Simultaneous determination of methyltestosterone and its metabolite in fish by gas chromatography-mass spectrometry

Abstract

Methyltestosterone is one of the banned drugs in aquaculture, and it should be monitored in food-producing animals. 17α-Methyl-5β-androstane-3α,17β-diol, as the main metabolite of methyltestosterone in vertebrates, could be used as another marker for controlling the administration of methyltestosterone, due to its high residual concentration and slow elimination rate. In this study, an analytical method based on gas chromatography-mass spectrometry (GC-MS) was developed and validated for the simultaneous determination of methyltestosterone and its main metabolite in fish. After pretreatment by liquid–liquid extraction with n-hexane and solid phase extraction with C₁₈ and NH₂ columns, the target analytes in the muscle tissues were extracted and concentrated, and the influence of the sample matrix was eliminated. Then, the prepared samples were separated and detected with GC-MS in the selected ion monitoring (SIM) mode. Methyltestosterone-D₃ was chosen as the internal standard for quantitation. After optimization, the limits of detection for methyltestosterone and 17α-methyl-5β-androstane-3α,17β-diol were 20 μg kg⁻¹ and 15 μg kg⁻¹, respectively. The limits of quantitation were both 50 μg kg⁻¹. The calibration curves showed good linearity in the concentration range from 50.0 ng mL⁻¹ to 500.0 ng mL⁻¹. The correlation coefficients of methyltestosterone and 17α-methyl-5β-androstane-3α,17β-diol were more than 0.9990. The recoveries of the analytes in real samples were in the range of 99.7–116.6% with the relative standard deviation of 5.2–8.3%. The established method could meet the demand for simultaneous detection of methyltestosterone and its major metabolite, and it could be used to provide more information on the abuse of methyltestosterone in food-producing animals.

---

Introduction

Sex determination and differentiation of fish is a matter of great interest in aquaculture as sex control brings great commercial value.¹ For example, female Atlantic salmons always exhibit deterioration in flesh quality due to the ability of diversify energy towards the reproductive process,² and male Nile tilapias show better growth with a lower feed conversion rate than females.³ Researchers have found that fish sex is determined by the male-promoting dmrt1/amh/gsdf/androgen pathway and the female-promoting foxl2/cyp19a1a/estrogen pathway. These two pathways work together and antagonize each other.^4,5 When genes like cyp19a1a and foxl2 mutate and forfeit the original physiological function, the level of estrogen will decrease and the level of androgen will increase, inducing female-to-male sex reversal.^6,7

Methyltestosterone (MT), classified as an androgen and anabolic steroid (AAS), was initially identified in 1935 and subsequently introduced for medical applications in 1936.⁸ MT has also been employed in animal administration to induce changes in sex differentiation or stimulate growth.^9,10 For example, studies have indicated that MT possesses the capability to inhibit the expression of a crucial steroidogenic enzyme (P450arom), thereby resulting in the masculinization of Nile tilapia.¹¹ The levels of dmrt1 and foxl2 in the Orange-spotted grouper, a marine ray-finned fish species, exhibited significant alterations, resulting in evident masculinization traits upon administration of MT.¹² Nonetheless, it is imperative to acknowledge the potential adverse consequences associated with the misuse of MT in aquaculture.¹³ Firstly, the potential consequences of MT include the induction of deformities in food-producing animals and the difficulty for farmers to effectively manage sexual differentiation.^14,15 Secondly, the leakage and migration of MT from breeding sites can lead to its sorption onto soils and sediments.^16,17 Finally, the consumption of residual MT by food-producing animals poses significant human health risks, including the potential for thrombotic and carcinogenic effects.^18–20 Therefore, numerous nations have implemented restrictions on the utilization of MT in the domains of animal husbandry and aquaculture, except for its application for therapeutic intentions. In China, MT is classified as one of the proscribed substances for all food-producing animals during the breeding process, aiming to guarantee the safety of consumable animal-derived food items.

