Research Paper Mediators of Inflammation, 7, 269–274 (1998)

CA Corresponding Author Fax : (+1) 734 764 5135 Email: nlukacs@umich.edu THE cell-to-ce ll interactions durin g chron ic inflam m atory diseases likely con tribute to leukocyte accum ulation leading to increas ed path ology and organ dys function . In particular, the re is a paucity of in form ation relating to th e m aintenance of ch ronic fibrotic diseases . Usin g a lung fibroblas t line and enriched m onocyte populations , we have investigated the activational events wh ich con tribute to th e production of tw o C-C chem okines , m acrophage in flam m atory prote in-1 alpha (MIP-1 a ) and m onocyte chem oattractan t protein-1 (MCP-1), during fibroblastm onocyte in teractions. Neither the fibroblast ce ll line (16 lu) nor iso lated m onocytes alone produced s ign ifican t levels of MIP-1 a or MCP-1. However, when isolated m onocytes w ere layered on to 16 lu fibroblast m onolayers a s ignificant increase in MIP-1 a and MCP1 production w as observed. Th e use of fix ed ce ll populations in dicated th at th e MIP-1 a w as derived from m onocytes and MCP-1 from both cell populations. To ex am ine th e m olecules wh ich w ere required for chem okine production durin g th e interaction, specific antibodie s w ere used in th e co-cultures . Blockin g b 3-integrin interactions s ign ifican tly in hibited MIP-1 a production . In con trast, beta-in tegrin in te ractions had no e ffect on the MCP-1 production, w hile, neutralization of TNF s ignifican tly decreased MCP-1 production dur in g the co-culture . Th es e data in dicate th at fibroblast–m onocyte in te ractions in duce chem okine production th rough differen t m echan ism s and a com bination of th ese response s m ay contribute to th e m aintenance of th e m ononuclear ce ll accum ulation durin g dis ease progress ion.


Introduction
The ac cumulation of leukoc ytes during infe ctious diseases is p aramount for host prote c tion. How ever, persiste nt leukoc yte influx during chronic inflammatory disease s appe ars to be a driving forc e be hind tissue pathology. 1 ,2 In partic ular, the c onsiste nt re c ruitment of monoc ytes to fibrotic tissue has be en identified as a hallmark of dise ase p rogre ssion. 3 ,4 The me chanisms w hich re gulate le ukocyte re cruitme nt are quite c omplex and are initiat ed by adhesion molecules on the endothe lial surfac e. 5 ,6 Onc e adhere d to the endothelial surface the c ells c an then migrate into the inflamed tissue follow ing chemotactic gradie nts. The chemokines appear to mediate the selec tive re c ruitment of c ell pop ulations during inflammatory dise ase s. 7 ,8 Chemokines are divide d into tw o distinct familie s by sequenc e similarity and by function. The C-x -C family (IL-8 family or alpha che mokines) are primarily chemotactic for ne utrophils, w here as the C-C che mokine family (MCP-1 family or be ta chemokines) are primarily chemotac tic for monoc ytes and lymphocyte s. How ever, the se broad divisions are now breaking dow n and it appears that both classe s of chemokines are important mediators in both ac ute and chronic inflammation. One of the more pote nt C-C family chemokine s, mac rophage inflammatory prote in-1 alpha (MIP-1a ), is made p rimarily by leukoc yte populations and not by stromal or struc tural ce lls. In c ontrast, a sec ond me mber of the family, monocyte chemoattra ctant prote in-1 (MCP-1), can be made by both immune and non-immune ce lls.
Once the cells have move d into the tissue, the y continue to interact by ce ll-to-c ell contac t w ith struc tural ce lls w ithin the tissue or organ. In earlie r studies, it has be e n demonstrate d that monocyte / macrophage inte raction w ith various c ell populations (endothe lial ce lls, synovial fibroblasts ) can drive the production of che mokine s and may possibly be a me chanism for maintainin g the pers is te nt leukoc yte influx obse rved during chronic diseas es. 9 -1 1 The me chanisms w hich gove rn the se inte ractions appear to be re gulated by adhesive inte ractions be tw een the leukoc yte and the tissue cells. These me chanisms, although intuitive , may be differe ntial de pending on the state of the inflammatory re sponse , the cells involved, the che mokine s p roduc ed, and the tissue loc ation of the re sponse.
In the pre sent studie s, the re sults demonstrate that tw o C-C family chemokines, MIP-1a and MCP-1, are produce d during monocyte -fibroblast interac tions. The tw o che mokines, how ever, are differe ntially re gulated during this re sponse : MIP-1a is made by only the monoc yte s and depe nde nt upon b 3-inte grin ligation, w he re as MCP-1 is made by both cell populations and the me chanism w as not depe nde nt upon be ta-integrins, 1-3 but rather on c ytokine (TNF) activation. These studie s demonstrate nove l mechanisms of re gulation of C-C chemokine production.

