Paralytic shellfish toxins and seabirds: evaluating sublethal effects, behavioral responses, and ecological implications of saxitoxin ingestion by common murres (Uria aalge)

Abstract

Paralytic shellfish toxins (PSTs), including saxitoxin (STX) and its congeners, are neurotoxins that can be produced during harmful algal blooms and cause illness or death in humans, fish, seabirds, and marine mammals. Since 2014, multiple large-scale seabird mortality events have occurred in Alaska waters, with STXs detected in some carcasses. To investigate the sublethal behavioral and ecological effects of STX on seabirds, we conducted captive dosing trials with common murres (Uria aalge). We gavaged purified STX (dehydrated STX dihydrocholoride, STX-diHCl) or an Alexandrium catenella culture extract into murres, monitored behavioral responses and recovery times, and assessed tissue concentrations in individuals that died or were euthanized. Using a modified up-and-down dose-finding scheme, we estimated a median effective dose (ED₅₀) of 89 μg STX-equivalents (eq)·kg⁻¹ for STX-diHCl and 366 μg STX-eq·kg⁻¹ for the A. catenella extract based on ecologically relevant behavior. Differences between the ED₅₀ estimates could reflect uncertainties in toxin equivalency factors for PST congeners, which are based on studies using purified toxins in mice and may vary across taxa or toxin matrices. Post-dosing concentrations of STX varied by tissue type across individuals, with quantifiable levels ranging from 3–379 μg STX-eq·100g⁻¹. Evidence of biotransformation of STX in A. catenella extract-dosed birds was observed. We also measured the chronic effects of dosing with sublethal levels of STX-diHCl over seven-days, which resulted in lower fish intake among treatment birds compared to controls (−187 g·day⁻¹). This investigation improves our understanding of the ecological effects of PSTs on seabird health.

1. Introduction

Saxitoxins are a suite of naturally occurring paralytic neurotoxins produced during harmful algal bloom (HAB) events. Saxitoxin (STX) and its analogs, collectively referred to as paralytic shellfish toxins (PSTs), are known to cause illness or death in a broad range of species and have been implicated in mortality events of fish, marine birds, and mammals worldwide (Geraci et al., 1989; Lefebvre et al., 2004, 2005; Ben-Gigirey et al., 2021; Van Hemert et al., 2022). In marine environments, STXs are primarily produced by members of the dinoflagellate genera Alexandrium, Gymnodinium, and Pyrodinium during bloom events and are passed through the marine food web via vector organisms such as filter feeding invertebrates and planktivorous fish (Shumway, 1995; Wiese et al., 2010). Over the past few decades, HAB events have increased in frequency and intensity in many parts of the world, in some cases due to changing climatic conditions (Glibert et al., 2014; Wells et al., 2020; Hallegraeff et al., 2021). In the Alaska arctic, the primary STX producer is Alexandrium catenella. Although various environmental factors can influence the timing and severity of Alexandrium bloom events, surface water temperature has often been identified as a contributing factor (Moore et al., 2008; Vandersea et al., 2018; Tobin et al., 2019). Changes in HAB occurrence are particularly relevant to marine waters in the Gulf of Alaska and the Bering and Chukchi Seas in the North Pacific, where anomalously warm sea surface temperatures have been recorded in recent years (Walsh et al., 2018; Litzow et al., 2020; Jones et al., 2023). These increases in temperature, combined with further predictions of warming in northern regions have led to expectations that HABs, particularly those producing STXs, may pose an emerging threat to ecosystem health in Arctic marine environments (Walsh et al., 2018; Anderson et al., 2022).

Since 2014, multiple large-scale seabird mortality events have been documented in Alaska marine waters (Piatt et al., 2020; U.S. Geological Survey, 2020; Jones et al., 2023), the largest being an unprecedented die-off of common murres (Uria aalge) in the winter of 2015-2016 during which as few as 460,000 and as many as 4 million birds are estimated to have perished (Piatt et al., 2020, Renner et al., 2024). Many carcasses examined from this and other mortality events were severely emaciated, with starvation largely identified as the proximate cause of death (Jones et al., 2019; Piatt et al., 2020; Van Hemert et al., 2021). Research into the possible roles of HAB toxins in these die-offs has shown mixed results. Saxitoxin was routinely detected at concentrations ranging from detectable but not quantifiable to 63.3 μg·100g⁻¹ in opportunistically collected seabird carcasses (Smith et al., 2022), indicating widespread exposure, but there has been little direct evidence of STX as a major contributor to mortality. Saxitoxin was detected in 36% of common murre carcasses collected in the Gulf of Alaska in 2015–2016 (Van Hemert et al., 2020) and quantifiable levels of STXs were reported in 88% of northern fulmars (Fulmarus glacialis) from a die-off in the Bering and Chukchi Seas in 2017 (Van Hemert et al., 2021). Additionally, there have been several small, isolated mortality events where STX has been detected in seabird tissues and tentatively or definitively assigned as the cause of death (Nisbet, 1983; Shearn-Bochsler et al., 2014; Starr et al., 2017; Ben-Gigirey et al., 2021; Van Hemert et al., 2022). Determining the role that STXs played in these die-offs has proven difficult, however, with logistical constraints and the transient nature of blooms making sample collection challenging. Furthermore, a lack of information remains on the toxicity and pharmacokinetics of STXs in seabirds largely due to most information coming from opportunistically collected carcasses with unknown toxin exposure and timing variables making interpretation of results difficult.

Few controlled exposure studies have examined the effects of STXs on avian species, with most information resulting from field observations or derived from opportunistically collected carcasses. An exposure trial determined the median lethal dose (LD₅₀) for STX in birds using mallards (Anas platyrhynchos) as a model species with an LD₅₀ estimate of 167 μg·kg⁻¹ (Dusek et al., 2021). The investigators also provided information on tissue distribution and time course for STX and detection times after ingestion. These results provided crucial information for interpreting STX concentrations in opportunistically collected seabird carcasses and allowed researchers to draw initial conclusions about the potential effects of STX on certain species (Rattner et al., 2022; Van Hemert et al., 2022). Many questions remain, however, about the broader ecological effects that STXs may have on seabird populations, including potential sublethal effects of STX ingestion and adverse effects of longer-term, repeated exposure to low levels of toxin.

To address the dearth of information on STX toxicity in seabirds and to allow for better interpretation of toxin profiles detected in wild birds, we conducted a multi-stage dosing trial with common murres (hereafter, murres), a seabird species that is ubiquitous in waters around Alaska and has been subjected to multiple, unprecedented mortality events over the past decade, accompanied by reproductive failure and population decline (Gibble et al., 2018; Piatt et al., 2020; Schoen et al., 2022; Renner et al., 2024). The primary objective of this study was to determine the sublethal effects of STX ingestion by murres and to identify ecologically relevant thresholds of STX exposure by estimating the median effective dose (ED₅₀) of STX and related congeners. Additional objectives were to (1) examine the concentration and distribution of STXs in tissues collected from murres after ingestion of known doses of toxin, (2) evaluate the behavioral responses of murres to ingestion of sublethal levels of STX, (3) determine the patterns, if any, of biotransformation of STX congeners by seabirds after ingestion of a crude A. catenella extract, and (4) evaluate the effects of chronic, low-dose ingestion of STX on murres’ food intake and ability to forage. Results from this study will facilitate interpretation of STX concentrations in tissues of wild birds and provide information about the ecological consequences of PST exposure for seabirds in northern regions.

