doi:10.1093/brain/awu358 BRAIN 2015: 138; 398–413 | 398

Pivotal role of choline metabolites in

1, 2,3, 1 1 Thomas Skripuletz, * Arndt Manzel, * Karoline Gropengießer, Nora Scha¨fer, Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 Viktoria Gudi,1 Vikramjeet Singh,1,4 Laura Salinas Tejedor,1,4 Stefanie Jo¨rg,3 Anna Hammer,3 Elke Voss,1 Franca Vulinovic,1 Diane Degen,1 Rebecca Wolf,3 De-Hyung Lee,3 Refik Pul,1 Darius Moharregh-Khiabani,1 Wolfgang Baumga¨rtner,4,5 Ralf Gold,6 Ralf A. Linker3 and Martin Stangel1,4

*These authors contributed equally to this work.

Neuroprotective approaches for regeneration have not been successful in clinical practice so far and compounds that enhance remyelination are still not available for patients with . The objective of this study was to determine potential regenerative effects of the substance cytidine-50-diphospho (CDP)-choline in two different murine animal models of multiple sclerosis. The effects of exogenously applied CDP-choline were tested in murine glycoprotein-induced experimental autoimmune encephalomyelitis. In addition, the cuprizone-induced mouse model of de- and remyelination was used to specifically test the hypothesis that CDP-choline directly increases remyelination. We found that CDP- choline ameliorated the course of experimental autoimmune encephalomyelitis and exerted beneficial effects on myelin, and . After cuprizone-induced demyelination, CDP-choline effectively enhanced myelin regeneration and reversed motor coordination deficits. The increased remyelination arose from an increase in the numbers of proliferating oligo- dendrocyte precursor cells and oligodendrocytes. Further in vitro studies suggest that this process is regulated by kinase C. We thus identified a new mechanism to enhance central nervous system remyelination via the choline pathway. Due to its regen- erative action combined with an excellent safety profile, CDP-choline could become a promising substance for patients with multiple sclerosis as an add-on therapy.

1 Department of Neurology, Hannover Medical School, 30625 Hannover, Germany 2 Ruhr-University Bochum, International Graduate School of Neuroscience, 44801 Bochum, Germany 3 Department of Neurology, University Hospital Erlangen, 91054 Erlangen, Germany 4 Centre for Systems Neuroscience, 30559 Hannover, Germany 5 Department of Pathology, University of Veterinary Medicine Hannover, 30559 Hannover, Germany 6 Department of Neurology, St Josef Hospital, Ruhr-University Bochum, 44791 Bochum, Germany Correspondence to: Prof. Dr. Martin Stangel, Department of Neurology, Hannover Medical School, Carl-Neuberg-Str-1, 30625 Hanover, Germany E-mail: [email protected] Keywords: remyelination; demyelination; oligodendrocyte; EAE; neuroinflammation Abbreviations: APC = anti-adenomatus polyposis coli; EAE = experimental autoimmune encephalomyelitis; OPC = oligodendrocyte precursor cell; PLP = proteolipid protein

Received April 24, 2014. Revised October 21, 2014. Accepted October 21, 2014. Advance Access publication December 18, 2014 ß The Author (2014). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected] CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 399

EAE is an animal model that resembles some immuno- Introduction pathological and neurobiological aspects of multiple scler- Multiple sclerosis is characterized by inflammation, demye- osis (Linker et al., 2002). Although EAE is an excellent lination, gliosis and axonal damage in the CNS (Compston model to study CNS inflammation and , and Coles, 2008). Remyelination is the natural repair assessment of remyelination may be complicated in a set- mechanism of demyelination and can be a highly efficient ting with concommittant demyelination and remyelina- process, but it may be incomplete or even fail in multiple tion. In this context, the findings in EAE should be sclerosis (Franklin and ffrench-Constant, 2008). Possible confirmed and supplemented in another model of demye- explanations for remyelination failure of demyelinated lination, preferentially in a toxic model of demyelination. intact axons include defects in recruitment/proliferation of Thus, we additionally employed the cuprizone-induced new oligodendrocyte precursor cells (OPCs) and/or differ- mouse model of de- and remyelination (Skripuletz et al., entiation into mature myelin producing oligodendrocytes 2013) to specifically test the hypothesis that CDP-choline Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 (Kotter et al., 2011; Hagemeier et al., 2012; de Castro directly increases remyelination. The cuprizone model is et al., 2013). Demyelinated axons are vulnerable to degen- widely accepted to study de- and remyelination processes eration and it was proposed that remyelination protects in the CNS. In contrast to models of focal-induced demye- against progressive axonal damage and thus long-term dis- lination in the cuprizone model no breakdown of the ability in patients with multiple sclerosis (Franklin and blood–brain barrier occurs, which allows us to analyse ffrench-Constant, 2008; Kotter et al., 2011). For this the pathomechanisms of remyelination bypassing possible reason, therapeutic promotion of remyelination represents interferences of peripheral immune cells (Gudi et al., an attractive option for preventing axonal loss and disease 2014). Besides, we have used different treatment protocols progression. Although many factors involved in remyelina- in both models. tion have been identified in animal models of multiple scler- osis, therapeutic approaches to increase the numbers of OPCs and oligodendrocytes to support myelin repair are Materials and methods still not available in the human scenario. Recently, CDP-choline (citicoline, cytidine 50-diphospho- EAE induction and CDP-choline choline) has been proposed to have neuroprotective and regenerative functions (Hurtado et al., 2011). CDP-cho- treatment protocol line, a naturally occurring endogenous nucleoside, is an For induction of EAE, mice received subcutaneous injections at 35–55 intermediate compound in the major pathway for the the tail base of 200 mg MOG in phosphate-buffered saline biosynthesis of the phospholipid phosphat- (PBS) emulsified in an equal volume of complete Freund’s ad- idylcholine (lecithin) (Kent and Carman, 1999; Adibhatla juvant containing Mycobacterium tuberculosis H37RA (Difco) at a final concentration of 1 mg/ml. Intraperitoneal injections et al., 2006). Endogenous formation of CDP-choline from of 200 ng pertussis toxin (List Biochemicals) were given at the choline is a limiting step in the synthesis of phosphatidyl- time of immunization and 48 h later. choline (Weiss, 1995). In the CNS choline is preferentially Mice either received CDP-choline (500 mg/kg) or saline used for the synthesis of acetylcholine, which may limit the (control group) by daily oral gavage beginning on the day of amount of choline for phosphatidylcholine production immunization until termination of the experiment. In add- (Conant and Schauss, 2004). Exogenous CDP-choline dis- itional experiments, CDP-choline (500 mg/kg) or saline in con- perses in the organism mainly in the form of its degrad- trol mice was given in therapeutic regimens with treatment ation products cytidine and choline that readily cross the starting at the onset of disease, corresponding to Day 7 after blood–brain barrier and enter the various biosynthetic MOG-immunization and after the first week of disease, cor- pathways that use CDP-choline (Conant and Schauss, responding to Day 15 after immunization. Mice were weighed and scored for signs of disease on a daily 2004). In various animal models of CNS injury and neu- basis. Disease severity was assessed using a scale ranging from rodegenerative CNS , exogenously administered 0 to 5 (Kleinewietfeld et al., 2013): 0, normal; 1, limp tail; 2, CDP-choline enhanced membrane repair and neuronal gait ataxia; 3, paraparesis; 4, tetraparesis; 5, death. functions (Adibhatla and Hatcher, 2005). Yet, the poten- tial beneficial effects of CDP-choline on de- and remyelina- tion in animal models for multiple sclerosis were not Cuprizone-induced demyelination evaluated to date. and CDP-choline treatments Here we investigated the effects of exogenously applied Experimental toxic demyelination was induced by feeding mice CDP-choline in murine myelin oligodendrocyte glycopro- a diet containing 0.2% cuprizone (bis-cyclohexanone oxaldi- tein (MOG)-induced experimental autoimmune encephalo- hydrazone, Sigma-Aldrich Inc) mixed into a ground standard myelitis (EAE) and in the cuprizone model of toxic rodent chow (Skripuletz et al., 2011). The cuprizone diet was induced dymyelination. The aim of our experiments was maintained for 5 weeks. Animals were treated with CDP- to use different animal models that each mimic some as- choline (500 mg/kg) or sham every day beginning on the day pects of the complex mechanisms of multiple sclerosis. of cuprizone feeding until termination of the experiment. 400 | BRAIN 2015: 138; 398–413 T. Skripuletz et al.