Up to now, not only the chemical additives themselves but also their metabolites are getting more and more attention due to their potential to prove the contamination or safety of food. For instance, both methylated and hydroxylated metabolites of phenylarsenic chemicals were considered as the promising target chemicals for detecting the contamination of some chemical warfare agents in fish.²¹ For robenidine hydrochloride (ROBH) which was an authorized feed additive for the prevention of coccidiosis in animal species, 4-chlorobenzoic acid (PBCA) was its predominant metabolite in the grass carp, and the elimination half-life of PBCA was 8 times that of ROBH. Thus, the detection results of PBCA could serve as reference data for the safety and reasonable use of ROBH in fish.²² Victoria Pure Blue BO (VPBO), a prohibited drug in aquaculture, was metabolized into deethyl-leuco-VPBO with longer residue time in rainbow trout, which could be used as a new residue marker for predicting the VPBO drug treatment of aquacultured species.²³

For MT, seven metabolites have been found in vertebrates, arising from hydroxylation, reduction, isomerization, and rearrangement.^24–27 17α-Methyl-5β-androstane-3α,17β-diol (M2) is one of its most prominent metabolites, owning high concentration and long half-life.²⁸ In the urine of the heifer, the concentration of M2 is 19 times that of MT even after 7 days of administration.²⁹ According to the pharmacokinetic research of MT and M2 in Nile tilapia, the elimination of M2 is extremely slow, and its residue concentration is high. The peak concentration (C_max) and elimination half-life (T_1/2b) of M2 were 5.604 μg mL⁻¹ and 362.7 h, while C_max and T_1/2b of MT were 1.032 μg mL⁻¹ and 18.4 h, respectively.³⁰ Therefore, in order to strengthen the supervision and inspection of the use of MT in food-producing animals, it is very necessary to simultaneously detect MT and its metabolites.

In the field of food safety, more attention has been paid to the detection of methyl-testosterone monomers in meat products and dairy products. Several methods have been developed to determine MT in food-producing animals.^31,32 Among them, mass spectrometry (MS) is the most effective and efficient detection technique because of its high sensitivity and resolution. In addition, the sample pretreatment and separation technology are also needed for the qualitative and quantitative analysis due to the complexity of the biological sample matrix. Regal et al. extracted MT from cow hair with acetonitrile, and then the determination was carried out by high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) in multiple reaction monitoring (MRM) mode after the hydroxylamine derivatization.³³ MT in muscle tissues could be directly determined by HPLC-MS/MS after the sample pretreatment of liquid–liquid extraction (LLE) and solid phase extraction (SPE).³⁴ In terms of MT metabolites, two main compounds including M2 were measured by stable isotope analysis based on gas chromatography-mass spectrometry (GC-MS), in which M2 was derivatized with N-methyl-N-trimethylsilyl-trifluoroacetamide/dithioerythriol/ammonium iodide first.³⁵

To the best of our knowledge, there was no method available for simultaneous qualitation and quantitation of MT and its main metabolites in food-producing animals. In this present research, a GC-MS method without any derivatization was established and validated for the simultaneous analysis of residual MT and M2 in fish muscle. LLE with n-hexane and SPE with C₁₈ and NH₂ columns were used to eliminate the influence of the sample matrix and concentrate the target compounds. The established method would provide more information to aid in further monitoring the administration of MT in food-producing animals.

Experimental

Reagents and materials

Methanol and acetonitrile were acquired from TEDIA (HPLC, Fairfield, USA). Acetone and n-hexane were purchased from Aladdin (HPLC, Shanghai, China). Methyltestosterone-D₃ (D-MT, 100 μg mL⁻¹), M2 (100 μg mL⁻¹), and β-glucuronidase/arylsulfatase (≥85 000 units per mL, ≤7500 units per mL) were purchased from Sigma-Aldrich (St. Louis, MO. USA). Methyltestosterone (MT, 98.3%) was purchased from Beijing Manhage Bio-tech Co., Ltd (Beijing, China). Diatomite (98%) was obtained from Macklin Biochemical Co., Ltd (Shanghai, China). C₁₈ solid-phase extraction columns and NH₂ solid-phase extraction columns were obtained from Agela Technologies (Tianjin, China).