2. Methods

2.1. Study species:

To allow for ease of handling and to ensure no prior exposure to PSTs, murres used in our study were hatched from eggs we collected from two separate nesting colonies located in southcentral Alaska in July 2021 (Fig. 1). To ensure viability, eggs collected from nests were placed in modified cases that kept the temperature at 37.5 °C and humidity at 60–90% during transport. Eggs were transported to hatching facilities at the Alaska SeaLife Center in Seward, Alaska, within 12 h of collection where they were transferred to incubators until hatch. After hatch, murre chicks were reared and cared for following Alaska SeaLife Center protocols (AZA Charadriiformes Taxon Advisory Group, 2018) until the beginning of the study, approximately 180 days post-hatch. Egg collection was conducted under United States Fish and Wildlife Service (USFWS) Alaska Maritime National Wildlife Refuge research and monitoring activity special use permit # 74500-21-015. All portions of this study, including egg collection, rearing, dosing, handling, recovery, and euthanasia were reviewed and approved by the U.S. Geological Survey Institutional Animal Care and Use Committee (IACUC; 2020-03) and Alaska SeaLife Center IACUC protocol R21-12-03.

Fig 1.

Map of common murre (Uria aalge) colonies in Cook Inlet, Alaska, where eggs were collected (stars) in July 2021. Map was created using ArcGIS 10.3 (ESRI) with a base map from the Alaska Department of Natural Resources (2017) available at https://arcgis.dnr.alaska.gov/arcgis/rest/services/OpenData/Physical_AlaskaCoast/MapServer.

2.2. Toxin Sources

Purified STX was provided by the U.S. Food and Drug Administration (FDA, College Park, Maryland) in the form of dehydrated saxitoxin dihydrochloride (STX-diHCl). Methods of production, purification, and verification are described in Dusek et al. (2021). Prior to administration, STX was resuspended into solution by adding pH 3.7 HCl to yield a 1 mg·mL⁻¹ solution. The solution was vortexed immediately following resuspension and kept at 4 °C for the duration of the study.

To produce the crude extract of A. catenella cells used in our study, a clonal STX-producing strain of A. catenella, IC-8 B1 (hereafter A. catenella extract), was established by germinating cysts isolated from a sediment sample collected in September 2018 from the Chukchi Sea, Alaska (70.9725 N, 163.5642 W; Anderson et al., 2021). Seed and production cultures of this isolate were maintained at 15 °C on a 14:10h light:dark cycle (ca. 250 μmol photons·m⁻²·sec⁻¹ irradiance provided by cool white spectrum LED bulbs) using sterilized Vineyard Sound (Cape Cod, Massachusetts) seawater (salinity ~ 32 PPT) enriched with f/2 nutrients (Guillard and Ryther, 1962) as modified by Anderson et al., (1994) through the addition of Na₂SeO₃ and the reduction of CuSO₄·5H₂O, each to final concentrations of 10⁻⁸ M. No silicate was added. For large-scale toxin production, 17-L cultures were grown in multiple 20-L carboys over the course of several weeks. These cultures were bubbled with air provided by an aquarium pump to help minimize CO₂ depletion and pH change. When the cultures reached mid-exponential stage growth, they were harvested by pressurizing the carboy and forcing the contents through a sampling port into a 10-μm Nitex net (Wildlife Supply Company^®, Yulee, Florida). The net filtrate was then refiltered through 15-μm Nitex sieves to ensure more complete cell capture, and the resulting cell concentrate was rinsed into four, 250-mL conical centrifuge bottles that were spun at 3000 x g for 10 minutes at 4 °C. The resulting four-cell pellets were combined into tared 50-mL conical centrifuge tubes. Homogenized cell count samples were collected at this point, and the 50-mL tubes were centrifuged as above, the supernatant removed, and the cell pellets transferred to 15-mL conical tubes. The cell pellets were consolidated into one tube via centrifugation and depending on the size of the resulting cell pellet, between 3 and 5 mL of 1.7 mM hydrochloric acid was added. The acidified cell pellets were sonified using a Branson 250 Sonifier (Branson Ultrasonics, Danbury, Connecticut) equipped with a micro tip and set to an amplitude of 40% for 1 minute while the sample was in an ice water bath. To further assist with toxin extraction, all cell pellets were frozen and thawed 3x prior to combining all cell pellets into one 50-mL tube. A subsample was then removed for toxin analysis using a modification of the post-column derivatization method described in Oshima (1995) and Anderson et al. (1994). The combined extract was frozen at −20 °C prior to the start of the ED₅₀ portion of the study, after which it was thawed and kept refrigerated at 4 °C for the duration of the study.

2.3. Dosing scheme: ED₅₀

We used a modified up-and-down dosing scheme described by Bruce (1985) to estimate the ED₅₀ value for both STX-diHCl and our crude A. catenella extract. This method is designed to minimize the number of animals required to determine the median dose for a designated outcome, which in the case of our study, was the appearance of behavioral changes that could have fatal consequences in the wild (i.e., ecologically relevant). Doses of the A. catenella extract were calculated based on total STX toxicity equivalence factors (TEFs, expressed as μg·kg⁻¹ STX-equivalents [eq]), which express the toxicities of PST congeners as a ratio to STX (Munday et al., 2013; Selwood et al., 2017). All doses are described as μg per kg of bird weight or μg·kg⁻¹ (or μg STX-eq·kg⁻¹ for doses of the A. catenella extract) and each bird only received a single dose. The starting dose for STX-diHCl was chosen based on the results of Dusek et al. (2021) with several stepwise dose reductions to minimize the possibility of unintended lethal effects. The first individual received a dose of 25.2 μg·kg⁻¹ of STX-diHCl. Doses for subsequent birds moved up or down following a constant factor of 0.12 on a log dose scale (Bruce, 1985) given the appearance or absence of ecologically relevant behavioral changes based on a modified observational battery (Fig. S1) in the dosed bird immediately preceding it. Dosing was conducted daily between 0845 and 0930 Alaska Daylight Time (AKDT). To ensure complete absorption of the toxin, murres were fasted after their last feeding the previous evening at approximately 1730 AKDT. Birds were selected from a designated pen chosen at random. Each bird was weighed to the nearest gram immediately after capture from their outdoor enclosure and the subsequent STX dose was calculated using this mass. Birds were transported indoors and remained inside a plastic kennel prior to dosing. Toxin was delivered to each bird using a 12ga rubber gavage tube inserted to the level of the proventriculus. Prior to insertion, the gavage tube was primed with 4.5 mL pH 3.7 HCl to facilitate placement of the gavage tube (Dusek et al., 2021). Once in place, toxin was pipetted into the gavage tube and the tube flushed with 9 mL pH 3.7 HCl. Birds were immediately placed back in a plastic kennel for a period of 10 min post-dosing to ensure the absence of acute toxic effects. After this 10 min period, birds were returned to their outdoor enclosures for the remainder of the observation period.

