In vitro and in vivo metabolite profiling of valnemulin using ultraperformance liquid chromatography-quadrupole/time-of-flight hybrid mass spectrometry.

Valnemulin, a semisynthetic pleuromutilin derivative related to tiamulin, is broadly used to treat bacterial diseases of animals. Despite its widespread use, metabolism in animals has not yet been fully investigated. To better understand valnemulin biotransformation, in this study, metabolites of valnemulinin in in vitro and in vivo rats, chickens, swines, goats, and cows were identified and elucidated using ultraperformance liquid chromatography-quadrupole/time-of-flight hybrid mass spectrometry (UPLC-Q/TOF-MS). As a result, there were totally 7 metabolites of valnemulin identified in vitro and 75, 61, and 74 metabolites detected in in vivo rats, chickens, and swines, respectively, and the majority of metabolites were reported for the first time. The main metabolic pathways of valnemulin were found to be hydroxylation in the mutilin part (the ring system) and the side chain, oxidization on the sulfur of the side chain to form S-oxides, hydrolysis of the amido bond, and acetylization in the amido of the side chain. In addition, hydroxylation in the mutilin part was proposed to be the primary metabolic route. Furthermore, the results revealed that 2β-hydroxyvalnemulin (V1) and 8α-hydroxyvalnemulin (V2) were the major metabolites for rats and swines and S-oxides (V6) in chickens.

Introduction

Valnemulin is an antimicrobial with a pleuromutilin chemical structure similar to that of tiamulin, which acts as an inhibitor of bacterial protein synthesis.[1−3] Valnemulin, tiamulin, and retapamulin are the pleuromutilin derivates on the market, and their chemical structures are shown in Figure 1. Owing to its effective antimicrobial activities against mycoplasma and brachyspira, valnemulin has been widely used in veterinary practice, including as treatment for swine dysentery, swine enzootic pneumonia, and mycoplasma infection in poultry and cattle.[4−10] Additionally, in humans valnemulin has been reported to treat resistant mycoplasma infection in immunocompromised patients.[11]

Figure 1

Chemical structures of valnemulin, tiamulin, retapamulin, and pleuromutilin.

The extensive use of valnemulin in veterinary medicine has drawn considerable attention. Recently, several studies on the absorption, distribution, and pharmacokinetics of valnemulin after oral administration in different species of animals have been reported, including in rats, chickens, ducks, dogs, and pigs.[9,10,12−14] These reports showed that valnemulin was well absorbed after oral administration and widely distributed to tissues. In addition, valnemulin was also excreted rapidly, mostly via bile and feces.[9,10,15] Despite its widespread use, unfortunately, the metabolism of valnemulin in animals has not been fully investigated. According to a report from the European Agency for the Evaluation of Medicinal Products (EMEA), after oral administration of 3H-labeled valnemulin, more than 22 different metabolites were detected in the plasma, liver, urine, and feces of rats. However, the chemical structures of the metabolites detected in rats have not been clarified.[9,10] After pigs were orally given valnimulin at 25 mg/kg bw, twice a day, for 7.5 days, 11 metabolites were detected in bile and 6 of these were also found in the liver; some metabolites were identified by using LC-MS and 1H NMR.[9] The results showed that all of these metabolites retained the intact valnemulin skeleton and were oxidized either in the side chain or in the pleuromutilin ring, and no epoxides were detected. As pleuromutilin and tiamulin have a pleuromutilin ring similar to that of valnemulin, their metabolism could provide some basic information for the metabolic identification of valnemulin. The incubation of pleuromutilin derivatives with hepatic cytochrome P-450 enzymes showed that pleuromutiline were extensively metabolically hydroxylated in the 1β-, 2β-, and 8α-positions.[16] After incubation of tiamulin with the liver microsomes of pigs, 5 metabolites of tiamulin were detected; 3 major metabolites were isolated and structurally elucidated to be 8α-hydroxytiamulin, 2β-hydroxytiamulin, and N-deethyltiamulin by using LC-MS, 1H NMR, and 13C NMR.[17]

In the past decades, due to the significant effect on the prevention and treatment of livestock diseases, antimicrobial agents have been widely used in clinical veterinary. Unfortunately, the accumulation of drug residues and metabolites in eggs, meat, and milk has led to potential hazards for humans.[18−20] Thus, the investigation of the metabolism of antimicrobial agents in food animals becomes very necessary. Valnemulin is one of the antimicrobial agents in animal husbandry. However, the literature on the metabolic research of valnemulin in animals is limited. Therefore, the aim of the present study is to investigate in detail the in vivo and in intro metabolism of valnemulin in different species and to provide the metabolic differences among different species. Valnemulin was separately incubated with rat, chicken, swine, goat, and cow liver microsomes, and the incubation mixtures were then analyzed using a sensitive UPLC-Q/TOF-MS method. In addition, the metabolites of valnemulin in in vivo rats, chickens, and swines were also investigated. As a result, many metabolites were identified, and the proposed metabolic pathways of valnemulin are presented in this paper. This work will contribute to comprehensively clarifying the differences among in vitro systems of different species and will give useful data for the explanation of the residues of valnemulin among different species and, moreover, will further facilitate food safety evaluation of valnemulin.

