You are here
Metabolic gene expression changes in astrocytes in Multiple Sclerosis cerebral cortex are indicative of immune-mediated signaling
Brain, Behavior, and Immunity (Available online 30 April 2015)
- Astrocytic ANLS and GGC genes are downregulated in MS gray matter tissue.
- MS NAGM gene expression is indicative of active Toll-like/IL-1B signaling.
- IL-1B treatment of astrocytes leads to ANLS and GGC downregulationin vitro.
- Inflammasome activation in astrocytes downregulates ANLS and GGCin vitro.
- Immune system activation leads to ANLS downregulationin vivo.
Emerging as an important correlate of neurological dysfunction in Multiple Sclerosis (MS), extended focal and diffuse gray matter abnormalities have been found and linked to clinical manifestations such as seizures, fatigue and cognitive dysfunction. To investigate possible underlying mechanisms we analyzed the molecular alterations in histopathological normal appearing cortical gray matter (NAGM) in MS. By performing a differential gene expression analysis of NAGM of control and MS cases we identified reduced transcription of astrocyte specific genes involved in the astrocyte–neuron lactate shuttle (ANLS) and the glutamate–glutamine cycle (GGC). Additional quantitative immunohistochemical analysis demonstrating a CX43 loss in MS NAGM confirmed a crucial involvement of astrocytes and emphasizes their importance in MS pathogenesis. Concurrently, a Toll-like/IL-1β signaling expression signature was detected in MS NAGM, indicating that immune-related signaling might be responsible for the downregulation of ANLS and GGC gene expression in MS NAGM. Indeed, challenging astrocytes with immune stimuli such as IL-1β and LPS reduced their ANLS and GGC gene expressionin vitro. The detected upregulation ofIL1Bin MS NAGM suggests inflammasome priming. For this reason, astrocyte cultures were treated with ATP and ATP/LPS as for inflammasome activation. This treatment led to a reduction of ANLS and GGC gene expression in a comparable manner. To investigate potential sources for ANLS and GGC downregulation in MS NAGM, we first performed an adjuvant-driven stimulation of the peripheral immune system in C57Bl/6 micein vivo. This led to similar gene expression changes in spinal cord demonstrating that peripheral immune signals might be one source for astrocytic gene expression changes in the brain.IL1Bupregulation in MS NAGM itself points to a possible endogenous signaling process leading to ANLS and GGC downregulation. This is supported by our findings that, among others, MS NAGM astrocytes express inflammasome components and that astrocytes are capable to release Il-1βin-vitro. Altogether, our data suggests that immune signaling of immune- and/or central nervous system origin drives alterations in astrocytic ANLS and GGC gene regulation in the MS NAGM. Such a mechanism might underlie cortical brain dysfunctions frequently encountered in MS patients.
Abbreviations: ANLS - astrocyte–neuron lactate shuttle, GGC - glutamate–glutamine cycle, NAGM - normal appearing cortical gray matter.
Keywords: Brain energy homeostasis, Gray matter pathology, Inflammasomes, Astrocyte–neuron lactate shuttle, Multiple Sclerosis, Normal appearing gray matter, Cerebral cortex.
Multiple Sclerosis (MS) is a chronic inflammatory demyelinating disease of the central nervous system (CNS). MS predominantly affects young adults and leads to substantial disability in a high proportion of patients ( Compston and Coles, 2002 ). Pathologically, MS is characterized by multiple demyelinated plaques in the white and gray matter. Besides that, diffuse white and gray matter abnormalities in non-lesional normally myelinated areas have been gaining increasing attention (Graumann et al, 2003, Stadelmann et al, 2008, Zeis et al, 2008, and Zeis et al, 2009). In the case of the non-lesional normal appearing gray matter (NAGM) of MS patients little is known about the molecular changes which possibly precede gray matter lesion formation (Dutta et al, 2006 and Dutta et al, 2007) and potentially contribute to clinical features encountered in MS. Of those, gray matter pathology has been linked with physical disability and cognitive impairment ( Pirko et al., 2007 ) as well as fatigue (Roelcke et al, 1997, Filippi et al, 2002, and Niepel et al, 2006), which are common manifestations of MS (Wishart and Sharpe, 1997, Blinkenberg et al, 2000, Lazeron et al, 2000, Amato et al, 2004, Benedict et al, 2004, Morgen et al, 2006, Sanfilipo et al, 2006, and Houtchens et al, 2007).
One of the most abundant cell types of the gray matter is the astrocyte, which fulfills many important tasks ensuring brain function ( Sofroniew and Vinters, 2010 ). A key assignment of astrocytes is the removal of neurotransmitters released by neurons via the glutamate–glutamine cycle (GGC; Schousboe et al., 2014 ). The disturbance thereof is strongly suggested to be involved in the pathogenesis of neurological disorders ( Seifert et al., 2006 ). A reduction of the astroglial excitatory amino acid transporter 2 (EAAT2) was reported in neuromyelitis optica (NMO; Hinson et al., 2008 ), an inflammatory demyelinating disease earlier thought to be a variant of MS ( Brosnan and Raine, 2013 ). Further, in amyotrophic lateral sclerosis (ALS), a decreased glutamate transport ( Rothstein et al., 1992 ) as well as a selective loss of EAAT2 was shown ( Rothstein et al., 1995 ). This suggests that a failure of the GGC by astrocytes might be crucial for excitotoxic damage and the subsequent pathogenic process in ALS ( Seifert et al., 2006 ).
Increasing evidence indicate that astrocytes are also important contributors to CNS metabolism ( Sofroniew and Vinters, 2010 ). In 1994, Pellerin and Magistretti proposed an astrocyte–neuron lactate shuttle (ANLS) based on existing experimental data ( Pellerin and Magistretti, 1994 ). This model includes the following sequence of molecular events: Following increased synaptic activity glutamatergic neurons release the neurotransmitter glutamate into the synaptic cleft. Glutamate is avidly taken up by the astrocytes surrounding the synaptic cleft, via specific glial glutamate transporters (EAAT1 and EAAT2). EAATs co-transport glutamate with sodium ions increasing intracellular sodium concentration in the astrocyte and activating the energy dependent Na+/K+ ATPase pump (specifically the recruitment of the Na+/K+ ATPase alpha 2 subunit, ATP1A2). The corresponding hydrolysis of ATP leads to activation of astrocytic glycolysis, i.e. the degradation of glucose to pyruvate, which is then converted to lactate via lactate dehydrogenase (LDH). Lactate is then released via astrocytic monocarboxylate transporters (MCT1 and 4) into the extracellular space and from there taken up by the neurons (via MCT2 transporter), where it serves as an energy substrate following its intracellular conversion to pyruvate ( Fig. 1 B;Sofroniew and Vinters, 2010 and Pellerin and Magistretti, 2012). This activity dependent supply of energetic metabolites from astrocytes to neurons is further enabled by a gap-junction mediated astrocyte network allowing intercellular trafficking of metabolites throughout the astrocyte syncytium ( Rouach et al., 2008 ). Recently, it was shown that the provision of lactate to neurons is important for long- as well as short-term memory (Newman et al, 2011 and Suzuki et al, 2011). This suggests that astrocyte dysfunction might play an important role in the reported cognitive impairments of MS patients (for review see Ferreira, 2010 ).
Inflammasomes are signaling platforms activated by signs of tissue damage and metabolic dysfunction (e.g. extracellular ATP), which lead to the maturation of cytokines such as IL-1β and thereby drive inflammation ( Schroder and Tschopp, 2010 ). They are composed of an inflammasome sensor molecule (usually a Nod-like Receptor, NLR) and a caspase effector, often coupled via a signaling adaptor (apoptosis-associated speck-like protein containing CARD, ASC). Inflammasomes can be activated by pathogen associated molecular patterns (PAMPs) and/or danger-associated molecular patterns (DAMPs; Schroder and Tschopp, 2010 ). Activation of inflammasomes and secretion of active IL-1β requires two sequential steps; “signal 1” induces the expression of the inactive cytokine precursor (pro-IL-1β), while “signal 2” triggers its caspase-1-dependent processing via inflammasomes ( Schroder and Tschopp, 2010 ). In innate immune cells, “signal 1” is usually provided by pro-inflammatory signaling by TLRs, or by IL-1β itself, whereas “signal 2” might among others be provided by host-derived molecules indicative of injury e.g. extracellular ATP ( Schroder and Tschopp, 2010 ). A recent publication by Kawana et al., 2013 showed that in MS active demyelinating lesions astrocytes express all necessary components of the NLRP3 inflammasome ( Kawana et al., 2013 ). Further, animals lacking Casp1, Nlrp3 or Asc develop a mild clinical course of in experimental autoimmune encephalomyelitis (EAE;Furlan et al, 1999, Gris et al, 2010, and Inoue et al, 2012), suggesting an involvement of inflammasomes in MS pathogenesis.
