The gut-associated lymphoid system: nature and properties of the large dividing cells

The origin and antigen-dependent distribution of IgA-containing cells in the intestine

The aims of this paper were to establish the origin of cells producing IgA antibody to cholera toxoid in the lamina propria of the small intestine and to define the role of antigen in their distribution. The use of Thirty-Vella loops made it possible to restrict antigenic challenge to a defined segment of the intestine in rats which had been primed i.p. with toxoid in Freund's complete adjuvant. The anti-toxin-containing cells (ACC) which appeared in the draining thoracic duct lymph after challenge of a loop were almost all of IgA specificity and their numbers were proportional to the length of intestine exposed to antigen. The abolition of this cellular response which occurred when Peyer's patches (PP) were removed from a loop before challenge and the failure of mesenteric lymphadenectomy significantly to affect the response indicated that ACC originated exclusively from PP. Cell transfer studies showed that although nonrecirculating large lymphocytes gave rise to ACC in the lamina propria, some of the recirculating small lymphocytes also developed subsequently into ACC. Counts of ACC in the lamina propria of challenged loops were consistently greater than in nonchallenged loops although some ACC were always present in the latter. However, a time-course study on the appearance of ACC in the lamina propria of cannulated rats given a single dose of thoracic duct lymphocytes from immunized donors demonstrated that ACC continued to accumulate and persist in challenged loops but only appeared transiently in nonchallenged loops. These transients did not migrate from the lamina propria back into the lymph and they presumably died in situ. The increase in the number of ACC in loops which had been challenged with antigen was probably due both to cell division in the lamina propria and to the development of new ACC from recirculating lymphocytes which had been recruited into the loop. Thus, the cells which give rise to intestinal ACC can migrate into the lamina propria independently of antigen, but antigen has a profound effect on the location, magnitude, and persistence of the response.

Immunity to intestinal parasites

Intestinal parasites are common in man and animals and can cause severe disease. Knowledge of immunity to such infections is limited and comes largely from studies using laboratory host-parasite systems. Understanding how immunity can operate, and why it often does not, is not only of intrinsic interest but necessary for the development of immunological methods of control.

Selective induction of an immune response in human external secretions by ingestion of bacterial antigen

Ingestion of capsules which contained killed Streptococcus mutans by four healthy human subjects led to the appearance of specific antibodies in external secretions. Salivary and lacrymal antibodies were detected within 1 wk of ingestion and continued to increase throughout a 14-day immunization period, with a gradual decline during the 2 ensuing months. A second period of immunization resulted in a pronounced increase of specific antibody levels which occurred earlier than in the primary immunization period and reached peak levels by day 10. No change was detected in serum antibody levels throughout either immunization period. The antibody activity in all secretions was associated with the immunoglobulin A class, as determined by immunochemical analyses. These data indicate that ingestion of bacterial antigens selectively stimulates the immune response in secretions.

Development of the IgA system in the mammary gland

1) Lymphoblasts in gut-associated lymphoid tissue, committed to the production of IgA, can home to the mammary glands of syngeneic mice and differentiate there into IgA-containing plasmablasts. The phenomenon is limited to near term and lactating recipients. 2) The ability of lymphocytes originating in gut-associated lymphoid tissue and sensitized to intestinal antigens to migrate to the mammary gland can account for the specificity of milk IgA toward intestinal microorganisms and the consequent passive protection offered to suckling infants. 3) The secretory immune system of the mammary gland is apparently under hormonal control since mammotropic hormones given to virgin females can induce morphological and functional characteristics seen naturally only during pregnancy and lactation. Examples are increased numbers of IgA plasma cells and the ability to trap their circulating precursors taken from mesenteric lymph nodes.

Origin of IgA-secreting plasma cells in the mammary gland

Lymphoblasts from the mesenteric lymph nodes (MN) of mice home to the mammary glands of syngeneic recipients late in pregnancy and during lactation, and within hours of transfer most can be shown to contain IgA. Homing does not occur in virgins, in early pregnancy, or after weaning. Homing MN lymphoblasts are sensitive to antiserum to IgA plus complement, but not to other class-specific antisera. Thus, lymphoblasts in MN with the potential to home to the mammary gland are already committed to IgA synthesis and bear surface IgA before reaching their destination. These results explain observations, made by others, of specific IgA antibodies and IgA plasma cells in milk and colostrum after oral immunization. Under natural conditions it is likely that IgA precursor cells, after stimulation in the gut-associated lymphoid tissue by intestinal antigens, migrate to the mammary gland where they secrete antibodies which constitute an important defense mechanism of the newborn. In the absence of lactation, these cells probably form part of the normal traffic to the lamina propria of the small intestine.

