Regional and Cellular Codistribution of Interleukin lß and Nerve Growth Factor rnRNA in the Adult Rat Brain: Possible Relationship to the Regulation of Nerve Growth Factor Synthesis Christine E. Bandtlow: Michael Meyer, Dan Lindholm. Matthias Spranger, Rolf Heumann, and Hans Thoenen Max Planck Institute far Psychialry. Depanmem of Neurochemistry, 8033 Martinsried, Federnl Republic ofGennany

Abstract. We have found a regionaJ distribution of IL 1/3 mRNA aod IL I activilY in the nonnaJ adult rat brain, which reveals al least partially a colocalization with nerve growth faelor (NGF) . The predominantly neuronal signal patterns were found over the granule cells of the dentale gy rus, the pyramidal cells of the hippocampus, the granule cells of the cerebellum, the granu le aod periglomcrular cells of the olfactory bulb. aod over dispersed cells of the ventromedial hypothalamus aod of the frontal cortex, In these areas also the highest levels of IL I aetivity were observed. In the striatum and septum mueh lower levels of IL I ß mRNA and IL I aetivity (shown for the striatum) ,

TH E physiological funclion of nerve growth faetor (NGF)' for specific populations of neurons of the pe. ripheral and central nervous systems is weil docu­

mented. NGF is essential for the development and main­tenance of function of the peripheral sympathetic and neura1 erest-deri ved sensory neurons (see Levi-Montalcini and An­geietti , 1968; Thoenen and Barde, 1980; Green and Shooler, 1980) . In the central nervous system a similar function of NGF has recently been established for cholinergic neurons of the basal forebrain nuclei (see Thoenen CI al.. 1987; Whiuemore and Seiger. 1987). Whereas the function of NGF as aretrograde messenger transferring infonnation from the field of innervation to the NGF-responsive innervat· ing neurons was unambiguously established , very Iitt1e is known about the mechanisms involved in the regionally differential regulation of NGF synthesis. In the peripheral nervous system, we have recently demonstrated that lesion­ing ofthe seiatic nerve leads to a dramatic increase in NGF synthesis by nonneuronal cells (Heumann et al. , 1987 a,b) . It oould also be demonstrated thai at least the long-Iasting

Address correspondcncc IOC. E. Bandllow. Brain Research Insliu.ile . Uni· versilY of Zurich. CH·8029. Zurich. Swilzerland.

I. Abbm:iafiOl/ uSM 1'1 {his pap~r: NGF. ne rve growth fador.

most likely synthesized by gl ial cells, eould be deter· mined. IL Iß--expressing cells were mainly found in brain regions that also synthesize NGF mRNA as shown by in situ hybridization . NGF mRNA could be demonstrated over pyramidal eells of the hip· pocampus, granule cells of the dentate gyrus, periglomerular cells of the olfaetory bulb and over prefrontal cortex neurons. These data indicate that IL 1.6, among other faetors, might also play a regulatory role in the synthesis of NGF in the CNS, as has been demonstrated in the peripheral nervous system (Lind· holm , D., R. Heumann, M. Meyer, and H. Thoenen . 1987. Na/ure (Lond.). 330:658-659).

augmentation of NGF synthesis is mediated by immigrat· ing macrophages. Of the many cylokines synthesized and re­leased by macrophages IL I was shown to be the predomi­nantly responsible agent for the regulation of the reactive synthesis of NGF (lindholm et al., 1987, 1988). Interest­ingly, IL )· Iike activity has also been found in the nonnal rat brain (Nieto--Sampedro and Bennan, 1987), which increases after injury. However, the site of synthesis and the physiolog· ica1 functions are unlcnown . More recently, we have demon· strated that IL-l induces tbe syntbesis of NGF in astrocytes in vitra and increases also the NGF mRNA levels in hip­pocampus after intraventricular injections (Spranger et al. , 1990). These results suggest that also in the CNS Il-I has the potential to regulate NGF syntbesis. However, from these experiments it could not be deduced whether Il I is also involved in the physioJogical regulation of NGF syn­thesis.

We now demonstrate that IL 1,B mRNA and IL 1 activiry are partially oolocalized with NGF mRNA in certain brain regions of the normal adult rat. In situ hybridization experi. ments show a predominantly neurona1 expression ofboth the IL Iß mRNA and the NGF mRNA , which indicates that IL Iß might indeed play a functional role in the regulation of NGF synthesis in the CNS.

Materials and Methods

Cell Culture

Primary Bmin CU/fU rtS. Cerebri of newborn rats were processc:d by Ihe method of McCanhy and DeVellis (1980). CeHs were steded in 7j-cm FaI­roll flash and culti\1lted in DME medium (Gibco Laboralorics. Paisley. Srotland) with. 10% FeS (Gibco Laboratories) for 2 wk 31 10% e(h. lhe medium was change<! every fourth day. The cell monolayer was laken off withO.l% trypsinl5 mM EDTA (SigmaChemica! Co .• SI. Louis. MO). Iril­uraled, and replated on polY-L·lysine (Sigma Chemieal Co,) coated rou r­we il Greiller dishes with. 10,000 cells/welL 2 d befOTe!,lst fOT in SilU hybrid­ization the cells weTe switched 10 serumless DME medium (Spranger CI al .. 1990).

