Evaluation of cross-generational exposure to microplastics and co-occurring contaminants on embryonic and larval behavior in fathead minnows, Pimephales promelas

Experimental overview

Figure 1 shows an overview of the experimental design. To assess the potential for cross-generational effects of MPs, we first exposed male and female P. promelas in the adult (F0) generation to a clean water control (no MPs or EE2), pure MPs (MP_Virgin), or MPs in combination with one of two environmentally relevant concentrations of EE2 (MP_EE2 10: 10 ng/L or MP_EE2 50: 50 ng/L). The concentrations of EE2 used in the study were selected because they are environmentally relevant (Aris, Shamsuddin & Praveena, 2014; Klaic & Jirsa, 2022) and allow for comparison with previous studies (Swank et al., 2022; Carter & Ward, 2024). We designated F0 adults in the clean water control treatment as F0⁻, and adults in the MP treatments as F0⁺; the negative indicating no exposure to MPs or EE2, and the positive indicating that there was exposure to MPs and/or EE2. Fish used for breeding during the experiment were allowed to spawn before exposure to ensure that gamete production occurred during the exposure period. After spawning, the fish were housed separately by sex and underwent one of the four treatments for 30 days. After this period, males and females were paired for up to 14 days and eggs were collected. We recorded several metrics of reproductive success, including the numbers of pairs that produced eggs, the latency to spawning, and clutch size.

The effects of parental exposure were assessed on embryonic behavior at day 4 post-fertilization (dpf) using established metrics for embryonic toxicity testing (von Hellfeld et al., 2023). Immediately after testing, each clutch was split into two approximately equal-sized groups. One half of the clutch did not receive any further exposure to MPs and was labeled as (F1⁻). Given the rapid embryogenesis and larval development of fathead minnows, it is likely that in nature offspring experience water quality conditions similar to those encountered by the breeding population. Therefore, the other half of each clutch continued with the same treatment as the F0 generation and was labeled (F1⁺). For each of the three exposure treatments (MP_Virgin, MP_EE2 10, MP_EE2 50), two experimental groups of larvae were created: F0⁺/F1⁻, wherein only the F0 generation was exposed to MPs but not the larvae; F0⁺/F1⁺, wherein both the F0 and F1 generations were exposed to MPs. Control larvae were designated as F0⁻/F1⁻, wherein neither the F0 nor the F1 fish were exposed to MPs. For F1⁺ larvae, dietary exposure to MPs began within 24 h of hatching and continued for 21 days; these larvae were fed a mixture of Artemia nauplii and MPs. F1⁻ larvae were fed only Artemia nauplii. Subsets of larvae were dissected at 14 and 21 dpf to ensure consumption of MPs. Larvae were tested for both swimming performance and space use in open-field trials at 14 and 21 dph (days post-hatch). Swimming performance is a critical predictor of foraging success and predator evasion (Fuiman et al., 2006; Nowicki, Miller & Munday, 2012), and space use, particularly the propensity for individuals to enter the center of an open area (Carlson & Langkilde, 2013), is a well-established proxy for risky behavior and boldness (Brown & Braithwaite, 2004). We thus selected these behavioral endpoints because of their relevance to organismal fitness under natural conditions. All treatment and welfare procedures were performed in compliance with Ball State University’s Institutional Animal Care and Use Committee (Approval # 1142896-17).

General animal housing and care

A total of 570 fish were used for this study (130 adult breeding minnows (65 males and 65 females), and 440 larval minnows from 34 produced clutches). Prior to the start of the experiment, 6-month-old, reproductively mature fathead minnows (Pimephales promelas) were obtained from an aquatic culturing facility (Aquatic Biosystems, Fort Collins, CO, USA). We used P. promelas for this study because they are ubiquitous in North American freshwater systems (Page & Burr, 2011), have well-defined reproductive and developmental cycles (United States Environmental Protection Agency, 1987), and are extensively used in behavioral ecotoxicology studies (Ankley & Johnson, 2004), including those involving estrogenic EDCs (e.g., Panter, Thompson & Sumpter, 1998). The number of adult minnows used reflected the number of pairings required to obtain a minimum of six genetically unique clutches per treatment (range: 6–11 clutches per treatment), to minimize the likelihood that variation among treatments was biased by parental characteristics. Upon arrival, adult fish were separated by sex and housed in a 530-L living stream (Frigid Units Inc, Living Stream System, Toledo, OH, USA) until use. Fish were kept under a 16:8-h light: dark cycle and at a water temperature of 23 °C to mimic summer breeding conditions. Water quality and the health of adult fish and subsequent offspring were monitored daily. Criteria established for euthanization of adult or larvae before the end of the experiment included evidence of significant injury, or illness as evidenced by symptoms such as lethargy, swimming or floating sideways, non-responsiveness, or laying on the bottom of the tank. However, no fish exhibited signs of injury or illness during the experiment. Fish were fed Artemia nauplii ad libitum twice a day, supplemented with blood worms once a day.

