Metabolic Disruption in Salmonids Following Co-Exposure to Road Runoff Contaminants 6PPD-Quinone and 9,10-Anthraquinone

Abstract

Components of road runoff have been correlated to adverse health outcomes in fish and aquatic invertebrates residing in urban water bodies. Recently, a transformation product of the antioxidant chemical added to car tires, N-(1,3-Dimethylbutyl)-N’-phenyl-p-phenylenediamine (6PPD-quinone, 6PPD-q), was found to induce urban stormwater mortality syndrome (URMS) in salmonids. Exposure to 6PPD-q has the potential to increase the sensitivity of salmonids to other co-occurring contaminants commonly detected in surface waters, such as the polycyclic aromatic hydrocarbon 9,10-anthraquinone (AQ). To investigate the mechanisms of 6PPD-q toxicity, fish were exposed to established sublethal concentrations of 6PPD-q and AQ, both separately and in combination, over a five-day period. Non-targeted MS-based metabolomics analysis was conducted on liver samples. Using in-house and open-source spectral libraries, 260 metabolites were confidently annotated in liver tissues. This study provides mechanistic insights into 6PPD-q and AQ toxicity in salmonids, highlighting oxidative stress, mitochondrial dysfunction, and impaired detoxification as key drivers of sublethal effects. Findings also suggest that AQ may mitigate 6PPD-q toxicity by enhancing antioxidant defenses and xenobiotic metabolism. Additionally, quantitative RNA sequencing of liver tissues revealed species-specific transcriptomic responses to chemical exposures, with coho salmon exhibiting more pronounced differential gene expression relative to chinook and rainbow trout. Transcriptomic analysis identified key modulation of genes involved in apoptosis, immune response, vascular permeability, and energy metabolism, deepening mechanistic understanding of 6PPD-q toxicity.

Graphical Abstract

1. Introduction

Salmon are a keystone species in the Pacific Northwest, playing a vital ecological role through their semelparous life cycle and sustaining other keystone predators such as bears and orcas (Garibaldi and Turner, 2004; Holtgrieve and Schindler, 2011; Masoner et al., 2019a). Beyond their ecological role, salmon hold deep cultural and spiritual significance for Indigenous communities, contributing to the social fabric, spiritual vitality, and ecological equilibrium across the northwestern United States and British Columbia (Bottom et al., 2009; Thompson et al., 2020). However, Pacific salmonid populations (Oncorhynchus spp.) are declining due to anthropogenic stressors, including climate change, habitat loss, and urban stormwater runoff (Siegel and Crozier, 2019; Wilson et al., 2022). Urbanization has intensified stormwater runoff by increasing impervious surfaces such as roads, roofs, and parking lots, which disrupt natural hydrological processes (Burns et al., 2012; Walsh et al., 2012). This runoff has been linked to pre-spawn mortality in coho salmon (Oncorhynchus kisutch), a phenomenon associated with exposure to toxic organic contaminants in urban waterways (Scholz et al., 2011). When mortality coincides with stormwater events, it is referred to as urban runoff mortality syndrome (URMS) (McIntyre et al., 2018). Symptoms include surface swimming, loss of equilibrium, and mouth gaping, often leading to death within hours of exposure (Scholz et al., 2011). The strong correlation of URSM with runoff events has driven further investigation into toxic components of urban runoff (McIntyre et al., 2018).

Urban stormwater runoff contains a complex mixture of chemicals, including car derived compounds, polycyclic aromatic hydrocarbons (PAHs), pesticides, metals, phthalates, and additional compounds from household, commercial, agricultural, and industrial uses (Masoner et al., 2019b; Mayer et al., 2024; Wicke et al., 2021). Tire wear particles and microplastics are also frequently detected in urban runoff and have been shown to cause adverse effects in aquatic species (Chibwe et al., 2022; Cunningham et al., 2022; Werbowski et al., 2021). Tire wear particles emanate from the friction occurring between rubber tires and roadways. Stormwater runoff can translocate tire wear particles from roadways to nearby sediments or water bodies, which are distributed further by natural water flow (Wagner et al., 2018). Chemical components in tire rubber can leach from the solid particles into the surrounding aquatic environment (Johannessen et al., 2021). Many of the components in urban runoff are not routinely monitored or have not yet been identified. Recently, a chemical additive in vehicle tires, N-(1,3-dimethylbutyl)-N′-phenyl-p-phenylenediamine (6PPD) and its oxidized transformation product, 6PPD-quinone (6PPD-q), were implicated as the causative agent in URMS (Tian et al., 2022, 2021). 6PPD is added to tires to serve as an antidegradant and antiozonant to increase lifespan. Subsequent testing revealed the 24-hour LC₅₀ for juvenile coho salmon is just 95 ng/L and is within the reported concentrations of 6PPD-q in tire wear particle leachate (250 – 1000 mg/L), road runoff (80 – 19000 ng/L), surface waters (6 – 3500 ng/L), snowmelt (80 – 270 ng/L), and drinking water (2.5 ng/L) (Cao et al., 2022; Challis et al., 2021; Chen et al., 2023; Johannessen et al., 2021, 2022; Tian et al., 2021, 2022; H.-Y. Zhang et al., 2023).

Following the identification of 6PPD-q, its toxicity and modes of action across various species have been extensively studied. The most sensitive species at the juvenile stage are considered to be coho followed by coastal cutthroat trout (O. clarkii clarkii), brook trout (Salvelinus fontinalis) and rainbow trout (O. mykiss) (Brinkmann et al., 2022; Di et al., 2022; Prarthana Shankar et al., 2024). 6PPD-q is substantially less toxic to other salmonids tested including chinook salmon (O. tshawytscha), sockeye salmon (O. nerka), Atlantic salmon (S. salar) and brown trout (S. trutta) (Foldvik et al., 2022). Some of the proposed modes of action for 6PPD-q include interaction with cytochrome P450s (CYPs), mitochondrial dysregulation, and permeation of the blood brain barrier (Blair et al., 2021; Hua et al., 2024; Mahoney et al., 2022; S.-Y. Zhang et al., 2023). Despite recent research efforts, the mechanism by which 6PPD-q is toxic to salmonids has not yet been fully elucidated.