Additional control animals were fed with normal chow and magnification within the margins of a 1/16 mm2 grid in indi- received CDP-choline or sham as described above. vidual lesions and counts were normalized to cells/mm2. For To analyse the effects of short CDP-choline treatments on further quantification of inflammation and demyelination, digi- remyelination additional five treatment protocols were applied. tal frames were acquired. Affected areas within a section and Mice were treated with CDP-choline (500 mg/kg) or sham total area of that section were determined using ImageJ (NIH). every day between Weeks 3–4, Weeks 3–5.5, Weeks 4–5, After Mac-3 or haematoxylin and eosin staining, infiltration Weeks 4–5.5 and Weeks 5–5.5. was additionally assessed as infiltrated area in mm2 per mouse Motor coordination was assessed by the beam walking test and numbers of Mac-3 positive infiltrates per section are given (Skripuletz et al., 2010). Mice were measured for the ability to as inflammatory index (Eugster et al., 1999). Demyelination is traverse a narrow beam to reach an enclosed safety box. Mice given as per cent affected area per total area. received two consecutive trials and the latency to traverse the In cuprizone-treated animals, brain sections between bregma beam was recorded for each trial (cut-off time 60 s). In the 0.82 mm and bregma 1.82 mm [according to the mouse results the mean score of the two trials is given. atlas by Paxinos and Franklin (2001)] were analysed.

Quantification of glial cells was performed for mature oligo- Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 dendrocytes (Nogo-A, APC), proliferating OPC (double stain- Histology, immunohistochemistry ing with Olig-2/Ki-67, NG2/Ki-67), activated microglia (RCA-1, and electron microscopy Mac-3), proliferating microglia (double staining with RCA-1/ Ki-67), and astrocytes (GFAP). Immunopositive cells with Histology, immunohistochemistry and electron microscopy identified nucleus (counterstaining with Mayer’s hemalum so- were performed as previously described (Skripuletz et al., lution for immunohistochemistry and DAPI for immunofluor- 2013). Mice were perfused with 4% paraformaldehyde (PFA) escence) were counted in the central part of the corpus and spinal cord and spleen (in EAE experiments) or brains callosum in an area of at least 0.185 mm2 using a magnifica- (in cuprizone experiments) were removed, postfixed in 4% tion of 400 (Olympus BX61) as previously described PFA and paraffin embedded. Spinal cord sections were sub- (Skripuletz et al., 2013). In results counted cells are expressed jected to haematoxylin and eosin stain, Bielschowsky’s silver as number of cells per mm2. impregnation and immunohistochemical stainings. Brain sec- The extent of demyelination in the corpus callosum was tions underwent Luxol Fast blue stain and immunohistochem- analysed by light microscopy (Olympus BX61) as described ical stainings. For immunohistochemistry the following previously (Skripuletz et al., 2008). For determination of primary antibodies were used: for myelin proteolipid demyelination in the corpus callosum myelin proteolipid pro- protein (PLP) (mouse IgG, 1:500, Serotec), myelin basic pro- tein, MBP, MOG, CNPase, and Luxol Fast blue stained brain 0 tein (MBP) (mouse IgG, 1:500, Covance), cyclic nucleotide 3 sections were scored in a blinded manner by three observers phosphodiesterase (CNPase) (mouse IgG, 1:200, Millipore), and graded on a scale from 0 (complete demyelination) to 3 and myelin oligodendrocyte glycoprotein (MOG) (1:2 hybri- (normal myelin). doma supernatant, generous gift from C. Linington), for acti- vated microglia Mac-3 (rat IgG, 1:500, BD Pharmingen) and the lectin ricinus communis agglutinin 1 (RCA-1) (1:1000, RNA isolation and real-time biotinylated, Vector Laboratories), for astrocytes glial fibrillary quantitative reverse acidic protein (GFAP) (mouse IgG, 1:200, Millipore), for oligo- dendrocytes Nogo-A (rabbit IgG, 1:750, Millipore) and anti- transcriptase PCR adenomatus polyposis coli (APC; mouse IgG; 1:200, For PCR analysis corpus callosum was dissected from whole Calbiochem), for OPC Olig-2 (rabbit IgG, 1:500, Millipore), brains under a light microscope. Total RNA was then ex- NG2 (1:200, rabbit IgG, Millipore), and PDGFR (goat IgG, tracted from the tissue using the RNeasyÕMini Kit (Qiagen) 1:500, R&D Systems), for T cells CD3 (rat IgG1, 1:200, as previously described (Gudi et al., 2011). expressions Serotec), for neurites the neurofilament antibodies SMI31 of insulin-like growth factor 1 (Igf1), hepatocyte growth factor and SMI32 (mouse IgG, 1:1000, Covance), and for prolifer- (Hgf), fibroblast growth factor (Fgf2), leukaemia inhibitory ation Ki-67 (mouse IgG, 1:50, BD Pharmingen). factor (Lif), ciliary neurotrophic factor (Cntf), brain-derived neurotrophic factor (Bdnf) and platelet-derived growth factor alpha (Pdgf ) were analysed in the corpus callosum after 4.5, Quantification of glial reactions and 5 and 5.5 weeks of cuprizone fed mice treated with CDP- determination of de- and choline or sham and in age-matched normal fed control mice. The ÁÁCt method was used to determine differences remyelination in the expression between cuprizone fed mice and age-matched In EAE spinal cord cross-sections (cervical, thoracic and control mice. Changes in mRNA expression levels were calcu- lumbar) were analysed using a light microscope (Olympus lated after normalization to hypoxanthin phosphoribosyltrans- BX51) by a blinded observer as described previously ferase (Hprt1) and glyceraldehyde 3-phosphate dehydrogenase (Herrero-Herranz et al., 2008). For each staining, four lesions (Gapdh). in two sections within each spinal cord segment were analysed. Silver impregnated sections were analysed at 630 magnifica- Flow cytometry and ELISA tion using a 25-point eyepiece. Axonal densities were counted within lesions and were given as number of grid points cross- Spleens and CNS (brain/spinal cord) of CDP-choline treated ing axons as described previously (Linker et al., 2011). T cells and control mice were removed on Day 9–10 post-EAE induc- and macrophages/microglia were counted at 400 tion and mechanically dissociated into single cell suspensions. CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 401

