The endocannabinoid system and its therapeutic exploitation in multiple sclerosis: Clues for other neuroinflammatory diseases.

(1)

Review article

The endocannabinoid system and its therapeutic exploitation in multiple sclerosis: Clues for other neuroin ﬂammatory diseases

Valerio Chiurchiù

^a,b

, Mario van der Stelt

^c

, Diego Centonze

^d,e,1

, Mauro Maccarrone

^a,b,

*

^,1

aDepartmentofMedicine,CampusBio-MedicoUniversityofRome,ViaAlvarodelPortillo21,00128Rome,Italy

bEuropeanCenterforBrainResearch/IRCCSSantaLuciaFoundation,ViadelFossodiFiorano64,00143Rome,Italy

cDepartmentofMolecularPhysiology,LeidenUniversity,Einsteinweg55,2333CC,Leiden,TheNetherlands

dDepartmentofSystemsMedicine,TorVergataUniversityofRome,ViaMontpellier1,00133Rome,Italy

eIRCCSIstitutoNeurologicoMediterraneoNeuromed,ViaAtinense18,86077Pozzilli,IS,Italy

ARTICLE INFO

Articlehistory:

Received29April2017

Receivedinrevisedform23October2017 Accepted28October2017

Availableonlinexxx

Keywords:

Cannabinoid Clinicaltrials Endogenouslipids Inﬂammation Neurodegeneration Neurotransmission

ABSTRACT

Multiplesclerosisis themostcommoninﬂammatorydemyelinatingdiseaseofthecentralnervous system, caused byan autoimmune response against myelin that eventually leads to progressive neurodegeneration and disability. Although the knowledge on its underlying neurobiological mechanisms has considerably improved, there is a still unmet need for new treatment options, especially fortheprogressive formsof thedisease. Bothpreclinical andclinical data suggestthat cannabinoids,derivedfromtheCannabissativaplant,maybeusedtocontrolsymptomssuchasspasticity andchronicpain,whereasonlypreclinicaldataindicatethatthesecompoundsandtheirendogenous counterparts,i.e.theendocannabinoids,mayalsoexertneuroprotectiveeffectsandslowdowndisease progression.Here,wereviewthepreclinicalandclinicalstudiesthatcouldexplainthetherapeuticaction of cannabinoid-based medicines,as well asthemedical potential of modulatingendocannabinoid signalinginmultiplesclerosis,withalinktootherneuroinﬂammatorydisordersthatsharecommon hallmarksandpathogeneticfeatures.

Contents

1. Introductiontoneuroinﬂammationandmultiplesclerosis ... 00

2. AglimpsetotheeCBsystem ... 00

2.1. Cannabinoidsandcannabinoidreceptors ... 00

2.2. Endogenousligandsofcannabinoidreceptorsandtheirmetabolicroutes ... 00

2.3. Otherreceptors ... 00

3. Therapeuticexploitationoftheendocannabinoidsysteminmultiplesclerosis ... 00

3.1. PreclinicalexvivoandinvivostudiesofeCBsignalinginMS ... 00

3.2. eCBsignalingfrompreclinicalmodelsofMStoclinicalapplications ... 00

4. eCBsignalingasapotentialtherapyforotherneurodegenerativediseaseswithaninﬂammatorycomponent ... 00

4.1. Alzheimer’sdisease ... 00

4.2. Parkinson’sdisease ... 00

4.3. Huntington’sdisease ... 00

Abbreviations:2-AG,2-arachidonoylglycerol;Ab,b-amyloid;GABA,g-aminobutyricacid;AD,Alzheimer’sdiseases;ALS,amyotrophiclateralsclerosis;AEA,anandamide;

EAE,autoimmuneencephalomyelitis;BBB,bloodbrainbarrier;CDB,cannabidiol;CB,cannabinoidreceptor;CNS,centralnervoussystem;COX,cyclooxygenase;DAGL, diacylglycerollipase;CB,endocannabinoids;FAAH,fattyacidamidehydrolase;HD,Huntington’sdisease;IFN,interferon;IL,interleukin;MS,multiplesclerosis;MAGL, monoacylglycerol lipase; NAPE-PLD, N-acylphosphatidyl-ethanolamine-speciﬁc phospholipase D; PD, Parkinson’s disease; PPAR, peroxisomeproliferator-activated receptors;PP,primaryprogressive;PR,progressiverelapsing;RR,relapsing-remitting;MRI,resonanceimaging;SP,secondaryprogressive;SOD,superoxidedismutase;THC, tetrahydrocannabinol;Th,-helper;TRPV1,transientreceptorpotentialvanilloid1;TNF,tumournecrosisfactor.

* Correspondingauthorat:DepartmentofMedicine,CampusBio-MedicoUniversityofRome,ViaAlvarodelPortillo21,00128Rome,Italy.

E-mailaddress:m.maccarrone@unicampus.it(M.Maccarrone).

1 Equallyseniorauthors.

xxx–xxx ContentslistsavailableatScienceDirect

Progress in Neurobiology

j o u r n a lh o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / p n e u r o b i o

https://doi.org/10.1016/j.pneurobio.2017.10.007

(2)

4.4. Amyotrophiclateralsclerosis ... 00

5. Conclusionsandfuturedirections ... 00

Competingﬁnancialinterests ... 00

Acknowledgements ... 00

References ... 00

1.Introductiontoneuroinﬂammationandmultiplesclerosis

Although the central nervous system (CNS) has been long consideredanimmune-privilegedsite,mainlyduetothepresence ofthebloodbrainbarrier(BBB),immuneactivitiesoccurandare sometimes ncesessary for neuronal function and host defence (Banks, 2015).Any insultof the brain is associated withacute inflammation,characterizedbyendothelialcellactivation,tissue oedemaand release of inflammatorymediators thatprovide to eliminatetheinsultandrestorebrainfunction.However,neuro- inflammationiswidelyregardedasachronicprocessthatinvolves both sustained activation of resident glial cells (microglia and astroglia)andrecurrentinfiltrationofperipheralleukocytesand solubleinflammatorymediators.Microglial cells,either derived frommesenchymalmonocyteprecursorsofthemesodermorfrom non-hematopoietic microglial precursors in the yolk sac, both entering the brain during embryonic and fetal stages, are the immune sentinels of the CNS. Microglial cells are extremely heterogeneous,becausetheycanexistinmanydifferentformsand activation states, from neuro-protective to neuro-destructive (Nayaketal.,2014).Microglialcellsarecriticalforthedevelopment andsurveillanceoftheCNS,shapingtheimmunephenotypeofthe brainandcriticallymodulatingneuroinflammation(Mosseretal., 2017).AstrocytesarebyfarthemostabundantcellsintheCNS, where they regulatevirtually everyphysiological process,from physically forming the BBB and giving nutritional support to sustaining neurotransmitters turnover, synaptic plasticity and immune functions. Indeed, immune activation of astrocytes in response to signals released by injured neurons or activated microglia leads to the so-called astrogliosis, a hallmark of neuroinflammation,whichleadstoglialscarsthatpreventaxonal regeneration(Jensenetal.,2013).Duringneuroinflammationalso immune cells of both innate (monocytes/macrophages and dendritic cells) and adaptive (Tand B lymphocytes) immunity arepersistentlyrecruitedtothebrain,andreleaseinflammatory cytokines and chemokines that exacerbate neuroinflammation (SchwartzandBaruch,2014).Asamatteroffact,neuroinflamma- tionoftenmayleadtoneurodegeneration,axonallossandsynaptic dysfunction, and thus it is typically associated with several neurological disorders, of which multiple sclerosis (MS) is the prototypicalexample.MSisaprogressive,chronicneurodegener- ative disease, which affects approximately 2.3 million people worldwide (Browne et al., 2014). It is the most common neurological disorder in young adults, and is regarded as an autoimmunediseaseinwhich inflammationleadstodemyelin- ationoftheaxonsintheCNS.AlthoughtheaetiologyofMSisstill unknown,itisalmostunanimouslybelievedthatbothgeneticand environmentalcomponentsplayacentralroleindiseaseonsetand development(Hafleretal.,2007;HuynhandCasaccia,2013).The MSprevalenceratioofwomentomen(2.3–3.5:1)hasincreased markedlyduringthelastdecadesandthisrapidincreaseprobably reflectsadifferentialgenderresponsetounidentifiedchangesin environmentornutrition(Harboetal.,2013).MSischaracterized byaseriesofepisodicacuteattacks(referredtoasrelapses)and remissions,butitgraduallyleadstoprogressiveneurodegenera- tionanddeteriorationofneurologicfunctionwithoutanyfurther remission.AlthoughthedifferentclinicalcoursesofMS(alsocalled

