Inhibition of Fin Regeneration in Fathead Minnow (Pimephales promelas) by a Potent Synthetic Glucocorticoid and Development of Adverse Outcome Pathway 334

Abstract

Despite structural and functional conservation across vertebrate species, the glucocorticoid receptor has been minimally studied in comparison to other biological targets for endocrine-disrupting compounds in aquatic systems. Because prolonged use of pharmaceutical glucocorticoids in humans has been linked to osteoporosis and impaired bone growth, we hypothesized that the ability of teleost fish to regenerate fins following damage may be inhibited by exposure to synthetic glucocorticoids in the environment. In the present study, we examined fin regeneration following a 7 days waterborne exposure of juvenile fathead minnows (Pimephales promelas) to the synthetic glucocorticoids, fluticasone propionate and dexamethasone. Expression of several biologically relevant gene products (sgk1, tdgf1, runx2a, lef1, shha, and tsc22d3) was measured in paired caudal fin and whole-body tissues. Fluticasone propionate and dexamethasone significantly impaired fin regeneration at measured water concentrations of 2.62 μg/L and 4.62 mg/L, respectively. Changes in gene expression indicated disruption of intercellular communication in the Wnt/β-catenin and bone morphogenetic protein (BMP) signaling pathways after exposure to 4.86 μg/L fluticasone propionate. Upregulation of tsc22d3, a transcription factor responsible for suppression of anti-inflammatory response, may be the plausible cause of repressed cellular signaling. These findings advance the development of adverse outcome pathway 334—Glucocorticoid Receptor Activation Leads to Impaired Fin Regeneration—and elucidate both the mechanistic relationship between activation of the glucocorticoid receptor by fluticasone propionate and inhibition of fin regeneration, which could plausibly reduce individual fitness in aquatic systems.

Graphical Abstract

INTRODUCTION

Historically, efforts to understand the occurrence and ecotoxicity of endocrine-disrupting compounds (EDCs) have predominantly focused on estrogen, androgen, thyroid, and steroidogenesis pathways.^1,2 In contrast, biological responses involving glucocorticoid receptor agonism have been minimally studied in aquatic organisms. We recently examined the global occurrence of synthetic glucocorticoids, used widely as anti-inflammatory medications, in freshwater systems, and identified data gaps in monitoring efforts and aquatic toxicity for these compounds.³ Since the first use of pharmaceutical glucocorticoids in the 1940s, drug discovery and refinement have progressed exponentially, resulting in chemicals with much higher binding affinity and target specificity.^4,5 Recently, we completed a review of existing in vitro analyses and observed in vitro potency of synthetic glucocorticoids to range over 6 orders of magnitude, and identified fluticasone propionate (FCP) as the most potent compound known to occur in aquatic environments at predicted concentrations of as low as 3.27 ng/L.^3,6 We further identified aquatic hazards of synthetic glucocorticoids,^3,6 which highlighted the importance of understanding potential impacts of these substances to nontarget species.⁴

In mammals, it has been well established that synthetic glucocorticoid exposure causes degradation to the skeletal system, with glucocorticoid-induced osteoporosis occurring due to decreased bone formation coupled with increased bone degradation.⁷ Additionally, glucocorticoids impair bone growth and regeneration following a fracture.⁸ Unlike mammals, teleost fishes are able to regenerate amputated limbs–fins–as damage to fins is common throughout their life cycle.⁹ In ecological systems, inhibition or delayed regeneration of fins could plausibly reduce the ecological fitness of an individual, making affected fish less capable of obtaining food, avoiding predation, attracting a mate, and/or migrating within their home range.¹⁰ The molecular events of fin regeneration have been well characterized, and feature processes involving multiple intracellular communications such as sonic hedgehog, bone morphometric, and wnt/β-catenin signaling pathways.^11,12 However, the effects of glucocorticoid exposure on the regenerative process have received less attention.