CNS mononuclear cells were isolated by percoll gradient cen- last 3 h. Cells were then washed in PBS and stainining for trifugation as described (Linker et al., 2010). Splenocytes and A2B5 and BrdU was performed as described previously CNS mononuclear cells were analysed by multicolour flow (Kotsiari et al., 2010). The labelled cells were analysed using cytometry using fluorochrome conjugated surface marker anti- a fluorescent microscope and the proliferation index was cal- bodies. Anti-CD3 (clone 17A2), anti-CD4 (clone RM4-5), anti- culated by dividing number of BrdU-positive cells with total CD8a (clone 53-6.7), anti-CD11b (clone M1/70), anti-CD19 number of A2B5 positive cells. A minimum of 300 cells was (clone 1D3), anti-CD16/32 (Fc-block, clone 2.4G2) anti-CD80 counted in six to eight fields. In addition, OPC proliferation (16-10A1), anti-MHCII (clone AF6-120.1), anti-Ly6c was analysed after staining with the proliferation marker Ki- (clone AL-21), and anti-NK1.1 (clone PK 136) were all pur- 67. After isolation, OPCs were plated at 8 104 cells per chased from BD. Anti-CD62L (clone MEL-14), anti-CD44 12 mm coverslip and maintained at 37 C, 5% CO2 for 4 (clone IM7), anti-CD45 (clone 30-F11), anti-PDCA-1 days. Every second day fresh medium was added to the cells. (clone 129C1), anti-B220 (clone RA3-6B2), and anti-F4/80 Further cells were treated with CDP-choline for 24 h and were (clone BM8) were purchased from Biolegend. Flow cytometry stained for the cell proliferation marker Ki-67. Briefly, after was performed on a FACS Canto II (BD) and results were incubation cells were washed in PBS and were incubated Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 analysed using FACSDiva software (BD). with A2B5 antibody solution for 1 h. After washing twice in For the measurement of cytokines in mouse spleen cell pri- PBS cells were fixed with 4% PFA. Permeabilization of cells mary cultures, splenocytes were prepared and cultured as was achieved by adding ice-cold methanol for 20 min at 4C. above with media alone or in the presence of 20 mg/ml of Cells were incubated with Ki-67 antibody (BD biosciences) for MOG35–55 peptide or 1.25 mg/ml ConA. Supernatants were 90 min followed by washing in PBS. Further, cells were incu- harvested after 48 h. IFNG, IL17, or IL10 concentrations bated with fluorescent labelled goat anti-mouse IgM 555 and were determined by ELISA, as described (Crowther, 1995). goat anti-mouse IgG1 488 antibody (1:500, Invitrogen) for For the measurement of corticosterone, serum was collected 90 min. Thereafter, cells were washed three times in PBS and on Day 10 post-EAE induction. Concentrations were deter- deionized water and mounted with MowiolÕ (Calbiochem) so- mined by ELISA, as described by the manufacturer (Abnova). lution containing nuclear stain 40,6-diamidino-2-phenylindole (DAPI, Invitrogen). Cells were examined under fluorescent T cell proliferation assay microscope and a minimum 300 cells was counted in six visual fields of coverslip. Proliferation index was calculated Single cell suspensions from spleens of CDP-choline treated by dividing number of A2B5 and Ki-67 double positive cells and control mice were prepared on Day 9 after EAE induction with total of A2B5 positive cells. (Linker et al., 2002). Cells (2 105) were seeded in 96-well microtitre plates (Nunc) in 100 ml medium with addition of antigen or mitogen. Concentrations were 20 mg/ml for MOG, Oligodendrocyte precursor cell and 1.25 mg/ml for Concanavalin A (ConA; Sigma-Aldrich). differentiation and migration assays Triplicate cultures were maintained at 37C in a humidified atmosphere with 5% CO2 for 56 h and harvested following Differentiation of OPCs was determined by double staining for a 16 h pulse with 0.2 mCi/ well 3H-dT (tritiated thymidine, the OPC marker A2B5 and the oligodendrocyte marker GalC Amersham-Buchler). Cells were collected on fibreglass filter as described previously (Kotsiari et al., 2010). OPCs were paper with a 96-well harvester (Pharmacia), and radioactivity plated at a density of 7 104 cells per poly-L-lysine coated was measured with a 96-well Betaplate liquid scintillation glass coverslip for 1 h. Non-adherent cells were removed and counter (Pharmacia). cells were left overnight. To allow differentiation, cells were cultured in N2B3 medium, N2B3 medium with 0.1 mM CDP- choline or N2B3 medium with 1 mM CDP-choline. After 48 h, Oligodendrocyte precursor cell cells were double-stained for A2B5 (1:2) and GalC culture (Galactocerebroside, 1:2, hybridoma supernatant, clone IC-07, European collection of cell cultures). Both antibodies Primary cultures of OPC were prepared from neonatal were applied together on living cells for 40 min at 37C. Sprague-Dawley rat brains as previously described (Kotsiari Cells were fixed with 4% PFA and incubated with secondary et al., 2010). OPC culture purity was analysed by staining antibodies Alexa FluorÕ 555 goat anti-mouse IgM (1:500) for with A2B5, a marker for OPC and was between 75–85% as A2B5 and Alexa FluorÕ 488 goat anti-mouse IgG3 (1:500) for quantified by fluorescence microscopy. For western blotting GalC. Positive cells were counted in 10 random visual fields experiments A2B5 positive OPC were purified by anti-A2B5 per coverslip using a magnification of 200. The differenti- microbeads as described by the manufacturer (Miltenyi ation index represents the ratio of mature oligodendrocytes biotech). (GalC positive) to OPC (A2B5 positive). Results were normal- ized to the control condition. Oligodendrocyte precursor cell Migration of OPCs was performed through a poly-L-lysine coated membrane with the use of a Boyden chamber. Three proliferation assays different conditions were studied: B104 medium, B104 OPC proliferation was measured by the detection of medium with 0.1 mM CDP-choline and B104 medium with 5-bromo-20-deoxyuridine (BrdU) (Roche diagnostics GmbH) 1 mM CDP-choline. On the lower wells 28 ml of each condition incorporation into the DNA. OPCs were treated with different was added and at least three replicates per condition were concentrations of CDP-choline (0.01–1 mM) together with performed. A total amount of 8 104 cells resuspended in 10 mM of BrdU for 24 h, with BrdU being present for the 50 ml of one of the three conditions was applied to the 402 | BRAIN 2015: 138; 398–413 T. Skripuletz et al. corresponding upper wells. Afterwards, the cells were incu- presence of CDP-choline (0.01–1 mM) and culture super- bated for 16 h at 37C. After the migration time, the cells natants harvested and stored at 80C for cytokine ELISA. were fixed and stained with the Diff-Quik Staining Kit (dade Supernatants were processed immediately for nitric oxide Behring) and migrating cells were counted at a magnification measurements (Griess reagent kit; Promega). Phagocytosis of 200. The migration index represents the ratio of migrat- was measured by stimulating macrophages with 100 ng/ml ing OPCs in each condition to migrating OPC in B104 lipopolysaccharide in the presence or absence of CDP-choline medium. Results were normalized to the control condition. (0.01–1 mM) for 18 h before addition of a 50-fold excess of opsonized FluoresbriteÕ carboxylate YG 0.75 Micron Microspheres (Polysciences, Inc.). After 1 h, 3 h and 24 h of Western blotting incubation with fluorescent beads, cells were subjected to Western blotting was carried out as described previously flow cytometry and phagocytic macrophages were detected in (de Oliveira et al., 2008). Briefly, cells were washed with the FITC channel. As a negative control, cells were analysed cold PBS and lysed in lysis buffer (42 mM Tris-HCl, 1.3% immediately after addition of beads. Migration of macro- SDS, 6.5% glycerin and 100 mM sodium orthovanadate, pro- phages was measured in a modified Boyden chamber assay Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 tease and phosphatase inhibitors). Before electrophoresis, using Transwell inserts with a 5 mm porous bottom Laemmli buffer (5% mercaptoethanol, 10% glycerol, 2% (Corning). Macrophages were loaded onto the migration SDS, 65 mM Tris HCl and bromophenol blue) was added to chamber in RPMI containing 0.5% FCS. The lower chamber the samples. Samples were separated by SDS-PAGE and trans- was loaded with RPMI containing 10% FCS. In the test con- ferred onto a polyvinylidene fluoride (PVDF) membrane dition, both compartments contained 1 mM CDP-choline. In (Millipore). The membrane was incubated with rabbit anti- the negative control condition, both compartments contained phospho-PKC (pan) (1 mg/ml; Cell Signaling) and mouse RPMI-1640 supplemented with 0.5% FCS. After 14 h of mi- anti-b- (1:3000; Santa Cruz biotechnology) antibodies. gration, cells remaining in the upper compartment were For detection horseradish peroxidase (HRP)-conjugated removed with a cotton swab and cells adhering to the trans- secondary antibodies goat anti-rabbit IgG (1:3000; R&D surface were removed by accutase treatment (adhering frac- systems) or goat anti-mouse IgG (1:5000; R&D Systems) tion). Adhering fraction and cells contained in the lower cham- were used followed by chemiluminescence (ECL) reagents ber (transmigrating fraction) were stained for macrophage (GE Healthcare). markers F4/80 and CD11b and quantified by flow cytometry.

