Reduced Freshwater Mussel Juvenile Production as a Result of Agricultural and Urban Contaminant Mixture Exposures

Freshwater mussels provide invaluable ecological services but are threatened by habitat alteration, poor water quality, invasive species, climate change, and contaminants, including contaminants of emerging concern (CECs). Contaminants of emerging concerns are well documented in aquatic environments, including the Great Lakes Basin, but limited information is available on how environmentally relevant mixtures affect freshwater mussel biology throughout their varied life stages. Our main goal was to assess mussels' reproductive output in response to exposure to agricultural and urban CEC mixtures during glochidial development through juvenile transformation and excystment focusing on how exposure duration and treatment affect: (1) the number of glochidia prematurely released by brooding females, (2) glochidial transformation through host‐fish excystment, and (3) the number of fully metamorphosed juveniles able to continue the lifecycle. Mussels and host fish were exposed to either a control water (CW), control ethanol (CE), agriculture CEC mixture (AM), or urban CEC mixture (UM) for 40 and 100 days. We found no effect from treatment or exposure duration on the number of glochidia prematurely released. Fewer partially and fully metamorphosed AM juveniles were observed during the 100‐day exposure, compared with the 40‐day. During the 40‐day exposure, CW produced more fully metamorphosed individuals compared with CE and UM, but during the 100‐day exposure AM produced more fully metamorphosed individuals compared with the CW. There was reduction in fully metamorphosed juveniles compared with partially metamorphosed for CE and UM during the 40‐day exposure, as well as in the CW during the 100‐day exposure. These results will be important for understanding how mussel populations are affected by CEC exposure. The experiments also yielded many insights for laboratory toxicology exposure studies. Environ Toxicol Chem 2024;43:1112–1125. © 2024 The Authors. Environmental Toxicology and Chemistry published by Wiley Periodicals LLC on behalf of SETAC. This article has been contributed to by U.S. Government employees and their work is in the public domain in the USA.


INTRODUCTION
North America has the second greatest freshwater mussel species richness worldwide with over 300 known species (Graf & Cummings, 2007), and consequently it is home to more endemic species than other continents (Graf & Cummings, 2021).Historically, these vast mussel assemblages have provided invaluable ecological services to the watersheds in which they reside, including sequestration and processing of nutrients, stabilization of stream beds, and water quality improvements (Strayer, 2014;Vaughn & Hakenkamp, 2001;Vaughn & Spooner, 2006).Stressors such as habitat alteration, poor water quality, introduction of nonnative species, and climate change threaten freshwater mussel populations (Ferreira-Rodríguez et al., 2019).In addition, increased anthropogenic activities have introduced contaminants (i.e., contaminants of emerging concern [CECs]) to aquatic environments, posing a potential hazard to mussels, among other aquatic species.
It has been well established that CECs are widespread in aquatic environments and are present in complex mixtures (e.g., Bradley et al., 2017;Corsi et al., 2019;Elliott et al., 2018).Furthermore, land use within a watershed strongly influences which CECs may be present (Baldwin et al., 2016;Kiesling et al., 2019).For example, watersheds in the Great Lakes Basin characterized as predominately urban are generally associated with CECs such as pharmaceuticals, flame retardants, and industrial contaminants, while predominately agricultural watersheds are generally associated with herbicides, insecticides, and fungicides (Elliott et al., 2018).Despite the complex co-occurrence of CECs in the environment, little is known about the impacts that such mixtures may have on freshwater mussels.
Freshwater mussel habitat requirements, unique lifecycle, and sessile nature make them particularly vulnerable to CEC exposure at numerous life stages (Cope et al., 2008), but the effects of CEC exposure on freshwater mussels have not been widely assessed.Adults may be exposed via contact with sediments and/or during water syphoning, which could impact gamete production (Rzodkiewicz et al., 2022).Glochidia may be exposed to CECs through a variety of routes, including maternal transfer and marsupial gill exposure during development, host-fish transfer during encystment, and direct contact in the water column.Finally, juveniles may be exposed to CECs in much the same ways as adults, through sediment and water syphoning (Cope et al., 2008).These multiple exposure routes at all life stages make it complex but important to understand how CECs impact the mussel lifecycle and ultimately population levels (Woolnough et al., 2020).
Classic toxicology indicates that mussels are highly sensitive to some chemicals, yet sublethal effects from exposure to CECs are not well documented (Woolnough et al., 2020).For example, lethal testing of free-living glochidia (prior to infestation on host fish) and newly transformed juveniles suggests freshwater mussels are among the most sensitive of aquatic species to ammonia and metals (Augspurger et al., 2003;Wang et al., 2017).Studies of organic chemicals (including CECs) suggest reduced sensitives, compared with ammonia and metals, with sublethal effects such as behavior and growth impacts (e.g., Bringolf et al., 2010;Hazelton et al., 2013;Wang et al., 2017).However, studies that focus on sublethal effects of CEC exposure in invertebrates are limited.Instead, most studies focus on sublethal effects in vertebrates (mostly fish) because modes of action and biological processes are better understood.For example, CECs, such as bisphenol A (antioxidant, plastic additive), nonylphenol (surfactant), and atrazine (herbicide), have estrogenic properties that can negatively affect development or reproductive success among other biological processes in vertebrates (Faheem & Bhandari, 2021;Saravanan et al., 2019;Tillitt et al., 2010).While these studies provide important information regarding potential sublethal effects of CECs on vertebrates, it is unclear if similar effects may be expected in invertebrates such as freshwater mussels.
Research suggests that exposure to elevated concentrations of antidepressants may increase activities such as luring and burrowing in freshwater mussels, which may have consequences for survival (Hazelton et al., 2014).Yet, the impact of these sublethal effects on population-scale effects such as reduced reproduction and recruitment is poorly understood.Furthermore, exposure experiments typically focus on exposure to one CEC at a single life stage and are conducted at concentrations ranging orders of magnitude higher than typical environmental concentrations.Such experiments, while necessary for sensitivity screening and establishment of water quality standards, do not reflect scenarios in which organisms are exposed to complex CEC mixtures throughout their long life.The consequences of such real-world exposures are largely unknown (Woolnough et al., 2020).
In the present study, our main goal was to assess freshwater mussel reproductive response to exposure to agricultural and urban CEC mixtures during glochidial development through juvenile transformation and excysment.We focused on how 40-and 100-day exposure experiments and treatment type affect: (1) the number of glochidia prematurely released by brooding females, (2) glochidial transformation through hostfish excysment, and (3) the number of fully metamorphosed juveniles able to continue the lifecycle.We provide ecological context for these results and the potential impacts CECs may have on these vulnerable taxa.