Preparation of standard solutions

The stock solution of MT was prepared at the concentration of 2.00 mg mL⁻¹ with methanol. Then, it was further diluted by methanol to the concentration of 10.0 μg mL⁻¹. The stock solutions of internal standard and M2 were prepared by diluting D-MT and M2 with methanol to 10.0 μg mL⁻¹, respectively. For plotting the standard curve, the standard solutions were prepared at the concentrations of 50.0, 100.0, 200.0, 400.0, and 500.0 ng mL⁻¹ by diluting the appropriate volume of the stock solutions with methanol. These solutions were all stored at −20 °C.

Pretreatment of samples

The homogenized fish muscle (1.0 ± 0.1 g) was weighed into the mortar. Each sample was added with the stock solution of D-MT (5 μL) and diatomite (1.0 g), then the mixture was thoroughly ground to disperse the tissue efficiently. The sample was mixed with acetonitrile (5 mL) and β-glucuronidase/arylsulfatase (60 μL) by vortex and then incubated at 37 °C for 8 h.³⁹ After that, the sample was centrifuged at 4000 rpm for 5 min at 4 °C, and the supernatant was collected. The above extraction step with acetonitrile was repeated twice. The supernatant was gathered and distilled to remove the solvent by rotary evaporator. The sample was redissolved with acetonitrile-saturated n-hexane (1 mL) and then centrifuged at 10 000 rpm for 5 min at 4 °C. The underlying solution was transferred into a centrifuge tube with 10 mL of water.

For clean-up by SPE, C₁₈ columns were successively activated with methanol and water, and NH₂ columns were activated with acetonitrile. The sample was loaded on the C₁₈ column first, and rinsed with water, acetone-aqueous solution (1 : 4, v/v), and n-hexane, respectively. After that, the C₁₈ SPE column was connected in series with the NH₂ column, and the solid-phase extraction system was eluted with acetonitrile (6 mL). In this process, the flow rate was controlled within 3 mL min⁻¹. The eluate was collected and dried by a rotary evaporator. Then, methanol (1 mL) and anhydrous sodium sulphate were sequentially added and mixed. Finally, the supernatant was filtered by a membrane filter (0.22 μm) and applied for GC-MS analysis.

GC-MS instrument and conditions

The experiment was carried out on a QP2020 gas chromatography-mass spectrometer (Shimadzu, Kyoto, Japan). A fused-silica capillary column Shimadzu SH-Rxi-5MS (30 m × 0.25 mm × 0.25 μm) was used for separation. The injection volume was 1 μL, and the injection mode was splitless. The injector temperature was 280 °C. Helium gas (99.999%) was used as the carrier gas at a constant flow rate of 1.0 mL min⁻¹. After optimization of separation, the initial temperature of the column was set at 170 °C for 2 min. It was increased at 16 °C min⁻¹ to 250 °C, then increased at 5 °C min⁻¹ to 280 °C, and maintained for 2 min. For MS detection, the electron impact ion source was operated with an ionization energy of 70 eV at 230 °C. Selected ion monitoring (SIM) mode was performed for the quantitation of analytes.

Results and discussion

Selection of the targets for analysis

We used the software, named XenoSites and ADMET predictor, to predict all the potential metabolites and their toxicity of MT. As shown in Table 1, the reduction reaction products may be the main metabolites. The metabolite M2, as the primary reduction metabolite, has the most significant cardiotoxicity and acute toxicity. Then, we used the ADMET predictor and Chemprop to predict the half-life in vivo and liver microsomal clearance of MT and M2, respectively. The half-life period of MT is 7.4 h, and its microsomal clearance is 161.2 μL min⁻¹ × 10⁶ hepatocytes. The half-life period of M2 is 8.2 h, and its microsomal clearance is 49.8 μL min⁻¹ × 10⁶ hepatocytes. It means that M2 has a longer retention time and lower clearance rate. Thus, M2 was chosen as another marker to monitor the use of MT in this work.

