Downloaded from
`
`jpet.aspetjournals.org
`
` at ASPET Journals on October 5, 2015
`
`0022-3565/04/3111-92–98$20.00
`THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
`Copyright © 2004 by The American Society for Pharmacology and Experimental Therapeutics
`JPET 311:92–98, 2004
`
`Vol. 311, No. 1
`69682/1169709
`Printed in U.S.A.
`
`GHB (␥-Hydroxybutyrate) Carrier-Mediated Transport across
`the Blood-Brain Barrier
`
`Indranil Bhattacharya and Kathleen M. K. Boje
`Department of Pharmaceutical Sciences, School of Pharmacy and Pharmaceutical Sciences, University at Buffalo, Buffalo,
`New York
`Received April 12, 2004; accepted June 1, 2004
`
`ABSTRACT
`␥-Hydroxybutyrate (sodium oxybate, GHB) is an approved ther-
`apeutic agent for cataplexy with narcolepsy. GHB is widely
`abused as an anabolic agent, euphoriant, and date rape drug.
`Recreational abuse or overdose of GHB (or its precursors ␥-bu-
`tyrolactone or 1,4-butanediol) results in dose-dependent cen-
`tral nervous system (CNS) effects (respiratory depression, un-
`consciousness, coma, and death) as well as tolerance and
`withdrawal. An understanding of the CNS transport mecha-
`nisms of GHB may provide insight into overdose treatment
`approaches. The hypothesis that GHB undergoes carrier-me-
`diated transport across the BBB was tested using a rat in situ
`brain perfusion technique. Various pharmacological agents
`were used to probe the pharmacological characteristics of the
`transporter. GHB exhibited carrier-mediated transport across
`the BBB consistent with a high-capacity, low-affinity trans-
`porter; averaged brain region parameters were Vmax ⫽ 709 ⫾
`
`214 nmol/min/g, Km ⫽ 11.0 ⫾ 3.56 mM, and CLns ⫽ 0.019 ⫾
`0.003 cm3/min/g. Short-chain monocarboxylic acids (pyruvic,
`lactic, and ␤-hydroxybutyric), medium-chain fatty acids (hex-
`anoic and valproic), and organic anions (probenecid, benzoic,
`salicylic, and ␣-cyano-4-hydroxycinnamic acid) significantly in-
`hibited GHB influx by 35 to 90%. Dicarboxylic acids (succinic
`and glutaric) and ␥-aminobutyric acid did not inhibit GHB BBB
`transport. Mutual inhibition was observed between GHB and
`benzoic acid, a well known substrate of the monocarboxylate
`transporter MCT1. These results are suggestive of GHB cross-
`ing the BBB via an MCT isoform. These novel findings of GHB
`BBB transport suggest potential therapeutic approaches in the
`treatment of GHB overdoses. We are currently conducting
`“proof-of-concept” studies involving the use of GHB brain
`transport inhibitors during GHB toxicity.
`
`␥-Hydroxybutyric acid (GHB), an endogenous neuromodu-
`lator (Cash, 1994) was synthesized by H. Laborit in the early
`1960s as a GABA mimetic agent (Laborit, 1964). GHB was
`studied for potential use as an anesthetic agent; however, its
`adverse effects outweighed its therapeutic effects. In the late
`1980s, interest in GHB was rekindled for use in sleep disor-
`ders. Recently, the Food and Drug Administration granted
`orphan drug status to GHB (sodium oxybate; Xyrem) as a
`controlled substance with restricted distribution for the
`treatment of narcolepsy with cataplexy. GHB is currently
`under investigation for potential therapeutic use in alcohol
`and opioid withdrawal (Gallimberti et al., 2000), and in other
`conditions such as depression, anxiety, and fibromyalgia
`(Scharf et al., 1998; Ferrara et al., 1999).
`However, GHB derives notoriety from its current popular-
`ity as a recreational drug of abuse. GHB and its analogs
`
`This work was supported in part by National Institutes of Health Grant
`DA14988.
`Article, publication date, and citation information can be found at
`http://jpet.aspetjournals.org.
`doi:10.1124/jpet.104.069682.
`
`(␥-butyrolactone and 1,4-butanediol) are currently abused for
`their recreational and pleasurable properties (heightened
`sexual pleasures, stress reduction, sedative, antianxiety, and
`antidepressant effects; http://www.projectghb.org/) by dance
`club attendees (rave parties); anabolic effects by body build-
`ers; and disinhibitory and sedative effects by sexual preda-
`tors (Ingels et al., 2000; Nicholson and Balster, 2001; Okun et
`al., 2001). Of interest, GHB’s physiochemical properties (col-
`orless, odorless, and slightly salty taste) have been exploited
`as an “ideal” date rape drug (ElSohly and Salamone, 1999;
`Smith, 1999).
`The surge in GHB (as well as ␥-butyrolactone and 1,4-
`butanediol) abuse by the drug counterculture has led to a
`substantial increase in drug overdoses and fatalities (Okun
`et al., 2001; Zvosec et al., 2001). Adverse events associated
`with GHB overdose include seizures, respiratory depression,
`and impaired consciousness leading to coma and death. Pres-
`ently, the treatment of GHB overdose includes empirical
`interventions and symptomatic treatments (Nicholson and
`Balster, 2001; Okun et al., 2001).
