Determination of Gentamicin: Development and Validation of a Sensitive UPLC-MS/MS Assay According to the European Medicines Agency Guideline

Abstract

Background/Objectives: Gentamicin (GEN) is an aminoglycoside antibiotic used in veterinary medicine to treat infections caused mainly by Gram-negative bacteria. GEN is a mixture of pharmacologically active components, known as isoforms. The objective was to develop and validate a sensitive, accurate, and precise Ultra-Performance Liquid Chromatography with triple quadrupole mass detector (UPLC-MS/MS) method to quantify the different GEN isoforms in pig plasma and feces using streptomycin as an internal standard. Methods: Solid-phase extraction (SPE) was carried out. A high-strength silica (50 × 2.1 mm, 1.8 µm) column was used for chromatographic separation and a mobile phase of 0.26% HFBA in water (A) and acetonitrile (B) was delivered in a gradient with a flow rate of 0.5 mL/min. The column temperature was 40 °C and the sample injection volume was 30 µL. Results: The method showed good selectivity and specificity, with no interfering peaks. Calibration curves were linear in the range from 0.05 to 0.3 µg/mL for all isoforms in both matrices. Within- and between-run precision and accuracy were satisfactory for the lower limit of quantification (LLOQ), with coefficients of variation (CV) ≤ 13.4% and deviations ≤ 116.5% in plasma and CV ≤ 12.3% with deviations ≤ 101.7% in feces. No carry-over was observed, and analyte stability was confirmed under different storage conditions. Conclusions: The method development fulfilled all validation criteria established by the European Medicine Agency Guideline (EMA/CHMP/ICH/172948/2019). Moreover, the applicability of the method in clinical practice was demonstrated by the quantification of GEN in plasma and feces samples from pigs.

1. Introduction

Aminoglycoside antibiotics are used in veterinary medicine for the treatment of animals with infections caused primarily by Gram-negative bacteria and, less frequently, Gram-positive bacteria. With respect to their origin, they can be classified into two groups: natural, such as neomycin, streptomycin (STR), or gentamicin (GEN), and semi-synthetic, like amikacin [1,2,3]. Their mechanism of action involves the binding of the drug to components of bacterial cell membranes, resulting in the displacement of divalent cations and increased membrane permeability. This allows the entry and accumulation of aminoglycosides into the cell and the subsequent inhibition of bacterial protein synthesis by binding to 16S rRNA on the 30S subunit of ribosomes [4,5,6].

GEN obtained from Micromonospora purpurea is used in veterinary medicine to treat respiratory, genitourinary, and gastrointestinal infections and also bone, joint, and skin infections caused by sensitive bacteria in different animal species [2]. In pigs, it is specifically indicated for digestive infections caused by Escherichia coli and for the prevention and treatment of swine dysentery caused by Brachyspira hyodysenteriae. Swine dysentery is a severe mucohemorrhagic enteric disease that has a large impact on pig production and causes important losses due to mortality and suboptimal performance. Therefore, it is of great importance to have adequate and effective therapeutic tools [1,4,7,8]. Currently, no commercial vaccine is available, so its control is based on the use of a limited number of effective antimicrobials, with pleuromutilins, macrolides, and lincosamides being the most commonly used. In recent years, the treatment of swine dysentery has been hampered by an increase in antimicrobial resistance to the mentioned antibiotics, as reported in different countries with high pig production [9,10], proposing GEN as a good therapeutic tool for the treatment and prevention of Brachyspira diseases in these animals.

GEN is a mixture of pharmacologically active components, known as isoforms. These isoforms differ slightly in their chemical structure [primarily in the degree of methylation of the amino group at the 6 carbon of the amino-hexose (purpurosamine) ring]. Each one contributes in a different way to the overall antibacterial activity and pharmacokinetic profile of the drug [6,11,12,13]. The relative proportions of these isoforms can influence both the therapeutic efficacy and toxicity of the formulation [12]. Given these differences, it becomes essential to accurately identify and quantify each gentamicin isoform in pharmaceutical formulations, especially in veterinary preparations used in food-producing animals. This not only ensures therapeutic consistency but also aligns with regulatory requirements for quality control.