“types”or“phenotypes”)weredeﬁnedin1996,theInternational

AdvisoryCommetteonClinicalTrialsofMS,basedonadvancesin the understanding of the disease process in MS and MRI technology, classiﬁed MS intoonly four independentsubtypes:

(i) relapsing-remitting (RR), deﬁned by unpredictable relapses withfullrecoveryorwithsequelae;(ii)primaryprogressive(PP), which progressescontinuously fromtheonset withoutattacks;

(iii)secondaryprogressive(SP),whichfollowsinitialRRandthen progresseswithdeclinewithoutremissions;and(iv)progressive relapsing(PR),characterizedbyasteadydeclineonsetwithsuper- imposedattacks(Fig.1).TheRR-MSisthemostprevalentformand accountsforapproximately85%ofallcases(CompstonandColes, 2008;Lublinetal.,2014).However,theCommitteealsorecognized the clinically isolated syndrome (CIS) as a first episode of neurologicsymptomscausedbyinflammationanddemyelination thatmustlastatleast214handthatischaracteristicofMSeven thoughdoesnot yetmeetthecriteria for itsdiagnosisbecause patientswithCISmight alsonotendupdeveloping MS(Lublin et al., 2014).MS pathogenesis and pathophysiology have been extensivelystudied,especiallyin theexperimental autoimmune encephalomyelitis(EAE)mousemodel,andarethoughttoinvolve initially the disruption of the immune system and of central myelin-producingcells.InthecourseofMS,damagetotheBBBand over-activationofbrainmicroglialeadtoasubstantialinfiltration ofautoreactivelymphocytes,causingoligodendrocytedeathand axonaldamageandultimatelyresultingindemyelination,synaptic alterationandneuronalloss(CompstonandColes,2008;Duttaand Trapp,2011;Calabreseetal.,2015;Mahadetal.,2015)(Fig.2).

Thoughtheimmune-mediatedneuroinflammationhypothesis hasdominated MS research for over 50 years, recent evidence seems to point to a neurodegenerative and microglia-centered process,accordingtowhichMSisprimarilyaneurodegenerative disease that starts in the brain, and then develops becauseof inflammation(Kassmannetal.,2007;Lassmannetal.,2012).This hasledtothecurrent“inside-out”and“outside-in”modelsofMS immunopathogenesis, whereby in the first model the immune response that destroys myelin and leads to BBB breakdownis drivenbyadysfunctionofbraincells,whereasinthesecondmodel a dysfunction residing in the periphery leads to BBB damage, myelindisruptionandaxonaldeath(TsunodaandFujinami,2002;

Stysetal.,2012).

The immune-mediatedattacksare mainlydrivenbycells of adaptiveimmunity,namelymyelin-speciﬁcandself-reactiveCD8 amd CD4 T-cells (T-helper 1 and T-helper 17), with a key contributionofB-cellsthatproducehighlevelsofautoantibodies and that have been recently shown to contribute to neuro- degenerationandcorticaldemyelination,especiallyformeningeal ectopicBcellfollicles(Fraussenetal.,2016).Duringtheseattacks, themyelinsheathisdamaged,thusimpairingaxonalconduction and the correct communication between differentparts of the nervoussystem(CompstonandColes,2008;Gandhietal.,2010;

Mahadetal.,2015).Theseprocessesaresustainedbyasubsequent recruitmentofcellsof theinnate immunityfromtheperiphery thatfurtheramplifytheactivationofpathogenicT-cellsandthe destructionofneuronsandoligodendrocytes.Inaddition,thereisa permanent activation of resident microglia and astrocytes that furtherpotentiatetheneuroinﬂammatoryresponsebyproducing proinﬂammatorymediators(Gandhietal.,2010;Chiurchiù,2014) (Fig.2).Oligodendrocytes areresponsiblefor myelinproduction

xxx–xxx

(3)

Fig.2. ImmunopathogenesisofMS.Injurytoblood-brainbarrier(BBB)and/orhyperactivationofbrainmicroglialcellsleadtoamassiveinﬁltrationinthecentralnervous systemofautoreactiveTcells(CD8andCD4T-helpercells),Bcellsandantigen-presentingcells(APCs),includingmonocytes/macrophages.Inthebrain,T cellsandmonocytes/

macrophagesreleasecytokinesandchemokinesthatdamagetheoligodendroglialcells,andhencethemyelinsheath.Concomitantly,Bcells(plasmacells)producemyelin- speciﬁcantibodies,whichformmembrane-attackcomplexesthatfurtherdamagemyelinsheath.Theseprocessesalsoleadtoanover-activationofmicrogliaandastrocytes, whichfurtherpotentiatetheinﬂammatoryresponse.

Fig.1.TheimmunologicalbasisofthedifferentclinicalformsofMS.

xxx–xxx

(4)

and generate new myelin in a process termed remyelination, especiallyintheearlyphasesofthedisease.Remyelinationisone of the reasons why symptoms tend to decrease or disappear temporarilyinRR-MS(Peferoenetal.,2014).However,oligoden- drocytesareunabletocompletelyrebuildthemyelinsheathand repeatedattacksleadtolessandlesseffectiveremyelination,until scar-likeplaques(scleroses)builduparoundthedamagedaxons withsubsequentaxonalloss(Cambronetal.,2012).Thesescars, usuallymorethan10asvisualizedbymagneticresonanceimaging (MRI),determinea widerangeof symptoms,includingphysical disability,fatigue,andcognitiveimpairment(Polmanetal.,2011).

Approximately10yearsafterdiseaseonset,frequentrelapsesand remissionsgivepacetoaprogressiveclinicalaggravation.Suchan evolution from relapsing-remitting toprogressive forms of MS certainlyinvolvesthelossofoligodendrogliaand neuronsmost likelyduetoseveralconcurrentevents,including(i)disruptionof peripheralimmune tolerance and a subsequent shift to innate immunity (plaques contain more monocyte/macrophages and dendriticcells andfewerTcellsinchronicdisease),(ii)reduced neurotrophicabilitiesofastrocytesonneurons,(iii)increasedloss ofoligodendrocyteprecursorcellsand(iv)toxiceffectsonaxons (Gandhietal.,2010;Reynoldsetal.,2011)(Fig.1).

