Neurochem. Int. Vol. 12, No. 1, pp. 53-59, 1988 0197-0186/88 $3.00 + 0.00 Printed in Great Britain. All rights reserved Copyright © 1988 Pergamon Journals Ltd GAMMA HYDROXYBUTYRATE DISTRIBUTION AND TURNOVER RATES IN DISCRETE BRAIN REGIONS OF THE RAT PHILIPPE VAYER*, JEAN-DANIEL EHRI-IARDTt, SERGE GOBAILLE*, PAUL MANDEL* and MICHEL MAITRE* *Centre de Neurochimie du CNRS et INSERM U44, 5 rue Blaise Pascal, 67084 Strasbourg C6dex, France and tLaboratoire de Spectrom6trie de masse, Institut de Pharmacologie, 11 rue Humann, 67085 Strasbourg C6dex, France (Received 12 June 1987; accepted 20 July 1987) Al~traet---Gamma-hydroxybutyric acid and trans-gamma-hydroxycrotonic acid levels have been deter- mined in 24 regions of the rat brain after sacrifice by microwave irradiation. Concentration ranges are from 4pmol/mg protein (frontal cortex) to 46pmol/mg protein (substantia nigra) for gamma- hydroxybutyric acid and from 0.4 pmol/mg protein (striatum) to 11 pmol/mg protein (hypothalamus) for trans-gamma-hydroxycrotonic acid. It appears that gamma-hydroxybutyric acid levels correlate well with GABA distribution in the same region. However this correlation is not evident with regard to the distribution of the gamma-hydroxybutyric acid synthesizing enzyme, specific succinic semialdehyde reductase. Using the antiepileptic drug, valproate which strongly inhibits gamma-hydroxybutyric acid release and degradation, we estimated the turnover rate of this compound in six regions of the rat brain. Turnover numbers ranged from 6.5 h -I in hippocampus to 0.76 h -I in cerebellum. Gamma-hydroxybutyrate (GHB) in rat brain is pri- marily a GABA metabolite (Anderson et al., 1977; Gold and Roth, 1977). During the last 6 years, much evidence has been presented which is in favour of a transmitter role for this substance in brain (Maitre and Mandel, 1984). In particular, characterisation of its synthesis (Cash et al., 1979; Rumigny et al., 1980; Rumigny et al., 1981a; Rumigny et al., 1981b), release (Maitre et al., 1983), transport (Benavides et aL, 1982a), turn-over (Gold and Roth, 1977), binding to synaptosomal membranes (Benavides et al., 1982b; Snead and Liu, 1984) and degradation (Vayer et al., 1985b) clearly suggest important functions in neuro- mediation. This role is also supported by a discrete regional distribution and ontogeny (Snead and Morley, 1981). However, no precise distribution in the brains of animals killed by microwave irradiation has yet been described. Indeed, taking into account the reported kinetics and characteristics of gamma- hydroxybutyrate binding, transport and degradation, it is important to know with precision the distribution pattern and levels of this molecule in brain. These results could constitute, together with studies of local turn-over rate, an index of regional GHBergic func- tional activity. In the present study, using microwave irradiation to kill the animals and negative ion mass spec- trometry to measure regional endogenous GHB lev- els, we describe the distribution of GHB in 24 regions of the rat brain. In parallel, a similar study has been made concerning trans-gamma-hydroxycrotonic acid (HCA), a structural analogue of GHB, whose presence in rat brain as an endogenous substance has recently been demonstrated (Vayer et al., 1985a). After treatment of rats with valproate, a time and dose dependent accumulation of GHB is described in the six regions of the rat brain investigated. This rapid accumulation which is due to an inhibition of release and degradation of GHB, has been used as a method to calculate regional turnover numbers for this substance in the rat brain. EXPERIMENTAL PROCEDURES 1. Determination of GHB regional levels Male Wistar rats (300-350 g) were sacrificed by focused microwave irradiation for 1.7 s (Piischner Apparatus 10 kW output 75%). Brains were removed and 8 slices were cut at the following coordinates with respect to lambda (zero point): (0, +1); (+1, +3.5); (+3.5, +5.5); (+5.5, +9.5); (+9.5, +11.5); (+11.5, +15.5); (-1,-2); (-2,-8). 24 areas were dissected out on a glass plate over liquid 53
`
`Ranbaxy Ex. 1038
`IPR Petition - USP 9,050,302
`
`

`
`54
`
`PHILIPPE VAYER et al.
