Document 3Q7xBj55j4OJZdjqj8Ma6Bd33

8 pp ANALYTICAL BIOCHEMISTRY 118, 336 343 (1981) AR226-1440 Characterization of Fluorinated Metabolites by a Gas Chromatographic-. Helium Microwave Plasma Detector--The Biotransformation of 1 H, 1 H, 2H, 2H-Perfluorodecanol to Pertluorooctanoate DONALD F. !IAGEN,* JON BELISLE,* JAMES D. JOHNSON,t AND P. VENKATESWARLUt *Ceniral Research Laboraranes, tR,ker Laborator,ec. The, and tC'ommercial Chemicals Division, 3M Company, 3M Center, St Paul, Muine3vio 55144 Received June 16, 1981 A gas chromatographic technique utilizing a microwave-sustained helium plasma detector (C C/MPD) was developed to study the biotransformation of lh,lH,2H,211-perlluorndecanol (C1,F,,CH,CH,OH) in adult male rats. The metabolic products resulting from a single oral dose ofthe alcohol were isolated from the blood plasma by an extraction technique, derivatized, and characterized by the GC/MPO system with confirmation by fluorine NMR Four fluorine- containing metabolites were detected by the fluorine-specific channel of the eiement-selcct,vc MPI) and one of these was shown to he perfluorooctanoate (C,F0CO0 ) Appearance of perfluorooctanoate as a metabolite and an additional observation of concomitant e1e~ationin plasma and urinary inorganic fluoride suggests that the biotransformation of the above alcohol to the perfluorooctanoate invohes defluorination of the CF, group adjacent to the CM, group in the parent compound. In recent years, a variety of analytical techniques for differentiating and quantitating the inorganic fluoride and organic fluorifle in biological samples has been described by Taves (1,2), Venkateswarlu et at (3--5), and Belisle and Hagen (6). However, characterization and determination of the specific organic compound(s) in the samples is more informative than determination of total organic fluorine. Based on fluorine NMR data, Guy et a!. (7,8) concluded that perfluorooctanoic acid or a similar compound was present in a fraction prepared from a large pool of human plasma samples A quantitative microanalytical method, based on GC1 with electron-capture detector, was reported by Belisle and Hagen (9) specifically for perfluorooctanoate in blood plasma though the analysis gave no evidence for the presence of perfiuorooctanoate It should be emphasized that this research (9) describes Abbreviations used' OC, gas chromatogiaphy, MPD, microwave plasma detector, FID. flame ionization detector an analytical method for the determination of perfluorooctanoate and is not a study of perfluorooctanoate levels ~n normal human plasma since only a few public donor samples were analyzed. Gas chromatography is well suited for the complex matrices of hiological extracts and the electron-capture detector, which gives a measure of specificity for the halogenated species present, is perhaps the most sensitive and widely used detector for compounds of this type. This detector, however, lacks sal ficient specificity when large quantities of interfering coextractants are present. We are studying the metabolism of various fluorinecontaIning compounds, and this paper WIll present our study of the biotransformation of lfJ,lH,2H,2H-perfluorodecannl in male rats using the element-specific gas chromatography/microwave plasma detector (GC/ MPD) system Specific characterization of fluorinated compounds will be described in this work. The detector plasma (not to be confused 0003-2697/Si/i 80336-O8$02 00/0 336 copynihi C 91i by Academic Press, ins. All righis oi aprod icRofl nan y form resenod RECEIVED OPPT NCIC 2003 Oct 12 11:50AM I H, IJf,2H,2H-PERFLUORODECANOL BIOTRANSFORMATION 337 with blood plasma) is an extremely energetic ionization source of metastable atoms, molecules, ions, and electrons generated in a microwave cavity. The plasma generates atomic line emission