Microglia cell culture and functional Macrophage polarization assays Bone marrow cells were obtained from tibias and femurs of 10–20 week old C57BL/6 mice as described previously Microglia were prepared from neonatal Sprague-Dawley rat (Muccioli et al., 2011) and maintained at a density of brains as previously described (Kotsiari et al., 2010). The 2 106 cells/cm2 in RPMI-1640 supplemented with 50 mMß- effect of CDP-choline on the release of nitrite, TNF and mercaptoethanol and 5% Panexin BMM (PanBiotech) and IL10 in lipopolysaccharide (Escherichia coli; Sigma Aldrich) 10 ng/ml recombinant mouse M-CSF. After 10 days, bone activated microglia was determined. The amount of nitrite marrow-derived macrophages were further cultured in the was measured with Griess reagent as described (Singh et al., presence or absence of 1 mM CDP-choline under polarizing 2012). TNFA and IL10 were measured in cell culture super- conditions as follows; M0, no cytokines; M1, 10 ng/ml LPS/ natants with enzyme immunoassay kits as described by the 10 ng/ml IFN-G; M2, IL4/IL13 20 ng/ml (recombinant cyto- manufacturer (BD biosciences). kines were from R&D systems). After 24 h of polarization, For generation of CDP-choline treated microglia super- expression of the M1 marker Nos2/iNos and Tnfa as natants, microglia were treated with different concentrations well as M2 marker genes Arg1/Arginase1 and Chi3l3/YM-1 of CDP-choline (0.01–0.1 mM) in serum free Dulbecco’s mod- was analysed by real-time PCR using pre-developed ified Eagle’s medium (4.5 mg/ml glucose, Gibco) containing TaqManÕ primer/probe assays (Life technologies). Expression 1% ITS supplement (BD biosciences) for 12 h. Further, cells of Actb served as internal control. were washed and incubated with fresh medium for 12 h. Cell supernatants were collected, centrifuged and stored at 80C until used. The effect of microglia supernatants on OPC pro- Statistical analysis liferation was examined by Ki-67 staining. Statistical analysis was performed using ANOVA followed by the Fisher-protected least-significant difference test for post hoc Macrophage cell culture and comparison if appropriate. EAE experiments were analysed functional assays using the non-parametric Mann-Whitney test. All data are given as arithmetic means standard error of the mean Primary macrophages were obtained from 10–20 week old (SEM). EAE data are additionally presented as median inter- C57BL/6 mice by peritoneal lavage and maintained at densities quartile ranges. For EAE data, a power calculation was per- of 5 105 to 2 106 cells/cm2 (depending on assay) in com- formed (G*Power Freeware, Version 3.1.5). To detect a one plete RPMI-1640 containing 10% heat inactivated foetal calf point difference on a 1–5 point scale with an alpha error prob- serum (FCS) at 37C under 5% CO2 atmosphere and 100% ability of 50.05 and a beta of 40.95 and standard deviations relative humidity. For cytokine and reactive oxygen species representing up to 33% of the mean, the minimal number measurements, cells were stimulated with 100 ng/ml lipopoly- needed to analyse is n = 8 per group resulting in an actual saccharide (LPS; Sigma-Aldrich) for 6–48 h with or without the power of 0.951. P-values are given in the results while group CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 403 comparisons derived from post hoc analysis are provided in (Supplementary Fig. 1C and G). After haematoxylin and the figures (*P 5 0.05; **P 5 0.01; ***P 5 0.001). eosin and Mac-3 staining, there was no difference in infil- trating macrophages/microglia between both groups on Days 15 and 27 post-immunization, neither regarding cell Results densities, number of infiltrates, nor the extent of infiltrated area (Supplementary Fig. 1D and H and data not shown). Beneficial effects of CDP-choline To closely investigate potential effects of CDP-choline on in EAE the immune system, including T cell responses, we per- formed multicolour flow cytometry of splenic leukocytes We investigated the therapeutic potential of CDP-choline in shortly after the onset of disease (Day 10 post-immuniza- MOG-induced EAE in C57BL/6 mice. First, CDP-choline tion). CDP-choline did not impact on the abundance of was preventively applied and mice either received CDP-cho- splenic CD3 positive T cells or frequencies of CD4 positive line (treated group) or saline (EAE-control group) every day helper T cells and CD8 positive cytotoxic T cells Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 beginning on the day of immunization until the end of (Supplementary Fig. 2A). Moreover, T cell-bound expres- the observation period. CDP-choline was highly potent sion of the activation markers CD62L and CD44 was to ameliorate MOG-EAE (Fig. 1A and Supplementary identical in treated and EAE-control mice (Supplementary Fig. 1A). Over the course of the disease mice in the EAE- Fig. 2B and C). CDP-choline did not affect frequencies of control group suffered from mild paraparesis, while treated CD19 positive B cells, NK1.1 positive NK cells, animals only displayed a limp tail (P 5 0.01, on Day 27 and NK1.1/CD3 positive NK-T cells (Supplementary Fig. post-immunization). The large difference between groups in 2D–F). Upon antigen-specific restimulation of splenocytes disease severity persisted until the end of the observation from CDP-choline treated and EAE-control mice in vitro, period. Next, we tested CDP-choline in therapeutic regi- production of the cytokines IFNG, IL10, and IL17 as mens in vivo, with treatment starting at the onset of dis- well as T cell proliferation did not differ between ease, corresponding to Day 7 after MOG-immunization the CDP-choline treated and EAE-control group (Fig. 1B and Supplementary Fig. 1B) and after the first (Supplementary Fig. 2G–J). Flow cytometry of splenic and week of disease corresponding to Day 15 after immuniza- CNS resident myeloid cells on Day 10 after EAE induction tion. Treatment from the onset of disease was still effective, did not reveal any effects of CDP-choline treatment: there but had a smaller potency (Fig 1B, P 5 0.05 on Day 23 were neither effects on the abundance of CD11b-positive post-immunization). Untreated mice displayed moderate cells, including Ly6c low monocytes, nor on CD11c/ paraparesis whereas CDP-choline treated mice only PDCA-1/B220 positive plasmacytoid dendritic cells showed mild gait ataxia. In contrast, treatment after Day (Supplementary Fig. 3A–C). Splenic CD11b-positive mye- 15 post-immunization was ineffective in ameliorating clin- loid cells did not change their surface expression of ical symptoms of EAE. MHCII and CD80 under CDP-choline treatment To dissect mechanisms of CDP-choline action in EAE, (Supplementary Fig. 3D and E). Likewise, CNS resident spinal cord cross sections from preventively treated and CD11b-positive microglia/macrophages from CDP-choline EAE-control mice were subjected to histopathological ana- treated and control mice expressed the same amounts of lyses. On Day 15 post-immunization, immunohistochemistry CD11b, CD80, and CD45 (Supplementary Fig. 3F–H). revealed higher numbers of Olig-2 positive oligodendroglia Furthermore, serum corticosterone levels were not changed in CDP-choline treated mice (Fig. 1D and E; P 5 0.05) between the CDP-choline treated and EAE-control group whereas numbers of mature Nogo-A positive oligodendro- on Day 10 post-immunization (Supplementary Fig. 2K). cytes were not different (Fig. 1C). As the marker Olig-2 In summary, CDP-choline effectively ameliorated the dis- stains mature oligodendrocytes and OPCs, higher numbers ease course of MOG-EAE especially at earlier phases of the of Olig-2 positive cells in CDP-choline treated mice are disease before onset of irreversible neurodegeneration with- likely caused by expansion of OPCs. Accordingly, on Day out modulation of myeloid cells and T cell responses. The 27 post-immunization, numbers of mature Nogo-A positive beneficial effects of CDP-choline were accompanied by oligodendrocytes were increased in lesioned white matter of higher numbers of oligodendrocytes and myelination CDP-choline treated mice (Fig. 1F and H; P 5 0.05). This which may be the result of an increased OPC generation. was accompanied by a higher degree of myelination as measured by CNPase staining (Fig. 1G and H; P 5 0.05). In addition, axonal densities and axonal integrity were pre- Regenerative effects of CDP-choline served in the CDP-choline treated group as quantified by after cuprizone-induced Bielschowsky’s silver stain (Supplementary Fig. 1E and F; demyelination P 5 0.01) and assessed by SMI31/32 immunostaining (Supplementary Fig. 1E), respectively. Opposed to evident In EAE the beneficial effect could be partly due to T cell changes in the glial and neuronal compartment, no change independent effects on the immune system. We therefore in quantities of CD3-positive CNS infiltrating T lympho- analysed direct effects of CDP-choline on myelin and glia cytes was observed on Days 15 and 27 after EAE induction in the cuprizone model of toxic de- and remyelination 404 | BRAIN 2015: 138; 398–413 T. Skripuletz et al. Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021