2.4. Behavioral monitoring and observation

Behavioral observations of dosed murres during the ED₅₀ phase began immediately after dosing and continued for an 8-h period. Passive behavioral observations were recorded at 10 min intervals for the first 90 min after dosing, followed by 30 min intervals for the remainder of the 8-h observation period. Observations at each time interval included each murre’s location, behavior, and a determination of whether the behavior was abnormal. Abnormal behaviors were identified based on previous studies (Silvagni, 2003; Dusek et al., 2021), expert opinion of husbandry personnel, and real-time comparison to control birds (n = 4) (Fig. S1). To determine the possible effects of STX ingestion on a murre’s ability to catch and consume prey, we conducted two foraging trials at specified time intervals after each dosing event. Approximately 90 min post-dosing, juvenile coho salmon (Oncorhynchus kisutch) were released into the pen containing the dosed murre and four control birds. Responses were recorded immediately following release of fish into each pen and continued for a duration of 15 min. In addition to the passive observations and foraging trial response, each dosing event included separate, twice daily observation periods using scan sampling to establish activity baseline (AB) behavior for murres. Scan samples followed a structured 30-min observation period, with observers recording behaviors of treatment and control birds at two-min intervals and following the same scheme as the passive behavior observations. The first AB observation period was conducted 30 min following the return of each dosed bird to the outdoor enclosures. The second AB was completed just prior to the second foraging trial event. Finally, recovery time from effects of toxin ingestion, tracked as the time elapsed between when abnormal behavior was identified in treatment birds post-dosing and when behavior became indistinguishable from control birds, was compared across all doses for both toxin types.

2.5. Chronic dosing trial

The chronic dosing portion of our study was conducted following the conclusion of the ED₅₀ trial and a 14-day grace period to allow for clearance of any residual STXs from dosed birds. This portion of our experiment took place over the course of seven days and was designed to more closely mimic how seabirds would be exposed to STX in the wild by examining cumulative effects of longer-term ingestion of STX using dosed fish. We calculated doses based on the outcome of our ED₅₀ testing using the mean weekly mass of the lowest 25^th percentile of murres to avoid unintended lethal effects on smaller individuals. Twice-daily “scale sessions” were used to offer dosed or non-toxic fish to murres as reward for stepping on a digital scale. This allowed for continuous monitoring of body mass and for husbandry personnel to ensure controlled ingestion of STX-dosed fish. Purified STX-diHCl was injected into the coelom of capelin (Mallotus villosus) or silversides (Menidia menidia) immediately prior to each scale session and the dosed fish offered to each treatment bird during AM and PM sessions. Observations were recorded in real-time during each scale session followed by a 30-min observation period to track response to the toxin. Activity baseline observations and foraging trials followed the same timing and methods as described for the ED₅₀ trial. Body weights of each murre were recorded twice daily at each scale session and total fish intake, measured as the difference between the mass of fish provided versus consumed, was tracked for each pen. All observations, body mass measurements, and total fish intake were measured for seven days before and after the chronic trial.

2.6. Tissue analysis and saxitoxin testing

Tissues, including serum, muscle, liver, upper gastrointestinal tract, small intestine, large intestine, heart, lung, brain, kidney, and spleen, were collected from murres that suffered acute toxicity or were euthanized at the conclusion of the study. Regurgitant and fecal samples were collected opportunistically throughout all stages of our study, and whole blood was drawn from all euthanized murres and opportunistically from birds that suffered acute toxicity post dosing. Murre carcasses were frozen immediately after death and kept at −20 °C until tissues were collected. Serum from whole blood samples was separated on the same day of collection and kept at −20 °C until subsequent analysis. After the study concluded, murre carcasses were thawed and tissues collected on the same day to prevent multiple freeze/thaw cycles. All tissues were extracted and screened for the presence of STXs following methods described in Lawrence et al. (2005) and later modified as described in Van Hemert et al. (2020). Up to 5 g of tissue was extracted and tested using commercially available enzyme-linked immunosorbent assay (ELISA) kits (Gold Standards Diagnostics; Warminster, Pennsylvania) with slight modifications. The manufacturer’s reported limit of detection for the ELISA is 0.015 ng STX-eq·mL⁻¹. Using this assay, Van Hemert et al. (2020) consistently detected and quantified STX-eq down to 1–2 μg·100g⁻¹ tissue. Samples from murres with STX concentrations ≥10 μg·100g⁻¹ by ELISA were analyzed via high-performance liquid chromatography (HPLC) to determine STX congener profiles. After extraction with acetic acid, Supelco® Supeclean^™ LC-18 (Sigma-Aldrich; St. Louis, Missouri) were used for the C18 cleanup procedure and, if necessary, Phenomenex Strata X-CW Polymeric weak cation exchange columns (Phenomenex; Torrance, California) were used for N-1-hydroxylated toxins. HPLC analysis used the precolumn oxidation procedure described by Lawrence et al. (2005) and Van Hemert et al. (2022).

2.7. Statistical analysis

We used centered isotonic regression to analyze all dose-response outcomes from the ED₅₀ portion of our study (Oron and Flournoy, 2017). Total fish consumption for the chronic study was analyzed using repeated-measures analysis of variance (ANOVA), structured as mixed-effects models using total fish consumed as a response. We included status (Treated vs Control), period (Before, During, and After STX exposure), and interaction between time period and status as fixed effects, and random effects to account for variation among daily sampling events within each period (Day) and variation in the physical characteristics of individual murres that comprised each pen (Pen). All isotonic regression models, ED₅₀ estimates, and accompanying 95% confidence intervals (CI) were produced using the cir package (Oron, 2024) in the R programming language (R Core Team, 2023). We used the lmerTest package (Kuzetsnova et al., 2017) in Program R (R Core Team, 2023) to fit mixed-effects models and compare average fish consumption among periods and status groups using least-square means. Effects were determined to be statistically significant at α = 0.05. All data produced during or referenced in this study are available in Smith (2024).

3. Results

3.1. ED₅₀ estimation

A total of 26 murres received doses of either STX-diHCl (n = 13) or the A. catenella extract (n = 13) during the ED₅₀ portion of our study (Table 1), 23 of which recovered from dosing, with three individuals suffering acute death (Table 1). We estimated two separate ED₅₀ values for STX-diHCl: the first was based on the appearance of ecologically relevant behavioral changes (ED_50E) following our modified observational battery (Figure S1), and the second was based on the appearance of any clinical response to the toxin (ED_50C). We estimated an ED_50E of 89 μg STX·kg⁻¹ (95% CI 72–103 μg STX·kg⁻¹) and an ED_50C of 45 μg STX·kg⁻¹ (95% CI 21–69 μg STX·kg⁻¹). When attempting to calculate the ED₅₀ value based on ecologically relevant behavior for the crude A. catenella extract, birds at the maximum toxin concentration (415.99 μg STX-eq·kg⁻¹; n = 2) suffered acute toxicity, which resulted in the calculation of an LD₅₀ value for the crude extract of 366 μg STX-eq·kg⁻¹ (95% CI 309–399 μg STX-eq·kg⁻¹) based on measured toxicity of the native toxin profile and assuming no hydrolyzation of the less toxic saxitoxins (e.g. C1C2 and Gonyautoxin (GTX)5; Leal and Cristiano 2022). Due to limitations on the amount of A. catenella extract available, we were only able to estimate an LD₅₀ for this suite of toxins and not a secondary value based on any clinical response.

Table 1.

Results of a modified up-and-down dosing trial where common murres (Uria aalge) were dosed with saxitoxin dihydrochloride (STX-diHCl) or a crude Alexandrium catenella extract (IC-8 B1). Empty behavioral response columns indicate a bird that was dosed for euthanasia and no observations were recorded. Time of death (TOD) measured as time (hr:min) elapsed since toxin ingestion. Dose is expressed as μg·kg⁻¹ for STX-diHCl and μg STX-equivalents·kg⁻¹ for the A. catenella extract.

