Biological Responses of Pacific Herring Embryos to Crude Oil Are Quantifiable at Exposure Levels Below Conventional Limits of Quantitation for PAHs in Water and Tissues Authors: John P. Incardona, Tiffany L. Linbo, James R. Cameron, Barbara L. French, Jennie L. Bolton, Jacob L. Gregg, Carey E. Donald, Paul K. Hershberger, Nathaniel L. Scholz Published: 2023 ABSTRACT Pacific herring (Clupea pallasii), a cornerstone of marine food webs, generally spawn on marine macroalgae in shallow nearshore areas that are disproportionately at risk from oil spills. Herring embryos are also highly susceptible to toxicity from chemicals leaching from oil stranded in intertidal and subtidal zones. The water-soluble components of crude oil trigger an adverse outcome pathway that involves disruption of the physiological functions of cardiomyocytes in the embryonic herring heart. In previous studies, impaired ionoregulation (calcium and potassium cycling) in response to specific polycyclic aromatic hydrocarbons (PAHs) corresponds to lethal embryolarval heart failure or subtle chamber malformations at the high and low ends of the PAH exposure range, respectively. Sublethal cardiotoxicity, which involves an abnormal outgrowth (ballooning) of the cardiac ventricular chamber soon after hatching, subsequently compromises juvenile heart structure and function, leading to pathological hypertrophy of the ventricle and reduced individual fitness, measured as cardiorespiratory performance. Previous studies have not established a threshold for these sublethal and delayed-in-time effects, even with total (∑)PAH exposures as low as 29 ng/g of wet weight (tissue dose). Here, we extend these earlier findings showing that (1) cyp1a gene expression provides an oil exposure metric that is more sensitive than typical quantitation of PAHs via GC–MS and (2) heart morphometrics in herring embryos provide a similarly sensitive measure of toxic response. Early life stage injury to herring (impaired heart development) thus occurs below the quantitation limits for PAHs in both water and embryonic tissues as a conventional basis for assessing oil-induced losses to coastal marine ecosystems. Introduction Pacific herring have long been a focal species for studying crude oil toxicity to marine fish, dating back to the ecosystem-scale natural injury assessments that followed the 1989 Exxon Valdez oil spill in Prince William Sound, Alaska.1−5 In the decades since, studies using oiled gravel effluent methods to generate controlled chemical leaching established broadly quantitative exposure–response relationships for biological end points at the molecular, cellular, organ, and whole animal scales. Delayed-in-time toxicity has also been a major focus area, whereby sublethal embryonic abnormalities propagate through hatching and early larval development, with subsequent reductions in individual fitness in surviving juveniles.6−8 These herring studies, combined with similar findings from a wide range of fish species globally,9−13 identified the heart and cardiovascular system as the primary target of acute and latent petroleum developmental toxicity. Importantly, end points that measure cardiac function or cardiac morphology, which are closely intertwined, show statistically robust and highly repeatable dose–response relationships for measures of polycyclic aromatic hydrocarbons (PAHs) in both exposure waters and embryonic tissues.14−16 Despite these major leaps forward in understanding how fish early life stages respond to spilled oil, the exposure dimension of natural resource injury assessments has changed very little over the past 30 years. In the early 1990s, environmental levels of PAHs in Prince William Sound water, sediments, and organismal tissues were typically quantified by gas chromatography/mass spectrometry (GC/MS).2,17−19 Similar methods were used to characterize exposure following more recent spills, with examples including the 2002 Prestige spill in Spain,20 the 2007 Hebei Spirit spill in South Korea,11 the 2010 Deepwater Horizon disaster in the Gulf of Mexico,10,12 and the 2015 Refugio Beach spill in Southern California.21 In the present study, we explore whether these conventional analytical methods for PAHs and other petroleum-derived compounds are sufficiently sensitive to determine thresholds for injury in fish, particularly “damage” predicated on novel forms of developmental cardiotoxicity including fitness losses that are delayed-in-time. Fertilized herring embryos normally develop to hatch over a period of about 2 weeks (12–15 days at ∼9–10 °C). The heart tube forms initially in a lateral orientation and produces irregular contractions by 4.5 days postfertilization (dpf). A regular heartbeat is established the following day, and cardiac organogenesis occurs thereafter (6–10 dpf), whereby the heart rotates into an anterior-posterior orientation, the atrioventricular canal becomes clearly demarcated, and the atrium folds in dorsally at the atrioventricular junction (chamber looping). The ventricle remains a mostly linear tube until hatch, whereafter the chamber begins a rapid posterior outgrowth or ballooning. During the early larval developmental phase, as ballooning progresses, the simple single-layer myocardium of the ventricle