Toxicogenomic Evaluation of Long-term Hepatic Effects of TCDD in Immature, Ovariectomized C57BL/6 Mice

Abstract

Acute exposure to hepatotoxic doses of 2,3,7,8-tetrachloro- dibenzo-p-dioxin (TCDD) in mice is characterized by differential gene expression that can be phenotypically anchored to elevated levels of serum alanine aminotransferase, increased relative liver weights, hepatic steatosis, inflammation, and hepatocellular necrosis. Unlike most studies that focus on acute exposure effects, this study evaluated the long-term effects of a single oral gavage of 30 μg/kg TCDD at 1, 4, 12, 24, 36, and 72 weeks postdose in ovariectomized C57BL/6 mice. Hepatic TCDD levels were almost completely eliminated by 24 weeks with a calculated half-life of 12 days. Hepatic gene expression analysis identified 395 unique differentially expressed genes between 1 and 12 weeks that decreased to ≤ 8 by 72 weeks, consistent with the minimal hepatic TCDD levels. Hepatic vacuolization, characteristic of short-term exposure, subsided by 4 weeks. Similarly, TCDD-elicited hepatic necrosis and inflammation dissipated by 1 week. Collectively, these results suggest that TCDD-elicited histologic and gene expression responses can be correlated to elevated hepatic TCDD levels, which, once eliminated, elicit minimal hepatic gene expression and histologic alterations.

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) and related compounds are ubiquitous environmental contaminants that elicit a broad spectrum of responses in a cell-, tissue-, sex-, age-, and species-specific manner (Poland and Knutson, 1982). These responses include wasting syndrome, tumor promotion, teratogenesis, immunotoxicity, modulation of endocrine systems, and hepatotoxicity, the majority of which are mediated by the aryl hydrocarbon receptor (AhR), a member of the basic-helix-loop-helix-PAS (bHLH-PAS) transcription factor family (Denison and Heath-Pagliuso, 1998; Poland and Knutson, 1982). Ligand binding to the cytoplasmic AhR leads to translocation to the nucleus where it forms a heterodimer with the AhR nuclear translocator (ARNT), another bHLH-PAS family member. This heterodimer binds specific DNA elements, termed dioxin response elements (DREs) in the regulatory regions of target genes leading to changes in gene expression (Denison et al., 2002; Hankinson, 1995).

Other studies also suggest that the AhR plays a functional role in a variety of cellular processes through nongenomic mechanisms (Barhoover et al., 2010; Tappenden et al., 2011; Tian et al., 2002). For example, the AhR has been shown to bind DNA and elicit differential gene expression in a DRE-independent manner (Beischlag et al., 2008; Denison et al., 2011; Dere et al., 2011; Huang and Elferink, 2012; Tanos et al., 2012). Moreover, the AhR localizes to the mitochondria, where it plays a functional role in energy metabolism (Tappenden et al., 2011). The AhR also regulates cellular inflammatory responses through modulation of intracellular calcium concentration and activation of cytosolic phospholipase A2 and SRC kinase activity without heterodimerization with ARNT (Dong et al., 2010; Matsumura, 2009).

The role of AhR/ARNT signaling in mediating the toxic and biochemical responses to TCDD is well established. Studies report decreased susceptibility to TCDD toxicity in AhR-null mice (Gonzalez and Fernandez-Salguero, 1998; Peters et al., 1999; Vorderstrasse et al., 2001) and in mice with low-affinity AhR alleles (Okey et al., 1989). Han/Wistar rats display a 1000-fold resistance to TCDD-induced lethality compared with Long-Evans rats, due to a 38–amino acid deletion within the transactivation domain of the AhR (Pohjanvirta et al., 1999). Mice possessing mutations in the AhR nuclear localization signal, DRE-binding domain, or harboring a hypomorphic ARNT allele also fail to exhibit TCDD toxicities (Bunger et al., 2003; Walisser et al., 2004). Although the mechanism of AhR/ARNT-mediated gene expression is well established, the specific gene expression changes and disrupted pathways involved in the toxic and biochemical effects elicited by TCDD and related compounds remain incompletely characterized.

Several studies have described the disposition of TCDD following acute oral exposure (Diliberto et al., 1995, 1996; Wang et al., 1997). Similarly, microarray studies have examined hepatic gene expression responses to TCDD in an effort to further elucidate its mechanisms of toxicity (Boverhof et al., 2005; Fletcher et al., 2005; Tijet et al., 2006; Vezina et al., 2004). The present complementary study concurrently monitored hepatic disposition, histopathology, and gene expression following a single acute oral exposure to TCDD. Our goal was to characterize the relationship between hepatic TCDD levels, and long-term alterations in hepatic gene expression and histopathology.

MATERIALS AND METHODS

Animal handling.