CerrMIlOT Granule Ce/Is. Granule cdls were prepared from cerebella of7-d..:old TalS by trypsi n treatment (0.1% IVQlIvoll. 15 min: Balns CI al. , 1988; Hallen CI al. , 1988), The teils werecultured fOT 3 d on polY'L-lysine­coated diskes (5 ,..glml) in a rnediurn consisting of one part ofEagle's ba!;a.l rnediurn with Earle's salls (Gibco Laboratories). Orte part of Ham's F12

rnediurn with 33 rnM glucose and 2 rnM glu tarnine and tlle following sup­plements: 25 ,..glrnl insulin, 100 ,..glrnl transferrin. 20 nM progesterone. 100,..M putn:scine. 30 nM seleniurn (Bonenstein and Sato, 1979). Cytosine arabirooside (10 ,..M) was mutinely added 10 tlle cu ltun:s to inhibil Ihe growth of oonneumnal cells. Granule cells wen: identified by thei r §Il!all

sizc aod by the extensive oUlgrowth ofneurites during the first days of cu!­ture (Bal3.Z.'l et a1., 1988; Hatten CI al.. 1988).

AcliWJltd MfKrophagu. Aclivaled rat pcriloneal macmphagcs werc ob­laincd as dcscribcd by Lindl>olm el al. (1988) and Heumann CI al. (1987b) .

Raising 0/ Polyclonal Rat TL Iß Antibodies Thc NHl-tenninal peptide of malure rat IL Iß (amino acids 117- 143 ac­cording 10 Nishida et al. [19881 with an additional lyrosine at position 144) was synthesj~ed on a peptide synthesizer (model431A: Applied Biosystems. Fester City. CA) using Pmoc cherniSlry. New Zealand While Rabbits we rt imrnunizcd with native. uncoupled peptide by standard pmtocols. In ELiSA lests using Ihe peptide adsorbed to plastic as a solid phase (Flanders el al.. 1988) the scrum had a liter of 10 .... after Ihe th ird booster injection.

Figure J. Schematic presenlatioßs of Ihe localization of IL-Iß mRNA (0) and NGF mRNA (b) in Ihe adult rat brain. The ~ilCs of synthesis are dOlled and rep­resent Ihe hippocampus (CAl- CA3). cerebellar granule cells (GrA ), granule und periglomerular cclls ofl he olfactory bulb. neurons in Ihe frontal conex. and neurons of Ihe venlromedial hypothala­mus (VMH ). Not shown are hybridiza­tions of IL Iß mRNA in the striatum und septum. Abbreviat ions and d rawings af­ter Pellegrino er a1. (1979)_

Fjgure 2. Results from in situ hybridization of rat brain sagiual scctions 10 IL 1/40 riboprobes. Positive IL Iß hybridizalions are shown for the hippocampus (a) , the granule cells of Ihe cerebellum (c). pcriglomerular cells of the olfactory bulb (e) with the corresponding phase contrast photomicrograph (f), neurons of the hypothalamus (g). and neurons of the frontal cone1l (h). No specific labeling is seen on sections hybridized with the sensc IL il40~ probe (b and d) . Positive labeling is also shown over glial cells in the septum U) and stria­IUm (I) wilh the corresponding phase contrast pictures (j and k). Background labeling is shown of a scction through the striatum hybridized with the sense IL 1/40- riboprobe (rn). All sections wcre counterstained with cresyl-violet. Exposure time was 5.5 wk at 4"C. Bars. (a-d) I ßlßl; (e-rn) 50 ~.

.' ..

'4' • c

•• • .' •

.. ,

.'

.;

, . , -.

'~.'

Figure 3. Higher magnifications of the positive IL lß mRNA hy· bridization signals found over granulecells ofthe dentale gyrus (a). the cerebellum (h), and the periglomerular cells of the olfactory bulb (c), indicating a differential expression of the IL Iß mRNA. Note, that the Purkinje cells (arrows) in the cerebellum are not la­beled (h). Exposure time was 4 wk at 4°C. Bar, 100 ~m .

Costimulator Assay Jor IL I Activity After transcardiat perfusion of adult Wistar rats (180-200 g) with Dul­becco's modified Eagle's medium (Sigma Clu:mieal Co.). brains were rap­idly disse<:ted. e om:sponding brain regionsofeaeh animal were pooled and store<! at - 70"C. SampIes wen: homogenized in 500 ~ I/]()(} mg wet weight ofbuffi:r H (SO mM Naet. ]() mM Hepes. 1 mM EDTA 20 ~M l-mtrcaptD­ethanol, 0.1 mM PMSF, 2 "glml Pepstatin A. pH 7.0) by 10 strokes in a glass Teflon homogenizer. After n:moval of membranes by sequential low­speed (12.000 g, 5 min. 4°C) aAd high-speed centrifugation (TLA 100.3 rotor; SO.ooo rpm, 15 min, 4"C, Beckman Instruments. Fullenon, CA). sampIes wen: analyzcd br FPLC using a Superose 12 column (HR 10130, Phannacia Fine Clu:micals, Piseataway. NJ). Elution was performed with 150 mM NaCI, 50 mM Na2HPO., pH 7.4, collecling 500 ..,1 fraclions. Myoglobin and lysozyme (Sigma Chemical Co. ) wen: used as molccular weight standards. IL I activity..as asscssed for each fraClion al a one-Ienlh dilution as described by LeMoal et al. (1988) using thc EU cell clone as indicator of IL 2 reccptor expn:ssion . Thc assay was modified 10 a Ihn:e­stcp procedun: applying biOlinylated second antibodies and stn:ptavidin ­coupled perollidase (Amersham Chemieals), The sensilivily limit was "'10-6 U per assay ofhuman n:combinant IL 113 (Biogen, Geneva. Swilzer­land) used as an interna! standard. As an addilional positive conlrol dilu­tions of condilioned medium from activated rat macrophages were assayed. Neither NGF (up 10 50 nglml) nor IL-6 (up to 200 Ufml) were aclive in this assay. Analysis by the described method or I U of recombinant human IL 113 in 5% FCS resulled in 70-85% recovery of act ivity. To exclude specific diffi:rences in inhibition or degradation or endogenous IL I activily. sampIes or brain regions were homogenized after addition or rttombinam IL 113 in lQ-fold execss ovcr the average of endogenous IL I activity. In all castS Ihe recovery was 70%

