Craniofacial abnormalities and altered wnt and mmp mRNA expression in zebrafish embryos exposed to gasoline oxygenates ETBE and TAME

Abstract

Gasoline additives ethyl tert butyl ether (ETBE) and tertiary amyl methyl ether (TAME) are used world wide, but the consequence of developmental exposure to these hydrophilic chemicals is unknown for aquatic vertebrates. The effect of ETBE and TAME on zebrafish embryos was determined following OCED 212 guidelines, and their toxicity was compared to structurally related methyl tert-butyl ether (MTBE), which is known to target developing vasculature. LC50s for ETBE and TAME were 14 mM [95% CI = 10 to 20] and 10 mM [CI = 8 to 12.5], respectively. Both chemicals caused dose dependent developmental lesions (0.625 to 10 mM), which included pericardial edema, abnormal vascular development, whole body edema, and craniofacial abnormalities. The lesions were suggestive of a dysregulation of WNT ligands and matrix metalloproteinase (MMP) protein families based on their roles in development. Exposure to 5 mM ETBE significantly (p ≤ 0.05) decreased relative mRNA transcript levels of mmp-9 and wnt3a, while 2.5 and 5 mM TAME significantly decreased wnt3a, wnt5a, and wnt8a. TAME also significantly decreased mmp-2 and -9 mRNA levels at 5 mM. ETBE and TAME were less effective in altering the expression of vascular endothelial growth factor-a and -c, which were the only genes tested that were significantly decreased by MTBE. This is the first study to characterize the aquatic developmental toxicity following embryonic exposure to ETBE and TAME. Unlike MTBE, which specifically targets angiogenesis, ETBE and TAME disrupt multiple organ systems and significantly alter the mRNA transcript levels of genes required for general development.

1. INTRODUCTION

Gasoline oxygenates were originally added to gasoline as octane enhancers following the removal of lead in the 1970s (Ancillotti and Fattore, 1998). Later, oxygenates were used to meet the standards of the Clean Air Act in 1990 (Ahmed, 2001; US EPA, 1990). Methyl tert butyl ether (MTBE) [CH₃OC(CH₃)₃] was a popular additive due to its blending properties and storage stability. Two structurally related chemicals, ETBE [CH₃CH₂OC(CH₃)₃] and TAME [CH₃OC(CH₃)₂CH₂CH₃], are also used as fuel oxygenates. Despite similarities in chemical characteristics (Table 1), the limitations put on the use of MTBE in the United States did not extend to ETBE and TAME, which are still used today (Coons, 2009; van Wezel et al., 2009). Neither, ETBE nor TAME have been extensively studied for their potential as developmental toxicants.

Table 1.

Chemical characteristics of Gasoline Oxygenates MTBE, ETBE, and TAME

Chemical	Structure	Molecular Formula	Molecular Weight (g/mol)	Boiling/Melting Point (C°)	Vapor Pressure (mm Hg)	Solubility (per 100g of H20)
Methyl tert-butyl ether (MTBE)		C5H12O	88.15	55/−108	245	4.8g
Ethyl tert-butyl ether (ETBE)		C6H14O	102.18	73/−94	130	1.2g
Tertiary amyl methyl ether (TAME)		C6H14O	102.18	85/−80	75	1.2g

We previously reported MTBE induces vascular lesions in developing Danio rerio, including pooled blood in the common cardinal vein (CCV), cranial hemorrhages, and abnormal intersegmental vessels (ISV), while other tissues appear to develop normally (Bonventre et al., 2011). The vascular lesions occur following an exposure during the 6-somites to prim-25 stages of development, and are associated with a decrease in mRNA transcripts of vascular endothelial growth factor (vegf) -a and -c and vascular endothelial growth factor receptor 2 (vegfr2) at the 21-somite stage. Disruption of vegf and vegfr during development is detrimental to embryogenesis, as it results in dysregulation of angiogenesis, the growth of blood vessels (Carmeliet et al., 1996; Ferrara et al., 1996; Shalaby et al., 1995). The available aquatic toxicity data on ETBE and TAME are limited to acute toxicity studies on fish, daphnia and algae. The MSDS for ETBE reports the LC50 for adult Cyprinodon variegatus as >2500 mg/L (24.5 mM), as no published studies have provided the data to establish a true 96 hr LC50 for ETBE (Chevron Philips, 2010). Huttunen et al. (1997) reported no toxic effects to Oncorhynchus mykiss treated up to 100 mg/L [0.97 mM] TAME, where as the American Petroleum Institute (1995) reported the LC50 for adult Oncorhynchus mykiss to be 580 mg/L (5.7 mM) TAME. Characterization of ETBE and TAME toxicity to developing aquatic models has not been previously reported.