spectra for the elements present in the compounds eluted from the GC column. Each element (C, H, Cl, F, 5, etc.) has a discrete emission spectrum and, using the described detector, one can monitor a representative line. For example, channel A can record chlorine, channel B can record fluorine, channel C can record sulfur, all simultaneously from the same Cl-, F-, and S-containing compound. This specific detection for each element simplifies the interpretation of a complex. multicomponent GC chromatogram, EXPERIMENTAL Gas chromatographic/microwave plasma detector analyses. Derivatized and nonderivatized reference compounds and samples were analyzed on the following chromatographic systems. Hewlett--Packard Models 7620 and 5830 gas chromatographs equipped with a flame ionization detector (FID) and the MPD were used with a 4060 splitter at the column exit between the FID:MPD, The carrier gas was helium at 25 mI/mm with an injection port temperature of 150C A 12 ft (1/8 in, o.d.) stainless-steel column packed with 20% DC-200 (methyl silicone) + 10% Bentone 34 (diatomaceous earth) on 80/90-mesh C-22A support was used for the GC/MPD analyses. The column was fitted for on-column injection of sample and programmed from 60 to 200C at 15C/mm. The microwave-sustained helium plasma detector is the Model MPD-850 from Applied Chromatography Systems Limited, Luton, England A Hewlett--Packard 3354 computer is interfaced to the amplifier array (Fig. I), Helium from the column is split between the RD detector and a transfer line to the microwave cavity. Makeup helium at 25 mI/mm is utilized after the splitter to optimize flow velocity through the transfer line. A microwave generator at 2450 Mhz supplies the energy (approximately IOU Vt) to the cavity. A vacuum system maintains the pressure of the cavity at 4--S mm Hg and oxygen or nitrogen at 0 2 mI/mm is introduced into the He carrier gas at the inlet of the cavity to prevent carbon buildup on the optical walls of the plasma tube. The plasma tube is a 0.1-mm id. 1/4-in. o.d. quartz tube 15.2 cm in length. The highly energetic helium plasma is initiated with a Tesla coil and is normally maintained without extinguishing throughout the day, Microliter quantities of solvent would completely extinguish the plasma so the cavity is fitted with a bypassvalve arrangement to shunt solvent around the quartz cell. The helium plasma is of sufficient energy to completely ionize the eluting component from the column, that is, metastable helium and energetic electrons in this plasma ionize each compound in the GC eluant to its respective elements transforming them to excited states; atomic line spectra result and the elemental response is independent of molecular configuration. The light emitted from a selected area of the plasma plume is monitored by the spectrom- eter where separate secondary slits and photomultiplier detectors are situated at the appropriate spectral positions for individual and simultaneous element monitoring. Fluorine is monitored at 685.60 nm. Large amounts of carbon generate a carbon continuum which can give an interference signal on noncarbon elemcntal channels. A "ghost" correction is electronically applied via an amplifier which constantly monitors the carbon emission and supplies a negative correction signal to the individual elemental channels. This limits the dynamic range of the detector for trace analysis if interfering components are not adequately separated from the peak of interest. The sensitivity ranges from 0.0! to 1 ng/s for various elements. The improved separation offered by GC capillary columns, coupled to the MPD, 338 PLOTTER 1 ~NAL HAGEN ET AL SPECTROMETER MICROWAVE GENERATOR POWER METER .1 VACUUM PUMP FIG I Schematic diagram of the GC/MPD_550 data system. FID, flame ionization MI'D. microwave