All experiments for imrnuooad50rption were done wilh an IgG fraclion (called B-lgG) pn:pan:d br protcin A-Sepharose chromatography from the bleeding after thc third boosl. The immunoad50rption columns conlained 50;.1 of packe<! pn:swollen prolein A-Sepharose. lOO..,g of B-lgG or, for control columns, the same amount of commerciaJ rabbit fgG (Miles Labora­tories, Naperville. IN) in 600 "I of 150 mM Tris, pH 8.8 were passed twicc over Ihe eolumn. For immunoad50rplion the column was washed with 1 ml of culture mfflium used for EL-4 cells (RPMI 1640. 5% FeS, 20 pM 2-mercaptoethanol. 12 mM Hepes: R-medium). so "I of the moSI active (raction from Ihc preceding Superose 12 chromatography step wen: pas.scd over the immunoadsorption or the control column. Thereafter the column was washed Iwice with 50 "I of R-medium and the flow through as weil as the washes wen: pooled and applied on a second column of Ihe same Iype. Thc flowthrough and IIu: washes of this colllmn wen: pooled and lested 3t a one-fifth dilution in the EL-4 costimulator assay.

In Situ Hybridi1.otion Adull Wislar rats (180--200 g) wen: ether anaethesized and transcardially pcrfused with prewannc:d 4% paraformaldehydclPBS (Fluka AG; Buchs, Switzcrland) for 30 min. Brains wen: removed. immersed in the fixative for 2 h, befon: being incubated in 15 % slicrosefPBS overnighl at 4°C. Brains were then frozcn alKl IO-"m sagittal s.ections wen: collccted on (50 "glml; Sigma Chemieal Co. ) cooled polY-l· lysinc glass slides. Cultun:s \\-'eR:

tn:ated wilh the same fixative for 30 min beforc Ihorough washing with PBS. Slides and cultUfeS wen: funher processed as described by Bandllow CI al . (l987). including an acetylalion Siep beron: prehybridization (Hayashi el al. , 1978).

Conditionsfor lL lß-speeific ProlHs. To get IL lj3-spccific probes of a defined length and seque nce a synthctic oligonucleotide of 40 bases (base 777- 816 ofthe mouse IL 1 fJ cDNA; Gray el al. [1986]) was subcloned into the Bluescript vcctor (Promega BiolCCh, Madison, WI). Thc sequcnce for theoligonucleotide was se lected by the following erileria: (a) low homology 10 mouse IL lar to ensure specific hybridization 10 [L IfJ; (b) high homology to human IL-Iß to incn:asc the probability Ihat mouse and rallL 113 is de­tccted by the probe: and (c) low homology to 28S ribosomal RNA to n:duce nonspecitic binding. In vitro l'S_I&beled sense (IL:- I/y40- ) and antisense (lL-I/Y40+) transcripts of linearized plasmids wen: made as n:commended by the manufacturer (Promega BiOlec). In comparison, we also used )'S­labeled single-strand DNA probes obtaincd br revcrse transcription of non­labeled sense alKl antisense transcripts. These probes resulted in a similar specific activity of g x 10' cpm/..,g and were used ulKler the same condi­tions as for cRNA probes, with the exception thatthc RNasc A digestion step was omitted . Single ~tranded RNA or DNA probes wen: di luted 10 30,000 cpmJ,,1 in the fol1owing hybridizalion buffer: 50% fo rmamide,

Figure 4. LocaliUltion of IL 1,8 mRNA in primary brain cultures of newborn rats after 2 d in vilro under low serum condilions (0.5% FeS). Dark field (a), the corresponding bright field (b) , and phase-conlrasl (c) photomicrographs showing a se­lective labeling of ""5 % of Ihe cells.

28 s-

18 S ­IL-Iß -

28 S

18 S IL-Iß - _ _

Figur/' 5. Nonhem blot of10lal RNA from purificd cerebellar gran­ule cells and of RNA from activated rat macrophages. hybridiz.ed with Ihe ll P_labeled IL Iß/45Q riboprobc. Cclls had bcen in cul ­ture fOT 4 d: RNA was clltracted and processcd as dcscribcd in Materials and Methods. Positions of the ribosomal 18 and 28S RNA arc indicated. The filters wcrc RNase-lreated and the ex­posure time for neuronal RNA (feft) was four limes longer than that ofactivated macrophages (righr). Roughly 5 ~g total neuronal RNA was loaded in all lanes. whereas total macrophage RNA was ap­plied 10 a lwo-time dilUlioß series of 5 J,tg.

