228
Chem. Res. Toxicol. 1988, 1, 228-233
Metabolism of [14C]MPTP in Mouse and Monkey Implicates MPP’, and Not Bound Metabolites, as the Operative Neurotoxin Song-cheng Yang, Jan N. Johannessen, and Sanford P. Markey* Laboratory of Clinical Science, National Institute of Mental Health, Bethesda, Maryland 20892 Received February 23, 1988
The distribution, identification, and binding of [phenyl-14C]-or [methyl-14C]MPTP metabolites have been determined in brains of mouse and monkey exposed to toxic doses of MPTP. The distribution of radiolabeled metabolites was heterogeneous, with levels of MPP+ 1-100 hmol/L in dissected and homogenized monkey brain tissues. MPP+ constituted >98% of all tissue radioactivity remaining at 1-3 days in the monkey and was identified in both cortical and striatal tissue. The relevance of the 2% of unextractable (“bound”) radiolabeled metabolite was assessed in mouse brain by using pargyline or mazindol pretreatments which block dopamine depletion. The amount of binding increased rather than decreased when [phenylJ4C]MPTP was used along with pargyline or mazindol but was unchanged when [methyl-14C]MPTPwas employed. This demonstrates that bound metabolites are inversely correlated to neurotoxicity as well as being N-demethylated. Two extractable metabolites, demethylated M P T P (PTP) and l-methyl-4phenyl-2-pyridone, were found a t 30-min survival times in mouse brain and probably derive from peripheral metabolism of MPTP. At 4 h, mouse brain profiles of extractable metabolites resembled those from monkey brain, containing MPP+ as the predominant (>go%) constituent. The similarity of MPP+ concentrations in mouse and monkey brain homogenates with those concentrations of MPP+ known to produce biological effects in vitro, along with the inverse relationship between bound metabolites and neurotoxicity, supports the intermediacy of MPP+ as the operative neurotoxin.
Introduction Since the initial observations on the clinical course of the parkinsonian syndrome produced in young drug addicts by the neurotoxin l-methyl-4-phenyl-l,2,3,6-tetrahydropyridine (MPTP) (1,2), attention has been focused on the mechanism of selective nigral cell death (3-5).The temporal progression of symptoms in man or nonhuman primates develops over a period of several days, with locomotor responses deteriorating for 10-15 days, ameliorating over 6-8 weeks, and stabilizing by about 12 weeks (6). This temporal pattern suggested that MPTP or one of its metabolites might be retained in the brain for long periods. In fact, our prior studies demonstrated that MPTP is oxidized in the brain to 1-methyl-4-phenylpyridinium ion (MPP+),which is retained in monkey brain with a half-life of about 10 days (7). Further, blockade of MPTP oxidation to MPP+ eliminated neurotoxicity in mouse (7, 8) or monkey (9). However, several questions have persisted regarding MPTP-MPP+ neurotoxicity. First, the observation that covalently bound MPTP metabolites could be detected in vitro led to postulates that similar metabolites may be formed in vivo and had escaped prior detection (10). Second, the proven intermediacy of a reactive dihydro(MPDP’) pyridine, l-methyl-4-phenyl-1,Pdihydropyridine provided a mechanism for the formation of such covalently bound intermediates (21). Third, prior studies of [3H]MPTP metabolism had indicated that there were minor quantities of unidentified polar metabolites (7,12).Taken together, these observations led us to synthesize [phenyl-14C]MPTPin order to test the relevance of MPTP metabolite binding to neurotoxicity and to determine the
identity of brain metabolites other than MPP+.
