Embryonic Crude Oil Exposure Impairs Growth and Lipid Allocation in a Keystone Arctic Forage Fish.

As Arctic ice recedes, future oil spills pose increasing risk to keystone species and the ecosystems they support. We show that Polar cod (Boreogadus saida), an energy-rich forage fish for marine mammals, seabirds, and other fish, are highly sensitive to developmental impacts of crude oil. Transient oil exposures ≥300 μg/L during mid-organogenesis disrupted the normal patterning of the jaw as well as the formation and function of the heart, in a manner expected to be lethal to post-hatch larvae. More importantly, we found that exposure to lower levels of oil caused a dysregulation of lipid metabolism and growth that persisted in morphologically normal juveniles. As lipid content is critical for overwinter survival and recruitment, we anticipate Polar cod losses following Arctic oil spills as a consequence of both near-term and delayed mortality. These losses will likely influence energy flow within Arctic food webs in ways that are as-yet poorly understood. Published by Elsevier Inc.

Introduction

Crude oil contains polycyclic aromatic hydrocarbons (PAHs) that are cardiotoxic. Three-ringed PAH families (e.g., phenanthrenes) enriched in crude oil block K+ and Ca2+ ion conductances in cardiomyocytes, disrupting the normal rhythmic pumping of the heart (Brette et al., 2014, Brette et al., 2017). When this occurs in oil-exposed fish embryos, disruption of cardiac function leads to abnormal heart development (Incardona, 2017, Incardona and Scholz, 2016). Although cardiocirculatory defects alone would be sufficient to impact growth, more recent findings indicate that reduced cardiac function during embryonic and early larval development has other indirect effects that may be equally if not more consequential for individual fitness. Specifically, recent advances in RNA sequencing of oil-exposed Atlantic haddock (Melanogrammus aeglefinus) embryos identified alterations in the expression of genes involved in lipid metabolism (Sørhus et al., 2017). This suggests that disruption of bioenergetics during early development may be a prominent mechanism underlying latent impacts on fish growth and survival at later life stages.

Oil spill science in marine systems has thus far focused on fish species with distinct ecophysiological characteristics (Incardona and Scholz, 2016). This includes nearshore and pelagic species spawning in cold northern waters (Carls et al., 1999, Incardona et al., 2015) and rapidly developing sub-tropical species (Incardona and Scholz, 2018). In general, cold water species or those with strong cold tolerance are more sensitive to oil-induced toxicity (Edmunds et al., 2015, Incardona et al., 2014, Incardona et al., 2015, Morris et al., 2018, Sørensen et al., 2017, Sørhus et al., 2016). Although common morphological and functional abnormalities are usually evident shortly after embryonic exposure, delayed reductions in growth and juvenile survival have been documented in pink salmon exposed to low concentrations of oil that did not cause externally visible malformation (Heintz, 2007, Heintz et al., 2000). These effects on growth could reflect a latent and lasting dysregulation of lipid metabolism. If so, this would have important consequences for global marine fisheries because management paradigms are premised on a positive relationship between juvenile bioenergetics and successful recruitment to adult populations (Bouchard et al., 2017, Copeman et al., 2017, Heintz et al., 2013).

Polar cod (Boreogadus saida) is a circumpolar species and one of the most abundant and important forage fish in Arctic ecosystems (Mueter et al., 2016). Significant changes in the distribution and abundance of cod would likely be highly disruptive to Arctic food webs, especially ice-obligate species (Bluhm and Gradinger, 2008, Choy et al., 2017, Vihtakari et al., 2018). Polar cod embryos exposed to a water-soluble fraction of crude oil from oiled gravel columns showed reduced hatching success and increased frequencies of malformations at very low total (∑)PAH concentrations (≤1 μg/L) (Nahrgang et al., 2016). In the present study we exposed Polar cod embryos using an environmentally realistic oil dispersion and investigated the latent consequences for two critical survival periods. These included “first-feeding success,” wherein newly hatched larvae consume their yolk reserves and begin feeding (Houde, 2008, Laurel et al., 2011), and “overwintering success,” a process that is critically dependent on the size-at-age and energy density (lipid) of juvenile fish entering their first winter (Copeman et al., 2008, Siddon et al., 2013, Sogard and Olla, 2000). To emulate environmentally realistic post-spill oil exposures, relatively low levels of oil droplets were generated using a previously described dispersion system (Nordtug et al., 2011), with nominal oil loadings (by mass) of 100, 300, and 900 μg/L. Dispersed oil was released into exposure tanks for only 3 days of the ~40-day embryonic development period, during mid-embryogenesis after the heart first began beating (20 days post-fertilization, dpf; Figure S1). Embryos were then transferred to new vessels and incubated in clean seawater to hatch (from 27–50 dpf) before a final transfer to flow-through tanks supplied with live prey for grow-out (to 193 dpf). We describe the toxicity of crude oil to the developing heart and craniofacial structures of cod embryos. We also characterize how trace and transient embryonic oil exposures affect growth rate and lipid content at later life stages.