0.6 M Naet, \0 mM Tris, pH 7D, 0.05% yeasl!otal RNA (Sigma Chcmical Co.), 2x Denhardfs. 0.05 % inorganic sodium pyrophosphate. \0 mM me­thionirn:. 10 JlM nonlabeled thio-UTP (New England Nudear, Dreieich, FRG), 20 mM ß-mercapwethanol. Prehybridizalion (2 h) and hybridization (12 h) were carried out 31 45·C. Thereafter. cultures and slides were rinse<! several limes with 2x SSC and washed three times for 30 min in 0.2x SSC at 4O' C. Cells and sections were then treated with RNase A (15pglmJ in 0.5 M NaCI. 10mM Tris. pH 8. 1 mM EDTA) at 3JOC for JO min, fol1owed by a wash in RNase buffer for 15 min. before thcy were finally washed in O.lx SSC for 15 min al 55·C. Cuhure dishe.s and slides wen: then deo hydrated, air dried. dippe<! in Kodak NTB 3 emulsion and nposed at 4 "C. Autorad iographs wen: oounterstained wilh cn:syl violel after developmenl.

Com1itiorufor NGF-specific pralles. A fragment of J 18 bp from the cod­ing reg ion of Ihe mouse NGF cDNA (amino acid 188- 227: Seon et ai.. 1983) was subcloned lnto the Blue.scripl vt:C1or (Promega Bioiec). Nonla­beled sense and antisense transcripts were reverse transcribed in the pre~­eoee of llS-labeled dGTP. Hybridizatio n was caTried out in the above de­scribed buffer at SO"C. Washing step-s wen: pcrformed thrcc limes in O.2x SSC at SOoC fOT 30 min and linally in O.lx SSC, 25% formamide al 50·C. Slides were funher proccssed as dcscribed above.

Northern Blot Analysis

Total cellular RNA from c ultured cerebellar granule cells was i$Clated and fuMher proccssed for Northern biO! analysis as described by Lindholm et aJ . (1988). Bridly,RNA was glyoxylated and elCClrophorcsed through a 1.2 % agarose gel. After lransfer of RNA tO nylon membranes (Hybond: Amer­sham Chemicals) hybridiu.lion was caTried out at 6O'C in the presence of 50% formamide as described previously (Heu mann et al.. 1987a). The IL I clone used for Nonhem blots was prepared by the polymerase chain reae · tion (PCR) using oligonudeotides oonstructed against I"'" regions of the mouse IL-Iß cDNA (amino acids 53- 58 and 2(17- 213 for the sense and anti ­sense oligonudeotides, respcctivt:ly) (Young and Sylvester, 1989). The c DNA used was obtained by reverse transcription of RNA from activated ral macrophages. The rallL Iß clone (450 bp) was further 5ubcloncd into the bluescript vector and sequenced and lZP_labeled cRNA fun-ofl" tran ­scdpu were used at a ooncentration of 2 x 1CJ6 cpm/ml for Nonhern bIO! hybridiutio n. After hybridiu.tion the filters were washed at 55°C in 2x SSC for 30 min an<! exposed tO x-ray films. The sp«ificity of the rat IL Iß probe was shown by RNase A treatment of the Nonhem blots and by comparison wilh RNA from activated macrophages.

Results

In Situ Hybridization Ta determine the cellular localization of the specific expres­sion ofthe IL l{j and NGF mRNA, saginal sections of adult rat brains were hybridized with llS-labeled antisense DNA-

or cRNA - probes (IL lfi/40"'; NGFIl IS+). Control sections were hybridized under the same condilions with llS-labeled sense DNA- or cRNA- probes OL Ity40-; NGF/118-). The autoradiographic results are summarized in Fig. I, by sehe­matic presentations ofthe siles ofll l{j and NGF expression.

Localization oilL 1(3 mRNA

Cell bodies hybridizing 10 the Il Ity40+ probe were found in the hippocampal pyramidal celllayer, the granule cells of the dentate gyrus (Fig. 2, a and b), Ihe granule cells of the cerebellum (Fig. 2, c and d), the granuleand periglomerular cells of the olfactory bulb (Fig. 2, e and j), disperse cells of the anterior hypothalamus (Fig. 2 g), and of the frontal cortex (Fig. 2 h). A higher magnification ofthe labeling pat­lern in the dentale gyrus suggests a differential expression of the IL 1{3 mRNA since not all cell bodies ofthe granule cells are labeled (Fig. 3 a) . The same is even more evident for the periglomerular cells of the olfactory bulb, where the signal is present only over a few cells (Fig. 3 cl . In contrast, in the cerebellum, where exclusively granule cells were labeled. the grain distribution was intense and equal (Fig. 3 b). A consistent but much weaker labeling was delected over the striatum and septum (Fig. 2, i-rn). However, in these brain areas the grains were homogenously distributed in between the neurons which are spared, suggesting that most probably glial cells are responsible far the signal. In eomparison, no labeling occurred in brain regions rieh in oligodendrocytes such as the corpus callosum (data not shown), indicating that oligodendrocytes do not contribute 10 the specific IL l{j sig­nal in situ. These results areconsistent with in situ hybridiza­tion experiments done on unstimulated primary cultures of newborn rat brains. In these cultures, consisting of oligoden­drocytes, astrocytes, fibroblasts and microglia cells, only a very small number of cells ( ...... 5%) express IL l{j under se­rum free conditions (Fig. 4). Although il was not possible to identify the labeled cells unequivocally, they could weil present amoeboid microglia eells since the number of la­beled cells correlated with the number of microglia cells found under these culture conditions by immunocytochemi­cal slaining (Spranger et al. , 1990). The neuronal identity of IL 1(3-containing cells was further assessed by Northern blot analysis of cultured cerebellar granule cells from 7-d-old rats. These purified eells produce a typical 1.6-1.7 kb IL 1(3 transcript (Fig. 5) as described for macrophages (Giulian et al., 1986). However, as demonstrated in Fig. 5, the cerebel­lar granule cells contain a much lower copy number of IL 1(3 as compared with activated macrophages. The presenee of IL l{j in cultured cerebellar granule eells was additionally confirmed by in situ hybridizslion as shown in Fig. 6, and reft.ects the sensitivity of the detection system.