Female C57BL/6 mice, ovariectomized by the vendor on postnatal day (PND) 20 and all having body weights (BWs) within 10% of the average BW, were obtained from Charles River Laboratories (Raleigh, NC) on PND 25. Immature mice are more responsive to TCDD and have been used in previous studies facilitating comparisons with other published data sets (Boverhof et al., 2005; Forgacs et al., 2012; Kopec et al., 2008, 2010a,b, 2011). Mice were ovariectomized to negate potential interactions with estrogens produced by the developing ovaries as animals become reproductively mature. Animals were housed in polycarbonate cages containing cellulose fiber chips (Aspen Chip Laboratory Bedding, Northeastern Products, Warrensberg, NY) in a 23°C HEPA-filtered environment with 30%–40% humidity and a 12-h light/dark cycle (0700 h–1900h) and were allowed free access to deionized water and Harlan Teklad 22/5 Rodent Diet 8640 (Madison, WI). All procedures were performed with the approval of the Michigan State University Institutional Animal Care and Use Committee.

Animal treatment.

On PND 28, acclimatized animals were weighed and a stock solution of TCDD (provided by S. Safe, Texas A&M University, College Station, TX, > 98% purity determined by gas chromatography) was diluted in sesame oil (Sigma, St Louis, MO) to achieve the desired dose based on the average weight. Mice were orally gavaged with 0.1ml of sesame oil for a nominal dose of 0 (vehicle control) or 30 μg/kg BW of TCDD using 1.5-inch feeding needle with a 2.25-mm ball end (Cadence Science, Lake Success, NY). The dose of 30 μg/kg TCDD maximally induces Cyp1a1 after 24h (Boverhof et al., 2005; Kopec et al., 2010a), elicits moderate hepatotoxic effects (e.g., steatosis with immune cell infiltration), while avoiding overt toxicity that might contribute to wasting and a lost in BW. A minimum of five animals were treated per time point and mice for each time point were housed in separate cages. Mice were sacrificed at 1, 4, 12, 24, 36, and 72 weeks after dosing by cervical dislocation. Tissue samples were removed, weighed, flash frozen in liquid nitrogen, and stored at −80°C. The right lobe of the liver was fixed in 10% neutral buffered formalin (Sigma) for histological analysis. All histological processing was performed at Michigan State University Investigative HistoPathology Laboratory (humanpathology.msu.edu/histology), using a modified version of previously published procedures (Sheehan and Hrapchak, 1980). Briefly, fixed liver tissues were sectioned and processed in ethanol, xylene, and paraffin using a Thermo Electron Excelsior tissue processor (Waltham, MA). Tissues were embedded in paraffin with Miles Tissue Tek II embedding center and then sectioned at 5 μm with a rotary microtome. Sections were placed on glass slides, washed twice in xylene for 5min, followed by four quick washes in ethanol, and rinsing in water. Slides were placed in Gill 2 hematoxylin (Thermo Fisher Scientific, Waltham, MA) for 1.5min followed by 2–3 quick dips in 1% glacial acetic acid and rinsed with running water for 2–3min. Slides were then rinsed in ethanol and counterstained with 1% eosin Y-phloxine B solution (Sigma) followed by multiple rinses in ethanol and xylene. Coverslips were attached using synthetic mounting media (Flo-Texx, Lerner Laboratories, Pittsburgh, PA).

Quantification of hepatic TCDD levels and half-life calculation.

Liver samples were processed in parallel with lab blanks (hexane – Caledon distilled in glass grade) and a reference sample (WMF-01 Certified Reference fish tissue) or background sample (chicken liver purchased in Guelph, ON) at Wellington Laboratories Inc. (Guelph, Ontario, Canada). Samples were weighed, spiked with ¹³C₁₂-2,3,7,8-TCDD, and digested in hydrochloric acid. Extracts were cleaned using a multilayer silica column cleanup and concentrated. Immediately prior to injection on the high-resolution gas chromatograph/high-resolution mass spectrometer (HRGC/HRMS, Hewlett Packard 5890 Series II HRGC interfaced to a VG 70SE HRMS) system, ¹³C₁₂-1,2,3,4-TCDD injection standard was added. Chromatographic separations were carried out on a 60-m DB5 (0.25-mm internal diameter [ID], 0.25-μm film thickness) column in constant flow mode (helium, 1ml/min). All injections were 2 μl using splitless injection. HRMS was operated in EI+ selective ion recording mode at a mass resolving power of 10,000 or greater.

Hepatic TCDD elimination half-life was determined using a linear regression of log₁₀ transformed TCDD levels across time to generate a best-fit model using GraphPad Prism 5 (GraphPad Software, Inc., La Jolla, CA). Elimination half-life was calculated assuming a simple one-compartment model (liver) described by the equation C_t = C₀e^−Ket.

RNA isolation.

Frozen liver samples (~70mg) were transferred to 1.0ml of TRIzol (Invitrogen, Carlsbad, CA) and homogenized using a Mixer Mill 300 tissue homogenizer (Retsch, Germany). Total RNA was isolated according to the manufacturer’s protocol with an additional phenol:chloroform extraction (Sigma), resuspended in RNA storage solution (Ambion Inc., Austin, TX), quantified (A₂₆₀), and assessed for purity by determining the A₂₆₀/A₂₈₀ ratio and by visual inspection of 1 μg on a denaturing gel.

cDNA microarray experimental design.

Custom mouse cDNA arrays, consisting of 13,361 features, representing 7884 unique genes (Unigene build no. 144), were used for gene expression analysis. Differential gene expression was assessed using an independent reference design (Yang and Speed, 2002), with TCDD-treated samples cohybridized with time-matched vehicle controls. A minimum of three biological replicates with two independent labelings of each sample (incorporating a dye swap) were performed for each time point.

cDNA microarray analysis.