Microplastics, chemical solutions, and treatments

In this study, we used virgin, spherical, transparent polyethylene microplastics (Cospheric, Santa Barbara, CA, USA). We chose polyethylene for this study because it is abundant in surface waters and is frequently consumed by aquatic organisms (Azizi, Khoshnamvand & Nasseri, 2021). The particles ranged in size from 150–180 µm in diameter, which is both environmentally relevant (Conkle, Báez Del Valle & Turner, 2018) and smaller than the maximum prey size consumed by P. promelas larvae (United States Environmental Protection Agency, 1987).

Fish were exposed to one of three dietary MP treatments in this study (i.e., pure MPs (MP_Virgin), MPs associated with 10 ng/L EE2 (MP_EE2 10), or 50 ng/L EE2 (MP_EE2 50)), or to a clean-water control (no MPs or EE2). An EE2-only group was not included for two reasons; first, our primary goal was to determine whether there was evidence of behavioral differences in fish exposed to MPs with and without co-occurring contaminants (Nobre et al., 2024). Second, it was suggested that differences in EE2 exposure pathways (e.g., Al-Ansari et al., 2013) could affect equivalency of exposure of EE2-only vs EE2-MP exposure groups. Estrogen-associated MPs for the MP_EE2 10 and MP_EE2 50 treatments were generated before the start of the experiment following previously described methods (Swank et al., 2022; Carter & Ward, 2024). In brief, appropriate amounts of 1-ug/mL EE2 stock solution were aliquoted into 1-L beakers of aged water. The MPs were placed into 75-micron mesh bags and immersed for 72 h. Solutions were freshly made and replaced every 24 h to account for any chemical degradation. After 72 h, the microplastics were removed from solution and allowed to air dry, after which they were aliquoted into falcon tubes and frozen at −20 °C until use. A similar protocol was followed for the MP_Virgin treatment, except that MPs were immersed in a solution of cleaned, aged water only.

Adult dietary exposure protocol and breeding

Adult minnows were introduced into 6-L tanks in an aquatic animal housing system (Pentair Aquatic Habitats, Apopka, FL, USA). The fish were housed in groups of 3–4, separated by sex. One exposure treatment was conducted at a time to prevent cross-contamination among tanks and because of space restrictions. Treatments were completed in the order control, MP_Virgin, MP_EE2 50, MP_EE2 10 to eliminate the potential for inadvertent residual EE2 exposure by control and MP_Virgin fish. Three exposure replicates per sex (i.e., six exposure tanks in total) were conducted for the control group, three exposure replicates per sex were conducted for the MP_Virgin treatment (six total tanks), seven exposure replicates were conducted for MP_EE2 10 treatment (14 total tanks), and five exposure replicates were conducted for MP_EE2 50 treatment (10 total tanks). Fish undergoing exposure treatment received MPs twice per day via an established dietary exposure protocol (Swank et al., 2022; Carter & Ward, 2024), with exposures separated by at least 7 h to ensure motivation to forage. At the start of each exposure event, the flow of water to the housing unit was turned off, and each tank was provided with two mL of live Artemia nauplii mixed with microplastics at a concentration of 100 MP/L. We selected this level of MP exposure based on reported estimates of MP concentrations in surface water samples from natural waterbodies (Burns & Boxall, 2018). The fish were permitted to forage naturally for 30 min. An air stone in each tank ensured that the microplastics remained in suspension (Carter & Ward, 2024). At the end of 30 min, the air stones were removed, and excess microplastics were removed using a microinvertebrate dip net. The flow of water to the housing unit was restored. Water turnover ensured that any residual remaining microplastics were removed within ∼10 min, which were collected in a 75-micron mesh sock and discarded.

