progress report human intestinalintraepithelial lymphocytesperipheral blood lymphocytes isolated...

Gut, 1986, 27, 972-985

Progress report

Human intestinal intraepitheliallymphocytes

Of all the various cell types present within the lamina propria, it isremarkable that only lymphocytes predominantly gain access to theepithelium.' It seems likely that intestinal intraepithelial lymphocytes havean important immune function, and that this function is distinct from thatof the lamina propria lymphocyte, although lymphocytes from bothcompartments (epithelium and lamina propria) presumably interact intheir immune functions. Considerable functional and morphological datahave been derived concerning the intraepithelial lymphocytes since thistopic was reviewed by Ferguson in 1977.2 It is my intention to review datain regard to the intraepithelial lymphocytes largely published since 1977, toprovide some new data, to compare the function and structure of theintraepithelial lymphocytes with that of the lamina propria lymphocyte,and to compare (when appropriate) the structure and function of bothintraepithelial lymphocytes and lamina propria lymphocytes with that ofperipheral blood lymphocytes. The role of the human intraepitheliallymphocytes will be compared and contrasted to that of the rodentintraepithelial lymphocytes when pertinent. Intraepithelial lymphocyteshave been described since 1847 and have been the subject of many curiousand speculative publications. For those interested in details of earlierstudies of the intraepithelial lymphocytes, the reviews by Otto in 19733 andby Douglas and Weetman in 1975 are highly recommended.

Morphology

These cells are located within the epithelium, above the basal lamina, andbetween epithelial cells-that is, interepithelial cell and intraepithelial inlocation. Presence of lymphocytes within epithelial cells (emperipolesis)has not been demonstrated. Intraepithelial lymphocytes tend to be foundlargely at the base of epithelial cells, below epithelial nuclei.'I In thenormal intestine these cells are largely medium sized (5-9 [tm in diameter),contain sparse cytoplasmic organelles including mitochondria, lysosomalgranules, ribosomes (both isolated and polyribosomes), endoplasmicreticulum (rough predominating over smooth by a ratio of 2:1), and a smallGolgi complex.5 The nucleus contains large amounts of heterochromatinand lesser amounts of euchromatin. Nucleoli are often present. Thecytoplasm tends to be less dense than that of adjacent epithelial cells,contains organelles indicated above, and endocytic vesicles are oftenpresent along the plasma membrane. The intraepithelial lymphocytetends to be rounded to elongated in shape, and in active inflammatorystates often has numerous elongated cytoplasmic projections which make

972

on August 27, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.27.8.972 on 1 A

ugust 1986. Dow

nloaded from

http://gut.bmj.com/

Human intestinal intraepithelial lymphocytes

intimate contact with adjacent epithelial cell membranes. Morphology ofthe intraepithelial lymphocytes in diseased intestine will be described inlater sections. In the non-inflammed intestine, the intraepithelial lympho-cytes tend to be tightly compressed between epithelial cells. Intraepitheliallymphocytes do not form junctional contacts (desmosomes and/or tightjunctions) with adjacent epithelial cells.The nature of intraepithelial lymphocyte 'granules' may be of import-

ance in defining intraepithelial lymphocyte function, and thus somebackground material in regard to lymphocyte granules is reviewed. Rodentintraepithelial lymphocytes have very large granules, some of whichcontain histamine, and there is conflicting evidence that rodent intra-epithelial lymphocytes may be mast cell derived.8l" There is no solidevidence that human intraepithelial lymphocytes are mast cell derived andtheir granules are small in size.12 The granules are membrane bound andare clearly lysosomes. Confusion could be avoided if they were calledlysosomes, not granules. The custom of calling heterogeneous andhomogeneous dense bodies, when found in lymphocytes, as granulesrather than lysosomes is so deeply embedded in the scientific literature,however, that both terms must be utilised.The granules are azurophilic in Giemsa and Wright stained preparations,

and are stained with alcian blue (pH 2.2) and are metachromatic withtoluidine blue (pH 4). 12 The lysosomal nature of lymphocyte granules wasdocumented in 1965 by showing that these granules contained acidphosphatase.13 Lymphocyte granules have also been shown to containbeta-glucuronidase and alpha-naphthyl acetate esterase.14 15 Structuresthat are the site of both alpha-naphthyl acetate esterase and acidphosphatase activity represent the Gall body, a cluster of primarylysosomes surrounding a lipid droplet.15 The distribution of lysosomes inlymphocytes is of limited value in defining their function. The majority ofhuman T cells which possess surface receptors for IgM contain Gall bodieswhich stain for alpha-naphthyl acetate esterase in a localised or dot-likefashion while the majority of human T cells which possess surface receptorsfor IgG have lysosomes scattered diffusely throughout the cytoplasm.16Lymphocytes do not contain peroxidases, an enzyme found in the

myelo-monocytic series.15 17 All subclasses of T lymphocytes examinedhave been shown to contain at least, in part, lysosomal granules includingTg and Tm cells17; Leu 3a+ cells18; Leu 7+ and (Leu 11+) cells'9; Leu 3+,7+, M1+ cells18; Leu 2a+, T8+ cells20; and lymphokine activated killer cells(Leu 1+, T8+, T3+)21.B-lymphocytes do not contain granules and have very little acid

phosphatase and alpha-naphthyl acetate esterase activity.15 It is likely thatthe presence of lysosomal granules (including multivesicular bodies) isnecessary for the cytolytic activity of lymphocytes, whether the activity isantibody mediated or natural killer mediated.22 Indeed the presence ofgranules may represent a marker for all human cytolytic lymphocytes.22