Figure 1 Beneficial effects of CDP-choline on MOG-EAE. CDP-choline ameliorated clinical symptoms and histopathology of MOG-EAE. Preventive (A, n = 22 per group) and therapeutic (B, n = 20 control, n = 24 CDP-choline) regimens of CDP-choline treatment were able to exert beneficial effects on the disease course. To dissect mechanisms of CDP-choline action in EAE, preventively treated and control mice were subjected to histopathological analyses. Histopathological analyses, performed on Day 15 after EAE induction, revealed no difference in numbers of Nogo-A positive mature oligodendrocytes (C) between groups, whereas higher numbers of Olig-2 positive cells (D and E) were found after treatment with CDP-choline in EAE lesions (n = 6 per treated group and n = 8 for naı¨ve mice). On Day 27 after immunization, higher numbers of Nogo-A positive oligodendrocytes (F and H, n = 6 per group) and an increased myelinated area (CNPase stain; G and H, n = 6 per group) were found in CDP-choline treated mice. Bars represent mean SEM, *P 5 0.05, **P 5 0.01, ***P 5 0.001. p.i. = post-immunization; NAWM = normal appearing white matter. CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 405 where the peripheral immune system is of no importance in 67. After 4 and 4.5 weeks of cuprizone treatment mature lesion induction (Kipp et al., 2009; Skripuletz et al., 2013). oligodendrocytes in the corpus callosum have almost com- Mice responded to cuprizone treatment as expected, show- pletely vanished (Fig. 3B). Thereafter, oligodendrocytes ing marked demyelination of the corpus callosum during began to reappear. Analogous to myelin re-expression the the 5 weeks of cuprizone feeding (Fig. 2B and C). No dif- density of Nogo-A and APC positive oligodendrocytes was ference was found between treated and control groups significantly higher in mice treated with CDP-choline com- showing that CDP-choline did not influence demyelination pared to cuprizone controls (for both, P 5 0.0001; Fig. 3B– in this model. D). These results show that CDP-choline exerts beneficial Yet, CDP-choline led to profound effects on remyelina- effects on remyelination associated with increased numbers tion: at Week 5.5 (0.5 week of remyelination) mice treated of oligodendrocytes. with CDP-choline already presented high densities of new During the period of oligodendrocyte loss high numbers of myelin proteins throughout the complete corpus callosum. new proliferating OPCs (Olig-2 and Ki-67 double positive For each of the markers studied in immunohistochemical cells) entered the corpus callosum, reaching a peak at Week Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 (PLP, MBP, CNPase, MOG) and histochemical stainings 4–4.5 followed by a decrease. At the peak of proliferation, (LFB Luxol Fast blue), CDP-choline treated mice showed CDP-choline significantly increased the numbers of prolifer- significantly higher amounts of new myelin compared to ating OPCs in the corpus callosum (P 5 0.0001; Fig. 3E the cuprizone controls (for all P 5 0.0001; Fig. 2B–G). At and F). The results were confirmed by the double staining that time point, no or only a low amount of new myelin for NG2/Ki-67 (P = 0.02; Supplementary Fig. 6A and B). was expressed at the edges of the corpus callosum in cupri- This result indicates that the higher numbers of mature oligo- zone control mice (Fig. 2C). At Week 6, strong remyelina- dendrocytes found in CDP-choline treated mice during tion was observed in both cuprizone groups and no remyelination might be a result of higher numbers of prolif- significant difference between the groups was seen any erating OPCs found 1.5 weeks earlier. more. Glial reactions were further analysed by imunohistochem- In a next step, ultrastructual investigations were per- istry for microglia/macrophages (staining for Mac-3 and formed for quantitative analysis of myelinated axons and RCA-1) and astrocytes (staining for GFAP). During demye- confirmed the beneficial effects of CDP-choline on remyeli- lination no CDP-choline treatment effects were found on nation. CDP-choline significantly increased the number of microglia/macrophage numbers (Supplementary Fig. 7B myelinated axons compared to the cuprizone controls and C). When strong remyelination occurred, numbers of (P 5 0.001; Fig. 2H and I). Additionally, CDP-choline microglia were significantly lower in CDP-choline treated induced lower G-ratios of the myelinated axons when com- mice compared to cuprizone controls (for both P 5 0.0001; pared with cuprizone controls indicating a better remyeli- Supplementary Fig. 7B and C). For proliferating microglia nation (P 5 0.001; Fig. 2J). In summary, CDP-choline (RCA-1 and Ki-67 positive cells) and astrogliosis, no CDP- effectively promoted myelin regeneration in the CNS. choline treatment effects were found during the entire de- To confirm the beneficial effects of CDP-choline on and remyelination phase (Supplementary Fig. 7D and E). remyelination, the beam walking test was performed to To identify factors involved in early remyelination, the analyse effects on motor coordination. During demyelin- mRNA expression of selected factors that have been asso- ation, disturbances of motor coordination were seen as ciated with OPC proliferation was analysed in the corpus the time taken by mice to traverse a narrow beam increased callosum. No significant CDP-choline effects were found compared to mice fed with normal chow. No difference for Igf1, Hgf, Fgf2, Lif, Cntf, Bdnf and Pdgfa mRNA ex- was seen between cuprizone groups. Analogous to higher pression after 4.5, 5 and 5.5 weeks of cuprizone feeding regeneration at Week 5.5, mice treated with CDP-choline (Supplementary Fig. 8). presented decreased disturbance of motor coordination when compared to the cuprizone controls (Fig. 2K; CDP-choline enhances remyelination P 5 0.0001). Thus, our results show that CDP-choline not only increased remyelination but also reversed demye- in a time window that dependens on lination-associated motor coordination impairment. OPC proliferation We next tested CDP-choline treatment in different time Impact of CDP-choline on OPC periods to evaluate additional mechanistic effects on oligo- proliferation dendrocytes and remyelination. CDP-choline was given in five different temporal treatment schemes (between Weeks To further evaluate the mechanisms of CDP-choline on 3–4, 3–5.5, 4–5, 4–5.5 and 5–5.5) during cuprizone treat- remyelination, oligodendrocytes and OPCs were investi- ment and remyelination. gated in the corpus callosum. Mature oligodendrocytes CDP-choline increased remyelination and oligodendro- were analysed using immunohistochemical staining for cyte numbers only when given between Weeks 4 and 5 of Nogo-A and APC, whereas proliferating OPCs were cuprizone treatment (treatment protocols: Weeks 3–5.5, labelled by double staining for Olig-2/Ki-67 and NG2/Ki- 4–5, and 4–5.5) (Fig. 4, for PLP P 5 0.01; for APC 406 | BRAIN 2015: 138; 398–413 T. Skripuletz et al. Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021

Figure 2 CDP-choline accelerates remyelination after cuprizone-induced demyelination. CDP-choline increased remyelination in the corpus callosum of mice 0.5 weeks after cuprizone withdrawal. Mice were fed with cuprizone for 5 weeks to analyse demyelination. After 5 weeks cuprizone was withdrawn and mice were allowed to remyelinate. The extent of myelination was judged by scoring of PLP (B and C), MBP (D), LFB (E), CNPase (F), and MOG (G) stained brain sections. In CDP-choline treated mice at Week 5.5, electron microscopy of the corpus callosum presents higher numbers of myelinated axons (H and I) and lower G-ratios of the myelinated axons (J) indicating a higher thickness of myelin when compared with cuprizone controls. K represents decreased motor coordination deficits in the beam walking test of CDP-choline treated mice during remyelination at Week 5.5. In A the box marks the area of the brain coronal section that represents the middle part of the corpus callosum of which camera images were taken. Bars represent mean SEM, **P 5 0.01, ***P 5 0.001, n = 6 per time point and group (histology and immunohistochemistry), n = 3 per treated group (electron microscopy). CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 407 Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021

Figure 3 CDP-choline increases OPC proliferation and the number of mature oligodendrocytes. Increased numbers of Nogo-A (B and D) and APC (C and D) positive mature oligodendrocytes in CDP-choline treated mice were found at Week 5.5 in the corpus callosum. At Week 4 of cuprizone feeding, CDP-choline increased numbers of Olig-2/Ki-67 positive proliferating OPCs (E and F). In A the box marks the area of the brain coronal section that represents the middle part of the corpus callosum of which camera images were taken. Bars represent mean SEM, ***P 5 0.001, n = 6 per time point and group. 408 | BRAIN 2015: 138; 398–413 T. Skripuletz et al. Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021

Figure 4 Effects of CDP-choline on remyelination are dependent on OPC proliferation. Treatment with CDP-choline between Weeks 3–5.5, 4–5, and 4–5.5 increased remyelination as shown by myelin proteolipid protein (PLP) staining (A and C) and numbers of APC positive oligodendrocytes (B and D). No effects were found when CDP-choline treatment was performed between Weeks 3–4 and 5–5.5. These results show that CDP-choline exerts its effects on oligodendrocytes and remyelination in a time window between Weeks 4 and 5 of cuprizone treatment. During this time window OPC proliferate in the corpus callosum in the cuprizone model. Bars represent mean SEM, *P 5 0.05, **P 5 0.01, n = 6 per group.

