2,3,7,8-Tetrachlorodibenzo-p-dioxin exposure disrupts development of the visceral and ocular vasculature

Abstract

The aryl hydrocarbon receptor (AHR) has endogenous functions in mammalian vascular development and is necessary for mediating the toxic effects of a number of environmental contaminants. Studies in mice have demonstrated that AHR is necessary for the formation of the renal, retinal, and hepatic vasculature. In fish, exposure to the prototypic AHR agonist 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induces expression of the AHR biomarker cyp1a throughout the developing vasculature and produces vascular malformations in the head and heart. However, it is not known whether the vascular structures that are sensitive to loss of AHR function are also disrupted by aberrant AHR activation. Here, we report that TCDD-exposure in zebrafish disrupts development of 1) the subintestinal venous plexus (SIVP), which vascularizes the developing liver, kidney, gut, and pancreas, and 2) the superficial annular vessel (SAV), an essential component of the retinal vasculature. Furthermore, we determined that TCDD exposure increased the expression of bmp4, a key molecular mediator of SIVP morphogenesis. We hypothesize that the observed SIVP phenotypes contribute to one of the hallmarks of TCDD exposure in fish – the failure of the yolk sac to absorb. Together, our data describe novel TCDD-induced vascular phenotypes and provide molecular insight into critical factors producing the observed vascular malformations.

1. Introduction

The vasculature transports oxygen, nutrients, hormones, metabolic products, and immune cells to tissues throughout the body (reviewed in Pugsley and Tabrizchi, 2000). Accordingly, proper vascular development and continuous vascular innervation is essential for the health of all organ systems and is necessary for survival. The aryl hydrocarbon receptor (AHR) is a highly conserved ligand-activated transcription factor known to have endogenous functions in vascular development during mammalian organogenesis. Complete loss of AHR function in mice disrupts vascular development in the liver, eye, and kidney (Lahvis et al., 2000; Choudhary et al., 2015). In addition, Ahr null mice have a number of age-related cardiovascular phenotypes including vascular hypertrophy and cardiomyopathy, further suggesting that AHR is important for both cardiovascular development and homeostasis (Fernandez-Salguero et al., 1997; Vasquez et al., 2003). Given the role of AHR in vascular development, it has long been proposed that the developing vasculature is a primary target for xenobiotic compounds that act as exogenous AHR ligands (Hankinson, 1995; Schmidt and Bradfield, 1996; Prasch et al., 2003, 2006). Consistent with this hypothesis, exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD, dioxin), a potent AHR agonist and a global environmental contaminant, induces vascular expression of cytochrome P4501A (cyp1a), a widely accepted biomarker of AHR activation (medaka: Cantrell et al., 1996, lake trout: Guiney et al., 1997, killifish: Toomey et al., 2001, and zebrafish: Andreasen et al., 2002). In zebrafish, TCDD exposure induces cyp1a mRNA in the endothelium as early as 24 hours post fertilization (hpf), preceding any signs of toxicity (Andreasen et al., 2002). Previous research also demonstrated that TCDD exposure produces a number of aberrant vascular phenotypes during development in multiple fish species, including: regression of the yolk vasculature (medaka; Cantrell et al., 1996), retrobulbar capillary hemorrhaging (lake trout; Guiney et al., 2000), decreased blood perfusion in the dorsal midbrain (Dong et al., 2002), inhibited growth and remodeling of the common cardinal vein (zebrafish; Bello et al., 2004), and malformation of the prosencephalic artery (zebrafish; Teraoka et al., 2010).

Although the vasculature throughout the body is composed of endothelial cells, mural cells, and extracellular matrix, the vasculature is not homogenous (Aird, 2012); rather, it is specialized to the organs and tissues it innervates. Correspondingly, the molecular mechanisms mediating vascular development vary between tissues (Aird, 2012). As result of these tissue-specific microenvironments, toxicant exposures can produce different phenotypes in different vascular beds. Given the tissue-specific vascular phenotypes observed in Ahr null mice, we sought to determine whether the vascular structures affected by AHR loss-of-function were also altered by toxicant-induced AHR activation in the zebrafish model. We examined whether TCDD exposure disrupts the development of 1) the subintestinal venous plexus (SIVP), which vascularizes the developing liver, kidney, gut, and pancreas, and 2) the superficial annular vessel (SAV), a central and critical component of the retinal vasculature. We demonstrate that TCDD exposure disrupts development of both the SIVP and the SAV and causes excessive growth and incorrect patterning of the SIVP. In the eye, TCDD inhibited closure of the SAV and disrupted vessel morphology.

Since TCDD exposure also impairs cardiac output, which is an important epigenetic regulator of vascular morphogenesis, we examined whether the observed TCDD-induced vascular phenotypes were secondary to changes in circulation (Carney et al., 2006b; Gebala et al., 2016). We manipulated cardiac function pharmacologically and genetically, and found that the TCDD-induced phenotypes associated with the SIVP development and SAV morphology, with the exception of SAV closure, were not secondary to changes in circulation. We, therefore, examined whether TCDD exposure disrupted the expression of signaling pathways known to be necessary for SIVP development and observed whole embryo mRNA expression upregulation of bmp4, but not acvr1 or vegfaa. Our findings uncover novel TCDD-induced vascular phenotypes and demonstrate that TCDD exposure alters the expression of key molecular mediators of vascular morphogenesis in the SIVP.