Detailed protocols for cDNA microarray preparation, labeling of the cDNA probe, sample hybridization, and washing can be found at dbzach.fst.msu.edu/interfaces/microarray.html. PCR amplified DNA was robotically arrayed onto epoxy-coated glass slides (Schott-Nexterion, Duryea, PA) using an Omnigrid arrayer (GeneMachines, San Carlos, CA) equipped with Chipmaker 2 pins (Telechem) at the Research Technology Support Facility (rtsf.msu.edu/genomics_inst_arrayservices.html). Total RNA (30 μg) was reverse transcribed in the presence of Cy3- or Cy5-dUTP (Amersham, Piscataway, NJ) to create fluorescently labeled cDNA, which was purified using a Qiagen PCR purification kit (Qiagen, Valencia, CA). Cy3 and Cy5 samples were mixed, vacuum dried, and resuspended in hybridization buffer and hybridized on the array under a lifterslip (Erie Scientific Company, Portsmouth, NH) for 18–24h in a 42°C water bath. Slides were then washed, dried by centrifugation, and scanned at 635nm (Cy5) and 532nm (Cy3) on an Affymetrix 428 Array Scanner (Santa Clara, CA). Images were analyzed for feature and background intensities using GenePix Pro 5.0 (Molecular Devices, Union City, CA).

Microarray data normalization and analysis.

All microarray data were normalized using a semiparametric approach (Eckel et al., 2005). Model-based t values were calculated from normalized data, comparing treated and vehicle control responses per time point. Empirical Bayes analysis was used to calculate posterior probabilities (P1(t) values) of activity on a per gene and time point basis using the model-based t value (Eckel et al., 2004). A P1(t) cutoff of 0.9999 combined with |fold change| > 1.5 was used to identify a subset of differentially expressed genes to initially focus analysis and data interpretation on the most reproducible differentially regulated genes. Normalization and empirical Bayes analysis were performed using SAS v9.1 (SAS Institute, Cary, NC) and R v2.0.1. The complete long-term microarray data set can be found in Supplementary table 1. A published study with mice treated with 30 μg/kg of TCDD for 2, 4, 8, 12, 18, 24, 72, and 168h (Boverhof et al., 2005) was used as a comparator. Overlapping time points (168h or 1 week) were included to ensure reproducibility between the experiments.

Quantitative real-time PCR.

Quantitative real-time PCR (QRT-PCR) was performed as previously described (Boverhof et al., 2005). Gene names, official gene symbols, Gene IDs, forward and reverse primer sequences, and amplicon sizes are provided in Supplementary table 2. PCR amplification was conducted on an Applied Biosystems PRISM 7000 Sequence Detection System. cDNAs were quantified using a standard curve approach and the copy number of each sample was standardized to the geometric mean of ActB, Gapdh, and Hprt to control for differences in RNA loading, quality, and cDNA synthesis (Vandesompele et al., 2002). For graphing purposes, the relative expression levels were scaled such that the expression level of the time-matched controls was equal to one.

Functional annotation, hierarchical clustering, and statistical analysis.

Annotation and functional categorization of differentially expressed genes was performed using the Database for Annotation, Visualization, and Integrated Discovery (DAVID) (Dennis et al., 2003). Annotation clusters were considered significant with an enrichment score of ≥ 1.3, which is equivalent to a −log p value of 0.05. Hierarchical clustering was generated with MultiExperiment Viewer (MeV v. 4.6.0) using the TM4 microarray software suite (Saeed et al., 2003). Statistical analysis, unless otherwise defined, was performed using SAS v9.1. Data were analyzed using ANOVA followed by Tukey’s post hoc test. Differences between treatment groups were considered significant when p < .05.

RESULTS

Organ and BWs

TCDD significantly increased absolute and relative liver weights (RLWs) by 46 and 35%, respectively, at 1 week postexposure (Table 1), consistent with our acute short-term complementary study (Boverhof et al., 2005). RLW was also increased (12%) at 4 weeks but was not statistically significant. No additional treatment-related alterations in the uterus, lung, heart, kidney, spleen, and mammary gland were noted. Although wasting syndrome is a hallmark of TCDD toxicity, no decreases in terminal BW or BW gain were noted at any time point. However, there was a BW gain at 36 weeks (Table 1), which may be due to the lower initial BWs in the 36-week TCDD group (average BW 9.2g, not shown in Table 1) relative to the 36-week vehicle group (average BW 11.9g, not shown in Table 1) and therefore was not considered treatment related. A number of animals (three vehicle, five treated) also developed dermatitis on the upper back with three cases requiring euthanasia. C57BL/6 mice are susceptible to dermatitis, which may have an environmental or behavioral etiology (Andrews et al., 1994).

Table 1.