Results

Buoyancy of Polar Cod Embryos Enhances Exposure to Dispersed Oil

Oil droplets rose to the surface to form slicks within the first 24 h of oil release. During a subsequent 4-day washout with clean seawater, buoyant embryos remained in the exposure tanks, in contact with the surface oil as it dissipated. To characterize exposure chemically, PAHs were measured in exposure water collected directly from the tank inlets at the exposure start (day 0) and day 1 and from the water column at exposure days 1, 3, and 7 (Table S1; Data S1). To determine any potential increase in exposure from contact with the surface slick, embryos were also placed in 500-μm mesh chambers (which allowed passage of dispersed oil droplets) in the water column below the slick. Tissue concentrations of PAHs and levels of PAH-induced cyp1a mRNA were measured in both submerged and floating, surface-oriented embryos (overall experimental design shown in Figure S1).

Over the 3 days of oil release, ∑PAH42 concentrations in the water column (Table 1) ranged from 0.65 ± 0.20 to 1.1 ± 0.1 μg/L for the low concentration (100 μg/L oil load), 2.6 ± 0.3 to 4.0 ± 0.1 μg/L for the medium concentration (300 μg/L oil load), and 11.9 ± 2.3 to 18.3 ± 2.7 μg/L for the high concentration (900 μg/L oil load). Control seawater contained very low background levels of PAHs (0.11 μg/L; Table 1). Visual inspection of embryos floating at the surface showed oil droplets adhering to the chorions (eggshell) within the first 24 h of oil release, with sufficient binding to remain attached through collection and transfer into methylcellulose for microscopic imaging (Figure S2). By day 3 of oil release, droplets were observed on chorions at even the low exposure concentration, with continued accumulation at the higher concentrations (Figure 1). Consequently, measurements of embryo-associated PAHs were conducted separately for floating embryos to compare with artificially submerged embryos to elucidate effects of direct contact with surface slicks (Table 1; Data S1). By exposure day 3, the ∑PAH in submerged embryos ranged from 67 ± 7 to 832 ± 85 ng/g wet weight and in floating embryos from 517 ± 94 to 4195 ± 650 ng/g. This represented an increase of 8, 9, and 5 times at the low, medium, and high exposure concentrations, respectively. We also measured tissue PAHs after oil release was stopped and embryos remained in contact with the dissipating surface slicks (Table 1; exposure day 7). Although embryo-associated ∑PAH had declined for the low and medium exposures (284 ± 100 and 4259 ± 937 pg/embryo, respectively), ∑PAH continued to rise for the high dose (11,343 ± 1898 pg/embryo). After 4 days incubation in clean seawater, tissue ∑PAH levels dropped by about 75% for all doses, consistent with metabolic depuration (Table 1).

Table 1

Quantification of Exposure by Measures of Water and Embryo ∑PAH42 and cyp1a mRNA Induction

