The embryos of teleost fish are exquisitely sensitive to the toxic effects of exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). However, several lines of evidence suggest that adults are less sensitive to TCDD. To better understand and characterize this difference between early life stage and adults, we exposed zebrafish (Danio rerio) to graded TCDD concentrations at different ages. The LD50 for embryos exposed at 1 day post-fertilization (dpf) was more than an order of magnitude lower than it was for juveniles exposed at 30 dpf. The latency between exposure and response also increased with age. Embryo toxicity was characterized by marked cardiovascular collapse and heart malformation, whereas juveniles exposed at 30 dpf had no detectable cardiovascular toxicity. In juveniles, the effects of TCDD exposure included stunted growth, altered pigmentation, and skeletal malformations. Furthermore, the transcriptional profile produced in hearts exposed to TCDD as embryos had very little overlap with the transcriptional changes induced by TCDD at 30 dpf. The early cardiotoxic response was associated with fish exposed prior to metamorphosis from the larval to the adult body plan at approximately 14 dpf. Our results show conclusively that the developmental stage at the time of exposure controls the toxic response to TCDD.
Agonists for the aryl hydrocarbon receptor (AHR) include halogenated aromatic hydrocarbons (HAHs), a family of chemicals that includes polychlorinated dibenzo-p-dioxins, dibenzofurans, and polychlorinated biphenyls (Abel and Haarmann-Stemmann, 2010; Hankinson, 1995; Ramadoss et al., 2005; White and Birnbaum, 2009). These widespread environmental contaminants cause a range of toxic effects, including carcinogenesis, developmental and reproductive defects, and death (Nguyen and Bradfield, 2008). Salmonids are especially sensitive to HAHs: lake trout (Salvelinus namaycush) at early life stages are currently among the most sensitive vertebrates known (Elonen et al., 1998; Tanguay et al., 1999; Van den Berg et al., 1998; Walker and Peterson, 1994; Walker et al., 1996). The prototype HAH used to study toxicity is 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). TCDD binding causes AHR to form a heterodimer with aryl hydrocarbon receptor nuclear transporter (ARNT) in the nucleus, where it regulates gene transcription (Jones et al., 1986; Nguyen and Bradfield, 2008). AHR and ARNT are both necessary for TCDD toxicity (Prasch et al., 2006; Tanguay et al., 2003).
The zebrafish (Danio rerio) has proven to be a useful model for studying TCDD toxicity (Carney, et al., 2006b). In zebrafish, TCDD causes a spectrum of toxic effects that matches that observed in lake trout: pericardial and yolk sac edema, heart malformation, and craniofacial and skeletal defects (Henry et al., 1997). Although the molecular mechanism remains unclear, the cardiotoxicity produced by TCDD has been extensively characterized in the zebrafish embryo (Antkiewicz et al., 2005, 2006; Canga et al., 1988; Heideman et al., 2005). TCDD exposure leads to decreased cardiac output, a reduction in peripheral blood flow, pericardial edema, and unlooping of the two chambered zebrafish heart (Antkiewicz et al., 2005; Carney, et al., 2006b; Henry et al., 1997). AHR activation by TCDD or other agonists also disrupts atrioventricular and bulboventricular valve formation and halts bulbus arteriosus growth (Grimes et al., 2008; Mehta et al., 2008). At the molecular level, TCDD has been shown to downregulate transcripts associated with cell proliferation. This effect on gene expression precedes an actual reduction in cardiomyocyte proliferation and number (Carney, et al., 2006b). Ultimately, toxicity progresses to complete heart failure with ventricular standstill and cessation of blood flow. All of these effects can be observed within the first 4 days post-fertilization (dpf).
Work with trout and other fish species indicates that adults are considerably less sensitive to the toxic effects of TCDD than early life stage embryos (Peterson et al., 1993; Walker and Peterson, 1994). It is possible that AHR activation disrupts processes important for development that are no longer critical as organogenesis is completed. The difficulty in following the growth and development of species such as trout has made it difficult to examine this in detail, and we know little about how sensitivity to TCDD changes during growth and development. To address this, we took advantage of the zebrafish model to systematically document changes in TCDD sensitivity throughout the developmental process.