U.S. EPA’s ToxCast high throughput screening program includes multiple assays capable of detecting glucocorticoid receptor-mediated bioactivity (e.g., ATG_GRE_CIS; ATG_GR_TRANS; NVS_NR_hGR; Tox21_GR_BLA_Agonist) and there are commercial assays available for detecting GR agonism.¹³ However, the ecological relevance of these activities has not been well established through the development of adverse outcome pathways (AOPs) linking GR agonism to adverse effects in aquatic organisms. Thus, we have begun to assemble evidence from the literature and conduct targeted studies to evaluate a hypothesized AOP linking GR agonism to impaired fin regeneration in fish (https://aopwiki.org/aops/334). Initial steps were previously taken to examine the taxonomic domain of applicability for molecular initiation event (MIE) 122, Activation, Glucocorticoid Receptor, by elucidating cross-species susceptibility to these compounds using in silico, in vitro, and in vivo methodologies to characterize potential hazards to aquatic vertebrates.¹⁴ Previous experiments on embryonic zebrafish have shown that several synthetic glucocorticoids inhibit fin regeneration,^15,16 although the specific biochemical and physiological responses leading to this inhibition have not been clearly elucidated. Consequently, these initial efforts to develop an AOP linking GR agonism to impaired fin regeneration relied on limited information from one species (zebrafish) and life stage (dechorionated embryo) at a single nominal treatment level (1 μM) of the tested glucocorticoids that was not analytically verified.

The hypothesized mechanistic connection between GR activation and impaired fin regeneration was expanded and enhanced through consideration of biological plausibility based on existing literature employing mammalian cell lines to examine intracellular signaling^15–17 as well as the detailed understanding of molecular biology and physiology of fin regeneration.^18–20 Collectively, these data led to the hypothesis that exposure to glucocorticoids perturbs the expression of cripto-1 (also called tdgf1 or teratocarcinoma-derived growth factor 1), inhibiting the activin signaling pathway and preventing the regeneration of fin tissue. The objective of the present study was to test this hypothesis by exposing an ecologically relevant model organism, the fathead minnow (Pimephales promelas), to multiple treatment levels of FCP and then examining gene expression and fin regeneration responses to GR agonism. Testing fathead minnows at a juvenile life stage also provided a means to evaluate the hypothesis that the proposed AOP would be relevant to additional species of fish and postembryonic life stages.

MATERIALS AND METHODS

Study Chemicals.

FCP and dexamethasone (DEX) were selected as test chemicals for the present study. As noted above, FCP was the most potent synthetic glucocorticoid detected in environmental samples based on a previous review of available exposure data.⁶ DEX, while less potent, is widely used as a prototypical GR agonist, and in vitro GR-mediated bioactivity is often expressed in dexamethasone-equivalents.^21–23 FCP (CASRN 80474–14–2; ≥98% purity) and DEX (CASRN 50–02–2; >99% purity) were purchased from Sigma-Aldrich (St. Louis, MO) and Cerilliant (Round Rock, TX), respectively. DEX stocks were prepared by directly dissolving the compound in filtered, UV-treated water from Lake Superior. Specifically, pretreatment of Lake Superior water involves passing water into a sand filter, UV treatment, a five-micron bag filter, and a degasser prior to being delivered to the laboratories and culture unit, where water is aerated in a head box before reaching organisms. Due to poor solubility, stock preparation of FCP consisted of dissolving the compound in acetone and evaporating under nitrogen to coat a 21 L glass carboy to increase the surface area and saturation of the compound. Following complete evaporation, filtered, UV-treated Lake Superior water was added to bring chemicals into solution at or near a saturated concentration. Deuterated analytical standards, FCP-D5 (Toronto Research Chemical Inc., North York, ON, Canada) and DEX-D4 (Cayman Chemical, Ann Arbor, MI), were employed for analytical verification of treatment levels by liquid chromatography mass spectroscopy (LC-MS), as further detailed below.