Indirect immunofluorescent identification of 19S immunoglobulin-containing cells in the intestinal mucosa of Xenopus laevis

The lymphoid structures in the gastrointestinal tract of immunized and non-immunized adult Xenopus laevis were studied by light and fluorescent microscopy. Serial sections stained with May-Grunnwald Giemsa showed that lymphoid aggregations and scattered lymphoid cells are present along the whole digestive tract. The aggregations are few and rather small in the oesophagus and stomach, they are particularly voluminous in the duodenum. An indirect immunofluorescent technique using antisera against Xenopus laevis 19S and 7S immunoglobulins and their heavy polypeptide chains, revealed the presence of immunoglobulin-containing cells in the duodenal region. The oesophagus and stomach were devoid of these. It has been shown that the immunoglobulins produced in the duodenal lamina propria are formed of both heavy and light polypeptide chains, and that the heavy chains are of the 19S H-type.

Transfer of immunity to Trichinella spiralis in the mouse with mesenteric lymph node cells: time of appearance of effective cells in donors and expression of immunity in recipients

Cells capable of transferring immunity to Trichinella spiralis, i.e. of accelerating adult worm expulsion, were present in the mesenteric lymph nodes of mice infected for 4, 6 or 8 days, but not in mice infected for only 2 days. The time-course of worm expulsion in mice infected on the day of transfer was similar in recipients of day 4 or day 8 cells, expulsion becoming marked only when the recipients had been infected for at least 6 days. Transfer of cells 4 or 6 days after infection did not result in an accelerated worm expulsion; transfer 1 or 2 weeks before infection did not enhance the level of immunity in recipient mice. In contrast to the results obtained with mesenteric lymph node cells (MLNC) on immunity was transferred when recipients were given spleen cells taken from donors infected for 8 days. It is suggested that MLNC do not cause worm expulsion directly, but cooperate with another component of the host's defence mechanism. Accelerated expulsion in recipients of cells was accompanied by a premature decline in fecundity of female worms. Evidence is presented to show that worm expulsion and impaired reproduction may represent independent aspects of the immune response to T. spiralis.

The immunological consequences of nematode infection

Nematode infections in the gut induce a strong immune response which is rapidly detected parenterally. The response is thymus-dependent and long-lasting and involves both antibodies and cell-mediated reactions. The immunological response to unrelated antigens, tumours and other infectious organisms is altered in animals infected with nematodes. Both antibodies and sensitized lymphocytes participate in the immune response which affects the nematodes themselves, and characteristically the lymphocyte-dependent step cannot act in lactating animals and is neither induced nor able to act in young animals. Present evidence suggests that, despite their well-known association with helminth infections, parasite rejection from the gut does not require the participation of IgE antibodies, mast cells or eosinophils. Homing of lymphoblasts from the mesenteric lymph node or thoracic duct lymph to the small intestine is increased in rats and mice infected with Nippostrongylus brasiliensis or Trichinella spiralis and the increase is antigenically non-specific. In mice infected with T. spiralis this increase is represented mainly by thymus-derived lymphoblasts.

The natural history of the cells producing IgA in the gut

The IgA-secreting cells in the lamina propria of the small intestine are derived from large lymphocytes which enter the blood by way of the thoracic duct and then migrate into the gut where they complete their differentiation into plasma cells. Three aspects of this cellular traffic have been examined in rats. 1. The cells in thoracic duct lymph which give rise to IgA-secreting cells in the lamina propria are among those which carry surface IgA. Blast cells lacking surface immunoglobulin migrate mainly into the Peyer's patches and do not contribute to the IgA response. 2. Studies on a secondary antibody response to cholera toxoid, in which the challenge was given into a Thiry-Vella loop, showed that the antibody-containing blast cells in thoracic duct lymph were derived from Peyer's patches. The mesenteric nodes contributed little, if anything, to the cellular response in the lymph. 3. The idea that secretory component is a signal for the emigration of large lymphocytes from the blood into the lamina propria lacks experimental support. Secretory component does not bind to the IgA on the surface of thoracic duct cells. On the other hand, antigen in the gut may play an important part in immobilizing large lymphocytes in the lamina propria once they have migrated.

