Neurochemical Research, Vol. 16, No. 9, 1991, pp. 965-974
`
`An Overview of 'Y-Hydroxybutyrate Catabolism: The Role
`of the Cytosolic NADP+-Dependent Oxidoreductase EC
`1.1.1.19 and of a Mitochondrial Hydroxyacid-Oxoacid
`Trans hydrogenase in the Initial, Rate-Limiting Step in
`This Pathway
`
`Elaine E. Kaufrnan1 and Thomas Nelson1
`
`(Accepted April 22, 1991)
`
`'Y-Hydroxybutyrate (GHB) is a naturally occurring
`compound present in micromolar concentration in both
`brain (1,2) and in peripheral tissues (3). This endoge(cid:173)
`nous compound is remarkable in that pharmacological
`doses of 200-500 mg/kg produce marked behavioral and
`electroencephalographic changes (4), a profound de(cid:173)
`crease in cerebral glucose utilization (5), an increase in
`striatal dopamine levels ( 6) and a decrease in body tem(cid:173)
`perature (7). High doses of GHB have also been reported
`to protect neurons (8) and intestinal epithelium (9) against
`cell death resulting from experimental ischemia. Behav(cid:173)
`ioral changes are not seen with doses of less than 30 mg/
`kg, but low doses stimulate the release of prolactin, growth
`hormone and cortisol (10,11), and doses of 5-10 mg!kg
`result in an increase in body temperature (12). These
`actions, and the discovery of high affinity binding sites
`for GHB in the central nervous system (13), suggest that
`GHB may have a biological function. Both the origin of
`endogenous GHB and its catabolism are, therefore, of
`considerable interest.
`This review will cover the early work on the deg(cid:173)
`radative pathway for GHB and the discovery of a dual
`pathway for the initial step in the oxidative catabolism
`of GHB. The factors which regulate the activity of the
`
`1 Laboratory of Cerebral Metabolism, National Institute of Mental Health,
`United States Public Health Service, Department of Health and Hu(cid:173)
`man Services, Building 36, Room lA-05, Bethesda, Maryland 20892
`* Special issue dedicated to Dr. Louis Sokoloff.
`Abbreviations used in this paper: GHB, ')'·hydroxybutyrate; SSA,
`succinic semialdehyde; DTI, dithiothreitol.
`
`enzymes in these pathways, and as a consequence, reg(cid:173)
`ulate tissue levels of GHB are also discussed.
`Walkenstein et al. (14) established that GHB is
`largely disposed of, in vivo by oxidation to C02 and
`water. These investigators could not find the 14C label
`from GHB in succinate in urine, but they did find that
`the label could be trapped in hippuric acid in the urine
`of animals treated with sodium benzoate as might be
`expected if GHB were undergoing 13-oxidation. They
`therefore proposed that GHB was metabolized by 13-ox(cid:173)
`idation (14). Mohler et al. (15) and Doherty et al. (16),
`however, assayed citric acid cycle intermediates in the
`tissues of animals given [14C]GHB and demonstrated
`that the carbon skeleton of GHB indeed does enter the
`citric acid cycle as succinate rather than as acetyl-CoA
`as would be expected if GHB were being oxidized through
`the 13-oxidation pathway.
`They proposed the following pathway:
`1) GHB ~ succinic semialdehyde
`2) Succinic semialdehyde --7 succinate
`3) Succinate ~ ~ ~ ~ Co2 + H20
`The discovery of a metabolic disease in which GHB
`and succinic semialdehyde (SSA) are markedly elevated
`(17) in both blood and urine due to a block in SSA
`dehydrogenase (18) added evidence to support a degra(cid:173)
`dative pathway in which GHB is oxidized to SSA, which
`in turn is oxidized to succinate.
`At the time the pathway shown above was pro(cid:173)
`posed, it was known that the enzymes of the citric acid
`cycle catalyzed the reactions in step (3) and that SSA
`
`965
`
`0364-3190/91/0900-0965$06.50/0 © 1991 Plenum Publishing Corporation
`
`Ranbaxy Ex. 1017
`IPR Petition - USP 9,050,302
`
`

`
`966
`
`Kaufman and Nelson
`
`dehydrogenase (the enzyme missing in patients with OHB
`aciduria (18)) catalyzed the reaction in step (2). Step (1)
`is always depicted as an essential part of this scheme.
`However, at the time this pathway was proposed, an
`enzyme or enzymes which could catalyze this step had
`not been identified.