`
`Structures
`Substantia nigra
`Posterior hypothalamus
`Superior colliculi
`Pineal gland
`Anterior hypothalamus
`Anterior thalamus
`Inferior colliculi
`Posterior thalamus
`Olfactory bulbs
`Amygdala
`
`Table 1. Concentrations of GHB and HCA in dissected brain areas
`GHB levels HCA levels
`(pmol/mg
`(pmol/mg
`protein)
`protein)
`46 1 6
`6.9 1 0.3
`42 1 3
`ll 1 1
`40 1 9
`n.d.
`40 1 4
`n.d.
`39 1 7
`3 0 1 0.4
`36 1 3
`4.7 1 1.2
`32 1 3
`4.5 1 0.5
`30 1 4
`n.d.
`27 1 9
`5.1 1 0.5
`24 1 7
`4.1 1 0.1
`21 1 5
`4 2 1 0.7
`I9 1 3
`2.8 1 0.5
`18 1 l
`0.4 1 0.1
`l5 1 4
`l 9 1 0.2
`1511
`47109
`
`Cortex
`
`nitrogen, and stored in liquid nitrogen. The extraction
`and derivatization of GHB and HCA were carried out as
`previously described (Ehrhardt et al., 1987) with the follow-
`ing modifications. Homogenization of the various brain
`regions was performed in a mixture of ethanol~water
`(80:20 v/v) at 0°C containing the
`internal
`standards
`([2,3-d2]GHB and
`beta-methyl
`trans-gamma-hydroxy-
`crotonic acid (MHCA)). After centrifugation (20 min
`50,000g)
`supematants were derivatized with penta-
`fluorobenzyl
`bromide
`and N-ter.-butyldimethylsily1-N-
`methyltrifluoroacetamide
`in
`acetonitrile
`as
`previously
`reported (Ehrhardt et al., 1988). The derivatized products
`were submitted to mass spectrometry analysis in an LKB
`2091 mass spectrometer modified for negative ion detection
`and equipped with a Pye 104 gas chromatograph. Chro-
`matographic and detection conditions are identical to those
`previously described (Ehrhardt et al., 1988). Measurements
`were performed
`by detection of characteristic
`ions
`m/z = 217
`for GHB, m/z = 219
`for
`[2,3-d2]GHB,
`m/z = 215 for HCA and m/z =229 for MHCA.
`Protein was measured by the Lowry method (Lowry et al.,
`1951) on the oentrifugated pellets after solubilization in 2 M
`NaOH. A schematic representation of the GHB distribution
`pattern was carried out using an arbitrary colour code.
`Diagrams of brain regions were computerized (Victor VPC2
`equipped with a 8087 coprocessor) using a data tablet and
`a graphic program (D-Calc from JPK-Conseil).
`2. Turnover rate estimation
`
`54 PHILIPPE VAYER et al. nitrogen, and stored in liquid nitrogen. The extraction and derivatization of GHB and HCA were carried out as previously described (Ehrhardt et al., 1987) with the follow- ing modifications. Homogenization of the various brain regions was performed in a mixture of ethanol-water SN (80:20v/v) at 0°C containing the internal standards Hyp ([2,3-d2]GHB and beta-methyl trans-gamma-hydroxy- SC crotonic acid (MHCA)). After centrifugation (20min Pin 50,000g) supernatants were derivatized with penta- Hya fluorobenzyl bromide and N-ter.-butyldimethylsilyl-N- THa methyltrifluoroacetamide in acetonitrile as previously IC reported (Ehrhardt et al., 1988). The derivatized products THp OB were submitted to mass spectrometry analysis in an LKB A 2091 mass spectrometer modified for negative ion detection PCx and equipped with a Pye 104 gas chromatograph. Chro- CCx matographic and detection conditions are identical to those Rad + m previously described (Ehrhardt et al., 1988). Measurements CP.GP were performed by detection of characteristic ions Pit m/z =217 for GHB, m/z =219 for [2,3-dJGHB, S m/z = 215 for HCA and m/z = 229 for MHCA. TCx ppCx Protein was measured by the Lowry method (Lowry et al., pCx 1951) on the centrifugated pellets after solubilization in 2 M Hid NaOH. A schematic representation of the GHB distribution Hiv pattern was carried out using an arbitrary colour code. OCx Diagrams of brain regions were computerized (Victor VPC2 Ent.Cx equipped with a 8087 coprocessor) using a data tablet and OT a graphic program (D-Calc from JPK-Conseil). C M 2. Turnover rate estimation pFCx Valproate (sodium salt) was injected IP into male Wistar FCx rats (300-350 g) at doses 200, 400, 600 mg/kg before sacrifice ppCx by focused microwave irradiation under the