Proc. Natl. Acad. Sci. USAVol. 90, pp. 4414-4418, May 1993Medical Sciences

Production of tumor necrosis factor and other proinflammatorycytokines by human mononuclear phagocytes stimulated withmyelin P2 protein

(interleukin 1/interleukin 8/interleukin 12/experimental allergic neuritis)

PIERLUIGI BARON*t, GABRIELA CONSTANTIN*t, ANNALISA D'ANDREAt, DIEGO PONZIN§, ELIO SCARPINIt,GUGLIELMO SCARLATOt, GIORGIO TRINCHIERI1, FILIPPO RossI*, AND MARCO A. CASSATELLA*1*Institute of General Pathology, University of Verona, 37134 Verona, Italy; tDepartment of Neurology, Dino Ferrari Center, University of Milan, 20100 Milan,Italy; §Fidia Research Laboratories, 35031 Abano-Terme (PD), Italy; and tThe Wistar Institute, Philadelphia, PA 19104

Communicated by Giuseppe Attardi, January 28, 1993

ABSTRACT In this study we examined the effect ofmyelinP2 protein on some proinflammatory functions exerted byhuman mononuclear phagocytes. Northern blot analysis dem-onstrated that P2 protein selectively induced in monocytes andmacrophages mRNA accumulation of tumor necrosis factor(TNF), interleukin 113 (IL-10), and interleukin 8 (IL-8) in atime-dependent manner. Natural killer stimulating factor (IL-12) mRNA and protein secretion was strongly induced bylipopolysaccharide but not by P2 protein. Supernatants har-vested from P2-stimulated monocytes contained signicantamounts of TNF, IL-118, and IL-8, whereas those from mac-rophages contained only TNF and IL-8. The effect of the P2protein on TNF and IL-8 mRNA accumulation and secretionwas not affected by polymyxin B, which, on the other hand,almost completely abolished the effect of lipopolysaccharide.Finally, P2 protein did not directly trigger hydrogen peroxiderelease but, through the induced release of TNF, potentiatedmonocyte respiratory burst capability. Since P2 protein is theantigen responsible for the induction of experimental allergicneuritis, these findings identify a potential mechanism involvedin the infammatory reaction and myelin damage during ex-perimental allergic neuritis.

P2 is a 14-kDa basic protein found in myelin and in the cytosolof Schwann cells and oligodendrocytes. P2 protein is thecrucial antigen responsible for the immune events associatedwith experimental allergic neuritis (EAN) (1), an autoimmunedemyelinating disease of the peripheral nervous system(PNS) proposed as an animal model of the acute humanGuillain-Barrd syndrome (2).Mononuclear phagocytes are the predominant inflamma-

tory cell type in the PNS during EAN and their role in theeffector phase and amplification of the disease has been wellestablished (3, 4). Even though previous studies showed thatduring EAN macrophages phagocytize the myelin sheath (5)and inflict myelin damage by releasing inflammatory media-tors such as reactive oxygen intermediates (ROIs), arachi-donic acid metabolites, and hydrolases (6, 7), the mecha-nisms by which these processes occur are unknown.The present study was designed to examine whether mono-

nuclear phagocytes generate potentially injurious mediatorsupon interaction with purified P2 protein. We observed thatP2 protein induced the production of tumor necrosis factor(TNF), interleukin 13 (IL-1p), and interleukin 8 (IL-8), didnot stimulate ROI release from phagocytes, and activatedmonocyte respiratory burst capability. Our findings showthat specific myelin components can regulate inflammatoryand immunological responses and add further insights on the

role of phagocytic cells in the effector mechanisms duringEAN.

