INTERNAI ION:V. J(NRNAI. or 1.I IROSY^

Volume 67, Number 1l'rinrrd in lhe U.S.A.

(ISSN 0148-916X)

Roles of Tumor Necrosis Factor-a and TransformingGrowth Factor-• in Regulating Intercellular Adhesion

Molecule-1 Expression on Murine PeritonealMacrophages Infected with M. leprae 1

Toshiaki Shimizu, Win Win Maw, and Haruaki Tomioka'

Adhesion molecules expressed on im-munocompetent cells play important rolesin cellular interactions for the developmentof immunological responses ( ,4). Intercellu-lar adhesion molecule-1 (ICAM-1) plays arole in lymphocyte trafficking by allowingICAM-1-bearing endothelial cells to inter-act with leukocyte function-associated anti-gen-1 (LFA-1) on lymphocytes, and therebyfacilitate the migration of lymphocytes tosites of inflammation along endothelialcells ("). Moreover, the interaction of LFA-1 with ICAM-1 is criticai for effective cel-lular interactions of T cells with antigen-presenting cells, including macrophages(M(I s), and the consequent activation ofresting T cells (12. 'S ''). Indeed. ICAM-1augments T-cell proliferation and is pivotalin recall antigen responses of T cells to pu-rified protein derivative of Mvcobacteriun(tuberculosis ( 3"). Moreover, a recent studyof Lopez Ramirez, et al. (") showed thatthe ICAM-1 expression on THP-1 humanmonocyte-derived cell line was potently in-creased by stimulation with M. tuberculo-sis. In addition, the ia vivo study by Sulli-van, et al. (") concerning M. leprae infec-tion revealed that granulomas from bothtuberculoid leprosy and lepromatous lep-rosy displayed ICAM-1 expression. Thisimplies that ICAM-1 plays important rolesin enhancing the immune and inflammatory

' Received for publication on 25 November 1997.Accepted for publication in revised form on 10December 1998.

T. Shimizu, M.S.; W. W. Maw, M.D., Ph.D.; H.Tomioka, Ph.D., Department of Microbiology and 1m-munology, Shimane Medical University, Inumo, Shi-mane 693, Japan.

Reprint requests to Dr. 11. Tomioka at the above ad-dress or FAX 81-853-20-2145.

responses to M. leprae. Therefore, it is ofinterest to know the profiles of ICAM-1 ex-pression by M(1)s after stimulation with Al.leprae. In the present study, we found thatICAM-1 expression on murine peritonealM(s was increased during days 1 to 3 af-ter stimulation with M. leprae because ofautocrine upregulation by tumor necrosisfactor-alpha (TNF-a). The ICAM-1 ex-pression thereafter gradually decreased, re-turning to the normal levei by day 7 due tothe downregulating effect of transforminggrowth factor-beta (TGF-(3) derived fromM. /eprae-stimulated Mis themselves.

MATERIALS AND METHODSOrganisms. M. leprae were purified

from the liver of an infected armadillo(#825) obtained from G. P. Walsh, ArmedForces Institute of Pathology, Washington,DC, U.S.A., by Percoll gradient centrifuga-tion according to the method of Draper(IMMLEP protocol 1/79) with slight modi-fications (21 ). This M. leprae strain 825 con-tained 30% of living cells on the basis offluorescein diacetate and ethidium bromide(FDA/EB) staining and 10" of the organ-isms contained 7.4 pmoles of ATP, indicat-ing good viability of the bacterial prepara-tion. In some experiments, M. leprae organ-isms were collected and purified frominfected foot pads of BALB/c nude mice, aspreviously reported Cl. When 10` of thesenude mouse M. leprae (strain Thai-53)were inoculated in the 7H 12B medium ofthe BACTEC 460 Tuberculosis System andcultured according to Franzblau ("), thegrowth index (GI) value increased from O to122 and 267 at day 4 and day 11, respec-tively, indicating good viability of this bac-terial preparation. When BALB/c rude

36

67, 1^ Shimizu, et al.: ICAM-1 Regulation^ 37

mice were infected with 10' of M. lepraeThai-53 into the hind foot pad, the organ-isms steadily grew at the sites of infection,reaching bacterial loads larger than 10'bacilli per foot pad within 1 year after in-fection (data not shown). This also indi-cates good viability of the bacterial prepara-tion used.

Mice. Seven- to ten-week-old, maleBALB/c mice purchased from Japan CleaCo., Osaka, Japan, were used.

Special agents for Md) culture. Re-combinant mouse TNF-a, ultrapure naturalhuman TGF-(31 and mouse anti-humanTGF43 monoclonal antibody (mAb) (alsospecific to mouse TGF-(3) were purchasedfrom Genzyme Co., Cambridge, Massa-chusetts, U.S.A. Rat anti-mouse TNF-amAb was obtained from Pharmingen Co.,San Diego, California, U.S.A. One unit ofTNF-a is defined as the amount required tomediate half-maximal cytotoxicity of L929cells. All of these agents contained nopreservative, such as sodium azide, andwere found to be essentially free from LPScontamination by the Limulus J Single Test(Wako Pure Chemical Industry Co., Osaka,Japan).

Medium. Medium RPMI 1640 (NissuiPharmaceutical Co., Tokyo, Japan) supple-mented with 25 mM HEPES, 2 mM gluta-mine, and 5% heat-inactivated fetal bovineserum (FBS) (M. A. Bioproducts, Walk-ersville, Maryland, U.S.A.) was used forcell culture unless otherwise specified. Themedium and FBS used were found to befree from LPS contamination by the Limu-lus J Single Test.