Nominal Oil Loading	Water Column Total ∑PAHa (μg/L)	Egg Incubation Position	Day 3 Tissue ∑PAH (ng/g Wet Weight)	Day 3 Tissue ∑PAH (pg/Embryo)	cyp1a mRNA Induction (Fold Change)	Interpolated Day 3 Tissue ∑PAH (pg/Embryo)b	Percent Day 3 ∑PAH Bioavailable	Day 7 Tissue ∑PAH (pg/Embryo)	4 Days Post-exposure Tissue ∑PAH (pg/Embryo)	Percent Depurated 4 Days Post Exposure
0	0.11 ± 0.03	Floating	43 ± 21	102 ± 50	0.78 ± 0.15			13 ± 7	28 ± 5
		Submerged	17 ± 2	40 ± 5	1.3 ± 0.2
100 μg/L	1.1 ± 0.1	Floating	517 ± 94	1,232 ± 223	62 ± 11	1,378	100	284 ± 100	81 ± 18	72
		Submerged	67 ± 7	161 ± 17	11 ± 3
300 μg/L	4.0 ± 0.1	Floating	2,358 ± 390	5,448 ± 901	69 ± 10	1,529	28	4,259 ± 937	973 ± 214	77
		Submerged	272 ± 89	629 ± 206	30 ± 12
900 μg/L	18.3 ± 2.7	Floating	4,195 ± 650	9,858 ± 1,527	115 ± 9	2,565	26	11,343 ± 1,898	2,864 ± 362	75
		Submerged	832 ± 85	1,956 ± 210	88 ± 14

Tissue concentrations are means (±SEM) of samples from four replicate tanks.

Measured at exposure day 3.

Calculated from the linear equation relating measured ∑PAH to cyp1a mRNA induction in submerged embryos (Figure S3).

Figure 1

Oil Droplets Bind to the Polar Cod Chorion

(A–C) Higher-magnification views compare embryos from the control surface (A) and low (B) and high (C) exposures on the day of transfer to clean water.

(D) Quantification of adhered droplets (mean ± S.E.M.) from representative images of single embryos (4 per replicate, 16 total per treatment) at exposure day 3 (end of oil release into the water column).

Arrowheads indicate oil microdroplets. Scale bar is 200 μm.

Levels of cyp1a mRNA reflect a response to cell-internal PAHs within tissues. Therefore, quantitative measures of cyp1a induction can distinguish the internalized bioavailable fraction of PAHs from external PAHs retained in oil droplets adhering to the chorion. Using fold-change relative to controls, cyp1a induction was linear in submerged embryos but non-linear and approaching saturation in floating embryos (Figure S3). Estimated true internal PAH concentrations were calculated for floating embryos by interpolation using the linear equation for submerged embryo cyp1a induction (Table 1). This demonstrated that surface slick exposure and oil droplet binding to the chorion led to increased PAH uptake in floating embryos. Based on the levels of cyp1a induction, contact with surface oil increased the internal concentration of ∑PAH 8.6 times for the low oil load and 2.4 times for the medium load, and 1.3 times for the high oil load (Table 1). At the low oil load, this means that 100% of the measured embryo-associated PAH was bioavailable, whereas roughly 25%–30% was bioavailable at the higher loads. These findings demonstrate that even small oil spills can have outsized impact for highly buoyant Polar cod eggs through the formation of surface slicks. For simplicity and consistency in referring to oil concentrations, hereafter, toxicity endpoints in floating embryos (below) are simply related to “low,” “medium,” and “high” exposures based on cyp1a-based (“true”) internal concentrations.

Transient Exposure during Mid-organogenesis Is Sufficient to Produce Latent Cardiac and Craniofacial Defects

At the end of exposure (27 dpf), embryos displayed dose-dependent slowing of the heart rate (bradycardia; Figure 2A, bottom line) and an increase in frequency of abnormal heart rhythms (Figure 2B). Measuring atrial contractions in digital videos showed that the duration between heartbeats increased in a dose-dependent manner from 3.0 ± 0.1 to 4.0 ± 0.2 s/beat (Figure 2B, top line) and that at a relatively sharp dose threshold, interbeat duration became irregular (atrial fibrillation) (Figure 2B, bottom line; Figures 2C–2F). Many individuals at the high dose showed, in addition to irregular rate, retrograde conduction from the atrioventricular junction (Video S1). Cardiac function was assessed again in hatching stage yolk-sac larvae at 42 dpf. Despite likely depuration of PAHs over the intermittent 15 days in clean water, larvae displayed a persistent dose-dependent bradycardia (Figure 2A, top line) but no rhythm irregularities. Consistent with a reduction in cardiac output, larvae accumulated pericardial and yolk sac edema. Edema was dose-dependent in severity, quantified by scoring contact between the cardiac chambers and pericardial membrane in lateral view digital videos (Figure 3A). Because the tissue PAH depuration data indicate a half-life of ≲ 3 days, PAHs were most likely nearly completely eliminated by this point. Therefore, the persistent bradycardia likely reflects electrophysiological remodeling of cardiomyocytes in response to oil exposure, rather than acute impacts on ion conductances.