Terminal BW, BW Gain, Absolute Liver Weight, and RLW Following Treatment With Vehicle or 30 μg/kg of TCDD

Sacrifice time	Treatment	Terminal BW (g)	BW gain	Absolute liver weight (g)	RLW
1 week	Vehicle	13.49±2.62	1.29±0.13	0.717±0.187	0.054±0.002
	TCDD	14.35±0.97	1.36±0.11	1.045±0.084*	0.073±0.003*
4 weeks	Vehicle	19.35±1.15	1.78±0.14	1.131±0.143	0.058±0.005
	TCDD	18.79±1.45	1.76±0.23	1.208±0.034	0.065±0.005
12 weeks	Vehicle	23.58±1.96	2.38±0.38	1.144±0.121	0.048±0.004
	TCDD	22.98±1.55	2.20±0.20	1.195±0.119	0.052±0.004
24 weeks	Vehicle	28.83±1.80	2.81±0.10	1.323±0.132	0.046±0.004
	TCDD	30.98±3.58	2.96±0.27	1.463±0.254	0.047±0.003
36 weeks	Vehicle	30.10±3.32	2.52±0.21	1.249±0.232	0.042±0.006
	TCDD	34.00±4.79	3.68±0.33*	1.399±0.131	0.041±0.002
72 weeks	Vehicle	38.44±7.72	3.94±1.04	1.556±0.192	0.042±0.015
	TCDD	36.93±5.51	3.56±0.68	1.685±0.214	0.046±0.013

Notes. Values represent averages ± SD for at least four animals in each group. BW gain is represented as terminal BW divided by BW prior to dosing. RLW is liver weight normalized to the terminal BW. Abbreviations: BW, body weight; RLW, relative liver weight.

*p < .05 (TCDD vs time-matched vehicle).

Histopathology

TCDD elicited severe cytoplasmic vacuolization, indicative of lipid accumulation/steatosis (Kopec et al., 2010b), in the periportal and midzonal regions at 1 week (Figs. 1A and 1B), with multiple instances of hepatocellular necrosis, consistent with histopathological changes at 168h (Boverhof et al., 2005). Hepatocellular vacuolization was almost completely subsided by 4 weeks (Supplementary figs. 1A and 1B), consistent with no significant increases in RLW. Multiple foci of extramedullary hematopoiesis (EMH) were seen in control and TCDD-treated mice as early as 1 week. EMH occurs when hematopoietic tissue develops outsides primary (medullary) sites. It is more frequent in rodents than in humans and more prevalent in mice than rats with greater frequency in females (Thoolen et al., 2010). EMH occurrence increased in incidence and severity over time, irrespective of treatment (Figs. 1C and 1D), suggesting it is not AhR mediated. In addition, some TCDD-treated mice showed rare examples of periportal hepatitis (inflammation) at 72 weeks (Supplementary figs. 1C and 1D), which diminished relative to earlier time points, suggesting it dissipates with time.

Fig. 1.

Representative hematoxylin and eosin–stained histopathology results from vehicle and TCDD-treated livers. After 1 week of exposure (A) vehicle-treated mouse liver showed no histological changes, whereas (B) TCDD-exposed mouse liver showed hepatocellular vacuolization (dotted circle) and multiple foci of inflammation (solid arrows). At 72 weeks posttreatment EMH was present in the (C) vehicle and (D) TCDD-treated mice (dashed arrows). Bars = 50 μm.

Hepatic Levels of TCDD and TCDD Half-Life

Hepatic TCDD levels increased steadily through early time points peaking at 72h, followed by a decline to background levels within 24 weeks (Table 2) (Boverhof et al., 2005), consistent with another report following a single dose of TCDD (Abraham et al., 1988). Although TCDD levels were lower at 168h (87.53±26.86 ppb) compared with 1 week (143.67±5.13 ppb), total hepatic levels expressed as the percent of administered TCDD dose (i.e., accounting for liver weight) were not significantly different at 168h (30.1±10.1%) and 1 week (40.3±5.4%), indicating good reproducibility between the studies. Linear regression using a single-compartment model calculated a hepatic TCDD half-life of 12.0 days, with a 95% confidence interval of 11.3–12.9 days. The model, model intercept, and treatment duration variable were significant (p < .0001) with a coefficient of determination (R ²) of 0.964.

Table 2.

Hepatic TCDD Levels and Percentage of Administered Dose

Time point	Study	TCDD levels in hepatic tissue (ppb)	Percent of administered dosea
Vehicle		0.0633±0.0555	NA
2 h	Boverhof et al. (2005)	62.267±7.336*	10.743±1.515
4 h		103.833±29.759*	18.660±3.935
8 h		153.667±40.624*	24.906±4.025
12 h		135.067±49.900*	28.078±11.306
18 h		167.667±32.021*	32.740±7.316
24 h		191.000±86.157*	47.745±25.694
72 h		213.667±16.197*	61.577±22.293
168 h		87.533±26.864*	33.087±10.114
1 week	Long term	143.667±5.132*	40.269±5.432
4 weeks		13.373±3.491*	4.576±1.175
12 weeks		0.553±0.306	0.183±0.094
24 weeks		0.013±0.004	0.006±0.002
36 weeks		ND	NA
72 weeks		ND	NA

Note. Abbreviations: NA, not applicable; ND, not determined, as levels had returned to background.

^aPercent of administered dose was calculated as: absolute liver TCDD weight (ng) divided by total TCDD administered per mouse (ng) multiplied by 100%. Total liver ng TCDD is equivalent to TCDD hepatic levels (in ng/g or ppb) × absolute liver weight (in g), whereas total ng TCDD administered per mouse takes into account the TCDD dose (30 μg/kg or ng/g) multiplied by the average mouse weight at the time of dosing (g).