METHODS
In 2018, we conducted two partial lifecycle laboratory experiments (40-and 100-day durations) in which brooding freshwater mussels (Lampsilis cardium), their host fish (Micropterus salmoides), and L. cardium parasitic glochidia were exposed to CEC mixtures representative of typical mixtures found in predominately urban and agricultural watersheds (Figure 1).

Experimental design
Brooding adult L. cardium were collected by hand via mask and snorkel from the Grand River in Lyons, Michigan in early July 2018.L. cardium are a common mussel species found throughout the Great Lakes Basin and have been previously used to represent mussels with similar profiles that are threatened or endangered (Woolnough et al., 2020).Adult L. cardium (length = 90 ± 1.5 mm) were transported to Central Michigan University (CMU), Mount Pleasant, Michigan, where they were housed in aerated aquaria.L. cardium were housed in individual aerated aquaria connected to separate recirculating systems dependent on the assigned treatment (Table 1) and were fed a mixture of Nannochloropsis spp.and shellfish diet (0.5 mL/individual; Reed Mariculture; Wang et al., 2007).The mussels collected for the 100-day exposure were acclimated 2 weeks prior to the start of exposure, and mussels collected for the 40-day exposure were acclimated for 60 days prior to exposure.Juvenile M. salmoides (total length = 132.5 ± 15.4 mm, mass = 25.5 ± 0.8 g) were purchased from Stoney Creek Fish Hatchery, Grant, Michigan in July 2018, and acclimated in two living streams at CMU for approximately 2 weeks prior to exposure.Micropterus salmoides were individually separated into 3-L aquaria connected to recirculating systems by treatment (Table 1).Exposure of M. salmoides to CEC mixtures began on day 1, when mussel exposures began, and continued throughout inoculation of aquaria with glochidia and during glochidia metamorphosis (Figure 1).Twice weekly, each fish was fed 1 mL of a mixture of three blood worm cubes and one brine shrimp cube (Hikari Bio-Pure) in 250 mL of dechlorinated water.
Specialized polycarbonate aquaria with 118 µm mesh filters (Aquaneering) were used to enable enumeration of glochidia partial and full transformation to juveniles.Polypropylene holding tanks for creating and recirculating treatment water were also used.Systems were flushed with dechlorinated carbon filtered water for 1 week prior to initiating treatments, and treatments were run for 3 days prior to adding the animals.Water samples were collected from the aquaria after the initial flush with dechlorinated carbon filtered water and before animals were added.These samples were evaluated for CECs to provide baseline conditions for both contaminants and water quality as per ASTM International guidance (ASTM International, 2013).None of the target CECs were detected in baseline samples.
Treatments were designed to replicate agricultural and urban mixtures representative of CEC exposures in the Great Lakes Basin as determined by surveys conducted by the US Geological Survey (USGS) and the US Fish and Wildlife Service (Baldwin et al., 2016;Choy et al., 2017;Elliott et al., 2017Elliott et al., , 2018)).The CEC concentrated stock mixtures were produced at the USGS National Water Quality Laboratory by diluting chemicals in ethanol for ease of dosing and to mirror previously conducted companion studies on fish (Swank et al., 2021).Agricultural mixtures included herbicides, pesticides, and flame retardants, and urban mixtures included several pharmaceuticals, a corrosion inhibitor, and a surfactant (Table 2; Elliott et al., 2017).For the present study, we analyzed the effect of four treatment types: dechlorinated water control (CW), ethanol control (CE), and ecologically relevant doses of the agricultural mixture (AM) and urban mixture (UM).To make exposure water, holding tanks were renewed four times daily with carbon-filtered dechlorinated water and respective stock CEC mixtures.Water flow was designed to target treatment concentrations and allow for full water exchange four times daily during pre-infestation and three times daily postinfestation (ASTM International, 2013).Ethanol was used as a solvent carrier in the CEC stock mixtures at a concentration of 0.5 µL of ethanol per liter of water (EtOH/L).An ethanol control (CE; 0.5 µL EtOH/L) was included to determine if the solvent   (2019), andGill (2019).Water samples were collected two times per week from the exposure aquaria and analyzed for target CECs by the USGS National Water Quality Laboratory.Substrate was not included in the exposure aquaria to reduce the risk of CEC adsorption to sediment particles (Arias-Estévez et al., 2008).Basic water quality measures were monitored following ASTM International protocols (ASTM International, 2013).A dissolved oxygen probe was used to measure dissolved oxygen (%) and temperature (°C).HACH (Loveland, Colorado) test strips were used to measure ammonia (mg/L), total chlorine (ppm), free chlorine (ppm), total hardness (mg/L), total alkalinity (ppm), and pH two times a week to ensure chemical consistency and to maintain any observed responses on L. cardium and M. salmoides were the result of CEC exposure (Sparks & Strayer, 1998;Wang et al., 2007).Vivarium lights were fixed to turn off at 22:00 each night until 06:00 the following day.

