gene Expression products. For amplification, diluted cDNA was combined with a reaction mixture containing TaqMan universal PCR Master Mix (Catalog No. 4304437, Applied Biosystems, Foster City, CA, USA) according to the manufacturer’s instructions. Samples were analyzed in duplicate, and a sample without reverse transcriptase was included with each plate to detect contamination by genomic DNA. Amplification was carried out as follows: (1) 50°C, two minutes (for uracil-N-glycosylase incubation); (2) 95°C, ten minutes (denaturation); and (3) 95°C, fifteen seconds, 60°C, thirty seconds (denaturation/amplification) for forty cycles. Fold increases or decreases in gene expression were determined by quantification of cDNA from target samples relative to a calibrator sample of respective controls on day 1 and day 2. The 18S RNA gene was used as the endogenous control for normalization of initial RNA levels. To determine this normalized value, 2-(Δ ΔCt) values were compared between target and calibrator samples, where the changes in crossing threshold (ΔCt) = CtTarget gene - Ct18S RNA, and Δ ΔCt = ΔCtcontrol - ΔCttarget.

Results

Model for Gene Expression Studies

In male rats exposed to 500 ppm COS for six hours on days 1 and 2, no morphological lesions were evident in any areas of the brain examined. However, in male rats exposed to 500 ppm COS for six hours per day for three days, bilaterally symmetrical necrosis was present in the posterior colliculi in 100% of rats (Table 1 and Figure 1). Endothelial cell hypertrophy and a variable degree of acute parenchymal hemorrhage in the posterior colliculi were present in a subset of animals. In rats exposed to 500 ppm COS for four days or longer, the anatomic distribution of the severe lesions was much more diverse, with necrosis in the frontoparietal cortex, retrosplenial cortex, thalamus, hippocampus region CA1, red nucleus, nucleus of the lateral lemniscus, anterior olivary nucleus, and the putamen, in addition to the posterior colliculi (Table 1 and Figure 2).

The lack of morphological lesions on days 1 and 2 of exposure and bilaterally symmetrical necrosis in all animals on day 3 of exposure in the posterior colliculi provided an opportunity to focus our gene expression studies on the posterior colliculi to determine molecular pathways that may be predictive of acute lesions observed on day 3 of COS exposure. Gene expression analysis was performed on the posterior colliculi on days 1 and 2 of exposure, prior to the onset of necrosis.

Neuronal Degeneration and GFAP Assessment

On day 5 there was a striking astrocytic response, as indicated by IHC GFAP positivity, and, by days 8 and 10, the astrocytic response was primarily at the periphery of the area of necrosis in the posterior colliculi (Figure 3A).

Small amounts of silver-impregnated debris indicative of neuronal degeneration were detected in the posterior colliculi beginning on day 5 in a subset of the animals examined. On day 8, a large amount of silver-stained debris was present in the posterior colliculi in all animals examined (Figure 3B), and on day 10, silver-stained debris was present in the medial geniculate nucleus, parietal cortex, and caudate/putamen in addition to the posterior colliculi.

Oligo Microarray Analysis

Statistical analysis of the gene expression values revealed a total of 195 transcripts in the day 1 exposure sample and 215 transcripts in the day 2 exposure sample that met the criteria for further analysis (fold change ≥ 1.3, p ≤ .01). These transcripts represented a total of seventy-nine annotated genes from the day 1 exposure sample, and 104 annotated genes from the day 2 exposure sample, of which sixty of seventy-nine and seventy-three of 104 were up-regulated and nineteen of seventy-nine and thirty-one of 104 were down-regulated, respectively. See Supplemental Data Tables 1 and 2 for the list of annotated genes from each time point that were dysregulated. A fair number of altered transcripts were related to DNA damage, apoptosis, vascular mediators, inflammation, heat shock proteins, and neurotransmission (Figure 4). The cellular heterogeneity present in the tissue was reflected in the gene expression changes. To address our hypothesis that COS neurotoxicity is caused by a complex mechanism including DNA damage, vascular mediators, disruption in energy metabolism, and inflammation, we focused our gene expression studies in these areas.