Atanyrate,allthesenoxiouseffectsappeartobecausingthe widespread,corticalandsubcorticalgreymatteratrophysincethe earlystagesofthediseaseasobservedbyMRIandanalysisofpost- mortemMSpatientsamples(Mandolesietal.,2015;Musellaetal., 2016). Neurodegeneration leads to permanent and increasing disability,withspasticityandpainasimportanthallmarksofthe disease.

CurrenttherapiesforMSfocusonthepreventionofrelapsesin theearlystage,mainlystoppingtheimmuneattackswithdrugs thatdirectlyinhibitcellactivationandthereleaseofinﬂammatory mediators,or thatprevent leukocyterecruitmentinto theCNS.

However, almost no drugs that prevent or slow down the

occurrence of neurodegeneration and disease progression are yetavailable, withtheexception ofthe veryrecentlyapproved humanized monoclonal antibody Ocrelizumab (Ocrevuz¹) for primaryprogressiveMSpatients(Montalbanetal.,2017).Approval ofNabiximol(Sativex¹,a1:1mixtureof

D

⁹-tetrahydrocannabinol (THC)andcannabidiol(CBD),bothisolatedfromCannabissativa)to treat MS-induced spasticity was recently obtained in several countries (Zettl et al., 2016; Maccarrone et al., 2017). Since preclinicaldatashowthatcannabinoidsalsoexertanti-inflammatory properties and neuroprotective effects (Chiurchiù et al., 2015a), it is hypothesized that these substances may limit progressiveneurodegenerationinMS(Pryceetal.,2003;Jackson etal.,2005;Wittingetal.,2006).Here,wewillreviewtheroleof cannabinoidsandtheirendogenouscounterparts(i.e.,theendocannabinoids,eCBs)inMS,theirmodeofactionandpreclinicaland clinical significance, as well as how their pharmacological modulation could be useful to treat other neuroinflammatory diseases.

2.AglimpsetotheeCBsystem

2.1.Cannabinoidsandcannabinoidreceptors

Forover5000years,Cannabis(Cannabissativa)haslongbeen usedforrecreationalandmedicinalpurposesformorethan5000 years. Cannabis extracts likemarijuana contain more than 500 naturalcompounds,outofwhichover110belongtocannabinoids (Maccarroneetal.,2017).Althoughthefirstisolatedplant-derived cannabinoids were cannabinol and cannabidiol (CBD), the identification of the major psychoactive constituent THC was achievedlateron(Table1)(Adams,1942;GaoniandMechoulam, 1964), establishing the current field of cannabinoid research.

Severalcannabinoidshavedistinctbiologicalactivity,yetTHCand CBD were particularly studied due to their broad spectrum of

Table1

Major(endo)cannabinoids,andmainmetabolicenzymesofeCBswitharoleinneuroinﬂammation.

Name(abbreviation) Chemicalstructure

D⁹-Tetrahydrocannabinol(THC)

Cannabidiol(CBD)

N-Arachidonoylethanolamine(Anandamide,AEA)

2-Arachidonoylglycerol(2-AG)

BiosyntheticenzymeofAEA Intracellularlocalization

N-acylphosphatidylethanolamines(NAPE)-speciﬁcphospholipaseD(NAPE-PLD) Membrane-associated

Biosyntheticenzymesof2-AG Intracellularlocalization

Diacylglycerollipasea(DAGLa) Membrane-associated

Diacylglycerollipaseb(DAGLb) Membrane-associated

DegradingenzymeofAEA Intracellularlocalization

Fattyacidamidehydrolase(FAAH) Membrane-associated(mainlyER)

Degradingenzymeof2-AG Intracellularlocalization

Monoacylglycerollipase(MAGL) Membrane-associatedandcytosolic

xxx–xxx

(5)

action, including analgesic and anti-inflammatory effects. THC stereoselectively acts in the CNS (Mechoulam et al., 1988) by bindingtoaspecifictargetidentifiedasthetype-1cannabinoid receptor (CB1) (Devane et al., 1988). Few years later a type-2 cannabinoidreceptor(CB2)wasdiscoveredinperipheralB-andT- lymphocytes(Munroetal.,1993).THCbindstobothCB1andCB2

with high afﬁnity (Huffman, 2000; Mahadevan et al., 2000), whereasCBD,anon-psychoactivecomponent,showslittleafﬁnity forbothreceptors(Mechoulametal.,2007).Uponactivationboth receptorsinhibittheformationofthesecondmessengercAMP,and modulateothersignaltransductionpathwayssuchasextracellular regulatedkinases,

b

-arrestin,andionchannels.Thereisadebate ontheexpressionlevelsofCB1andCB2indifferentcelltypesand itssubcellularlocalisation,butthereisageneralconsensusthat CB1iswidelyexpressedwithintheCNS incorticalneuronsand interneurons, astrocytes, oligodendrocytes and oligodendrocyte precursor cells, aswell as in several leukocytesinﬁltrating the brain(NavarreteandAraque,2010;Galve-Roperhetal.,2013).CB1

regulates cognitive, memory and motor functions as well as analgesiaandsynapticplasticity.Onthecontrary,theexpressionof CB2 in cells of the CNS is more controversial,inasmuch as its presenceforalongtimewasexclusivelyassociatedtomicroglia, butmorerecentlyitwasalsodocumentedinbrainstemneurons andastrocytesuponcellularactivationbyaninsultorinﬂamma- tion (Van Sickle et al., 2005; Onaivi et al., 2006; Atwood and Mackie, 2010; et al., 2015a, 2015b; et al., 2015a, 2015b).

Furthermore, a substantial proportion of CB1 in the brain is intracellular, being found in vesicles (Leterrier et al., 2004), mitochondria(Bénardet al.,2012)orlysosomes(Rozenfeldand Devi,2008).Interestingly,thehumanbrainhasmoreCB1thanany otherGprotein-coupledreceptor.

2.2.Endogenousligandsofcannabinoidreceptorsandtheirmetabolic routes

Unsurprisingly, in view of the lipophilic nature of THC, the endogenous ligands of CB receptors,i.e. the endocannabinoids (eCBs),arederivativesofthelipidarachidonicacid(AA).Themost important eCBs are N-arachidonoylethanolamine (anandamide, AEA) and 2-arachidonoylglycerol (2-AG) (Table 1). These two endogenousligandsofCB1andCB2arethebest-studiedmembers ofanever-growingfamilyofcompounds(Maccarroneetal.,2014;

Mechoulam et al., 2014) that include N-acylethanolamines, 2- monoacylglycerols,andnowalsopeptidiceCBs(pepcans)(Bauer etal.,2012).Thislargefamilyalsoincludestherecentlydiscovered ethanolaminesof

v

-3 (n-3)fattyacid, namelyN-eicosapentae- noylethanolamine(EPEA) and N-docosahexaenoyl-ethanolamine (DHEA)(Artmannetal.,2008;Lucanicetal.,2011).EPEAandDHEA act indeed as CB1/CB2 agonists (Brown et al., 2010), but their functionstillneedstobeclariﬁed.Thereareadditionalcompounds thatdonotbindtoCB₁and/orCB₂,butareprocessedbythesame synthesizinganddegradingenzymesasauthenticeCBs;theseso- called “eCB-like” substances include N-palmitoylethanolamine (PEA)andN-oleoylethanolamine(OEA)(Uedaetal.,2013).