`The blood-brain barrier (BBB) maintains brain homeosta-
`
`ABBREVIATIONS: GHB, ␥-hydroxybutyrate; BBB, blood-brain barrier; MCT, monocarboxylate transporter; BA, benzoic acid; CHC, ␣-cyano-4-
`hydroxycinnamic acid; DPH, 1,6-diphenyl-1,3,5-hexatriene; OAT, organic anion transport.
`
`92
`
`Ranbaxy Ex. 1027
`IPR Petition - USP 9,050,302
`
`

`
`Downloaded from
`
`jpet.aspetjournals.org
`
` at ASPET Journals on October 5, 2015
`
`sis by restricting the movement of molecules based on size,
`charge, hydrogen bonding potential, and lipid solubility
`(Pardridge, 1997; Saunders et al., 1999). Whereas many com-
`pounds penetrate the BBB by passive diffusion, many other
`agents undergo active influx or efflux by transport proteins
`(Tamai and Tsuji, 2000).
`An understanding of the transport mechanisms of GHB
`across the BBB may provide insight into rationale treatment
`approaches for GHB toxicity. A review of the older literature
`suggests that GHB may cross the BBB by a transport pro-
`tein. Roth and Giarman (1966) reported that preadministra-
`tion of ␤-hydroxybutyrate to rats resulted in decreased brain
`and blood concentrations of exogenously administered GHB.
`A careful reexamination of Roth and Giarman’s data reveals
`a decrease in the GHB brain-to-blood ratio in the presence of
`␤-hydroxybutyrate (0.438 without ␤-hydroxybutyrate, n ⫽
`3–4 versus 0.323 without ␤-hydroxybutyrate, n ⫽ 3–4).
`␤-Hydroxybutyrate was recently identified as a substrate for
`the monocarboxylate transporter (MCT)
`(Enerson and
`Drewes, 2003). GHB also inhibits MCT substrate uptake in
`erythrocytes and cardiac myocytes (Poole and Halestrap,
`1993). In addition, there is substantial evidence of MCT
`expression at the BBB (Kang et al., 1990; Terasaki et al.,
`1991; Kido et al., 2000).
`Considered in toto, this evidence suggests the hypothesis
`that GHB undergoes MCT carrier-mediated transport across
`the BBB. Using an in situ brain perfusion technique, we
`report that GHB undergoes both carrier-mediated transport
`and passive diffusion and that the carrier-mediated pro-
`cesses are pharmacologically inhibited by known inhibitors of
`MCT.
`
`Materials and Methods
`Chemicals. Male Sprague-Dawley rats (250–350 g) were pur-
`chased from Harlan (Indianapolis, IN). [3H]GHB (specific activity,
`35.5 Ci/mmol]) and [14C]benzoic acid ([14C]BA; specific activity, 60
`mCi/mmol) were purchased from Moravek Biochemical (Brea, CA).
`All test compounds (␣-cyano-4-hydroxycinnamic acid; CHC), probe-
`necid, GABA, succinic acid, glutaric acid, L-lactic acid, glycine, su-
`crose, and the sodium salt forms of GHB, BA, salicylic acid, ␤-hy-
`droxybutryic acid, hexanoic acid, valproic acid, and pyruvic acid were
`purchased from Sigma-Aldrich (St. Louis, MO). Soluene 350 and
`Soluscint O were purchased from PerkinElmer Life and Analytical
`Sciences (Boston, MA) and National Diagnostics (Atlanta, GA), re-
`spectively. Ketamine and xylazine were purchased from J.A. Web-
`ster (Sterling, MA).
`In Situ Rat Brain Perfusion Protocol. All procedures involving
`animals were approved by the University of Buffalo Institutional
`Animal Care and Use Committee. The transport of GHB across the
`BBB was quantified using the in situ rat brain perfusion. This
`technique was first developed by Takasato et al. (1984) and later
`modified, which included a change in the surgical procedure and
`perfusion flow rate (Allen and Smith, 2001). Briefly, adult male
`Sprague-Dawley rats (250–350 g) were anesthetized using a mixture
`of ketamine (90 mg/kg) and xylazine (9 mg/kg), administered intra-
`muscularly and placed on a warming pad to maintain body temper-
`ature. Electrocardiograms were continuously monitored (Snap-Mas-
`ter, version 3; Hem Data Corporation, Southfield, MI) throughout
`the surgical procedure. The thoracic cavity of the rat was opened and
`the left common carotid artery exposed. This was followed by the
`ligation of the external carotid artery and cauterization of the supe-
`rior thyroid and occipital arteries. The common carotid artery was
`cannulated proximal to the bifurcation of the external and internal
`carotid arteries with a 25-gauge hypodermic needle affixed to poly-
`
`Transport of GHB across the Blood-Brain Barrier
`
`93
`
`ethylene-50 tubing [filled with physiological perfusate: 128 mM
`NaCl, 24 mM NaHCO2, 4.2 mM KCl, 2.4 mM NaH2PO4, 1.5 mM
`CaCl2, 0.9 mM MgSO4, and 9 mM glucose (oxygenated with 95%, 5%
`O2/CO2, 37°C, pH ⬃7.4) (Mahar Doan et al., 2000)]. A syringe con-
`taining oxygenated perfusate was attached to the cannula. The left
`ventricle of the heart was quickly severed to arrest blood flow to the
`brain. The left common carotid artery was ligated below the cannula
`insertion point. In situ brain perfusion was immediately initiated at
`a flow rate of 10 ml/min with a perfusion pump (model 55-4150;
`Harvard Apparatus Inc., Holliston, MA). This technique resulted in
`the perfusion of the left cerebral hemisphere.