Fathead Minnow Culture.

Juvenile fathead minnows were cultured at the U.S. Environmental Protection Agency Great Lakes Toxicology and Ecology Division (EPA GLTED, ACUP Eco25–10–004). Adults were maintained under a continuous flow (300 mL/min) of filtered, UV-treated Lake Superior water with the following general water qualities and characteristics: temperature 25 ± 1 °C; pH 7.4–8.2; dissolved oxygen 6.4–7.9 mg/L; ammonia <1.0 mg/L; hardness 44–46 mg/L as CaCO₃ and alkalinity 40–44 mg/L as CaCO₃. The photoperiod was 16:8 light/dark under bulbs with a visible light intensity of approximately 800 lm. Adult fish were housed in a tank containing 20 L of water at a density of 40 fish per tank. Fish were fed frozen adult brine shrimp (Artemia) twice daily. Juvenile fish were housed in similar conditions at a density of 100 fish per 20 L of water with a minimum flow rate of 150 mL/min. Juveniles were fed twice daily with newly hatched Artemia twice per day ad libidium.

Fin Regeneration Experiments.

To determine the range of activity for FCP, a range-finding experiment was conducted in which we determined a lowest observed effect concentration (LOEC) for the inhibition of fin regeneration of 8.9 μg/L (Supporting Text S1 and Figures S2 and S3 and Table S2). Using results from the preliminary study, statistical power analysis was performed to determine the effective sample size (n = 12) for the definitive study.²⁴

The definitive experiment was conducted to determine whether exposure to FCP and DEX significantly alters fin regeneration. To initiate the experiment, juvenile fathead minnows (~30 days posthatch) were sedated using a nonlethal tricaine methanesulfonate²⁵ (buffered MS-222 at 100 mg/L; Syndel, Ferndale, WA). The caudal fin of each fish was clipped at the fork, and fish were immediately transferred to a communal container with fresh Lake Superior water for recovery (Figure 1). Every eighth fish was sedated, but not clipped, to serve as the unclipped control group. Following recovery, fish were randomly placed into tanks containing 10 fish per tank at a density of 1 fish/L, and exposed to either FCP (2, 4, 8, 16, and 32 μg/L; designed to bracket our LOEC from the range-finding study) or DEX positive control (4 mg/L) for 7 days. Treatments were performed in duplicate tanks (n = 20 fish per treatment, total), and Lake Superior water was used as a control for both clipped and unclipped fish. Tanks received a continuous flow (44 mL/min) of water. Water quality (temperature, dissolved oxygen) was measured daily, and the fish were fed newly hatched Artemia nauplii twice daily. Stock solutions and tanks were sampled daily for analytical verification of test chemical concentrations. After 7 days, fish were anesthetized using buffered MS-222 prior to measuring weight, and imaging. Images of fish alongside a millimeter scale were taken using a Nikon SMZ1270 stereo microscope (Nikon Instruments Inc., Melville, New York) at 1x zoom. Caudal fin and the remainder of the body (referred to as whole-body in subsequent text) tissues were each collected, flash frozen in liquid nitrogen, and stored at −80 °C for subsequent analyses.

Figure 1.

Unclipped caudal fin (A) indicating where clipping occurred, perpendicular to the body at the fork and (B) a caudal fin directly after clipping on day 0. Every eighth fish was anesthetized but did not undergo fin clipping to serve as an unclipped control group.

Fin measurement was performed on the archived images using ImageJ (Version 1.54).²⁶ A random number generator was used to manipulate the file names of each image for blinded measurement. The scale of measurement was set to 100 pixels/mm using the reference in each photo. Two independent researchers conducted a double-blind analysis recording the standard length of the fish and total area of the caudal fin of each individual (Figure S1, N = 160). Results were derandomized following measurement. To account for variability of fish size affecting the area of the caudal fin, a ratio of caudal fin area/standard length was calculated prior to statistical analysis.

