Rapid determination of thyroid hormones in blood plasma from Glaucous gulls and Baikal seals by HybridSPE®-LC-MS/MS

A rapid hybrid solid phase extraction (HybridSPE®) protocol tailored to liquid chromatography–electrospray ionization tandem mass spectrometry (LC–ESI–MS/MS) analysis, was developed for the determination of four thyroid hormones, L-Thyroxine (T4), 3,3,5-triiodo-L-thyronine (T3), 3,3,5-triiodo-L-thyronine (rT3) and 3,3diiodo-L-thyronine (T2) in blood plasma from Glaucous gulls (Larus hyperboreus) and Baikal seals (Phoca sibirica). The use of target analyte specific 13C internal standards allowed quantification to be performed through the standard solvent calibration curves and alleviated the need to perform quantification with matrix match curves. The relative recoveries were 100.0–110.1 % for T4, 99.1–102.2 % for T3, 100.5–108.0 % for rT3, and 100.5–104.6 % for T2. The matrix effects ranged from − 1.52 to − 6.10 %, demonstrating minor signal suppression during analysis. The method intra-day precision (method repeatability, RSD %, N = 5, k = 1 day) and inter-day precision (method reproducibility, RSD %, N = 10, k = 2 days) at the 1 ng/mL concentration of fortification were 8.54–15.4 % and 15.4–24.8 %, respectively, indicating acceptable chromatographic peak stabilities for all target THs even at trace level concentrations. The method limit of detection (LOD) for T4, T3, rT3 and T2 was 0.17, 0.16, 0.30 and 0.17 ng/mL, respectively. The HybridSPE® protocol was simple and rapid (<1 min) upon application, while the HybridSPE® cartridge did not require (as in classical SPE cartridges) any additional equilibration nor conditioning step prior sample loading. A total of 46 blood plasma samples, 30 samples collected from Glaucous gulls and 16 samples collected from Baikal seals, were analyzed for thyroid hormones to demonstrate the applicability of the developed method in these wildlife species. The concentrations of T4 and T3 in blood plasma from the Glaucous gulls were 5.95–44.2 and 0.37–5.61 ng/mL, respectively, whereas those from Baikal seals were 3.57–46.5 and 0.45–2.07 ng/mL, respectively. In both species, rT3 demonstrated low detection rate, while T2 was not detected. Furthermore, cross-array comparison between the HybridSPE®-LC-MS/MS protocol and an established routine radioimmunoassay (RIA) kit-based method was performed for T4 and T3 concentrations from selected Baikal seal plasma samples.

Disruption of normal thyroid function in animals is linked to pollution exposures [1,4], and the concentration profiles of blood circulating THs are commonly used to diagnose actual thyroid disorders [4,5]. Increased serum rT 3 concentrations, but also specific concentration profiles of THs are linked to adverse health effects [5,6]. Traditionally, immunoassay (IA)-based methods were the main choice for bioanalysis of THs, and thus far, they remain the predominant technique of analysis [6][7][8]. Currently, THs measurements in clinical routines are regularly performed by the radioimmunoassay technique (RIA) [6,8], which is a sensitive and rapid technique, but it has limitations with respect to specificity and accuracy [9][10][11][12]. As liquid chromatography (LC)-tandem mass spectrometry (MS/MS) technique evolved during the past 20 years, an increased number of bioanalytical LC-MS/MS methods were developed for THs analysis [9,13]. The LC-MS/MS technique allowed for THs quantification with enhanced specificity, reproducibility, and sensitivity [14][15][16]. A few gas chromatography-mass spectrometry (GC-MS) methods are also reported, but with limited applicability due to often laborious derivatization processes [16].
The THs are mostly determined in blood serum or plasma and they are bound to proteins (e.g., 99.9 and 99.7 % of T 4 and T 3 , respectively), while a small percentage (< 0.5 %) of those remain free (unbound) [16]. For the measurement of total concentrations of THs (bound and unbound hormone chemical species), the protein-TH binding sites are disrupted through protein precipitation methods, which destroy the binding ability of the proteins through denaturation [13,16]. Strong acid reagents (e.g., HCl), salts and organic solvents are reported to yield clear supernatants with effective denaturation of proteins for the determination of THs [13,16]. However, the major bioanalytical challenge for THs analysis is the light, heat and chemical stability, which can be significantly impacted when exhaustive sample preparation protocols are applied [13,17]; it is documented that T 4 can be degraded or biotransformed (due to remaining deiodinase activity in the samples) to T 3 and rT 3 , while T 3 can be further converted to T 2 [3,6,18]. In particular, the transformation of T 4 to T 3 has a strong impact on T 3 concentrations, which are significantly (even 2 orders of magnitude) lower than those of T 4 in the blood of vertebrates [19]. Moreover, another major challenge in LC-MS/MS bioanalysis of THs is the uncertainty of quantification due to matrix effects that derive mainly from the high endogenous protein and phospholipid content of serum and plasma in contrast to the natural occurring trace concentrations of THs [13,16].
In this study, a rapid methodology that used the hybrid solid phase extraction (HybridSPE®) technique tailored to LC-electrospray (ESI+)-MS/MS was developed for the analysis of the total concentrations of four THs, T 4 , T 3 , rT 3 and T 2 . Blood plasma concentrations and profiles of those THs were determined in actual samples from Glaucous gulls (Larus hyperboreus) and Baikal seals (Phoca sibirica). The method performance characteristics and the applicability of the HybridSPE® method was extensively demonstrated. The advantage of the HybridSPE® technique is that the extracted biological matrix is rapidly processed, and the gross amounts of proteins and phospholipids are removed, significantly decreasing the impact of the matrix during analysis [20]. In this context, three different sample preparation techniques, including liquid-liquid (LLE), solid phase (SPE) and HybridSPE® extraction were assessed in ultrapure water and bovine blood serum. The T 4 and T 3 concentrations that were determined in the blood plasma from Baikal seals by the HybridSPE® method, were also cross-array compared with those concentrations determined by an established routine RIA kit-based method to further demonstrate measurement validity.