`After a perfusion period of 15 to 60 s, rats were sacrificed by
`decapitation. Brains were removed, placed on ice chilled glass plates,
`and the left hemispheres were dissected into the following brain
`regions: the cortices (frontal, parietal, and occipital), hippocampus,
`striatum, and thalamus/hypothalamus. Dissected tissue samples
`were placed in pretared liquid scintillation vials, weighed, and sol-
`ubilized overnight with 0.8 ml of Soluene 350 at 50°C. Five milliliters
`of Soluscint O was added, and the samples were analyzed by liquid
`scintillation counting using a 1900CA liquid scintillation analyzer
`(PerkinElmer Life and Analytical Sciences). The counting efficien-
`cies for 3H and 14C were 0.61 and 0.95, respectively. An aliquot of the
`perfusion fluid was similarly assayed by liquid scintillation counting
`to verify the perfusate analyte concentration.
`In separate experiments, the capillary depletion technique was
`used to determine the distribution of [3H]GHB between brain vas-
`culature and brain parenchyma (Triguero et al., 1990). For experi-
`ments where high concentrations of GHB or inhibitors (⬎5 mM) were
`required, the sodium chloride concentration of the perfusate was
`adjusted to maintain physiological osmolality.
`Experimental Protocols: Linear Influx of GHB. Pilot studies
`were first performed to determine linear permeability conditions,
`i.e., the time course over which [3H]GHB influx was linear and
`unidirectional (Takasato et al., 1984; Smith, 1999; Mahar Doan et
`al., 2000). Animals (n ⫽ 3–4) were perfused with [3H]GHB (0.028
`␮M; 1.0 ␮Ci/ml) for 15, 30, 45, or 60 s and sacrificed. Brain regions
`were assayed for [3H]GHB as described previously. Based on these
`studies (see Results), a 30-s perfusion period was selected for all
`subsequent studies.
`GHB Concentration-Dependent Study. GHB concentration-
`dependent influx studies were performed to assess the extent of
`saturable transport in the presence of the following concentrations of
`unlabeled GHB in separate groups of rats: 0.028 mM (n ⫽ 4), 28 ⫻
`10⫺5 mM (n ⫽ 3), 5 ⫻ 10⫺3 mM (n ⫽ 3), 0.05 mM (n ⫽ 3), 0.1 mM (n ⫽
`3), 0.5 mM (n ⫽ 3), 1 mM (n ⫽ 3), 10 mM (n ⫽ 3), 20 mM (n ⫽ 4), 30
`mM (n ⫽ 4), and 40 mM (n ⫽ 4). Brain tissues were assayed for
`[3H]GHB as described previously.
`Substrate Inhibitor Studies. Substrate inhibitor studies were
`performed to determine the substrate specificity of the GHB trans-
`porter and to aid in the pharmacological characterization of the
`transporter. The compounds used for the inhibition studies were
`selected based on their known transport characteristics and/or chem-
`ical structures. Test inhibitors (1–20 mM) were individually coper-
`fused with [3H]GHB (0.028 ␮M; 1.0 ␮Ci/ml). Short-chain monocar-
`boxylic acids (L-lactic, pyruvic, and ␤-hydroxybutyrate acids),
`dicarboxylic acids (succinic and glutaric), medium-chain fatty acids
`(hexanoic and valproic acids), and organic acids (benzoic and salicylic
`acids) were coperfused at 20 mM. CHC, a specific inhibitor of the
`MCT, was coperfused at 1 mM. Other organic anions that were
`tested for inhibitory effects on GHB transport included GABA (10
`mM) and probenecid (10 mM). Negative controls for transport in-
`cluded substances that undergo minimal to moderate passive diffu-
`sion, e.g., sucrose (20 mM) and glycine (20 mM). The concentrations
`of potential inhibitor compounds were selected either based on Km
`values (if known) or the limit of solubility in the perfusate.
`Inhibition of Benzoic Acid Transport. BA, a known substrate
`for MCT at the BBB (Kido et al., 2000), was used to further probe the
`role of MCT in GHB transport across the BBB. [14C]BA (8.33 ␮M; 0.5
`
`

`
`Downloaded from
`
`jpet.aspetjournals.org
`
` at ASPET Journals on October 5, 2015
`
`lyze the CLin data, comparing the control (without inhibitors) versus
`test (with inhibitor). Depending on these results, statistical signifi-
`cance (p ⫽ 0.05) of the CLin data were assessed using the appropriate
`Student’s t test (with equal or unequal variances) to test for statis-
`tically significant differences (p ⬍ 0.05).
