Document 8KVZdYoJDOR401nkYv890Kwy

8 PP ANALYTICAL BIOCHEMISTRY 118, 336 343 (1981) AR226-1440 Characterization of Ftuorinated Metabolites by a Gas ChromatographicHelium Microwave Plasma Detector-- The Biotransformation of 1H, 1H,2/-/,2H-PerfluorodecanoI to Perfluorooctanoate D o n a l d F. H a g e n ,* J o n B e l is l e ,* J a m e s D . J o h n s o n ,! a n d P . V en r a t e s w a r l i4 *Central Research Laboratories, ^Riker Laboratories, Inc, and \Commercial Chemicals Division, 3M Company, 3M Center, S t Paul, Minnesota 55144 Received June 16, 1981 A gas chromatographic technique utilizing a microwave-sustained helium plasma detector (GC/MPD) was developed to study the biotransformation of l/f,lH,2ff,2//-pernuorodecanol (CaF,7CHzCH20H) in adult male rats. The metabolic products resulting from a single oral dose of the alcohol were isolated from the blood plasma by an extraction technique, denvatized, and characterized by the GC/M PD system with continuation by fluorine NMR Four fluorinecontaining metabolites were detected by the fluorine-specific channel of the element-selective MPD and one of these was shown to be perfluorooctanoate (CnF'n C O O ) Appearance of perfluorooctanoate as a metabolite and an additional observation of concomitant elevation in plasma and urinary inorganic fluoride suggests that the biolransformation of the above alcohol to the perfluorooctanoate involves defluorination of the CF2group adjacent to the CH3 group m the parent compound. In recent years, a variety of analytical techniques for differentiating and quantitat ing the inorganic fluoride and organic fluo rine in biological samples has been described by Taves (1,2), Venkateswarlu et al. (3-5), and Belisle and Hagen (6). However, char acterization and determination of the spe cific organic compound(s) in the samples is more informative than determination of total organic fluorine. Based on fluorine NMR data, Guy et a l (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 mtcroanalytical method, based on GC1with electron-capture detector, was reported by Belisle and Hagen (9) specifi cally for perfluorooctanoate in blood plasma though the analysis gave no evidence for the presence of perfluorooctanoate It should be emphasized that this research (9) describes 1Abbreviations used1GO, gas chromatogiaphy, MPD, microwave plasma detector, FID. flame ionization de tector an analytical method for the determination of perfluorooctanoate and is not a study of perfluorooctanoate levels in normal human plasma since only a few public donor samples were analyzed. Gas chromatography is well suited for the complex matrices of biological 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 suf 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 l//,l//,2//,2if-perfluorodecano! in male rats using the element-specific gas chroma tography/microwave plasma detector (GC/ MPD) system Specific characterization of fluorinated compounds will be described in this work. The detector plasma (not to be confused 0003-2697/81/180336-08*02 00/0 Copyright 1981 by Academic Press, Inc All rights of reproduction m uny form resected 336 RECEIVED OPPT NCIC 2003 Oct 12 11:50AM !H, l//,2//.2tf-PERFLUORODECANOL BIOTRANSFORMATION 337 with blood plasma) is an extremely energetic ionization source of metastable atoms, mol ecules, 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, S, etc.) has a discrete emission spectrum and, using the described detector, one can monitor a rep resentative line. For example, channel A can record chlorine, channel B can record fluo rine, channel C can record sulfur, all si multaneously from the same C1-, F-, and S-containing compound. This specific de tection for each element simplifies the in terpretation 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 chromato graphic systems. Hewlett-Packard Models 7620 and 5830 gas chromatographs equipped with a flame ionization detector (FD) and the MPD were used with a 40-60 splitter at the column exit between the FID.