Migration of neutrophils through epithelial monolayers

FORUM

References 1 CHOW, R H, VON RUDEN,L and NEHER,E (1992)Natare 356, 60-63 2 ALVAREZDE TOL[DO, G., FERNANDEZ-CHACON,a and FERNANDEZ, J. M (1993) Nature 363, 554-558 3 HEUSER,I E. and REESE,T. S (1973) ] Cell8ml 57, 315--344 4 CECCAREUJ,B., HURLEUT,W P and MAURO,A (1973)/ Cell Biol 57, 499-524 $ CECCAREIII,B, GROHOVAZ, F and HURLBUT,W P (1979) I CellBJol.El, 178-192 6 TORRI-TARELU,F., GROHOVAZ, F., FESCE,R and CECCARELU, B. (1985) ]. Cell Biol. 101, 1386-1399 7 TORRI.TARELU,F, HAIMANN, C_and CECCARELLI,B. (1987) I. NeurocytoL 16, 205-214 8 VALTORTA,F, IAHN, R, FESCE,R, GREENGARD,P and CECCARELLI,B. (1988)/ CellBJol 107, 2717-2727 9 TORRI-TARELLI,F., VILLA,A., VALTORTA,F., DE CAMILLI, P, GREENGARD,P. and CECCARELLI,B. (1990) I. Cell Bml 110,

449--459 10 TORRI-TARELLI,F., BOSSI,M., FESCE,R, GREENGARD,P. and VALTORTA,F (I 992) Neuron 9, 1143-1153 11 FERNANDEZ,J. M, NEHER, E and GOMPERTS,B. D. (1984) Nature 312, 453-455 12 RRECKENRIDGE,L. and ALMERS,W. (1987) Nature 328,

Migration of neutrophils through epithelial

monolayers Jam

es

L M ad ai:a

Acute bacterial infections are a major challenge to epithelial linings ttlat interface indirectly with the external world. Such infections are in part tbugbt by neutrophlls, which phagocytose and destroy pathogens. Neutrophils arrive at the in~ection sites by emigrating from small blood vessels and subsequently interacting with and transmigrating across columnar epithelia. There is substantial inte+est itt defining the mechanisms and fimctional consequences of neutrophil-epithelial interactions. Model systems reveal that vecific molecular events are required for appropriate neutrophii--epithelial interactions a:td, as a result of these interactions, neutrophils may reversibly modtdate diverse epithelial

timctions. t'~

© 1994 ElsevierSciencettd 0962-89241941507O0

814 817 !3 HELM,C A, ISRAELACHVILI,J N and McGUIGGAN, P M. (1992) Biochemistry31, 1794--1805 14 NANAVATI,C., MARKIN,V S, OBERHAUSER,A. F and FERNANDEZ, J. M. (1992) Bmphys_]. 63, 1118-1132 IS OBERHAUSER,A F, MONK, J. R and FERNANDEZ,J M (1992) Biophys J 61,800-809 16 THOMAS,Let oL (1988) Soente 242, 1050-1053 17 SOLLNER,T. et al (1993) Nature 362, 318-324 18 BENNETT,M. K and SCHELLER,R H (1993) Proc NatlAcad. 50 USA90, 2559-2563 19 SCHIAVO,G_ etal (1992) EMBO] 11, 3577-3583 20 SCHEKMAN,R (1992) Curt OpJn CellBiol 4, 587-592 2.1 SCHIAVO,G. etal (1992) Nature 359, 832-835 22 BLASI,J. et aL (1993) Nature 365, 160-163 23 MONCK, J. R and FERNANDEZ,J_M (i 992) I Cell Bml 119, 1395-1404 24 TSE,F. W., IWATA, A. and ALMERS,W (1993) J. Cell Bml 121, 543-552 25 NANAVATI,C and FERNANDEZ,J. M (1993) Saence259,

963-965 26 NEHER,E (1993)Nature 363, 497--498 27 MELDOLESI,J. and CECCARELLI,8 (1991) Phdas Trans Roy Sac Lend. B 296, 55-81