Both AEA and 2-AG are often found together, but their individual levels vary between species, tissues, developmental stages and pathophysiological conditions. Within the CNS and amidstglialandneuronal cells,AEAand2-AGactasretrograde messengers,i.e.theyarereleasedfromsomataand/ordendritesof neurons, and then act on afferent axon terminals or nearby astroglial processes to inhibit neurotransmitter release (Ohno- Shosaku et al., 2001; Wilson and Nicoll, 2001). Depending on neurontype glutamatergicor

g

-aminobutyric acid(GABAergic), activationof CB₁ mayresults in inhibition or activation of the neuronal circuit. Unlike other neurotransmitters, eCBs are not storedinvesicles,butareproduced“ondemand”(i.e.,whenand

whereneeded)upondifferentbiologicalstimuli,toactparacrinally orautocrinally.AtleastﬁvedifferentpathwayscangenerateAEA (DiMarzoetal.,1994;Leungetal.,2006;Liuetal.,2006),themost studiedbeingthereleaseofAEAfrommembraneprecursorsviaN- acylphosphatidyl-ethanolamine(NAPE)-speciﬁcphospholipaseD (NAPE-PLD)(Okamotoetal.,2004).CleavageofAEAintoAAand ethanolamine ismainlyoperatedby fattyacidamidehydrolase (FAAH)(Cravattetal.,1996).Ontheotherhand,2-AGisformedby the action of two diacylglycerol lipases, DAGL

a

and DAGL

b

(Bisogno et al., 2003), and is primarily degraded into AA and glycerolbymonoacylglycerollipase(MAGL)(Dinhetal.,2002)and toalesserextentby

a

/

b

-hydrolasedomain(ABHD)6andABHD12 (Blankmanetal.,2007).Furthermore,AEAand2-AGcanalsobe metabolized bycyclooxygenase-2 (COX-2), several lipoxygenase isozymes and by cytochrome P450, generating oxidized com- pounds like prostaglandin-ethanolamides and glyceryl esters, hydroxy-anandamidesandhydroxyeicosatetraenoyl-glycerols,respectively.AllthesesubstancescanbindtoCB1andCB2,andare endowedwithdistinct,yetunclear,biologicalactivities(vander Steltetal.,2002;RouzerandMarnett,2011).AlleCBscanalsobe inactivatedthroughtheiruptakebyapurported“endocannabinoid membranetransporter”(EMT),whosemolecularidentiyhasyetto beidentifiedbutwhosepresencehasbeensuggestedbyseveral reports as well as by its in vitro and in vivo pharmacological targeting(Chiccaetal.,2017).Overall,theactivityandfunctionof eCBsiscontrolledbytheirendogenoustone,whichdependsona finely regulated balance between biosynthetic and degradative pathways.Atvariancewiththispicturemorerecentlyithasbeen demonstrated that AEA can be stored in lipid droplets (adipo- somes)andboundtointracellulartransporters(Maccarroneetal., 2010). These additional players seem to be relevant, because neurophysiologicalactionsofeCBsdependonthespecificlocation oftheproteinsthatsynthesize,transport,bindanddegradethem, altogethercalled“eCBsystem”(Maccarroneetal.,2014,2015).The complexityanddistinctdistributionofeCBsystemcomponentsin pre- and post-synaptic neurons, as well as in microglia and astrocytes,isschematicallydepictedinFig.3.

2.3.Otherreceptors

InadditiontoCB1andCB2,cannabinoidsandeCBscanexert their effects engaging other non-CB receptors, including the transientreceptorpotentialvanilloid1(TRPV1)channel,expressed mainlyinperipheraltissues,includingperipheralsensoryneurons, epithelialandendothelialcellsaswellasimmunecells(Xiaetal., 2011).Othertargetsareperoxisomeproliferator-activatedrecep- tors(PPAR)

a

and

g

(PistisandMelis,2010)thatbelongtoafamily of nuclear receptors whose role is mainly to control lipid metabolism; and orphan G protein-coupled receptor GPR55 (Moriconi et al., 2010). The existence of such supplementary moleculartargetssupporttheviewthatanextensionoftheterm

“cannabinoidreceptor” shouldreallybeencouraged.The multi- plicityofcannabinoidsandeCBsatinteractingwithmanyother non-CBreceptorsandtheevidencethatcannabinoidreceptorsare abletobindtootherligandsotherthancannabinoidsand eCBs endorsesuchview.

3.Therapeuticexploitationoftheendocannabinoidsystemin multiplesclerosis

3.1.PreclinicalexvivoandinvivostudiesofeCBsignalinginMS

Toestablishwhetheradrugisreadyforclinicaltrials(theso- calledmovefrombenchtobedside)involvesextensivepreclinical studiesthatyieldpreliminaryefﬁcacy,pharmacokinetic,safetyand toxicological information. The starting point of such studies is

xxx–xxx

(6)

often the evidence of a signiﬁcant alteration of activity or expression of a certain molecular target (often a protein) in diseasemodels,orinaffectedpatients,eitherinvitro(experiment and analysis performed outside a living organism) or ex vivo (experiment performed inside and analysis outside a living organism).In thisline, accumulatedexvivo evidenceshoweda widealterationofCB1andCB2expressioninseveralexperimental modelsofMSandinpatientsaffectedbydifferentclinicalformsof thedisease(Benitoetal.,2003,2007;Cabranesetal.,2005;Loría etal.,2008;Jean-Gillesetal.,2009;Chiurchiùetal.,2013;Sánchez López et al., 2015; Maccarrone et al., 2017). This observation stronglysuggestsafunctionalroleforcannabinoidreceptorsinMS pathology, and has stimulated preclinical studies directed at modulating their activity by in vivo administration of plant cannabinoids,syntheticagonistsorantagonists(Table2).Theﬁrst such preclinical study demonstrated an amelioration of both tremorandspasticityinmicesufferingfromchronicrelapsingEAE uponinvivotreatmentwithTHC,methanandamide(astableand non-hydrolizableAEAanolog),andsyntheticCB1orCB2agonists (Bakeretal.,2000).SelectiveantagonistsofCB1orCB2inhibited theseeffects, indicating thattheeCBsystem mayparticipatein the control of MS. In addition to symptom management, cannabinoids also conferred neuroprotection via CB1, slowing down the neurodegenerative process that eventually leads to chronicdisability(Pryceet al.,2003).Furthermore,theycaused immunosuppression-mediatedameliorationofsymptomprogres- sion and promoted remyelination also in a viral model of MS (CroxfordandMiller,2003;Arévalo-Martínetal.,2003).SuchCB

receptor-mediated neuroprotection and interference with MS progression were paralleled by downregulation of adhesion moleculesandsubsequentinhibitionofTlymphocyteinﬁltration andmicroglialresponses(Mestreetal.,2009;Zhangetal.,2009;

Kozelaetal.,2011),aswellasbyrestorationofself-toleranceto myelinantigen(Arévalo-Martínetal.,2012).Selectiveactivation ofCB2inautoreactiveT-cellswascrucialtocontrolinflammation in EAE (Maresz et al., 2007)by reducing highly pathogenicT- helper 17 differentiationand immune cellaccumulation in the CNS, including activated macrophages/microglia (Kong et al., 2014). In line, selective activation of CB2 on T lymphocytes isolatedfromMSpatientsinhibitedcellproliferationandimmune responseswithoutinducingcelldeath(Malfitanoetal.,2013).The involvement of CB receptors was confirmed not only by pharmacological modulation of theiractivity but also by their genetic deletion. Accordingly, mice whose CB1 was genetically ablatedpoorlytoleratedinflammatoryandexcitotoxicinsultsand developed substantial neurodegeneration (Pryce et al., 2003;