Peritoneal Meos. Peritoneal exudatecells (PECs) were collected with Hanks'balanced salt solution (HBSS) containing2% FBS from mice given an intraperitonealinjection of 1 ml of peptone-starch 4 daysbefore harvest. After washing with 2%FBS-HBSS by centrifugation at 200 x g x 5min, the PECs were suspended in a smallvolume of the culture medium. Ten ml eachof PEC suspension in 5% FBS-RPMI 1640at the cell density of 5 x 10"/ml was pouredonto a 90-mm cell culture plate (CorningGlass Works Co., Corning, New York,U.S.A.) which was overlaid with 14-mmplastic sheets (about 20 sheets) (Wako PureChemical Industry Co., Osaka, Japan). Af-ter a 2-hr incubation at 37°C in a CO, incu-

bator (5% CO 3 -95% humidified air), the re-sultant sheets were removed, thoroughlyrinsed with 2% FBS-HBSS, and immersedfinto 1 ml of culture medium in a 16-mmcell culture well (24-wells; Corning). Thismethod gave more than 90% pure M(I) cul-tures, with potent pinocytic ability for neu-tral red and phagocytic activity for latexbeads.

Microscopic counting for ICAM-1-positive M 1 s. M(I) monolayer cultures onthe plastic sheets were cultured in 1.0 ml of5% FBS-RPMI 1640 in 16-mm culturewells (Corning) in the presence or absenceof 1 x 10 7/ml of M. leprae or latex beads(3.0-pm diameter; Sigma Chemical Co., St.Louis, Missouri, U.S.A.) at 37°C in a CO,incubator for up to 14 days. At intervals, the-M(I) monolayer cultures were removed,washed with phosphate buffered saline(PBS) containing 02% bovine serum albu-min (BSA), fixed with 1% paraformalde-hyde, incubated in 1% BSA-PBS at roomtemperature for 30 min, and then reactedwith rat anti-mouse ICAM-1 mAb (Seika-gaku Corp., Tokyo, Japan) for 1 hr. Afterrinsing with 0.1% BSA-PBS, the M 1 s werereacted with alkaline phosphatase-conju-gated anti-rat Ig mAb (Seikagaku) for 1 hrand then washed with 0.1% BSA-PBS.Color development was achieved by usingNBT-BCIP substrate (nitroblue tetra-zolium/5'-bromo-4-chloro-3-indolylphos-phate) and the proportion of blue-stainedM(I)s (ICAM-1-positive cells) was enumer-ated by microscopic counting ('a).

Cytokine measurement. The 1-, 3- or7-day culture fluids of M I s with or withoutM. leprae infection were measured forTNF-a and TGF-f3 concentrations as previ-ously described. Briefly, Immulon 4 plates(Dynatech Laboratories, Chantilly, Vir-ginia, U.S.A.) were coated with a captureAb for each cytokine using rat anti-mouseTNF-a mAb (Pharmingen) and mouse anti-human TGF-(3 mAb (also specific to mouseTGF-(3) (Genzyme). Biotinylated rat anti-mouse TNF-a mAb (Pharmingen) andchicken anti-human TGF-(3 mAb (R & DSystem, Inc., Minneapolis, Minnesota,U.S.A.) were used as the detecting Abs. Af-ter binding of alkaline phosphatase-conju-gated streptavidin (Life Technologies Co.,Gaithersburg, Maryland, U.S.A.) in the caseof TNF-a assay and alkaline phosphatase-

-oo -

60

.}r

O) 50

0)O 40Q)

oa-

30 _1

20 1

Qo

10

oO^1^3

Cultivation time(days)

1714c^ 3

Cultivation time(days)

o0

.ó07.

01)

6)

No

1 00

80

GO

38^

International Journal aíLeprosv^ 1999

FIG. I. Profiles of ICAM-I expression by h1. lep-rae 825-infected M(I)s during cultivation in mediumwith or without addition of exogenous "INF-a. =Uninfected M(I)s were cultured in lhe absence of TNF-a;

= M. leprae-infected M ( I)ti were cultured in lhe ab-sence of "INF-a;ell = M. /eprae-infected M(Ds werecultured in the presence of TNF-a added at 500units/1111 on day 0. Each bar indicates mean ± S.E.M.(N = 2).

conjugated rabbit anti-chicken/turkey IgGAb (Zymed Laboratories Inc., San Fran-cisco, California, U.S.A.) in the case ofTGF-(3 assay, color development wasachieved using p-nitrophenyl phosphatetablets (Sigma) as the substrate. Capture Aband detecting Ab for each cytokine had dif-ferent epitope specificities from each other.The present ELISA method was capable ofmeasuring the active form TGF-(3 in Md)culture fluids in a specific fashion.

RESULTSICAM-1 expression on M. leprae-in-

fected MOs during cultivation. M shownin Figure 1, when MDs were infected withM. leprae at day O (hatched bars), theICAM-1 expression of cultured Mis, interms of increase in the ratio of ICAM-1-positive Mdas measured by the microscopicmethod, rapidly increased reaching thepeak in the early phase (during day 1 to day3) of cultivation and thereafter decreased,returning to the normal levei by day 7. Alower increase in the ICAM-1 expressionwas also noted for uninfected (unstimu-lated) MDs in the same period of time(open bars). In this experiment, residualnumbers of attached MDs on a plastic cul-ture sheet during cultivation after M. lep-

Etc. 2. Effeels of exogenous TNF-a on ICAM-1expression by uninfected M(I)s during eultivalion.M(Ds were cullivaled in the presence (III) or absence( 1) of TNF-a which was added at 500 units/1111 ondays 0, 2, and 6 during eultivatioa. Each bar indicatesmean ± S.E.M. (N = 3). * Significnlly larger thanvalue of M(Ds cultured in absence of TNF-a (p <0.05;Student's t test).

rae-infection (hatched bars) were changedas follows when the day O value was fixedas 100% (N = 5): day 1, 97.6 ± 5.5%; day 3,91.0 ± 7.1%; day 7, 78.9 ± 6.9%. This indi-cates that the reduction of ICAM-1 expres-sion in the middle phase (day 7) of M(13 cul-tivation after M. /eprae-infection was notdue to M. /eprae-mediated cytotoxicity.Figure 1 also shows that the addition ofTNF-a at the concentration of 500 units/mlovercame the middle phase (day 7) down-regulation of the ICAM-1 expression(shaded bar), indicating that ICAM-1downregulation in the middle phase wasdue in pari to a defect of TNF-a. In thiscontext, separate experiments showed thatTNF-a, a crucial ICAM-1 upregulating cy-tokine ("• 2") displayed its ICAM- I upregu-lating effect on MOos as follows.