The quantification of GEN in plasma and feces is essential when this aminoglycoside is used for the treatment of swine dysentery caused by Brachyspira hyodysenteriae. Measuring plasma concentrations allows for the assessment of systemic absorption. Also, the determination of its concentration in feces is crucial to confirm that therapeutic levels are maintained at the site of infection (intestinal lumen). This is especially relevant for enteric disease, where the antimicrobial effect must be exerted locally. Moreover, fecal quantification provides insight into the drug’s persistence in the gastrointestinal tract and its potential environmental excretion, with implications for gut microbiota disruption and the development of antimicrobial resistance. In this context, Therapeutic Drug Monitoring (TDM) becomes particularly relevant, as it enables the evaluation of whether drug concentrations achieved in biological matrices are appropriate to ensure efficacy while avoiding toxicity [14,15]. Applying TDM principles to gentamicin therapy in pigs allows for a more accurate interpretation of plasma and fecal levels, ensuring that sufficient concentrations reach the intestinal lumen to treat swine dysentery. Incorporating TDM into the analytical and clinical assessment reinforces the need for precise quantification methods such as the one proposed in this study, thereby improving dosage optimization and contributing to responsible antimicrobial use. Therefore, simultaneous determination of GEN in both matrices is key for optimizing therapeutic efficacy, minimizing adverse effects, and promoting the responsible use of antimicrobials in veterinary medicine.

Several analytical methods have been developed for the determination and quantification of gentamicin in plasma, urine, and tissues (muscle, liver, and kidney) in different animal species, such as rabbits [16], mice [17], cattle [3,18,19,20], goats [21], horses [11], dogs [12], and pigs [6,18,22,23]. These studies have mainly employed chromatographic techniques, including HPLC with fluorescence detection (HPLC-FLD) [22], HPLC with diode array detection (HPLC-DAD) [12], HPLC with UV detection (HPLC-UV) [11], as well as HPLC coupled to mass spectrometry (HPLC/MS) [6,17,18] and HPLC coupled to tandem mass spectrometry (HPLC/MS/MS) [3,19,20,21]. Regarding feces, analytical methods have been reported for the determination of several antibiotics in pig feces. Specifically, methods have been developed for the determination of fluoroquinolones (HPLC-FLD [24]); HPLC-DAD [25]), sulfonamides (LC-MS/MS [26]; HPLC-FLD [27]), tetracyclines (LC-MS/MS [28]), and cephalosporins (HPLC-MS/MS [29]). However, despite its relevance, no validated method specifically for the determination of gentamicin in pig feces was found in the literature.

Thus, the objective of this study was to develop and validate a sensitive Ultra-High-Performance Liquid Chromatography coupled to tandem mass spectrometry (UPLC-MS/MS) assay for the determination and quantification of GEN isoforms in pig plasma and feces using an appropriate extraction procedure.

2. Results and Discussion

Under the UPLC-MS/MS conditions described in the Section 3, the retention time of GEN in plasma was 3.30 min for isoform C1 and 3.21 min for C2C2a, while in feces samples, it was 3.16 min for isoform C1 and 3.10 min for C2C2a. The retention time for the internal standard (IS) (STR) in plasma and feces was 1.70 min (Figure 1C and Figure 2C).

2.1. Selectivity

After having tested blank plasma and black feces from six different individual sources, no interfering peaks from endogenous components were detected at the retention time of the different isoforms of GEN and IS, as shown in Figure 1 (plasma chromatogram) and Figure 2 (feces chromatogram). Therefore, the method used meets the selectivity criterion indicated by the European validation guideline [30].

2.2. Specificity

No significant interference at the retention times of GEN C1 isoform and GEN C2C2a isoform was found in the chromatograms of plasma and feces blank samples, with no signal exceeding 20% of the analyte response at the LLOQ and 5% of the IS response in the LLOQ sample, as requested in the EMA guideline [30].

2.3. Matrix Effect

To assess the matrix effect, three replicates of QC2 (0.15 µg/mL) and QC4 (0.225 µg/mL) were analyzed. Each sample was prepared from the matrix of each of the six different piglets (lots). No alteration in the analyte response was found due to interfering or unidentified compounds in the sample matrix, and the accuracy was always within ±15% of the nominal concentration of QC2 and QC4 of both isoforms, and CV (%) (precision) was not greater than 15%. The matrix factor is expressed as the mean ± standard deviation and CV (%). In plasma, the GEN C1 isoform was 1.150 ± 0.044 (2.4%) for QC2 and 1.235 ± 0.032 (2.6%) for QC4, while the GEN C2C2a isoform values were 1.220 ± 0.054 (4.4%) and 0.820 ± 0.014 (1.7%) for QC2 and QC4, respectively. In feces, the GEN C1 showed values of 0.983 ± 0.034 (3.5%) for QC2 and 0.802 ± 0.016 (2.1%) for QC4, and the GEN C2C2a isoform showed values of 0.762 ± 0.029 (3.8%) for QC2 and 0.789 ± 0.038 (4.8%) for QC4.