In addition to chemicals such as 6PPD-q, tire formulations include thousands of other proprietary components such as heavy metals, polycyclic aromatic hydrocarbons (PAHs), and plasticizers to create functional tires (Mayer et al., 2024). PAHs originate from both tire wear and vehicle exhaust, making them ubiquitous in urban environments (Gbeddy et al., 2020). Many oxy-PAHs are also known to be carcinogenic and/or mutagenic (Clergé et al., 2019). One such oxy-PAH, 9,10-anthraquinone (AQ), is abundant in vehicle emissions and forms through the atmospheric oxidation of anthracene and co-occurs with 6PPD-q in urban surface waters (Awonaike et al., 2021; Valduga et al., 2024). PAHs are known to exert toxicity through activation of the aryl hydrocarbon receptor (AhR) and modulation of cytochrome P450 enzymes, which play a key role in detoxification (Hollenberg, 2002; Honda and Suzuki, 2020; Mallah et al., 2022). PAHs such as AQ can inhibit CYP expression, inhibition of CYP expression may reduce the clearance of 6PPD-q, potentially enhancing its toxicity. Given the widespread presence of both 6PPD-q and PAHs in urban waters, their combined effects may alter toxicity in salmonids (Liu and Sayes, 2024).

This study, for the first time, investigates the mechanisms of 6PPD-q toxicity alone and in combination with AQ, a representative oxy-PAH, to better understand their interactive effects on aquatic organisms. AQ was selected due to its widespread environmental occurrence and relatively low individual toxicity, which makes it an ideal candidate for assessing how co-exposure with 6PPD-q may influence toxicity without confounding effects from strong AQ-driven responses. Liquid chromatography-mass spectroscopy (LCMS)-based metabolomics and RNA-sequencing transcriptomics analyses of liver tissue were utilized to elucidate sublethal toxicity and mechanisms of action of road runoff pollutants among rainbow trout (O. mykiss), coho (O. kisutch), and chinook (O. tshawytscha), which were chosen to represent an array of sensitivity to 6PPD-q and fill current literature data gaps. Liver tissues were chosen based on high metabolic activity of CYP expression present in liver cells (Rui, 2014). Integrating metabolomics and transcriptomics enables the identification of key biomarkers of exposure and effect, while also uncovering potential adverse outcome pathways crucial for understanding the toxicological impact of chemicals on organisms. The present study aimed to 1) generate hypotheses that further describe the mechanisms driving the toxicity of 6PPD-q and explore drivers of species-specific differences in sensitivity; and 2) understand if co-exposures to other common aquatic pollutants such as AQ exacerbate the toxicity of 6PPD-q.

2. Materials and Methods

2.1. Standards and Reagents

Native 6PPD-Q standards (powdered 95.6 % purity; solution in acetonitrile, 100 μg/mL) were purchased from HPC Standards (Atlanta, GA). 9,10-anthraquinone (powdered 97 % purity) was purchased from Sigma Aldrich (St. Louis, MO). Acetonitrile (LCMS-grade), water (LCMS-grade), and formic acid (HPLC grade) were purchased from ThermoFisher Scientific (Waltham, MA, USA).

2.2. Fish Husbandry

Fish were housed in the Aquatic Animal Health Laboratory (AAHL) at Oregon State University approved by the Institutional Animal Care and Use Committee (IACUC-2022–0260). Details are described in SI Text S1.

2.3. Salmon Exposures to Road Runoff Contaminants

Range-finding exposures were conducted to determine sublethal concentrations for each juvenile fry salmon species. Exposure-response experimental design included a 120-hour fasted control tank, a 120-hour fed control tank, and three chemical treatment groups: a 120-hour 1 μg/L 6PPD-q exposure tank (6Q), a 120-hour 10 μg/L AQ exposure tank (AQ), and a 120-hour both 1 μg/L 6PPD-q and 10 μg/L AQ co-exposure tank (CE) for rainbow trout and chinook. Coho sublethal concentrations were lower due to heightened sensitivity and thus were exposed to 0.001 ug/L 6PPD-q. Further details are described in SI Text S2.

2.4. Water Chemical Analysis

Water from rainbow trout (RT) exposure tanks was sampled in triplicate directly after chemical spiking (initial) and at the end of the exposure period (final) at 120-hours for each biological replicate. Water samples were collected in 2 mL autosampler vials from the fasted control tank and both 6PPD-q and AQ exposure tanks. Water samples were analyzed to quantify 6PPD-q and AQ using liquid chromatography coupled to high-resolution mass spectrometry. Further details for methods and results are described in SI Text S3, Text S8, Figure S2, Table S1, and Table S3.

2.5. Metabolomics Analysis

Salmonid livers were utilized for analysis of small molecules using LCMS-based metabolomics. Details are described in SI Text S4. Briefly, 10 mg of flash-frozen liver tissue was added to Eppendorf vials with 40 μL of 1.4 mm ceramic beads, 300 μL of chilled methanol:water (80:20) and 3 μL isotope labeled metabolite internal standards. Samples were homogenized, centrifuged and stored overnight. Before analysis, samples were centrifuged again and transferred to autosampler vials. Liquid-chromatography mass spectrometry (Shimadzu Nexera paired to a Sciex 7600 ZenoTOF) data dependent acquisition was performed on all samples in both positive and negative electrospray ionization (ESI). Pooled QCs and instrument blanks were injected every 8 samples and samples were randomized prior to injection.