Validation of Oligo Microarray Analysis

Validation of the microarray results was performed using quantitative RT-PCR on a subset of the genes identified. Although there was often variation in the magnitude of the fold change, there was generally good correlation in directionality between the two techniques, particularly with transcripts with a higher fold change (Table 2). There was less agreement between the techniques for genes with a low fold change and for genes at the first time point, which may reflect innate differences between individual animals, heterogeneity in the early response to COS exposure, the small number of animals at each time point, or differences in the gene region amplified by the RT-PCR primers and the gene sequence on the array. The correlation increased for many of the genes by the second time point.

Functional Classification of Differentially Expressed Genes

Gene ontologies and the cellular pathways to which the genes belonged were determined using a combination of Rosetta Resolver, Ingenuity Pathways Analysis software and manual literature and database searches, including Online Mendelian Inheritance in Man (OMIM), PubMed, GOPubMed, and the University of California–Santa Cruz Genome Database. The functional classification of the annotated genes, their fold change, and their accession numbers are summarized in Supplemental Tables 1 and 2. Briefly, the most common molecular functions of the differentially expressed genes include transcription factor or transcription regulator activity, cell cycle regulation, signal transduction, inflammation, cellular metabolism, and growth factors. In addition, several other categories that contained smaller numbers of altered genes, but that are potentially significant in their functional implications, were present, including angiogenesis, apoptosis, and coagulation.

The Ingenuity Pathways gene networks that contained the largest numbers of focus genes and had the highest Ingenuity relevance scores were: cellular assembly and organization, cell growth and proliferation, cell movement, cell-to-cell signaling, cell morphology, cell development, cell death, gene expression, DNA replication and recombination, and small molecule biochemistry. The gene networks identified at the two time points were very similar (data not shown); however, more focus genes were identified at the second time point and the cell death gene network received a higher rank.

Discussion

Carbonyl sulfide–induced necrosis occurs as early as three days following exposure and, in the posterior colliculi, has a 100% incidence in rats exposed to 500 ppm for three days, six hours per day. Gene expression profiling on the posterior colliculi of exposed animals before the onset of morphological changes (days 1 and 2) provides evidence for a complex pathogenetic mechanism involving DNA damage, vascular alteration, and inflammation. These changes may represent the most significant initial events that are linked to cell injury in the posterior colliculi and other neuroanatomical sites following COS exposure.

The analysis of the gene expression data identified clusters of differentially regulated genes that collectively implicate certain factors and pathways that are likely playing a role in the tissue damage. Two transcription factors that were strongly up-regulated were Kruppel-like factor 4 (KLF4) and B-cell translocation gene 2 (BTG2). Both of these genes are induced by or involved in the regulation of DNA damage repair and/or the G1/S phase cell cycle transition. Kruppel-like factor 4 is an upstream regulator of p53 that, through its transcription targets such as p21, regulates cell cycle progression and cellular differentiation (Rowland and Peeper 2006; Yoon et al. 2003; Zhang et al. 2000). B-cell translocation gene 2 is a member of the BTG/Tob family that is p53-inducible during genotoxic stress and has strong antiproliferative properties. B-cell translocation gene 2 has also been implicated in neuronal differentiation and may play a role in increasing neuronal survival during periods of stress (Corrente et al. 2002; el-Ghissassi et al. 2002; Rouault et al. 1996; Tirone 2001). Two other genes that were up-regulated implied activation of p53, including Tumor Protein p53 Inducible Nuclear Protein 1 (TP53INP1) on day 1 of COS exposure and Ubiquitin Specific Peptidase 7 (USP7) on day 2 of COS exposure. Tumor Protein p53 Inducible Nuclear Protein 1 has been reported to be p53 dependent and to regulate P53 dependent apoptosis (Okamura et al. 2001). Ubiquitin Specific Peptidase 7 de-ubiquitinates p53, thereby promoting its persistence and activation (Cummins and Vogelstein 2004; Li et al. 2002).