In this report, we demonstrate that the acute cardiotoxic effects produced by TCDD in embryos are completely absent in fish exposed as juveniles. The zebrafish heart is most sensitive to TCDD at approximately 48–96 hpf and then becomes increasingly resistant to acute cardiotoxicity as the fish moves through the larva stage. This period of sensitivity to TCDD cardiotoxicity does not coincide with the initial development of the heart: the heart is formed prior to 48 hpf. Analysis of heart morphology and peripheral blood flow shows that sensitivity to cardiotoxicity ends at a point that coincides with larval metamorphosis. This metamorphosis alters fins, pigmentation, and muscle to transform the larval shape into the adult body plan. The onset of metamorphosis occurs when the larva is approximately 5 mm in length (Brown, 1997; Parichy et al., 2009). Under our rearing conditions, this takes place at approximately 2 weeks post-fertilization.
Because AHR is a transcriptional regulator, changes in messenger RNA produced by TCDD are likely to be important in controlling cellular responses. Microarray experiments comparing transcript levels in embryonic and juvenile hearts reveal drastic differences, with only minor overlaps, in the transcriptional responses to TCDD. This indicates a fundamental shift during development in the way that heart cells respond to TCDD activation of AHR.
MATERIALS AND METHODS
Experiments were performed using wild type AB zebrafish (Danio rerio), Tg(cmlc2:GFP) (generously supplied by C.G. Burns), or albino Tg(cmlc2:GFP). Eggs were obtained from spawning adult zebrafish maintained according to procedures in The Zebrafish Book (Akimenko et al., 1995). Zebrafish eggs were collected within 4 h of spawning.
3H-TCDD exposure for dose response and body burden experiments.
Zebrafish eggs, larvae, or juveniles were exposed to either 3H-TCDD or vehicle alone. These waterborne exposures were conducted in glass scintillation vials for 1 h. The acetone vehicle concentration in the treatment water was 0.10% in the dose-response experiment and 0.15% in the time-course studies. For mass determination, fish were placed on nylon mesh on top of absorbent tissue until surface sheen was gone to remove water without desiccating the tissue prior to weighing. For the dose-response experiments, zebrafish were exposed to graded concentrations of 3H-TCDD (0, 0.05, 0.1, 0.2, 0.4, and 0.6 ng TCDD/ml of treatment water). For the time-course experiments, the TCDD concentrations during the waterborne exposure were 0.2 and 0.7 ng/ml. Immediately following the 1-h exposure, the zebrafish were rinsed in TCDD-free water and transferred to housing aquaria. To measure TCDD uptake, five zebrafish from each group were transferred from these holding vessels to scintillation vials and hour after the end of the TCDD exposure, euthanized by an overdose of MS222, and solubilized in 1 ml Soluene 350. Then, 10 ml Hionic-Fluor was added, and the vials were stored overnight in the dark. Liquid scintillation counting was performed (Packard model 1900TR) using the tSIE quench correction method. Disintegrations per minute were converted to doses based on the specific activity of 3H-TCDD.
TCDD exposure for microarray and morphological/functional assessments.
Zebrafish eggs, larvae, or juveniles were statically exposed in water to either TCDD (1 ng/ml) or vehicle (dimethyl sulfoxide [DMSO]) in 5 or 20 ml glass scintillation vials for 1 h. After exposure, fish were rinsed in TCDD-free water and placed in housing aquaria. Juvenile exposures (30 dpf) were performed on fish measuring greater than 1 cm in length to control for differences in developmental maturity (Parichy et al., 2009).
Whole hearts were obtained from TCDD- or vehicle-exposed Tg(cmlc2:GFP) juveniles at 2 and 12 hpe, using 5 juvenile hearts per replicate and 3 replicates per time point. Hearts were placed in RNAlater on ice, and total RNA was isolated using the Qiagen RNeasy Mini kit (Qiagen, Valencia, CA) according to the manufacturer’s protocol for isolation of RNA from heart tissue. Total RNA (1 μg) was used to synthesize complimentary DNA and biotin-labeled complimentary RNA (cRNA) using the One-Cycle Target Labeling and Control Reagents kit (Affymetrix, Santa Clara, CA); the cRNA (15 μg) was fragmented and hybridized against Affymetrix Zebrafish GeneChip Arrays according to the manufacturer’s protocol. Following hybridization, the arrays were stained using streptavidin-phycoerythrin on an Affymetrix Fluidics Station 400 using protocol EukGEWS2v4 then scanned with an Agilent Gene Array Scanner (Agilent Technologies, Santa Clara, CA). Transcript changes were determined using Arrayassist software (Stratagene, La Jolla, CA).
Determination of red blood cell perfusion rate.