In our study, we discovered a downregulation of astrocyte specific genes involved in the ANLS and the GGC in the MS NAGM. Additionally, a gene expression signature indicative of activated Toll-like/Il-1β signaling as well as an upregulation ofIL1Bwas detected. To investigate a potential interconnection between Toll-like/Il-1β signaling and the downregulation of astrocyte GGC and ANLS genes we treated pure astrocyte cultures with IL-1β, LPS (a TLR4 agonist), ATP (as a potential endogenous danger/degeneration signal) and with ATP/LPS (inflammasome activation protocol for Il-1β release). These treatments led to a similar downregulation of genes as foundin vivoin the MS NAGM, demonstrating an interconnection between the Toll-like/IL-1β signaling and the astrocytic ANLS as well as GGC genes. We further demonstratedin vivoin C57Bl/6 mice that such an astrocyte reaction in the MS NAGM can be a consequence of peripheral immune-system signaling. On the other hand, a downregulation of GGC and ANLS genes might also be a consequence of an endogenous brain reaction. Supporting this hypothesis, we demonstrated thatIL1Bis upregulated in the MS NAGM. Further, astrocytes were shown to express the necessary inflammasome proteins in the MS NAGM and to be capable of releasing Il-1βin-vitro. Altogether, this indicates that in MS, immune-or danger related signals induce ANLS and GGC changes in astrocytes of the NAGM. As astrocyte dysfunction was shown to play an important role in the reported cognitive impairments of MS patients (for review see Ferreira, 2010 ), ANLS or GGC changes might contribute to the encountered cognitive impairments in MS.
2. Material and methods
2.1. Tissue collection
MS and control brain tissue samples were provided by the UK Multiple Sclerosis Tissue Bank (UK Multicentre Research Ethics Committee, MREC/02/2/39), funded by the Multiple Sclerosis Society of Great Britain and Northern Ireland (registered charity 207495). All brain tissues were routinely screened by a neuropathologist to confirm diagnosis of MS and to exclude other confounding pathologies ( Reynolds et al., 2011 ). Tissues used for this study were immunohistochemically analyzed for MOG (myelin marker), NeuN (neuronal marker) and CD68 (microglial marker). Tissues showing demyelination ( Supplementary Fig. 1 ), neuronal degeneration or strong microglia activation were not included in the study. In total 31 tissue samples from 17 control as well as 31 tissue samples from 18 MS cases were used for the study ( Table 1 ).
|Patient||Gender||Age at death||p.m. time||Disease duration||MS type||Cause of death||IHC||qRT-PCR||Array||In-situ|
|C1||m||69||10||Acute cardiac death||X (1)|
|C2||f||59||22||Acute pancreatitis||X (1)|
|C3||m||89||9||Cancer of the rectum||X (1)||X (1)|
|C5||f||85||9||Cancer of the esophagus||X (1)||X (1)||X (1)|
|C7||m||73||21||Cardiogenix shock||X (1)|
|C8||m||77||26||Carcinoma of the lung||X (1)||X (1)|
|C9||m||64||18||Cardiac failure||X (1)||X (4)||X (1)|
|C10||f||84||24||Congestive cardiac failure, ischemic heart disease||X (1)||X (1)||X (2)|
|C11||m||35||22||Carcinoma of the tongue||X (4)||X (2)||X (2)||X (1)|
|C12||f||78||33||Myeloid leukemia||X (1)|
|C13||f||60||13||Ovarian cancer||X (1)||X (1)||X (2)|
|C14||m||75||17||Aspiration pneumonia||X (4)||X (1)||X (1)||X (1)|
|C15||m||88||22||Prostate cancer, bone metastases||X (1)||X (1)|
|C16||f||88||20||Bronchopneumonia||X (1)||X (1)|
|C17||m||84||5||Bladder cancer, pneumonia||X (2)||X (3)||X (1)||X (1)|
|Multiple Sclerosis patients|
|MS1||f||56||8||31||SPMS||Breast carcinoma, pneumothorax||X (1)||X (1)||X (1)|
|MS2||f||58||16||22||PPMS||Peritonitis, Multiple Sclerosis||X (1)||X (1)|
|MS3||f||78||18||33||SPMS||Myocardial infarction, acute abdomen||X (1)|
|MS4||f||58||6||21||PPMS||Bronchopneumonia||X (3)||X (1)||X (1)||X (1)|
|MS5||f||20||17||17||PPMS||Pulmonary embolus, pneumonia||X (2)||X (1)|
|MS6||f||69||11||31||SPMS||Acute pyelonephritis, Multiple Sclerosis||X (1)||X (1)|
|MS7||f||78||9||47||PPMS||Lung infection||X (2)||X (1)||X (1)|
|MS8||f||78||5||42||SPMS||Metastatic carcinoma of bronchus||X (4)||X (3)||X (3)||X (1)|
|MS9||f||54||22||20||SPMS||Bronchopneumonia||X (2)||X (2)||X (2)|
|MS10||m||53||12||11||SPMS||Urinary tract infection, Multiple Sclerosis||X (1)||X (1)||X (1)|
|MS11||m||38||19||17||RPMS||Aspiration pneumonia, pulmonary edema||X (1)||X (1)||X (1)|
|MS12||f||50||12||SPMS||Multiple Sclerosis||X (1)||X (1)||X (1)|
|MS13||m||37||12||27||PPMS||Intestinal obstruction, Multiple Sclerosis||X (1)||X (2)||X (1)|
|MS14||f||78||24||47||SPMS||Aspiration pneumonia, Multiple Sclerosis||X (2)||X (1)||X (1)||X (1)|
|MS15||f||49||12||27||PPMS||Bronchopneumonia, Multiple Sclerosis||X (1)||X (2)|
|MS16||m||45||24||25||SPMS||Bronchopneumonia, Multiple Sclerosis||X (1)||X (1)|
|MS17||f||47||16||17||SPMS||Septicaemia, bronchopneumonia, Multiple Sclerosis||X (1)||X (2)||X (1)||X (1)|
|MS18||m||45||9||17||SPMS||Multiple sclerosis||X (1)|
2.2. RNA profiling and statistical analysis
Total RNA was isolated using the RNeasy Kit (Qiagen, Hilden, Germany) together with Qiazol as lysis reagent. RNA quality control was done by using the Nanodrop Spectrophotometer (Thermo Scientific, Wilmington, USA) and by RNA gel electrophoresis or by the RNA 6000 Nano total RNA Kit on an Agilent 2100 Bioanalyzer. Degraded (RIN < 6) and/or contaminated (260/280 nm ratio < 1.5; 230/280 nm ratio < 1.5) samples were excluded from the study. Gene expression profiling using our custom microarrays with 1176 representative cDNA sequences was performed as described before (Graumann et al, 2003 and Zeis et al, 2008). Quantification of differential hybridization signal intensities was done with the AtlasImage™ 2.0 software program. Normalization of the gene expression array was performed by using the 10% trimmed mean as suggested for Human 1.2 Arrays from Clontech ( Kroll and Wolfl, 2002 ). Microarray data analysis was performed with Partek Genomics Suite software. Differentially expressed genes were identified by performing a three-way ANOVA (disease, patients, tissue type). For this study, genes with ap-value of <0.05 (t-test on disease) and a mean fold difference of >1.2 were considered to be significantly differentially expressed between MS and controls. Significantly differentially expressed genes were further analyzed with the use of IPA (Ingenuity Systems, www.ingenuity.com ). IPA upstream regulator analysis was performed to detect possible transcriptional regulators (TR) leading to the detected differential gene expression in MS NAGM. An activationz-score was calculated by IPA showing whether a potential TR was in an activated (z-score > 2) or inhibited (z-score < −2) or uncertain state.
2.3. Quantitative RT-PCR and statistical analysis
Real-time RT-PCR was performed in 384-well plates (VWR, Radnor, USA, Cat.No. 82006-678) using the ABI ViiA Fast Real-Time PCR system (Applied Biosystems, Life technologies Ltd., Paisley, UK). Primer sequences ( Supplementary Table 1 ) were designed from unique sites over exon–intron junctions to prevent amplification of genomic DNA. Real-time RT-PCR was performed according to the manufacturer’s protocol using the KAPA SYBR Fast Universal 2x qPCR Master mix (KAPA Biosystems, Woburn, USA). RNA amounts were calculated with relative standard curves for all mRNAs of interest. Optimal normalization factor for each experiment was identified among a set of detectable housekeeping genes by the NormFinder Software ( Andersen et al., 2004 ). Data of the qRT-PCR expression analysis are shown as the mean of relative mRNA levels with error bars representing the standard deviation (SD). For all experiments, statistical significance was determined using the student’st-test,* p< 0.05,**p < 0.0001. Unless otherwise stated, all values were adjusted to the corresponding control group set to 1.
Cryostat sections (10 μm) from fresh frozen tissue blocks or cell culture dishes were fixed for 10 min in 10% formalin. For myelin staining, sections were delipidated after fixation with methanol at −20 °C for 8 min. Tissue sections and dishes were blocked for 2 h at room temperature (RT) in blocking buffer (PBS pH 7.4, 0.2% Triton X-100, 0.1% Tween-20, 5% normal donkey serum) and incubated with primary antibodies in blocking buffer overnight at 4 °C. Tissue sections and dishes were then treated with 0.3% hydrogen peroxide in PBS for 20 min. Sections subjected to immunofluorescence were instead incubated for 1 h in 10 mM CuSO4, 50 mM CH3COONH4, pH 5.0. Next, sections and dishes were incubated with secondary antibodies in blocking buffer for 1 h at RT. Sections for peroxidase stainings were processed with the Vectastain ABC Kit (Vector Laboratories, Burlingame, USA) as described by the manufacturer. Finally, immunofluorescence sections and dishes were counterstained with DAPI and mounted using Fluorosave (Calbiochem, Merck Millipore, Merck KGaA, Darmstadt, Germany). Antibodies used for immunohistochemical and immunofluorescence stainings are listed in Table 2 .