*Significantly different from vehicle, p < .05.

Long-term Differential Gene Expression

A total of 402 differentially expressed genes were identified between 1 and 72 weeks (|fold change| > 1.5, P1(t) > 0.9999) compared with 540 unique gene expression changes between 2 and 168h (|fold change| > 1.5, P1(t) > 0.9999) (Boverhof et al., 2005). The number of gene expression changes peaked at 1 week and subsequently decreased to zero within 24 weeks (Fig. 2). At 36 weeks, Rad51b and Lpin2 were modestly induced (up to 1.8-fold), whereas suppression of Rpl3, Apaf1, Agt (−1.7-fold) and induction of Hsd17b10 (2.0-fold), Acox1 (1.8-fold), Trmt1 (1.7-fold), and Ighm (4.1-fold) were observed exclusively at 72 weeks. Hierarchical clustering of differentially expressed genes in both data sets (Fig. 3A) revealed that 168h and 1 week time points clustered together, as expected. Additional comparison between 168h (Boverhof et al., 2005) and at 1 week identified 100 overlapping genes with a Pearson’s r correlation coefficient of 0.943 indicating good reproducibility across the two experiments (Fig. 3B). When filtering criteria were relaxed (from |fold change| > 1.5, P1(t) > 0.9999 to |fold change| > 1.2, P1(t) > 0.9), the overlap increased to 265 genes with a Pearson’s correlation of 0.913 (Fig. 3C).

Fig. 2.

Number of genes identified as differentially expressed in TCDD-treated mice relative to the time-matched vehicle controls. Same filtering criteria (|fold change| > 1.5, P1(t) > 0.9999) were used to obtain the number of differentially expressed genes in the short-term (light gray bars, 2–168h; Boverhof et al., 2005) and long-term TCDD study (dark gray bars, 1–72 weeks). TCDD-elicited differential gene expression was maximal after 1 week of exposure and then decreased to zero at 24 weeks, followed by negligible gene expression changes at 36 and 72 weeks, respectively.

Fig. 3.

(A) Hierarchical clustering (average linkage, Spearman ranked correlation) of the union of the differentially expressed genes in the short-term (2–168h; Boverhof et al., 2005) and long-term TCDD studies was generated using MeV software (gene tree was cropped out for simplicity). (B) Comparison of the 168h (Boverhof et al., 2005) and 1 week differential gene expression at stringent (|fold change| > 1.5, P1(t) > 0.9999) criteria identified 100 overlapping genes that were correlated in terms of fold change with a Pearson’s r correlation coefficient of 0.943. (C) When the filtering criteria were relaxed (|fold change| > 1.2, P1(t) > 0.9), the overlap between 168h and 1 week time points increased to 265 genes with significant correlation, suggesting good experimental reproducibility.

Functional Annotation of Differential Gene Expression

The 402 differentially expressed genes were functionally associated with oxidoreductase activity and glutathione metabolism, carbohydrate, glycerol and lipid metabolism, immune response, DNA/RNA binding, as well as protein folding and apoptosis (Table 3). Within the oxidoreductase category, “AhR battery genes” were induced through 4 weeks (Tiparp and Nqo1 up to 5.1-fold) and 12 weeks (Cyp1a1 up to 32.2-fold) postdose. However, the induction of the vast majority of redox-related genes only occurred within the first week of treatment. Consistent with hepatic fat accumulation, TCDD induced the expression of lipid transport, processing, and metabolism genes, including Cd36 (4.5-fold), Lpl (2.3-fold), Fabp4 (1.7-fold), Fabp5 (3.7-fold), and the suppression of Apoa1 (−2.5-fold). TCDD also suppressed gene expression involved in gluconeogenesis, carbohydrate, and glycerol metabolism between 1 and 4 weeks, including Dak (−2.6-fold), Gpd2 (−2.2-fold), Pck1 (−2.6-fold), Gpcpd1 (−1.6-fold), and Got1 (−1.6-fold). In addition, TCDD elicited hepatic inflammation, which included the induction of antigen processing and presentation genes such as Cd74 (2.9-fold), H2-Aa (2.9-fold), H2-Ab1 (3.1-fold), H2-Eb1 (3.1-fold), and H2-Bf (2.1-fold). Almost all differentially expressed genes in the long-term study were also differentially expressed between 2 and 168h (Boverhof et al., 2005) (Table 3). The few unique long-term differential expression responses included lipid-processing genes (Acox1, Hsd17b10), RNA-binding genes (Rpl3 and Trmt1), as well as the apoptosis-associated gene, Apaf1.

Table 3.