Juvenile (30 dpf) Tg(cmlc2:GFP) zebrafish were exposed to TCDD or vehicle as described above. At 2 weeks post-exposure, fish were placed in 4% methylcellulose and their caudal fins imaged with videomicroscopy. The number of red blood cells passing through a defined point within a caudal vessel in 10 s was recorded for each animal.
BrdU incorporation assay.
Juvenile (30 dpf) Tg(cmlc2:GFP) zebrafish were exposed to TCDD or vehicle as described above then placed in aquaria containing 10nM BrdU for 1 week. Fish were transferred to fresh water containing 10nM BrdU daily. Hearts were isolated after 1 week and stained to determine BrdU incorporation. Briefly, following isolation, hearts were rinsed in PBS then fixed at room temperature in 4% paraformaldehyde for 2 h. Following fixation hearts were rinsed twice with PBST (PBS + 0.1% Tween-20) and digested in collagenase (1 mg/ml in PBS) for 10 min at room temperature. Hearts were washed to remove collagenase and blocked in Roche Blocking Buffer containing 10% fetal calf serum and 1% DMSO in PBST for 30 min at room temperature. Hearts were then incubated overnight at 4°C with anti-BrdU antibody (1:100) in DNA Digestion Mix (DNase I (2 U/ml), ExoIII (10 U/ml), Rhodamine Phalloidin (1:50), ExoIII buffer, 1% bovine serum albumin). After washing in PBST, hearts were incubated with Alexa Fluor 633 antibody (1:800) for 2 h at room temperature, washed in PBST, and imaged with a confocal microscope. Cells that colocalized with GFP-labeled cardiomyocytes and Alexa Fluor 633–labeled BrdU were counted using Image J software.
Determination of TCDD effects on embryonic and larval heart morphology.
Albino Tg(cmlc2:GFP) zebrafish were exposed to TCDD or vehicle at three different time points (0–4 hpf, 7 dpf, and 12 dpf) then placed in aquaria and allowed to develop for 4 days. At 4 days post-exposure, fish were placed in 4% methylcellulose and imaged ventral side up using an Optronics MicroFire camera mounted on a Leica MZ16 stereomicroscope. Bright field and fluorescent images were taken and merged to show changes in heart morphology in situ.
Sensitivity to TCDD Decreases With Age
To assess changes in TCDD sensitivity during the course of development, we used dose-response experiments to measure the potency of TCDD at different embryonic, larval, and juvenile life stages. Figure 1 shows the results of dose-response experiments measuring mortality in zebrafish exposed to varying TCDD concentrations at 0, 1, 3, 10, and 25 (dpf). Fish were exposed for 1 h to radiolabeled TCDD in water so that the body burden and mortality could be determined. The results show a steady decrease in sensitivity to TCDD as the age at time of exposure increases. For zebrafish exposed to TCDD immediately after fertilization, the LC50 for TCDD was approximately 0.6 ng/g; however, this increased to approximately 2 ng/g in fish exposed at day 3 and to 7 ng/g for fish exposed at 10 dpf. The developing fish continued to become resistant, with an LC50 of 25 ng/g for exposure at day 25, the latest exposure age in the experiment. This represents approximately a 40-fold decrease in sensitivity.
TCDD Uptake Is Lowest During Early Development
We were initially concerned that decreases in the sensitivity to waterborne-TCDD exposure could be simply due to a decrease in TCDD uptake, as the fish assumed a more adult-like morphology with skin and scales. However, we observed that the body burden of radiolabeled-TCDD in zebrafish after our lab-standard 1-h exposure protocol actually tended to increase with age at exposure (Fig. 2). With this waterborne exposure, zebrafish exposed at 25 dpf reached a body burden roughly threefold to fivefold higher than the zebrafish exposed at 1 dpf. We also observed a substantial increase in body burden between 1 and 3 dpf, a period coinciding with hatching and loss of the egg chorion as a barrier to uptake. We conclude that resistance to TCDD with age is not due to a reduction of TCDD uptake.
The Latency to Response Lengthens With Age
The dose-response curve shown in Figure 1 does not completely capture the changes in TCDD sensitivity that we observed as fish developed and grew. In addition to changes in the LC50, we found that the latency period between exposure and response changed with age at exposure. In the experiment shown in Figure 3, zebrafish at varying ages were exposed for 1 h to TCDD at either a low (0.2 ng/ml) or high dose (0.7 ng/ml). Cumulative mortality was then followed over time, with the horizontal bars showing the period of time at dosing and the time at 50% mortality. The light bars indicate the 0.2 ng/ml exposure, and the dark bars indicate the higher concentration. The size of the fish measured as average wet weight is plotted as an indicator of growth during this period.
