RESEARCH ARTICLE

Distinct Modulation of Human Myeloidand Plasmacytoid Dendritic Cells by

Anandamide in Multiple Sclerosis

Valerio Chiurchi�u, PhD, MS,1,2 Maria Teresa Cencioni, PhD, MS,2

Elisa Bisicchia, PhD, MS,1,2 Marco De Bardi, MS,2 Claudio Gasperini, MD,3

Giovanna Borsellino, PhD, MD,2 Diego Centonze, MD,2,4

Luca Battistini, PhD, MD,2 and Mauro Maccarrone, PhD, MS2,5

Objective: The immunopathogenesis of multiple sclerosis (MS) has always been thought to be driven by chronicallyactivated and autoreactive Th-1 and Th-17 cells. Recently, dendritic cells (DCs) have also been thought to signifi-cantly contribute to antigenic spread and to maturation of adaptive immunity, and have been linked with diseaseprogression and exacerbation. However, the role of DCs in MS pathogenesis remains poorly understood.Methods: We compared the level of cytokine production by myeloid DCs (mDCs) and plasmacytoid DCs (pDCs) inhealthy subjects and MS patients, following in vitro stimulation of Toll-like receptors 7/8. We also evaluated theeffect of the main endocannabinoid, anandamide (AEA), in these DC subsets and correlated cytokine levels withdefects in the endocannabinoid system.Results: mDCs obtained from MS patients produce higher levels of interleukin-12 and interleukin-6, whereas pDCsaccount for lower levels of interferon-a compared to healthy subjects. AEA significantly inhibited cytokine productionfrom healthy mDCs and pDCs, as well as their ability to induce Th-1 and Th-17 lineages. Moreover, we found that inMS only pDCs lack responsiveness to cytokine inhibition induced by AEA. Consistently, this specific cell subsetexpresses higher levels of the anandamide hydrolase fatty acid amide hydrolase (FAAH).Interpretation: Our data disclose a distinct immunomodulatory effect of AEA in mDCs and pDCs from MS patients,which may reflect an alteration of the expression of FAAH, thus forming the basis for the rational design of newendocannabinoid-based immunotherapeutic agents targeting a specific cell subset.

ANN NEUROL 2013;00:000–000

Dendritic cells (DCs) seem to play an important role

in the initiation and progression of multiple sclero-

sis (MS).1,2 The mechanisms that drive the conflicting

immunological outcomes of tolerance and immunity

have been the subject of intense scrutiny in recent years,

most recently in terms of how the various DC subsets

take part in these events.2–5 To date, in MS only a few

studies have directly investigated the 2 main DC subsets,

that is, myeloid DCs (mDCs) and plasmacytoid DCs

(pDCs), showing atypical numbers in periphery, altered

phenotype, and dysfunctional interaction with T cells;

however, their exact involvement in MS is not well

understood.6–8 Over the past decade, the endocannabi-

noid system (ECS) has emerged as a potential target for

MS management. Growing evidence suggests that endo-

cannabinoids may be neuroprotective during central

nervous system inflammation, and that these molecules

are endowed with important suppressive effects on

immune functions.9–14 Although their anti-inflammatory

role is getting a foothold among scientists, little evidence

View this article online at wileyonlinelibrary.com. DOI: 10.1002/ana.23875

Received Mar 17, 2012, and in revised form Jan 14, 2013. Accepted for publication Feb 20, 2013.

Address correspondence to Dr Maccarrone, Center of Integrated Research, Campus Bio-Medico University of Rome, Via Alvaro del Portillo 21, 00128

Rome, Italy. E-mail: 00128 m.maccarrone@unicampus.it

From the 1Department of Biomedical Sciences, University of Teramo, Teramo, Italy; 2European Center for Brain Research/Santa Lucia Foundation,

Rome, Italy; 3Department of Neurosciences, San Camillo Hospital, Rome, Italy; 4Department of Neurosciences, Tor Vergata University of Rome, Rome,

Italy; and 5Center of Integrated Research, Campus Bio-Medico University of Rome, Rome, Italy.

Additional Supporting Information may be found in the online version of this article.

VC 2013 American Neurological Association 1

shows these purported immunosuppressive effects among

specific immune cell populations of human blood. In previ-

ous studies, we have shown that MS patients display

increased levels of anandamide (AEA) in cerebrospinal fluid,

lymphocytes, and plasma, compared to healthy controls,15

and we found that AEA was able to strongly inhibit human

Th-1 and Th-17 immune responses.16 Furthermore, it was

recently reported that AEA downregulates interleukin

(IL)223 and IL-12 release by microglial cells in a viral

model of MS by a mechanism that also involves endoge-

nous production of IL-10.17 In this study, we sought to

assess the response of mDCs/pDCs obtained from MS

patients and healthy subjects to stimuli that mimic viral

infection, not only in terms of cytokine production but also

in terms of expression of relevant elements of the ECS.

