Application of Immunoaffinity Column Cleanup to Aflatoxin M1 Determination and Survey in Cheese

1011

Research Note

S. DRAGACCI* and J. M. FREMY Downloaded from http://meridian.allenpress.com/jfp/article-pdf/59/9/1011/1666175/0362-028x-59_9_1011.pdf by guest on 03 March 2020

Ministere de ['Agriculture, Centre National d'Etudes Veterinaires et Alimentaires (CNEVA Paris), Unite Toxines Microbiennes, 43 Rue de Dantzig, 75015 Paris, France

(MS# 95-220: Received 21 August 1995/Accepted 9 February 1996)

ABSTRACT lady the occurrence of AFMj in their final products. For that purpose, reliable methods to determine the possible contami- Milk products such as cheeses may be contaminated by nation of cheeses are necessary. aflatoxin M1 when manufactured with milk from dairy cattle that In previous studies, the effectiveness of immunoaffinity have consumed aflatoxin Bl-contaminated feeds. The usefulness of columns for extracting AFMj from naturally contaminated immunoaffinitycolumns to determine aflatoxinM1 content in many kinds of cheeses with very good recoveries is demonstrated. The and spiked cheeses was demonstrated (5) and confirmed (2). In this report, an application of the immunoaffinity method- analysis of aflatoxin Mj in a 1990 to 1995 limited survey in France shows that the occurrence of this mycotoxin in cheeses is rather ology to a wide variety of cheeses is described. In addition, a infrequent. With the exception of samples from 1989 to 1990 when limited survey on the occurrence of AFM j in cheeses has aflatoxin B)-contaminated maize meals were incidentally imported been undertaken over the last 6 years and the data are also to supplement dairy cattle feed, very few samples were found with presented here. above 0.200 f.lgof aflatoxin M1 per kg of cheese, the maximum acceptable level. MATERIALS AND METHODS

Key words: Aflatoxin M j, immunoaffinitycolumn, cheese Cheese samples For spiking experiments, various kinds of cheeses (cottage- Aflatoxin M1 (AFM1) is the main aflatoxin B] metabo- like, ripened, blue, hard, and spiced cheeses) were chosen and lite found in the milk of lactating animals that have ingested obtained from the retail market. For the survey, samples for routine contaminated feed. AFM 1 was recently placed into group 2B analysis came directly from dairy manufacturers from 1990 to as "an agent possibly carcinogenic to humans" by !ARC (4). mid-1995. Part of this survey was accomplished with the help of Since AFM] is known to be stable throughout the manufac- one French veterinary laboratory from Calvados. This laboratory ture of dairy products, the monitoring of AFM] contents belongs to the French network of laboratories involved in the need to be enhanced. It is noteworthy that during cheese national monitoring program for aflatoxinM1 for which proficiency testing is organized by the authors' laboratory each year. processing the binding of AFMj with casein may result in a concentration of this toxin (by two- to fivefold) in the manufactured products (11). Determination of AFMj in cheese France produces more than 1,000,000 tons of cheeses a Before 1992, the analysis of AFM1 in surveyed cheeses was year (1994: 1,541,000 tons), which are distributed under conducted as follows. AFM1 was extracted from samples with approximately 400 trademarks. National monitoring pro- chloroform followed by a cleanup through silica-gel columns. The grams carried out since 1976 in order to control the domestic AFM1 contents were then measured with a high-pressure liquid production of raw milk have been shown to be very chromatography (HPLC) system, the detector being a fluorometer effective, since almost none of the milk produced has (3). exceeded the French tolerance level of 0.050 Ilg of AFM j per After 1992, for routine analysis and for spiking experiments, liter (1). Nevertheless, in addition to these yearly surveys, the method used was as previously described (2). Briefly, AFMj more and more dairy manufacturers wish to monitor regu- was extracted by mixing roughly cut samples of cheeses with dichloromethane. The extracts were then evaporated and the residues suspended in methanol and water. After washing with n-hexane to eliminate fats, the aqueous phases were passed through * Author for correspondence. Tel: 33 1 55762178; Fax: 33 1 55762706. immunoaffinity columns (AflaPrep M, Rhone Poulenc Diagnos- 1012 DRAGACCI AND FREMY tics) following the manufacturer's instructions. The acetonitrile shown in Table 1. Samples are classified as cottage-like, eluates were then analyzed with an HPLC connected to a fluorom- soft, hard, blue, and spiced cheeses. Some of them are from eter according to Tuinstra et al. (8). The limit of detection was milk of goat or ewe origins. Most of the samples displayed 0.010 f!g of AFMj per kg of cheese. For quantification, a 0.050 to no residual contamination. The recovery rates after spiking 0.500 ng of AFMj calibration curve was used. The chromatograms at 0.200 !lg AFM/kg were found in a range from 63.1 to were very clear with few extraneous peaks, which improves the 114%, most of them (17 of 22) having a 70 to 100% accuracy and precision of the assays. recovery. No relation between the rates of recovery and the milk origins or the types of cheeses was observed. We only Spiking procedure noticed that in cottage-like cheeses or in ripened cheeses, Samples were spiked by mixing them with a known volume of very high recoveries (90 to 100%) with high repeatibility standard AFMj solution previously diluted in acetonitrile plus were obtained. This could be due (i) to an easier mixing of water (1;3). The spiking amount was chosen at a level of 0.200 f!g these products compared with hard cheeses and (ii) to a more of AFMj per kg of cheese to mimic the possible contamination of cheese made from milk contaminated at the tolerance level (i.e., simple matrix compared to that of blue cheeses. Besides, as 0.050 f!g/liter). we have pointed out earlier J2), the way in which the milk Downloaded from http://meridian.allenpress.com/jfp/article-pdf/59/9/1011/1666175/0362-028x-59_9_1011.pdf by guest on 03 March 2020 All assays were made in duplicate. proteins were precipitated (during rennet and/or lactic fermentation) to prepare cheeses had no effect on the