2.6. Data Processing and Statistical Analysis

Metabolomics data processing was performed using MS-DIAL, version 4.0.7 (Tsugawa et al., 2015). An in-house metabolite library generated using the IROA Mass Spectrometry Library of Standards and open source spectral libraries were used for metabolite identification Statistical analyses were performed using the open-source tool MetaboAnalyst and R studio (Pang et al., 2022; R Core Team, 2023). Further details are described in SI Text S5.

2.6. RNA Sequencing and Analysis

Fish liver tissues were used for quantitative RNA-sequencing. Full details are described in SI Text S6. RNA was extracted using the MagMAX™ mirVana™ kit (Thermo Fischer Scientific). Library preparation and cDNA synthesis of liver samples were performed using the Quant-Seq mRNA Seq Library Prep Kit for Illumina (FWD v2) (Lexogen, Cat no./ID: 191.96). Sequencing occurred on the Illumina NextSeq 2000. Genes with at least 10 counts in at least two samples were analyzed for for differential gene expression between treatments (DESeq2). Gene ontology (GO) terms were established using PANTHER version 19.0 (Thomas et al., 2022), DAVID version 6.8 (Huang et al., 2009; Sherman et al., 2022) and g:profiler version e113_eg59_p19_f6a03c19 (Kolberg et al., 2023).

3. Results and Discussion

3.1. Range Finding Exposures

Range-finding experiments aligned with the acute toxicity reported in the literature. In RT, the 24-hour exposures did not result in mortalities below 0.1 μg/L and 1 μg/L. However, at 10 μg/L, two-thirds of the exposed RT after 24 hours (Fig. S1B). In contrast, CH exhibited no mortality or URMS symptoms following 6PPD-q exposure at concentrations up to 10 μg/L, consistent with previous studies (Chow et al., 2019; Greer et al., 2023a). RT appear to have moderate sensitivity, with a reported LC₅₀ of 1.96 μg/L (Brinkmann et al., 2022), while CH show significantly less sensitivity, with a reported LC₅₀ of 67.3 μg/L (Lo et al., 2023). From the range-finding exposures, concentrations ≤ 1 μg/L were determined to be sublethal to RT and CH under experimental conditions. The CO were exposed to a wider scheme of concentrations from 0.001 μg/L to 10 μg/L and exhibited the highest sensitivity to 6PPD-q, with mortality occurring at all concentrations above 0.001 μg/L (Fig. S1B). Partial mortality was observed at 0.01 – 1 μg/L, and 100% mortality was observed at 1 μg/L and 10 μg/L. The reported 24-hour median lethal concentration (LC₅₀) for juvenile CO salmon ranges from 0.084 – 0.095 μg/L, depending on experimental conditions (Greer et al., 2023b; Tian et al., 2022). Sublethal exposure concentrations were determined to be ≤ 0.001 μg/L for CO. In both CO and RT, mortality was preceded by characteristic URMS symptoms including surface swimming and mouth gaping (Fig. S1A).

Since the identification of 6PPD-q in 2021, numerous aquatic species have been tested, yet salmonids remain the most sensitive. Comprehensive summaries of 6PPD-q toxicity across species are available in several review papers (Chen et al., 2023; Hua and Wang, 2023; Mayer et al., 2024). Given the heightened vulnerability of CO salmon, conservation efforts should prioritize stormwater treatment in CO habitats to mitigate 6PPD-q runoff from urban areas. Meanwhile, the absence of toxicity in CH salmon at concentrations up to 10 μg/L suggests that protective measures targeting CO could provide broad ecological benefits without unnecessary restrictions on less sensitive species. Previous studies have investigated the lethality of 6PPD-q in salmonids (Brinkmann et al., 2022; Di et al., 2022; Greer et al., 2023a). While lethality is a critical toxicological endpoint, sublethal effects must also be considered to further refine hypotheses regarding the compound’s mode of action. Consequently, this study was designed to assess concentrations that did not produce overt phenotypic effects in juvenile salmonids. For the five-day exposures, fish were exposed to 1 μg/L of both AQ and 6PPD-q, except for CO, which were exposed to 0.001 μg/L 6PPD-q due to their higher sensitivity. To optimize the observation of molecular effects preceding URMS symptoms and mortality, sublethal concentrations were prioritized over maintaining consistent exposure levels across species. These concentrations are environmentally relevant, as surface water samples from urban areas have reported levels ranging from 1 to 230 ng/L (Helm et al., 2024; Jaeger et al., 2024; Tian et al., 2022).

3.2. LCMS-Based Metabolomics

3.2.1. Metabolomics Data Quality and Reproducibility

The number of detected features remained consistent, varying only by 5% across species and treatments, demonstrating the reproducibility of the extraction method and data acquisition process (Fig. S3). Despite inherent baseline differences in metabolomes between species, the consistent extraction efficiencies across species confirm that interspecies trends and variations in metabolites among exposure treatments can be reliably compared. Additionally, there were no statistically significant differences between treatment groups or species when comparing the average peak areas of the internal standard surrogate L-Tyrosine (ring-13C6) (Fig. S4). This reproducibility provides confidence that the observed trends within each sample are attributable to the experimental treatments rather than variations in extraction efficiency or instrument performance. Data were further analyzed to explore effects of the treatments in salmonids relative to controls. Although fasting can influence the metabolome, principal component analysis (PCA) analysis revealed minimal differences between fed and fasted controls, indicating that fasting had little impact on the overall metabolic profile (Fig. S7 and S8). Thus, using the fasted control as the baseline provides a reliable foundation for assessing treatment-related metabolic changes (Text S7). On average, 7,935 ± 416 unique features were identified across all extractions (Fig. S3). Using an in-house library and open-source spectral libraries, 260 distinct metabolites were annotated across all treatments. Of these annotated metabolites, 23% were classified as a level 1 confidence matching MS, MS² and retention time information according to the Schymanski scale (Schymanski et al., 2014). Additionally, 59% were classified as a level 3 confidence matching MS and MS² but not retention time. In summary, 82% of the annotated metabolite matches were high confidence, confirmed, or probable structure matches.