In 1974, Guy-Grand showed in the rat that the vast majority ofintraepithelial lymphocytes were T cells and that, while both B cells and Tcells home to the gut, only T cells home to the epithelium.23 In 1976,Meuwissen et al showed in human colon and intestine that the vast majorityof intraepithelial lymphocytes were T cells, defined by immunoperoxidaselocalisation of intraepithelial lymphocytes using an anti-T cell globulin.24

973



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

This finding has been repeatedly confirmed with immunoperoxidase andimmunofluorescent localisation of a variety of monoclonal antibodies to Tcells.12 20 25-32 In vitro data indicating the presence of B cells within theintraepithelial lymphocyte population must be discounted.33 The B cellsmust have been contaminants derived from the lamina propria or fromlymphoid follicles.26 Plasma cells are not found within the epitheliumexcept possibly in the lymphoepithelial complex found in Peyer'spatches."4 Macrophages and mast cells are occasionally found within theepithelium while epithelial PMN leucocytes and eosinophils are quiteunusual except in inflammatory states. 12 26 34

Intraepithelial lymphocytes, virtually all of which are T cells, do notexpress class II histocompatibility antigens (HLA-DR, Ia-like), even ininflamatory states.20 28 3032 Moreover, 80 to 90% of intraepithelial Tcells express the phenotype of suppressor cells (T8, Leu 2a), while only 10to 20% express the helper phenotype (T4, Leu 3a).20 27 2932 In contrast,lamina propria lymphocytes are predominantly of the helper phenotype(60-70% T4', Leu 3a+) 20 27 29S:2 As expected, intraepithelial lympho-cytes express Class I HLA antigens29 but do not express T cell activationantigen (Tac),30 31 complement receptors (c3b),3° ii and rarely expressantigens characteristic of natural killer cells (HNK-1, Leu 7, Leu 11, T 10,Mi).20 30 32

In summary, intraepithelial lymphocytes are T cells predominantly ofthe suppressor phenotype. They are not activated, and do not express classII MHC antigens, and even though they often have the morphology ofnatural killer and/or killer cells (large granular lymphocytes), they do notbear surface antigens characteristic of natural killer cells. Intraepitheliallymphocyte populations do not contain B cells. Macrophages, mast cells,and PMN and eosinophilic leucocytes constitute a distinct minority of thiscellular compartment in the normal intestine.

Quantification of intraepithelial lymphocytes is inappropriately a subjectof controversy.35 Two methods of quantification have been proposed, oneis to quantify the number of intraepithelial lymphocytes in relation to thenumber of epithelial cells36 and the other is to quantify the number ofintraepithelial lymphocytes in relation to mucosal volume as determined bythe length of underlying muscularis mucosae-that is, areal density.6 Incoeliac sprue there is an increase in number of intraepithelial lympho-cytes/100 epithelial cells while there is a decrease in number of intra-epithelial lymphocytes per unit area (millimeters of muscularis mucosae)when compared with normal intestine.37 The latter (areal density) methodwould be the most appropriate method for quantification of intraepitheliallymphocytes if the immune response were directed towards all portions ofthe intestinal mucosa, while the former (intraepithelial lymphocytes/epithelial cells) would be appropriate if the immune response were largelydirected towards the epithelium. Morphologically, the immune response inmost intestinal diseases of unknown aetiology, or pathogenesis, is directedboth towards the epithelium (vide infra) and towards the lamina propria. Itis my opinion that generally quantification of intraepithelial lymphocytesshould be in relation to the epithelium, not in relation to areal density, andthat both epithelial and lamina propria be quantified separately. In thefinal analysis, any method of quantification of intraepithelial lymphocytesat present is arbitrary.

974 Dobbins



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/


The number of intraepithelial lymphocytes/100 epithelial cells in normaljejunum ranges from 9-39 with a mean of 21.1 with a standard deviation of7.5.2 We found a mean of 20 intraepithelial lymphocytes/100 epithelial cellsin normal jejunum,5 13/100 epithelial cells in normal ileum and 5/100epithelial cells in normal colon.32The cellular immune content of the lamina propria is strikingly

different from that of the epithelium, and there are clearly differences inimmune function of intraepithelial lymphocytes and of lamina proprialymphocyte.26 Lamina propria lymphocytes consist predominantly of thehelper phenotype (T4, Leu 3a) and with a smaller component of thesuppressor phenotype (T8, Leu 2a), the ratio of help to suppression beingsimilar to that of peripheral blood, 2:1. 20 27 29-32 Plasma cells are profuselypresent in the lamina propria with ratios of IgA:IgM:IgG containing cellsof 20:3:1. Macrophages, fibroblasts, and occasional mast cells andleucocytes are present throughout the normal lamina propria. T cells foundin the lamina propria are occasionally activated on the basis of expressionof Class II MHC antigens and Tac antigen,3032 and NK cells (Leu 7+,and/or Leu 11+) are far more easily detected in the lamina propria than inthe epithelium.20 30 32 A detailed description of the function of laminapropria lymphocytes is beyond the scope of this review. A comparison oflamina propria lymphocyte function to that of intraepithelial lymphocytesfunction will be given when appropriate in the next section.

