REVIEWS

CANNABINOID-BASED DRUGS
AS ANTI-INFLAMMATORY
THERAPEUTICS
Thomas W. Klein
Abstract | In the nineteenth century, marijuana was prescribed by physicians for maladies
ranging from eating disorders to rabies. However, as newer, more effective drugs were
discovered and as the potential for abuse of marijuana was recognized, its use as a therapeutic
became restricted, and only recently has its therapeutic potential been re-evaluated. Recent
studies in animal models and in humans have produced promising results for the treatment of
various disorders — such as obesity, cancer, and spasticity and tremor due to neuropathology
— with drugs based on marijuana-derived cannabinoids. Moreover, as I discuss here, a wealth
of information also indicates that these drugs have immunosuppressive and anti-inflammatory
properties; therefore, on the basis of this mode of action, the therapeutic usefulness of these
drugs in chronic inflammatory diseases is now being reassessed.

CANNABINOID RECEPTORS Cannabis sativa — also known as marijuana — is the drugs with a high potential for abuse; however, this
G-protein-coupled receptors most frequently used illicit drug in the United States, now seems to be changing. In the past few years,
for ∆9-tetrahydrocannabinol, particularly among young people, with up to 46.1% of numerous publications have reported the potential use
its synthetic analogues and
those aged 17–18 being users. Although marijuana- of marijuana-based medicines for the treatment of dis-
endocannabinoids. They
have been identified in most usage rates are lower than those for legal drugs, such as eases ranging from cancer to glaucoma1–8, and one
vertebrate phyla. Two subtypes cigarettes and alcohol, they are significantly higher than compound, SR141716A (also known as rimonabant or
are known: cannabinoid those for other illicit drugs, such as cocaine and ecstasy. Acomplia; Sanofi-Synthélabo)9, has been widely publi-
receptor 1 (CB1) and CB2. As a therapeutic, marijuana has been recognized for cized as the next wonder drug for promoting weight
ENDOCANNABINOIDS
centuries, and in the nineteenth century, it was recom- loss and smoking cessation. Here, I review the effects of
Endogenous agonists for mended as an analgesic, muscle relaxant, appetite stim- marijuana and related compounds on the adaptive and
cannabinoid receptors that are ulant and anticonvulsant1. Its therapeutic applications innate immune systems and provide some perspective
present in animals. They are continued to grow such that, by the early twentieth cen- concerning the therapeutic potential of marijuana-
metabolites of eicosanoid
tury, therapy with marijuana was used to ease the symp- derived cannabinoids and related compounds for the
fatty acids.
toms of a broad spectrum of diseases, ranging from treatment of chronic inflammatory diseases.
rheumatism and epilepsy to tetanus and gonorrhoea.
However, the popularity of marijuana as a therapeutic Cannabinoid-based drugs and receptors
Department of Medical declined as newer, more effective drugs were discovered The main psychoactive component in marijuana is the
Microbiology and for these ailments and as the potential for abuse of mari- ‘classical’ cannabinoid THC10 (FIG. 1a). Similar to its
Immunology, University juana was recognized, so by the early 1940s, it was synthetic analogues (FIG. 1b,c), it functions by activating
of South Florida College of removed from use as a pharmaceutical in the United specific cell-surface CANNABINOID RECEPTORS, which are
Medicine, Tampa, Florida
33612, USA.
States. From that time until the 1980s, marijuana and its normally engaged by a family of endogenous ligands
e-mail: tklein@hsc.usf.edu derivatives, such as ∆9-tetrahydrocannabinol (THC), — the ENDOCANNABINOIDS (FIG. 1g,h). Compounds that
doi:10.1038/nri1602 have been viewed as both scientific curiosities and illegal bind these receptors induce CANNABIMIMETIC responses

400 | MAY 2005 | VOLUME 5 www.nature.com/reviews/immunol

© 2005 Nature Publishing Group
 REVIEWS

a b c in vivo and in vitro11. Research on the structure–activity

HO
relationships of THC led to the synthesis of ‘non-
COOH classical’ cannabinoid analogues. The most widely
OH OH OH studied of this group is CP55,940 (FIG. 1d), a high-affinity
ligand with potent biological effects that are mediated
through binding both cannabinoid receptor 1 (CB1)
O O O and CB2. Another class of cannabimimetic agents is
the aminoalkylindoles, such as WIN55,212-2 (FIG. 1e),
Classical cannabinoid: Synthetic derivative of THC: Synthetic derivative of THC-11-oic
THC HU-210 acid: ajulemic acid which also has high affinity for both receptors and has
potent activity. By contrast, some derivatives of this
class, such as JWH-015 (FIG. 1f), have been shown to
d e f selectively bind CB2. Such derivatives might be of thera-
peutic value, because psychoactive effects are mediated
through binding CB1 and not through binding CB2.
Other plant-derived cannabinoids (such as cannabid-
O
OH iol) and synthetic derivatives (such as HU-211; also
O
known as dexanabinol; Pharmos Corporation) bind CB1
OH
N and CB2 with very low affinity, so they have low
O N cannabimimetic activity and might function by binding
N
different receptors, such as the NMDA (N-METHYL-D-ASPARTATE)
5
RECEPTOR or other unknown receptors. Furthermore, a
OH
O derivative of THC-11-oic acid, ajulemic acid (also known
Non-classical cannabinoid: Aminoalkylindole: Aminoalkylindole derivative: as CT-3) (FIG. 1c), has low affinity for CB2 (REF. 12) but has
CP55,940 WIN55,212-2 JWH-015 anti-inflammatory activity13, which might be mediated
through disruption of the arachidonic-acid cascade14
or through activation of peroxisome-proliferative-
g h activated receptor-γ (PPAR-γ)15. Finally, receptor antag-
O
onists have been synthesized, such as SR141716A and
CONHCH2CH2OH OH
O
OH
SR144528 (REF. 11) (FIG. 1i,j), which inhibit or reverse the
biological effects of CB1 and CB2 agonists. These antag-
onists have been used experimentally to determine the
Endocannabinoid: arachidonoylethanolamide Endocannabinoid: 2-arachidonoylglycerol
receptor-binding activities of various putative agonists
and are now also being used clinically to inhibit CB1
activity. For example, SR141716A is given to suppress
i j appetite during the treatment of obesity9.
The naturally occurring endocannabinoids are pro-
O N
O duced by the cleavage of membrane fatty acids16,17, in par-
NH
NH ticular arachidonic acid, and have varying specificities for
N the two cannabinoid receptors. Arachidonoylethano-
N N
Cl N lamide (AEA; previously known as anandamide) is an
Cl Cl
endogenous fatty-acid amide and was the first endo-
cannabinoid to be discovered18 (FIG. 1g). It has less intrinsic
Cl activity when bound to CB2 than to CB1, and it has also
CB2 antagonist: SR144528
been shown to bind VANILLOID RECEPTORS, which are ligand-
CB1 antagonist: SR141716A
gated cation channels that are sensitive to capsaicin and
Figure 1 | Drugs based on marijuana-derived cannabinoids are divided into various related analogues19. Several other endocannabinoids have
groups. There are four groups of cannabinoids or cannabimimetic drugs that have also been described, including 2-arachidonoylglycerol
cannabinoid-receptor-binding activity. The first group contains the ‘classical’ cannabinoids, (2-AG)20 (FIG. 1h) and 2-arachidonylglyceryl ether (also
which are tricyclic dibenzopyran derivatives. They occur naturally in the marijuana plant — known as noladin ether)21, with the former being a full
for example, ∆9-tetrahydrocannabinol (THC) (a). Also, there are synthetic analogues of these agonist of CB2 and the latter showing a higher affinity
natural compounds — for example, HU-210 (b). These cannabinoids bind relatively non-
for CB1. The endocannabinoids are produced by various
selectively to both cannabinoid receptor 1 (CB1) and CB2, with HU-210 having a higher affinity
than THC. Ajulemic acid (c) is another analogue; it is a classical cannabinoid derived from
cells, including cells of the immune system and the
THC-11-oic acid. Ajulemic acid binds both cannabinoid receptors but has a lower affinity brain (discussed later).
for CB2. The second group contains the non-classical cannabinoids, which are synthesized In addition to the two cannabinoid receptors that
analogues that lack the dihydropyran ring of THC. This group contains compounds such have been characterized22,23, there might be others, par-
as CP55,940 (d). The third group contains the aminoalkylindoles, such as WIN55,212-2 (e). ticularly receptors that interact with the endocannabi-
Derivatives of WIN55,212-2, such as JWH-015 (f ), have been shown to selectively bind noids, which would be consistent with the observation
CB2. The fourth group contains the endocannabinoids, which are eicosanoid compounds
that AEA also binds vanilloid receptors19. CB1 and CB2
rather than cannabinoid compounds. The most studied members of this group are
arachidonoylethanolamide (g) and 2-arachidonoylglycerol (h). Finally, several cannabinoid- are seven-transmembrane, G-protein-coupled receptors
receptor antagonists have been synthesized, and the most widely studied are SR141716A (i), that are coupled to Gi or Go heterotrimeric proteins
which selectively binds CB1, and SR144528 ( j), which selectively binds CB2. and adenylyl cyclases, but other second messengers and