3.2.2. Metabolomic Profiles of Salmonid Species Exposed to 6PPD-q and AQ

PCA illustrates a comparison of the metabolomic profiles among the species by highlighting differences in the expression of shared features. All species liver samples clustered together in the control and AQ treatments, demonstrating similar metabolic profiles. However, PCA plots of liver metabolomes revealed distinct separation between coho from the other two species in treatments containing 6PPD-q, including individual exposure and co-exposure regimes (Fig. 1). This divergence suggests a systemic metabolomic shift in CO in response to 6PPD-q. Fig. S5 reports PLS-DA plots and their associated important features driving differences between the metabolic profiles of the species in each treatment group. Among the metabolites with the highest variable importance in projection (VIP) scores when comparing the control species were palythine, glucuronolactone, ophthalmic acid and nicotinate, all of which are important for detoxification. This suggests different salmon species likely inherently possess varying levels of metabolites available for protective mechanisms against xenobiotics. For example, palythine serves as a photoprotective agent by absorbing UV radiation and exhibits antioxidant properties (Lawrence et al., 2018). Ophthalmic acid (OPH) was the highest VIP score in the 6Q treatment and the fifth most important feature amongst the controls (Fig. S5). OPH is a ubiquitous metabolite and modulates glutathione (GSH) utilization and transport, serving as a biomarker for oxidative stress (Schomakers et al., 2024). In both control and 6PPD-q exposed fish, OPH levels in salmon species were highest in CO, followed by RT, and lowest in CH. An elevation in OPH may result in competitive inhibition of GSH transport into the mitochondria, likely leading to mitochondrial dysfunction due to the depletion of antioxidative GSH (Schomakers et al., 2024).

Fig. 1.

Principal component analysis generated in MetaboAnalyst of annotated metabolites from each treatment to compare metabolomic profiles between species for livers. Metabolite intensities were normalized with log-transformation and Pareto scaling. Each point represents the 260 annotated metabolites for an individual fish replicate. Shaded regions represent the 95% confidence intervals.

3.2.3. 6PPD-q Inhibits Metabolic Activity in Salmonid Species

A heatmap of all modulated metabolites with Log₂ fold change (log₂FCs) from controls can be found in Supplementary Fig. S6. Hierarchical clustering analysis explores the variations between groups and modulated metabolites in Fig. S6 revealed CO and CH were most similarly impacted by 6PPD-q exposure than RT, despite their observed different sensitivity to 6PPD-q at the phenotypic level. Similarly, RT exposed to 6PPD-q also revealed an overall depletion of metabolic activity, however not to the same degree as CO and CH. The mortality observed in RT but not CH—despite less severe global metabolite depletion—suggests that trout may be uniquely sensitive to disruptions in specific energy pathways rather than broad metabolic suppression. RT may rely more heavily on aerobic metabolism and may be less resilient to impairments in glycolysis and the TCA cycle compared to CH, which may possess compensatory mechanisms that allow survival despite extensive metabolite loss.

A subset of modulated metabolites are reported in Fig. 2 as a heatmap of the log₂FC between treatments and controls. Many different amino acids were downregulated in response to 6PPD-q exposure (Fig. 2A). Amino acids are essential for protein synthesis, metabolism, immune function, neurotransmission, and hormone production (Ling et al., 2023). Additionally, metabolites involved in cellular respiration processes such as glycolysis and the TCA cycle were impacted in response to 6PPD-q exposure. Several different amino acids regulate metabolites essential for the TCA cycle, consequently, the depletion of amino acids likely inhibits energy production (Ling et al., 2023). In a recent study, 6PPD-q also decreased the glycolysis metabolite pool and TCA cycle intermediates in human neurons (Fang et al., 2024). These findings suggest that 6PPD-q disrupts fundamental biochemical pathways, potentially impairing energy metabolism and cellular function across multiple biological systems.

Fig. 2.

Heatmap of individual metabolite log₂ fold changes from control categorized by; (A) amino acids, (B) fatty acids and lipids, and (C) intermediate metabolites involved in cellular respiration. Treatments: 6PPD-q (6Q), 9,10-Anthraquinone (AQ), co-exposure (CE).

Quinones like 6PPD-q are highly reactive molecules that bind to cellular nucleophiles, such as amino acids, proteins, and DNA, disrupting cellular functions (Bolton et al., 2000). Quinones can also induce oxidative stress by binding to cellular reducing agents like GSH and NADH (Krylova et al., 2016). Furthermore, fatty aid depletion in salmon exposed to 6PPD-q also reduces the amount of Acetyl Co-A available for the TCA cycle via oxidative metabolism (Jiang et al., 2020, p. 9). Fatty acid depletion may be a result of the increase in their metabolism. Oleic acid (OA), a crucial monounsaturated fatty acid, was the metabolite with the largest log₂FC in CO exposed to 6PPD-q. OA is a precursor to Acetyl-CoA, metabolized by CYP450s via beta-oxidation (Houten and Wanders, 2010). Beta-oxidation is associated with increased oxidative stress and mitochondrial dysfunction by increasing the electron load on the ETC, producing damaging superoxide (Schönfeld and Wojtczak, 2008). Additionally, OA may play an antioxidative protective role in hepatocytes (Santa-María et al., 2023). The log₂FC decreases in OA in response to 6PPD-q treatment were −4.4, −4.0, and −3.3 in CO, RT, and CH, respectively. The OA depletion matches the salmonid species’ acute 6PPD-q sensitivity trend. These decreases in OA may prevent the activation of anti-inflammatory cytokines or inhibition of proinflammatory cytokines (Santa-María et al., 2023).