In vitro studies of intraepithelial lymphocytes

Enriched populations of intestinal and/or colonic intraepithelial lympho-cytes can be collected and analysed in vitro.12 26' The intestinalepithelium with its associated lymphocytes is separated from the under-lying lamina propria by incubation in a calcium-magnesium free mediumcontaining 0-7mM ethylenediaminetetraacetate.26 3 The epithelial cellsare then removed by washing through nylon mesh and glass wool columns.The yield of intraepithelial lymphocytes is approximately 106 cells/g oftissue with a viability of 88%.26 These isolated cells, when tested in vitro,are likely to reflect in vivo function although some caution must beexercised in interpretation of the in vitro data.39 39 The isolation proceduremay deplete the T8+ cell population.31 A mechanical procedure forisolation of intraepithelial lymphocytes12 yields cells with an appearanceand function similar to cells collected using chemical methods.26 Most ofthe isolated cells are T cells (E rosette positive, T 11+, t3+), and expressthe phenotype associated with cytotoxic-suppressor T cells (T8+). Isolatedintraepithelial lymphocytes contain few B cells (sIg-, cIg-, B1-), macro-phages and monocytes (NSE-, Ml-), and almost no (<1%) natural killercells (Leu 7-).26 Approximately 80% of the cells are medium sizedlymphocytes and 20% are small sized lymphocytes.12 About 20% of theisolated lymphocytes contain small lysosomal granules which are azuro-philic with Giemsa staining.'2 26

Isolated intraepithelial lymphocytes have a low rate of spontaneousreplication and respond little, or not at all, to a variety of non-specificmitogens.'2 26 Addition of autologous macrophages, and addition ofinterleukin-2 to cultures of intraepithelial lymphocytes and mitogen doesnot initiate DNA synthesis.26 In contrast, lamina propria lymphocyte and

975



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

peripheral blood lymphocytes isolated under similar conditions respondedvigorously to such mitogens.26 These cells do not mediate natural killeractivity, nor do they mediate killer cell activity.'2 38 Intraepitheliallymphocytes, like lamina propria lymphocyte, can respond to mitogens tobe cytotoxic for Chang cells, but in contrast with lamina proprialymphocyte, are not cytotoxic for chicken erythrocytes.38 Addition ofgamma interferon, or of interleukin-2, to the culture medium fails toinduce natural killer activity by isolated intraepithelial lymphocytes, nordoes addition of antagonists of histamine or prostaglandins.12The intracytoplasmic granules stain with alcian blue (for acidic muco-

polysaccharides) and are metachromatic with toluidine blue (pH 4.0), andcan incorporate radioactive sulphate-all features in common with mastcells.12 Unlike mast cells, they contain no histamine and their granules arequite dissimilar to mast cell granules seen at electron microscopy.'2 26Finally, the mast cell yield in isolated intraepithelial lymphocytes collec-tions is minimal compared with lymphocyte yield.40

Intraepithelial lymphocytes have been shown to possess immuno-regulatory activity when added to co-cultures of peripheral blood lympho-cytes in a pokeweed-mitogen driven immunoglobulin synthesis system.26The effect of intraepithelial lymphocytes was the same for both autologousor heterologous assay systems, resulting in enhancement of IgA synthesisand ofIgM synthesis. When increasing numbers of intraepithelial lympho-cytes were added to peripheral blood lymphocytes, enhancementdisappeared and suppression of immunoglobulin synthesis was observed.26Similar results have been observed (using lamina propria lymphocytes)when increasing numbers of T cells were added to a fixed number of B cellsin a similar assay system,41 although Elson has shown that the predominantlamina propria response is that of help.42The specific role of intraepithelial lymphocytes remains to be defined.

The cells are largely T cells of the suppressor phenotype, and in commonwith other suppressor effector cells (in contrast with cytotoxic effectors)respond poorly to mitogens.43 They do not mediate cytotoxic functions invitro, even though one-fourth of intraepithelial lymphocytes have themorphology of large granular lymphocytes, cells that characteristicallymediate natural killer function. Even though intraepithelial lymphocytegranules bear some similarity to mast cell granules, intraepitheliallymphocytes are clearly not mast cells, nor do they appear to be mast cellderivatives. Suppressor T cells are reactive with class I (HLA-A, B, C)antigens, antigens strongly expressed by intestinal epithelium.29 It isconceivable that intraepithelial lymphocytes play an important role in thelocal immune response by reacting to epithelial cells altered by viralinfection or other foreign antigens.2 Increased numbers of intraepitheliallymphocytes are present ip coeliac sprue, cow's milk intolerance, parasiticdiseases, and graft-vs-host disease. The intestinal immune response ispredominantly a helper response, particularly when lamina proprialymphocyte are assessed.42 It is possible that the epithelial suppressor cellsdampen this helper response, particularly when exuberant, although thereare no data to support this speculation at present.

Cerf-Bensussan et al have recently shown that isolated rat intraepitheliallymiphocytes secreted a factor (probably gamma interferon) capable ofinducing Ia antigen (HLA-DR in man) expression by an intestinal

Dobbins976



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/


epithelial cell line, IEC 17.25 Importantly, lymphocyte proliferation wasnot essential for the secretion of the Ia-inducing factor. These data suggestthat intraepithelial lymphocytes may be involved in modulating someepithelial cell functions, and specifically Ia antigen expression. Thispermits the speculation that Ia antigen positive epithelial cells have theability to present antigen to T cells.2' Further, growth of the epithelial cellline was inhibited by supernatants from Con A stimulated intraepitheliallymphocytes, indicating that intraepithelial lymphocytes may modulateepithelial cell growth.

Origin and traffic of intraepithelial lymphocytes

Lymphocytes isolated either from efferent lymph of gut associatedmesenteric nodes, and returned intravenously after radiolabelling to thesame animal will preferentially recirculate through those regions fromwhich they were first isolated. There are considerable data available inlaboratory animals (but none in man) that provides a reasonableunderstanding of the mechanisms of this selective migration, or 'homing',of lymphocytes to gut associated lymphoid tissue (GALT).