In mammalian systems, developmental toxicity studies report similar concentrations for effects with MTBE, ETBE, and TAME. Gestational exposure to MTBE via inhalation showed increased fetal resorption in rats and mice at 2500 ppm (28.4 moles/m³ air), with no effect on the dam (Conaway et al., 1985). In another study, mice exposed to 4000 ppm (45.4 moles/m³ air) and 8000 ppm (90.8 moles/m³ air) MTBE exhibited a significant increase in fetal toxicity, post implantation loss, altered sex ratio, craniofacial abnormalities, and skeletal variations, with mild maternal toxicity, even though similar effects were not observed in rabbits (Bevan et al., 1997). Finally, gavaged doses of 500-1500 mg/kg (5.7-17 moles/kg) of MTBE to pregnant Sprague Dawley rats did not result in treatment related effects in the pups (Kozlosky et al., 2012). Developmental inhalation studies for ETBE have not been published. However, rat and rabbit developmental studies report no adverse effects to fetuses exposed prenatally up to 1,000 mg/kg/day (9.8 moles/kg/day) ETBE (Asano et al., 2011, reviewed in de Peyster, 2010). TAME gestational inhalation studies reported NOAELs to be 1500 ppm (14.7 moles/m³ air) for rat fetuses and 250 ppm (2.5 moles/m³ air) for mice fetuses (Welsh et al, 2003). The mice in this study exhibited significantly reduced fetal body weights, increased cleft palate and enlarged lateral cerebral ventricles when exposed to 1500 ppm (14.7 moles/m³ air) and 3500 ppm (24.2 moles/m³ air) TAME. A direct comparison between the developmental effects of the three gasoline oxygenates is complicated due to the various paradigms used and the lack of any study designed to directly compare the three oxygenates.

The aims of the present studies were to (1) characterize developmental lesions associated with ETBE and TAME exposure in the zebrafish embryos, (2) compare and contrast the toxicity of MTBE to that of the replacement oxygenates, ETBE and TAME and (3) examine gene transcript levels of vegf, as well as multiple isoforms from the wnt and mmp families. WNT signaling plays a major role in early embryogenesis in the form of axis orientations, embryonic stem cell pluripotency, germ layer formation, cell fate specifications, and organogenesis (reviewed in Sokol et al., 2011). WNT ligands bind to cell surface receptors and activate signal transduction pathways that result in targeted gene activation. Matrix metalloproteinases (MMPs), in contrast, are a family of zinc-dependent proteins that cleave extracellular matrix (ECM) components, including collagen, laminin, gelatin, and elastin, to allow for tissue restructuring (reviewed in Zitka et al., 2010). Both WNTs and MMPs are important to normal embryonic development. We hypothesized that structurally related chemicals ETBE and TAME would have similar toxic effects on the zebrafish embryo due to similarities in chemical structure and characteristics. Our results demonstrate that although MTBE, ETBE, and TAME differ by only one methyl group, they caused significantly different developmental lesions and targeted different genes in normal embryogenesis pathways.

2. METHODS

2.1 Animal Handling

Transgenic zebrafish fli1-EGFPs (Fli1s), which express enhanced green fluorescent protein in all vascular endothelial cells (Lawson and Weinstein, 2002), were used for all experiments in order to visualize the developing vascular structures. Fli1s were obtained from the Zebrafish International Resource Center. Breeding stocks were housed in an Aquatic Habitat recirculating system under a 14:10 hour light:dark cycle. Water quality was maintained at <0.05 ppm nitrite, <0.2 ppm ammonia, pH between 7.2 and 7.7, and temperature between 26 and 28°C. Husbandry (#03-014) and embryonic exposure protocols (#08-025) were approved by the Rutgers University Animal Care and Facilities Committee. During exposures, embryos were incubated at 25 °C as previously described, and all embryos were selected at the same stage at the beginning of each experiment (Bonventre et al. 2011).

2.2 Chemicals

All chemicals were obtained from Sigma Aldrich: ethyl tert-butyl ether (ETBE) [purity 97.0 %], tertiary amyl methyl ether (TAME) [purity 97.0 %], and methyl tert-butyl ether (MTBE) [purity 99.9 %]. All chemical solutions were made the day of treatment with aerated egg water (60 μg/ml Instant Ocean in double distilled H₂O).

2.3 LC50 Studies

Exposures were performed in triplicate at nominal concentrations of 2.5, 5.0, 7.5, 10.0, 12.5, 15.0, 20.0, 25.0, 30.0, and 40.0 mM ETBE or TAME in sealed glass scintillation vials to avoid volatilization of the chemical (N = 3 vials per concentration, 15 embryos each vial). Exposure began at approximately 3 hours post fertilization (hpf) (approximately the 512 cell stage) and embryos were observed daily for mortality until 120 hpf (day 5 of development). ETBE and TAME LC50 studies were carried out separately and were repeated twice. LC50 and 95% confidence intervals were calculated at 120 hpf using the Litchfield-Wilcoxon method (Litchfield and Wilcoxon, 1949).

2.4 Dose Response Vial Studies

Following OECD 212 guidelines, embryos were exposed individually (N = 20 per treatment) in sealed 4 ml glass vials to static, non-renewal concentrations of 0.625, 1.25, 2.5, 5, and 10 mM nominal concentrations of ETBE or TAME. The concentrations used for the dose response studies were based on sub-lethal concentrations of MTBE (0.625 - 10 mM) established in Bonventre et al. (2011). Embryos were observed under light and fluorescence microscopy daily for the 5 day developmental period characteristic of development at 25°C, and eleutheroembryos (sac-fry) survival was observed at 3 day post hatch (dph). On day 3 and 5 (approximately the Pec Fin and Protruding Mouth stages of development), heart rate was measured for each embryo by manually counting the number of beats for one minute. Since heart rate was not measured in the original MTBE dose response studies, 20 embryos were exposed individually in vials to 10 mM MTBE, and the heart rate was measured as previously described. EC50 and 95% confidence intervals for lesion occurrence were determined using the Litchfield-Wilcoxon method (Litchfield and Wilcoxon, 1949). ETBE and TAME dose response studies were each performed twice.