recovery of extracted AFM I' RESULTS A 1990 to 1995 limited survey for the occurrence of AFM 1 in cheeses was conducted. Table 2 shows in detail the The efficiency of immunoaffinity cleanup on a wide AFM 1 contents found in cheeses in this survey. In 1990, 3 of variety of cheeses was demonstrated. The AFM j levels in 28 samples exhibited an AFM1 level above 0.200 !lg/kg. various samples before and after spiking (values from blank Since this period, no cheeses had more than 0.060 !lg/kg of samples having been subtracted) and the recoveries are AFMj•

TABLE I. AFMj levels in cottage-like, ripened, hard, blue, and spiced cheeses before and after spiking

Aflatoxin M j (flg/kgof cheese)a Recovery Cheese types Origin of milk Before spiking After spiking AFMj (0.200 flg/kg)b (%)

Cottage-like Faisselle cow Mimolette cow

C The limit of detection was 0.010 f!g of AFM1 per kg of cheese. IMPROVED ASSAY FOR AFLATOXIN M) IN CHEESE 1013

TABLE 2. French survey, 1990 to 1995,for aflatoxin M] (AFM]) contamination in cheeses

No. of samples containing AFM)

Year" No. of samples Limits of detection Limit-O.200 Ilg/kg (range) >0.200 Ilg/kg

Results from a French vet lab 1990-1994 121 1990-1993: 0.100 0 0 1994: 0.050 Results from CNEVAParis 1990 28 0.100 1 3 (0.12) (0.21-0.51 ) 1991 20 0.100 2 0 (0.10-0.19) 1992 27 0.010 10 0

(0.012-0.037) Downloaded from http://meridian.allenpress.com/jfp/article-pdf/59/9/1011/1666175/0362-028x-59_9_1011.pdf by guest on 03 March 2020 1993 13 0.010 0 0 1994 32 0.010 5 0 (0.027-0.059) 1995 (Jan. to June) 16 0.010 0 0 a Until 1992 some samples were analyzed using solid-phase extraction chromatography (7); since then, the current immunoaffinity procedure has been used: the limit of detection has been drastically lowered.