2.4. Calibration Curve and Range

Three calibration curves were obtained, and the characteristics of each one of these curves for plasma and feces for the isoforms C1 and C2C2a are summarized in

Table 1. Data from linear regression analysis of calibration curves.

	Plasma
	GEN C1 isoform	GEN C2C2a isoform
Calibration curve 1 (R2)	y = 1.6336 x + 0.0899 (0.9858)	y = 1.0892 x − 0.1066 (0.9773)
Calibration curve 2 (R2)	y = 1.2438 x + 0.1761 (0.9857)	y = 1.2836 x − 0.0897 (0.9893)
Calibration curve 3 (R2)	y = 1.5593 x + 0.0161 (0.9889)	y = 1.6617 x − 0.2323 (0.9903)
	Feces
	GEN C1 isoform	GEN C2C2a isoform
Calibration curve 1 (R2)	y = 1.9623 x + 0.0391 (0.9809)	y = 1.3492 x − 0.0786 (0.9846)
Calibration curve 2 (R2)	y = 1.5576 x − 0.0490 (0.9877)	y = 0.9564 x − 0.1029 (0.9906)
Calibration curve 3 (R2)	y = 1.5587 x − 0.0913 (0.9787)	y = 1.8146 x − 0.0848 (0.9886)

. There was a linear relationship between the nominal analyte concentration and the analytical response obtained in the calibration curves. Standard GEN solutions used were within the range 0.05–0.3 µg/mL. Good linearity was evidenced, and high coefficients of determination were obtained.

Back-calculated recoveries of each calibration concentration level were established from the calibration curve. All samples of each calibration curve were within ±20% of the nominal concentration at the LLOQ, and ±15% at all the other levels. The mean recovery of the isoforms C1 and C2C2a was 94.3 ± 5.50% for C1 and 94.9 ± 5.10% for C2C2a in plasma and 94.5 ± 2.98% for C1 and 95.4 ± 2.13% for C2C2a in feces (

Table 2. LLOQ, LOD, and recovery for C1 and C2C2a in plasma and feces.

	Plasma
	LLOQ (µg/mL)	LOD (µg/mL)	Recovery (%) ($\overline{\mathbf{x}}$ ± SD)
GEN C1 isoform	0.1155	0.0831	94.3 ± 5.50
GEN C2C2a isoform	0.1975	0.0950	94.9 ± 5.10
	Feces
	LLOQ (µg/mL)	LOD (µg/mL)	Recovery (%) ($\overline{\mathbf{x}}$ ± SD)
GEN C1 isoform	0.1155	0.0783	94.5 ± 2.98
GEN C2C2a isoform	0.1975	0.1030	95.4 ± 2.13

). The plasma LOD (limit of detectión) values were 0.0831 µg/mL and 0.0950 µg/mL for C1 and C2C2a, respectively, and 0.0783 µg/mL and 0.1030 µg/mL for C1 and C2C2a in feces.

2.5. Accuracy and Precision

Accuracy and precision were determined for the four QC levels (QC1, QC2, QC3, and QC4) for each of the GEN isoforms (C1 and C2C2a). The intra-day (within-run) and inter-day (between-run) precision and accuracy of the method in plasma and feces samples are shown in

Table 3. Intra- and inter-day accuracy and precision for the quality controls (QCs) in plasma.

		Intra-Day (Within-Run)						Inter-Day
		Batch 1		Batch 2		Batch 3		(Between-Run)
		CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)
GEN C1 isoform	QC1	13.4	92.2	10.6	102.0	12.9	88.5	12.5	95.5
	QC2	8.9	104.1	7.3	105.7	8.6	100.0	7.8	104.7
	QC3	5.1	106.1	5.9	96.0	4.9	101.9	6.9	102.7
	QC4	6.6	102.5	7.0	107.6	6.4	98.4	6.7	104.2
GEN C2C2a isoform	QC1	1.5	116.5	5.3	115.4	11.5	111.8	3.1	116.1
	QC2	7.9	98.3	6.5	103.3	7.5	94.4	7.3	100.0
	QC3	8.0	98.7	9.7	96.4	7.6	94.7	8.0	97.9
	QC4	8.8	101.4	4.9	89.3	8.4	97.3	9.5	97.3

and

Table 4. Intra- and inter-day accuracy and precision for the quality controls (QCs) in feces.