plasma detector. Spectral lines of up to eight different elements can simultaneously via the data system detector, and be monitored would be preferred since it would permit calculating elemental ratios (C/F, C/H) in addition to specific element detection. A capillary system was briefly evaluated for the characterization of the plasma extracts. In this case, a Hewlett--Packard 5840 chromatograph equipped with a 30-m fused silica methyl silicone capillary column (J and W Scientific, Rancho Cordova, Calif.) with an electron-capture detector was employed. The temperature program was 50Cfor 1 mm, then 2C/mm for 15 mm, and 20C/mm to 280C.A 2-~lsample injection was utilized with a 70/I split ratio. Animal dosing and sample colleciion. Thirty male Charles River CD rats (Charles River Breeding Lab., Wilmington, Mass.), 9 weeks old, were divided into groups of three (nine test groups and one control group). The rats were conditioned for 24 h to individual metal metabolism cages with free access to water and fasted overnight prior to dosing. The rats were allowed free access to Purina Ground Chow (Ralston Purina Company, St. Louis, Mo.) and water immediately after dosing with a single oral dose. The I H, IH,2H,2H-perfluorodecanol (Hoechst, Frankfurt, W, Germany) was analyzed to ensure its suitability for biotrans- formation studies and found to contain less than I ppm perfluorooctanoate and less than 5 ppm inorganic fluoride. Eight grams of the alcohol was suspended in 100 ml of pure Mazola corn oil. The mb~turewas resuspended with a tissue homogenizer before each dose to assure homogeneity. The rats were weighed immediately before dosing. The volume of the alcohol- corn oil suspension used was calculated to give an alcohol dosage of 400 mg/kg and that volume was administered to each rat with a 2ml glass syringe fitted with a stainless-steel intubation tube. Groups of three rats were sacrificed by exsanguination at I, 2, 6. 12, 24, 48, 96, 144, and 480 h postdosage. The rats were anesthetized with diethyl ether and ilJ,i H,2H,2H-PERFLUORODECANOL BIOTRANSFORMATION 339 blood was drawn from the descending aorta and immediately transferred to a heparinized tube. Plasma was prepared promptly by ~~ntrifugation. Urine was collected and stored frozen. Inorganic and organic fluorine analyses. ~gorganie fluoride in rat plasma and urine samples was determined by a microprocedure using the hanging-drop fluoride electrode (10). Total fluorine was also measured with the electrode, following reductive cleavage of the organic halogen with sodium biplienyl reagent (11,12) This method, as applicable to mierodetermination of fluorine in biological samples, will be the subject of a separate publication by one of the authors (P. V.). Organic fluorine present in the sample was calculated by subtracting inorganic guoride from the total fluorine. Extraction procedure. The extraction technique was similar to the one previously used to extract perfluorooctanoate from plasma (9); however, the 80% hexane/20% diethyl ether extractant was replaced with ether, a more polar solvent, to ensure a more complete extraction of metabolites. In several cases, it was necessary to poo1 plasma to obtain sufficient quantities for the GC/ MPD analysis. One milliliter of rat plasma was pipetted into a 50-mi polypropylene tube (DuPont 3284) followed by 5 ml of water and I ml reagent grade hydrochloric acid. The contents were extracted with 7 ml of ether (Baker, anhydrous grade) and centrifuged 3 mmn (1 l,000g) with transfer of the ether phase using a polyethylene dropper (Nalgene 6219) to a 10-mI polyallomer centrifuge tube (Nalgene 3119). The extract was concentrated under N2 at 50Cto about 1 ml. The extraction was repeated, this time using 6 ml of ether, and a third time using 5 ml of ether. The total extract