Localization 01 NGF mRNA

Similar labeling patterns as described for Il 1(3 were found fo r the distribution of NGF mRNA. As has been shown by others (Rennert and Heinrich, 1986; Ayer-Le Lievre el al. , 1988; Whittemore et al. , 1985) labeled cell bodies were found in the pyramidal cell layer of the hippocampus and in the granule celllayer of the dentale gyrus (Fig. 7 a) . Interest­ingly, in this particular region the labeling signal found for NGF strikingly resembles the pattern seen for IL l{j (see Fig. 2 a), suggesting thai both messages might be synthe-

on January 5, 2009 jcb.rupress.orgDownloaded from

on January 5, 2009 jcb.rupress.orgDownloaded from

0.0 0.16

0.' • L .. , 0 • 0.' - 0 .06 ,

~ '" " .. 0 ., 0.'

0

t 0 ., • 0.0 0.00

0 " " '" .. '" Froction (O.S ml)

Figure 8. Column~fraClionation of IL I activity. Frontal cortex ho­mogenate (of 40 mg wet wO has been applied on a Superose 12 column. Fractions thaI eluted between 14 and 17 kD have been measured (see Materials and Methods) for their IL I ac\ivity. Thc absorbance at 280 nrn was continuously monitored (dQ/led Une). All the IL 1 aClivity eluted as a single peak al 16 kD (straight line). Molccular weighl markers: myoglobin (M; 17 !cD) anti lysozyme (L; 14.7 kD).

sized within the same cells. Labeled neurons were also ob­served in the cerebral cortex, restricted over neurons in the frontal cortex, however, complete identification cf these cells was not possible (Fig. 7 c). In addition we could now detect cell bo<lies hybridizing with the NGF/1IS· DNA probe over the periglomerular cells of the olfactory bulb (Fig. 7 b), whereas no signal above background v.as seen in sections hybridized with the control probe NGF/118- . (Fig. 7 d). As seen for the IL I mRNA, the specific labeling over periglomerular cells varied from strongly positive neurons to negative ones, presumably reftecting differential NGF synthesis by individual neurons.

IL 1 Activity

The IL I assay we used was a costimulator assay measuring IL 2 receptor expression induced on EU cells by IL land a suboptimal dose of a phorbolester (LeMoal et al., 1988). Preliminary experiments showed that rat blood contains a phorbol ester independent activity (constituent of 30-40 \cD), which drastically interfered with the bioassay used. We therefore decided to perfuse the animals and to fractionate the tissue sampies by FPLC to separate other IL 2 receptor inducing activities as described by Tagaya et al. (1988). The peak activily found in various brain regions eluted slightly later than myoglobin (l7 kD) and before lysozyme (14 kD) al a position corresponding to "'16 \cD (Fig. 8). Forcompari­son >80% of the lL 1 activity of the conditioned medium from activated rat macrophage cultures eluted at the same position (data not shown) . Since macrophages release a vari­ety of factors that could possibly interfere with our assay sys­lern we take this finding as a proof for thc specificity of the method . Determination of the IL I activity in various brain regions after fraclionalion closely resembled the disuibulion pattern found by in silu hybridizalion. Although Ihe values of corresponding tissue sampies varied between different animals, Ihey reftect a regionally consistent distribution of the IL I activity in all examined rat brains. The highest activ­ity levels were found in the olfactory bulb, hippocampus, cerebellum, and the fronlai cortex, reaching maximal levels

of 10-3 U/mg wet weight (Table I). In comparison. IL t ac­tivity found in conditioned medium from activated rat mac­rophages (per \()6 cells) was "'50()....1.OOO U. A 1,OOO-fold lower, but consistent level was observed in the striatum (Ta­ble I), where in situ hybridizations have indicated that proba­bly glial cel1s are the IL Iß-specific source. The specificity of these values was further proven by the fact that rat spccific polyclonal antibodies raised against IL I could adsorb up to 93% ofthe IL 1 activity found in frontal cortex or cerebel­lum (Table I) .

