Multiple Sclerosis Resource Centre

Welcome to the Multiple Sclerosis Resource Centre. This website is intended for international healthcare professionals with an interest in Multiple Sclerosis. By clicking the link below you are declaring and confirming that you are a healthcare professional

You are here

GABA transport and neuroinflammation are coupled in multiple sclerosis: Regulation of the GABA transporter-2 by ganaxolone

Neuroscience, pages 24 - 38


Interactions between neurotransmitters and the immune system represent new prospects for understanding neuroinflammation and associated neurological disease. GABA is the chief inhibitory neurotransmitter but its actions on immune pathways in the brain are unclear. In the present study, we investigated GABAergic transport in conjunction with neuroinflammation in models of multiple sclerosis (MS). Protein and mRNA levels of γ-amino butyric acid transporter 2 (GAT-2) were examined in cerebral white matter from MS and control (Non-MS) patients, in cultured human macrophages, microglia and astrocytes, and in spinal cords from mice with and without experimental autoimmune encephalomyelitis (EAE) using western blotting, immunocytochemistry and quantitative real-time polymerase chain reaction (qRT-PCR). GABA levels were measured by HPLC. The GAT-2′s expression was increased in MS patients’ (n = 6) white matter, particularly in macrophage lineage cells, compared to Non-MS patients (n = 6) (p < 0.05). Interferon-γ (IFN-γ) stimulation of human macrophage lineage cells inducedGAT-2 expression and reduced extracellular GABA levels (p < 0.05) but soluble GABA treatment suppressedHLA-DRα,GAT-2 andXBP-1/s expression in stimulated macrophage lineage cells (p < 0.05). Similarly, the synthetic allopregnanolone analog, ganaxolone (GNX), repressed GAT-2, JAK-1 and STAT-1 expression in activated macrophage lineage cells (p < 0.05).In vivoGNX treatment reducedGat-2,Cd3ε,MhcII, andXbp-1/s expression in spinal cords following EAE induction (p < 0.05), which was correlated with improved neurobehavioral outcomes and reduced neuroinflammation, demyelination and axonal injury. These findings highlight altered GABAergic transport through GAT-2 induction during neuroinflammation. GABA transport and neuroinflammation are closely coupled but regulated by GNX, pointing to GABAergic pathways as therapeutic targets in neuroinflammatory diseases.

Abbreviations: BBB - blood–brain barrier, DMSO - dimethyl sulphoxide, EAE - experimental autoimmune encephalomyelitis, FBS - fetal bovine serum, GAT-2 - γ-amino butyric acid transporter 2, GNX - ganaxolone, HFAs - human fetal astrocytes, HFMs - human fetal microglia, HPLC - high-performance liquid chromatography, IFN-γ - interferon-γ, MBP - myelin basic protein, MDMs - monocyte-derived macrophages, MEM - minimal essential medium, MOG - myelin oligodendrocyte glycoprotein, MS - multiple sclerosis, PBLs - peripheral blood lymphocytes, PBMCs - peripheral blood mononuclear cells, PBS - phosphate-buffered saline.

Key words: multiple sclerosis, experimental autoimmune encephalitis, neuroinflammation, γ-amino butyric acid, γ-amino butyric acid transporter-2, ganaxolone.


There is a growing recognition that innate and adaptive immune processes are involved in neuroinflammation and subsequent neurodegeneration, which contribute to Alzheimer’s and Parkinson’s diseases, HIV-associated neurocognitive disorders and multiple sclerosis (MS) (Eikelenboom et al, 2002, Frohman et al, 2006, Noorbakhsh et al, 2006, and Tansey et al, 2008). MS is the prototype neuroinflammatory disease of the CNS defined by inflammatory demyelination and axonal loss ( Stadelmann, 2011 ). Defects in GABA signaling have been implicated in MS and other inflammatory neurodegenerative diseases (Demakova et al, 2003 and LeWitt et al, 2011). GABA has been reported also to regulate lymphoid and macrophage lineage cell activation (Stuckey et al, 2005 and Nigam et al, 2010). GABA’s reported effects on the immune system include the modulation of pro-inflammatory cytokine and chemokine production (Reyes-Garcia et al, 2007 and Tian et al, 2011) and diminished T cell proliferation (Shiratsuchi et al, 2009, Dionisio et al, 2011, and Soltani et al, 2011). GABA’s actions have been studied in the MS animal model, experimental autoimmune encephalitis (EAE) in which GABA-specific therapeutics displayed differential outcomes, highlighting the complexity of the GABAergic system ( Carmans et al., 2013 ).

The GABA (reuptake) transporters (GATs) are essential regulators of extra- and intra-cellular GABA levels ( Conti et al., 2004 ). Moreover, GATs’ expression is induced in macrophages and lymphocytes following immune activation, presumably leading to enhanced GABA cellular reuptake (Bhat et al, 2010 and Dionisio et al, 2011). Neuroactive steroids are produced within the nervous system ( Mellon et al., 2008 ) and bind to both intracellular steroid receptors and/or cell surface neurotransmitter receptors (Rupprecht et al, 1993, Robel and Baulieu, 1994, and Rupprecht, 1997). Several neuroactive steroids are positive allosteric modulators of the GABAAreceptor (GABA-A-R), including allopregnanolone, or they act to antagonize the GABA-A-R, such as DHEA-S (Rupprecht, 1997, Falkenstein et al, 2000, Hosie et al, 2006, and Mellon et al, 2008). Allopregnanolone is a cholesterol-derived neuroactive steroid, which is synthesized by neurons and glia (Rupprecht et al, 1993, Belelli and Lambert, 2005, and Houtchens, 2007). Importantly, allopregnanolone levels were decreased in cerebral white matter from MS patients and treatment with allopregnanolone reduced neuroinflammation and demyelination in EAE ( Noorbakhsh et al., 2011 ). Ganaxolone (GNX) is a synthetic analog of allopregnanolone ( Reddy and Woodward, 2004 ) and binds to a GABA-A-R site ( Carter et al., 1997 ). Studies in neurons indicate that this interaction enhances chloride (Cl) permeability of the GABA-A ionophore receptor complex, which results in hyperpolarization of the post-synaptic membrane (Carter et al, 1997, Jorgensen, 2005, and Hosie et al, 2006). GNX has been investigated as an anti-epileptic drug in animal studies (Carter et al, 1997, Liptakova et al, 2000, and Reddy and Rogawski, 2010) and in human studies for epilepsy treatment ( Reddy, 2010 ) although its effects on the immune system remain unknown. Herein, we report the induction of γ-amino butyric acid transporter 2 (GAT-2) (SLC6A12) in the brains of MS patients, particularly on macrophage lineage cells, and also in animals with EAE. GNX suppressed GAT-2 expression in macrophage lineage cells leading to reduced neuroinflammatory gene expression with reduced disease severity in EAE.

Experimental procedures

Human brain tissue samples

Frontal brain white matter was collected at autopsy, with consent, from age and sex-matched MS subjects (n = 6; relapsing-remitting (RR-MS;n = 3), secondary progressive MS (n = 3)) and Non-MS patients (n = 6; sepsis, myocardial infarction, cancer, HIV/AIDS, all without neuropathological lesions), as previously reported ( Deslauriers et al., 2011 ). These studies were approved by the University of Alberta Ethics Committee (Pro00002291).

Cell cultures

Human monocyte-derived macrophages (MDMs) and peripheral blood lymphocytes (PBLs) were prepared from blood collected from healthy donors as previously described (Tsutsui et al, 2004 and Acharjee et al, 2011). Briefly, human peripheral blood mononuclear cells (PBMCs) were purified from healthy blood with Histopaque (Sigma–Aldrich, Oakville, ON, Canada) ( Power et al., 1998 ). PBLs were isolated from PBMCs by removing adherent macrophages and maintained in RPMI 1640 medium with 15% fetal bovine serum (FBS) with phytohemagglutinin-P (PHA-P) (5 μg/ml) (Sigma–Aldrich, Oakville, ON, Canada) for 3 days. Cells were then harvested and cultured with anti-human anti-CD3 (5 μg/ml)-coated plates supplemented with recombinant human IL-2 (103 U/ml) (PeproTech, Rocky Hill, NJ, USA) for an additional 24 h with either GNX (100 μM) or dimethyl sulphoxide (DMSO) (Sigma–Aldrich, Oakville, ON, Canada) vehicle control in equal volume. MDMs were isolated from PBMCs, differentiated for 1 week ( Maingat et al., 2009 ). MDMs were pretreated with GNX (100 μM) or DMSO in minimal essential medium (MEM) containing 10% FBS (Life Technologies, Burlington, ON, Canada), 1% penicillin/streptomycin (Life Technologies, Burlington, ON, Canada) and 1%l-glutamine (Life Technologies, Burlington, ON, Canada), followed by recombinant human interferon-γ (IFN-γ) (400 U/mL) (PeproTech, Rocky Hill, NJ, USA) exposure for 24 h. Treated MDMs were lysed in either TRIzol® reagent (Life Technologies, Burlington, ON, Canada) followed by total RNA extraction or Laemmli buffer with 0.1% β-mercaptoethanol (Sigma–Aldrich, Oakville, ON, Canada) for protein isolation.