Table 1 Prediction of methyltestosterone metabolites and toxicity analysis

Metabolites	Probability	Molecule	hERG toxicity (%)	Acute toxicity (LD50 mg kg^−1)	Mutagenicity toxicity (%)
Reduction metabolite	High	CC12CCC(=O)C=C1CCC3C2CCC4(C3CCC4(C)O)C (MT)	0.03	2363	0.03
		CC12C(C)(CCC1C(CCC3C4(C)CCC(O)C3)C4CC2)O (M2)	0.25	2386	0.03
		CC12C(C)(CCC1C(CCC(C3)C4(C)CCC3=O)C4CC2)O	0.25	1625	0.03
		CC12C(C)(CCC1C(CCC3=CC(CCC43C)O)C4CC2)O	0.21	686	0.08
Dehydrogenation metabolite	Medium	CC12C(C)(CCC1C(CCC3=CC(CCC43C)=O)C4CC2)[O–]	0.42	1031	0.03
		CC12C(C)(CCC1C([CH–]CC3=CC(CCC43C)=O)C4CC2)O	0.45	1480	0.03
		CC12C(C)(CCC1C(C[CH–]C3=CC(CCC43C)=O)C4CC2)O	0.42	1369	0.03
		CC12C(C)(CCC1C(CCC3=CC(C[CH–]C43C)=O)C4CC2)O	0.40	1625	0.03
		CC12C(C)(CCC1C(CCC3=CC([CH–]CC43C)=O)C4CC2)O	0.42	472	0.03
		CC12C(C)(CCC1C([CH–][CH–]C3=CC([CH–][CH–]C43C)=O)C4CC2)[O–]	0.40	1625	0.03
Oxygenation metabolite	Medium	CC12C(O)(CCC1C(CCC3=CC(CCC43C)=O)C4CC2)O	0.03	1813	0.03
		CC12C(C)(C(O)CC1C(CCC3=CC(CCC43C)=O)C4CC2)O	0.03	2281	0.03
		CC12C(C)(CC(O)C1C(CCC3=CC(CCC43C)=O)C4CC2)O	0.03	1464	0.03
		CC12C(C)(CCC1C(C(O)CC3=CC(CCC43C)=O)C4CC2)O	0.03	2262	0.03
		CC12C(C)(CCC1C(CC(O)C3=CC(CCC43C)=O)C4CC2)O	0.03	1031	0.03
		CC12C(C)(CCC1C(CCC3=CC(CC(O)C43C)=O)C4CC2)O	0.03	2262	0.03
		CC12C(C)(CCC1C(CCC3=CC(C(O)CC43C)=O)C4CC2)O	0.03	1369	0.03
		OC12C(O)(C(O)C(O)C1C(C(O)C(O)C3=CC(C(O)C(O)C43O)=O)C4C(O)C2O)O	0.03	1542	0.03

---

Selection of the sample pretreatment method

As is well known, fish is a complex biological sample. Thus, the determination of residues in fish always requires sample pretreatment to meet the analysis requirements, such as extraction, purification, and concentration. Meanwhile, the internal standard also needs to be added before pretreatment to improve the accuracy of quantification. Here, D-MT was chosen as the internal standard. As shown in Fig. 1, there is only a slight difference in the chemical structure of MT, M2, and D-MT. Thus, they possess similar physicochemical properties. We referred to the sample pretreatment procedure in the national standard method which specifies the methods of sample preparation, determination and confirmation by LC-MS/MS of 7 androgen residues in foodstuffs of animal origin and validated it.³⁶

Fig. 1 Chemical structures of MT, M2, and D-MT.