|Antibody||Company||Cat.Nr.||Used for IHC||Dilution||Used for WB||Dilution|
|anti-MOG (Clone Z12)||Kindly provided by Prof. R. Reynolds||–||X||1:200|
|anti-MCT1||Kindly provided by Dr. Merenzhinskaya, Washington DC, USA||–||X||1:500||X||1:1000|
|anti b-Actin (AC-74)||Sigma–Aldrich||A2228||X||1:5000|
|gt-a-rb-IRDye 800||Rockland Immunochemicals, Inc.||611-132-002||X||1:10000|
2.5. Astrocyte cultures
Primary cultures of cerebral cortical astrocytes are prepared from newborn (1- to 2-days-old) OF1 mice (Charles River Laboratories, L’Arbresle, France) as previously described (Allaman et al, 2004 and Gavillet et al, 2008). Culture experiments were approved by the Veterinary Office of the State of Vaud. For IL-1β treatments, culture medium was renewed 24 h before cytokine treatment induced by addition of IL-1β at 0.25 ng/ml (R&D systems, Abington, UK) in the culture medium ( Gavillet et al., 2008 ). Treatment with IL-1β was performed on confluent 21-day-old cultures for 24 h. For ATP/LPS induced inflammasome activation, astrocytes were incubated in serum-free DMEM (D5030) supplemented with 5 mM glucose, 44 mM NaHCO3and 10 mL/L of an 100× antibiotic–antimycotic solution (Invitrogen, Life technologies Ltd., Paisley, UK) (DMEM) 24 h before treatments. Cells were then prestimulated with LPS (10 ng/ml) and 4 h later 5 mM ATP was added for an additional 1 h or 24 h. At the end of the treatment cells were harvested and total RNA was isolated using the RNeasy kit (Qiagen, Hilden, Germany). RNA quality control was done by using the Nanodrop Spectrophotometer (Thermo Scientific, Wilmington, USA) and by the RNA 6000 Nano total RNA Kit on an Agilent 2100 Bioanalyzer. Degraded (RIN < 6) and/or contaminated (260/280 nm ratio < 1.5; 230/280 nm ratio < 1.5) samples were excluded from the study. For protein analysis cells were lysed in 62.5 mM Tris–HCl pH 6.8, 50 mM DTT and 0.3% SDS, sonicated for 10 s, boiled for 5 min at 100 °C and stored at −80 °C.
Using this culture procedure, >99% of the cells were immunoreactive for the astrocytic marker glial fibrillary acidic protein (GFAP), and microglia (CD68) were almost complete absent (<0.01%). No significant cell death upon any treatment was detected by the calcein survival test (data not shown).
2.6. Western blot analysis
Protein samples from the astrocyte cultures were subjected to gel electrophoresis using a 4–12% NUPAGE gradient gel (Invitrogen, Life technologies Ltd., Paisley, UK). After blotting on 0.2 μm Immun-Blot PVDF membranes (BIO-RAD, Hercules, USA), blots were incubated with blocking buffer (3% TOP-BLOCK in TBS) for 1 h at RT. Membranes were incubated with primary antibodies in blocking buffer overnight at 4 °C. After several washing steps membranes were incubated with secondary fluorescence-labeled antibody for 1 h at RT, washed four times with TBST and scanned using the Odyssey Western System (LICOR Systems, Lincoln, USA). Antibodies used for Western blot analysis are listed in Table 2 .
2.7. CX43 analysis
CX43 analysis was performed on cerebral cortex using 19 gray matter tissue blocks from 12 control cases and 24 NAGM tissue blocks from 15 MS cases ( Table 1 ). Average staining intensity was measured for 20 segments throughout the cortical layers (an example is given in Fig. 3 A and C). For every tissue block all possible gyri were measured and an average staining intensity for all 20 segments for every tissue block was calculated. Intensity values of each block were normalized to the staining intensity of the non-lesioned adjacent normal appearing white matter (NAWM; showing no significant differences between control and MS patients) to correct for possible minor staining artefacts. Values from different blocks from the same patients were then combined and an average value was calculated. Statistical analysis over the 20 segments was then calculated by performing a two-way ANOVA (Disease, Segments). To further assess a possible reduction of CX43 in control gray matter tissue, a hypothetic “evenly stained” group was created by calculating the average intensity for all 20 segments by using data from all tissue blocks without a reduction in CX43 immunoreactivity. Differences between this virtual group and control and MS cases were then evaluated by using the Holm-Sidak method, with alpha = 5.000%.
2.8. In situ hybridization
Synthetic digoxigenin-labeled riboprobes (cRNA) were generated from recombinant pCRTMII-Topo® plasmid containing a 723 bp cDNA insert of humanMCT1sequence (5′-11-733-3′) and a 836 bp cDNA insert of humanIPAFsequence (5′-50-865-3′). Transcription was done from both sides with either SP6 or T7 RNA polymerase, generating antisense or sense (control) cRNA probes. Further, synthetic digoxigenin-labeled riboprobes (cRNA) were generated from recombinant pBSKSII plasmid containing a 823 bp cDNA insert of humanCASP4sequence (5′-56-878-3′), a 130 bp cDNA insert of humanASCsequence (5′-600-729-3′), a 539 bp cDNA insert of humanNALP1sequence (5′-2541-3079-3′), and a 713 bp cDNA insert of humanNALP3sequence (5′-2693-3405-3′). Transcription was done from both sides with either T3 or T7 RNA polymerase, generating antisense or sense (control) cRNA probes. In situ hybridization was performed on 14 μm cryosections of freshly frozen tissues as described previously (Schaeren-Wiemers and Gerfin-Moser, 1993 and Graumann et al, 2003). In situ hybridization signal was revealed by alkaline phosphatase with BCIP and NBP as substrate. Immunohistochemistry was performed as described above. Antibodies used for immunohistochemical stainings are listed in Table 2 .
For determination of Il-1β secretion, astrocytes were stimulated with LPS, ATP or ATP/LPS for either 1 or 24 h (see above for treatments conditions). Il-1β concentration was measured by ELISA using the ELISA MAX Deluxe Set Mouse Il-1β Kit (Biolegend, San Diego, USA) according to the manufacturer’s instructions.
2.10. CFA-induced peripheral immune stimulation in C57/Bl6 mice
8 weeks old C57Bl/6 mice were bred in-house and kept together in groups of five. Four independent experiments were made with 5 (pilot study) or 10 mice per treatment group. Mice were anaesthetized with 3% isoflurane and immunized subcutaneously in the flanks with either 0.1 ml PBS with 0.1 ml complete Freund’s adjuvant (CFA) containing 0.4 mg Mycobacterium tuberculosis (CFA/PT) or with 0.2 ml PBS (Control). 200 ng pertussis toxin was injected intraperitonally on the day of immunization and two days later. Injections were made in the SPF animal facility. Animals were euthanatized using carbon dioxide and subsequent decapitation. Spinal cord tissue was dissected at day 14 after immunization. All experiments were approved by the Veterinary Office of the State of Basel. Sample size was estimated by using the following assumptions derived from earlier experiments (Difference of means = 20%; common standard deviation = 30%; alpha = 0.05; power = 80%). Both, male and female animals were randomly assigned to the treatment groups, and the analysis was performed by a nonblinded investigator. In total, 35 control and 34 CFA/PT injected mice were analyzed. One animal of the CFA/PT group was excluded due to bad RNA quality. No adverse events were noted during all experiments.
3.1. Metabolic genes are downregulated in chronic MS NAGM
To identify molecular alterations in cortical MS NAGM, a differential gene expression analysis was performed. 8 control and 13 MS cases ( Table 1 ) were investigated using a custom microarray containing 1176 representative cDNA sequences. In MS NAGM, we found a significantly decreased expression of genes involved in the ANLS and in the GGC; for example the monocarboxylate transporter 1 (MCT1; Solute carrier family 16, member 1, SLC16A1) and the excitatory amino acid transporter 2 (EAAT2; GLT-1; solute carrier family 1, member 2, SLC1A2; Fig. 1 A and B). Additionally, a downregulation of connexin 43 (CX43, gap junction protein, alpha 1, GJA1) was detected ( Fig. 1 A and B).
In order to validate and substantiate our observations we analyzed additional NAGM tissues of up to 14 control and 17 MS cases by qRT-PCR. We further enlarged our analysis by additional genes functionally connected to the ones identified by the microarray, e.g. genes from the ANLS to which MCT1 belongs to, as well as additional genes of the GGC ( Fig. 1 A and B). Results from the qRT-PCR confirmed the transcriptional downregulation ofMCT1,CX43,EAAT2, glutamine synthetase (GLUL) and glucose transporter type 1 (GLUT1; solute carrier family 2 (facilitated glucose transporter, member 1, SLC2A1). Additionally, we detected a significant downregulation of N-system amino acid transporter 1 (NAT1; solute carrier family 38, member 3, SLC38A3) and ATPase, Na+/K+ transporting, alpha 2 polypeptide (ATP1A2). In contrast, an upregulation of lactate dehydrogenase A (LDHA) was found. All genes of interest ( Fig. 1 A) were further evaluated whether their expression pattern is influenced by confounding factors such as post-mortem delay time, age at death, disease duration, clinical course or gender. Neither significant correlation nor differential expression due to any of these confounding factors could be identified.