2. Materials and methods

2.1. Zebrafish husbandry

Embryos were obtained from adult zebrafish (Danio rerio) housed and maintained according to methods described by Westerfield (2000). The following transgenic lines were used: Tg(flk1:DsRed2)^pd27 (Kikuchi et al., 2011), Tg(fli1a:EGFP)^y1 (Lawson and Weinstein, 2002) and Tg (gata1a:DsRed)^sd2 (Traver et al., 2003). Eggs were collected within 1 hour of spawning and fertilized eggs were placed into a petri dish with egg water (0.03 % Instant Ocean Sea Salts). All procedures involving zebrafish were approved by the Animal Care and Use Committee of the University of Wisconsin-Madison or Brown University, and adhered to the National Institute of Health’s “Guide for the Care and Use of Laboratory Animals.”

2.2. TCDD exposure

At 4 hpf, zebrafish embryos were exposed in egg water to either TCDD (1 ng/mL) or vehicle (0.1 % dimethyl sulfoxide, DMSO) for 1 h in 4 mL glass scintillation vials while gently rocking. Ten embryos were present per ml of dosing solution with a total of 40 embryos in each vial. After 1 h exposure to TCDD or vehicle, embryos were rinsed three times with fresh egg water and placed into petri dishes. Embryos were raised in egg water containing 0.003 % 1-phenyl-2-thiourea (PTU, Sigma) to inhibit formation of pigment. Water changes were made daily.

2.3. Morpholinos and 2,3-butanedione 2-monoxime (BDM) treatment

Cardiac troponin T2 (tnnt2) morpholino (MO) was obtained from Gene Tools (Philomath, OR). Embryos were microinjected with prepared MO solutions (4 ng/embryo) at the 1–2 cell stage, as previously described (Sehnert et al., 2002; Plavicki et al., 2014; Yue et al., 2015). The Gene Tools standard control morpholino (MO) was used as a control. The MO sequences are as follows: tnnt2, 5’ CATGTTTGCTCTGATCTGACACGA 3’ (Sehnert et al., 2002) and control MO, 5’ CCTCTTACCTCAGTTACAATTTATA 3’. Embryos injected with tnnt2 MO were screened for complete absence of heartbeat at 48 hpf, prior to use in experiments.

For experiments involving 2,3-butanedione 2-monoxime (BDM, Sigma), a 15 mM BDM solution was prepared with egg water containing 0.003 % PTU. At 60 hpf, embryos were placed into new petri dishes, with 10 embryos per dish, and maintained in either regular egg water with PTU or in the 15 mM BDM solution until assessment at 72 hpf.

2.4. Live confocal imaging

Embryos were anesthetized in 0.02 % tricaine (MS-222, Sigma) and mounted in glass bottom petri dishes using 1.2 % low melting point agarose (Sigma). Confocal imaging was performed on an inverted Nikon A1R scanning laser confocal microscope and analyzed using NIS-Elements AR4.30 software. To assess developing ocular vasculature, embryos were mounted in lateral orientation and z-stacks were acquired at 54, 60, and 72 hpf. Z-stacks spanned 77–134 μm with z-sections collected at 2.85 μm intervals. To assess the developing SIVP, embryos were mounted in dorsal-lateral orientation on the left and right sides, and z-stacks were acquired at 48, 54, 60, and 72 hpf with z-sections collected at 2.85 μm intervals. Maximum intensity projections and volume depth color-coding were generated using NIS-Elements AR4.30 software and images were processed using Adobe Photoshop.

2.5. Analysis of the subintestinal venous plexus (SIVP)

SIVP development was assessed using z-stacks collected from live confocal imaging of a double transgenic line carrying reporters for the vasculature and erythrocytes (Tg(fli1a:EGFP; gata1a:DsRed)). Quantification of area, length, and number of compartments in the SIVP was modelled after the methods described by Goi and Childs (2016). Area of the SIVP was defined as the area spanning from the posterior cardinal vein (PCV) to the ventral edges of the SIVP, indicated by expression of the endothelial reporter (Tg(fli1a:EGFP)). Length of the SIVP was defined as the perpendicular distance from the PCV to the ventral apex of the SIVP. To evaluate patterning of the SIVP, the number of compartments that formed between secondary vessels in the SIVP were counted. For consistency, area, length, and number of compartments were assessed between the same 7 somites in all samples. Schematic examples are shown in Fig. 2A’–C’.

Fig. 2.

TCDD exposure increases the area, length, and number of compartments in the SIVP. SIVP development was assessed at 48, 54, 60, and 72 hpf in DMSO- and TCDD-treated embryos and larvae. TCDD-exposure increased the average area of the left SIVP (A) and number of compartments (C) at 48–72 hpf. The average length of the left SIVP (B) was significantly greater in TCDD-treated embryos and larvae at 54, 60, and 72 hpf. Significance is indicated with an asterisk (ANOVA, p ≤ 0.05). (A’-C’) Examples of measurements are shown in a DMSO-treated Tg(fli1a:EGFP; gata1a:DsRed) zebrafish. Yellow dashed box shows region of assessment. All images are orientated dorsal-laterally with anterior to the left. Scale bars =100 μm. Abbreviations: posterior cardinal vein (PCV).