`
`Results
`Pilot GHB BBB Transport Studies. Pilot studies re-
`vealed a linear influx of [3H]GHB (0.028 ␮M) into various
`brain regions over 60 s. Figure 1 presents a representative
`concentration-time course for the hippocampus. Similar lin-
`ear time courses were observed for other brain regions (data
`not shown). A 30-s perfusion time was chosen for additional
`single time point studies because this was within the linear
`region, indicating a predominant influx with minimal efflux.
`The capillary depletion technique was performed as de-
`scribed previously (Triguero et al., 1990) to investigate the
`capillary sequestration of [3H]GHB. The [3H]GHB distribu-
`tion volume (n ⫽ 3 rats) in the homogenate, supernatant, and
`pellet (capillary) fractions were 0.071 ⫾ 0.009, 0.067 ⫾ 0.008,
`and 0.003 ⫾ 0.001 cm3/g, respectively, indicating that less
`than 5% of the total [3H]GHB was sequestered within the
`capillaries. Moreover, there were no significant differences in
`the [3H]GHB CLin determined with versus without the cap-
`illary depletion procedure (with capillary depletion, 0.119 ⫾
`0.027 cm3/min/g, n ⫽ 3; without capillary depletion, 0.075 ⫾
`0.052 cm3/min/g, n ⫽ 4). These results with GHB are consis-
`tent with the behavior of small, hydrophilic, nonpositively
`charged molecules, such as sucrose and urea (Triguero et al.,
`1990), which undergo minimal capillary sequestration. Ac-
`cordingly, the capillary depletion step was not performed in
`subsequent experiments.
`For all brain regions and concentrations, GHB CLin values
`were at least 40-fold lower than the cerebrovascular flow
`values obtained from the literature (Takasato et al., 1984;
`Allen and Smith, 2001), which suggests that GHB BBB
`transport is flow independent (data not shown).
`GHB BBB Transport Concentration Dependence.
`Figure 2 shows GHB concentration-dependent influx for two
`representative regions: the hippocampus and parietal cortex.
`Similar data were observed for the other brain regions (data
`
`Fig. 1. Time course of GHB influx into rat hippocampus after perfusion
`with 0.028 ␮M [3H]GHB. Filled circles represent mean ⫾ S.E.M. (n ⫽ 3–4
`rats). Similar results were observed for other brain regions (data not
`shown).
`
`94
`
`Bhattacharya and Boje
`
`␮Ci/ml) was perfused in presence or absence of 20 mM (GHB or BA)
`or 40 mM GHB for 30 s. Brain tissues were harvested and assayed
`for [14C]BA as described previously.
`Self-Association Studies. 1,6-Diphenyl-1,3,5-hexatriene (DPH)
`was used as a fluorescence probe for substance self-association in the
`perfusate. Sodium dodecyl sulfate, with a critical micellar concen-
`tration of 0.83 mM (Kumar Sau et al., 2002), was used as a positive
`control. Four microliters of a freshly prepared solution of DPH (5
`mM, solubilized in tetrahydrofuran) was added to separate test tubes
`containing either 2 ml of valproic or hexanoic acids (10–50 mM) or
`sodium dodecyl sulfate (0–5 mM). The solutions were kept in dark
`for 30 min before analysis. Fluorescence measurements were per-
`formed using a PTI fluorometer (Photon Technology International,
`Lawrenceville, NJ). The excitation and emission wavelengths were
`360 and 430 nm, respectively. The study was repeated in triplicate.
`Data Analysis. GHB influx clearance (CLin, cubic centimeters per
`minute per gram) for unidirectional transfer was obtained by fitting
`eq. 1 to the time course data using WinNonlin Pro version 2.1
`(Pharsight, Cary, NC):
`
`(1)
`
`⫽ CLin䡠T ⫹ Vvasc
`
`Q C
`
`where Q (dpm per gram) represents the quantity of radiotracer in the
`brain region normalized for wet brain tissue weight, C (dpm per
`milliliter) represents the perfusion fluid concentration of [3H]GHB, T
`(minutes) is the time of perfusion, and Vvasc (milliliters per gram)
`represents the volume of the cerebrovascular capillary bed for each
`brain region. Vvasc data were previously determined in our labora-
`tory using [3H]inulin for each brain region (10⫺3 cm3/g (n ⫽ 5):
`thalamus/hypothalamus, 7.97 ⫾ 0.96; hippocampus, 10.0 ⫾ 2.08;
`striatum, 7.59 ⫾ 0.85; frontal cortex, 6.16 ⫾ 0.78; occipital cortex,
`6.08 ⫾ 0.52; and parietal cortex, 6.97 ⫾ 1.19).
`For perfusion studies involving single time point analysis (a 30-s
`perfusion period, determined at different concentrations of GHB),
`CLin data were converted to cerebrovasculature permeability surface
`area products (PA, cubic centimeters per minute per gram) using eq.