Bird ID	Toxin Type	Dose	Abnormal Behavior	Ecol-Relevant Behavior	Outcome (TOD)
UA45	STX-diHCl	25.2	No	No	Recovered
UA49	STX-diHCl	33.3	No	No	Recovered
UA54	STX-diHCl	43.9	No	No	Recovered
UA59	STX-diHCl	57.8	Yes	No	Recovered
UA68	STX-diHCl	76.2	Yes	No	Recovered
UA43	STX-diHCl	100.5	Yes	Yes	Recovered
UA51	STX-diHCl	76.2	Yes	No	Recovered
UA55	STX-diHCl	100.5	Yes	Yes	Acute toxicity (0:18)
UA63	STX-diHCl	76.2	Yes	No	Recovered
UA52	STX-diHCl	57.8	Yes	No	Recovered
UA57	STX-diHCl	43.9	Yes	No	Recovered
UA50	STX-diHCl	33.3	Yes	No	Recovered
UA45	STX-diHCl	25.2	No	No	Recovered
UA67	A. catenella Extract	34.6	Yes	No	Recovered
UA41	A. catenella Extract	45.6	Yes	No	Recovered
UA47	A. catenella Extract	60.1	Yes	No	Recovered
UA46	A. catenella Extract	79.3	Yes	No	Recovered
UA58	A. catenella Extract	104.5	Yes	No	Recovered
UA66	A. catenella Extract	137.8	Yes	No	Recovered
UA65	A. catenella Extract	181.6	Yes	No	Recovered
UA40	A. catenella Extract	239.4	Yes	No	Recovered
UA42	A. catenella Extract	315.6	Yes	No	Recovered
UA48	A. catenella Extract	416.0	Yes	Yes	Acute toxicity (0:15)
UA56	A. catenella Extract	315.6	Yes	No	Recovered
UA64	A. catenella Extract	416.0	Yes	Yes	Acute toxicity (0:11)
UA67	A. catenella Extract	315.6	Yes	No	Recovered
UA49	STX-diHCl	89.0	-	-	Euthanized (1:07)
UA54	STX-diHCl	89.0	-	-	Euthanized (1:24)
UA46	A. catenella Extract	104.5	-	-	Euthanized (0:24)
UA66	A. catenella Extract	104.5	-	-	Euthanized (0:34)

3.2. Tissue results

We tested a total of 10 tissue types from each murre that died acutely (n = 3) or was euthanized (n = 4) at the end of the study. We also screened serum (n = 5), feces (n = 4), and regurgitant (n = 2) for the presence of STX and its analogs using ELISA and HPLC methods. Concentrations of STX derived from ELISAs varied by toxin and tissue type across individuals. We did not detect STX in serum, brain, or muscle tissues from any individuals (Table 2). Digestive tissues from all individuals, consisting of upper gastrointestinal tissue (UGI), small intestine, and large intestine, contained quantifiable levels of STX ranging from 2.6–52.2 μg·100g⁻¹ (Table 2). We found the highest STX concentrations in fecal and regurgitant samples (up to 379.0 μg·100g⁻¹; Table 2) from euthanized birds. Remaining tissues (liver, heart, lung, kidney, and spleen) contained varying but comparatively lower concentrations of STX depending on dose (Table 2).

Table 2.

Concentrations of saxitoxin (μg·100 g⁻¹) resulting from enzyme-linked immunosorbent assay (ELISA) analysis of tissues from common murres (Uria aalge) that suffered acute toxicity resulting in death or were euthanized at the conclusion of the median effective dose (ED₅₀) dosing study. Individuals received either purified saxitoxin dihydrochloride (STX-diHCl) or an equivalent dose of a crude Alexandrium catenella extract (IC-8 B1). Abbreviations are as follows: BDL = below detection limits, DBNQ = detectable but not quantifiable, UGI = upper gastrointestinal tract, Sm Intestine = small intestine, Lg intestine = large intestine, NT = not tested. Dose is expressed as μg STX·kg⁻¹ for STX-diHCl and μg STX-equivalents (eq)·kg⁻¹ for the A. catenella extract. Tissue values are expressed as μg STX-eq·100 g⁻¹.

Bird ID	UA49	UA54	UA55	UA46	UA66	UA48	UA64
Toxin type	STX-diHCl	STX-diHCl	STX-diHCl	A. catenella Extract	A. catenella Extract	A. catenella Extract	A. catenella Extract
Dose	89	89	100.5	104.5	104.5	415.9	415.9
Tissue
Serum	BDL	BDL	BDL	BDL	BDL	NT	NT
Muscle	BDL	BDL	BDL	BDL	BDL	BDL	BDL
Liver	BDL	BDL	DBNQ	BDL	BDL	1.8	BDL
UGI	3.6	3.4	26.4	2.6	7.2	29.4	23.3
Sm Intestine	22.0	4.2	20.9	13.7	4.8	37.0	52.2
Lg Intestine	18.5	42.5	23.7	49.0	13.1	43.3	31.7
Heart	BDL	BDL	2.4	BDL	BDL	1.8	BDL
Lung	BDL	BDL	3.3	BDL	BDL	5.8	1.2
Brain	BDL	BDL	BDL	BDL	BDL	BDL	BDL
Kidney	0.7	BDL	2.0	1.7	0.6	4.2	1.5
Spleen	BDL	BDL	BDL	BDL	BDL	14.3	BDL
Feces	379.0	198.4	NT	66.5	179.6	NT	NT
Regurgitant	235.0	145.5	NT	NT	NT	NT	NT

3.3. STX congeners and bioconversion

The crude A. catenella extract used in our study contained a suite of STX congeners that were quantified using HPLC prior to the start of the dosing trial. The original toxin profile consisted of six individual or epimer pair groups of toxins (Fig 2): C1C2 (55.6%), Saxitoxin (18.9%), GTX5 (11.7%), GTX2,3 (7.3%), Neosaxitoxin (NEO; 4.6%), and GTX1,4 (1.9%). Subsequent HPLC analysis of tissues from two murres that suffered acute toxicity and death at 415.99 μg STX-eq·kg⁻¹ and two that were euthanized after ingesting 104.50 μg STX-eq·kg⁻¹ showed considerable changes to congener profiles (Table 3; Fig. 2-3). In these samples, there were decreases in percent mass of C1C2 toxins and corresponding increases in amounts of GTX1,4 and STX, with changes being relatively uniform in tissues from all four individual murres. (Fig. 2-3). The purified STX-diHCl appeared to undergo very little bioconversion during ingestion by murres. This toxin was certified 99.5% pure when shipped, and congener profiles in the tissues of three murres showed only STX in tissues, with small amounts of decarbamoylsaxitoxin (dcSTX), NEO, and GTX1,4 detected in feces and regurgitant samples (Table 3; Fig. 2-3).

Fig 2.

Paralytic shellfish toxin profiles from tissues analyzed from four individual common murres (Uria aalge) that died acutely or were euthanized after being dosed with the crude Alexandrium catenella extract (IC-8 B1). Bars represent total toxin concentrations of each congener in tissues reported as saxitoxin (STX) equivalents (μg·100g⁻¹ STX-eq). Pie charts depict the overall toxin profiles in all tissues combined (% mass) by individual bird. STX = Saxitoxin, GTX = Gonyautoxin, NEO = Neosaxitoxin, dcSTX = decarbamoylsaxitoxin, dcGTX = decarbamoylgonyautoxin.

Table 3.