Gene Expression (Reverse Transcriptase–Polymerase Chain Reaction; RT-PCR).

Total RNA was extracted from individual caudal fin and whole-body tissue samples using RNeasy Mini RNA Purification Kits (Qiagen; Hilden Germany) according to the manufacturer’s protocol with DNase digestion. Following extraction, the quantity and quality of RNA were determined using a Nanodrop ND- 1000 (Thermo Fisher Scientific; Waltham, MA). All samples had A_260nm/A_280nm ratios >1.95 and A_260nm/A_230nm ratios >1.53 and were diluted to 5 ng/μL prior to analysis.

Target genes for fin regeneration pathways were selected (Table S1) from previous literature.^{12,15,27–29} Specifically, mRNA sequence information for selected genes was obtained from the National Center for Biotechnology Information (NCBI). Exon boundaries were determined using CLC Sequence Viewer (Qiagen; Hilden, Germany), and segments of conserved sequence information were selected as optimal primer locations. The primer and Taqman probe design were carried out using Integrated DNA Technologies’ (IDT) PrimerQuest Tool. Amplification efficiencies for each assay were evaluated to ensure efficiency between 90 and 110%. Sequence information for each gene-specific primer set is reported in Supporting Information (Table S1).

One-step RT-PCR was carried out using a QuantStudio 6 Pro Real-Time PCR system (Thermo Fisher Scientific; Waltham, MA). Each reaction consisted of 2.5 μL Luna Probe One-Step RT-qPCR 4x Mix (New England Biosciences), 0.5 μL of each assay sample, 4.5 μL of RNase-free water, and 2 μL of template RNA for a final well volume of 10 μL. Final primer and probe concentrations were 900 and 250 nM, respectively. Cycling parameters were carried out as per manufacturers guidelines as follows: (1) carryover prevention 25 °C for 30 s, (2) reverse transcription (RT) at 55 °C for 10 min, (3) initial denaturation at 95 °C for 1 min, followed by 40 cycles of denaturation at 95 °C for 10 s and extension at 60 °C for 1 min. Transcript levels were normalized to the ribosomal protein L8 (rpl8) reference gene using the 2^−ΔΔCT method.³⁰

Exposure Verification.

Water samples taken throughout the experiment were separated with a Thermo Scientific Vanquish (UHPLC) system interfaced with a Thermo Scientific Hypersil GOLD column (50 × 2.1 mm²; 1.9 μm) held at 30.0 °C. Each sample (2 μL) was injected onto the column and separated via gradient elution at 0.4 mL/min using optima grade water premixed with 0.1% formic acid (A) and optima grade Methanol (B) as eluants. Gradient elution was conducted by holding at 5% B for 0.25 min, followed by a gradual increase to 95% B from 0.25 to 4.00 min. B was held at 95% until 5 min, followed by a rapid decrease to starting conditions by 5.01 min, which was maintained through the end of run time at 7 min.

Following chromatographic separation, samples were analyzed with a Thermo Scientific Orbitrap ID-X mass spectrometer with heated electrospray ionization (H-ESI) in the positive mode. Source parameters included 3500 V spray voltage, 50 Arb sheath gas, 10 Arb Aux gas, 1 Arb sweep gas, an Ion transfer tube at 325 °C, and the Vaporizer temp at 350 °C. Orbitrap parameters included 60,000 resolution power, AGC target set to standard, and Injection time set to auto. Full scan from 100 to 800m/z was utilized to detect analytes of interest, and data generated were processed using TraceFinder software v5.0 (Thermo Fisher Scientific, Waltham, MA). Reporting limits were defined as the lowest concentration in the standard curve, 1 μg/L for each compound.

Statistical Analyses.