NATURE REVIEWS | IMMUNOLOGY VOLUME 5 | MAY 2005 | 4 0 1

© 2005 Nature Publishing Group
REVIEWS

CANNABIMIMETIC signalling components are also involved in their activ- chemically similar metabolites — such as the endo-
∆9-Tetrahydrocannabinol ity24,25. The tissue distribution of CB1 and CB2 accounts cannabinoid AEA — are produced and released by acti-
(THC)-like in pharmacological for the well-known psychotropic and peripheral effects vated immune cells32 (FIG. 2). In vitro studies have shown
terms. A compound is usually
of cannabinoids (reviewed in REFS 11,25–28). CB1 is abun- that stimulation with bacterial lipopolysaccharide (LPS)
accepted to be cannabimimetic
if it produces four characteristic dant in the central nervous system (CNS), in particular increases the production of AEA and 2-AG by immune
effects of THC in an in vivo assay in the basal ganglia, cerebellum, hippocampus and cor- cells (FIG. 2), including macrophages33, peripheral-blood
known as the ‘mouse tetrad tex, where the receptors and endocannabinoids are mononuclear cells (PBMCs)34 and dendritic cells35. In
model’. These effects are implicated in retrograde neurotransmitter regulation of addition, LPS-activated PBMCs show reduced expres-
hypomotility, hypothermia,
analgesia and a sustained
synaptic transmission28,29. CB1 is also expressed in the sion of the AEA-degrading enzyme fatty-acid amide
immobility of posture periphery, as is CB2, where they are mainly restricted to hydrolase (FAAH)36. Given that inflammatory responses
(catalepsy). immune cells and tissues (discussed later). must be tightly controlled to avoid extensive tissue dam-
age37 and to allow a return to homeostatic conditions,
Endocannabinoids in innate immunity FAAH-mediated degradation is one of the mechanisms
Production of endocannabinoids. Microbial pathogens for the inactivation of endocannabinoids. In addition,
that invade the tissues are recognized by host cells and as-yet-uncharacterized endocannabinoid membrane
host factors that trigger the activation of both innate transporters seem to facilitate both the release and the
and adaptive immune responses30. Activation of the subsequent uptake of endocannabinoids by neurons and
inflammatory response to infection largely depends on glial cells, thereby contributing to the regulation of endo-
the release of pro-inflammatory cytokines and chemo- cannabinoids. So, immune cells from humans and ani-
kines. However, in addition to cytokines and other mals increase the production of endocannabinoids in
proteins, various metabolic products of immune cells — response to LPS and in response to activation by other
including membrane fatty acids, such as arachidonic stimuli. Additional studies are needed to determine the
acid31 — have also been implicated in the inflammatory range of immune and microbial stimuli that induce
response to infection. It is therefore not surprising that endocannabinoid production and to further define the
mechanisms that regulate this effect.
Bacterial infection Upregulation of expression
of cannabinoid receptors
Endocannabinoids as chemotactic agents. After recogni-
tion of an invading pathogen, the release of cytokines
Pathogens
Cannabinoid and chemokines by cells involved in the innate immune
receptor Recruitment of
immune cells
response triggers an influx of lymphoid and myeloid
Dendritic cell cells from the blood to the site of infection38. Recent
TLR evidence indicates that the endocannabinoid 2-AG also
induces the migration of various cell types (FIG. 2). For
example, in transwell cultures, 2-AG attracted human
Macrophage eosinophils39, as well as Raji B cells40 and mouse bone-
marrow-derived dendritic cells41. Furthermore, CB2
seems to be overexpressed in several myeloid leukaemias,
Lymphocyte and these leukaemic cells are induced to migrate after
stimulation with 2-AG42. It therefore seems that endo-
cannabinoids might be involved in cellular migration by
functioning as chemotactic agents, together with
Release of pro-inflammatory cytokines and proteins that have classically been defined
cytokines, chemokines and NK cell
endocannabinoids such as as chemokines. It is interesting to note that a similar
AEA and 2-AG function has been reported for opioids, such as mor-
Eosinophil
phine, which have neuroimmune functions in common
with cannabinoids43.
Blood B cell
Expression of cannabinoid receptors. In addition to
T cell
the induction of endocannabinoid expression and the
Monocyte subsequent chemotaxis of immune cells, there is evi-
dence that activation of immune cells by LPS or other
Figure 2 | The endocannabinoid system and innate immunity. Bacteria stimulate
stimuli also modulates the expression of the cannabi-
lymphocytes, dendritic cells (DCs) and macrophages — through pattern-recognition receptors, noid receptors CB1 and CB2 by these cells (FIG. 2). For
such as Toll-like receptors (TLRs) — to release cytokines and chemokines, which attract example, the initial report on the cloning of CB2 also
leukocytes to the site of infection. It is now known that stimulation of these cells also induces showed that stimulation of the human pro-myelocytic
the release of endocannabinoids — such as arachidonoylethanolamide (AEA) and leukaemia cell line HL-60 with phorbol 12-myristate
2-arachidonoylglycerol (2-AG) — that are also chemotactic for leukocytes. Leukocytes — including 13-acetate (PMA) caused an increase in the level of
T cells, B cells, eosinophils, natural killer (NK) cells, DCs and macrophages — invade the tissues,
promoting the elimination of the microorganisms and the development of an adaptive immune
mRNA encoding CB2 (REF. 23); similar results were also
response. These activated cells seem to upregulate the expression of both types of cannabinoid obtained in mouse splenocyte cultures stimulated
receptor — cannabinoid receptor 1 (CB1) and CB2 — and these are then available for with CD40-specific antibody 44. In addition to CB2, the
participation in immune regulation. levels of mRNA encoding CB1 were also increased in

402 | MAY 2005 | VOLUME 5 www.nature.com/reviews/immunol

© 2005 Nature Publishing Group
 REVIEWS

both Jurkat T cells 45 and mouse splenocytes46 after this treatment was also neuroprotective and resulted
stimulation. However, other studies have shown that in a better clinical outcome56. Although HU-211 does
stimulation decreased the expression of these recep- not bind cannabinoid receptors and is therefore not
tors46,47, which most probably reflects differences in cannabimimetic, it seems to function as an NMDA-
the stimulatory substance used or the cell type studied. receptor antagonist, thereby preventing excitotoxicity
LPS has also been shown to modulate cannabinoid- and neuronal death57. In another mechanistically
receptor expression, although findings from separate complex mouse model, treatment with WIN55,212-2
studies are inconsistent, with some reports showing decreased tissue damage after myocardial ischaemia–
upregulation of receptor expression48 and others showing reperfusion injury58. Treatment of animals before
downregulation35,44,49. ischaemia and reperfusion considerably reduced the
Consistent with a possible role for endocannabinoids size of the infarct, and this was paralleled by lower
and CB2 in chemotaxis and other immune-activation levels of production of IL-1β and CXC-chemokine
events, a recent study of mouse brain microglial cells ligand 8 (CXCL8) in the injured tissue. In human
showed that, under chemotactic conditions, the expres- studies, lung alveolar macrophages removed from
sion of CB2 was specifically localized to the leading edge marijuana smokers were compromised in their ability
of the cell50, indicating that there is receptor modulation to produce TNF, granulocyte/macrophage colony-
similar to that observed for immunoreceptors during stimulating factor and IL-6 in response to LPS stimula-
chemotaxis or during the formation of the immuno- tion59 (TABLE 1). It therefore seems that cannabinoids
logical synapse51. Taken together, these studies indicate can inhibit the production of TNF and other cytokines
that microbial antigens or other stimuli that induce in several different models and by several different
immune activation influence cannabinoid-receptor mechanisms (TABLE 1), not all of which depend on
expression by immune cells; however, the factors that interaction with cannabinoid receptors.
are involved in increased or decreased expression are far As well as suppressing the production of cytokines,
from understood, and little is known about the molec- cannabinoids have been shown to increase the produc-
ular regulation of the genes encoding these receptors in tion of cytokines (including TNF, IL-1, IL-6 and IL-10)
immune cells. Further research is needed to take full when they are administered together with bacteria
advantage of manipulating this potentially important or other antigens60–62, or in some cases, when cannabi-
immunomodulating system for therapeutic purposes noids are administered alone63,64. So, in vivo, cannabinoids
(discussed later). might either suppress or enhance the production of
these pro-inflammatory agents, depending on either the
Cannabinoids in inflammation nature of the pro-inflammatory stimulus or the type of
Cannabinoids modulate cytokine production. In the cannabinoid used (TABLE 1).
mid-1980s, it was shown that mouse cells treated with
NMDA RECEPTOR the drug THC produced decreased levels of type I inter- Cannabinoids modulate inflammatory-cell migration.
(N-methyl-D-aspartate ferons (IFN-α and IFN-β) after stimulation with LPS or The effect of cannabinoids has also been analysed in
receptor). NMDA is a synthetic
polyinosinic–polycytidylic acid (polyI:C)52,53, providing the experimental allergic encephalomyelitis (EAE)
amino acid with affinity for
NMDA receptors, which the first evidence that cannabinoids might modulate model of multiple sclerosis (a neuroinflammatory dis-
mediate excitatory effects in the cytokine production. Many subsequent studies have ease), and they have been shown to suppress disease
brain when they are stimulated shown that cannabinoids, for the most part, suppress progression65,66 and inflammatory reactions66. In one
by endogenous ligands such as the production of cytokines in innate and adaptive study, mice were induced to develop EAE and given
glutamic acid. Overstimulation
can lead to neuronal
immune responses, both in animal models and in WIN55,212-2 alone or WIN55,212-2 and then either
excitotoxicity. human cell cultures26,27,54. Their suppression of pro- a CB1 or CB2 antagonist every 4 days. After several
inflammatory cytokine and chemokine production indi- weeks, disease progression was assessed, together with
VANILLOID RECEPTORS cates that these drugs might have anti-inflammatory leukocyte–endothelial-cell interactions, using intra-
Cation channels that are
effects and could therefore be used for the treatment of vital microscopy. Treatment with WIN55,212-2 sup-
expressed by nerve sensory
fibres and are involved in the chronic inflammatory diseases. Consistent with this, pressed the rolling and adhesion of venous leukocytes
perception of pain. These serum levels of tumour-necrosis factor (TNF) and and improved neurological function in mice with
receptors are ligand-, proton- interleukin-12 (IL-12) were shown to be decreased in EAE, compared with control animals; furthermore,
and heat-activated and are mice that were primed by infection with Propioni- the suppressive effect was attenuated by co-treatment
targets for capsaicin — the hot
component of chillies.
bacterium acnes (Corynebacterium parvum) and stimu- with a CB2 antagonist but not a CB1 antagonist, indi-
lated with an injection of LPS (conditions that promote cating that CB2 is involved in these processes66. The
DIAPEDESIS optimum cytokine upregulation) then treated with the reduction in DIAPEDESIS presents a paradox, because
The last step in the synthetic THC derivative HU-210 or the aminoalkyl- cannabinoids have been shown to enhance chemo-
leukocyte–endothelial-cell
indole WIN55,212-2 (REF. 55). The cannabinoids also pro- taxis and chemokine production (as discussed ear-
adhesion cascade. This cascade
includes tethering, triggering, tected these mice from the lethal effects of LPS in this lier). Although the levels of cytokines and adhesion
tight adhesion and model, and this protection might have resulted from, molecules were not measured in this study, it is possi-
transmigration. Diapedesis is the at least in part, a concomitant drug-induced increase ble that cannabinoid-mediated suppression of leuko-
migration of leukocytes across in the levels of the regulatory cytokine IL-10 (TABLE 1). cyte adhesion results from an inhibition of T helper 1
the endothelium, which occurs
by squeezing through the
TNF production was also suppressed in the brain of (TH1)-cell cytokine production: for example, of IFN-γ,
junctions between adjacent rats subjected to closed head injury, after treatment which facilitates transendothelial cell trafficking 67,68
endothelial cells. with the non-psychoactive cannabinoid HU-211, and (discussed later).