The procedure of sample pretreatment is divided into two steps, including LLE and SPE. In the first step, the sample was digested by β-glucuronidase/arylsulfatase due to their ability to degrade extracellular matrix and reduce the sulfuric acid conjugates of MT into natural androgens. Then, the analytes were extracted by acetonitrile three times. Of course, some impurities are inevitably co-extracted. It should be noted that it is necessary to remove some of the lipid impurities before the clean-up of the sample through SPE, as the excess lipid will bind to the alkyl chain in the C₁₈ SPE column, preventing the retention of MT, D-MT, and M2 on it. As we know, n-hexane is always used for removing lipids. However, n-hexane may not have the ability to eliminate sufficient lipids from the sample when the sample is rich in lipids. Repeated use of n-hexane may result in loss of the analytes. Therefore, after being extracted by acetonitrile and concentrated by rotary evaporation, the sample was re-dissolved with acetonitrile-saturated n-hexane, and stored at 4 °C for at least 6 h. It will coagulate the fat in the solution into a white solid, which is easy to be removed. In the second step of SPE, the C₁₈ column was combined with the NH₂ column to remove the liposoluble impurities with weaker polarity than MT and M2, and the strong polarity impurities. To evaluate the efficiency of the SPE procedure, the standard solutions of MT, M2, and D-MT dissolved in acetonitrile and blank matrix were directly loaded on SPE columns for test, and their recovery rates were analyzed in triplicate. The obtained data were all in the range of 93.2–112.6%, which met the requirement of sample pretreatment. It means that the proposed SPE procedure will not be affected by the presence of solvent and impurities in the sample.

GC-MS analysis programs

Temperature-programmed GC was employed for the separation due to the complicated matrix of fish muscle. Meanwhile, the full scan mode of MS was used at first for the identification of the targets and obtaining their specific mass-to-charge ratio (m/z) for SIM mode. After the optimization of GC-MS conditions, the retention time, qualitative and quantitative ions of the analytes were summarized in Table 2.

Table 2 The retention time, qualitative and quantitative ions of MT, M2, and D-MT

Analyte	Retention time (min)	Quantitative ion (m/z)	Qualitative ion 1 (m/z)	Qualitative ion 2 (m/z)
M2	11.11	230	235	135
D-MT	12.63	305	306	287
MT	12.66	302	303	284

---

Specificity

Specificity refers to the ability to unambiguously evaluate an analyte in the presence of possible components, such as impurity and matrix. As shown in Fig. 2, M2 was first detected with GC-MS analysis in SIM mode. D-MT and MT were nearly co-eluted. However, it does not affect their qualitative and quantitative analysis because of their different m/z. No other peaks with adjacent retention times of the targets were observed in the fish samples. All of the results indicated good specificity of this method.

Fig. 2 SIM profiles of (A) standard solution, (B) fish sample spiked with MT, M2, and D-MT.

Limit of detection (LOD) and limit of quantitation (LOQ)

LOD and LOQ of each analyte were calculated based on the signal-to-noise ratio (S/N) of 3 : 1 and 10 : 1, respectively. Given that the real sample might exist the matrix effect, LOQ was also validated with the sample matrix. LODs of MT and M2 were determined to be 20 μg kg⁻¹ and 15 μg kg⁻¹, respectively. LOQs of MT and M2 were both 50 μg kg⁻¹.

Linearity

The linearity of an analytical method refers to its ability to obtain test results proportional to the concentration or amount of analyte in the sample. In this work, D-MT was used as the internal standard to eliminate the potential interference and provide the most robust compensation for the impacts from the matrix effect, sample pretreatment, and measuring instruments. The ratio of the peak area of MT and M2 to D-MT was used for the calibration and calculation of samples. The calibration curves of MT and M2 were exhibited in Fig. 3. Within the concentration range of 50.0 to 500.0 ng mL⁻¹, the correlation coefficients of MT and M2 were 0.9992 and 0.9996, respectively, indicating good linear relationships.

Fig. 3 Calibration curve of MT and M2 (n = 3).

Accuracy and precision

To further demonstrate the feasibility of the proposed analytical methodology, the recoveries of MT and M2 in fish muscle were measured at three spiked levels. The average recovery and relative standard deviation (RSD) were used to evaluate the accuracy and precision. The results in Table 3 showed the recoveries were in the range of 99.7–116.0% for MT, and 105.3–116.6% for M2. The RSD values were all below 8.3%.