same conditions as described above 20 min after injection. In parallel experi- ments, the time dependent accumulation of GHB following a valproate dose of 400 mg/kg was determined from 0 to 90 min at chosen intervals. In every case, GHB levels were determined by the method previously described in 6 regions of the rat brain: dorsal hippocampus, pineal gland, hypo- thalamus, cerebellum, striatum, temporo-parietal cortex. Turnover rates were estimated by following the kinetics of accumulation of GHB in the linear part of the different curves obtained. RESULTS 1. GHB levels in discrete rat brain areas The amount of GHB and HCA in 24 regions of the rat brain are indicated in Table 1. Taking into account the very low level of GHB in blood (Nelson et al., 1981), the errors induced by the variable blood contamination among regions is insignificant when compared to the unavoidable experimental vari- ations. The richest structures for GHB are substantia nigra, hypothalamus, superior colliculi and pineal gland, whereas the cortical regions are relatively poor. The distribution pattern of HCA follows in general that of GHB, except for olfactory tubercles, septum and dorsal hippocampus which contain rela- tively high amounts of HCA. In contrast striatum, hypothalamus, thalamus and pineal gland contain a Table 1. Concentrations of GHB and HCA in dissected brain areas GHB levels HCA levels (pmol/mg (pmol/mg Abbreviations Structures protein) protein) Substantia nigra 46 _+ 6 6.9 ± 0.3 Posterior hypothalamus 42 ± 3 11 + 1 Superior colliculi 40 ± 9 n.d. Pineal gland 40 ± 4 n.d. Anterior hypothalamus 39 ± 7 3.0 ± 0.4 Anterior thalamus 36 ± 3 4.7 + t~2 Inferior colliculi 32 ± 3 4.5 ± 0.5 Posterior thalamus 30 ± 4 n.d. Olfactory bulbs 27 ± 9 5. I + 0.5 Amygdala 24 ± 7 4. I ± 0. I Parietal ~, Cingulate f Cortex 21 ± 5 4.3 ± 0.7 Raphe 19±3 2.8-+0.5 Striatum 18 + I 0.4 ± 0. I Pituitary gland 15 ± 4 1.9 ± 0.2 Septum 15 ± 1 4.7 ± 0.9 Temporal ] Prepiriform ~- Cortex 13 ± 2 2.6 ± 0.3 Piriform J Dorsal hippocampus 12 t 1 6.2 + 1.2 Ventral hippocampus 12 ± 3 1.6 ± 0.5 Occipital "~ EntorhinalJ Cortex 12 _+ 2 2.6 ± 0.7 Olfactory tubercles 9 ± I 4.8 + 0.4 Cerebellum 8 ± 2 1.4 ± 0.2 Medulla oblongata 8 + 1 1.3 ± 0. I Prefrontal cortex 6 ± 1 1.2 ± 0.3 Frontal ] Prepiriform.f Cortex 4.0 -+ 0.5 0.7 + 0. I Values are the mean _+ SEM of six different rats (n.d. = not detected). lower proportion of HCA compared to GHB. In 18 regions (among the 24 regions studied, i.e. if we elimi- nate posterior hypothalamus, olfactory tubercles, striatum, pineal gland, superior colliculi and posterior thalamus) a Student t-test reveals a linear correlation between HCA and GHB levels (r = 0.70 with P < 0.01). Figure 1 shows the general distribu- tion pattern of GHB in rat brain using an arbitrary colour code. This representation illustrates the het- erogeneous distribution of GHB in the brain regions investigated. 2. Time course of GHB accumulation in rats treated with valproate (400mg/kg IP) Rats were treated with valproate (400 mg/kg IP) and killed by microwave irradiation at time 0 to 90min every 15 minutes. Then, GHB levels were determined in cerebellum, pineal gland, hypo- thalamus, striatum, temporo-parietal cortex. For hip- pocampus, the accumulation of GHB is more rapid and GHB levels were determined every 5 min for 45 min. In all regions investigated, valproate induced a rapid and strong increase of GHB level, except for cerebellum where accumulation is much lower. GHB content in all regions increases linearly for 20 rain in
`
`Abbreviations
`SN
`Hyp
`SC
`Pin
`Hya
`THa
`[C
`THp
`OB
`A
`5%’;
`Raphe
`Rad + m
`Striatum
`CP.GP
`Pituitary gland
`Pit
`Septum
`S
`Temporal
`TCx
`Prepiriform Cortex
`ppCx
`Piriform
`pCx
`6.2 1 1.2
`12 1 1
`Dorsal hippocampus
`Hid
`l 6 1 0.5
`12 1 3
`Ventral hippocampus
`Hiv
`2.6 1 0.7
`12 1 2
`Cortex
`gfifcx
`4 ll 1 0.4
`9 1 l
`Olfactory tubercles
`OT
`1 4 1 0.2
`8 1 2
`Cerebellum
`C
`I 3 1 0.]