First, as shown in Figure 2, a marked in-crease in the ICAM-1 expression was alsoobserved for uninfected McDs, when treatedwith exogenous TNF-a (500 units/ml),which was added on days 0, 2, and 6, in theearly phase of cultivation (days 1 to 3) butnot in the middle phase (day 7). Thus, in thecase of uninfected M(s, exogenous TNF-acould increase only their early phaseICAM-1 expression, differing from the caseof M. /eprae-infected McDs (Fig. 1). Sec-ond, as shown in Figure 3, the addition of

TNF-a production (pg/ml)''^TGF-f3 production (pg/mU''

Day 7 Day 1 Day 3 Day 7

20± 10 50± 10 360 ± 10 30 ± 1040 ± I() 50± 10 490 ± 10' 420 ± 30

ND 50 ± I0 640 ± 10' 600 ± 20'40 ± 2 ND ND ND

^l(y)rue^Addition

^

infection^of Ah

+^—+ Anti-TNF-a+ Anti-TGF-13

20±2

ND'

Day I^Day 3

20±1010 ± 8

ND

67, 1^ Shinti;u, et al.: ICAM- I Regnlatiott^ 39

GO -

• 50

CD

a)

30oo.

<

•

ino

O^1^3

Cultivation time(days)

Fio. 3. Gffects of anti-TNF-a Ah on ICAM-I ex-pression hy Al. lepra(' 'fhai-53-infected M(I)s. 1 1 =Uninfected M(Os were cultured in the ahsence of anti-'INI -a Ah; r I = A/. /ryrue-infected M(I)s were cul-ture(' in the ahsence of anti "I'NI -u Ah; • = M. lrp-rue-infectetl M(I)s were cultured in the presence ofanti-]'NF-u Ah which was added at 10 Ltg/ml to theculture mediam on days (1. 2, and 6 during cultivation.Gach har indicates ntean ± S.G.M. (N = 3). ' Signifi-cantly smaller than the value of M(I)s cultured in theahsence of anti-TNF-a Ah (p <0.05; Student's t test).

anti-TNF-u Ah (10 ic/m1) caused the re-duction of the ICAM-1 expression by M./eprae-infected M(I)s and Chis effect wasmost marked at day 3, confirnting that M(1)-derived endogenous TNF-a plays the mostimportant role in the early phase (aroundday 3) increase in the ICAM-1 expressionby M. /eprae-infected M(1)s.

The above lindings indicate central rolesof TNF-a in ICAM-1 upregulation of M.

/eprae-infected M(I)s. However, as indi-cated in The Table only small amounts ofTNF-a were released from M. (eprae-in-fected M(I)s isto the culture medium duringcultivation periods for up to 7 days. It thusappears that the majority of the TNF-amolecules were produced by M. !eprae-in-fected M(I)s in a membrane-bound form,molecular weight (MW) of 26 kD, but notin a released forni (MW of 17 kD) (' `). Thismembrane-bound form of TNF-a seems toplay a crucial role in the ICAM-1 upregula-tion of M. /eprae-infected M(I)s.

Another important finding in Figure 1 isthat TNF-a failcd to overcome the latephase (day 14) downregulation of ICAM-1expression. This phenomenon may be ex-plained as follows. First, ICAM-1 reductionin the Iate phase seems to be due in part tonutrient deticiency in the M(I) culturemedium, as previously demonstrated for M.aviam complex-infected M(Is ("'). Second,it is plausible that, in the middle-to-Iatephase of cultivation, M. /eprae-infectedNI(1)s produced certain factors possessingan ICAM-1 downregulatory activity, as de-scribed below.

Role of. TGF-(3 in the middle-to-latephase downregulation ot' ICAi 1-1 ex-pression by M. !eprae-infected M(I)s.M(I)s stimulated with mycobacterial organ-isms produce TGF-(3, an immunosuppres-sive and M(l)-deactivating cytokine ( 4 . ").We previously found that the in litro TGF-Pproduction by murine peritoneal M(I)s wasinitiated after day 3 and continued for up to

ao

Ti II': T,v)I.t.. Produ tion of "I'NF-a anel TGF-fl by M. leprae-in/ceted MO.r durint., culti-vation in the presence or al)selu e ef anti-TNF-a or curti-TGF j Ah."

MÕs infected with ,l/. lepra(' Thai-53 were cultured ia mediam with or without addition of indicated anti-cytokineAbs. Teu µ g each of anti "FNF-a or anti TGF-L$ Ah was a^ltdded to the cabine mediam (1.0 nil) on days O. 2, and 6. Atintervals, eulture fluids sere measured for "INF-a and TFG-13 eoneentrations hy ELISA.

'' The mean±SF.N1.(N=3).Lower than the detectahle limit for TNF-a (I O pJnil) or TGF-13(IO pg/ml).Significantly I:uger than the value of uninfeeted Mós (p <0.005; Student's t test).ND = Not deternrined.

'•"^larger than the value of AI. /eprae-infected Mms cultured in the absence of anti-cytokine Ah(' p <0.(105; p <0.05: Student's t test).