2.5 Alcian Blue stain for cranio-facial abnormalities

Embryos (N=10) were exposed as described in the dose response vial studies and were collected at hatch (5 days post fertilization). Alcian Blue staining and measurements were carried out as described in Hillegass et al. (2008). Briefly, embryos were fixed in paraformaldehyde and then incubated in 30% peroxide to decrease pigmentation. Alcian Blue staining was performed overnight, and destained in acidified ethanol, before being transferred through a gradient of glycerol, up to 100% glycerol, for craniofacial measurements. Pictures were of each embryo taken with a Scion camera, and craniofacial measurements were performed using Adobe Photoshop and converted to μm.

2.6 Analysis of mRNA expression by Quantitative Polymerase Chain Reaction

The effect of ETBE or TAME on the relative mRNA transcript levels of key genes involved in early development was determined with qPCR. In order to directly compare these studies with the previously published MTBE zebrafish studies (Bonventre et al., 2011), embryos were exposed to 5 mM ETBE or TAME. Due to the lower LC50 for TAME, an additional concentration of 2.5 mM TAME was also used for the mRNA expression studies. Embryos were exposed to a static non-renewal treatment of 0 (control) or 5 mM ETBE, 2.5 mM TAME, or 5 mM TAME until 24 hpf (21-somites) in glass scintillation vials. Three individual samples were set up for each treatment group (3 biological replicates), and each sample consisted of approximately 50 pooled embryos. The experiment was repeated three separate times for each chemical (3 experimental replicates). Studies for the different chemicals were performed independently, each with its own control at the time of the experiment.

Real-time PCR sample preparation was carried out as described in Bugel et al. (2010). Briefly, mRNA was isolated using TRIzol® Reagent (Invitrogen), checked for quality (A260/280), and DNase treated with the DNA-free™ kit (Ambion®). Reverse transcription to produce cDNA was performed with iScript™ (Bio-Rad) and real-time PCR was performed with SYBER green qPCR methods from Bio-Rad. Primer sets were selected to work with the same qPCR protocol: 35 cycles of 95°C for 15 sec and 60°C for 1 minute. All genes were normalized to the housekeeping gene, 28s ribosomal RNA (Delaunay et al., 2000), and the relative mRNA levels were determined using standard curves. The primer sets for mmp-2 and mmp-9 were previously published in Hillegass et al. (2008), and for mmp-13 in Hillegass et al. (2007). Primer sets for vegf-a and vegf-c were previously published in Bonventre et al., (2011). Primer sets for wnt3a and wnt8a were created using IDT Primer Quest. All the primer sets are listed in Table 2.

Table 2.

Primer sequences used for real-time PCR

Gene	Accession #	Sequences	Product size	Reference (if applicable)
vegf-a	NM_131408	fwd 5′-TGCTCCTGCAAATTCACACAA-3′	85 bp	Bonventre et al., 2011
		rev 5′-ATCTTGGCTTTCACATCTGCAA-3′
vegf-c	NM_205734	fwd 5′- ATAAACCACCCTGCGTGTCTGTCT-3′	132 bp	Bonventre et al., 2011
		rev 5′ - TCCTTGCTTGACTGGAACTGTGA - 3′
wnt3a	NM_001007185	fwd 5′ – ATGGTGTCCCGAGAGTTTGCTGAT- 3′	134 bp
		rev 5′- AAGCCCGTGACACTTGCATTTCAG -3′
wnt8a	BC164176	fwd 5′ – GGACTACATGGAACTGAAGG -3′	130 bp
		rev 5′- CTGTCTCAATCCTCCTCTCTT – 3′
mmp-2	NM_198067	fwd 5′-AGCTTTGACGATGACCGCAAATGG -3′	224 bp	Hillegass et al., 2008
		rev 5′-GCCAATGGCTTGTCTGTTGGTTCT-3′
mmp-9	NM_213123	fwd 5′-AACCACCGCAGACTATGACAAGGA-3′	89 bp	Hillegass et al., 2008
		rev 5′-GTGCTTCATTGCTGTTCCCGTCAA-3′
mmp-13	BC065591	fwd 5′-ATGGTGCAAGGCTATCCCAAGAGT-3′	289 bp	Hillegass et al., 2007
		rev 5′-GCCTGTTGTTGGAGCCAAACTCAA-3'

2.8. Statistical analyses

Statistical analyses were performed using the SigmaPlot version 11 computer software package. Differences in lesion occurrence were determined using the Chi—Squared Analysis and Fisher Exact Test. The effect of treatment on heart rate at each time point was determined with a One-way Analysis of Variance (ANOVA) with the Holm-Sidak post hoc for ETBE or TAME, and a t-test for 10 mM MTBE. Differences in craniofacial structures were determined using a One-way ANOVA with the Holm-Sidak post hoc or an ANOVA on the Ranks with the Dunn’s post hoc test when equal variance was not achieved. Outliers were removed using Grubbs Test for Detecting Outliers on GraphPad. The Student t—test was used to determine significance between relative mRNA levels in treatment and control groups for each experimental replicate for MTBE and ETBE. Since the two concentrations of TAME were run concurrently, a One-way ANOVA with the Holm-Sidak post hoc was used to analyze the difference between control and treatments. The probability level for statistical significance was p ≤ 0.05 for all studies.