After the 30-day exposure period, one male and fish female fish were paired in each tank and permitted to breed for a maximum of 14 days. Pairs were discarded from the experiment if they did not produce a clutch during this timeframe, and the failure to breed was recorded. Ten breeding pairs were used for each of the control and MP_Virgin treatments, all of which produced eggs. Twenty-eight pairs were used for the MP_EE2 10 treatment (17 of which did not spawn) and 17 pairs were used for the MP_EE2 50 treatment (9 of which did not spawn). The tanks were checked for eggs twice a day, and clutches were immediately removed when found and placed into 800-mL glass jars containing an air stone and 600 mL of aged, aerated water. Parental fish were only used once, to produce one clutch. After breeding, the fish were euthanized via an overdose of MS-222, measured for standard length, and preserved in 95% ethanol. The latency to spawning (i.e., days since pairing) and total number of eggs laid in the clutch were also recorded.

Embryonic development and behavior

Clutches were maintained in clean water at room temperature (25 °C) until hatching and checked twice a day for mortality. All clutches collected from F0 breeding pairs in the control, MP_Virgin, and MP_EE2 50 groups survived to hatching. In the MP_EE2 10 group, five clutches died prior to hatching. Thus, the final number of experimental clutches used was 10, 10, 6, and 8 for the control, MP_Virgin, MP_EE2 10 and MP_EE2 50 groups, respectively. For each clutch we recorded the minimum latency to hatching competence as the number of days from spawning to the time that the first egg hatched. Because researchers were aware of the exposure status of the test subjects, embryonic behavior was assessed at 4 dpf using automated behavioral tracking software (DanioScope, Noldus) to prevent the potential for researcher bias. To begin a trial, 10 embryos from a clutch were placed into a clear, acrylic dish. For a few clutches, only 4 or 5 embryos were used due to a small number of individuals in the clutch. A thin layer of methyl cellulose on the bottom of the dish prevented the embryos from rolling out of view during the trial. Embryos were placed under a stereomicroscope (Stemi 508, Zeiss) and filmed with a microscope camera (Axiocam 208, Zeiss, Oberkochen, Germany) for 5 min. Embryonic activity during the trial was assessed as the percentage of time (activity %) the spent engaged in spontaneous activity (e.g., flexing and rolling within the chorion), and the total number of distinct locomotor bursts performed.

Larval dietary exposure protocol

After embryonic testing and just prior to hatching, each clutch was divided in half and the eggs were permitted to hatch. Beginning on the day of hatching and continuing for 21 days, one half of the F1 larvae from each clutch received continued F0 exposure (F0⁺/F1⁺ group) and the other half were maintained in clean water (F0⁺/F1⁻ group). Larvae were maintained in 800-mL jars equipped with an airstone under the same ambient environmental conditions as during the embryonic phase, and larval densities of F0⁺/F1⁺ and F0⁺/F1⁻ fish in jars were approximately equal for each clutch. Fish in the F0⁺/F1⁺ group underwent exposures twice daily for 30 min. At the start of an exposure, the larvae were provided with 100 uL of live Artemia nauplii mixed with microplastics of the appropriate treatment (MP_Virgin, MP_EE2 10, MP_EE2 50; 100 MP/L concentration). The F0⁻/F1⁻ group received 100 uL of Artemia nauplii only. After 30 min, the excess microplastics were removed using a microinvertebrate dip net and the bottom of each jar was vacuumed to remove debris. In addition, a 30% to 50% water change was performed daily to ensure that water quality remained high. To confirm that larvae ingested microplastics during natural foraging, we dissected a subset of randomly selected larvae (n = 5–29 individuals per treatment) on days 14 and on 21 under a stereomicroscope 30 to 60 min after feeding and counted the number of MPs observed in the reproductive tract.

Larval open-field trials

Larval open-field swimming trials were conducted at 14 and 21 dph. Trials were conducted in a circular, transparent, Plexiglas area (3.5-cm diameter) with 2-mm x 2-mm gridlines on the bottom. The arena walls were tinted black using a marker to reduce outside visual disturbances. The arena was placed on a dimmable LED light board to provide illumination and centered under a GigE camera (Basler, AG, Germany) positioned 13 cm above the surface of the water. Each trial was conducted in a darkened room to minimize visual disturbance to the larvae. At the beginning of each trial, a single larva was placed into the arena containing aged, aerated water and allowed to acclimate for 1–2 min. The activity of the larva was then recorded for 5 min, and swimming behavior was analyzed using automated behavioral tracking software (Ethovision XT, Noldus, Wageningen, the Netherlands, ver. 13.0), which prevents observer bias. No more than five larvae from each clutch were used for trials (range: 1–5), and each larva was only used once. Individuals were immediately euthanized after trials via an overdose of MS-222 and preserved in 95% ethanol. Trials were conducted in a quiet, darkened room at 25 °C to reduce the potential for disturbance.