The spaces of the renal tubules, the airways and the alimentary tract are largely lined by a continuous nlonolayer of cohtmnar epithelial cells, These epithelial cells rest on a basement membrane under which vessels el the microclrculation course through a COUlplex matrix. Each epithelial cell is circuntferentlally wrapped at Its apical pole by a complex of junctional interactions with nelghbouringcelk Such Junctions, including the Intercellular ttghl junction, seal the space between cells and thus, In concert w=th the apical membranes of epithelial cells, separate the underlying tissues from the external world, Not surprisingly, such sites are often challenged by pathogens, which evoke inflammatory events geared towards defending this surface. Neutrophils, which are able to phagoeytose and destroy many pathogens by secreting toxic substances into Jae resulting phagosomes, serve as a primary deicnce against bacterial pathogens tn our environm~.nt. In acute inflammatory diseases of the lung, kidney and intestine, defending neutrophlls transnugrate across colunmar epithelial monolayers. What directs neutrophll movement across epithelia? For mechanistic studies of th,s transmlgrahon pro-

cess, several investigators have used model systems comprising cultured monolayers of columnar epithelial cell lines, and human neutrophils purified from peripheral blood I-'4 (Fig, 1). Directed transepithelial neutrophil migration can be stimulated m such TRENDS IN CELL BIOLOGYVOL 4 JANUARY 1994

FORUM

I

model systems in wire by constructing transmonolayer gradients of chemotactic substances recognized by nemrophils. N-Formylated peptides such as f-MetLeu-Phe (fMLP) are use~l for this purpose. Since bacteria N-formylate peptides but eukaryohc cells do not, N-formyl-peptlde gradients radiating from bacterial sites and detected by N-formyl-peptide receptols on the neutrophll surface provide a classic means for neutrophils to track bacteria. However, despite the demonstration that flVILPgradients drive tran~epithelial neutrophii migration in vitro, the special charactenstics of epithelial surfaces compared wRh other hssue compartments may iequire other strategies to induce ~,irected neutrophil movement m rive. First, the lumenal compartments defined by these epithelial surfaces are often not aseptic; for example, high co_ncentrations of bacteria-derived N-formyl-peptldes are normally present in the alimentary tract. How would the trivial increment in N-formylated peptide radiating from a pathogen be differentiated from this vast background N-formylpeptide signal? Second, columnar epithelia, by their nature, restrict diffusmn of peptides across them. For example, we have found that a log higher N-formyl-peptide concentration gradient Is required to move neutrophils across a filter coated with a confluent epithelium, than across a filter itself. Thus this classical system for detecting bacteria may be less effective at epitheiial sites than, say, in a muscle, where peptide diffusion is essentially unrestricted. Indeed, further evidence obtained in vitro now suggests that the signals directing neutrophil movement across epithelia in response to pathogens may be complex. For example, when the pathogen Sal. momlla typhtmuriumattaches to the apical membrane of model intestinal epithelial monolayers, neutrophil transeplthellal mlgraUon Is effectively stimulated t~, but this response requites bacterial attachment, does not appear to he coordinated via the N-formyl-peptide receptor, and requires continuous epithelial, as well as bacterial, protein synthesis. Such observations suggest that epithelial cells may be able to signal to underlying neutrophils by novel means when they detect pathogens in the lumen, i.e. they may produce a transcellular chemotactlc signal. It has recently been reported that Sahnonella attachment to epithelial cells can alter the signalling cascades of eukaryotic cells ~6 and that such attachmer, t, when apical, can stimulate epithelial cells to synthesize and release, in a polarized fashion, the potent neutrophil chemotactic cytokine interleukin (IL) 8 ~s. However, this known chemotaxin does not appear to be responsible for the Sabnonella.induced transepithelial neutrophil migration ts and other transcellular chemotactic signals must be sought. Neutrophll space

migration

through

the paracellular

For conceptual pu[poses, columnar epithelia ate viewed by physiologists as consisting of two general compartments or'pathways': a transcellular pathway consisting of apical membrane, cytosol and baselateral membranes; and a paracellular pathway conTRENDSIN CELLBIOLOGYVOL 4 IANUARY1994

ral transmigration ~ "