`Isolation of a Cytosolic GHB Dehydrogenase. Two
`unusual oxidoreductases, one cytosolic (19) and the other
`mitochondrial (20), that catalyze the oxidation of OHB
`to SSA have now been isolated. The cytosolic enzyme,
`which will be referred to as ORB-dehydrogenase in this
`review, was first purified to homogeneity from hamster
`liver (19) and was found to be an NADP+-dependent
`oxidoreductase. A study of the substrate specificity of
`purified ORB-dehydrogenase revealed that D-glucuron(cid:173)
`ate and L~gulonate, the product of D-glucuronate reduc(cid:173)
`tion, were also good substrates (19). The physical
`characteristics, as well as the substrate and inhibitor
`specificity of this enzyme, indicate that the ability to
`catalyze the oxidation of OHB probably represents a pre(cid:173)
`viously unreported activity for the NADP+-dependent
`oxidoreductase (EC 1.1.1.19) commonly known as D(cid:173)
`glucuronte reductase (21). This enzyme may also be
`identical to the group of enzymes categorized in a 1985
`review by Cromlish et al. (22) as "ALR-1", the high
`Km aldehyde reductase or L-hexonate dehydrogenase.
`Although the oxidation of OHB catalyzed by this
`ORB-dehydrogenase proceeds at an easily measurable
`rate when assayed in vitro under optimal conditions, the
`very low activity found in the in vitro system under
`conditions simulating those in the cytosol raises the
`question of how, or indeed whether, this enzyme cata(cid:173)
`lyzes the oxidation of OHB in vivo. An answer to this
`question may have been found when it was discovered
`that OHB dehydrogenase could catalyze the reduction of
`o-glucuronate coupled to the oxidation of GHB (23) as
`shown below:
`OHB + NADP+ :;;:= SSA + NADPH + H+
`o-glucuronate + NADPH + H + :;;:= L-gulonte +
`NADP+
`The overall or "coupled" reaction is:
`GHB + o-glucuronate :;;:= SSA + L-gulonate
`When the kinetic constants for the coupled system were
`determined, it was found that they were more favorable
`to oxidation of GHB under conditions present in cytosol
`of most tissues th~n were those for the uncoupled sys(cid:173)
`tem.
`
`The time course of the coupled reaction (Figure 1)
`in which both products, SSA and NADPH, were mea(cid:173)
`sured, shows that in the presence of an adequate con(cid:173)
`centration of o-glucuronate, only a very small net amount
`of NADPH is formed even though SSA formation is
`
`. 10
`
`. 08
`
`i . 06
`
`"' :!l
`.
`';;
`" . 04
`E ,_
`
`0
`
`. 02
`
`0
`
`0
`
`SSR
`
`""
`
`NRDPH
`
`I
`
`10
`
`12
`
`14
`
`16
`
`18
`
`20
`
`Time
`( m 1 n)
`Fig. 1. Time course of succinic semialdehyde (SSA) and NADPH
`formation in the presence of D-glucuronate (23). The reaction mixture
`contains 10 mM GHB, 1.0 mM D-glucuronate, 0.1 mM NADP+, 80
`mM phosphate buffer, pH 7.6, enzyme and sufficient water to bring
`the volume to 1.0 mi. SSA and NADPH were determined as previously
`described (19).
`
`proceeding rapidly. This is in contrast to the control
`reaction mixture without glucuronate in which NADPH
`and SSA are formed in stoichiometric amounts (19). These
`results indicate that NADPH is being used for the re(cid:173)
`duction of D-glucuronate at the same rate at which it is
`being produced by the oxidation of GHB. This would
`account for the low steady state level of NADPH seen
`in Figure 1. The effect of increasing concentrations of
`D-glucuronate on the rate of oxidation of GHB to SSA
`in the presence of limiting amounts of NADP+ and in(cid:173)
`hibitory amounts of NADPH is shown in Figure 2. Un(cid:173)
`der these conditions, 2 mM o-glucuronate increased SSA
`formation 8-fold.
`Other important changes in the kinetic constants for
`this reaction occur in the presence of D-glucuronate (Ta(cid:173)
`ble 1). The Km (4.5 x 10- 4M) for the coupled reaction
`is fiVe-fold lower and the V max (1.52 J.Lmol/min/mg pro(cid:173)
`tein) is 1.8 times higher than in the uncoupled reaction.
`The effects of coupling and changes in pH on the rate
`of degradation of OHB are, however, more accurately
`described by changes in k (the first order rate constant
`for the reaction) than they are by changes in V max' The
`concentration of GHB in the tissues is much lower than
`Km and under these conditions the quantity V max!Km pro(cid:173)
`vides a good approximation of k (24). In the coupled
`reaction, the rate constant for GHB degradation, V maxi
`Km, is increased 9-fold as compared to the 1.8-fold in(cid:173)
`crease in V max at saturating concentrations of OHB and
`NADP+ (Table 1). The effect of o-glucuronate on the
`rate of the reaction is much greater at the very low sub-
`
`

`
`Enzymes Which Catalyze the Initial Step in GHB Catabolism
`
`967
`
`. 40
`
`.35
`
`~ .30
`
`~ .25
`' d:
`ill
`~ .15
`
`. 20
`
`. 10
`
`.05
`
`.4
`
`. 8
`
`I. 2
`
`!.&
`
`[0-Glucuronate) (mM)
`
`Fig. 2. The effect of D-glucuronate concentration on the rate of con(cid:173)
`version of GHB to succinic semialdehyde (SSA) in a reaction mix
`containing NADPH and a low concentration of NADP+ (23). Reaction
`mixture: 10 mM GHB, 0.01 mM NADP+, 0.02 mM NADPH, 80 mM
`phosphate, pH 7.6, enzyme, o-glucuronate as indicated and water to
`1 mi. The assay for SSA was carried out as previously described (19).