O^1^3^7

Cultivation time(days)

80

20

O^1^3^7

80

a) 60oN

i3 40oo-

20

U

40^ International Jourual o/ Leproso^ 1999

Cultivation time(days)

FIG. 4. Eifeets of anti-TGF-13 Ah on ICAM-1 cx-pression hy M. leprae 825-infected M(Ds. E = Unin-fected Md)s were cultured in the absence of anti-TCF13Ah; 0= M. Ieprue-infected Mct)s were cultured in theabsence of alui-TU:-1S Ah; • = M. /eproc-infectedM(hs were cultured in the presentes of anui-TGF-l) Ahwhich was added at 30 pg/ml to the culture medium ondays 0, 3, and 6 during cultivation. Each har indicatesmean ± S.E.M. (N = 3). " Signiticantly larger lima valueof Mis cultured in absence aí anti-TGF-(3 Ah (p <0.05;Studcnt's t test).

at least day 14 when the MeDs were infectedwith M. aviam complex ( " '). As indicated inThe Table, significam leveis of TGF-(3 wereproduced into the culture medium by Misregardless of M. leprae infection. However,it was also found that Al. leprae-infectedMus produced significantly larger amountsof TGF-(3 than did uninfected M(I)s. No-tably, anti-TNF-a Ab enhanced the TGF-(3production by M. leprae-infected Mis. In aseparate experiment, the addition of 20µg/ml of anti-TGF-(3 Ab caused an 82 ± 2%reduction of TGF-f3 accumulation isto cul-ture fluids of M. leprae-infected MEI?s onday 3.

In addition to these findings, it has alsobeen reported that TGF-P displayed down-regulatory effects on the interferon-gamma(IFN-y)-induced ICAM-1 expression by ratmicroglia cells, which are cells of M(1) lin-eage ("). It is thus likely that TGF-Í3 alsoplays some active role in ICAM-1 down-regulation observed in the case of M. lep-rae-infected Mis. As shown in Figure 4,anti-TGF-(3 Ab significantly blocked themiddle phase (day 7) downregulation ofICAM-1 expression by Mis infected withM. leprae. In this case, anti-TGF-(3 Ab was

FIG. 5. Effects of anti-TGF43 Ah on ICAM-1 cx-pression hy Mis stimulated with heat (121°C, 15min)-killed M. leprae 825. ^ = ICAM-1 expressionhy normal McDs hefore stimulation with heat-killed M.leprae; 111. leprae-stimulated Md)s were culturedin the absence of anti-TGF-13 Ah; • = M. Ieprae-stintulated M4)s were cultured in the presence of anti-TGF-13 Ab which was added at 30 pg/ml to culturemedium on days 0, 3, and 6 during eultivation. Eachhar indicates mean ± S.E.M. (N = 4). '"' Signiticantlylarger lima value of McDs cultured in absence of anti-TGF-13 Ah (p <0.05; Student's t test).

added at 30 pg/mI on days 0, 3, and 6. Inseparate experiments, anti-TGF-(3 Ab at 30µg/ml was found to neutralize 12 ng/ml ofTGF-(3, which is much larger than theamounts of TGF-(3 produced by M. leprae-infected M 1 s (The Table).

Figure 5 shows the effects of anti-TGF-(3Ab on the ICAM-1 expression by M(bsstimulated with heat-killed (121°C, 15 min)M. leprae. In this case, anti-TGF-(3 Ab sim-ilarly blocked the middle phase (day 7)downregulation of ICAM-1, thereby indi-cating that profiles of ICAM-1 expressionin Mis stimulated due to infection with liv-ing M. leprae organisms were essentiallythe same as those in M(1 s stimulated withheat-killed M. leprae.

Effects of addition of TNF-a or anti-TGF-(3 Ah on ICAM-1 expression by la-tex beads-phagocytosing Mcbs. In orderto know whether or not TNF-a-mediatedupregulation and TGF-(3-mediated down-regulation of Md) ICAM-1 expression ob-served in Figures 1, 4, and 5 was specifi-cally caused by M. leprae infection, we ex-amined the effects of the addition of TNF-aand anti-TGF-P Ab on the ICAM-1 expres-sion by Mis stimulated by phagocytosis oflatex beads. For this purpose, McDs were

60

50

20

67, 1^S/iilnizu, et al.: ICAM-I Re,ç>ulation^ 41

O^1^3^7

Cultivation time(days)

Fic. 6. Profiles of ICAM-1 expression hy latexheads-phagocytizing M(hs during cultivation. 1 1 =Day 0: ICAM-1 expression by normal M<0s bcforc la-tex phagocytosis. Days 1, 3, 7: ICAM-1 expression hycontrol McUs without phagocytizing latex particles af-ter cultivation in medium without addition; = la-tex beads-phagocytizing MNs cultured in mediamwithout addition; I. and

[] = latex beads-phago-

cytizing M(I)s cultured in presence of cithcr TNF-a( me) or anti-TGF-13 Ah (1;t j). TNF-a and anti-TGF-(3 Ah added to culturc medium at 500 units/mi and 10

respcctively, on days 0, 2, 6 during cultivation.Each bar indicates mean ± S.E.M. (N = 3). * Signili-cantly largcr than the valuc of control M(I)s (p ,0.05;Student's 1 test).

cultured with 3-pm latex beads which areknown to be taken up by MDs throughphagocytosis. As shown in Figure 6, theICAM-1 expression by latex beads phago-cytizing MDs did not reduce so markedly atday 7 as in the case of M. leprae-infectedMis (Figs. 1, 3, and 4). Moreover, neitherTNF-a nor anti-TGF-(3 Ab significantly af-fected the ICAM-1 expression by latexbeads phagocytizing Mis.

DISCUSSIONThe present study indicated that stimula-

tion of murine peritoneal M(I)s with M. lep-rae caused an upregulallon of ICAM-1 ex-pression by the M(I)s in the early phase(days 1 to 3) of cultivation. However, theICAM-1 expression thereafter rapidly de-creased, and returned to the normal leveiby day 7. This profile of change in M^ICAM-1 expression during the course ofcultivation seems to be peculiar to M. lep-rae-infected McIs, since such marked de-crease in the ICAM-1 expression at day 7was not observed for M(I)s stimulated dueto phagocytosis of latex beads (Fig. 6). Themiddle phase (day 7) downregulation of

ICAM-1 expression appeared to be due inpari to a defect of M(I)-derived TNF-a hav-ing upregulating effects on M(I) ICAM-1expression ( 14 ''"), since supplementary ad-dition of TNF-a prevented the middle (day7) phase reduction of ICAM-1 expression(Fig. 1).