Peyer's patches are an enriched (probably exclusive) source of pre-cursor's of IgA - secreting plasma cells found in the intestinal laminapropria while the presence of mesenteric nodes is not required forpopulation of the lamina propria.45 Antigenic stimulation is not essential tohoming, but has a profound effect on the number and distribution of cellsin the lamina propria.45 There is a large increment of plasma cells in the gutat the site of locally applied antigen. Not only do Peyer's patch derivednon-recirculating large lymphocytes give rise to plasma cells in the laminapropria but Peyer's patch derived recirculating small lymphocytes alsocontribute. The increase in plasma cells in the lamina propria afterantigenic challenge is because of both recruitment from the circulatinglymphocyte pool and proliferation in the lamina propria of these newlyrecruited cells.45 Cells obtained from mesenteric lymph nodes also localisepreferentially in GALT.46 Guy-Grand, and colleagues showed that both Bcells and T cells take part in preferential homing to GALT, and thatantigen does not direct homing in that homing to fetal mucosa (unexposedto antigen) paralleled that to antigen exposed mucosa.23 4 Peripherallymph node blast cells, immunised by an antigen also present in the gut, donot home to the gut.23 In later studies, Guy-Grand also showed that thepresence of T cells in the gut are the result of antigenic stimulation in thatsuch cells were absent in germ free mice, but showed that the gut homingmechanism was independent of antigen in that it was the generation of Tblasts within Peyer's patches that conferred the gut homing property.8Only T cells enter the epithelium and these cells have a very low mitoticrate.8 Husband showed that initial appearance of lymphoblasts in thelamina propria is independent of antigen, but that there was a selectiveantigen associated accumulation of specific plasma cells at a later stage ofthe local response.48 He found no specialised site (vide infra) within thelamina propria for lymphocytes to enter the tissue, and that the migratorylymphocytes accumulated around the crypt region in immunised intestine,a portion of the accumulation being accounted for by proliferation withinthe lamina propria.48

977



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

In summary, T cells (and B cells) arising as a result of antigenicstimulation in Peyer's patches populate the gut lamina propria andepithelium. These cells are normally processed both in the spleen49 and inmesenteric lymph nodes, but neither the spleen nor mesenteric lymphnodes are essential to homing of Peyer's patch derived cells to the gut.Antigen is not essential to homing but greatly expands the accumulationand proliferation of cells within the gut, and determines distribution ofhoming. There is no evidence that intraepithelial lymphocytes (in contrastwith lamina propria lymphocyte proliferate in situ.The localisation of B cells and T cells to Peyer's patches is mediated by

interactions between these cells and the endothelial cells lining post-capillary high endothelial venules (HEV).50 51 There is preferentialadherence of Peyer's patch lymphocytes to high endothelial venules inPeyer's patches over those in peripheral lymph nodes. Further, there ispreferential binding of B cells to Peyer's patch high endothelial venulesand of T cells to peripheral node high endothelial venules, while B and Tcells bind equally well to mesenteric node high endothelial venules.Finally, T cells of the helper phenotype localise more efficiently in Peyer'spatches whereas cells of helper phenotype and of suppressor phenotypelocalise equivalently in peripheral lymph nodes.5' The specificity ofmigration of lymphocytes appears to be determined by selective recogni-tion of organ specific determinants on endothelial cells ofhigh endothelial venules, specialised venules that mediate the exitof migrating lymphocytes from the blood, and which appear tobe an important factor in determining the character of local immuneresponses.5' Selective lymphocyte entry, however, is probably only oneof many factors controlling the distribution of lymphocytes, - that is,this mechanism is operative in Peyer's patches but is absent in the laminapropria.The factors that determine the highly exclusive entry of T suppressor

cells into the epithelium have not been determined. There is clearmorphologic evidence of lymphocyte migration from the lamina propriainto the epithelium and back again, and virtually no evidence of extrusionof lymphocytes into the gut lumen. 1 5 52-54 Marsh studied proliferation andmigration of intraepithelial lymphocytes in mice after intraperitonealinjections of tritiated thymidine.5' He found labelled lymphocytes crossingthe basal lamina for seven days after injection, that they divided at a rate of1% per hour, and circulated through the epithelium returning to thelamina propria at a rate of three intraepithelial lymphocytes/1000 epithelialcell nuclei per hour.55 No intraepithelial lymphocytes were observedadjacent to or extending through tight junctions of epithelial cells. Eventhough lymphocytes do not enter the lumen from normal epithelium, thereis an intriguing and well documented report showing in adult rats thatlymphocytes placed in the intestinal lumen can traverse the epithelial layerand gain access to host tissues and produce a graft-vs-host response-thatis, they are 'naturally transplanted'. 6 It is unlikely that such an event takesplace in adult man, but may play a role in lymphocyte transfer from motherto offspring during lactation (breast milk contains both lymphocytes andmacrophages) and nursing.56 For example, there are reports of the transferof specific reactivity to tuberculin in human infants breast fed by tuberculinpositive mothers.

978 Dobbins



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/


Intraepithelial lymphocytes in disease states

Intraepithelial lymphocytes are markedly decreased in germ free animals.8When expressed as a ratio to number of epithelial cells (see previous page),intraepithelial lymphocytes are clearly increased in number in severaldisease states. They are increased in coeliac sprue, tropical sprue, and inthe coeliac sprue lesion found in dermatitis herpetiformis.2 7 0 37 Failureto find a clear increase of intraepithelial lymphocytes in an intestinal lesionotherwise compatible with that of untreated coeliac sprue should castserious doubt upon the diagnosis.2 Intraepithelial lymphocytes areincreased, but not consistently so, in other inflammatory states of theintestine such as in the stasis syndromes.2 Intraepithelial lymphocytes areincreased in the small intestine, but not in the colon of subjects withacquired immunodeficiency syndrome, both in the presence and absence ofdiarrhoea and/or of obvious pathogens.58 Intraepithelial lymphocytes arenot increased in inflammatory bowel disease.2 2 31 32