Materials and Methods Radiolabeled [phenyl-14C]MPTP (42.6 mCi/mM) was prepared from [14C8]bromobenzeneas previously described (13)and was assayed by high-performance liquid chromatography (HPLC) as being >96% radiochemically pure. [methyl-14C]MPTP (50 mCi/mM) was purchased from ARC (St. Louis, MO) and was assayed by HPLC as being >98% radiochemically pure. Monkey Experiments. Radiolabeled M P T P (200-280 WCi) was injected into the right internal carotid artery of three female rhesus monkeys (weight, 3-4 kg) housed in Horsfal isolation chambers. Two monkeys treated with [phenyl-14C]MPTP were killed 1 and 3 days after injection as described previously (13). The third monkey treated with [methyl-14C]MPTP was killed 10 days after injection. The total amount of M P T P used (0.16 to 0.38 mg/kg) produces unilaterally parkinsonian monkeys when administered by carotid infusion (14). Mouse Experiments. [phenyl-’%]MPTP diluted with MPTP was injected ip into adult male C57B16 mice in glass metabolic chambers with gas traps (13). After specific times, mice were killed by spinal dislocation and decapitated, and the brains were removed and frozen. All 14C injected was quantitatively recovered by washing the chamber with 0.1 N HC1, accumulating ‘%02 evolved in traps, and digesting portions of weighed and homogenized tissues. Sample Preparation. Brain tissue samples were weighed and homogenized in four volumes of distilled water. Aliquots of homogenate were treated with 10 volumes of ethanol, vortexed, and centrifuged a t 3000 rpm for 15 min. The ethanol supernatant was transferred to another tube, and the tissue precipitate reextracted with another 10 volumes of ethanol. For assessing bound radioactivity, this procedure was repeated five times and then the residual pellet processed as described for whole tissue. The combined ethanol extracts were acidified with several drops
T h i s article not subject to U.S. Copyright. Published 1988 by t h e American Chemical Society
Metabolism of MPTP Table I. Extractable Radioactivity of Monkey 359 Right Front Cortex extractable % radioactivity remainder ethanol extraction dpm" 70 extracted 1 17902 f 290 92.26 2 1000 f 16 97.41 66.5 3 140 f 4 98.13 27.8 4 57 f 1 98.43 16.0 5 29 f 2 98.58 9.6
" Values are dpm/20 mg of tissue, mean f SEM (n = 3). Table 11. Unextractable Radioactivity of Monkey Cortex unextractable radioactivity % total survival dpm/20 mg tissue monkey time, day of tissuea radioactivity 1.42 359 1 276 f 11 (n = 3) 1.80 308 3 154 f 3 (n = 2) 350 10 116 f 14 (n = 3) 17.80 aMean f SEM, determined by Protosol digestion of residue following five ethanol extractions. of 1.5 N HC1 and evaporated to dryness under nitrogen. For HPLC, dried ethanol extracts were dissolved in the appropriate mobile phase, centrifuged a t 3000 rpm for 10 min, and filtered through Centrifree (Amicon) filters at 4000 rpm for 20 min. For gas chromatography-mass spectrometry (GC-MS), dried ethanol extracts from 20-100 mg of tissue were dissolved in 0.5 mL of 0.4 N HC104,vortexed vigorously, and centrifuged at 3000 rpm for 10 min. Supernatants were transferred to screw-capped glass tubes to which 0.1 mL of saturated Na2C03and 0.05 mL of acetic anhydride were added. The tube was vortexed for several seconds, 2 mL of ethyl acetate added, and the tube vortexed thoroughly for 15 s and centrifuged a t 3000 rpm for five min. The ethyl acetate extracts were transferred to another glass tube, evaporated to dryness under nitrogen, and reconstituted in 25 pL of ethyl acetate for injection. High-Performance Liquid Chromatography. Separations were performed by using a 5-wm C-18 reverse-phase column (25 cm) with a mobile phase of acetonitrile/O.l M sodium acetate (60/40, v/v) containing 0.1% triethylamine (final p H of sodium acetate-triethylamine adjusted to 5.6 before mixing with acetonitrile). Fractions were collected a t 15-9 intervals at a flow rate
Chem. Res. Toxicol., Vol. 1, No. 4 , 1988 229 of 1 mL/min and then mixed with 4-mL aliquots of scintillation fluid (Biofluor, New England Nuclear) for counting. Injected radioactive metabolites were quantitatively recovered from HPLC. Gas Chromatography-Mass Spectrometry. Analyses were performed by using a Finnigan 3200 electron ionization quadrupole mass spectrometer with a Teknivent data system. The glass GC column (1.8 X 2 mm i.d.) was packed with 3% Poly 1-110 on 80/lOO mesh Gas Chrom Q (Alltech). T i s s u e Radioactivity. Weighed portions or aliquots of aqueous tissue homogenates were digested in glass tubes with Protosol (New England Nuclear) a t 55 " C for 16 h using 1 mL for 50-100 mg of tissue. The solutions were decolorized with 0.1 mL of 30% hydrogen peroxide a t 55 O C for 30 min, cooled, and mixed with scintillation fluid (Econofluor-2, New England Nuclear).