3.2. MCT1, CX43 and EAAT2 are expressed in astrocytes in MS NAGM
MCT1, CX43 as well as EAAT2 are expressed by astrocytes in human brain (Milton et al, 1997, Rouach et al, 2002, Pellerin et al, 2005, Pierre and Pellerin, 2005, and Chiry et al, 2006). To confirm their expression pattern in MS NAGM we performed immunohistochemistry and immunofluorescence co-localization studies. MCT1 staining was strongest on blood vessels and, additionally, a diffuse staining of the neuropil was evident ( Fig. 2 A, inset shows a blood vessel and the surrounding neuropil at higher magnification). Immunofluorescence co-localization microscopy of MCT1 and GFAP demonstrates MCT1 expression in human cortical astrocytes ( Fig. 2 B, arrows). The expression pattern of CX43 in MS NAGM was variable within the cortical layers ( Fig. 2 C, inset shows a single CX43-positive astrocyte at higher magnification). Immunofluorescence co-localization microscopy of CX43 and GFAP confirmed the exclusive cellular expression of CX43 in human astrocytes ( Fig. 2 D, arrow). Finally, the glutamate transporter EAAT2 was also detected to be expressed exclusively in astrocytes of the MS NAGM ( Fig. 2 E and F).
Currently, a controversial discussion is taking place in the scientific community concerning the localization of MCT1 expression in brain tissue. As IHC stainings of MCT1 showed an overall diffuse staining of the neuropil, we have performed an additional in situ hybridization analysis to further validate the cellular expression pattern of MCT1 within the human cerebral cortex. We found a widespread MCT1 expression in many different cell types such as in endothelial cells of blood vessels ( Fig. 2 G and H, arrowhead), in astrocytes identified by immunohistochemistry for GFAP ( Fig. 2 H, arrows), in a subpopulation of cortical neurons identified by NeuN ( Fig. 2 I, arrow) and in Olig2-positive cells ( Fig. 2 J, arrows). Altogether, our data confirm the expression of MCT1, CX43 and EAAT2 in astrocytes in the MS NAGM.
3.3. CX43 protein reduction in MS NAGM is characterized by its loss in individual astrocytes
Reduced expression ofCX43as well as the inhomogeneous CX43 protein expression pattern in MS NAGM ( Fig. 2 C) prompted us to investigate the expression pattern of CX43 in control and MS cortex in more detail. Whereas the NAGM of most of the MS cases was characterized by a selective absence of CX43 in individual astrocytes ( Fig. 3 A and B), such a CX43 loss was found to a much lesser extent in the gray matter of control cases ( Fig. 3 C and D).
The reduction of CX43 was not homogenous throughout the cortical layers. Instead the reduction was observed in individual astrocytes within layers III–V with single astrocytes still expressing CX43 (e.g. Figs. 3 A, B and 2 C inset). In some cases the reduction of CX43 was so profound that only few Cx43 positive astrocytes remained as a thin rim in layer I and layer VI ( Fig. 3 B as an example). To quantify the CX43 reduction in MS NAGM, we measured CX43 staining intensity over all layers of the gray matter from 12 control and 14 MS cases ( Fig. 3 E and F). For each case, we measured the average CX43 staining intensity of 20 segments reaching from layer I to layer VI ( Fig. 3 A and C) and normalized it to the adjacent white matter staining intensity to account for minor staining artifacts (no statistical significant difference was detected between control and MS white matter CX43 staining). Statistical analysis over the segments from MS NAGM versus controls revealed that the CX43 expression was significantly more reduced in MS NAGM than in control gray matter ( p < 0.0001, Two-way ANOVA) with an average reduction of about 32% throughout all layers ( Fig. 3 E). If compared to a virtual “evenly stained” group, generated from tissues with an even CX43 staining throughout all layers (e.g. Fig. 3 C), MS NAGM but also control gray matter were statistically significantly different (p < 0.0001 for both, Holm-Sidak’s multiple comparisons test; Fig. 3 F). As in a study of human autopsy tissue additional factors may affect CX43 immunoreactivity, we investigated whether the degree of CX43 reduction correlates with post-mortem delay time, age at death, disease duration or gender. No significant correlation or difference was found due to these possible confounding factors (data not shown). A comparable immunohistochemical staining pattern as for CX43 was also detected for EAAT2 ( Fig. 2 E). This suggests that EAAT2 expression might be reduced in individual astrocyte as observed for CX43. Taken together, our analysis demonstrates that CX43, the major gap junction protein expressed by astrocytes and involved in the formation of the astrocyte syncytium, is downregulated in MS NAGM most prominently in layers III–V. A recent study identified a higher CX43 expression in MS cortical NAGM in comparison to control samples ( Markoullis et al., 2014 ). Whether this discrepancy is due to the smaller sample number investigated or to the heterogeneous expression of CX43, which we also observed in control cases, cannot be verified.
3.4. Differential gene expression in MS NAGM is indicative of immune-related signaling
Simultaneously to the downregulation of metabolic genes in MS NAGM, the inflammasome associated cytokine, interleukin-1beta (IL1β), was significantly upregulated in MS ( Fig. 4 A). In keeping with this, Ingenuity pathway upstream analysis (Ingenuity Systems, www.ingenuity.com ) of the microarray data identified a Toll-like/IL-1β signaling expression signature to be present in the differential gene expression pattern of MS NAGM (IL1 group activationz-score 4.748, IL1B activationz-score 3.538; LPS/TLR4 activationz-score 4.441; Fig. 4 A). Further, additional immune-related signaling signatures were found in the differential gene expression pattern of MS NAGM (e.g. TNF, IL2, IFNG; Fig. 4 A). Together, this strongly suggests that gene expression in MS NAGM is influenced by immune-related signaling.
3.5. ANLS and GGC genes in astrocytes are downregulated upon immune-related signaling
The upregulation ofIL1Bas well as the immune-related signaling signatures and the downregulation of ANLS and GGC genes in MS NAGM lead to the question whether these phenomenons are interconnected. Therefore, we investigated whether immune-related signaling directly affect astrocyte ANLS and GGC gene expression. For that, we chose IL-1β, which was upregulated in MS NAGM and whose signaling signature was detected in the MS NAGM, and the mediator of the identified second most significant signaling signature, the TLR4 agonist LPS ( Fig. 4 A). We treated mouse primary cortical astrocyte cultures with either 0.25 ng/ml IL-1β or 10 ng/ml LPS for 24 h. qRT-PCR analysis of treated astrocytes revealed significant transcriptional downregulation ofCx43,Glul,Ldhb,Glut1,Nat1(IL-1β and LPS treatment), andEaat1andAtp1a2(LPS treatment only),whereas the expression ofEaat2and the astrocyte specific monocarboxylate transporterMct4(Solute carrier family 16, member 3, Slc16a3) (both upon LPS treatment) were significantly upregulated ( Fig. 4 B). Western blot analysis as well as immunofluorescence stainings confirmed strong downregulation of Cx43 on the protein level ( Fig. 4 C and F). In the case of Mct1, a significant upregulation could be detected ( Fig. 4 B) whereas its protein expression was rather the opposite ( Fig. 4 C). The unchanged or slightly upregulated expression of EAAT2 and the induction of MCT1 mRNA expression in pure astrocyte cultures might be due to the lack of their normal cellular counterparts such as neurons and oligodendrocytes in the cerebral cortex. Our findings show that cultured astrocytes exposed to immune-related stimuli such as IL-1β and LPS alter the expression of their ANLS and GGC genes substantially.
3.6. ANLS and GGC genes are differentially expressed upon inflammasome activation
The upregulation ofIL1Bin MS NAGM suggests priming or even activation of inflammasomes in the tissue. To investigate whether inflammasome activation can also take place in astrocytes and whether this influences gene expression of the ANLS and GGC components, we treated cortical astrocyte cultures with ATP alone, mimicking possible tissue damage in MS, as well as with ATP plus LPS in a similar manner as described for inflammasome activation (Schroder and Tschopp, 2010 and Zhou et al, 2011). In general, treatment with ATP or ATP/LPS led to significant downregulation ofCx43,Ldhb,Glut1,Nat1,Eaat1andMct4as well asAtp1a2if treated by ATP/LPS ( Fig. 4 D). In contrast,Glulexpression was significantly upregulated by both treatments ( Fig. 4 D) as well asAtp1a2if treated with ATP alone. WhereasEaat2andMct1where downregulated by ATP treatment, ATP/LPS treatment led to an upregulation of these genes. ForMct1, this is in line with the data obtained from the LPS and IL-1β treatment ( Fig. 4 B). Western blot analysis as well as immunofluorescence stainings confirmed the strong downregulation of Cx43 on the protein level ( Fig. 4 E and F), and also Mct1 protein expression was slightly but significantly downregulated ( Fig. 4 E). Taken together, our data demonstrate that astrocytes challenged with inflammasomal triggers respond by changing the expression of genes involved in the ANLS and GGC.
3.7. Peripheral inflammation signals reduce Mct1 and Cx43 expression in the spinal cord of C57/Bl6 mice
In MS, immune-related signaling might origin from immunological processes (e.g. lesion formation) in regions far away of the investigated MS NAGM tissue. Therefore, we asked the question whether signaling from the activated immune system can lead to a response in the gray matter tissue similar to that observed in MS NAGM brain tissue and in astrocyte cultures. To test this hypothesis we treated C57Bl/6 mice with a subcutaneous injection of 100 μl CFA with 100 μl PBS in the base of the tail, and twice pertussis toxin i.p. according to the standard EAE protocol. Injection of myelin auto-antigen was omitted to exclude possible T cell infiltration and to limit the immunological reaction to the outside of the CNS. In order to detect CNS gene expression changes also due to possible signaling by primed T cell whose accumulation in the meninges peaks around d11-12 after EAE induction ( Kivisakk et al., 2009 ), spinal cord tissues were dissected at day 14 after CFA injection. Gene expression in cervical spinal cord was then analyzed by qRT-PCR. This protocol, which does not lead to the development of demyelinating lesions in the CNS, caused a small but persistent significant downregulation ofMct1andCx43( Fig. 5 A). No changes in gene expression were seen at that stage forEaat1and2,Glul,Ldha,Ldhb,Atp1a2,Nat1as well asGlut1(data not shown). Whether the expression of these genes has already normalized after the CFA injection or whether they will be changed later or not at all has to be further elucidated. However, our data show that a CNS-unspecific inflammation in the periphery can lead to alterations of CNS metabolic gene expression.