I I

l-~a©.Aateral-tr~-apical transm~gratL~_._~

I

~,Aplcal-to-basolat I~LP FIGURE 1

Assaysof transepithelialmigration of neutrophds. Using simplemethods4, epithelialmonolayerscan be grown on permeablesupports in the usual configuration (left) or as invertedmonolayers(right). Neutrophilscan then be layeredby gravily on the surfaceof monolayersand induced to transmigratein the apical-to-basolateral,or the morephysiologicallyrelevantbasolateral-toapical, direction by establishingtransepithelialgradientsof chemntacticsolutes such as fMLP. Inhibitorsof transmigration(suchas antibodiesrepresentedby the 'Y') can the-,1be interfacedwith this assay.Adapted, with permission,from Ref.9 sisting of the apical intercellular tight junction and the subjunctional paracellular space :'.]° (Fig. 2). Analy~c-"utilizing canine renal (MDCK) and human intestinal ('['84) epRhehal monolayers mdicated that neutrophlls cross epitheha via the paracellular route 2,s,7. This route of movement dictates that neutrophils, which have a diameter averaging 10-12 IZnl, have to move through a space bounded by a -25 lain length of epithelial lateral plasma membrane. Thus, migration through the paracellular space of columnar epithelia hkely necessitates repeated rounds of extension and attachment/detachment of neutrophils on the epRhehal cell lateral membrane. Indeed. neutrophils migrating across epitheha are found to develop focally close membrane appositions with epithelial cells at sites where the underlying cytoskeleton of both cell types is modifieds. Recent analyses of cell adhesion requirements In transepithelial migration Indicate that one member of the neutrophil surface [32integrln family, CDI Ib/CDIS,

Apical membrane

;rnl:'ibCr°dll Lateral l membrane_.~

•

'Kisses'

)I~ . . . . "

Transcellular pathway

"v' -

} 1

X,,j~--

Perijunctional actomyosin ring

Paracellular pathway

FIGURE 2

Pathwaysacrosscolumnar epithelia.Columnarepitheliamay be viewedas consisting of transcellularand paracel~ularpathways,the latter havingapical circumferentialt,ght junctions as the rate.hmitmgseal Neutrophdsm;grateacross such epRhehavia the paracellularpathwayand, becauseof the substantialheight of theseepithelialcells,multiple neutrophd-ep=thehallateralmembrane interactionsare likelyto be requiredfor such r,eutroph,Imovement

FORUM

i !iit is reqmred for transepithelial migration in vitro, while the remaining members of this family, CD1 la/CD 18 and CDllc/CD18, may have limited importance 9,~z. Thus, for example, neutrophiln obtmned from patients with a disorder characterized by the lack of surface I~z mtegrin fall to migrate across intestinal epithelia9. It is currently not dear what epithelial receptors for CD1 lb/CD18 (such as ICAM-1) are recognized dunng transepithelial migration 9,~z. In vivo, neutrophil migration across columnar epithelia occurs in the basolateral-to-apicaldirection, which requires modification of the usual t~ssue culture geometry in which epithelia are studied (Fig. 1). For such studies, we have used the human intestinal epithelial cell line T84 since it has highly polarized apical and basolateral domains and is analogous in structural and functional phenotype to epithelia of the crypt, the malor site at wfuch neutrophil transmigration occurs in the intestine. Using this model, a consistent bias, based on polarity, was found to exist in the efflcmncy of neutrophil transepithelial migration under any given set of conditions: migration in the physiological direction was quantitatively 10-20 times greater than that in the reverse, apical-to.basolateral, direction 9. Such data suggest that polarity factors, such as limited availability of initial binding sites on the apical as opposed to the basolateral membrane and the sequence of binding signals, are likely to play a role in directing neutrophil movement across columnar epithelia. The efficlencles of transepithellal migration can also be modulated by cytokine.lnduced alterations in gene expression by epithelial cells. For example, interferon y (IFN-y) is a dominant cytoklne In the intestinal mucosa (due, In pnrt, to the presence of Intraeplthelial lymphoid cells), and exposure of intestinal epithelial models to IFN.y upregulates neutrophll movement into the monolayers In response to standard chemotactlc driving forces ~2. However, the effects on neutrophil migration elicited by IFN.7 are highly polarlzedtZ; treatment of the monolayer with IFN.,/enhances the ability of neutrophiis to penetrate this paracellular space from both the apical and basolateral sides, but migration across the epithelium is impeded in the basolateral. to.apical direction and promoted in the apical.to. basolateral direction. Since chronic exposure to IFN-ywould be a signal associated with chronic states of inflammation, such cytoklne effects might contribute to epithelial defence by prolonging the residency time of the neutrophll between the epithelial ceils. In the IFN-7 upregulated state, CDI lb/CDlgbased binding to epithelial cells is again required, whereas the contributions of CDIla/CD18 and CD1 lc/CD18 are not apparently significant tz, Other cytoklnes, such as IL-I, that dramatically affect endothelial-neutrophil Interactions are relatively ineffective in modulation of transepithelial mi. gration of neutrophils across model epitheli# z. The migration of neutrophils through the paracellular pathway requires that neutrophils cross the intercellular tight iunctions that form a belt around each epithelial cell at its apex. These junctions are the rate-limiting barrier that restricts passive solute 6

