Regulation of prostanoid synthesis in microglial cells and effects of prostaglandin E2 on microglial functions

633,

isa

inghetti,

rancesca Aloisi

’ prostunoid lcvcis are normally low but can increase after ischcmia and during inflammatory and infectious diseases. can iGl!Ct brain function in SCVUill ways. In particular, prostaglandin E, (PGE,) might cxcrt both immunodepressivc and proinHammatory actions. The present short review focuses on the regulation of prostanoid synthesis in microgliol cultures and on the possible role of PGE, in the down-regulation of microglial activation induced by lipopolysaccharide (LPS). Our studies were carried out using puritied mouse or rat microglial cultures. LPS induced a dose-dependent expression of the inducible isoform of cjclooxygenase (COX-2). both in neonatal and adult microglial cultures. In the latter, the inducibility of COX-2 increased with time in culture. paralleling the acquisition of a more ‘activated’microglial phenotype. and appeared to account for the time-dependent increase in the PGE, /TXB, production ratio. The LPS-induced COX-2 expression and prostanoid production were down-regulated by potentially neurotoxic agents, such as nitric oxide (NO), the proinflammatory cytokine IFN-y (which acted both directly and indirectly, through its NO-inducing activity) and the HIV regulatory protein tat. On the other hand, COX-2 expression was up-regulated by the maurophage-dcactivz\ting cytokinc TGF$l. by exogenous PGE, itself, which acted through EP2 receptors linked to cyclic AMP generation, and by non steroidal anti-inflammatory drugs. Interestingly, PGE, utilized the same EP2 receptor-mediated signal transduction mechanism to down-regulate the expression of the inducible NO synthasc and the production of NO. Lurgcly, but not exclusively, through its effect on cyclic AMP, PGE, can also: i) dcprcss the expression of majo! histocompatibility complex class II antigens and of the costimulatory molecule B7-2; ii) down-rcgulatc TNF and up-regulate IL-IO microglinl production: iii) inhibit microglial IL- I2 secretion. Thcsc observations, together with literature data on in vivo models of central nervous system (CNS) diseases. suggest a ncuroprotcctive role of PGE, in pathological conditions. 0 So&%! franqaisc dc biochimic L‘Ibiologic mol&ulair~ / Elscvicr, Paris High

~~(~Sli~i~~)i~ Iwds

mierogliu / cyclooxyg~~~usc / cy tokincs / prustqlandin

/ ~~~~~n~~~tioli

Micmglinl cells cm bc considcrcd as resident brain macrophuges. They arc diffusely prcscnt in our brain to protect. it from infective, toxic, ischemic niadother l,yp~4of insults, but, may contrihu& to the cstablishmcnL ml maintenance of brain chimgu when the irlagnitudc of the microglial rcaclion surpasses a critical threshold lcvcl (see [ L-71 for reviews). The most known microglial functions (phagocytosis, antigen prcscntation, secretion of cytokines, growth factors, coagulation and complement factors, extracellular matrix, ncurotoxins, nitrogen and oxygen radicals, lipid mcdialors etc., cxprcssion of a set of macrophage markers) are displayed when the cells arc activated by approzi:Correspondence and reprints Abbrw’utims: AA, arachidonic acid: cAMR cyclic AMP; OX, cyclooxygcnase; dbcAMP, dibutyryl cyclic AMP; iNOS, inducible nitric oxide synthase; IFN, interferon; IL, intcrlcukin; LPS, lipopolysaccharidc; NSAIDs, non steroidal anti-inllammatory drugs; PG, prostaglandin; TX, thromboxane; TNF, tumor nccrosis factor.