3.2.4. AQ Exposure enhances metabolic activity in sensitive salmonid species

In the environment, stormwater runoff and urban water bodies contain complex chemical mixtures, which may contribute to the high acute toxicity of 6PPD-q (Masoner et al., 2019b). PAHs can both inhibit and induce CYP450 enzymes, some studies suggest anthraquinones inhibit CYP450 metabolizing enzymes (Liu et al., 2021), which may reduce the detoxification efficiency of 6PPD-q when fish are co-exposed to both compounds. On the other hand, induction of CYP450 enzymes by PAHs could lead to an upregulation of receptors and enzymes available for detoxification of 6PPD-q. AQ exposure triggered widespread metabolite enrichment, 82–97% upregulated across species (Fig. S6) Anthraquinones are known to activate antioxidant response elements and phase II enzymes like glutathione S-transferases [76, 77], which could enhance detoxification capacity and stimulate compensatory metabolic activity. This upregulation may counteract oxidative stress, leading to enriched metabolite profiles even in co-exposed CO and RT. CO and RT co-exposure treatments also revealed an enrichment of modulated metabolites, 99% and 100%, respectively. In CO and RT, AQ may upregulate detoxification enzymes that concurrently metabolize 6PPD-q, reducing its bioaccumulation while maintaining metabolic activity. However, 99% of metabolites were downregulated in the CH co-exposure. As the least sensitive species, CH may allocate resources to survival-critical functions such as ion regulation and neuronal signaling over metabolic activity, manifesting as depletion in affected pathways. This pattern aligns with ecological risk models where acute tolerance correlates with reduced sublethal biomarker responsiveness (Rohr et al., 2016). These results indicate chinook may have an evolutionary adaptation to withstand brief, high-intensity exposures (e.g., seasonal contaminant pulses) however, this might come at the cost of impaired chronic mixture response mechanisms. Further research is needed to understand the complexity of roadway runoff mixtures and their toxicological effects to better address salmon mortality.

3.3. Salmonid Liver Transcriptomics GO Analysis

3.3.1. Differential Gene Expression and Functional Categorization in Salmonid Livers

More differentially expressed genes (DEGs) were identified in CO livers (175) in response to all treatments compared to RT (26) and CH (12) (Fig. S9). The salmonids used in the present study exhibited high transcriptomic variability, requiring an infeasibly high number of replicates to control for variance according to power analysis (Fig. S10). However, DEGs identified despite high transcriptomic variance likely play important roles in the molecular modulations involved in the toxicity of 6PPD-q. GO analysis revealed DEGs involved in 23 different functional categories across all species and treatments (Fig. 3). Categories with the most DEGs in CO across treatments included apoptosis and cell death, immune response, metabolism and biosynthesis, and ubiquitination (Fig. 3).

Fig. 3.

Salmon liver log2FC DEG heatmap generated using RStudio grouped by gene ontology (GO) analysis molecular function categories, species, and treatment. GO analysis was performed using a combination of data outputs from PANTHER, DAVID, and g:Profiler software to categorize genes based on molecular function. Treatments: 6PPD-q (6Q), 9,10-Anthraquinone (AQ), co-exposure (CE).

3.3.2. Vascular Permeability Disruptions in Coho Salmon Exposed to 6PPD-q

CO GO terms differentially regulated by 6PPD-q included pathways related to molecular stress and immune response, consistent with findings in exposed embryos (Greer et al., 2023b). Genes linked to vascular growth (platelet derived growth factor – PDGFA Rab11 family-interacting protein 3 – RAB11FIP3) and hemostasis/anticoagulation (Annexin 5 – ANXA5, coagulation factor VII – F7, (Protein S – PROS1) were also affected by sub-phenotypic exposure. Indication of altered vascular permeability in response to sublethal 6PPD-q exposure was observed in the present study and in Greer et al. (2023), as evidenced by perturbations in genes involved in hemostasis, anticoagulation, and vascular endothelial growth. Greer et al. (Greer et al., 2023b) proposed that an inflammatory response may drive increased vascular permeability, leading to capillary leak syndrome (CLS) and contributing to symptoms of urban runoff mortality syndrome (URMS). Furthermore, apoptotic and endothelial cell death pathways, previously linked to CLS [63], were also activated in the present study. These findings support the hypothesis that URMS symptoms and mortality in 6PPD-q-exposed salmon result may be connected to CLS.

3.3.3. 6PPD-q induces cellular stress and immune responses in coho salmon

Two genes in the present study linked to apoptosis regulation were differentially expressed in CO salmon exposed to 6PPD-q. PSMC2 was downregulated, and evidence suggests inhibition of PSMC2 promotes apoptosis (Zheng et al., 2022). Similarly, PARP14, a well-established anti-apoptotic regulator (Iansante et al., 2015), was also downregulated in CO exposed to 6PPD-q. Several genes associated with other cellular stress responses were differentially regulated in CO exposed to 6PPD-q. Heat Shock Protein Family A Member 1 Like (HSPA1L) is involved in stress and apoptotic mechanisms (Cheng et al., 2024; Daugaard et al., 2007). HSPA1L was upregulated in response to 6PPD-q, which likely indicates a cellular defense response to stress. UBIAD1 is a nonmitochondrial CoQ10 biosynthetic enzyme that has been shown to provide protection against oxidative damage (Mugoni et al., 2013). UBIAD1 was upregulated the CO6Q treatment, suggesting a cellular response to mitigate ROS. Additionally, DDB1 And CUL4 Associated Factor 6 (DCAF6) was also upregulated the CO6Q treatment. DCAF6 is utilized by the CRL4 E3 ligase complex to target proteins for ubiquitination (Sang et al., 2015). DCAF6 upregulation may suggest an increased protein degradation response to endoplasmic reticulum ER stress (Lin et al., 2008). Similarly, HECT and RLD ubiquitin protein ligase family member 1 (HERC1) was upregulated in CO6Q, indicating an increased need for protein ubiquitination (Montes-Fernández et al., 2020). Furthermore, endoplasmic reticulum to nucleus signaling 1 (ERN1), also referred to as IRE1, was also upregulated. ERN1 is part of the unfolded protein response (UPR) as a protective mechanism under metabolic and ER protein misfolding stress (Lindholm et al., 2017). Functions of ATP-dependent proteins in the ER are decreased in response to reduced glucose, leading to proteostasis dysfunction and activating the UPR. Metabolic signaling is part of the UPR response, and research has shown that the activation of ERN1 can lead to the reduction of glycolysis and mitochondrial respiration (van der Harg et al., 2017). Increased transcription of genes involved in apoptosis, oxidative stress protection, and ubiquitination all indicate 6PPD-q induces cellular stress and elicits compensatory transcription at sublethal concentrations (Cosulich and Clarke, 1996; Maxwell et al., 2021; Rogers, 2020).

