Normal Structure,Function,and Histology of Mucosa-Associated Lymphoid Tissue

Abstract

The mucosa-associated lymphoid tissue (MALT) initiates immune responses to specific antigens encountered along all mucosal surfaces. MALT inductive sites are secondary immune tissues where antigen sampling occurs and immune responses are initiated. Effector sites, present as diffuse lymphoid tissue along all mucosal surfaces are the sites of IgA transport across the mucosal epithelium. Though there are many differences between inductive sites in various organs, they all contain the same basic compartments—follicles, interfollicular regions, subepithelial dome regions, and follicle-associated epithelium. The morphologic differences between MALT and other secondary lymphoid tissues, between the MALT sites of differing anatomic locations, and species differences among laboratory animals are described. The morphologic changes in MALT associated with aging, route of nutrition, and genetic mutation (i.e., the nude and SCID mutations) are also discussed. MALT tissues comprise the mucosal immune system which can function independently of the systemic immune system and are, therefore, an important and often overlooked aspect of immunopathology.

Introduction

About half the lymphocytes of the immune system are in the Mucosa-associated lymphoid tissue (MALT) (Croitoru and Bienenstock, 1994). MALT is situated along the surfaces of all mucosal tissues. Its most well-known representatives are gut-associated lymphoid tissue (GALT), nasopharynx-associated lymphoid tissue (NALT), and bronchus-associated lymphoid tissue (BALT); however, conjunctiva-associated lymphoid tissue (CALT), lacrimal duct-associated (LDALT), larynx-associated (LALT) and salivary duct-associated lymphoid tissue (DALT) have also been described. The main function of MALT is to produce and secrete IgA across mucosal surfaces in antigen specific, T_h2-dependent reactions, though T_h1 and cytotoxic T-cell mediated reactions can also occur, the later resulting in immunotolerance (Gormley et al., 1998; Kiyono and Fukuyama, 2004).

MALT can be functionally divided into effector sites and inductive sites. GALT, BALT and NALT, CALT in mice, dogs, (Giuliano et al., 2002) and baboons (Astley et al., 2003), and DALT in cynomolgus macaques (Nair and Schroeder, 1986; Sakimoto et al., 2002) are inductive sites. Inductive sites contain secondary lymphoid tissues in which IgA class switching and clonal expansion of B-cells occurs in response to antigen specific T-cell activation. After activation and IgA class switching, T- and B-cells migrate from inductive sites to effector sites. Effector sites are present in all mucosal tissues as disseminated lymphoid tissue diffusely distributed throughout the lamina or substantia propria (Yan et al., 2003). In effector sites, secretory IgA, or S-IgA (2 IgA molecules joined by a J-chain and bound to secretory component, an epithelial cell membrane receptor) is secreted across the mucosal epithelium (Pabst, 1987). In rodents, hepatocytes also produce secretory component, so S-IgA also enters the gut lumen via the bile duct (Pabst, 1987). The cellular composition of effector sites includes T-cells, the majority of which are CD4⁺, IgA plasma cells with fewer IgG and IgM plasma cells, and few B-cells, dendritic cells, and macrophages (Pabst, 1987; Kelsall and Strober, 1999; MacDonald, 2003). Though MALT sites are anatomically separated, they are functionally connected in what has been termed the “common mucosal immune system,” so that antigen presentation and B-cell activation at one mucosal site can result in IgA secretion at mucosal sites of different organs (Bienenstock et al., 1999; Hiroi et al., 1998; Kiyono and Fukuyama, 2004; Kuper et al., 2002). Furthermore, the mucosal immune system can act independently of the systemic immune system making evaluation of MALT an important aspect of immunopathology (Kuper et al., 2002).

Intraepithelial lymphocytes, T-lymphocytes found amid epithelial cells of all mucosal tissues, are another component of MALT. They are predominantly CD8⁺ and are relatively abundant, typically with 1 lymphocyte per 4–6 epithelial cells (Kelsall and Strober, 1999; Kuper et al., 2002; MacDonald, 2003; Pabst et al., 2005). They are somewhat unique, particularly in rodents, in that a large proportion of them express the γδ TCR and many express the CD8αα homodimeric form of the CD8 receptor rather than the αβ TCR and CD8αβ heterodimer found in most other CD8⁺ T-cell populations (Eberl and Littman, 2004; Kelsall and Strober, 1999; Pabst et al., 2005; Saito et al., 1998). They also uniquely express the α_Eβ₇ integrin, which is thought to be important for sequestering these lymphocytes in the epithelium (Kelsall and Strober, 1999; MacDonald, 2003). Evidence suggests that these cells are generated in extrathymic locations in the intestinal mucosa, but this remains controversial (Saito et al., 1998).