3. RESULTS

3.1 ETBE and TAME Mortality Studies

The calculated LC50s and confidence intervals for ETBE and TAME were based on nominal test concentrations. Embryos were exposed to 2.5, 5.0, 7.5, 10.0, 12.5, 15.0, 20.0, 25.0, 30.0, 40.0 mM of either ETBE or TAME to establish LC50s for each compound. The LC50 studies were repeated twice for each compound and an LC50 was determined for each study separately (Table 3). The percentages presented here are the average percent of embryos dead in each treatment group from both LC50 studies. By day 1, (24 hrs post exposure), 100% of embryos were dead in the 30 and 40 mM treatments of ETBE. Development in 30 mM ETBE terminated development at the bud stage, while embryos in the 40 mM treatment terminated earlier than that, as indicated by total cell necrosis. At the end of a 5 day exposure to ETBE, the average percent of embryos that died in the 5.0 and 7.5 mM treatments was 20%, 30-35% in 10 and 12.5 mM, 40% in 15mM, greater than 60% in 20 mM, and 100% in 25, 30 and 40 mM ETBE. Less than 10% of embryos died in 2.5 mM ETBE by day 5. The lesions that were observed during the LC50 studies with ETBE included enlarged hearts, cranial hemorrhages, and yolk dysmorphogenesis. The calculated LC50s and confidence intervals (CI) for both studies were similar (Table 3): 16 mM (CI = 14.9 — 17.2) and 12 mM (CI= 10.9 — 13.2).

Table 3.

Calculated LC50 and EC50s for ETBE and TAME

	0-120 hpf ETBE Exposure (mM)	95% CI	0-120 hpf TAME Exposure (mM)	95% CI
Mortality (LC50) Study 1¥	16	14.9 - 17.2	11.2	10.2 - 12.3
Mortality (LC50) Study 2¥	12	10.9 - 13.2	8.7	8.1 - 9.3
Total Lesions (EC50)♯	1.7	1.1 - 2.5	1.6	1.1 - 2.3
Individual Lesions (EC50)♯
Pericardial Edema	9.5	6.1 - 14.8	6	4.1 - 8.8
Pooled Blood in   CCV	4.2	2.9 - 6.1	4.1	3.1 - 5.5
Cranial   Hemorrhage	10	4.9 - 20.6	4.8	2.8 - 8.2
Abnormal ISV	8.5	5.4 - 13.3	3.8	2.2 - 6.4
Edema	7.5	5.1 - 11.1	9	5.9 - 13.8

^¥Group exposure: N =15, 3 replicates per concentration

^♯Individual exposure: Average of 2 individual studies, N=20 for each concentration

The same concentration range was used for the TAME LC50 studies. By day 1, 100% of embryos exposed to 25, 30 and 40 mM TAME were dead. Seventy-five percent of embryos exposed to 20 mM TAME were dead by day 1 and 100% were dead by day 3. At the end of 5 days, less than 10% embryos died in 2.5 and 5 mM, 20% in 7.5 mM, 58% in 10 mM, 80% in 12.5 mM, 91% in 15 mM, and 100 % in 20, 25, 30 and 40 mM TAME. The lesions that were observed during the TAME LC50 studies included stage delay (beginning at 10 mM), craniofacial abnormalities, no circulating blood cells, serpentine body structure, decreased pigmentation, and yolk dysmorphogenesis. The calculated LC50s and confidence intervals (CI) for both studies were similar (Table 3): 11.2 mM (CI = 10.2 — 12.3) and 8.7 mM (CI= 8.1 — 9.3).

3.2 ETBE and TAME Dose Response Studies

The purpose of the dose response studies was to compare the toxicity of ETBE and TAME to developing zebrafish to the previously reported toxicity of MTBE (Bonventre et al. 2011). Embryos were exposed to 0.625 - 10 mM ETBE or TAME, and the lesions associated with exposure were characterized and reported in Tables 3 and 4. Both ETBE and TAME caused a significant increase (p ≤ 0.05) in mortality at 5 days post fertilization (dpf) in the 10 mM treatments with an average of 30.8% and 45%, respectively. Significant increases in lesion occurrence began at 1.25 mM for both ETBE (42.5%) and TAME (45%). The EC50 for all (total) lesions with continuous embryonic exposure is 1.7 mM [95% CI = 1.1 — 2.5] ETBE and 1.6 mM [1.1-2.3] TAME (Table 3).

Table 4.

Percent of embryos exhibiting lesions with ETBE exposure.

	Control	0.625mM ETBE	1.25mM ETBE	2.5mM ETBE	5.0mM ETBE	10.0mM ETBE
All Lesions	10 (± 5)	25 (± 15)	42.5 (± 2.5)*	56.3 (± 3.7)*	90 (± 10)*	85 (± 15)*
Death (by 120hpf)	0 (± 0)	0 (± 0)	5 (± 5)	5 (± 0)	7.5 (± 2.5)	30.8 (± 0.8)*
Hatch (at 120hpf)	97.5 (± 2.5)	95 (± 0)	90 (± 10)	95 (± 2.5)	90 (± 0)	22.9 (± 7.1)*
3dph Survival	97.5 (± 2.5)	96 (± 0)	91 (± 10)	97.5 (± 2.5)	90 (± 0)	2.6 (± 2.6)*
Reduced Circulating RBCs	0 (± 0)	0 (± 0)	2.5 (± 2.5)	10 (± 0)	37.5 (± 7.5)*	65 (± 25)*
Pericardial Edema	2.5 (± 2.5)	0 (± 0)	0 (± 0)	10 (± 5)	27.5 (± 17.5)	52.5 (±7.5)*
Pooled Blood in CCV§	2.5 (± 2.5)	5 (± 5)	10 (± 10)	25 (± 5)	55 (± 25)*	72.5 (± 22.5)*
Cranial Hemorrhage§	2.5 (± 2.5)	20 (± 15)	20 (± 5)	22.5 (± 2.5)	52.1 (± 17.5)*	30 (± 5)*
Abnormal ISVs§	2.5 (± 2.5)	0 (± 0)	10 (± 5)	15 (± 5)	17.5 (± 12.5)	65 (± 10)*
Edema	2.5 (± 2.5)	2.5 (± 2.5)	15 (± 10)	30 (± 10)	27.5 (±17 5)	55 (± 0)*
Yolk Dysmorphogenesis	0 (± 0)	0 (± 0)	0 (± 0)	0 (± 0)	15 (± 10)	65 (± 5)*