Rossietal.,2011a;Musellaetal.,2014).Similarly,specificdeletion ofCB2onencephalitogenicTcellsincreasedproliferation,reduced T-cell apoptosis, elevated production of proinflammatory cyto- kinesandworsenedthesymptoms.Thisscenariosuggestedthat activation of CB2 on T-cells by eCBs was sufficient to control autoimmuneinflammation,and possiblytoslowdownprogres- sionofthedisease(Mareszetal.,2007).Thesestudiesindicatethe involvementofbothCB1andCB2inthetherapeuticefficacyofthe eCB system modulation. They also suggest a differential role played by these two receptor sub-types, whereby control of Fig.3.AmodernviewoftheeCBsystemintheCNS.eCBsignalingisorchestratedbytargetreceptors(CB1,CB2,GPR55,TRPV1andPPARs),biosynthetic(NAPE-PLDandDAGL) anddegradative(FAAH,MAGL,ABHD6andABHD12)enzymes,transmembranetransportmechanisms(liketheputativeEMT),intracellulartraffickingbyAITslikefattyacid bindingproteins,heatshockprotein70,andFAAH-likeAEAtransporter,aswellasbystorageorganelles(adiposomesorlipiddroplets).Altogether,theseproteinsregulatethe endogenoustoneofeCBs,andhencetheirbiologicalactivity.ForsomeoftheelementsoftheeCBsystemadistinctdistribution,bothintracellularlyandamongpre-andpost- synapticneurons,microgliaandastrocytes,hasbeendocumented.Itshouldbenotedthat,unlikeothereCB-bindingreceptors,CB1appearstobelocatedincholesterol- enrichedmembranemicrodomainstermedLRs,andthatCB2isexpressedinneuronsmainlyuponbraininjury.Moreover,theroleinneuroinflammatorydiseasesofeCB systemelementsshownherebutnotlistedinTable1remainstobeelucidated.

Abbreviations:ABHD6/12,a-b-hydrolasedomain6/12;AITs,AEAintracellulartransporters;CB1/CB2,G-proteincoupledtype-1andtype-2cannabinoidreceptors;DAGL, diacylglycerollipasea/b;eCBs,endocannabinoids;EMT,putativeendocannabinoidtransmembranetransporter;FAAH,fattyacidamidehydrolase;GPR55,G-proteincoupled receptor55;LRs,lipidrafts;MAGL,monoacylglycerollipase;NAPE-PLD,N-acylphosphatidylethanolamine-speciﬁcphospholipaseD;PPARs,peroxisomeproliferator- activatednuclearreceptors;TRPV1,transientreceptorpotentialvanilloid1channels.

xxx–xxx

(7)

spasticityismostlymediatedbyCB1,whileCB2seemstomodulate overactive and exacerbated immune responses by (Pryce and Baker,2007;Palazuelosetal.,2008).

MuchalikeCB1andCB2,metabolicenzymesofeCBshavebeen foundtobesigniﬁcantlyalteredinbothMSpatients(Bakeretal., 2001;Cabranesetal.,2005;Benitoetal.,2007;Centonzeetal., 2007a,b;DiFilippoetal.,2008;Jean-Gillesetal.,2009;Chiurchiù etal.,2013;SánchezLópezetal.,2015)andanimalmodelsofthe disease(Bakeretal.,2001;Pryceetal.,2003;Wittingetal.,2006;

Loríaetal.,2008)(Table2).ThesealterationsofeCBsandoftheir metabolismarestillsomewhatcontroversial,mostlikelybecause of different experimental models used and variations in MS patients’recruitmentandstratification.Despitethesecontrover- sies, theanti-inflammatoryand neuroprotective role of eCBs is widelydocumentedinbothinvivoandexvivoexperimentsoncells obtainedfromeitherexperimentalmodelsofMSorfromauthentic MS patients(Wittinget al.,2006; Mestre et al., 2005; Ortega- Gutiérrezetal.,2005;Loríaetal.,2010).Exogenousadministration ofeCBsimprovedcontrolofspasticity(Bakeretal.,2000,2001)and protected neurons from inflammatory damage possibly by engagingCBreceptorsonmicroglial cells (Eljaschewitschetal., 2006).Additionally,inhibitionofendogenousAEAuptakeandofits degradationbyFAAHwasshowntoamelioratemotorsymptoms (Bakeretal.,2001;deLagoetal.,2004).Thisimprovementwas

associatedwithreducedinﬂammatoryresponsesinthespinalcord (Mestreetal.,2005)alongwithdownregulationofmacrophage and microglial function (Mestre et al., 2005; Ortega-Gutiérrez et al., 2005). In a virus-induced model of MS, AEA inhibited microglial activation by lowering interleukin (IL)-23 and IL-12 release(Correaetal.,2011),aswellasIL-1

b

andIL-6(Hernangó- mez et al., 2012). Inhibition of the last two pro-inﬂammatory cytokineswasmediatedbyamechanisminvolvingCB2receptor- mediatedrecoveryoftheinteractionbetweenCD200(expressedin neurons)anditsreceptor(expressedinmacrophages/microglia).

SuchaninteractionappearstobedisruptedinMS,thusexposing neurons to macrophages/microglia-induced neurotoxicity and cytokine release (Hernangómez et al.,2014).Recently, we have demonstratedthattheanti-inflammatoryroleofAEAinMSiscell type-specific,sinceitinhibitscytokineproductiononlyinmyeloid dendritic cells (Chiurchiù et al., 2013). Instead, plasmacytoid dendritic cells were unresponsive to AEA due to a significant upregulation of FAAH compared tohealthy subjects(Chiurchiù et al., 2013). Interestingly, chronic and long-term inhibition of FAAH (via genetic ablation) resulted in clinical remission and improved long-termoutcome in EAE mice(Webb etal., 2008).

Further,theacutephaseofMSwasamelioratedbyincreasedlevels of2-AGthatinhibitedspasticitywhengivenat10mg/kg(Baker etal.,2001)andcausedadelayedonsetinacuteandchronicEAE Table2

AlterationsofdistinctelementsoftheeCBsystem,andtheirroleininﬂammationandneurodegenerationinMS.

ECS element

Model Sample Variation Effects Reference

AEA ChronicEAE Brain,spinalcord " Earlyinhibitionofspasticity Bakeretal.(2001)

LewisEAErats Brain # Worseningofdiseasedevelopmentand

neurologicalimpairment

Cabranesetal.(2005)

RR-MS Autopsiedbrain " Microglia-inducedneuroprotection Eljaschewitschetal.(2006)

EAEandRR-MS patients

Brain,CSF,plasma,Tcells " Neuroprotection Centonzeetal.(2007a)

RR-MS,SP-MS CSF # – DiFilippoetal.(2008)

RR-MS,PP-MP, SP-MS

Plasma " Diseaseprogression Jean-Gillesetal.(2009)

RR-MS Tcells,Bcells,NKcells " – SánchezLópezetal.(2015)

NAPE-PLD/

FAAH

EAEandRR-MS patients

Brain,CSF,plasma,Tcells "NAPE-PLDand

#FAAH

Neuroprotection Centonzeetal.(2007a)

SP-MS Plasma #FAAH Diseaseprogression Jean-Gillesetal.(2009)

RR-MS mDCandpDC #FAAHinmDC

and"inpDC

Lackofimmunoregulation Chiurchiùetal.(2013)

RR-MS Tcells,Bcells,NKcells ! – SánchezLópezetal.(2015)

2-AG ChronicEAE Brain,spinalcord " Lateinhibitionofspasticity Bakeretal.(2001)

LewisEAErats Brain # Worseningofdiseasedevelopment

andneurologicalimpairment

Cabranesetal.(2005)

RR-MSpatients CSF ! – Centonzeetal.(2007a)

RR-MS,SP-MS CSF # – DiFilippoetal.(2008)

TMEV-IDD Spinalcord " – Loríaetal.(2008)

RR-MS Tcells,Bcells,NKcells IncreasedinNK cells

– SánchezLópezetal.(2015)

DAGL/

MAGL

EAE – – InhibitionofMAGLameliorates

EAEprogression

Hernández-Torresetal.