Results Bound Metabolites. Exhaustive, repeated extractions of resuspended monkey cortex homogenate with fresh ethanol indicated that 1.5-2.5% of the total tissue radioactivity was unextractable (Table I). Cortical tissue was used in these studies because regional dissection did not show any differences in distribution, and the quantity of cortical tissue permitted replicate determinations. Unextractable radioactivity persisted with respect to monkey survival time, which probably reflects the clearance of MPP+ relative to a covalently bound metabolite (Table 11). Similarly in the mouse, small but consistent amounts of bound radioactivity were measured throughout the brain 4 h after [phenyl-14C]-(Table 111) or [methylJ4C]MPTP (Table IV). To determine whether this bound radioactivity was related to neurotoxicity, groups of mice were protected by pretreatment with agents which are known to block neurotoxicity. Either the monoamine oxidase inhibitor pargyline (10 mg/kg i.p., 16 h prior) or the dopamine uptake inhibitor mazindol (30 mg/kg, 15 min prior) was administered before radiolabeled MPTP using a regimen which reduced significantly MPTP-induced dopamine depletion (Tables I11 and IV). Unextractable radioactivity in the pargyline or mazindol pretreated mouse brains was variable but was significantly higher in several brain regions when [phenyl-14C]MPTPwas used (Table
Table 111. Unextractable "Bound" [ ~ ~ ~ I I ~ ~ - ' ~ Cin] Mouse M P T PBrain 4 h after pretreatment (time) nonea pargyline (16 h)" brain areas dpm/2O mg % total dpm/20 mg % total 17 f 8.5 1.24 f 0.6 186 f 70 cerebellum 15.24 f 5.6 1.7 f 1.7 0.12 f 0.12 393 f 1lC 35.02 f 4.5 medulla & pons cerebral cortex & olfactory bulb, striatum 27 f 7 1.49 f 0.21 691 f 217b 46.1 f 15.3 46 f 29 hypothalamus, ventral tegmentum 22 f 14 1.14 f 0.7 3.15 f 1.9 hippocampus, midbrain, thalamus, cortex 20 f 7 0.99 f 0.26 42 f 20 2.5 f 1.2
Treatment mazindol (15 min)" dpm/20 mg % total 396 f 268 18.12 f 10.6 214 f 139 14.98 f 10.2 101 f 7* 6.15 f 1.3 228 f 130 17.12 f 12.6 139 f 40 7.26 f 2.6
"Values are dpm/20 mg of tissue as mean f SEM (n = 3) and expressed as % total tissue radioactivity. [Phenyl-14C]MPTP(546 pCi/kg, 2.21 mg/kg, ip) was given 16 h after pargyline (10 mg/kg, ip) or 15 min after mazindol (30 mg/kg, ip). Indicates p < 0.05 student's t test. cIndicates p < 0.01 Student's t test. Table IV. Unextractable "Bound" [metl~yl-'~C]MPTP in Mouse Brain 4 h after Treatment pretreatment (time) none" pargylinen (16 h) mazindoP brain area dpm/20 mg 70 total dpm/20 mg 70 total dpm/20 mg cerebellum 142 f 17 10.25 f 1.3 111 f 16 13.2 f 5 3 182 f 18 medulla & pons 113 f 6 6.15 f 0.47 107 f 21 9.8 f 0.7b 96 f 5 cerebral cortex & olfactory bulb, striatum 145 f 4 7.7 f 0.1 121 f 9 155 f 18 11.6 f 1.6 hypothalamus, ventral tegmentum 8f4 1.3 f 1 13 i 6 1.9 f 0.9 19 f 1 2 125 f 6 27.3 f 17 139 f 4 6.5 f 0.2 202 f 15 hippocampus, midbrain, thalamus, cortex
(15 min) % total 23.9 f 2 3 14.9 f lC 19.7 f 1.4b 7.3 f 3.6 20.3 f O.lc
" Values are dpm/2O mg tissue as mean f SEM (n = 3) and expressed as % total tissue radioactivity. [methyl-'*C]MPTP (347 NCi/kg, ip) was given after pargyline (10 mg/kg, ip) or mazindol (30 mg/kg, ip). bIndicates p < 0.05 Student's t test. "Indicates p < 0.01 Student's t test.