The arena was divided into two zones for video imaging analysis; a 1.6-cm diameter central zone in the middle of the arena with a zone-exit threshold of 2 mm, and a surrounding outer zone that included the edges of the arena. Spatial measurements were calibrated for each video before analysis using a 2-mm arena grid. Metrics of swimming performance that we measured included the total distance moved during the trial (mm), mean swimming velocity (mm/s), and the duration of trial time spent actively swimming (s). Space use was measured and compared among treatments as the duration of trial time spent in the central zone (s) and the number of visits made to the central zone. At the end of the experiment, all remaining individuals not used in trials were euthanized via an overdose of MS-222.

Statistical analyses

Non-parametric Kruskal–Wallis tests were used to compare differences among groups in the duration of embryonic development, clutch size, and time to spawn. Spearman rank correlations were used to evaluate the relationship between female body length and clutch size, and among behavioral variables. Preliminary analyses indicated that our measures of embryonic behavior were significantly and positively correlated (Spearman r = 0.95, P < 0.001), as were larval behavioral variables for swimming performance (Spearman r = 0.98–0.99, all P < 0.001) and behavioral space use (Spearman r = 0.88–0.92, all P < 0.001) on testing days 14 and 21. We therefore used principal component analysis (PCA) to combine embryonic behavioral variables into one behavioral index score for use in statistical analysis, and similarly used PCA to combine variables for larval swimming performance and space use, respectively, on each of days 14 and 21. Principal component analysis yielded one significant component for each model. The components explained more than 96–98% of the variation in embryonic behavior and larval swimming performance, respectively, and 73–76% of the variation in space use.

Generalized estimating equation models (GEE) were conducted to assess differences among treatments in embryo activity, with PCA score used as the dependent variable and parental treatment set as the fixed factor. Generalized estimating equations were also used to analyze larval swimming performance and space use across treatments. For these analyses, PCA scores were used as the dependent variable and parental treatment and larval exposure (continued exposure after hatching or not) were included as fixed factors, as well as the interaction between these two factors. Data collected on days 14 and 21 were analyzed separately, and all models were fit with a robust estimator for the covariance matrix and an independent working correlation structure. Clutch identification was specified as the subject variable, and individual was specified as the within-subject variable to account for correlations among individuals originating from the same clutch. In cases where significant factor effects were found, least significant difference (LSD) pairwise post-hoc tests were conducted to evaluate the differences among factor levels. No a priori criteria were set for excluding data from analysis and all data were included, with the necessary exception that adult breeding pairs that didn’t produce a clutch were excluded from F1 analysis. Analyses were conducted in R (ver 4.3.1) or SPSS (ver 29, IBM Corp., Armonk, NY, USA).

Preliminary dissections

Preliminary dissections confirmed that larvae did ingest microplastics via the dietary exposure protocol. Microplastics were observed in the gastrointestinal tract of randomly selected larvae on both days 14 and 21. On day 14, 20% (1/5) of larvae in the F0⁺/F1⁺ MP_Virgin group contained MPs; 41% (11/27) of larvae in the F0⁺/F1⁺ MP_EE2 10 group contained MPs; and 40% (6/15) of larvae in the F0⁺/F1⁺ MP_EE2 50 group contained MPs. Of those with MPs, mean (± SD) numbers of particles were as follows: F0⁺/F1⁺ MP_Virgin: 1 ± 0, F0⁺/F1⁺ MP_EE2 10: 2.09 ± 1.51, and F0⁺/F1⁺ MP_EE2 50: 1.67 ± 1.63. On day 21, 56% (15/27) of F0⁺/F1⁺ MP_Virgin larvae had MPs in their digestive tract; 41% (12/29) of F0⁺/F1⁺ MP_EE2 10 larvae contained MPs; and 54% (14/26) of F0⁺/F1⁺ MP_EE2 50 larvae contained MPs at the time of dissection. The average numbers of MPs in the F0⁺/F1⁺ MP_Virgin, MP_EE2 10, and MP_EE2 50 groups on day 21 were 3.8 ± 5.61, 5.08 ± 9.80, and 6.36 ± 7.94 MPs, respectively. We previously confirmed that reproductively mature fish readily ingest MPs during dietary exposures (Swank et al., 2022), so the F0 breeding fish were not dissected for this study.