		Intra-Day (Within-Run)						Inter-Day
		Batch 1		Batch 2		Batch 3		(Between-Run)
		CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)
GEN C1 isoform	QC1	3.9	83.3	6.5	85.7	2.9	90.0	5.1	85.2
	QC2	12.6	98.5	7.6	93.1	3.2	90.7	10.2	97.1
	QC3	3.4	94.4	5.1	101.9	6.5	96.7	6.3	98.5
	QC4	0.2	95.0	8.2	98.4	7.4	99.2	7.0	97.6
GEN C2C2a isoform	QC1	12.3	96.5	4.0	100.3	6.1	101.7	9.2	98.8
	QC2	11.0	94.1	6.0	89.9	1.6	87.6	8.9	93.0
	QC3	3.0	96.3	3.3	100.9	2.5	97.7	4.3	98.4
	QC4	10.3	105.1	6.1	93.9	8.7	102.4	10.6	98.6

, respectively.

As shown in Table 3, in plasma samples and considering the GEN C1 isoform, within-run precision ranged from 10.6 to 13.4%, with an accuracy ranging from 88.5 to 102.0% for QC1 (LLOQ). On the other hand, precision for QC2, QC3, and QC4 ranged from 4.9 to 8.9%, with an accuracy from 96.0 to 107.6%. For the GEN C2C2a isoform, within-run precision and accuracy for QC1 ranged from 1.5 to 11.5% and from 111.8 to 116.5%, respectively. Within-run precision and accuracy for QC2, QC3, and QC4 ranged from 4.9 to 9.7% and from 89.3 to 103.3%, respectively.

In feces samples (Table 4) and considering the GEN C1 isoform, within-run precision and accuracy for the LLOQ ranged from 2.9 to 6.5% and from 83.3 to 90.0%, respectively. For QC2, QC3, and QC4, within-run precision ranged from 0.2 to 12.6% and accuracy ranged between 90.7% and 101.9%. In the case of the GEN C2C2a isoform, within-run precision ranged from 4.0 to 12.3% and accuracy from 96.5 to 101.7% for the LLOQ. As shown in Table 4, within-run precision and accuracy for QC2, QC3, and QC4 ranged from 1.6 to 11.0% and from 87.6 to 105.1%, respectively.

Regarding between-run precision and accuracy in plasma and feces samples, taking into account the GEN C1 isoform, this ranged from 6.7 to 12.5% with an accuracy from 95.5 to 104.7% in plasma and from 5.1 to 10.2% with an accuracy from 85.2 to 98.5% in feces. Considering GEN C2C2a, between-run precision and accuracy in plasma ranged between 3.1 and 9.5% and 97.3 and 116.1%, respectively. In feces samples, these values ranged between 4.3 and 10.6% for precision and between 93.0 and 98.8% for accuracy.

The percentages obtained comply with the EMA criteria [30]. Similarly, in terms of precision (% CV) within and between runs, it did not exceed the values of 15% (20% at the LLOQ).

Therefore, the method used meets the accuracy and precision criterion described in the European ICH Guideline M10 for bioanalytical method validation and study sample analysis [30] for both isoforms, C1 and C2C2a.

2.6. Carry-Over

No residual analyte was detected from a preceding sample that remained in the analytical instrument when blank plasma and blank feces samples were injected after the ULOQ (0.3 µg/mL) and the highest quality control (QC4, 0.225 µg/mL) in each run. The results complied with the validation guideline criteria, which indicates that the analyte response in the blank samples should not be greater than 20% of the analyte response at the LLOQ and 5% of the response for the IS.

2.7. Stability

The stability of GEN isoforms C1 and C2C2a, as well as the IS, in plasma and feces was evaluated under different times and storage conditions using QC2 and QC4. Concentrations were calculated by using the calibration curve obtained on the day of the analysis. In all cases, the acceptance criteria were fulfilled, with an accuracy that ranged from 85.1 to 114.8% in plasma samples and from 95.3 and 114.7% in feces (

Table 5. Stability of quality controls (QC2 and QC4) under different storage conditions in plasma.

		GEN C1 isoform				GEN C2C2a isoform
Temperature (°C)	Time	QC2		QC 4		QC2		QC 4
		CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)
−20	3 days	6.1	110.8	6.5	112.4	6.8	87.5	6.4	107.7
	7 days	2.3	111.4	2.8	114.3	3.0	86.9	2.4	104.6
	15 days	3.4	113.0	7.2	113.2	7.6	88.8	3.6	97.6
	1 month	2.3	111.7	2.8	112	2.9	90.9	2.4	92.3
	2 months	4.3	100.7	2.6	111.3	2.7	88.2	4.5	85.4
	4 months	3.7	109.5	3.3	108.3	3.5	86	3.9	91.9
4	24 h before extraction	1.0	102.9	0.7	109.8	0.6	91.2	1.0	94.7
	24 h after extraction	3.0	114.8	2.5	112.6	2.4	92.4	3.1	97.7
	48 h before extraction	1.7	112.2	3.9	107.0	3.8	89.9	1.8	91.2
	48 h after extraction	8.2	102.9	6.7	113.5	6.5	89.2	8.6	101.1
25	24 h before extraction	3.6	107.4	3.5	113.3	3.4	85.1	3.8	104.8
	24 h after extraction	5.2	114.3	11.0	106.5	4.6	99.6	5.5	86.3

and

Table 6. Stability of quality controls (QC2 and QC4) under different storage conditions in feces.