Mussel and fish exposure
All laboratory experiments followed a protocol approved by the CMU Institutional Animal Care and Use Committee (IACUC, approval number 17-11), and adapted from ASTM International guidelines for early life-stage exposures (ASTM International, 2013).Partial mussel lifecycle exposures were conducted with L. cardium and M. salmoides undergoing the same treatments concurrently, including 40-and 100-day exposure durations (Figure 1).For the 40-day exposure, fish and mussels were exposed to CECs in individual aquaria for 13 days prior to infestation of larval, non-metamorphosed glochidia released into host-fish water via standardized methods (Table 1-3; Yeager & Neves, 1986;Yeager & Saylor, 1995;Zale & Neves, 1982).Fish were then exposed to CECs within their separate aquaria for an additional 27 days during the juvenile transformation period (September-November 2018).For the 100-day exposure, fish and mussels were exposed independently to CECs for 60 days prior to infestation of fish with glochidia (Table 1).Fish were then exposed to CECs within their separate aquaria for an additional 40 days during the juvenile mussel transformation period (July-November 2018).

Data collection
Premature glochidia release.Parturition events, or releases of premature glochidia, are not common effects of handling and husbandry stress in L. cardium (Barnhart et al., 2008).To evaluate if parturition was associated with chemical exposure, as previously seen with CECs (Bringolf et al., 2010), female mussels were monitored for glochidial releases starting on day 0 of the exposure through day 13 (40-day exposure) or day 60 (100-day exposure), coinciding with the infestation date.
Prematurely released glochidia that were visible in the aquaria were counted by collecting a subsample.In addition, every 2 to 4 days, glochidia that were captured by a 118-µm mesh filter attached to the outflow polyvinyl chloride pipe were enumerated.Glochidia were counted using a Leica EZ4W stereomicroscope (Leica Microsystems) and the subsample volume was recorded.These glochidia were prematurely released and not available for fish infestation or recruitment into the population.
Fish aquaria infestation with glochidia.Mussels were assessed for the presence and viability of glochidia.Mussels from the 100-day exposure had limited gravidity on infestation day (exposure day 60).Only CW had more than one gravid female (n = 3), with two animals being used because of space constraints.The 40-day exposure had higher numbers of gravid animals on infestation day (exposure day 13) and were used to the maximum extent that space would allow (Table 1).Gravid L. cardium marsupial gills were flushed with deionized water using a 23-gauge syringe to collect eggs and glochidia that were enumerated using the stereomicroscope.Glochidia were tested for viability using a salt test in which approximately 100 glochidia per female were subjected to a saline solution and examined under a microscope (Fritts et al., 2014).The viability of glochidia from all assessed mussels was >95%.Viable glochidia were separated evenly into Petri dishes based on the number of fish available for infestation within each treatment type.An average count was obtained by averaging three (1 × 1 cm) grids and multiplying by 36 (number of grids on a Petri dish) prior to infestation of M. salmoides of the same treatment, similar to Zale and Neves (1982).Glochidia were then poured into each aquarium while the flow was paused for 45 min, after which unattached glochidia were filtered out of the system.We estimated the attachment rate by compiling the total number of glochidia and juveniles counted throughout the duration of the experiment exposures.
Glochidia transformation.Each host-fish aquarium and outflow filter were checked every other day for glochidia.Fully and partially metamorphosed juveniles were counted using the stereomicroscope.Fully metamorphosed juveniles were active and moving, with a shell growth line, foot, and formation of a gut while partially metamorphosed juveniles were not actively moving, and only one or two of the other criteria were met (Yeager & Saylor, 1995).Data on both fully and partially metamorphosed juveniles were analyzed to account for the natural reduction in numbers moving onto the next lifecycle stage and variation in the number of juveniles to undergo a full metamorphosis despite any effects of the treatment.