Discussion The only information on a possible involvemenl of JL 1(3 in the regulation of NGF synthesis in the CNS is based on the augmentation of NGF mRNA in the hippocampus after in­Iraventricular injections of IL 1ß (Spranger el al., 1990). However, these experiments did not provide evidence whether IL Iß alsoplays a role in the regulation ofNGF syn­thesis under physiological conditions. We have now iden­lified the cdl types and areas in the adult rat brain that syn­thesize IL Iß mRNA as weli as NGF mRNA, and found a regional colocalization of both mRNAs. NGF was shown 10 be regional1y dislributed in the mammalian brain, with the highest levels in areas innervated by magnocellular neurons present in the basal forebrain (Korsching et al. , 1985; Shel­ton and Reichardl, 1986; Whittemore et al. , 1986). Due to the limiting amounts of NGF mRNA a precise cellular local­ization has been successfully demonstrated only over the py­ramidal cel1 layer of the hippocampus, granule cells of the dentate gyrus and over neurons found in the neocortex (Ren­nert and Heinrich , 1986; Ayer-Le Lievre et al., 1988; Whiltemore et al. , 1988). These results are consistent with RNA blot analysis showing that hippocampus and neoconex have the highest NGF mRNA levels (see Korsching, 1987). The olfactory bulb is another area known to express NGF mRNA in relatively high quantities (see Korsching, 1987). In this paper we have shown by in situ hybridization that the periglomerular cells are the specific source of NGF mRNA in this region. Interestingly, periglomerular cells are inter­neurons Iying in the glomerular layer which reeeives cholin­ergic input (Halasz and Sheperd. 1983), whichdemonstrates again that the target cells synthesize NGF. Of panicular in-

Tubte I. iL-i Activity in Different Brain Regions ofThree individual Animals

IL L activily

Region Animal Animal Animal Aven,g~ SD

10-' V/mg wer wr

Olfactory 1.25 I.S 0.1 0 .95 0.75 Frontal cortex' 0.25 0.5 1.25 0.67 0.52 Hippocampus 1.0 0.5 0.5 0.67 0.29 Cerebellum' 0.5 0.13 0.5 0.37 0.21 Siriatum 0.0 0.0025 0 .0005 0 .0025 0.0025

• Q".. frontal rortex and Olle cerebellum wen: analyzed by lhe immunoadsorb· tion protocol givcn in M~lerials and Methods. Comparcd with I"" conlrol coluJTlII$ 93'1 (cerebellum) and 90% (frontal conex) of Ihe IL I aCli~i'Y were adwrbed. Tis$ucS have beeil fr~ctionated by gel fillfllion and IL-l aclivity was analyzed as described in Materials and Mdhods.

terest was the observation that the regional distribution found for NGF coincides with the determination of IL I activity. In addition , in situ hybridization demonstrated that IL 1(3 is predominantly expressed by neurons, which were shown to be positive for NGF. However, IL Iß was also present in other brain regions such as the hypothalamus, striatum, and septum, where the levels of NGF mRNA are at best at the limit of detectability (Whittemore et al., 1986), suggesting that IL-I(3 must have a different or additional funclion in these tissues (Breder et al. , 1988). It is of interest that ex­cept for the hypothalamus glial cells are most likely respon­sible for the expression of IL 1ß mRNA in these brain tis­sues. However, it was not possibJe to unequivocally identify whether astrocytes andlor microglial cells contain the IL 1 ß mRNA. 80th cell types have been shown to synthesize IL-I in vitro (Fontana et al. , 1982; Giulian et al., 1986) but !here is no evidence for a production under physiological condi­lions. Although we could demonstrate a partial colocaliza­lion of IL lß and NGF, suggesting that IL 1(3 might be in­volved in the regulation of NGF synthesis also in !he CNS, the physiological relevance of!he observed results are ham­pered by the fact that hydropathicity profiJesofIL 1(3 precur­sor apparently lack any regions that have sufficient hydropho­bicity and length to qualify as a signal peptide (Lomedico el al. , 1984). Whether IL J can be secreted despite these apparent shortcomings remains an open question. Recenl reports have shown that only monocytes secrete IL I in both the precursor and the mature form (Hazuda et a1., 1988), whereas lransfecled fibroblast celliines were not able to re­lease any of the two forms , despite a high level expression oflhe IL Iß precursor within the cells (Voung etaJ., 1988). 11 is therefore Iikely that the used fibroblast cell lines lack some ceJlular factor(s) that allow the secretion and release of the intracellular IL 1(3. To evaluate the function of IL 1 in the regulation ofNGF synthesis under physiological con­ditions in the CNS it remains 10 be established whether neu­rons use arelease mechanism similar to Ihe one described for rnonocytes and how neuronal IL 1 synthesis and release is regu1ated .

We wisli to tliank M. Scllörnig for perfusion oflhe ralS, Drs. A. Lawcllky

and T. Hünig for the EL4 ccil line , A. Hohn for lhe plasmid NGF/I18, B. Hengerer for his hclp in cloning the IL 1,8/115, W. Risau for syntlicsis

o f the peptide . and Dr. P . Carroll for improving the English.

Received forpublication 12 Junuary 1990 und in ~vised form 9 May 1990.

R~futnces

Aycr·Lc Utvre, C .• L. Olsen, T. Ebtndal, A. Seiger. alld H. !>emon. 1988. Exp~ss ion of the ß·nerve growlh faclor 8ene in the hippocampal neurons, Scitnu (W/lsh. DC). 240:1339- 1341.