`
`strate concentrations found in vivo than in the standard
`in vitro assay where V max conditions are used. In the
`coupled reaction the Km for NADP+ is decreased from
`2 x l0- 5 to 1.4 x 10-6M and the inhibition produced
`by NADPH (Ki = 7 x 10- 6M) has been eliminated
`(Figure 3). GHB can now be oxidized in the presence
`of an otherwise extremely inhibitory concentration of
`NADPH. All of the kinetic constants· for the coupled
`reaction are closer to the tissue concentration range shown
`in Table I (25) than are those for the uncoupled reaction.
`The rate of GHB oxidation is pH-dependent. Earlier
`work had shown that the pH optimum for the cytosolic
`oxido-reductase under V tnax conditions was 9.0 (19), but
`
`at a more physiological pH (7.0-7.2) the enzyme was
`only half as active. The pH optimum was dependent on
`GHB concentration in both the coupled and uncoupled
`reactions and in both cases was above 8.0 when satu(cid:173)
`rating concentrations of substrate were used (26). As the
`concentration of GHB decreases toward levels found in
`vivo the pH optimum for the coupled reaction shifts toward
`pH 7.0 (26). A plot of V max!Krn against pH (Figure 4)
`shows that, at substrate concentrations near those found
`in vivo, the pH optimum approaches the intracellular
`pH, i.e. 7.5 for the uncoupled reaction and 7.0 or lower
`for the coupled reaction (26). Vayer et al. (27) subse(cid:173)
`quently reported a pH optimum of 8.0 under different
`conditions from those described above. In their system
`the oxidation of GHB catalyzed by the cytosolic oxi(cid:173)
`doreductase was coupled to both the reduction of D(cid:173)
`glucuronate and the transamination of SSA to form GABA.
`GHB dehydrogenase is inhibited by a number of
`products of intermediary metabolism (Table II) which
`includes the ketone bodies, a'.-ketoglutarate and branched
`0'.-ketoacids derived from amino acid degradation as well
`as degradation products of phenylalanine (26). As has
`been found with certain aldehyde reductases (28,29),
`anticonvulsants such as barbiturates, diphenylhydantoin
`and valproate are good inhibitors of GHB dehydrogenase
`(30). In addition, GHB dehydrogenase is inhibited by
`salicylates (30)-
`GHB dehydrogenase like lysozyme, ribonuclease
`and a number of other proteins (26), may contain disul(cid:173)
`fide bridges which are essential for its activity. It is
`inhibited by compounds such as 13-mercaptoethanol and
`dithiothreitol (DTI), which can reduce disulfide bonds
`(26). DTT has the most pronounced effect; addition of
`2.5 mM DTT produces an 85% inhibition of the activity
`
`Table I. The Effect of D-Glucuronate on the Kinetic Constants for -y-Hydroxybutyrate (GHB), NADP+, and NADPH
`
`GHB
`NADP+
`NADPH
`GHB
`V m,./K,GHB
`
`Kinetic Constants
`
`Uncoupled assay
`
`2.3 X 10·3 M
`2 X 10-5 M
`7 X 10-6 M
`0.83
`0.36
`
`Coupled assay
`
`4.5 X lQ-<Mh
`1.4 X 10-6 M<
`No inhibition
`1.52"
`3.38
`
`Tissue concentration
`
`(range between brain,
`liver, kidney, muscle)"
`0.1-5 X 1Q-S M
`2-11 X 10" 6 M
`1-30 X 10-5 M
`
`"The tissue concentrations of GHB are from reference (3). The molar concentration of NADP+ and NADPH in the various tissues were calculated
`from data taken from reference (25).
`•1 mM D-glucuronate.
`<2 mM D-glucuronate.
`dlJ.mol/min/mg protein.