Immunopathology of coeliac disease

Coeliac disease may be defined as a condition in which there is an abnormal jejunal mucosa with loss of villi, which improves morphologically after treatment with a gluten-free diet. Pathologically, there is damage to the jejunal enterocytes, with hyperplasia of crypt cells so that overall enteropoiesis is increased. On conventional or scanning electron microscopy the enterocytes are markedly abnormal. Histochemically, the normal punctate appearance of the lysosomes is lost and sensitive lysosomal enzyme assays on mucosal biopsy samples using isopycnic centrifugation techniques show that there is an increase in total lysosomal activity with reduction in lysosomal latency. Studies following gluten feeding in patients whose mucosa has returned to normal after treatment with a gluten-free diet show that pathological abnormalities appear within 4--8 hours of gluten challenge. Complement together with extracellular IgM can be demonstrated in the lamina propria, suggesting the formation of immune complexes. In untreated coeliac disease there is a significant reduction in serum levels of C3 and C4. There is also evidence indicating the presence of immune complexes in the serum. Coeliac disease may therefore be an intestinal model of an immune complex disease, in which an antigen derived from gluten reacts with an antibody formed locally in the gut, fixing complement and causing damage to the enterocyte by activation of lysosomes.

The secretory IgA system of the gut

Most commonly, humoral immunity manifested in the gastrointestinal tract of mammals is due to the presence of secretory IgA antibodies. Antibody specificities have been detected in the secretory IgA of gut secretions to a wide range of naturally occurring viral and bacterial components and to test antigens such as chemically modified proteins. Much of the IgA found in gut secretions is synthesized and secreted locally by the abundant plasma cells of the lamina propria. Development of methods for establishing local protective immunity in the gut requires knowledge of the origins of these plasma cells and of the whereabouts of their precursors when they are susceptible to antigen-driven proliferation and/or maturation. The Peyer's patches have been shown to contain a population of B lymphocytes especially rich in precursors for IgA plasma cells and in cells which can repopulate gut lamina propria with such IgA plasma cells. The Peyer's patches also appear to 'sample' gut antigens, in that small amounts of antigens are passed intact through their dome epithelial cells. Recent experiments bearing on the origins, differentiation and maturation, antigen sensitivity, migration and lodging of precursors for gut IgA plasma cells are discussed. We use the following three systems: (1) congenic transfer of cells from different murine lymphoid cell sources or mixtures of these (CB20 leads to BALB/c or BALB/c leads to CB20) and the use of allo-antisera to IgA allotypic determinants to assess their potential to impart an adoptive IgA antibody response to the recipient and to repopulate its histocompatible lamina propria with IgA plasma cells; (2) clonal precursor analysis (the method of Klinman) both to enumerate antigen-sensitive cells in different tissues of mice and to evaluate their potential to generate plasma cells making particular isotypes and idiotypes of antibodies; (3) use of pairs of Thiry-Vella loops in rabbits, each member either bearing or lacking a Peyer's patch, and quantitation of antibodies of each isotype and of total secretory IgA to assess the response of each loop with the time after local immunization. The results from all three systems provide strong evidence for the importance of Peyer's patches in supplying cells responsible for local humoral immunity and suggest both a differentiative pathway for IgA precursors and their whereabouts when antigen may cause the expansion of a population of specific cells.

Subpopulations of T-lymphocytes in Peyer's patches: sensitivity to antilymphocyte serum and adult thymectomy

Anti-lymphocyte (ALS) treatment or adult thymectomy of the donor have been shown to depress respectively the cell proliferation and the cytotoxicity in the graft-versus-host (GVH) reaction. A quantitative assay and the histological criteria of the GVH reaction have been used to demonstrate that all the known subpopulations of T-lymphocytes involved in the GVH reaction are present in the Peyer's patches as well as in the spleen and mesenteric lymph nodes in the rat.

Colostral cell-mediated immunity and the concept of a common secretory immune system

Historically colostrum and milk have been thought to confer immunity on the neonate only by virtue of their immunoglobulin content. Recently we have observed that colostrum also contains viable T lymphocytes capable of expressing cell-mediated immunity in vitro and have employed techniques of lymphocyte culture to elucidate the local nature of mammary tissue immunity at the T-cell level. The results indicate that the activity of colostral lymphocytes appears not to represent the total immunological experience of the mother but that they may contain reactive clones beneficial for the suckling. Colostral immunity appears to depend upon sensitizing events within the intestine and respiratory tract, followed by the migration of lymphoid precursors to the breast, suggesting a relationship between the expression of immunity at various secretory surfaces.