High-performance liquid chromatography (HPLC) analysis of total concentrations of saxitoxin (STX) and its associated congeners in tissues collected from common murres (Uria aalge) dosed with saxitoxin dihydrochloride (STX-diHCl) or a crude Alexandrium catenella extract (IC-8 B1). Total paralytic shellfish toxin (PST) values and individual congeners reported as saxitoxin equivalents (STX-eq; μg·100g⁻¹). HPLC analysis only occurred on selected tissues that had quantifiable STX via enzyme-linked immunosorbent assay (ELISA) testing. Tissue values are expressed as μg STX-eq·100 g⁻¹. Abbreviations are as follows: GTX = Gonyautoxin, NEO = Neosaxitoxin, dcSTX = decarbamoylsaxitoxin, dcGTX = decarbamoylgonyautoxin, BDL = below detection limits, UGI = upper gastrointestinal tract, Sm Intestine = small intestine, Lg intestine = large intestine.

Bird ID/Tissues	Total	dcGTX2,3	C1C2	dcSTX	GTX2,3	GTX5	STX	GTX1,4	NEO
STX-diHCl
UA49
UGI	4.0	BDL	BDL	BDL	BDL	BDL	4.0	BDL	BDL
Sm Intestine	18.5	BDL	BDL	BDL	BDL	BDL	18.5	BDL	BDL
Lg Intestine	24.4	BDL	BDL	BDL	BDL	BDL	24.4	BDL	BDL
Regurgitant	369.9	BDL	BDL	14.3	BDL	BDL	313.0	39.5	3.1
Feces	354.3	BDL	BDL	2.0	BDL	BDL	340.5	10.7	1.1
UA54
UGI	2.4	BDL	BDL	BDL	BDL	BDL	2.4	BDL	BDL
Sm Intestine	4.2	BDL	BDL	BDL	BDL	BDL	4.2	BDL	BDL
Lg Intestine	16.2	BDL	BDL	BDL	BDL	BDL	16.2	BDL	BDL
Regurgitant	144.1	BDL	BDL	0.8	BDL	BDL	141.7	BDL	1.6
Feces	170.3	BDL	BDL	1.0	BDL	BDL	157.4	11.3	0.7
UA55
UGI	22.9	BDL	BDL	BDL	BDL	BDL	22.9	BDL	BDL
Sm Intestine	17.3	BDL	BDL	BDL	BDL	BDL	17.3	BDL	BDL
Lg Intestine	21.2	BDL	BDL	BDL	BDL	BDL	21.2	BDL	BDL
Lung	1.5	BDL	BDL	BDL	BDL	BDL	1.5	BDL	BDL
A. catenella Extract
UA46
UGI	2.7	BDL	0.2		0.3	0.1	1.6		0.5
Sm Intestine	52.1	BDL	0.7	0.1	1.5	0.4	11.4	37.9
Lg Intestine	83.9	0.3	3.6	0.6	7.7	1.8	55.0	12.5	2.3
Feces	71.4	BDL	4.8	0.3	6.0	1.8	32.3	20.5	5.8
UA66
UGI	4.5	BDL	0.2	BDL	0.4	0.1	2.8	BDL	0.9
Sm Intestine	5.3	BDL	0.2	BDL	0.4	0.1	2.8	BDL	1.8
Lg Intestine	8.0	BDL	0.7	0.1	1.1	0.2	5.4	BDL	0.7
Kidney	21.4	BDL	0.0	BDL	BDL	0.0	0.3	20.6	0.4
Feces	223.7	0.6	14.2	1.0	19.2	6.0	118.1	40.0	24.5
UA48
Liver	8.8	BDL	0.1	BDL	BDL	0.1	1.1	7.1	0.4
UGI	44.2	BDL	2.5	0.3	4.3	1.1	26.0	7.3	2.9
Sm Intestine	45.3	BDL	1.1	0.5	2.2	1.7	22.2	16.4	1.2
Lg Intestine	55.8	0.4	2.9	0.6	5.4	2.5	32.0	8.1	4.1
Heart	12.0	BDL	0.1	BDL	0.4	0.1	11.1	BDL	0.3
Lung	2.5	BDL	0.4	0.0	0.0	0.2	1.5	BDL	0.4
Kidney	28.6	BDL	0.1	0.1	0.4	0.3	2.4	24.8	0.6
Spleen	6.1	BDL	0.5	BDL	1.2	0.4	3.9	BDL	BDL
UA64
UGI	36.0	0.4	2.6	0.3	3.7	0.8	17.8	7.7	2.7
Sm Intestine	63.5	BDL	1.4	0.8	3.1	2.3	39.3	15.3	1.3
Lg Intestine	33.7	BDL	1.8	0.4	3.1	1.6	18.2	5.8	2.7
Kidney	25.8	BDL	BDL	BDL	BDL	0.1	0.5	24.8	0.4

Fig 3.

Toxin profiles from high-performance liquid chromatography (HPLC) analysis of a crude Alexandrium catenella extract (IC-8 B1) (left; % molar) and the mean toxin profiles recovered from tissues of four common murres (Uria aalge) (right; % mass) that died acutely or were euthanized. STX = Saxitoxin, GTX = Gonyautoxin, NEO = Neosaxitoxin, dcSTX = decarbamoylsaxitoxin, dcGTX = decarbamoylgonyautoxin. ¹dcSTX and dcGTX2,3 represent less than 1% of detected congeners across murre tissues.

3.4. Behavioral results

The relative frequency of observed behaviors differed significantly between dosed and control birds in the chronic study (χ² = 18.99, df = 3, p = 0.0003), in that dosed birds spent less time resting compared to control birds (Table 4). In the ED₅₀ study, birds did not differ significantly in their behaviors across treatments overall (χ² = 5.47, df = 10, p = 0.86), but the distribution of behaviors varied during periods in which abnormal behaviors were observed (χ² = 78.09, df = 10, p <0.0001). Birds dosed with STX-diHCl were observed standing on land more often than control birds in the chronic study (treatment: 44%, control: 18%; Table 4). These behavioral differences in resting and standing rates were also observed in the ED₅₀ study during the periods in which abnormal behaviors were also observed, with much lower rates of resting in dosed (A. catenella extract: 18%, STX-diHCl: 27%) than in control birds (47%), and much higher rates of standing in dosed (A. catenella extract: 55%, STX-diHCl: 50%) than in control birds (19%; Table 4).

Table 4.

Percentage of observed time intervals of different common murre (Uria aalge) behaviors for both the median effective dose (ED₅₀) and chronic dosing trials. Murres were dosed with saxitoxin dihydrochloride (STX-diHCl) or a crude Alexandrium catenella extract (IC-8 B1). Behaviors rarely observed (less than 5% of the time) are not reported. Numbers in parentheses represent the percent of time intervals for each behavior that occurred when dosed murres behavior was designated as abnormal.