Several immune system response genes were differentially regulated in CO exposed to 6PPD-q, including genes such as nucleotide-binding leucine-rich repeat protein 12 (NLRP12), complement factor (CFH), complement component 2 (C2), TAP Binding Protein Like (TAPBPL), and DExD/H-box helicase 58 (DDX58) (Fig. 3 and 5). CFH was downregulated in the CO6Q treatment, which could lead to activation of the alternative complement system, a key component of the innate immune system (Noris and Remuzzi, 2013). Moreover, complement component 2 was upregulated in the COCE exposure group, which can be activated in response to inflammatory signals. These markers of immune activation in response to the co-exposure may suggest exacerbation of 6PPD-q toxicity when co-exposed to a PAH. The combination of these DEGs involved in the complement pathway may also indicate a compensatory response to the inhibition of the lectin pathway via activation of the alternative pathway. DDX58 was upregulated in the CO 6PPD-q exposure, which is a key player in the innate immune system and triggers the interferon signaling pathways to induce an antiviral response (Prasov et al., 2022). Additionally, NLRP12 was upregulated in the CO6Q and COCE treatment groups and negatively regulates the NF-kB signaling pathway; activation of this pathway induces an inflammatory response by increasing the secretion of the cytokine interleukin (IL)-1β and IL-18 (Tunçer Çağlayan et al., 2014). Moreover, (IL)-1β was the gene most significantly upregulated transcript in RT embryos exposed to 6PPD-q in Greer et al. (2023). Exposure to 6PPD-q may itself be recognized as non-self by innate immune system pattern recognition receptors and activating innate immune pathways, but it is more likely that exposure to 6PPD-q leads to tissue damage that elicits a pro-inflammatory cascade as a protective response. A study evaluating sublethal 6PPD-q exposure in C. elegans also observed an activation of genes involved in innate immunity (Wang and Wang, 2024). Very few genes were differentially expressed in response to road contaminants in CH liver samples; of the five DEGs across all treatments, two are involved in immune system activation. SOCS3 was upregulated in all treatments and complement component 3 (C3) was upregulated in response to the CE treatment in CH. Similarly, only one gene - UNC93B1 was differentially expressed in RT exposed to CE. C3 and UNC93B1 immune response genes cannot be attributed to 6PPD-q exposure, as they were only observed in the co-exposure with AQ. The data suggests CO exhibited more robust immune system changes in response to 6PPD-q compared to the other two species studied.

Figure 5.

Metaboanalyst joint pathway enrichment analysis with both significant DEGs and metabolites input from coho and rainbow trout livers. Pathways represented with a circle were significantly impacted by chemical exposure (p ≤ 0.05, according to a Holm-Bonferroni adjustment). Treatments: 6PPD-q (6Q), 9,10-Anthraquinone (AQ), co-exposure (CE).

3.4. Mechanistic Insights into 6PPD-q Toxicity

One of the primary advantages of investigating endogenous small molecules is that it allows for the generation of hypotheses about mechanisms that cannot be easily observed by identifying relationships between pollutants, transcriptional, and translational responses (Gil-Solsona et al., 2021; Koh and Hwang, 2019). Several hypotheses for 6PPD-q toxicity include mitochondrial dysfunction, blood-brain barrier disruption, cardiorespiratory effects, and variable detoxification rates (Blair et al., 2021; Mahoney et al., 2022; Montgomery et al., 2023a; Selinger et al., 2025). A molecular summary of sublethal 6PPD-q toxicity observed in coho in this study is depicted in Fig. 4. Further discussion of the compensatory induction of mitochondrial respiration genes, ROS production, inhibition of 6PPD-q metabolism, and subsequent stress/immune response follows.

Fig. 4.

Proposed 6PPD-q mechanism of toxicity in CO salmon based on gene ontology and metabolite pathway analysis. Sublethal 6PPD-q toxicity in CO includes compensatory induction of mitochondrial respiration genes, depletion of glycolysis and TCA cycle metabolites, increased expression of mitochondrial transport and ETC components and xenobiotic metabolizing enzymes and associated immune/stress responses. Figure created with BioRender.com.