`8 1 1
`Medulla oblongata
`M
`l 2 1 0.3
`6 1 l
`Prefrontal cortex
`pFCx
`0 7 1 01
`4.0 1 0.5
`IF,::l':i‘::0m Cortex
`iféx
`Values are the mean 1 SEM of six different rats (n.d. = not detected).
`
`13 1 2
`
`2.6 1 0.3
`
`lower proportion of HCA compared to GHB. In 18
`regions (among the 24 regions studied, i.e. if we elimi-
`nate posterior hypothalamus, olfactory tubercles,
`striatum,
`pineal
`gland,
`superior
`colliculi
`and
`posterior thalamus) a Student t-test reveals a linear
`correlation between HCA and GHB levels (r = 0.70
`with P < 0.01). Figure 1 shows the general distribu-
`tion pattern of GHB in rat brain using an arbitrary
`colour code. This representation illustrates the het-
`erogeneous distribution of GHB in the brain regions
`investigated.
`
`2. Time course of GHB accumulation in rats treated
`with valproate (400mg/kg IP)
`
`Rats were treated with valproate (400 mg/kg IP)
`and killed by microwave irradiation at
`time 0 to
`90 min every 15 minutes. Then, GHB levels were
`determined in
`cerebellum, pineal gland, hypo-
`thalamus, striatum, temporo-parietal cortex. For hip-
`pocampus, the accumulation of GHB is more rapid
`and GHB levels were determined every 5min for
`45 min. In all regions investigated, Valproate induced
`a rapid and strong increase of GHB level, except for
`cerebellum where accumulation is much lower. GHB
`
`content in all regions increases linearly for 20 min in
`
`Valproate (sodium salt) was injected IP into male Wistar
`rats (300~350 g) at doses 200, 400, 600 mg/kg before sacrifice
`by focused microwave irradiation under the same conditions
`as described above 20 min after injection. In parallel experi-
`ments, the time dependent accumulation of GHB following
`a valproate (lose of 400mg/kg was determined from 0 to
`90 min at chosen intervals. In every case, GHB levels were
`determined by the method previously described in 6 regions
`of the rat brain: dorsal hippocampus, pineal gland, hypo-
`thalamus, cerebellum, striatum,
`temporo-parietal cortex.
`Turnover rates were estimated by following the kinetics of
`accumulation of GHB in the linear part of the different
`curves obtained.
`
`RESULTS
`
`1. GHB levels in discrete rat brain areas
`
`The amount of GHB and HCA in 24 regions of the
`rat brain are indicated in Table 1. Taking into
`account the very low level of GHB in blood (Nelson
`et al., 1981), the errors induced by the variable blood
`contamination among regions is insignificant when
`compared to the unavoidable experimental vari-
`ations. The richest structures for GHB are substantia
`
`nigra, hypothalamus, superior colliculi and pineal
`gland, whereas the cortical
`regions are relatively
`poor. The distribution pattern of HCA follows in
`general that of GHB, except for olfactory tubercles,
`septum and dorsal hippocampus which contain rela-
`tively high amounts of HCA. In contrast striatum,
`hypothalamus, thalamus and pineal gland contain a
`
`

`
`pFCx 06 ® cc~. --.-- FCx ® f CP © pHtd CCx • Hya@ PCx CCx Hid ~ ® DE Ent. Cx -"SN -J @ PCx / ® \TCx ® GHB (pmo lo/m~ pPotein) no-6 I. - 14 I /4 - 22 I ~2 - ~8 I 28 - 37 137 - O O" ,.¢ t~r O t:k o < Fig. 1. Distribution of GHB in rat brain represented by an arbitrary colour code. Stereotaxic coordinates of sections (with respect to lambda-zero point-): A (+ 11.5, + 15.5); B (+ 9.5, + 11.5); C (+ 5.5, + 9.5); D (+3.5, +5.5); E (+ 1, +3.5); F (0, + 1); G (ventral view); H (dorsal view).
`
`

`
`55
`
`PHILIPPE VAYER er al.