`2:
`
`(2)
`
`PA ⫽ ⫺F 䡠 ln冉1 ⫺
`
`冊
`
`CLin
`F
`
`where F (cubic centimeters per minute per gram) is the perfusion
`fluid flow through each region; these values were obtained from the
`literature (Takasato et al., 1984; Allen and Smith, 2001).
`GHB mass transfer influx data (Jin, nanomoles per minute per
`gram) were calculated by eq. 3:
`
`Jin ⫽ PA 䡠 C
`
`(3)
`
`where C (millimolar) is the total perfusate concentration of GHB
`(labeled and unlabeled).
`To determine the saturability of GHB BBB influx, parameter
`estimates of Vmax, Km, and CLns were obtained by iterative nonlinear
`regression analysis (WinNonlin Pro version 2.1; Pharsight) using eq.
`4:
`
`Jin ⫽
`
`Vmax 䡠 C
`Km ⫹ C
`
`⫹ CLns䡠C
`
`(4)
`
`where Vmax (nanomoles per minute per gram) is maximal transport
`rate of GHB influx, Km (millimolar) is the Michaelis-Menten half-
`saturation constant, CLns (cubic centimeters per minute per gram) is
`the nonsaturable clearance representing passive diffusion, and C
`(millimolar) is the total concentration of GHB (labeled and unla-
`beled). A weighting scheme was used for the nonlinear regression
`analysis (iterative reweighting, 1/Y2
`predicted).
`Statistical Analysis. Statistical analyses were performed using
`SAS version 8.0 (SAS Institute, Cary, NC). A two-sample test for
`equal or unequal variance (Fischer’s test) was initially used to ana-
`
`

`
`Transport of GHB across the Blood-Brain Barrier
`
`95
`
`Downloaded from
`
`jpet.aspetjournals.org
`
` at ASPET Journals on October 5, 2015
`
`Fig. 2. Concentration dependence of GHB influx into hippocampus (A) and parietal cortex (B) over 0 to 0.1 mM (insets) or 0 to 40 mM (large graphs).
`The solid line represents the fit of eq. 4 to influx data, the dashed line represents the computer estimated saturable influx, and the dotted line
`represents the computer estimated nonsaturable (passive diffusion) influx. Filled circles represent mean ⫾ S.E.M. (n ⫽ 3–4 rats). Similar results were
`observed for other brain regions (data not shown).
`
`not shown). Michaelis-Menten BBB transport parameters for
`each region are shown in Table 1. Because the concentration-
`BBB influx data were pooled from multiple animals and
`subjected to nonlinear regression analysis as a single data
`set, the regional parameter estimates cannot be statistically
`compared against each other due to an inability to estimate
`the true variability associated with each parameter estimate.
`However, it seems that the cortices show greater Vmax esti-
`mates relative to the other regions; this is perhaps due to the
`higher capillary density of the cerebral cortex relative to
`other regions (Klein et al., 1986).
`GHB BBB Transport Inhibition Studies. Tables 2 and
`3 illustrate the effects of various compounds on the BBB
`transport of GHB. Consistent with the concentration-depen-
`dent transport studies, self-inhibition of GHB (40 mM) influx
`was observed. Short-chain monocarboxylic acids are known
`substrates of MCT (Enerson and Drewes, 2003). L-Lactic acid
`(three-carbon backbone; C3), pyruvic acid (C3), and ␤-hy-
`droxybutyric acid (four-carbon backbone; C4) each signifi-
`cantly inhibited [3H]GHB BBB influx transport (Table 2).
`
`TABLE 1
`Brain regional parameter estimates of GHB transport at the BBB
`Values are computer estimates obtained through nonlinear regression analysis.
`
`Brain Region
`
`Hippocampus
`Striatum
`Frontal cortex
`Parietal cortex
`Occipital cortex
`Thalamus/hypothalamus
`
`CLns
`Km
`Vmax
`nmol/min/g ⫻ 101 mM cm3/min/g ⫻ 10⫺2
`41.3
`7.68
`1.33
`31.0
`3.97
`2.43
`59.9
`7.98
`2.34
`119
`22.4
`1.82
`152
`21.5
`0.89
`22.5
`2.62
`2.99
`
`The dicarboxylic acids succinic acid (C4) and glutaric acid
`(five-carbon backbone; C5), did not inhibit [3H]GHB BBB
`transport (Table 2), although an unexplained stimulation of
`[3H]GHB BBB transport was observed for succinic acid in the
`striatum and thalamus/hypothalamus regions. The medium
`chain fatty acids, hexanoic acid (six-carbon backbone; C6)
`and valproic acid (eight-carbon backbone; C8), significantly
`inhibited [3H]GHB BBB transport (Table 2). The inhibition of
`[3H]GHB BBB transport by other organic anions is shown in
`Table 3. Benzoic acid and salicylic acid significantly inhibited
`[3H]GHB uptake. CHC, a specific inhibitor of MCT (Wang et
`al., 1996; Enerson and Drewes, 2003), showed significant
`inhibition of [3H]GHB BBB transport. Probenecid, which has
`a broad specificity for multiple transporters (Deguchi et al.,
`1997), significantly inhibited [3H]GHB BBB influx. GABA
`did not significantly inhibit [3H]GHB transport.