DISCUSSION ACKNOWLEDGMENTS

The use of an immunoaffinity technique to extract The authors thank the veterinary laboratory of Calvados, France, for AFM[ from many types of cheeses was demonstrated. The providing dam on contamination of cheeses skillfully analyzed by Mr. Bouchaud. scope of the provisional International Dairy Federation

(IDF) Standard for the determination of AFM1 in milk using REFERENCES immunoaffinity cleanup might then be extended to other milk products like cheeses. Thus, the full method (immuno- 1. Dragacci, S., and 1. M. Fremy. 1993. Contamination du lait par I'AFM1: resu1mts de quinze annees de surveillance. Sci. Aliments 13:711-722. affinity plus HPLC-fluorimetry) (8) could be suitable to 2. Dragacci, S., E. Gleizes, J. M. Fremy, and A. A. G. Candlish. 1995. monitor milk and milk products for AFM[ as well. Use of imrnunoaffinity chromatography as a purification step for the The survey conducted in France between 1990 and determination of AFM) in cheeses. Food Addit. Contam. 12:59-65. 1995 shows that some cheeses may contain significant 3. Fremy, J. M., and B. Boursier. 1981. Rapid determination of AFM) in dairy products by reverse-phase liquid chromatography. J. Chro- amounts of AFM,. In 1990, three samples were higher than matogr.219:156-161. the value considered by some European countries as the 4. IARC (International Agency for Research on Cancer). 1993. Some maximum acceptable level for cheeses (i.e., 0.20 to 0.25 naturally occurring substances: food items and constituents, heterocy- !lg/kg) (9). Indeed, between 1989 and 1992, France was clic aromatic amines and mycotoxins, p. 245-395. In !ARC mono- graphs on the evaluation of carcinogenic risks to hUmans, vol. 56. facing some occurrences of AFM[ in raw milk due to the IARC Scientific Publications, Lyon, France. consumption by dairy cattle of imported feed (com-gluten 5. Sharman, M., A. L. Patey, and J. Gilbert. 1989. Application of an feed) containing an unexpected aflatoxin B[level. In 1989 to immunoaffinity column sample clean-up to the determination of AFM1 in cheese. J. Chromatogr. 474:457--461. 1990, 3% of the monitored French production of raw milk 6. Taguchi, S., S. Fukushima, T. Sumimoto, S. Yoshida, and T. Nish- was above the French recommended value (i.e., 0.05 !lg/ imune. 1995. Aflatoxins in food collected in Osaka, Japan, from 1988 liter) (1). So the contamination of cheeses manufactured in to 1992. J. AOAC Int. 78:325-327. the winter of 1989 to 1990 could be clearly related to that of 7. Trucksess, M. W., and S.W. Page. 1986. Examination of imported cheeses for AFM\. J. Food Prot. 49:632-633. lactating cows. In the following years, no more contamina- 8. Tuinstra, L. G., A. H. Roos, and J. M. P. Van Trijp. 1993. IDF tion at such levels was observed. collaborative study on the determination of AFM) in milk powder This limited survey does not intend to predict the actual using immunoaffinity column. J. AOAC Int. 76: 1249-1254. occurrence of AFM[ in French cheeses. But, in accordance 9. Van Egmond, H. P. 1989. Current situation on regulations for mycotoxins. Overview of tolerances and status of standard methods of with other published reports from countries where monitor- sampling and analysis. Food Addit. Contam. 6:139-188. ing for aflatoxin Bl in feed and AFM[ in milk does exist (6, 10. Van Egmond, H. P. 1989. AFM): occurrence, toxicity, regulation, p. 7, 10), it seems that the contamination of cheeses by AFM[ 11-56. In H. P. Van Egmond (ed.), Mycotoxins in dairy products. did not appear to be a threat to public health at that time. Elsevier Science Publishers, New York. 11. Yousef, A. E., E. H. Marth. 1989. Stability and degradation of AFMIo Close attention should still be paid by manufacturers in this p. 127-161. In H. P. Van Egmond (ed.), Mycotoxins in dairy products. field. Elsevier Science Publishers, New York.