MATERIALS AND METHODSPuriflcation and Culture of Monocytes and Monocyte-

Derived Macrophages (MDMs). Peripheral blood mononu-clear cells were obtained from blood buffy coats of healthydonors (8) and were plated, depending on the type of exper-iment, in 24-well tissue culture plates (Nunc) at 2 x 106 per0.5 ml per well, in 6-well tissue culture plates (Nunc) at 15 x106 per ml per well, or in Petri dishes (Greiner, Nilrtigen,F.R.G.) at 10 x 107 per 6 ml. After 1 h at 37°C, nonadherentcells were removed by several washings with phosphate-buffered saline (PBS). The remaining adherent cells (>95%peroxidase-positive mononuclear cells) were cultured for upto 8 days in RPMI-1640 medium containing antibiotics andheat-inactivated 5% human serum, pooled from 10-15 do-nors. Bovine or rabbit P2 protein, purified according toPonzin et al. (9), or lipopolysaccharide (LPS; 10 ng/ml; fromEscherichia coli, serotype 026.B6; Sigma) was added tocultures of either 3-day-old monocytes or 7-day-old MDMs.Cell-free supernatants were harvested after the indicatedincubation times, spun at 8000 x g for 30 sec, and stored at-70°C, whereas the remaining adherent cells were eitherextracted for total RNA (when 6-well plates or Petri disheswere used) or lysed with 1 M NaOH for protein determination(when 24-well plates were used) (10). The content of LPS inthe P2 protein stock (3 mg/ml) solutions, and in all buffersand reagents used for this work, was <0.0125 unit/ml (asdetermined by the Limulus amebocyte lysate assay fromAssociates of Cape Cod).RNA Isolation and Northern Blot Analysis. Total RNA was

extracted from adherent phagocytes and analyzed as de-scribed (11). mRNAs for human TNF, IL-1p, IL-8, naturalkiller stimulating factor (NKSF/IL-12) p40 and p35, and actinwere detected by autoradiography after hybridization ofnylon filters (Schleicher & Schull) with 32P-labeled cDNAfragments (Multiprime Kit; Amersham).

Cytokine Assays. Antigenic TNF was determined by usinga double-ligand immunoassay (50 pg/ml detection limit),recently developed in our laboratory (12). Extracellular an-tigenic IL-8 was measured by a specific ELISA, having adetection limit of 20 pg/ml (13). IL-1p was determined eitherby ELISA [with a kit from Biochrom (Berlin); 50 pg/ml

Abbreviations: LPS, lipopolysaccharide; TNF, tumor necrosis fac-tor; PMA, phorbol 12-myristate 13-acetate; IFN, interferon; IL,interleukin; EAN, experimental allergic neuritis; PNS, peripheralnervous system; MDM, monocyte-derived macrophage; NKSF,natural killer stimulating factor; ROI, reactive oxygen intermediate;mAb, monoclonal antibody.lTo whom reprint requests should be addressed.

4414

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

Proc. Natl. Acad. Sci. USA 90 (1993) 4415

detection limit] or by a double-determinant RIA developedwith monoclonal antibodies (mAbs) 609 and 206 (5 pg/mldetection limit), kindly donated by M. A. Cousin (14).NKSF/IL-12 was determined by a double-determinant RIAdetecting the p70 heterodimers and the p40 free chains witha detection limit of 10 pg/ml, as described (15).H202 Assay. H202 release was assayed by the fluorimetric

measurement of the horseradish peroxidase-dependent oxi-dation of homovanillic acid (16). Briefly, monocyte or MDMmonolayers cultivated in 24-well trays, preincubated or notpreincubated for 24 h with LPS (10 ng/ml) or interferon y(IFN-y; 100 units/ml), were washed twice with PBS andstimulated with bovine or rabbit P2 protein (1-50 ,ug/ml) orphorbol 12-myristate 13-acetate (PMA; 100 ng/ml) dissolvedin 0.5 ml of a standard reaction mixture (16). After 60-90 minof incubation at 37°C, aliquots of the reaction mixture fromtriplicate wells per condition were collected and analyzed ina Perkin-Elmer LS-5 fluorometer.