This paper focuses on the normal morphology of the inductive sites of BALT, GALT, and NALT, and these terms will be used to refer to these secondary lymphoid tissues. The functional compartments of these tissues are the lymphoid follicles, the interfollicular region, the subepithelial dome region, and the overlying follicle-associated epithelium, or lymphoepithelium, which contains M-cells (Figure 1). M-cells are specialized epithelial cells within the follicle-associated epithelium that transport microorganisms, and macro- and soluble molecules from the intestinal lumen to the subepithelial dome region giving the lymphoid tissues access to luminal antigens (Gebert and Pabst, 1999; Kelsall and Strober, 1999). They are difficult to distinguish light microscopically but have a distinct ultrastructural appearance. The apical surface of M cells is characterized by small, irregular, disorganized microvilli with a poorly developed brush border or microfolds rather than the dense brush border of absorptive enterocytes (Gebert and Pabst, 1999; Kelsall and Strober, 1999). The basal membrane of M cells is invaginated, forming a pocket that typically contains one or more lymphocytes (T-cells or B-cells) and occasionally macrophages (Gebert and Pabst, 1999; Pabst, 1987). Since there are no specific histochemical or immunohistochemical markers for M cells, clusters of lymphocytes in the follicle-associated epithelium may be the only light microscopic clue to the presence of an M cell. Other common features of MALT include the lack of afferent lymphatics and medullary regions and the presence of high endothelial venules (Figure 5) within the interfollicular regions. The cellular components of MALT include B-cells, CD4⁺ and CD8⁺ T-cells, antigen-presenting dendritic cells, macrophages, and, occasionally, mast cells and eosinophils in the interfollicular region. Thus, they contain all the cell types necessary to initiate an immune response.

Gut-Associated Lymphoid Tissue

Several types of lymphoid nodules have been described in the intestine, including Peyer’s patches, isolated lymphoid follicles, cryptopatches, and, in the large intestine, lymphoglandular complexes. Peyer’s patches and lymphoglandular complexes are the primary inductive sites in the gut, but the functions of the isolated lymphoid follicles and cryptopatches are unclear.

Peyer’s patches (Figures 1–5) are randomly distributed throughout the mucosa and submucosa of the gastrointestinal tract but are of greatest density in the jejunum and are oriented along the anti-mesenteric border (Schuurman et al., 1994). The Peyer’s patch follicles in rodents typically contain 6–12 basally located germinal centers, which are more numerous in Peyer’s patches than in either NALT or BALT (Haley, 2003). B-cell areas are larger than T-cell areas, which is reflected in the T cell:B cell ratio of 0.2 (Sminia and Kraal, 1999). The size, number, distribution, and composition of Peyer’s patches may vary depending on species or strain. For example, Peyer’s patches are generally smaller in Fischer 344 rats than in Wistar rats (Bruder et al., 1999). The ratio of CD4⁺ to CD8⁺ T-cells of 5.0 is approximately 2-fold higher in Peyer’s patches than in NALT or BALT (2.4 and 2.6, respectively) (Sminia and Kraal, 1999). In Lewis rats, however, the ratios are nearly equal (Kuper et al., 1992).

Isolated lymphoid follicles, also located on the antimesenteric border of the small intestine, have been identified in the rat, mouse, rabbit, and guinea pig (Hamada et al., 2002; Lorenz and Newberry, 2004; Pabst et al., 2005). They are smaller than Peyer’s patches with an average diameter of 150 microns and appear as barrel-shaped lymphoid aggregates in shortened intestinal villi and are typically associated with a single dome (Pabst et al., 2005). They are very similar to Peyer’s patches having 1–2 B-cell follicles that may contain germinal centers and a small population of CD4⁺ T-cells, an overlying follicle-associated epithelium with M-cells, and scattered dendritic cells with few macrophages, (Lorenz and Newberry, 2004; Pabst et al., 2005). Up to 200 isolated lymphoid follicles per mouse have been identified, though this number may vary depending on mouse strain. Hamada et al. (2002) identified 150–200 per mouse in BALB/cA/Jcl and 100–150 per mouse in C57BL/6J/Jcl. Isolated lymphoid follicles are not present in neonatal mice, becoming consistently detectable in the duodenum and proximal jejunum by light microscopy at 7 days of age in BALB/cA/Jcl and at 25 days of age in C57BL/6J/Jcl (Hamada et al., 2002). According to Lorenz and Newberry, C57BL/6 mice have fewer and smaller isolated lymphoid follicles that are located predominantly in the distal small intestine and are not confined to the antimesenteric border (Lorenz and Newberry, 2004).

