Nitrofuran antibiotic metabolites detected at parts per million concentrations in retina of pigs—a new matrix for enhanced monitoring of nitrofuran abuse

Abstract

Nitrofuran metabolite residues AOZ, AMOZ, AHD and SEM were detected at parts per million concentrations in retina of pigs fed therapeutic doses of nitrofuran antibiotics. Discovery of this residue depot may allow widespread technology transfer to laboratories lacking LC-MS/MS thus improving global monitoring of these drugs.

---

Therapeutic or prophylactic use of nitrofuran antibiotics in food-producing animals is prohibited in the European Union¹ because of potentially carcinogenic and mutagenic effects on human health.² These broad-spectrum bactericidal drugs were previously used widely for treatment and prevention of various gastrointestinal infections and as growth promoters in livestock. Nitrofurans are metabolised rapidly in vivo but stable tissue-bound metabolites are formed. Fragments of these metabolites may be released by mild acid hydrolysis and monitored as marker residues.³ For example, 3-amino-2-oxazolidinone (AOZ) is monitored as a marker residue for the drug furazolidone. 3-Amino-5-morpholinomethyl-2-oxazolidinone (AMOZ) is the marker residue for furaltadone, 1-aminohydantoin (AHD) for nitrofurantoin and semicarbazide (SEM) for nitrofurazone. EU Member States are required to monitor compliance with the ban on nitrofurans through their annual national residues control plans. Inspection laboratories previously analysed muscle, kidney or liver tissues for nitrofuran metabolites using HPLC with UV detection.⁴ This was superceeded by LC-MS.⁵ Tandem LC-MS/MS has now become the state-of-the-art for nitrofuran detection.⁶ LC-MS/MS can quantify nitrofuran metabolites at parts per billion (ppb) concentrations in tissue and satisfies the EU Minimum Required Performance Limit for nitrofuran metabolites⁷ of 1.0 µg kg⁻¹.

A global nitrofuran “crisis” came to light in 2002–03 when nitrofuran metabolites were reported in poultry and aquaculture products from many countries including Thailand, Vietnam and Brazil, in Portuguese poultry meat, and in pork meat purchased in Italy, Portugal and Greece. During 2003–04, notifications via the EU Rapid Alert System for Food and Feed included findings of nitrofurans in Chinese rabbit meat, Italian pork, and poultry meat products from Argentina, Romania and Bulgaria. Nitrofurans also continue to be reported in fish, prawns, honey and egg powder from various countries. Trade restrictions arising from such findings prompted many food producers and regulatory authorities to instigate nitrofuran testing schemes. However, many laboratories do not have access to LC-MS/MS instrumentation and are unable to detect the low ppb levels of nitrofurans commonly found in edible tissues.

Seventy two 8 week old piglets were divided into 4 groups and housed in concrete floored pens. Four nitrofuran medicated feeds were prepared using furazolidone, furaltadone, nitrofurantoin and nitrofurazone (Sigma-Aldrich, UK) at 400 mg kg⁻¹ feed, the recommended therapeutic dose for furazolidone⁸ prior to its ban in 1995. Each group was fed medicated feed ad libitum for 10 days. On the tenth day, medicated feed was removed, pens were thoroughly swept out and the feeding areas hosed out with water. All pigs were then fed conventional unmedicated feed ad libitum for the remainder of the study. Fresh water was available at all times, pigs were fed once each morning and faeces swept from pens daily. At 0, 1, 2, 3, 4 and 6 weeks following withdrawal of medicated feed 3 pigs from each group were euthanized by captive bolt and tissue samples removed and stored at −20 °C. Both eyes from each animal were dissected and the retina/choroid layer was separated from the sclera prior to freezing. For the purposes of this communication, the term retina is used for a combination of the retina and vascular choroid layers, both of which contain the pigment melanin which is implicated in the accumulation of drug residues.