Materials and Methods

Chemicals

The valnemulin standards were obtained from the Council of Europe’s European Pharmacopoeia. S-oxidized valnemulin (metabolite V6) was synthesized at the College of Science, China Agricultural University (Beijing, China). Acetonitrile and formic acid (HPLC grade) were purchased from Fisher Chemical Co. (USA). Water was purified using a Milli-Q system (Millipore, USA). β-Nicotinamide adenine dinucleotide phosphates (NADPH) was purchased from Roche Chemical Co. (Beijing, China). All other chemicals and reagents were of the highest analytical grade available.

Animals

Twelve Wistar rats (weight 200–250 g, six males and six females) were purchased from Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China). Twelve Avian chickens (weight 1.0–1.2 kg, six males and six females) were purchased from Beijing Huadu Co. Ltd. (Beijing, China). Twelve Changbai swines (weight 20–25 kg, six males and six females), eight goats (weight 20–25 kg, four males and four females), and four cows (weight 150–250 kg, two males and two females) were purchased from Zhoukou Village Nursery (Beijing, China). All animals, after acclimatizing for 1 week in an animal room with standardized temperature (25–28 °C), humidity (50–60%), and 12 h light/dark cycle conditions before the experiment, were fed different commercial standard diets and given water ad libitum. All care and handling of animals was performed with approval of the Institutional Authority for Laboratory Animal Care.

Preparation of Liver Microsomes

Following a 12 h fast, animals (three male and three female rats, chickens, and swines; two male and two female goats; and one male and one female cow) were exsanguinated. Hepatic microsomes were prepared as a procedure reported in previous studies.[21] The livers were rapidly removed and washed several times with ice-cold 0.1 mol/L PBS (pH 7.4) and 0.05 mol/L Tris-HCl buffer (pH 7.4) to wash away residual blood, gently blotted, and weighed. The pooled livers were minced and immediately homogenized with 3 volumes of ice-cold 0.05 mol/L Tris-HCl buffer (pH 7.4) containing 1 mmol/L EDTA and 0.25 M sucrose using a glass homogenizer. The liver S9 fraction was obtained by centrifuging the homogenate at 10000g for 20 min at 4 °C, and then the supernatant was collected and centrifuged at 100000g for 60 min at 4 °C. The microsomal pellets were suspended in 0.05 mol/L Tris-HCl buffer, and the protein content of the microsomes was estimated according to the method of Lowry et al. using crystalline bovine serum albumin as standard.[22] To evaluate the activities of CYP enzymes in liver microsomes prepared in this work, the carbon monoxide (CO) difference spectrum and the metabolism of probe drug coumarin were applied. The liver microsomes of rat, chicken, swine, goat, and cow were incubated with coumarin, and then the metabolites of coumarin were analyzed by UPLC-Q/TOF-MS. The results showed that the maximum absorption wavelength of microsomes was 450 nm, but not 420 nm, which was consistent with the results of Pearce et al.[23] The CO difference spectrum of CYP in rat, chicken, pig, goat, and cow liver microsomes is shown in the Supporting Information (Figure 9). In addition, coumarin could be metabolized to 7-hydroxycoumarin in liver microsomes, indicating the systems had the activities of drug metabolism enzymes.

Incubation of Valnemulin with Liver Microsomes

Valnemulin was separately incubated with liver microsomes of rats, chickens, swines, goats, and cows under the same conditions. The incubation mixture, in 0.05 M Tris-HCl buffer (pH 7.4), consisted of 2 mg protein/mL liver microsomes, 100 mM valnemulin, and 1 mM NADPH with the total volume of 500 mL. After 2 h of incubation at 37 °C in a metabolic shaker, the reaction was terminated by adding 500 μL of ice-cold acetonitrile. After volution and centrifugation at 12000 rpm at 4 °C for 15 min, the supernatant was filtered through a 0.22 μm microbore cellulose membrane into an autosampler vial and analyzed by UPLC-Q/TOF-MS for identification of metabolites. Parallel controls include the absence of NADPH and incubation without valnemulin, respectively. All experiments were conducted in triplicate.