Data analysis
CEC exposure concentrations.Summaries of CEC concentrations measured within exposure aquaria were calculated, including detection frequency, average concentration, and standard deviation.When censored values (i.e., values below the detection limit) were reported for CECs, averages and standard deviations were estimated using the regression on order statistics method with the censtats() function within the NADA package (Lee, 2020).Averages and standard deviations could only be estimated if the percentage of values that were censored was <80%.
Premature glochidia release.We assessed the effect of treatment during the 100-day exposure on brooding females' premature glochidia release using a negative binomial zero-inflated generalized linear mixed model (R package "glmmTMB"; Brooks et al., 2017).Because of the short time frame (13 days) for data collection during the 40-day exposure, there was insufficient data for analyses.The global model included the number of glochidia released as the response variable with treatment and day of exposure as fixed effects and the mussel's aquarium identifier as the random effect; interactions were included.We conducted model selection using the MuMIn package in R (Bartoń, 2022) and determined the best model based on Akaike information criterion (AIC) and R 2 values.The AIC values were similar, so we chose the model with the greatest explanatory power (as determined by the conditional R 2 ) to best describe the observed effects (James et al., 2013); the model selected included only treatment type as a fixed effect.We ran a Type II Wald χ 2 test (R package "car") to further assess the influence of the fixed effect variables (Fox & Weisberg, 2018), and a pairwise comparison of the fixed effects was conducted using estimated marginal means (R packages "emmeans" and "graphics"; Lenth, 2023; R Core Team, 2022).
Glochidia transformation.The effects of treatment and exposure duration (40-day vs. 100-day) on glochidia transformation were analyzed using the number of partially and fully metamorphosed juveniles transformed from each fish.We used a negative binomial zero-inflated generalized linear mixed model (R package "glmmTMB"; Brooks et al., 2017) to analyze the effects of treatment type, exposure duration, and metamorphose state (partial and full metamorphosis) on the number of glochidia to metamorphose.All three predictor variables were included as interactions, with the fish identifier as a random effect.An offset representing an estimate of glochidia to initially attach to gills was included to account for different numbers of glochidia that attached to host fish among treatments.We performed model selection on the global model to determine the fit with the most explained variance and predictive power (R package "MuMIn"; Bartoń, 2022).A pairwise comparison of the predictor variables was conducted using estimated marginal means to compare effects (R packages, "emmeans," "graphics"; Lenth, 2023; R Core Team, 2022).This pairwise comparison pulled from the negative binomial model resulting in log scale (base 10) output comparisons.The total number of juveniles was further organized by the female mussel that produced those glochidia.Analyses and results are included in the Supporting Information, but because of low sample sizes, will not be discussed in detail.All analyses were conducted using R (R Core Team, 2022; Ver 4.2.1) and R Studio (Posit team, 2023; Ver 2023.6.0.421); figures were created using ggplot2 (Wickham, 2016).Data used in analyses are available in a USGS data release (Richard et al., 2023).
CEC concentrations in exposure tanks varied throughout the exposure, with average concentrations that ranged from −96% to 200% of the nominal concentrations (Table 2).Several CECs were detected in CE samples: metolachlor, tris(2-butoxyethyl) phosphate, 4-nonylphenol, N,N-diethyl-mtoluamide, bisphenol A, and desvenlafaxine.In most instances, the average concentrations detected in CE samples were substantially less than those in AM or UM samples; two exceptions were 4-nonylphenol and bisphenol A (Table 2).Furthermore, unforeseen issues with recovery rates and standard stability affected the determination of bisphenol A, 4-nonylphenol, and estrone concentrations in most or all samples.Despite the variation in concentrations and detections in CE samples, average concentrations in the exposure aquaria provided sufficient difference among treatments while representing realistic chronic exposures that mussels may experience in nature.

Premature glochidia release
Treatment did not affect the amount of glochidia that were prematurely released during the 100-day exposure.The AIC and R 2 values were similar across all models evaluated, 4330 to 4334 and approximately 0.38, respectively (Supporting Information, Table S2).The model with the highest R 2 value did not include treatment as a significant variable.Although the model with the next highest R 2 value did include treatment as a significant variable for explaining premature glochidia release, no differences among treatments were observed (Table 3, Figure 2, and Supporting Information, Table S3).However, mussels in the CE treatment were significantly more likely to release no glochidia on any given day (i.e., more zero counts of glochidia, p < 0.05; Table 3).