P 5 0.0001). In this time window, OPCs proliferate in the migration assays were performed. Incubation of OPCs corpus callosum in the cuprizone model, while no signifi- with CDP-choline significantly enhanced their proliferation cant OPC proliferation occurs between Weeks 3–4 and 5– as measured by BrdU incorporation (P = 0.03; Fig. 5A 5.5. Thus, these results suggest that CDP-choline increases and B). The increased proliferation of OPCs after CDP- numbers of oligodendrocytes via increasing the prolifer- choline treatment was confirmed by staining of OPCs ation of OPCs. with the proliferation marker Ki-67 by omitting BrdU (P = 0.035; Supplementary Fig. 5A). As shown in Involvement of protein kinase C in Supplementary Fig. 5C, CDP-choline did not influence the CDP-choline mediated OPC differentiation of OPC into mature oligodendrocytes. In addition, no effects of CDP-choline were found on OPC proliferation migration (Supplementary Fig. 5D). To further study the mechanisms of CDP-choline on OPCs, Within cells CDP-choline can be converted directly additional cell culture proliferation, differentiation and with diacyglyceride (DAG) into phosphatidylcholine, CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 409 a major component of the cell membrane (Hjelmstad and Bell, 1991; McMaster and Bell, 1994). Thus, we have used Discussion an inhibitor of the enzyme phospholipase C (PLC), which is The concept that remyelination protects against progressive required for the activation of DAG (Fukami et al., 2010). axonal injury and consequently diminishes long-term dis- The PLC inhibitor did not influence the basic proliferation ability in patients with multiple sclerosis is the subject of rate of OPC (Fig. 5C) whereas the addition of the PLC increasing interest and attention within the neurosciences inhibitor to OPC cultures before adding CDP-choline com- (Franklin and ffrench-Constant, 2008; Kotter et al., pletely reversed the additional proliferative effects of CDP- 2011). In this context, new aspects of the underlying mech- choline (Fig. 5C). Control experiments with phosphatidyl- anisms of remyelination and the role of OPCs in these choline (end product in the CDP-choline pathway) showed processes are being discovered continuously. Yet despite no effects on OPC proliferation (Fig. 5D). As diacyglyceride advances in understanding the oligodendrocyte biology in is also an essential mediator for the activation of protein remyelination, regenerative compounds for patients with Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 kinase C (PKC) and protein kinase C is involved in regulat- multiple sclerosis are still not available. CNS remyelination ing cell proliferation (Reyland, 2009; Fukami et al., 2010), is the result of migration, proliferation and differentiation three different protein kinase C inhibitors were used along of OPCs. These processes need to be orchestrated by many with CDP-choline in OPC cultures. A protein kinase C signals and failure at any stage might lead to remyelination inhibitor dosis that did not influence the basal proliferation failure. Strategies to enhance remyelination are preferen- of OPCs was chosen. All protein kinase C inhibitors com- tially based on promoting OPC differentiation into new pletely reversed the additional proliferative effects exerted myelin building oligodendrocytes. Some of the mechanisms by CDP-choline (Fig. 5E). To determine whether protein by which the inhibition of OPC differentiation is regulated kinase C phosphorylation was required for CDP-choline have recently been described. These include modulation of induced OPC proliferation, the cells were treated with S1P receptors by FTY720 (Miron et al., 2008) and CDP-choline and immunoreactivity for phospho-protein phosphodiesterase 7 (Medina-Rodriguez et al., 2013). kinase C was investigated by western blotting. Results Inhibiting LINGO1 presents another option to enhance show that treatment of OPCs with CDP-choline did remyelination in patients with multiple sclerosis, as it was not increase the phosphorylation of protein kinase C found to negatively regulate remyelination by inhibiting (Supplementary Fig. 5B). These results suggest that the OPC differentiation (Mi et al., 2005). Treatment with CDP-choline induced OPC proliferation requires protein LINGO1 antagonists exerted beneficial effects on remyeli- kinase C, however, phosphorylation of protein kinase C nation by induction of OPC differentiation in vitro and in seems not to be affected. Our results provide strong animal models of multiple sclerosis (Mi et al., 2009) and evidence for a protein kinase C-mediated effect of CDP- clinical trials with the aim to prove the effects of LINGO1 choline to increase the proliferation rate of OPCs. blocking in multiple sclerosis patients are ongoing. In our study, we identified the substance CDP-choline for its potent regenerative effects. CDP-choline ameliorated the disease course and exerted beneficial effects on myelin, CDP-choline does not influence the oligodendrocytes, and axons in EAE. Furthermore, after function of microglia/macrophages cuprizone-induced demyelination CDP-choline effectively in vitro enhanced remyelination and reversed motor coordination deficits. The increased regeneration arose from an increase Macrophages and microglial cells are known to be in the numbers of proliferating OPCs, which resulted in involved in CNS repair mechanisms (Kotter et al., 2011). higher numbers of new oligodendrocytes. Several earlier We have thus analysed the effects of CDP-choline on studies indicated that CDP-choline may have the capacity microglia and macrophages in cell culture experiments. to protect cells, such as , and that cellular prolifer- CDP-choline did not exert any effects on TNFA, IL10, ation may be stimulated by this substance (Mykita et al., and nitrite release in microglia and macrophages 1986; Hurtado et al., 2011; Gutierrez-Fernandez et al., (Supplementary Fig. 4D–F and 5 E–G). In addition, CDP- 2012). However, the substance was not tested for its neu- choline did not exert any effects on macrophage phagocyt- roprotective and regenerative effects in demyelinating dis- osis, macrophage migration, macrophage polarization eases such as multiple sclerosis. (M1/M2), and the production of further cytokines (IL6, To date, several potentially neuroprotective approches IL1B, MCP1, MIP1 ) (Supplementary Fig. 4). have been described in animal models mimicking multiple Furthermore, the supernatants of CDP-choline stimulated sclerosis (Aktas et al., 2010; Tselis et al., 2010). These in- microglia had no effects on OPC proliferation clude activation of the Nrf2 antioxidant pathway (Linker (Supplementary Fig. 5H). Thus, our results indicate that et al., 2011), sodium channel blockers (Meuth et al., 2010), CDP-choline does not drive its regenerative effects via matrix metalloproteinase inhibitors (Maier et al., 2007), microglia and the effects on OPCs are driven directly via and glutamate antagonists (Smith et al., 2000). Although promotion of OPC proliferation. these substances may to some extent exert , 410 | BRAIN 2015: 138; 398–413 T. Skripuletz et al. Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021

Figure 5 Effects of CDP-choline on OPC proliferation in vitro. CDP-choline increased the proliferation rate of OPCs in cell culture experiments as measured by BrdU incorporation (A, n = 4). In B exemplary stained proliferating OPC (green BrdU, red A2B5) are shown after incubation with CDP-choline compared to controls (nuclei were counterstained with DAPI). Inhibition of the enzyme phospholipase C (PLC),