begins to form inward projections, or trabeculae, that ultimately make up the inner spongy myocardium of the mature heart. By juvenile stages, the heart is mostly spongy myocardium, surrounded by a multilayered compact myocardium. As in zebrafish,22 the overall shape of the heart is established in late larvae. Thereafter, the proliferative addition of new cells simply increases the heart size as the animal grows to adulthood. Because heart function and form are inseparable during development, perturbation of cardiomyocyte function in oil-exposed embryos leads to altered morphogenetic trajectories, which in turn lead to delayed-in-time impairments at later life stages. Based on average molecular weights of component PAHs, crude oil-derived mixtures with total aqueous (∑)PAH concentrations in the nanomolar range disrupt both K+ and Ca2+ ion fluxes that regulate cardiomyocyte action potentials and excitation-contraction coupling.23 Consequently, abnormal contractility leads to altered fluid dynamics within the forming cardiac chambers, in turn directly and indirectly affecting key cell shape changes and migrations that further refine the ongoing shaping of the heart.8,9,24,25 As has been detailed extensively in zebrafish,26−28 the molecular and cellular pathways that integrate normal contractile function with morphogenesis have multiple direct and indirect links with intracellular Ca2+ handling. Consequently, the phenotypic effects of oil exposure on herring heart development closely mirror the phenotypic effects of genetic disruption of these pathways in zebrafish.8,24 In both species, defects in ventricular ballooning and trabeculation lead to nonlethal impairments of embryolarval heart form and pathological hypertrophy at later life stages.7,8,29 At the same time, with complete loss of function for some genes in zebrafish (e.g., encoding contractile proteins30) or exposure to the higher end of oil concentrations,6,25,31,32 severe cardiac defects lead to embryolarval mortality. For crude oil exposure, this variation in specific end points depends on the effective PAH dosing for water and tissues. Prior studies in herring covered a tissue ∑PAH dose range of 29–8100 ng/g wet weight,6−8 which corresponds to roughly 0.14–39 μM for the sake of comparison to the aforementioned in vitro physiological studies. Those rapid, direct effects of crude oil mixtures on critical cardiomyocyte ion fluxes occurred at a very low ∑PAH IC50 dose range of 0.17–0.29 μM.23 This falls at the extreme low end of the embryonic tissue dose range for robust cardiotoxicity in terms of dysregulating both the form and function of the developing heart. Across an upper dose range of ∑PAH 620–8100 ng/g, acute exposures during early to middle stages of cardiac organogenesis (8 dpf) led to a dose-dependent reduction in heart rate (bradycardia), an increase in heartbeat irregularity, and a reduction in chamber contractility.6 Tissue ∑PAH concentrations of ≥3000 ng/g (∼14 μM) were lethal to embryos by the hatching stage. Conversely, exposures to ∑PAH at concentrations ≤240 ng/g (∼1.2 μM) did not significantly reduce hatch rates, allowing assessment of downstream impacts on heart development.7,8 Across this lower dose range of 29–240 ng/g wet weight, oil exposures produced a robust and dose-dependent reduction in the degree of ventricular ballooning measured at hatch.8 The chamber did eventually expand posteriorly by the 2nd week posthatch, but defects in the trabeculation process were subsequently evident in larvae at 67 dpf, followed by hypertrophy of the spongy myocardium in early juveniles at 125 dpf. Consistent with these delayed anatomical defects, herring embryos that survived a ∑PAH body burden of 29 ng/g (∼0.14 μM) and were subsequently grown out in clean seawater for 9 months showed subtle alterations in juvenile ventricular shape and consequent reductions in swimming performance.7 Consequently, transient and trace exposures to mixtures containing PAHs trigger a pathological hypertrophic remodeling of the developing heart, initiated later in life secondary to delayed ventricular ballooning and defective trabeculation in planktonic larvae. We recently observed the same effect on ventricular ballooning in a distantly related species with a very different final heart structure, Atlantic haddock.13 This species showed a similar sensitivity, with an effect at the lowest tested ∑PAH concentrations of 0.1 and 27 ng/g in water and tissues, respectively. Importantly, this was observed with a geologically distinct crude oil (Norwegian Sea Heidrun), using an entirely different exposure route: whole oil droplets dispersed into the water column.13 In both herring and haddock, the induction of mRNA for the PAH-metabolizing enzyme cytochrome P4501A (CYP1A) was quantified at these low-level exposures. However, CYP1A plays no role in crude oil cardiotoxicity as demonstrated by gene knockdown studies in zebrafish, and instead plays a protective role in detoxification.33,34 Nevertheless, measures of cyp1a mRNA are clearly readily quantifiable exposure metrics for bioavailable PAHs in tissues. Across 3 decades of oil spill research on marine forage fish, the whole-organ developmental impacts observed at the hatching stage (and later) are