Mononuclear cell isolation
Pe riphe ral blood w as draw n into a hep arinize d syringe from healthy voluntee rs, diluted 1:1 in normal saline, and mononuc le ar c ells sep arated by de nsity gradie nt c entrifugation. The re covere d c ells w ere w ashe d thre e time s w ith RPMI 1640. The PBMs w ere then layere d onto a de nsity gradie nt (1.068 g/ml) for the enrichme nt of monocyte s (Atlanta Biologic als, Atlanta, GA). The isolate d c ells w e re then w as he d, cytospun onto a glass slide, stained w ith Diff-Quik (Bax te r, Mc Gaw, IL) and diffe re ntially counted. The purity of the monocyte s from the gradient w ere cons is te ntly be tw e en 75 and 80% monocyte s w ith the re mainder lymphoc ytes.

16lu fibroblast cultures
The transformed fibroblast c ell line , 16 lu, w as obtain ed from ATCC (CCL 204) and cultured as re quired in Eagle's minimum essential me dium w ith non-essential amino ac ids and Earle's BSS w ith 10% fetal calf serum. The ce lls w ere grow n to ne ar confluent monolaye rs in six -w ell tissue c ulture plates and fresh media ap plie d prior to utilizing them in the assays.

Blocking cellular interactions
To demonstrate that ce ll-to-c ell interac tions w ere re quired for produc tion of the chemokines, fix e d ce ll populations w e re use d (4% paraformalde hyde for 5 min). Subse que ntly, bloc king antib odies to be tainte grins (Chemicon, Teme cula, CA) and/or adhesion molecules (R&D Syste ms, MN) w e re use d to block adhesion of the tw o c ell pop ulations. The antib odies w ere used at a c once ntration of 5 m g/ml. In addition, blocking polyclonal antib odie s to IL-1 and TNF w ere also use d at a 1: 200 dilution in culture to ex amine inflammatory cytokine ne tw orks.

Statistical analysis
Data are ex pre ssed as means ± SEM. Data that appeared statistic ally signific ant w e re c ompare d by ANOVA for comparing the me ans of multiple groups, and c onside re d significant if P values w ere less than 0.05.

Results
Fibroblast: monocyte coculture induces C-C family chemokines in an adherence-dependent mechanism To de te rmine w hether monoc yte interac tions w ith fibroblast induce d an ac tivational event, enriche d monocyte s w e re layere d onto 16 lu fibroblast monolayers and c ulture d for 24 h at 37°C. The data demonstrate s that ne ither monocyte s nor fibroblasts by themselves p roduc ed substantial levels of MIP-1a ( Fig. 1) or MCP-1 (Fig. 2). How ever, w hen the tw o ce ll populations w ere c ulture d together a syne rgis tic inc re as e in both chemokines w as obse rved. In contrast, e nriched lymphoc yte p opulations (~80%) demonstrated little inc re ase in the tw o chemokine s w hen added to the fibroblasts (data not show n). To dete rmine w hich c ell pop ulations w ere re sponsible for the che mokine production, one of the c ell populations w as fix ed w ith 4% paraformaldehyde prior to the coc ulture proce dure. MIP-1a production w as observe d only w hen the fibroblasts w ere fix ed and monocyte s w ere not, indic ating that only the monocytes w e re re sponsible for the MIP-1a produc tion (Fig. 1), follow ing pre vious studies. In contrast, Antibodies (5 ug/ml) * inc re as ed MCP-1 p roduction w as also obse rved only w hen fibroblast population w as fix ed, indicating that viable monoc ytes ne eded to be p re se nt to inc ude MCP-1 production during the ce ll-to-c ell interaction (Fig. 2). The se studies e stablishe d that an activational e ve nt occ urred follow ing interaction be tw e en the tw o c ell populations.