`•first order rate constant when [S] < < Km
`Data in this table are from reference (23)
`
`

`
`+NADPH
`
`Table II. Kinetic Constants for the Inhibition of GHB(cid:173)
`Dehydrogenase (26,30)
`
`Kaufman and Nelson
`
`CONTROL
`
`968
`
`"-
`
`E
`
`'
`~
`IS
`!!' 1/v
`
`~
`
`ill
`6
`
`-E
`
`
`"-
`
`Inhibitor
`
`DL-13-Hydroxybutyrate
`Acetoacetate
`a-Ketoglutarate
`p-Hydroxyphenylpyruvate
`Phenylacetate
`Pyruvate
`a-Ketoisovalerate
`a-Ketoisocaproate
`DL-a-keto-13-methyl n-Valerate
`Valproate
`Salicylate
`
`K; (mM)
`
`Uncoupled
`reaction
`
`Coupled
`reaction
`
`4.4
`
`0.6
`
`0.4
`
`0.2
`
`3.9
`3.0
`1.1
`1.0
`0.5
`7.0
`0.33
`0.17
`0.06"
`0.057
`0.115
`
`Assays for the coupled reaction and the uncoupled reaction were car(cid:173)
`ried out as described (26).
`"It should be noted that the K; has only been reported for the DL(cid:173)
`mixture.
`
`not been previously inhibited by sulfhydryl reducing
`agents.
`When the activity of GHB dehydrogenase is being
`measured, the inclusion of reducing compounds such as
`13-mercaptoethanol or DTT in the assay mixture as has
`been reported by some laboratories (27) could lead to
`incorrect results.
`Identification of a Mitochondrial Transhydrogenase
`Which Catalyzes the a-Ketoglutarate-Dependent Oxi(cid:173)
`dation of GHB. The discovery of GHB dehydrogenase
`made it possible to complete a sequence of reactions
`leading from GHB to C02 and H20. Development of a
`polyclonal antibody to the GHB dehydrogenase then made
`it possible to determine whether or not there were ad(cid:173)
`ditional enzymes in the cytosol or in other subcellular
`fractions that could catalyze the initial, and probably
`rate-limiting, step in the oxidation of GHB to C02 (31).
`The antibody, which inhibited ""' 95% of the ability of
`the cytosolic fraction to oxidize GHB to SSA, failed to
`inhibit approximately 60% of this activity in kidney ho(cid:173)
`mogenate (Figure 5). These results suggested that there
`was at least one additional enzyme which catalyzed the
`conversion of GHB to SSA. The mitochondrial fraction,
`which had previously been shown to lack GHB dehy(cid:173)
`drogenase activity (19), could oxidize [14C]GHB to 14C02
`and could catalyze the pyridine nucleotide-independent
`oxidation of GHB to SSA (31). The mitochondrial en(cid:173)
`zyme was subsequently isolated and shown to be a hy(cid:173)
`droxyacid-oxoacid transhydrogenase which catalyzed the
`following reaction (20):
`GHB + a-ketoglutarate ~ SSA + a-hydroxyglutarate.
`
`18
`
`20
`
`30
`
`40
`
`50
`
`Fig. 3. The effect of D-glucuronate on the kinetics of the oxidation
`of GHB with NADP+ as the variable substrate (23). In the absence of
`D-glucuronate, NADPH formation was measured; in the presence of
`D-glucuronate, succinic semialdehyde (SSA) formation was measured.
`• = control, • = control with 0.02 mM NADPH, o = control with
`2.0 mM D-glucuronate, o = control with 0.02 mM NADPH and 2.0
`mM D-glucuronate.
`
`A.
`
`vmax , Krn
`
`0.24
`
`0.22
`
`coupled
`
`0.20
`
`- .::
`E
`y: ~ 0.18
`~ .s 0.16
`E ~
`>
`
`0.14
`
`0.12
`
`0.10
`7.0
`
`uncoupled
`
`7.5
`pH
`
`8.0
`
`E
`~ 0.8
`E
`
`0 :f'-~/\
`~ 2E 0.6 >if
`OOL_j
`
`'
`
`I
`
`8
`pH
`
`10
`
`Fig. 4. The effect of pH on the ratio of V mflm and on V mox for the
`oxidation of GHB by GHB dehydrogenase. The data for Figure 4A
`were taken from reference (26). The data in Figure 4B are from ref(cid:173)
`erence (19). VmJ~ = the first order rate constant, k, when [S] < <
`Km•
`
`of the purified enzyme and 2.0 mM DIT inhibits 84%
`of the activity found in rat liver cytosol. The inhibition
`produced by DTT can be partially reversed by the ad(cid:173)
`dition of H20 2 or completely reversed by oxidized glu(cid:173)
`tathione. Preincubation of GHB dehydrogenase with 2.5
`mM DTT before it is added to the reaction mixture (and
`thereby diluted 50 fold) does not inhibit the enzyme.