There is some semanlic confusions in the ~it~rat~~~:c about what people mean by ‘activated’ microglin. In order to understand the in vitro experimental system that WCand many others have used, it may help to consider schematically the existence of three functional states of microglia (fighrveI): i) a ‘resting’ state, in which the cells arc apparcnlly inactive and most micro lial functions arc down-rcgulaled; ii) a ‘responsive’ slate (that some may wish to call activated, or differentiated, or dcdiffcrcntiatcd or maturation state), in which a set of intracellular and surface antigens are up-regulated: the cells are now ready to respond to a variety of stimuli, but functionally they are still 4qcly inactive; iii) an ‘effector’ state, that may have diticrent dcgrecs and may vary as a funcCon of the challenging stimulus. In this stale the cells actively respond to a variety of stimuli by floodin extraccllular milieu with numerous substances, such as

intuaeallularsntig4ine (e Q Ebl

vimsnta)

intracellular and surface antigens

identical. LPS is a cl in?~rl~u~i~(IL)-1 is e

eri

n vitro:

brain thw prostanoids wcire produced in comp;lrnhlo amounts 1IS]. Further oxpcrimcnts perFormed using adult micr@isl cultures provided some athcr interesting intiarmatian, due to the peculiarity sf these cultures. Indeed, during the fitxt 1-2 days in vitro, adult microgliul cells maintain ;arn~~h~~~~i~uland antigcnic phcnotypc similar ts that af ‘resting”, ramilied micro& in viva After another cnuple of duys, the c~11sspontunsously azquirc an umc id ?~~~~h~l~~~~ and repass intrzlcellulur and surface untigcns typical of activated (WCwould say ‘responsive’) rni~r~~~i~f321. At ncithcr stage in vitro did the cells express d~t~tu~l~ amounts of CQX-2. In these adult cells -2 was scarcely inducible by LPS in l-2-day culs it massively induced in 4-day cultures. in X-2 expression was paralleled by a dramatic increase in the production of PC%,, while that of TXB, remained essentially unchanged 1331.The massive

~sopr~~tere~oI[ I 5 1 Forskolin 1351 PGE, 1351 Adenosine [ 161 TGF-(3 [36] TNF*:” NSAIDs [ 351

“activated” (4.day cultures)

\\

TXA-2 -._ 7 surcide

’

Prostanoid formalion in LPS-trcalcd adult microglia in eulturi. In cultures of microglial cells obtained from adult rat brains 1331, COX-2 inducibility upon exposure to LPS appears to be related to the degree of activation occurring spontaneously with time in culture. The greater LPS-induced expression ot COX-2 observed in older cultures is accompanied by an increased capacity to synthesize PGE,, while the synthesis of another prostanoid, TXA2, remains unchanged. The increased output of PGE, may be due to the higher availability of tbc substrate PGH,. PGH, may take preferentially the route of PGE, synthase, which cxhit%ts a higher K,,, for the substrate than TX synthase. Other explanations are also possible (see text). Modified from 171.

output of the prostaglandin, not accompanied by an incrcascd rclcasc of TX&, may bc largely cxplaincd by ths incrcasc in COX-2 activity. In fact, in conditions of intcnsc enzymatic activity, the COX-2 product PGH, is cxpcctcd to siWriItc INOW easily TX synthasc rhun PGE synthasc, since the K,,, of PGE synthasc for the substrate INIS been reporlcd to 1)~’ higher than that of TX synthasc. 0th~ factors, how~vcr, may ~~~~tr~~~~t~~ lo the r~~~tlts obtuincd: for example, PGH2 can bc converted to non-cnaymatically, contributing to the accumulation of lhc prostaglandin. Moreover, a change with time in culture in the compartmcntntion of cnzymcs or substrates may take place. On the other hand, the suicide of TX synthasc dcsoribcd in other systems as a consequence of its catalytic activity has not been studied in tnicroglia [figlcrt~2). A massive increase in the production of PG& in 4-day, as compared to 2-day cultures, was also observed when zymosan, rather than LPS, was used as a microglia activating stimulus [ 321. The predominant synthesis of PGE* in activated adult microglial cells may have important conscqucnccs, in view of the function of this prostaglandin in inflammation and immunoregulation. -2 expression: 3. Substances regulating C studies on neonatal microglial cultures Using neonatal microglial cultures, the LPS-induced microglial COX-2 expression and prostanoid production