Cryptopatches, lymphoid aggregates found in the intercryptal lamina propria of the small intestine, are small aggregates of T-cells and dendritic cells with an average diameter of 80 μm (Eberl and Littman, 2004; Mowat and Viney, 1997; Pabst et al., 2005). They form after weaning and are far more numerous than isolated lymphoid follicles and Peyer’s patches with up to 1700 per adult mouse (Eberl and Littman, 2004). Studies suggest that cryptopatches are primary lymphoid tissue in which extrathymic generation of intraepithelial lymphocytes occurs, though this remains controversial (Eberl and Littman, 2004; Saito et al., 1998).

Lymphoglandular complexes (Figures 6 and 7) in the colon resemble Peyer’s patches, however, they are smaller and have fewer follicles with smaller germinal centers (Owen et al., 1991). Also, crypts extending into the colonic submucosa that are lined by follicle-associated epithelium and surrounded by lymphoid tissue may occasionally be evident (Figure 6) (Owen et al., 1991). In the mouse distal colon, lymphoglandular complexes are randomly distributed with an average of 1.4 patches per centimeter of colon (Owen et al., 1991). Lymphoglandular complexes in the proximal colon of the mouse and throughout the colon in the rat are oriented toward the antimesenteric border (Deasy et al., 1983; Morfitt and Pohlenz, 1989; Perry and Sharp, 1988). In mice, at least one lymphoglandular complex, the rectal patch, can consistently be found within 10 mm of the anus (Owen et al., 1991). The cecal patches, large aggregates of lymphoid follicles in the proximal cecum opposite the ileocecal valve, also have a consistent location (Owen et al., 1991). In Fisher rats, the “proximal colonic lymphoid tissue” is a lymphoid nodule that is consistently present in the submucosa of the proximal colon roughly at the distal end of the first fifth of the colon (Crouse et al., 1989).

Dogs have a total of 26–29 Peyer’s patches (HogenEsch and Hahn, 2001). They have two types of Peyer’s patches, as opposed to rats and mice that have uniform Peyer’s patches (Haley, 2003). In the jejunum and upper ileum, the Peyer’s patches of dogs are smaller and more discrete (similar to those of mice and rats), while in the terminal ileum, there is a 26–30 cm long Peyer’s patch that completely encircles the distal 6–10 cm of the ileum and narrows proximally to a 1-cm-wide band on the antimesenteric border (Haley, 2003; HogenEsch and Hahn, 2001). The ileal Peyer’s patch has small domes and interfollicular regions and an inconspicuous corona relative to those of the duodenum and jejunum (HogenEsch and Felsburg, 1992). The duodenal Peyer’s patches differ in that they have intrafollicular invaginations of the dome epithelium (HogenEsch and Felsburg, 1992). The dome regions in dogs contain more plasma cells than the dome regions in rats or mice (Haley, 2003). Dogs have pinpoint to >2 mm diameter lymphoid nodules throughout the gastric lamina propria, most numerous in the fundic region, but these are not associated with a follicle-associated epithelium (HogenEsch and Hahn, 2001; Kolbjornsen et al., 1994). In rhesus macaques, ileal Peyer’s patches are larger than those in the jejunum, duodenum, or colon (Veazey et al., 1997). In comparison to rat Peyer’s patches, those of Baboons are smaller and they lack the IgA⁺ centroblasts associated with the high endothelial venules as described by Spencer et al. (Spencer et al., 1986). Rabbits are unique in that they have an appendix, represented as a large collection of lymphoid nodules at the ileocecal valve, as well as a sacculus rotundus, which is a large, circumferential Peyer’s patch in the terminal ileum (Haley, 2003).

