N E W S A N D V I E W S

Inflammatory bowel diseases (IBDs) such asulcerative colitis and Crohn’s disease affectsover a million people in the United States1,with an estimated indirect cost from workloss of $3.6 billion annually2. Many of theseindividuals suffer from pain, diarrhea andpoor ability to digest their food, and in up tohalf of those with IBD, the disease is so severethat it ultimately requires surgery to removethe affected bowel segment.

Despite recent therapeutic advances andimproved understanding of the underlyingpathologies, patients with IBD are often resist-ant to treatment, justifying the continuedsearch for new therapeutic approaches.Although the mechanisms underlying ulcera-tive colitis and Crohn’s disease are different,they share one pathological feature: chronicinflammation. In a recent issue of the Journal ofClinical Investigation, Massa et al. provide evi-dence that stimulation of cannabinoid recep-tors protects against colonic inflammation3.

As their model, the authors induced bowelinflammation in mice by treatment with dif-ferent chemical agents, an approach com-monly used to explore endogenousprotective mechanisms and to screen poten-tial therapeutic agents. The authors beganwith an experiment testing chemical agentsin mice lacking the CB1 subtype of cannabi-noid receptor; such agents induced moresevere colitis in the CB1 knockout mice thanin wild-type mice. Moreover, pretreatmentof wild-type mice with a CB1 antagonistcaused a similar increase in the inflamma-tory response.

The authors then performed the converseexperiments. They found that either treat-ment of wild-type mice with a CB1 receptoragonist or genetic ablation of fatty acid ami-dohydrolase (FAAH), the enzyme thatdegrades the endogenous cannabinoid ago-nist anandamide4, reduced inflammation in

response to chemicals. The authors alsoobserved an increase in the expression of CB1receptors in intrinsic neurons of the inflamedmouse colon. Together, these findings suggestthat anandamide counteracts inflammation.Another endocannabinoid, 2-arachidonoyl-glycerol, is unlikely to be involved because itsabundance in tissue is unaffected by geneticablation of FAAH5.

These findings may offer a new therapeuticapproach to IBD. CB1 receptors in the brainmediate the addictive psychological effects ofmarijuana, so treating a chronic disease with a

drug that directly stimulates CB1 receptorswould be socially objectionable. Recent evi-dence indicates, however, that drugs which tar-get endogenous cannabinoids may not havethe same potential for abuse. A potent FAAHinhibitor, for example, elicits cannabinoid-likeantianxiety effects in mice without producingmany of the other behavioral effects of psy-choactive cannabinoids6.

The results of Massa et al. would thereforewarrant testing FAAH inhibitors as anti-inflammatory agents in the mouse model, as aprelude to clinical trials in IBD. It should be

678 VOLUME 10 | NUMBER 7 | JULY 2004 NATURE MEDICINE

Cannabinoids cool the intestineGeorge Kunos & Pál Pacher

Cannabinoids inhibit motility and secretion in the intestine. They are now assigned the additional task of curbingexcessive inflammation, suggesting that drugs targeting the endogenous cannabinoid system could be exploited forinflammatory bowel disease.

George Kunos and Pál Pacher are at the National

Institute on Alcohol Abuse and Alcoholism,

National Institutes of Health, Bethesda, Maryland

20892-8115, USA.

e-mail: gkunos@mail.nih.gov or

pacher@mail.nih.gov

Mucosalepithelium

Laminapropria

Muscularismucosae

Submucosa

Smoothmuscle layer

Myentericplexus(neurons)