`
`120-
`
`100 -
`E’
`eo—
`e
`/
`E 40$/E I—I\I
`:1’
`
`20
`
`£1
`
`O
`
`4
`20
`
`‘E
`
`l
`I
`6O
`40
`Time (min)
`
`l
`80
`
`I
`100
`
`Fig. 2. Time course of GHB accumulation in dorsal hippo-
`campus (Q) and in hypothalamus (I) of rats treated with
`valproate (400 mg/kg IP). Results are the mean 1 SEM
`from three animals.
`
`least for 60 min in the other
`hippocampus and at
`regions tested. Then a plateau was reached. more
`rapidly in hippocampus (Fig. 2).
`
`3. Dose—response curve of valproate
`
`in hypothalamus were determined
`GHB levels
`after
`treatment with
`valproate
`20 minutes
`(0—200~400—600 mg/kg IP). As shown in Fig. 3.
`GHB levels increased linearly in function of valproate
`doses. As the LD50 for valproate in rats is about
`800 mg/kg, higher doses than 600 mg/kg were not
`tested. Twenty minutes after 400 mg/kg valproate.
`the increase of GHB in hypothalamus is around 75%.
`
`For valproate dosages of 400 mg/kg, the drug con-
`centration in brain after l5 min is around l400,uM
`(about 20 times the K for GHB dehydrogenase
`inhibition). Thus GHB degradation is completely
`inhibited.
`In fact, we have demonstrated the total
`inhibition of SSR,
`in vitro with lmM valproate
`(Rumigny er al., 1980). Additional effect on GHB
`accumulation for 600 mg/kg valproate appears to be
`due to an enhanced inhibition of release from pre-
`synaptic stores (Fig. 3).
`
`4. Estimation of GHB turnover time
`
`After valproate 400 mg/kg IP, the time dependent
`accumulation of GHB in the different regions tested
`was used to determine the rate of GHB synthesis by
`the difference between GHB level after treatment and
`the control level. Usually, the linear GHB increase
`after 20 min treatment was used to calculate the rate
`
`In pineal gland.
`of GHB accumulation (Table 2).
`hypothalamus. temporo-parietal cortex and striatum,
`turnover times are of the same order of magnitude.
`In hippocampus. turnover is much more rapid but
`much more lower in cerebellum.
`
`In the different regions tested. the accumulation of
`GHB starts rapidly (except for cerebellum) and is
`linear in every case for at least 20 min. At an early
`time, curves exhibit only one slope. Linear regression
`analysis of initial accumulation of GHB gives the
`following turnover times (Table 2): Hippocampus
`9min, pineal gland 35 min. hypothalamus 34 min.
`temporo-parietal cortex 47 min, striatum 30 min and
`cerebellum 78 min. Turnover time in hippocampus is
`about 8 times shorter than in cerebellum and about
`
`3 to 4 times lower compared to other regions.
`
` Table 2.
`Initial rate of
`Turnover Turnover
`GHB synthesis
`time
`number
`(pmol.imin,fmg
`(min)
`(h ')
`protein)
`Brain region
`9 1 3
`6.5
`L34 1 0.30
`Dorsal hippocampus
`30 1 5
`2.0
`0.60 1 0.08
`Striatum
`35 117
`1.7
`l.l4 1 0.25
`Pineal gland
`34 1 5
`1.7
`L16 1 0.20
`Hypothalamus
`47 1 l0
`1.2
`0.32 1 0.03
`Temporo-parietal cortex
`
`Cerebellum 0.7 0.10 1 0.02 78 1 14
`
`
`The animals received 400 mg/kg valproate [P at starting time. Then
`three animals were killed every 15 min for 90 min and GHB was
`determined in dissected brain regions. For dorsal hippocampus.
`three animals were killed every 5 min for 20 min. All values are
`means 1 SEM. Control GHB concentrations were determined in
`animals receiving saline.
`Turnover time:
`time necessary to renew the whole compartment.
`Turnover number: number of times the compartment was renewed
`per hour.