2.6. Analysis of the superficial annular vessel (SAV)

SAV development and morphology was assessed using z-stacks collected from live confocal imaging of embryos from the endothelial reporter line Tg(kdrl:DsRed2). A researcher blinded to the age and treatment of samples scored for the incidence of complete SAV closure. To assess morphology of the SAV, the diameter of the SAV vessel was measured at multiple points between the dorsal radial vessel (DRV), nasal radial vessel (NRV), and ventral radial vessel (VRV). Given that the posterior side of the SAV (i.e., VRV-DRV) does not form in all of the samples, only the vessel diameters on the anterior side of the SAV (i.e., DRV-NRV and NRV-VRV) were measured. To control for variability stemming from different developmental stages and/or treatments, we took advantage of the ring-like structure of the SAV to select comparable points for consistently measuring vessel diameter. The length of SAV between the DRV and VRV was arbitrarily divided into 20 sectors by 21 lines that radiated from a “center”. The “center” point was defined as the base of the hyaloid vascular system, where the various branching vessels coalesce into the singular hyaloid artery. NIS-Elements software 3D volume rendering allowed for sample z-stacks to be digitally oriented such that the “center” consistently aligned with the center of the pre-sumptive SAV ring. The point where each of the 21 radiating lines intersected with the SAV was where vessel diameter was measured. From these measurements the average vessel diameter and variation in vessel diameters (indicated by coefficient of variation) were calculated for each sample and grouped by treatment. Schematic examples are shown in Fig. 6A’–B’.

Fig. 6.

TCDD exposure increases the variation in vessel diameters measured along the SAV. The diameter of the SAV vessel was measured at multiple points between the DRV-NRV-VRV in embryos exposed to DMSO or TCDD (1 ng/mL) at 4 hpf. A detailed description of the method for vessel diameter measurements is described in the Materials and Methods section. Representative examples of (A) DMSO- and (B) TCDD-treated fish are shown. (A’, B’) Schematic diagrams of vessel measurements in (A) and (B), respectively. Vessel diameters are aligned in a column to illustrate the variation in diameters along the length of SAV that was assessed. (C) The average SAV vessel diameter is not significantly different between DMSO- and TCDD-treated fish. (D) However, the variation in vessel diameter is significantly greater in TCDD-treated embryos and larvae at 60 and 72 hpf, as indicated by higher coefficients of variation. Significance indicated by asterisk (ANOVA, p ≤ 0.05). All images are orientated laterally with anterior to the left. Abbreviations: superficial annular vessel (SAV), dorsal radial vessel (DRV), nasal radial vessel (NRV), ventral radial vessel (VRV).

2.7. Gene expression analysis

qRT-PCR was performed for DMSO (0.1 %) and TCDD (1 ng/mL) treated zebrafish embryos collected at 72 hpf. Each data point represents of a pool of 20 embryos from independent dosing experiments. Pooled embryos were flash frozen in liquid nitrogen. RNA isolation and purification were performed using the RNeasy Mini Kit (Qiagen). cDNA synthesis was performed using oligo-dT-primed reverse transcription of 100 ng/μl of total RNA using the SuperScript IV First-Strand Synthesis System (Invitrogen). Zebrafish-specific TaqMan probes designed for TaqMan gene expression assays were used to detect the following genes, with 5 ng of starting cDNA: bmp4, vegfaa, acvr1, cyp1a, sox9b, and actb1 (reference gene). Amplification and cycle threshold (Ct) determination based on fluorescent accumulation was performed using the Viia 7 Real-Time PCR System (Applied Biosystems).

2.8. Statistics

One-way analysis of variance (ANOVA) followed by Tukey’s HSD test was used to compare DMSO- and TCDD-treated embryos for incidence of SAV closure, average SAV vessel diameter, variation in SAV vessel diameter, area of SIVP, length of SIVP, and number of SIVP compartments at ages ranging from 48 to 72 hpf. Significance was set at p ≤ 0.05. Student’s t-test was used to compare SIVP area, length, and number of compartments in embryos exposed to BDM and tnnt2 morphants, with their respective controls, at age 72 hpf. Significance was set at p ≤ 0.05. For assessments of ocular vasculature: n = 15 for DMSO- and TCDD-treated embryos; n = 6 for control and BDM-treated embryos; n = 6 for control and tnnt2 morphants. For assessments of SIVP on left and right sides: n = 5–7 for DMSO- and TCDD-treated embryos; n = 4 for control and BDM-treated embryos; n = 4–5 for control and tnnt2 morphants. For qRT-PCR experiments, Ct values were subtracted from the corresponding stable reference gene, actb1, and fold changes represent normalization to the DMSO-treatment group. Statistics were performed using unpaired two-tailed t-tests. All statistical analyses were done using GraphPad Prism statistics software.

3. Results

3.1. TCDD exposure leads to aberrant growth and incorrect patterning of the SIVP

Zebrafish develop a subintestinal venous plexus (SIVP) that is structurally and functionally similar to the vitelline veins in mouse (Goi and Childs, 2016; Crawford et al., 2010). The SIVP consists of a bilateral vascular network that is intimately involved in the absorption of nutrients from the yolk sac during embryogenesis and early larval development. As embryogenesis progresses, the SIVP remodels to provide blood flow to the digestive tract, kidney, and pancreas, as well as contributes to the development of hepatic sinusoids and the portal veins (Goi and Childs, 2016; Isogai et al., 2001).