`The BBB influx of [14C]BA, a known MCT substrate, was
`significantly inhibited (p ⬍ 0.05) by unlabeled BA (20 mM)
`and GHB (40 mM) as shown in Table 4. Interestingly, 20 mM
`GHB did not significantly inhibit [14C]BA, suggesting that
`GHB has a lower affinity for the transporter than benzoic
`acid.
`The inhibition of [3H]GHB influx by a number of com-
`pounds required the use of high concentrations (20 mM) of
`inhibitors. Several control studies were performed to ascer-
`tain that the observed inhibition of [3H]GHB influx was not
`due to nonspecific physicochemical interactions that would
`impede [3H]GHB access to the transporter. Sucrose and gly-
`cine were selected as negative controls as these compounds
`undergo minimal to moderate passive diffusion. Neither su-
`crose (20 mM) nor glycine (20 mM) inhibited [3H]GHB up-
`
`

`
`Downloaded from
`
`jpet.aspetjournals.org
`
` at ASPET Journals on October 5, 2015
`
`96
`
`Bhattacharya and Boje
`
`TABLE 2
`Effect of various unlabeled compounds on influx clearance of 关3H兴GHB across different brain regions
`Each value represents mean ⫾ S.E.M.
`
`Inhibitor
`
`关3H兴GHB (control)a
`Short-chain monocarboxylic acids
`Pyruvicb
`Lacticb
`␤-Hydroxybutyricb
`␥-Hydroxybutyricc
`
`Dicarboxylic acids
`Succinicb
`Glutaricb
`Medium-chain fatty acids
`Hexanoicb
`Valproicb
`
`Hippocampus
`
`Parietal Cortex
`
`Occipital Cortex
`
`Frontal Cortex
`
`Striatum
`
`Thalamus/
`Hypothalamus
`
`Percentage of Control
`
`100 ⫾ 13.3
`
`100 ⫾ 6.00
`
`100 ⫾ 5.14
`
`100 ⫾ 7.51
`
`100 ⫾ 10.2
`
`100 ⫾ 12.6
`
`38.3 ⫾ 13.8*
`61.4 ⫾ 2.31†*
`51.5 ⫾ 7.46*
`75.8 ⫾ 7.86
`59.8 ⫾ 3.66*†
`
`51.4 ⫾ 16.9*
`75.3 ⫾ 2.47*
`72.2 ⫾ 2.12†*
`87.8 ⫾ 7.57
`72.7 ⫾ 3.16*
`
`52.0 ⫾ 12.5*
`71.5 ⫾ 5.62*
`64.8 ⫾ 3.61*
`85.5 ⫾ 5.77
`72.7 ⫾ 5.59*
`
`46.6 ⫾ 12.6*
`70.5 ⫾ 8.76*
`67.9 ⫾ 4.51*
`87.7 ⫾ 5.04
`74.9 ⫾ 3.95*
`
`48.3 ⫾ 18.2*
`87.6 ⫾ 4.84
`115 ⫾ 12.1
`87.1 ⫾ 9.98
`75.6 ⫾ 13.8
`
`38.3 ⫾ 13.3*
`87.4 ⫾ 5.79
`90.5 ⫾ 10.5
`80.6 ⫾ 11.7
`86.6 ⫾ 6.69
`
`85.6 ⫾ 9.24
`84.4 ⫾ 7.17
`
`93.3 ⫾ 9.89†
`101 ⫾ 11.5
`
`108 ⫾ 8.43
`101 ⫾ 11.5
`
`119 ⫾ 6.85
`97.7 ⫾ 17.8
`
`152 ⫾ 6.41*
`127 ⫾ 25.3
`
`219 ⫾ 38.8*
`122 ⫾ 12.7
`
`Conc.
`
`mM
`
`20
`20
`20
`20
`40
`
`20
`20
`
`43.5 ⫾ 3.56*
`38.3 ⫾ 6.08*
`11.8 ⫾ 4.83*†
`49.7 ⫾ 12.7*
`44.8 ⫾ 8.96*
`24.2 ⫾ 13.1*
`* P ⬍ 0.05; significantly different from control by Student’s t test (with equal or unequal variances).
`† P ⬍ 0.05; variance significantly different from control by Fisher’s variance test.
`a n ⫽ 8 for control group.
`b n ⫽ 4 for treatment group and cn ⫽ 6 for treatment group.
`
`20
`20
`
`40.4 ⫾ 2.44†*
`39.9 ⫾ 9.22*
`
`66.4 ⫾ 5.37*
`58.3 ⫾ 12.0*
`
`45.1 ⫾ 8.88*
`37.3 ⫾ 16.2*
`
`TABLE 3
`Effect of various unlabeled compounds on influx clearance of 关3H兴GHB across different brain regions
`Each value represents mean ⫾ S.E.M. (n ⫽ 8 for control group, n ⫽ 4 for treatment group).