Subjects and Methods

MS PatientsWe collected peripheral blood from 48 patients (34 females and

14 males), aged 18 to 68 years (mean 5 38.6 years), admitted to

the neurological clinic of the University Hospital Tor Vergata or of

the San Camillo Hospital of Rome, and later diagnosed as suffer-

ing from relapsing–remitting MS (RR-MS). The Disability Status

Scale scores were always <3, and the diagnosis of MS was estab-

lished at the end of the diagnostic protocol by clinical, laboratory,

and magnetic resonance imaging (MRI) parameters, and matched

published criteria.18 All RR-MS patients were treatment-naive, and

blood collection was performed at least 1 month after the last cor-

ticosteroid therapy. In total, 30 subjects were studied during exac-

erbations (Tor Vergata Hospital), and 18 during remissions (San

Camillo Hospital). Exacerbation was defined as the development

of new symptoms or worsening of a pre-existing symptom, con-

firmed by neurological examination, lasting at least 48 hours, and

occurring after a period of stability of �30 days. MS patients who

were clinically stable for at least 3 months prior to enrollment,

and who did not present Gd-enhancing lesions on MRI, were

considered to be in remission. Twenty-eight healthy subjects

(HSs), matched for age with MS patients (18 females and 10

males, mean age 5 34.1 years), with no history of autoimmune or

degenerative diseases of the central or peripheral nervous system,

were also enrolled in this study. All the subjects gave their written

informed consent to the study. The ethics committees of Tor Ver-

gata Hospital and of San Camillo Hospital approved the study.

Cell Preparation and PurificationPeripheral blood mononuclear cells (PBMCs) were isolated, after venous

puncture, from HSs and MS patients, and were separated by density gra-

dient over Ficoll-Hypaque (Pharmacia, Uppsala, Sweden), according to

standard procedures.16 For the identification and subsequent purification

of mDCs and pDCs, PBMCs from both HSs and MS patients under-

went a 5-color staining and then were sorted using a MoFLo Cell Sorter

(Beckman Coulter, Fullerton, CA). The correct identification of the dif-

ferent cell populations required multiparametric flow cytometry. There-

fore, the monocyte population was excluded as CD141 and HLA-

DR1, regardless of the different monocyte subsets expressing CD161.

HLA-DR2 and CD14dim granulocytes were also excluded, based also

on physical parameters. DCs were then gated as HLA-DR1 cells and

lineage-negative (Lin2) cells by excluding CD31 T lymphocytes,

CD191 B lymphocytes, and CD561 NK cells. Thus, HLA-DR1Lin2

cells were further divided into CD1231CD11c2 pDCs and

CD1232CD11c1 mDCs. This gating strategy allowed us to analyze

with confidence these 2 subsets of innate immune cells.19 In the Table, a

list of all the antibodies used for flow cytometry staining is given.

Quantitative Reverse Transcriptase PolymeraseChain ReactionRNA was extracted from sorted mDCs and pDCs by using the

RNeasy extraction kit (Qiagen, Crawley, UK) according to the

manufacturer’s instructions. Total RNA (350ng) was used for

reverse transcriptase (RT) reaction, performed by means of the

iScript TM cDNA Synthesis kit (Bio-Rad, Hercules, CA). The

reaction was performed using the following quantitative RT po-

lymerase chain reaction (qRT-PCR) program: 25�C for 5

minutes, 42�C for 30 minutes, 85�C for 5 minutes. The target

transcripts were amplified by means of a Light Cycler 480 sys-

tem (Roche Applied Science, Indianapolis, IN), using the fol-

lowing primers: human CB2R F1 (50-TCAACCCTGTCATCT

ATGCTC-30) and human CB2R R1 (50-AGTCAGTCCCAA

CACTCATC-30); human fatty acid amide hydrolase (FAAH)

F1 (50-CCCAATGGCTTAAAGGACTG-30) and human FAAH

R1 (50-ATGAACCGCAGACACAAC-30).20 b-Actin was used

as housekeeping gene for quantity normalization. One microli-

ter of the first strand of the cDNA product was used for ampli-

fication (in triplicate) in a 20ll reaction solution, containing

10ll of iQTM SYBR Green Supermix (Bio-Rad) and 1lmol of

each primer. The following PCR program was used: 95�C for 3

minutes, 40 amplification cycles at 95�C for 10 seconds, 56�C

for 20 seconds, and 72�C for 30 seconds. Fold-change was

determined by using the 22DDCT method.20

Intracellular Cytokine ProductionFreshly isolated PBMCs from HSs or MS patients were left

untreated or were pretreated with 2.5lM AEA, alone or in the

presence of the selective CB2R agonist JWH-015 or CB2R an-

tagonist SR144528, or of the FAAH inhibitor URB597. To

allow cytokine synthesis by mDCs and pDCs, cells were then

stimulated with 10ng/ml R848 for 5 hours, in the presence of

10lg/ml brefeldin A. At the end of the incubation, cells were

stained at the cell surface with specific antibodies, fixed with

4% paraformaldehyde (PFA) for 10 minutes, and then stained

intracellularly with cytokine-specific antibodies in 0.5% sapo-

nin, at room temperature (see Table). Intracellular cytokine pro-

duction by mDCs and pDCs was analyzed by flow cytometry

(FACS-Cyan ADP, Beckman Coulter). For each analysis,

300,000 events were usually acquired and viable cells were ana-

lyzed using Flowjo software (TreeStar, Ashland, OR).