Statistical analysis of mRNA results and fin regeneration was carried out using GraphPad Prism Software (v10.0.2; Boston, MA). Responses were examined for normality by Shapiro-Wilk’s Test and homogeneity by the Brown-Forsythe test. If the assumptions of normality and homogeneity were met, a one-way analysis of variance (ANOVA) was conducted, followed by a Dunnett’s multiple comparison post hoc test to determine the statistical differences between treatments and the clipped control group. If samples were normally distributed with unequal variance, a Welch’s ANOVA was conducted, followed by a Dunnett’s T3 post hoc test to determine significant differences between treatments and control groups. An α = 0.05 was employed, and significant responses were identified at p < 0.05 (*), although in some instances for highly significant results p < 0.01 (**) or <0.001 (***) was also reported. Additionally, a Spearman’s Correlation test was completed using IBM SPSS (Armonk, NY; v29.0.1.1) to explore relationships among treatment levels and gene expression.

RESULTS AND DISCUSSION

Analytical Verification of Treatment Levels.

Measured mean (SD, n = 8) concentrations of FCP were 2.62 (±0.25), 4.86 (±0.38), 10.27 (±0.76), 16.95 (±2.58), and 34.56 (±4.34) μg/L, and the analytically verified mean concentration of DEX was 4.62 (±1.14) mg/L. Control samples were below the reporting limit (1 μg/L) for both compounds. Due to an error in the delivery system on day four, concentrations of DEX and the highest treatment of FCP were higher than expected. On average, measured concentrations were 18% higher than nominal values, but no treatment groups overlapped in concentration throughout the duration of the study (Figure S4). Full details of measured analytical values can be found in the Supporting Information (Table S3).

Fin Regeneration.

Double-blind analysis of regeneration of caudal tissue identified a significant reduction in fin regeneration (p < 0.001) across all FCP and DEX treatments compared to the clipped control group, and a negative correlation between caudal fin area/standard length and test chemical concentration (ρ = −0.457, p < 0.001; Figure 2). Notable differences in fish fins were visibly apparent with increasing concentrations of the FCP (Figure 3D,E). The unclipped control group was also significantly different from the clipped control (p < 0.001; Figure 3B–C), indicating that the time frame required for total regeneration is longer than our study length. Additionally, we observed that when caudal fins did partially regenerate, malformed rays were present in 75% fish exposed to FCP or DEX (Figure 3D–F) compared to 10% of the clipped control group. Inhibited or improper fin regeneration may greatly impact the ability of a fish to swim.¹⁰ A deficit in swim performance reduces ecological fitness, resulting in a lack of mobility which affects the ability, for example, of an individual to avoid predation, obtain food or prey, or complete a migration path resulting in decreased reproduction rates.³¹ Any or all of these outcomes could contribute to an overall decrease in population. It should be noted that the LOEC for inhibition of fin regeneration for the definitive experiment was lower than that determined in the range-finding experiment, perhaps due to employing a larger sample size. Thus, the no observed effect concentration (NOEC) for the effects of FCP on fin regeneration cannot be reliably determined.

Figure 2.

Ratio of caudal fin area/standard length among control (white), fluticasone propionate (gray), or dexamethasone (black) treatment groups. The Pearson correlation coefficient appears as ρ with a significance level (p) below. Asterisks indicate significant difference compared to the clipped control (***p < 0.001).

Figure 3.

Results from caudal fin imaging of the unclipped control group (A) or after 7 days of regeneration following exposure to either fluticasone propionate (FCP) or dexamethasone (DEX), (B) clipped control, (C) 2.62 μg/L FCP, (D) 16.95 μg/L FCP, (E) 34.56 μg/L FCP, or (F) 4.62 mg/L DEX.

Analysis of Gene Expression.