Values in the table are percentages (± range) based on the average number of embryos in each category from replicate studies. (N=20)

^*Significantly different from control (P ≤ 0.05). § MTBE-induced vascular lesion

Hatching by day 5 was significantly decreased at 10 mM ETBE where only 22.9% of embryos were hatched, compared to 97.5% hatched in control (Table 4, 5). All concentrations of TAME decreased hatching, but the decrease was statistically significant at both 5 mM (67.5%) and 10 mM (0%). Three day post hatch survival, based on the control hatch time, was significantly reduced with 10 mM of both ETBE (2.6%) and TAME (0%). Embryos exposed to 10 mM TAME did not hatch by 3 dph (8 dpf), the last day of observation. These embryos were severely edemic, dysmorphic, and 90% did not exhibit any signs of circulating red blood cells in the blood (Figure 1f). Exposure to 10 mM TAME also caused a significant stage delay in 100% of the embryos. At 24 hpf, when untreated embryos were at the 21-somite stage, embryos exposed to 10 mM TAME were at the 12-somites, and as a result, were smaller and underdeveloped over the entire development period (5 days). Embryos exposed to 10 mM ETBE also exhibited a significant increase in edema (55% at 10 mM) and a reduction in circulating red blood cells (65%).

Table 5.

Percent of embryos exhibiting lesions with TAME exposure.

	Control	0.625mM TAME	1.25mM TAME	2.5mM TAME	5.0mM TAME	10.0mM TAME
All Lesions	7.5 (± 7.5)	20 (± 15)	45 (± 15)*	70 (± 5)*	92.5 (± 2.5)*	90 (± 10)*
Death (by 120hpf)	2.5 (± 2.5)	0 (± 0)	2.5 (± 2.5)	2.5 (± 2.5)	7.5 (± 7.5)	45 (± 35)*
Hatch (at 120hpf)	92.5 (± 2.5)	87.5 (± 2.5)	85 (± 10)	82.5 (± 7.5)	67.5 (± 2.5)*	0 (± 0)*
3dph Survival	97.5 (± 7.5)	87.5 (± 7.5)	87.5 (± 7.5)	92.5 (± 2.5)	67.5 (± 2.5)	0 (± 0)*
Reduced Circulating RBCs	0 (± 0)	2.5 (± 2.5)	10 (± 0)	25 (± 10)	65 (± 15)*	90 (± 10)*
Pericardial Edema	0 (± 0)	2.5 (± 2.5)	12.5 (± 2.5)	12.5 (± 2.5)	40 (± 7.5)*	72.5 (± 27.5)*
Pooled Blood in CCV§	2.5 (± 2.5)	7.5 (± 2.5)	17.5 (± 2.5)	27.5 (± 2.5)	72.5 (± 2.5)*	27.5 (± 22.5)
Cranial Hemorrhage§	2.5 (± 2.5)	10 (± 5)	25 (± 10)	40 (± 10)*	32.5 (± 5)*	7.5 (± 7.5)
Abnormal ISVs§	0 (± 0)	12.5 (± 2.5)	20 (± 5)	32.5 (± 12.5)*	65 (± 5)*	15 (± 15)
Yolk Edema	0 (± 0)	0 (± 0)	5 (± 0)	10 (± 0)	50 (± 0)*	90 (± 10)*
Yolk Dysmorphogenesis	0 (± 0)	0 (± 0)	0 (± 0)	0 (± 0)	92.5 (± 7.5)*	62.5 (± 37.5)*

Values in the table are percentages (± range) based on the average number of embryos in each category from replicate studies. (N=20)

^*Significantly different from control (P ≤ 0.05). § MTBE-induced vascular lesion

Figure 1.

The effect of ETBE and TAME on morphological endpoints in zebrafish embryos. AC: Representative pictures of embryos at 120 hpf for control (A), 10 mM ETBE (B), and 10 mM TAME (C). D-F: Representative pictures of embryos at 3 dph for control (D), 10 mM ETBE (E), and 10 mM TAME (F). PE = Pericardial Edema, CH = Cranial Hemorrhage, WBE = Whole Body Edema, MH = Miscellaneous Hemorrhage.

Other common lesions induced by ETBE and TAME included pericardial edema (PE), pooled blood in CCV, cranial hemorrhages, abnormal ISVs, and dysmorphic yolks (Table 4, 5). Embryos exposed to 10 mM ETBE exhibited significantly more PE than controls, 52.5% verses 2.5% (background control). TAME caused a significant increase in PE at 5 mM (40%) and 10 mM (72.5%) as compared to control embryos. Both ETBE and TAME caused a significant increase in the pooling of blood in the CCV, which occurs on 2 dpf (Bonventre et al., 2011). Cranial hemorrhages occurred in at least 20% of the embryos exposed to any concentration of ETBE, but was significantly increased in 5 mM (55%). TAME caused a significant increase of cranial hemorrhages (40%) at 2.5 mM. Effects on the heart were also observed with ETBE and TAME, including a significant decrease in heart rate on day 3 and 4 of development. On day 3, average heart rate was significantly reduced by ETBE (5 and 10 mM) and TAME (1.25 — 10 mM). On day 4, average heart rate was similarly reduced by ETBE (5 and 10 mM) and TAME (1.25, 5 and 10 mM). Embryos treated with 10 mM MTBE had a significantly reduced average heart rate as compared to control on day 3, but by day 4, average heart rate in MTBE animals was no longer significantly different from control. In addition to an effect on heart rate, 10-30% of the embryos in 10 mM treatments of ETBE or TAME exhibited tube hearts or enlarged hearts, but the effect were not significant. Morphological abnormalities of the heart were not observed with MTBE treatment. ETBE also caused an increase, though not significant, in hemorrhages in locations other than the common cardinal vein and the cranium, specifically in different regions of the tail (Figure 1e).