DC/T-Cell CoculturesHighly purified mDCs and pDCs obtained from healthy con-

trols were tested for their ability to activate allogenic naive

ANNALS of Neurology

2 Volume 00, No. 00

CD41CD25lowCD45RA1CD271 T cells (1 3 105/well; see

Table), obtained from 3 unrelated healthy controls. Sorted

mDCs and pDCs were left untreated or were challenged with

10ng/ml R848 and/or AEA for 24 hours, and then were cul-

tured with naive T cells for 6 days into 96-well round-bottom

plates at 37�C at a DC:T-cell ratio of 1:2 and 1:10 in 200ll

final volumes of X-VIVO 15 medium (Lonza, Walkersville,

MD). After 6 days, cell supernatants were analyzed for inter-

feron (IFN)-c and IL-17 secretion by enzyme-linked immuno-

sorbent assay (Immunological Sciences, Rome, Italy), according

to the manufacturer’s instructions.

Confocal MicroscopyHighly purified pDCs were sorted with a MoFlo Cell Sorter (Beck-

man Coulter), following surface staining with anti–CD14-PE,

anti–HLA-DR-Pacific Blue, anti–CD3/CD19/CD56-PE-Cy7,

anti–CD11c-Alexa Fluor 700, and anti-CD123-V450. pDCs were

seeded on Chamber Slides (Lab-Tek; Electron Microscopy Sciences,

Hatfield, PA) at 5 3 104 cells/well, fixed with 3% PFA for 10

minutes, and then stained with anti-CB2R or anti-FAAH specific

antibodies. The localization of CB2R and FAAH was visualized by

confocal microscopy (TCS SP; Leica Wetzlar, Germany), perform-

ing image acquisition through the LAS AF program (Leica).

Statistical AnalysisParametrical statistical analysis (mean and standard deviation) was

performed through Prism 5.0 (GraphPad Software, San Diego,

CA) using standard methods. Significant differences were calcu-

lated using the nonparametric Mann–Whitney U test, and linear

regression analysis was performed using the Pearson correlation

test. For all tests, p values< 0.05 were considered significant.

Results

Differential Cytokine Production from Healthyand MS DC SubsetsWe first studied the production of typical cytokines of

CD1231CD11c2 pDCs and CD1232CD11c1 mDCs

in 20 HSs and in 38 RR-MS patients by using flow

cytometry. Both groups were stimulated with resiquimod

(R848), a potent and selective agonist of Toll-like recep-

tors 7/8 (TLR7/8).19 These receptors are abundant in

most innate immune cells, including monocytes, mDCs,

and pDCs.4 Therefore, we chose to use a single stimula-

tory agent like R848, to observe a multicellular response

through a multicolor flow cytometry staining that

enabled us to discriminate each cell population of inter-

est. The gating strategy is detailed in the Subjects and

Methods section. Figure 1 shows that unstimulated

mDCs and pDCs did not produce significant amounts

of cytokines. However, when stimulated with R848,

mDCs from both HSs and MS patients showed a similar

tumor necrosis factor (TNF)-a production, whereas

R848-stimulated mDCs from MS patients displayed a 2-

fold increase in IL-12p40 and IL-6 compared to HSs,

suggesting hyperactivation of mDCs during MS.

TABLE. Antibodies Used for Myeloid Dendritic Cell, Plasmacytoid Dendritic Cell, and Naive CD4 T-Cell Stain-ing and for Analysis of Intracellular Cytokine Production

Antibody Manufacturer Dilution

CD14 APC-Cy780/PE eBioscience (San Diego, CA) 1:60

Lin (CD3, CD19, CD56) PE-Cy7 Beckman Coulter (Fullerton, CA) 1:60

HLA-DR ECD/Pacific Blue Beckman Coulter 1:100

CD11c PerCP Cy5.5/Alexa Fluor 700 eBioscience 1:100

CD123 v450 eBioscience 1:60

CD80 PE BD Biosciences (Franklin Lakes, NJ) 1:60

CD86 APC Miltenyi Biotec (Bergisch Gladbach, Germany) 1:60

CD83 PE-Cy7 BD Pharmingen (San Jose, CA) 1:100

CD4 e780 eBioscience 1:50

CD25 PE Dako (Carpinteria, CA) 1:40

CD45RA PE-Cy7 Biolegend (San Diego, CA) 1:40

CD27 APC Miltenyi Biotec 1:50

TNF-a PE Miltenyi Biotec 1:100

IFN-a FITC Miltenyi Biotec 1:60

IL-12 p40 APC Miltenyi Biotec 1:60

IL-6 FITC BD Pharmingen 1:40

Chiurchi�u et al: Anandamide Modulates DCs in MS

Month 2013 3

Conversely, the analysis of pDCs showed that the num-

ber of TNF-a–producing pDCs did not differ between

MS patients and HSs, whereas R848-stimulated, IFN-a–

secreting pDCs were 2-fold lower in MS patients than in

HSs. In cells from MS patients, we consistently measured

elevated levels of IL-12 and IL-6 production by mDCs

and reduced production of IFN-a by pDCs, regardless of

the phase of the disease and of the donor’s gender.