Preparation of in Vivo Samples

After a 12 h fast and free access to water, animals (three male and three female rats, chickens, and swines) were administered a single dose of valnemulin (100 mg/kg, dissolved in water) by oral gavage. The urine and feces samples were individually collected prior to the experiment and during 0–24, 24–48, and 48–72 h postdose. All of the samples were frozen at −20 °C until analysis. Each fecal sample was well homogenized before weighing. Ten milliliters of ethyl acetate was added to 2 g of feces, and then the mixture was vortexed for 5 min and centrifuged at 9000g at 4 °C for 10 min. The organic layer was transferred into a new polypropylene tube, and the extraction of the feces was repeated following the same procedure. The resulting supernatants were pooled and evaporated to dryness with a gentle nitrogen flow at 45 °C. The residue was redissolved in 1.0 mL of acetonitrile/water (15:85, v/v), and then the resulting solution was loaded onto a preconditioned activated Oasis HLB cartridge (3 mL, 60 mg); the cartridge should be activated with 3 mL of methanol and then preconditioned with 3 mL of water. Furthermore, the cartridge was washed with 6 mL of water, and then the metabolites were eluted with 3 mL of methanol. The eluate was evaporated to dryness under nitrogen at 45 °C, and the residues were eventually reconstituted in 1.0 mL of acetonitrile/water (15:85, v/v). The redissolved solution was centrifuged at 12000g at 4 °C for 15 min. Finally, the supernatant was filtered through a 0.22 μm microbore cellulose membrane into an autosampler vial and analyzed by UPLC-Q/TOF-MS for the identification of metabolites. As to the urine samples, 10 mL of urine from rats and pigs was loaded onto a preconditioned activated Oasis HLB cartridge (3 mL, 60 mg). Afterward, the rest of the steps were performed as described for feces samples.

Instrumental Conditions

Identification of valnemulin metabolites was conducted using an ACQUITY UPLC system (Waters Co., USA) coupled with a hybrid Q/TOF-MS SYNAPT HDMS (Waters, UK). Chromatographic separation of valnemulin metabolites was performed on the ACQUITY ultraperformance liquid chromatography system with ambient temperature, using an Acquity BEH RP18 column (50 mm × 2.1 mm i.d., 1.7 μm particle size) (Waters, USA). The mobile phase of 0.1% formic acid in water (solvent A) and 0.1% formic acid in acetonitrile (solvent B) was pumped at a flow rate of 0.3 mL/min. The gradient elution program was as follows: 0–3.0 min, 5–15% solvent B; 3.0–16.0 min, 15–40% solvent B; 16.0–18.0 min, 40–70% solvent B; 18.0–19.0 min, 70–100% solvent B; 19.0–20 min, 100% solvent B; 20.0–21.0 min, 100– 5% solvent B; 21.0–22.0 min, 5% solvent B. The injection volume was 10 μL.

Mass spectrometric analysis was operated using a hybrid Q/TOF-MS SYNAPT HDMS (Waters, UK). Typical source conditions for maximum intensity of precursor ions were as follows: capillary voltage, 3.0 kV; source temperature, 120 °C; desolvation temperature, 300 °C; cone gas (N2) flow rate, 10 L/h; desolvation gas (N2) flow rate, 600 L/h. The argon pressure in the collision cell was 3.5 × 10–3 mbar. For parent drug and metabolites, ESI source was operated in the positive ionization mode. Data were acquired from 100 to 700 Da and centroid during acquisition using an internal reference comprising a 1 ng/μL solution of leucine enkephalin infused at 50 μL/min, generating a reference ion at m/z 556.2771 in ESI positive ionization mode. Low-energy data were acquired at a collision energy of 5 eV and high-energy data using a ramped collision energy of 10–45 eV.

Data Processing

Data processing was carried out using MetabolynxXS software (version 4.1), which could automatically identify metabolites by comparing the samples with the controls. A threshold of 5 mDa was set as a limit to filter the processed data for calculation of possible elemental composition. The mass window was 0.05 Da for both expected and unexpected mass chromatograms, and the absolute area was 50. The intensity threshold of spectrum condition was 10%. The availability of accurate masses and elemental composition data on precursor and all fragment ions in each MS/MS spectrum greatly enhanced confidence in the ion structure assignments.

Results and Dicussion

Structural Assignment of Product Ions of Valnemulin

Because it is critical to figure out the fragmentation pathways of valnemulin, UPLC-Q/TOF-MS was applied initially to analyze the protonated valnemulin. The product ion mass spectrum of valnemulin is shown in Figure 2A, which was used not only as a reference to aid in the interpretation of product ions of the metabolites but also to examine the high resolution and mass accuracy of the instrument. With the help of formula predictor software, the formulas, the observed and calculated masses, and the mass errors of protonated valnemulin and its fragment ions are listed in Table 1. Relatively good accuracy was obtained with values between the measured and calculated masses ranging from −0.9 to 1.1 mDa (from −6.9 to 4.6 ppm), which powerfully enhanced the reliability for the structures of product ions.