`
`Percentage of Control
`
`Inhibitor
`
`关3H兴GHB (control)
`Organic anions
`Salicylic
`Benzoic
`CHC
`Probenecid
`GABA
`Negative controls
`Sucrose
`Glycine
`
`Conc.
`
`mM
`
`20
`20
`1
`10
`10
`
`Hippocampus
`
`Parietal Cortex
`
`Occipital Cortex
`
`Frontal Cortex
`
`Striatum
`
`Thalamus/
`Hypothalamus
`
`100 ⫾ 13.3
`
`100 ⫾ 6.00
`
`100 ⫾ 5.14
`
`100 ⫾ 7.51
`
`100 ⫾ 10.2
`
`100 ⫾ 12.6
`
`31.9 ⫾ 7.05*
`29.2 ⫾ 8.23*
`58.7 ⫾ 5.73*
`18.4 ⫾ 5.54*
`65.8 ⫾ 7.35
`
`50.7 ⫾ 5.48*
`44.6 ⫾ 6.18*
`72.2 ⫾ 7.78*
`33.8 ⫾ 1.09†*
`87.5 ⫾ 2.87
`
`50.7 ⫾ 5.48*
`46.5 ⫾ 6.59*
`72.2 ⫾ 7.77*
`28.7 ⫾ 2.69*
`83.8 ⫾ 6.50
`
`48.0 ⫾ 7.72*
`40.1 ⫾ 5.97*
`72 ⫾ 8.12*
`31.4 ⫾ 0.55*†
`80.1 ⫾ 8.59
`
`111 ⫾ 2.56†
`104 ⫾ 7.10
`
`57.9 ⫾ 5.23*
`56.4 ⫾ 15.7*
`89.9 ⫾ 18.5
`51.5 ⫾ 6.81*
`94.9 ⫾ 12.2
`
`110 ⫾ 17.8
`106 ⫾ 17.0
`
`54.6 ⫾ 15.1*
`47.3 ⫾ 10.8*
`89.2 ⫾ 7.22
`45.2 ⫾ 3.28*†
`112 ⫾ 9.09
`
`86.6 ⫾ 12.3
`100 ⫾ 20.1
`
`92.8 ⫾ 4.78†
`111 ⫾ 1.58
`106 ⫾ 3.16
`113 ⫾ 15.7
`88.7 ⫾ 12.7
`105 ⫾ 10.1
`* P ⬍ 0.05; significantly different from control by Student’s t test (with equal or unequal variances).
`† P ⬍ 0.05; variance significantly different from control by Fisher’s variance test.
`
`20
`20
`
`TABLE 4
`Effect of various unlabeled compounds on influx clearance of 关3H兴GHB across different brain regions
`Each value represents mean ⫾ S.E.M.
`
`Percentage of Control
`
`Inhibitor
`
`Hippocampus
`
`Parietal Cortex
`
`Occipital
`Cortex
`关14C兴BA (control)
`100 ⫾ 4.65
`100 ⫾ 3.67
`100 ⫾ 6.09
`64.6 ⫾ 7.09*
`58.1 ⫾ 3.96*
`59.7 ⫾ 6.12*
`关14C兴BA ⫹ BA 20 mMa
`关14C兴BA ⫹ GHB 20 mMb
`86.2 ⫾ 4.70
`88.1 ⫾ 4.29
`89.3 ⫾ 6.16
`关14C兴BA ⫹ GHB 40 mMa
`76.1 ⫾ 3.37*
`73.7 ⫾ 2.16*
`70.9 ⫾ 1.11*†
`* P ⬍ 0.05; significantly different from control by Student’s t test (with equal or unequal variances).
`† P ⬍ 0.05; variance significantly different from control by Fisher’s variance test.
`a n ⫽ 4 for control group and n ⫽ 3 for treatment group.
`b n ⫽ 4 for control group and n ⫽ 9 for treatment group.
`
`Frontal Cortex
`
`Striatum
`
`100 ⫾ 3.79
`53.3 ⫾ 2.88*
`89.1 ⫾ 4.32
`68.4 ⫾ 1.18*
`
`100 ⫾ 6.42
`67.3 ⫾ 7.52*
`91.1 ⫾ 7.29
`96.5 ⫾ 7.26
`
`Thalamus/
`Hypothalamus
`100 ⫾ 5.20
`54.2 ⫾ 4.08*
`84.6 ⫾ 5.20
`66.1 ⫾ 2.70*
`
`take (Table 3), suggesting that high millimolar concentra-
`tions of substances do not necessarily physiochemically
`interact with [3H]GHB to sequester it from access to the
`transporter.
`Another potential artifact that might explain the inhibi-
`tion [3H]GHB influx by medium-chain fatty acids could be
`the entrapment of [3H]GHB in self-associative structures
`formed by the fatty acids. Such self-associative structures
`would have the effect of reducing [3H]GHB influx clearance.