In coeliac sprue, the T4:T8 ratio of intraepithelial lymphocytes isunchanged when compared with normal (O*08 vs 0.19), but the proportionof T8+ intraepithelial lymphocytes expressing Leu 1 (a pan-T cell antigenAis significantly increased when compared with control (56% vs 320).3Heterogeneity of Leu 1+ expression was not found in the T8+ laminapropria lymphocyte population. The intraepithelial lymphocytes in coeliacsprue were not activated in that they failed to express HLA-DR and Tac,the latter antigen expressed by T cells after stimulation in culture. Onlyvery rare natural killer (HNK-1+) cells were found.30

Intraepithelial lymphocytes are not increased in number in inflammatorybowel disease (Crohn's colitis, Crohn's ileitis, and ulcerative colitis) nor isthere a change in the ratio of T8+ cells to T4+ cells when compared withnormal.2 24 31 32 Further, T8+ intraepithelial lymphocytes co-expressedLeu 1+ in a ratio similar to control (one-third positive) and were notactivated (HLA-DR-, Tac).3' Colonic epithelium in active ulcerativecolitis and coeliac disease showed variable expression of HLA-DR whereasnormal colonic epithelium, non-involved mucosa from diseased specimens,and mucosal epithelium in inactive disease did not.3' 32 Both normal andinflammatory bowel disease small intestinal epithelium express HLA-DR.20 59 There is a significant decrease in intraepithelial lymphocytes(when compared with normal ileum) in ileal coeliac disease, the changebeing largely because of a decrease in T8+ intraepithelial lymphocytes. 2The T4:T8 ratio in ileal coeliac disease was 0 37±5, in normal margins ofdisease tissue, 0-22+1-0, and in control ileum, 0-13+0- 1.32 The inductionof colonic epithelial cell expression of HLA-DR in inflammatory boweldisease has led to the hypothesis that there is an autoimmune mechanism ininflammatory bowel disease in which the epithelium presents antigen toT4+ intraepithelial lymphocytes and possibly to T4+ lamina proprialymphocyte.60 Both graft-vs-host disease and naturally occurring immun-ological stimuli can, however, induce the expression of Ia antigen (HLA-DR-like) in rat intestinal epithelium61 indicating that epithelial expressionof Ia (DR) is secondary to a variety of immunological stimuli.61 B cellsand natural killer cells (Leu 7+) do not enter the epithelium in inflam-matory bowel disease.31 32 Chiba et al isolated intraepithelial lymphocytesfrom diseased specimens of inflammatory bowel disease patients and

979



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

found no significant differences in cytotoxic capabilities of disease tissueintraepithelial lymphocytes vs normal tissue intraepithelial lympho-cytes.38

In untreated Whipple's disease, the number of intraepithelial lympho-cytes/100 epithelial cells remains in the normal range.3 5 At electronmicroscopy, Whipple intraepithelial lymphocytes were similar to normalintraepithelial lymphocytes.5 Intraepithelial lymphocytes in Whipple'sdisease were slightly larger in size, had an increased number of lysosomalgranules, and an increased content of ribosomes when compared withnormal intraepithelial lymphocytes. There was a clear increase in intra-epithelial eosinophils, PMN leucocytes, and macrophages (the latter twocells often containing ingested Whipple bacilli) in untreated Whipple'sdisease. Clearly leucocytes other than lymphocytes are involved in theepithelial immune response in Whipple's disease.5 Curiously, there was afour-fold increase in number of intraepithelial lymphocytes in the singlepatient with Whipple's disease without intestinal involvement.5

Intraepithelial lymphocytes have not been quantified in human graft-vs-host disease but are clearly increased in murine graft-vs-host disease.62 Theintraepithelial lymphocytes count rose within 24 hours of induction ofgraft-vs-host disease while increases in crypt length occurred within threedays. The data indicated that measurements of mucosal architecture andintraepithelial lymphocytes counts could be used to quantify mucosal cellmediated immune reactions in appropriate experimental conditions.62Intraepithelial lymphocytes are implicated as effectors in the damage torectal crypt epithelium in human graft-vs-host disease.63 In this morph-ological study, lymphocytes were shown to be the predominant cellinfiltrating the epithelium and to have multiple features of cytolyticcells,-that is, indentation of adjacent epithelial cells by narrow cyto-plasmic probes, extension of brood pseudopods to the nuclear membranesof epithelial cells, and protoplasmic surrounding of desmosomes.63 Suchmorphologic features suggested that the lymphocyte-to-epithelial cellcontacts represented the recognition phase of alloimmune T-lymphocytecytolysis.63 Epithelial injury and lymphocytic infiltration predominated inthe bases of crypts in mild graft-vs-host disease and extended to the surfaceepithelium in severe graft-vs-host disease.High intraepithelial lymphocytes counts have been reported in children

with giardiasis2 and in murine giardiasis.2 64 There are limited data in themurine model that intraluminal lymphocytes may promote clearance ofgiardia.64 Finally, we have morphological evidence that intraepitheliallymphocytes are involved in the immune response to microsporidialinvasion of intestinal epithelial cells in AIDS.65 There is a clear increase inintraepithelial lymphocytes number and many of these intraepitheliallymphocytes contain remnants of microsporidial spores and trophozoites.There are little other data2 in regard to intraepithelial lymphocyterole in response to intestinal parasites although it is likely that intra-epithelial lymphocytes are actively involved in the immune response toparasites.

Of mice and men

Throughout this paper, frequent reference has been made to intraepithelial

980 Dobbins



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/


lymphocytes morpholog2y and function in rodents and other smalllaboratory animals' 41l 25 34 44-56 61 62 64 while the major emphasis hasbeen upon the human intraepithelial lymphocytes. The Table compares themajor features of human and rodent intraepithelial lymphocytes, featuresderived from the references cited above and from several other importantpublications concerning rodent intraepithelial lymphocytes.6669 The mostnotable distinction between the intraepithelial lymphocytes of the twospecies is that rodent intraepithelial lymphocytes are more likely to begranulated and are capable of mediating all cytotoxic functions whilehuman intraepithelial lymphocytes are exceedingly limited in their cytotoxiccapabilities. It is suspected that the human intraepithelial lymphocytes willbe shown to have cytotoxic capabilities similar to those of the rodent whenappropriate experimental conditions are devised.