NATURE REVIEWS | IMMUNOLOGY VOLUME 5 | MAY 2005 | 4 0 3

© 2005 Nature Publishing Group
REVIEWS

Table 1 | Cannabinoid effects on pro-inflammatory cytokines

Cannabinoid Receptor Cell or tissue type Cytokine stimulant or Effect Reference
inflammation model
Mice
WIN55,212-2 ND Spleen LPS and Decreases TNF and IL-12 55
or HU-210 Propionibacterium acnes* and increases IL-10
THC ND Macrophage cell line LPS Decreases TNF 122
(RAW264.7)
HU-211 NMDA receptor Brain Closed head injury Decreases TNF 56
WIN55,212-2 CB2 dependent Heart Ischaemia–reperfusion Decreases IL-1β and CXCL8 58
THC, AEA ND Macrophage cell line LPS Decreases IL-6 123
or 2-AG (J774)
THC ND Spleen Legionella pneumophila Increases TNF and IL-6 60
THC ND Peritoneal macrophages LPS Increases IL-1α and IL-1β 61
Humans
Marijuana ND Lung alveolar LPS Decreases TNF, GM-CSF and IL-6 59
smoking macrophages
Ajulemic ND Peripheral-blood and LPS Decreases IL-1β 100
acid synovial monocytes
2-AG CB2 dependent Pro-myelocytic leukaemia 2-AG Increases CXCL8 and CCL2 64
cell line (HL-60)
CP55,940 CB2 dependent Pro-myelocytic leukaemia cell CP55,940 Increases TNF, CXCL8, CCL2 63
line (HL-60) (CB2 transfected) and CCL4
Rats
AEA, 2-AG, CB1 and CB2 Microglial cells LPS Decreases TNF 124
WIN55,212-2 independent
or HU-210
*Propionibacterium acnes was previously known as Corynebacterium parvum. AEA, arachidonoylethanolamide; 2-AG, 2-arachidonoylglycerol; CB, cannabinoid receptor;
CCL, CC-chemokine ligand; CXCL8, CXC-chemokine ligand 8; GM-CSF, granulocyte/macrophage colony-stimulating factor; IL, interleukin; LPS, lipopolysaccharide; ND, not
determined; NMDA, N-methyl-D-aspartate; THC, ∆9-tetrahydrocannabinol; TNF, tumour-necrosis factor.

An increasing number of studies indicate that response to the drug was biased towards the production
cannabinoids and related compounds modulate pro- of TH2 cells. Cannabinoid receptors were shown to be
inflammatory cytokines in various systems, ranging involved in the effects of THC, and it is possible that
from infection to tissue injury. So far, the mechanisms of signalling through these G-protein-coupled receptors
this modulation are unclear, but mechanisms that suppresses the expression of IL-12, as has been shown in
involve cannabinoid receptors, as well as other mecha- other systems72. In addition to this infection model, the
nisms, seem to be involved. For therapeutic applications, TH-cell-biasing effect was also shown in a mouse model
compounds that function either by binding CB2 or by of lung cancer, in which IFN-γ production was decreased
by-passing cannabinoid receptors altogether would be of but IL-10 and transforming growth factor-β (TGF-β)
benefit, because CB1-mediated psychoactive side-effects production were increased73. The presence of higher lev-
would be diminished. els of these cytokines indicates that THC either biases
the immune response towards TH2 cells or activates
Cannabinoids modulate T helper cells regulatory T cells, either T regulatory 1 (TR1) cells or
Cannabinoid treatment has been shown to suppress TH3 cells74,75. These regulatory T-cell populations have
both innate immunity and adaptive immunity, and the been shown to produce IL-10 and TGF-β, respectively,
effects of cannabinoids on humoral and cellular immu- so they could be the source of these cytokines following
nity have been extensively reviewed26,27,69. So, here I treatment with THC. THC was also shown to bias
focus on the more recent studies that indicate that TH-cell differentiation towards TH2 cells during an allo-
cannabinoids have a TH-cell biasing effect, in which geneic response to human peripheral-blood T cells76.
TH1-cell activity is suppressed and TH2-cell activity is Furthermore, PBMCs isolated from marijuana smokers
increased (TABLE 2). The first indications of this effect can proliferated less and produced less IL-2 in response to
be traced to the observation that treatment with THC mitogens, but they produced more IL-10 and TGF-β
induces suppression of both TH1-cell activity and cell- than PBMCs isolated from non-marijuana smokers,
mediated immunity in mice infected with Legionella indicating a TH-cell bias towards TH2 cells77 (TABLE 2).
pneumophila70. Additional studies showed that treat- Animal models of multiple sclerosis have also shown that
ment of mice with THC not only decreased the pro- treatment with cannabinoids attenuates cell-mediated
duction of IFN-γ and the levels of IL-12 and the IL-12 immunity and TH-cell activity, concomitant with reduc-
receptor (IL-12R) but also increased the production ing disease symptoms. For example, in a model in which
of IL-4 (REF. 71) (TABLE 2), showing that the immune demyelinating disease is induced by Theiler’s murine

404 | MAY 2005 | VOLUME 5 www.nature.com/reviews/immunol

© 2005 Nature Publishing Group
 REVIEWS

encephalomyelitis virus, cannabinoid-receptor ligands responses, their therapeutic potential is now being
— such as WIN55,212-2 and JWH-015 — were shown evaluated on the basis of their immunomodulatory
to ameliorate disease progression by reducing the and anti-inflammatory actions.
delayed-type hypersensitivity response, by decreasing
IFN-γ production78 and microglial-cell activation, and Nervous-tissue inflammation. Injection of THC was
by suppressing the number of CD4+ T cells infiltrating initially shown to suppress the neurological signs and
the brain79. symptoms of EAE in rat and guinea-pig models89
From these studies, it seems that cannabinoids bias (TABLE 3). A marked reduction in inflammation in the
the immune response away from TH1-cell immunity CNS of cannabinoid-treated animals was observed, and
and that cannabinoid receptors are involved in this this was later suggested to result from a THC-induced
process. It is possible that signalling through cannabi- increase in the serum levels of the steroid cortico-
noid receptors expressed by T and B cells, as well as by sterone, which is known to be anti-inflammatory and
antigen-presenting cells, suppresses the expression of immunosuppressive90. However, this explanation might
TH1-cell-promoting cytokines and increases the expres- be an oversimplification, because, as described earlier,
sion of TH2-cell-promoting cytokines. Indeed, treat- cannabinoids attenuate demyelinating disease in these
ment with cannabinoids has been shown to alter the models by suppressing the differentiation of TH cells
expression of specific transcription factors: the expres- into TH1 cells78,79, and this suppression can be uncou-
sion of the TH2-cell-promoting transcription factor pled from THC-induced steroid mobilization in treated
GATA-binding protein 3 (GATA3) has been found to be animals91. Another consequence of neurodegeneration
increased by treatment with THC80,81, whereas the pro- is neuropathic or central pain, which results from scle-
duction of IL-2-expression-promoting transcription rotic plaques that affect pain pathways in the CNS92.
factors is suppressed by treatment with cannabinoids82. These lesions and central pain are observed in various
The selective suppression of TH1-cell immunity by these chronic conditions, including trauma, type 2 diabetes
drugs supports their potential use in the treatment of and multiple sclerosis. Although cannabinoids have a
chronic inflammatory diseases. well-recognized analgesic effect that is mediated by sig-
nalling through CB1 (REF. 93), compounds that selec-
Therapeutics for inflammatory diseases tively bind CB2 and act in the periphery have recently
Recent work has led researchers to consider that been shown to suppress experimentally induced central
cannabinoid-based drugs have therapeutic potential for pain94. The aminoalkylindole derivative AM1241 was
the treatment of a variety of disorders. For example, the shown to reverse tactile and thermal hypersensitivity
CB1 antagonist SR141716A has been shown to effec- produced in rats by spinal nerve ligation, and this effect
tively promote weight loss and smoking cessation in was blocked by a CB2 antagonist, but not a CB1 antago-
preclinical and clinical trials9,83,84. Furthermore, cannabi- nist, and occurred in CB1-deficient mice. This almost
noids seem to be neuroprotective in models of inflamma- completely CB2-dependent effect has been proposed to
tory neurodegenerative disease, possibly because some occur in the periphery because of the paucity of these
derivatives function as inhibitors of NMDA receptors receptors in the CNS; the mechanism of action has
and as antioxidants5,7 and because they function to con- been suggested to be indirect, with the drug acting on
trol spasticity and tremor that result from neurodegen- local mast cells and immune cells to suppress the release
eration85. These drugs have also been proposed for the of mediators (such as histamines and TNF) that can
treatment of tumours, owing to their antiproliferative sensitize primary afferent neurons, thereby rendering
and pro-apoptotic effects2,86–88. However, as cannabi- them more sensitive to pain94. This anti-inflammatory
noids can also modulate innate and adaptive immune mechanism has yet to be established; however, it is