Glochidia transformation by host-fish drop-off
The model with the lowest AIC score (941) that explained the most variance (R 2 = 0.95) in transformed juvenile counts included treatment type, exposure duration, metamorphose state, the interaction between the three factors, and an offset of the total number of glochidia estimated to have attached during infestation (Table 4 and Supporting Information, Table S4).Tukey's post hoc pairwise comparisons were conducted to answer our questions regarding the effects of exposure duration, treatment type, and ability to fully metamorphose on juvenile production and, consequently, the glochidia's ability to move on to the next stage of the lifecycle.
Effects of exposure duration.There were few differences between exposure duration for partially and fully metamorphosed juveniles (Supporting Information, Table S5).For both metamorphose stages, there was a significant effect of exposure duration (40 vs. 100 days) on AM juveniles.Relatively fewer partially and fully metamorphosed juveniles were observed during the 100-day exposure, compared with the 40-day, suggesting a negative influence from a longer exposure time for this treatment type (Figure 3A-D and Supporting Information, Table S5).Effects of treatment type.Within the 40-day exposure, there was no difference among the proportion of glochidia that partially metamorphosed for all treatment types (p > 0.05; Supporting Information, Table S5).However, some differences were observed for fully metamorphosed individuals.There were nearly 100 times more fully metamorphosed individuals in CW compared with CE and 33 times more fully metamorphosed individuals compared with UM (p < 0.05; Figure 3A,C,E and Supporting Information, Table S5).Similar to the 40-day exposure, no significant differences were observed among treatments for the number of partially metamorphosed individuals during the 100-day exposure.With respect to fully metamorphosed juveniles, nearly twice as many were observed in CW compared with AM (p < 0.05; Figure 3B,D,F and Supporting Information, Table S5).
Partial versus fully metamorphosed juveniles.Apparent reductions in the ability of glochidia to fully transform were observed in both the CE and UM treatments.Within the 40-day exposure, 40 times more partially metamorphosed juveniles were observed compared with fully metamorphosed in CE and 11 times more partially metamorphosed were observed compared with fully in UM, suggesting many glochidia were unable to fully metamorphose (p < 0.05; Figure 3A,C,E and Supporting Information, Table S5).
During the 100-day exposure, nearly 10 times more partially metamorphosed juveniles were observed in CW, compared with fully metamorphosed juveniles (p < 0.05; Figure 3B,D,F).There was no significant difference between partially and fully metamorphosed juveniles in the AM, CE, and UM treatments (Figure 3B,D,F and Supporting Information, Figure S1).Furthermore, no UM fully metamorphosed juveniles were produced during the 100-day exposure experiment (Figure 3B,D,F, Supporting Information, Figure S1, and Supporting Information, Table S5).