Figure 2

Accurate MS/MS spectra (A) and proposed fragmentation pathways (B) of valnemulin.

Table 1

Elemental Composition, Measured and Calculated Masses, Double-Bond Equivalents (DBE), and Corresponding Mass Errors of Valnemulin and Its Fragment Ions

elemental composition	measured mass (Da)	calculated mass (Da)	DBE	mass errors (mDa)	corresponding mass errors (ppm)
C31H53N2O5S+	565.3668	565.3675	6.5	–0.7	–1.2
C20H31O2+	303.2315	303.2324	5.5	–0.9	–3.0
C20H29O+	285.2230	285.2218	6.5	1.2	4.6
C11H23N2O3S+	263.1421	263.1429	1.5	–0.8	–3.0
C11H21N2O2S+	245.1335	245.1324	2.5	1.1	4.5
C9H19N2O+	171.1504	171.1497	1.5	0.7	4.1
C6H14NO2S+	164.0736	164.0745	0.5	–0.9	–5.5
C6H11O2S+	147.0483	147.0480	1.5	0.3	2.0
C4H10N+	72.0807	72.0813	0.5	–0.5	–6.9

As shown in Figure 2A, the main fragment ion at m/z 263.1421 corresponded to cleavage of the ester bond from the protonated valnemulin. Meanwhile, m/z 303.2315 can also be generated with cleavage of the ester bond. Subsequently, m/z 285.2230 was generated with the loss of H2O from m/z 303.2315. In addition, m/z 263.1421 generates m/z 164.0736 by losing C5H9O, which accorded with cleavage of the acidamide bond. Afterward, m/z 147.0483 was generated by the loss of NH3 from the m/z 164.0736. In addition, m/z 171.1504 can be generated by cleaving the bond of S–C in the absence of C2H4O2S. According to the results above, the fragmentation pathways of valnemulin are proposed in Figure 2B.

Identification of Valnemulin Metabolites in Swine Liver Microsomes

To rapidly characterize the metabolism of valnemulin, incubation of swine liver microsomes with valnemulin and NADPH was initially carried out. The samples and the controls were first analyzed by UPLC-Q/TOF-MS. Subsequently, MetabolynxXS data-processing software utilizing MDF (mass defect filtering) technique was used to identify metabolite ions. As a result, seven potential metabolites were listed by the software. In addition, the accurate extracted ion chromatograms (EIC) (the window was 0.05 Da) of the potential metabolites V1–V6 (m/z 581.3624) and V41 (m/z 466.2991) are shown in the Supporting Information (Figure 10). To elucidate the molecular structure of the potential metabolites, the MS/MS experiments were conducted. Furthermore, the accurate MS/MS spectra of these metabolites are indicated in Figure 3. With the assistance of chromatographic behaviors, accurate mass measurements, and basic rules of drug metabolism, the chemical structures of metabolites were elucidated or clarified.

Figure 3

MS/MS spectra of the metabolites of valnemulin detected in in vitro rats, chickens, swines, goats, and cows.

Metabolites V1–V5

Metabolites V1–V5 showed the same [M + H]+ ions at m/z 581.3620 ([C31H53N2O6S]+), which was 16 Da higher than the protonated valnemulin at m/z 565.3675, suggesting that they were isomers of mono-oxidized metabolites of valnemulin. Their difference was the position of hydroxylation. Fragment ions from MS/MS spectra of V1–V5 were m/z 263, 171, 164, and 147, which were identical to valnemulin, implying that the side chains of these metabolites were not oxidized. Moreover, the fragment ions at m/z 319 and 301 of V1–V5 were 16 Da higher than the ions at m/z 303 and 285 of valnemulin, respectively, further suggesting that the hydroxylation takes place in the mutilin part (the ring system) of valnemulin. Owing to the complex structure of valnemulin with many similar moieties, it would be extremely difficult to unambiguously identify the exact position of hydroxylation in the mutilin part solely using mass spectrometry. NMR analysis would be ideal for such compounds, but it may be difficult to generate sufficient amounts of metabolites. However, earlier studies have shown that the derivates of pleuromutilin were often metabolized by hydroxylation on the 1β-, 2β-, and 8α-positions. Therefore, it is probable that valnemulin and tiamulin can also be hydroxylated on these positions, which was confirmed by later research conducted by Lykkeberg et al.[17] After incubation of tiamulin with the liver microsomes of pigs, two hydroxylated metabolites were detected as 8α-hydroxytiamulin and 2β-hydroxytiamulin by using LC-MS, 1H NMR, and 13C NMR.[17] In the present research, more than five metabolites hydroxylated in the mutilin part (V1–V5) were found, and two (V1 and V2) had larger amount output. On the basis of the analysis, metabolites V1 and V2 were probably 2β-hydroxyvalnemulin and 8α-hydroxyvalnemulin, respectively. As to the trace metabolites V3–V5, it was hardly possible to identify the exact position of hydroxylation in the mutilin part.