Human fetal astrocytes (HFAs) and microglia (HFMs) were prepared from 15–19-week-old fetal brains obtained, with consent (approved by the University of Alberta Ethics Committee, Pro00027660), as previously described (Zhu et al, 2009 and Vivithanaporn et al, 2010). Briefly, fetal brain tissues were dissected, meninges were removed and digested for 30 min with 2.5% trypsin (Life Technologies, Burlington, ON, Canada) and 2 mg/ml DNase I (Roche Diagnostics, Mannheim, Germany) and passed through a 70-μm cell strainer (BD Biosciences, Mississauga, ON, Canada). Cells were washed twice with centrifugation at 14,000 rpm for 10 min and plated in T-75 poly-l-ornithine-coated flasks (Sigma–Aldrich, Oakville, ON, Canada) at 6–8 × 107 cells/flask with media. Following a week of incubation (37 °C at 5% CO2), adherent cells (HFA) were separated from suspension cells (HFM) and re-plated for primary astrocytic or microglia differentiation (80% confluence). HFAs and HFMs were maintained in MEM containing 10% FBS (Life Technologies, Burlington, ON, Canada), 1% penicillin/streptomycin (Life Technologies, Burlington, ON, Canada), 1%l-glutamine (Life Technologies, Burlington, ON, Canada) and 10% dextrose (Life Technologies, Burlington, ON, Canada).


C57BL6 female mice were purchased from the Jackson Laboratory and maintained in the Health Sciences Laboratory Animal Services facility of the University of Alberta under conventional housing conditions. All experiments were approved by the University of Alberta Animal Care Committee (AUP00000317).

GNX preparation

GNX (5α-pregnan-3β-methyl-3α-ol-20-one) (Steraloids Inc., Newport, RI, USA) was solubilized in DMSO and further diluted in 1× phosphate-buffered saline (PBS) (Life Technologies, Burlington, ON, Canada) and stored at 4 °C. Active therapeutic concentrations of GNX were derived from previous studies ( Ram et al., 2001 ). For intraperitoneal (i.p.) injections, GNX and DMSO were diluted in PBS to concentrations of low dose (15 mg/kg) and high dose (50 mg/kg), adjusted to physiological pH (7.4) and filter sterilized prior to administration. Forin vitrocell cultures, GNX and DMSO were further diluted in MDM or PBL MEM media (100 μM).

EAE induction

Ten to twelve- week-old C57BL6 female mice were injected subcutaneously in each hind leg with 50 μg of myelin oligodendrocyte glycoprotein peptide (MOG35–55peptide) (Peptide Synthesis Facility, University of Alberta, Edmonton, AB, Canada) emulsified in complete Freund’s adjuvant (Sigma–Aldrich, Oakville, ON, Canada) supplemented with 5 mg/mol heat-killed mycobacteria H37 RA (Difco Laboratories, Detroit, MI, USA) (1:1) ( Ellestad et al., 2009 ). Animals received 0.3-μg i.p. injections of pertussis toxin (List Biological Laboratories, Campbell, CA, USA) on the day of induction with MOG35–55and were then boosted 48 h later. Control animals were immunized with complete Freund’s adjuvant and 1× PBS (1:1) and boosted with pertussis toxin, as above. Animals were assessed daily for EAE severity (behavioral score) for 30 days using a 0–14 rating scale ( Ellestad et al., 2009 ). EAE mice received i.p. injections of GNX (15 mg/kg and 50 mg/kg) daily beginning on the day of EAE induction (Day 0). Control animals received daily i.p. injections of DMSO in matched volumes to those used in the GNX-treated animals.

High-performance liquid chromatography (HPLC)

Fluorometric HPLC was adapted from a previously described protocol (Grant et al, 2006 and Maingat et al, 2009). Cell culture supernatants were mixed to a 60× final concentration with 100% methanol then centrifuged at 12,000gfor 4 min at 4 °C. A 1:1 (5 μl) aliquot of the supernatant was mixed with derivatizing agent (2 mg ofN-isobutyryl-l-cysteine and 1 mg o-phthaldialdehyde dissolved in 100 μl of methanol, followed by addition of 900 μl of 0.1 M sodium borate buffer), and then was placed into a sample management system (Waters Alliance 2690XE; Waters Corporation, Milford, MA, USA). HPLC separation was achieved on a Symmetry C18 column (4.6 × 150 mm; 3.5 μm particle diameter) coupled with a guard column of the same stationary phase (Waters Corporation, Milford, MA, USA). The column heater was set to 30 °C and the sample cooler was held at 4 °C. To separate the derivatized amino acids of interest, a gradient was established from equal parts of solvents A (850 ml of 0.04 M sodium phosphate buffer and 150 ml of methanol, pH 6.2) and B (670 ml of 0.04 M sodium phosphate buffer, 555 ml of methanol, and 30 ml of tetrahydrofuran, pH 6.2) to only solvent B by ∼45 min, with a flow rate of 500 μl/minute. The run time was 60 min for column washout and equilibrium, and 30 min to elute all compounds. A Waters 2475 fluorescence detector (Waters Corporation, Milford, MA, USA) was used to quantify the eluted derivatized GABA (excitation, 344 nm; emission, 433 nm).

Western blotting

Human MDMs were plated in 12 well plates until 80% confluent, pre-treated with IFN-γ (400U/ml) for 1 h, followed by GNX (100 μM) or DMSO (comparable volume) for 24 h. THP-1 cells were plated at 2.0 × 105 cells/well in 12 well plates and treated under the same conditions. Cells were then lysed with Laemmli buffer with 0.1% β-mercaptoethanol and boiled at 95 °C for 10 min. Proteins from whole-cell lysates were separated using polyacrylamide gel electrophoresis and protein fractions were transferred to a nitrocellulose membrane overnight (Bio-Rad, Mississauga, ON, Canada) ( Zhu et al., 2009 ). The membrane was blocked with 5% milk for 1 h and labeled with monoclonal mouse anti-STAT-1 (p86/p90) (1:1000, Cell Signaling Technology, Boston, MA, USA), monoclonal mouse anti-p-STAT-1 (Tyr701) (1:500, Abcam, Eugene, OR, USA), monoclonal mouse anti-JAK-1 (1:1000, Cell Signaling Technology, Boston, MA, USA) and polyclonal rabbit anti-GAT-2 (1:1000, GeneTex, Inc., Irvine, CA, USA) overnight at 4°. The immunolabeled membrane was then probed with secondary peroxidase-conjugated goat anti-rabbit IgG (1:500, Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) for 2 h. β-Actin-HRP was used as a loading control (1:1000, Santa Cruz Biotechnology, Inc., Dallas, TX, USA). Membranes were developed with Pierce ECL Western blotting substrate (Fisher Scientific, Ottawa, ON, Canada) and exposed on film (Canon Canada, Inc., Mississauga, ON, Canada).

Immunofluorescence and immunohistochemistry

MDM or THP-1 cultures were stimulated with IFN-γ (400 U/ml) for 1 h followed by GNX (100 μM) or DMSO for 24 h. Cells were fixed with 2% paraformaldehyde (Fisher Scientific, Ottawa, ON, Canada) for 30 min, washed 3× with PBS + 2% FBS, blocked with Odessey blocking buffer (Li-cor, Guelph, ON, Canada) for 2 h and incubated overnight at 4° with polyclonal rabbit anti-GAT-2 (1:1000, Abcam, Eugene, OR, USA) and monoclonal mouse anti-HLA-DR (1:100) (Abcam, Eugene, OR, USA) primary antibodies. Following overnight incubation, fluorescent conjugated secondary antibodies (Alexa488 goat anti-mouse (Invitrogen, Eugene, ON, Canada) and Alexa Fluor 680 goat anti-rabbit (Invitrogen, Eugene, ON, Canada) were incubated for 2 h and washed 3× with PBS + 2% FBS. Cells were washed and mounted with gelvatol on microscope slides and imaged on an upright fluorescent microscope (Axioskop2; Carl Zeiss MicroImaging Inc., Thornwood, NY, USA) Spinal cords isolated from untreated and treated animals (n = 4) were formalin-fixed (4%) and paraffin-embedded. The subsequent sections were de-paraffinized, hydrated, and boiled in 0.01 M trisodium citrate buffer (pH 6) for 10 min for antigen retrieval. Sections were incubated in 0.3% hydrogen peroxide for 20 min. Sections were blocked with Odyssey blocking buffer (Li-cor, Guelph, ON, Canada) for 2 h at room temperature then probed with polyclonal rabbit anti-CD3 (1:100, Abcam, Eugene, OR, USA), monoclonal mouse anti-MHCII (1:50, eBioscience, San Diego, CA, USA), monoclonal mouse anti-STAT-1 (1:1000, Cell Signaling Technology, Boston, MA, USA), polyclonal rabbit anti-GAT-2 (1:1000, Abcam, Eugene, OR, USA), polyclonal rabbit anti-GFAP (1:500, BD Biosciences, Mississauga, ON, Canada) and monoclonal mouse myelin basic protein (MBP) (1:500, Abcam, Eugene, OR, USA) antibodies were used to detect lymphocyte infiltration, MHC Class II activation, GAT-2 and MBP abundance. Bielschowsky’s silver staining was performed for axonal identification, as described previously ( Tsutsui et al., 2004 ). Formalin-fixed (4%), paraffin-embedded autopsy white matter from MS and Non-MS samples were prepared as above. Immunolabeling with polyclonal rabbit anti-GAT-2 (1:1000, Abcam, Eugene, OR, USA) antibodies and monoclonal mouse anti-HLA-DRα (1:100, Abcam, Eugene, OR, USA) were used to detect co-localization of GAT-2 expression in MHC Class II+cells conjugated to DAB or alkaline peroxidase, anti-rabbit or anti-mouse secondary antibodies, respectively (Life Technologies, Burlington, ON, Canada). DAB peroxidase substrate kit (Vector Laboratories, Burlington, ON, Canada) or alkaline phosphatase substrate kit (Vector Laboratories, Burlington, ON, Canada). Sections were dehydrated with xylene, mounted with acrytol (Leica Biosystems, Buffalo Grove, IL, USA) and viewed under a light microscope (Axioskop2; Carl Zeiss MicroImaging Inc., Thornwood, NY, USA). Immunofluorescence images were quantified as mean pixel intensity using ImageJ software ( Schneider et al., 2012 ).