Consistent with the data described earlier, as growth progressed, the fish became resistant to the 0.2 ng/ml TCDD exposure. More than half of the fish exposed to this concentration at 25 dpf survived to the termination of the experiment at 58 days. In contrast, this concentration was lethal when the fish were exposed as embryos.
The latency to 50% mortality for fish exposed to the higher dose fell into recognizable groups. Fish exposed during the first 5 days of life all appeared to die around day 10–12, regardless of the time of dosing. This suggests that TCDD toxicity prevented them from passing a developmental hurdle. Zebrafish exposed at 7 and 17 dpf had a latency to 50% mortality of approximately 5 days. Thereafter, we observed an increase in latency, reaching 15 days for fish exposed to the high dose at 37 dpf. For reference, approximate wet weights and developmental stages are also shown in the figure. The shift in latency corresponds with the metamorphic transition of larval fish to the juvenile stage.
Age and TCDD Cardiotoxicity
The most prominent effect of TCDD after embryonic exposure is heart failure, with pericardial and yolk sac edema (Antkiewicz et al., 2005). We found that this response diminished with age of exposure. In order to follow changes in heart morphology in larval fish, we used an albino strain that carries a cmlc2::GFP reporter, specifically labeling heart muscle with GFP (Huang et al., 2003). The fish were exposed to TCDD just after fertilization (0 dpf), or as larva (7 and 12 dpf), and ventral images of the heart muscle were recorded with fluorescence microscopy 4 days after exposure (Fig. 4). TCDD exposure at 0 dpf produced a pronounced, and previously well-characterized, elongation defect in the atrium, with no looping remaining (Antkiewicz et al., 2005, 2006). Larval exposure at 7 dpf produced a less severe defect, but the heart was partially unlooped, and the atrium was still clearly affected. The heart defects in fish exposed at 12 dpf were far more modest than those in fish exposed at 0 or 7 dpf. In particular, the atrium was less distended, and the heart appeared more compact, with a large degree of overlap between atrium and ventricle. Nonetheless, the ventricle was notably smaller and did not occlude the atrium, which remained visible and moderately elongated. Mild edema was also present in both 7 and 12 dpf exposed larvae (data not shown).
Juvenile zebrafish (30 dpf) exposed to a lethal concentration of dioxin (1 ng/ml) by waterborne exposure, showed no outward evidence of pericardial edema at 1, 2, and 3 weeks post-exposure. Although non-cardiac–related endpoints of dioxin toxicity were evident within a week of exposure, we found that the hearts were morphologically normal even 2 weeks after exposure (Fig. 5A). The TCDD-exposed juveniles began to die at 2–3 weeks post-exposure, yet the hearts appeared unaffected. This lack of effect on heart morphology has also been observed by others in adult fish injected with a lethal dose of TCDD (Zodrow et al., 2004). We conclude that the heart becomes resistant to TCDD toxicity during the maturation from the larval to juvenile stage.
We also looked for more subtle signs of cardiac toxicity. TCDD halts cardiomyocyte proliferation in exposed embryos (Antkiewicz et al., 2005). We therefore used BrdU labeling to determine whether TCDD causes an effect on cell proliferation in juvenile hearts. Juveniles were exposed to TCDD or vehicle at 30 dpf as described above and were grown in the presence of BrdU to label cells entering S-phase. Consistent with the lack of morphological change, we were unable to detect a change in BrdU incorporation in cardiomyocytes exposed to TCDD (Fig. 5B).
The heart defects produced in the embryo are closely associated with decreased circulation (Antkiewicz et al., 2005; Belair et al., 2001; Carney, et al., 2006a; Dong et al., 2004; Henry et al., 1997). We monitored blood flow through caudal fin vessels in juveniles exposed to TCDD at day 30: in these juveniles, the caudal fin vessels remain visible, allowing assessment of blood flow. Even 2 weeks after the exposure at 30 dpf, we observed no TCDD effect on blood flow at the periphery (Fig. 6). This is in marked contrast with the effect in embryos exposed immediately after fertilization, in which blood flow completely stops at 4 days post-exposure (4 dpf) (Belair et al., 2001; Prasch et al., 2006). Belair et al. (2001) also noted decreased blood flow in TCDD-exposed larva (5 dpf) followed by increased mortality at 10 dpf.