Behavior	ED50 Study			Chronic Dosing Study
	Control	A. catenella Extract	STX-diHCl	Control	STX-diHCl
Resting	47	40 (18)	39 (27)	57	31
Standing	19	18 (55)	24 (50)	18	44
Swim/Dive	14	14 (3)	15 (0)	12	14
Preen	8	13 (6)	8 (0)	9	7
Foraging	6	7 (3)	4 (0)	0	0
Other	0	1 (9)	1 (15)	0	0

Abnormal behaviors were observed in murres dosed with STX-diHCl at concentrations above our estimated ED_50C (45 μg·kg⁻¹), and in all birds that received doses of the A. catenella extract (Table 1). Regurgitation combined with head shaking was the most common clinical and abnormal response to STX ingestion and was observed in birds dosed with both toxin treatments. Uneasy or unsettled posture that presented as a constant up-and-down body movement or unwillingness to rest was another common abnormal behavior and would typically precede regurgitation. Murres also showed varying levels of wing abduction (distal wing paralysis) throughout our study, ranging from almost imperceptible in low doses to nearly complete wing paralysis in individuals receiving doses at or above the ED_50E value. In more severe cases, paralysis of the legs and head were also observed. Abnormal behaviors typically occurred immediately after dosing, with an average recovery time of 30 min (SD = 26). There was no apparent relationship between recovery time and dose concentration except highest doses for STX-diHCl (>89 μg·kg⁻¹; Fig. 4). At the highest dose of the STX-diHCl administered in our study (100.45 μg·kg⁻¹), one bird suffered acute death and another took more than four times longer than average to recover (Fig. 4).

Fig 4.

Recovery time of common murres (Uria aalge) dosed with a crude Alexandrium catenella extract or saxitoxin dihydrochloride (STX-diHCl). Red-dashed lines depict median effective dose (ED₅₀) estimates for each toxin type (89 μg·kg⁻¹ STX-diHCl and 366 μg·kg⁻¹ IC-8 B1 Alexandrium catenella extract STX-equivalents) and gray-dashed line represents our clinical response to the toxin (ED_50c) estimate for STX-diHCl (45 μg·kg⁻¹).

3.5. Chronic dosing results

Based on results of our ED_50E trial, we chose to set our STX dose for the chronic dosing trial at 76.2 μg·kg⁻¹, resulting in a final volume of 59.1 μg of STX-diHCl being injected into each capelin or silverside. Murres in the treatment group ingested between 0 and 154.4 μg of STX-diHCl daily depending on individual body mass and willingness to accept dosed fish at each scale session (Table 5). Total ingested toxin per bird ranged from 409 to 905 μg over the 7-day study duration, with multiple individuals ingesting dosed fish twice daily (Table 5). Body weights of all birds in the treatment group decreased slightly over the course of the study (≥ −5.5%), while all control group birds experienced modest body weight increases over the same period (1.1 to 3.6%; Table 5). Results of a Spearman rank correlation test showed no significant association between weight loss by murres and total STX intake (r = −0.112, p = 0.729, n = 12). Our repeated measures ANOVA results showed that treatment murres consumed less fish per day during the trial as compared to control birds with a significant interaction effect of Time:Status on total forage consumption (F_(2,62.5) = 5.09, p = 0.008). Comparison of least-squares means showed a reduction of overall fish intake by treatment murres during the seven days of STX exposure of 187 g·day⁻¹ per day as compared to control murres (Table S1; Fig. 5). Following the conclusion of the study, total fish intake for both groups rebounded to slightly higher than before STX exposure (Fig. 5).

Table 5.

Total daily intake (μg) of saxitoxin dihydrochloride (STX-diHCL) by common murres (Uria aalge) during a chronic dosing study and resulting body mass of each individual at the start and end of the study.

Bird ID	Day 1	Day 2	Day 3	Day 4	Day 5	Day 6	Day 7	Total STX Intake (μg)	Starting Mass (g)	Ending Mass (g)	Mass change (%)
Treatment
UA43	82	163	81	82	0	84	83	576	720	708	−1.7
UA45	143	74	72	74	0	73	72	507	827	823	−0.5
UA47	81	84	83	83	0	82	83	498	728	718	−1.5
UA51	135	70	69	69	0	140	69	552	876	853	−2.7
UA58	145	147	73	72	147	147	76	806	817	782	−4.4
UA65	150	152	150	148	154	0	151	905	787	782	−0.6
UA67	132	138	136	69	71	71	140	758	894	846	−5.5
UA68	136	137	137	133	138	139	70	889	871	842	−3.3
UA501	0	0	147	153	77	75	78	529	800	760	−5.0
UA521	0	0	133	68	69	69	70	409	896	850	−5.1
UA561	0	0	150	150	149	148	75	671	807	792	−1.9
UA571	0	0	154	78	157	0	78	468	763	758	−0.7
Control
UA40	-	-	-	-	-	-	-	0	814	836	+2.7
UA41	-	-	-	-	-	-	-	0	794	802	+1.1
UA42	-	-	-	-	-	-	-	0	810	839	+3.6
UA44	-	-	-	-	-	-	-	0	839	860	+2.5

¹Individuals started in control group but were converted to treatment after day 2 of chronic study.

Fig 5.

Total fish consumption (g/day) of control and treatment common murres (Uria aalge) before, during, and after a chronic dosing trial. During the trial, prey items that had been intracoelomically injected with 59.1 μg of saxitoxin dihydrochloride (STX-diHCl) were offered to treatment murres twice daily (AM and PM). Each time period (before, during, and after) consisted of seven days. For each group, the middle horizontal line in the box represents the median value, the box represents the interquartile range, the vertical lines represent the full extent of the values, and the dots represent the outliers.

Abnormal behaviors during the chronic dosing trial consisted exclusively of regurgitation. Of the 104 total incidents in which murres ingested toxic fish, 52% resulted in regurgitation. Murres that ingested only dosed fish regurgitated a higher proportion of time (60%) whereas birds that ingested additional offered fish regurgitated less often (44%). The average time from ingestion of toxin-containing fish to regurgitation was 14.5 min.

4. Discussion

4.1. ED₅₀ values and ecological effects of STX

Many studies have attempted to determine the role, if any, PSTs might play in seabird mortality events ( Shumway et al., 2003; Starr et al., 2017; Ben-Gigirey et al., 2021; Van Hemert et al., 2020, 2022). Most studies have relied on data gathered from opportunistically collected carcasses or lethally sampled birds, making interpretation of tissue concentrations, or assignment of cause of death difficult due to unknown variables. The estimation of an LD₅₀ value for STX in birds produced by Dusek et al. (2021) provided the first information on lethal thresholds for this toxin in avian species, but still left unanswered questions regarding sublethal effects and specific physiological variables, particularly in seabirds. The present study estimated an ED₅₀ value of 89 μg·kg⁻¹ (95% CI 72–103 μg·kg⁻¹) for purified STX-diHCL and 366 μg STX-eq·kg⁻¹ (95% CI 309–399 μg STX-eq·kg⁻¹) for a naturally occurring A. catenella extract composed of a broad suite of STX toxins. The accompanying tissue results for STX and other PST concentrations in combination with known levels of toxin ingestion provides more precise data for the interpretation of toxin concentrations found in wild seabirds (Tables 1-2). Overall, the combined ED₅₀ estimates, subsequent tissue distribution data, behavioral data, and chronic toxicity described in this study provide valuable information about the ecological effects that PSTs might have on seabirds, a question that has eluded researchers in the past decade.