Flow cytometry

Human MDMs were prepared as above with surface phenotype labeling of activated MDMs with anti-HLA-DR and GAT-2 and gating was assessed by total percentage (%) of the HLA-DR+/GAT-2+population compared to secondary only controls. Primary antibodies included PE-conjugated mouse anti-HLA (1:100, eBioscience, San Diego, CA, USA), polyclonal rabbit anti-GAT-2 (1:200, Abcam, Eugene, OR, USA; immunofluorescence), polyclonal rabbit anti-GAT-2 (1:100, GeneTex, Irvine CA; Western blot) and PE-conjugated anti-rabbit secondary IgG (1:500, eBioscience, San Diego, CA, USA) antibodies. Mouse splenocytes were harvested, permeabilized with permeabilization buffer (eBioscience, San Diego, CA, USA) and stained with efluor450-conjugated mouse anti-T-bet (1:100, eBioscience, San Diego, CA, USA) or PerCPCy5.5-conjugated mouse anti-GATA3 (1:100, eBioscience, San Diego, CA, USA). Cells were fixed, washed with PBS + 2% FBS and subsequently stained with PerCP-conjugated mouse anti-CD3ε (1:1000, eBioscience, San Diego, CA, USA) and APC-conjugated mouse anti-CD4 (1:1000, eBioscience, San Diego, CA, USA). CellTrace CFSE™ cell proliferation kit (Molecular Probes, Eugene, OR, USA) was used for proliferation assays, according to the manufacturer’s protocol. In all experimental procedures, anti-CD3ε, anti-CD4, anti-GAT-2, anti-HLA, anti-Tbet, and anti-GATA3 antibodies were compensated with compbeads (BD Biosciences, Mississauga, ON, Canada) for single color compensation controls. The immunolabeled cells were analyzed on BD FACSCanto flow cytometer (BD Biosciences, Mississauga, ON, Canada) using FACSDiva software (BD Biosciences, Mississauga, ON, Canada) for acquisition and FlowJo for analysis (Tree Star Inc., Ashland, OR, USA).

Cell proliferation assay

Splenocytes were isolated from EAE induced C57BL/6 mice at day 30 post induction and stained with Cell Trace CFSE cell proliferation kit (Invitrogen, Eugene, OR, USA) (1 μM), according to the manufacturer’s protocol ( Ellestad et al., 2009 ). Briefly, total splenocytes were seeded at 2 × 105 cells/well in a 96-well round-bottomed plate with or without MOG3555(20 μg/ml) peptide for 5 days, followed by subsequent fixation and flow cytometric analysis, as above. Dulbecco’s modification of Eagle’s medium (DMEM) supplemented with 10% FBS (Life Technologies, Burlington, ON, Canada), 1% penicillin/streptomycin (Life Technologies, Burlington, ON, Canada), 2 mMl-glutamine (Life Technologies, Burlington, ON, Canada) and 1% nonessential amino acids (Life Technologies, Burlington, ON, Canada) were used for cell culture.

Quantitative real-time RT-PCR

Human white matter (normal appearing white matter), HFAs, HFMs, hMDMs and murine lumbar spinal cords were homogenized in TRIzol® reagent (Life Technologies, Burlington, ON, Canada) and total RNA was isolated and purified using RNeasy mini columns (Qiagen, Alameda, CA, USA) (Noorbakhsh et al, 2006, Antony et al, 2007, and Ellestad et al, 2009). Complementary (cDNA) was prepared by oligo(dT)-primed reverse transcription of mRNA using either Superscript II or Superscript III reverse transcriptase (Invitrogen, Eugene, ON, Canada) according to the manufacturer’s recommended protocols. Quantitative real-time PCR (qRT-PCR) was performed using Bio-Rad iQ 2× SYBRgreen supermix (Bio-Rad, Mississauga, ON, CA) on the iQ5 cycler (Bio-Rad, Mississauga, ON, CA) according to the manufacturer’s recommended protocols. All data were analyzed using the ΔΔCt method and described as relative fold change (RFC). Specific oligonucleotide primer sequences are provided in Table 1 .

Table 1 Species-specific PCR primers

Oligonucleotide name Sense Antisense

Statistical analyses

All statistical tests were performed using GraphPad InStat version 3.0 (GraphPad Software, La Jolla, CA, USA) with both parametric and nonparametric comparisons.pvalues of <0.05 were considered significant ().


Increased GAT-2 in MS white matter

Previous studies have implicated GABA’s actions on leukocytes in the regulation of neuroinflammation (GAT-2Bhat et al, 2010 and Carmans et al, 2013), prompting the investigation of GABA transporter expression in cerebral white matter.GAT-2transcript levels were significantly increased in white matter specimens from MS compared to Non-MS patients whileGAT-1andGAT-3transcript levels were similar between groups ( Fig. 1 A). Western blotting showed increased GAT-2 immunoreactivity in MS (n = 2) compared to Non-MS patients’ (n = 2) white matter lysates ( Fig. 1 B). Immunohistochemical detection of MHC Class II (HLA-DRα) in cerebral white matter sections from Non-MS, MS (normal appearing white matter, MS-NAWM) and MS (demyelinating lesion, MS-Lesion) patients ( Fig. 1 C, top panels) disclosed minimal immunoreactivity in Non-MS white matter, occasional microglial cells in MS-NAWM and numerous immunopositive cells in the MS-Lesion white matter. Similarly, GAT-2 immunoreactivity was minimal in Non-MS white matter, but detected on rare cells and as background in MS-NAWM, but apparent in cells and interstitial tissue in MS-Lesion white matter; GAT-2 (brown) was co-localized with MHC Class II (blue) immunoreactivity in MS-Lesion ( Fig. 1 C inset (i), bottom panels). These data suggested that GAT-2 was up-regulated in MS white matter, particularly brain macrophage lineage cells.


Fig. 1 GAT-2 expression in cerebral white matter. (A)GAT-2was increased within white matter of MS brains compared to Non-MS samples. Other GABA reuptake transporters,GAT-1andGAT-3did not differ between groups. (B) Immunoblotting of white matter revealed GAT-2 immunoreactivity in tissue from MS white matter (WM) but not in Non-MS WM. (C) MHC Class II (top panels) and GAT-2 (bottom panels) immunoreactivity in Non-MS white matter, MS normal appearing white matter (NAWM) and MS demyelinating lesion (MS-Lesion) in white matter. (i) MS-Lesion white matter showing cellular co-localization of MHC Class (blue) and GAT-2 (brown). Data presented as mean ± SEM (Student’s unpairedt-test,p < 0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

IFN-γ dependent induction of GAT-2 in human macrophages

As GAT-2 induction was most evident on brain macrophage lineage cells expressing MHC Class II in MS lesions, we investigated cultured human microglia, MDMs and astrocytes, all of which are capable of expressing MHC Class II. Following IFN-γ stimulation, increasedMHC-IIexpression was observed in all cell types ( Fig. 2 A).GAT-2was also induced in stimulated MDMs but not in astrocyte or microglial cultures ( Fig. 2 A). To verify these transcriptional changes, MHC Class II and GAT-2 immunoreactivity was examined in human MDMs revealing their expression was minimal in unstimulated MDMs ( Fig. 2 B, D) but IFN-γ stimulation significantly increased the immunoreactivity for both proteins in MDMs ( Fig. 2 C, D). In addition to increased expression of GAT-2, MDMs exposed to IFN-γ also showed decreased extracellular GABA levels in culture supernatants compared to unexposed MDMs’ supernatants ( Fig. 2 E). To explore the direct cellular effects of GABA, the human macrophage lineage (THP-1) cell line was stimulated with recombinant human IFN-γ and then treated with GABA (at physiologically relevant concentrations). These studies disclosed that GABA (300 ng/ml) exposure significantly reduced the expression of bothHLA-DRα( Fig. 2 F) andGAT-2( Fig. 2 G) in stimulated MDMs. Moreover, the endoplasmic reticulum stress marker,XBP-1/s, showed reduced expression in stimulated MDMs exposed to GABA ( Fig. 2 H). Thus, GAT-2 was induced in MDMs by IFN-γ, which was associated with reduced extracellular GABA levels. Moreover, increased extracellular GABA prevented IFN-γ-mediated induction of MHC Class II, GAT-2 and the ER stress indicator, XBP-1s.