Table 2 | Cannabinoid effects on adaptive immunity and T helper cells

Cannabinoid Receptor Cell or tissue type Cytokine stimulant or Effect Reference
inflammation model
Mice
THC ND Spleen Legionella pneumophila Decreases IFN-γ and IgG2a 70
THC CB1 and CB2 Spleen Legionella pneumophila Decreases IL-12 and IL-12R 71
dependent and increases IL-4
THC CB2 dependent Spleen Tumour model Decreases IFN-γ and 73
increases IL-10 and TGF-β
WIN55,212-2 ND Spleen Theiler’s murine Decreases IFN-γ 78
encephalomyelitis virus
Humans
THC CB2 dependent Peripheral-blood T cells Allogeneic dendritic cells Decreases IFN-γ 76
Marijuana ND Peripheral-blood Phytohaemagglutinin Decreases IL-2 and 77
smoking mononuclear cells and concanavalin A increases IL-10 and TGF-β
CB, cannabinoid receptor; IFN-γ, interferon-γ; IL, interleukin; IL-12R, IL-12 receptor; ND, not determined; TGF-β, transforming growth factor-β;
THC, ∆9-tetrahydrocannabinol.

NATURE REVIEWS | IMMUNOLOGY VOLUME 5 | MAY 2005 | 4 0 5

© 2005 Nature Publishing Group
REVIEWS

interesting that, in a recent small clinical trial, the THC- symptoms and disease progression of multiple sclerosis
based drug dronabinol (Marinol; Unimed Pharma- and other neurodegenerative diseases. These mecha-
ceuticals, Inc.) was shown to have a small but clinically nisms might include the inhibition of TH1-cell responses
relevant analgesic effect on central pain in patients with in the CNS, owing to increased steroid production and
multiple sclerosis; this is possibly associated with, in other mechanisms, and the suppression of neuro-
part, suppression of release of these sensitizing mediators immune mediators of neuropathic pain that are locally
through a CB2-dependent mechanism95. From these released (FIG. 3). In addition to these mechanisms, the
findings and other data, it seems that cannabinoids have action of cannabinoids in reducing spasticity, which
several mechanisms of action for the attenuation of does not seem to be mediated by inflammatory

Table 3 | Preclinical and clinical studies examining the anti-inflammatory effects of cannabinoids
Preclinical model Drugs Outcome or mechanism Clinical trials Drug Outcome Refs
Nervous-tissue inflammation
EAE in rats and guinea pigs THC Decreased CNS inflammation – – – 89
EAE in rats THC Reduced disease progression – – – 90
and increased corticosterone
production
Theiler’s murine WIN55,212-2 Decreased DTH responses and – – – 78
encephalomyelitis-virus- IFN-γ production
induced EAE in mice
Theiler’s murine- WIN55,212-2, Decreased CD4+ T-cell influx – – – 79
encephalomyelitis-virus- ACEA or to CNS
induced EAE in mice JWH-015
Neuropathic pain in rats AM1241 Decreased nerve hypersensitivity Central pain in Dronabinol Analgesia 94,95
(CB2 involved) and pro- patients with
inflammatory-mediator production multiple sclerosis
Closed head injury in mice 2-AG Decreased brain oedema and Cerebral pressure and HU-211 Favourable 100,101
improved clinical outcome clinical outcome in
(CB1 involved) patients with head injury;
Phase III trial for HU-211 No efficacy 102
patients with traumatic
brain injury
Closed head injury in rats HU-211 Decreased TNF production and – – – 56
neuropathology
Inflammatory bowel disease
LPS-induced gastrointestinal ACEA or Decreased gastrointestinal – – – 107
transit in rats JWH-133 transit (CB2 involved) and
pro-inflammatory-mediator
production
Chemically induced colitis HU-210 Decreased colonic inflammation – – – 108
in mice (CB1 involved) and inflammatory-
cell influx
Arthritis
Leukocyte influx in mice Ajulemic acid Decreased granulocyte influx, – – – 13
and adjuvant-induced joint inflammation and
arthritis in rats prostaglandin production
PBMCs and synovial-fluid Ajulemic acid Decreased IL-1β production – – – 110
monocytes
Collagen-induced arthritis Cannabidiol Decreased arthritis, cell-mediated – – – 112
in mice immunity, and IFN-γ and TNF
production
Collagen-induced arthritis HU-320 Decreased arthritis, cell-mediated – – – 114
in mice immunity and TNF production
Vascular inflammation
LPS-induced hypotension SR141716A Decreased hypotension (2-AG – – – 115
in rats and CB1 mediate hypotension)
Myocardial ischaemia– WIN55,212-2 Decreased tissue injury and – – – 58
reperfusion injury in mice cytokine production (CB2 involved)
Septic shock in mice HU-211 Decreased lethality and TNF – – – 117
production
ACEA, arachidonoyl-2′-chloroethylamide; 2-AG, 2-arachidonoylglycerol; CB, cannabinoid receptor; CNS, central nervous system; DTH, delayed-type hypersensitivity;
EAE, experimental allergic encephalomyelitis; IFN-γ, interferon-γ; IL-1β, interleukin-1β; LPS, lipopolysaccharide; PBMC, peripheral-blood mononuclear cell;
THC, ∆9-tetrahydrocannabinol; TNF, tumour-necrosis factor.

406 | MAY 2005 | VOLUME 5 www.nature.com/reviews/immunol

© 2005 Nature Publishing Group
 REVIEWS

changes96, has been analysed in clinical trials and has gastrointestinal transit was associated with the attenua-
been shown to have efficacy in reducing spasms and tion of inflammatory changes, such as suppressing
stiffness97–99. the levels of cyclooxygenases and inducible nitric-oxide
Several studies have shown that endocannabi- synthase, indicating that prostaglandins and/or nitric
noids (such as AEA and 2-AG), as well as synthetic oxide mediate the effects of CB2 agonists on gastro-
cannabinoids that bind cannabinoid receptors with intestinal transit107.Anti-inflammatory effects of cannabi-
very low affinity (such as HU-211), are neuroprotec- noids were also observed in a mouse model of chemically
tive after traumatic brain injury. The mechanisms that induced colitis108. Both colonic-tissue ulceration and the
mediate the protective effects involve cannabinoid- inflammatory response were attenuated by cannabinoid
receptor function, as well as NMDA receptors, inhibi- treatment, as well as in FAAH-deficient mice, as assessed
tion of GLUTAMATERGIC SYNAPTIC TRANSMISSION, and reduction histologically and by measuring tissue myeloperoxidase
of TNF production and oxidative stress5. For example, activity (which correlates with neutrophil influx). By
levels of endogenous 2-AG were found to be markedly contrast, ulceration and inflammation were increased
increased in mice with closed head injury, and adminis- in mice deficient in CB1 and in mice treated with a CB1
tration of 2-AG following injury reduced brain antagonist. From these studies, it seems that cannabi-
oedema (a consequence of inflammation) and led to noids regulate the tissue response to inflammation in
a better clinical outcome 100 (TABLE 3). The effect of the colon. It is possible that this regulation occurs on
2-AG was attenuated by pretreatment with a CB1 two levels: the first, involving the smooth-muscle
antagonist, indicating the involvement of cannabinoid response to pro-inflammatory mediators that affect
receptors. However, receptor-independent mecha- gastrointestinal transit time; and the second, involving
nisms also seem to be involved. For example, admin- the direct suppression of pro-inflammatory-mediator
istration of HU-211 attenuated neuropathology in production (FIG. 3). Further studies will be needed to
rats with closed head injury and decreased TNF pro- determine the relative contribution of each of these
duction in the brain, indicating that the suppression possibilities.
of cytokine production by HU-211 was at least partly
responsible for the neuroprotection 56. Similarly, Arthritis. Several cannabinoids have been shown to be
administration of HU-211 to patients with severe anti-inflammatory in animal models of arthritis. The
closed head injury was initially shown to result in a first of these is the dimethylheptyl homologue of the
better clinical outcome101 but was recently reported natural C. sativa plant product THC-11-oic acid13. This
not to have any clinical efficacy102. As well as TNF, derivative, known as ajulemic acid, binds cannabinoid
other pro-inflammatory mediators, such as the IL-1R receptors with very low affinity. It was initially tested for
antagonist, might also be involved in the neuroprotec- its effects on the AIR-POUCH INFLAMMATORY RESPONSE in mice
tive effects of cannabinoids103. So, it seems that and on adjuvant-induced arthritis in rats13 (TABLE 3).
cannabinoids and endocannabinoids regulate some of Daily feeding of ajulemic acid suppressed leukocyte
the inflammatory aspects of brain injury and that this accumulation after injection of IL-1β and TNF into the
occurs by both cannabinoid-receptor-mediated and pouch. In the adjuvant-induced arthritis model, chronic
non-cannabinoid-receptor-mediated mechanisms. It administration (every third day) of ajulemic acid attenu-
GLUTAMATERGIC SYNAPTIC is possible that these drugs reduce brain oedema and ated inflammation in the joints, compared with animals
TRANSMISSION other aspects of neuroinflammation by inhibiting that did not receive ajulemic acid. In terms of the mecha-
Glutamic acid is the main NMDA receptors, by functioning as antioxidants and nism of action, ajulemic acid was shown in in vitro stud-
excitatory transmitter in the
central nervous system, where
by reducing the levels of pro-inflammatory cytokines ies to suppress prostaglandin production to a greater
it mediates fast synaptic in the brain (FIG. 3). extent than the anti-inflammatory drug INDOMETHACIN13.
transmission. It is released from In other studies, ajulemic acid was shown to be more
the terminal of a glutamatergic Inflammatory bowel disease. Inflammatory bowel dis- potent than common non-steroidal anti-inflammatory
nerve, crosses the synaptic cleft
ease, which includes ulcerative colitis and Crohn’s disease, drugs in suppressing adjuvant-induced arthritis, with
and acts on postsynaptic
receptors. affects millions of individuals. Better therapies are needed less gastrointestinal ulceration occurring 109, and it
to control the chronic inflammation that is associated was shown to decrease the LPS-induced production
AIR-POUCH INFLAMMATORY with this disease, which leads to increased intestinal of IL-1β in cultures of human PBMCs or synovial-
RESPONSE motility and faecal transit, resulting in pain, diarrhoea fluid monocytes110. So, ajulemic acid might have thera-
An experimental model of acute
inflammation. Skin pouches are
and poor ability to digest food3. The endocannabinoid peutic potential for patients with arthritis or other
established on the backs of mice, system seems to be involved in the inhibition of intestinal chronic inflammatory diseases, and it could have fewer
by subcutaneous injection of air motility, because functional CB1 is expressed in the side-effects than conventional therapies; however,
on several consecutive days. human ileum and colon104, and its expression is increased there is some controversy concerning its psychoactive
Subsequently, inflammation
during inflammation105. Furthermore, AEA and 2-AG potential111.
is induced by injection of
interleukin-1β and tumour- can be present in high levels in the gut105 and can inhibit The second non-psychoactive component of C. sativa
necrosis factor into the pouch colonic propulsion in mice106. Also, activation of CB2 was that has anti-inflammatory potential is cannabidiol112,113.
cavity. recently shown to attenuate LPS-induced increases in This drug was tested in a mouse model of collagen-
gastrointestinal transit107 (TABLE 3), and CB1 agonists were induced arthritis, and when administered either intra-
INDOMETHACIN
A cyclooxygenase inhibitor and
shown to be suppressive — but only for basal, and not peritoneally or orally (every day), beginning at the
thereby a non-steroidal anti- for LPS-induced, gastrointestinal transit. Interestingly, first signs of arthritis, it effectively blocked progres-
inflammatory drug. CB 2-agonist-mediated inhibition of LPS-induced sion of the disease. It seemed to function as an