This concept is supported by the follow-ing findings obtained in separate experi-menta. First, exogenous TNF-a increasedthe ICAM-1 expression on uninfected M(I)s(Fig. 2). Second, anti-TNF-a Ab abolishedthe M. leprae infection-induced increase inMttt expression of TNF-a in the early phase(day 3) of cultivation (Fig. 3). However, un-der the present experimental condition, wedetected only low leveis of TNF-a releasedfrom M. leprae-infected M(Ds roto the cul-ture medium (The Table). This may meanthe individual M. leprae-infected MDs pro-duced TNF-a mainly in a membrane-boundform with a MW of 26 kD but not in a re-leasing form with a MW of 17 kD ( 13 27),and that such membrane-bound TNF-amolecules transfer their stimulatory signalsto other M(I)s through cell-to-cell contact.Alternatively, it is also possible that MO-derived endogenous TNF-a could inducethe ICAM-1 upregulation of M. leprae-in-fected Mc1s even at very low concentra-tions.

It appears that an immunoregulatory cy-tokine, TGF-(3 ( 31 ) also plays an importantrole in the ICAM-1 downregulation of theM(I)s, since anti-TGF-(3 Ab overcame themiddle phase (day 7) downregulation ofthe ICAM-1 expression by M(I)s infectedor stimulated with M. leprae (Figs. 4 and5). In this context, anti-TGF-(3 Ab failed toincrease the TNF-a released by M. leprae-infected MDs (The Table). Therefore,TGF-[3-mediated downregulation of Md)ICAM-1 expression is not attributable tothe reduction of M(1 TNF-a-producingability due to TGF43 action. Moreover, al-though a significam levei of the TGF-(3 pro-duction by M. leprae-infected MOS was ob-served at day 3, ICAM-1 downregulationwas not observed at this time point (TheTable versus Figs. 1 and 4). This suggeststhat TGF-(3 could not effectively suppressthe ICAM-1 expression by M(I)s whichwere actively producing ICAM-1 upregu-lating factors, including TNF-a. Moreover,it seems that M(1) factors other than TGF43

42^ International Journal o/ Le'prosy^ 1999

also contribute to the middle-to-late phasedownregulation of the ICAM-1 expression.For instance, the antiinflammatory cytokineinterleukin I() (IL-10) (`- ' ") and, also,prostaglandin E, may play some roles in theICAM-1 downregulation. Indeed, our re-cent study on M. ariwn complex-infectedM(Ds revealed that the middle (day 7) tolate (day 14) phase ICAM-1 expression byM. ariunt complex-infected M(1)s is po-tently suppressed by these M(1) factors in anautocrine fashion (Tomioka, et al.: manu-script submitted). Thus, it is of interest toexamine the participation of these and otherunknown ICAM-1 downregulating factorsin the late-phase reduction of Mc1 ICAM-1expression, with special reference to theircollaborating effects.

It is somewhat enignlatic to note thatboth exogenous TNF-a and anti-TGF-í3Abs fully overcame the middle phase (day7) downregulation of ICAM- I expressionon M. /eprae-infected MdDs (Figs. 1 and 4).Although the precise reason for this situa-tion is unknown, it may be explained asfollows. First, it is thought that by receiv-ing high leveis of activating signals ofTNF-a Meps might become resistant toICAM-1 expression downregulatòry ef-fects of TGF43. If this was the case, the ad-dition of exogenous TNF-a in excessamounts might cause the effective recoveryof M(I) ICAM-1 expression in the middlephase (day 7) by protecting M. leprae-in-fected MDs from TGF-f3-mediated ICAM-1downregulating effects despite the fact thatsignificantly high amounts of TGF43 wereproduced and released by the M 1 s finto theculture medium at the same time point. Sec-ond, it is also thought that TNF-a-mediatedactivating signals suppress TGF-(3 produc-ing ability of M. /eprae-infected MeDs, sinceanti-TNF-a Ab was found to decrease McpTGF-(3 production (The Table). Therefore,TNF-a treatment of M. /eprae-infectedMeDs might antagonize TGF43-mediatedICAM-1 downregulation by reducing theproduction of Md) TGF-(3, causing a recov-ery of MD ICAM-1 expression.

Notably, the profiles of ICAM- I expres-sion in MeDs stimulated with living Al. lep-rae organisms were essentially the same asthose in Mis stimulated with heat-killedbacilli (Figs. 4 and 5). This indicates thatheat-stable components of M. leprae play

roles not only ias the early-phase upregula-tion ICAM- 1 expression but also in themiddle-to-Iate phase downregulation ofICAM-1 expression. First, lipoarabino-niannan (LAM) is une of lhe major candi-dates for the ICAM-1-modulating compo-nents of M. leprae, since mycobacterialLAM is capahle of upregulating M(IICAM-1 expression (' '). It lias been re-ported that arahinosylated LAM of M. tu-berc alosi.c induces Md> production of pra-inllammatory cytokines, including TNF-aand IL- 1 ( 1 ' which are lhe major me-diators of ICAM-I upregu lation ( 11 14 ").Nevertheless, ias the present study, M. iep-rae-infected Md)s did not produce de-tectable amounts of TNF-a in the culturemedium (The Table). In this context,Adams, et al. ( 1 ) reported that LAM fromM. leprae having niannose capping did notinduce TNF-a production by M(I)s, as inthe case of mannosylated LAM from viru-lent M. íuberculosis Erdman strain(LAM A ) (`'). However, they also foundthat mannosylated LAM A could triggerM(I) antimicrobial activity and elevatedleveis of Md> nitric oxide production, andthat this triggering capacity of LAMwas abrogated by an anti-TNF-a Ab, inc^t -cating that TNF-a is involved in the induc-tion of Md> activation by LAM I . No-tably, LAM did so without miucingelevated leveis of TNF-a in the supernatantmedium. As discussed above, it thus ap-pears that the LAM of M. leprae triggeredMchs to produce TNF-a in a membrane-bound forni (MW = 26 kD) ("), and suchmembrane-bound TNF-a molecules trans-ferred their stiniulatory signals from oneM(I) to another Mch through cell-to-cellcontact. Indeed, it lias been demonstratedthat activated M(I)s could kill a TNF-a-sensitive tumor cell line in the absence ofany demonstrable TNF-a in the culture su-pernatants and this killing was attributed tothe presence of cell surface-associatedTNF-a ("). In addition, segments of M. lep-rae cell-wall skeleton, such as mycolyl-ara-binogalactan-peptidoglycan complex, pro-tein-peptidoglycan complex, and muramyldipeptide, may play roles in the upregula-tion of ICAM-1 expression by M. leprae-infected MclDs since these substances areknown to trigger elevated production ofTNF-a (').