Table 3 The average recovery rates and RSD values of MT and M2 in fish (n = 6)

Spiked concentration (μg kg^−1)	Average recovery (%)		RSD (%)
	MT	M2	MT	M2
50.0	104.4	105.3	5.2	6.8
150.0	99.7	116.6	8.2	8.3
400.0	116.0	115.7	6.3	5.6

---

Matrix effect

The matrix effect of MS may severely impair sensitivity and accuracy of the method. In this study, we evaluated the potential matrix effect by comparing the MS response with and without the presence of blank sample matrix. Six different fish species (Basa, Mackerel, Capelin, Salmon, Cod, and Flounder) were chosen for the purpose of examining. The analytes were spiked at two levels of concentration. To investigate the potential ion-suppression or ion-enhancement effect, matrix factors (MF) of each target were calculated. The results in Table 4 showed that the average MF of MT and M2 was in the range of 0.95–1.01, and the RSD values were all below 6.0%, suggesting little matrix effect.

Table 4 The matrix effect of GC-MS (n = 6)

Spiked concentration (μg kg^−1)	Matrix factor (MF)		RSD (%)
	MT	M2	MT	M2
100.0	0.99	1.01	6.0	3.5
300.0	1.00	0.95	3.4	4.1

---

Comparison with other methods

In Table 5, we summarized the instrument, pretreatment, sensitivity, repeatability, and accuracy in this work together with those obtained in other literatures. Recent research findings indicate that LLE exhibits superior recovery rates compared to SPE, refrigeration, and dispersed solid phase methods for the extraction of methyltestosterone from biological samples.^37,39 Nevertheless, it is worth noting that LLE being used alone tends to introduce more impurities and result in a heightened interference signal.³⁸ Consequently, we propose the integration of LLE with SPE in this work to enhance recovery rate and eliminate the background interference. In addition, MT and its main metabolite are simultaneously determined by GC-MS rather than LC-MS/MS. It means we offer another robust method for their qualitative and quantitative analysis in complex biological samples. Although the sensitivity of the proposed method is not good enough, it is mainly due to the scanning mode of MS being SIM. We believe that it could be improved by using GC-MS/MS for the determination.

Table 5 Comparison of the present method with other methods for the determination in animal samples

Method	Pretreatment	Sample	Target	Recovery (%)	LOD (μg kg^−1)	RSD (%)	Ref.
LC-MS/MS	LLE	Fish, shrimp, crab, bullfrog and mussel	MT	86.4–104.6	0.4	0.45–7.31	37
LC-MS/MS	LLE	Fish and shrimp	MT	73–115	2	2.1–7.0	39
LC-MS/MS	LLE	Fish and shrimp	MT	96–105	0.4	96–105	32
LC-MS/MS	QuEChERS	Pork, chicken and beef	MT	80–96	0.4	1–5	40
GC-MS	LLE, SPE	Fish	MT, M2	99.7–116.6	MT: 20, MT: 15	5.2–8.3	Present method

---

Application to real samples

The developed method was applied to test MT and M2 in crucian carp and silver carp bought from the supermarket. Although there were no residues observed in the real samples, the method could still have significant application for the determination of MT and M2 in food of animal origin.

Conclusions

MT, as a banned drug for food-producing animals, is needed to be strictly monitored and supervised. Although more and more methods for the detection of methyltestosterone have been developed, the research and application of its metabolites in the field of food safety have not been paid much attention. In this research, the main metabolite of MT was chosen as another marker for controlling the abuse of MT in the field of food analysis. A simple GC-MS method coupled with traditional sample pretreatment was established and validated for the determination of MT and M2 in fish. Compared with previous work, MT and its metabolite were simultaneously detected for the first time, and there was no need for any derivatization before the analysis of GC-MS. The established method was sensitive, accurate, and precise. The LOQs of MT and M2 were both 50 μg kg⁻¹. The recoveries of real samples were all in the range of 99.7–116.6% with the RSDs between 5.2% and 8.3%. Obviously, the analysis of methyltestosterone and its metabolites in food has some feasibility, which is something we want to emphasize and present repeatedly in this work. It will help to provide more information on the safety of food of animal origin.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was financially supported by the National Key Research and Development Program of China (Grant No. 2019YFC1605400), and the Natural Science Foundation of Jiangsu Province (Grant No. BK20200567).