Sample collection
For the method development and validation, matrix standards were prepared from bovine serum samples (surrogate pool matrix), as previously performed [21]. Baikal seal (n = 16) plasma samples were collected from the Mishikha region, Russia, which is located at the southern part of Lake Baikal in April 2011 (permission for the sampling was provided by Angara-Baikal Directorate of the Russian Fisheries Agency, ref. 000024). Sampling and handling of the seals were performed in accordance with the Federal Law on fisheries and conservation of aquatic biological resources of the Russian Federation. The sample set consisted of two adult females (8 and 12 years old) and 14 pups. In total, 6 pups were female and the remaining 8 were male (and the age of all pups ranged from 1 to 2 months). The seals were euthanized for sampling of tissues for other research purposes prior to blood sampling from the heart with a 20 mL heparinized syringe. The blood samples were directly transferred into Vacuette© heparinized tubes followed by centrifugation (3000 rpm, 10 min) to collect the plasma. Subsequently, the plasma samples were transferred to cryovials and were immediately snap frozen into liquid nitrogen (− 196 • C), and thereafter transferred to − 70 • C until analysis. Glaucous gull (n = 30) plasma samples were obtained in 2017 and 2018 from Isfjorden, which is near Longyearbyen, Svalbard, Norway. Permissions for the sampling were provided by the Governor of Svalbard (ref. 17/00414-2 and 17/ 00414-6). The sample set consisted of 18 female and 12 male adults. Sampling and handling of the sea birds was performed in accordance with the regulations of the Norwegian Animal Welfare Act. Blood samples were collected primarily from the jugular vein, and plasma was obtained and stored according to the procedures specifically indicated for Baikal seals.