`
`The formation of self-associative structures was studied us-
`ing fluorescence. DPH, a fluorescence probe for self-associa-
`tion, inserts itself into the self-associated structure, resulting
`in an increased fluorescence signal. The positive control so-
`dium dodecyl sulfate showed a steep concentration-depen-
`dent increase reaching a plateau phase at higher concentra-
`tions, consistent with the formation of micelles saturated
`with DPH probe (data not shown). Valproic and hexanoic
`acids, medium-chain fatty acids, did not show any concentra-
`
`

`
`Downloaded from
`
`jpet.aspetjournals.org
`
` at ASPET Journals on October 5, 2015
`
`tion increase in fluorescence with increasing concentrations
`(data not shown), suggesting that these compounds do not
`self-associate. These results rule out a fatty acid inhibition
`mechanism based on self-association.
`
`Discussion
`A careful reexamination of data published in 1966 by Roth
`and Giarman (1966) revealed a decrease in the GHB brain-
`to-blood ratio in the presence of ␤-hydroxybutyrate. Because
`␤-hydroxybutyrate is an MCT substrate (Enerson and
`Drewes, 2003) and GHB inhibits MCT in erythrocytes and
`cardiac myocytes (Poole and Halestrap, 1993), it was hypoth-
`esized that GHB undergoes carrier-mediated transport via
`BBB MCT.
`Using a rat in situ brain perfusion preparation, the present
`study observed that the kinetics of GHB BBB influx is char-
`acterized as a saturable, carrier-mediated process (average
`Km of ⬃11 mM, range of 2.62–22.4 mM among the various
`brain regions; average Vmax of ⬃709 nmol/min/g, range of
`225-1520 nmol/min/g range among the various brain regions)
`and a nonsaturable, diffusional process (average CLns of
`⬃0.019 cm3/min/g, range of 0.0089–0.0299 cm3/min/g).
`These Km and Vmax values are consistent with low-affinity,
`high-capacity transport and are comparable with those val-
`ues observed for other substrates of the MCT (Pollay and
`Stevens, 1980; Kido et al., 2000) and the medium-chain fatty
`acid transporter (Adkison and Shen, 1996). At low GHB
`concentrations (⬃1 mM), the saturable transport pathway
`contributes an estimated 67 to 89% of the total influx trans-
`port across the various rodent brain regions. In humans
`(where endogenous plasma GHB concentrations are typically
`less than 10 ␮M; Fieler et al., 1998), it is likely the GHB
`carrier-mediated process will predominate over the diffu-
`sional process, i.e., the Vmax/Km is 3.3-fold greater than CLns
`assuming that the average Vmax and Km estimates in rats are
`similar to humans.
`The characteristics of the BBB transport protein responsi-
`ble for GHB influx was pharmacologically probed using a
`diverse set of potential inhibitors. Short-chain monocarboxy-
`lic acids known to be transported by the BBB MCT1, e.g.,
`lactic, pyruvic, and ␤-hydroxybutyrate (Tildon and Roeder,
`1988; Kang et al., 1990; Enerson and Drewes, 2003), signif-
`icantly inhibited [3H]GHB influx clearance, suggesting that
`MCT1 contributes to BBB GHB transport. Salicylic acid [an
`MCT and organic anion transport (OAT) substrate], benzoic
`acid (a well known MCT substrate; Kang et al., 1990; Ter-
`asaki et al., 1991; Kido et al., 2000), and CHC (a specific
`inhibitor of MCT; Wang et al., 1996; Enerson and Drewes,
`2003] showed significant inhibition of GHB influx, again
`implicating a MCT isoform in the BBB transport of GHB.
`Because MCT substrates inhibited GHB influx, it was nec-
`essary to determine whether GHB inhibited the BBB trans-
`port of a known MCT substrate across the BBB, e.g., benzoic
`acid. GHB, as well as unlabeled benzoic acid, significantly
`inhibited [14C]benzoic acid influx. The mutual inhibitory in-
`teraction of GHB and benzoic acid on each other’s influx
`implicates a role of MCT in the BBB transport of GHB.
`Presently, it is known that the endothelial cells of the BBB
`express MCT1 but little MCT2. However, review of literature
`suggests that other isoforms of MCT (MCT6 and 7) are also
`expressed in the brain (Price et al., 1998), with the trans-
`
`Transport of GHB across the Blood-Brain Barrier
`
`97
`
`porter localization yet to be defined. Thus, the possibility that
`more than one isoform of MCT could be involved in GHB
`transport across the BBB cannot be ignored.
`Substrates of other transporters were studied for their
`effects on GHB transport. The dicarboxylic acids glutaric and
`succinic, which are substrates of OAT and not MCT (Lee et
`al., 2001), did not inhibit GHB influx clearance, thereby
`implying that OAT is not likely involved in GHB influx.
`The role of the medium-chain fatty acid transporter in
`GHB transport was studied using hexanoic and valproic ac-
`ids (Adkison and Shen, 1996). Both valproic and hexanoic
`acids significantly inhibited GHB brain influx, which may
`implicate a fatty acid transporter for GHB influx. Adkison
`and Shen (1996) observed that valproic acid inhibited the
`BBB influx of MCT substrates, but MCT substrates failed to
`inhibit valproic acid BBB influx. This lack of mutual inhibi-
`tion suggested that valproic acid, although not transported
`by MCT, can interact with MCT to prevent substrate trans-
`port.
`Probenecid inhibits a wide array of transport