Fig. 2 Immune activation diminishes GABA levels and induces GAT-2 expression in human macrophage lineage cells. (A) Transcript levels ofHLA-DRαandGAT-2in primary human cell cultures activated with IFN-γ (100U/ml (HFA and HFM) and 500 U/ml (MDM) for 24 h revealed that bothHLA-DRαandGAT-2transcript levels were inducible in human HFA, HFM and MDM cultures stimulated with IFN-γ. MDMs stimulated with IFN-γ showed increased expression ofGAT-2andHLA-DRαtranscripts compared to unstimulated controls (n = 3, repeated 3×). (B, C) MDMs stimulated with IFN-γ displayed increased expression of GAT-2 (red)/HLA-DRα (green) immunofluorescence compared to untreated controls (n = 3, repeated twice). (D) Images were quantified (mean pixel intensity) for median GAT-2, MHC Class II, and Merge immunofluorescence. (E) MDM culture supernatants were harvested and GABA levels were measured by HPLC (n = 3, repeated twice). (F-H) IFN-γ-activated macrophage lineage cell (THP-1) cultures showed reducedHLA-DRα,GAT-2andXBP-1/sfollowing GABA treatment (n = 6, 1×). Data presented as mean ± SEM (Mann-Whitney U test, Student’s unpairedt-test,p < 0.05). Scale bar = 10 μm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

GNX suppresses inflammatory gene expression in human macrophage lineage cells

Since GABA was found to inhibit IFN-γ dependent responses in macrophages, we examined the effects of the positive allosteric GABA-A-R modulator, GNX in human MDMs. IFN-γ stimulated MDMs receiving GNX treatment displayed a reduction inHLA-DRα( Fig. 3 A) andGAT-2( Fig. 3 B) transcript levels, relative to DMSO (vehicle)-treated MDMs. Flow cytometric analyses confirmed that IFN-γ stimulated HLA-DRα+expression in human MDMs (treated with DMSO) with increased GAT-2 expression but GAT-2 was suppressed with GNX treatment (GAT-2+/HLA-DRα+%) ( Fig. 3 C). Likewise confocal imaging of MDMs showed that IFN-γ stimulated cells exhibited robust GAT-2 and HLA-DRα expression ( Fig. 3 D, F), which was significantly reduced by GNX treatment ( Fig. 3 E, F). Given that IFN-γ acts through the JAK-1/STAT-1 pathway, the expression ofSTAT-1was examined in IFN-γ-stimulated MDMs and THP-1 cells, revealing that GNX treatment suppressedSTAT-1transcript levels in both cell types ( Fig. 3 G). In addition, JAK-1 and STAT-1 immunoreactivity was increased in IFN-γ-stimulated THP-1 cells but GNX treatment suppressed JAK-1 and STAT-1 immunoreactivity, compared to DMSO-treated controls ( Fig. 3 H). Thus, GNX treatment reduced HLA-DRα and GAT-2 expression together with diminished JAK-1 and STAT-1 abundance in human macrophage lineage cells stimulated with IFN-γ.


Fig. 3 GAT-2 expression in activated human macrophage lineage cells is suppressed by ganaxolone. (A, B) Treatment of IFN-γ stimulated MDMs with GNX showed reductions inHLA-DRαandGAT-2transcripts. (C) Flow cytometric analysis of HLA-DRα+MDMs expressing GAT-2+showed reduced populations with GNX-pretreatment (1 h) followed by IFN-γ stimulation (500U/ml, 24 h) compared to DMSO-pretreated controls (n = 3, repeated twice). (D, E) Immunofluorescence of GAT-2 (red) and HLA-DRα (green) co-localization of MDMs with DMSO or GNX-pretreatment (100 μM, 1hr) that were subsequently stimulated with IFN-γ (500 U/ml, 24 h) (n = 3, repeated twice). (F) Images were quantified (mean pixel intensity) for median GAT-2, MHC Class II, and Merge immunofluorescence. (G) HumanSTAT-1expression in MDM and THP-1 cells showed a marked reduction with GNX treatment (n = 3, repeated thrice). (H) Western blotting of THP-1 cultures pretreated with GNX (100 μM, 1 h) or DMSO (equal volume) and IFN-γ (500 U/ml, 24 h) probed with anti-JAK-1, anti-STAT-1 and anti-β-Actin antibodies. GNX-treated MDM displayed a reduced expression of STAT-1 (86/91: c-terminus) and JAK-1 compared to DMSO-treated controls (n = 3, repeated 3×). Data presented as mean ± SEM (Student’s unpairedt-test,p < 0.05). Scale bar = 10 μm. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

GNX treatment prevents EAE-associated neurobehavioral deficits

To extend the above findings to anin vivomodel, EAE was induced in C57BL6 mice with MOG3555and treated daily by intraperitoneal (i.p.) injections with a vehicle control (DMSO) or GNX. Controls included PBS with complete Freund’s adjuvant and without MOG3555, and with daily GNX or DMSO treatments. Among the EAE groups, neurological deficits were evident by day 11 post-induction and peaked by day 17 post-induction ( Fig. 4 A). Treatment of EAE mice with GNX at low (15 mg/kg) or high (50 mg/kg) doses reduced the EAE-associated neurological deficits compared to the DMSO-treated EAE group, although the therapeutic benefit was greater with the higher dose of GNX ( Fig. 4 A). Animals receiving either GNX concentration also showed significant reductions in the maximum severity score. Again, the therapeutic effect was greatest in the high-dose-treated group ( Fig. 4 B). Thus, GNX exerted beneficial effects on EAE severity in a concentration-dependent manner.


Fig. 4 In vivoGNX treatment reduces EAE-associated neurobehavioral deficits. (A) GNX treatment at both 15 mg/kg and 50 mg/kg reduced behavioral scores from day 17 post-induction onward over time. (B) Similarly, GNX also reduced maximum disease severity in EAE animals at both doses. Data presented as median (Kruskal–Wallis post hoc analysis,p < 0.05).

Reduced inflammatory gene expression in the spinal cord during EAE with GNX treatment

The corresponding molecular changes were assessed in lumbar spinal cords from control and EAE animals receiving GNX treatment revealing thatMhc-II( Fig. 5 A),Gat-2( Fig. 5 B),Cd3ε( Fig. 5 C) andXbp-1/s( Fig. 5 D) andF4/80( Fig. 5 E) transcript expression was induced in EAE animals compared to healthy control animals but these RNA levels were significantly decreased in EAE/GNX-treated animals at day 16 post-induction of EAE (peak severity). Conversely,Gfapwas minimally activated in EAE animals at day 16 with negligible effects mediated by GNX treatment ( Fig. 5 F). Indeed, these GNX-mediated effects persisted for all genes at day 30 post-induction ( Fig. 6 A–D).


Fig. 5 GNX reduces inflammatory gene andGat-2transcript levels in EAE spinal cords. (A–E)MhcII I-Eα,Gat-2,Cd3ε,Xbp-1/s and F4/80transcript levels were increased in the lumbar spinal cords from DMSO-treated EAE animals, GNX suppressed the relative expression of all genes from day 16 isolated lumbar spinal cords (n = 4–6). (F) GNX treatment did not affectGfapexpression in EAE animals. (G–I) GNX-treated EAE animals exhibited lower behavioral scores that positively correlated with reduced RFC transcript expression levels (n = 4–6). Data presented as mean ± SEM (Bonferroni post hoc analysis,p < 0.05 for RT-PCR analysis) and regression analysis (rhovalue calculation,p < 0.05) for correlation analysis.


Fig. 6 GNX treatment in EAE animals reduced transcript abundance ofMhc II,Gat-2andF4/80at day 30 post induction of disease. (A–C)Mhc II,Gat-2 and F4/80transcripts were induced in EAE/DMSO-treated animals lumbar spinal cords, with a reduction in activation observed in EAE/GNX-treated at day 30 (n = 6). (D) Transcript level ofCd3εobserved in spinal cords of EAE mice showed an induction in EAE/DMSO animals with minimal to no significant difference observed in EAE/GNX animals at day 30 post induction (n = 6). Data presented as mean ± SEM, (Bonferroni post hoc analysis,p < 0.05).

Neurobehavioral deficits also showed significant (positive) correlations with several host transcript levels includingMhc-II( Fig. 5 G),Gat-2( Fig. 5 H),Cd3( Fig. 5 I) when comparing their severity scores with the expression levels at day 16 post-EAE induction. These data indicated thatGat-2was increased in EAE in conjunction with several immune genes that were correlated with neurobehavioral deficits, recapitulating immunomolecular features of MS. However, these effects were ameliorated in EAE by GNX treatment.