Age at Exposure Is Associated With Qualitative Changes in TCDD Toxicity
The decrease in sensitivity to TCDD with age was associated with qualitative changes in toxicity endpoints. Fish exposed just after fertilization demonstrated previously reported signs of toxicity, primarily characterized by heart malformations, pericardial and yolk sac edema, craniofacial malformations, and the absence of swim bladder inflation (Fig. 7). However, when TCDD exposure was delayed until the larval stage at 7 or 12 dpf, the appearance of fish at 4 days post-exposure was grossly normal: swim bladder inflation was normal; jaw malformation and edema were hardly visible. However, the bodies were often distorted, with a tendency for sharp bending in the trunk.
In contrast to the rapid and pronounced edema produced in the embryo, the toxicity in fish exposed at 30 dpf was slower to appear. Juvenile zebrafish (30 dpf) exposed to a lethal concentration of TCDD showed a pronounced wasting syndrome, accompanied by erythema around the gills, shortening of the snout, and erosion of the operculum (Fig. 8). As toxicity progressed, we also observed degeneration of the fins, which was characterized by fragmentation of the unpaired and the caudal fins. In several instances, loss of the pelvic fins was observed (Fig. 8), and less frequently loss of the caudal fin was also observed (data not shown).
TCDD exposure is generally associated with hyperpigmentation in fishes (Kleeman et al., 1988); however, we observed reduced pigmentation in juvenile zebrafish exposed to TCDD at 30 dpf and examined 3 weeks after exposure (Fig. 8). Melanophores, cells constituting the primary lateral stripes, were generally smaller, fewer in number, and more dispersed. The secondary ventral stripe (2V) was almost completely absent in TCDD-exposed juveniles. Interestingly, we also observed a decrease in the number and size of xanthophores that compose the lateral light stripes between the dark stripes.
Although the 30 dpf zebrafish juveniles appeared to be resistant to cardiotoxicity, TCDD continued to produce the stunting and craniofacial defects observed in the embryo (Henry et al., 1997). Juvenile zebrafish exposed to TCDD at 30 dpf, and measured at 14 days after exposure, were shorter than controls treated with DMSO (Fig. 9). As with the embryos, the effects on the craniofacial region were more pronounced than in the rest of the body: the snout length relative to total body length was significantly reduced by TCDD exposure. The results in Figures 4 and 5 show that cardiotoxicity persists into the larvae exposed at 12 dpf cardiotoxicity but is absent in juveniles exposed at 30 dpf. Although cardiotoxicity is gone by 30 dpf, other forms of toxicity found in the embryo persist into later stages. A summary of these changes is presented in Table 1.
|Toxic Responses||Exposed at 0 dpf||Exposed at 30 dpf|
|Decreased peripheral blood flow||Y||N|
|Decreased cardiac myocyte proliferation||Y||N|
|Swim bladder inflation impaired||Y||N|
|Toxic Responses||Exposed at 0 dpf||Exposed at 30 dpf|
|Decreased peripheral blood flow||Y||N|
|Decreased cardiac myocyte proliferation||Y||N|
|Swim bladder inflation impaired||Y||N|
Note. Zebrafish were statically exposed in water to either TCDD (1 ng/ml) or vehicle (DMSO) in 20 ml glass scintillation vials for 1 h either as embryos (within 4 h of fertilization) or as juveniles (30 dpf). After exposure, fish were rinsed in TCDD-free water and placed in housing aquaria as described in the Materials and Methods section.
Transcriptional Responses to TCDD in Embryonic and Juvenile Hearts
TCDD at parts per billion levels rapidly disrupts normal development of the embryonic heart, but delaying the exposure by only 30 days seems to protect the heart entirely. This is both striking and puzzling. Because AHR regulates transcription, one hypothesis is that the juvenile heart no longer responds to TCDD at the transcriptional level: the AHR in the juvenile heart might not be functional. Alternately, TCDD-activated AHR may produce the same initial transcriptional responses in both the embryo and juvenile hearts, but this no longer produces toxic downstream responses in the juvenile heart.
Carney et al. (2006a) used microarrays to examine the transcriptional response to TCDD in hearts from zebrafish embryos at 1, 2, 4, and 12 h post-exposure (hpe). We conducted closely parallel experiments using juveniles exposed at 30 dpf, measuring expression at the 2-h point to capture potential direct targets of AHR and the 12-h point to identify later responses.