LLE, SPE and HybridSPE®-based sample preparation protocols
Three different sample preparation techniques were assessed for optimal TH extraction: LLE; SPE; and HybridSPE®. Pre-and postextraction matrix (bovine serum) spikes, fortified with 25 ng/mL of target THs, were prepared (similarly to previous works [20,22,23]) to calculate recoveries and matrix effects for every extraction technique. Method blanks were used for monitoring background contamination.
For the LLE, a generic protocol that involved ethylacetate as the organic solvent was used with minor modifications [24]. Aliquots of 250 μL of matrix were transferred into 15-mL PP tubes. The samples were fortified with 300 μL of 1.0 M ammonium acetate. Thereafter, the samples were extracted 3 consecutive times with 3 mL of ethyl acetate each time (3 × 3 mL). For each extraction, the mixture was shaken in an oscillator shaker for 45 min, and thereafter centrifuged (3500 rpm) for 10 min in room temperature. The supernatants were combined, and 1 mL of Milli-Q water was added. The mixture was centrifuged again, and the supernatant was transferred into another PP tube and concentrated to near dryness under a gentle nitrogen stream (N 2 ). Finally, 1 mL of (1:1 v/v) MeOH:Milli-Q water was added, vortex mixed, and transferred for LC-MS/MS analysis.
For the SPE, the extraction of the target THs was assessed by both, Oasis® HLB and Strata™-X-CW cartridges. Aliquots of 150 μL of matrix were transferred into 15-mL PP tubes. The samples were fortified with 300 μL of 1.0 M ammonium acetate. Both Oasis® HLB and Strata™-X-CW cartridges were conditioned with 3 mL of MeOH and 3 mL of acidified Milli-Q water (adjusted to pH < 3, acidified with HCl solution). Thereafter, the samples were passed through the cartridges, washed with 3 mL of Milli-Q water, and then dried under vacuum for 10 min (8-10 psi). The target THs were eluted with 3 mL of MeOH and 3 mL of MeOH with 2 % (v/v) ammonium hydroxide from the Oasis® HLB and Strata™-X-CW cartridges, respectively. All eluents were concentrated to near dryness under a gentle nitrogen stream (N 2 ). The solvents were reconstituted with 300 μL of (1:1 v/v) MeOH:Milli-Q water, vortexmixed, and transferred for LC-MS/MS analysis. For the HybridSPE®, two protein precipitation agents, MeOH with 1 % ammonium formate (w/v) and ACN with 1 % formic acid (v/v) were assessed for optimal THs extraction and cartridge functionality. In each case, aliquots of 150 μL of sample were transferred into 15-mL PP tubes and 450 μL of the respective precipitation agent were added. The samples were vortex-mixed and centrifuged (3500 rpm) for 10 min in room temperature. The supernatants were collected and passed directly through the pre-washed HybridSPE® cartridges. They were pre-washed with 1 mL MeOH, which was pipetted on the cartridges, while the vacuum on the SPE manifold was kept between 8 and 10 psi. The extracts, ~450 μL, were collected and transferred directly for LC-MS/MS analysis.

LC-ESI(+)MS/MS analysis
The chromatographic separation was carried out using an Acquity UPLC I-Class system (Waters, Milford, USA) coupled to a triple quadrupole mass analyser (QqQ; Xevo TQ-S) with a ZSpray ESI ion source (Waters, Milford, USA). The injection volume was 2 μL and the mobile phase flow rate was 200 μL/min. The mobile phase comprised Milli-Q water (0.1 % v/v formic acid) (A) and MeOH (B). The gradient elution started at 50 % (v/v) A, decreased to 20 % A within 4.0 min, then further decreased to 0 % A within 2.0 min (6.0 th min), and reverted back to 50 % A at the 6.0 th min that was further held for 2.0 min for a total run time of 8.0 min. The specific MS/MS parameters are presented in Table S1. Nitrogen was used as both the desolvation and cone gas, which were set at a flow rate of 800 and 150 L/h, respectively. The electrospray ionization voltage was set at +2.8 kV. The desolvation and source temperatures were set at 500 and 150 • C, respectively. The calibration standard solutions were prepared in (1:1 v/v) MeOH:Milli-Q water.

HybridSPE®-LC-MS/MS method validation
The calibration of the ESI method was verified by injecting solvent calibration standards at concentrations of 0.25-50.0 ng/mL (0.25, 0.50, 1.00, 2.50, 5.00, 10.0, 25.0 and 50.0 ng/mL). For every target analyte, a specific 13 C 6 -labelled internal standard was used, and quantification was performed with the internal standard method [22,23]. The method limits of detection (LODs) and quantification (LOQs) were estimated for each target analyte (from the respective 13 C 6 -labelled internal standard) as 3 and 10 times the signal from the baseline noise (S/N ratio) in bovine serum matrix, respectively. Precision was assessed through repeatability and reproducibility experiments. For method repeatability (method intra-day precision) experiments, the bovine serum was fortified at two concentrations (10 and 20 ng/mL) and three replicate analyses (N = 3) were prepared for each concentration. The precision of the method was also extensively assessed at a low fortified concentration; method repeatability and reproducibility (method intra-and inter-day precision, respectively) were assessed in bovine serum at 1 ng/mL with 5 replicates prepared during the same day (N = 5 × 1) and with 10 replicates prepared in-between two consecutive days (N = 5 × 2), respectively. Since the method intra-day precision at the low fortified concentration (1 ng/ mL) was assessed in-between two different days, the results were provided for two days (day 1 and day 2). For instrumental repeatability (instrumental intra-day precision) experiments, solvent matrix was fortified at three concentrations (1, 2 and 5 ng/mL), and three replicate analyses (N = 3) were prepared for each fortified concentration. The accuracy (trueness) was evaluated through recovery experiments at the fortified concentrations of 10 and 20 ng/mL of the target THs; absolute and relative recoveries % (as defined in [20]) were calculated in three replicates (N = 3) in both concentrations.