(continued) CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 411 effects on remyelination were not found. CDP-choline is our results, we firmly believe that effects on OPCs are an characterized by a very beneficial side effect profile not important mechanism of CDP-choline action after a demye- only in animal models, but also in humans (Adibhatla linating attack. However, it is possible that this mechanism and Hatcher, 2005). CDP-choline has been studied in is not exclusive. 411 000 humans and showed some beneficial effects in In the cuprizone model CDP-choline not only increased , Alzheimer’s and Parkinson’s dis- the numbers of myelinated axons, but the newly formed eases, learning and memory disorders, amblyopia, and myelin sheaths were also thicker as compared to cuprizone (Adibhatla and Hatcher, 2005). Side effects controls. The effects on remyelination are not mediated via were rare and consisted mainly of stomach pain, diarrhoea, increasing the production of growth factors that are known and headaches (Conant and Schauss, 2004; Davalos and to have an impact on OPC proliferation. Instead, our Secades, 2011). The sodium salt of CDP-choline, the form results provide a new mechanism of remyelination via the used in clinical trials, is sold as a drug in Europe (Conant choline metabolism. In the CNS, choline is used for the Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 and Schauss, 2004). Due to this excellent safety profile and biosynthesis of acetylcholine and the cell membrane easy availibility, CDP-choline is a promising substance, phospholipid phosphatidylcholine (lecithin) (Weiss, 1995; which may be of particular interest as an add-on therapy Clement and Kent, 1999; Hunt et al., 2001). In the synthe- in multiple sclerosis patients. sis of phosphatidylcholine the availability of choline might In the present study we could not find any neuroprotec- be limited during pathological conditions in which regen- tive effects of CDP-choline in two different animal models erative processes require more substrate. Thus, exogenously of multiple sclerosis. However, we identified the substance applied CDP-choline is an interesting tool to promote these CDP-choline to enhance CNS remyelination via a novel processes as it disperses in the organism mainly in the form mechanism mediated by the choline metabolic pathway. of its degradation products choline and cytidine, which Recently, clinical trials in acute ischaemic stroke did not cross easily the blood–brain barrier (Conant and Schauss, reveal beneficial effects of CDP-choline (Davalos et al., 2004; Garcia-Cobos et al., 2010). 2012), which might be explained by a completely different In our work, exogenous CDP-choline was effective to mechanism of lesion induction. In contrast to stroke, mul- increase the proliferation of OPCs in vivo as well as tiple sclerosis lesions are accompanied by demyelination in vitro. In the synthesis of phosphatidylcholine, and subsequent repair processes. This remyelination is the CDP-choline first needs to be produced from choline and result of migration and proliferation of OPCs that differen- cytidine as an intermediate compound. In the next step, tiate into mature oligodendrocytes and build new myelin CDP-choline is converted together with diacyglyceride sheaths (Franklin and ffrench-Constant, 2008). Our results into phosphatidylcholine by the enzyme CPT (Hjelmstad from the cuprizone model and preventive MOG-EAE show and Bell, 1991; McMaster and Bell, 1994). Inhibition of that CDP-choline promotes remyelination via increasing the enzyme phospholipase C, which is essential for the ac- numbers of OPCs and thus oligodendrocytes. In EAE, treat- tivation of diacyglyceride (Fukami et al., 2010), completely ment with CDP-choline from the onset of disease (Day 7 reversed the enhancement of proliferation of CDP-choline after immunization) was still effective. In contrast, treat- demonstrating that diacyglyceride is required in this pro- ment with CDP-choline after the peak of disease (Day 15 cess. The end product phosphatidylcholine did not show after immunization) was ineffective in ameliorating clinical any effect on OPC proliferation, further supporting the symptoms of EAE. Since the maximum of proliferating role of diacyglyceride in CDP-choline enhanced OPC pro- OPCs occurs before Day 15 in the MOG-EAE (Herrero- liferation. At the same time, diacyglyceride is also an essen- Herranz et al., 2008), the results from the CDP-choline tial mediator for the activation of protein kinase C that is treatment at the peak of disease are consistent with an involved in the regulation of cell proliferation (Reyland, effect of CDP-choline on proliferating OPCs. In view of 2009; Fukami et al., 2010). Inhibition of protein kinase

Figure 5 Continued which is required for the activation of the enzyme diacyglyceride (DAG), reversed the additional proliferative effects induced by CDP-choline (C, n = 4). Phosphatidylcholine (end product in the CDP-choline pathway) did not induce any effects on OPC proliferation (D, n = 5). DAG is essential for the activation of protein kinase C (PKC) and three different inhibitors (Bis = bisindolylmaleimide IV: Calph = calphostin C: Chelery = chelerythrine chloride) were tested with CDP-choline in OPC cultures. All PKC inhibitors completely reversed the additional proliferative effects of CDP-choline (E, n = 5). Bars represent mean SEM, *P 5 0.05, **P 5 0.01, ***P 5 0.001. The schematic overview (F) shows that CDP-choline is converted directly with DAG to phosphatidylcholine and CMP (cytidine-mono-phosphate) in a reaction catalyzed by CPT (1,2-diacylglycerol cholinephosphotransferase). DAG is a product of the hydrolysis of PIP2 (phosphatidylinositol 4,5-bisphosphate) by the membrane-bound enzyme PLC (phospholipase C) and is an essential physiological activator of protein kinase C. Our results show that the inhibition of PLC and thereby DAG reverses the CDP-choline effect on proliferation. DAG is essential for protein kinase C activation and it can be suggested that CDP-choline exerts its effects on proliferation via DAG and thus via protein kinase C activation, as we could show that inhibition of PKC activity reversed the CDP-choline effects on proliferation and phosphatidylcholine had no effects on proliferation. PIP2 = phosphatidylinositol 4,5-bisphosphate. 412 | BRAIN 2015: 138; 398–413 T. Skripuletz et al.

C completely reversed the additional proliferative effects Davalos A, Alvarez-Sabin J, Castillo J, Diez-Tejedor E, Ferro J, induced by CDP-choline. Thus, our results provide strong Martinez-Vila E, et al. Citicoline in the treatment of acute ischaemic stroke: an international, randomised, multicentre, placebo-controlled evidence for a protein kinase C-mediated effect of CDP- study (ICTUS trial). Lancet 2012; 380: 349–57. choline to increase the proliferation rate of OPCs. Davalos A, Secades J. Citicoline preclinical and clinical update 2009– However, it is possible that this mechanism is not exclusive 2010. Stroke 2011; 42: S36–9. and might not occur in adult OPCs. de Oliveira AC, Candelario-Jalil E, Bhatia HS, Lieb K, Hull M, In conclusion, our results identified CDP-choline as a Fiebich BL. Regulation of prostaglandin E2 synthase expression in activated primary rat microglia: evidence for uncoupled regulation drug with remyelination promoting functions with a new of mPGES-1 and COX-2. Glia 2008; 56: 844–55. mechanism of action. We believe that CDP-choline is a Eugster HP, Frei K, Bachmann R, Bluethmann H, Lassmann H, promising substance for multiple sclerosis patients and Fontana A. Severity of symptoms and demyelination in MOG- should be further explored for its regenerative effects in induced EAE depends on TNFR1. Eur J Immunol 1999; 29: 626–32. humans. Franklin RJ, ffrench-Constant C. Remyelination in the CNS: from