`Under these conditions, the sulfhydryl groups are prob(cid:173)
`ably reoxidized by molecular oxygen. Little or no stim(cid:173)
`ulatory effect is seen when compounds such as H20z,
`oxidized glutathione and cystamine, all of which can
`oxidize sulfhydryl groups, are added to enzyme that has
`
`

`
`Enzymes Which Catalyze the Initial Step in GHB Catabolism
`
`969
`
`100
`
`80
`
`r
`t=
`~ 60
`u
`
`"" ~ "" :; 40
`"
`
`f-
`
`1> e RAT
`0 A HAMSTER
`
`20
`
`4
`
`16
`
`32
`
`RATIO OF ANTIBODY TO CYTOSOL OR TO HOMOGENATE (V:V)
`
`Fig. 5. Titration of GHB dehydrogenase activity in cytosol and ho(cid:173)
`mogenate with the antibody to GHB dehydrogenase (31). The con(cid:173)
`version of [V4C]GHB to 14C02 was used as a measure of activity in
`the experiments in which homogenate was used. When cytosol was
`used, the conversion of GHB to SSA was measured spectrophoto(cid:173)
`metrically. The cytosol and homogenate used in this experiment were
`derived from the same amount of kidney. Antibody was added to the
`reaction mixture in the amounts indicated.
`
`The oxidation of GHB by the mitochondrial enzyme was
`found to be completely dependent on the presence of a(cid:173)
`ketoglutarate. The reaction was reversible. The mito(cid:173)
`chondrial enzyme also catalyzed the conversion:
`a-hydroxyglutarate + SSA ~a-ketoglutarate + GHB
`with the two products being formed in stoichiometric
`amounts. The substrate specificity is shown in Table III
`and the kinetic constants for the principal substrates in
`Table IV (20). The assumption that this enzyme is a
`transhydrogenase was confirmed when it was shown by
`gas chromatographic-mass spectroscopy that a deuterium
`on the hydroxyl-bearing carbon of one of the optical
`isomers of DL-'y-deutero-')'-hydroxybutyrate was trans(cid:173)
`ferred to the ketone-bearing carbon of a-ketoglutarate.
`This hydroxyacid-oxoacid transhydrogenase which has
`
`been found in the soluble fraction of mitochondria from
`liver, kidney and brain has been partially purified (20).
`Comparison of the Cytosolic and Mitochondrial En(cid:173)
`zymes which Oxidize GHB to SSA. With the discovery
`of the mitochondrial transhydrogenase it became appar(cid:173)
`ent that there was a dual pathway for the initial step in
`the oxidative pathway for GHB. Both cytosolic GHB
`dehydrogenase and the mitochondrial enzyme are oxi(cid:173)
`do:-eductases, but of remarkably different types. The cy(cid:173)
`tosolic enzyme is an NADP+-dependent dehydrogenase,
`whereas the mitochondrial enzyme is a pyridine nucleo(cid:173)
`tide-independent, a-ketoglutarate-dependent transhydro(cid:173)
`genase. These two enzymes, despite some striking
`differences, have one property in common: the activity
`of each of these enzymes is regulated by coupling the
`oxidation of GHB, the hydroxyacid, to the simultaneous
`reduction of an oxoacid. Though GHB dehydrogenase
`can function in the uncoupled state, conditions in the
`cytosol of most tissues (Table I) would not be favorable
`to the uncoupled reaction. The mitochondrial enzyme,
`on the other hand, has an absolute requirement for an
`oxoacid, which suggests that the metabolism of GHB by
`the mitochondrial enzyme would depend on the steady
`state levels of citric acid cycle intermediates and of a(cid:173)
`ketoglutarate in particular. The relative activities (V max
`and v) of these two enzymes in brain and kidney are
`shown in Table V. The activities of the two enzymes
`were measured under V max conditions; the rate of GHB
`oxidation (v) at average tissue GHB concentrations could
`then be calculated (Table V). How well these values
`correspond to measurements of GHB catabolism in vivo
`can be determined by examination of the half-life of
`GHB in brain. Doherty et al. (16) and Mohler et al. (15)
`have both reported a half-life (t11z) for labeled GHB in
`brain of 5 min. If one assumes that the rate of disposal
`of GHB follows first order kinetics, then from the fol-
`
`Table III. The Relative Rates of the Transhydrogenase Reaction With Other Hydroxyacids and Oxoacids (20)
`
`Hydroxy acid
`(with SSA as co-substrate)
`
`D-a-Hydroxyglutarate
`L-a-Hydroxyglutarate
`DL-13-Hydroxybutyrate
`L-13-Hydroxybutyrate
`D-13-Hydroxybutyrate
`DL-a-Hydroxybutyrate
`D-Lactate
`L-Lactate
`D-Malate
`L-Malate
`
`Relative
`Rate
`
`100
`0
`43
`45
`0
`0
`0
`1.2
`0
`0
`
`The data in this table are derived from reference (20)
`
`Oxoacid
`(with GHB as co-substrate)
`
`a-Ketoglutarate
`a-Ketoadipate
`Oxalacetate
`Pyruvate
`a-Ketobutyrate
`Acetoacetate
`13-Ketoglutarate
`13-Ketoadipate
`
`Relative
`Rate
`
`100
`24
`18.5
`8.0
`4.3
`0
`0.5
`0
`
`

`
`970
`
`Kaufman and Nelson
`
`Table IV. Kinetic Constants for the Partially Purified Hydroxyacid(cid:173)
`Ox?acid Transhydrogenase (20)
`
`Substrate
`
`"f·hydroxybutyrate
`L-(3-hydroxybutyrate
`a-Hydroxyglutarate
`a-Ketoglutarate
`Succinic semialdehyde
`
`M
`3.0 X 10- 4
`3.0 X 10-3
`4.2 X 10-4
`1.8 X 10- 4
`4.6 X 10-6
`
`lowing expression k, the first order rate constant, can be
`calculated.