To investigate whether TCDD exposure disrupts development of the SIVP, zebrafish embryos with fluorescent endothelial [Tg(fli1a:EGFP)] and erythrocyte [Tg(gata1a:DsRed)] reporters were exposed to either DMSO (vehicle control, 0.1 %) or TCDD (1 ng/mL, waterborne) at 4 hpf. As expected, TCDD-treatment upregulated cyp1a mRNA expression indicating AHR activation and downregulated sox9b expression in 72 hpf zebrafish larvae (Supplemental Fig. S1). The development of the right and left SIVP was assessed at 48, 54, 60, and 72 hpf (Fig. 1 and Supplemental Fig. S2). By 48 hpf, endothelial cells in DMSO-treated controls coalesced to form the primitive architecture of the SIVP, which extends ventrally from the posterior cardinal vein (PCV) (Fig. 1A). Over time, the SIVP expands in area and length (Fig. 1B–D), with additional interconnected secondary vessels forming a vascular plexus consisting of several “compartments.” The primary subintestinal vein (SIV), which at this stage runs mostly parallel to the PCV, was also identifiable (Fig. 1A). Blood began to perfuse the SIVP at approximately 60 hpf (Fig. 1C) and by 72 hpf, secondary vessels were remodeled to form a basket-like structure delimited by veins (Goi and Childs, 2016) (Fig. 1D). In TCDD-exposed embryos, the SIVP appeared to sprout normally and spreads ventrally over the yolk (Fig. 1E). However, it quickly expanded to cover an area beyond that of aged-matched controls (Fig. 1F–H). The primary SIV in TCDD-exposed embryos was poorly defined and the secondary vessels appeared disorganized. Additionally, there was an increased number of compartments that developed and persisted between secondary vessels (Fig. 1H).

Fig. 1.

TCDD exposure causes aberrant growth and patterning of the subintestinal venous plexus (SIVP). (A-H) Double transgenic embryos carrying markers for the endothelium (Tg(fli1a:EGFP), green) and erythrocytes (Tg(gata1a:DsRed), red) were exposed to either DMSO (0.1 %) or TCDD (1 ng/mL, waterborne) at 4 hpf. SIVP development was assessed at 48, 54, 60, and 72 hpf. Maximum intensity projections from live confocal imaging of the left SIVP are shown. (A-D) The SIV and SIVP form normally in DMSO-treated control embryos. (E-H) The SIVP in TCDD-treated embryos is overgrown, covering a greater area and distance than similarly aged DMSO-treated controls, and the primary SIV is poorly defined. Representative images are shown (n = 5 to 7). All images are orientated dorsal-laterally with anterior to the left. Scale bars =100 μm. Abbreviations: posterior cardinal vein (PCV) and subintestinal venous plexus (SIVP).

We quantified the observed SIVP phenotypes in TCDD-treated embryos by measuring the area, length, and number of compartments at 48, 54, 60, and 72 hpf (Fig. 2), and compared these endpoints with age-matched DMSO-treated controls. Our approach to quantification was based on the methodology described by Goi and Childs (2016) and is illustrated in a DMSO-treated sample in Fig. 2. At 54, 60, and 72 hpf, we observed a significant increase in the area, length, and number of compartments in the left and right SIVPs in TCDD-treated embryos when compared to DMSO-treated controls (Fig. 2A–C and Supplemental Fig. S2; p ≤ 0.05). The length of the right SIVP was significantly higher in TCDD-treated embryos slightly earlier at 48 (Supplemental Fig. S2; p ≤ 0.05). Together, our findings indicate that both the growth and remodeling of the SIVP were altered by TCDD exposure.

3.2. TCDD-induced effects on the SIVP are not phenocopied by impaired circulation

TCDD exposure results in reduced cardiac output beginning at approximately 60 hpf, and ultimately, culminates in complete circulatory collapse (Antkiewicz et al., 2005). Therefore, we sought to simulate TCDD-induced circulatory impairment by manipulating cardiac contraction (Bartman et al., 2004; Plavicki et al., 2014). Using a tnnt2 morpholino (MO), we genetically inhibited cardiac contraction and assessed how complete absence of circulation affected SIVP development (Fig. 3) (Sehnert et al., 2002). As a complementary approach, we simulated the reduction in cardiac input in TCDD-exposed embryos by pharmacologically inhibiting contraction at 60 hpf using the non-selective myosin ATPase inhibitor 2,3-butanedione monoxime (BDM). Neither approach completely replicated the vascular phenotypes caused by TCDD exposure. The left SIVP in tnnt2 MO-injected embryos was reduced in length and had fewer compartments when compared to MO-injected control embryos (Fig. 3E–F and Table 1; p ≤ 0.05) Overall, there was a trend toward smaller SIVP area and length in BDM-treated and tnnt2 MO-injected embryos when compared to their respective controls, which was in direct contrast to what was observed in TCDD-treated embryos (Table 1). At 72 hpf, the SIVP in control MO embryos was reduced in area compared to the other controls, though circulation and patterning appeared normal (Fig. 3E). Both tnnt2 MO injections and BDM exposure resulted in vessels that lacked identifiable lumens (Fig. 3D, F), whereas the vessel lumens were still clearly present in TCDD-treated embryos (Fig. 3B). We observed the same effects of inhibited or impaired circulation on the development of the right SIVP in embryos (Supplemental Fig. S3 and Table S1; p ≤ 0.05). Our findings indicate that the SIVP phenotypes observed following TCDD exposure are not secondary to TCDD-induced reductions in circulation.

Fig. 3.

TCDD-induced effects on the SIVP are not secondary to TCDD-induced reductions in circulation. The development of the left SIVP was compared at 72 hpf in Tg(fli1a:EGFP; gata1a:DsRed) control larvae (DMSO, no treatment BDM control, or control MO) and experimental larvae (TCDD; 15 mM BDM beginning at 60 hpf, or tnnt2 MO). (A-F) Maximum intensity projections from live confocal imaging are shown. At 72 hpf, the SIVP developed normally in DMSO-treated larvae (A) and larvae with no BDM treatment (C) larvae. The SIVP in TCDD-treated larvae is overgrown and is poorly perfused (B). In contrast, the SIVP in BDM-treated larvae (D) has a similar area, length, and number of compartments as their respective no BDM treatment controls (C). The SIVP in control MO larvae is notably reduced in size compared to other controls (A, C vs. E); however, the SIVP in the tnnt2 MO is even more diminished and appears poorly organized (F). In treatments where circulation was inhibited by BDM or tnnt2 MO the vessels of the SIVP are smooth, condensed, and lack visible lumens (D, F), which is dissimilar to the TCDD-treated SIVP where the lumen of primary and secondary vessels is still visible (B). Images are representative, oriented dorsal-laterally with anterior to left. Scale bars =100 μm. Abbreviations: posterior cardinal vein (PCV).