		GEN C1 Isoform				GEN C2C2a Isoform
Temperature (°C)	Time	QC2		QC 4		QC2		QC 4
		CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)	CV (%)	Accuracy (%)
−20	3 days	5.8	111.4	3.3	105.3	4.1	98.0	7.6	114.7
	7 days	2.2	100.5	4.2	105.8	12.8	97.0	3.4	114.3
	15 days	3.2	109.2	4.1	107.3	0.8	99.4	8.5	109.4
	1 month	2.2	110.6	5.5	106.1	2.9	101.9	3.3	103.3
	2 months	4.1	111.1	0.8	95.7	4.6	98.8	3.1	95.6
	4 months	3.5	112.7	2.9	104.0	7.8	96.3	3.9	102.9
4	24 h before extraction	2.9	97.7	8.2	97.7	9.2	102.2	0.7	106.1
	24 h after extraction	2.8	109.1	3.6	109.1	3.5	103.5	2.7	109.4
	48 h before extraction	1.6	106.6	9.2	106.6	3.3	100.7	4.2	102.1
	48 h after extraction	7.8	97.8	3.5	97.8	4.2	99.9	7.2	113.2
25	24 h before extraction	3.5	102.1	4.6	107.6	8.2	95.3	3.8	106.9
	24 h after extraction	5.0	108.5	7.8	108.5	3.6	111.6	5.1	96.7

).

2.8. Robustness

The analytical methods were robust, in terms of the flow rate (±0.1 mL/min), pH of the mobile phase (±0.2), and gradient slope (5.2 vs. 5 min and 14 vs. 13.5 min). The method remained stable and reproducible, with no significant deviations in retention time, peak shape, or analyte response. Moreover, the relative standard deviation (RSD) values for all tested parameters remained below 5%, confirming the robustness and reliability of the developed analytical method.

2.9. Method Application

The applicability of the method was evaluated by analyzing plasma and feces samples obtained from three healthy pigs treated with GEN. Six plasma samples were collected at different time points, and three feces sample was collected at 0, 24, and 48 h after oral administration of GEN. Retention times for the different GEN isoforms in plasma and feces samples were consistent with those obtained during method validation. Moreover, no potentially interfering peaks at the retention times of the isoforms of GEN were observed in the analyzed samples, as shown in Figure 3. Plasma concentrations ranged from 8.91 ± 0.62 µg/mL at 0.5 h to 0.54 ± 0.16 µg/mL at 3 h for the GEN C1 isoform. For the GEN C2C2a isoform, plasma concentrations ranged from 13.78 ± 0.30 µg/mL at 0.5 h to 0.69 ± 0.12 µg/mL at 3 h. Always keeping in mind that at time 0, the concentration was 0 µg/mL. In feces, GEN isoforms were not detected either on the day of administration or 48 h post-administration. The concentration determined at 24 h was 8.93 ± 0.36 µg/mL and 15.21 ± 0.27 µg/mL for the GEN C1 and GEN C2C2a isoforms, respectively. The results obtained confirmed that the UPLC-MS/MS method developed was suitable for the determination of the GEN components (C1 and C2C2a).

The development of this analytical method based on UPLC-MS/MS for the determination of GEN in plasma and feces represents a significant contribution, as it combines high sensitivity and specificity in the separation of the different GEN isoforms. UPLC-MS/MS provides narrower peaks and greater chromatographic efficiency, allowing shorter analysis times through the use of sub-2 µm stationary phase particles, compared to previously reported methods for plasma and tissue samples from pigs, based on HPLC, HPLC-MS, HPLC-MS/MS, or HPLC-FLD [6,18,22,23], where retention times are longer. Furthermore, the lower consumption of solvents during the extraction phase and the lower production of waste constitute significant economic and environmental advantages.

3. Materials and Methods

3.1. Chemicals and Reagents

The reference analytical standard of GEN was purchased from Sigma-Aldrich (Merck, Darmstadt, Germany). The GEN used [Gentamicin sulfate salt hydrate, Vetranal^TM (Merck, Darmstadt, Germany)] is composed of different isoforms: C1 (23.1%) and C2C2a (39.5%). STR was used as an internal standard (streptomycin sulfate salt, purity 98%, Sigma-Aldrich (Merck, Darmstadt, Germany)).