Balazs. R" O. S. Jorgensen, and N. Hack. 1988. N-Methyl·o-aspartale pm­moles the survival of cc~bel1ar granule ~ells in cultu",. Ncurosrit nu, 27:437-451.

B.andtlow. C. E., R. Htumann, M. E. Schwab. ~nd H. Tl>oenen. 1987. Cellular localization of nerve growlh factor synthesis by in si tu hybridization. EMBO (Eur. M/ll. Bio!. Org(JII.) J. 6:891-899.

Bottenste in, J. E., and G. M. S~to. 1979. Growlh of a neuroblastom.a cel! line in serum.f",e $IIpplemented medium. Proc. Nml. AauJ. Sei. USA. 76:~ 14-~17 .

Brcder, C. 0 ., C . A. Oina",1I0, and C. B. Saper . 1988 . Interleukin-I im­muoo",sctive innervation oflhe human liypoth • .Iamus. Scienct ( W<lJ'h . DC). 240:321-323.

Flanders. K. C" A. B. Roberts. N. Ung, B. E. FleurdeJys, and M. B. Sporn. 1988. Antibodies to peptide determlnanlS in lransfonning growth faclor beUl and thcir applieations. Biocht mislry. 27:739-746.

Fontana, A" F. Kristensen, R. Dubs, O. Gem!a. and E. Weber. 1982. Produc· tion of pmstaglandin E md interleukin I-like facIDrs by cultured Ulrocytes

and C6 glioma ccll$. J. JrnmulfOl. 129:2413-24]9, Giulian. 0" T . J . Baker, allCl L. B. La~hman . ]986. Interleukin·l o f lheccnlraJ

nervous system is produced by amcboid microg]ia . J . EJ;p. Mtd, 164:$94- 604.

Gray, P. W. , O. GlaiSler. E. Chen, V. O. Goedde] , an.:! D. Pennica. ]986. Two intcrleukin I genes in the rnoose: cloning an.:! cxp",55ion of Ihe cDNA for murine inlerleu k.in lß. 1. Im_lWl. 131:3644-3648.

G",cne , L. A .. and E. M. Shootcr. ]980. The nerve growlh faclor: Biochemis­try. synthcsis, and mechanism of acton. Amlu. R~v. Nt .. rosci. 3:353-402 .

Haläsz. N .. and G. M. Sheperd. 1983. Neuroehemislry ofthe vertebrate olfac· tory bulb. Ntllroscimu. 10:579-619.

Hauen. M. E., M. Lyneh , R . E. Rytlcl, D. MosclteJli, and O. B. Rif1cin. 1988. ]n vitro neurite extension by granule neurons is dependcm upon aSlrogliaJ ­derivcd fibroblul growth faclo r . [H". Bio!. 125:280-289.

Hay.ashi. 5., I. C. Gillam, A. D. Oclaney. and G. M. Tener . t978. Acety lation of chromorome squashes of drosophila mcJallOgaster decrcascs Ihc back· ground in autoT3diographs from hybridization with '''I-]abclcd RNA. J. HiJ1och~m. Cyfoclr~",. 26:677-679.

Hullda. D. J .• J . C . Lee. and P. R. Young. 1988. Thc kinetics of inlCrleukin 1 5eC",lion from activated monocytes. J . 810/. Chtm. 17;8473-8479.

Hcumann, R. , S. Korsching. C. E. Randtlow, and H. Thocnen. 1987a. Changes of nerve growth fac:tor synthesi$ in nonneurona! eells in ",sponse to sciallc nerve transcction. J. Ct U Biol. 104:]623- 1631.

Heumann. R., O. Lindholm, C. E. Bandllow. M. Meyer, M. J . Radeke, T. P. Misko, E. Shooter, and H. Thoenen. ]987b. Oiffe",ntial ",gulation of mRNA encoding nerve growth factor and its receptor in rat so;:iatie nerve dur­ingdevclopmc:nl, degeneralionand ",generation: roJe ofmacrophages. Proc. Nil/I. Acad. Sci. USA. 84:8735 - 8739 .

Korsching, S. ]986. The role of nerve growth factor in the CNS. TrmJs Ntu· rosei. 9:570- 573.

Korso;:hing. S .. G. Auburger. R. Heumann. J. ScOIl. and H. Thocnen. 1985 . Levels of nerve growth factor and its mRNA in tlie ccntral ncrvous !iystem oflhe rat com:late with cholinergic innervation. EMBO (Eur. Mol. BIo!. 0,.. g/l"') J . 4: 1389-1393.

LcMoal. M. A. , M. Stoock, J. M. Cavaillon, H. Robson McDona]d, an.:! P. Tryffa-BilChi. 1988. A H:nJitive. IL-2-dcpcndent. a$$ay for IL·1. J. 1m­mUIWI. Mt lhodJ. 107:23-30.

Lcvi-Montalcini. R., and P . U. Angele!! i. 1968. Nerve growth factor. Plrysio/. Rtv. 48:$34-569.

Lindholm. D., R. Heumann, M. Meyer, and H . Thocnen. 1987. lnter]euk.in-l regu latC$ synthesis ofnerve grOwlh fsetm in TIOn-neuronal cells ofrat seiatie nerve. Nil/ure (lmuJ.J. 330:6~8-6~9.