Table 1.

Effects of TCDD, BDM, or tnnt2 Morpholino Treatment on Morphologic Development of the Left Subintestinal Venous Plexus (SIVP) in Zebrafish Larvae at 72 hpf.

	SIVP Area (103 μm2)		SIVP Length (μm)		SIVP Compartments (number)
DMSO†	48.1	±1.7	182	±5	14	±1
TCDD†	58.9	±1.8*	223	±7*	27	±1*
No BDM‡	55.6	±2.5	190	±7	14	±1
BDM‡	43.5	±6.9	152	±18	13	±1
Control MO‡	31.7	±0.7	126	±4	17	±1
tnnt2 MO‡	26.3	±2.5	95	±11*	11	±2*

^ΔAll values are mean ± SE (n = 4–7).

^*Significantly different from respective control (p ≤ 0.05).

^†Analysis done by ANOVA followed by Tukey’s HSD test.

^‡Analysis done by Student’s t-test.

3.3. TCDD exposure inhibits closure of the superficial annular vessel (SAV) in the eye

To investigate the effects of TCDD exposure on developing ocular vasculature, we used the previously described dosing paradigm and assessed SAV morphology at different stages of development (Fig. 4). We focused on development of the superficial vascular system, which is stereotypically patterned and easily characterized. During normal development, the SAV develops into a ring-like structure that connects flow from three radial vessels (dorsal radial vessel, DRV; nasal radial vessel, NRV; ventral radial vessel, VRV). At 54 hpf, the SAV formed between the DRV-NRV and NRV-VRV (Fig. 4A). Endothelial tip cells extended from the DRV and VRV (Fig. 4A–B, blue arrowheads) and eventually coalesced to form a complete SAV structure by 72 hpf (Fig. 4C, asterisk). In TCDD-treated embryos at 54 hpf, the anterior side of the SAV formed between the DRV-NRV and NRV-VRV (Fig. 4D). Endothelial tip cells were apparent (Fig. 4, blue arrows). However, unlike in DMSO-treated controls, in TCDD-exposed embryos these tips frequently failed to connect and, by 72 hpf, a mature SAV structure was not formed in the majority of TCDD-treated embryos (Fig. 4F). This observation was quantified by having a researcher blinded to both sample age and treatment score for closure of the SAV in DMSO- and TCDD-treated embryos at 54, 60, and 72 hpf (Fig. 4H). The percent of DMSO-treated control embryos with SAV closure increased over time, such that by 72 hpf 93 ± 7% of embryos were scored with successful closure of the SAV. In contrast, by 72 hpf only 47 ± 13 % of TCDD-treated embryos showed successful SAV closure, which was significantly lower than similarly-aged DMSO-treated controls (Fig. 4H; p ≤ 0.05).

Fig. 4.

TCDD exposure inhibits closure of the superficial annular vessel (SAV) in the eye.

Transgenic Tg(kdrl:DsRed2) embryos were exposed to either (A–C) DMSO (0.1 %), or (D–F) TCDD (1 ng/mL, waterborne) and the development of the superficial vascular system was assessed at 54, 60, and 72 hpf. (A–F) Maximum intensity projections of live confocal imaging are shown. kdrl expression is depth color-coded to distinguish between deep vs. shallow structures (see key, G). (A’-F’) Schematics show the developing superficial (orange) and hyaloid (gray) vascular systems in (A–F), respectively. At 52 hpf, radial vessels of the superficial system connect to the dorsal, anterior, and ventral portions of the SAV in both DMSO- and TCDD-treated fish (A, A’; D, D’). The growing tips of the SAV are indicated by blue arrowheads. (B, B’) In DMSO-treated larvae, the growing tips extend and eventually coalesce. (C, C’) At 72 hpf the SAV is complete and forms a ring-like vessel that connects flow between all three radial vessels of the superficial system (asterisk). In contrast, by 72 hpf the SAV is incomplete in TCDD-treated larvae and the vessel tips remain apart (F, F’; blue arrowheads). (H) The incidence of SAV closure is significantly decreased in TCDD-treated larvae at 72 hpf (ANOVA, p ≤ 0.05; significance indicated by asterisk). Representative images are shown (n = 15). All images are oriented laterally with anterior to the left. Scale bars =50 μm. Abbreviations: dorsal radial vessel (DRV), nasal radial vessel (NRV), ventral radial vessel (VRV).

3.4. TCDD exposure disrupts morphological development of the SAV

To further characterize how TCDD exposure alters SAV development, we examined vessel morphology in detail (Fig. 5). High magnification confocal micrographs revealed striking differences in vessel shape and texture. At 72 hpf, smooth endothelial lumens form the SAV in DMSO-treated embryos (Fig. 5A–A”’). In contrast, at 72 hpf, the SAV in TCDD-treated embryos had a rough texture and an irregular shape (Fig. 5B–B”’). Filopodia appeared to extend randomly outwards (Fig. 5B’, B”), while in some regions the lumens appeared highly constricted (Fig. 5B”’). Of note, the TCDD-treated sample in Fig. 5 was an example of a completely formed SAV, and indicates that parallel comparisons could be made of all SAV regions between DMSO- vs. TCDD-treated samples. Even in cases where the SAV closed successfully in TCDD-treated embryos, the vessel was rougher and meandered in shape, which would likely increase turbidity of blood flow and decrease circulatory efficiency.