Differential regulation of C-C chemokines during monocyte-fibroblast interactions
To dete rmine the me chanism of chemokine production during c ell-to-cell inte raction, antib odies specific for inflammatory cytokines and adhe sion molec ule s w ere utilize d. It has pre viously be e n demonstrate d that TNF and IL-1 are strong induc ers of chemokine production in multiple ce ll type s. It w as intere sting that e ve n though fibroblasts are know n to p roduc e MCP-1 and likely the main source of MCP-1, no inc re as e w as obse rved w he n the fix e d monocytes w ere layere d onto them. When antib odie s to TNF and IL-1 w ere added into the c oc ulture no de cre ase in MIP-1a w as observe d (Fig. 3). In contrast, w he n MCP-1 production w as ex amine d, tre atment of the coc ultures w ith anti-TNF significantly de cre ase d the p roduc tion of MCP-1, w he re as anti-IL-1 only slightly inhibited (Fig. 4). These latter studies indic ate a differe ntial re gulation of the se tw o C-C family che mokines. Nex t, since the che mokine produc tion w as differentially inhibited by fix ation of one or the othe r ce ll populations or by separation by transw ells, the role of adhesion mole cules w as ex amined. Blocking antibodie s to be ta-inte grins (b 1,2,3) or adhesion molecules (ICAM-1, VCAM-1) w ere adde d to the tw o ce ll populations (5 m g/ml) for 15 min p rior to combining them. The data indicate d that only antibodies to b 3-integrins signific antly inhibite d MIP-1a , w hile antibodie s to othe r b -integrins (b 1, b 2) or adhesion molecules (ICAM-1, VCAM-1) had no significant affe ct on MIP-1a produc tion (Fig. 5). In contrast, none of the antibodies to be ta-inte grins or adhesion mole cules (ICAM-1, VCAM-1) inhibited MCP-1 (Fig. 6). Inte restingly, antib odie s to b 1-integrins appe are d to inc re as e the production of MCP-1. Altogether, the se data suggest that e ve n though the tw o C-C family che mokines are produc ed during the adhe sion e ve nt the y are diffe re ntially re gulated. MIP-1a appe ars to be depende nt upon adhesion-me diate d pathw ays, w here as MCP-1 appears to re quire cytokine signaling pathw ays for its produc tion. Thus, e ven though these tw o C-C chemokine s have ove rlapping functions in v itro , the y app ear to be diffe re ntially induce d during the ce ll-to-c ell interactions .