-MPD, The carrier gas was helium at 25 ml/min with an injection port temperature of !50C 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/M PD analyses. The column was fitted for on-column injection of sample and pro grammed from 60 to 200C at 15C/min. The microwave-sustained helium plasma detector is the Model MPD-850 from Ap plied Chromatography Systems Limited, Luton, England A Hewlett-Packard 3354 computer is interfaced to the amplifier array (Fig. 1). Helium from the column is split between the FID detector and a transfer line to the microw'ave cavity. Makeup helium at 25 ml/min is utilized after the splitter to optimize flow velocity through the transfer line. A microwave generator at 2450 MHz supplies the energy (approximately 100 W) to the cavity. A vacuum system maintains the pressure of the cavity at 4-5 mm Hg and oxygen or nitrogen at 0 2 ml/min is intro duced 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 i.d. 1/4-in. o.d. quartz tube 15.2 cm in length. The highly energetic he lium plasma is initiated with a Tesla coil and is normally maintained without extinguish ing 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 suf ficient energy to completely ionize the elut ing 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 trans forming 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 pho tomultiplier detectors are situated at the ap propriate spectral positions for individual and simultaneous element monitoring. Flu orine is monitored at 685.60 nm. Large amounts of carbon generate a carbon con tinuum which can give an interference signal on noncarbon elemental channels. A "ghost" correction is electronically applied via an amplifier which constantly monitors the car bon emission and supplies a negative cor rection 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 0.01 to 1 ng/s for various ele ments. The improved separation offered by GC capillary columns, coupled to the MPD, 338 HAGEN ET AL Fig i Schematic diagram of the GC/MPD-850 data system, FID, flame ionization detector, and MPD, microwave plasma detector, Spectral lines of up to eight different elements can be monitored simultaneously via the data system would be preferred since it would permit calculating elemental ratios (C /F, C /H ) in addition to specific element detection, A cap illary system was briefly evaluated for the characterisation of the plasma extracts. In this case, a Hewlett-Packard 5840 chro matograph 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 50C for 1 min, then 2C/min for 15 min, and 20C/min to 280C, A 2-f sample injection was utilized with a 70/1 split ratio. Animal dosing and sample collection. 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 l/f,l/f,2/T,2/f-perfluorodecanoI (Hoechst, Frankfurt, W, Germany) was an alyzed to ensure its suitability for biotrans formation studies and found to contain less than 1 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 mixture was resus pended 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 vol ume was administered to each rat with a 2 ml glass syringe fitted with a stainless-steel intubation tube. Groups of three rats were sacrificed by exsanguination at 1,2, 6, 12, 24, 48, 96, 144, and 480 h postdosage. The rats were anesthetized with diethyl ether and 1/Altf,2#,2//-PERFLUORODECANOL BIOTRANSFORMATION 3J9 blood was drawn from the descending aorta and immediately transferred to a heparined tube. Plasma was prepared promptly by CCIitnfugalion. Urine was collected and stored frozen, Inorganic and organic fluorine analyses. Inorganic fluoride in rat plasma and urine samples was determined by a microproce dure using the hanging-drop fluoride elec trode (10). Total fluorine was also measured with the electrode, following reductive cleav age of the organic halogen with sodium bi phenyl reagent (11,12) This method, as ap plicable to microdetermination 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 sam ple was calculated by subtracting inorganic fluoride 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 sev eral cases, it was necessary to pool plasma to obtain sufficient quantities for the GC/ MPD analysis. One milliliter of rat plasma was pipetted into a 50-ml polypropylene tube (DuPont 3284) followed by 5 ml of water and 1 ml reagent grade hydrochloric acid. The con tents were extracted with 7 ml of ether (Baker, anhydrous grade) and centrifuged 3 min (ll,000g) with transfer of the ether phase using a polyethylene dropper (Nalgene 6219) to a 10-ml polyallomer centri fuge tube (Nalgene 3119). The extract was concentrated under N 2 at 50C to 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-ml 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 sup plied in literature from the Aldrich Chem ical Company. After intermittent swirling of the tube (5 min), the contents were trans ferred with the above plastic dropper to a 1-ml volumetric flask and brought to volume with one ether rinse of the 10-ml tube. This sample was analyzed via GC/electron cap ture. The contents of tube B were concentrated under N2 to about 50 1, and 50 diazo methane reagent added. After 5 min, the contents were transferred with the plastic dropper to a 1-ml Reacti-Vial (Pierce 