Routine radioimmunoassay analysis of T 4 and T 3
The analysis of T 4 and T 3 was performed according to previous work [21,22,25] with minor modifications. Briefly, 25 and 100 μL of Baikal seal plasma was added to the test kit tubes of T 4 and T 3 , respectively, with the specific radioactive antigen (1 mL of 125 I Total T 4 and 125 I Total T 3 , respectively), vortex mixed and incubated in a water bath at 37 0 C for 2 h. After incubation, the tubes were aspirated, and the reactivity was measured using the γ-scintillation counter. Calibration curves and concentrations of THs in the samples were calculated by the embedded software (SpectraWorks™ Spectrum Analysis Software, Packard Instrument Company, Connecticut, USA) of the γ-scintillation counter. The quality control/quality assurance (QC/QA) of measurements was established by the analysis of the standard reference material, Lyphocheck Immunoassay Plus Control Levels 1, 2 and 3 (BioRad, California, USA). Samples were run in triplicates (N = 3) and the repeatability was RSD % (Relative Standard Deviation %) < 10%. LODs were set in accordance to those reported by the kit protocols: 0.07 (T 3 ) and 2.50 ng/ mL (T 4 ).

Data analysis and statistical treatment
LC-MS/MS data was acquired with the MassLynx v4.1 software, while quantification processing was performed with TargetLynx (Waters, Milford, USA). Excel (Microsoft, 2018) was used for t-test and general descriptive statistics. Principal component analysis (PCA) was performed by the statistical software R (https://www.R-project.org; 2018). Data analysis did not include censored data (i.e., non-detects; NDs). Concentrations were reported as ng/mL. The statistical significance was set at p < 0.05.

Extraction efficiency of the different techniques
The extraction efficiency of the applied sample preparation protocols that were based on three different sample preparation techniques (SPE, LLE and HybridSPE®) is presented in Fig. S1. In the SPE-based protocols, the recoveries from the Oasis® HLB cartridge were 22.7-63.2 % in Milli-Q water and 53.4-74.2 % in serum, and the recoveries from the Strata™-X-CW cartridge were 29.0-52.7 % in Milli-Q water and 29.2-57.0 % in serum. The two SPE polymeric sorbents (Strata™-X-CW and Oasis® HLB) exhibit hydrophobic, π-π, and hydrophilic intermolecular forces on the target THs. Due to the polar group of N-vinylpyrrolidone present in the Oasis® HLB sorbent, there is an increased contribution of hydrogen bonding [26]. The Strata™-X-CW cartridge is designed to increase the retention of basic compounds through a carboxylic acid functional group [26]. The classes with an amine group, such as THs, can be positively charged in an acidic environment, and consequently, promote cation exchange interactions with the cartridge. THs obtain a conjugated molecular system that is prone to be strongly retained by the cartridges through π-π interactions. Overall, the Oasis® HLB sorbent functioned significantly better in the biological matrix, and consequently, it was preferred over the Strata™-X-CW sorbent (Fig. S1); this observation agrees with previous reports [27]. Both SPE-based protocols demonstrated recoveries with higher standard deviations (SDs) when compared to those from the HybridSPE®-based protocols (Fig. S1). The applied LLE-based protocol demonstrated higher recoveries than those from the SPE-based protocols in the Milli-Q water matrix (Fig. S1). However, the serum recoveries from the LLE-based protocol were lower compared to those in Milli-Q water, but were found similar to those from the SPE-based protocols (Fig. S1). The SDs of the T 3 recoveries from the LLE technique were also lower than those from the SPE technique (Fig. S1). The larger variations (SDs) in the recoveries of T 3 in relation to the other target THs were attributed to the concentration step to near dryness.
The HybridSPE® technique has so far been applied in a limited number of sub-trace concentration bioanalytical applications [20,28]. Even though it was previously reported tested for human serum analysis of THs, it was not extensively applied [13]. It can be concluded that this technique is still not commonly reported in literature in contrast to SPE and LLE-based protocols. The HybridSPE® cartridge consists of a zirconium packed-bed, low porosity filter, 0.2 μm hydrophobic frit assembly [20]. The application of HybridSPE® consists of few, relatively simple and rapid steps: the sample is mixed with a suitable organic solvent, loaded onto the cartridge, filtered on-cartridge (through the cartridge frit) and the sample is ready for analysis. Thus, the samples need less than a minute (<1 min) to pass through the cartridge, avoiding the time exhaustive typical SPE and LLE processes that can compromise the stability of THs, while sample throughput is maximized [13,20]. The optimal results with the application of HybridSPE® were obtained for THs when ACN (1 % v/v formic acid) was used as the precipitation agent. This agent demonstrated absolute recoveries (±SD) % of 43.9 (±2.21) % for T 4 (Fig. S1). The absolute recoveries (%) of the HybridSPE® technique were lower than those from the other two techniques; however, they demonstrated significantly lower uncertainties (lower SDs). The recoveries in the HybridSPE® extraction were not statistically different (t-test) when 1, 3 and 5 % v/v formic acid was added to ACN. The second precipitation agent that was assessed for HybridSPE® extraction, MeOH (1% w/v ammonium formate), provided recoveries < 8% for all THs, but even then, the actual SDs of those showed good precision (maximum recovery SD: ± 5.61%). Overall, the low uncertainty (low SDs) of the TH absolute recoveries with the HybridSPE® protocol was attributed to the rapid sample preparation that does not impact the stability of THs, but also to the lack of an evaporation and reconstitution step (in contrast to the SPE-and LLEbased protocols) that introduces by-default higher uncertainties [29].