movement through the paracellular space TMand thus contribute substantially to the so-called barrier funcnon of epithelia. The permeability and structure of tight lunctions are highly plastic, being modulated by a host of pharmacological and physiological stimuli ~a. Given the highly regulated nature of the tight junction it is perhaps surprising that it is unlikely to be the major site of adhesion between neighbounng cells. For example, the tight junction is unable to maintain cell--cell associations if unaided by other junctional interactions between cells such as that afforded by E-cadherin-based homotyptc adhesion. Thus, it is not surprising that transmigrating neutrophfls have little difficulty crossing tight junctions. This is also true of those neutrophils that are genetically deficient in producing products of the respiratory burst s, products that we might suspect to be required if the neutrophil were to open the junction by chemical means. When neut-ophils cross tight junctions in large numbers, increases in junchonal permeability can be reachly measured by flux or electrical assays. Such disrupted junctions do have the ability to reseal after conditions driving transepithelial migration are removed. Even more impressive, however, ]s tile ability of the Junctions to maintain permeability seals when transmigration densities are

[ , . ~ - " ~ Neutrophll$ ~ I /W I f ~ __J-PI -""1

~ # Adenoslnq

-

\ Ih

/

~

~

O/

I 5'-ectonu¢leotidese

Neutrophll migration induced by the presence of chemotactlc agents in the crypt lumen FIGURE) Model of neutmphil interactionswith intestinalepithehum during transepithelialmigration, and the effectsof the neutrophil-derivedsecretagogueS'-AMP. Neutroph.ls transmigratethe columnarepitheliallining of the intestinal crypt via the paracellularpathwayand subsequentlycollect, at high density, in the lumen of the intestinalcrypts, Natural lumenal neutrophil-activatingfactorsinduce the regulated releaseof 5'-AMPwhich subsequentlyelicitsCI- secretionthe transportmechanismby which cryptsflush themselveswith isotonicfluid (seetext). Adapted,with i ,ermission,from Ref.13 TRENDSIN CELL610LOGYVOL. 4 JANUARY1994

FORUM

lOW7. This latter characteristic calls to mind the ability of the tight junctions of Sertoli cells to maintain their barrier function through complex structural rearrangements even while being penetrated by sperm cells ~9. Translocated neutrophils may interact w i t h the apical domain of epithelia

TABLE 1 - COKlrRAsTs BETWEEN ENDOTllELIAL AND EPITHELIAL NEUTROPHIL TRANSMIGRA110N Cell type

Role for ILl, TNF, LPS

IFN-y

L-selectin

-H-

-H-

+/--

-H-

-

-

++

_

ICAM-1

CD11a/1R (post ILl)

Endothelia

++

Epithelia (T84)

-

Diseases such as cystitis, pyeloEndotheha! data have been recently revaewedin Refs21 and 22. For epithehal (T84) data see nephritis, pneumonia and enteritis are Refs 9 arid i2. characterized by massive migration of neutrophils into the urinary space, the broncho-alveolar space and the intestinal lumen, grins? Do the attachment/de-adhesion events accomrespectively, where they encounter new and unique panying transepithehal migration involve adhesion microenvironments. The relatively high concensteps that are independent of ~2 integrin? H o w does trations of endotoxm and N-formyl-peptides (see an epltheha] cell signal to an underlying neutrophd above) that naturally occupy the intestinal lumen are that a pathogen has attached to its apical membrane? potent synergistic activators of neutrophils. When Do neutrophil-epithelia] interactions modulate ionactivated, neutrophils release a plethora of biologi- transport pathways in addition to Ci- secretion? A cally active compounds. At least in the intestinal site, better understanding of the answers to th2se quesafter transmigration neutrophils appear able to pro- tions should provide new insights into the pathovide signals to the apical epithelial membranes to genesis and treatment of inflammatory diseases. modulate fundamental epithelial characteristics such as vectorial movement of salt and water ¢,~,~°,l~(Fig. 3). Recently, neutrophils have been shown to elicit electrogenic C]- secretion both from T84 cell model epithelia and primary isolated m a m m a l i a n intestinal crypt epithelia. This CI-secretory activity is the specific transport pathway by w h i c h mucosal sinfaces are hydrated and, in the e×treme, the process

by which pathological transport such as sec~'etory dhlrrhoea (dmrrhoea due to active fluid secretion as opposed to that due to malabsorptlon) is elicited'". To drive this response, neutrophlls release, In regu-