was then concentrated to about 1 ml and divided into two portions by splitting the sample between two 10-mi Nalgene tubes (tubes A and B). About 0.3 ml diazomethane in diethyl ether (toxic) was added to tube A. Appro- priate precautions must be taken in its usage as described in previous work (9). The method of preparation from Diazald is supplied in literature from the Aldrich Chemical Company. After intermittent swirling of the tube (5 mm), the contents were transferred with the above plastic dropper to a 1-mi volumetric flask and brought to volume with one ether rinse of the 10-mi tube. This sample was analyzed via GC/eiectron capture. `Fl'ie contents of tube B were concentrated under N to about 50 M1. and 50 M1 diazomethane2 reagent added. After 5 mm, the contents were transferred with the plastic dropper to a I-mI Reacti-Vial (Pierce 13221) and brought to 100 M1 final volume This sample was analyzed with the GC/MPD system. RESULTS AND DISCUSSION The inorganic fluoride and organic fluorine levels in plasma of rats sacrificed at different intervals are shown in Table 1. Inorganic fluoride levels in plasma and in selected urinary samples were significantly higher in the experimental animals than in the controls. This observation suggested a defluorination step in the biotransformation of the above perfluorodeeanol. Selected samples of the rat plasma were then chosen for further characterization of the organofluoro compounds present. The effect of methylation is shown in Fig. 2 where the same sample (Rat 17,24 h postdosage) was run on the GC/MPD system before and after derivatization with diazomethane, As a result of methyiation, at least three additional fluorine-containing components were observed in the chromatogram. The single peak in the nonderivatized extract eluting at 10.8 mm was shown to be the original alcohol (C3F~,CH2CH2OH)by retention-time matching. Since one can expect the biotransformation of an alcohol functional group into a carboxylic acid, the new peaks appearing only after derivatiza- 340 HAGEN ET AL tion with diazomethane were assumed to be A acids. `rwo possible metabolites of the al- cohol were suggested, namely, C8FITCOOH and C8FI7CH2COOH. Perfluorononanoic acid (C8F17COOH) was purchased from Riedel-de Haen (via the American Hoechst Corp., Somerville, N. J.). 2H.211-Perfluo- rodecanoic acid (C~F7CH2COOH) was kindly supplied by the American Hoechst Corporation. These two reference acids were TABLE I INORGANIC FLUORIDE AND ORGANIC FLUORINE ____ CONCENTRATiONS IN RAT PLASMA Rat plasma concentration 0 (pM) Time, postdosage Rat (h) No. Organic fluorine 1norganic fluoride Control <SO <5 2 <50 cS 3 <50 <5 i 4 80 25 5 140 35 6 140 40 2 in 350 40 ii i 10 30 i2 250 30 6 7 410 80 5 660 70 9 470 35 12 13 360 20 14 250 30 IS 430 30 24 16 890 65 17 960 20 Is 600 30 48 19 500 to 20 280 5 21 590 5 96 22 320 <5 23 320 5 24 250 5 144 25 250 <5 26 390 0 27 250 5 20 days 29 60 <5 30 130 <5 3) 90 <5 pM 0019 ppm. 4 I I I I I I 6 7 i S a ii 12 MINUTES Fie. 2 Effect of esterification (A) Rat 17(24 h postdosage) plasma eKtract nonesterified, (B) A after esterifloation, (C) 111,iB,2H,2H-perfluorodecariol reference. (a) lH,lH,2H,2H-perfiuorodecanol (C 1- CH C'H 01-fl, 1 (y17) 2121.2H1-per(x) perfluoroocianoate (C,F75C00); fluorodecanoate (C5F1,CH,C00), (a) unidentifted metabolite analyzed after diazomethane derivatization separately and also were used to "spike" rat plasma isolates for retention-time matching. Figure 3 illustrates the chromatograms obtained. The new peak appearing at 7.5 mm is the methyl ester of perfluorononanoic acid and the peak at 8.9 mm is the methyl ester of 2H,211-perfluorodeeanoic acid. These data, therefore, indicate that perfluorononanoic acid was not present as a metabolite while the peak at 8.9 mm matches the retention time of