NATURE REVIEWS | IMMUNOLOGY VOLUME 5 | MAY 2005 | 4 0 7

© 2005 Nature Publishing Group
REVIEWS

Cannabinoid-receptor dependent Cannabinoid-receptor independent

Neurodegeneration Inflammatory bowel disease Brain injury Vascular inflammation Arthritis

↑ Vasodilation
↑ Corticosteroids ↓ Gastrointestinal transit ↓ Oxidative stress ↓ Prostaglandins
↓ Myeloperoxidase
↓ TH1 responses ↓ Prostaglandins ↓ NMDAR binding ↓ Cell-mediated immunity
↓ CXCL8
↓ Neuroimmune ↓ Nitric oxide ↓ Pro-inflammatory ↓ IL-1β
↓ IL-1β
mediators ↓ Colonic inflammation cytokines ↓ TNF
↓ TNF

Figure 3 | Anti-inflammatory effects of cannabinoid-based drugs. Symptoms of neurodegeneration are attenuated by

treatment with cannabinoid-based drugs. This process is mediated, at least in part, through binding of cannabinoid receptors.
It involves an increase in corticosteroid release and a decrease in both T helper 1 (TH1)-cell responses in the central nervous
system and in neuroimmune-mediator production, including histamine and tumour-necrosis factor (TNF). Similarly, brain
oedema that occurs after injury is suppressed by treatment with cannabinoid-based drugs. This process is also mediated, at
least in part, through binding of cannabinoid receptors. It involves a decrease in oxidative stress, NMDA (N-methyl-D-aspartate)
receptor (NMDAR) binding and cytokine production, including TNF. The pathophysiology of inflammatory bowel disease is
also suppressed by cannabinoids. This involves binding to cannabinoid receptors, and there is a concomitant decrease in
gastrointestinal-tract faecal transit, prostaglandin and nitric-oxide production, and colonic-tissue inflammation. The
inflammation that is associated with arthritis is suppressed by cannabinoids. This occurs through cannabinoid-receptor-
independent mechanisms, which involve a decrease in prostaglandin production and cell-mediated immunity, together with a
decrease in cytokine production, including interleukin-1β (IL-1β) and TNF. Finally, vascular inflammatory disease might be partly
regulated through the binding of cannabinoids to cannabinoid receptors and through vasodilation effects. This occurs together
with a decrease in local pro-inflammatory-mediator production, including myeloperoxidase, CXC-chemokine ligand 8 (CXCL8),
IL-1β and TNF, through cannabinoid-receptor-dependent or -independent mechanisms.

immunosuppressant, because cells from the draining injury in mice, as measured by infarct size, was consider-
lymph node of treated mice showed reduced cell- ably reduced by pretreatment with WIN55,212-2, and
mediated immune functions and IFN-γ production in the protective effect was attenuated by a CB2 antagonist
response to stimulation with antigen. Furthermore, but not a CB1 antagonist58. Treatment with WIN55,212-2
treatment with cannabidiol blocked the LPS-induced also reduced the levels of myeloperoxidase, IL-1β and
increase in serum TNF and other immune responses. CXCL8 in injured tissue, indicating that it suppresses
In a subsequent study, similar results were obtained at the mobilization of pro-inflammatory mediators. The
lower doses using a more potent dimethylheptyl authors of this study speculated that WIN55,212-2
derivative of cannabidiol114. These data indicate that might have exerted its effect by binding CB2 at the cell
plant-derived cannabinoids and their synthetic deriv- surface of macrophages in the injured tissue58. There is
atives are anti-inflammatory and immunosuppressive; also a third mechanism in that cannabinoid effects that
their mechanisms of action are independent of are independent of both CB1 and CB2 have been shown
cannabinoid receptors and are mediated, in part, by to attenuate septic shock117. The non-psychoactive
suppression of pro-inflammatory-cytokine production cannabinoid HU-211, when given as single pretreatment
by lymphocytes and macrophages (FIG. 3). dose, reduced mouse and rat lethality in response to a
challenge with LPS. In addition, similar to effects in
Vascular inflammation. Many of the features of cardio- models of closed head injury, HU-211 markedly sup-
vascular disease and atherosclerosis include the elements pressed TNF production, which might explain the atten-
and mechanisms of the inflammatory cascade, and this uation of septic shock. So, it seems that cannabinoids
contributes to thrombosis and tissue destruction. This is have both a hypotensive effect that is CB1 mediated and
also true of the pathophysiology of septic shock, in an anti-inflammatory effect that is either CB2 mediated
which pro-inflammatory mediators are released into the or independent of cannabinoid receptors (FIG. 3). This
blood in excess, leading to thrombosis, vasodilation, cap- complexity provides both challenges and opportunities
illary leakage and organ-system failure. Recent evidence in the management of cardiovascular diseases.
indicates that cannabinoids might influence vascular
inflammatory disease in several ways. One way is as a Conclusions
mediator of vasodilation, particularly LPS-induced Several general principles emerge from the studies
hypotension4,115. Levels of the endocannabinoids AEA that are discussed in this Review. The first is that
and 2-AG were found to increase in the sera of patients endocannabinoids are expressed by immune cells and
with endotoxic shock116, and CB1 antagonists were that cannabinoid receptors at the surface of immune
shown to prevent LPS-induced hypotension in animals, cells are activated after infection or immune stimulation.
indicating that endocannabinoids and CB1 might func- The consequences of this for the immune response are
tion as an endogenous system that promotes LPS- not fully understood, but regulation of cellular chemo-
induced hypotension115 (TABLE 3). In addition to this taxis seems to be involved. Furthermore, because
effect of CB1, there is also evidence of a second way in cannabinoid receptors are G-protein-coupled receptors
which vascular inflammatory disease might be influ- (similar to receptors for chemokines and lipid media-
enced, because CB2 is involved in vascular changes dur- tors), their ligation during immune stimulation proba-
ing inflammation. Myocardial ischaemia–reperfusion bly leads to regulation of gene products that are

408 | MAY 2005 | VOLUME 5 www.nature.com/reviews/immunol

© 2005 Nature Publishing Group
 REVIEWS

required for immune-cell function. The second general which are known to be important for immune-cell
principle is that the main immune targets of cannabi- function121. The third general principle involves the
noid-based drugs involve the suppression of cytokines suppression, by marijuana-based drugs, of the chronic
and cell-mediated immunity, through cannabinoid- inflammatory response and the subsequent attenu-
receptor-dependent and -independent mechanisms. ation of disease processes and symptoms. These
The cannabinoid-receptor-dependent mechanisms anti-inflammatory effects are undoubtedly, in part,
most probably involve G-protein signalling and the associated with the ability of these drugs to suppress the
regulation of numerous cytokine genes. Regarding expression of cytokines, as well as other endogenous
cannabinoid-receptor-independent mechanisms, in the pro-inflammatory mediators. In addition, these drugs
case of endocannabinoids, their effects could be medi- might also function by increasing the production of
ated through vanilloid receptors, which are known to anti-inflammatory mediators. Further elucidation of the
bind these compounds118, or they might result from effect of marijuana-based drugs on pro-inflammatory
the binding of as-yet-uncharacterized receptors119. and anti-inflammatory mechanisms will provide the
Cannabinoid-receptor-independent mechanisms might basis for the formulation of more effective drugs for
also involve effects on lipid-raft structure and function120, the management of chronic inflammatory diseases.