3.3 Craniofacial Abnormalities

Abnormal craniofacial structures were observed in the higher concentrations (5 and 10 mM) of the ETBE and TAME dose response studies. In order to quantify the observation, cartilaginous structures in the embryo were stained with Alcian Blue, and the lengths of three cranial structures were measured: interocular distance, long jaw length, and ceratohyal cartilage length. Since craniofacial abnormalities were not observed with MTBE, only 10 mM MTBE treated embryos were stained and measured for comparison. Both ETBE and TAME caused a significant change in the lengths of the long jaw and ceratohyal cartilage (Table 6). In 10 mM TAME treated animals, the formation of both structures was almost completely impeded, with embryos exhibiting only rudimentary and incomplete formation of long jaws and ceratohyal cartilages (Figure 2d). Both structures were also shorter in embryos exposed to 5 mM TAME when compared to controls. At 0.625 mM TAME, embryos exhibited significantly longer ceratohyal cartilage as well as longer long jaws, though the length of the long jaw was not significantly different from control. TAME also significantly decreased interocular distance at 10 mM. ETBE significantly decreased the lengths of the long jaw and the ceratohyal cartilage at both 5 and 10 mM, though to a lesser degree than TAME (Figure 2c). The distribution of the CCL measurements for ETBE and TAME are presented in Figure 3a and b. MTBE did not have an effect on the formation of any of the cartilaginous cranium endpoints used in this study.

Table 6.

Exposure to ETBE and TAME results in craniofacial abnormalities in zebrafish

	mM	Interocular Distance	Long Jaw Length	Ceratohyal Cartilage Length
Control	0	156.5 (± 21.4)	480.3 (± 20.9)	294.8 (± 26.6)
MTBE	10.0	157.8 (± 24.9)	479.6 (± 32.2)	315.3 (± 31.7)
	0.625	151.0 (± 11.3)	462.8 (± 8.9)	292.3 (± 17.4)
	1.25	145.5 (± 7.7)	466.6 (± 15.6)	297.0 (± 8.3)
ETBE	2.5	145.6 (± 5.3)	463.5 (± 6.2)	289.7 (± 13.3)
	5.0	148.4 (± 6.0)	433.3 (± 12.6)*	265.3 (± 8.2)*
	10.0	156.6 (± 10.6)	309.3 (± 33.1)*	170.6 (± 36.3)*
	0.625	172.1 (± 31.2)	506.0 (± 40.5)	342.4 (± 21.2)*
	1.25	152.7 (± 18.9)	480.4 (± 25.3)	308.5 (± 25.6)
TAME	2.5	164.8 (±16.5)	483.8 (± 24.5)	318.1 (±18.0)
	5.0	154.0 (± 24.8)	379.7 (± 59.3)*	227.7 (± 46.6)*
	10.0	95.0 (± 44.7)*	n.a.	n.a.

Numbers represent the average length of the structure (μm) from an N of 10. Significance determined by ANOVA (p ≤ 0.05).

^*Significantly different from control. LJL and CCL for 10 mM TAME are represented as n.a. because the structures were not formed in the embryos and could therefore not be measured.

Figure 2.

Alcian blue stain of craniofacial structures in the 5 dpf embryo. A-D: Representative pictures of embryos at 5 dpf for control (A), 10 mM ETBE (B) 10 mM MTBE (C), and 10 mM TAME (D). Craniofacial measurement description in panel E.

Figure 3.

Box plot of the ceratohyal cartilage lengths all doses measured for (A) ETBE and (B) TAME. * denotes significantly decreased from control, ¥ denotes significantly increased from control.

3.4 mRNA studies

At the 21-somites stage, 5 mM TAME induced a statistically significant decrease in vegf-c transcript levels (1.94 fold change from control). Although not significant, 5 mM ETBE reduced mRNA transcript levels of vegf-a (1.42 fold change from control) and vegf-c (1.69 fold), and 5 mM TAME decreased vegf-a (1.64 fold). Conversely, 2.5 mM TAME increased both vegf -a and -c transcript levels (1.74 and 1.30 fold), though not significantly. In comparison, 5 mM MTBE significantly decreases both isoforms greater than 2 fold (Table 7). ETBE decreased both wnts and all three mmps, but statistically decreased transcript levels of mmp-9 (2.78 fold) and wnt3a (4.07 fold). The expression of both wnt isoforms was significantly decreased by both 2.5 and 5.0 mM TAME. Both mmp-2 and mmp-9 were significantly decreased more than 3.5 fold by 5.0 mM TAME. All three mmps were decreased by 2.5 mM TAME, but none significantly (Table 7). MTBE did not significantly alter the transcript levels of any of the wnt or mmp genes tested. MTBE exposure decreased mmp-9 greater than 2.5 fold, but the effect was not significant (P = 0.081 for all three replicates).

Table 7.

Comparison of the relative fold changes from control at 21-somites.