Lindholm, 0., R. Heumann. B. Hengerer, and H . Thocnen. 1988. Interleukin ] increases stability and lranscriplion of mRNA encodinll nerve growlh fac· IDr in cultured ral fibroblasu. J . Biol. Clrem. 263:]63411- 16351.

Lomcdico, P. T . , U. Gubler , C . P. He!lmann, M. Oukovicli. J. G. Giri, Y.·C. E. Pan. K. Collier. R. ~mionow. A. O. Chua. and S. Miul. 1984. Cloning arid exp",ssion cf murine interleukin-l cDNA in Es~richi/l coIi. Na/ure (LonJ.). 3]2:458-462.

McCarthy , K. 0 ., and 1. DcVelli $. 1980. Preparation of sepanue IStrogliaf and oligodendroglial ce!! cullures from rat brain li5$lIe. J. Ctll8iol. 85:890-902.

Ni~to-Samp"dro. M., an<! M . A. Berman. 1987.lnterleukin·l·like ac1ivity in Tat brain: SOUTCes, targets and effectof injllT)' . J. Ne .. rosei. Res. 17:2]4- 219 .

Nishida. T .. T . Hirato, N. Nisino , K. Mizuno, Y. Sekiguchi, M. Takano, K. Kawai. S. Nuai. and Y. Hirai . 1988. In Monokines an<! Other Lymphocytic C~tokine~. Alln R. Lin. [oc .. New York .

Pellegrino, t.. J., A. S. PelJegrino, and A. J. Cusllman. 1979. A Ste reota~ic Atlas of the Rat Brain. Plenum Publishing Corp .. New York.

Rennert, P. 0 ., and G. He inrich . 1986. Nerv~ growth fllClor mRNA in brain: locaJization by in siru hybridization . Bioch~m. Biophys. Ru. Commu". 138:g13- 816.

ScOIl, J. , M . Selby. M. Urdea. M. Quaroga. G . I. Bell, and W . J . Runer. 1983. Isolation and nuclCOlide sequence of a cONA encoding the precurrol o f mouse nerve llrowth {aclor. Naturt (Lond.). 302:S38-540.

Shelton. D. L ., and L. F. Reicliardt . 1984. Studies on OIe exp",ssion of tlle ß-nerve 8rowth factor (NGF) gene in the centr.1 nervous system: level of up"'Ulon anll regional distribution ofNGF mRNA suggestth3t NGF fune­tions as a trophic factor for several distiocl populations of neurons. Proc. Nil/I, Acad. Sci. USA. 83:2714- 2718.

Sprangt r, M .. O. Lindholm. C. E. Randtlow. R. Heumann , and H. Thocnen. 1990. Regulation of nerve growth fa<;tor syntliesis in the rat ccnua! nervous system: compariron between Ihc effccts of interteukin-I{J am! various growth factors in 3SU'OCyte eU)lu", s and in vivo. Eur. J. Nt .. roJet. 2:69-76.

Tagaya. Y., M. Okada, K. Sugie, T . Kasahara. N. Kondo, 1. Hamuro, K. Mat-5ushima. C. Oina",llo. IndJ. Yodoi . 1988 . IL-2 reccptor(p~5)rrac·intlucing faclOI. Puritication anti ~haraClerizalion ofadult T -cdl ]eukemia-dcrived fac­tor. J. Imm""ol. 140:2S 14_2620.

TIIocncn. H., am! Y. -A. Barde. ]980. Physiology ofnerve growth factor. Ph)'s­;01. Rtv. 60:1284-1335.

Thocnen, H. , C. E. Bandtlow, anti R. Hellmann. 1987 . Thc physiologieal fIIne­tion of nerve growlh fac:tor in the centra l nervous system: comparison with tlie periphery. Rtv. PhysioI. Biochtm. PhDfm<lcol. 109:]45-178.

Whittemo",. S. R ., lind A. Seiger. 1987. Thc uprcssion ]ocalWllion of beta· ncr;e growth faClor in the central nervous system. Brain. Rts. Rt .... 12:439--464.

Whinemore, S. R., T . Ebendal, L. Lärl<fors, L. Olson, A. Seiger, I. Ström­berg, and H. Persson. 19S6, Dcvclopmcntal and regi01Ul1 expression or tl-nerve Srowlb factor messenser RN'" and protein in tbc rat ~nlTal ~rvO\ls system. Pro<:. Nml. Acad. Sci. USA. S3:811-S21.

Wbillcmore, S. R .• P. L. Friedman. D. LarhamrrulT. H. Persson, M. Gonules­Carvajal. and V. R. Holets. 19S5. Rat tI·nervt growth facIDT scquencc and

si!c or syn!hesis in lhe adult hippocampus. J, Neurosci. Res. 20:403-410. Young. P . R . . Ind D. Sy lvcstcr. 1989. Cloning of rabbil in!erleukin-ltl :

difl'cn:ntial evolution of IL· la aoo IL-ltl proteins. P'Qlein Eng. 2:545- 551. Young, P. R .. D. J. Huudl , and P. L. Simon. 19S5. Human inlcrleukin Itl

is not secreted from ham~ter tibrobLast$ wben expressed conslitulively from transfected cDNA. J . Cell Biol. 107:447- 456 .