230 Chem. Res. Toxicol., Vol. 1, No. 4,1988
Yang et al.
Table V. Extractable Radioactivity in Monkey Brain tissue concentration" injected doseb right frontal cortex right caudate pCi mg dpm/20 mg pmol/L dpm/20 mg pmol/L ~
monkey
survival, day
359 308 350
1 3 10
279.5 215.3 200
1.13 0.874 0.692
20403 9504 703
10.8 5 0.37
~~~
187613 6645
99.2 3.5
Concentration calculated from sDecific activitv of MPTP and assuming 1 mg of tissue occupies a volume of 1 pL. Injected into right internal carotid artery. 1mo 1
0
500
24m
1
?
i L,
-'
1 2w
0 1 0
I
I
20
40
Fraction Number
-
I
60 15 seclfraction
MPP+
J
4 hrs
1
80
Figure 1. Radiochemical detection of HPLC eluate of ethanol extracts from two monkeys treated with [14C]MPTP.Upper panel is the chromatogram of extract from cortical tissue 1 day after treatment with [phenyl-14C]MPTP;lower panel is the chromatogram of extract from caudate 10 days after treatment with [methyl-14C]MPTP. 111). In contrast, there was no difference between control or pretreated mice in unextractable radioactivity when [nethyl-14C]MPTPwas injected (Table IV). Extractable Metabolites. In monkey brain homogenate extracts, a single radioactive metabolite which cochromatographed with MPP+ was detected on HPLC (Figure 1). Identical profiles were obtained from monkeys sacrificed after 1,3, or 10 days in extracts of either cortical or striatal regions. Only MPP+ was detected whether methylJ4C- or phenyl-14C-labeled MPTP was administered. The distribution of extractable radioactivity changed with survival time of the monkey, with higher concentration persisting in caudate than in cortex (Table V). The HPLC profile of radiolabeled metabolites was determined from extracts of brains of mice sacrificed 30 min or 4 h after i.p. [phenyl-14C]MPTPbecause of the rapid elimination of radiolabel from mouse brain in contrast to monkey brain. There were at least six radiolabeled compounds detected at the shorter survival time (Figure 2a). In addition to MPTP and the previously identified MPP+
0-
0
I
I
I
I
20
40
60
80
FRACTION NUMBER
Figure 2. Radiochemical detection of HPLC eluate of ethanol extracts from mice treated with [phenyl-'%]MPTP. Upper panel is the chromatogram of extract from whole brain homogenate obtained 30 min after treatment; lower panel is the chromatogram of extract from whole brain homogenate 4 h after treatment. Identifications are based upon coelution of standards and mass spectrometry of the extracts.
Figure 3. Mass spectra of components detected in acetylated extract of mouse brain (upper) and authentic standards (lower) of l-acetyl-4-phenyl-l,2,3,6-tetrahydropyridine (left) and 1methyl-4-phenyl-2-pyridone (right). (7), two components were identified by comparison with authentic standards and by gas chromatography-mass spectrometry (Figure 3a, b) as l-methyl-4-phenyl-2-
Chem. Res. Toxicol., Vol. 1, No. 4, 1988 231
Metabolism of MPTP Table VI. Distribution and Concentration of Radioactivity in Mouse Tissue after [pl~enyl-'~C]MPTP~ 48 h
tissue brain liver kidney adrenal lung heart spleen intestine blood carcass excreted or exhaled
t issuebof dpm/mg tissueb ( n = 2) % totalc (n = 1) % totalc expt 1 expt 2 expt 1 expt 2 186 1.01 15 15 0.09 0.04 844 721 281 380 227 117 946 138 206
12.48 3.41 0.03 0.84 0.51 0.24 27.23 1.07 48.85 4.34
38 158 317 144 103 143 222
371 219 649 172 97 132 192
58
55
0.57 0.62 0.08 0.39 0.22 0.15 7.36 0.07 9.26 81.18
2.49 0.45 0.09 0.19 0.05 0.05 3.69 0.03 4.91 88.01
"Administered MPTP doses were: 24 mg/kg, 142 pCi/kg to 30 min mouse; 30 mg/kg, 159 and 298.6 pCi/kg to mice in expt 1 and expt 2, respectively. bRadioactivity is the average of two determinations of tissue homogenates. Total radioactivity was calculated by summing all radioactivity found in each experiment.