Fig. 5.

SAV morphology is abnormal in TCDD-treated larvae. (A-B) Maximum intensity projections from live confocal imaging of Tg(kdrl:DsRed2) embryos. kdrl expression is depth color-coded (see Fig. 4G for key). (A’-A”’) High magnification images of the boxes in (A) of different regions of the SAV in a DMSO-treated larva show that the vessel is uniformly smooth throughout and has a clearly defined lumen. (B’-B”’) In contrast, high magnification images of the boxes in (B) of the SAV in a TCDD-treated larva show a rough vessel surface and tortuous shape with varying lumen thickness at different points along the structure. Representative images are shown. All images are oriented laterally with anterior to the left. Scale bars =25 μm. Abbreviations: dorsal radial vessel (DRV), nasal radial vessel (NRV), ventral radial vessel (VRV).

To quantify the morphological changes observed in the SAV of TCDD-treated embryos, a system was devised to measure vessel thickness at comparable, regular intervals along the anterior portion of the SAV (see Materials and Methods for a detailed description). An example of how measurements were taken from a DMSO- and TCDD-treated embryo is shown in Fig. 6A–A’ and Fig. 6B–B’, respectively. The mean vessel diameter between DMSO- and TCDD-treated samples was not significantly different at 54, 60, or 72 hpf (Fig. 6C). However, there was a significant difference in the variation of vessel diameters in each sample (Fig. 6D). The variation in vessel diameters was described using the coefficient of variation (calculated as the standard deviation divided by sample mean, expressed as a percentage). At 60 and 72 hpf, the variation in vessel diameters was significantly higher in TCDD-treated samples compared to age-matched DMSO controls. Together, these findings indicate that TCDD exposure impaired closure of the SAV in the developing superficial vascular system of the eye and altered SAV morphology.

3.5. Inhibiting circulation recapitulates failed SAV closure, but yields fundamentally different SAV morphology

Since blood flow is an important epigenetic factor that influences vascular morphogenesis, we investigated whether the effects observed in the SAV could be recapitulated when circulation was repressed or inhibited (Fig. 7). As with the SIVP experiments, circulation was inhibited pharmacologically using BDM (Bartman et al., 2004; Plavicki et al., 2014) or genetically by injection of tnnt2 MO (Sehnert et al., 2002). At 72 hpf, in all of the control groups, the SAV closed normally and vessels appeared smooth with clearly visible lumens (Fig. 7A, C, E; asterisk). The SAV failed to close in BDM-treated and tnnt2 MO-injected embryos at a frequency comparable to what was observed in TCDD-treated larvae at 72 hpf (Fig. 7B, D, F). However, the SAV vessel morphology in TCDD-treated embryos differed from what was observed in BDM-treated and tnnt2 MO embryos. As described above, the SAV in TCDD-treated embryos appeared rough, winding in shape, with a visible, albeit occasionally constricted, lumen. In contrast, the vessels in BDM-treated and tnnt2 MO-injected embryos appeared smooth, dense, and lacked a defined luminal space (compare Fig. 7B–D, F). In addition, the superficial radial vessels (DRV, NRV, VRV) in tnnt2 MO embryos were severely reduced and, in some cases, completely degenerated (data not shown). Overall, these data suggest that the observed SAV phenotypes cannot be solely attributed to secondary effects from TCDD-induced reductions in circulation.

Fig. 7.

TCDD-induced effects on the SAV are not completely secondary to TCDD-induced depression of circulation. The development of the superficial vascular system in Tg(kdrl: DsRed2) embryos exposed to different treatments was assessed at 72 hpf: (A) DMSO or (B) TCDD exposure at 4 hpf; (C) raised in normal conditions (i.e., No BDM) or (D) raised in 15 mM BDM beginning at 60 hpf; microinjected with (E) Control MO or (F) tnnt2 MO. (A-F) Maximum intensity projections from live confocal imaging are shown with depth color-coding (see Fig. 4G for key). The SAV is closed and developed normally in DMSO-treated, No BDM, and Control MO embryos (A, C, E; asterisk). In contrast, the SAV is does not close in larvae with reduced or absent circulation (B, D, F; blue arrowheads). The morphology of vessels in TCDD-treated embryos is different than in BDM-treated and tnnt2 MO embryos. The TCDD-treated SAV has a rough surface, tortuous shape, and visible lumen (B). In contrast, when circulation was inhibited beginning at 60 hpf by BDM (D) or continuously in tnnt2 MO (F) all vessels in the superficial system developed a smooth texture, appearing dense, and lacking visible lumens. In addition, the hyaloid system is significantly underdeveloped. Images are representative, oriented laterally with anterior to the left. Scale bars =100 μm. Abbreviations: dorsal radial vessel (DRV), nasal radial vessel (NRV), ventral radial vessel (VRV), hyaloid system (H).

3.6. TCDD exposure upregulates mRNA expression of bmp4, but not other key vascular markers vegfaa or acvr1 in whole embryos

As previously described by Mehta et al. (2008), TCDD-treated embryos have increased mRNA expression of bmp4, which is a critical factor in regulating vascular development (Lowery and deCaestecker, 2010). Given this, and the observed TCDD-induced phenotypes in SIVP and SAV development, we examined whole embryo mRNA expression of bmp4, the BMP receptor acvr1, and another key vasculogenesis marker vegfaa (Supplemental Fig. S4) in DMSO- or TCDD-treated 72 hpf embryos. While bmp4 mRNA expression was significantly upregulated as expected, TCDD exposure however did not appear to alter expression of acvr1 or vegfaa. This suggests that the TCDD-induced phenotypes on the developing SIVP and SAV vasculature may not be solely attributed to aberrant expression of typical endothelial growth factor pathways.