Metabolite V6

The [M + H]+ ion of metabolite V6 (m/z 581.3627, C31H53N2O6S+) was 16 Da higher than the protonated valnemulin. The main product ion at m/z 279 was also 16 Da higher than the ion at m/z 263 of valnemulin, implying that the side chain of V6 was alternatively hydroxylated or oxidized at the sulfur. The key fragment ion at m/z 171 (loss of C2H4O3S from m/z 279) of V6 further suggests that the side chain was oxidized at the sulfur to form S-oxides. In addition, V6 showed the same retention time and fragment ions on the MS/MS spectrum as the synthesized S-oxidized valnemulin. Consequently, metabolite V6 was identified as S-oxidized valnemulin.

Metabolite V41

Metabolite V41 was eluted at 8.85 min with the [M + H]+ ion at m/z 466.2995 (C26H44NO5S+), which was 99 Da less than the protonated valnemulin at m/z 565.3675, implying that the amido bond of the side chain was hydrolyzed. In addition, the major fragment ion at m/z 164 from metabolite V41, corresponding to cleavage of the ester bond, was also 99 Da less than the main fragment ion at m/z 263 of valnemulin, further suggesting that the amido bond was hydrolyzed.

Metabolism of Valnemulin in Rat, Chicken, Swine, Goat, and Cow Liver Microsomes

Considering that metabolic enzymes and metabolic rates were different among various species, the quantitative and qualitative evaluation of metabolites of valnemulin may present some distinctions. Therefore, it was essential and urgent to investigate the metabolism of valnemulin in rats, chickens, swines, goats, and cows, which will be very helpful for its application in pharmacology and clinical use in veterinary practice. Figure 9 (Supporting Information) shows the accurate EICs of valnemulin metabolites formed in liver microsomes of rats, chickens, goats, and cows. The results reveal the same number of valnemulin metabolites for incubation with liver microsomes of different species. Except that metabolite V4 was not detected in cow liver microsomes, seven metabolites (V1–V6 and V41) in swine liver microsmoes were also observed in the other four species. On the basis of these results, the possible metabolic pathways of valnemulin in liver microsomes were hydroxylation, S-oxidation, and hydrolysis, which are proposed in Figure 4.

Figure 4

Metabolic pathways of valnemulin in rat, chicken, swine, goat, and cow liver microsomes.

However, as to the amount of metabolites, there were some differences in the five species. To better analyze the species differences in the metabolism of valnemulin, a semiquantitative analysis was conducted in in vitro systems using the peak areas of extracted ion chromatograms (the window was 0.05 Da) about metabolites; the results are summarized in Figure 5. The results indicated that the ability to metabolize valnemulin in rat liver microsomes was higher than that of the other four species, because the residual level of the unchanged valnemulin in rats was the lowest and there were many metabolites with larger production detected in rats. As to the other four animals, their abilities to metabolize valnemulin were roughly equivalent. The present results also suggest that hydroxylation in the mutilin part was the main metabolic pathway because abundant 2β-hydroxyvalnemulin (V1) and 8α-hydroxyvalnemulin (V2) were produced during the incubation. Moreover, quantitative species differences in the metabolism of valnemulin among the five species were observed. For metabolite V1, the rank order of metabolism ability was rat > goat > pig > chicken > cow and that for metabolite V2, rat > cow > pig > chicken > goat. In addition, the amount of metabolite V1 produced in the liver microsomes of rats, chickens, swines, and goats was greater than that of V2, whereas in cow the amount of metabolite V1 was less than that of V2. Additionally, rat and chicken liver microsomes produced considerable yields of metabolites V4 and V5. The marked quantitative differences in the formation of valnemulin metabolites among the various animal species were perhaps in relation to the levels of the involved enzymes in the liver. As to the S-oxidized (V6) and hydrolyzed metabolites (V41), they were minor products in liver microsomes of the five species. In the absence of the NADPH, only two metabolites (V1 and V2) were detected with minor production in liver microsomes. The amount of metabolites V1 and V2 in the presence of the NADPH was 50-fold higher than that in the absence of the NADPH. These results indicate that NADPH-cytochrome P450 oxidase may be involved in hydroxylation in the mutilin part because NADPH was absolutely essential for this reaction.