13221) and brought to 100 gl final volume This sample was analyzed with the GC/MPD system, RESULTS AND DISCUSSION The inorganic fluoride and organic fluo rine levels in plasma of rats sacrificed at different intervals are shown in Table 1. In organic fluoride levels in plasma and in se lected 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 perfluorodecanol. Selected sam ples 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 post dosage) was run on the GC/MPD system before and after derivatization with diazo methane. As a result of methylation, at least three additional fluorine-containing compo nents were observed in the chromatogram. The single peak in the nonderivatized ex tract eluting at 10.8 min was shown to be the original alcohol (CgFnCPUCHjOH) by retention-time matching. Since one can ex pect 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 acids. Two possible metabolites of the al cohol were suggested, namely, C8FnCOOH and CgF|7CH2COOH. Perfluorononanoic acid (CsF,7COOH) was purchased from Riedel-de Haen (via the American Hoechst Corp., Somerville, N. J.). 2/7,2/APcrfluorodecanoic acid (CsF |7CH 2COOH) was kindly supplied by the American Hoechst Corporation. These two reference acids were TABLE 1 Inorganic Fluoride and Organic Fluorine Concentrations in Rat Plasma Rat plasma concentration Time, postdosage (hi Ral No. Control 2 6 12 24 43 96 144 20 days 1 2 3 4 5 6 10 It 12 7 8 9 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 30 31 " I /xM = 0 0 1 9 ppm. Organic fluorine <50 <50 <50 80 140 140 350 110 250 410 660 470 360 250 430 890 960 600 500 280 590 320 320 250 250 390 250 60 130 90 Inorganic fluoride <5 <5 <5 25 35 40 40 30 30 80 70 35 20 30 30 65 20 30 10 5 5 <5 5 5 <5 J 5 <5 <5 <5 MINUTES Fig. 2 Effect of esterification (A) Rat 17 (24 h post dosage) plasma extract tionestenfied, (B) A after ester ification, (C) 1H, 17f,2//,27i-perfiuorodecanol reference, (a) lt/,lf/',2//',2if-perfluorodeeanol (Oat'iiCH-CILOH), (x) perfluorooctanoate (CTF,sCOCr); (y) 2H,7Ji-pufluorodecanoate (CsF,7CHjCOO~), (a) unidentified me tabolite analyzed after diazomethane denvatization separately and also were used to ``spike" rat plasma isolates for retention-time matching. Figure 3 illustrates the chromatograms ob tained. The new peak appearing at 7.5 min is the methyl ester of perfluorononanoic acid and the peak at 8.9 min is the methyl ester of 2H,2//-peTfluoTodecanoic acid. These data, therefore, indicate that perfluoronon anoic acid was not present as a metabolite while the peak at 8.9 min matches the re tention time of methyl 2//,2//-perfluorodecanoate. A German patent (13) indicated that the methyl ester of 27/,27/-perfluorodecanoic acid is readily (as least chemically) con verted to the corresponding unsaturate, C7F[sCF=CHCOOCH3 via defluorination. It was speculated that this unsaturate might also be a metabolite of the alcohol and there fore was synthesized according to the above patent and shown by fluorine/NMR to be JJT,l//,2//,2/F-PERFLUORODF,CANOL BIOTRANSFORMATION 341 principally by volatility. The 6.8-min peak was therefore lower boiling than the refer ence 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 2//,2J/-perfluorodecanoic acid in the rat plasma extract. Ophaug and Singer (14) have studied the metabolism of perfluorooctanoic acid in fe male rats and concluded that perfluorooctanoate is not further metabolized. In the Griffith and Long (15) study of ammonium perfluorooctanoate, vastly different levels (about 100-fold) of organic fluorine in the serum of male vs female rats were reported. Fui 3. Rat plasma extracts before and after the ad dition of two known reference compounds (A) Rats 8 and 17, pooled (6 h and 24 h postdosage, respectively), (B) A spiked with perfluorononanoic acid (CMF nCOOH) and 27f,2/f-perfluorodecanoic acid (C8F,7CHjCOOH). (b) Perfluorononanoale (CjFnCOO"), (x) perfluorooctanoate (CjF isCOO ); (y) 2f/,2//-perfluorodccanoate {CsF|jC H 2COO"); ( z ) unidentified metabolite greater than 90% in purity, This unsaturate was not well resolved from the 2H,2H-psr~ 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 unsatu rated reference. As shown, the unsaturate elutes slightly ahead of methyl dihydroperfluorodecanoate. In related work, the capil lary column (described earlier) provided the expected greater resolution and indicated a component whose retention time matched that of the reference unsaturate (Fig. 5). Thus, the peak eluting at 10.0 min on the packed column (Fig. 3) is not the unsaturate but perhaps a derivative thereof. The first eluting fluorine-containing com pound at 6.8 min