Nitric oxide [?3] IFN-y [34]* HIV-tat** IL-10 [36] Lipocortin I*** Glucocorticoids***

*Tested also in adult derived microglial cultures (Slepko, Minghetti and Levi, unpublished). *:‘fiPola/./l, Minghetti and Levi, in preparation. ***Minghctti, Nicolini, PolalLi, Grcco, Perretti. Parcnte and Levi,

submitted.

was shown to be regulated by ~1 variety of substances (ftrhk I) 1IS, 16, M-36]. It is worth noting that COX-2 was upregulated by potentially ncuroprotective substances, such as transforming growth factor-l% adcnosinc and PGE, itself. Interestingly, tumor necrosis

factor (TNF), whose role in neuroprotection or gcncration is still controversial, also upregulate Oddly enough, non-steroidal anti-inflarnmat (NSAIDs) upregulated the cxpressio concentrations abrogating prostanoid p in apparent contrast with the effect On the other hand, potentially

o

no intcrfcron-y (IFN-y), NO and the h cicncy virus lX!&jklli~tO~y protein trrl( Levi, in preparation) down-rcgulatcd COX-2 and prostanoid production, Hcrc WC: shall ~(~1~~~~9tr~~t~on lhc cffccts of NO and of CXO~CNOCS P(‘iT-:? and WC shall briefly disculgs the problem of NSAIDs,

The rqgulation of COX-2 by NO has a intcrest, since the rnetabalic palhweys leadia production of prostanoids and NO arc induced during microglial activation. Both rated NO s~~bst~l~~t~al~y produced and cxogenously t, scvcral NO inhibited the expression of -2 cxpr~ssion. donors almost abrogated LP Consistently, inhibitors of nitric o dc synthasc (NW3 up-regulated COX-2 [ 341. Intcrcsti y, while NO downrcgulatcd COX-2 cxprcssion, PG dow~9~rc~ulatcdthe expression of the inducible isoform of NOS (iNOS) I371+ which suggests the cxistcncc of an unusual control of the COX and NOS metabolic pathways whereby the cndproducts of each pathway, rather than competing with each other or neutralizing each other’s actions, act at the level of the parent enzymes, thus limiti gcncration. Several observations support that the stimulatory effect of exogenous induced CQX-2 expression and prostanoid production was mediated by cyclic AMP (CAMP), whose synthesis was

the activation of EP2 r~~~~~~rs~ stimulation of protein kinase A. have an opposite effect on the expre iNCS through the transduction syste The consequent elevation in the p depnzss neurotoxicity and promo E, biological effects on microglia Another aspect of concerns the regulation of the expression of B7 antigens (391. This bears to the problem of the control of the antigen presenting function of micro lia. It is known that there are several equipments for a cell to act us antigen

El formation, and

Critical for sustaining T ccl1 proliferation and cytokine secretion 1401whereas CTLA-4 has an inhibitory role in T lockade of the interaction between 28 on T cells bas been reported to p~vent or at~~eliorateautoimmune diseases 142). Microglial cells can be induced to express B7 molecules by appr~p~ate stimuli, and regulation of the expression of these molecules may modulate the ability of microglia to present antigens to T cells. With this b it is inter~stit~gthat PC%?, that was sh depress the expression of class II major histocompatibility was able to inhibit

the clRct sf PC%&an the production of IL-10 and TNF. IL-16 is an it~mun~m~ulat~ry cytokine, produced hy Th2 cells, B cells and macrophagie cells including microglia, and plays a major role in inhibiting inflammatory and Tl~l~n~ediatedimmune responses 1441.TNF is a proinllammatory cytnkine that. alone or in concert with other cytokincs, is believed to bc respe,nsible for cellular datnage or pt-oteetion in the CNS (see [4Sl for editorial on this issucb. In mouse n~i~t~~l~al cultutxs stimulattcd with LPS, Ez cnhanccd the expression and secretion of IL- 10 while it depressed that al’ TNE Also in this case, the effects of the p~staglandin were mimicked by CAMP elcvatiny agents (Aloisi 61 al., in preparation). The last set of experiments that WC would like to mention concerns the inhibitory etYcct of PC& on the expression and production of IL- 12. IL- 12 is a 75 kDa heterodimcric protein. composed of two chains (~35 and