DISCUSSION
Our study suggests that reproductive impairment may occur during partial life-stage exposures of ecologically relevant CEC  mixtures.Even minor impairments by stressors such as CECs may cause population declines, particularly in vulnerable species such as those that are threatened or endangered.Our study begins to address the complexity of mussel reproduction by evaluating a more natural exposure of CEC mixtures throughout the brooding, glochidia release, and glochidia transformation stages of the freshwater mussel lifecycle.Furthermore, our novel approach provides valuable lessons on which to build future chronic environmental studies.
It should be noted that the CE, AM, and UM treatments included a solvent carrier of 0.5 µg/L of EtOH/L.Although this concentration is 200 times below ASTM International guidance for chronic mussel assessments (ASTM International, 2013) and 40 times less than that recommended in a review of solvent carrier effects for aquatic toxicology (Hutchinson et al., 2006), unexpected results in our study bring into question the possibility of ethanol as a confounding factor.Additional research is warranted to determine the extent that such solvent  S5), total counts (C, D; Supporting Information, Table S5), and log10 transformation rate (juveniles/ total glochidia and juvenile count; E, F).AM = agriculture mixture; CE = control ethanol; CW = control water; UM = urban mixture.
concentrations may have on chronic mussel assessments.Despite this, the results from our study provide a baseline for informing future partial lifecycle exposures of freshwater mussels to CECs.
Our study showed reduction in the proportion of juveniles produced as a result of exposure to CEC mixtures (Figure 4) and highlights the value of understanding the impacts of exposure to brooding mussels, particularly at ecologically relevant doses.Premature release of glochidia may remove individuals from recruitment into the population.Because L. cardium is generally not associated with premature glochidia release as a result of handling stress (Barnhart et al., 2008), we evaluated parturition to determine if exposure of brooding females to ecologically relevant CEC mixtures leads to recruitment loss, as has been documented in acute elevated fluoxetine exposures (3-3000 µg/L; Bringolf et al., 2010).In our study, parturition occurred across treatments and exposure duration, with no discernible patterns in timing, number of glochidia released, or treatment effects.This mirrors chronic 28-day exposures of brooding females to fluoxetine concentrations above environmental concentrations (0.37-29.3 µg/L), which also did not elicit significant parturition in brooding females (Hazelton et al., 2013).The high variability of premature glochidia releases, in our study, did not inform later transformation success.For example, CW prematurely released a similar number of glochidia compared with treatment mussels.However, a higher proportion of CW juveniles fully transformed compared with CE and UM during the 40-day exposure and compared with AM during the 100-day exposure.A better understanding of the complex general and reproductive biology of mussels is needed to assess the effects that exposure variables such as life stage, chemical mode of action, and exposure duration and concentration may have on mussel population dynamics.
Although reductions in juvenile production were observed across all treatments, UM exposures resulted in the overall lowest proportion of partially and fully transformed juveniles.The complex chemical mixtures associated with urban landscapes have long been considered a contributing factor in mussel declines, with molecular stress responses elicited in adults exposed to in situ wastewater effluent for as little as 14 days (Falfushynska et al., 2014;Gillis et al., 2014).Similarly, reduced production (40-day exposure) or no production (100-day exposure) of juveniles in our study suggests exposure to representative urban mixtures may contribute to recruitment loss.As with our results, chronic (28-day) exposures of brooding mussels to perfluoroalkyl acid concentrations ranging from 1 to 100 μg/L had a reduced probability of glochidia metamorphosing to juveniles (Hazelton et al., 2012).This warrants further understanding of chemical modes of action as well as the complex processes associated with glochidia to juvenile transformation.The complex nature of the mixtures with multiple modes of action does not allow us to resolve the action that elicited the reduced juvenile production, but it suggests urban sites with a constant influx of CECs may be important to understand prioritizing conservation of mussel populations.
One key aspect of transformation is the poorly understood host-fish-glochidial relationship.Understanding the effect of host-fish health under natural scenarios may be important for both fish and mussel management.Alterations in host-fish immune response and resulting mussel transformation impacts from CEC exposure have not to our knowledge been evaluated.Yet studies evaluating host-fish immunity from previous glochidia infestation have been well documented (Barnhart et al., 2008;Dodd et al., 2006).Alterations in host immune response from prior glochidia infestation leads to premature shedding of glochidia, encapsulation irregularities, and delays, leading to increased glochidia mortality and reduced transformation success (Dodd et al., 2006).Limited host availability through acquired immunity may lead to further population declines (Rashleigh & DeAngelis, 2007).Natural glochidia infections or contaminant impacts, particularly pharmaceuticals that are designed to alter immune responses, may lead to reduced transformations among CEC-exposed fish and glochidia, as in our study.
Exposures to AM in our study had similar juvenile production compared with CW during the 40-day exposure experiment; however, a higher proportion of glochidia fully transformed compared with CW during the 100-day exposure.This supports previous research that indicates mussels are relatively tolerant of agricultural contaminants with acute lethality testing shown to be near water solubility limits for early life stages (Bringolf et al., 2007).Despite the apparent tolerance, chronic and sublethal assessments suggest impairments in juveniles, such as delayed growth and reduced survival, are likely (Bringolf et al., 2007).A therapeutic effect of agriculture chemicals such as altered food availability or quality may, in part, explain the minimal transformation impacts of short-term exposure to agricultural chemicals observed in our study.For example, microbial shifts in mussels (assessed on the same individuals used in the present study) may lead to metabolic changes affecting both behaviors and health (Gill et al., 2022), which ultimately could lead to population changes.Such shifts in nutrient availability and alteration of microbes with agricultural chemical exposure are likely to impact the energetics available for health maintenance and reproduction in currently unknown ways.However, agricultural chemicals have been shown to alter food quality and quantity with shifts in sediment bacterial and algal communities (Fairchild et al., 2009;Seghers et al., 2003;Strayer, 2014).Strayer (2014) indicates that agricultural locations that experience eutrophication often reach a plateau where benefits in growth due to high food availability are limited because of shifts in poorer nutrient quality algal species.Regardless of the reason for the dampening of reproductive impacts seen in AM, even these minor reductions in transformation may have large implications for mussel populations, especially vulnerable populations including threatened or endangered species.Furthermore, modeling by Rzodkiewicz et al. (2023) indicates that even common species may be extirpated from locations at current agricultural chemical concentrations if minor reductions occur in the wild.However, the mechanisms leading to reductions are currently unknown.
Additional exposures, beyond in vivo (maternal and marsupial), may have contributed to the overall reproductive loss observed in our study.Exposure during the free glochidia stage, immediately after release from the marsupial gills by the brooding female and directly before encystment on host fish, is possible and might be a contributor (Cope et al., 2008).This free glochidia stage is used in toxicology assessments to base mussel sensitivity to a chemical (Augspurger et al., 2003;Wang et al., 2017) because it is easily reproducible and provides valuable acute effects concentrations on which to base protective water quality criteria.However, because this stage only lasts minutes to days (Bauer & Wächtler, 2001), additional information on chronic and multiple life-stages effects, such as partial life-stage exposures, may be necessary to address the overall species risk.Rapid encapsulation may further limit water exposure during the parasitic stage, with the encystment occurring within hours of infestation providing some degree of protection (Jacobson et al., 1997;Rach et al., 2006).Exposure may occur during active transformation when encysted glochidia, as obligate parasites, derive nutrients from their host fish (Cope et al., 2008;Fisher & Dimock, 2002).Our study, which exposed host fish throughout the exposure period (40 or 100 days) mimicked this natural scenario and may have contributed to observed effects.During this period of metamorphosis, glochidia are the most vulnerable, undergoing complex processes that transform simple parasitic larvae into fully functioning juvenile mussels (Chumnanpuen et al., 2011;Fisher & Dimock, 2002).Unknown potential disruptions from chemical or other stressors may arrest development (partial transformation) and reduce transformation success, as observed in our study.
The period of transformation may be a critical time point during mussel development (Rzodkiewicz et al., 2023).We analyzed partial transformation to better capture where recruitment loss might occur during this complex process.Histological and molecular assessments of transforming glochidia show an increase of RNA, DNA, and protein production first enabling glochidia to derive nutrients from the host fish and later a secondary stage of progressively more advanced organ development until excystment from host fish is achieved (Fisher & Dimock, 2002).Such intense proliferation of protein production allows multiple avenues for interrupted pathways from CECs or other stressors, causing a cascade of adverse outcomes resulting in arrested development, including partial transformations.Because the proportion of glochidia that partially transformed did not differ among treatments, we hypothesize that all glochidia that infested fish in our study were fit enough to begin metamorphosis.Conversely, glochidia from the UM and CE exposures had the lowest partial and fully transformed juveniles, suggesting glochidia were either unable to begin transformation or were lost early in the process.Partial transformation was observed throughout the assessed transformation period, indicating that possible arrested development occurred, or possible delay of development, which has been associated with other invertebrates exposed to contaminants (Pisa et al., 2015).