Also supporting a role for DNA damage in COS lesion pathogenesis is the strong up-regulation of Growth Arrest and DNA Damage-inducible 45, gamma (GADD45g), at the second time point. Growth Arrest and DNA Damage-inducible 45, gamma is a sensitive, but not specific, marker of DNA damage caused by a variety of insults, including exposure to methyl methanesulfonate (MMS) or UV radiation, and is an important regulator of cell survival (Jung et al. 2000; Zhang et al. 1999). Recently, GADD45g has also been recognized as being up-regulated in response to other stimuli, such as pro-inflammatory cytokines (Jung et al. 2000). The above gene expression changes suggest that DNA damage with p53 activation, caused by COS either directly or indirectly (e.g., inflammation, ischemia), is playing a role in the pathogenesis of the lesions.

Several pro-apoptotic genes were up-regulated at one or both time points including several genes mentioned above in reference to DNA damage, such as GADD45g, USP7, and TP53INP1, and also Transglutaminase 2 (TGM2) and Receptor Interacting Serine Threonine Kinase 3 (RIPK3). Transglutaminase 2 is an enzyme responsible for the cross-linking of proteins at glutamine residues and is a potent effector protein of apoptosis (Fesus et al. 1987; Piacentini et al. 1991). It was modestly up-regulated at the first time point but strongly up-regulated at the second time point. Ubiquitin Specific Peptidase 7 de-ubiquitinates proteins, most notably p53, thereby stabilizing it and indirectly promoting apoptosis (Cummins and Vogelstein 2004; Li et al. 2002). Receptor Interacting Serine Threonine Kinase 3 is a serine-threonine kinase that associates with Tumor Necrosis Factor (TNF) Receptor 1 (TNFR1) and TNF Receptor Associated Factor 2 (TRAF2) and can induce apoptosis via a region near its C-terminus after appropriate stimulation (Kasof et al. 2000; Sun et al. 1999). The above gene expression changes are consistent with activation of the apoptosis cascade. Morphologically, the COS lesions had been previously categorized as necrosis (Morgan et al. 2004); however, it is now widely recognized that apoptosis and necrosis represent the ends of a continuum and that many factors influence the biochemical and morphological characteristics of a given lesion, including cell energy status and the prior biochemical status of the cell (Martin 2001; Nicotera et al. 1998; Nicotera et al. 2000).

Heat shock proteins are fundamental to maintaining cellular homeostasis in response to a variety of cellular insults, including thermal stress, oxidative stress, and toxic exposure. They function as molecular chaperones that control protein folding and prevent the aggregation of proteins. Overexpression of heat shock proteins is strongly neuroprotective in several animal models of neurological disease (Cummings et al. 2001; Shen et al. 2005). In this study, the expression level of Heat Shock Protein 70 kDa (HSPA1A) was altered at both time points, displaying a decrease in expression level on day 1 and a strong increase in expression level on day 2. The increase in heat shock protein expression on day 2 is an expected response to DNA damage and apoptosis on day 1 of COS exposure.

At one or both time points, several genes were differentially expressed that are important regulators of endothelial function, angiogenesis, and coagulation, collectively implicating these pathways in COS lesion pathogenesis. These genes include connective tissue growth factor (CTGF), cysteine-rich angiogenic inducer 61 (CYR61), vascular endothelial growth factor C (VEGFC), thrombomodulin (THBD), a disintegrin and metallopeptidase with thrombospondin type 1 motif 1 (ADAMTS1), and platelet factor 4 (PF4).