Previous efforts characterizing glucocorticoid receptor agonism in fathead minnow examined sgk1 (serum/glucocorticoid-regulated kinase 1 isoforms X1–X4) expression in liver tissues.^14,23,27 Somewhat surprisingly, we did not observe a consistent comparable response in sgk1 expression (Figures 4A and 5A) in either caudal fin or whole-body tissue samples. Upregulation in caudal fin tissue was observed in the 16.9 μg/L of FCP treatment (p > 0.05), but no significant observations were made in any other treatment group, including the DEX positive control group. Variability in observations between the various fathead minnow studies may be due to exposure duration or life-stage differences. For example, significant down-regulation of hepatic sgk1 was observed in adult, male fathead minnows following a 96 h exposure to DEX,^14,27 while the present study employed juvenile fathead minnows and recorded observations following 7 days of exposure.

Figure 4.

Gene expression results for (A) sgk1, (B) tdgf1, (C) lef1, (D) run2xa, (E) shha, and (F) tsc22d3 in caudal fin tissue following exposure to fluticasone propionate (FCP) or dexamethasone (DEX). When a significant correlation between concentration and gene expression was present, the Pearson Correlation Coefficient appears as ρ with a significance level (p) below. The red line represents a 0-fold change, and asterisks indicate significant differences compared to the clipped control (* p < 0.05, ** p < 0.01, ***p < 0.001).

Figure 5.

Gene expression results for (A) sgk1, (B) tdgf1, (C) lef1, (D) run2xa, (E) shha, and (F) tsc22d3 in whole body tissue following exposure to fluticasone propionate (FCP) or dexamethasone (DEX). When a significant correlation between concentration and gene expression was present, the Pearson Correlation Coefficient appears as ρ with a significance level (p) below. The red line represents a 0-fold change, and asterisks indicate significant differences compared to the clipped control (* p < 0.05, ** p < 0.01, ***p < 0.001).

To elucidate potential transcriptional mechanisms inhibiting fin regeneration, multiple genes were examined that had been assessed in previous studies.^{12,15,32–34} For example, Garland et al.¹⁵ examined cripto-1 (also referred to as tdgf1) with the hypothesis that expression of cripto-1 is essential for the inhibition of fin regeneration to occur. However, in the current study, no significant differences in tdgf1 expression in either caudal fin or whole-body tissues were observed (Figures 4b and 5b). In caudal fin tissue, there was a nonsignificant negative relationship between tdfg1 and FCP treatments (ρ = −0.346, p = 0.09), and no significant differences (p > 0.05) were detected between the clipped control and treatment levels (Figure 4B). However, a positive dose–response relationship for tdgf1 (ρ = 0.265, p = 0.048) was observed in whole body tissues using Pearson correlation analysis (Figure 5B).

Although no observations of significant induction of tdgf1 were made, the present mRNA analysis with the juvenile fathead minnows was done 7 days postamputation of the fin rather than at 24 h, the time point previously examined in dechorionated zebrafish embryos.¹⁵ Consequently, life stage, exposure duration, and/or species-specific differences may be responsible for the differential observations. The cripto-1 gene codes for a multifunctional embryonic protein that is re-expressed during inflammation, wound repair, and the malignant transformation of tumors in humans.³⁵ In Teleost, cripto-1 function varies between tissue types, although induction does not seem to vary much between species.³⁶ It is possible that the cartilaginous composition of embryonic fin rays compared to bone in juvenile fish may be a factor in observed differences. Previous research in dechorionated embryonic zebrafish also found that induction of cripto-1 was dependent on the structure of the synthetic glucocorticoid tested, primarily the 17C side chain.¹⁵ However, fish were exposed to only a single nominal concentration (1 μM). Thus, it is entirely possible that the structure of the 17C side chain confers more or less potency, but this was not examined.