3.4.1. Chemical Exposures Disrupt Key Pathways Involved in Energy Production, Immune Function, and Cellular Homeostasis

Modulated metabolite HMDB IDs and DEGs were used as input for integrated joint-pathway analysis. Pathways significantly affected by chemical exposure are presented in Fig. 5. A total of 18% of the annotated metabolites did not have corresponding HMDB IDs, as a result, only 82% of the 260 annotated metabolites were used for pathway analysis. Metabolites were predominantly depleted in response to 6PPD-q and enriched following AQ exposure. There were no pathways significantly impacted specific to AQ exposure alone. Co-exposures effects varied, with CO and RT showing primarily metabolite enrichment, while CH exhibited metabolite depletion. One pathway, ABC transporters, was impacted across all treatments and species compared to control. ABC transporters export xenobiotics out of the cell via hydrolysis, as well as import nutrients and other molecules into the cell (Miller, 2010). Several pathways involving essential amino acids were significantly impacted in the 6Q and CE groups including alanine, aspartate, glutamate, histidine, cystine, glycine, serine and threonine metabolism. All these amino acids play important roles in energy production (Chandel, 2021a), and their depletion can inhibit essential cellular functions leading to impaired immune function, metabolic disorders, and cell death (Ling et al., 2023). Additionally, the β-Alanine metabolism pathway was impacted in response to 6PPD-q exposure. β-alanine is a non-essential amino acid with many different metabolic functions. There is evidence beta-alanine can improve oxidative metabolism by increasing glucose transport, lipid metabolism, and mitochondrial biosynthesis (Schnuck et al., 2016). The decrease in metabolite intermediates involved in β-alanine metabolism may lead to deficient energy output. β-alanine metabolism was significantly affected in RT and CH exposed to 6PPD-q, but not CO; however, this is likely due to the lower exposure concentration used for CO in the present study. Folate transport and metabolism and one-carbon pool by folate pathways were significantly impacted in CH exposed to 6PPD-q. Folate is essential for many different metabolic processes including amino acid synthesis and mitochondrial protein translation, and the impairment of these processes can inhibit mitochondrial respiration and cell proliferation (Zheng and Cantley, 2019). Folate metabolism impairment has also been linked to diseases such as anemia and tumor metastasis (Zheng and Cantley, 2019). Purine metabolism was impacted by 6Q and CE exposure, in CO and CH. Purines such as adenine and guanine play a crucial role in energy metabolism and immune system signaling (Pedley and Benkovic, 2017). Purines decreased in this study included xanthine, hypoxanthine, and guanosine monophosphate, however these metabolites were elevated in the spleens of mice exposed to 6PPD-q (Jia et al., 2025). Additionally, fatty acid biosynthesis was affected by 6PPD-q exposure in all species, further represented by the depletion of fatty acids and lipids in Fig.2. This suggests the inhibition of lipid synthesis, potentially impacting membrane formation, energy storage, or cellular signaling (Khan et al., 2024). These findings highlight widespread disruptions to key metabolic pathways involved in energy production, immune function, and cellular maintenance.

3.4.2. Sublethal 6PPD-q Inhibits Metabolic Activity and Induces Compensatory Mitochondrial Responses in Coho

Energy production begins with glucose metabolism via glycolysis and the tricarboxylic acid (TCA) cycle, yielding the reducing equivalents nicotinamide adenine dinucleotide (NADH) and flavin adenine dinucleotide (FADH_₂). These molecules then transfer electrons to the electron transport chain through a series of redox reactions (Jonckheere et al., 2012). Glycolysis is the first step of cellular respiration and crucial for ATP production under anaerobic conditions (Chandel, 2021b). Several glycolysis intermediates were downregulated in all species exposed to 6PPD-q including glucose-6-phosphate, fructose-6-phosphate, glycerol 3-phosphate, and dihydroxyacetone phosphate (Kanehisa et al., 2021a). Similarly, intermediates of the TCA cycle: citrate, succinate, fumarate, and malate (Kanehisa et al., 2021b) were also downregulated in all species exposed to 6PPD-q alone. The citrate cycle plays a vital role in transforming the breakdown products of carbohydrates, fats, and proteins into high-energy electron carriers which drive ATP production via oxidative phosphorylation (Arnold and Finley, 2022). The decrease in cellular intermediates indicates disruption of energy production in salmonids exposed to 6PPD-q. Additionally, the detection of OPH highest in CO and lowest in CH may indicate a lack of antioxidative protection and potential mitochondrial dysfunction (Schomakers et al., 2024). Decreased cellular energy production has many molecular consequences including inhibition of muscle contraction, nerve impulses and biosynthesis (Liu et al., 2025).

In addition to metabolic evidence, several DEGs involved in mitochondrial function were differential expressed in response to 6PPD-q exposure. Metastasis-associated in colon cancer 1 (MACC1) was upregulated in response to 6PPD-q, its overexpression has been shown to inhibit mitochondrial ATP production (Lisec et al., 2021). MACC1 was also found to enhance the uptake of glucose, glutamine and pyruvate in metabolic pathways in response to their depletion (Lisec et al., 2021). B4GALT1 was upregulated, which plays an important role in the metabolism of galactose to glucose-1-phosphate, a precursor for glycolysis (Conte et al., 2021). Mitochondrial pyruvate carrier (MPC2) was upregulated, suggesting an increased need for mitochondrial pyruvate uptake (Halestrap, 1978; Papa and Paradies, 1974). Additionally, isocitrate dehydrogenase 1 (IDH1) expression was upregulated. IDH1 is responsible for decarboxylating isocitrate to α-ketoglutarate, a crucial step in the TCA cycle, indicating its upregulation may be in response to low levels of α-ketoglutarate available for use in the TCA cycle (Bergmeyer, 1974). Furthermore, genes involved in complexes I (NDL4 and NDUFAB1) and IV (COX5A) of the electron transport chain (ETC) were also upregulated. The upregulation of NDUFAB1 has been shown to enhance mitochondrial bioenergetics and reduce oxidative stress via the ETC (Giannos et al., 2022; Hou et al., 2019). There has been evidence of mitochondrial dysfunction in response to 6PPD-q in several other studies and organisms including RT liver cells (Mahoney et al., 2022), C. elegans (Hua et al., 2024), and human neuronal cells (Fang et al., 2024). In the present study, for the first time, an upregulation of genes involved in mitochondrial respiration in CO in response to 6PPD-q was observed. Interestingly, Hua et al. (2024) observed the inhibition of mitochondrial complexes I and III in C. elegans in response to lifespan reducing concentrations of 6PPD-q (Hua et al., 2024). It is likely that the sub-phenotypic concentrations of 6PPD-q used in the present study elicited an induction of mitochondrial genes as a compensatory response to counter act the depletion of metabolites necessary for energy production.