In order to compare the response in juvenile hearts to the embryonic transcriptional response, we compiled the array data from our new experiments together with the embryo array data from Carney et al. (2006a). We selected those transcripts that were altered by TCDD by ≥ twofold with p < 0.05 in either the embryonic or juvenile hearts. A heat map generated from these results is shown in Figure 10.
A cluster of genes (Cluster 1) was induced in both juvenile and embryonic hearts. These consisted of AHR gene battery transcripts such as cyp1a and tiparp. This shows that AHR was present, activated by TCDD, and capable of inducing gene expression in both embryo and juvenile hearts. Beyond this cluster, the expression patterns were distinctly different between the embryonic and juveniles hearts. At 2 hpe, a set of transcripts appeared to be upregulated in the juvenile, whereas not in the embryo. At 12 hpe, the transcripts in clusters 4 and 5 were downregulated in the embryo but not in the juvenile.
Venn diagrams showing the overlap of gene sets are shown in Figure 10. At two hpe, only 45 genes were affected in common between the embryonic and juvenile hearts; at 12 hpe, less than 20% of the 317 transcripts affected in the embryo were also among the 164 transcripts altered in the juvenile hearts. The genes represented in the heat map and Venn diagrams are listed in Supplementary table 1.
It has long been thought that fish are more sensitive to the toxic effects of TCDD at early life stages (Peterson et al., 1993; Walker and Peterson, 1994). Using zebrafish, we were able to follow changing sensitivity through growth and development in a detailed, qualitative, and quantitative manner. We found that the LC50 steadily increases with age. In addition, the signs of TCDD toxicity are markedly different between fish exposed immediately after fertilization and those exposed as juveniles or later. Most notably, the acute cardiotoxicity disappears as fish metamorphose into the adult form. Fish develop more impenetrable skin and scales as they age; however, rather than decreasing, we found that TCDD uptake increases somewhat with age at exposure.
The hearts of premetamorphic zebrafish are susceptible to acute TCDD toxicity that alters normal heart shape and function and culminates in death within days of exposure. Zebrafish exposed as embryos displayed the most severe phenotype, but exposure of premetamorphic larvae still produced heart defects, including a smaller ventricle and an elongated atrium. While these structures were affected, other structures once formed did not appear to regress in response to TCDD exposure. Embryonic exposure produces disruption of valve formation and reduced development of the bulbus arteriosus (Grimes et al., 2008; Mehta et al., 2008). Once formed in the larvae, these structures persisted when exposed to TCDD. Exposure to TCDD beyond 5 dpf, a period after valve and outflow tract formation have already occurred, appears to bypass these developmental endpoints of toxicity, and produces a less severe but equally fatal phenotype.
We know that basic heart formation up to approximately 48 hpf is completed even in the presence of TCDD (Antkiewicz et al., 2005). Furthermore, when TCDD is added after a fully looped two chambered heart is formed, cardiotoxicity is still observed (Carney, et al., 2006a). Therefore, TCDD cardiotoxicity cannot be due to disruption of basic pattern formation or early organogenesis. Instead, AHR activation must be affecting a process that is important to the maturation or maintenance of the heart during the period between initial organogenesis and the acquisition of the adult body plan.
An interesting attribute of fish, anurans, and some other vertebrates is their ability to use passive oxygen diffusion to survive throughout the process of embryogenesis and early larval development (Wells and Pinder, 1996). In zebrafish, this mechanism can no longer sustain larvae after ∼14 dpf (Jacob et al., 2002; Rombough, 1998, 2002). Zebrafish with mutations that compromise heart function succumb at approximately 7–10 days (Kopp et al., 2005; Stainier et al., 1996). We hypothesize that the passive diffusion of oxygen allows TCDD-exposed embryos and larvae to survive until reaching a period of crisis between 10 and 14 dpf. Larvae exposed at 17 dpf are completely dependent on an active circulation and had the shortest latency to mortality.
TCDD cardiotoxicity was completely dependent on age at exposure: embryos were exquisitely sensitive; juveniles showed no signs of cardiotoxicity. In other studies, adult zebrafish injected with lethal high doses of TCDD showed no signs of altered heart morphology (Zodrow et al., 2004) (R.E.P. and W.H. unpublished data). We speculate that in the juvenile, TCDD-affected processes may be completed, or perhaps have become unimportant to the organism, and are thus no longer targets for toxicity. An alternative idea is that the initial signals generated by TCDD-activated AHR are different between the embryonic and juvenile heart cells. There is a continuum between these models: the completion of a developmental step might change the molecular responses to AHR activation while simultaneously abolishing a downstream molecular process that is disrupted by TCDD. Since the fish heart tends to increase in efficiency with age (Mirkovic and Rombough, 1998; Schwerte et al., 2006), it is possible that the juvenile heart can compensate for the effects of TCDD in a way that premetamorphic embryos and larvae cannot.