Figure 5

Peak areas of extraction ion chromatograms on valnemulin metabolites detected in rat, chicken, swine, goat, and cow liver microsomes.

Metabolite Profiles of Valnemulin in in Vivo Rat, Chicken, and Swine

Seven metabolites were detected and characterized in liver microsomes of different species. To investigate the correlation of in vitro findings with in vivo metabolism, animals (rats, chickens, and pigs) were orally dosed with valnemulin, and urine and feces were collected at different times (0–24 and 24–48 h). Then, these biological samples and the control samples after pretreatment were injected into the UPLC-Q/TOF-MS for analysis. In addition to the unchanged valnemulin, totald of 75, 61, and 74 metabolites of rats, chickens, and pigs were identified, respectively. Among the identified metabolites, 7 known metabolites have been confirmed in the in vitro research, and 68 new metabolites were found in vivo. In addition, the accurate extracted ion chromatograms (EIC) of valnemulin metabolites detected in vivo are shown in the Supporting Information (Figure 11). We assume that they were valnemulin metabolites because comparison of the samples with the control samples as well as the agreement between the accurate mass measurement in MS spectra and predicted formula calculation within 10 ppm. The predicted elemental compositions, exact masses, and mass errors of the metabolites are indicated in the Supporting Information (Table 2). The exact and measured masses agree to within <5 ppm, providing support for the proposed elemental compositions of the metabolites.

Identification of Valnemulin Metabolites in Rats, Chickens, and Pigs after Oral Administration

The metabolites were proposed by the MetabolynxXS software, and then the structures were confirmed using the MS/MS mode. As shown in Table 2 (Supporting Information), there were totally 75 metabolites of valnemulin discovered in rats, chickens, and swines, and the majority are being reported for the first time. In addition, the MS/MS spectra of the detected metabolites are shown in Figure 6. Using Q-TOF, the accurate masses for the protonated molecular ions of the metabolites and their product ions in MS and MSMS were obtained, respectively. The measured accurate masses, mass errors, fragment ions, and retention times for these proposed metabolites are presented in Table 2. In all cases, the structures of metabolites and their fragment ions were rapidly and reliably characterized on the basis of the determined elemental composition and accurate MS/MS spectra. Previously, seven metabolites (V1–V6 and V41) have been confirmed in liver microsomes, whereas the other metabolites detected in vivo need further identification.

Figure 6

MS/MS spectra of the metabolites of valnemulin detected in in vitro rats, chickens, and swines.

Metabolite V7

The [M + H]+ ion of metabolite V7 (m/z 581.3627, C31H53N2O6S+) was 16 Da higher than that of the protonated valnemulin. The main product ion at m/z 279 was also 16 Da higher than the ion at m/z 263 of valnemulin, implying that the side chain of V7 was hydroxylated. Furthermore, the ions at m/z 164 and 147 of metabolite V7 were identical to those of valnemulin, suggesting that the position of hydroxylation should be located at the terminus of the side chain.

Metabolites V8–V26

The protonated molecule ions of metabolites V8–V26 were at m/z 597.3569 (C31H53N2O7S+), which were 16 Da higher than the metabolites V1–V7 at m/z 581.3620, implying that they were further hydroxylated or oxidized products of metabolites V1–V7. The fragment ion at m/z 295 from metabolite V8 was 16 Da higher than the ion at m/z 279 of metabolites V6 and V7, implying that the side chain of V8 was not only hydroxylated but also oxidized at the sulfur in the side chain. Furthermore, the ion at m/z 187, generated by cleaving the S–C bond, was 16 Da higher than the ion at m/z 171 of metabolite V6, further proving the above point. Similar characteristic product ions were shown in the eight metabolites V9–V16 with different retention times, suggesting that they were the isomers. Fragment ions at m/z 263, 171, 164, and 147 from metabolites V9–V16 were identical to those of valnemulin, implying that their side chains were not hydroxylated. In addition, the fragment ions at m/z 335 and 317 of metabolites V9–V16 were 16 Da higher than the ions at m/z 319 and 301 of metabolites V1–V5, respectively, further implying that there was a couple of hydroxylated sites in the mutilin part of metabolites V9–V16.

Compared with metabolite V6, similar ions at m/z 279 and 171 were also detected in metabolites V17–V21, suggesting that oxidization at the sulfur occurs in the side chain. As to the fragment ions at m/z 279, 261, and 243 of metabolite V7, they were also available in metabolites V22–V26, implying that their side chains were hydroxylated. In addition, the ions at m/z 319 and 301 of metabolites V17–V21 and V22–V26 were identical to those of metabolites V1–V5. This indicated that the hydroxylation also occurs in the mutilin part of these metabolites.