RESULTSCytokine mRNA Expression and Production in Response to

P2 Protein and LPS. To determine the effect of myelin P2protein on the gene expression and secretion of cytokines bymononuclear phagocytes, monocytes and MDMs were cul-tured for different times with P2 protein (10 ,ug/ml) or withLPS. Northern blot analysis of the two experiments reportedin Fig. 1 A and B shows that in resting monocytes the mRNAexpression for TNF, IL-8, IL-1f3, and p40/IL-12 was absentor very low. Steady-state mRNA levels of TNF, IL-8, andIL-1,8 were clearly induced after culture with either bovine P2protein (Fig. 1A) or rabbit P2 protein (Fig. 1B), though to alower extent with the latter. Maximal accumulation of TNF,IL-8 (Fig. 1A), and IL-1p (not shown) mRNA in response toP2 protein was observed at 6 h and was followed by adecrease at later times. In comparison with P2, the levels ofTNF, IL-8, and IL-1l3 transcripts induced upon stimulationwith LPS were higher (see densitometry), and in the case ofIL-8 they continued to rise over the 24-h time period (Fig.1A). Furthermore, only LPS was able to induce NKSF/IL-12p40 (Fig. 1B) and NKSF/IL-12 p35 (not shown) mRNAaccumulation. mRNA accumulation of p47-phox, one of the

._-ECDVCi

s= q

(0 N

cn

c\A _- C))

D3 aAj Q )

C9ww0, c., L N 5 a - o

TNF 2 o TNF

*15 .C12

_ ~IL-8 ' 1t IL-1 E

I8a) CC)

`3 5.n4

p47- 3 L-12/phox 2

L3

- JSL ( L

L "_ _ " ~Actin E EActin

cytosolic components of the NADPH oxidase superoxide-generating system, was strongly enhanced by LPS, as re-ported (17), and it was increased also by the P2 protein at the24-h time point (Fig. 1A). Fig. 1C shows that in MDMstreated for 6 h with bovine P2 protein and LPS, the patternof expression of TNF, IL-8, IL-1,8, and NKSF/IL-12 p40mRNA was substantially similar to that observed in mono-cytes.

Determination of cytokine protein levels in cell-free super-natants harvested from P2- and LPS-treated monocyte orMDM cultures revealed that unstimulated cells did not re-lease TNF, IL-8, IL-1,3, and NKSF/IL-12 at levels signifi-cantly higher than the detection limit of the assays (Fig. 2).Monocytes stimulated with P2 protein released significantamounts of TNF, IL-8, and IL-lp but no NKSF/IL-12 (Fig.2A), whereas MDMs secreted only TNF and IL-8 (Fig. 2B).The concentrations of IL-8 and IL-1,8 in the supernatants ofP2-stimulated cells continued to increase up to 24 h, whereasthose of TNF reached a maximum level at 6 h and thendecreased (not shown). After stimulation with LPS, the levelsof immunoreactive TNF, IL-8, and IL-1,8 detected in thesupernatants of monocyte or MDM cultures followed similarkinetics but were on average 6-20 times higher than thoseobtained after P2 treatment. In addition, unlike P2, LPS alsoinduced the secretion of high amounts of NKSF/IL-12 (Fig.2). The pattern of cytokine release described was extremelyreproducible in every experiment, and overall results werequalitatively the same even though they were expressed as afunction of the adherent protein content per well (see, forexample, the experiments shown in Figs. 3 and 5).

In additional experiments, P2 protein and different con-centrations of LPS were preincubated with polymyxin Bsulfate (10 ,g/ml) before addition to MDMs. Polymyxin Bbinds to the lipid A moiety of endotoxin and prevents mostof LPS biologic activities (18). Polymyxin B almost com-pletely abolished mRNA accumulation (not shown) and se-cretion of TNF and IL-8 induced by all doses of LPS tested,whereas it had no effect on cytokine production induced bythe P2 protein (Fig. 3).

Effect of P2 Protein on Monocyte and MDM Capability ToProduce H202. To determine if P2 could directly induce theproduction of ROIs, we stimulated monocytes and MDMs

B

15

12

E ) L N :::

:3 EL D~ o

C-C

E (O.2 (0

TNF 2

2.5

" ~~~~~~.,°2n_ IL-8 a) _

-o 1.61I

--1.621

co

IL-12/ 3tSp40

*"_ Actin ~~/LPS P2

ActinI~~

Monocytes Macrophages

FIG. 1. Effect of myelin P2 protein and LPS on mRNA expression for TNF, IL-8, IL-1p3, NKSF/IL-12 p40, and p47-phox in humanmonocytes and macrophages. Northern blots of total RNA (8 ,g per lane) purified from two different preparations of monocytes (A and B) andone ofMDMs (C), hybridized with the indicated cDNA probes, are shown. Mononuclear phagocytes were cultured with P2 protein (10 pg/ml)and LPS (10 ng/ml) for the indicated times. The histograms show a quantitative representation of hybridizations obtained from densitometricanalysis of the exposed films. The values on the histograms are plotted as arbitrary densitometric units after normalization on the basis ofhybridization to actin mRNA. The results shown are representative of three experiments performed with each cell type.