Since GNX is assumed to function by way of binding to the GABA-A-R, the expression levels of the principalGABA-A-Rsubunitsα1andβ1were examined, showing a decrease in the expression of theGABA-A-R α1andβ1subunits in EAE mice at day 16 after induction while GNX exerted no effects on the expression level of these subunits at either day 16 or day 30 ( Fig. 7 A, B). BothIl-1β( Fig. 7 C) andIl-6( Fig. 7 E) exhibited increased expression in EAE animals’ spinal cords at day 16, but surprisingly GNX treatment enhanced the activation of both genes at this time point, although GNX had no apparent effects at day 30 ( Fig. 7 D, F).


Fig. 7 Spinal cords isolated from EAE/DMSO animals show reducedGaba-A-Rexpression with pro-inflammatory activation. (A, B) TheGaba-A-Rsubunitsα1andβ1were decreased at peak severity in EAE/DMSO-treated animal spinal cords, whereby GNX treatment did not affect the expression level of these subunits. No difference was noted at day 30 post-induction in either of theGaba-A-Rsubunit expression levels (n = 4–6) (C, E) Pro-inflammatory gene expression,Il-β and Il-6are increased in EAE/DMSO-treated animals showing an induced expression of both of these cytokines at peak severity (day 16) in EAE/GNX-treated animals (n = 4–6). (D, F). There was no difference noted with bothIl-β and Il-6at day 30 post-induction (n = 4–6). Data presented as mean ± SEM (Bonferroni post hoc analysis,p < 0.05).

T-bet+ T cells are suppressed by GNX treatment in EAE

As GNX treatment suppressed the inflammatory environment in spinal cords of EAE mice, it was imperative to define the impact of GNX on lymphocyte activation outside of CNS in EAE animals. EAE/GNX-treated mice showed a reduction in the T-bet+CD4+population (3.3%) derived from splenocytes isolated at day 16, compared to the EAE/DMSO-treated group (24.3%) ( Fig. 8 A). Furthermore, CTL/GNX-treated animals showed a reduced abundance of the T-bet+CD4+population (1.60%) compared to the CTL/DMSO-treated animals (5.96%) ( Fig. 8 A). Since total CD3ε transcripts were reduced in the CNS of GNX-treated EAE animals ( Fig. 5 C), we also examined total T-bet/GAPDH transcript abundance in the CNS, observing a reduction of T-bet transcripts at day 16, but not at day 30 of EAE/GNX-treated animals ( Fig. 8 B). Since GNX appeared to reduce T-bet+CD4+populations in the periphery ( Fig. 8 A), we next determined if GNX treatment reduced T cell proliferation in response to an antigen-specific challenge (MOG). Splenocytes isolated at day 30 post-induction that were stimulated with the MOG3555peptide showed similar CFSElowpercentages of CD4+T cell populations from both EAE-induced mice treated with DMSO or GNX ( Fig. 8 C). Moreover, both GATA3+CD4+(Th2) and Tbet+CD4+(Th1) subsets exposed to MOG3555showed CFSElowproliferation profiles that were similar in both EAE groups ( Fig. 8 D, E) suggesting GNX did not affect MOG-specific clonal expansion of Th1- or Th2-polarized cells. Thus, GNX’s effects appeared limited to T-bet expression (transcript and protein) at peak severity, which regulates transcription of pro-inflammatory genes, such as IFN-γ, as well as, potentially reducing total Th1 cell numbers during peak disease, both within the CNS and in the periphery.


Fig. 8 Spleens and lumbar spinal cords isolated from GNX-treated EAE animals showed reduced Th1 (T-bet+) lymphocyte expression. (A) T-bet+cells within the CD4+(T-bet+/CD4+) population (%) of EAE mice treated with GNX showed a reduction in T-bet+/CD4+cells% compared to DMSO-treated EAE controls (n = 3 per group). (B) Lumbar spinal cords from EAE animals at both D16 and D30 post-EAE induction showed reducedT-bet/gapdhcompared with the DMSO-treated EAE group. (C) Splenocytes isolated at day 30 post induction with or withoutex vivoMOG3555challenge, stained with CFSE and cultured for 5 days showed no reduction in proliferation among EAE animals treated with GNX or DMSO as measured by CD4+populations that were CFSElow(%,n = 3 per group). (D and E) CFSElowcells that were CD4+and expressed T-bet+and GATA3+populations (%) were analyzed and showed no difference in profiles (n = 3 per group). Data presented as mean ± SEM (Student’s unpairedt-test,p < 0.05).

Diminished neuropathology in EAE with GNX treatment

Spinal cords from healthy (CTL/DMSO or CTL/GNX) and EAE mice with DMSO or GNX treatments showed that MHC Class II-immunopositive cells were not detected in the CTL/DMSO- and CTL/GNX-treated animals’ spinal cords while EAE/DMSO animals exhibited abundant Class II immune-labeled cells within the parenchyma and perivascular regions, compared to EAE/GNX animals that showed substantially less MHC Class II immunoreactivity ( Fig. 9 A). Total STAT-1 was evident in EAE/DMSO-treated animals’ spinal cords but displayed diminished in expression in EAE/GNX-treated animals ( Fig. 9 A, insets (i and ii)). GAT-2 immunoreactivity was apparent in perivascular cells among CTL/DMSO and CTL/GNX animals’ spinal cords although GAT-2 immunoreactivity was markedly enhanced in the spinal cords of the EAE/DMSO animals in contrast to the EAE/GNX-treated animals (arrows), which displayed similar GAT-2 immunoreactivity to that observed in the control groups ( Fig. 9 B). GAT-2 (brown) and MHC Class II (blue) co-localized immunoreactivity was present in cells resembling macrophages ( Fig. 9 B, insets (i and ii)) in both EAE/DMSO and EAE/GNX spinal cords. Bielchowsky (silver) staining to evaluate axonal viability revealed that CTL/DMSO and CTL/GNX animals showed abundant axonal staining within the spinal cords ( Fig. 9 C). Among EAE/DMSO mice, there was a striking paucity of axons while EAE/GNX-treated animals’ spinal cords showed numerous axons ( Fig. 9 C). CD3+cells were not evident in CTL/DMSO and CTL/GNX animals but numerous CD3+cells were detected in the dorsal column white matter of EAE/DMSO animals with reduced CD3+lymphocytes observed in EAE/GNX-treated animals ( Fig. 9 D). Similarly, MBP immunoreactivity was readily observed in CTL/DMSO and CTL/GNX animals’ spinal cords but EAE/DMSO-treated animals showed reduced MBP immunoreactivity in spinal cord white matter (dorsal column, arrow), which was comparatively preserved in EAE/GNX-treated animals ( Fig. 9 E). Astrocyte and microglia activation was assessed by GFAP and Iba-1 immunoreactivity, respectively, showing increased immunolabeling of both cell types in the EAE/DMSO-treated animals’ spinal cords, which was minimally diminished with GNX treatment ( Fig. 9 F, G, respectively). Nonetheless, these morphological findings were congruent with the molecular and neurobehavioral data above, in that among EAE animals exhibiting neurological deficits there was increased GAT-2 expression with concurrent inflammatory changes, which were reduced by GNX treatment.


Fig. 9 Spinal cords from EAE animals treated with GNX display reduced neuroinflammation with preserved myelin and axons. (A) MHC Class II immunoreactivity was minimally detected in Control (CTL)-treated cervical spinal cords with DMSO or GNX but EAE animals showed increased MHC Class II immunoreactivity that was suppressed by GNX treatment. (A (i) and (ii) insets) STAT-1 immunoreactivity was increased in EAE/DMSO-treated animals showing reduced expression in EAE/GNX-treated animals (B) GAT-2 immunoreactivity was evident in CTL animals but was increased in EAE animals, especially in perivascular regions while GNX reduced its expression. (B (i) and (ii)) MHC Class II (blue) co-localization with GAT-2 (brown) showing reduced immunoreactivity of both markers in EAE/GNX-treated animals compared to EAE/DMSO animals (C) Silver staining showed numerous axons in CTL animals but fewer axons in the EAE/DMSO compared with the EAE/GNX animals. (D) CD3ε immunoreactivity was rarely detected in CTL groups, whereas there were numerous CD3ε-immunopositive T cells observed in the EAE/DMSO animals, although CD3ε immunodetection was reduced by GNX treatment. (E) MBP immunoreactivity was evident in the CTLs but markedly reduced in the EAE with some preservation in the GNX-treated EAE cervical spinal cords. (F) GFAP immunoreactivity was observed in EAE/DMSO-treated animals at day 16 post-induction, showing no difference in immunoreactivity in EAE/GNX-treated animals, while CTL/DMSO or CTL/EAE animals displayed minimal GFAP activation in both animals groups. (G) Iba-1 immunoreactivity was observed in EAE/DMSO-treated animals at day 16 post-induction, showing no difference in immunoreactivity in EAE/GNX-treated animals. CTL/DMSO or CTL/EAE animals displayed minimal Iba-1 activation in both animals groups (n = 3 per group) (Original magnification 400×; insets 1000×). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)


The present study represents the first report of increased GAT-2 levels in MS white matter, which were chiefly localized in activated brain macrophage lineage cells expressing MHC Class II, in demyelinating lesions. Complementing these observations, exposure of human macrophages to IFNγ, induced GAT-2 expression with a concurrent reduction in extracellular GABA levels. Supplementing extracellular GABA levels suppressed the expression of GAT-2 and several immune genes in IFNγ-stimulated macrophage lineage cells. Likewise, the positive allosteric modulator of the GABA-A-R, GNX, suppressed GAT-2 expression in IFNγ-activated macrophage lineage cells in conjunction with repressing both the JAK-1/STAT-1 signaling pathway and downstream pro-inflammatory gene expression that was apparent in bothex vivoandin vivomodels. These findings highlight GABA’s ability to regulate GAT-2 levels as well as immune activation, which was also reflected by GNX’s capacity to inhibit inflammation, likely acting via intracellular steroid receptors and perhaps the GABA-A-R. Thus, the previously recognized immune effects of GABA are affected by GAT-2 abundance, which is regulated by a synthetic neuroactive steroid with downstream effects on pro-inflammatory mechanisms implicated in MS.