appeared to be an ester since it required diazomethane derivatization to render it GC volatile. The column utilized in this case separates components Fig, 4 The identification of the unsaturate (C iF|,C F=C H CO O ") in addition to 2//,27/-pcrfiluorodecanoate {CjF,7CH2COO' ) in rat plasma extracts. This represents a separate experiment where six rats were given a single dose of i //, I//2/y,2//-perfluorodecanol and sacrificed 6 h later as described in the exper imental section, the plasma from the six rats was pooled (A) Rat plasma, 6 h postdosage, (B) A spiked with the unsaturate (C7Fi5CF=C H C O O ` ). (w) 2-Hydroperfluoro-2-decenaie (C7F MC F =C H C 00~), (x) perfluo rooctanoate (C,F!<COO ), (y) 2H,2//-perfluorodecanoate (C8F r.CH2COO"); (z) unidentified metabolite. 342 HAGEN ET AL abolic derivative is possible in that the small est molecular metabolite (CyFijCOO- ) ob served in this work has one less CP2 group than the starting alcohol. The metabolism of fluorine-containing and perfluoro long-chain acids has not been studied in great detail. It is known that the perfluoroheptyl chain (specifically perfluorooctanoatc) is metabolically stable (14-16) while w-fluorocarboxylic acids (17) undergo /3oxidation. The high inorganic fluoride level in the plasma (Table 1) and formation of perfluorooctanoate suggest the overall re- F ig S. GC/elcctron-capture detector of the pooled plasma (6 h postdosage), see Fig. 4A for the same sam ple on MPD and note the simplified chromatogram using the MPD-spccific fluorine detector (A) Rat plasma, 6 h postdosage: (B) A spiked with the unsaturate (C \F,,C F = CHCOO ) (w) 2-Hydroperfluoro-2decenate (CjFuC F=C H CO O _), (x) perfluorooctanoate (C^FuCOO ), (y) 2//,2//-peifluorodecanoate (CsFnCHjCOO- ). Note the unidentified fluorine con taining metabolite (z) is not apparent with the electroncapture detector Furthermore, while the MPD responds to the total fluorine content of a compound per the em pirical formula, the electron-capture detector has dif ferent sensitivities for different compounds. In this senes, the sensitivity is C.F .CF= CHCOOCH^ > C ,F,,CFjCH3COOCH, > C jFhCFjCH jCH jOH The fluorine-containing biotransforma tion product of the alcohol eluting at 10.0 min 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 idiwurs, Fio 6. Rat plasma extracts at various times, post dosage (A) Rat 10 (2 h postdosage, (B) Rat 9 (6 h postdosage), (C) Rat 20 (48 h postdosage) (x) Perfiuorooctanoate {C,F^COO"); (y) 2 /f;2//-perfUiorodecanoate (C sF itC H C O O - ); ( z) unidentified metabo lite. IH,lW,2//,2/7-PERFl.UORODECANOL BIOTRANSFORMATION 343 action, C7F |5CF2CH2CH 7OH - I//, IW,2W,3//-pcrfluorodeL4tiul C7F,sCOi + 2 HF. pcrfluornoctancMie 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 simpli fied 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 detec tor with the improved separation capabilities of a capillary system and this is under development at this time. MPD plots of the fluorine channel re sponse are shown for various rats (Fig. 6). Note that the 2H,1 ff-perfluorodecaaoic acid ester is the predominant component in the 2-h-postdosage rat. The chromatograms in Fig. 6 illustrate the progressive biotransfor mation of the alcohol to periluorooctanoate. Differences in the rate of biotransformation of the alcohol between rats were observed. ACKNOWLEDGMENTS We wish to acknowledge the contributions of Vicki Bunnclle, Richard A Newmark, Robert A Prokop, and Robert E Ober REFERENCES 1 Taves, D R. (1968) Nature (London) 217, 1050-- 2 Taves, D. R (1968) \alure (London) 220, 582-- 583 3. Venkateswarlu, P , Singer, L , and Armstrong, W D. (1971) Anal Biochem 42, 350-359 4. Venkateswarlu, P (! 975)/in/ Biochem 68,512 521 5 Venkateswarlu. P (1975) Biochem Med 14,363 377 6 Bclisle, J., and Hagen, D. F (1978) Anal Biochem 87, 545-555. 7. Guy, W S , Taves, D R , and Brey, W. S (1976) Biochemistry Involving 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. Belisle, J , and Hagen, D F (1980) Anal Biochem. 101, 369 376 10 Venkateswarlu, P, (1975) Clm Chim. A d a 59, 277-282 11 Liggett, L. M (1954) Anal Chem 26,748-750 12 Clark, L C , Wesseler, E P , Miller, M L , and Kaplan, S (1974) Microvasc Re.s 8, 320-340 13. German Patent 27 42 6B5 14 Ophaug, R. H , and Singer, L (1980) Proc Soc Exp Biol Med 163, 19-23 15 Griffith, F D , and Long, J. E. (1980) Amer Ind Hyg Assoc J 41, 576-583 16 Ubel, F. A., Sorenson, S D , and Roach, D E. (.1980) Amer [nd Hyg Assoc, J 41, 5&4-S&9. 17 Saunders, B. C. (1972) in Carbon-Fluorine Com pounds--Chemistry, Biochemistry and Biologi cal Activities, pp 13 15, Elsevier, Amsterdam.