Medical Sciences: Baron et al.

w2

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

4416 Medical Sciences: Baron et al.

A Monocytes B Macrophages TNF release

120

120 IL-8

E 90LE020 1 IL-1:

00 IL-113

E 400 p0.003

NKSF/IL-12200

16001E0 1200 _I

/P2 LPS

U]._0

to

-

So

901

601

30112

8-

4 P<000010 MM

80O

NKSF/IL-12

P2 LPS

FIG. 2. Effect ofmyelin P2 protein and LPS on TNF, IL-8, IL-1,3,and NKSF/IL-12 release by human monocytes and MDMs. Mono-cytes (A) and MDMs (B) were stimulated for 6 h with P2 protein (10jug/ml) or LPS (10 ng/ml). Supernatants were harvested frommonocytes cultured in 6-well plates (n = 5) and from MDMs culturedin 24-well plates for TNF and IL-8 determinations (n = 5) or in Petridishes for IL-1l3 and NKSF/IL-12 determinations (n = 4). Meanvalues ± SEM of duplicate assays performed with supernatantscollected and pooled from three wells for each condition are shown.P values of statistically significant (Student's t test for paired data)differences between P2-treated and untreated cells are indicated.n.s., Not significant.

with P2 protein and measured the release of H202. UnlikePMA, which is a potent stimulus ofthe oxidative burst, dosesof P2 protein ranging from 1 to 50 ,.g/ml (Fig. 4 and data notshown) did not induce production of H202 by human phago-cytes. P2 protein was also ineffective in inducing H202production by MDMs primed by LPS and IFN-y, whichresponded to PMA stimulation with enhanced H202 release(Fig. 4 Lower). However, pretreatment ofmonocytes with P2for 24 h enhanced their ability to produce H202 in responseto PMA in a manner similar to LPS or TNF (see Fig. 5Lower).

Effect of Anti-TNF Antibodies on P2-Induced Release ofIL-8 and P2-Dependent Activation of Respiratory Burst Ca-pability. To determine whether TNF rapidly produced inresponse to P2 stimulation could be responsible for elicitingIL-8 production (19) and respiratory burst activation, we

performed studies in which neutralizing antibodies to TNF(B.154.2) (20) were added to P2-, LPS-, and TNF-stimulatedmonocytes. Anti-TNF antibodies, over a 24-h time course,reduced by 61% the production of IL-8 in response to P2protein, whereas it slightly (20%) affected the release of IL-8stimulated by LPS (Fig. 5 Upper). Anti-TNF antibodiesneutralized almost completely the release of IL-8 induced byconcentrations ofTNF (5 ng/ml) similar to those observed tobe released by P2-stimulated monocytes, whereas an isotype-matched antibody (2B10) (21) had no effect in any conditions.Under the same experimental conditions, anti-TNF antibod-

En

8)0S.4

C4wE

60

40

20

E| -polymyxinI +polymyxin

/ I ng 10 ng 10 ,ug/ml

IL-8 release

/ 1 ng 10 ng

LPS

__ug m

__

FIG. 3. Effect of polymyxin B sulfate on TNF and IL-8 releaseby human MDMs in response to P2 protein and LPS. Eight-day-oldMDMs were cultured in 24-well plates with P2 protein (10 pg/ml) ordifferent doses of LPS in the presence or absence ofpolymyxin B (10pg/ml). After 6 h of incubation, supernatants from three differentwells per each condition were pooled for cytokine ELISAs. Meanvalues of duplicate assays from a representative experiment (out ofthree performed with similar results) are shown. The protein contentdetected in the wells from MDMs incubated with polymyxin B (mean± SEM = 52 + 3 1tg, n = 15) was not statistically different from thatof untreated wells (48 ± 5).

ies also abrogated the ability ofTNF and P2 protein, but notof LPS, to potentiate monocyte capability to produce H202in response to PMA (Fig. 5 Lower).