Nasopharynx-Associated Lymphoid Tissue

NALT (Figures 8–12), found in rats, mice, hamsters, and non-human primates, is composed of paired lymphoid aggregates in the caudoventral portion of the left and right nasal passages at the entrance to the nasopharyngeal duct (Figure 8) (Spit et al., 1989). It is visible in the caudal portions of level II and throughout level III of the nasal cavity as sectioned for toxicological studies performed by the National Toxicology Program (Herbert and Leininger, 1999). NALT is considered the rodent equivalent to Waldeyer’s ring, the oro- and nasopharyngeal lymphoid tissues (tonsils) found in humans and some other species (Heritage et al., 1997). Though similar in appearance, there are a number of differences between NALT and Peyer’s patches. There are fewer intraepithelial lymphocytes In NALT (Sminia and Kraal, 1999). The relative sizes of the B- and T-cell areas are roughly equal to each other in NALT, which is reflected in the higher T:B cell ratio of 0.9 (Sminia and Kraal, 1999). Plasma cells are present predominantly in the connective tissues deep to the NALT (i.e., away from the nasal passage) (Sminia and Kraal, 1999). The efferent lymphatic vessels and high endothelial venules extend deep into the interfollicular region to the follicular margins, the former confined to the basilar portion (Bienenstock and McDermott, 2005; Kuper et al., 2002; Sminia and Kraal, 1999). Though macrophages and dendritic cells are scattered throughout, dendritic cells in the subepithelial dome region are less numerous than in Peyer’s patches (Sminia and Kraal, 1999). There are no significant species differences in NALT, except that NALT in nonhuman primates is more extensive than in rodents, extending to the lateral surface of the nasal cavity (Haley, 2003).

Bronchus-Associated Lymphoid Tissue

There is a great deal of species variability in BALT. For example, BALT is normally absent in dogs, cats, and Syrian hamsters (Brownstein et al., 1980; Pabst and Gehrke, 1990). Rabbits typically have the most BALT in regard to the number of BALT sites, followed by rats, guinea pigs, and mice (Pabst and Gehrke, 1990). BALT is absent in germ-free pigs, while germ-free rats have BALT, though much less than in their conventionally reared counterparts (Bienenstock et al., 1973; Pabst and Gehrke, 1990). The situation in mice (and humans) is somewhat more controversial. There are conflicting reports on the presence of BALT in germ-free mice (i.e., in the absence of antigenic stimulation). Bienenstock and McDermott have reported the presence of BALT in germ-free mice (Bienenstock and McDermott, 2005; Bienenstock et al., 1999), but others report that BALT is not present in germ-free in mice (Moyron-Quiroz et al., 2004; Seymour et al., 2006). Furthermore, Pabst and Gehrke reported the presence of BALT in only 43% of 4-month-old SPF mice (Pabst and Gehrke, 1990), so the presence of BALT in mice in general appears to be rather variable. Regardless of this issue, BALT can be induced in mice by exposure to pathogens, and this BALT, referred to as inducible BALT, or iBALT, by some, has characteristics similar to BALT in other animals (Moyron-Quiroz et al., 2004).

BALT (Figures 13–16) is randomly distributed along the airways but is most consistently located at sites of bronchial tree bifurcation and is usually located between a bronchus and an artery (Figure 13). Of the 3 main MALT sites, BALT is the most divergent. The follicle-associated epithelium of BALT contains the fewest intraepithelial lymphocytes (Sminia and Kraal, 1999). Germinal centers and tingible body macrophages are less common in BALT than either NALT or Peyer’s patches (Sminia and Kraal, 1999). Follicular dendritic cells, identified by ED5 staining, have been detected in GALT and NALT, but not BALT (Kuper et al., 1992). The divisions between the B- and T-cell areas are less conspicuous than in Peyer’s patches and NALT and, in the rat, there does not appear to be any consistent spatial relationship between the two regions nor to surrounding structures (e.g., bronchus or artery) (Breel et al., 1988; Sminia and Kraal, 1999). Similarities do exist, however. For example, the relative sizes of the Band T-cell areas in BALT are roughly equal, and the T:B cell ratio of 0.7 is comparable to that of NALT (Sminia and Kraal, 1999). The general structure and cellular composition of BALT is also similar to NALT (Bienenstock and McDermott, 2005).

Tonsils

Tonsils are secondary lymphoid organs located in the oro- and nasopharynx of most species except rodents. Dogs have 4 sets of tonsils, the lingual (diffuse lymphoid tissue at the caudodorsal base of the tongue), palatine (on the lateral wall of the pharynx just caudal to the palatoglossal arch), pharyngeal (on the roof of the nasopharynx), and tubal (at the openings of the auditory tubes) tonsils. Primates have at least 3 sets of tonsils (in addition to NALT as described above), the lingual, palatine, and pharyngeal, and may have tubal tonsils (as do humans and all other domestic animals except rodents), but descriptions of tubal tonsils in these species are lacking. Two types of tonsils have been described, those with crypts (follicular tonsils) and those without. Tonsilar crypts are blind, often branched invaginations of the surface epithelium into the submucosal lymphoid tissue. Tonsils without crypts typically have a slightly folded surface epithelium and bulge into the oro-or nasopharynx (Banks, 1993). They are structurally similar to Peyer’s patches but differ in the overlying epithelium.