Figure 2

Cardiac Dysfunction Immediately Following Embryonic Oil Exposure

(A) Dose-dependent reduction of heart rate measured at the end of exposure (lower line) and at hatch after 15 days in clean seawater (upper line).

(B) Dose-dependent increases in heartbeat duration (upper line) and irregularity (lower line) measured at the end of exposure.

(C–F) Graphical representations of heart rhythms derived from digital videos, used to generate data shown in (B). Wave forms represent atrial motion; baselines drift because of embryo movement during video acquisition.

Values are means (±SEM) from end of exposure (A, bottom line, and B) and are based on video analyses of 10 representative embryos per replicate tank (n = 4 replicates per treatment) and 30 larvae per tank at hatch.

Figure 3

Quantification of Pericardial Edema and Craniofacial Malformation in Post-hatch (Yolk Sac) Larvae

(A) Representative images at top show criteria for scoring presence of pericardial edema by contact of the ventricular (V) or atrial (A) chambers with the pericardial membrane (PCM, arrowheads). Edema was scored as present if one chamber did not contact the pericardial membrane (gray and dark gray; green, no edema). Scale bar is 200 μm. Plot values are means (±SEM) for the three categories based on multiple larvae (n = 30) from each replicate (n = 4) for each treatment.

(B) Representative images at top show categorization of craniofacial defects as mild (yellow), moderate (orange), or severe (red) based on loss of anterior jaw structures. Values are means (±SEM) based on imaging for 30 representative individuals per replicate tank (n = 4 replicates per treatment). Scale bar is 400 μm.

Asterisks in both indicate p < 0.01 for post hoc means comparison with control by Dunnett's Test following ANOVA.

Video S1. Embryonic Heart Beats of Polar Cod Boreogadus saida, Related to Figure 2

Ectopic pacemaker and retrograde contraction from the atrioventricular region in Polar cod embryos immediately after oil exposure. A pair of 30-s video clips show the beating looping-stage heart for representative embryos from control and high-dose (900 μg/L oil) exposures. Embryos are viewed ventrally through the chorion with the atrium (A) oriented to the video's right and ventricle (V) near the midline. AVC, atrioventricular canal.

The effects of oil exposure on overall gross morphology of yolk sac larvae were remarkably similar to those observed in other closely related gadids, Atlantic haddock (Melanogrammus aeglefinus) and Atlantic cod (Gadus morhua) (Figure S4). Polar cod larvae showed craniofacial defects, characterized by a loss of jaw structures progressing from the upper jaw/basal neurocranium to the lower jaw with increasing dose (Figure 3B). Although not quantified, an apparent reduction in melanophore pigment cells at the high dose was noted (Figure S4D). There was reduced hatch success; in addition, oil-exposed embryos were significantly smaller at hatch (standard length) and had proportionately smaller eyes (Figures 4A–4C). Other structures and metrics (e.g., myotome depth) were not affected (Figure 4D).

Figure 4

Acute Extracardiac Impacts of Oil Exposure on Polar Cod Embryos

Increasing oil exposure decreased hatching success (A) but had no significant effect on cumulative hatch rate (B). The size of cod larvae at hatch (C) and eye diameter (D) decreased with oil exposure, whereas body depth (myotome) remained constant (D). Values are means (±SEM) of total counts per replicate tank (A and B) or 20 individuals/tank (C and D). Letters indicate statistically different groups determined by ANOVA with Tukey-Kramer HSD test for post hoc means comparisons.

Embryonic Oil Exposure Leads to Persistent Derangements in Larval-Juvenile Lipid Metabolism, Even in Morphologically Normal Survivors

Surviving larvae exposed to oil also experienced higher rates of abnormalities that would likely impair swimming and foraging ability, including the fluid imbalance represented by pericardial and yolk sac edema, axial deformation, and jaw deformities (e.g., Figure S4). We therefore monitored survival and growth after yolk sac larvae were placed on a previously established dietary regimen (Koenker et al., 2018). To profile changes in bioenergetics, lipid classes were measured throughout the experiment, beginning with embryonic exposure. Lipids were subsequently measured in newly hatched yolk sac larvae, through larval growth, and in juveniles. In addition, further latent effects on cardiac morphology and function were assessed.