DISCUSSIONThis study demonstrates that myelin P2 protein is able toinduce the release of TNF, IL-13, and IL-8 by culturedhuman monocytes and MDMs. The effect of P2 protein onTNF, IL-1,3, and IL-8 production is regulated at the level ofmRNA accumulation, and, at least with monocytes, it wasobtained with P2 proteins purified from two different specieshaving homology of -95%. Although P2 protein did nottrigger the production of ROIs by mononuclear phagocytes,regardless of their stage of maturation or activation, itpotentiated the oxidative burst capability of monocytes in amanner equivalent to that of LPS or TNF. Interestingly, themRNA levels of p47-phox, one of the NADPH oxidasecomponents that presumably plays a key role in the mecha-nism of activation of macrophage respiratory burst (17, 22),were also increased in monocytes treated with P2 protein.However, use of a neutralizing anti-TNF mAb revealed thatthe P2-induced activation of respiratory burst and part of theP2-induced release of IL-8 were mediated by the TNFinduced by P2 stimulation. Whether the P2 protein interactswith a specific receptor on phagocytes or is simply internal-ized by fluid-phase pinocytosis is not known yet.The following observations unequivocally show that.the

effects attributed to the P2 protein were not influenced by

Proc. Natl. Acad Sci. USA 90 (1993)

F-I

. m

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

Proc. Natl. Acad. Sci. USA 90 (1993) 4417

16donorA

12 |M donor B31I

Monocytes

81

41

resting P2 lpg/rnl P2 10g/ml

Macrophages

IL--- release. 5000C\Q

14000

o 30000.cL 2000

-_ 1000K,- oo

tw

7--

MaTNFIimAXb(TRLmlAib:

Wm]medium LPS I) TNFPMA-- sL1 rn ulated 11202 prodc:tion

.

0

r-

ES

C)

E._1-

c

/ P2 PMA / PPMA

LPS-treated IFNvtreated

FIG. 4. Effect of myelin P2 protein on H202 release by humanmonocytes and MDMs. Monocytes and MDMs, cultured in 24-wellplates, were stimulated for the times indicated with P2 protein (10,ug/ml with MDMs) or PMA (100 ng/ml). MDMs were also previ-ously activated by an exposure to LPS (10 ng/ml) or IFN--y (100units/ml) for 24 h (Lower). H202 release was measured as described(16). Results are from representative experiments (out of threeperformed with each cell type) giving similar results.

trace levels of endotoxin present in our P2 stocks: (i) LPS,but not P2, was able to increase the mRNA levels encodingfor NKSF/IL-12 and to induce its secretion; (ii) maximalexpression of IL-8 mRNA in monocytes stimulated with P2protein was detected at 6 h, whereas in cells stimulated withLPS it was detected at 24 h (this paper; ref. 23); (iii)polymyxin B almost completely suppressed production ofcytokines in response to 1-100 ng of LPS per ml, but it hadno effect on P2 protein stimulation; and (iv) anti-TNF mAbhad a weak influence on the capability ofLPS either to inducethe release of IL-8 or to activate macrophage respiratoryburst, in agreement with previous observations (23, 24), but,as mentioned above, significantly inhibited some of theeffects determined by the P2 protein.

Production of mononuclear phagocyte-derived TNF, IL-13, and IL-8 following P2 stimulation could represent animportant mechanism involved in demyelination duringEAN. TNF is a pleiotropic proinflammatory cytokine (25)capable of inducing myelin lysis and oligodendrocyte necro-sis in vitro (26), and it is known to be involved in thedevelopment of experimental allergic encephalomyelitis andmultiple sclerosis lesions (27). IL-1l3 may promote recruit-ment of leukocytes by increasing their adhesion to endothe-lial cells, thus favoring the onset of the inflammatory process(28). IL-8 is a cytokine that typically mobilizes and activatespolymorphonuclear leukocytes (PMNs) (29) but that also actsas a chemoattractant for about 10% of human peripheralblood T lymphocytes (30). The significance of the present invitro studies is supported by the fact that in the early stagesofEAN, vesicular disintegration of myelin without the direct