	5 mM MTBE	5 mM ETBE	2.5 mM TAME	5 mM TAME
vegfa	2.14*	1.42	1.74	1.64
vegfc	2.21*	1.69	1.30	1.94*
wnt3a	1.90	4.07*	2.29*	2.27*
wnt8a	1.86	1.75	3.80*	3.66*
mmp-2	1.03	1.83	1.74	3.72*
mmp-9	2.54	2.78*	1.29	3.54*
mmp-13	1.35	1.89	1.54	1.45

Values in the table are averaged fold changes from 3 studies. Each study consisted of 3 to 4 biological replicates. Arrows indicate an increase or decrease of expression at 21-somites.

^*denotes significantly different from control.

4. DISCUSSION

Historically, chemicals with similar structures were presumed to have similar toxic effects. The hypothesis that ETBE and TAME would be toxic to developing vasculature of zebrafish embryos was based on the characterization of MTBE toxicity in both piscine and mammalian model systems (Longo, 1995; Bonventre et al., 2011; Kozlosky et al., 2012). MTBE, ETBE, and TAME are chemically similar (Table 1), and on a concentration basis, did not have vastly different LC50s in zebrafish embryos. The toxicity of the gasoline oxygenates differed in sublethal effects and target tissues. ETBE and TAME are four times more lipophilic than MTBE, and therefore have the potential to be more readily taken up by the embryo. However, the calculated LC50 for ETBE and TAME were one third and one half that of MTBE, respectively. The greater toxicity observed with ETBE and TAME was due to more than their lipophilicity alone. Similarly, different lesions observed at the same concentrations indicated that the toxicities of the three chemicals resulted from different biochemical mechanisms.

We are the first to report developmental LC50s and effects for ETBE and TAME in an aquatic finfish. The LC50s of MTBE, ETBE, and TAME are within the same range, but TAME had the lowest LC50 and the steepest mortality curve. Based on LC50s alone (Table 3 and Bonventre et al., 2011), TAME (~9.5 mM) and ETBE (~14.2 mM) were more toxic than MTBE (~19.5 mM). While vascular lesions commonly induced by MTBE were observed with ETBE and TAME exposure, other non-vascular lesions were also present (Table 4, 5). ETBE and TAME significantly induced whole body edema, craniofacial abnormalities, and both had a greater effect on cardiac development, based on occurrence of pericardial edema and decreased heart rate. The effective concentrations at which the three chemicals caused the various developmental lesions observed with the different studies are shown in Table 8. Most of the lesions represented within the table were not observed with MTBE in the range of concentrations tested, and were therefore represented as >10 mM MTBE. Based on the calculated LC50s and the effective concentrations, ETBE toxicity was more similar to MTBE, than was TAME. ETBE exposed embryos exhibited dose dependent vascular lesions similar to those of MTBE exposed embryos (Table 3). However, ETBE hemorrhages occurred in parts of the body that were not observed in MTBE exposed embryos, specifically in the medial caudal region (Figure 1e). Furthermore, all the lesions present in TAME exposed embryos, but not with MTBE (edema, yolk dysmorphogenesis, craniofacial abnormalities), were also present with ETBE exposure, though generally at higher concentrations of ETBE.

TABLE 8.

Concentrations at which the gasoline oxygenates induced significant lesions at the doses tested (0.625 to 10 mM)

	MTBE (mM)	ETBE (mM)	TAME (mM)
Developmental Delay	> 10	> 10	10
Delayed Hatch (by 120 hpf)	> 10	10	10
Survival to 3 days post hatch	> 10	10	5
Reduced Circulation	5	5	5
Pericardial Edema	10	10	5
Heart Rate at 72hpf (beats/min)	10	5	1.25
Heart Rate at 96hpf (beast/min)	>10	5	1.25
Pooled Blood in CCV	5	5	5
Cranial Hemorrhages	10	5	2.5
Abnormal ISVs	5	10	2.5
Edema	> 10	10	5
Yolk Dysmorphogenesis	>10	10	5
Craniofacial Abnormalities	> 10	5	0.625, 5#

Based on data from dose response studies.

^#At 0.62 5mM TAME there was a significant increase in craniofacial abnormalities, while at 5 mM TAME there was a significant decrease.

Reduced circulating red blood cells (RBCs) was the only lesion in which MTBE, ETBE, and TAME were equally effective at inducing (5 mM). However, since a reduction in circulating formed elements in the blood can be caused by decreased RBC production, increased RBC depletion, and hemorrhages that remove cells from circulation, the physiological mechanism by which all three induce this lesion may differ, making it a less specific lesion from which to draw conclusions. In contrast, cardiac-specific lesions, including tube hearts and decreased heart rates, were more severe in ETBE and TAME than in MTBE, which only caused a decrease in heart rate on day 3. The effect on the heart development may be a consequence of the severe edema observed with ETBE and TAME. Alternatively, ETBE and TAME may induce a direct affect on cardiac development which could then lead to edema or other secondary effects, as has been previously reported with dioxin (Antkiewicz et al., 2005) and PAHs (Incardona et al., 2004).

The lesion unique to TAME in our studies was developmental delay. TAME significantly delays embryogenesis at 10 mM, while ETBE and MTBE do not. Coupled with the fact that TAME induces all the observed lesions at one-half and one-quarter the concentrations of ETBE and MTBE, and has an LC50 approximately one-half that of MTBE, the conclusion is that TAME is significantly more toxic to developing embryos and poses a potentially greater risk in a spill scenario. The data for the LC50 and dose response studies suggest the order of toxicity to developing zebrafish embryos for the three oxygenates to be TAME > ETBE > MTBE.