TLR7/8-Induced Immunomodulation by AEA inmDCs and pDCs from HSs and MS SubjectsWe next asked whether the differences between HSs and

MS patients in the responsiveness of mDCs and pDCs

to R848 could also reflect a distinct response to the bio-

logically active endocannabinoid AEA. Of note, in all

subsequent experiments with AEA, only patients in exac-

erbation were analyzed. We found that, when R848-

stimulated mDCs from HSs were pretreated with AEA,

production of TNF-a was inhibited by �70%, IL-12p40

production was reduced by �60%, and IL-6 dropped by

�90%, compared to untreated cells (Fig 2A–C). Simi-

larly, AEA was also able to significantly suppress the

R848-induced cytokine production by mDCs obtained

from MS patients, although such an immunosuppression

was less potent, and reduced the production of IL-12p40

and IL-6 by �50% and that of TNF-a by �30% (see

FIGURE 1: Cytokine production from R848-stimulated dendritic cell (DC) subsets in healthy subjects (HSs) and multiple sclerosis(MS) patients during exacerbation and remission. Peripheral blood mononuclear cells (PBMCs) (1 3 106 cells/well) from HSs(n 5 28) and MS patients (n 5 48) were left unstimulated or were challenged with R848 for 5 hours, and then were stainedintracellularly with anti–tumor necrosis factor (TNF)-a, anti–interferon (IFN)-a, anti–interleukin (IL)212, or anti–IL-6 antibodies.Levels of intracellular cytokine production were analyzed by flow cytometry in the different cell populations, as detailed in Sub-jects and Methods. Data are means 6 standard deviation. According to physical parameters, PBMCs were appropriately gated.Monocytes were excluded by gating on CD141 and HLA-DR1 cells. DCs were gated on HLA-DR1 cells but excluding cells thatexpressed the lineage (Lin) antibodies (CD31, CD191, CD561). Within HLA-DR1Lin2 cells, myeloid DCs (mDCs) were character-ized as CD11c1CD1232, whereas plasmacytoid DCs (pDCs) were characterized as CD1231CD11c2. ECD = Energy CoupledDye (phycoerythrin-Texas Red conjugate); FSC = Forward Light Scatter; SSC-H = Side Light Scatter.

ANNALS of Neurology

4 Volume 00, No. 00

Fig 2E–F). Cannabinoid agonists bind to CB1R and

CB2R, and cells of the immune system mainly express

CB2R. To address the role of CB2R in mediating the

effects induced by AEA, mDCs from HSs and MS

patients were also pretreated with the selective synthetic

CB2R agonist JWH-015, alone or in combination with

its selective CB2R antagonist SR144528. JWH-015 was

also able to exert a significant suppression of TNF-a, IL-

6, and IL-12p40 production by mDCs of both HSs and

MS patients, resembling the effect of AEA. The JWH-

015–induced effect on mDCs was completely abolished

by SR144528, thus suggesting the engagement of CB2R

in the modulation of the inflammatory responses trig-

gered by TLR7/8. We next assessed the effect of AEA on

pDCs from HSs and MS patients. As shown in Figure 3,

AEA-pretreated pDCs from HSs showed a significant

�45% inhibition of TNF-a production and a �30% in-

hibition of IFN-a production, compared to R848-stimu-

lated pDCs. However, cytokine production by MS pDCs

was not inhibited by AEA, inasmuch as AEA could exert

only a weak (�15%) inhibition of TNF-a and IFN-a.

These findings suggest a loss of responsiveness to AEA of

pDCs in MS. Moreover, as expected these unresponsive

pDCs did not respond to JWH-015 nor to SR144528.

Any possible involvement of the other AEA-binding

receptors CB1R and TRPV1 in mediating these AEA-

induced effects was ruled out by pretreating cells with

the selective synthetic CB1R agonist ACEA and/or antag-

onist SR141716A (Supplementary Fig 1), or with the

TRPV1 antagonist capsazepine (Supplementary Fig 2).

In addition, we stimulated mDCs and pDCs with the

selective TLR7 ligand imiquimod (IMQ). We found that

AEA-pretreated pDCs from HSs showed a significant

�50% inhibition of TNF-a production (Supplementary

Fig 3A) and a �40% inhibition of IFN-a production

(see Supplementary Fig 3B), compared to IMQ-stimu-

lated pDCs. However, AEA could not inhibit TNF-aand IFN-a production by pDCs obtained from MS

patients (see Supplementary Fig 3 C, D). The latter find-

ing, together with the evidence that IMQ-stimulated

mDCs were unable to produce detectable levels of TNF-

a and IL-12p40 in either HSs or MS patients (data not

shown), suggests that the lack of responsiveness to AEA

of pDCs from MS patients depends on activation

through TLR7, and not TLR8. Given the distinct effect

of AEA in modulating cytokine production in each DC

FIGURE 2: Effect of anandamide (AEA) on cytokine production in activated myeloid dendritic cells (mDCs) from healthy sub-jects (HSs) and multiple sclerosis (MS) patients during exacerbation. Peripheral blood mononuclear cells (1 3 105 cells/well)from HSs and MS patients were either left untreated or treated with 2.5lM AEA and/or CB2R agonist or antagonist (each at1.0lM). Cells were then stimulated with R848 for 4 hours, and were stained intracellularly with anti–tumor necrosis factor(TNF)-a, anti–interleukin (IL)212 or anti-IL6 antibodies. Levels of intracellular cytokine production within mDCs were analyzedby flow cytometry, as detailed in Subjects and Methods.

Chiurchi�u et al: Anandamide Modulates DCs in MS

Month 2013 5

subset, we also ascertained whether the same substance

could modulate the activation status of DCs. Thus, we

assessed the surface expression of key markers of DCs.