Several other recent studies have observed results corroborating mitochondrial impairment and a decrease in metabolic activity. A recent study investigating metabolic alterations in RT following 6PPD-q exposure identified dysregulated metabolites linked to mitochondrial dysfunction, disrupted energy metabolism, and enhanced vascular permeability (Ankley et al., 2025). Additionally, analysis of mouse primary dopaminergic neurons found 6PPD-q reduced mitochondrial respiration capacity, elevated ROS, and decreases in mitochondrial metabolite intermediates (Fang et al., 2024). The study corroborates the present findings that 6PPD-q’s mechanism of action is likely related to dysregulation of energy metabolism. Furthermore, URMS symptoms can be attributed to mitochondrial disruption-induced oxidative stress, cell death, and respiratory failure. Tilapia lake virus collapses mitochondrial membrane potential and oxidative damage in fish gill cells rich in mitochondria, inhibiting oxygen uptake and ATP production (Raksaseri et al., 2023). This mitochondrial disruption sparks fish to gasp at the water surface for oxygen, as seen in salmon exposed to 6PPD-q. Additionally, mitochondrial disruption related to hypoxia and excess ROS production has been linked to mouth gaping and surface swimming in killifish (Du et al., 2016). A previous study suggests the uncoupling of oxidative phosphorylation in the ETC from 6PPD-q exposure in RT gill cells, mimicking hypoxia and inducing surface gasping (Mahoney et al., 2022).

3.4.3. Xenobiotic metabolism

Observed for the first time, a metabolizing enzyme was found to be differentially regulated in salmonid species exposed to 6PPD-q. CYP3A274 was downregulated in response to 6PPD-q exposure in CO liver tissue (Fig. 3). It has been established that similar CYP450 enzymes, such as CYP1A1, 3A4, and 2C19 play crucial roles in the metabolism of 6PPD-q to hydroxylated metabolites in humans (S.-Y. Zhang et al., 2023). This transformation is proposed to be a detoxification step, as OH-6PPD-q is observed at higher concentrations in species that do not present symptoms of 6PPD-q toxicity (Montgomery et al., 2023b). Additionally, a study found OH-6PPD-q was non-toxic to coho embryo cells and its formation was inhibited in combination with a general CYP450 inhibitor, supporting the hypothesis that OH-6PPD-q is a detoxification product via CYP450s (Nair et al., 2025). Correlating to these previous findings, the present study observed a significant reduction in CYP3A274 expression in CO, but not RT or CH. This trend may indicate an inhibition of the detoxification of 6PPD-q in more sensitive species, leading to increased toxic effects compared to species with higher CYP450 activity. Additionally, CYP3A27 was downregulated in the CE treatment in RT, indicating 6PPD-q and PAH chemical co-exposure may inhibit expression of CYP450 metabolizing enzymes. Furthermore, there is evidence that immune responses can suppress metabolizing enzymes expression via oxidative stress and cytokine signaling (Renton, 2004). The observed upregulation of NLRP12 can activate interleukin (IL) pathways, inhibiting CYP450 expression (Jover et al., 2002; Xun et al., 2021). Further research is necessary to elucidate fish biotransformation pathways of 6PPD-q and potential CYP450 inhibition mechanisms.

4. Environmental Implications

This study advances the understanding of 6PPD-q toxicity by exploring its molecular mechanisms and the effects of co-exposure with other common quinone pollutants. It provides crucial insights into interspecies differences, highlighting the heightened sensitivity of CO salmon compared to CH and RT. The study also suggests that in species like CH, Phase 1 metabolizing enzymes may facilitate detoxification of 6PPD-q. In contrast, incomplete detoxification in CO salmon leads to mitochondrial dysfunction and increased vulnerability. These findings, consistent with prior research, reinforce the need for species-specific toxicity thresholds in environmental risk assessments (Magnuson et al., 2020; Meador et al., 2020). While the molecular initiating event of 6PPD-q is not yet elucidated, the present study hypothesizes: 1) URMS symptoms and mortality in coho salmon are likely driven by excessive ROS accumulation resulting from 6PPD-q redox cycling and enhanced electron transport chain activity, exacerbated by impaired detoxification due to downregulation of CYP450 metabolizing enzymes; and 2) AQ exposure modulates 6PPD-q toxicity in salmon by inducing antioxidant and xenobiotic metabolism pathways, which may enhance detoxification and reduce oxidative stress. Efforts to mitigate 6PPD-q toxicity should prioritize CO habitats while integrating strategies to address the cumulative impacts of complex pollutant mixtures in aquatic ecosystems. This research underscores the complexity of stormwater mixtures and their combined toxicological effects. Notably, co-exposure to compounds like anthraquinone was shown to modulate the toxic response to 6PPD-q, adding another layer of complexity to environmental risk assessments. Future studies using targeted qPCR analysis to quantify CYP450 enzymes, ATP, and ROS would advance the knowledge of 6PPD-q toxicity. Further studies should also explore sublethal and chronic effects, particularly in the context of the metabolomic and transcriptomic alterations documented here, to better understand the long-term ecological risks posed by 6PPD-q.

Synopsis.

This study reveals species-specific 6PPD-q toxicity mechanisms and highlights co-exposure effects, informing risk assessments for runoff pollutant mixtures.

Highlights.

• Depletion of cellular energy intermediates occurs after 6PPD-q exposure in salmonids.

• Co-exposure with AQ shifts 6PPD-q metabolomic effects in a species-dependent manner.

• Coho showed greater gene expression changes than chinook or trout after 6PPD-q exposure.

• Mitochondrial compensatory gene induction follows sublethal 6PPD-q exposure in coho.

• Immune and vascular genes are upregulated in coho at sublethal 6PPD-q concentrations.

• First evidence that CYP3A274 expression is suppressed by 6PPD-q exposure in coho.

Data Availability

Metabolomics data are publicly available through the Metabolomics Workbench (Project DOI: 10.21228/M84V61). All remaining data are available from the corresponding author upon reasonable request.