1. Tomida, I., Pertwee, R. G. & Azuara-Blanco, A. 18. Devane, W. A. et al. Isolation and structure of a brain 36. Maccarrone, M. et al. Lipopolysaccharide downregulates
Cannabinoids and glaucoma. Br. J. Ophthalmol. constituent that binds to the cannabinoid receptor. fatty acid amide hydrolase expression and increases
88, 708–713 (2004). Science 258, 1946–1949 (1992). anandamide levels in human peripheral lymphocytes.
2. Guzman, M. Cannabinoids: potential anticancer agents. An important breakthrough in the field was the Arch. Biochem. Biophys. 393, 321–328 (2001).
Nature Rev. Cancer 3, 745–755 (2003). discovery of an endocannabinoid. This paper describes References 32–36 discuss the metabolism
3. Kunos, G. & Pacher, P. Cannabinoids cool the intestine. the isolation of the first endocannabinoid to be found, of endocannabinoids in immune cells.
Nature Med. 10, 678–679 (2004). AEA (previously known as anandamide). 37. Wyss-Coray, T. & Mucke, L. Inflammation in
4. Mendizabal, V. E. & Adler-Graschinsky, E. Cannabinoid 19. Zygmunt, P. M. et al. Vanilloid receptors on sensory nerves neurodegenerative disease — a double-edged sword.
system as a potential target for drug development in the mediate the vasodilator action of anandamide. Nature Neuron 35, 419–432 (2002).
treatment of cardiovascular disease. Curr. Vasc. Pharmacol. 400, 452–457 (1999). 38. Moser, B., Wolf, M., Walz, A. & Loetscher, P. Chemokines:
1, 301–313 (2003). 20. Mechoulam, R. et al. Identification of an endogenous multiple levels of leukocyte migration control. Trends
5. Mechoulam, R., Panikashvili, D. & Shohami, E. 2-monoglyceride, present in canine gut, that binds to Immunol. 25, 75–84 (2004).
Cannabinoids and brain injury: therapeutic implications. cannabinoid receptors. Biochem. Pharm. 50, 83–90 39. Oka, S. et al. 2-Arachidonoylglycerol, an endogenous
Trends Mol. Med. 8, 58–61 (2002). (1995). cannabinoid receptor ligand, induces the migration of EoL-1
6. Baker, D. & Pryce, G. The therapeutic potential of 21. Hanus, L. et al. 2-Arachidonyl glyceryl ether, an endogenous human eosinophilic leukemia cells and human peripheral
cannabis in multiple sclerosis. Expert Opin. Investig. agonist of the cannabinoid CB1 receptor. Proc. Natl Acad. blood eosinophils. J. Leukoc. Biol. 76, 1002–1009 (2004).
Drugs 12, 561–567 (2003). Sci. USA 98, 3662–3665 (2001). 40. Rayman, N. et al. Distinct expression profiles of the
7. Baker, D., Pryce, G., Giovannoni, G. & Thompson, A. J. 22. Matsuda, L. A., Lolait, S. J., Brownstein, M. J., Young, A. C. & peripheral cannabinoid receptor in lymphoid tissues
The therapeutic potential of cannabis. Lancet Neurol. Bonner, T. I. Structure of cannabinoid receptor and functional depending on receptor activation status. J. Immunol.
2, 291–298 (2003). expression of the cloned cDNA. Nature 346, 561–564 (1990). 172, 2111–2117 (2004).
8. Di Marzo, V., Bifulco, M. & De Petrocellis, L. The 23. Munro, S., Thomas, K. L. & Abu-Shaar, M. Molecular 41. Maestroni, G. J. The endogenous cannabinoid
endocannabinoid system and its therapeutic exploitation. characterization of a peripheral receptor for cannabinoids. 2-arachidonoyl glycerol as in vivo chemoattractant for
Nature Rev. Drug Discov. 3, 771–784 (2004). Nature 365, 61–65 (1993). dendritic cells and adjuvant for TH1 response to a soluble
9. Black, S. C. Cannabinoid receptor antagonists and obesity. References 20–23 describe the isolation of protein. FASEB J. 18, 1914–1916 (2004).
Curr. Opin. Investig. Drugs 5, 389–394 (2004). endocannabinoids that were discovered subsequently 42. Alberich Jorda, M. et al. The peripheral cannabinoid
10. Gaoni, Y. & Mechoulam, R. Isolation, structure, and partial to AEA and the isolation of the cannabinoid receptors, receptor Cb2, frequently expressed on AML blasts,
synthesis of an active constituent of hashish. J. Am. Chem. CB1 and CB2. either induces a neutrophilic differentiation block or confers
Soc. 86, 1646–1647 (1964). 24. McAllister, S. D. & Glass, M. CB1 and CB2 receptor- abnormal migration properties in a ligand-dependent
This is a seminal paper that describes the isolation mediated signalling: a focus on endocannabinoids. manner. Blood 104, 526–534 (2004).
and synthesis of THC. Prostaglandins Leukot. Essent. Fatty Acids 66, 161–171 The many putative roles of cannabinoid receptors
11. Howlett, A. C. et al. International Union of Pharmacology. (2002). in immune-cell biology remain undefined. This paper
XXVII. Classification of cannabinoid receptors. Pharmacol. 25. Klein, T. W. et al. The cannabinoid system and immune indicates that these receptors might have a role
Rev. 54, 161–202 (2002). modulation. J. Leukoc. Biol. 74, 486–496 (2003). in leukaemia.
For those wanting more on the chemistry and biology 26. Klein, T., Newton, C. & Friedman, H. Cannabinoid receptors 43. Szabo, I. et al. Heterologous desensitization of opioid
of the cannabinoid system, this is a recent and and immunity. Immunol. Today 19, 373–381 (1998). receptors by chemokines inhibits chemotaxis and enhances
authoritative review. 27. Berdyshev, E. V. Cannabinoid receptors and the regulation the perception of pain. Proc. Natl Acad. Sci. USA
12. Rhee, M. H. et al. Cannabinol derivatives: binding to of immune response. Chem. Phys. Lipids 108, 169–190 99, 10276–10281 (2002).
cannabinoid receptors and inhibition of adenylylcyclase. (2000). 44. Lee, S. F., Newton, C., Widen, R., Friedman, H. & Klein, T. W.
J. Med. Chem. 40, 3228–3233 (1997). 28. Schlicker, E. & Kathmann, M. Modulation of transmitter Differential expression of cannabinoid CB2 receptor mRNA
13. Zurier, R. B. et al. Dimethylheptyl-THC-11 oic acid. release via presynaptic cannabinoid receptors. Trends in mouse immune cell subpopulations and following B cell
A nonpsychoactive antiinflammatory agent with Pharmacol. Sci. 22, 565–572 (2001). stimulation. Eur. J. Pharmacol. 423, 235–241 (2001).
a cannabinoid template structure. Arthritis Rheum. 29. Christie, M. J. & Vaughan, C. W. Cannabinoids act 45. Daaka, Y., Friedman, H. & Klein, T. W. Cannabinoid receptor
41, 163–170 (1998). backwards. Nature 410, 527–530 (2001). proteins are increased in Jurkat, human T-cell line after
14. Burstein, S. H. The cannabinoid acids: nonpsychoactive 30. Akira, S. & Takeda, K. Toll-like receptor signalling. mitogen activation. J. Pharmacol. Exp. Ther. 276, 776–783
derivatives with therapeutic potential. Pharmacol. Ther. Nature Rev. Immunol. 4, 499–511 (2004). (1996).
82, 87–96 (1999). 31. Noverr, M. C., Erb-Downward, J. R. & Huffnagle, G. B. This report was among the first to show that
15. Liu, J., Li, H., Burstein, S. H., Zurier, R. B. & Chen, J. D. Production of eicosanoids and other oxylipins by cannabinoid-receptor expression changes with
Activation and binding of peroxisome proliferator-activated pathogenic eukaryotic microbes. Clin. Microbiol. Rev. variations in immune-cell activation.
receptor γ by synthetic cannabinoid ajulemic acid. 16, 517–533 (2003). 46. Noe, S. N., Newton, C., Widen, R., Friedman, H. & Klein, T. W.
Mol. Pharmacol. 63, 983–992 (2003). 32. DiMarzo, V., DePetrocellis, L., Sepe, N. & Buono, A. Anti-CD40, anti-CD3, and IL-2 stimulation induce contrasting
Cannabinoid-based drugs might function through Biosynthesis of anandamide and related acylethanolamides changes in CB1 mRNA expression in mouse splenocytes.
mechanisms other than binding cannabinoid in mouse J774 macrophages and N18 neuroblastoma cells. J. Neuroimmunol. 110, 161–167 (2000).
receptors, such as through binding PPAR-γ, Biochem. J. 316, 977–984 (1996). 47. Gardner, B. et al. Autocrine and paracrine regulation of
as described in this paper. 33. DiMarzo, V. et al. Biosynthesis and inactivation of the lymphocyte CB2 receptor expression by TGF-β. Biochem.
16. Piomelli, D. The molecular logic of endocannabinoid endocannabinoid 2-arachidonoylglycerol in circulating Biophys. Res. Commun. 290, 91–96 (2002).
signalling. Nature Rev. Neurosci. 4, 873–884 (2003). and tumoral macrophages. Eur. J. Biochem. 264, 258–267 48. Klein, T. W., Newton, C., Zhu, W., Daaka, Y. & Friedman, H.
Cannabinoids are best known for their psychoactive (1999). ∆9-Tetrahydrocannabinol, cytokines and immunity to
effects on the brain. This Review, together with 34. Maccarrone, M., Bari, M., Battista, N. & Finazzi-Agro, A. Legionella pneumophila. Proc. Soc. Exp. Biol. Med.
references 28 and 29, discusses these mechanisms Endocannabinoid degradation, endotoxic shock and 209, 205–212 (1995).
in detail. inflammation. Curr. Drug Targets Inflamm. Allergy 1, 53–63 49. Carlisle, S., Marciano-Cabral, F., Staab, A., Ludwick, C. &
17. Sugiura, T., Kobayashi, Y., Oka, S. & Waku, K. Biosynthesis (2002). Cabral, G. Differential expression of the CB2 cannabinoid
and degradation of anandamide and 2-arachidonoylglycerol 35. Matias, I. et al. Presence and regulation of the receptor by rodent macrophages and macrophage-like cells
and their possible physiological significance. Prostaglandins endocannabinoid system in human dendritic cells. in relation to cell activation. Int. Immunopharmacol. 2, 69–82
Leukot. Essent. Fatty Acids 66, 173–192 (2002). Eur. J. Biochem. 269, 3771–3778 (2002). (2002).