Each pig yielded less than 200 mg of retina. Consequently, samples (<100.0 mg) were weighed into glass tubes and results corrected for sample mass following analysis. Pooled bovine retina (100 mg) known to be free of nitrofurans was used as negative control tissue. Negative retina (100 mg) was also fortified at 5 mg kg⁻¹ using a mixed standard (50 µl) containing AOZ, AMOZ, AHD and SEM (10 µg ml⁻¹ in methanol) to act as recovery controls. All samples, controls and standards were fortified with a 1 µg ml⁻¹ mixed internal standard solution (150 µl) containing isotopically labelled D4-AOZ, D5-AMOZ and ¹³C¹⁵N2-SEM supplied by Witega Laboratorien Berlin-Adlershof (Germany). Calibration standards were prepared using aliquots of 1 or 10 µg ml⁻¹ mixed standard solutions. To all tubes were added 0.1 M hydrochloric acid (5 ml) and 100 mM 2-nitrobenzaldehyde in DMSO (150 µl). Tubes were incubated overnight at 60 °C on a spiral roller mixer to facilitate hydrolytic release of the nitrofuran metabolites from tissue proteins and their conversion to stable nitrophenyl derivatives. Use of a laboratory homogeniser to disrupt such small retina samples prior to derivatisation ran the risk of significant loss of sample within the machinery. Retina tissue is sufficiently thin (<0.5 mm) to allow diffusion of the acid through the film of tissue and solubilisation of the nitrofuran metabolites. Gentle mixing at 60 °C also helped to partially disrupt the tissue during incubation.

Following derivatisation, samples were cooled to room temperature and adjusted to pH 7.4 ± 0.2 with 0.1 M dipotassium hydrogen orthophosphate (5 ml) and 1 M sodium hydroxide (0.4 ml). Liquid–liquid extraction was performed using 2 × 5 ml ethyl acetate, mixing for 2 min followed by centrifugation for 10 min at 2000 rpm. The organic phases were combined and evaporated to dryness at 45 °C under nitrogen. Residues were re-dissolved in methanol : water (50 ∶ 50 v/v; 5 ml), transferred to HPLC microvials (200 µl) and stored at 4 °C. Immediately prior to LC-MS/MS analysis, microvials were centrifuged at 13000 rpm for 10 min to remove particulate material.

An Agilent 1100 Series HPLC system (Agilent Technologies, USA) coupled to a Quattro Ultima^® Platinum tandem mass detector (Micromass, UK), operating under MassLynx^® software, were used for sample analysis. The mass spectrometer operated in electrospray positive mode and data acquisition was in multiple reaction monitoring mode. The ions monitored are listed in Table 1. The source settings were as follows: capillary voltage 2.45 kV, RF lens 25.0 V, source temperature 120 °C, desolvation temperature 350 °C, cone nitrogen gas flow 80 l h⁻¹, desolvation gas flow 670 l h⁻¹. Argon was used as the collision gas and the multiplier was operated at 650 V. The cone voltage (35 to 60 V) and collision energy (7 to 16 eV) changed during analysis depending on the analyte. The HPLC system was equipped with a Luna C18(2) 3 µm, 2.0 × 150 mm column (Phenomenex, UK). A binary gradient mobile phase was used at a flow rate of 0.2 ml min⁻¹, solvent A being 0.5 mM ammonium acetate and methanol (80 ∶ 20 v/v mix), solvent B being methanol. Sample injection volume was 10 µl. Analyte concentrations in samples were calculated by comparing the ratio of an analyte base peak response to its appropriate internal standard response with the same ratio in calibration curve standards. Quantification of AHD utilised the D4-AOZ internal standard.

Table 1 LC-MS/MS precursor/product ion combinations monitored in MRM ESI positive mode for the analysis of 4 nitrofuran metabolites

Compound	Precursor ion (m/z)	Product ions (m/z)
2-NP-AOZ	236.3	104.3, 134.3
2-NP-AMOZ	335.3	262.3, 291.4
2-NP-AHD	249.3	104.3, 134.2
2-NP-SEM	209.2	166.2, 192.2
2-NP-D4-AOZ	240.3	134.3
2-NP-D5-AMOZ	340.3	296.4
2-NP-^13C^15N2-SEM	212.2	195.3

---

Depletion of nitrofuran metabolites from porcine retina during the 6 weeks withdrawal period is illustrated in Fig. 1. Nitrofuran metabolites accumulate in retina at concentrations (approximately 1 mg kg⁻¹ AHD, 11 mg kg⁻¹ AOZ, 18 mg kg⁻¹ SEM and 100 mg kg⁻¹ AMOZ) an order of magnitude greater than in edible tissues such as muscle. Depletion half lives over the 6 week period were: AOZ 3.8 weeks, SEM 3.8 weeks, AMOZ 2.4 weeks and AHD 1.9 weeks. AHD exhibits the lowest concentrations (0.1 mg kg⁻¹ remaining after 6 weeks withdrawal) and the shortest half life, suggesting this is the least stable of the nitrofuran metabolites in vivo. By contrast, AMOZ concentrations are remarkably high, 14 mg kg⁻¹ remaining in retina after 6 weeks. AOZ and SEM demonstrate similar concentrations and depletion characteristics (3 and 5 mg kg⁻¹ respectively remaining after 6 weeks).