GABA is in constant efflux/influx equilibrium that is regulated by action potentials, electrochemical concentration gradients, and its re-uptake transporters (GATs) (Ruiz-Tachiquin et al, 2002, Cordeiro et al, 2003, and Jorgensen, 2005). In the present study, we focused on the mechanism involving the GABA re-uptake transporter, GAT-2, because its expression was increased in white matter from persons with MS and furthermore, GAT-2 was localized in HLA-DRα+cell populations that were confined to regions in which active demyelination was observed. GAT-2 is a transmembrane spanning receptor that contains an intracellular carboxy-terminus containing two putative phosphorylation sites (Goncalves et al, 1999 and Soudijn and van Wijngaarden, 2000). Possible adaptor protein interaction(s) with GAT-2 might mediate GAT-2’s signaling pathways and downstream transcriptional involvement with genes involved in neuroinflammation.

GAT-2 is the least abundant GAT in the human nervous system despite being constitutively expressed on neurons and astrocytes (Conti et al, 1999 and Christiansen et al, 2007). It has also been immunodetected in meninges and endothelial cells of the CNS, as well as, outside the CNS in immune cells (Conti et al, 2004 and Bhat et al, 2010). It is noteworthy that, although not statistically significant, IFN-γ stimulation induced an increase in GAT-2 expression within astrocytes; thus astrocytes, the chief cell type within the CNS, might also participate in the regulation of GABA-mediated regulation of inflammation. To extend our findings of GAT-2 induction in MS brain tissue, GAT-2 was induced in IFN-γ stimulated macrophages with ensuing high levels of MHC Class II expression; indeed, the increase of GAT-2 in MDMs was associated with decreased GABA levels in supernatant from these cells. Thus, a loss or reduction of extracellular GABA through GAT-2 induction might augment inflammation within the brain at sites of active demyelination. To confirm the importance of GABA in inhibiting active inflammation we treated macrophage lineage cells with soluble GABA at physiological concentrations ( Ferkany et al., 1978 ) while also stimulating the same cells with IFNγ. Soluble GABA was able to reduce both MHC Class II and GAT-2 expression, supporting our hypothesis that the regulation of GABAergic signaling is important in controlling inflammation.

Assuming GABA homeostasis is dysregulated in MS, GABA replacement would maintain homeostasis and restrict inflammation. However, soluble GABA cannot cross the blood–brain barrier (BBB) but GNX is BBB-permeable. To investigate GNX’s impact on inflammation, we pre-treated IFN-γ-activated MDMs with GNX and found that indeed both GAT-2 and MHC Class II subunits are suppressed at both the transcript and the protein levels. To determine the mechanism of action by which GNX exerted its effects, human macrophage lineage cells were pre-treated with GNX or DMSO with subsequent IFN-γ stimulation, revealing a reduction in total JAK-1/STAT-1 abundance; thus GNX pre-treatment suppressed IFN-γ stimulation of macrophage lineage cells. STAT-1 can activate transcription of the Class II Transactivator (CIITA), which regulates MHC Class II subunit expression (Linhoff et al, 2001, Nickerson et al, 2001, and Barbaro Ade et al, 2002). Not surprisingly, diminished STAT-1 was observed in IFN-γ-stimulated THP-1 cells that were treated with GNX accompanied by reduced (downstream) MHC Class II expression. Furthermore, thein vivoEAE studies showed that both STAT-1 protein and downstream MHC Class II (I-Eα) subunit expression were both suppressed with GNX treatment, suggesting an overall suppression of the JAK-1/STAT-1 signaling pathway. The outcome of suppressing MHC Class II+cells would result in reduced peptide presentation and lymphocyte engagement, thereby preventing the formation of MOG-specific, auto-inflammatory helper T cells, which play integral roles in MS-defined pathogenesis.

The JAK-1/STAT-1 pathways are not limited to macrophages and thus, other possible targets for GNX-suppressive effects are the circulating MOG-specific T helper cells of the T-bet lineage (Th1), as STAT-1 is involved in the transcriptional regulation of T-bet ( Oestreich and Weinmann, 2012 ). The Th1 (T-bet+CD3+) population in EAE has been directly correlated with MS disease severity ( Yeh et al., 2011 ). GNX treatment showed a marked reduction in T-bet+CD4+cells in spleens isolated from GNX-treated EAE animals, compared to EAE/DMSO controls. Furthermore, animals not immunized with the MOG3535peptide, which did not develop disease (CTL/DMSO or CTL/GNX), also showed a reduced percentage of the T-bet+CD4+population with GNX treatment compared to DMSO-treated CTL animals. Furthermore, the associated Th1 (T-bet+) population was increased in EAE/DMSO lumbar spinal cords, with a reduction of this marker in EAE/GNX-treated animals at day 16. These results confirm that the systemic T-bet+Th1 immune response is dampened with GNX treatment in the presence or absence of the auto-antigen in both the CNS and the periphery at peak disease.

To investigate the antigen-specific, adaptive immune response to MOG3555in GNX-treated EAE animals, anex vivoproliferation assay was performed using CFSE as an indicator of cell proliferation. Since GNX treatment resulted in reduced T-bet+Th1 pro-inflammatory lymphocyte profiles at day 16, it was important to determine if GNX was functionally interfering with the clonal expansion of Th1 cells.Ex vivoproliferation in response to the cognate (MOG3555) antigen confirmed that antigen-specific responses to MOG3555displayed no differences between the total CD4+population in spleens of mice treated with DMSO or GNX isolated at day 30. These latter observations suggested that lymphocyte expansion in response to the MOG3555antigen was similar in both EAE/DMSO and EAE/GNX-treated animals. Further investigation within the individual lymphocyte profiles revealed no differences between CFSElowproliferating T-bet+CD4+(Th1) and GATA3+CD4+(Th2) populations compared to DMSO-treated EAE animals, indicating that there was no difference in the adaptive immune profiles expressed between EAE/GNX-treated animals and EAE/DMSO animals. Daily GNX treatment delayed the progression of Th1 polarization in EAE animals at day 16, measured by reduced T-bet+CD4+lymphocyte profiles but this effect was not apparent at day 30 post-induction, measured by theex vivoproliferation assay. Despite similar responsiveness to the MOG3555peptide, measured byex vivocell division, EAE/GNX-treated animals at day 30 were generally healthier than EAE/DMSO animals. A potential explanation for this finding might be that GNX was able to directly repress the transcription of T-bet in CD4+cells in doing so, reduced the proportion of Th1 cells detected in the CNS of EAE/GNX-treated animals compared to EAE/DMSO-treated animals. Nonetheless, GNX-treatment conferred protection in mice with EAE as it limited the proportion of T-bet+cells expressed in the CNS at peak severity. Of relevance, the roles of other T-bet expressing cell types such as, CD8+ T lymphocytes, were not studied but might also be affected by GNX.

Like soluble GABA-treated macrophage lineage cells, GNX-treatmentin vivoalso exerted direct effects on immune gene transcription. For example, the biomarker for endoplasmic reticulum stress,XBP-1/s, which is increased in MS brains and associated with inflammation, was repressed. These findings highlight the capacity for GNX to act through several mechanisms including, positive allosteric modulation of the GABA-A-R and selective transcriptional repression. Moreover, GNX’s natural analog, allopregnanolone, has been linked to promoting myelination as it is one of the principal neuroactive steroids secreted from oligodendrocytes in the CNS and Schwann cells in the PNS (Robel and Baulieu, 1994, Rupprecht, 1997, and Belelli and Lambert, 2005). This raises other possible therapeutic targets for GNX, as disease severity in EAE/GNX-treated animals was abrogated with preservation of myelin and axonal integrity, confirmed by immunohistological analyses of spinal cord sections by silver staining and MBP immunoreactivity. These results suggest that GNX treatment prevented demyelination and given that demyelination is the pivotal event leading to the clinical manifestations of MS, GNX treatment might represent a valuable therapeutic approach to prevention and possible reversal of this neuroinflammatory-driven disease process.