The reagents and solvents used for drug extraction and analysis were acetonitrile LC-MS grade (Fisher Scientific, Madrid, Spain), ammonium acetate (GPR RECTAPUR VWR, Leuven, Belgium), methanol (LiChrosolv Merck, Darmstadt, Germany), isopropanol (HiPerSolv CHROMANORM VWR, Fontenay sons Bois, France), formic acid (Sigma-Aldrich, Steinheim, Germany), hydrochloric acid (Analyticals Carlo Erba, Milano, Italy), trichloroacetic acid (TCA) (AnalaR NORMAPUR, VWR, Leuven, Belgium), sodium chloride (AnalaR NORMAPUR VWR, Leuven, Belgium), sodium hydroxide 1N (Panreac Quimica S.A., Barcelona, Spain), hydrochloric acid 1M (Analyticals Carlo Erba, Milano, Italy), disodium EDTA (Sigma-Aldrich, Steinheim, Germany) and heptafluorobutyric acid (HFBA) (Sigma-Aldrich, Steinheim, Germany).

A Millipore Milli-Q gradient water purification system was used to obtain HPLC grade water (18.2 MΩ·cm resistivity at 25 °C; TOC < 5 ppb). Solid-phase extraction was carried out by using Oasis HLB 1cc 30 mg cartridges (Waters Corporation, Mildford, MA, USA).

3.2. Animals and Experimental Procedures

Six healthy female pigs (5 weeks old) with a weight range between 9 and 10 kg were used. The study was carried out in the experimental farm of the Veterinary Faculty (University of Leon, Leon, Spain). Animals were determined to be clinically normal by physical examination. Before starting the experiment, pigs were allowed to acclimatize to their environment to minimize stress and were maintained in an adequately ventilated building, having free access to drinking water and food. The Institutional Animal Care and Use Committee (University of Leon) approved the study (OEBA-ULE-018-2023).

Heparinized tubes (Vacutainer^®, sodium heparin, BD, Plymouth, UK) were employed to collect samples from the jugular veins. Once obtained, blood samples were centrifuged (1500 rpm, 20 min), and plasma was stored at −20 °C until analysis. In the case of feces samples, these were collected on the same day as the blood samples and stored at −20 °C until processing.

3.3. Preparation of Stock, Calibration, and Quality Control Working Solution

1.

GEN stock solution containing a mixture of the isoforms C1 (23.1%) and C2C2a (39.5%) (2 mg/mL) was prepared in HPLC-grade water. An independent solution for IS (2 mg/mL) was also prepared.

2.

Calibration working solutions containing GEN solution (0.5, 1, 1.5, 2, 2.5, and 3 µg/mL) and IS (1 µg/mL) were then prepared by diluting an appropriate volume of the stock solution in 10 mL of HPLC-grade water.

3.

Quality control working solutions (QC) were prepared at 4 concentration levels, according to the European validation guideline [30]:

• QC1 (LLOQ, lower limit of quantitation): 0.5 µg/mL, which contained 0.1155 µg/mL of GEN C1 isoform and 0.1975 µg/mL of GEN C2C2a isoform.
• QC2 (LOW, three times the LLOQ): 1.5 µg/mL, which contained 0.3465 µg/mL of GEN C1 isoform and 0.5925 µg/mL of GEN C2C2a isoform.
• QC3 (MED, between 30 and 50% of the calibration curve range): 1.75 µg/mL, which contained 0.4023 µg/mL of GEN C1 isoform and 0.6913 µg/mL of GEN C2C2a isoform.
• QC4 (HIGH, at least 75% of the upper calibration curve range): 2.25 µg/mL, which contained 0.5197 µg/mL of GEN C1 isoform and 0.8888 µg/mL of GEN C2C2a isoform.

Stock solutions, calibration working solutions, and quality control working solutions were prepared daily.

3.4. Preparation of Analysis Samples

The following types of samples were used:

1.

Blank samples: biological matrix (plasma or feces) without GEN and IS (1 mL).

2.

Zero samples: blank sample (0.9 mL plasma or 1 g feces) with 0.1 mL IS.

3.

Calibration standards: 0.9 mL plasma or 1 g feces spiked with 100 µL of each calibration working solution. Thus, the concentrations of the calibration samples were 0.05, 0.1, 0.15, 0.2, 0.25, and 0.3 µg/mL for GEN solution and 0.1 µg/mL for IS.

4.