Discussion
The pers iste nt production of chemokines during an inflammatory re sponse is like ly re quire d to maintain the c onstant influx of leukocyte s w hich ac companies the development of chronic disease. 1 3 -15 The mechanism w hich w as ex amined in this study w as w hether the c ell-to-ce ll interaction of monoc ytes w ith fibroblasts w as sufficie nt to drive the production of che mokines. The produc tion of C-C family chemokine s, MIP-1a and MCP-1, may be one of the me chanisms w hich maintain chronic and/or fibrotic lesions . Pre vious studie s in fibrotic human dise as es have indicate d that leukocyte acc umulation at the site of fibrotic e pisode s is ne c essary to initiate , maintain , and progress the pathological manifestations w ithin the affec te d tissue or organ. The produc tion of the che mokines in the pre sent study w as depe ndent on the interac tion of the tw o cell populations, as individual ce ll p opulations produce d re lative ly little or no chemokine . We have also ide ntified that che mokine produc tion w as differe ntially re gulate d. MCP-1 appeared to be partially depe nde nt upon monocyte -de rive d inflammatory cytokine production, TNF, and not on ce ll adhesion molec ule s. In contrast, MIP-1a w as induc ed by cell-to-ce ll interac tions depende nt up on b 3-integrin ligation. Since b 3-inte grins have be e n show n to bind to matrix protein compone nts, this inte rac tion w ould be very ap propriate w ithin fibrotic le sions to maintain the chemokine (MIP-1a ) production. The se data be gin to define differe nc es in ce llular interac tion pathw ays w hich may allow mainte nance and progre ssion of chronic diseases. Previous studies have de monstrated chemokine production during c ell-to-ce ll interactions , 9,1 0 but the diffe re ntial re gulation of the chemokine s w ithin the se studie s is striking. This differe nc e of re gulation may lie on the source of the chemokines. MIP-1a ap peare d to be de rive d primarily from the monocyte population, w here as MCP-1 c an be e lic ite d from both the fibroblast and monoc yte populations. Altoge the r, these re sults may indicate w hy it may be so diffic ult to modulate diseas e p henotype s in chronic ailme nts e ve n in the abse nce of any apparent inc iting agent.
FIG. 5. Adhesion pathways mediated production of MIP-1a during fibroblast-monocyte interaction. Monoclonal antibodies to specific adhesion molecules or integrins were added to fibroblast-monocyte co-culture at time 0 at a concentration of 5m g/ml. After 24h of co-culture the culture supernatant was harvested and MIP-1a was measured using a specific ELISA. Data represents mean ± SE of three repeat experiments. Significant decreases in MIP-1a were observed only in cultures treated with anti-b 3-integrin. *P<0.05.
FIG. 6. Blockade of adhesion pathways do not alter production of MCP-1 production during fibroblast-monocyte interaction. Monoclonal antibodies to specific adhesion molecules or integrins were added to fibroblast-monocyte co-culture at time 0 at a concentration of 5m g/ml. After 24 h of co-culture the culture supernatant was harvested and MCP-1 was measured using a specific ELISA. Data represents mean ± SE of three repeat experiments. *P<0.05.
The induction of MIP-1a production by c ell adhesion events has be e n pre viously re ported by our laboratorie s. How e ve r, it app ears that the me chanism is differe nt depe nding upon the c ell-type w hich the monocyte binds. In a pre vious study e ndothelial cells w ere used and the me chanism w as a b 2-inte grin/ ICAM-1-mediated mechanism. 9 In the pre sent study, fibroblasts w ere utilize d and it app eare d that b 3-inte grins playe d a more important role during the interaction for MIP-1a production. The re ason for this differe nc e may lie in not only the cell type, but w here the cells are normally locate d. Endothelial c ells w hich me diated MIP-1a produc tion via ICAM-1 interac tions are the initial ce ll type that the monocyte c ontac ts during inflammatory e ve nts prior to migration into the inflamed tissue. This initial ac tivational e vent may se t up the induc tion of chemokines by monoc ytes for production at site s of inflammation. Onc e at the inflamed area, stromal cells, such as fibroblasts, c ould then maintain the activate d state of the monocyte/ macrophage for c ontinue d chemokine production via ce ll-to-c ell interaction. This latte r mechanism appears to be induced by b 3-integrin-mediated events. The ligand(s) for this inte grin is primarily matrix prote ins, a produc t that fibroblasts are w e ll suited to produce . 16 ,1 7 The induction of MCP-1 produc tion by TNF and IL-1 has be en de monstrated by many laboratories from multiple cell types. 1 8 -24 Intere stingly, TNF ap pears to be a major factor for inducing and maintainin g fibrotic re sponses, even in the absenc e of an appare nt inciting agent. 25 ,26 In chronic fibrotic dise as es, such as idiopathic fibrosis (IPF) or liver c irrhosis, the pe rsiste nt production of TNF and re cruitme nt of mononucle ar ce lls is maintain ed. [27][28][29] In addition, MCP-1 appears to play a role in chronic fibrotic e vents in animal models of fibrosis. 3 0 Inte re stingly, MCP-1 w as first identifie d as a c ompete nce factor during fibroblast activation and has be e n show n to p artic ipate in c ollagen ge ne activation w hich c an c ontribute to the ove rall p athology w ithin fibrotic le sions. 3 0,3 1 Intere stingly, w hen monocyte s w ere fix e d prior to the addition to fibroblast, no MCP-1 w as induced, thus de picting the ne ed for monocyte -de rive d fac tors (TNF) to induc e MCP-1 production. In c ontrast, MIP-1a w as induce d via dire ct monocyte adhe re nc e to fix e d fibroblasts.
The inc re ased production of MCP-1 via the cellular adhesion follow ing TNF activation may re pre sent a ke y me chanism in progre ssion of chronic fibrotic diseases.
The data in the pre sent study sugge st that the inte rac tion of fibroblasts w ith monoc ytes initiat es che mokine production. The initiation of C-C family che mokines during this adhe sion e ve nt app ears to utilize diffe re nt me chanisms of ac tivation. Ye t, the overall re sult is inc re ased MIP-1a and MCP-1 production w hich may be utilize d to maintain pers iste nt leukoc yte acc umulation and ce llular activation w ith-out other initiatin g cytokines or fore ign p athoge ns. These mechanisms may help to c ontribute to the mainte nanc e of chronic fibrotic dise as es, re sulting in significant pathogenic changes.