By 86 dpf (43 days post hatch), there were no fish surviving from the high-exposure treatment. Fish from the low- and medium-exposure groups still displayed cardiac defects (Figure S5) despite surviving at relatively high numbers (albeit significantly lower than control groups, p < 0.001). Although there was a slight decreasing trend in the size of the ventricle, this was not significant when normalized to standard length among controls (Figure S5D). However, there was a significant dose-dependent reduction in the ventricular aspect ratio, indicating a relationship between oil exposure and rounder ventricles. In addition, oil-exposed larvae had reduced trabeculation (Figure S5C). Although neither normalized ventricular size nor heart rate were different at this point, oil-exposed larvae showed persistent dose-dependent edema, likely evidence of reduced cardiac output (Figure S6). Although 100% of the medium-exposure group showed edema at this stage, 65% of larvae exposed to the low dose were grossly normal.

Relationships between morphology, physiological function (e.g., fluid balance), metabolism, growth, and survival were revealed by repeated sampling over the next ~150 days. Lipid class profiling demonstrated prolonged impacts on lipid metabolism that changed in character from the initial rapid larval growth stage to post-metamorphic juveniles. Although lipid classes in embryos were not different across treatments at the end of the transient oil exposure (Figure 5A), at hatch oil-exposed yolk sac larvae had elevated levels of three lipid classes on a per-dry weight basis; triacylglycerols (TAGs), free fatty acids (FFAs), and sterols (STs). Throughout feeding stages, the latent impacts of embryonic oil exposure were observable in larvae and juveniles by way of poor survival, altered lipid content (shown for TAGs and normalized to dry weight), and decreased growth (Figures 4B and 4C). By 66 dpf, relative elevations of TAGs in oil-exposed larvae became exacerbated, with the high-dose larvae showing little change in TAG content over the preceding 20 days, remaining 3.7 times higher than controls (Figure 5C). By 87 dpf, as animals completely shifted to exogenous feeding, no high-dose fish survived and TAG levels were at their lowest but not significantly different among the control and remaining oil-exposed groups. By 123 dpf, all larvae in the medium-exposure group were dead. Presumably, fish in the high- and medium-dose groups succumbed to sequelae stemming from severe fluid imbalance represented by persistent edema. In contrast, large numbers of larvae from the control and low (100 μg/L) exposures completed flexion and survived into the juvenile stage. At this point (123 dpf), both control and oil-exposed (low-dose) fish were at pre-flexion stage and showed evidence of TAG accumulation (Figure 5C), but exposed fish showed significantly lower levels. Compared with controls, low-dose survivors had 46% less TAG at 123 dpf, 46% less TAG at 156 dpf (flexion stage), and 31% less TAG by termination of the experiment at 193 dpf, at which point fish had transformed into early juveniles. Importantly, animals that survived oil exposure were significantly smaller than unexposed controls at the 123 and 156 dpf time points, but were not different in dry weight by 193 dpf, despite having a lower TAG content (Figure 5B).

Figure 5

Latent (Delayed in Time) Impacts of Embryonic Oil Exposure on Polar Cod

(A) Newly hatched larvae exposed to oil had elevated Triacylglycerol (TAG), Free Fatty Acids (FFA), and Total Lipids (TL), but there was no significant difference in Sterols (ST) or Polar Lipids (PL) classes among treatments.

(B) The dry mass of surviving larvae transiently exposed to oil as embryos was significantly lower than controls between 66 and 156 days post fertilization (dpf).

(C) The lipid density (triacylglycerol, TAG) of oil-exposed embryos was initially higher during the larval period (hatch to flexion; 45 and 67 dpf) but was lower in oil-exposed fish during the juvenile period (123–193 dpf). Dry mass and lipid density values are means (±SEM) per fish based on replicate pooled or individual fish depending on the time of sampling (see Transparent Methods). Letters indicate statistically different groups determined by ANOVA with Tukey-Kramer HSD test for post hoc means comparisons.