readily related to the preceding degree of cardiac dysfunction caused by oil-derived PAHs. However, previous studies have not fully characterized the lower end of the exposure–response range for sublethal cardiotoxicity, with the goal of establishing a no-effect concentration. Here, we use oiled gravel preparations from a prior study, further weathering experimental PAH generator columns to produce even lower exposure levels for embryonic herring, that is, spanning the limits of quantitation (LOQ) for PAHs in water and tissues. We used the PAH-inducible expression of cyp1a as a tissue-dose exposure metric and cardiac chamber anatomy at hatch as a biological effect measure. Compared with the most widely used, conventional chemical analyses for PAHs in both water and tissues, the biological measures were more sensitive. Materials and Methods Production of Herring Embryos Ripe herring were captured by the Washington Department of Fish and Wildlife on April 20, 2017, by gill net from the Cherry Point stock in northern Puget Sound, WA, and the study carried out at the United States Geological Survey’s Marrowstone Field Station. Seawater for all aspects of the study was drawn from 150 m offshore in high current Admiralty Inlet (remote northern Puget Sound) at 17 m below mean lower low tide and treated with sand filtration and UV sterilization. Intact herring were transported on ice and dissected upon receipt for immediate fertilization. Testes were dissected from 7 males (21.6 ± 5.9 g average weight) and ovaries from 17 females (28.4 ± 6.7 g). Testes were macerated and pooled to provide sperm to fertilize all the eggs stripped and pooled from the ovaries (483 g of source tissue from the 17 females) using the poly(vinyl alcohol) method35 to prevent clumping and control adherence while evenly distributing onto 16 21 × 24 cm sheets of 1 mm nylon mesh. Fertilized eggs were incubated overnight by suspension of mesh sheets in 4 ft diameter flow-through seawater tanks and assessed for fertilization rates before transfer to column effluent tanks. Mean fertilization success was ∼90% (n = 16 sheets, 89.4 ± 5.6%) Oil Exposure To produce oiled gravel effluents with relatively low dissolved PAH concentrations, we continued weathering gravel from an earlier study.8 As described previously, gravel was coated with Alaska North Slope crude oil at three dosing levels (0.25, 0.5, and 1.0 g oil/kg gravel) and loaded into identical PVC columns (∼1000 cm3 of gravel per column), each plumbed to flow into 114 L glass aquaria in which herring embryos were exposed. Prior to the current study, the gravel was used in 20168 to generate effluents at a flow rate of 3.6 L/h for a total of 26 days at 10.5 ± 0.2 °C, then stored at −20 °C for 1 year. For this study, columns were reactivated and subsequently weathered for an additional 9 days at 3.6 L/h (9.9 ± 0.2 °C) before initiating the exposure reported here. For exposure, 4 sheets with adherent embryos were randomly selected from the pool of 16 to suspend in each of the column effluent tanks (one tank per dosing level, three oil levels plus clean gravel control). Standpipes kept the steady-state volume of the exposure tanks at ∼100 L, which were placed within larger plastic baths plumbed with high flow ambient seawater for temperature control during incubation. Column flow rates were targeted for 7.2 L/h and adjusted on a daily basis; final average flow rates were 6.7 ± 0.3 L/h for control, 7.1 ± 0.2 for 0.25, 6.9 ± 0.8 for 0.5, and 7.2 ± 0.4 for 1.0 g/kg oil. Temperature, pH, and dissolved oxygen were monitored daily and were (control) 10.0 ± 0.3 °C, 7.7 ± 0.3, and 9.3 ± 0.1 mg/L; (0.25 g/kg oil) 9.9 ± 0.2 °C, 7.7 ± 3, and 9.4 ± 0.1 mg/L; (0.5 g/kg oil) 9.9 ± 0.2 °C, 7.7 ± 0.3, and 9.3 ± 0.1 mg/L; and (1.0 g/kg oil) 9.9 ± 0.2 °C, 7.7 ± 0.3, and 9.4 ± 0.1 mg/L, respectively. At 10 dpf (exposure day 9) embryos were transferred to 760 L hatching tanks with clean seawater with flow rates of 2–4 L/min. Hatching initiated the following day (11 dpf) and was considered complete by 16 dpf. Hatching was quantified by counting empty eggshells in randomly imaged areas of each mesh sheet (n = 8 or 12). Analytical Chemistry 200 mL water samples (one per treatment) were collected by glass pipet from the center of aquaria on exposure days 1 and 9, stabilized with 20 mL dichloromethane and stored at 4 °C until extracted and analyzed by GC–MS as detailed elsewhere.8,36 For body burden analyses, 100 embryos were randomly selected from the initial fertilized pool and from each exposure group at 10 dpf (exposure day 9), flash frozen on liquid nitrogen, and stored at −80 °C. Total mass of samples averaged 0.122 ± 0.052 g (s. d., N = 5). Subsequent microscale tissue extraction and GC–MS/MS analysis followed established methods.37,38 LOQ for PAHs were based on the GC/MS response areas of calibration standards. For water analyses, calibration standards range from 0.001 to 3.00 μg/mL. With instrument injection volumes of 2 μL, the lowest quantifiable mass was 0.006 ng. With the concentration of extracts, the actual LOQ was 0.002 μg/L. If a sample analyte was not detected or had a response area less than that of the lowest concentration calibration standard, it was reported as less than the LOQ values (