Connective tissue growth factor and CYR61 are strong pro-angiogenic extracellular matrix (ECM)-associated members of the CCN protein family and have numerous effects on endothelial cells, including cell growth, proliferation, adhesion, and migration (Brigstock 2002). Vascular endothelial growth factor C, like CTGF and CYR61, is pro-angiogenic, stimulates endothelial cell growth, and modulates vascular permeability (Joukov et al. 1996). Thrombospondin type 1 motif 1 is a secreted endopeptidase involved in ECM remodeling and in the inhibition of angiogenesis. It is normally expressed at low levels in the central nervous system (CNS) and has recently been recognized to be up-regulated in response to cerebral ischemia (Cross et al. 2006). Thrombomodulin is an endothelial-specific receptor that has varied effects on several related pathways, including potent inhibition of the coagulation system, activation of the fibrinolytic system, and suppression of inflammation. Its expression on endothelial cells is regulated by a variety of pro-inflammatory and coagulation factors, such as histamine, VEGF, and thrombin (Van de Wouwer and Conway 2004). Based on the above gene expression changes, it is clear that concomitant with the initial DNA damage, there is a striking vascular response. Whether this response is directly related to COS exposure or is secondary to the tissue damage is not clear.

In addition to its role in angiogenesis, CTGF, as the name suggests, has been strongly implicated in the production of ECM and is commonly up-regulated during wound repair in many tissues. Recently, it has been recognized in the CNS as one of the earliest markers of astrocyte activation, detectable in the cytoplasm within hours of the onset of damage, and it has been implicated as being an important component of CNS damage repair and glial scarring in response to several types of injury, including traumatic, ischemic (Conrad et al. 2005; Schwab et al. 2000) and, now, toxic forms of damage. Consistent with the up-regulation of CTGF, additional evidence of astrocyte activation was present through the up-regulation of glial fibrillary acidic protein (GFAP) on day 2 of COS exposure. Glial fibrillary acidic protein is an intermediate filament protein whose expression is classically up-regulated in response to a variety of insults. Morphological evidence of astrocytic activation with increased GFAP expression was present beginning on day 5 of the study, consistent with the gene expression changes.

After the initial cellular damage due to COS exposure, a significant inflammatory cell response develops within days that is predominantly composed of mixed mononuclear cells, presumably resulting from a combination of resident microglial cell activation and inflammatory cell recruitment from the blood. Inflammation, although necessary for the clearance of damaging agents and cellular debris, can also contribute to or exacerbate neuronal damage and ultimately, neuronal cell loss. Having a more complete understanding of the gene expression changes underlying inflammation resulting from a variety of different causes may lead to identification of new targets for therapeutic intervention. The gene expression data in this study contain several up-regulated genes that are likely playing a role in the initiation of the inflammatory response to COS-induced damage.

Zinc Finger Protein 36 (ZFP36) and CCAAT/Enhancer Binding Protein delta (CEBPD), two transcription factors/regulators that are progressively up-regulated at both time points, are important modulators of inflammatory responses in a variety of tissues. Zinc Finger Protein 36 is a post-transcriptional regulator protein shown to modulate the expression of various pro-inflammatory proteins including Tumor Necrosis Factor-α (TNF-α) and Interleukin-2 (IL-2). It accomplishes this task through binding to an AU-rich repeat at the 3′-untranslated region and thereby inducing mRNA degradation. It is increasingly recognized as an important mechanism of modulating the intensity of an inflammatory response (Anderson et al. 2004; Ogilvie et al. 2005).

CCAAT/Enhancer Binding Protein delta is a transcription factor expressed in a variety of tissues that has a primary role in inflammatory responses. It has been implicated in the activation and differentiation of macrophages in response to various stimuli, and possibly also microglial cells, since they are of the macrophage lineage. Up-regulation of CEBPD has also been reported in astrocytes in response to various acute and chronic stimuli, including pro-inflammatory cytokines and β-amyloid plaques in Alzheimer’s disease patients (Cardinaux et al. 2000; Li et al. 2004). Our study provides further evidence that CEBPD is an important component of the neuroinflammatory response, resulting from a variety of insults.

Several other pro-inflammatory genes were up-regulated at one or both time points, including Chemokine (C-C Motif) Ligand 2 (CCL2), Colony Stimulating Factor 2 receptor beta 1 (CSF2RB1), Alpha 2 Macroglobulin, CD32, Intercellular Adhesion Molecule 2 (ICAM2), and several complement components (C1qB, C1qR1, and C4-2), among others.