TAME induced craniofacial abnormalities in zebrafish embryos at a lower concentration than MTBE or ETBE (Table 6). In the zebrafish, 10 mM MTBE did not cause craniofacial abnormalities (Table 8). It is possible that higher concentrations of MTBE could result in effects on craniofacial development, although based on the morphological observations reported in Bonventre et al. (2011), MTBE did not induce cranial structures deformities in treatments as high as 25 mM MTBE (Bonventre et al., 2011). Similarly, craniofacial abnormalities, including cleft palate, were previously reported for mammalian developmental models with MTBE and TAME exposure, but not with ETBE (Bevan et al., 1997, Welsh et al., 2003; Asano et al., 2011). The craniofacial structures measured in our study are similar to cleft palate formation in mammals in that they involve comparable molecular mechanisms, specifically the involvement of both WNTs and MMPs.

The bimodal effect of TAME on long jaw and ceratohyal cartilage lengths in the embryos, significantly increased at 0.625 mM and significantly decreased at 5 mM, suggests multiple mechanisms or targets of TAME in the developing embryo (Table 6, Figure 3b). The inhibition of growth by TAME at 5 and 10 mM may be mediated, at least in part, by the significant down regulation of wnt3a and wnt8a by both 2.5 and 5 mM TAME at 21-somites (Table 7). While WNTs play multiple roles in development, WNT signaling in the development of the palate well known. Palatopathogenesis has been associated with WNT inhibition, knockouts, and mutations in rodent and human studies. Brugmann et al. (2007) and Mani et al. (2010) identified a role for WNT signaling in the development of vertebrate facial structures by mapping out the regions of WNT activity in the snout of Wnt-reporter TOPgal and BATgal mice. Correlations between single nucleotide polymorphisms (SNPs) in different WNT isoforms and cleft palate in Brazilian, Chinese, and Polish populations have been reported (Mostowska et al. 2012; Menezes et al., 2010; Yao et al., 2010; Chiquet, et al. 2008). Both ETBE and TAME significantly decrease wnt3a, which has been shown in multiple SNP studies to be associated with cleft palate formation.

The decreased mRNA transcripts of mmp-9 by ETBE and both mmp-2 and mmp-9 by 5 mM TAME may also play a role in the cranial cartilage defects. The temporospatial distribution of MMPs, including MMP-2, MMP-9, and MMP-13 have been shown to play a critical role in secondary palate formation during gestational day 12, 13, and 14 in mice (Morris-Wiman et al., 2000). In zebrafish, craniofacial defects induced by glucocorticoids were associated with altered expression and activity of MMP-2, MMP-9, and MMP-13 (Hillegass et al., 2008; Hillegass et al., 2007). Furthermore, the interaction between WNT and MMP pathways may also affect cranial facial formation in the presence of ETBE or TAME. WNT3a protein was shown to upregulate the expression of MMP-2 mRNA in a human cell culture invasion model, a relationship that may be compromised in an embryo exposed to ETBE or TAME, which both significantly decrease mRNA levels of wnt3a (Planutiene et al. 2011).

While MTBE significantly decreases both vegf-a and vegf-c expression, neither mRNA transcript level was significantly reduced by ETBE, and TAME significantly reduced only vegfc. These results further support MTBE’s specific toxicity, and suggest the gasoline oxygenates act via different mechanisms. VEGF is essential to the formation of blood vessels and lymphatic vessels. Multiple splice variants and homologous proteins bind to extracellular portions of tyrosine kinase receptors, VEGFRs (Habeck et al., 2002; Bahary, et al., 2007; Covassin, et al., 2006). All VEGF ligands and receptors are required in assorted combinations for endothelial cell differentiation and migration, but each has varying capacities as factors in vessel growth. VEGF-c is known to play an important role in lymphangiogenesis (Enholm et al., 2001). The significant reduction in vegf-c transcripts at 5 mM TAME, along with the severe edema observed in these animals, may indicate a disruption in lymphangiogenesis. Further studies would be necessary to determine if a direct relationship between TAME and the lymphatic system exists beyond an effect on VEGF-c.

5. CONCLUSIONS

In summary, short term embryo-sac fry toxicity studies (OECD 212) with nominal concentrations of ETBE and TAME (0.625 mM to 10 mM) yielded a dose-response relationship for multiple developmental lesions. At equal molar concentrations, ETBE and TAME were more toxic than MTBE based on the calculated LC50s and EC50s. MTBE toxicity is specific to angiogenesis (Bonventre et al., 2011), while ETBE and TAME toxicity disrupted the development of multiple organ systems, including the heart, skeletal system, and water regulation. ETBE and TAME were less effective in altering the mRNA expression of vegf-a and vegf-c compared to MTBE, which significantly decreases the expression of both genes (Bonventre et al., 2011). The significant decrease of wnts and mmps by ETBE and TAME, but not by MTBE, suggests a possible mechanism for the general developmental dysfunction observed with ETBE and TAME, and further supports the specificity of MTBE as an anti-angiogenic compound. Although the addition of a methyl group does not drastically alter the chemical characteristics between MTBE, ETBE and TAME (Table 1), the toxicity of the three oxygenates on developing zebrafish was different. While MTBE, ETBE and TAME elicit some of the same vascular lesions, the addition and placement of a methyl group in ETBE and TAME resulted in more toxic compounds with different target organ systems. MTBE remains unique in its ability to specifically target the developing vasculature, even when compared to two structurally related chemicals.