FIG. 5. Effect of anti-TNF antibody on P2-stimulated IL-8 pro-duction and on P2 activation of monocyte respiratory burst. IL-8release and H202 production were investigated in the same monocytepreparation cultured in 24-well plates. Cells were stimulated with P2protein (10 .g/ml), LPS (10 ng/ml), orTNF (5 ng/ml) with or withoutan anti-TNF antibody (5 ,L.g/ml) or control (CTRL) antibody. At the24-h time point, supematants were harvested for IL-8 measurement,and then adherent cells were incubated with PMA (100 ng/ml) totrigger H202 production. A representative experiment (out of threeperformed with similar results) is shown.

proximity of invading cells is known to occur (31), and PMNsand lymphocytes are present within the nerve of P2-immunized animals (32). Therefore, P2 protein, derived frommyelin membrane altered during the primary stages of thedisease, could stimulate macrophage production of TNF,IL-1,B, and IL-8. These cytokines might increase prolifera-tion, activation, and chemotaxis of other inflammatory cellsor act to degrade myelin directly or sustain a myelin-destructive reaction. Recently (4), monocytes isolated frompatients with Guillain-Barrd syndrome were found to releaselarger amounts of H202 than monocytes from patients withother noninflammatory, nondemyelinating disease. Our ob-servations that P2 protein induces the release of TNF, IL-8,and IL-1,3 provide new insights on the role of macrophagesin the effector mechanisms of inflammatory demyelinatingneuropathies and warrant further analysis on the effects ofother myelin constituents. The results may also be relevantto the events that occur during the course of Walleriandegeneration in the PNS, in which myelin degradation isaccompanied by macrophage activation, IL-1 secretion, andresultant secretion in the distal stump by nonneuronal cells ofnerve growth factor, presumably contributing to subsequentregeneration (33). Demyelination initiated or sustained bycytokines secreted by macrophages suggests several strate-gies for intervention-for instance, use of pharmacologicinhibitors or neutralizing antibodies, which might be able toreduce inflammation and thus provide a specific approgchtoward treatment of inflammatory demyelinating diseases.

We thank Dr. G. Berton for his critical reading of this paper andDr. C. S. Raine (Albert Einstein College of Medicine, Department ofNeuropathology, New York), Dr. I. Lindley, and Dr. M. Ceska

C)

1-

0S

oll0

9

0S

Medical Sciences: Baron et al.

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1

4418 Medical Sciences: Baron et al.

(Sandoz Pharmaceutical, Vienna) for kindly providing us with rabbitP2 protein, IL-8 cDNA, and IL-8 ELISA, respectively. This workwas supported by Grants from Ministero della 'Universita' e delleRicerca Scientifica e Tecnologica (fondi 40%), Consiglio Nazionaledelle Ricerche (Grant 92.02777.CT04), "Dino Ferrari Center," andthe Italian Multiple Sclerosis Society and in part by U.S. PublicHealth Service and National Institutes of Health Grants CA 10815,CA 20833, CA 32898, and CA 40256. G.T. and M.A.C. are recipientsof a North Atlantic Treaty Organization Travel award.

1. Kadlubwoski, M. & Hughes, R. A. C. (1979) Nature (London)277, 140-141.

2. Ropper, A. H. (1992) N. Engl. J. Med. 326, 1130-1136.3. Raine, C. S. (1985) Handbook ofClinical Neurology (Elsevier,

Amsterdam), Vol. 3, Rev. Ed., pp. 429-466.4. Hartung, H. P. & Toyka, K. V. (1990) Ann. Neurol. 27, 57-63.5. Lampert, P. W. (1969) Lab. Invest. 20, 127-138.6. Hartung, H. P., Schafer, B., Heininger, K. & Toyka, K. V.