Also known as CCL13 or Macrophage Chemoattractant Protein 1 and 4, CCL2 is a member of the β-chemokine (C-C) family. Whereas members of this family of proteins are principally involved in the recruitment and chemoattraction of monocytes and macrophages, effects on apoptosis, the cell cycle, and the production of soluble inflammatory products, such as free radicals, nitric oxide, and cytokines, have also been demonstrated (Ambrosini and Aloisi 2004; Ubogu et al. 2006). β-Chemokines can be produced by a variety of resident cell types in the CNS, including neurons, astrocytes, and microglia and have been increasingly implicated in playing important roles in neuroinflammatory responses, in particular the autoimmune conditions multiple sclerosis in humans and its widely studied animal model, experimental autoimmune encephalomyelitis (Banisor et al. 2005). The strong up-regulation in this study is consistent with a broad role for this class of proteins in responding to CNS damage, and it is likely that CCL2 is providing an early signal for inflammatory cell recruitment and activation in the damaged areas of the brain.

Colony Stimulating Factor 2 receptor beta 1 is the beta chain of the granulocyte macrophage colony-stimulating factor (GM-CSF) receptor. Granulocyte macrophage colony-stimulating factor is a potent activator of microglial cells and has been implicated in increasing neuronal survival (Meuer et al. 2006; Schermer and Humpel 2002). The up-regulation of CSF2RB1 is consistent with the marked microglial cell response observed at later time points after COS exposure (Morgan et al. 2004).

Alpha 2 Macroglobulin (A2M) is a plasma pan-proteinase inhibitor that, as an acute phase protein, is commonly up-regulated in response to various pro-inflammatory stimuli. In addition to the well-described effects on inflammation and the coagulation/fibrinolytic systems, neuroprotective properties on CNS neurons by the activated form of the protein have recently been described. The proposed mechanism of action is via decreasing the calcium influx into the cell that results from stimulation of the NMDA receptor, potentially decreasing the likelihood of excitotoxic neural damage (Qiu et al. 2002). In this study, the up-regulation of A2M is most likely associated with the inflammatory and vascular responses with a further protective effect on the surviving neuronal population.

Collectively, our gene expression data provided insight into the potential mechanism of and responses to the cellular damage due to COS exposure and identified a number of cellular pathways that predict the neuronal degeneration, vascular response, inflammatory changes, and gliosis subsequently present in the tissue. Most notably, several genes involved in the response to DNA damage and G1/S checkpoint regulation are up-regulated compared to controls, including the transcription factors KLF4 and BTG2, the p53 stabilizer USP7, the p53-dependent gene TP53INP1, and the pro-apoptotic gene GADD45g. Up-regulation of these genes suggests that DNA damage is playing a role in the pathogenesis of COS neurotoxicity. Other important pathways that have been perturbed by exposure to COS implicate apoptosis, vascular response, and inflammation, which fit with the morphological changes that subsequently develop as a result of exposure.

Other hypotheses have been proposed for carbonyl sulfide neurotoxicity (Morgan et al. 2004; Nutt et al. 1996; Sills et al. 2005). Previous studies by our group using a range of COS concentrations in rats have shown decreases in mitochondrial cytochrome oxidase c activity in the posterior colliculus at time points of three weeks or more of exposure to COS, suggesting that inhibition of oxidative phosphorylation may also contribute to neuronal cell death. In our gene expression studies, all genes related to mitochondrial function including cytochrome oxidase c were evaluated, and there was no difference in expression levels between controls and COS-exposed animals. This finding suggests that decreases in cytochrome oxidase c activity may be a later response to COS neurotoxicity.

This is the first study characterizing the gene expression changes in the brain following exposure to the neurotoxicant COS and contributes a great deal of information regarding the cellular pathways that are activated in response to this form of toxic damage. As the gene expression changes in the central nervous system in response to various insults have begun to be characterized, considerable overlap in gene expression patterns has been recognized resulting from diverse injuries. This finding is consistent with the CNS being capable of a limited range of potential responses and suggests that through careful study of these patterns, critical cellular pathways and effectors can be identified to increase our overall understanding of CNS disease pathogenesis.