3.8. Inflammasome components are expressed in human cerebral cortex
The upregulatedIL1Bexpression, together with the found IL-1β signaling signature in the MS NAGM raised the question whether inflammasome activation might participate in MS NAGM alterations. Therefore, we first investigated if specific NLR inflammasome sensor molecules (NLRC4, NLRP1, NLRP3), the inflammasome adaptor ASC, and caspase effectors (CASP1, CASP4) are expressed in control and MS NAGM. In situ hybridization analysis demonstrated that these inflammasome components are expressed throughout the human cortex ( Fig. 5 B). We could demonstrate that CASP4, ASC, NLRC4, NLRP1, and NLRP3 were expressed in astrocytes ( Fig. 5 B, arrows). This indicates that astrocytesin vivoare capable to elicit inflammasome activation.
3.9. Cortical astrocytes are capable to release IL-1β upon inflammasome activation
The upregulation ofIL1Bin MS NAGM further provides evidence that immune-related signaling leading to a disturbance in astrocyte metabolic processes may take place in the tissue itself. AsIL1Bwas upregulated and astrocytes expressed the necessary components for inflammasome activation we investigated whether these cells could be induced to trigger inflammasome-dependent IL-1β release. The expression profiling of NAGM indicated thatIL1Bgene expression was induced in MS patients as compared healthy controls ( Fig. 4 A), suggesting that “signal 1”of the activation cascade has already occurred in MS NAGM. We thus hypothesized that LPS would also trigger expression of theIl1bgene in primary astrocytes, as is already established for immune cells. Indeed, LPS strongly upregulatedIl1bgene expression, compared to control astrocyte cultures whereIl1bgene expression was below the detection limit (data not shown).
We then treated cortical astrocyte cultures with LPS alone (signal 1 alone), ATP alone (signal 2 alone), or LPS plus ATP (both signals 1 and 2) as described for Nlrp3 inflammasome activation (Schroder and Tschopp, 2010 and Zhou et al, 2011) and measured IL-1β release. While LPS strongly induced the expression of theIl1bgene in primary astrocytes (data not shown), IL-1β release was weak and present only after 24 h treatment ( Fig. 5 C). Treatment with ATP alone did neither induceIl1bexpression nor IL-1β release. However, strong IL-1β release, although only minorIl1binduction, required cell stimulation with both LPS and ATP ( Fig. 5 C), which is in line with studies in immune cells.
Beside lesions, which are considered to be the pathological hallmark of MS, numerous molecular alterations have been identified in the so-called normal-appearing tissue (Graumann et al, 2003, Lindberg et al, 2004, Dutta et al, 2006, Dutta et al, 2007, Zeis et al, 2008, and Zeis et al, 2009). Among these processes, alterations of the gray matter are of particular interest since they appear to be associated with clinical deficits and disease course of MS. Magnetic resonance imaging (MRI) studies for example revealed a strong correlation between gray matter atrophy and future disability progression (Amato et al, 2007 and Fisher et al, 2008). Gray matter pathology has further been linked with physical disability and cognitive impairment ( Pirko et al., 2007 ) as well as fatigue (Roelcke et al, 1997, Filippi et al, 2002, and Niepel et al, 2006), which are common manifestations of MS (Wishart and Sharpe, 1997, Blinkenberg et al, 2000, Lazeron et al, 2000, Amato et al, 2004, Benedict et al, 2004, Morgen et al, 2006, Sanfilipo et al, 2006, and Houtchens et al, 2007). However, to date the precise pathological alterations underlying these manifestations is not known. To investigate molecular mechanisms possibly underlying these impairments in MS we analyzed the molecular alterations in histopathological normal appearing gray matter in MS.
Transcriptional profiling revealed that MS NAGM is associated with the downregulation of gene transcripts involved in the GGC and in the ANLS ( Fig. 1 B), an experimentally-based framework proposed by Magistretti and coworkers (Pellerin and Magistretti, 1994, Bittar et al, 1996, Pellerin et al, 1998, and Magistretti et al, 1999). In the ANLS concept, neuronal activity is coupled to astrocytic glucose utilization involving an activation of aerobic glycolysis in astrocytes and lactate consumption by neurons ( Pellerin et al., 2007 ). Reduction of the ANLS might lead to a decrease in activity-dependent lactate delivery to neurons, and thus to a decrease in neuronal activity as suggested in the study of Dutta and coworkers ( Dutta et al., 2006 ). Such a mechanism was hypothesized as a possible cause for central or mental fatigue ( Ronnback and Hansson, 2004 ), which is observed in the majority of MS cases. Astrocyte–neuron lactate transport is also required for memory formation (Newman et al, 2011 and Suzuki et al, 2011). Hence, chronic reduction of this particular energy supply, as suggested by our findings, might lead to some of the observed cognitive impairments in MS (for review see Ferreira, 2010 ). Further, they might also have an impact on synaptic activity in the cerebral cortex especially in layer III–V, the major in- and output of the cerebral cortex. Exactly in these regions we detected the strongest reduction of CX43 and potentially of the glutamate transporter EAAT2. A similar reduction of the astroglial EAAT2 was also found in neuromyelitis optica (NMO; Hinson et al., 2008 ), an inflammatory demyelinating disease earlier thought to be a variant of MS ( Brosnan and Raine, 2013 ). Changed expression of CX43 as well as EAAT2 was suggested to be crucially involved in memory processes (Frisch et al, 2003, Stehberg et al, 2012, and Poletti et al, 2014). In summary, a reduced ANLS and GGC gene expression in the MS NAGM might lead to many of the encountered clinical features of MS such as fatigue, cognitive impairment or physical disability as discussed before.
The biological activity of IL-1β is controlled by signaling complexes called inflammasomes ( Schroder and Tschopp, 2010 ), which are emerging as central drivers of innate immune function ( Martinon et al., 2009 ). Concomitant to ANLS and GGC gene downregulation, the expression of the gene encoding the pro-inflammatory cytokine IL-1β was upregulated. Further, and in keeping withIL1Bgene induction,in silicoupstream regulator analysis revealed gene signatures indicative of Il-1β signaling, suggesting innate immune activation in MS NAGM. This is supported by reports showing elevated expression of the purinergic receptor P2X ATP-gated ion channel 7 (P2X7R), mediating ATP-dependent NLRP3 activation, in the spinal cord of MS patients ( Yiangou et al., 2006 ). Also, elevated levels of the NLRP3 agonist uric acid were found in the CSF of MS patients ( Amorini et al., 2009 ). As the receptor for IL-1β is ubiquitously expressed among all brain cell types ( Allan et al., 2005 ), a chronic exposure of gray matter tissue to IL-1β in MS patients as suggested by our differential gene expression analysis may be a crucial factor driving MS pathogenesis. In rats, intracerebral administration of IL-1β led to oligodendrocytes apoptosis ( Fan et al., 2009 ), and subarachnoid injection of proinflammatory cytokines such as TNF and IFNγ lead to subpial demyelination ( Gardner et al., 2013 ). Chronic IL-1β release, as suggested by the global upregulation ofIL1Bin MS NAGM, may be further pathologically relevant as it was reported to lead to neuronal death mediated by astrocytes ( Thornton et al., 2006 ).
The concomitant presence of immune-related signaling signatures, potential inflammasomal activation as well as downregulation of ANLS and GGC genes in MS NAGM suggests that these phenomenons are interconnected. Treatment experiments of cultured cortical astrocytes demonstrated that this is indeed the casein-vitro. We found that in all treatments made (IL-1β, ATP, LPS and ATP/LPS), genes from the ANLS as well as the GGC have been downregulatedin vitro. This demonstrates a direct link between immune-related signaling and the expression of ANLS and GGC genes in astrocytes, resembling our observations for MS NAGMin vivo. These findings are in line with reports showing that pro-inflammatory cytokines can modulate the astrocytic metabolic phenotype ( Gavillet et al., 2008 ). This suggests that in MS, immune-related signaling, either from brain exogenous or endogenous sources lead to a reduction in ANLS and GGC genes. Finally, this might result in cognitive impairments as discussed before. A possible impact of immune-related signals on cognition is further supported by the findings of long-term cognitive impairment among survivors of severe sepsis ( Iwashyna et al., 2010 ). Additionally, peripheral inflammation was reported to acutely impair human spatial memory ( Harrison et al., 2014 ).
Although gene expression regulation of treated astrocytes resembles gene regulation in MS NAGM, some genes showed a different expression behavior. This is mostly true for MCT1 and EAAT2. The use of pure astrocyte cultures to investigate immune mediated changes in their gene expression offers methodological advances. However, they are likely to exhibit different properties than thosein vivo( Halim et al., 2010 ). We suspect that in a controlled culture environment with constant glutamate supply, downregulation of EAAT2 might be overruled. In the case of MCT1, Nijland et al., 2014 showed that MCT1 is strongly upregulated in MS white matter lesions whereas its expression is reduced in perilesional areas ( Nijland et al., 2014 ). We therefore speculate that MCT1 expression in astrocytes is dependent on their reactivity state. Our treatment might, in the case of MCT1, activate astrocytes to a higher degree so that MCT1 gets upregulated as e.g. in MS white matter lesions. In contrast, astrocytes of the NAGM might be activated to a lesser extent, downregulating MCT1 as shown in MS perilesional areas ( Nijland et al., 2014 ).