Functional Categorization and Expression of Select Differentially Expressed Genes Following Long-term Exposure to 30 μg/kg TCDD

Functional category	Gene ID	Gene name	Gene symbol	Regulation	1 week	4 weeks	12 weeks	24 weeks	36 weeks	72 weeks	2–168h (Boverhof et al., 2005)a
Oxidoreductase activity/glutathione metabolism	72535	Aldh1b1 aldehyde dehydrogenase 1 family, member B1	Aldh1b1	▲	1.5	1.4	NC	NC	NC	NC	Yes
	109857	Carbonyl reductase 3	Cbr3	▲	1.7	NC	NC	NC	NC	NC	Yes
	13057	Cyba cytochrome b-245, alpha polypeptide	Cyba	▲	1.6	NC	NC	NC	NC	NC	Yes
	12858	Cytochrome c oxidase subunit Va	Cox5a	▲	1.6	NC	NC	NC	NC	NC	Yes
	13076	Cytochrome P450, family 1, subfamily a, polypeptide 1	Cyp1a1	▲	16.3	32.2	3.3	NC	NC	NC	Yes
	14858	Glutathione S-transferase, alpha 2 (Yc2)	Gsta2	▲▼	3.0	1.9	−2.0	NC	NC	NC	Yes
	14860	Glutathione S-transferase, alpha 4	Gsta4	▲	1.8	NC	NC	NC	NC	NC	Yes
	14872	Glutathione S-transferase, theta 2	Gstt2	▲	1.7	NC	NC	NC	NC	NC	Yes
	18104	NAD(P)H dehydrogenase, quinone 1	Nqo1	▲	2.0	1.9	NC	NC	NC	NC	Yes
	99929	TCDD-inducible poly (ADP-ribose) polymerase	Tiparp	▲	5.1	1.6	NC	NC	NC	NC	Yes
	22235	UDP-glucose dehydrogenase	Ugdh	▲	2.1	2.0	NC	NC	NC	NC	Yes
	22436	Xanthine dehydrogenase	Xdh	▲	1.8	1.5	NC	NC	NC	NC	Yes
Lipid metabolism	66082	Abhydrolase domain containing 6	Abhd6	▲▼	2.3	NC	NC	NC	NC	−1.6	Yes
	11430	Acyl-coenzyme A oxidase 1, palmitoyl	Acox1	▲	NC	NC	NC	NC	NC	1.8	No
	11606	Angiotensinogen (serpin peptidase inhibitor, clade A, member 8)	Agt	▼	NC	NC	NC	NC	NC	−1.7	Yes
	11806	Apolipoprotein A-I	Apoa1	▼	−2.5	−2.0	NC	NC	NC	NC	Yes
	12491	CD36 antigen	Cd36	▲	4.5	2.5	NC	NC	NC	NC	Yes
	11770	Fatty acid–binding protein 4, adipocyte	Fabp4	▲	1.7	NC	NC	NC	NC	NC	Yes
	16592	Fatty acid–binding protein 5, epidermal	Fabp5	▲	NC	NC	3.7	NC	NC	NC	Yes
	15108	Hsd17b10 hydroxysteroid (17-beta) dehydrogenase 10	Hsd17b10	▲	NC	NC	NC	NC	NC	2.0	No
	15446	Hydroxyprostaglandin dehydrogenase 15 (NAD)	Hpgd	▲	1.6	NC	NC	NC	NC	NC	Yes
	15450	Lipase, hepatic	Lipc	▼	−1.5	NC	NC	NC	NC	NC	Yes
	64898	Lipin 2	Lpin2	▲	NC	1.6	NC	NC	1.6	NC	Yes
	16956	Lipoprotein lipase	Lpl	▲	2.3	2.2	NC	NC	NC	NC	Yes
Carbohydrate/glycerol metabolism	225913	Dihydroxyacetone kinase 2 homolog (yeast)	Dak	▼	−2.6	−1.6	NC	NC	NC	NC	Yes
	14718	Glutamate oxaloacetate transaminase 1, soluble	Got1	▼	−1.7	NC	NC	NC	NC	NC	Yes
	14571	Glycerol phosphate dehydrogenase 2, mitochondrial	Gpd2	▼	−2.1	−2.1	NC	NC	NC	NC	Yes
	74182	Glycerophosphocholine phosphodiesterase GDE1 homolog (Saccharomyces cerevisiae)	Gpcpd1	▼	−1.6	NC	NC	NC	NC	NC	Yes
	18534	Phosphoenolpyruvate carboxykinase 1, cytosolic	Pck1	▼	−2.5	NC	NC	NC	NC	NC	Yes
Immune response	12010	Beta-2 microglobulin	B2m	▼	−1.5	NC	NC	NC	NC	NC	Yes
	16149	CD74 antigen	Cd74	▲	2.9	1.5	NC	NC	NC	NC	Yes
	12628	Complement component factor h	Cfh	▼	−1.7	NC	NC	NC	NC	NC	Yes
	14960	Histocompatibility 2, class II antigen A, alpha	H2-Aa	▲	2.9	1.6	NC	NC	NC	NC	Yes
	14961	Histocompatibility 2, class II antigen A, beta 1	H2-Ab1	▲	3.1	NC	NC	NC	NC	NC	Yes
	14969	Histocompatibility 2, class II antigen E beta	H2-Eb1	▲	3.1	1.6	NC	NC	NC	NC	Yes
	14962	Histocompatibility 2, complement component factor B	H2-Bf	▲	2.1	1.8	NC	NC	NC	NC	Yes
	16019	Immunoglobulin heavy constant mu	Ighm	▲	NC	NC	NC	NC	NC	4.1	Yes
	17195	Mannose-binding lectin, serum (C)	Mbl2	▼	−1.7	NC	NC	NC	NC	NC	Yes
DNA and RNA binding/ protein folding/apoptosis	11783	Apoptotic peptidase activating factor 1	Apaf1	▼	NC	NC	NC	NC	NC	−1.7	No
	15505	Heat shock 105kDa/110kDa protein 1	Hsph1	▲	2.0	NC	NC	NC	NC	NC	Yes
	19363	RAD51 homolog B	Rad51b	▼▲	−1.9	−4.6	NC	NC	1.8	NC	Yes
	27367	Ribosomal protein L3	Rpl3	▼	NC	NC	NC	NC	NC	−1.7	No
	212528	tRNA methyltransferase 1	Trmt1	▲	NC	NC	NC	NC	NC	1.7	No