We tried several experiments to pinpoint the age at which fish become resistant to TCDD cardiotoxicity, but at approximately 14 dpf found increasing heterogeneity in response: some fish were clearly affected, whereas others were clearly unaffected. This is a natural period of population heterogeneity. Zebrafish metamorphose into the adult form between 14 and 21 dpf, with some changing earlier than others. During this time, the growing fish separate into distinct size classes; probably due to a self-propagating positive feedback loop of food capture and growth that amplifies initial differences. TCDD-induced cardiotoxicity was consistently associated with the larval form and was not observed in fish that had completed metamorphosis. Sensitivity to cardiotoxicity appears to end at metamorphosis.
Changes in Cardiac Gene Expression With Age
We thought that the transcriptional response to AHR/ARNT activation in the heart might change with age, producing a different pattern of genes responding to TCDD in the heart. Alternatively, the initial transcripts produced by activated AHR/ARNT might remain unchanged but would then no longer produce the same downstream cellular responses.
The former idea is supported by our microarray data: the transcriptional responses to TCDD were quite different between the embryonic heart and the juvenile heart. We conclude that the genes affected by TCDD do change with time. However, a common set of genes remains that responded to TCDD in both the embryonic and juvenile hearts. Most of these represent members of the AHR gene battery but some other genes were also affected in common between the two tissues. Although the overall transcriptional patterns are clearly different, we cannot exclude the possibility that the initial transcript changes that cause toxicity in the embryo still occur in the juvenile, but no longer are toxic.
In the embryo, the number of transcripts altered in the heart by TCDD exposure moves rapidly from 110 at 2 h to 317 at 12 h. This is consistent with the idea that by 12 hpe, secondary responses to the initial toxic insult are underway. At 12 hpe, embryonic exposure produces a marked downregulation of a set of genes in Clusters 4 and 5. This is termed the cell cycle gene cluster (CCGC) because it is highly enriched in cell cycle–related genes (Carney, et al., 2006a; Chen et al., 2008). Associated with the decreased expression of CCGC genes, by 12 hpe cardiomyocyte proliferation in the TCDD-exposed embryonic heart has halted (Carney, et al., 2006a; Chen et al., 2008). In the juvenile heart, we observed neither the decrease in CCGC expression nor a halt in proliferation. In contrast, the response in the juvenile diminished with time after exposure: 205 genes affected at 2 h and 164 at the 12-h point. The prominent downregulation of the CCGC simply did not occur in the juvenile heart. This suggests that the response to AHR activation has changed on a very fundamental level in the juvenile heart cells (Supplementary table 1).
It is unknown whether the downregulation of the CCGC in the embryonic heart is directly controlled by AHR-ARNT or whether it is a secondary response. Cardiac output is reduced in the embryonic heart within 4 h of TCDD exposure, preceding the decrease in CCGC expression. The observed decrease in proliferation could be triggered as the result of compromised cardiac function. In order to understand how CCGC responsiveness is altered in embryonic versus juvenile hearts, we will first need to unravel the mechanism underlying the proliferative defect during early development. AHR activation has been reported to increase or reduce cell proliferation depending on the cellular context, and a direct effect of the AHR on cell cycle regulation in the embryonic heart cannot be ruled out at this time.
This difference in response to AHR activation between tissues is not unprecedented. Beedanagari et al. (2009) showed that in TCDD-exposed HepG2 cells, AHR activation, DNA binding, and recruitment of p300 were not sufficient to induce CYP1B1, whereas in MCF-7 cells, the gene was readily induced by TCDD. Similarly, the simplest explanation for the change in cardiac transcriptional response following TCDD exposure in embryos and juveniles is an epigenetic one; however, instead of a change between tissues, we observe a change in AHR response within the same tissue across time. During the course of normal development, the needs and demands placed on the heart change and the transcriptome responds in a dynamic fashion (Ton et al., 2002) (R.E.P. and W.H., unpublished data). It is likely that these gene expression changes are accompanied by chromatin modifications that alter the occupancy of transcription factors in regulatory regions and also the availability of cofactors required for the productive activation or repression of target genes (Vallaster et al., 2012). Additional support for developmental changes altering AHR responsiveness and differences in proliferation potential is provided by Laiosa et al. (2010). The authors of that work noted stage-specific changes during thymocyte development, wherein late stage thymocytes (TN3 and TN4) show a reduced proliferation following TCDD exposure, whereas early stage thymocytes (TN1 and TN2) show no change in S-phase cells. One must conclude that the genes regulated by TCDD-activated AHR depend on the intracellular environment and that the intracellular environment, including different chromatin regions, can change with development and differentiation.