Immune-related signaling in MS NAGM can be of brain exogenous but also endogenous origin. Downregulation of Cx43 and Mct1 in thein vivoadjuvant driven peripheral immunization experiment provides evidence that ANLS gene expression alterations in the MS NAGM might result from peripheral events and subsequent immune-related signaling to the brain tissue. Recent investigations detected TNF and IFNγ in inflamed meninges and in cerebrospinal fluid of MS cases ( Gardner et al., 2013 ), which could act as such proinflammatory stimuli. Further, Dujmovic et al., 2009 demonstrated elevated IL-1β levels in the CSF and sera of MS patients ( Dujmovic et al., 2009 ). As it was shown that cytokines can be transported through the blood brain barrier (for review see Banks et al., 1995 ), continuous and/or repetitive inflammatory events, such as lesion formation, might lead to the observed downregulation of ANLS and GGC genes in MS NAGM. However, immune-related signaling in MS NAGM might also be of endogenous origin. This notion is supported by the finding of a CX43 loss in the middle of the NAGM in layer III-V. Immune-related signaling from exogenous sources as e.g. from the meninges as recently suggested (Magliozzi et al, 2010 and Gardner et al, 2013) would most probably lead to a gradient of CX43 loss from layer I towards layer VI as e.g. shown for neuronal loss ( Magliozzi et al., 2010 ). Loss of CX43 due to immune-related signaling from blood vessels would further be expected to strike the whole NAGM rather than a demarcated zone in layer III–V. An endogenous signaling is also supported by the upregulated expression ofIL1Bin the MS NAGM itself, lacking immune cell infiltration as a source for this gene expression. Finally, besides microglia known to express inflammasomes and release cytokines and chemokines ( Benveniste, 1997 ), an intrinsic signaling by activated astrocytes might be the source of MS NAGM ANLS and GGC downregulation as we could show that astrocytes themselves express the necessary inflammasome components and can be stimulated to release IL-1β.
Although recent progress has been made in MS treatment to date ( Kappos et al., 2010 ), clinical disability continues to rise in MS patients over time. Our study demonstrates that in MS patients, astrocytes display a differential ANLS and GGC gene expression phenotype that is a likely cause for encountered clinical features in MS such as fatigue, cognitive impairment or physical disability. Further, we posit that the observed astrocyte differential gene expression in MS can be caused by immune-related signaling (e.g. inflammasomes) of exogenous but also endogenous origin. Our findings highlight the importance of subtle gray matter changes in MS pathogenesis, and the need to develop therapies to interfere with the responsible molecular pathways (e.g. IL-1β) to ameliorate clinical symptoms of MS.
Conflict of interest
The authors declare no competing financial interests.
This study was supported by the National Multiple Sclerosis Society of the United States of America (Grant RG 3583A1 and RG 4249A2), the Swiss Multiple Sclerosis Society, by the Novartis Stiftung (all to NSW), and by FNRS Grant 31003A-130821/1 and /2, by the National Center of Competence in Research (NCCR) “SYNAPSY – financed by the Swiss National Science Foundation (n° 51AU40_125759) to PJM and by the Panacée Foundation.
Study concept and design: NSW, qRT-PCR, microarray analysis, immunofluorescence, immunohistochemistry and analysis, Western blot analysis: TZ, immunohistochemistry, in situ hybridization: MG, primary astrocyte culture experiments: IA and PM, inflammasome related experiments: KS, JT; drafting of the manuscript: TZ, IA, KS, PM, and NSW.
We thank Prof. Dr. Richard Reynolds (UK Multiple Sclerosis Tissue Bank, Charing Cross Hospital London, UK) and Prof. Dr. Markus Tolnay (Pathology, University Hospital Basel, Switzerland) for providing human brain tissues and Dr. Jochen Kinter for tissue collection and experimental contribution. Further, we thank the DBM Bioinformatics Core Facility, Dr. Robert Ivanek and Dr. Florian Geier for statistical support and analysis. We also thank Profs. Stephen Miller (Northwestern University Medical School, Chicago, USA), Britta Engelhardt (Robert-Kocher Institute, University of Bern, Switzerland), Wendy Macklin (University of Colorado School of Medicine, Denver, USA) and Wolfgang Brück (University Hospital, Göttingen, Germany) for fruitful discussions, and Roland Martin (Neurology, University Hospital Zürich, Switzerland) and Richard Ransohoff (Cleveland Clinic Lerner College of Medicine, Cleveland, USA) for critical reading the manuscript.
- Allaman et al., 2004 I. Allaman, L. Pellerin, P.J. Magistretti. Glucocorticoids modulate neurotransmitter-induced glycogen metabolism in cultured cortical astrocytes. J. Neurochem.. 2004;88:900-908
- Allan et al., 2005 S.M. Allan, P.J. Tyrrell, N.J. Rothwell. Interleukin-1 and neuronal injury. Nat. Rev. Immunol.. 2005;5:629-640 Crossref
- Amato et al., 2004 M.P. Amato, M.L. Bartolozzi, V. Zipoli, E. Portaccio, M. Mortilla, L. Guidi, G. Siracusa, S. Sorbi, A. Federico, N. De Stefano. Neocortical volume decrease in relapsing-remitting MS patients with mild cognitive impairment. Neurology. 2004;63:89-93 Crossref
- Amato et al., 2007 M.P. Amato, E. Portaccio, B. Goretti, V. Zipoli, M. Battaglini, M.L. Bartolozzi, M.L. Stromillo, L. Guidi, G. Siracusa, S. Sorbi, A. Federico, N. De Stefano. Association of neocortical volume changes with cognitive deterioration in relapsing-remitting multiple sclerosis. Arch. Neurol.. 2007;64:1157-1161 Crossref
- Amorini et al., 2009 A.M. Amorini, A. Petzold, B. Tavazzi, J. Eikelenboom, G. Keir, A. Belli, G. Giovannoni, V. Di Pietro, C. Polman, S. D’Urso, R. Vagnozzi, B. Uitdehaag, G. Lazzarino. Increase of uric acid and purine compounds in biological fluids of multiple sclerosis patients. Clin. Biochem.. 2009;42:1001-1006 Crossref
- Andersen et al., 2004 C.L. Andersen, J.L. Jensen, T.F. Orntoft. Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets. Cancer Res.. 2004;64:5245-5250 Crossref
- Banks et al., 1995 W.A. Banks, A.J. Kastin, R.D. Broadwell. Passage of cytokines across the blood-brain barrier. NeuroImmunoModulation. 1995;2:241-248 Crossref
- Benedict et al., 2004 R.H. Benedict, D.A. Carone, R. Bakshi. Correlating brain atrophy with cognitive dysfunction, mood disturbances, and personality disorder in multiple sclerosis. J. Neuroimaging. 2004;14:36S-45S Crossref
- Benveniste, 1997 E.N. Benveniste. Role of macrophages/microglia in multiple sclerosis and experimental allergic encephalomyelitis. J. Mol. Med.. 1997;75:165-173 Crossref
- Bittar et al., 1996 P.G. Bittar, Y. Charnay, L. Pellerin, C. Bouras, P.J. Magistretti. Selective distribution of lactate dehydrogenase isoenzymes in neurons and astrocytes of human brain. J. Cereb. Blood Flow Metab.. 1996;16:1079-1089 Crossref
- Blinkenberg et al., 2000 M. Blinkenberg, K. Rune, C.V. Jensen, M. Ravnborg, S. Kyllingsbaek, S. Holm, O.B. Paulson, P.S. Sorensen. Cortical cerebral metabolism correlates with MRI lesion load and cognitive dysfunction in MS. Neurology. 2000;54:558-564
- Brosnan and Raine, 2013 C.F. Brosnan, C.S. Raine. The astrocyte in multiple sclerosis revisited. Glia. 2013;61:453-465 Crossref
- Chiry et al., 2006 O. Chiry, L. Pellerin, F. Monnet-Tschudi, W.N. Fishbein, N. Merezhinskaya, P.J. Magistretti, S. Clarke. Expression of the monocarboxylate transporter MCT1 in the adult human brain cortex. Brain Res.. 2006;1070:65-70 Crossref
- Compston and Coles, 2002 A. Compston, A. Coles. Multiple sclerosis. Lancet. 2002;359:1221-1231 Crossref
- Dujmovic et al., 2009 I. Dujmovic, K. Mangano, T. Pekmezovic, C. Quattrocchi, S. Mesaros, N. Stojsavljevic, F. Nicoletti, J. Drulovic. The analysis of IL-1 beta and its naturally occurring inhibitors in multiple sclerosis: the elevation of IL-1 receptor antagonist and IL-1 receptor type II after steroid therapy. J. Neuroimmunol.. 2009;207:101-106 Crossref