F0 reproduction and early life history

We did not find any significant effects of exposure on the latency to spawning in F0 breeding pairs (X² = 5.44, df = 3, P = 0.14), clutch size (X² = 7.27, df = 3, P = 0.06), or the duration of development before first evidence of hatching competence (X² = 7.71, df = 3, P = 0.05) (Fig. 2). However, both the spawning latency and the developmental time to hatching competency showed a U-shaped trend (Figs. 2A, 2C). On average, the time to spawning for reproductively mature F0 fish in the control, MP_Virgin, MP_EE2 10, and MP_EE2 50 treatments was 7.6 ± 3.7 days, 5 ±  4.2 days, 4.3 ± 4.3 days, and 9.9 ± 6.9 days. The duration of development to the earliest evidence of hatching competence in the control, MP_Virgin, MP_EE2 10, and MP_EE2 50 treatment groups was 6 ± 1.2 days, 5 ± 1.2 days, 4.7 ± 0.8 days, and 6 ± 1.1 days, respectively. No significant correlations were found between female standard length and clutch size for any treatment (F0 control: r = −0.26, P = 0.47; MP_Virgin: r = −0.02, P = 0.96; MP_EE2 10: r = −0.26, P = 0.61; MP_EE2 50: r = −0.54, P = 0.16). However, clutch size showed a linear, negative, dose-dependent trend (Fig. 2B), with clutches in the control group being on average approximately 26% larger than those in the MP_Virgin group, and 31% and 58% larger than clutches in the MP_EE2 10 and MP_EE2 50 groups, respectively.

Embryonic activity

Sample sizes for embryos tested from the four treatments are as follows: n = 94, 83, 55, 70 embryos from the F0⁻ control group (from 10 clutches), F0⁺ MP_Virgin (eight clutches), F0⁺ MP_EE2 10 (six clutches), and F0⁺ MP_EE2 50 (eight clutches). We found a significant overall main effect of parental treatment on embryonic activity (X² = 18.88, df = 3, P < 0.001). post-hoc tests indicated that embryos whose parents were exposed to MP_EE2 10 had reduced locomotor activity compared to offspring from parents in the control, MP_Virgin and MP_EE2 50 groups (all P < 0.007; Fig. 3).

Larval swimming performance

Parental (F0⁺) exposure to MPs significantly influenced swimming behavior of F1 offspring on both day 14 (X² = 11.46, df = 3, P = 0.009) and day 21 (X² = 19.44, df = 3, P < 0.001). Pairwise post-hoc tests revealed that on both days, F0⁺ larvae from MP-exposed parents swam further, faster, and spent more time engaged in swimming activity than larvae from control parents (F0⁻) (Figs. 4A, 4B). The magnitude of difference between control and exposed fish was generally more pronounced at day 21, with offspring from MP-exposed parents swimming approximately 35% to 43% farther and 35% to 42% faster compared to non-exposed embryos during performance trials (F0⁺/F1⁻ and F0⁺/F1⁺ groups averaged for each MP treatment). We did not detect a significant effect of continued post-hatch larval exposure (F1⁺) on swimming performance at either day 14 (X² = 0.73, df = 1, P = 0.39) or day 21 (X² = 1.44, df = 1, P = 0.23), or evidence of an interaction between these two factors (day 14: X² = 1.33, df = 2, P = 0.52; day 21: X² = 0.06, df = 2, P = 0.97), suggesting few additive effects of indirect exposure during gametogenesis and direct exposure after hatching.

Larval space use

We did not find a significant effect of parental (F0⁺) exposure on how individuals used the arena space on either day 14 (X² = 6.89, df = 3, P = 0.08) or day 21 (X² = 5.71, df = 3, P = 0.13), nor a significant interaction between parental treatment and larval exposure on either day (day 14: X² = 3.57, df = 2, P = 0.15; day 21: X² = 3.45, df = 2, P = 0.18). On day 14, larvae from MP-exposed parents who received continued exposure visited and spent more time in the center of the center of the arena compared to those exposed only via parental gametes (X² = 7.16, df = 1, P = 0.007), but this effect disappeared by day 21 (X² = 2.71, df = 1, P = 0.10) (Figs. 4C, 4D).