Of note, all activation markers were increased in R848-

stimulated mDCs (Fig 4A–G) and pDCs (Fig 5A–G)

obtained from both groups of individuals compared to

unstimulated cells, except for CD86 in mDCs, whereas

particularly in MS patients expression levels were already

high in the unstimulated samples (see Fig 4B, F). How-

ever, AEA did not significantly modulate any activation

marker in DC subsets, suggesting that immunomodula-

tion by this endocannabinoid mainly occurs at the level

of cytokine inhibition rather than at that of activation

marker downregulation. Furthermore, because DCs are

unique in their potent ability to activate and polarize na-

ive T cells into specialized T helper cells,2–5 we also ana-

lyzed the direct effect of AEA on R848-dependent mDC

and pDC polarization of naive CD4 T cells into Th-1

and Th-17 cells (Fig 6). In the absence of TLR ligand,

no polarization of mDC- and pDC-dependent naive

CD4 T cells was observed. The presence of TLR ligand

enhanced the ability of both mDCs and pDCs to polar-

ize naive CD4 T cells into Th-1 and Th-17, because

levels of their signature cytokines (IFN-c and IL-17,

respectively) were significantly increased at both DC:T-

cell ratios. In addition, we observed that mDCs were

more effective than pDCs in inducing Th-1 and Th-17

cells. Interestingly, AEA significantly suppressed mDC-

dependent Th-1 and Th-17 induction, both at 1:2 and

1:10 coculture ratios. In addition, AEA equally inhibited

pDC-dependent Th-1 and Th-17 polarization when used

at a 1:10 coculture ratio. Instead, no pDC-dependent

induction of either T helper cell subsets was observed at

a pDC:naive CD4 T-cell coculture ratio of 1:2.

Differential Expression of ECS on mDCs andpDCs in Healthy and MS SubjectsNext, we asked whether the functional differences

between mDCs and pDCs from HSs and MS subjects

might reflect a different expression of crucial elements of

the ECS that regulate the biological activity of AEA,

namely CB2R and the AEA-degrading enzyme FAAH.

qRT-PCR analysis showed a �40-fold increase in CB2R

expression in mDCs obtained from MS patients com-

pared to those from HSs (Fig 7). pDCs from patients

also displayed an enhanced CB2R expression compared

FIGURE 3: Effect of anandamide (AEA) on cytokine production in activated plasmacytoid dendritic cells (pDCs) from healthysubjects (HSs) and multiple sclerosis (MS) patients during exacerbation. Peripheral blood mononuclear cell (1 3 105 cells/well)from HSs and MS patients were either left untreated or treated with 2.5lM AEA and/or CB2R agonist or antagonist (each at1.0lM). Cells were stimulated with R848 for 4 hours, and then were stained intracellularly with anti–tumor necrosis factor(TNF)-a or anti–interferon (IFN)-a antibodies. Levels of intracellular cytokine production within pDCs were analyzed by flowcytometry, as detailed in Subjects and Methods.

ANNALS of Neurology

6 Volume 00, No. 00

to pDCs from HSs, although such an increase was less

marked (�6-fold), overall suggesting a hyperactivation of

the ECS in MS. Interestingly, we observed that the

expression of FAAH in both DC subsets from HSs and

MS patients was significantly higher in HS pDCs than

in MS pDCs. Conversely, MS pDCs displayed a �60-

fold increase in FAAH levels with respect to HS pDCs,

indicating that pDCs from patients have a better ability

to metabolize AEA. The increase of FAAH expression in

pDCs from MS patients was also confirmed by means of

confocal microscopy. MS pDCs showed a significant

(�30%) increase of FAAH expression with a clear local-

ization within the cytoplasm, whereas HS mDCs showed

only a minor expression without any clear subcellular

localization (Fig 8). Because MS pDCs possessed high

levels of FAAH, we sought to inhibit this enzyme by pre-

treating highly purified pDCs cells with the potent and

irreversible FAAH inhibitor URB5979; then, cells were

treated with AEA, and finally they were challenged with

R848. It is noteworthy that URB597-pretreated MS

pDCs showed a �1.5-fold inhibition of TNF-a produc-

tion compared to R848-treated MD pDCs (Supplemen-

tary Fig 4), whereas treatment with URB597 alone was

ineffective (data not shown). Taken together, these data

show that pDCs from MS patients improve their respon-

siveness to AEA only upon inhibition of their FAAH

activity.

Discussion

This is the first study reporting a detailed characteriza-

tion of cytokine release from mDCs and pDCs in MS,

as well as the immunomodulatory effect of the main

endocannabinoid AEA on these DC subsets in health

and disease. Our data show that both cell populations

bear an altered cytokine profile in MS patients compared

to age-matched HSs, and that pDCs from MS patients

are unresponsive to AEA, likely due to an altered expres-

sion of key elements of the ECS. The observed increase

in IL-12p40 and IL-6 from mDCs, and the concomitant

decrease of IFN-a from pDCs in MS patients are in

agreement with the accumulating evidence of an altered

DC phenotype/function in MS, where mDCs show an

increased expression of activation markers, an enhanced

production of IL-12p40 in response to IFN-c and lipo-

polysaccharide, and an increased secretion of Th1 cyto-

kines6; conversely, pDCs exhibit a decreased or delayed

expression of activation markers, an altered functionality

in terms of T-cell proliferation and generation of

FIGURE 4: Effect of anandamide (AEA) on activation marker expression in activated myeloid dendritic cells (mDCs) fromhealthy subjects (HSs) and multiple sclerosis (MS) patients during exacerbation. Peripheral blood mononuclear cells (1 3 105

cells/well) from HSs and MS patients were either left untreated (CTRL) or treated with 2.5lM AEA. Cells were then stimulatedwith R848 for 4 hours, and were stained with anti-CD80, anti-CD86, anti-CD83, or anti–HLA-DR antibodies. Levels of surfaceexpression of activation markers within mDCs were analyzed by flow cytometry, as detailed in Subjects and Methods.M.I.F. 5 medium intensity of fluorescence.