In addition to tdgf1, expression of lef1 (lymphoid enhancer-binding factor 1 isoforms X1-X2), runx2a (RUNX family transcription factor 2a isoforms X1-X2), and shha (sonic hedgehog signaling molecule a) was examined based on the importance of these genes in various signaling pathways involved in fin regeneration.^11,12 It was hypothesized that one or more of the genes would be upregulated in regenerating fathead minnow tissues. In whole body tissues, significant upregulation in the expression of lef1 (p < 0.05), runx2a (p < 0.001), and shha (p < 0.01) was observed at concentrations of FCP as low as 4 μg/L (Figure 5C–E). Conversely, no dose–response relationships or significant differences in expression of the three genes between clipped control and glucocorticoid-treated fish were noted in caudal fin tissue (Figure 4C–E). Both lef1 and shha modulate and are dependent on Wnt/β-catenin pathways, which also play a major role in epidermal maintenance.¹² Therefore, observed increases in lef1 and shha expression in whole-body tissues (Figure 5C,5E), but not the caudal fin tissue (Figure 4C,E) suggests a potential failure in intracellular communication in regenerating tissue. Furthermore, fin regeneration relies on bone BMP signaling, regulating osteocyte differentiation and proliferation by inducing runx2a transcription.^11,37 The expression of runx2a in whole-body tissues (Figure 5D) in tandem with the lack of expression in caudal fin tissue (Figure 4D) implies a lack of bone morphogenetic protein (BMP) signaling in the epidermal wound.

The significant upregulation of tsc22d3 (TSC22 domain family, member 3 isoforms X1-X2) in caudal fin and whole-body tissues at the lowest concentrations of FCP supports a plausible cause of impaired fin regeneration (Figures 4F and 5F). Increases in tsc22d3 have been associated with intensified cell cycle progression and proliferation of osteocytes in cancerous tumors.^29,38–40 Other studies have also suggested that tsc22d3 may inhibit the expression of inflammatory genes following induction by glucocorticoid receptor agonism.^32–34 In fish, expression of tsc22d3 may inhibit inflammatory response in caudal fin tissue, resulting in a lack of cellular signaling and downstream inhibition of regenerating damaged tissues.

Our results suggest a similarity in biological pathways between bone regeneration in humans and fin regeneration in the fathead minnow. It is well-known that exposure to glucocorticoids alters the growth, regeneration, and degradation of bones in humans and other mammals.^7,8,41,42 Through the use of in silico and in vitro modeling, it has been demonstrated that the structure and likely functionality of the glucocorticoid receptor are highly conserved across vertebrate species.¹⁴ Such observations are potentially relevant to experimentation, because fish have been employed as toxicology models and as animal alternatives for mammalian models due to their lowered financial cost and maintenance, higher fecundity, and anatomic and physiological homology.^43–45

The molecular pathways involved in the inhibition of bone growth and regeneration in humans are not well-defined from a biochemical and physiological perspective. Therefore, the approach of biological “read across” from mammalian systems to aquatic species has been challenging. In the current study, the teleost fathead minnow model is used to enhance understanding of the potential mechanisms underlying the linkage between glucocorticoid receptor agonism and inhibition of fin regeneration. If the underlying mechanisms are relatively well conserved between fish fin regeneration and human bone growth/regeneration, AOP 334 (or at least several of the key events (KEs) and key event relationships (KERs) associated with it) may have relevance to other vertebrate classes, including humans. KEs and KERs related to AOP334 (Activation, Glucocorticoid Receptor Leading to Inhibition, Fin Regeneration) have the potential to support cross-species extrapolation of the effects of glucocorticoid receptor agonists. However, taxonomic differences in functionality for diverse pathways do exist among species.⁴⁶ As a result of gene duplication and evolutionary divergence, teleost fish possess multiple glucocorticoid receptor isoforms and a mineralocorticoid receptor, all of which have an affinity for glucocorticoids.^47–50 Consequently, further research would be needed to fully elucidate taxonomic domains of applicability for this pathway perturbation. Additionally, time-dependent sampling and mRNA analyses could reveal more specific mechanisms in which fin regeneration is being inhibited.

Adverse Outcome Pathway Development.