67, 1^Shimizu, et al.: ICAM-1 Regulation^ 43

Second, mycobacterial LAM also in-duces Iv1(1) production of immunoregulatorycytokines, including TGF-f3 and IL-10 ( 7 . 21which have ICAM-1 downregulating func-tions ( 23 . 32), suggesting the roles of LAM inthe ICAM-1 downregulation of M. leprae-infected Mis in the middle-to-late phase ofcultivation. In addition, it is also likely thatother cellular components, such as heat-sta-ble proteins, play roles in the ICAM-1 mod-ulation since purified protein derivative(PPD) of M. tuberculosis is reported to up-regulate MO production of TNF-a, IL-1,and, moreover, TGF-f3 (I". 2 ` ) ).

Moncada, et al. ( 17 ) reported that the epi-dermis of leprosy patients with inflamma-tory skin diseases exhibited an increasedICAM-1 expression. They also found a lackof ICAM-1 expression in the epidermis, es-pecially karatinocytes (cells with Mc13-likefunctions), of lepromatous patients. It is ofinterest to know whether or not the defec-tive expression of ICAM-1 in the lesionaiepidermis of lepromatous leprosy patientsis mediated by TGF-(3. In this context, wepreviously found sequential increase of tis-sue leveis of TNF-a, IL-10 and TGF-(3 inspleens during the course of M. arriam com-plex infection in mice ("). That is, the in-crease in the TNF-a levei was observedaround week 2 after infection, followed byan elevated levei of IL-10 during weeks 2 to4, and the increase of TGF-f3 levei was sub-sequently observed after week 4 for at leastup to week 8. Further studies are currentlyunderway to examine the roles of ICAM-1downregulating cytokines, including TGF-(3and IL-10 as mediators of defective ICAM-1 expression in skin lesions of lepromatousleprosy patients.

SUMMARYProfiles of intercellular adhesion mole-

cule-1 (ICAM-1) expression on murineperitoneal macrophages (Mcl)s) infectedwith Mvcobacterium leprae during cultiva-tion were examined with special referenceto the regulatory effects of tumor necrosisfactor-alpha (TNF-a) and transforminggrowth factor-beta (TGF43). When Miswere infected with M. leprae or stimulatedwith heat-killed M. leprae at day 0, theirICAM-1 expression, measured in terms ofthe ratio of Mis positively stained withanti-ICAM-1 antibody (Ab), rapidly in-

creased, peaking during days 1 to 3 andthereafter fell, returning to the normal leveiby day 7. The addition of TNF-a or anti-TGF-(3 Ab inhibited the middle phase (day7) downregulation of ICAM-1 expres-sion, although the late-phase (day 14)downregulation of ICAM-1 was not pre-vented by them. M. leprae-infected Misreleased small amounts of TNF-a and sig-nificant amounts of TGF-f3 roto the culturemedium. This may indicate that M. leprae-infected Mis produced the majority ofTNF-a in a membrane-bound form. Alter-natively, endogenous TNF-a might upregu-late M(I) ICAM-1 expression even at verylow concentrations. In any case, these find-ings indicate the central roles of TNF-a andTGF (3 in the early phase upregulation andthe middle-to-late phase downregulation,respectively, of ICAM-1 expression by M./eprae-infected M1s.

RESUMENSe examinei el efecto regulatorio dcl factor de

necrosis minorai alfa (TNFa) y del factor de crec-imiento transformante beta (TGF(3) sobre la expresiónde la molécula de adhesión intercelular ICAM- I enmacrófagos peritoneales murinos infectados con My-cobac ter ium leprae.

Los macrófagos infectados con Al. leprae o estim-ulados cons M. lepras muerto por calor en el dia O,mostraron un rápido incrmeneto en la expresión deICAM-1 que fue máxima entre los dias 1-3 y luegodecayó, retornando a su nivel normal havia el dia 7. Laadición de TNFa o de un anticuerpo anti-TGF-(3 in-hibió la fase media (día 7) de la expresión de ICAM-1pero no modificó la fase tardia (día 14). Los macrófa-gos infectados con M. leprae liberaron pequefias canti-dades de TNFa y cantidades significativas de TGF-(3en el medio de cultivo. Esto puede indicar que losmacrófagos infectados por M. leprae produjeron lamayor parte dei TNFa en una forma unida a mem-branas.

Alternativamente, el TNFa endógeno podría regu-lar la expresión de ICAM-1 aun a muy hajas concen-traciones. En cualquier caso, estos hallazgos indicanque el TNF-a y el TGF-f3 participar en la regulaciónde la expresión de ICAM en ias fases temprana ytardia, respectivamente, de la infección por M. leprae.

RÉSUMÉOn a examiné le mode d'expression des molécules

d'adhésion intercellulaire de type 1 (ICAM-l) chezdez macrophages (M I s) infectés iu nitro par Mv-cubacterium leprae, en tenant compte en particulier dela regulation d'expression exercée par le tumor necro-sis factor-alpha (TNF-a) et le transforming growthfactor-beta (TGF-(3).