methyl 2H,2H-perfluorodecanoate. A German patent (13) indicated that the methyl ester of 2H,2H-perfiuorodecanoic acid is readily (as least chemically) converted to the corresponding unsaturate, C7F15CF=CHCOOCH3 via defluorination. It was speculated that this unsaturate might also be a metabolite of the alcohol and therefore was synthesized according to the above patent and shown by fluorine/NMR to he I H, i H,2H,2H-PERFLUORODF.CANOL BIOTRANSFORMATION 341 S2 C S S 7 8 9 10 MINUTES principally by volatility. The 6.8-mm peak was therefore lower boiling than the reference sample of methyl perfluorononanoate and matched the retention time of the methyl ester of perfluoroctanoic acid (PCR, Incorporated, Gainesville, Fla.). Fluorine/ NMR data confirmed the presence of perfluorooctanoic acid and 2H,2H-perfluorodecanoic acid in the rat plasma extract. Ophaug and Singer (14) have studied the metabolism of pertluorooctanoic acid in female rats and concluded that perfluorooctanoate is not further metabolized. In the Griffith and Long (15) study of amrnonium perfluorooctanoate, vastly different levels (about 100-fold) of organic fluorine in the serum of male vs female rats were reported. FiG 3. Rat plaslna extracts before and after the ad- A dition of two known reference compounds (A) Rats 8 X and 17, pooled (6 h and 24 h postdosagc, respectively), (B) A spiked with perfiuorononanoic acid (CSFI,COOH) and 2H,2H-perfluorodeeanoie acid (C0F1~CH~COOH). (b) Perfluorononanoate (C F C0O), (x) perifuorooc- 8 21H7,2H-perfluorodecanoato tanoate (C,F17C00 )~(y) (C5F1,CH5COO); (z) ~nidentifmleetdahohte I~~ 2 I ~~ 2 2 greater than 90% in purity. This unsaturate = 0 El was not well resolved from the 2H,2H-per- z fluorodecanoic acid ester on the packed col- umn. Figure 4 illustrates the chromatograms obtained for derivatized rat plasma extract before and after spiking with this unsaturated reference. As shown, the unsaturate elutes slightly ahead of methyl dihydroperfluorodecanoate. In related work, the capillary column (described earlier) provided the JLiJL 2 I 9 ~, I I I I II II 5 6 7 I 0 II II Ii 11 expected greater resolution and indicated a ~NU~ES component whose retention time matched that of the reference unsaturate (Fig. 5) Thus, the peak eluting at 10.0 mm on the packed column (Fig. 3) is not the unsaturate FiG, 4 The identIfication of the unsaturate (C,F1,CF=CHCOO-) in addition to 2H,2H-perfluorodecanoate (C F CH C00) in rat plasma extracts. 0 17 2 This represents a separate experiment where SIX rats ~re given a stogIe dose of IH1H2/i,2H-perfluorode- but perhaps a derivative thereof. canol and sacrificed 6 h later as described in the exper- The first eluting fluorine-containing cornpound at 6.8 mm appeared to be an ester since it required diazomethane derivatization to render it GC volatile. The column utilized in this case separates components iinental section, the piasma from the Six rats was pooled (A) Rat plasma, 6 h postdosago, (B) A spiked with the unsaturate (C,F CFCHCOO'). (v.) 2-Hydroperfluoro-2-decenaic 1(C5 ,F,,CF=CHCOO), (x) pertIuo- rooctanoate (CIFI,COO ), (y) 2H,2H-perfluorodeca- abate (C~F1-CH5COOi;(a) unidentified metabolite. 342 HAGEN ET AL abolic derivative is possible in that the small- est molecular metabolito (C7F15C00) ob- S C served in this work has one less CF2 group a than the starting alcohol. 