`
`56 PHILIPPE VAYER etal. 12° f t 100 ~ "~ 80 /t/ i i 20 [~ I 2o I J 40 60 Time (rain) J _J 80 100 Fig. 2. Time course of GHB accumulation in dorsal hippo- campus (@) and in hypothalamus (11) of rats treated with valproate (400mg/kg IP). Results are the mean _+ SEM from three animals. hippocampus and at least for 60 min in the other regions tested. Then a plateau was reached, more rapidly in hippocampus (Fig. 2). 3. Dose-response curve of valproate GHB levels in hypothalamus were determined 20 minutes after treatment with valproate (0--200~400~600 mg/kg IP). As shown in Fig. 3, GHB levels increased linearly in function of valproate doses. As the LDs0 for valproate in rats is about 800mg/kg, higher doses than 600mg/kg were not tested. Twenty minutes after 400mg/kg valproate, the increase of GHB in hypothalamus is around 75%. 100 80 EO rn Z 20 CD I F 200 400 VaLproate {mglkg IP] I 600 Fig. 3. Effect of valproate on GHB levels in hypothalamus. Rats were treated with different doses of valproate. GHB levels were determined 20rain after treatment with valproate. Each point represents the mean value + SEM from three animals. For valproate dosages of 400 mg/kg, the drug con- centration in brain after 15 min is around 1400/~M (about 20 times the /(1 for GHB dehydrogenase inhibition). Thus GHB degradation is completely inhibited. In fact, we have demonstrated the total inhibition of SSR~ in vitro with l mM valproate (Rumigny et al., 1980). Additional effect on GHB accumulation for 600 mg/kg valproate appears to be due to an enhanced inhibition of release from pre- synaptic stores (Fig. 3). 4. Estimation of GHB turnover time After valproate 400 mg/kg IP, the time dependent accumulation of GHB in the different regions tested was used to determine the rate of GHB synthesis by the difference between GHB level after treatment and the control level. Usually, the linear GHB increase after 20 min treatment was used to calculate the rate of GHB accumulation (Table 2). In pineal gland, hypothalamus, temporo-parietal cortex and striatum, turnover times are of the same order of magnitude. In hippocampus, turnover is much more rapid but much more lower in cerebellum. In the different regions tested, the accumulation of GHB starts rapidly (except for cerebellum) and is linear in every case for at least 20 min. At an early time, curves exhibit only one slope. Linear regression analysis of initial accumulation of GHB gives the following turnover times (Table 2): Hippocampus 9min, pineal gland 35 min, hypothalamus 34min, temporo-parietal cortex 47 rain, striatum 30 rain and cerebellum 78 min. Turnover time in hippocampus is about 8 times shorter than in cerebellum and about 3 to 4 times lower compared to other regions. Table 2, Brain region Initial rate of GHB synthesis Turnover Turnover (pmol/min/mg time number protein) (min) (h ') Dorsal hippocampus 1.34 + 0.30 9 _+ 3 6.5 Striatum 0.60 + 0.08 30 + 5 2.0 Pineal gland 1.14 ± 0.25 35 + 17 I. 7 Hypothalamus 1.16 _+ 0.20 34 ± 5 1.7 Temporo-parietal cortex 0.32 _+ 0.03 47 ± 10 1.2 Cerebellum 0.10_+0.02 78 ± 14 0.7 The animals received 400 mg/kg valproate IP at starting time. Then three animals were killed every 15 min for 90 rain and GHB was determined in dissected brain regions. For dorsal hippocampus, three animals were killed every 5 min for 20 min. All values are means _+ SEM. Control GHB concentrations were determined in animals receiving saline. Turnover time: time necessary to renew the whole compartment. Turnover number: number of times the compartment was renewed per hour.
`
`
`
`GHB(pmot/mgprotein)
`
`-
`
`60-
`
`O
`
`Egg
`
`IX
`
`l
`l
`4OO
`2OO
`valproate (mg/kg IP)
`
`I
`6OO
`
`Fig. 3. Effect of valproate on GHB levels in hypothalamus.
`Rats were
`treated with different doses of valproate.
`GHB levels were determined 20 min after treatment with
`valproate. Each point represents the mean value1SEM
`from three animals.