`
`k = 2.3 log 2 I t112
`At a brain concentration of 2.3 nmol/g the rate of dis(cid:173)
`posal of GHB, calculated from the tt12 would be 0.32
`nmol!min/g. The combined rate of catabolism in brain
`by the two enzymes was found to be 0.18 nmol!min/g
`brain (Table V). Since the rate of disposal of GHB cal(cid:173)
`culated from the tissue half-life must include both the
`rate of catabolism of GHB and the rate of loss from brain
`by transport to the plasma, the rates calculated from the
`half-life in brain and from the combined activity of the
`two catabolic enzymes (Table V) are in reasonable
`agreement.
`Effects of Modulators of GHB Dehydrogenase In
`Vivo. One approach used to determine the magnitude of
`the contribution of GHB dehydrogenase to the metabo-
`
`lism of GHB in vivo was to give rats compounds which
`inhibit the activity of the enzyme in vitro (30). If GHB
`dehydrogenase plays a significant role in the metabolism
`of GHB in a tissue, inhibition of the enzyme in vivo
`would be expected to result in increased tissue concen(cid:173)
`trations of GHB. Two metabolic intermediates, pheny(cid:173)
`lacetate and a-ketoisocaproate, and two drugs, valproate
`and salicylate, all potent inhibitors of GHB dehydroge(cid:173)
`nase with Ki values in the range of 10- 4 to 10-5 M,
`were administered to rats. The effects of a-ketoisoca(cid:173)
`proate and phenylacetate (Table VI) and valproate and
`salicylate (table VII) on tissue contents of GHB varied
`depending on both the tissue and the compound. All of
`these compounds increased the concentration of GHB in
`one or more of the tissues assayed, in some cases by
`200%. These results are in agreement with those of Snead
`(32) who found elevated levels of GHB in brain follow(cid:173)
`ing administration of sodium valproate. Although each
`of these compounds also inhibits SSA dehydrogenase
`(30), which catalyzes the second step in the degradative
`pathway, it is very unlikely that the degree to which this
`enzyme would be inhibited is sufficient to elevate the
`concentration of SSA enough to cause the increase in
`tissue levels of GHB. The values fqr the Ki for valproate,
`for GABA transaminase (29) and for SSA dehydroge(cid:173)
`nase (29,30) are approximately 10- 2 to 10- 3M. These
`Ki values are several orders of magnitude higher than
`the K; for GHB dehydrogenase (30). From these data it
`was calculated (30) that sodium valproate at a tissue
`
`Table V. Relative Activity of the Degradative Enzymes for GHB in Brain and Kidney
`
`Enzyme
`
`GHB dehydrogenase
`
`hydroxyacid-oxoacid
`transhydrogenase
`
`GHB dehydrogenase
`
`hydroxyacid-oxoacid
`trans hydrogenase
`
`vrn<X
`nmol!min/gram
`tissue
`
`Brain
`10.1
`
`16.6
`
`Kidney
`324
`
`172
`
`Km
`(M)
`
`2.5 X 10-3
`4.5 X 1Q-4:j:
`
`3 X 10-4
`
`2.2 X 10- 3
`4.5 X 10- 4:j:
`
`3 X 10- 4
`
`Tissue
`concentration
`nmol GHB/
`gram tissue
`
`2.3
`
`28.4
`
`Calculated rate of
`oxidation of GHB at
`average tissue
`concentrations
`v = nmol!min/gram•
`
`0.01
`0.051 (coupled)
`
`0.126
`
`4.2
`19.2 (coupled)
`
`14.9
`
`b
`Vm"" • [S] h
`·
`[S]
`= su strate concentra Ion.
`*The velocity (v) has been calculated from the equatiOn: v =
`t.
`w ere
`K., + [S]
`V max data is from reference ( 41 ), K.n data are from references (23) .and (20), tissue concentrations reference (3).
`tKm for GHB when the oxidation of GHB is coupled to the reductiOn of D-glucuronate (23).