Chiurchi�u et al: Anandamide Modulates DCs in MS

Month 2013 7

FIGURE 5: Effect of anandamide (AEA) on activation marker expression in activated plasmacytoid dendritic cells (pDCs) fromhealthy subjects (HSs) and multiple sclerosis (MS) patients during exacerbation. Peripheral blood mononuclear cells (1 3 105

cells/well) from HSs and MS patients were either left untreated (CTRL) or treated with 2.5lM AEA. Cells were then stimulatedwith R848 for 4 hours, and were stained with anti-CD80, anti-CD86, anti-CD83, or anti–HLA-DR antibodies. Levels of surfaceexpression of activation markers within pDCs were analyzed by flow cytometry, as detailed in Subjects and Methods.M.I.F. 5 medium intensity of fluorescence.

FIGURE 6: Effect of anandamide (AEA) on myeloid dendritic cell (mDC) and plasmacytoid dendritic cell (DCs) cocultures withnaive CD4 T cells from healthy subjects (HSs). Highly purified mDCs and pDCs from HSs (n 5 4) were either left untreated ortreated with 2.5lM AEA and/or stimulated with R848 for 24 hours, and then were cocultured with naive CD4 T cells at 1:2 and1:10 ratios for 6 days. Interferon (IFN)-c and interleukin (IL)217 release was analyzed by enzyme-linked immunosorbent assay,as detailed in Subjects and Methods.

ANNALS of Neurology

8 Volume 00, No. 00

regulatory T cells,7 and a reduced IFN-a production in

response to TLR9 agonist.21 In the present investigation,

we have chosen R848 as an ideal stimulatory agent,

because it simultaneously stimulates both DC subsets

through 2 distinct TLRs; pDCs express high levels of

TLR7 and do not express TLR8, whereas mDCs express

TLR8 and very low levels of TLR7.4,22 Overall, the dif-

ferences observed here between MS patients and HSs

help to better understand the role of mDC and pDC

subsets during the disease, and suggest that microbial

stimuli may indeed be involved in the alteration of

immune responses and in triggering MS attacks. Interest-

ingly, to date there are no specific therapies to target

innate immune cells (and specifically DCs) in MS. Given

recent evidence of the ECS as a novel therapeutic target

for autoimmune and chronic inflammatory diseases, and

the protective role of CB2R in such disorders,9–17 our

findings on the immunomodulatory effects of AEA on

both DC subsets are of major interest. Additionally, our

data show for the first time that AEA is immunosuppres-

sive on activated mDCs and pDCs in healthy subjects,

and that it is also able to inhibit cytokine production by

mDCs from patients, whereas in these subjects cytokine

production by pDCs is unaffected by AEA treatment.

The observed AEA-induced inhibition of TNF-a, IL-

12p40, and IL-6 production by mDCs from MS patients

is of crucial importance in preventing priming and acti-

vation of autoreactive T cells, as well as of NK cells.

This is particularly relevant for IL-12 p40, which shares

a subunit with IL-23. IL-12 and IL-23 are indeed hetero-

dimers, each composed of a common chain (p40) and a

distinct chain (IL-12 p35 and Il-23 p19, respectively). IL-

12 is a potent inducer of IFN-c and stimulates naive

CD41 T differentiation along a Th1 lineage; IL-23 pro-

motes and stabilizes IL-17 production by CD41 T cells

and, therefore, induces a Th17 phenotype. Both Th1 and

Th17 phenotypes have been linked to MS pathogenesis.23–

25 In line with our previous findings of a CB2R-mediated

AEA inhibition of IFN-c and IL-17 from highly purified

CD41 T cells,16 the present results document an unprece-

dented suppressive effect of AEA on specific innate

immune cells, which accounts for the ability of this endo-

cannabinoid to interfere with the IL-12–IL-23–IL-17 axis.

Such an AEA-mediated immunosuppression also influences

T-cell responses. In addition, the ability of healthy mDCs

(and to a lesser extent, healthy pDCs) to skew Th-1 and

Th-17 lineages is much reduced by AEA.

pDCs are type I IFN-producing cells that bear

unique characteristics, which enable them to both regu-

late immunity and to promote the development of toler-

ance.26 The reduced production of IFN-a by pDCs of

MS patients, and the failure of AEA to modulate this

DC subset in MS patients, seem to reflect the immune

responses at sites of inflammation in MS plaques, where

autoreactive T cells are more abundant and Tregs are less

numerous.27,28 Furthermore, we suggest that pDC

FIGURE 7: Endocannabinoid system expression in highly purified myeloid dendritic cells (mDCs) and plasmacytoid dendriticcells (pDCs) from healthy subjects (HSs) and multiple sclerosis (MS) patients during exacerbation. Peripheral blood mononuclearcell from HSs (n 5 6) and MS patients (n 5 6) were stained with anti–CD14-APC-Cy780, anti–HLA-DR-ECD, anti–CD3/CD19/CD56-PE-Cy7, anti–CD11c-PerCP/Cy5.5, or anti–CD123-V450 antibodies, and then were sorted by means of flow-activated cellsorting–MoFlo. Purified mDCs and pDCs were washed in phosphate-buffered saline, and mRNA was isolated and analyzed byquantitative reverse transcriptase polymerase chain reaction using specific primers for CB2R and fatty acid amide hydrolase(FAAH). A.U. 5 arbitrary units.