TNFα

TNFαchemokines

CB1/CB2

CB1/CB2

Andandamide

CB1

CD14

TOLL4LPS

Bacteria

PMN Mastcell

Macrophage

ACh ortachykinins

Ly

a b

Figure 1 Cellular source and proposed targets of anti-inflammatory endocannabinoids in IBD. (a)Cross-section of inflamed bowel with leukocyte infiltration (polymorphonuclear, lymphocytes,macrophages, mast cells). (b) In macrophages, LPS induces the production of TNFα and chemokines(such as MIP-2 and CXCL-8) as well as anandamide. Anandamide is released to act as an autocrinemediator to inhibit TNFα and chemokine production via CB1 or CB2 receptors or both. Activation ofCB1 and CB2 receptors may similarly inhibit TNFα production in mast cells, with these effectsresulting in decreased leukocyte infiltration and inflammation. Paracrine activation of CB1 receptors onextrinsic and intrinsic enteric neurons inhibits acetylcholine and tachykinin release, respectively,resulting in inhibition of gut motility. These effects are amplified by treatment with a FAAH inhibitor,which prevents the breakdown of anandamide.

Deb

orah

Mai

zels

©20

04 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine

N E W S A N D V I E W S

noted, however, that the effect of FAAHinhibitors may not be entirely due to elevatedlevels of anandamide, as they also elevate thelevel of noncannabinoid substrates such asoleoylethanolamide6 or prostamides generatedvia COX-2 (ref. 7).

Historically, marijuana has been used totreat diarrhea and has been advocated for thetreatment of a variety of other gastrointesti-nal problems, including Crohn’s disease3.More recent pharmacological studies haveclearly established that cannabinoids inhibitgastrointestinal motility and secretion by act-ing on CB1 receptors located on the termi-nals of both intrinsic and extrinsicsubmucosal neurons8. When administered tomice with chemically induced enteritis,cannabinoids also reduce inflammation9 andfluid accumulation10 in the gut. In these lat-ter studies, high levels of anandamide and 2-arachidonoylglycerol as well as increasedexpression of CB1 receptors have beendetected in the inflamed intestines.

The novelty in the findings of Massa et al.3 isthe active protective role of the endocannabi-noid system, as indicated by the altered inflam-matory response of mice lacking CB1 receptorsor FAAH. Furthermore, CB1 activationreversed the electrophysiological signs ofsmooth muscle irritability and, at the sametime, blunted the increase in tissue myeloper-oxidase activity, a measure of leukocyte infil-tration. These observations suggest thatendocannabinoids protect the gut not only bydecreasing bowel motility but also by inhibit-ing the inflammatory process itself.

The nature of the anti-inflammatory effect ofendocannabinoids, however, remains to be elu-cidated. The exaggerated increase in myeloper-oxidase activity in the inflamed bowel of CB1knockout mice3 indicates that in wild-typemice endocannabinoids actively inhibit leuko-cyte infiltration caused by the chemical treat-ment. This may be due to inhibition of therelease of chemokines and proinflammatorycytokines, such as TNFα. TNFα has beenimplicated in the pathogenesis of IBD11

, itslikely source being activated macrophages andmast cells12. Cannabinoids suppress TNFαrelease from both cell types13,14 (although theCB1 antagonist SR141716 was also reported tosuppress TNFα release in a different model ofbowel injury15). In a mouse model of myocar-dial ischemia-reperfusion injury, cannabinoidsattenuated the increase in myeloperoxidaseactivity via CB2 receptors16. The anti-inflam-matory potential of CB2 receptors in the gutshould therefore be explored.

Infiltrating macrophages in the inflamedbowel may be not only the target but alsothe source of endocannabinoids. Bacterial

components, such as lipopolysaccharide(LPS), are known to powerfully induceanandamide synthesis in macrophagesthrough a pathway dependent on CD14 andNF-κB5. Anandamide may thus act as anautocrine-paracrine mediator to limit therelease of cytokines by macrophages andmast cells and the neuronal release of thetachykinins and acetylcholine that controlbowel motility (Fig. 1).

An important feature of IBD is the alteredtissue response to enteric bacteria. Mutationsin the gene CARD15 (also called NOD2),which result in deficient activation of NF-κBin response to LPS, are associated with sus-ceptibility to Crohn’s disease17. A defect inNF-κB signaling is also expected to result indeficient activation of anandamide produc-tion given that NF-κB is required for LPS-induced anandamide synthesis 5. Bowelinflammation can also increase FAAH activ-ity8, resulting in further reductions in anan-damide levels.