`
`

`
`Enzymes Which Catalyze the Initial Step in GHB Catabolism
`
`971
`
`Table VI. Effects of a-Ketoisocaproate and of Phenylacetate on Tissue Levels of GHB (30)
`
`Saline
`infusion
`
`nmol GHB/g
`tissue
`
`2.6 ± 0.3 (5)
`27.8 ± 3.2 (5)
`22.2 ± 3.0 (6)
`
`Tissue
`
`Brain
`Kidney
`Muscle
`
`a-Ketoisocaproate infusion
`
`Phenylacetate infusion
`
`nmol GHB/g
`tissue
`
`3.2 ± 0.1 (3)
`55.5 ± 8.2 (4)b
`46.4 ± 9.9 (4)b
`
`Percent of
`control
`
`123
`198
`209
`
`nmol GHB/g
`tissue
`
`6.1 ± 1.0 (4)"
`18.4 ± 1.3 (4)b
`16.7 ± 7.1 (4)
`
`Percent of
`control
`
`235
`66
`75
`
`Phenylacetate (1.0 M) was given intravenously as an initial bolus of 1.5 ml followed by a constant infusion of= 2.0 ml!h for 2 h. a-Ketoisocaproate
`(0.5 M) was given intravenously as an initial bolus of 0.6 ml followed by a constant infusion of = 2.0 ml!h for 2 h. At the end of the infusion
`the animals were killed and the tissues were removed and assayed for GHB as described (30). All values are means ± SEM, numbers of animals
`are in parentheses.
`"p < 0.01
`bp < 0.05
`
`Table VII. Effects of Salicylate and of Valproate on Tissue Levels of GHB (30)
`
`Saline
`(control)
`
`nmol GHB/g
`tissue
`
`2.9 ± 0.5 (6)
`2.6 ± 0.2 (4)
`34.5 ± 5.7 (6)
`30.0 ± 4.1 (4)
`
`Tissue
`
`Brain
`
`Kidney
`
`Salicylate
`
`Valproate
`
`nmol GHB/g
`tissue
`
`5.8 ± 0.5 (4)"
`
`52.8 :t 2.5 (5)b
`
`Percent of
`control
`
`200
`
`153
`
`nmol GHB/g
`tissue
`
`Percent of
`control
`
`3.7 ± 0.2 (4)"
`
`23.7 ± 3.8 (4)
`
`142
`
`79
`
`Sodium valproate (100 mg/kg i.p.) was given 2 hr prior to decapitation; Sodium salicylate (500 mg/kg i.p.) was given 1 hr prior to decapitation.
`Tissues were removed and assayed as described (30). All values are means ± SEM, numbers of animals are in parentheses.
`"p<0.01
`bp<0.05
`
`level of 0.3 mM, a reasonable tissue concentration ex(cid:173)
`pected in therapeutic use, would produce an 85% inhi(cid:173)
`bition of GHB dehydrogenase but only a 5% inhibition
`of SSA dehydrogenase and a negligible inhibition of
`GABA transaminase. It has been reported that sodium
`valproate inhibits the conversion of SSA to GHB by an
`aldehyde reductase (EC 1.1.1.2) (29) which may be sim(cid:173)
`ilar to, or identical to, GHB dehydrogenase (EC 1.1.1.19).
`However, inhibition of GHB synthesis would not explain
`the increased tissue levels of GHB seen following
`administration of sodium valproate.
`The kinetic constants which have been determined
`for GHB dehydrogenase in vitro suggest that, in the un(cid:173)
`coupled state, the enzyme would be more likely to syn(cid:173)
`thesize GHB from SSA than to oxidize it. Indeed the
`identical enzyme, n-glucuronic reductase, is classified
`as an aldehyde reductase. It is only in the coupled state
`that the kinetic constants become favorable to the oxi(cid:173)
`dation of GHB catalyzed by this oxidoreductase under
`in vivo conditions. In kidney, D-glucuronate is likely to
`be one of the aldehydes coupled to the oxidation of GHB.
`
`In this organ mya-inositol oxygenase, the enzyme that
`synthesizes D-glucuronate, has been found to exist as a
`complex with n-glucuronate reductase (GHB dehydro(cid:173)
`genase) and is believed to transfer the newly formed n(cid:173)
`glucuronate directly to the oxidoreductase (33). Further
`proof that this oxidoreductase is probably identical to
`GHB dehydrogenase came from the finding that 95% of
`the n-glucuronate reductase activity in kidney cytosol
`and 80% of that activity in brain cytosol could be titrated
`with the polyclonal antibody to GHB dehydrogenase (31).
`What is not known is the extent to which the reaction is
`coupled in vivo and whether or not there are aldehydes
`other than D-glucuronate that can serve this role.