Implications and potential future research
This novel approach not only provides additional support for CECs impacting mussel recruitment, but also highlights the need for research that more accurately mimics the exposure scenarios that these long-lived animals experience in the wild.Such shifts from shorter duration early life-stage exposures will require standardization and additional sublethal endpoints, as observed in our study.
Solvent usage.Influences of exposure experiment and type produced unexpected results in our study.Reduction in juvenile transformation was unexpectedly high in ethanol controls.The ASTM International procedures for early life-stage exposures was used as guidance because such chronic research is currently unavailable.Under these guidelines, ethanol is permitted as a solvent carrier with no greater than 100 µL EtOH/L.Our ethanol concentration was 0.5 µL EtOH/L, 200 times less than these recommendations (ASTM International, 2013), yet it resulted in reduced reproduction in all experiments.This suggests that under chronic exposures, ethanol may account for some of the observed responses.In a review of solvent usage during toxicity testing, Hutchinson et al. (2006) suggested solvent usage may influence effects with both indirect (food and water quality) as well as direct measurable organismal effects (reduced growth, fecundity/reproduction, and endocrine disruption biomarkers alterations).In the extensive review (Hutchinson et al., 2006), the authors suggest chronic exposure may be particularly impactful with current acceptable solvent concentrations above observable effects for some organisms, suggesting concentrations at or below 20 µL/L for carrier solvents.Our study indicates further refinement may be needed for chronic mussel exposures because unquantified sublethal alterations resulted in reduced transformation success at levels as low as 0.5 µL EtOH/L.The authors acknowledge such a carrier effect may have impacted the effects and/or the degree of the effects seen in the present study.However, the literature suggests that impacts of CECs are probable.Further refinement of solvent use in chronic exposures would allow resolution between effects associated with the CEC solvent carrier and the CECs themselves.
Exposure duration.Exposure duration could be evaluated and weighed against biological timing of exposure, exposure cost, and mussel husbandry constraints.The percentage of gravid females available to donate glochidia was lower in our 100-day study, suggesting extended captivity under laboratory conditions may have been a factor.However, CEC exposure duration may have also impacted mussel gravidity, with the largest number of gravid females in CW when compared with treatment animals.In our study, exposure duration did not significantly affect the number of glochidia that transformed across all treatments, except AM.Our study showed even short-term exposures to CECs, as little as 13 days in vivo, caused impairments to reproduction with reductions in juvenile transformation, particularly in UM and CE treatments for which juvenile production was the lowest.Exposure duration and husbandry constraints could benefit from further attention to enable more accurate exposure scenarios in these long-lived species.
Variability.Variability is very high both annually and among mussels, and does necessitate an additional number of mussels to validate findings.For example, fecundity is highly variable within species.Similar sized individuals can have mean differences of up to two times higher egg counts from within a site and nearly seven times higher egg counts between sites (Haag, 2013).All endpoints assessed had high variability, which may be attributable to long protracted spawning (up to 4 months) and a lengthy window of glochidial release with several peaks throughout the spring and summer in long-term brooders such as those in our study (reviewed in Haag, 2012).Such factors coupled with limited mussel numbers available for analyses made interpretation difficult.Repeated studies and higher sample numbers could capture the variability and enable population level interpretations.
Reproductive behaviors.High variability among treatments, study exposure duration, and mussels is common.In our study, luring behaviors were not indicative of reproductive success, with no qualitative differences between parturition or reproductive readiness.Additional factors associated with mussels' complex lifecycle are largely unknown, and laboratory conditions may have altered luring in our study, including unknown pheromone cues from host fish in our recirculating water system and/or lack of sediment.In-stream luring is highly variable and not correlated to population size or later evidence of recruitment (Jones & Neves, 2011).Parturition is also likely not a valuable endpoint to assess because we did not see correlation to total numbers of juveniles produced.Even though some treatments and individuals did have many parturition events, there was no discernible pattern to these events, nor was it predictive of glochidia donor mussel.
Biomarkers.For chronic studies, such as ours, that mimic environmental exposures, subtle changes in sublethal biomarkers may more accurately indicate if the stressor elicited a response.However, we currently lack both the biological understanding and methods to quantify many sublethal but potentially population-limiting responses.For example, as part of our study, additional behavior data were collected for qualitative assessments, as well as health endpoints such as glycogen, fatty acid, and hemolymph analyses.These endpoints are often collected to provide insights into mussel health (Newton & Cope, 2006).However, in our study, these endpoints exhibited alterations that were highly variable, making their application to addressing CEC effects limited.Further biological associations and population level impacts of these endpoints as well as other emerging biomarkers such as proteomics and other molecular endpoints are largely unknown.Development of baseline values for biomarkers through increased knowledge of mussel biology, reproduction, and population impacts would be helpful to assess environmentally relevant CEC exposure and other chronic stressor effects.