(2014)

CB1 LewisEAErats Brain # Worseningofdiseasedevelopmentand

neurologicalimpairment

Cabranesetal.(2005)

P-MS Plasma " Diseaseprogression Jean-Gillesetal.(2009)

MSplaques Neurons,oligodendrocytes, inﬁltratedTcells

" Diseaseprogression Benitoetal.(2007)

RR-MS Tcells,Bcells,NKcells "inTcells – SánchezLópezetal.(2015)

CB2 TMEV-IDD Spinalcord " – Loríaetal.(2008)

P-MS Plasma " Diseaseprogression Jean-Gillesetal.(2009)

MSplaques InﬁltratedTcells,astrocytes, microglia

" Diseaseprogression Benitoetal.(2007)

RR-MS mDCandpDC "inmDCand

!inpDC

Lackofimmunoregulation Chiurchiùetal.(2013)

RR-MS Tcells,Bcells,NKcells Increasedin Bcells

– SánchezLópezetal.(2015)

CSF,cerebrospinalﬂuid;EAE,experimentalautoimmuneencephalomyelitis;mDC,myeloiddendriticcells;RR,relapsing-remitting;P,progressive;pDC,plasmacytoid dendriticcells;PP,primaryprogressive;SP,secondaryprogressive;TMEV-IDD,Theiler'smurineencephalomyelitisvirus-induceddemyelinatingdisease.",increase;#, decrease; !,unchanged.

xxx–xxx

(8)

modelswhen administeredintraperitoneallyat100

m

g.Thelast effectwasprobablymediatedbyrecruitmentofanti-inﬂammatory macrophages(Lourbopoulosetal.,2011).

BasedontheseeCB-mediatedneuroprotectiveactions,selective inhibitors of FAAH and MAGL, that degrade AEA and 2-AG respectively, were used in vivo to increase the tone and thus potentiate the effects of these eCBs. The blockade of AEA degradationbyseveralpotentandselectiveFAAHinhibitorswas infactabletocontrolspasticityinBiozziABHmice(Pryceetal., 2013),anothermousemodelofMSthatreproducesbothrelapsing- remittingcourseandsecondaryprogressiveneurologicalaspects (Amoretal.,2005).Likewise,MAGLinhibitorscontrolledspasticity (Pryce et al., 2013), preserved myelin integrity by suppressing oligodendrocyteexcitotoxicityand microglialactivation(Bernal- Chicoet al.,2015), and slowed the clinical progression of EAE (Hernández-Torresetal.,2014;Brindisietal.,2016).

Several preclinical studies have recently shown that eCB signalling controls synaptopathy, another hallmark of MS and neuroinflammation. Loss of homeostaticcontrol of neurotrans- missionandsubsequentsynapsedysfunctionisdeeplyaffectedby inflammation.Excitatory(glutamatergic)andinhibitory(GABAer- gic)synapses areprivileged sites of action for eCBs and act in concertwithproinflammatorycytokines(inparticularIL-1

b

and tumournecrosisfactor-

a

,TNF-

a

).Neuroinﬂammationunbalances synaptictransmissiontowardsexcitation,thatultimatelyresultsin excitotoxic neurodegeneration (Mandolesi et al., 2015), and activation of CB1 dampens the enhancing effects of TNF-

a

on postsynapticglutamatereceptor expressionand function(Rossi et al., 2011b), and reduces the frequency of spontaneous glutamate-mediated synaptic currents in striatal neurons, thus contrasting the effects of IL-1

b

on glutamate release from presynapticnerveterminals(Musellaetal.,2014).Inthiscontext, geneticdeletionofCB1inglutamatergicorGABAergicneuronsof EAE mice provided a conclusive demonstration that distinct subsets of these receptors are involved in both modulation of TNF-

a

-mediatedpostsynapticexacerbationofglutamatecurrents, andIL-1

b

-dependentpresynaptic increaseof glutamate release inducedinGABAergicprojectionneuronsofthestriatum(Musella et al.,2014).Moreover, IL-1

b

inhibitsGABA synapses, and this effectfurthercontributes totheexcitotoxicdamage ofneurons duringEAEand MS(Mandolesietal.,2015).In keepingwithits homeostaticroleinneuroinﬂammatorydisorders,CB1onGABAer- gicnerveterminalsisselectivelydownregulatedinthestriatumof EAEmice,sothatincreasedeCBlevelsintheseanimalsonlyreduce excitatorytransmissionwithoutaffectingGABAsignalling (Cen- tonzeetal.,2007a,b).

Ontheotherhand,AEAcanalsobindtothetransientreceptor potentialvanilloid type-1(TRPV1) ion channel (Zygmuntet al., 1999), whose activation may negatively impact on neuronal survival.TRPV1alsomodulatesthesynapticeffectsofTNF-

a

andof IL-1

b

,hinderingthesynapticdeﬁcitsexertedbyTNF-

a

inthepeak phaseofEAEand,inthechronicstages,enhancingIL-1

b

-induced GABAergicdefects(Musumecietal.,2011).Thesedatasuggestthat theaversive or protectiveeffects of TRPV1 on MS neuronsare strictlydependentontheinflammatorymilieu.Althoughitisstill unclear which is the main cell type responding to central inflammationviaeCBrelease,neuralprecursorcellsappeargood candidates,becausetheyrespondtoinflammatorystimuliandto sensitized TRPV1 receptors (Stock et al., 2014), and limit inflammatorysynapticdamagethroughAEArelease(Buttietal., 2012).ThemechanismsbywhichtheseelementsoftheeCBsystem modulatethepresynapticandpostsynapticperturbationsinduced byinflammationaredepictedinFig.4.

Takentogether,preclinicaldatastronglysuggestthatmodula- tion of the eCB system is actively involved not only in the management of symptoms but also in neuroprotection and

suppressionofinflammation.Theseobservationsaregenerating considerable enthusiasm for bringing cannabinoid/eCB-based medicines into clinical practice for the treatment of both relapsing-remittingandprogressiveformsofMS.Afurthersupport tothisperspectivecomesalsofromtheexperimentaldemonstra- tionthatplant-derivedcannabinoidslimitinflammation-induced neurodegenerationthat drivesprogressivedisabilityduringEAE (Pryceetal.,2015).Forexample,intwodifferentmurinemodelsof MS, Sativex slowed downMS progression by improving motor activity,reducingCNSinfiltratesandmicroglialactivity,together withpromotingmyelinrepair(Feliúetal.,2015)andimprovingthe neurologicaldeficitsandcellinfiltrateswithintheCNS.Thiseffect wasattributedtoTHCactivationofCB1receptorsandnottoCBDin theEAE model (Moreno-Martetetal., 2015), while in theviral model both THC and CBD were effective. However CBD, the cannabinoidthatisdevoidofpsychoactiveactivityandpotentially safe, has also been investigated as an effective alternative for alleviating neuroinflammation and neurodegeneration (Maccar- rone et al., 2017).CBD ameliorates EAE clinicalsymptoms and reducesaxonaldamage andinflammation,aswell asmicroglial activation, T-lymphocyterecruitmentin thespinal cord(Kozela et al., 2011) and protection of oligodendrocyte progenitorcells fromneuroinflammation-inducedapoptosis(Mechaetal.,2012).