- Dutta et al., 2007 R. Dutta, J. McDonough, A. Chang, L. Swamy, A. Siu, G.J. Kidd, R. Rudick, K. Mirnics, B.D. Trapp. Activation of the ciliary neurotrophic factor (CNTF) signalling pathway in cortical neurons of multiple sclerosis patients. Brain. 2007;130:2566-2576 Crossref
- Dutta et al., 2006 R. Dutta, J. McDonough, X. Yin, J. Peterson, A. Chang, T. Torres, T. Gudz, W.B. Macklin, D.A. Lewis, R.J. Fox, R. Rudick, K. Mirnics, B.D. Trapp. Mitochondrial dysfunction as a cause of axonal degeneration in multiple sclerosis patients. Ann. Neurol.. 2006;59:478-489 Crossref
- Fan et al., 2009 L.-W. Fan, H.J. Mitchell, L.-T. Tien, P.G. Rhodes, Z. Cai. Interleukin-1β-induced brain injury in the neonatal rat can be ameliorated by α-phenyl-n-tert-butyl-nitrone. Exp. Neurol.. 2009;220:143-153 Crossref
- Ferreira, 2010 M.L. Ferreira. Cognitive deficits in multiple sclerosis: a systematic review. Arq. Neuropsiquiatr.. 2010;68:632-641 Crossref
- Filippi et al., 2002 M. Filippi, M.A. Rocca, B. Colombo, A. Falini, M. Codella, G. Scotti, G. Comi. Functional magnetic resonance imaging correlates of fatigue in multiple sclerosis. Neuroimage. 2002;15:559-567 Crossref
- Fisher et al., 2008 E. Fisher, J.C. Lee, K. Nakamura, R.A. Rudick. Gray matter atrophy in multiple sclerosis: a longitudinal study. Ann. Neurol.. 2008;64:255-265 Crossref
- Frisch et al., 2003 C. Frisch, M. Theis, M.A. De Souza Silva, E. Dere, G. Sohl, B. Teubner, K. Namestkova, K. Willecke, J.P. Huston. Mice with astrocyte-directed inactivation of connexin43 exhibit increased exploratory behaviour, impaired motor capacities, and changes in brain acetylcholine levels. Eur. J. Neurosci.. 2003;18:2313-2318 Crossref
- Furlan et al., 1999 R. Furlan, G. Martino, F. Galbiati, P.L. Poliani, S. Smiroldo, A. Bergami, G. Desina, G. Comi, R. Flavell, M.S. Su, L. Adorini. Caspase-1 regulates the inflammatory process leading to autoimmune demyelination. J. Immunol.. 1999;163:2403-2409
- Gardner et al., 2013 C. Gardner, R. Magliozzi, P.F. Durrenberger, O.W. Howell, J. Rundle, R. Reynolds. Cortical grey matter demyelination can be induced by elevated pro-inflammatory cytokines in the subarachnoid space of MOG-immunized rats. Brain. 2013;136:3596-3608 Crossref
- Gavillet et al., 2008 M. Gavillet, I. Allaman, P.J. Magistretti. Modulation of astrocytic metabolic phenotype by proinflammatory cytokines. Glia. 2008;56:975-989 Crossref
- Graumann et al., 2003 U. Graumann, R. Reynolds, A.J. Steck, N. Schaeren-Wiemers. Molecular changes in normal appearing white matter in multiple sclerosis are characteristic of neuroprotective mechanisms against hypoxic insult. Brain Pathol.. 2003;13:554-573
- Gris et al., 2010 D. Gris, Z. Ye, H.A. Iocca, H. Wen, R.R. Craven, P. Gris, M. Huang, M. Schneider, S.D. Miller, J.P. Ting. NLRP3 plays a critical role in the development of experimental autoimmune encephalomyelitis by mediating Th1 and Th17 responses. J. Immunol.. 2010;185:974-981 Crossref
- Halim et al., 2010 N.D. Halim, T. McFate, A. Mohyeldin, P. Okagaki, L.G. Korotchkina, M.S. Patel, N.H. Jeoung, R.A. Harris, M.J. Schell, A. Verma. Phosphorylation status of pyruvate dehydrogenase distinguishes metabolic phenotypes of cultured rat brain astrocytes and neurons. Glia. 2010;58:1168-1176
- Harrison et al., 2014 N.A. Harrison, C.F. Doeller, V. Voon, N. Burgess, H.D. Critchley. Peripheral inflammation acutely impairs human spatial memory via actions on medial temporal lobe glucose metabolism. Biol. Psychiatry. 2014;76:585-593
- Hinson et al., 2008 S.R. Hinson, S.F. Roemer, C.F. Lucchinetti, J.P. Fryer, T.J. Kryzer, J.L. Chamberlain, C.L. Howe, S.J. Pittock, V.A. Lennon. Aquaporin-4-binding autoantibodies in patients with neuromyelitis optica impair glutamate transport by down-regulating EAAT2. J. Exp. Med.. 2008;205:2473-2481 Crossref
- Houtchens et al., 2007 M.K. Houtchens, R.H. Benedict, R. Killiany, J. Sharma, Z. Jaisani, B. Singh, B. Weinstock-Guttman, C.R. Guttmann, R. Bakshi. Thalamic atrophy and cognition in multiple sclerosis. Neurology. 2007;69:1213-1223 Crossref
- Inoue et al., 2012 M. Inoue, K.L. Williams, M.D. Gunn, M.L. Shinohara. NLRP3 inflammasome induces chemotactic immune cell migration to the CNS in experimental autoimmune encephalomyelitis. Proc. Natl. Acad. Sci. U.S.A.. 2012;109:10480-10485 Crossref
- Iwashyna et al., 2010 T.J. Iwashyna, E.W. Ely, D.M. Smith, K.M. Langa. Long-term cognitive impairment and functional disability among survivors of severe sepsis. JAMA. 2010;304:1787-1794 Crossref
- Kappos et al., 2010 L. Kappos, E.W. Radue, P. O’Connor, C. Polman, R. Hohlfeld, P. Calabresi, K. Selmaj, C. Agoropoulou, M. Leyk, L. Zhang-Auberson, P. Burtin, F.S. Group. A placebo-controlled trial of oral fingolimod in relapsing multiple sclerosis. New Eng. J. Med.. 2010;362:387-401 Crossref
- Kawana et al., 2013 N. Kawana, Y. Yamamoto, T. Ishida, Y. Saito, H. Konno, K. Arima, J. Satoh. Reactive astrocytes and perivascular macrophages express NLRP3 inflammasome in active demyelinating lesions of multiple sclerosis and necrotic lesions of neuromyelitis optica and cerebral infarction. Clin. Exp. Neuroimmunol.. 2013;4:296-304
- Kivisakk et al., 2009 P. Kivisakk, J. Imitola, S. Rasmussen, W. Elyaman, B. Zhu, R.M. Ransohoff, S.J. Khoury. Localizing central nervous system immune surveillance: meningeal antigen-presenting cells activate T cells during experimental autoimmune encephalomyelitis. Ann. Neurol.. 2009;65:457-469 Crossref
- Kroll and Wolfl, 2002 T.C. Kroll, S. Wolfl. Ranking: a closer look on globalisation methods for normalisation of gene expression arrays. Nucleic Acids Res.. 2002;30:e50
- Lazeron et al., 2000 R.H. Lazeron, D.W. Langdon, M. Filippi, J.H. van Waesberghe, V.L. Stevenson, J.B. Boringa, D. Origgi, A.J. Thompson, M. Falautano, C.H. Polman, F. Barkhof. Neuropsychological impairment in multiple sclerosis patients: the role of (juxta) cortical lesion on FLAIR. Mult. Scler.. 2000;6:280-285
- Lindberg et al., 2004 R.L. Lindberg, C.J. De Groot, U. Certa, R. Ravid, F. Hoffmann, L. Kappos, D. Leppert. Multiple sclerosis as a generalized CNS disease–comparative microarray analysis of normal appearing white matter and lesions in secondary progressive MS. J. Neuroimmunol.. 2004;152:154-167 Crossref
- Magistretti et al., 1999 P.J. Magistretti, L. Pellerin, D.L. Rothman, R.G. Shulman. Energy on demand. Science. 1999;283:496-497 Crossref
- Magliozzi et al., 2010 R. Magliozzi, O.W. Howell, C. Reeves, F. Roncaroli, R. Nicholas, B. Serafini, F. Aloisi, R. Reynolds. A Gradient of neuronal loss and meningeal inflammation in multiple sclerosis. Ann. Neurol.. 2010;68:477-493 Crossref
- Markoullis et al., 2014 K. Markoullis, I. Sargiannidou, N. Schiza, F. Roncaroli, R. Reynolds, K.A. Kleopa. Oligodendrocyte gap junction loss and disconnection from reactive astrocytes in multiple sclerosis gray matter. J. Neuropathol. Exp. Neurol.. 2014;73:865-879
- Martinon et al., 2009 F. Martinon, A. Mayor, J. Tschopp. The inflammasomes: guardians of the body. Annu. Rev. Immunol.. 2009;27:229-265 Crossref
- Milton et al., 1997 I.D. Milton, S.J. Banner, P.G. Ince, N.H. Piggott, A.E. Fray, N. Thatcher, C.H.W. Horne, P.J. Shaw. Expression of the glial glutamate transporter EAAT2 in the human CNS: an immunohistochemical study. Mol. Brain Res.. 1997;52:17-31 Crossref
- Morgen et al., 2006 K. Morgen, G. Sammer, S.M. Courtney, T. Wolters, H. Melchior, C.R. Blecker, P. Oschmann, M. Kaps, D. Vaitl. Evidence for a direct association between cortical atrophy and cognitive impairment in relapsing-remitting MS. Neuroimage. 2006;30:891-898 Crossref
- Newman et al., 2011 L.A. Newman, D.L. Korol, P.E. Gold. Lactate produced by glycogenolysis in astrocytes regulates memory processing. PLoS ONE. 2011;6:e28427