References

1. R. H. Devlin and Y. Nagahama, Aquaculture, 2002, 208, 191–364.
2. F. Piferrer, A. Beaumont, J. C. Falguiere, M. Flajshans, P. Haffray and L. Colombo, Aquaculture, 2009, 293, 125–156.
3. M. T. Ridha, International Aquatic Research, Islamic Azad University, Tonekabon Branch, 2011, vol. 3.
4. X. Zhang, M. Li, H. Ma, X. Liu, H. Shi, M. Li and D. Wang, Endocrinology, 2017, 158, 2634–2647.
5. M. H. Li, H. H. Yang, M. R. Li, Y. L. Sun, X. L. Jiang, Q. P. Xie, T. R. Wang, H. J. Shi, L. N. Sun, L. Y. Zhou and D. S. Wang, Endocrinology, 2013, 154, 4814–4825.
6. Q. Zhang, D. Ye, H. Wang, Y. Wang, W. Hu and Y. Sun, Endocrinology, 2020, 161.
7. D. N. Jiang, H. H. Yang, M. H. Li, H. J. Shi, X. B. Zhang and D. S. Wang, Mol. Reprod. Dev., 2016, 83, 497–508.
8. N. T. Shahidi, Clin. Therapeut., 2001, 23, 1355–1390.
9. R. T. Sparks, B. S. Shepherd, B. Ron, N. H. Richman III, L. G. Riley, G. K. Iwama, T. Hirano and E. G. Grau, Comp. Biochem. Physiol., B, 2003, 136, 657–665.
10. M. Nagaraju and G. S. Devi, Ann. Romanian Soc. Cell Biol., 2021, 25, 20608–20616.
11. R. K. Bhandari, M. Nakamura, T. Kobayashi and Y. Nagahama, Gen. Comp. Endocrinol., 2006, 145, 20–24.
12. Q. Lyu, J. Hu, X. Yang, X. Liu, Y. Chen, L. Xiao, Y. Liu, Q. Wang, J. Chen, M. Huang, Z. Yu, H. Yang, H. Shi, Y. Zhang and H. Zhao, Gen. Comp. Endocrinol., 2019, 274, 26–36.
13. N. Mlalila, C. Mahika, L. Kalombo, H. Swai and A. Hilonga, Environ. Sci. Pollut. Res., 2015, 22, 4922–4931.
14. A. M. Budd, Q. Q. Banh, J. A. Domingos and D. R. Jerry, J. Mar. Sci. Eng., 2015, 3, 329–355.
15. J. A. Beardmore, G. C. Mair and R. Lewis, Reproductive Biotechnology in Finfish Aquaculture, 2001, 283–301.
16. S. K. Ong, P. Chotisukarn and T. Limpiyakorn, Water, Air, Soil Pollut., 2012, 223, 3869–3875.
17. S. Biswas, C. Shapiro, W. Kranz, T. Mader, D. Shelton, D. D. Snow, S. L. Bartelt-Hunt, D. Tarkalson, S. Van Donk and T. C. Zhang, J. Soil Water Conserv., 2013, 68, 325–336.
18. I. Thiblin, H. Garmo, M. Garle, L. Holmberg, L. Byberg, K. Michaelsson and R. Gedeborg, Drug Alcohol Depend., 2015, 152, 87–92.
19. S. Chang, A. B. Munster, J. Gram and J. J. Sidelmann, Semin. Thromb. Hemost., 2018, 44, 734–746.
20. C. L. G. Rivero-Wendt, A. L. Miranda-Vilela, I. Domingues, R. Oliveira, M. S. Monteiro, M. A. M. Moura-Mello, R. Matias, A. Soares and C. K. Grisolia, J. Environ. Sci. Health, Part A, 2020, 55, 1321–1332.