CONCLUSION
The present study highlights the complexities of conducting freshwater mussel partial lifecycle exposures to CECs.Because of these complexities, more similar studies would help to better understand the threat that CEC exposure poses to sensitive freshwater mussels throughout their unique lifecycle.We observed no effect from treatment or exposure duration on the number of glochidia that were prematurely released, but we did see reductions in the number of juveniles produced in certain treatments.More juveniles were produced by mussels exposed to CW and AM treatments with CE and UM.This novel laboratory exposure yielded many insights for future toxicology studies, including considerations for solvent treatment effect, host-fish exposure, and evaluation and development of sublethal endpoints.The results from the present study may provide information to support management decisions such as the development of protective water quality standards, the prioritization of relocation and restoration efforts, the optimization of hatchery practices, and the timing of mussel relocation/release to reduce CEC exposure and increase reproductive success.Such efforts are necessary to ensure freshwater mussels continue their vital roles in healthy ecosystems.
Supporting Information-The Supporting Information is available on the Wiley Online Library at https://doi.org/10.1002/etc.5844.

FIGURE 1 :
FIGURE 1: Timeline of the 40-and 100-day exposures of the freshwater mussel, Lampsilis cardium, and the fish, Micropterus salmoides, to contaminant of emerging concern mixtures.

FIGURE 2 :
FIGURE 2: Glochidia counts for 40-and 100-day exposures of the freshwater mussel, Lampsilis cardium, and the fish, Micropterus salmoides, to contaminant of emerging concern mixtures.Glochidia counts were not affected by treatment during the 100-day exposure experiment; limited data from the 40-day exposure inhibited comparisons.AM = agriculture mixture; CE = control ethanol; CW = control water; UM = urban mixture.

FIGURE 4 :
FIGURE 4: Log10-transformed average counts of Lampsilis cardium glochidia inoculated into Micropterus salmoides aquaria, estimated average number of glochidia that infested fish gills, and average number of glochidia that partially and fully metamorphosed during the 40-day (A) and 100-day (B) exposure experiments.AM = agriculture mixture; CE = control ethanol; CW = control water; UM = urban mixture.

TABLE 1 :
Number of freshwater mussels (Lampsilis cardium) and fish (Micropterus salmoides) in each exposure duration and treatment a Number of L. cardium evaluated, each mussel had glochidia present.

TABLE 2 :
Detection frequencies (top number; n = 75) and mean concentrations (±standard deviation) of contaminants of emerging concern in control water (CW), control ethanol (CE), agriculture (AM), and urban (UM) exposure aquaria Mean concentrations and standard deviations were estimated for contaminants detected in <100%, but more than 20% of samples.Detection frequencies are given as percentages; concentrations are nanograms per liter.bdl = mean concentration could not be estimated because there were too few quantifiable detections; nd = not detected in any sample.

TABLE 3 :
Negative binomial zero-inflated generalized linear mixed model (R 2 = 0.391) assessing treatment type on the number of prematurely released glochidia during the 100-day exposure experiment

TABLE 4 :
Negative binomial generalized linear mixed model for partially and fully metamorphosed juvenile mussels (plain pocketbook, Lampsilis cardium) by host fish (largemouth bass, Micropterus salmoides) for 40-and 100-day exposure experiments