Murres that ingested STX-diHCl doses ≥89 μg·kg⁻¹ exhibited behavioral responses, including pronounced paralysis of the wings, legs, and head, that could have fatal consequences in a natural setting, where adverse weather conditions or other factors such as risk of predation present more severe obstacles to recovery than were simulated in our experiment. Importantly, the associated 95% CI of our ED_50E estimate (72–103 μg·kg⁻¹) falls within the CIs of the LD₅₀ estimate for mallards (69–275 μg·kg⁻¹; Dusek et al., 2021) and 1 of 2 birds in the present study dosed at 100.5 μg·kg⁻¹ suffered acute death. Taken together, these results indicate there may be a narrow threshold between a lethal dose of STX for murres and an ecologically relevant, yet survivable dose. Our dosing methodology may also have influenced this ED₅₀ estimate. For example, LD₅₀ values in mice can fluctuate depending on the route of toxin administration and time since last feeding (Munday et al., 2013; Finch et al., 2023). Researchers found a nearly two-fold increase in oral LD₅₀ values in non-fasted compared to fasted mice, and a nearly three-fold increase when mice were given access to food during dosing (Finch et al., 2023). Murres in our study were fasted overnight prior to each dosing event, resulting in delivery of toxins to an empty stomach. This situation could occur during the breeding season in the wild, when birds are alternating incubation duties and foraging bouts; however, it is unlikely that birds would be fasted for such periods during other timeframes. Additionally, birds in the wild would be ingesting a prey matrix along with STXs, which may reduce the acute effects of STXs. These considerations, when taken in context with our ED₅₀ estimates, may greatly increase what would be an ecologically relevant dose of STX for murres in the wild.

Given our ED_50E of 89 μg·kg⁻¹, we calculated that it would require an STX concentration in excess of 1,780 μg·100g⁻¹ for a single 5 g forage fish (murres in Alaska generally eat forage fish between 5–10 g; Piatt et al., 2020) to elicit an ecologically relevant response in a healthy murre (assumed to be around 1 kg in the Gulf of Alaska; Piatt et al., 2020). Paralytic shellfish toxins have been detected in the prey of seabirds in Alaska (forage fish: 20%; marine invertebrates: 54%; Van Hemert et al., 2020). Concentration of PSTs in forage fish have generally been well below the 1,780 μg·100g⁻¹ threshold during seabird die-offs (216 μg·100g⁻¹ Shearn-Bochsler et al., 2014; 97 μg·100g⁻¹ Nisbet, 1983; 360 μg·100g⁻¹ Levasseur et al., 1996; 494.1 μg·100g⁻¹ Van Hemert et al., 2022). However, during a mortality event in the St. Lawrence Estuary, which killed marine fish, birds, and mammals, PSTs measured in sand lance (Ammodytes spp.) were near (1,321 μg·100g⁻¹ in viscera) and above (1,922 μg·100g⁻¹ in stomach contents) our estimated threshold to elicit an ecologically relevant response in murres (Starr et al., 2017). When considered alongside the propensity for murres to regurgitate prey containing STXs and the potential for increased toxic thresholds when toxic prey are consumed with additional non-toxic prey, makes the likelihood of murres experiencing acute toxicity from ingesting a single prey item unlikely. Future experiments using modified dosing techniques employing the use of multiple prey matrices and fasted versus non-fasted seabirds could fill some of these information gaps.

The large difference in ED₅₀ values between STX-diHCl and the A. catenella extract used in our study was unexpected, considering our dosing methodology and timeline. Lethal thresholds for PSTs can fluctuate across taxa and even within the same species, but a more than 300% increase in thresholds, as observed between toxin types in our study, was unexpected (Munday et al., 2013; Dusek et al., 2021). All doses of the A. catenella extract were calculated based on total STX TEFs (Munday et al., 2013; Selwood et al., 2017). These values have largely been based on intraperitoneal injections using a mouse bioassay, a method that has proven useful due to its convenience and ability to minimize the amount of dosing material needed, although recent work has questioned the validity of these methods for determination of PST congener TEFs, indicating a need for revising these values, particularly for oral routes of administration (EFSA, 2009; Munday et al., 2013; Selwood et al., 2017; Rattner et al., 2022). Inconsistencies in TEF calculations may account for some of the difference in our ED₅₀ estimates. Additional work comparing the toxicities of A. catenella extracts in future dosing studies may allow for a better understanding of these toxins in their native form.

4.2. STX tissue concentrations

Tissue concentrations of STXs in our samples from ELISA and HPLC analysis followed trends observed in previously conducted laboratory and field-based studies (Starr et al., 2017; Dusek et al., 2021; Smith et al., 2022) (Tables 1-2). The highest levels of STXs were found in digestive tract tissue, with quantifiable levels detected in all UGI, small intestine, and large intestine samples (Table 2). Brain, serum, and muscle samples had no detectable levels of STX, while liver, heart, lung, and kidney contained low, but still quantifiable levels in some individuals (Table 2). Gastrointestinal tissue from field-based studies is commonly used as a testing matrix for STX exposure in seabirds ( Starr et al., 2017; Van Hemert et al., 2020, 2021; Smith et al., 2022) and has proven to be effective as a screening tool.

Our results further support the use of GI tissue as a testing matrix to screen for PST exposure due to the consistency of STX detection and the apparent timing of its detectability, which ranged from as little as 11 min to over 90 min post-ingestion (Table 1). Similar to the conclusions drawn by authors of previous studies, caution should be exercised in using stand-alone tissue concentrations for the determination of saxitoxicosis as the cause of death in unknown die-off events (Dusek et al., 2021). The lack of clear association between high gastrointestinal (GI) tissue concentrations and associated STX doses, combined with the rapid and full recoveries from STX doses near the ED₅₀ threshold observed in our study make establishing causality from tissue concentrations alone difficult.

The lack of detection of STXs in serum collected from murres is notable (Table 2), indicating that blood draws may not effectively assess STX exposure in seabirds. Quantification of STX from whole blood sampled from mallards in a previous dosing study was also unsuccessful using both ELISA and HPLC (Dusek et al., 2021). Multiple studies focused on domoic acid (DA) exposure in marine mammals have used serum as a screening matrix (Lefebvre et al., 2010, 2016), and DA has been detected in serum from avian species in a previous dose-finding study (Silvagni, 2003). Differences in chemical structure and physiological methods of toxin excretion, however, may account for variation in detectability between DA and STX in serum using ELISA techniques. Moreover, case studies of human STX exposure events have successfully quantified STXs from serum collected from patients exposed through ingestion of toxic shellfish using HPLC (Gessner et al., 1997). These cases involved ingestion of considerable amounts of toxin, indicating that detectability in serum may only be possible after ingestion of large doses and, as such, would not be a useful method to monitor low-dose chronic STX exposure in free-living seabirds.

4.3. Behavioral response to STX ingestion

Regurgitation was the most prevalent abnormal behavior in response to both STX-diHCl and A. catenella extract ingestion. This behavior has been observed in other seabird species when exposed to STXs and has been hypothesized as a learned behavior for toxin avoidance (Coulson et al., 1968; Kvitek, 1991). Murres are not known to regurgitate as part of normal feeding or chick rearing behavior. Prey caught by adults is typically ingested underwater or shortly after surfacing (Sanford and Harris, 1967), and prey are delivered whole to rearing chicks (Ainley et al., 2021). Additionally, the absence of observed regurgitation in control birds in any phase of our study indicates that the repeated regurgitation is a direct response to STX ingestion.

In addition to regurgitation, dosed murres spent more time standing on land and less time resting on land or water during both the ED₅₀ and chronic phases of study (Table 4). This unsettled behavior may be due to multiple factors, including the onset of gastric distress or perhaps more mechanistically related paralytic effects of the toxin. This decrease in resting behavior associated with toxin ingestion could cause changes to preening maintenance or locomotion in wild murres, thus potentially increasing energetic demands (Ellis and Gabrielsen, 2002).