Fig. 1 Depletion of total nitrofuran metabolites from porcine retina.

The major significance of these very high retinal concentrations and long depletion half lives is that all four nitrofuran metabolites will be detectable at slaughter in retina of a pig fed therapeutic doses of nitrofuran antibiotics at any point in its lifetime. Given that LC-MS/MS is capable of detecting sub-ppb concentrations, it is also likely that exposure of an animal to contamination levels of nitrofurans, for example via housing-to-pig transfer, will also be detectable via retinal analysis. Furthermore, since retinal concentrations are so high, it may be possible to analyse whole eyeballs in a rapid test for nitrofurans with minimal sample preparation. We are currently investigating this possibility with eyes of chickens fed nitrofuran medicated diets.

The accumulation of drug residues in the eye has been well documented in the case of some β-adrenergic agonists such as clenbuterol.⁹ Other β-agonists such as ractopamine do not accumulate to the same extent.¹⁰ The pigment melanin has been identified as the site of binding of drugs in pigmented ocular tissues.¹¹ Melanin is found in the iris, ciliary body and the retinal and choroid layers of the eye in addition to hair, skin and certain regions of the brain. The precise mechanism of drug binding to melanin is unclear but depends on the chemistry of the drug. Polycyclic compounds with coplanar fused ring structures, steroids and hydrophobic primary amines are likely to accumulate in pigmented ocular tissues.^11,12 Furthermore, electronic, steric and acid/base considerations appear to influence the degree of accumulation.¹¹ Nitrofuran metabolites bind covalently to tissue proteins following metabolic degradation in vivo. However, the precise structures of these metabolites are unclear. It is therefore difficult to speculate on the reasons for the variable degree of accumulation of the nitrofuran metabolites in retina. Nevertheless, as with some β-agonists, retina appears to be a metabolic “dead end” tissue where nitrofuran metabolites accumulate at high concentrations and where they are not subject to the higher cellular turn-over and clearance rates of tissues such as liver.

Many residues testing laboratories do not have access to expensive LC-MS/MS instrumentation which has become the state-of-the-art for the detection of nitrofuran residues. Single LC-MS or HPLC-UV are not sufficiently sensitive to reach the 1 µg kg⁻¹ EU MRPL. However, the use of retina, which contains much higher concentrations than in currently monitored matrices, would allow many more laboratories throughout the world to embark on nitrofuran monitoring schemes using single MS or UV detection. This may be of particular benefit to countries in S.E. Asia where nitrofuran use has been widespread and has caused problems for trade relations with the EU. Furthermore, retinal analysis may allow detection of nitrofuran administration to animals at any point from birth to slaughter. The transfer of this technology could be widespread and fast, leading to a significant improvement in the global monitoring of nitrofuran use and a subsequent improvement in food safety.

Notes and references

1. Commission Regulation 1442/95, Off. J. Eur. Commun., 1995, L143, 26.
2. J. E. M. van Koten-Vermeulen, M. F. A. Wouters and F. X. R. Van Leeuwen, Report of the 40th Meeting of the Joint FAO/WHO Expert Committee on Food Additives, World Health Organisation, Geneva, 1993, pp. 85–123.
3. L. A. P. Hoogenboom, O. Tomassini, M. B. M. Oorsprong and H. A. Kuiper, Food Addit. Contam., 1991, 29, 185.
4. L. A. P. Hoogenboom, M. C. J. Berghmans, T. H. G. Polman, R. Parker and I. C. Shaw, Food Addit. Contam., 1992, 9, 623.
5. R. J. McCracken and D. G. Kennedy, J. Chromatogr. B, 1997, 691, 87.
6. A. Leitner, P. Zollner and W. Lindner, J. Chromatogr. A, 2001, 939, 49.
7. Commission Decision 2003/181, Off. J. Eur. Commun., 2003, L71, 17.
8. National Office of Animal Health Limited, Compendium of Data Sheets for Veterinary Products 1992–93. Datapharm Publications Ltd., UK, 346.
9. A. Malucelli, F. Ellendorff and H. H. D. Meyer, J. Anim. Sci., 1994, 72, 1555.
10. D. J. Smith, K. M. Ehrenfried, J. D. Dalidowicz and M. P. Turberg, J. Anim. Sci., 2002, 80, 2931.
11. R. M. J. Ings, Drug Metab. Rev., 1984, 15, 1183.
12. F-J. Leinweber, Drug Metab. Rev., 1991, 23, 133.

---