Metabolites V27–V40

Mass spectra of metabolites V27–V40 showed the same [M + H]+ ions at m/z 613.3530 (C31H53N2O8S+), and their product ions were similar to those of V8–V26 except for a 16 Da (an oxygen atom) mass value shift, suggesting that they were further hydroxylated or oxidized on the basis of metabolites V8–V26. The fragment ion at m/z 295 from metabolites V27–V31, which is identical to metabolite that of V8, hints that they were further hydroxylated on the mutilin part on the basis of metabolite V8. As to the fragment ion at m/z 279 of metabolites V32–V37, it was identical to metabolites that of V17–V26, indicating that their side chains were hydroxylated or oxidized and they were also hydroxylated in the mutilin part on the basis of metabolites V17–V26. In addition, compared with metabolites V9–V16, the similar ions at m/z 263, 171, 164, and 147 were also detected in metabolites V38–V40, suggesting that the hydroxylation was in the mutilin part on the basis of metabolites V9–V16.

Metabolites V42–V46

Metabolites V42–V46 (m/z 482.2940, C26H44NO6S+) were 16 Da higher than the metabolite V41, suggesting that they were hydroxylated or oxidized at the sulfur to form S-oxides on the basis of metabolite V41. Fragment ion at m/z 180 of metabolite V42 was 16 Da higher than ion m/z 164 of metabolite V41, implying that the side chain of metabolite V42 was oxidized at the sulfur. Compared with metabolite V41, the major fragment ions at m/z 164, 147, 129, and 101 were also detected in metabolites V43–V46. Furthermore, the fragment ions at m/z 319, 301, and 283 of metabolites V43–V46, which were 16 Da higher than the ions at m/z 303, 285, and 267 of metabolite V41, further suggest that the mutilin part of metabolites V43–V46 was hydroxylated.

Metabolites V47–V58

Metabolites V47–V58 (m/z 498.2875 C26H44NO7S+) were detected during 3.11–6.03 min, and their product ions were similar to those of V42–V46 except for a 16 Da (an oxygen atom) mass value shift, suggesting that they were further hydroxylated. The fragment ions at m/z 180 and 162 from metabolites V47–V49, which were identical to metabolite V42, hinted that they were hydroxylated on the mutilin part on the basis of metabolite V42. Compared with metabolites V43–V46, the major fragment ions at m/z 164, 147, 129, and 101 were also detected in metabolites V50–V58, whereas other fragment ions at m/z 335, 317, and 299 of metabolites V50–V58 were 16 Da higher than the ions at m/z 319, 301, and 283 of metabolites V43–V46. On the basis of the analysis, it was concluded that metabolites V50–V58 were further hydroxylated on the mutilin part of metabolites V43–V46.

Metabolite V59

Metabolite V59 showed the [M + H]+ ion at m/z 607.3775 (C33H55N2O6S+), which was 42 Da higher than that of the protonated valnemulin, implying that it was acetylated. The main fragment ion at m/z 305 of metabolite V59, which corresponded to cleavage of the ester bond, was 42 Da higher than the fragment ion m/z 263 from valnemulin. Furthermore, the fragment at m/z 164, generated by cleavage of the amido bond from the ion at m/z 305, further suggests that the acetylization occurs on the amido side chain.

Metabolites V60–V65

Metabolites V60–V65 (m/z 623.3721, C33H55N2O7S+) were 16 Da higher than metabolite V59. In addition, the fragment ion at m/z 321 from metabolite V60 was also 16 Da higher than the ion m/z 305 of metabolite V59, indicating that it was not only acetylated on the side chain but also oxidized at the sulfur. Fragment ions at m/z 305, 164, and 147 from metabolites V61–V65 were identical to the those of metabolite V59, implying that the side chain was not hydroxylated. Therefore, the mutilin part of metabolites V61–V65 should be hydroxylated. For metabolites V1–V4, it was hardly possible to know their exact position of hydroxylation in the mutilin part.

Metabolites V66–V75

Metabolites V66–V75 were detected during 7.56–10.98 min with the same [M + H]+ ion at m/z 639.3684, which was 16 Da higher than those of metabolites V60–V65. Compared with metabolite V60, the major fragment ions at m/z 321 and 213 were also detected in metabolites V66–V69, implying that the mutilin part was hydroxylated on the basis of metabolite V60. Likewise, fragment ions at m/z 305, 164, and 147 of metabolites V70–V75 were identical to those of metabolites V61–V65, implying that the mutilin part was further hydroxylated on the basis of metabolites V61–V65.