Immunoglobulin and specific-antibody synthesis in vitro by enteral and nonenteral lymphoid tissues after subcutaneous cholera immunization

An in vitro culture technique with synthesis of 14C-labeled protein has been used to study immunoglobulin and specific-antibody formation by spleen, mesenteric lymph nodes, Peyer's patches, and small intestine of rabbits, which were immunized twice subcutaneously with Vibrio cholerae lipopolysaccharide (LPS) and enterotoxin; saline-injected rabbits served as controls. Newly synthesized immunoglobulin G (IgG), IgA, and IgM were quantitated by liquid scintillation after their isolation by means of affinity chromatography from columns with immunoglobulin class-specific antibodies coupled to Sepharose. Specific antibodies were similarly measured after purification from gels derivatized with LPS or cholera toxin. The isolated antibodies had full biological activity as studied in protection tests. The immunization increased the overall IgM synthesis in the spleen. It also enhanced the production of IgA and IgG in Peyer's patches and of IgA in intestine. Significant synthesis of radiolabeled antibodies against both V. cholerae LPS and enterotoxin was found in spleen as well as in Peyer's patches of immunized animals. Titration with an enzyme-linked immunosorbent assay (ELISA) showed significant levels of IgG as well as IgA antibodies in incubation medium from all the studied tissues, whereas specific IgM was only found for spleen and mesenteric lymph nodes. Simultaneous tissue incubations at 37 degrees C and in an ice bath indicated that the major part of the antibodies registered with the ELISA represented de novo synthesis.

Studies on the lymphocytes of sheep. III. Destination of lymph-borne immunoblasts in relation to their tissue of origin

Lymph-borne immunoblasts were labeled in vitro with 125I[]iodo-deoxy-uridine, washed and returned by intravenous injection to the sheep from which they had been collected. Twenty h later the sheep were killed and the distribution of the immunoblasts was determined by assaying the radioactivity in various organs. Immunoblasts from the efferent lymph of peripheral somatic lymph nodes (PSLN) went mainly to the spleen, lungs and other PSLN, while immunoblasts from intestinal lymph went mainly to the small gut. This ability of intestinal immunoblasts to home to the gut was demonstrated also in the sterile environment of fetuses in utero; apparently the migratory behavior of immunoblasts, like that of small lymphocytes, is not primarily "antigen-driven". A technique was devised for the collection of peripheral (i.e. afferent to the mesenteric node) intestinal lymph which was found to contain 10-20 times the numbers of small lymphocytes that occur in the peripheral lymph from other tissues. Immunoblasts from peripheral intestinal lymph also homed to the gut. The immunoglobulin content of immunoblasts was studied by making detergent extracts of lymph cells, by applying immuno-peroxidase techniques to cell films and by investigating the incorporation of 14C-labeled amino acids into immunoglobulins by immunoblasts in vitro. Immunoblasts from both somatic and intestinal lymph contained and made IgG and IgM, but many intestinal immunoblasts contained and made IgA. It is not known whether this immunoglobulin mediates the extravasation of immunoblasts into the gut. Nonetheless, there is compelling evidence that there are two major migratory pathways for lymphoid cells; one through the gut-associated lymphoid tissue and the other through the somatic-splenic lymphoid tissues.

Naturally occurring secretory immunoglobulin A antibodies to Streptococcus mutans in human colostrum and saliva

Human colostrum, parotid saliva, and serum were assayed for the presence of naturally occurring antibodies to five serotypes of Streptococcus mutans. Appreciable levels of agglutinins to strains AHT, BHT, 10449, 6715, and LM-7 (groups a leads to e, respectively) were detected in normal colostrum and saliva, whereas relatively low levels were found in serum. No agglutinins could be detected in the colostrum or saliva of immunodeficient patients. Molecular sieve chromatography of the colostrum on Sephadex G-200 revealed agglutinin activity in the secretory immunoglobulin A (s-IgA)-rich fraction only. Titration of purified colostral s-IgA confirmed the IgA nature of this agglutinating activity. Indirect immunofluorescence tests with anti-s-IgA, -IgG, and -IgM revealed S. mutans specificity only in the s-IgA class. The presence of s-IgA antibodies to indigenous oral microorganisms in colostrum, as well as in saliva, suggests that antigenic stimulation occurs at a site remote from the oral mucosa.