Conclusion

Intraepithelial lymphocytes are a unique set of T cells that are pre-dominantly of the suppressor phenotype and which have a variedexpression of other T cell antigens, at least in coeliac sprue. Intraepitheliallymphocytes are not activated as determined by expression of HLA-DRantigen and Tac antigen and do not respond to non-specific mitogens. Theydo not have natural killer phenotype or function, although they resemblenatural killer cells in that they are often large granular lymphocytes.Intraepithelial lymphocytes in rodents have some similarity to mast cells,but there is little evidence in man that the intraepithelial lymphocytes isrelated to mast cells. Intraepithelial lymphocytes may play an immuno-regulatory role in man, possibly by suppressing the systemic immuneresponse to antigens that simultaneously promote a mucosal (laminapropria) helper immune response. The epithelium is ideally suited to thefunction of suppressor cells because of its strong expression of class I majorhistocompatibility complex antigens.29 The small population of helper cellswithin the epithelium may add further control over the functions ofsuppressor intraepithelial lymphocytes, especially in inflammatory condi-

Table In vitro studies of intraepithelial lymphocytes inman and rodent

Man Rodent

Granulated (LGL) 25% 25-90%MICC - T cell function + +ADCC - K cell function 0 + +SCMC - NK cell function 0 +++Suppressor T Cell 85% 75%Respond to mitogens 0 0Respond in MLR (MLC) ND +Influence Ig synthesis + NDLymphokine production ND +

ND - not determined, MICC - mitogen induced cellular cytotoxicity, ADDC - antibody dependentcellular cytotoxicity killer (K) cell function, SCMC - spontaneous cell mediated cytotoxicity natural killer(NK) function, MLR - mixed lymphocyte response, MLC - mixed lymphocyte cytotoxicity.

981



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

982 Dobbins

tions of the colon when the colonic epithelium is induced to express Class IImajor histocompatibility complex antigens.20 31 32

Supported by funding from the National Foundation for Ileitis and Colitisand from the Veterans Administration.

WILLIAM 0 DOBBINS III

Department of Internal MedicineUniversity of Michigan Medical School andVeterans Administration Medical CenterAnn Arbor, Michigan, USA.

Address for correspondence: William 0 Dobbins III, MD, Veterans Administration Medical Center, 2215 Fuller Road, AnnArbor, Michigan 48105, USA.

Received for publication 11 November 1985.

References

1 Marsh MN. Studies of intestinal lymphoid tissue. I. Electron microscopic evidence of'blast transformation' in epithelial lymphocytes of mouse small intestinal mucosa. Gut1975; 16: 665-82.

2 Ferguson A. Intraepithelial lymphocytes of the small intestine. Gut 1977; 18: 921-37.3 Otto HF. The interepithelial lymphocytes of the intestine. Morphological observationsand immunologic aspects of intestinal enteropathy. Curr Top Pathol 1973; 57: 81-121.

4 Douglas AP, Weetman AP. Lymphocytes and the gut. Digestion 1975; 13: 344-71.5 Austin LL, Dobbins WO III. Intraepithelial leucocytes of the intestinal mucosa in normalman and in Whipple's disease: a light and electron microscopic study. Dig Dis Sci 1982;27: 311-20.

6 Marsh MN. Studies of intestinal lymphoid tissue. Quantitative analyses of epitheliallymphocytes in the small intestine of human control subjects and of patients with celiacsprue. Gastroenterology 1980; 79: 481-91.

7 Marsh MN, Mathan M, Mathan VI. Studies of intestinal lymphoid tissue. The secondarynature of lymphoid cell "activation" in the jejunal lesion of tropical sprue. Am J Pathol1983; 112: 302-12.

8 Guy-Grand D, Griscelli C, Vassalli P. The mouse gut T lymphocyte, a novel type ofT cell.Nature, origin, and traffic in mice in normal and graft-versus-host conditions. J Exp Med1978; 148: 1661-77.

9 Schrader JW, Scollay R, Battye F. Intramucosal lymphocytes of the gut: Lyt-2 and Thy-1phenotype of the granulated cells and evidence for the presence of both T cells and mastcell precursors. J Immunol 1983; 130: 558-64.

10 Sredni B, Friedman MM, Bland CE, Metcalfe DD. Ultrastructural, biochemical, andfunctional characteristics of histamine-containing cells cloned from mouse bone marrow:tentative identification as mucosal mast cells. J Immunol 1983; 131: 915-22.

11 Tagliabue A, Befus AD, Clark DA, Bienenstock J. Characteristics of natural killer cells inthe murine intestinal epithelium and lamina propria. J Exp Med 1982; 155: 1785-96.

12 Cerf-Bensussan M, Guy-Grand D, Griscelli C. Intraepithelial lymphocytes of human gut:isolation, characterisation and study of natural killer activity. Gut 1985; 26: 81-8.

13 Parker JW, Wakasa H, Lukes RJ. The morphologic and cytochemical demonstration oflysosomes in lymphocytes incubated with phytohemagglutinin by electron microscopy.Lab Invest 1965; 14: 1736-43.

14 Ferrarini M, Cadoni A, Franzi AT, et al. Ultrastructure and cytochemistry of humanperipheral blood lymphocytes. Similarities between the cells of the third population andTG lymphocytes. Eur J Immunol 1980; 10: 562-70.

15 Zicca A, Leprini A, Cadoni A, Franzi AT, Ferrarini M, Grossi CE. Ultrastructurallocalization of alpha-naphthyl acid esterase in human TM lymphocytes. Am J Pathol 1981;105: 40-6.