Quality control samples were also prepared in plasma (0.9 mL) and feces (1 g) with 0.1 µg/mL for IS (0.1 mL) at 4 concentration levels:

• 0.05 µg/mL (QC1): 0.01155 µg/mL of GEN C1 isoform and 0.01975 µg/mL of GEN C2C2a isoform.
• 0.15 µg/mL (QC2): 0.03465 µg/mL of GEN C1 isoform and 0.05925 µg/mL of GEN C2C2a isoform.
• 0.175 µg/mL (QC3): 0.04023 µg/mL of GEN C1 isoform and 0.06913 µg/mL of GEN C2C2a isoform.
• 0.225 µg/mL (QC4): 0.05197 µg/mL of GEN C1 isoform and 0.08888 µg/mL of GEN C2C2a isoform.

All samples were fully thawed at room temperature.

3.5. Extraction Method

Solid-phase extraction (SPE) was used to determine GEN from plasma and feces samples [19]. Initial experiments were carried out by using different extraction solvents and solid-phase extraction cartridges. The best results were obtained with the following procedure:

Plasma samples: 7.5 mL of extraction buffer (ammonium acetate-TCA-NaCl-disodium EDTA) was added to 1 mL of plasma sample. Next, samples were shaken for a few minutes using a vortex (VWR International Eurolab S.L., Barcelona, Spain) and then centrifuged at 4000 rpm for 5 min. Finally, the supernatant was collected and the pH was adjusted to 6.5 using HCl (1 M) or NaOH (1 M).

Feces samples: as in plasma samples, 7.5 mL of the extraction buffer was added to 1 g of feces. The same process was followed as in the case of the plasma samples, repeating the process 3 times.

Subsequently, Oasis HLB 1cc 30 mg cartridges were used. The cartridges were conditioned with 1 mL of methanol and 1 mL of HPLC water. Then, 1 mL of supernatant obtained was added and the cartridges were washed with 1 mL of HPLC water. Thereafter, cartridges were eluted with 0.5 mL of elution solution (formic acid–isopropanol-reagent water (10/5/85, v/v/v)). A 250 µL aliquot of the eluate was collected and combined with 3 µL of HFBA. Finally, the resultant sample was analyzed by UPLC-MS/MS.

All these procedures were performed at room temperature.

3.6. UPLC-MS/MS Conditions

Samples were analyzed by Ultra-Performance Liquid Chromatography (UPLC) with a triple quadrupole mass detector (QQQ) in UPLC AcQuity Waters with triple quadrupole mass detector TQD AcQuity (Waters Corporation). Separation was performed by using an AcQuity UPLC HSS T3; 1.8 µm (2.1 mm × 50 mm) (Waters Corporation).

Table 7. UPLC-MS/MS conditions.

Mobile phase	Mobile phase A = ultrapure water with 0.26% HFBA Mobile phase B = acetonitrile with 0.26% HFBA
Flow rate	0.5 mL/min	Sampler temperature (°C)	8 °C
Column temperature (°C)	40 °C	Injection volume	30 µL
Gradient program	Time (min)	Mobile phase A	Mobile phase B
	0	80	20
	5.2	0	100
	9	0	100
	14	80	20
	17	80	20

shows the mobile phase, the chromatographic method conditions, and details of the different chromatographic gradient conditions used.

The analytes were detected in positive ion mode (ESI+) using multiple reaction monitoring (MRM). Quantification MRM transitions were m/z 478.2766 → 157.183 for GEN C1, m/z 464.2766 → 322.199 for GEN C2C2a, and m/z 582.2681 → 263.1675 for STR. Collision energies were 14 and 22 V for GEN C2C2a and GEN C1, respectively, and 32 for STR.

Under the Good Laboratory Practice (GLP) regulations, the study was conducted at our laboratory, LAFARLE (University of Leon, Spain), and in the Instrumental Techniques Laboratory (LTI) of the University of Leon.

3.7. Method Validation

The European Medicines Agency Guideline EMA/CHMP/ICH/172948/2019 [30] was followed to validate the method developed for plasma and feces samples. The parameters described below were established according to the description included in this guideline (selectivity, specificity, matrix effect, and calibration curve and range, accuracy, precision, carry-over, and stability).

3.7.1. Selectivity

Selectivity was assessed by analyzing the chromatograms of blank plasma and blank feces samples from six different individual sources to determine the interference in the retention times of the different isoforms of GEN (C1 and C2C2a) and the internal standard. According to the guideline [30], no significant interference from other compounds should be observed in blank samples at the retention times of both analyte and IS (not higher than 20% of the analyte response and not more than 5% of the IS signal both in the LLOQ samples for plasma and feces).

3.7.2. Specificity

Ten blank samples were used to assess specificity. According to the guideline [30], interference from other compounds should not exceed 20% of the analyte signal and 5% for the IS at the LLOQ [30].