The authors thank Dr. Kenneth Warren for helpful discussions and Gail Rauw for preparation and analyses of all HPLC samples and the University of Alberta Animal Services for excellent care of laboratory animals. JGW holds an Alberta Innovates-Health Solutions (AI-HS) Fellowship. CP holds a Canada Research Chair (CRC) (Tier 1) in Neurological Infection and Immunity. These studies were supported by the Multiple Sclerosis Society of Canada (CP). None of the authors have commercial interests or activities related to the contents of the present manuscript.


  • Acharjee et al., 2011 S. Acharjee, Y. Zhu, F. Maingat, C. Pardo, K. Ballanyi, M.D. Hollenberg, C. Power. Proteinase-activated receptor-1 mediates dorsal root ganglion neuronal degeneration in HIV/AIDS. Brain. 2011;134:3209-3221 Crossref
  • Antony et al., 2007 J.M. Antony, K.K. Ellestad, R. Hammond, K. Imaizumi, F. Mallet, K.G. Warren, C. Power. The human endogenous retrovirus envelope glycoprotein, syncytin-1, regulates neuroinflammation and its receptor expression in multiple sclerosis: a role for endoplasmic reticulum chaperones in astrocytes. J Immunol. 2007;179:1210-1224 Crossref
  • Barbaro Ade et al., 2002 L. Barbaro Ade, G. Tosi, G. Frumento, E. Bruschi, A. D’Agostino, M.T. Valle, F. Manca, R.S. Accolla. Block of Stat-1 activation in macrophages phagocytosing bacteria causes reduced transcription of CIITA and consequent impaired antigen presentation. Eur J Immunol. 2002;32:1309-1318
  • Belelli and Lambert, 2005 D. Belelli, J.J. Lambert. Neurosteroids: endogenous regulators of the GABA(A) receptor. Nat Rev Neurosci. 2005;6:565-575 Crossref
  • Bhat et al., 2010 R. Bhat, R. Axtell, A. Mitra, M. Miranda, C. Lock, R.W. Tsien, L. Steinman. Inhibitory role for GABA in autoimmune inflammation. Proc Natl Acad Sci U S A. 2010;107:2580-2585 Crossref
  • Carmans et al., 2013 S. Carmans, J.J. Hendriks, H. Slaets, K. Thewissen, P. Stinissen, J.M. Rigo, N. Hellings. Systemic treatment with the inhibitory neurotransmitter gamma-aminobutyric acid aggravates experimental autoimmune encephalomyelitis by affecting proinflammatory immune responses. J Neuroimmunol. 2013;255:45-53 Crossref
  • Carter et al., 1997 R.B. Carter, P.L. Wood, S. Wieland, J.E. Hawkinson, D. Belelli, J.J. Lambert, H.S. White, H.H. Wolf, S. Mirsadeghi, S.H. Tahir, M.B. Bolger, N.C. Lan, K.W. Gee. Characterization of the anticonvulsant properties of ganaxolone (CCD 1042; 3alpha-hydroxy-3beta-methyl-5alpha-pregnan-20-one), a selective, high-affinity, steroid modulator of the gamma-aminobutyric acid(A) receptor. J Pharmacol Exp Ther. 1997;280:1284-1295
  • Christiansen et al., 2007 B. Christiansen, A.K. Meinild, A.A. Jensen, H. Brauner-Osborne. Cloning and characterization of a functional human gamma-aminobutyric acid (GABA) transporter, human GAT-2. J Biol Chem. 2007;282:19331-19341 Crossref
  • Conti et al., 1999 F. Conti, L.V. Zuccarello, P. Barbaresi, A. Minelli, N.C. Brecha, M. Melone. Neuronal, glial, and epithelial localization of gamma-aminobutyric acid transporter 2, a high-affinity gamma-aminobutyric acid plasma membrane transporter, in the cerebral cortex and neighboring structures. J Comp Neurol. 1999;409:482-494 Crossref
  • Conti et al., 2004 F. Conti, A. Minelli, M. Melone. GABA transporters in the mammalian cerebral cortex: localization, development and pathological implications. Brain Res Brain Res Rev. 2004;45:196-212 Crossref
  • Cordeiro et al., 2003 J.M. Cordeiro, V.S. Silva, C.R. Oliveira, P.P. Goncalves. Aluminium-induced impairment of Ca2+ modulatory action on GABA transport in brain cortex nerve terminals. J Inorg Biochem. 2003;97:132-142 Crossref
  • Demakova et al., 2003 E.V. Demakova, V.P. Korobov, L.M. Lemkina. Determination of gamma-aminobutyric acid concentration and activity of glutamate decarboxylase in blood serum of patients with multiple sclerosis. Klin Lab Diagn. 2003;:15-17
  • Deslauriers et al., 2011 A.M. Deslauriers, A. Afkhami-Goli, A.M. Paul, R.K. Bhat, S. Acharjee, K.K. Ellestad, F. Noorbakhsh, M. Michalak, C. Power. Neuroinflammation and endoplasmic reticulum stress are coregulated by crocin to prevent demyelination and neurodegeneration. J Immunol. 2011;187:4788-4799 Crossref
  • Dionisio et al., 2011 L. Dionisio, M. Jose De Rosa, C. Bouzat, C. Esandi Mdel. An intrinsic GABAergic system in human lymphocytes. Neuropharmacology. 2011;60:513-519 Crossref
  • Eikelenboom et al., 2002 P. Eikelenboom, C. Bate, W.A. Van Gool, J.J. Hoozemans, J.M. Rozemuller, R. Veerhuis, A. Williams. Neuroinflammation in Alzheimer’s disease and prion disease. Glia. 2002;40:232-239 Crossref
  • Ellestad et al., 2009 K.K. Ellestad, S. Tsutsui, F. Noorbakhsh, K.G. Warren, V.W. Yong, Q.J. Pittman, C. Power. Early life exposure to lipopolysaccharide suppresses experimental autoimmune encephalomyelitis by promoting tolerogenic dendritic cells and regulatory T cells. J Immunol. 2009;183:298-309 Crossref
  • Falkenstein et al., 2000 E. Falkenstein, H.C. Tillmann, M. Christ, M. Feuring, M. Wehling. Multiple actions of steroid hormones – a focus on rapid, nongenomic effects. Pharmacol Rev. 2000;52:513-556
  • Ferkany et al., 1978 J.W. Ferkany, L.A. Smith, W.E. Seifert, R.M. Caprioli, S.J. Enna. Measurement of gamma-aminobutyric acid (GABA) in blood. Life Sci. 1978;22:2121-2128 Crossref
  • Frohman et al., 2006 E.M. Frohman, M.K. Racke, C.S. Raine. Multiple sclerosis – the plaque and its pathogenesis. N Engl J Med. 2006;354:942-955 Crossref
  • Goncalves et al., 1999 P.P. Goncalves, S.M. Meireles, M.G. Vale. Regulation of the gamma-aminobutyric acid transporter activity by protein phosphatases in synaptic plasma membranes. Neurosci Res. 1999;33:41-47 Crossref
  • Grant et al., 2006 S.L. Grant, Y. Shulman, P. Tibbo, D.R. Hampson, G.B. Baker. Determination of d-serine and related neuroactive amino acids in human plasma by high-performance liquid chromatography with fluorimetric detection. J Chromatogr B Analyt Technol Biomed Life Sci. 2006;844:278-282 Crossref
  • Hosie et al., 2006 A.M. Hosie, M.E. Wilkins, H.M. da Silva, T.G. Smart. Endogenous neurosteroids regulate GABAA receptors through two discrete transmembrane sites. Nature. 2006;444:486-489 Crossref
  • Houtchens, 2007 M.K. Houtchens. Pregnancy and multiple sclerosis. Semin Neurol. 2007;27:434-441 Crossref
  • Jorgensen, 2005 Jorgensen EM (2005), Gaba. WormBook, pp 1–13.
  • LeWitt et al., 2011 P.A. LeWitt, A.R. Rezai, M.A. Leehey, S.G. Ojemann, A.W. Flaherty, E.N. Eskandar, S.K. Kostyk, K. Thomas, A. Sarkar, M.S. Siddiqui, S.B. Tatter, J.M. Schwalb, K.L. Poston, J.M. Henderson, R.M. Kurlan, I.H. Richard, L. Van Meter, C.V. Sapan, M.J. During, M.G. Kaplitt, A. Feigin. AAV2-GAD gene therapy for advanced Parkinson’s disease: a double-blind, sham-surgery controlled, randomised trial. Lancet Neurol. 2011;10:309-319 Crossref
  • Linhoff et al., 2001 M.W. Linhoff, J.A. Harton, D.E. Cressman, B.K. Martin, J.P. Ting. Two distinct domains within CIITA mediate self-association: involvement of the GTP-binding and leucine-rich repeat domains. Mol Cell Biol. 2001;21:3001-3011 Crossref
  • Liptakova et al., 2000 S. Liptakova, L. Velisek, J. Veliskova, S.L. Moshe. Effect of ganaxolone on flurothyl seizures in developing rats. Epilepsia. 2000;41:788-793 Crossref