4. Discussion

Vascular development is essential for organogenesis, growth, maturation, and survival. TCDD exposure is known to adversely affect the development of vascular structures in multiple fish species (medaka, Cantrell et al., 1996; lake trout, Guiney et al., 1997; killifish, Toomey et al., 2001; and zebrafish, Andreasen et al., 2002). In this study, we demonstrated for the first time that TCDD exposure disrupts development of the SIVP, which innervates the liver, kidney, gut, and pancreas, as well as the SAV, which innervates the eye. Together, our findings indicate that organ-specific vascular structures disrupted by AHR loss-of-function (Lahvis et al., 2000; Choudhary et al., 2015) in mouse are also sensitive to aberrant AHR activation in zebrafish.

4.1. TCDD causes excessive growth and incorrect patterning of the SIVP

Development of the SIVP in zebrafish requires a series of discrete developmental events: angiogenic sprouting, endothelial cell migration, lumen formation, and remodeling (Lenard et al., 2015; Goi and Childs, 2016). The SIVP begins to form at approximately 30 hpf through the coalescence of endothelial cells that have migrated ventrally from the PCV (Goi and Childs, 2016). From here, the primary vessel (or primary SIV) forms and migrates ventrally across the surface of the yolk sac. Interconnecting secondary vessels form within the plexus, which initially generate a honeycomb-like appearance. By approximately 3 dpf, these compartments are remodeled to form the secondary vessels aligned in parallel, generating circulatory loops from the PCV to the primary SIV. In this study, we focused our assessment to the developmental stages prior to the overt divergence of left and right SIVP development that begins after 3 dpf. In DMSO-treated embryos, we observed the primary SIV growing and migrating ventrally (48 hpf), the formation and expansion of compartments between secondary vessels (54–60 hpf), and finally the remodeling of the SIVP (72 hpf). TCDD exposure affected all stages of SIVP development in both the left and right SIVP. As a result, both the left and right SIVP never acquire a more mature functional morphology. Together, the observed SIVP malformations may explain one of the hallmarks of TCDD exposure observed in fish species: the failure of the yolk sac to absorb (Hornung et al., 1998). These malformations may also contribute to later hepatic histopathological changes associated with TCDD-exposure, such as hepatocyte hypertrophy and lipidosis (Spitsbergen et al., 1991; Cantrell et al., 1996; Guiney et al., 1997; Henry et al., 1997; Yamauchi et al., 2006, Zodrow et al., 2004).

Previous reports indicate that common cardinal vein, mesencephalic artery, and the prosencephalic artery are sensitive to TCDD exposure. These structures are formed through vasculogenesis, the process by which endothelial cells come together to form nascent tubes that establish the primitive architecture of the vasculature. Our results indicate that angiogenesis and pruning, which are also essential steps in vascular development, are disrupted by TCDD exposure. The culmination of evidence suggests that endothelial cells are a specific target of TCDD-induced AHR activation. Although it was beyond the scope of this work, future genetic studies with endothelial-specific expression of constitutively-active AHR could be used to investigate the nuance between endothelial versus systemic effects of AHR activation in vascular development.

4.2. TCDD exposure disrupts the morphological development of the SAV

Due to the previously described effects of AHR loss-of-function or aberrant activation on choroidal neovascularization (Choudhary et al., 2015; Takeuchi et al., 2009), we examined the development of the SAV, the central ring-like structure of the superficial ocular system that connects flow between three radial vessels (DRV, NRV, VRV), and the hyaloid system via the hyaloid vein. We found that TCDD exposure significantly inhibited closure of the SAV. Vessel morphology was also affected, including alterations in general shape, texture, and increased variation in vessel diameter. In addition, there was significantly greater variation in SAV vessel diameters within samples of the TCDD-treated group at 60 and 72 hpf. This variability in vessel shape may disrupt laminar flow patterns within the vessel. Fluid modeling of vessels with varying degrees of stenosis indicates that as little as 30 % stenosis of vessel diameter can increase peak blood velocity and wall shear stress (Li et al., 2007). Alterations in shear stress or flow type can also affect mechanotransduction signaling. At present, it is unclear how this may further impact development in embryonic endothelial cells. In adult vascular systems, there is a growing body of evidence implicating the importance of normal blood flow patterns in vascular sprouting, remodeling, maintenance of arterial vs. venous identity, and regulation of Notch signaling (Jones, 2011a,b). It is therefore likely that the observed alterations in embryonic vessel morphology could impact downstream developmental events and have important physiological consequences.