3.7.3. Matrix Effect

In this case, the guideline states that accuracy should be within the range ±15% of the nominal concentration, and precision (expressed as % CV) should not be greater than 15% for each matrix lot [30]. The matrix effect between different independent sources was evaluated during method validation. In addition, the matrix factor was calculated, although it is not a parameter included in the guideline. The matrix factor was determined by calculating the ratio of peak areas of solutions in the presence of the matrix to the peak areas of solutions in the absence of the matrix [31].

3.7.4. Calibration Curve and Range

The calibration curve was established from the analysis of plasma and feces samples. Each calibration curve was obtained with a blank sample, a zero sample, and the six different concentration levels of calibration standards mentioned before, which included the LLOQ and the ULOQ (0.05–0.3 µg/mL) for the GEN solution. The concentration of IS was 0.1 µg/mL. The LLOQ was assessed in blank samples spiked with the lowest quality control working solution (0.5 µg/mL) and the ULOQ (3 µg/mL). Known concentrations of each GEN isoform against the ratio of area of each GEN isoform vs. IS were used to develop the linear regression analysis. Parameters (determination coefficient (R²), slope, and the intercept of the resulting calibration curves) were calculated without both blank and zero samples.

3.7.5. Accuracy and Precision

Within- and between-run analysis with QCs was carried out to define accuracy and precision. For the within-run (intra-day) analysis, 5 replicates of each of the 4 QCs were processed and evaluated the same day. To determine the between-run (inter-day) study, each of the 4 QC levels was processed 5 times in 3 runs and on 3 different days. Precision was expressed as CV. For both of these parameters (accuracy and precision), mean values should be within ±15%, except for the LLOQ, which may not be higher than 20% [30].

3.7.6. Carry-Over

Carry-over was assessed by the analysis of blank samples (plasma and feces) after the injection of the ULOQ (0.3 µg/mL) and the QC4 (0.225 µg/mL) in each run. The signal should not exceed 20% of the analyte’s response at the LLOQ and 5% of that of the IS.

3.7.7. Stability

The stability of the different GEN isoforms (C1 and C2C2a) in matrix samples (plasma and feces) was evaluated using low- and high-concentration QCs (QC2 and QC4) under different storage conditions (25 °C, 4 °C, and −20 °C) and over different periods of time.

3.7.8. Robustness

The robustness of the analytical method was evaluated by introducing minor and deliberate changes in chromatographic parameters, including the pH of the mobile phase, the flow rate, and the gradient program, to assess the method’s ability to maintain accuracy and precision under slightly altered conditions. The robustness of this method was determined in terms of % RSD.

3.8. Method Application

To verify the applicability of the method developed in clinical practice, GEN was measured in plasma and feces from three healthy 5-week-old female pigs weighing 9.0 ± 0.3 kg. For this purpose, GEN was orally administered at a dose of 2 mg/kg live weight and after a washout period by the intravenous route.

Blood samples were collected (6 mL, Vacutainer^®, sodium heparin, BD, Plymouth, UK) and centrifuged (1500 rpm, 20 min) to obtain plasma. Plasma was then frozen (−20 °C) and stored until processing. Sampling was performed at 0, 0.5, 1, 1.5, 2, and 3 h.

In the case of feces samples, these were collected the day of treatment and 24 and 48 h after oral administration of GEN and stored at −20 °C until processing. IS was added at the time of plasma and feces analysis.

3.9. Data Analysis

MassLynx V4.2 SCN989 (Waters Corporation, USA) software was employed for data acquisition and processing. Descriptive statistics (mean, standard deviation, and percentages) as well as the calibration curve parameters were calculated with Microsoft Excel (Microsoft Corporation, Redmond, WA, USA).

4. Conclusions

The development and validation of a UPLC-MS/MS quantification method for different gentamicin isoforms (C1 and C2C2a) in plasma and feces has been shown, being the first validated method for the determination of gentamicin in pig feces. The method is sensitive, accurate, and precise. The UPLC technique followed can be used in laboratories with standard equipment, with sample preparation being relatively simple. The clinical applicability of the developed method was also demonstrated. As previously mentioned, quantifying gentamicin in feces is important to ensure therapeutic concentrations at the site of infection, particularly in enteric diseases such as swine dysentery caused by Brachyspira hyodysenteriae, where local antimicrobial activity is essential. UPLC-MS/MS has emerged as a powerful analytical technique for separating and quantifying gentamicin isoforms due to its high resolution, speed, and sensitivity. Compared with traditional HPLC methods, UPLC appears to offer improved peak separation, allowing the detection of individual components even at low concentrations.