Notes. Abbreviation: Fold changes presented in the table met the filtering criteria of |fold change| > 1.5 and P1(t) > 0.9999. NC, no change.

^aGenes differentially expressed in the Boverhof et al. (2005) data set meeting the filtering criteria of |fold change| > 1.4 and P1(t) > 0.98.

Comparing the union of short- and long-term responses identified 192 overlapping genes (|fold change| > 1.5, P1(t) > 0.9999) that increased to 602 with relaxed filtering criteria (|fold change| > 1.2, P1(t) > 0.9) (Figs. 4A and 4B). Among the 37 unique long-term gene expression responses, DAVID identified functional clusters (albeit most with enrichment scores < 1.3 and therefore considered not significant) associated with cytoskeleton, nuclear lumen, actin binding, regulation of transcription, DNA binding, transport, and transition metal ion binding. The responses unique to short-term TCDD exposure were linked to mRNA processing, blood vessel development, redox, cellular response to stress, apoptosis, and transcription (Supplementary table 3).

Fig. 4.

Comparison of short-term (2–168h; Boverhof et al., 2005) and long-term TCDD-elicited gene expression responses. (A) The union of short-term and long-term gene expression responses was compared using stringent filtering criteria (|fold change| > 1.5, P1(t) > 0.9999) to identify 197 overlapping genes. (B) When filtering criteria were relaxed (|fold change| > 1.2, P1(t) > 0.9), the number increased to 602 overlapping differentially expressed genes.

Overall, TCDD-elicited long-term differential gene expression responses were associated with increases in RLW, hepatocellular vacuolization, and inflammation, and correlated with hepatic TCDD levels. Microarray analysis identified only a handful of modest gene expression changes, the majority of which were not altered at any other time point, consistent with the minimal hepatic TCDD levels between 24 and 72 weeks.

Verification of Microarray Responses

QRT-PCR was used to verify the differential expression of Cyp1a1, Tiparp, Nqo1, Lpl, Cd36, and Pck1 (Fig. 5). Pearson’s correlation coefficients of greater than 0.85 were calculated for each gene indicating a good agreement between the microarray and QRT-PCR data. The magnitude of induction for each gene at the 168h (Boverhof et al., 2005) and 1 week time points of the two studies were also similar. However, microarray data compression was evident for Cyp1a1 and Tiparp, likely due to limited fluorescence range of the dyes (0–65,535), which resulted in signal saturation. Cyp1a1, Tiparp, and Nqo1 expression were tightly linked to hepatic TCDD levels at every time point. Indeed, analysis of Cyp1a1, Tiparp, and Nqo1 expression and hepatic TCDD levels (between 2h and 24 weeks) calculated significant Pearson’s correlations of 0.646, 0.711, and 0.798, respectively (data not shown).

Fig. 5.

QRT-PCR verification of microarray results. The same RNA used for cDNA microarray analysis was examined by QRT-PCR. All fold changes were calculated relative to time-matched vehicle controls. Bars (left axis) and lines (right axis) represent QRT-PCR and cDNA microarray data, respectively, whereas the x-axis represents the time points. The numerical value in the top right corner of each graph indicates the Pearson’s correlation value (r) between the QRT-PCR and microarray results. Official gene symbol names are used and results are the average of four biological replicates. Error bars represent the SEM for the average fold change. Asterisk (*) represents p < .05 for QRT-PCR.

DISCUSSION

Several toxicogenomic studies have examined hepatic gene expression responses to TCDD after acute, short-term exposure or repeated subchronic or chronic exposure (Boverhof et al., 2005; Fletcher et al., 2005; Nault et al., 2013, in preparation; Tijet et al., 2006; Vezina et al., 2004). However, no study has examined the long-term decrease in differential gene expression following a single acute dose of TCDD. The present study concurrently monitored hepatic disposition, histopathology, and gene expression to investigate the long-term effects of TCDD in mice.

The highest hepatic TCDD levels were achieved at 72h, which decreased to control levels by 24 weeks with a calculated half-life of 12 days, comparable with other mouse studies. For example, Gasiewicz et al. (1983) reported a TCDD half-life of 11 days in C57BL/6 mice, whereas Birnbaum (1986) provides hepatic and adipose half-life estimates of 8.5 and 10.3 days, respectively. This is relatively shorter than the 21- to 22-day half-life in Han/Wistar and Long-Evans rats (Pohjanvirta et al., 1990), the 12–15 days in the Golden Syrian hamster (Olson et al., 1980), and the 30–94 days in the guinea pig (Gasiewicz and Neal, 1979; Olson, 1986). In adult humans, the half-life of TCDD has been estimated to be 1.3 years based on Victor Yushchenko data (Sorg et al., 2009), 5.8 years in a healthy volunteer (Poiger and Schlatter, 1986), and 7.1–11.3 years in Operation Ranch Hand veterans (Michalek et al., 1996; Pirkle et al., 1989; Wolfe et al., 1994), compared with 1.6 years in children from the Seveso incident (Kerger et al., 2006).