Evaluation of the intraepithelial lymphocyte count in the jejunum in childhood enteropathies

Intraepithelial lymphocyte counts were evaluated in 131 jejunal mucosal biopsies taken from children with a small intestinal enteropathy arising from a variety of causes including coeliac disease, (untreated, after gluten withdrawal, and during subsequent challenge), giardiasis, cow's milk protein intolerance, and 'intractable diarrhoea'. The counts were compared with those from the biopsies of children referred for investigation but in whom no gastrointestinal disease was demonstrated and from healthy siblings of children with coeliac disease, investigated during a family study. Children with coeliac disease showed a raised count which fell after gluten withdrawal as has been demonstrated by others in adults. Lymphocytic infiltration of the epithelium increased rapidly during gluten challenge in such children, while no change was seen in those children proven ultimately not to have coeliac disease by the usually recognized criteria. In other enteropathies the range of counts was wide, overlapping with both normal and coeliac groups and indicating the nonspecificity of lymphocytic infiltration of the gut epithelium. The findings are discussed in relation to their significance and to further avenues of investigation to determine their possible diagnostic value in confirming the diagnosis of coeliac disease during gluten challenge.

Peyer's-patch-associated synthesis of immunoglobulin in germ-free, specific-pathogen-free, and conventional mice

The synthesis of immunoglobulins G, A, and M has been studied in Peyer's patches together with closely associated intestinal mucosa and in small intestine distant from Peyer's patches in specific-pathogen-free (SPF) Swiss mice and conventional and germ-free C3H mice. Thissue fragments were cultured in vitro in medium containing 14C-labelled amino acids, and newly synthesized proteins were detected by radioimmunoelectrophoresis. Small intestine from SPF and conventional animals synthesized almost exclusively IgA. No immunoglobulin synthesis was detectable in germ-free intestine. In contrast, the Peyer's patches and associated mucosa of all the groups of mice synthesized IgG, IgA, and IgM. This observation is discussed in relation to the possible role of the Peyer's patches as a source of precursors for immunoglobulin-producing cells in the intestine.

B-lymphocyte heterogeneity: ontogenetic development and organ distribution of B-lymphocyte populations defined by their density of surface immunoglobulin

The density of total Ig and of IgM, IgG1, IgG2, and IgA on the surface of adult murine splenic B lymphocytes was measured using the technique of rapid flow microfluorometry. In addition, the density of total surface Ig and of IgM on B lymphocytes derived from adult bone marrow, lymph nodes, and Peyer's patches, and from neonatal spleen was determined. Adult spleen and lymph node B lymphocytes are characterized by the presence of a large population of cells with a low-to-intermediate density of total surface Ig, which is seen as a peak in the fluorescence profiles when these cells are labeled with fluorescein-conjugated (F1) anti-Ig. This peak is not detected when neonatal spleen or adult bone marrow are examined; the development of this peak among spleen cells occurs during the first 4 wk of life. Although the characteristic fluorescence intensity peak is not seen when adult spleen cells are labeled with Fl anti-mu, changes in the density of surface IgM do occur during the first few weeks of life and are detected as a decrease in the frequency of cells which have relatively large amounts of surface IgM. The differences seen in the fluorescence patterns of adult spleen cells labeled with Fl anti-Ig and Fl anti-mu may be due to the appearance of IgD on the surface of mature splenic B lymphocytes. This is supported by the similarity of the fluorescence profiles of adult bone marrow cells stained with Fl anti-Ig and Fl anti-mu, as the latter population of cells is reported to lack surface IgD.

Ingestion of Streptococcus mutans induces secretory immunoglobulin A and caries immunity

Ingestion of killed cells of a highly cariogenic strain of Streptococcus mutans induced specific antibodies in both saliva and milk but not in serum of gnotobiotic rats. These antibodies were associated with the immunoglobulin A class. When infected with Streptococcus mutans, orally immunized animals developed significantly fewer carious lesions than nonimmunized infected controls.

Cortisone sensitive T-cells in Peyers' patches

Pretreatment of donor lymphoid cells with cortisone has been shown to depress the T-cell subpopulation responsible for cellular proliferation in the GVH reaction. A quantitative assay as well as the histological criteria of the GVH reaction have been used in this study to demonstrate the presence of cortisone-sensitive T-cells within the Peyer's patches as well as in the spleen and mesenteric lymph nodes in the rat.