(1988) Ann. Neurol. 23, 453-460.7. Hartung, H. P., Schafer, B., Heininger, K., Stoll, G. & Toyka,

K. V. (1988) Brain 111, 1039-1059.8. Cassatella, M. A., Bazzoni, F., Flynn, R. M., Dusi, S., Trinch-

ieri, G. & Rossi, F. (1990) J. Biol. Chem. 265, 20241-20246.9. Ponzin, D., Menegus, A. M., Kirschner, G., Nunzi, M. G.,

Fiori, M. G. & Raine, C. (1991) Ann. Neurol. 30, 678-685.10. Bradford, M. M. (1976) Anal. Chem. 72, 248-254.11. Cassatella, M. A., Hartman, L., Perussia, B. & Trinchieri, G.

(1989) J. Clin. Invest. 83, 1570-1579.12. Zambello, R., Trentin, L., Bulian, P., Cassatella, M. A., Rai-

mondi, R., Chisesi, T., Agostini, C. & Semenzato, G. (1992)Blood 80, 1-5.

13. Ceska, M., Effenberger, F., Peichl, P. & Pursh, E. (1989)Cytokine 1, 136 (abstr.).

14. Fontaine, S., Damais, C., Lando, D. & Cousin, M. A. (1989)Lymph. Res. 8, 129-139.

15. D'Andrea, A., Rengaraju, M., Valiante, N. M., Chehimi, J.,Kubin, M., Aste, M., Chan, S., Kobayashi, M., Young, D.,

Nickbarg, E., Chizzonite, R., Wolf, S. F. & Trinchieri, G.(1992) J. Exp. Med. 176, 1387-1398.

16. Valletta, E. A. & Berton, G. (1987) J. Immunol. 138, 4366-4373.

17. Cassatella, M. A., Bazzoni, F., Aste Amezaga, M. & Rossi, F.(1991) Biochem. Soc. Trans. 19, 63-67.

18. Morrison, D. C. & Jacobs, D. M. (1976) Immunochemistry 133,813-820.

19. Matsushima, K., Morishita, K., Yoshimura, T., Lavu, S.,Kobayashi, Y., Lew, W., Appella, E., Kung, H. F., Leonard,E. J. & Oppenheim, J. J. (1988) J. Exp. Med. 167, 1883-1893.

20. Perussia, B., Chan, S. H., D'Andrea, A., Tsuji, K., Santoli, D.,Pospisil, M., Young, D., Wolf, S. F. & Trinchieri, G. (1992) J.Immunol. 149, 3495-3502.

21. Berton, G., Dusi, S., Serra, M. C., Bellavite, P. & Rossi, F.(1987) J. Biol. Chem. 264, 5564-5568.

22. Levy, R. & Malech, H. L. (1992) J. Immunol. 147, 3066-3071.23. DeForge, L. E., Kenney, J. S., Jones, M. L., Warren, J. S. &

Remick, D. G. (1992) J. Immunol. 148, 2133-2141.24. Szefler, S. J., Norton, C. E., Ball, B., Gross, J. M., Aida, Y.

& Pabst, M. J. (1989) J. Immunol. 142, 3985-3992.25. Beutler, B. & Cerami, A. (1989) Annu. Rev. Immunol. 7,

625-655.26. Selmaj, K. W. & Raine, C. S. (1988) Ann. Neurol. 23, 339-346.27. Powell Broome, M. & Steinman, L. (1992) in Tumor Necrosis

Factors: The Molecules and Their Emerging Role in Medicine,ed. Beutler, B. (Raven, New York), pp. 355-364.

28. Dinarello, C. A. (1988) FASEB J. 2, 108-115.29. Baggiolini, M. & Clark-Lewis, I. (1992)FEBS Lett. 307, 97-101.30. Larsen, C. G., Anderson, A. O., Appella, E., Oppenheim, J. J.

& Matsushima, K. (1989) Science 243, 1464-1466.31. Ballin, R. H. M. & Thomas, P. K. (1969) J. Neurol. Sci. 8,

1-18.32. Rosen, J. L., Brown, M. J., Hjckey, W. F. & Rostami, A.

(1990) Muscle Nerve 13, 629-639.33. Lindholm, D., Heumann, R., Meyer, M. & Thoenen, H. (1987)

Nature (London) 330, 658-659.

Proc. Natl. Acad Sci. USA 90 (1993)

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

1