Chiurchi�u et al: Anandamide Modulates DCs in MS

Month 2013 9

responsiveness to AEA in MS patients is mediated by

TLR7. In keeping with this view, we show that the

TLR7-specific ligand IMQ mimics the AEA-induced

effect in R848-activated DCs. In particular, IMQ-

induced cytokine activation was observed only in pDCs,

whereas mDCs did not respond to this ligand. Taken to-

gether, these results suggest that an altered response of

pDCs in MS might be due at least in part to a deregula-

tion of the ECS. pDCs, unlike mDCs, did not show a

marked upregulation of the protective CB2R during MS.

Also the remarkable elevation of FAAH levels in patho-

genic pDCs seems particularly relevant, because FAAH is

crucial in maintaining the proper in vivo tone of circulat-

ing AEA and related endocannabinoids.29,30 Although we

were unable to measure FAAH activity in mDCs and

pDCs, due to their paucity in human blood,4 it can be

proposed that during MS pDCs improve their ability to

degrade endocannabinoids via FAAH, a feature coupled

to a very low expression of the anti-inflammatory CB2R

on their surface. In support of this hypothesis, herein we

show that FAAH inhibition in pDCs from patients

restored the AEA-induced suppression of TNF-a. The

recent finding that pharmacological inhibition of

endocannabinoid degradation modulates the expression

of inflammatory mediators in the hypothalamus, follow-

ing an immunological stressor, further corroborates this

concept.31 Interestingly, inhibition of FAAH activity by

non–brain-penetrant molecules has been recently demon-

strated to have an antianalgesic effect, due to enhanced

AEA levels in the periphery.32 Although it seems neces-

sary to further explore the details of the role of ECS in

pDC functionality during MS, their failure to appropri-

ately control immune tolerance is likely to account for

the present observations. Notably, the failure of most tri-

als based on immunotherapies targeting adaptive immu-

nity seems to support our data and to prompt broader

strategies, which possibly target also innate immunity. In

conclusion, the findings of the present study not only

contribute to a better understanding of the role of DCs

in the immunopathogenesis of MS, but may also form

the basis for possible ECS-oriented therapies.

Acknowledgment

This investigation was partly supported by the Italian

Ministry of Health (Finalized Project, L.B.), Italian

FIGURE 8: CB2R and fatty acid amide hydrolase (FAAH) expression in highly purified plasmacytoid dendritic cells (pDCs) fromhealthy subjects (HSs) and multiple sclerosis (MS) patients during exacerbation. Peripheral blood mononuclear cells from HSs(n 5 4) and MS patients (n 5 4) were stained with anti–CD14-PE, anti–HLA-DR-Pacific Blue, anti–CD3/CD19/CD56-PE-Cy7, anti–CD11c-Alexa Fluor 700, or anti–CD123-V450 antibodies, and then were sorted by means of flow-activated cell sorting–MoFlo.Purified HLA-DR1CD1231CD11c2 pDCs were stained with anti-CB2R or anti-FAAH primary antibodies, and with Alexa Fluor633 secondary antibody. CB2R and FAAH expression was reported as medium intensity of fluorescence values on a Zeiss con-focal microscope. A.U. 5 arbitrary units.

ANNALS of Neurology

10 Volume 00, No. 00

Foundation of Multiple Sclerosis, FISM (grant 2010,

M.M.; grant 2011, D.C. and L.B.), by the French Soci-

ety of Multiple Sclerosis, ARSEP (grant 2011 to L.B),

and Tercas Foundation (grant 2009–2012, M.M.).

We thank Dr S. Fiore for her kind help in recruiting

MS patients.

Authorship

V.C. and M.T.C. contributed equally to the study. L.B.

and M.M. are equally senior authors.

Potential Conflicts of Interest

Nothing to report.

References1. Gandhi R, Laroni A, Weiner HL. Role of the innate immune system

in the pathogenesis of multiple sclerosis. J Neuroimmunol2010;221:7–14.

2. Zozulya AL, Clarkson BD, Ortler S, et al. The role of dendritic cellsin CNS autoimmunity. J Mol Med (Berl) 2010;88:535–544.

3. MacDonald KP, Munster DJ, Clark GJ, et al. Characterization ofhuman blood dendritic cell subsets. Blood 2002;100:4512–4520.

4. Schreibelt G, Tel J, Sliepen KH, et al. Toll-like receptor expressionand function in human dendritic cell subsets: implications for den-dritic cell-based anti-cancer immunotherapy. Cancer ImmunolImmunother 2010;59:1573–1582.

5. Shortman K, Heath WR. Immunity or tolerance? That is the ques-tion for dendritic cells. Nat Immunol 2001;2:988–989.

6. Karni A, Abraham M, Monsonego A, et al. Innate immunity in mul-tiple sclerosis: myeloid dendritic cells in secondary progressivemultiple sclerosis are activated and drive a proinflammatoryimmune response. J Immunol 2006;177:4196–4202.

7. Stasiolek M, Bayas A, Kruse N, et al. Impaired maturation andaltered regulatory function of plasmacytoid dendritic cells in multi-ple sclerosis. Brain 2006;129:1293–1305.