`
`

`
`Gamma-hydroxybutyrate distribution and turnover 57 DISCUSSION Conflicting reports exist about the possible post- mortem changes in brain GHB levels of animals sacrificed by decapitation. In particular, Eli and Cattabeni (1983) reported a lower level of GHB in the brain of animals dissected after irradiation in a microwave oven. For this reason, in order to avoid GHB increase during the time required for tissue dissection, the rats used in the present study were killed by microwave irradiation. In addition, sample preparation was performed by homogenization in a mixture of ethanol-water (80: 20 v/v), thus we could exclude any possible lactonization of GHB during the extraction procedure. GHB is present in all the 24 brain areas studied and is distributed unevenly. The ratio of the highest (substantia nigra and hypo- thalamus) to lowest concentrations (frontal cortex) is about 10 to 1. This distribution pattern of GHB in rat brain can be compared with the similar study made for GABA, the main precursor of GHB in brain, on approximately the same regions of rat brain killed by microwave irradiation (Balcom et al., 1975). The two distributions appear to be similar, the richest regions being in both cases substantia nigra, hypothalamus, colliculi and the poorest are cerebellum, medulla and cortex. The other regions contain intermediate con- centrations. Thus the ratio GABA/GHB is always about 1000, increasing respectively to 1600 and 2000 for cerebellum and medulla. The correlation between the regional distribution of GHB and the amount of its synthesizing enzyme, succinic semialdehyde reduc- tase (SSR2) seems to be more erratic (Rumigny et al., 1982). SSR2 specific activity and GHB levels are well correlated for hypothalamus, colliculi (high levels), amygdala, hippocampus, striatum (medium levels) and frontal cortex (low levels). However, no cor- relation seems to exist in cerebellum, pons medulla, septum and olfactory bulbs. If we compare the regional density of binding sites for GHB (Benavides et al., 1982b) and concentrations of this substance in the same region, no apparent correlation exists and this seems to indicate that GHB could have a different functional role in the different brain regions studied. The regional levels of GHB given in the present study are close to the values reported by Eli and Cattabeni (1983) which concern however only 5 regions of the brains of rats killed by microwave irradiation. The values given by Snead and Morley (1981) are somewhat higher. This discrepancy could be the result of the difference in the time required for dissection of the tissue after decapitation. Values for HCA regional concentrations are about 50 times lower than the value of 220 pmol/mg protein for total brain reported in a previous publication (Vayer et al., 1985a). Preliminary experiments have shown that this discrepancy cannot be attributed to the difference in the mode of sacrifice (decapitation versus microwave irradiation), in the extraction procedure (0.1 M for- mic acid versus 80% ethanol solution) or to the difference in the age of the rats used in the two studies. Although the method used in the present work for derivatization and quantification is some- what different from those previously used (Vayer et al., 1985a), it seems most probable that the first reported value for HCA was overestimated. It is possible that an interfering peak is the reason for the too high value reported for HCA. The actual amounts of HCA are about 1/10 of those for GHB in the same region. Thus, it seems that the relative importance of HCA in interfering with GHB func- tional activity in brain can be considered less credible. The first report on an accumulation of GHB following acute treatment by valproate was reported by Snead et al. (1980). Our results confirm these findings and describe regional variations in the kinet- ics of these accumulations. The GHB increase can be explained by the two following phenomena. First, we have reported that valproate inhibits GHB release evoked by 40 mM K ÷ from hippocampal and striatal slices preloaded with [3H]GHB. The IC5o are re- spectively 500 # M for hippocampus and 250 p M for striatum (Vayer et al., 1987). These values must be compared to the plasma level of valproate (837 pM) 1 h after 150 mg/kg IP (Patsolos and Lascelles, 1981). In addition, the distribution of valproate in brain 15 min after 200 mg/kg IP is homogenous at about 700/~M (Hariton et al., 1984). This effect of valproate on GHB release probably induces a rapid synaptic increase of this substance. Secondly, Kaufman et al. (1983) have described a strong inhibition of valproate on GHB dehydrogenase with a K~ of 60 p M. We have demonstrated that this GHB dehydrogenase is in fact the high K m aldehyde reductase (ALRI) which is identical to non specific SSR 1 and glucuronate reduc- tase (Vayer et al., 1985c). We measured a K~ of 80 #M for valproate on this enzyme activity (competitive inhibition). At the dosage of valproate used, GHB dehydrogenase inhibition is mainly responsible for the GHB accumulation (valproate concentration equal about 20 x K~ for GHB dehydrogenase in- hibition). Under these conditions, accumulation of GHB is mainly due to dehydrogenase inhibition at 400 mg/kg of valproate. The dose effect observed in GHB accumulation for valproate dosages higher
`
`

`
`58
`
`PHILIPPE VAYER et al.
`
`than 400 mg/kg is probably due to additional efi"ect
`on the release.