Discussion

Vulnerability of Polar Cod to Low-Level Exposure from Oil Spills

These findings have two key sets of implications. First, they have broad implications for understanding the long-term impact of oil spills on fish populations. Second, they have discrete and important implications for the vulnerability of Polar cod to spills in the Arctic. Studies conducted 20 years ago in the wake of the Exxon Valdez oil spill demonstrated that embryonic exposure to oil at levels insufficient to produce malformed larvae nevertheless led to reduced juvenile growth and marine survival (Heintz, 2007, Heintz et al., 2000). For pink salmon, these effects explained population-level impacts observed on the scale of individual streams in oiled habitat (Peterson et al., 2003, Rice et al., 2001). Although the effects of early exposure on later growth and survival have been presumed to be connected to developmental impacts on the heart, the precise mechanisms have remained elusive. Lipids are essential for both growth and early survival in marine fish larvae (Tocher, 2003), even more critically so at the extremely cold temperatures occupied by Polar cod (Copeman et al., 2017). Thus, the profound changes in lipid composition and dynamics observed here in Polar cod represent a probable mechanism for reduced growth and survival.

Owing to several life history characteristics, we anticipate Polar cod will be particularly vulnerable to future oil spills. First, as noted earlier, crude oil is relatively more toxic to cold water fishes. Second, their highly buoyant embryos (Laurel et al., 2018) make exposure to oil in surface slicks much more likely. Third, the adherence and accumulation of oil micro-droplets on the outside of the chorion suggests that even short-term contact with a slick could lead to a localized and prolonged release of dissolved PAHs. Consistent with the above, acute effects of spilled oil were magnified in Polar cod embryos, as evidenced by a higher cyp1a induction in embryos floating near the surface slick relative to those caged below in the water column. This effect was highly magnified at the lowest levels of dispersed oil released into the water column. Therefore, the transport and fate of oil in Arctic environments potentiates the exposure risk for Polar cod. Oil weathers more slowly in colder water and can be encapsulated in sea ice as it forms, suspending weathering and degradation until the oil is released during breakup. Oil is likely to accumulate and persist along the margins, in openings, and under sea ice, all rearing habitats for Polar cod embryos and larvae. Finally, the general importance of lipids to multiple aspects of fish early history stages, compounded by the special role of lipids in Polar cod, only increases this species' sensitivity.

Implications of Deranged Lipid Metabolism for Early Growth and First-Year Survival

In the natural environment, high amounts of TAGs can improve first-feeding success by reducing starvation risk during the transition from endogenous yolk-sac-based energy to the exogenous feeding stage (Laurel et al., 2008). Although TAGs and FFAs were significantly higher in animals exposed to the medium and high oil doses, the availability of these increased energy stores for metabolic activity is uncertain. Earlier studies on Atlantic haddock indicate oil exposure may interfere with the ability of larvae to mobilize lipids from the yolk during the critical transition from endogenous to exogenous feeding. Specifically, genetic upregulation of genes controlling intrinsic cholesterol and lipid biosynthesis and transport has been postulated to reflect deprivation in larval tissues as a consequence of the heart and circulatory system failing to deliver lipoproteins from the yolk (embryos) and intestine (larvae) (Sørhus et al., 2017). Our lipid results are from whole larval lipid pools, and the elevation of TAG, FFA, and ST in exposed larvae likely reflect both an upregulation of lipid synthesis in larval tissues as well as under-utilization of yolk lipids.

Higher concentrations of lipid in larval tissues has also previously manifested in histological findings such as liver lesions that have been measured in cultured fish both as a result of excessive fat storage or from exposure to hydrocarbons (Agamy, 2012). Furthermore, increased presence of FFAs are more indicative of fish stress (Mazeaud et al., 1977), and FFAs are generally not found in high levels in fish tissues except as a result of sample degradation (Parrish, 1988). The pairing of these acute findings of lipid metabolic disorder with other measures of larval physiology, morphology, and survival provide strong evidence that lipid synthesis genes are promising indicators, or biomarkers, for injury in fish embryos exposed to crude oils (Sørhus et al., 2017).

After flexion, surviving oil-exposed fish with poor cardiac function may have begun the juvenile phase with an energetic deficit, necessitating increased TAG utilization for compensatory growth and maintenance while leaving proportionally less TAG for storage in advance of the first winter. Basal metabolic processes in larvae are stage-specific (Peck and Moyano, 2016), and shifts in TAG levels can indicate a reduced use of TAG stored in yolk, the de novo synthesis of TAG in embryonic tissues deprived of yolk from the circulation, or both.