Comparison of Metabolism of Valnemulin in Rats, Chickens, and Swines

Table 2 lists the metabolites of valnemulin detected in rats, chickens, and swines. There were 72 and 62 metabolites detected in the urine of rats and swines, respectively, whereas in the feces 71 and 72 metabolies were detected, suggesting that valnemulin undergoes extensive metabolism. Because the urine and feces are excreted together in chickens, it was difficult to naturally separate them. Thus, the excreta (mixture of feces and urine) of chicken were analyzed together, and totally 61 metabolites were detected. To find the major metabolites of valnemulin in the animals, the area of the every peak corresponding to the metabolites of valnemulin was surveyed. Owing to the different ionizations based on ESI resource in mass spectrometry, the peak areas can only be used to roughly quantify the metabolites. The majority of metabolites of valnemulin were eliminated in 0–24 h, but there were a few metabolites detected in 24–48 h in three animals. Therefore, the excreta of three animals in the 0–24 h were used to roughly quantify the metabolites. The result is summarized in Figure 7, which shows that there were no major qualitative matabolites in the biotransformation of valnemulin between rats and pigs. Both main metabolites were 2β-hydroxyvalnemulin (V1) and 8α-hydroxyvalnemulin (V2), which was consistent with the in vitro results. However, the metabolite V6, S-oxides, was the major metabolite in chickens and accounts for approximately 42.1% of total metabolites’ peak area, which was very different from the in vitro results. This was because there were larger numbers of metabolic enzymes involved in valnemulin metabolism in vivo than in the liver microsomes. In addition, gastrointestinal microflora also participated in the metabolism of valnemulin, whereas, in vitro, the number of enzymes involved in drug metabolism was very limited in the liver microsomes. The above facts resulted in metabolic differences of valnemulin between in vivo and in vitro chickens. As to the metabolic differences of valnemulin among different species, it may be that metabolic enzymes in chickens were different from rats and pigs.[24,25] In addition, Figure 7 also reveals that the hydrolysis capabilities in three species were swines > rats > chickens, whereas they were chickens ≫ swines > rats for oxidation on the sulfur and rats > swines > chickens for acetylization.

Figure 7

Peak areas of extraction ion chromatograms on valnemulin metabolites detected in in vivo rats, chickens, and swines.

Metabolic Pathways of Valnemulin in Rats, Chickens, and Swines

According to the summary report from EMEA, valnemulin was extensively metabolized to more than 22 different metabolites when rats were orally dosed with 3H-labeled valnemulin.[9,10] After pigs were given oral doses of 25 mg/kg bw valnemulin, twice a day, for 7.5 days, 11 metabolites were detected in bile, and 6 of these were also found in the liver.[9] Those studies indicate that valnemulin was extensively metabolized in various animals. Although 3H-labeled valnemulin was used, unfortunately, it did not succeed in the identification of exact molecular structures of the detected metabolites. In addition, the main metabolites of valnemulin were not indicated yet. The EMEA report shows that all of these metabolites retain the intact valnemulin skeleton and were oxidized either in the side chain or in the pleuromutilin ring. In this study, metabolism of valnemulin in rats, chickens, and swines was investigated using UPLC-Q/TOF-MS technology, and more than 75 metabolites were detected. The results showed that the main metabolic pathways of valnemulin were hydroxylation in the mutilin part and the side chain, oxidization on the sulfur of the side chain to form S-oxides, hydrolysis of the amido bond, and acetylization in the amido of the side chain. Numerous hydroxylated metabolites of valnemulin were detected. In addition, many metabolites (V9–40) having the mutilin part double or triple oxidized were first detected in the present research. This suggests that the pleuromutilin ring tends to be further oxidized on the basis of metabolites V1–V5. Metabolite V6, in which the sulfur of the side chain was oxidized to form S-oxides, was also found in three species, which have been reported by EMEA.[9] In addition, the metabolites of hydrolysis (V41–V58), acetylization (V59–V75), and hydroxylation (V7) in the side chain were also found first in this study. On the basis of the metabolites identified, the proposed metabolic pathways of valnemulin in rats, chickens, and pigs are shown in Figure 8.

Figure 8

Metabolic pathways of valnemulin in in vivo rats, chickens, and swines.

In summary, UPLC-Q/TOF-MS was used to investigate the metabolic fate of valnemulin in in vitro rats, chickens, pigs, goats, and cows and in in vivo rats, chickens, and pigs. As a result, 7 and 75 metabolites of valnemulin were detected in vitro and in vivo, respectively, and most of them are reported for the first time. The results showed that the metabolic pathways of valnemulin were hydroxylation in the mutilin part (the ring system) and the side chain, and the former was the major metabolic route. In addition, the results also revealed that the metabolic pathways and the main metabolites of valnemulin in three species are different. All in all, the proposed metabolic pathways of valnemulin in rats, chickens, pigs, goats, and cows will give useful data for the explanation of the residue of valnemulin among different species and, moreover, will further facilitate food safety evaluation of valnemulin.