NATURE REVIEWS | IMMUNOLOGY VOLUME 5 | MAY 2005 | 4 0 9

© 2005 Nature Publishing Group
REVIEWS

50. Walter, L. et al. Nonpsychotropic cannabinoid receptors 71. Klein, T. W., Newton, C. A., Nakachi, N. & Friedman, H. 95. Svendsen, K. B., Jensen, T. S. & Bach, F. W. Does the
regulate microglial cell migration. J. Neurosci. 23, 1398–1405 ∆9-Tetrahydrocannabinol treatment suppresses immunity cannabinoid dronabinol reduce central pain in multiple
(2003). and early IFNγ, IL-12, and IL-12 receptor β2 responses sclerosis? Randomised double blind placebo controlled
Although CB2 has been reported to be expressed in to Legionella pneumophila infection. J. Immunol. 164, crossover trial. BMJ 329, 253 (2004).
the brain, this was the first report to indicate that it 6461–6466 (2000). 96. Baker, D. et al. Endocannabinoids control spasticity in a
might be expressed by brain microglial cells. 72. Braun, M. C. & Kelsall, B. L. Regulation of interleukin-12 multiple sclerosis model. FASEB J. 15, 300–302 (2001).
51. Bromley, S. K. et al. The immunological synapse. Annu. Rev. production by G-protein-coupled receptors. Microbes 97. Zajicek, J. et al. Cannabinoids for treatment of spasticity and
Immunol. 19, 375–396 (2001). Infect. 3, 99–107 (2001). other symptoms related to multiple sclerosis (CAMS study):
52. Blanchard, D. K., Newton, C., Klein, T. W., Stewart, W. E. & 73. Zhu, L. X. et al. ∆9-Tetrahydrocannabinol inhibits antitumor multicentre randomised placebo-controlled trial. Lancet
Friedman, H. In vitro and in vivo suppressive effects of immunity by a CB2 receptor-mediated, cytokine-dependent 362, 1517–1526 (2003).
∆9-tetrahydrocannabinol on interferon production by murine pathway. J. Immunol. 165, 373–380 (2000). 98. Vaney, C. et al. Efficacy, safety and tolerability of an orally
spleen cells. Int. J. Immunopharmacol. 8, 819–824 (1986). 74. Piccirillo, C. A. & Thornton, A. M. Cornerstone of peripheral administered cannabis extract in the treatment of spasticity
53. Cabral, G. A., Lockmuller, J. C. & Mishkin, E. M. tolerance: naturally occurring CD4+CD25+ regulatory T cells. in patients with multiple sclerosis: a randomized, double-
∆9-Tetrahydrocannabinol decreases α/β interferon Trends Immunol. 25, 374–380 (2004). blind, placebo-controlled, crossover study. Mult. Scler.
response to herpes simplex virus type 2 in the B6C3F1 75. Weiner, H. L. Induction and mechanism of action of 10, 417–424 (2004).
mouse. Proc. Soc. Exp. Biol. Med. 181, 305–311 (1986). transforming growth factor-β-secreting TH3 regulatory 99. Wade, D. T., Makela, P., Robson, P., House, H. & Bateman, C.
54. Klein, T. W., Lane, B., Newton, C. A. & Friedman, H. cells. Immunol. Rev. 182, 207–214 (2001). Do cannabis-based medicinal extracts have general or
The cannabinoid system and cytokine network. Proc. Soc. 76. Yuan, M. et al. ∆9-Tetrahydrocannabinol regulates specific effects on symptoms in multiple sclerosis? A double-
Exp. Biol. Med. 225, 1–8 (2000). TH1/TH2 cytokine balance in activated human T cells. blind, randomized, placebo-controlled study on 160
55. Smith, S. R., Terminelli, C. & Denhardt, G. Effects of J. Neuroimmunol. 133, 124–131 (2002). patients. Mult. Scler. 10, 434–441 (2004).
cannabinoid receptor agonist and antagonist ligands on 77. Pacifici, R. et al. Modulation of the immune system 100. Panikashvili, D. et al. An endogenous cannabinoid (2-AG) is
production of inflammatory cytokines and anti-inflammatory in cannabis users. JAMA 289, 1929–1931 (2003). neuroprotective after brain injury. Nature 413, 527–531
interleukin-10 in endotoxemic mice. J. Pharmacol. Exp. Ther. This report provided the first evidence from a human (2001).
293, 136–150 (2000). study to support the findings in animals that indicate 101. Knoller, N. et al. Dexanabinol (HU-211) in the treatment of
56. Shohami, E., Gallily, R., Mechoulam, R., Bass, R. & that marijuana-derived cannabinoids bias TH-cell severe closed head injury: a randomized, placebo-controlled,
Ben-Hur, T. Cytokine production in the brain following responses. Phase II clinical trial. Crit. Care Med. 30, 548–554 (2002).
closed head injury: dexanabinol (HU-211) is a novel 78. Croxford, J. L. & Miller, S. D. Immunoregulation of a viral 102. Pharmos Corporation. Dexanabinol did not demonstrate
TNF-α inhibitor and an effective neuroprotectant. model of multiple sclerosis using the synthetic cannabinoid efficacy [online], http://www.pharmoscorp.com/news/
J. Neuroimmunol. 72, 169–177 (1997). R(+)WIN55,212. J. Clin. Invest. 111, 1231–1240 (2003). pr/pr122004.html (2004).
Using an animal model of closed head injury, this 79. Arevalo-Martin, A., Vela, J. M., Molina-Holgado, E., 103. Molina-Holgado, F. et al. Endogenous interleukin-1 receptor
paper shows that non-psychoactive cannabinoids Borrell, J. & Guaza, C. Therapeutic action of cannabinoids antagonist mediates anti-inflammatory and neuroprotective
can suppress the production of pro-inflammatory in a murine model of multiple sclerosis. J. Neurosci. actions of cannabinoids in neurons and glia. J. Neurosci.
cytokines. 23, 2511–2516 (2003). 23, 6470–6474 (2003).
57. Doble, A. The role of excitotoxicity in neurodegenerative 80. Das, J. et al. A critical role for NF-κB in Gata3 expression 104. Manara, L. et al. Functional assessment of neuronal
disease: implications for therapy. Pharmacol. Ther. and TH2 differentiation in allergic airway inflammation. cannabinoid receptors in the muscular layers of human
81, 163–221 (1999). Nature Immunol. 2, 45–50 (2001). ileum and colon. Dig. Liver Dis. 34, 262–269 (2002).
58. Di Filippo, C., Rossi, F., Rossi, S. & D’Amico, M. 81. Klein, T. W. et al. Cannabinoid receptors and T helper 105. Izzo, A. A. et al. Cannabinoid CB1-receptor mediated
Cannabinoid CB2 receptor activation reduces mouse biasing. J. Neuroimmunol. 147, 91–94 (2004). regulation of gastrointestinal motility in mice in a model
myocardial ischemia–reperfusion injury: involvement 82. Herring, A. C. & Kaminski, N. E. Cannabinol-mediated of intestinal inflammation. Br. J. Pharmacol. 134, 563–570
of cytokine/chemokines and PMN. J. Leukoc. Biol. inhibition of nuclear factor-κB, cAMP response element- (2001).
75, 453–459 (2004). binding protein, and interleukin-2 secretion by activated 106. Pinto, L. et al. Endocannabinoids as physiological
59. Baldwin, G. C. et al. Marijuana and cocaine impair thymocytes. J. Pharmacol. Exp. Ther. 291, 1156–1163 regulators of colonic propulsion in mice. Gastroenterology
alveolar macrophage function and cytokine production. (1999). 123, 227–234 (2002).
Am. J. Respir. Crit. Care Med. 156, 1606–1613 (1997). 83. Foulds, J., Burke, M., Steinberg, M., Williams, J. M. & 107. Mathison, R., Ho, W., Pittman, Q. J., Davison, J. S. &
This is one of the few reports that shows a direct Ziedonis, D. M. Advances in pharmacotherapy for tobacco Sharkey, K. A. Effects of cannabinoid receptor-2
suppressive effect of marijuana smoking on immune- dependence. Expert Opin. Emerg. Drugs 9, 39–53 (2004). activation on accelerated gastrointestinal transit in
cell function. 84. Howlett, A. C. et al. Cannabinoid physiology and lipopolysaccharide-treated rats. Br. J. Pharmacol.
60. Klein, T. W., Newton, C., Widen, R. & Friedman, H. pharmacology: 30 years of progress. Neuropharmacology 142, 1247–1254 (2004).
∆9-Tetrahydrocannabinol injection induces cytokine-mediated 47, 345–358 (2004). 108. Massa, F. et al. The endogenous cannabinoid system
mortality of mice infected with Legionella pneumophila. 85. Baker, D. et al. Cannabinoids control spasticity and tremor protects against colonic inflammation. J. Clin. Invest.
J. Pharmacol. Exp. Ther. 267, 635–640 (1993). in a multiple sclerosis model. Nature 404, 84–87 (2000). 113, 1202–1209 (2004).
61. Zhu, W., Newton, C., Daaka, Y., Friedman, H. & Klein, T. W. The central effects of cannabinoids that bind CB1 The endocannabinoid system might control
∆9-Tetrahydrocannabinol enhances the secretion of might have a role in controlling spasticity and tremor and regulate intestinal inflammatory responses,
interleukin 1 from endotoxin-stimulated macrophages. in individuals with multiple sclerosis, as shown in as reported in this paper.
J. Pharmacol. Exp. Ther. 270, 1334–1339 (1994). animals with chronic, relapsing EAE. 109. Dajani, E. Z. et al. 1′,1′-Dimethylheptyl-∆8-
62. Smith, S. R., Terminelli, C. & Denhardt, G. Modulation of 86. De Petrocellis, L. et al. The endogenous cannabinoid tetrahydrocannabinol-11-oic acid: a novel, orally effective
cytokine responses in Corynebacterium parvum-primed anandamide inhibits human breast cancer cell proliferation. cannabinoid with analgesic and anti-inflammatory properties.
endotoxemic mice by centrally administered cannabinoid Proc. Natl Acad. Sci. USA 95, 8375–8380 (1998). J. Pharmacol. Exp. Ther. 291, 31–38 (1999).
ligands. Eur. J. Pharmacol. 425, 73–83 (2001). 87. Galve-Roperh, I. et al. Anti-tumoral action of cannabinoids: 110. Zurier, R. B., Rossetti, R. G., Burstein, S. H. & Bidinger, B.
63. Derocq, J. M. et al. Genomic and functional changes involvement of sustained ceramide accumulation and Suppression of human monocyte interleukin-1β production
induced by the activation of the peripheral cannabinoid extracellular signal-regulated kinase activation. Nature Med. by ajulemic acid, a nonpsychoactive cannabinoid.
receptor CB2 in the promyelocytic cells HL-60. Possible 6, 313–319 (2000). Biochem. Pharmacol. 65, 649–655 (2003).
involvement of the CB2 receptor in cell differentiation. 88. Bifulco, M. et al. A new strategy to block tumor growth 111. Burstein, S. & Zurier, R. B. Pain reduction and lack of
J. Biol. Chem. 275, 15621–15628 (2000). by inhibiting endocannabinoid inactivation. FASEB J. psychotropic effects with ajulemic acid. Arthritis Rheum.
64. Kishimoto, S. et al. 2-Arachidonoylglycerol, an endogenous 18, 1606–1608 (2004). 50, 4078–4079 (2004).
cannabinoid receptor ligand, induces accelerated In this paper, a novel mechanism is reported for 112. Malfait, A. M. et al. The nonpsychoactive cannabis
production of chemokines in HL-60 cells. J. Biochem. endocannabinoids in the inhibition of tumour growth. constituent cannabidiol is an oral anti-arthritic therapeutic in
(Tokyo) 135, 517–524 (2004). 89. Lyman, W. D., Sonett, J. R., Brosnan, C. F., Elkin, R. & murine collagen-induced arthritis. Proc. Natl Acad. Sci. USA
65. Pryce, G. et al. Cannabinoids inhibit neurodegeneration in Bornstein, M. B. ∆9-Tetrahydrocannabinol: a novel 97, 9561–9566 (2000).
models of multiple sclerosis. Brain 126, 2191–2202 (2003). treatment for experimental autoimmune encephalomyelitis. 113. Gallily, R. et al. γ-Irradiation enhances apoptosis induced
66. Ni, X. et al. WIN 55212-2, a cannabinoid receptor agonist, J. Neuroimmunol. 23, 73–81 (1989). by cannabidiol, a non-psychotropic cannabinoid, in cultured
attenuates leukocyte/endothelial interactions in an 90. Wirguin, I. et al. Suppression of experimental autoimmune HL-60 myeloblastic leukemia cells. Leuk. Lymphoma
experimental autoimmune encephalomyelitis model. encephalomyelitis by cannabinoids. Immunopharmacology 44, 1767–1773 (2003).
Mult. Scler. 10, 158–164 (2004). 28, 209–214 (1994). 114. Sumariwalla, P. F. et al. A novel synthetic, nonpsychoactive
67. Kawai, T. et al. Selective diapedesis of TH1 cells induced 91. Newton, C. A. et al. The THC-induced suppression of TH1 cannabinoid acid (HU-320) with antiinflammatory properties
by endothelial cell RANTES. J. Immunol. 163, 3269–3278 polarization in response to Legionella pneumophila infection in murine collagen-induced arthritis. Arthritis Rheum.
(1999). is not mediated by increases in corticosterone and PGE2. 50, 985–998 (2004).
68. Savinov, A. Y., Wong, F. S. & Chervonsky, A. V. IFN-γ affects J. Leukoc. Biol. 76, 854–861 (2004). References 112–114 use animal models to show the
homing of diabetogenic T cells. J. Immunol. 167, 6637–6643 92. Kingery, W. S. A critical review of controlled clinical trials anti-inflammatory potential of non-psychoactive
(2001). for peripheral neuropathic pain and complex regional pain cannabinoids.
69. Cabral, G. & Dove Pettit, D. Drugs and immunity: syndromes. Pain 73, 123–139 (1997). 115. Varga, K., Wagner, J. A., Bridgen, D. T. & Kunos, G.
cannabinoids and their role in decreased resistance to 93. Calignano, A., La Rana, G., Giuffrida, A. & Piomelli, D. Platelet- and macrophage-derived endogenous
infectious diseases. J. Neuroimmunol. 83, 116–123 (1998). Control of pain initiation by endogenous cannabinoids. cannabinoids are involved in endotoxin-induced
70. Newton, C. A., Klein, T. W. & Friedman, H. Secondary Nature 394, 277–281 (1998). hypotension. FASEB J. 12, 1035–1044 (1998).
immunity to Legionella pneumophila and TH1 activity 94. Ibrahim, M. M. et al. Activation of CB2 cannabinoid receptors 116. Wang, Y. et al. Simultaneous measurement of anandamide
are suppressed by ∆9-tetrahydrocannabinol injection. by AM1241 inhibits experimental neuropathic pain: pain and 2-arachidonoylglycerol by polymyxin B-selective
Infect. Immun. 62, 4015–4020 (1994). inhibition by receptors not present in the CNS. Proc. Natl adsorption and subsequent high-performance liquid
This was the first report to show that cannabinoids Acad. Sci. USA 100, 10529–10533 (2003). chromatography analysis: increase in endogenous
bias the immune response away from TH1-cell This study shows that cannabinoids might inhibit cannabinoids in the sera of patients with endotoxic shock.
responses. pain by binding CB2, as well as CB1. Anal. Biochem. 294, 73–82 (2001).