During initial efforts to develop AOP334, a putative pathway was outlined linking glucocorticoid receptor activation to the inhibition of fin regeneration (Figure 6A). Therefore, prior to initiating the fathead minnow glucocorticoid experiments described herein, our hypothesized AOP334 proposed a KER between cripto-1 expression and Activin signaling; this process is required for regeneration to occur.^12,17 Based on data from the present study, further refinement of AOP334 may improve the description of inhibition of fin regeneration while retaining much of the evidence presented in the initial AOP. Specifically, we suggest the substitution of “Increase, cripto-1 expression”, “Inhibition, Activin signaling”, and their respective upstream and downstream KERs with two new KEs we now believe to be the cause of inhibition; Decrease, Intercellular Signaling Pathways as suggested by the lack of lef1 and shha up regulation, and Decrease, Osteocyte Differentiation as suggested by lack of runx2a (Figure 6A,B).

Figure 6.

Diagrams showing the originally hypothesized adverse outcome pathway (AOP) 334: Activation, glucocorticoid receptor leading to inhibition, fin regeneration, and proposed changes to create key events which will better incorporate results from the present study. The molecular initiating event (MIE), key events (KEs), and adverse outcome (AO) are represented by their designated numbers from the AOP wiki (aopwiki.org).

By broadening specific key events, such as KE1759: “Increase, cripto-1 Expression” and KE1760: “Inhibition, Activin Signaling” to encompass data gathered in the current study, it is possible to propose a larger domain of applicability for AOP334. Specifically, a new KE titled “Decrease, Intercellular Signaling Pathways” would encompass decreases in Activin, Wnt/β-catenin, and BMP signaling previously and presently observed. Expanding the applicability of the KE would also allow for the addition of intermediate or parallel KEs to be included in the AOP such as Suppression of T Cell Activation (KE1702). Glucocorticoid receptor agonism is known to modulate inflammatory response via disruption of nuclear factor kappa B signaling and consequential reduction of neutrophil recruitment.^51,52 While this is a plausible mechanistic cause for the inhibited fin regeneration observed in the present study, additional analyses would be required to confirm the hypothesis. A second KE titled “Decrease, Osteocyte Differentiation” would allow for employment of the prior KE in other pathways and while still maintaining the structural integrity of AOP334. Additionally, the taxonomic and life-stage domains of applicability would expand from a single species at a single life-stage (i.e., embryonic zebrafish) to a full taxonomic infraclass–teleost–and multiple life-stages. These elements within AOP334 can be found in Figure 6B. Further efforts are necessary to complete proposed changes to AOP334 in the AOP Wiki (aopwiki.org), including further elucidating evidence for KE and KERs in multiple species.

Ecological Relevance.

The current study was notable in that it was unsuccessful in determining an NOEC at which exposure to FCP does not inhibit fin regeneration. A previous analysis of the occurrence of the glucocorticoid in surface waters and wastewater effluents revealed that the predicted 95th percentile of its concentration in surface water is 3.27 ng/L³. While this value is an order of magnitude lower than concentrations employed in the current study, environmental glucocorticoids and their respective degradation products exist as complex mixtures that could elicit additive effects.^53–56 In addition, effluent-dominated or effluent-dependent systems are often representative of worst-case scenarios for “down the drain” chemical exposure, including complex mixtures.⁵⁷ As climate change progresses, projections include an increase in arid and semiarid climates coupled with an influx of population transitions to urban areas, which will contribute to these worst-case scenarios becoming more frequent.^58,59 Ergo, further research is justified to develop a NOEC value and better understand the effects of complex mixtures of glucocorticoid agonists and other contaminants on fin regeneration.

Supplementary Material

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.est.5c02446.

Preliminary range finding experiment, fin measurement methods, water quality, gene assay information, and analytical analyses (PDF)

Data Availability Statement

Data are openly available at data.gov under the DOI: 10.23719/1531950.