- Niepel et al., 2006 G. Niepel, Ch.R. Tench, P.S. Morgan, N. Evangelou, D.P. Auer, C.S. Constantinescu. Deep gray matter and fatigue in MS: a T1 relaxation time study. J. Neurol.. 2006;253:896-902 Crossref
- Nijland et al., 2014 P.G. Nijland, I. Michailidou, M.E. Witte, M.R. Mizee, S.M. van der Pol, B. van Het Hof, A. Reijerkerk, L. Pellerin, P. van der Valk, H.E. de Vries, J. van Horssen. Cellular distribution of glucose and monocarboxylate transporters in human brain white matter and multiple sclerosis lesions. Glia. 2014;62:1125-1141 Crossref
- Pellerin et al., 2005 L. Pellerin, L.H. Bergersen, A.P. Halestrap, K. Pierre. Cellular and subcellular distribution of monocarboxylate transporters in cultured brain cells and in the adult brain. J. Neurosci. Res.. 2005;79:55-64 Crossref
- Pellerin et al., 2007 L. Pellerin, A.K. Bouzier-Sore, A. Aubert, S. Serres, M. Merle, R. Costalat, P.J. Magistretti. Activity-dependent regulation of energy metabolism by astrocytes: an update. Glia. 2007;55:1251-1262 Crossref
- Pellerin and Magistretti, 1994 L. Pellerin, P.J. Magistretti. Glutamate uptake into astrocytes stimulates aerobic glycolysis: a mechanism coupling neuronal activity to glucose utilization. Proc. Natl. Acad. Sci. U.S.A.. 1994;91:10625-10629 Crossref
- Pellerin and Magistretti, 2012 L. Pellerin, P.J. Magistretti. Sweet sixteen for ANLS. J. Cereb. Blood Flow Metab.. 2012;32:1152-1166 Crossref
- Pellerin et al., 1998 L. Pellerin, G. Pellegri, P.G. Bittar, Y. Charnay, C. Bouras, J.L. Martin, N. Stella, P.J. Magistretti. Evidence supporting the existence of an activity-dependent astrocyte-neuron lactate shuttle. Dev. Neurosci.. 1998;20:291-299 Crossref
- Pierre and Pellerin, 2005 K. Pierre, L. Pellerin. Monocarboxylate transporters in the central nervous system: distribution, regulation and function. J. Neurochem.. 2005;94:1-14 Crossref
- Pirko et al., 2007 I. Pirko, C.F. Lucchinetti, S. Sriram, R. Bakshi. Gray matter involvement in multiple sclerosis. Neurology. 2007;68:634-642 Crossref
- Poletti et al., 2014 S. Poletti, D. Radaelli, M. Bosia, M. Buonocore, A. Pirovano, C. Lorenzi, R. Cavallaro, E. Smeraldi, F. Benedetti. Effect of glutamate transporter EAAT2 gene variants and gray matter deficits on working memory in schizophrenia. Eur. Psychiatry. 2014;29:219-225 Crossref
- Reynolds et al., 2011 R. Reynolds, F. Roncaroli, R. Nicholas, B. Radotra, D. Gveric, O. Howell. The neuropathological basis of clinical progression in multiple sclerosis. Acta Neuropathol.. 2011;122:155-170 Crossref
- Roelcke et al., 1997 U. Roelcke, L. Kappos, J. Lechner-Scott, H. Brunnschweiler, S. Huber, W. Ammann, A. Plohmann, S. Dellas, R.P. Maguire, J. Missimer, E.W. Radu, A. Steck, K.L. Leenders. Reduced glucose metabolism in the frontal cortex and basal ganglia of multiple sclerosis patients with fatigue: a 18F-fluorodeoxyglucose positron emission tomography study. Neurology. 1997;48:1566-1571 Crossref
- Ronnback and Hansson, 2004 L. Ronnback, E. Hansson. On the potential role of glutamate transport in mental fatigue. J. Neuroinflammation. 2004;1:22 Crossref
- Rothstein et al., 1992 J.D. Rothstein, L.J. Martin, R.W. Kuncl. Decreased glutamate transport by the brain and spinal cord in amyotrophic lateral sclerosis. New Eng. J. Med.. 1992;326:1464-1468 Crossref
- Rothstein et al., 1995 J.D. Rothstein, M. Van Kammen, A.I. Levey, L.J. Martin, R.W. Kuncl. Selective loss of glial glutamate transporter GLT-1 in amyotrophic lateral sclerosis. Ann. Neurol.. 1995;38:73-84 Crossref
- Rouach et al., 2002 N. Rouach, E. Avignone, W. Meme, A. Koulakoff, L. Venance, F. Blomstrand, C. Giaume. Gap junctions and connexin expression in the normal and pathological central nervous system. Biol. Cell. 2002;94:457-475 Crossref
- Rouach et al., 2008 N. Rouach, A. Koulakoff, V. Abudara, K. Willecke, C. Giaume. Astroglial metabolic networks sustain hippocampal synaptic transmission. Science. 2008;322:1551-1555 Crossref
- Sanfilipo et al., 2006 M.P. Sanfilipo, R.H. Benedict, B. Weinstock-Guttman, R. Bakshi. Gray and white matter brain atrophy and neuropsychological impairment in multiple sclerosis. Neurology. 2006;66:685-692 Crossref
- Schaeren-Wiemers and Gerfin-Moser, 1993 N. Schaeren-Wiemers, A. Gerfin-Moser. A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes. Histochemistry. 1993;100:431-440 Crossref
- Schousboe et al., 2014 A. Schousboe, S. Scafidi, L.K. Bak, H.S. Waagepetersen, M.C. McKenna. Glutamate metabolism in the brain focusing on astrocytes. Adv. Neurobiol.. 2014;11:13-30
- Schroder and Tschopp, 2010 K. Schroder, J. Tschopp. The inflammasomes. Cell. 2010;140:821-832 Crossref
- Seifert et al., 2006 G. Seifert, K. Schilling, C. Steinhauser. Astrocyte dysfunction in neurological disorders: a molecular perspective. Nat. Rev. Neurosci.. 2006;7:194-206 Crossref
- Sofroniew and Vinters, 2010 M.V. Sofroniew, H.V. Vinters. Astrocytes: biology and pathology. Acta Neuropathol.. 2010;119:7-35 Crossref
- Stadelmann et al., 2008 C. Stadelmann, M. Albert, C. Wegner, W. Bruck. Cortical pathology in multiple sclerosis. Curr. Opin. Neurol.. 2008;21:229-234 Crossref
- Stehberg et al., 2012 J. Stehberg, R. Moraga-Amaro, C. Salazar, A. Becerra, C. Echeverria, J.A. Orellana, G. Bultynck, R. Ponsaerts, L. Leybaert, F. Simon, J.C. Saez, M.A. Retamal. Release of gliotransmitters through astroglial connexin 43 hemichannels is necessary for fear memory consolidation in the basolateral amygdala. FASEB J.l. 2012;26:3649-3657 Crossref
- Suzuki et al., 2011 A. Suzuki, A. Stern Sarah, O. Bozdagi, W. Huntley George, H. Walker Ruth, J. Magistretti Pierre, M. Alberini Cristina. Astrocyte-neuron lactate transport is required for long-term memory formation. Cell. 2011;144:810-823 Crossref
- Thornton et al., 2006 P. Thornton, E. Pinteaux, R.M. Gibson, S.M. Allan, N.J. Rothwell. Interleukin-1-induced neurotoxicity is mediated by glia and requires caspase activation and free radical release. J. Neurochem.. 2006;98:258-266 Crossref
- Wishart and Sharpe, 1997 H. Wishart, D. Sharpe. Neuropsychological aspects of multiple sclerosis: a quantitative review. J. Clin. Exp. Neuropsychol.. 1997;19:810-824 Crossref
- Yiangou et al., 2006 Y. Yiangou, P. Facer, P. Durrenberger, I.P. Chessell, A. Naylor, C. Bountra, R.R. Banati, P. Anand. COX-2, CB2 and P2X7-immunoreactivities are increased in activated microglial cells/macrophages of multiple sclerosis and amyotrophic lateral sclerosis spinal cord. BMC Neurol.. 2006;6:12 Crossref
- Zeis et al., 2008 T. Zeis, U. Graumann, R. Reynolds, N. Schaeren-Wiemers. Normal-appearing white matter in multiple sclerosis is in a subtle balance between inflammation and neuroprotection. Brain. 2008;131:288-303
- Zeis et al., 2009 T. Zeis, A. Probst, A.J. Steck, C. Stadelmann, W. Bruck, N. Schaeren-Wiemers. Molecular changes in white matter adjacent to an active demyelinating lesion in early multiple sclerosis. Brain Pathol.. 2009;19:459-466 Crossref
- Zhou et al., 2011 R. Zhou, A.S. Yazdi, P. Menu, J. Tschopp. A role for mitochondria in NLRP3 inflammasome activation. Nature. 2011;469:221-225 Crossref
a Neurobiology, Department of Biomedicine, University Hospital Basel, University of Basel, Hebelstrasse 20, CH-4031 Basel, Switzerland
b Laboratory of Neuroenergetics and Cellular Dynamics, Brain Mind Institute, Ecole Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland
c Department of Biochemistry, University of Lausanne, CH-1066 Epalinges, Switzerland
d Centre de Neurosciences Psychiatriques, CHUV, Département de Psychiatrie, Site de Cery, CH-1008 Prilly/Lausanne, Switzerland
e Division of Biological and Environmental Sciences and Engineering, KAUST, Thuwal, Saudi Arabia
f Institute for Molecular Bioscience, The University of Queensland, St Lucia 4072, Australia
1 Deceased March 22, 2011.
© 2015 Published by Elsevier B.V.