21. H. Niemikoski, D. Koske, U. Kammann, T. Lang and P. Vanninen, J. Hazard. Mater., 2020, 391, 122221.
22. Y. Liu, Y. Song, B. Cheng, J. Dong, N. Xu, S. Zhou, Q. Yang and X. Ai, Food Anal. Methods, 2020, 13, 516–529.
23. E. Dubreil, M. Laurentie, J. M. Delmas, M. Danion, T. Morin, D. Hurtaud-Pessel, A. Viel, P. Sanders and E. Verdon, Chemosphere, 2021, 262, 127636.
24. W. Schänzer, G. Opfermann and M. Donike, Steroids, 1992, 57, 537–550.
25. O. J. Pozo, P. Van Eenoo, K. Deventer, L. Lootens, W. Van Thuyne, M. K. Parr, W. Schanzer, J. V. Sancho, F. Hernandez, P. Meuleman, G. Leroux-Roels and F. T. Delbeke, Drug Metab. Dispos., 2009, 37, 2153–2162.
26. W. Schanzer and M. Donike, Anal. Chim. Acta, 1993, 275, 23–48.
27. D. Martinez-Brito, X. de la Torre, M. K. Parr and F. Botre, Rapid Commun. Mass Spectrom., 2020, 34, e8834.
28. D. Martinez-Brito, M. Iannone, M. A. Tatangelo, F. Molaioni, X. de la Torre and F. Botre, Rapid Commun. Mass Spectrom., 2020, 34, e8870.
29. M. Blokland, H. Van Rossum, H. Herbold, S. Sterk, R. Stephany and L. Van Ginkel, Anal. Chim. Acta, 2005, 529, 317–323.
30. E. Hong, G. Zheng, D. Meng, S. Wu, B. Ma, L. Ma and Y. Song, J. Anhui Agric. Univ., 2011, 38, 911–915.
31. X. Li, T. Yuan, T. Zhao, X. Wu and Y. Yang, J. Chromatogr. Sci., 2020, 58, 880–886.
32. J. M. Storey, S. B. Clark, A. S. Johnson, W. C. Andersen, S. B. Turnipseed, J. J. Lohne, R. J. Burger, P. R. Ayres, J. R. Carr and M. R. Madson, J. Chromatogr. B, 2014, 972, 38–47.
33. P. Regal, C. Nebot, B. I. Vazquez, A. Cepeda and C. A. Fente, Meat Sci., 2010, 84, 196–201.
34. P. S. Chu, M. Lopez, S. Serfling, C. Gieseker and R. Reimschuessel, J. Agric. Food Chem., 2006, 54, 3193–3198.
35. Y. Shinohara, K. Isurugi and T. Hashimoto, J. Chromatogr. B Biomed. Sci. Appl., 2000, 741, 271–278.
36. L. M. Lin, L. Z. Chen, M. Du, Q. H. Meng, Z. F. Yue, T. Peng, S. P. Pan and D. Y. Ba, Chinese National Standard, 2010, SN/T 2677-2010.
37. X. Zhang, C. Fang, X. Lou, G. Yang, C. Kong, Y. Shi and D. Huang, Microchem. J., 2022, 183, 108119.
38. M. P. Juhascik and A. J. Jenkins, J. Chromatogr. Sci., 2009, 47, 553–557.
39. W. Zheng, K. H. Yoo, J. M. Choi, D. H. Park, S. K. Kim, Y. S. Kang, A. M. Abd El-Aty, A. Hacimuftuoglu, J. Wang, J. H. Shim and H. C. Shin, Biomed. Chromatogr., 2019, 33, e4396.
40. M. Lopez-Garcia, R. Romero-Gonzalez and A. G. Frenich, J. Chromatogr. A, 2018, 1540, 21–30.

---

Footnote

† The first two authors contributed equally to this work.

---