  • Maingat et al., 2009 F. Maingat, P. Vivithanaporn, Y. Zhu, A. Taylor, G. Baker, K. Pearson, C. Power. Neurobehavioral performance in feline immunodeficiency virus infection: integrated analysis of viral burden, neuroinflammation, and neuronal injury in cortex. J Neurosci. 2009;29:8429-8437 Crossref
  • Mellon et al., 2008 S.H. Mellon, W. Gong, M.D. Schonemann. Endogenous and synthetic neurosteroids in treatment of Niemann-Pick Type C disease. Brain Res Rev. 2008;57:410-420 Crossref
  • Nickerson et al., 2001 K. Nickerson, T.J. Sisk, N. Inohara, C.S. Yee, J. Kennell, M.C. Cho, P.J. Yannie 2nd, G. Nunez, C.H. Chang. Dendritic cell-specific MHC class II transactivator contains a caspase recruitment domain that confers potent transactivation activity. J Biol Chem. 2001;276:19089-19093 Crossref
  • Nigam et al., 2010 R. Nigam, H. El-Nour, B. Amatya, K. Nordlind. GABA and GABA(A) receptor expression on immune cells in psoriasis: a pathophysiological role. Arch Dermatol Res. 2010;302:507-515 Crossref
  • Noorbakhsh et al., 2006 F. Noorbakhsh, S. Tsutsui, N. Vergnolle, L.A. Boven, N. Shariat, M. Vodjgani, K.G. Warren, P. Andrade-Gordon, M.D. Hollenberg, C. Power. Proteinase-activated receptor 2 modulates neuroinflammation in experimental autoimmune encephalomyelitis and multiple sclerosis. J Exp Med. 2006;203:425-435 Crossref
  • Noorbakhsh et al., 2011 F. Noorbakhsh, K.K. Ellestad, F. Maingat, K.G. Warren, M.H. Han, L. Steinman, G.B. Baker, C. Power. Impaired neurosteroid synthesis in multiple sclerosis. Brain. 2011;134:2703-2721
  • Oestreich and Weinmann, 2012 K.J. Oestreich, A.S. Weinmann. Master regulators or lineage-specifying? Changing views on CD4+ T cell transcription factors. Nat Rev Immunol. 2012;12:799-804
  • Power et al., 1998 C. Power, J.C. McArthur, A. Nath, K. Wehrly, M. Mayne, J. Nishio, T. Langelier, R.T. Johnson, B. Chesebro. Neuronal death induced by brain-derived human immunodeficiency virus type 1 envelope genes differs between demented and nondemented AIDS patients. J Virol. 1998;72:9045-9053
  • Ram et al., 2001 K. Ram, G.N. Lam, B. Chien. A high-performance liquid chromatography-tandem mass spectrometric method for the determination of pharmacokinetics of ganaxolone in rat, monkey, dog and human plasma. J Chromatogr B Biomed Sci Appl. 2001;751:49-59
  • Reddy, 2010 D.S. Reddy. Neurosteroids: endogenous role in the human brain and therapeutic potentials. Prog Brain Res. 2010;186:113-137 Crossref
  • Reddy and Rogawski, 2010 D.S. Reddy, M.A. Rogawski. Ganaxolone suppression of behavioral and electrographic seizures in the mouse amygdala kindling model. Epilepsy Res. 2010;89:254-260 Crossref
  • Reddy and Woodward, 2004 D.S. Reddy, R. Woodward. Ganaxolone: a prospective overview. Drugs Future. 2004;29:227-242 Crossref
  • Reyes-Garcia et al., 2007 M.G. Reyes-Garcia, F. Hernandez-Hernandez, B. Hernandez-Tellez, F. Garcia-Tamayo. GABA (A) receptor subunits RNA expression in mice peritoneal macrophages modulate their IL-6/IL-12 production. J Neuroimmunol. 2007;188:64-68 Crossref
  • Robel and Baulieu, 1994 P. Robel, E.E. Baulieu. Neurosteroids biosynthesis and function. Trends Endocrinol Metab. 1994;5:1-8 Crossref
  • Ruiz-Tachiquin et al., 2002 M.E. Ruiz-Tachiquin, E. Sanchez-Lemus, L.E. Soria-Jasso, J.A. Arias-Montano, A. Ortega. Gamma-aminobutyric acid transporter (BGT-1) expressed in human astrocytoma U373 MG cells: pharmacological and molecular characterization and phorbol ester-induced inhibition. J Neurosci Res. 2002;69:125-132 Crossref
  • Rupprecht, 1997 R. Rupprecht. The neuropsychopharmacological potential of neuroactive steroids. J Psychiatr Res. 1997;31:297-314 Crossref
  • Rupprecht et al., 1993 R. Rupprecht, J.M. Reul, T. Trapp, B. van Steensel, C. Wetzel, K. Damm, W. Zieglgansberger, F. Holsboer. Progesterone receptor-mediated effects of neuroactive steroids. Neuron. 1993;11:523-530 Crossref
  • Schneider et al., 2012 C.A. Schneider, W.S. Rasband, K.W. Eliceiri. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9:671-675 Crossref
  • Shiratsuchi et al., 2009 H. Shiratsuchi, Y. Kouatli, G.X. Yu, H.M. Marsh, M.D. Basson. Propofol inhibits pressure-stimulated macrophage phagocytosis via the GABAA receptor and dysregulation of p130cas phosphorylation. Am J Physiol Cell Physiol. 2009;296:C1400-C1410 Crossref
  • Soltani et al., 2011 N. Soltani, H. Qiu, M. Aleksic, Y. Glinka, F. Zhao, R. Liu, Y. Li, N. Zhang, R. Chakrabarti, T. Ng, T. Jin, H. Zhang, W.Y. Lu, Z.P. Feng, G.J. Prud’homme, Q. Wang. GABA exerts protective and regenerative effects on islet beta cells and reverses diabetes. Proc Natl Acad Sci U S A. 2011;108:11692-11697 Crossref
  • Soudijn and van Wijngaarden, 2000 W. Soudijn, I. van Wijngaarden. The GABA transporter and its inhibitors. Curr Med Chem. 2000;7:1063-1079 Crossref
  • Stadelmann, 2011 C. Stadelmann. Multiple sclerosis as a neurodegenerative disease: pathology, mechanisms and therapeutic implications. Curr Opin Neurol. 2011;24:224-229 Crossref
  • Stuckey et al., 2005 D.J. Stuckey, D.C. Anthony, J.P. Lowe, J. Miller, W.M. Palm, P. Styles, V.H. Perry, A.M. Blamire, N.R. Sibson. Detection of the inhibitory neurotransmitter GABA in macrophages by magnetic resonance spectroscopy. J Leukoc Biol. 2005;78:393-400 Crossref
  • Tansey et al., 2008 M.G. Tansey, T.C. Frank-Cannon, M.K. McCoy, J.K. Lee, T.N. Martinez, F.E. McAlpine, K.A. Ruhn, T.A. Tran. Neuroinflammation in Parkinson’s disease: is there sufficient evidence for mechanism-based interventional therapy?. Front Biosci. 2008;13:709-717 Crossref
  • Tian et al., 2011 J. Tian, J. Yong, H. Dang, D.L. Kaufman. Oral GABA treatment downregulates inflammatory responses in a mouse model of rheumatoid arthritis. Autoimmunity. 2011;44:465-470 Crossref
  • Tsutsui et al., 2004 S. Tsutsui, J. Schnermann, F. Noorbakhsh, S. Henry, V.W. Yong, B.W. Winston, K. Warren, C. Power. A1 adenosine receptor upregulation and activation attenuates neuroinflammation and demyelination in a model of multiple sclerosis. J Neurosci. 2004;24:1521-1529 Crossref
  • Vivithanaporn et al., 2010 P. Vivithanaporn, F. Maingat, L.T. Lin, H. Na, C.D. Richardson, B. Agrawal, E.A. Cohen, J.H. Jhamandas, C. Power. Hepatitis C virus core protein induces neuroimmune activation and potentiates Human Immunodeficiency Virus-1 neurotoxicity. PLoS One. 2010;5:e12856 Crossref
  • Yeh et al., 2011 W.I. Yeh, I.L. McWilliams, L.E. Harrington. Autoreactive Tbet-positive CD4 T cells develop independent of classic Th1 cytokine signaling during experimental autoimmune encephalomyelitis. J Immunol. 2011;187:4998-5006 Crossref
  • Zhu et al., 2009 Y. Zhu, D. Vergote, C. Pardo, F. Noorbakhsh, J.C. McArthur, M.D. Hollenberg, C.M. Overall, C. Power. CXCR3 activation by lentivirus infection suppresses neuronal autophagy: neuroprotective effects of antiretroviral therapy. FASEB J. 2009;23:2928-2941 Crossref


a Department of Medicine, University of Alberta, Edmonton, AB, Canada

b Department of Laboratory Medicine & Pathology, University of Alberta, Edmonton, AB, Canada

c Department of Psychiatry, University of Alberta, Edmonton, AB, Canada

lowast Corresponding author. Address: Division of Neurology, HMRC 6-11, University of Alberta, Edmonton, AB, Canada. Tel: +1-780-407-1938; fax: +1-780-407-1984.