410 | MAY 2005 | VOLUME 5 www.nature.com/reviews/immunol

© 2005 Nature Publishing Group
 REVIEWS

117. Gallily, R. et al. Protection against septic shock and the immunological synapse: lipid raft and tetraspan Competing interests statement
suppression of tumor necrosis factor α and nitric oxide microdomains. Immunol. Rev. 189, 136–151 (2002). The author declares no competing financial interests.
production by dexanabinol (HU-211), a nonpsychotropic 122. Fischer-Stenger, K., Dove Pettit, D. A. & Cabral, G. A.
cannabinoid. J. Pharmacol. Exp. Ther. 283, 918–924 ∆9-Tetrahydrocannabinol inhibition of tumor necrosis factor-α:
(1997). suppression of post-translational events. J. Pharm. Exp. Ther. Online links
118. Veldhuis, W. B. et al. Neuroprotection by the endogenous 267, 1558–1565 (1993).
cannabinoid anandamide and arvanil against in vivo 123. Chang, Y. H., Lee, S. T. & Lin, W. W. Effects of cannabinoids DATABASES
excitotoxicity in the rat: role of vanilloid receptors and on LPS-stimulated inflammatory mediator release from The following terms in this article are linked online to:
lipoxygenases. J. Neurosci. 23, 4127–4133 (2003). macrophages: involvement of eicosanoids. J. Cell. Biochem. Entrez Gene:
119. Breivogel, C. S., Griffin, G., Di Marzo, V. & Martin, B. R. 81, 715–723 (2001). http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene
Evidence for a new G protein-coupled cannabinoid receptor 124. Facchinetti, F., Del Giudice, E., Furegato, S., Passarotto, M. & CB1 | CB2 | CXCL8 | IL-1β | IL-6 | myeloperoxidase | NMDA
in mouse brain. Mol. Pharmacol. 60, 155–163 (2001). Leon, A. Cannabinoids ablate release of TNFα in rat microglial receptor | PPAR-γ | TNF
This paper indicates that there might be more than cells stimulated with lipopolysaccharide. Glia 41, 161–168 OMIM: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM
two cannabinoid receptors. (2003). inflammatory bowel disease | multiple sclerosis
120. Biswas, K. K. et al. Membrane cholesterol but not putative
receptors mediates anandamide-induced hepatocyte Acknowledgements FURTHER INFORMATION
apoptosis. Hepatology 38, 1167–1177 (2003). I express sincere appreciation of H. Friedman and C. Newton for National Institute of Drug Abuse:
This paper shows that lipid rafts might be affected years of collaboration, resulting in many novel findings. I also thank http://www.nida.nih.gov/Infofax/HSYouthtrends.html
by cannabinoid-based drugs. the National Institute on Drug Abuse (United States) and the Thomas Klein’s homepage:
121. Vogt, A. B., Spindeldreher, S. & Kropshofer, H. Clustering National Institute of Allergy and Infectious Diseases (United States) http://hsc.usf.edu/medicine/medmicro/klein.html
of MHC–peptide complexes prior to their engagement in for continued support. Access to this interactive links box is free online.

NATURE REVIEWS | IMMUNOLOGY VOLUME 5 | MAY 2005 | 4 1 1

© 2005 Nature Publishing Group