Because TCDD-activated AHR produces cardiotoxicity in the embryonic but not the juvenile heart, the genes affected only in the embryo could provide clues to the mechanism of embryo cardiotoxicity. We identified 11 transcripts induced in the embryo and not in the juvenile (Supplementary table 1). The list includes several candidates for future investigation.
Although there were major differences in the transcriptional response to TCDD in embryonic and juvenile hearts, there were also similarities. Using the same criteria as above, we identified 23 transcripts that were consistently altered in the hearts of TCDD-exposed animals at both developmental stages (Supplementary table 1). Nine of these were also altered following TCDD exposure in the developing zebrafish jaw (Xiong et al., 2008). We believe the remaining 14 genes represent core AHR target genes that are regulated in a heart specific but age-independent manner.
Other Forms of Toxicity
Early TCDD exposure prevented the formation/expansion of the swim bladder, a response that was uniformly associated with mortality. However, TCDD exposure had no effect on the organ once formed. Similar results were observed for structures such as the bulbus arteriosus and heart valves, which were affected if exposure occurred before, but not after, they were formed.
Exposure of post-metamorphic fish to TCDD produced fragile, degenerating fins, stunted growth, and shortened snouts; responses similar to those previously reported for other fishes (Kleeman et al., 1988). These toxicities are also seen in exposed embryos (Henry et al., 1997; Mathew et al., 2007). This suggests that for some tissues, TCDD continues to have an effect throughout life.
The craniofacial malformations produced by TCDD in early life stage zebrafish are caused by downregulation of sox9b. This in turn reduces expression of downstream Sox9b target genes such as col2a1 to disrupt cartilage development (Yan et al., 2002, 2005). It has recently been reported that exposure to TCDD just after fertilization in the medaka embryo impairs cartilage development of the larval (10 dpf) hypural complex in the caudal fin, concomitant with a decrease in col2a and sox9a gene expression (Dong et al., 2012). The disruption of noncranial cartilage implies a fundamental defect in the formation of cartilaginous structures by TCDD. Whether these same pathways are required for the maintenance of already existing structures in the juvenile fish is unknown. We did not examine gene expression in the craniofacial region or skeleton of TCDD-exposed zebrafish in this work, but it is possible that this mechanism of toxicity remains active throughout the life of the zebrafish and is responsible for the fin, snout, and operculum defects that we report here.
Changing Profile of Toxicity With Age: Distinct Toxic Responses
Within less than a month, juveniles become approximately 25-fold more resistant to TCDD than newly fertilized embryos, endpoints associated with toxicity change, and latency between exposure and death lengthens. In embryos, exposure to TCDD produces a catastrophic cardiotoxic response. This tends to obscure the skeletal deformity and growth defects that also occur at early time points. The cardiotoxicity appears to be a critical factor in the lethality in fish exposed early in development. However, the cardiotoxicity disappears with age, and in adults and juveniles, other responses such as skeletal deformity and wasting become prominent. These responses are slower in progression and may be less sensitive to TCDD. While somewhat speculative, this model explains the changes in sensitivity and latency that we observed with age.
Supported by the National Institutes of Health (NIH) grant R01 ES012716 from the National Institute of Environmental Health Sciences (NIEHS) (to W.H. and R.E.P.) and the University of Wisconsin Sea Grant Institute, National Sea Grant College Program, National Oceanic and Atmospheric Administration, U.S. Department of Commerce grant number NA 16RG2257, Sea Grant Project Numbers R/BT-20 and R/BT-22 (to W.H. and R.E.P.) and NIH grant T32 ES07015 from the NIEHS (to K.A.L.).
We thank Sean Severson, Dorothy Nesbit, J. Jeff Walker, and T. Van Zandt for technical assistance. The contents are solely the responsibility of the authors and do not necessarily represent the official view of the NIEHS, NIH. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.