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

Human intestinal intraepithelial lymphocytes 983

16 Grossi CE, Cadoni A, Zicca A, Leprini A, Ferrarini M. Large granular lymphocytes inhuman peripheral blood: Ultrastructural and cytochemical characterization of thegranules. Blood 1982; 59: 277-83.

17 Grossi CE, Webb SR, Zicca A, et al. Morphological and histochemical analyses of twohuman T-cell subpopulations bearing receptors for IgM or IgG. J Exp Med 1978; 147:1405-17.

18 Velardi A, Grossi CE, Cooper MD. A large subpopulation of lymphocytes with T helperphenotype (Leu-3/T4+) exhibits the property of binding to NK cell targets and granularlymphocyte morphology. J Immunol 1985; 134: 58-64.

19 Grossi CE, Ferrarini M. Morphology and cytochemistry of human large granularlymphocytes. In: NK cells and other natural effector cells. New York: Academic Press,1982: 1-8.

20 Cerf-Bensussan N, Schneeberger EE, Bhan AK. Immunohistologic and immunoelectronmicroscopic characterization of the mucosal lymphocytes of human small intestine by theuse of monoclonal antibodies. J Immunol 1984; 130: 2615-22.

21 Fiocchi C, Tubbs RR, Youngman KR. Human intestinal mucosal mononuclear cellsexhibit lymphokine-activated killer cell activity. Gastroenterology 1985; 88: 625-37.

22 Grossi CE, Zicca A, Cadoni A, Mingari M, Moretta A, Moretta L. Ultrastructuralcharacteristics of human T cell clones with various cytolytic activities. Eur J Immunol1983; 13: 670-77.

23 Guy-Grand D, Griscelli C, Vassalli P. The gut-associated lymphoid system: nature andproperties of the large dividing cells. Eur J Immunol 1974; 4: 435-43.

24 Meuwissen SGM, Feltkamp-Vroom TM, de la Reviere AB, von dem Borne AEG Kr,Tytgat GN. Analysis of the lympho-plasmacytic infiltrate in Crohn's disease with specialreference to identification of lymphocyte-subpopulations. Gut 1976; 17: 770-80.

25 Cerf-Bensussan N, Quaroni A, Kurnick JT, Bhan AK. Intraepithelial lymphocytesmodulate Ta expression by intestinal epithelial cells. J Immunol 1984; 132: 2244-52.

26 Greenwood JH, Austin LL, Dobbins WO III. In vitro characterization of human intestinalintraepithelial lymphocytes. Gastroenterology 1983; 85: 1023-24.

27 Janossy G, Tidman N, Selby WS, Thomas JA, Granger S. Human T lymphocytes ofinducer and suppressor type occupy different microenvironments. Nature 1980; 288: 81-4.

28 Selby WS, Janossy G, Jewell DP. Immunohistological characterisation of intraepitheliallymphocytes of the human gastrointestinal tract. Gut 1981; 22: 169-76.

29 Selby WS, Janossy G, Goldstein G, Jewell DP. T lymphocyte subsets in human intestinalmucosa: the distribution and relationship to MHC-derived antigens. Clin Exp Immunol1981; 44: 453-8.

30 Selby WS, Janossy G, Bofill M, Jewell DP. Lymphocyte subpopulations in the humansmall intestine. The findings in normal mucosa and in the mucosa of patients with adultcoeliac disease. Clin Exp Immunol 1983; 52: 219-28.

31 Selby WS, Janossy G, Bofill M, Jewell DP. Intestinal lymphocyte subpopulations ininflammatory bowel disease: an analysis by immunohistological and cell isolationtechniques. Gut 1984; 25: 32-40.

32 Hirata I, Berrebi G, Austin LL, Keren DF, Dobbins WO III. Immunohistologicalcharacterization of intraepithelial and lamina propria lymphocytes in control ileum andcolon and in inflammatory bowel disease. Dig Dis Sci (In press).

33 Bartnik W, ReMine SG, Chiba M, Thayer WR, Shorter RG. Isolation and characteriza-tion of colonic intraepithelial and lamina proprial lymphocytes. Gastroenterology 1980; 78:976-85.

34 Bhalla DK, Owen RL. Cell renewal and migration in lymphoic follicles of Peyer's patchesand cecum-an autoradiographic study in mice. Gastroenterology 1982; 82: 232-42.

35 Phillips AD. Epithelial lymphocytes in celiac-sprue. Gastroenterology 1981; 80: 1085-6.36 Ferguson A, Murray D. Quantitation of intraepithelial lymphocytes in human jejunum.

Gut 1971; 12: 988-94.37 Corazza GR, Frazzoni M, Gasbarrini G. Jejunal intraepithelial lymphocytes in coeliac

disease: are they increased or decreased? Gut 1984; 25: 158-62.38 Chiba M, Bartnik W, ReMine SG, Thayer WR, Shorter RG. Human colonic

intraepithelial and lamina proprial lymphocytes: cytotoxicity in vitro and the potentialeffects of the isolation method on their functional properties. Gut 1981; 22: 177-86.

39 Gibson PR, Hermanowicz A, Verhaar HJJ, Ferguson DJP, Bernal AL, Jewell DP.Isolation of intestinal mononuclear cells: factors released which affect lymphocyteviability and function. Gut 1985; 26: 60-8.



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

984 Dobbins

40 Fox CC, Dvorak AM, Peters ST, Kagey-Sobotka A, Lichtenstein LM. Isolation andcharacterization of human intestinal mucosal mast cells. J Immunol 1985; 135: 483-91.