44^ Inte iational foi trila! nf LeprosV^ 1999

Lorsquc des MdDs étaient infectés par M. /eprae ouhien siumilés par des Al. leprae tuécs par Ia ch a leur,I'expression dc ICAM-1, mcsuréc en termo de propor-tion de Mis marquês par des anticorps (Ac) anti-ICAM-1, augntcntait rlpidement, présentait un pio en-tre le premicr et Ic troisiêmc jour après I'infcclion etdiminuait cnsuitc, pour rcjoindre Ic nivcau basal d'cx-pression vcrs lc scptièmc jour. L'addition de TNF-a oud'anticorps anti-TGF-¡3 inhibait la phasc intcrmédiairc(scptièmc jour) dc rctour au nivcau basal d'cxpressiond'ICAM-1 par les M4rs. Cependant, le rctour tardif(quatorzièmc jour) au nivcau dc base d'cxpression deICAM-1 n'a pu ctrc prévenu par ces moléculas. LesMchs infectés par M. leprae ont relargués dans le mi-licu dc cultue unc pctitc quantité dc TNF-a et uncquantité plus importante dc TGF-13.

Cola pourrait suggérer que soit les M(1)s infectéspar M. leprae produiraient la majorité de Ieur TNF-adans une forme assossiée aux membranas, sois que lcTNF-a endogènc pourrait stinuller l'expression deICAM-1 même à três pctitc conccntration chcz lesmacrophages. En tout état de cause, ces trouvai1lessuggèrent un rôlc central de TNF-a et dc TGF-f3, dans last1nwlation précoce et dans le rctour au nivcau basal enphasc moycnnc et tardive, respectivement, de I'cxpres-sion de ICAM-1 par les McDs infectés par M. leprae.

Acknowledgment. This study was supported inpart hy a Grant-in-Aid for Scientific Rescarch from theMinistry of Education, Scicncc, and Culturc of Japan,and hy the United States-Japan Cooperativo, MedicalScicncc Program (Tuhcrculosis and Lcprosy Scction).

REFERENCES1. ADAMs, L. B., Fl1KUTONII, Y. and KRAUENeuut., J.

L. Regulation of nu?rine macrophage cffectorfunctions hy lipoarahinomannan from mycobacte-rial strains with different degrces of virulence. In-fect. Immun. 61 (1993) 4173-4181.

2. BARNES, P. F., CIt.vrrERJEE, D., ABRAMS, J. S., Lu,S., WANG, E., YANL\NiuRA, M., BRENNAN, P. J. andMODLIN, R. L. Cytokinc production induccd hyM)robaeterimn [o6e,rulosis lipoarahinomannan;rclationship to chcmical structure. J. Immunol.149 (1992) 541-547.

3. BARNES, P. F., CIIATTERJV.E, D., BRENNAN, P. J.,REA, T. H. and MODLIN, R. L. Tumor nccrosisfactor production in patients with lcprosy. Infect.Immun. 60 (1992) 1441-1446.

4. BLRMuoEz, L. E. Production of transforminggrowth factor-o by Mvcobacteriu u avium-in-

fectcd human macrophages is associatcd with un-responsivencss to IFN-y. J. Immunol. 150 (1993)1838-1845.

5. BRADBLRY, M. G. and MORENO, C. Effect oflipoarahinomannan and mycobacteria on tumornecrosis factor production by different populationsof rumine macrophages. Clin. Exp. 1111[1111'101. 94(1993) 57-63.

6. ClI,\FrratJI:E:, D., RonELRTS, A. D., Lowraa., K.,BRENNAN, P. J. and ORME, 1. M. Structural hasisof capacity of lipoarahinomannan to induce sacro-

tios of tumor nccrosis factor. Infect. Ilu mm. 60(1992) 1249-1253.

7. D:\Hi. , K. E., SuutArsuctit, II., HANnt:roN, B. D.,Ei.t.NE:R, J. J and Toosi, Z. Selectivo induction oftransforming growth factor ¡3 In human monocyteshy lipoarabinomannau of Mtrolacteriunm tubercu-

Insis. 1nica. Immun. 64 (1996) 399-405.8. Dy, Wv,vL M,vLEFYT, R., YssuL, H., RoNc,vRol.o,

M.-G., Suis, H. and Vinis, J. E. lntcrleukin-10.Curr. Opin. Immunol. 4 (1992) 314-320.

9. DI :CE:I[R, T., LouoANN-M,vrrHrs, M.-L. and Glr-t oEtD, G. E. Ccll-associatcd tumor nccrosis factor(TNF) as a killing mechanism of activated cyto-toxic macrophages. J. Im nwnol. 138 (1987)957-962.

10. DUSTIN, M. L., RorIII.IiN, R., BIIAN, A. K.,DINARELLO, C. A. and SPRING a , T. A. Inductionby IL 1 and intcrfcron-y: tissuc distribution, hio-chcmistry, and function of a natural adhcrcnccmoleculc (ICAM-1). J. Immunol. 137 (1986)245-254.

11. FRANZRLAU, S. G. Drug susceptihility testing ofAlvcnbacterium leprae in the BACTEC 460 Sys-tem. Antimicrob. Agents Chemother. 33 (1989)2115-2117.

12. HAGERTY, D. T. Interccllular adhesion molecule-1is necessary but not sufficicnt to activatc CD4' Tcens. J. Imnwnol. 156 (1996) 3652-3659.

13. KRU:GI.uR, M., PEiREZ, C., DE:F,\Y, K., ALBERT, I.and Lu, S. D. A novel forni, of TNF/cachectin is asurface cytotoxic transmcmhranc protcin: ramifi-cations for the complcx physiology of TNF. Cell53 (1988) 45-53.

14. Lovu:z R,sMiREZ, G. M., RONI, W. N., Ourou, C.,T,sLnoT, A., MARTINIUI:, F., CRONsTE.N, B. andREIt3MAN, J. Mvcobacterium tuberculosis altersexpression of adhesion molecules on monocyticcens. Infect. Immun. 62 (1994) 2515-2520.