Theseanti-inﬂammatoryandneuroprotectiveactionsofCBDare long-lasting (Mechaet al.,2013),counteract neuronalapoptosis (Giacoppo et al., 2015) and antagonistically interact with PEA (another non-psychotropic compound) in protecting against neurodegenerationandneuroinﬂammation(Rahimietal.,2015).

Thisisprobablyduetoacross-talkbetweentheCBD-inducedPI3K/

Akt/mTORpathwayand upregulationofthePEA-bindingPPAR

g

(Giacoppoetal.,2017).ThemostrecognizedeffectsofCBDarethe immunoregulatoryones,namely(i)suppressingproinﬂammatory Th17 responses,(ii) promoting Tcell exhaustion/tolerance,(iii) enhancingIFN-dependentanti-proliferativeprograms,(iv)ham- peringantigenpresentation,and(v)inducingantioxidantmilieu thatresolvesinﬂammation(Kozelaetal.,2016).

3.2.eCBsignalingfrompreclinicalmodelsofMStoclinical applications

A great amount of preclinical studies have convincingly demonstrated neuroprotective, neuromodulatory and immuno- modulatory role of cannabinoids and eCBs, and theirability to alleviatesymptomsand behaviouralabnormalitiesinmodelsof neurological diseases (Baker et al., 2001; Pryce et al., 2003;

Centonzeetal.,2007b;Chiurchiù,2014;Chiurchiùetal.,2015a).

Various cannabinoid preparations have also been examined in clinicalstudiestoassesstheirefﬁcacyinhumans.Thus,distinct pharmaceuticalpreparationsofcannabinoidagentshavebeenor arebeing testedin numeroushumandiseases and, besides the aforementionedSativex¹andCBD,theseincludenaturalcannabi- noidorphytocannabinoidssuchasBedrocan¹(19%THC,<1%CBD from Cannabis sativa), Bedrobinol¹ (12% THC,<1% CBD from Cannabissativa),Bediol¹(6%THC,7,5%CBDfromCannabissativa), Bedica¹(14%THC,<1%CBDfromCannabisindica),Bedrolite¹(19%

THC,<1%CBD from Cannabis sativa), and Bedropuur¹ (20–24%

THC,<1%CBDfromCannabis indica).In addition,also synthetic cannabinoids, such as Dronabinol (Marinol¹, a synthetic THC), Nabilone (Cesamet¹, a THC analog), and Levonantradol (a syntheticanalogofDronabinol)havebeenusedinhumans.These preparationscanbetakenperosmixedwithfoodormadeintotea, orbysublingualortopicaladministration,ortheycanbesmoked orinhaled.However,theseformulationsareextremelyheteroge- neous in nature, making the results of the clinical studies incomparable, and contributing to inconclusive evidence of cannabinoidefﬁcacyinmostclinicalstudies.

xxx–xxx

(9)

A recentlypublished systematic reviewanalyzed safety and efficacy results of randomized clinical trials that compared cannabinoids to placebo in three neurological disorders (MS, epilepsyandmovementdisorders)publishedinthepreceding50 years(1948–2013interval)(Koppeletal.,2014;GlossandVickrey, 2014). According to the method of the American Academy of Neurology, quality evidence was rated as A (established with strongevidenceaseffective,ineffectiveorharmful,supportedby2 ClassIstudies),B(establishedwithmoderateevidenceasprobably effective,ineffectiveorharmful,supportedby1ClassIstudy),C (establishedwithweakevidenceaspossiblyeffective,ineffective or harmful, supported by 1 Class II study), or U (insufficient, inadequate or conflicting data to make a determination). This ratherstringentanalysisapproachshowedthatefficacyevidence ofthepublishedstudiesisgenerallyweak,inclinicalconditions otherthanMS-associatedspasticityandneuropathicpain,andin thecaseofpreparationsotherthannabiximols(Koppeletal.,2014;

Maccarroneetal.,2017).

Asreportedinanotherrecentandrigoroussystematicreview and meta-analysis considering ten diverse clinical conditions, cannabis-basedmedicationsweregenerallyfoundtobeeffective forthemajorityoftheexploredclinicaloutcomes(Whitingetal., 2015). In the presented meta-analysis, only randomized trials comparingcannabinoidpreparationswithnotreatment,placebo orstandard carewereconsidered,and non-randomized clinical studieswereconsideredonlyifatleast25patientswereincluded (Whitingetal.,2015).Depressionwastheonlyclinicalconditions in which placebo was reported to be more effective than nabiximolsin threeclinicalstudies,whilecannabis-basedmed- icationswereconﬁrmedtobemoreeffectivethanplaceboonthe

primary outcomes of the selected studies in conditions like spasticityduetomultiplesclerosisorparaplegia,neuropathicpain, cancerpain,butalsoanorexiasecondarytoHIV/AIDS,nauseaand vomitingduetochemotherapy,sleepdisorders,anxiety,psychosis andTourettesyndrome(Whitingetal.,2015).

Despitethelimitationsandpoorqualityofmostclinicaltrials, thepositiveresultsofsomerigorouslyconductedclinicalstudies andtheacceptablesideeffectsofmostcannabinoidpreparations (themostcommonadverseeffectsaredizziness,drowsinessand disorientation), led to the approval of Nabiximol as the only cannabis-baseddrugavailableinCanada,NewZealandandseveral European andAsiancountries(butnotyetintheU.S.A.)forthe controlofspasticityandneuropathicpainassociatedwithMS.

Therearefewclinicaltrialsassessingefﬁcacyandpharmaco- logicalmodulationofcannabinoidsandeCBsasdiseasemodifying treatments(Table3).TheCAMSstudy(Zajiceketal.,2003,2005;

Katonaetal.,2005),failedtoreportanyeffectofcannabinoidson serumlevelsofproinﬂammatorycytokinesandonTcellactivation.

The CUPID study (Zajicek et al., 2013) showed that Marinol¹, (approvedasanantiemeticandanorexicdrug)hasnooveralleffect ontheprogressionofMSeveninalong-termfollow-upstudy(Ball etal.,2015).Furthermore,atwo-foldcrossoverstudymeasuring immune function in progressive MS patients revealed a pro- inﬂammatoryroleofDronabinol(Killesteinetal.,2003).Although MS progression and associated grey matter damage are both acceleratedin humansubjectscarrying ageneticvariantofCB1

withreducedmembraneexpression(Ramiletal.,2010;Rossietal., 2011a, 2013), clinical trials designed to investigate the neuroprotective effectof cannabinoid-based drugs didnot showany impact onmodulationof disease progression.In summary,the Fig.4.OverallschemeofthepresynapticandpostsynapticperturbationsmediatedbyproinﬂammatorycytokinesandmodulatedbyeCBsysteminEAE.IL-1bincreases glutamatereleaseatpresynapticterminalsandTNFinducesAMPAreceptorupregulation,resultinginenhancedglutamatetransmission.CB1contraststheeffectsofIL-1bby reducingthefrequencyofspontaneousglutamate-mediatedsynapticcurrentsonpresynapticterminals,converselyTRPV1ispermissiveforIL-1bsynapticeffectson glutamatetransmission.Atthepostsynapticsite,bothCB1andTRPV1restrainTNF-mediatedpotentiationonpostsynapticAMPAreceptor.Moreover,IL-1bpromotesthe inhibitionofCB1functiononGABAergicsynapses,thusmitigatingthereductionofGABArelease.Moreover,IL-1breducespostsynapticGABAAreceptorfunctionby promotingthedecreaseofGABAsignaling.Finally,TRPV1channelsarepermissiveforIL-1bsynapticeffectsatbothpre-andpostsynapticsites.

xxx–xxx

(10)

results obtained from different clinical trials suggest that cannabinoids areable to control spasticity in MS patients, but do not confer neuroprotection in progressive MS. Curiously, notwithstandingthepotentanti-inﬂammatoryandneuroprotec- tiveeffectsofCBDanditslowpsychoactivity,allclinicaltrialsfor MS entailformulations with mixtures of THC and CBD but no formulation with only CBD. GW pharmaceuticals is seeking approvalfrom theFDA tomake Epidiolex (an oral formulation of 99% pureplant-derived CBD) commercially available for the treatmentofseveralformsofepilepsiesandepilepsy-associated raresyndromes;however, noclinicaltrialswiththis drughave beenrunorarebeingplannedforMS.