`
`the specificity of action of
`The problem of
`valproate can be discussed. This drug is known to
`induce an increase of GABA and a decrease of
`
`58 PHILIPPE VAYER et al. than 400 mg/kg is probably due to additional effect on the release. The proble/n of the specificity of action of valproate can be discussed. This drug is known to induce an increase of GABA and a decrease of aspartate (Schechter et al., 1978; Van der Laan et al., 1979). However, the increase of GABA in total brain is only 13% 10 minutes after 400mg/kg valproate IP. This increase cannot significantly modify GHB syn- thesis and is thought to be due to the metabolism of succinic semi-aldehyde accumulated after SSADH inhibition. Moreover, it is generally accepted that valproate, at concentrations used in the present study, does not modify the activity of the biosynthetic pathway for GHB. GABA-T activity is only very slightly inhibited with a K~ of 9.5 to 18 mM (Maitre et al., 1978; Van der Laan et al., 1979) and specific succinic semialdehyde reductase, which synthetizes GHB (Rumigny et al., 1980; Rumigny et al., 1981a) is not affected by valproate. Thus it seems that the drug has minimal action on GHB biosynthesis, but inhibits strongly release and degradation of GHB as demonstrated by the very rapid accumulation of this substance. GHB accumulation does not seem to cause a feedback inhibition on GHB synthesis since no significative effect of this compound has been observed on its synthetic enzyme (Rumigny et al., 1980). Thus the use of valproate for the deter- mination of GHB turnover is entirely justified. How- ever, these measurements do not necessarily reflect the turnover rate at GHBergic synapses, but rather the overall rate of GHB synthesis in brain [i.e. GHB in neuronal cytosol and in synapses (Rumigny et al., 1981b)]. In conclusion, the turnover rate of GHB appears to be very rapid in dorsal hippocampus, a region rich in GHB binding sites (Benavides et al., 1982b), whereas in cerebellum, where GHB binding is very low, the turnover rate is much lower. Thus, studies on turnover rates are more accurate than steady state levels to determine the functional role of this sub- stance in brain. Turnover numbers of GHB in rat brain are similar to those of serotonin in raphe nucleus or in hippocampus (Necker and Meek, 1976), but higher than those for GABA in cortex, cere- bellum, striatum and hippocampus (Bernasconi et al., 1982). Finally our results are in good accordance with the turnover time of GHB in total brain (26 min) determined by Gold and Roth (1976) by the use of [3H]GABA as the GHB precursor. REFERENCES Anderson R. A., Ritzmann R. F. and Tabakoff B. (1977) Formation of gamma-hydroxybutyrate in brain. J. Neurochem. 28, 633-639. Balcom G. J., Lenox R. H. and Meyerhoff J. L. (1975) Regional gamma-aminobutyric acid levels in rat brain determined after microwave fixation. J. Neurochem. 24, 609-6 13. Benavides J., Rumigny J. F., Bourguignon J. J., Wermuth C. G., Mandel P. and Maitre M. (1982a) A high-affinity Na+-dependent uptake system for gamma-hydroxy- butyrate in membrane vesicules prepared from rat brain. J. Neurochem. 38, 1570-1575. Benavides J., Rumigny J. F., Bourguignon J. J., Cash C. D., Wermuth C. G., Mandel P., Vincendon G. and Maitre M. (1982b) High-affinity binding site for gamma- hydroxybutyrate in rat brain. Life Sci. 30, 953-961. Bernasconi R., Maitre L., Martin P. and Raschdorf F. (1982) The use of inhibitors of GABA-transaminase for the determination of GABA turnover in mouse brain regions: an evaluation of aminooxyacetic acid and gab- aculine. J. Neurochem. 38, 57-66. Cash C. D., Maitre M. and Mandel P. (1979) Purification from human brain and some properties of two NADPH- linked aldehyde reductases which reduce succinic semi- aldehyde to 4-hydroxybutyrate. J. Neurochem. 33, 116%1175. Ehrhardt J. D., Vayer Ph. and Maitre M. (1988) A rapid and sensitive method for the determination of gamma- hydroxybutyric acid and trans-gamma-hydroxycrotonic acid in rat brain tissue by gas chromatography-mass spectrometry with negative ion detection. Biomed. mass spectro., In press. Eli M. and F. Cattabeni (1983) Endogeneous gamma- hydroxybutyrate in rat brain areas. Postmortem changes and effects of drugs interfering with gamma-aminobutyric acid metabolism. J. Neurochem. 41, 524-530. Gold B. I. and Roth R. H. (1977) Kinetics of in vivo conversion of gamma [3H]aminobutyric acid to gamma [3H]hydroxybutyric acid by rat brain. J. Neurochem. 28, 106% 1073. Hariton C., Ciesielski L., Simler S., Valli M., Jadot G., Gobaille S., Mesdjian E. and Mandel P. (1984) Distribu- tion of sodium valproate and GABA metabolism in CNS of the rat. Biopharmaceut. Drugs Dispos. 5, 409-414. Kaufman E. and Nelson T. (1983) Inhibition of gamma- hydroxybutyrate dehydrogenase by aromatic or branched-chain acids. J. N