Kinetics of intraepithelial lymphocytes in the small intestine of thymus-deprived mice and antigen-deprived mice

The kinetics of lymphoid cells within the epithelium of the small gut has been studied in various thymus-deprived mice and in antigen-deprived mice by the use of 3H-thymidine injections and radioautography. In thymus-deprived mice--including adult thymectomized, thymectomized and irradiated, neonatally thymectomized, and nude mice - and in germ-free mice decreased numbers of intraepithelial lymphocytes (IL) were found. On the other hand, the radioautographic results indicated that the remaining IL populations included both newly formed and long-lived lymphoid cells in the same percentages as found in sham-operated controls and normal mice. It is concluded that although the presence of the thymus and the antigen content of the gut is of importance to the maintenance of the numbers of cells in the lymphoid populations of the intestinal wall, the basic kinetics of these cell populations are preserved in deprived mice.

Proliferative kinetics of large and small intraepithelial lymphocytes in the small intestine of the mouse

The proliferative kinetics of the intraepithelial lymphocytes (IL) of the mouse intestine have been evaluated. By inducing mitotic arrest it was found that large IL - constituting about 50% of the IL - showed a mitotic rate of 2.3. Autoradiographic results obtained after two different schedules of 3H-thymidine injections showed that 30% of the large IL were in DNA synthesis, and that the large IL were renewed at a rate comparable to that of blast cells from Peyer's patches, mesenteric lymph nodes and thoracic duct lymph. The small IL were renewed very rapidly compared to small lymphocytes of peripheral lymphoid tissues, although small lymphocytes with lifespans of several weeks were also present in the epithelial sheet. By the use of intestinal perfusion, in vivo, it was estimated that the loss of lymphocytes from intestinal villi into the lumen of the gut was negligible, and it is concluded that the most probable kinetic model for the majority of IL is: B and T lymphoblasts invade the epithelium and undergo mitosis. B lymphoblasts give rise predominantly to plasma cells, and T lymphoblasts give rise to small lymphocytes - probably long-lived - which reenter the circulation.

Factors which determine the accumulation of immunoblasts in gut and skin

The capacity of immunoblasts from two sources (1) peripheral lymph nodes draining the site of application of a contact sensitizer and (2) mesenteric lymph nodes from mice infected with the gut parasite T. spiralis to migrate to the gut and to inflamed skin sites were compared. The peripheral lymph node blasts readily entered skin sites in a non-specific way but failed to migrate to the gut even when inflammation was induced. By contrast, the mesenteric lymph node blasts readily migrated to the gut in normal mice and in increased amounts to the gut of mice infected with T. spiralis or inflamed with oral turpentine. A small proportion of mesenteric lymph node blasts did, however, migrate, non-specifically to the skin but in much smaller amounts than peripheral lymph node blasts. We conclude that the migration of immunoblasts to the gut has some specificity related to the source from which the cells were taken but little specificity with regard to intraluminal antigen.

Hypersensitivity reactions in the small intestine. 2. Effects of allograft rejection on mucosal architecture and lymphoid cell infiltrate

Small intestinal mucosa contains both thymus dependent and thymus independent lymphoid cells and thus has the capacity to act via humoral and cellular mechanisms as a site of local immunity and local hypersensitivity. Allograft rejection of mouse small intestine is a model of a local cell mediated reaction. The effects of this clearly defined, immunologically mediated damage villi, crypts, enterocytes, and lymphoid cell infiltrate have been assessed by comparing the morphology of rejecting allografts with that of isografts and normal small intestine of the same age. In rejection there is infiltration of the lamina propria with lymphocytes, hyperplasia of the crypts of Lieberkuhn, and an eventual sloughing off of the mucosa. Usually, but not always, there is villous atrophy and increased numbers of intraepithelial lymphocytes. However, the morphology of individual enterocytes remains normal throughout rejection and neither plasma cells nor polymorphonuclear leucocytes infiltrate the lamina propria before mucosal ulceration. These results show unequivocally that a local T cell mediated immune response causes villous atrophy and crypt hyperplasia in this animal model, and since there is no evidence of local enterocyte cytotoxicity, a lymphokine may be the link between the activated T cell and the effects on mucosal architecture. We suggest that a local CMI reaction may be the cause of villous atrophy, crypt hyperplasia, and malabsorption in many clinical and experimental conditions, including coeliac disease, food allergy, and intestinal infections.