As for drugs speciﬁcally targeting CB receptors, after the withdrawal of Rimonabant (a CB1 inverse agonist) for its potentiallyserious side effects, alternative strategies would be eithertodevelopnovelCB1ligandsthatareperipherallyrestricted (i.e.,unabletocrosstheblood-brainbarrier)and/orexhibitneutral antagonism, or to eventually bring CB2 agonists into clinical practice.Allthesestrategiesaresubjectofintenseresearch.Several clinicaltrialshavebeenreportedorareongoingwithMAGLand FAAHinhibitors(seeforarecentreviewChiccaetal.,2015).For example,theirreversibleMAGLinhibitorABX-1431hassuccess- fullycompletedphase1aclinicaltrialsandisnowbeingevaluated forTourettesyndrome(phase1b)andisscheduledforaphase2 studyinMSpatients.TheFAAHinhibitorPF-04457845exhibited alsoafavourablesafetyproﬁleinphase1and2clinicaltrials,but was ineffective in treating pain in osteoarthritis patients(Ahn et al., 2011; Li et al., 2012). The compound is currentlybeing evaluatedforthetreatmentofTourettesyndrome,fearresponses and cannabis withdrawal. The development of FAAHinhibitors sufferedamajorsetbackwiththephase1clinicaltrialofBIA10- 2474,whichledtothedeathofonevolunteerandmild-to-severe neurological symptoms in four others (Kerbrat et al., 2016).

Recently,itwasshownthatBIA10-2474isapromiscuousdrugthat inhibitsmultiple lipases,including Neuropathy TargetEsterase, thatarenottargetedbythehighlyselectiveandtheclinicallysafe FAAHinhibitorPF-04457845(vanEsbroecketal.,2017).BIA10- 2474,but not PF-04457845, produced substantialalterations in lipidnetworksinhumancorticalneurons,therebysuggestingthat BIA10-2474hasthepotentialtodysregulatemetabolicpathways inthenervoussystem(vanEsbroecketal.,2017).Itis,therefore, highlyunlikelythatinhibitionofFAAHcausedtheadverseeffects

in thehealthy volunteers.None of these drugs, norany other selectiveinhibitorofthedegradationoftheothermajoreCB,2-AG, hassofarbeenusedinclinicalpracticeinrelapsingorprogressive formsofMS.

4.eCBsignalingasapotentialtherapyforother

neurodegenerativediseaseswithaninﬂammatorycomponent

Despitetheaetiology(whereknown)andthepathogenesisof other neuroinflammatory diseases such as Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease (HD) and amyotrophiclateral sclerosis(ALS) are divergentand heterogeneous, it can be stated that neuroinflammation is a common hallmarks for all of them. Of note, all these disorders are characterized by a progressive loss of structure or function of neuronsfoundatspecificcircuitriesandbrainareas.Theyallshow subcellular histopathological similarities that include atypical protein deposits and cell death. In addition, the term neuroinflammation classically involves infiltration of peripheral im- munecellswithintheCNS,butitscurrentusehasexpandedto includethoseneurodegenerativediseaseswhere,althoughleuko- cyterecruitmentis notacritical feature,cellularandmolecular signs of inflammation (i.e., expression of cytokines and other inflammatorymediatorsaswellasglialactivation)notonlyare presentbutarelikelytocontributetodiseasepathogenesisand progression.

The role of the eCB system in these disorders has been extensively reviewed in recent papers (Chiurchiù et al., 2015a;

Centonzeetal.,2007b;DiMarzoetal.,2015;Fernández-Ruizetal., 2015).SimilarlytoMS,perturbationsofdifferentcomponentsof theeCBsystemhavebeenreported,beingCB1,CB2andeCBlevels eitherincreasedordecreasedindifferentbrainareas(especiallyfor CB1),speciﬁccelltypesordiseasestages(Centonzeetal.,2007a,b;

Westlake etal.,1994; Benito et al., 2003;Lastres-Becker et al., 2003;Wittingetal.,2004;Horneetal.,2013;Ramirezetal.,2005;

vanderSteltetal.,2006;Yiangouetal.,2006;Battistaetal.,2007;

Zhaoetal.,2008;Pisanietal.,2011;Jungetal.,2012;Barietal., 2013;Altamuraetal.,2015)(Table4).Forinstance,stimulationof CB1leadstoagenerallyreducedglutamate-inducedexcitotoxicity and improved neuronal viability (van der Stelt et al., 2001;

Marsicanoetal.,2003;Bilsland,2006),whereasanincreaseinthe expressionofCB2 isusuallyassociated withactivatedmicroglia Table3

Outcomesandresultsofcannabinoid-basedclinicaltrialsinMS.

Trialname(source) Studydesign Disease No. Formulation Duration Main/otherclinical outcomes

Results

CAMSstudyZajicek etal.(2003)

MC,PC,randomized RR-MS, P-MS

630 OralTHC,THC/

CBD,Placebo

15weeks Changeinoverall spasticityscores

Nosigniﬁcantimprovementof spasticity

Admissionforrelapses Reductionofadmissionsforrelapsesin thetreatedgroups

Killesteinetal.(2003) Two-foldcrossover P-MS 16 OralTHC/

Cannabissativa extract

4 weeks+4 weeks

Immunomodulation IncreaseofTNF-ainLPS-stimulated wholebloodinthegrouptreatedwith CannabissativaandincreasedIL-12p40 inpatientswithhighadverseevents score

CAMSstudyfollowup Zajiceketal.(2005)

MC,PC,randomized RR-MS, P-MS

502 OralTHC,THC/

CBD,Placebo 12 months

Efﬁcacyandlongterm safetyinMS

Smalleffectonmusclespasticityand disability,measuredbyEDSSandRMI Admissionforrelapses Noreductionofadmissionforrelapses Katonaetal.(2005) MC,PC,randomized P-MS 100 OralTHC,THC/

CBD,Placebo

13weeks Immunomodulation Noevidenceforinﬂuenceonserum levelsofIFN-g,IL-10,IL-12and C-reactiveprotein

CUPIDstudyZajicek etal.(2013)

MC,DB,PC,randomised, parallel-grouptrial

P-MS 498 Dronabinol/

Placebo

3years Efﬁcacyoftreatmentin slowingMSprogression

Nooveralleffectontheprogression

NeworenlargingT2 lesionsornewT1lesions

Nodifference

Brainatrophy Nodifference

Abbreviations:No.,numberofsubjects;MC,multicentric;PC,placebo-controlled;DB,double-blind;EDSS,expandeddisabilitystatusscale;P,progressive;RMI,rivermead mobilityindex;RR,relapsing-remitting.

xxx–xxx