In our study, we observed rapid recovery and continued foraging ability of murres that recently ingested STXs. Except for a single individual that survived a dose above our ED₅₀ estimate (Table 1; Fig. 4), all murres appeared to recover from toxic effects at a similar rate (approximately 30 min), with limited visual evidence of lingering effects. Wing abduction and paralysis that was observed in higher doses is a concerning effect for murres given that they are wing-propelled diving birds (Ellis and Gabrielsen, 2002; Ainley et al., 2021). We did not observe any reduced foraging ability for murres exposed to STX in our study, but it is unclear if this would translate to natural marine environments, where murres dive much deeper in pursuit of prey (Ainley et al., 2021). Additionally, our foraging trials occurred 90 min post dosing, and dosed murre behaviors returned to normal around 30 min post dosing. It is unclear whether a murre’s foraging ability would be reduced in the first 30 min after ingestion of STXs.

4.4. Biotransformation of STX congeners

The A. catenella extract used in our study was specifically chosen because it contains a higher proportion of di-sulfated toxins (C1C2) and lesser proportions of the mono-sulfated gonyautoxin (GTX5). These PSTs are weaker in toxicity when compared to other STX analogs (Wiese et al., 2010) and considered to be more labile, potentially converting to more potent derivatives when ingested (Asakawa et al., 1995; Bricelj and Shumway, 1998). These transformations are well documented in marine bivalve species, where chemical, bacterial, or enzymatic actions can alter toxin profiles containing less toxic sulfated C toxins and certain GTX analogs to more toxic carbamate constituents (e.g., GTXs, NEO and STX; Oshima et al., 1976, 1993; Cembella et al., 1994; Asakawa et al., 1995; Bricelj and Shumway, 1998; Jaime et al., 2007).

Work has been limited on the possible biotransformation processes in higher trophic level species. A study in white seabream (Diplodus sargus) documented transformation and elimination of PSTs using laboratory experiments (Costa et al., 2011). Another study reporting on a multispecies mortality event describes PST biotransformation in forage fish, seabird, and marine mammal tissues following a large bloom of A. tamarense (= A. catenella) (Starr et al., 2017). Documented shifts in toxin profiles from those determined in Alexandrium cells with reductions in di-sulfated C1C2 toxins as well as the carbamate toxins GTX1,4 were primarily coincident with increases in STX and NEO, as well as transformation from NEO to STX. These changes were specific to the tissue or gut content analyzed in sand lance, razorbill (Alca torda), and gray seals (Halichoerus grypus) and followed described or suspected trends of biotransformation (Starr et al., 2017). Starr et al. (2017) speculated that toxin transformation may continue as toxins proceed through the food web, and our results in murres support this hypothesis.

The considerable differences between the original toxin profile of our A. catenella extract and the resulting congener profiles isolated from bird tissues (Fig. 2-3) resemble patterns of biotransformation of PSTs observed in previous studies. The substantial reduction in C1C2 and GTX5 toxins from the native A. catenella profile and increases in STX, GTX1,4, and NEO (Fig. 3) shows that biotransformation is occurring at some point post-ingestion. This is further supported by the uniformity of these differences in congener profiles across all four murres that received doses of the A. catenella extract (Fig. 2). Methodological constraints, however, impede conclusive determination of these biotransformation processes. We are unable to make a direct mass comparison of toxin profiles, relying instead on a molar fraction calculated from HPLC analysis of the A. catenella extract, and percent mass back calculated from HPLC profiles from murre tissues. Additionally, the overall nature of our study design, which aimed to minimize unintended mortality of murres, hinders robust inference into the specific mechanisms and extent of PST biotransformation in our samples. Future work involving additional dosing experiments and possibly in vitro spike recovery experiments with seabird tissue and other A. catenella extracts may allow for more definitive conclusions on biotransformation of STXs by seabirds and other upper trophic level consumers. A better understanding of PST compositional changes as these toxins move through marine food webs will improve understanding of PST food-web transfer and help determine net toxicity in higher level marine consumers.

4.5. Chronic STX ingestion results

Our chronic dosing experiment aimed to closely resemble what seabirds might encounter in their natural environment, where murres, or other seabirds, consume multiple prey items potentially containing sublethal levels of STX. The possibility of cumulative toxic effects on seabirds resulting from repeated, sublethal exposure to PSTs is a concern that multiple researchers have raised in previous studies but is very difficult to ascertain from traditional, field-based collections (Shumway et al., 2003; Gibble et al., 2018; Piatt et al., 2020; Dusek et al., 2021; Van Hemert et al., 2020, 2022).

Murres in our study regularly consumed cumulative daily amounts of STX well above our estimated ED₅₀ value, with multiple instances of individuals consuming levels near the LD₅₀ threshold found in mallards (Table 5; Dusek et al., 2021). The tendency for murres to regurgitate fish that had been injected with toxin supports theories that regurgitation may act as a method of toxin avoidance, and greater tendency towards this reaction may be advantageous for murres in the wild (Kvitek, 1991; Shumway et al., 2003). The ability of murres to ingest near-lethal levels of STX in a relatively short (<6 h) timeframe and continually detect and subsequently regurgitate toxic fish indicates that seabirds may be able to tolerate exposure to higher concentration of STXs in the wild, and the idea of cumulative toxicity over specific time frames may be more complex than previously thought.

Despite these potential avoidance strategies, the significant reduction in total fish intake by treatment birds during our chronic study (Fig. 5) is notable when considering potential sublethal effects of chronic STX ingestion. Emaciation or poor body condition has regularly been identified as the cause of death during seabird die-offs in the Gulf of Alaska and Bering and Chukchi Seas (Piatt et al., 2020; Kaler et al., 2022; Jones et al., 2023; Renner et al., 2024). Murres maintain very high metabolic rates, particularly those that inhabit arctic or subarctic marine environments, where costs of thermoregulation are greater (Ellis and Gabrielsen, 2002). It is estimated that murres may need to consume as much as 56% of their body mass daily (560-g for a 1-kg murre) to meet energetic costs, equating to >60 high quality forage fish (Piatt et al., 2020). Given this context, the 187-g reduction in daily fish intake by treatment murres in our study may lead to serious ecological consequences in wild murres, especially if they are subject to additive effects of other environmental stressors.

5. Conclusions

This study determined behavioral responses of common murres to STX ingestion and identified thresholds for clinical and ecologically relevant responses. Our results show that murres can recover quickly from STX ingestion while maintaining foraging ability when toxin exposure is below 89 μg·kg⁻¹, but toxic effects may lead to drowning or acute death above this threshold. We detected STX in a broad range of tissues, with continued support for the use of GI tissues as screening matrices for PSTs in seabirds. Our comparison of purified STX and an A. catenella extract showed evidence for biotransformation in seabirds and raised questions about the use of TEFs for calculating doses through oral exposure. This may be especially true when multiple PSTs are involved, much like our A. catenella extract, as there may be unknown synergistic effects of the multiple PSTs as well as potential biotransformation effects. Finally, results from our chronic study show that murres can repeatedly detect and regurgitate fish containing STX, potentially acting as an avoidance method. However, murres that ingested sublethal doses of STX during our chronic study consumed significantly less fish, leading to reduced body weights, which indicate that murres may face nutritional challenges if PST exposure is prolonged. Many questions still exist about the effects of chronic, sublethal ingestion of PSTs on seabird populations and additional experimental and field-based studies are warranted. Overall, our data provide pertinent information for researchers and managers focused on seabird health and can inform observations during future mortality events.