Matrix effects and ion ratios (%) of THs analysis
The HybridSPE® protocols demonstrated significantly lower matrix effects than the SPE and LLE protocols (Table S2). The HybridSPE® protocol with ACN (1 % v/v formic acid) as the precipitation agent demonstrated matrix effects (%) of − 6.10, − 6.01, − 2.62 and − 1.52 % for T 4 , T 3 , rT 3 and T 2 , respectively. The matrix effects were found at the same order of magnitude when MeOH (1 % w/v ammonium formate) was used as the precipitation agent. The matrix effects were stronger for T 4 and T 3 in the LLE than the two SPE protocols, while the opposite was observed for rT 3 and T 2 (Table S2). Among the two SPE cartridges, only T 4 experienced significantly stronger matrix effects with Strata™-X-CW (− 28.7 %) than with Oasis® HLB (− 17.0 %). The SRM transition 778.0 > 324.2 m/z (T 4 ), which provided the highest sensitivity (from the two monitored SRMs) for T 4 (Table S1), was used successfully as the quantifier transition for all three matrices. However, the confirmation SRM transition, 778.0 > 732.1 m/z (T 4 ), suffered in the Baikal seal plasma matrix from coeluting matrix interferences (Fig. S2).

Analysis of Glaucous gull and Baikal seal plasma samples
A total of 46 plasma samples were analyzed, and the results are presented in Table 2. The highest detection rates were found for T 4 and T 3 followed by rT 3 . T 2 was not detected in any of the two species; previous studies performed also on the same species in blood matrices, did not report measuring or detecting concentrations of T 2 [6,21,31,32]. T 4 and T 3 are previously reported in plasma samples from Glaucous gulls analyzed using RIA, and the results are in good agreement with those found in this study (Table 3). All four studies, including the present study, reported higher T 4 concentrations in females as compared to males (Table 3). Minor differences in THs concentrations among the four studies can be attributed to physiological differences among the sampled individuals caused by age, reproductive and nutritional status, and the season of sampling. In addition, it is documented that persistent halogenated pollutants (e.g., chlorinated biphenyls) interfere with the THs homeostasis in Glaucous gulls, and differing concentrations of these pollutants among the studies can have influenced the reported TH concentrations [31,32].
For Baikal seals, THs were previously measured in blood serum using LC-MS/MS, and the results for T 3 and rT 3 agreed in general with those measured here in blood plasma (Table 4). Although the mean T 4 concentrations were ~ 2 fold higher in the serum samples of the male Baikal seals [6] than in the plasma samples analyzed here; the variation was large in both matrices. In the females, the T 4 concentrations in the serum and plasma were clearly within the same range ( Table 4). The mean T 3 concentrations were higher in the plasma samples than those previously reported in the serum samples. Even though the mean rT 3 concentrations were similar in the females in both studies (Table 4), the mean rT 3 Fig. 1. SRM ion chromatograms of the target THs and internal standards.