4.3. Potential targets of TCDD and AHR activation in the developing vasculature of the eye and intestine

TCDD induced developmental toxicity is mediated, in part, by aberrant activation of Ahr2 and subsequent misexpression of downstream targets (Carney et al., 2006a, 2006b; King-Heiden et al., 2012; Prasch et al., 2003, 2006). As the majority of the phenotypes observed in this study are not secondary to repressed circulation, we hypothesized that perturbations in vascular signaling pathways contribute to the aberrant vessel patterning phenotypes. We performed whole embryo qRT-PCR analysis for critical factors of vascular development: bmp4, vegfaa, and the BMP receptor acvr1. Bmp4 expression was significantly upregulated at 72 hpf, which is congruent with previous findings that TCDD exposure expands the bmp4 expression domain in the developing heart (Mehta et al., 2008). However, vegfaa and acvr1 mRNA expression were not significantly changed following exposure. Greater interrogation of the various vascular development pathways may help explain the molecular mechanism mediating the observed vascular phenotypes. For example, nuclear translocation of AHR following TCDD binding allows AHR to form a heterodimer with AHR nuclear translocator (ARNT). The AHR:ARNT complex binds to promoters in target genes to alter transcription. However, ARNT can also form heterodimers with other factors, such as hypoxia inducible factor 1 alpha or beta (HIF1a, HIF1b, respectively), which compete with AHR for ARNT binding (Prasch et al., 2004; Fritz et al., 2008). The HIF1a:ARNT pathway induces Vegf expression, but the presence of AHR has been shown to inhibit this process (Fritz et al., 2008). Therefore, the effects of TCDD exposure on the developing vasculature may be through an uncommon pathway or culmination of several dysregulated signaling events.

Previous work has shown that TCDD exposure expands the bmp4 expression domain in the developing heart (Mehta et al., 2008). An increase in BMP expression or increased sensitivity to BMP signaling in the vascular endothelium could encourage excessive growth of the SIVP. Alternatively, expression of various components of the Vegf signaling pathway could be misregulated by TCDD induced Ahr2 activation. For example, while not required for SIVP development, vegfr3/flt4 is also expressed in the SIVP vessels (Koenig et al., 2016). Misexpression of the complimentary ligand, vegfc could therefore promote excessive growth of the SIVP. Since loss of kdrl can reduce the number of cells in the SIVP (Koenig et al., 2016) the opposite may occur if kdrl was over-expressed, or if cues that guide vegfaa/b were increased. These effects could also induce incorrect patterning of the SIVP that was observed in TCDD-treated embryos.

In both the SAV and SIVP, the morphology of vessels in TCDD-treated embryos was altered. Specifically, in the SAV, the vessel appeared rigid, the shape was meandering, and there was significantly higher variation in vessel diameter within each sample. The mechanism by which TCDD exposure may be altering vessel shape and diameter is unclear. In adults, vessel morphology can be influenced by extracellular matrix proteins, vascular smooth muscle cells, endothelial nitric oxide synthase (eNOS), and blood flow (Jones, 2011a). These same factors may not be applicable to embryonic vasculature. Smooth muscle cells differentiate at later stages of development, and various studies have suggested conflicting effects of eNOS induction or inhibition during embryonic development (Jones, 2011a, b). Nevertheless, TCDD has been shown to decrease expression of sox9b, a regulator of extracellular matrix proteins, in the developing heart and jaw of zebrafish embryos (Hofsteen et al., 2013; Xiong et al., 2008). A similar pathway involving mis-regulation of extracellular matrix components via the Ahr2 pathway may be involved in creating the effects on embryonic vessel morphology observed here.

The vessel lumens in TCDD-treated embryos remained intact despite TCDD-induced reductions in circulation, which contrasted sharply with apparent loss of vessel lumens in BDM-treated and tnnt2 MO embryos. This effect was unexpected considering that other effects induced by TCDD exposure on vessel morphology seem to result in non-laminar flow conditions that can lead to vessel degeneration or possibly rupture (Jones, 2011a, b). Work from Zhong et al. (2011) may offer a clue to this paradox. Using small molecule screening, Zhong et al. (2011) discovered that activity of MEK1/2 and Vegfr2 worked in combination to support vessel integrity during zebrafish development. Recent studies highlight that Vegfa-Vegfr2 and BMP signaling pathways are key regulators in SIVP sprouting, growth, and patterning (Koenig et al., 2016; Goi and Childs, 2016). Vegfaa is expressed in the endoderm along the yolk, while its receptors Kdrl and Kdr are expressed in the SIVP vessels (Koenig et al., 2016). Loss of expression of Vegfa homologs (Vegfaa, Vegfab) or Vegfr2 receptors (Kdrl, Kdr) leads to reductions in SIVP growth and mispatterning, with the severity depending on the homolog (Koenig et al., 2016; Goi and Childs, 2016). Here we found that TCDD exposure does not compromise expression of vegfaa, at least when assessed at the whole embryo level. This suggests that the role of Vegfa-Vegfr2 in the SIVP growth is upheld despite TCDD exposure, and that some other mechanism may be contributing to vessel mispatterning. Other studies have indicated two-way crosstalk exists between AHR signaling and MAP kinase pathways (Puga et al., 2009). Although still uncertain, cumulatively these findings suggest a possible mechanism whereby TCDD induced Ahr2 activation supports lumen integrity through altered Vegf and MAP kinase signaling.

5. Conclusions

Our findings demonstrate that early embryonic TCDD exposure disrupts the development of the SIVP and SAV, essential vascular structures that supply blood to the developing liver, kidney, gut, pancreas, and retina. The mammalian counter parts to these vascular structures were previously shown to be affected by Ahr loss-of-function (Lahvis et al., 2000; Choudhary et al., 2015). Our findings indicate that indeed vascular beds that are sensitive to loss of Ahr function are also affected but aberrant Ahr activation. We have presented several potential molecular targets that may be involved in the TCDD-induced phenotypes in the developing SAV and SIVP. Though beyond the scope of this study, future studies have the opportunity to provide a thorough mechanistic understanding of TCDD-induced vascular malformations. Together, the findings from this study emphasizes the need to understand physiological versus pathophysiological Ahr activation and how modulation of Ahr signaling impacts vascular development and health.