Multiple factors contribute to the longer TCDD half-life in humans such as percent body fat. Humans have significantly more body fat (23%) relative to rats (7%) and mice (4%) (Reitz et al., 1996), leading to greater TCDD sequestration and slower elimination. Accounting for the lipid solubility of TCDD (log p value) also reduces the discrepancy in predicted human and rodent half-life estimates (Sarver et al., 1997). Age, smoking habits, and breast-feeding status are other determinants in human elimination rates (Milbrath et al., 2009). In addition, the half-life of TCDD decreases as exposure increases, suggesting dose-dependent elimination (Emond et al., 2006), which may also contribute to longer TCDD half-life in humans. Most rodent half-life estimates of ~30 days involve high TCDD doses (> 1 μg/kg) that induce maximal cytochrome P450 induction. In contrast, the half-life in highly exposed humans is ~90 days, whereas the half-life is significantly longer in low-dose human studies (e.g., Operation Ranch Hand). Moreover, the half-life doubled from ~100 to 200 days in physiologically based pharmacokinetic (PBPK) modeling studies that increased the adipose tissue compartment from 7 to 70% (Emond et al., 2006), further suggesting an interaction between body fat composition and TCDD dose that affects elimination.

Hepatic TCDD levels at 12 weeks postdose averaged 553 ppt TCDD with 83 unique differentially expressed genes. By 24 weeks, TCDD levels decreased to 13 ppt (within range of control mice) with no genes exhibiting differential expression. Hepatic TCDD levels in treated mice at 12 and 24 weeks are similar to published human concentrations. For example, mean TCDD adipose and serum levels in the general population (United States, Canada, Germany, and France; 1972–1999) are estimated to range between 2.0 and 19.8 ppt (Aylward and Hays, 2002), whereas average serum lipid TCDD levels in 250 exposed National Institute for Occupational Safety and Health (NIOSH) workers (1987–1988) ranged between 69 and 1410 ppt depending on exposure duration (< 1 to 15+ years) (Aylward et al., 2005). However, direct comparisons between mice and humans are confounded by species-specific AhR-mediated gene expression differences, different target tissue TCDD level comparisons (adipose and serum vs liver), and the activities of related compounds.

RLW, histopathology, and gene expression changes were associated with hepatic TCDD levels at earlier time points. As TCDD levels decreased to control levels at 24 weeks, RLW increases and histological changes returned to normal with negligible gene expression changes. Specifically, although TCDD levels were higher in treated animals and some differential gene expression persisted, histologically, control and treated samples were similar 4 weeks postdose. These results suggest the effects elicited by a single oral gavage of TCDD are reversible, and that the tissue levels of TCDD at 4 weeks are insufficient to elicit marked steatosis with immune cell infiltration.

Consistent with published studies, increases in RLW were associated with hepatic steatosis resulting from the differential expression of lipid transport, processing, and metabolism genes, whereas hepatic inflammation was associated with the induction of immune response genes 1 week postdose (Boverhof et al., 2005; Dere et al., 2011; Forgacs et al., 2012; Kopec et al., 2008, 2010a, 2011). Moreover, almost all gene expression changes between 1 and 12 weeks were also differentially expressed in the complementary short-term study (2–168h) (Boverhof et al., 2005), where hepatic TCDD levels were significantly greater.

By 72 weeks, hepatic TCDD levels were not detectable, with treated and control mice exhibiting similar RLWs, and no evidence of hepatocellular vacuolization in treated mice. Microarray analysis identified only seven differentially expressed genes, of which five were not significantly changed at any other time point in this study or previously reported to be differentially expressed (Boverhof et al., 2005). This included the modest induction of Hsd17b10 and Acox1, and the suppression of Agt, all of which are involved in lipid metabolism. The late differential expression of Hsd17b10, Acox1, and Agt may play a role in reestablishing hepatic lipid homeostasis following TCDD elimination. For example, Agt is regulated by hyperglycemia and linked to insulin-resistant obesity and type 2 diabetes (Gabriely et al., 2001). Further details of TCDD-elicited effects on hepatic lipid transport, processing, and metabolism have been described previously (Angrish et al., 2011, 2012, 2013; Dere et al., 2011; Forgacs et al., 2012).

Our results demonstrate that effects elicited by a single oral gavage of TCDD are reversible. Overall, hepatic differential gene expression could be anchored to changes in histopathology and correlated with hepatic TCDD levels. By 24 weeks, TCDD levels were comparable in control and treated mice with no significant differences in liver histopathology and only modest changes in gene expression. Ongoing complementary studies are examining the progression of hepatic toxicity and other systemic effects following repeated TCDD dosing.

SUPPLEMENTARY DATA

Supplementary data are available online at http://toxsci.oxford journals.org/.

FUNDING

National Institute of Environmental Health Sciences Superfund Basic Research Program (P42ES04911).