biology to therapy. Nat Rev Neurosci 2008; 9: 839–55. Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 Fukami K, Inanobe S, Kanemaru K, Nakamura Y. Phospholipase C is a key enzyme regulating intracellular calcium and modulating the Acknowledgements phosphoinositide balance. Prog Lipid Res 2010; 49: 429–37. Garcia-Cobos R, Frank-Garcia A, Gutierrez-Fernandez M, We thank I. Cierpka-Leja, S. Lang, K. Rohn, S. Seubert, Diez-Tejedor E. Citicoline, use in cognitive decline: vascular and and A. Niesel for excellent technical assistance. Further, we degenerative. J Neurol Sci 2010; 299: 188–92. Gudi V, Gingele S, Skripuletz T, Stangel M. Glial response during thank Trommsdorff GmbH & Co. KG Arzneimittel for cuprizone-induced de- and remyelination in the CNS: lessons donating CDP-choline. We are indebted to Ben Ettle, learned. Front Cell Neurosci 2014; 8: 73. Department of Molecular Neurology, Erlangen, Germany Gudi V, Skuljec J, Yildiz O, Frichert K, Skripuletz T, Moharregh- for help with confocal microscopy of EAE lesions. Khiabani D, et al. Spatial and temporal profiles of growth factor expression during CNS demyelination reveal the dynamics of repair priming. PLoS One 2011; 6: e22623. Gutierrez-Fernandez M, Rodriguez-Frutos B, Fuentes B, Funding Vallejo-Cremades T, Alvarez-Grech J, Exposito-Alcaide M, et al. CDP-choline treatment induces brain plasticity markers expression T.S., W.B. and M.S. are supported by the German Research in experimental animal stroke. Neurochem Int 2012; 60: 310–17. Foundation (DFG, FOR 1103, TP1b, STA 518/4-1, BA Hagemeier K, Bruck W, Kuhlmann T. Multiple sclerosis—remyelina- 815/10-2). R.A.L. holds an endowed chair supported by tion failure as a cause of disease progression. Histol Histopathol 2012; 27: 277–87. the Novartis Stiftung fu¨ r Therapeutische Forschung. Herrero-Herranz E, Pardo LA, Gold R, Linker RA. Pattern of axonal injury in murine myelin oligodendrocyte glycoprotein induced experimental autoimmune encephalomyelitis: implications for multi- ple sclerosis. Neurobiol Dis 2008; 30: 162–73. Supplementary material Hjelmstad RH, Bell RM. Molecular insights into enzymes of mem- Supplementary material is available at Brain online. brane bilayer assembly. Biochemistry 1991; 30: 1731–40. Hunt AN, Clark GT, Attard GS, Postle AD. Highly saturated endonuclear phosphatidylcholine is synthesized in situ and colocated with CDP-choline pathway enzymes. J Biol Chem 2001; 276: References 8492–9. Hurtado O, Lizasoain I, Moro MA. Neuroprotection and recovery: Adibhatla RM, Hatcher JF. Cytidine 50-diphosphocholine recent data at the bench on citicoline. Stroke 2011; 42: S33–5. (CDP-choline) in stroke and other CNS disorders. Neurochem Res Kent C, Carman GM. Interactions among pathways for phosphatidyl- 2005; 30: 15–23. choline metabolism, CTP synthesis and secretion through the Golgi Adibhatla RM, Hatcher JF, Larsen EC, Chen X, Sun D, Tsao FH. apparatus. Trends Biochem Sci 1999; 24: 146–50. CDP-choline significantly restores phosphatidylcholine levels by dif- Kipp M, Clarner T, Dang J, Copray S, Beyer C. The cuprizone animal ferentially affecting phospholipase A2 and CTP: phosphocholine model: new insights into an old story. Acta Neuropathol 2009; 118: cytidylyltransferase after stroke. J Biol Chem 2006; 281: 6718–25. 723–36. Aktas O, Kieseier B, Hartung HP. Neuroprotection, regeneration and Kleinewietfeld M, Manzel A, Titze J, Kvakan H, Yosef N, Linker RA, immunomodulation: broadening the therapeutic repertoire in mul- et al. Sodium chloride drives autoimmune disease by the induction tiple sclerosis. Trends Neurosci 2010; 33: 140–52. of pathogenic TH17 cells. Nature 2013; 496: 518–22. de Castro F, Bribian A, Ortega MC. Regulation of oligodendrocyte Kotsiari A, Voss EV, Pul R, Skripuletz T, Ragancokova D, Trebst C, precursor migration during development, in adulthood and in path- et al. Interferon-beta treatment normalises the inhibitory effect of ology. Cell Mol Life Sci 2013; 70: 4355–68. serum from multiple sclerosis patients on oligodendrocyte progenitor Clement JM, Kent C. CTP:phosphocholine cytidylyltransferase: proliferation. Neurosci Lett 2010; 485: 107–11. insights into regulatory mechanisms and novel functions. Biochem Kotter MR, Stadelmann C, Hartung HP. Enhancing remyelination in Biophys Res Commun 1999; 257: 643–50. disease—can we wrap it up? Brain 2011; 134: 1882–900. Compston A, Coles A. Multiple sclerosis. Lancet 2008; 372: 1502–17. Linker RA, Lee DH, Demir S, Wiese S, Kruse N, Siglienti I, et al. Conant R, Schauss AG. Therapeutic applications of citicoline for Functional role of brain-derived neurotrophic factor in neuroprotec- stroke and cognitive dysfunction in the elderly: a review of the tive autoimmunity: therapeutic implications in a model of multiple literature. Altern Med Rev 2004; 9: 17–31. sclerosis. Brain 2010; 133: 2248–63. Crowther JR. ELISA. Theory and practice. Methods Mol Biol 1995; Linker RA, Lee DH, Ryan S, van Dam M, Conrad R, Bista P, et al. 42: 1–218. Fumaric acid esters exert neuroprotective effects in CDP-choline promotes remyelination BRAIN 2015: 138; 398–413 | 413

neuroinflammation via activation of the Nrf2 antioxidant pathway. Mykita S, Golly F, Dreyfus H, Freysz L, Massarelli R. Effect of CDP- Brain 2011; 134: 678–92. choline on hypocapnic neurons in culture. J Neurochem 1986; 47: Linker RA, Maurer M, Gaupp S, Martini R, Holtmann B, Giess R, 223–31. et al. CNTF is a major protective factor in demyelinating CNS dis- Paxinos G, Franklin KBJ. The mouse brain in stereotaxic coordinates. ease: a neurotrophic cytokine as modulator in neuroinflammation. New York, Academic Press; 2001. Nat Med 2002; 8: 620–4. Reyland ME. Protein kinase C isoforms: Multi-functional regulators of Maier K, Merkler D, Gerber J, Taheri N, Kuhnert AV, Williams SK, cell life and death. Front Biosci 2009; 14: 2386–99. et al. Multiple neuroprotective mechanisms of minocycline in auto- Singh V, Voss EV, Benardais K, Stangel M. Effects of 2-chlorodeox- immune CNS inflammation. Neurobiol Dis 2007; 25: 514–25. yadenosine (Cladribine) on primary rat microglia. J Neuroimmune McMaster CR, Bell RM. Phosphatidylcholine biosynthesis via the Pharmacol 2012; 7: 939–50. CDP-choline pathway in Saccharomyces cerevisiae. Multiple mech- Skripuletz T, Gudi V, Hackstette D, Stangel M. De- and remyelination anisms of regulation. J Biol Chem 1994; 269: 14776–83. in the CNS white and grey matter induced by cuprizone: the old, Medina-Rodriguez EM, Arenzana FJ, Pastor J, Redondo M, the new, and the unexpected. Histol Histopathol 2011; 26: Palomo V, Garcı´a de Sola R, et al. Inhibition of endogenous 1585–97.

phosphodiesterase 7 promotes oligodendrocyte precursor differenti- Skripuletz T, Hackstette D, Bauer K, Gudi V, Pul R, Voss E, et al. Downloaded from https://academic.oup.com/brain/article/138/2/398/293983 by guest on 26 September 2021 ation and survival. Cell Mol Life Sci 2013; 70: 3449–62. Astrocytes regulate myelin clearance through recruitment of micro- Meuth SG, Bittner S, Ulzheimer JC, Kleinschnitz C, Kieseier BC, glia during cuprizone-induced demyelination. Brain 2013; 136: Wiendl H. Therapeutic approaches to multiple sclerosis: an update 147–67. on failed, interrupted, or inconclusive trials of neuroprotective and Skripuletz T, Lindner M, Kotsiari A, Garde N, Fokuhl J, Linsmeier F, alternative treatment strategies. BioDrugs 2010; 24: 317–30. et al. Cortical demyelination is prominent in the murine cuprizone Mi S, Miller RH, Lee X, Scott ML, Shulag-Morskaya S, Shao Z, et al. model and is strain-dependent. Am J Pathol 2008; 172: 1053–61. LINGO-1 negatively regulates myelination by oligodendrocytes. Skripuletz T, Miller E, Moharregh-Khiabani D, Blank A, Pul R, Nat Neurosci 2005; 8: 745–51. Gudi V, et al. Beneficial effects of minocycline on cuprizone induced Mi S, Miller RH, Tang W, Lee X, Hu B, Wu W, et al. Promotion of cortical demyelination. Neurochem Res 2010; 35: 1422–33. central nervous system remyelination by induced differentiation of Smith T, Groom A, Zhu B, Turski L. Autoimmune encephalomyelitis oligodendrocyte precursor cells. Ann Neurol 2009; 65: 304–15. ameliorated by AMPA antagonists. Nat Med 2000; 6: 62–66. Miron VE, Jung CG, Kim HJ, Kennedy TE, Soliven B, Antel JP. Tselis A, Khan OA, Lisak RP. Approaches to neuroprotective strate- FTY720 modulates human oligodendrocyte progenitor process gies in multiple sclerosis. Expert Opin Pharmacother 2010; 11: extension and survival. Ann Neurol 2008; 63: 61–71. 2869–2878. Muccioli M, Pate M, Omosebi O, Benencia F. Generation and labeling Weiss GB. Metabolism and actions of CDP-choline as an endogenous of murine bone marrow-derived dendritic cells with Qdot nanocrys- compound and administered exogenously as citicoline. Life Sci tals for tracking studies. J Vis Exp 2011. 1995; 56: 637–660.