0U The metabolism of fluorine-containing and perfluoro long-chain acids has not been studied in great detail. It is known that the U perfluoroheptyl chain (specifically perfluo. CF rooctanoate) is metabolically stable (14--16) while w-fluorocarboxylic acids (17) undergo fi oxidation. The high inorganic fluoride level in the plasma (Table 1) and formation of perfluorooctanoate suggest the overall re- a0 a, 0 A U 0 22 0 MINUTES Flu 5. GC/elcctron_capture detector of the pooled plasma (6 h posidosage), see Fig. 4A for the same sample on MPD and note the simplified chromatograrn using the MPD-spccific fluorine detector (A) Rat plasma, 6 h postdosage; (B) A spiked with the unsaturate (C,F,~CF=CHCOO) (w) 2-Hydroperfluoro-2- ), deconato (C,F,5CF=CHCOO), (x) perfluorooctan- onte (C~F,COO (y) 2t1,2H-peifluorodecanoate (C8F1,CH~COO).Note the unidentified fluorine cootaintng metabolite (a) is not apparent with the electroncapture detector Furthermore. while the MPD responds to the total fluorine content of a compound per the empirical formula, the electron-capture detector has dif- ferent sensitivities for different compounds. In this serIes, the Sensitivity is C~F,,CF=CHCOOCH~ > C,F1~CF5C1-I~COOCHC>,F,~CF7CH5CH,O1{ aaC U a 0 j~ The fluorine-containing biotransformation product of the alcohol eluting at 10.0 mm has not as yet been identified but it is speculated that it contains a carboxyl group since diazomethane renders it volatile for GC. The unsaturate has been identified (by GC retention data) and an additional met- I I I I I I I I 7 9 IS II Ii 13 MINUTES Fie 6. Rat plasma extracts at various times, post- dosage (A) Rat 10 (2 h postdosage, (B) Rat 9 (6 h postdosagc), (C) Rat 20 (48 h postdosage) (x) Perfluorooctanoate (C~F COOj; (y) 2H,2H-perfluoro- 15 decanoate (CsF Cl-!~COO);(z) unidentified metabohtc. 11 I H, I FI,2H,2H-PERFL.UORODECANOL BIOTRANSFORMATION 343 action, C7F5CF2CH2CH,OH -~ 1//,1H,211,2H-pErflaorodecanoI C7F5CO~+ 2 HF. periuor000Ia'041e Figure 5 illustrates the electron capture response obtained for the esterified plasma extract on the capillary column system and shows the complex mixture which is simplifled by monitoring only the fluorine content on the GC/MPD. It is obvious that an ideal system would combine the selectivity of the microwave-sustained helium plasma detector with the improved separation capabilities of a capillary system and this is under development at this time. MPD plots of the fluorine channel response are shown for various rats (Fig. 6). Note that the 2H,2H-perfiuorodecanoic acid ester is the predominant component in the 2-h-postdosage rat. The chromatograms in Fig. 6 illustrate the progressive biotransformation of the alcohol to perfluorooetanoate. Differences in the rate of biotransformation of the alcohol between rats were observed. ACKNOWLEDGMENTS We wish to acknowledge the contributions of Vicki Bunneile, Richard A Newrnark, Robert A Prokop, and Robert E Ober REFERENCES I Taves, D R. (1968) Nature (London) 217, 1050-- ~051 2 Taves, D. R (1968) p~atu,e(London) 220, 582-- 583 3. Venkateswarlu, P , Singer. L, and Armstrung, W D. (1971) Anal Riochem 42, 350--359 4. venk,deswarlu, P (1975) Anal Biochem 68, Si 2-- 521 5 Venkateswarlu. P (1975) Biochem Med 14, 368377 6 Belisle, J., and Hagen, D. F (1978) Anal Biochem 87, 545--555. 7. Guy, W S , Taves, P Rand Brey, W. S (1976) Biochemistry invoivlag Carbon-Fluorine Bonds, pp. 117-134, Amer Chem Soc , Washington. DC 8 Guy, W S. (1972) Fluorocompounds(s) of Human Plasma Analysis, Prevalence, Purification and Characterization, Ph.D thesis, University of Rochester, Rochester, N Y 9. Belislo, J and Hagen, 0 F (1980) Anal Rinchem. 101, 369 376 10 Venkateswarlu, P. (1975) din Chirn. Ac/a 59, 277--282 ii Liggett, L. M (1954) Anal Chem 26, 748--750 12 Clark, L C , Wesseler, E P, Miller, M L, and Kaplan, S (1974) ,l'f,croi,'avc Re,s 8, 320--340 13. German Patent 27 42 685 14 Ophaug, R. H, and Singer, L (1980) Proc Soc Exp Riot Med 163, 19--23 15 Grtffith, F 0 , and Long, J. E. (1980) Amer Faa' Hyg Assoc J 41, 576--583 16 Ubel, F. A., Sorenson, S 0, and Roach, 0 F. (t980) Amer IS Hvg Assoc. J `it, 5%4--5%9. 17 Saunders, B. C, (1972) in Carbon--Fluorine Com- pounds--Chemistry, Biochemistry and Biologi- cal Activities, pp 13 iS, Elsevier, Amsterdam.