15. M,\RI_iN, S. D. and SPRINGER, T. A. Purificd inter-cellular adhesion molecule-1 (ICAM-1) is a ligandfor lymphocyte function-associatcd antigen-1(LFA-1). Ccll 51 (1987) 813-819.

16. MAw, W. W., TON110K,v, H., S,vTo, K. and S,vrro, H.The expression of ICAM-1 on macrophages stim-ulated with Mvcobacterium avio'? complex and itscontrol by some regulatory cytokines. Kekkaku 71(1996) 561-567.

17. MONCADA, B., TORRES-ALVAREZ, M. B., GONZALEZ-

ANL\Ro, R., HUENTES-Artll\CADA, C., BARANDA, L.,DELGADO, S. P. and GARCIA, R. Lack of expres-sion of intercellular adhesion molecules ICAM-1in Icpromatous leprosy patients. Int. J. Lepr. 61(1993) 581-585.

18. MORENO, C., TAVERNE, J., MEHLERT, A., BATE, C.A., BREALEY, R. J., MEAGEiR, A., RooK, G. A. andPI.,vYF,vnt, J. H. Lipoarabinomannan from Mv-cobacterium tuberculosi.c induccd the production

67, 1^ S/rimizu, et al.: ICAM-I Regu/crtiou^ 45

of tumor necrosis factor from human and nu ninemacrophagcs. Clin. Exp. Immunol. 76 (1986)240-245.

19. Multlt,w, P. J., WANG, L., ONUIIZYK, C., TEPPiaZ, R.1. and Youx ;, R. A. T cell-derived 1L-10 antago-nizes macrophage lunction in mycobacterial in-fection. J. Immunol. 158 (1997) 315-321.

20. RurnI.t:IN, R., CzAikowSKI, M., O'NI a .L, M. M.,N1 \RLIN, S. D., M,VNoi.II, E. and MEIZt.uzzt, V. J.1nduction of intercellular adhesion molecule 1 onprimary and continuous ccli tines by pro-infl mlma-tory cytokines. J. Immunol. 141 (1988) 1665-1669.

21. SAno, H., Towoi:A, H. and SAIO, K. Purilicationof M. Ieprae with special reference to the effectsof purified M. Ieprae vaccines on host macrophageccll functions. Jpn. J. Lepr. 56 (1987) 101-109.

22. SEvi.t:IZ, I., APPEL, M., DIìVlSBAGGUI ïr, J.-P.,LLGRAND, P. and BAIZIZXvr, G. Modulation of nitricoxide production in RAW 264.7 cens by transfonu-ing growth factor-beta and interleukin-10: differen-tial effects on free and encapsulated immunomodu-lator. J. Leukoc. Biol. 62 (1997) 374-380.

23. $HKIKANT, P., WEBER, E., Ju.I.iNG, T. and BEN-VENISTE, E. N. Intercellular adhesion molecule-1gene expression by glial cens. Differcntial mecha-nisms of inhibition by IL-10 and IL-6. J. Immunol.155 (1995) 1489-1501.

24. SPIZINGER, T. A. Adhesion receptors of the nu-mune system. Nature 346 (1990) 425-434.

25. SULLIVAN, L., SANO, S., PuZnu:z, C., SALGAM, P.,MUELL1:IZ, C., HOI MAN, F., UYEMURA, K., REA, T.H., BLoout, B. R. and Mout.IN, R. L. Expressionof adhesion molecules in leprosy lesions. Infect.Immun. 59 (1991) 4154-4160.

26. TouIOKA, H., SAITO, H. and SATO, K. Evaluationof BACTEC 460 TB System for measurement ofia litro anti-Mvcobacterium Icprae activity of var-ious antimicrobials. Jpn. J. Lepr. 61 (1992)157-164.

27. ToNtoK,v, H., SATO, K., MAM:, W. W and SAITO, H.The role of tumor necrosis factor, interferon-y,transforming growth factor-o, and nitric oxide inthe expression of in ununosuppressive functions ofsplenic macrophages induced by Mvcobacteriumarium complex infection. J. Leukoc. Biol. 58(1995) 704-712.

28. To\ nm :A, H., SATo, K., Sltuuzu . T., SANO, C.,AKAKI, T., SArio, H., Ftuii, K. and HIDAKA, T. Ef-fects of benzoxazinorifamycin KRM-1648 on cy-tokine production at sites of Myrobacteriontanum complex infection induced in mice. Antimi-crob. Agents Chemother. 41 (1997) 357-362.

29. Toossi, Z., YouNG, T. G., AVERIt.L, L. E., HANtu.-Too, B. D., SnIRArsUGIII, H. and ELLNEIZ, J. J. In-duction of transforming growth factor (3l by puri-fied protein derivativa of Mvrobacterium tubercu-Iosis. Infect. Immun. 63 (1995) 224-228.

30. VAN SEVENTER, G. A., Suiutzu, Y., HoRGAN, K. J.and SiiAw, S. The LFA-1 ligand ICAM-1 pro-vides an important costinudatory signal for T cellreceptor-mediated activation of resting T cens. J.Immunol. 144 (1990) 4579-4586.

31. WAUi., S. M. Transforming growth factor beta(TGF

-13 ) in intlammation: a cause and a cure. J.

Clin. Immunol. 12 (1992) 61-73.32. WILLLNts, F., MARCHANT, A., DELVILLE, J.-P., Gim-

ARO, C., DELAUX, A., Vt t.u, T., DE BORI:, M. andGoI_DyIAX, M. Interlcukin-10 inhibits B7 and in-tercellular adhesion molecule-1 expression on hu-man monocytcs. Eur. J. Immunol. 24 (1994)1007-1009.

33. XLV), B.-G., ZIANG, G.-X., MA, C.-G. and LINK,H. Transforming growth factor-beta 1 (TGF-(31)-mediated inhibition of glial ccll proliferation anddownregulation of intercellular adhesion mole-cule-1 (ICAM-1) are interrupted by interferon-gamma (IFN-y). Clin. Exp. Immunol. 103 (1996)475-481.