pyridone and 4-phenyl-1,2,3,6-tetrahydropyridine (PTP). Four hours after treatment with [phenyl-14C]MPTP,the HPLC pattern of radioactive metabolites in brain extracts had changed considerably (Figure 2b). The major metabolite was identified as MPP+, and a more polar minor metabolite remains unidentified. The ethanol extracts from each mouse were also measured in order to ascertain whether mazindol or pargyline pretreatment altered the profile of extractable MPTP metabolites. The amount of MPP+ remaining in mouse brain tissue 4 h after [methylJ4C]MPTP was very small, but greatest in the mice without pretreatment (total 1723 dpm), approximately one-third less in the pargyline (1172 dpm), and one-eighth in the mazindol pretreated mice (203 dpm). The corresponding values obtained from a [phenyl-14C]MPTPexperiment were 1172 (control), 478 (pargyline), and 285 dpm (mazindol). These values were derived from single quantitative HPLC measurements of profiles similar to Figure 2b (data not shown). However, to obtain sufficient data for statistical comparisons, whole brain ethanol extracts were used without chromatographic fractionation. In these whole extracts, there was significantly less activity ( p < 0.01) in pargyline and mazindol pretreated mice than in controls receiving [ methyl-14C]MPTP. For methyl-14Ccontrols were 1548 dpm/20 mg f 608 SD, pargyline 882 f 367, and mazindol, 698 f 255. Distribution of Radiolabel in Mice. The tissue distribution of [phenyl-14C]MPTPin mice was measured in order to ascertain whether there is any organ which accumulated or concentrated MPTP metabolites. Table VI indicates that MPTP metabolites are not concentrated and accumulated in any tissue over a 48-h period, although higher concentrations of label were retained by adrenal relative to other tissues. Most significantly, the activity persisting in brain was negligible. The urinary excretion and exhalation of radiolabeled metabolites has been described elsewhere (13).
Discussion There are several alternative hypotheses regarding the mechanism of MPTP-induced neurotoxicity. Since the initial findings that MPTP is metabolized to MPP+, that relatively high levels of MPP+ persist in primate brain, and that neurotoxicity can be blocked either by inhibiting the oxidation of MPTP to MPP+ or the uptake of MPP+ into
dopaminergic terminals, the necessary intermediacy of MPTP oxidation has been firmly established. The enzymatic oxidative process involves the partly oxidized reactive intermediate MPDP+, and thus the involvement of covalently bound or other neurotoxic metabolites have been postulated ( 1 0 , l l ) . Trevor et al. have suggested that either MPDP+ (15, 16) or a ring-opened metabolite of MPDP+ such as 3-phenylglutaraldehyde (17) could bind covalently to a macromolecule and cause cell death as a consequence. The present study addresses these alternative hypotheses by quantifying covalent binding after 14C-labeled MPTP, determining the distribution and identity of metabolites, and confirming that the amount of MPP+ present in primate brain is sufficient for the production of neurotoxic effects. Successive extractions of aqueous brain homogenates with ethanol demonstrated the presence of "bound" radioactivity. These successive ethanol extractions recovered decreasing percentages of residual radioactivity from monkey brain (Table I) or mouse brain (data not shown). If extraction efficiency rather than binding were the determining factor, then repeated extractions would have continued to recover a constant percentage of the remainder from the resuspended pellet. The fact that extraction efficiency changed from 92% to 9.6% from the first to the fifth extract argues that the radioactivity persisting in the pellet differs from that which is extractable. Further, the concentration of bound radioactivity in monkey brain was persistent (Table 11). Because [phenylJ4C]MPTP was used in the 1and 3 day study, but methyl-14C in the monkey sacrificed after 10 days, no interpretation of the nature of the binding was possible, and further investigation was pursued in the mouse to permit larger numbers of animals to be studied by using both phenyl-14C and methyl-14Cradiolabels. With mice, it was possible to test whether conditions which block or reduce MPTP's depletion of striatal dopamine would reduce radiolabeled metabolite binding and thus correlate positively with neurotoxicity. In fact, as observed in Tables I11 and IV, the bound radioactivity either increased significantly after [phenyl-14C]MPTP (Table 111) or was unchanged in [methyl-lVIMPTP (Table IV) experiments in which