8. Huang YM, Xiao BG, Ozenci V, et al. Multiple sclerosis is associ-ated with high levels of circulating dendritic cells secreting pro-inflammatory cytokines. J Neuroimmunol 1999;99:82–90.

9. Zhang H, Hilton DA, Hanemann CO, Zajicek J. Cannabinoid re-ceptor and N-acyl phosphatidylethanolamine phospholipase D—evidence for altered expression in multiple sclerosis. Brain Pathol2011;21:544–557.

10. Rossi S, Bernardi G, Centonze D. The endocannabinoid system inthe inflammatory and neurodegenerative processes of multiplesclerosis and of amyotrophic lateral sclerosis. Exp Neurol2010;224:92–102.

11. Correa FG, Mestre L, Docagne F, et al. The endocannabinoidanandamide from immunomodulation to neuroprotection. Implica-tions for multiple sclerosis. Vitamin Horm 2009;81:207–230.

12. Centonze D, Battistini L, Maccarrone M. The endocannabinoidsystem in peripheral lymphocytes as a mirror of neuroinflamma-tory diseases. Curr Pharm Des 2008;14:2370–2342.

13. Baker D, Pryce G. The endocannabinoid system and multiple scle-rosis. Curr Pharm Des 2008;14:2326–2336.

14. Pandey R, Mousawy K, Nagarkatti M, Nagarkatti P. Endocannabi-noids and immune regulation. Pharmacol Res 2009;60:85–92.

15. Centonze D, Bari M, Rossi S, et al. The endocannabinoid systemis dysregulated in multiple sclerosis and in experimental autoim-mune encephalomyelitis. Brain 2007;130:2543–2553.

16. Cencioni MT, Chiurchi�u V, Catanzaro G, et al. Anandamide sup-presses proliferation and cytokine release from primary human T-lymphocytes mainly via CB2 receptors. PLoS One 2010;5:.e8688.

17. Correa F, Hernang�omez-Herrero M, Mestre L, et al. The endocan-nabinoid anandamide downregulates IL-23 and IL-12 subunits in aviral model of multiple sclerosis: evidence for a cross- talkbetween IL-12p70/IL-23 axis and IL-10 in microglial cells. BrainBehav Immun 2011;25:736–749.

18. Polman CH, Reingold SC, Banwell B, et al. Diagnostic criteria formultiple sclerosis: 2010 revisions to the McDonald criteria. AnnNeurol 2011;69:292–302.

19. Jensen SS, Gad M. Differential induction of inflammatory cyto-kines by dendritic cells treated with novel TLR-agonist and cyto-kine based cocktails: targeting dendritic cells in autoimmunity. JInflamm (Lond) 2010;27:7–37.

20. Pasquariello N, Catanzaro G, Marzano V, et al. Characterization ofthe endocannabinoid system in human neuronal cells and proteo-mic analysis of anandamide-induced apoptosis. J Biol Chem2009;284:29413–29426.

21. Bayas A, Stasiolek M, Kruse N, et al. Altered innate immuneresponse of plasmacytoid dendritic cells in multiple sclerosis. ClinExp Immunol 2009;157:332–342.

22. Kadowaki N, Ho S, Antonenko S, et al. Subsets of human dendriticcell precursors express different Toll-like receptors and respond todifferent microbial antigens. J Exp Med 2001;194:863–869.

23. Chen Y, Langrish CL, McKenzie B, et al. Anti-IL-23 therapy inhibitsmultiple inflammatory pathways and ameliorates autoimmune en-cephalomyelitis. J Clin Invest 2006;116:1317–1326.

24. Cua DJ, Sherlock J, Chen Y, et al. Interleukin-23 rather than inter-leukin-12 is the critical cytokine for autoimmune inflammation ofthe brain. Nature 2003;421:744–748.

25. McKenzie BS, Kastelein RA, Cua DJ. Understanding the IL-23-IL-17 immune pathway. Trends Immunol 2006;27:17–23.

26. Maldonado RA, von Andrian UH. How tolerogenic dendritic cellsinduce regulatory T cells. Adv Immunol 2010;108:111–165.

27. Venken K, Hellings N, Thewissen M, et al. CompromisedCD41CD25(high) regulatory T-cell function in patients with relaps-ing-remitting multiple sclerosis is correlated with a reduced fre-quency of FOXP3-positive cells and reduced FOXP3 expression atthe single-cell level. Immunology 2008;123:79–89.

28. Swiecki M, Colonna M. Unraveling the functions of plasmacytoiddendritic cells during viral infections, autoimmunity, and tolerance.Immunol Rev 2010;234:142–162.

29. Piomelli D. The endocannabinoid system: a drug discovery per-spective. Curr Opin Investig Drugs 2005;6:672–679.

30. Di Marzo V, Bisogno T, De Petrocellis L. Endocannabinoids andrelated compounds: walking back and forth between plant naturalproducts and animal physiology. Chem Biol 2007;14:741–756.

31. Kerr DM, Burke NN, Ford GK, et al. Pharmacological inhibition ofendocannabinoid degradation modulates the expression ofinflammatory mediators in the hypothalamus following an immu-nological stressor. Neuroscience 2012;204:53–63.

32. Clapper JR, Moreno-Sanz G, Russo R, et al. Anandamide sup-presses pain initiation through a peripheral endocannabinoidmechanism. Nat Neurosi 2010;13:1265–1270.

Chiurchi�u et al: Anandamide Modulates DCs in MS

Month 2013 11