The observed changes in growth and lipid content can be expected to directly impact the health and survival of individual cod and by extension, recruitment into forage fish populations that support the upper trophic levels of Polar food webs (Hop and Gjosaeter, 2013). Biological productivity at high latitudes supports large stocks of fish and mammals and is driven by a highly specialized lipid-based energy transfer system (Falk-Petersen et al., 1990, Falk-Petersen et al., 2009). Gadids in the higher Arctic are also reliant on sufficient lipid stores to survive the relatively longer period of overwintering (Copeman et al., 2017). For example, walleye pollock (G. chalcogrammus) in the Eastern Bering Sea represent one of the largest single species fisheries in the world. Survival through their first winter is strongly correlated with their energetic content at the beginning of fall and their eventual recruitment to the fishery 2–3 years later (Heintz et al., 2013). The ability to successfully forage and allocate energy to satisfy immediate demands for growth, as well as storage needs to offset future winter starvation stress, is a particularly challenging trade-off for juveniles facing size-dependent predation in their first year (Sogard, 1997). Many of the fish in the low-exposure treatment, with a ~25%–30% reduction in size and TAG density, would unlikely have sufficient reserves to successfully overwinter. Latent effects of oil exposure on individual survival and productivity and resulting impacts on fish populations and ecosystems should be incorporated into prospective risk assessments and retrospective injury assessments. However, projections of Polar cod losses following large oil spills will also need to consider other population drivers, including thermal habitat conditions, prey availability and quality, and top-down predation pressures (Brodersen et al., 2011).

Potential Interactions of Oil Exposure with Other Environmental Stressors in the Arctic

Importantly, the adverse health effects we report here are likely to be exacerbated by parallel environmental stressors in Arctic habitats. Among the latter, most notable are increasing seawater temperatures and co-exposure to UV radiation in natural sunlight. Polar cod embryos and larvae are highly sensitive to temperature (Koenker et al., 2018, Laurel et al., 2016), and thermal stress may potentiate oil toxicity in this species. Atmospheric warming is occurring at a more rapid rate in the Arctic than in temperate zones, and observational data indicate that ocean surface temperatures are increasing in every polar region (e.g., 0.5°C per decade in the Chukchi Sea [Timmermans and Proschutinsky, 2014]). These changes in atmospheric warming are driving the loss of Arctic sea ice (Jeffries et al., 2018). As a result, the area is much more accessible to shipping and oil exploration/extraction, which then increases the risk of oil spills (Foster et al., 2015). Lastly, the photoactivation of chemicals in crude oil and the bunker oils that currently fuel large vessels is well known to increase lethal toxicity to fish early life stages (Barron et al., 2003, Incardona et al., 2012). This will be a particularly important hazard for translucent and buoyant organisms that are within the zone of UV penetration in the upper water column, i.e., Polar cod.

Finally, as it relates to the potential oiling of cod spawning habitats, the response of fish in our low treatment (~1 μg/L water column ∑PAHs) indicates that individual PAHs are toxic in the parts-per-trillion concentration range. Although this very low exposure did not cause significant mortality through larval development, it was sufficient to disrupt embryonic cardiac function and produce a lasting dysregulation of lipid homeostasis well into the juvenile life stage. This latent form of injury to cod suggests that the toxic footprint of any future Arctic spills will be much larger in space and time than previously anticipated.

Limitations of the Study

The current study addresses vulnerability at one particular life stage, which has proven to be particularly sensitive in northern temperate species. The adherence of droplets to the egg chorion along with the long embryonic development times of Polar cod will likely result in exposure risks beyond what was measured. Population- and ecosystem-level impacts from laboratory studies are also difficult to predict, and the current study is unable to quantify injury level beyond the individual. Multi-stressors such as temperature and UV exposure may amplify the RNA-seq and phenotypic effects observed in the study. Mortality rates in larval fish are also largely driven by prey availability and predation, which would likely exacerbate mortality rates observed in the oil-exposed embryos due to unmeasured physiology, e.g., impairments to visual acuity and swim performance. Finally, although there was significant effort to simulate in situ oil exposure by way of a micro-droplet generator, this study does not account for complex ice dynamics and turbulence that may result in different exposure scenarios.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.