It is not known whether anandamide levelsare reduced and the activity of FAAHincreased in the intestines of patients withIBD, but this could be verified using biopsyspecimens. If confirmed, such changes maybe interpreted as the weakening of anendogenous protective mechanism, which

could be restored by preventing the break-down of anandamide with a FAAH inhibitor.Biopsy results from humans could furthersupport the case made by Massa et al. for tar-geting the endocannabinoid system to treatIBD, offering renewed hope to a much-suf-fering patient population.

1. Loftus, E.V. Gastroenterology 126, 1504–1517(2004).

2. Longobardi, T., Jacobs, P. & Bernstein, C.N. Am. J.Gastroenterol. 98, 1064–1072 (2003).

3. Massa, F. et al. J. Clin. Invest. 113, 1202–1209(2004).

4. Cravatt, B.F. et al. Nature 384, 83–87 (1996).5. Liu, J. et al. J. Biol. Chem. 278, 45034–45039

(2003).6. Kathuria, S. et al. Nat. Med. 9, 76–81 (2003).7. Weber, A. et al. J. Lipid Res. 45, 757–763 (2004).8. Hornby, P.J. & Prouty, S.M. Br. J. Pharmacol. 141,

1335–1345 (2004).9. Izzo, A.A. et al. Br. J. Pharmacol. 134, 563–570

(2001).10. Izzo, A.A. et al. Gastroenterology 125, 765–774

(2003).11. Palladino, M.A., Bahjat, F.R., Theodorakis, E.A. &

Moldawer, L.L. Nat. Rev. Drug Discov. 2, 736–746(2003).

12. Bischoff, S.C. et al. Gut 44, 643–652 (1999).13. Samson, M.-T. et al. J. Immunol. 170, 4953–4962

(2003).14. Fachinetti, F., Del Giudice, E., Furegato, S.,

Passarotto, M. & Leon, A. Glia 41, 161–168 (2003).15. Croci, T., Landi, M., Galzin, A.-M. & Marini, P. Br. J.

Pharmacol. 140, 115–122 (2003).16. Di Filippo, C., Rossi, F., Rossi, S. & D’Amico, M. J.

Leukocyte Biol. 75, 453–459 (2004).17. Hugot, J.-P. et al. Nature 411, 599–603 (2001).

NATURE MEDICINE VOLUME 10 | NUMBER 7 | JULY 2004 679

APCs size up antigensKeith S Bahjat & Stephen P Schoenberger

Intracellular antigens—from viruses, tumors, bacteria and othersources––pose a challenge to the immune response and to vaccinedevelopment. How do such antigens enter antigen-presenting cells? Threenew studies examine this process, implicating whole proteins instead ofsmall peptides as key intermediaries.

Keith S Bahjat and Stephen P Schoenberger are at

the La Jolla Institute for Allergy and Immunology,

Division of Cellular Immunology, 10355 Science

Center Drive, San Diego, California 92121, USA.

e-mail: sps@liai.org

Many of the pathogens that threatenhuman health do so from within the pro-tected environment of our own cells andtissues. An effective immune responseagainst such invaders often necessitates thedrastic step of killing the host cells thatharbor them. Cytotoxic CD8+ T lympho-cytes (CTLs) are especially suited to thistask, with an impressive range of methods

for inducing the death of target cells. CTLsrecognize target cells through major histo-compatibility (MHC) class I moleculesoccupied with 8–10-amino-acid peptidesderived from endogenous proteins.Initiation of CTL responses againstinfected target cells generally requires thatthe antigens presented at their surface mustfirst be acquired and presented to CTLs byhost antigen-presenting cells (APCs), suchas dendritic cells. This poorly understoodpathway of antigen trafficking and presen-tation is called ‘cross-priming’ and is essen-tial for the generation of CD8+ T-cellresponses against viruses, intracellular bac-teria, tumors and tissue antigens1,2.

©20

04 N

atur

e P

ublis

hing

Gro

up

http

://w

ww

.nat

ure.

com

/nat

urem

edic

ine