9 s

. . Chamak B.. Brain macrophages: Neurotoxic or neurolrophic effector cells. J. Leukocyte Biol. 56 ( I9941) 416-422. Perry V.H.. Lawson L.J., Reid D.M., Biology of the mononuclear phagocyte system of the central nervous system and HIV infection. J. Leukocyte Biol. 56 (1994) 399-406.

and IL-12 ~40 mRNA expression induced by IFN-]!/LPS inhibited the production of the bioactive protein IL- 12 nnd of IL- 12 p40. several previous experiments. and db mimicked the effect of the our observations show the existence of a complex interplay between inflammatory substances (including cytokines, NO and prostaglandins) in the rcgulation of COX-2 activity and scvcral microglial functions related to inflammatory and immune reactions (7J. AS mentioned at the begin,ling of this article, microglia may predominantly play a neuroprotrctivc or a ncurotoxic role depending on the type of stimulus and on the degree ol microglial activation. In this complex interplay, PGE, is emerging as a crucial micro&a deactivating substance, and seems to mcdiatc, generally through CAMP, actions directed towards limitation of inflammatory and immune responses and, possibly, towards ncuroprotcction. The hypothesis, however, is still largely based on in vitro observations and requires confirmation in animal modclh of brain diseases. The dclcction of increased cerebral levels of PGE, during the remitting phase of cxpcrimmtd ,

allergic

cllccphalomyelit~~

dkmc

by ~1 !:tnhle

148

1 ;md the suppression 01 this

PGE2 nnalog 1311 WC’consistent with i\ tacaahcial roic) of PGI& in vivol Such nc3uroprokcl~v(:rol may hc rclatccl (nt Icast; in part) to the CAMP clcvatin activity of 0~~”prostq$ndin (39, 37 1. indcd, also other cAMP clcvating agents have been r~~~~~r~~d IS cxcr~. a protcclivc rnlc in cxpcrimcntal brain discasc {see 171 fol more cxtmsivc discussion). The controversial results obtained in viva with NSAIDs, which in some instances have been reported to exert a ncuroprot.ecGve cffccl 13 I, 491 may be related to cffccts of these drugs other than COX inhibition. For example, NSAIDs were reported to intcrfcrc with the activation of transcription factors reguMing the cxprcssion (among others) of NQS and pro- or anti-inllammalory cylokincs (see [ 71).

Acknowledgments Supported Supericm

IO/A/H)

by

’

Prqjcct

di Sanitk

of the Itdim

on

kmw,

Mulliplc

Soluro~is

and by Pro~ccl

Ministry

of Health.

on

of’ tlw

Istituk)

AIDS (granl no.

MC Geer P.L., McGeer E.G., The inflammatory rcsponsc system of brain: Implications for therapy of Alzheimer and other ncurodcgenerative diseases. Brain Rcs. Rev. 2 I ( 1995) 195-2 18. Kreutybcrg G.W.. Microglip* a sensor for pathological the CNS, Trends Ncukosci. I9 (1996) 312-318.

events in

Moore S., Thanos S., The concept of microglia in relation 10 ccntrul ncrwus sy~m discasc and rcgcncration Prog. Neurobiol. 48 ( 1996) 44 l-460. Minghcui L.. Levi G., Microglia as cl‘tcclor cells in brain damage and repair: Ibcus on prosI;moids and nitric oxide, Prog. Ncuroblol. 53 (1998) 99-125.

Shcrin J.H., Scamm~211 TX,, ‘EIickcy ,B,, Imrtsvcnuus ~i~~~ys~c~~~~d~ Iike immunoracctivity in rat brain meningeal macrophagrs, J. Camp.

amino acid receptor?;, and neurologiad dis-

F,, Dutr&

N,, Bonnc C..

N,, Skiami N,, W~b~~ubi