mice were sacrificed 4 h after treatment. This experimental protocol was chosen in order to maximize the detection of the small amount of bound radioactivity present and to discern whether there was regional distribution by using the dissection procedures of Glowinski and Iversen applied to the mouse (18). Most of the administered radioactivity had left the mouse brain after 4 h, and there was considerable variance in the remainder. The amount of bound metabolite increased after pargyline or mazindol pretreatment in the [phenyl-14C]MPTP experiment both in absolute terms and as a percentage of total tissue radioactivity. This demonstrated that radiolabel binding was not merely a fixed portion of the injected radioactivity but that it differed qualitatively from the extractable radioactivity. The increase in binding following pargyline pretreatment reached significance in two brain areas after [phenyl-14C]MPTP but not after [methylJ4C]MPTP. Thus, it is probable that one of the bound metabolites is N-demethylated. Similarly, after mazindol pretreatment, binding increased significantly in at least one brain region with [ p h e n ~ l - ~ ~but C ] - not [methyl-14C]MPTP. Because the mechanisms by which pargyline and mazindol produce these increases in N-demethylated metabolite binding are not related, or are inversely related, to the phenomenon of neurotoxicity, these studies were not further pursued.
232 Chem. Res. Toxicol., Vol. 1, No. 4, 1988
The question of whether other, previously unidentified, minor metabolites may correlate with neurotoxicity was also addressed by using [14C]MPTP. In extracts of monkey brain from animals treated with either [methyl-14C]-or Iphenyl-14C]-MPTP,only MPP+ was detected in cortical or striatal regions, either 1,3, or 10 days after treatment. Reported observations of small quantities of metabolites more polar than MPP+ in studies of 3H-labeled MPTP must have been the result of 3H-exchangeor label loss (7, 12). Brain metabolites were extracted and characterized at shorter survival times in mice. At 30 min, MPTP, MPP+, l-methyl-4-phenyl-2-pyridone, and PTP were identified in extracts. The latter two compounds have been reported previously from in vitro liver homogenate metabolism of MPTP (11, 19) and probably appear in brain secondary to peripheral formation in the mouse liver. Neither metabolite is observed in in vitro brain homogenates (20). The des-methyl compound PTP has been tested in mice and does not have MPTP-like neurotoxicity (21,22), but no toxicity data have been published on the pyridone metabolite. Consistent with the observation that peripheral metabolism of MPTP is reflected in the brain is the distribution and concentration of MPTP radiolabeled metabolites throughout the mouse after [phenyl-14C]MPTP (Table VI). While at 30 min, MPTP has been widely distributed, radiolabeled metabolites were largely (81-88%) eliminated after 48 h. The radioactivity remaining was most concentrated in adrenal, kidney, liver, intestine, with lowest organ concentrations found in brain. I t might be argued that the experimental evidence discussed thus far is inconclusive proof that a bound or other MPTP metabolite is not involved in neurotoxicity. However, in the process of collecting this data, it was possible to gather further quantitative data on tissue levels of MPP+, demonstrating that it alone is present in sufficient concentration to produce the observed effects. These levels, when augmented by those reported previously and compared to concentrations found to produce cytotoxic effects in vitro, further implicate MPP+ as the operative neurotoxin. In the monkey data in Table V, tissue concentrations of 0.3-99 pM are indicated for survival times of 1-10 days. Because 98% of this radioactivity is MPP’, those concentrations are relevant to MPP+ concentrations which produce toxicity. From the literature it is possible to calculate striatal concentration in monkeys of 2.7 pM, 10 days after injection of 0.19 mg/kg (7,23), 1.7 pM, 13 days after injection after 4 daily 0.4 mg/kg doses (24), and 33 pM, 3 days after 4 iv doses of 2 mg/kg (25). These bulk tissue concentrations of 1-33 pM MPP+ in monkey striata are indicative of higher localized concentrations at the cellular level because of the nonuniform, heterogeneous distribution of radiolabel within tissue which is visable autoradiographically (7) and active uptake and concentration of MPP+ within mitochondria (26). The peak concentrations of MPP+ found in mouse brain are of a similar magnitude to those measured in monkey, but not as persistent. MPP+ was found to be