Table 1
Method intra-day precision (method repeatability, RSD %, N = 5, k = 1 day; results were provided for two individual days) and inter-day precision (method reproducibility, RSD %, N = 10, k = 2 days), calculated based on the external standard method, at the fortified concentration of 1 ng/mL (low level). concentration in the males was higher in the plasma samples analyzed here than in the previously reported mean concentration in serum. Similarly, to the Glaucous gulls, these minor differences in the THs concentrations between the two studies can be attributed to the differences in the physiological status of the sampled Baikal seals. Sex classification [females (f) and males (m)] based on the sum concentrations of THs and biometric characteristics was visualized through PCA (Figs. S3 and S4). In Glaucous gulls, the classification was more distinct than that observed in Baikal seals. This is due to that the Glaucous gull sample set involved only adults, where sex differences are already established on the TH profile, while the Baikal seal sample set involved mostly pups, where sex differences among those are not fully pronounced yet. However, the two adult female Baikal seal samples were separated distinctively from all the pups (Fig. S4), which agrees with findings from other seal species, e.g., grey seal pups (Halichoerus grypus) [33].
It is noteworthy that T 2 was not detected in any of the samples, whereas rT 3 was only detected in one Glaucous gull sample. Although concentrations of T 2 in healthy humans are reported to be 133 ± 15 [standard error of mean (SEM), N = 29] pg/mL [34], to our knowledge Table 2 Thyroid hormone concentrations (ng/mL) in plasma samples from Glaucous gulls and Baikal seals. * n.d.: not detected.
there are no reference concentrations available for seals or birds. With respect to rT 3 , there is also a lack of available reference concentrations in seals and birds. Thus, the analytical technique developed in the present study will allow also for rapid and accurate analysis of rT 3 and T 2 in serum and plasma samples, allowing for more detailed studies on the presence and role of these THs in wildlife species.
In the 14 Baikal seal plasma samples, in-which both two THs were detected, cross-array comparison οf T 4 and T 3 concentrations was performed between LC-MS/MS and RIA (Fig. 2). The mean concentrations did not demonstrate statistical differences between the two techniques for both T 4 (t-test; p = 0.61) and T 3 (t-test; p = 0.93). A significant correlation coefficient (r = 0.82; p = 0.0004) for T 4 suggests that both RIA and LC-MS/MS technique can be potentially used for the measurement of T 4 . However, when focusing on the lower concentration points of T 4 , the correlation remains poor (Fig. 2). For T 3 , no significant correlation coefficient was identified (r = -0.18; p = 0.53). The results on T 3 agreed with the previous work in Baikal seal serum samples [6], where the lack of T 3 correlation was attributed to possible immunoglobulin interference in RIA.

Conclusions
A rapid HybridSPE® method tailored to LC-MS/MS was developed for the determination of four THs in blood plasma from Glaucous gulls and Baikal seals. Target analyte specific 13 C internal standards were used for effective compensation of matrix effects and extraction losses during analysis. This allowed for quantification to be performed with the standard solvent calibration curves and alleviated the need to perform it with matrix match curves. The HybridSPE® protocol demonstrated minor signal suppression (matrix effects: − 1.52 to − 6.10 %) for all THs. The relative recoveries (%) ranged between 99.1 and 110.1% with excellent precision (RSD: 2.91-12.0 %). These low uncertainties obtained with the HybridSPE® protocol were partially attributed to the rapidity of the specific sample preparation protocol that does not impact the stability of THs. The method inter-day precision at the fortification concentration of 1 ng/mL was < 25 %, when calculated based on the external standard method, denoting acceptable chromatographic peak stabilities for all target THs even at traces. The method was applied successfully in 46 blood plasma samples from Glaucous gulls and Baikal seals, while cross-array comparison between LC-MS/MS and RIA was performed for T 4 and T 3 , which were those two THs detected in the majority of samples; all results were found in agreement to previous works demonstrating the highly selective and sensitive valid analysis of THs through the HybridSPE® technique. In the analyzed blood plasma samples, rT 3 demonstrated low detection rate, while T 2 was not detected. Nonetheless, the HybridSPE® technique allows for further in-depth studies on the presence and role of all four THs in wildlife species.  Table 4 Total concentrations of T 4 , T 3 and rT 3 reported in Baikal seals.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.