41 Clancy R, Cripps A, Chipchase H. Regulation of human gut B lymphocytes by Tlymphocytes. Gut 1984; 25: 47-51.

42 Elson CO, Machelski E, Weiserbs DB. T cell-B cell regulation in the intestinal laminapropria in Crohn's disease. Gastroenterology 1985; 89: 321-27.

43 Clement LT, Dagg MK, Landay A. Characterization of human lymphocyte subpopula-tions: alloreactive cytotoxic T-lymphocyte precursor and effector cells are phenotypicallydistinct from Leu 2+ suppressor cells. J Clin Immunol 1984; 4: 395-402.

44 Cahill RNP, Poskitt DC, Frost H, Trnka Z. Two distinct pools of recirculating Tlymphocytes: migratory characteristics of nodal and intestinal T lymphocytes. J Exp Med1977; 145: 420-8.

45 Husband AJ, Gowans JL. The origin and antigen-dependent distribution of IgA-containing cells in the intestine. J Exp Med 1978; 148: 1146-60.

46 Griscelli C, Vassalli P, McCluskey RT. The distribution of large dividing lymph node cellsin syngeneic recipient rats after intravenous injection. J Exp Med 1969; 130: 1427-51.

47 Parrott DMV, Rose ML. Migration pathways of T-lymphocytes in the small intestine. AdvExp Med Biol 1978; 107: 67-74.

48 Husband AJ. Kinetics of extravasation and redistribution of IgA-specific antibody-containing cells in the intestine. J Immunol 1982; 128: 1355-59.

49 Tseng J. Transfer of lymphocytes of Peyer's patches between immunoglobulin allotypecongenic mice: repopulation of the IgA plasma cells in the gut lamina propria. J Immunol1981; 127: 2039-43.

50 Chin Y, Carey GD, Woodruff JJ. Lymphocyte recognition of lymph node highendothelium. J Immunol 1982; 129: 1911-15.

51 Kraal G, Weissman IL, Butcher EC. Differences in in vivo distribution and homing of Tcells subsets to mucosal vs nonmucosal lymphoid organs. J Immunol 1983; 130: 1097-101.

52 Darlington D, Rogers AW. Epithelial lymphocytes in the small intestine of the mouse. JAnat 1966; 100: 813-30.

53 Meader RD, Landers DF. Electron and light microscopic observations on relationshipsbetween lymphocytes and intestinal epithelium. Am J Anat 1967; 121: 763-74.

54 Ropke C, Everett NB. Proliferative kinetics of large and small intraepithelial lymphocytesin the small intestine of the mouse. Am J Anat 1975; 145: 395-408.

55 Marsh MN. Aspects of proliferation and migration of epithelial lymphocytes in the smallintestine of mice. Gut 1975; 16: 674-82.

56 Seelig LL Jr, Billingham RE. Capacity of "transplanted" lymphocytes to traverse theintestinal epithelium of adult rats. Transplantation 1981; 32: 308-14.

57 Ogra SS, Weintraub D, Ogra PL. Immunologic aspects of human colostrum and milk. III.Fate and absorption of cellular and soluble components in the gastrointestinal tract of thenewborn. J Immunol 1977; 119: 245-8.

58 Kotler DP, Gaetz HP, Lange M, Klein EB, Holt PR. Enteropathy associated with theacquired immunodeficiency syndrome. Ann Intern Med 1984; 101: 421-8.

59 Hirata I, Austin LL, Blackwell WH, Weber JR, Dobbins WO III. Immunoelectronmicroscopic localization of HLA-DR antigen in control small intestine and colon and ininflammatory bowel disease. Dig Dis Sci. (In press).

60 Bottazzo GF, Pujol-Borrell R, Hanafusa T. Role of aberrant HLA-DR expression andantigen presentation in induction of endocrine autoimmunity. Lancet 1983; 2: 1115-8.

61 Barclay AN, Mason DW. Induction of Ta antigen in rat epidermal cells and gut epitheliumby immunological stimuli. J Exp Med 1982; 156: 1665-76.

62 Mowat AM, Ferguson A. Intraepithelial lymphocyte count and crypt hyperplasia measurethe mucosal component of the graft-versus-host reaction in mouse small intestine.Gastroenterology 1982; 83: 417-23.

63 Gallucci BB, Sale GE, McDonald GB, Epstein R, Shulman HM, Thomas ED. The finestructure of human rectal epithelium in acute graft-versus-host disease. Am J Surg Pathol1982; 6: 293-305.

64 Owen RL, Nemanic PC, Stevens DP. Ultrastructural observations on giardiasis in amurine model. Gastroenterology 1979; 76: 757-69.

65 Dobbins WO III, Weinstein WM. Electron microscopy of the intestine and rectum inacquired immuno-deficiency syndrome. Gastroenterology 1985; 88: 738-49.

66 Arnaud-Battandier F, Nelson DL. Immunologic characteristics of intestinal lymphoidcells of the guinea pig. Gastroenterology 1982; 82: 248-53.

67 Lyscom N, Brueton MJ. Intraepithelial, lamina propria and Peyer's patch lymphocytes of



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

Human intestinal intraepithelial lymphocytes 985

the rat small intestine: isolation and characterization in terms of immunoglobulin markersand receptors for monoclonal antibodies. Immunology 1982; 45: 775-83.

68 Mowat AM, Tait RC, Mackenzie S, Davies MDJ, Parrott DMV. Analysis of natural killereffector and suppressor activity by intraepithelial lymphocytes from mouse small intestine.Clin Exp Immunol 1983; 52: 191-8.

69 Nauss KM, Pavlina TM, Kumar V, Newberne PM. Functional characteristics oflymphocytes isolated from the rat large intestine. Gastroenterology 1984; 86: 468-75.



j.com/


ugust 1986. Dow

nloaded from

http://gut.bmj.com/

progress report human intestinalintraepithelial lymphocytesperipheral blood lymphocytes isolated...

Documents