lated fashion In response to sUmulaiion with agonists normally present In the colonic lumen, S'.AMP as a paracrine signal. This nucleotlde is not further metahollzed by the neutrophll surface. IJowever, epithelia, Including Intestinal epithelia, carry the ectoenzyme 5'-ectonucleottdase (CD73) on their stirface which, due to Its GPI anchor, is sorted to the apical domain ~,~(Fig. 3). This ectoenzyme thus provides a means for intestinal epithelial cells to convert the S'-AMP signal into adenosine ~~.Adenosine then activates the apical membrane receptors that initiate the salt and water secretory process:~. Since the volume flush represented by CI- secretion :s known to serve as a primitive defence that hmits duration of colonization by pathogens, transiocation of neutrophils in,u the lumenal compartment allows them to signal to epithelial cells and thereby to recruit them to participate in defence of the mucosal surface. Concluding remarks The process of m i g r a t i o n of neutrophils across epithelial cells differs in many respects to that of migration across endothelial layers (Table 1), a process that has recently become relatively well understood zl. Many crucial questions about the molecular nature and physiological sequelae of neutrophd-

epithelial ,_ellivteractions remain unanswered What are the epithelial receptors for neutrophd [~z roleTRENDS IN CELL BIOLOGYVOL 4 IANUARY 1994

References 1 CRAMER,E. B eta/_ (1986)/ CellBiol 102, 1868-1877 2 EVANS,C. W, TAYLOR,J E, WALKER,J- D and SIMMONS, N L (1983) Br J Exp Patho/64, 644-654 3 MADARA,J. Let ol (1992) J Chn Invest 89, 1938-1944 4 MADARA,J L, COLGAN, S. P, NUSRAT,A, DELP,C. and PARKOS,C. A (1992)/. T~ss Cult Meth. 14, 209-216 B MILKS,L C., CONYERS,G. P. and CRAMER,E (1986) j Cell Biol. 103, 2729-2738 6 NASH,S., PARKOS,C. A., NUSRAT,A., DELP,C. and MADARA,J. (1991) J Clln. Invest. 87, 1474-1477 7 NASH,S, STAFFORD,I. and MADARA, I. L. (1987)I. C/In Invest. 80, 1104=I113 B NASH,S., STAFFORD,J and MADARA,I, L. (1988) Lob. Invest. 59, 531=537 9 PARKOS,C. A., DELP,C., ARNAOUT,M. A and MADARA,J. L. (1991)/. C/In. Invest. 8B, 1605=1612 10 PARKOS,C. A., COLGAN, S. P, DELP,C., ARNAOUT,M. d MADARA,J L (I 992) J. Cell Biol. 117, 757-764 11 COLCAN,S., MATI'HEWS,J B., PARKOS,C. A., DELP,C., AWTREY, C and MADARA,I L.(1992)J LeukocyteBiol 52, 183-187 12 COLGAN,S. P., PARKOS,C, DELP,C, ARNAOUT,M A. and MADARA,J L (I 993) J. Cell B~ol 120, 785-798 13 MADARA,J L eta/ (1993)J Clin Invest 91,2320-2325 14 COLGAN,S. P,, SERHAN,C. S., DELP,C and MADARA,J. L J. C/in. Invest. On press) 1S McCORMICK,8., COLGAN, S P., I~ILLER,S I and MADARA, J L (1993)1. CellBioL 123, 895-908 16 BLISKA,I B, GALAN,J and FALKOW,S (1993) Cell73, 903-920 17 POWELL,D (1981)Am J Phys~ol 241, GZ75-GZ88 18 MADARA,J. L (1988) Cell 53, 497-498 19 BYERS,S and PELLETIER,R. M. (1991)m Tight luncvons (Cereljido,M, ed.), pp Z79-302, CRCPress 20 DONOWITZ,Ivl. and WELSH,M I (1987) m Physrologyoflhe Gastrointestinal Tract (Johnson,L R, ed ), pp. 1351-1388, RavenPress 21 OSBORN,L (1990) Cell62, 3-6 22. POBER,J S and COTRAN,R S (1990) TronspfanlaIK~,q SO, 537-544

JamesMadarais at the Harvard MedicalSchool (Brighamand Women's HOSpltat), Departmentof Pathology,7S FrancisSt, Boston, MA 02115, USA. Acknowledgements I gratefully acknowledgemy collaborators, R Mrsn,~and T. Patapoff (Genentech),and S. Miller and A. Arnaout (Massachusetts GeneralHosp=tal), aswell as lab. group members (C A. Parkos, S P Colgan, B A McCormick, G Strohmeler,W Lencerand S Nash)who camed out much of th~ work oted 7