`To test the extent to which the enzyme is coupled
`in vivo, the effect of i.v. administration of n-glucuronate
`on the half-life of [14C]GHB in plasma has been exam(cid:173)
`ined (30). The half-life of 14C-labeled GHB in plasma
`reflects the overall rate of GHB metabolism by the body
`tissues since urinary loss of the compound is insignifi(cid:173)
`cant (34). The effect of the administration of L-gulonate,
`the product of D-glucuronate oxidation and a competitive
`
`

`
`972
`
`Kaufman and Nelson
`
`inhibitor, was also examined (30). The results shown in
`Figure 6 demonstrate that the half-life of GHB in plasma
`is shortened by D-glucuronate. The ability of o-glucu(cid:173)
`ronate to decrease the half-life of GHB in plasma indi(cid:173)
`cates that GHB dehydrogenase is not completely coupled
`in vivo. The half-life of GHB in plasma is lengthened
`by L-gulonate as would be expected from a competitive
`inhibitor of the enzyme. The effect of inhibitors of GHB
`dehydrogenase on tissue content and plasma half-life of
`GHB and the effect of D-glucuronate on the plasma half(cid:173)
`life both suggest that this enzyme may play an important
`role in the metabolism of GHB in vivo.
`NADPH is another endogenous inhibitor of GHB
`dehydrogenase (Ki = 7 x 10- 6 M) in the uncoupled
`state. The results of the experiment in which o-glucu(cid:173)
`ronate was administered to rats have shown that the re(cid:173)
`action was not completely coupled since the rate of
`metabolism could be stimulated by D-glucuronate. In the
`absence of maximal coupling, changes in the ratio of
`NADPH/NADP+ such as would occur with low ambient
`oxygen concentration might be expected to decrease the
`activity of GHB dehydrogenase and thereby increase the
`tissue concentration of GHB. The data in Table VIII
`show that in rats exposed to 5.6% oxygen for 2 hours
`the concentration of endogenous GHB is increased in
`brain from 3.0 to 10.7 nmol/gm of brain (a 3.6-fold
`increase) and in kidney from 24.6 to 57.0 nmol/gm of
`kidney (a 2.3-fold increase). This change in the redox
`
`:O.JOO
`
`~ :,000
`<
`
`;2:
`~
`-,
`
`100
`
`COMPOUND
`
`INFUSFn
`
`SALINE
`DGLUCUROr<.ATE
`LGLILONATE
`
`{6)
`{2)
`12)
`
`ll12iori 14C]GHBIM•nl
`
`-nean"!: S.EM
`60 ± 3 6
`81 = 0 5. *
`!
`40
`9
`
`*
`
`t:>·······ll L GULONATE
`
`o---o D GLUCURONATL
`
`.. "
`
`0
`
`0
`
`20
`
`40
`
`60
`
`80
`
`100
`
`120
`
`140
`
`160
`
`180
`
`200
`
`TIME {mw)
`
`Fig. 6. The effect of D·glucuronate and L·gulonate on the half-life
`(T 1/z) of GHB in plasma (30). Animals that received D-glucuronate
`received a bolus containing 240 mg followed by an infusion of 320
`mg!h for 3 h. Animals receiving L-gulonate received a bolus of 333
`mg and an infusion of 556 mg!b for 3 h. The t'/2 for the disappearance
`of [14C]GHB from plasma was determined as described. *p < 0_05;
`**p < 0.025.
`
`Table VIII. The Effect of Oxygen Concentration on Tissue Levels
`of GHB In Vivo
`
`Normal 0 2
`
`Low 0 2 (5.6%)
`
`nmol GHB/g
`tissue
`
`nmol GHB/g
`tissue
`
`Percent of
`Control
`
`Brain
`Kidney
`Muscle
`
`3.0 ± 0.15 (3)
`24.6 ± 4.2 (3)
`19.5 ± 4.2 (3)
`
`10.7 ± 0.74 (3)
`57.0 ± 13.4 (3)
`18.0 ± 2.3 (3)
`
`357*
`232
`92
`
`Animals were exposed to either room air or 5.6% oxygen for 2 hr
`before decapitation. Tissues were removed and assayed as have been
`described (3). Results are expressed as means ± SEM.
`*p = 0.0005
`
`state would also favor the synthesis of GHB by SSA
`reductase, an NADPH-dependent enzyme (35).
`Fate ofthe Carbon Skeleton ofGHB. Although GHB
`is a short chain acid and might be expected to undergo
`[3-oxidation (14), the markedly elevated plasma SSA and
`GHB concentrations in patients with GHB aciduria (17,18)
`provide evidence that [3-oxidation cannot be the major
`route of degradation. Inasmuch as the major portion of
`GHB proceeds through oxidation to SSA, the fate of the
`SSA formed either by GHB dehydrogenase or the mi(cid:173)
`tochondrial transhydrogenase must be considered. Sev(cid:173)
`eral investigators have isolated metabolic products formed
`in vivo after administration of labeled GHB (15,16,27,36-
`39). The label is distributed in citric acid cycle inter(cid:173)
`mediates and in aspartate, glutamate and GABA. Several
`reports indicate that the specific activity of GABA is
`higher than that found in glutamate. Therefore, it has
`been suggested (37,39) that the GHB skeleton may be
`converted to GABA by GABA transaminase instead of
`traversing a more circuitous route leading through the
`citric acid cycle and glutamate decarboxylase. The work
`cit