There are two types of epithelium overlying the tonsils: reticular and non-reticular (Belz and Heath, 1995). Reticular epithelium is spongy in appearance and is located over the apices of lymphoid follicles (Belz and Heath, 1995). It contains M-cells, numerous lymphocytes, macrophages and dendritic cells, and may contain a few granulocytes (Belz and Heath, 1995; Bernstein et al., 1999). The non-reticular epithelium, which separates islands of reticular epithelium, is stratified squamous in the oropharynx (palatine and lingual tonsils) and respiratory in the dorsal and some lateral surfaces of the nasopharynx (pharyngeal and tubal tonsils). The palatine tonsil is located in the palatine fossa, partially covered by the semilunar fold.

Nonpathologic Factors Affecting MALT Morphology

Age

In general, MALT is relatively undeveloped at birth with low cellularity. Upon stimulation after birth by exposure to environmental antigens, especially the developing gut microflora, T-cell areas expand and become fully populated and follicular germinal centers develop (as previously stated, germinal centers are frequent in GALT, but relatively infrequent in NALT and BALT). There are several differences in the ontogeny of the MALT sites. In fetal animals, BALT is generally absent, but is present by 4 days of age, initially as a few lymphocytes within a reticular stroma (Hameleers et al., 1989; Pabst and Gehrke, 1990; Sminia and Kraal, 1999). In contrast, Peyer’s patches are present before birth and NALT is present at birth (Sminia and Kraal, 1999). T- and B-cell regions do not develop in BALT until 3–4 weeks of age, whereas distinct B- and T-cell areas are discernible at 10 days of age in both NALT and GALT (Hameleers et al., 1989; Sminia and Kraal, 1999). Last, Peyer’s patches and NALT are initially composed of T-cells, whereas B-cells are the first lymphocytes to populate BALT (Sminia and Kraal, 1999).

Though much of the work on the effects of aging on the mucosal immune system has focused on changes in antibody production, evidence suggests that aging also alters lymphocyte proliferative capacity (Taylor et al., 1992; Schmucker et al., 2001; Schmucker, 2002). With age, the number of Peyer’s patches does not change significantly, however, in aged mice, but not rats, the Peyer’s patches do have fewer T-cells in the interfollicular and parafollicular regions relative to younger animals (Kawanishi and Kiely, 1989; Schmucker, 2002; Schmucker et al., 2001). The number of antibody-containing plasma cells in the Peyer’s patches of mice and lamina propria of rats intraduodenally immunized with cholera toxin decreases with age (Haaijman et al., 1977; Taylor et al., 1992). There may also be decreased numbers of B-cells and changes in the distribution of lymphocyte subsets in the intestine, but evidence regarding changes in the number of dendritic cells is conflicting (Schmucker, 2002; Schmucker et al., 2001). None of these studies, however, reported detectable, light microscopic changes, so even though cell numbers may be decreased, whether or not this is detectable morphologically is unknown.

In dogs, Peyer’s patch precursors can be seen as small aggregates of lymphocytes in intestinal villi in 45 day old fetuses, but lymphoglandular complexes first appear in the cecum 1 week after birth (HogenEsch and Hahn, 2001). In 50-day-old fetuses, the Peyer’s patches are much more well developed with domes and submucosal follicles, however, germinal centers do not form until 1 week of age (HogenEsch and Hahn, 2001). The ileal Peyer’s patch expands rapidly after birth, reaching maximal size by 6 months of age (HogenEsch and Hahn, 2001). Upon reaching sexual maturity, the follicles of the canine ileal Peyer’s patch become markedly reduced in size (HogenEsch and Hahn, 2001).

Route of Nutrition

Administration of total parenteral nutrition (TPN) has been shown to affect GALT, decreasing lymphocyte numbers in both rats and mice (Tanaka et al., 1991; King et al., 1997). In mice, both B and T cells in Peyer’s patches and lamina propria and the intraepithelial lymphocytes decreased significantly after 2 days of receiving TPN with the maximal decrease in the lamina propria occurring after 2 days and in the Peyer’s patches after 3 days (King et al., 1997). These Peyer’s patches and lamina propria changes were shown to be reversible in mice after returning to enteral nutrition and returned to control levels after 2 days (King et al., 1997). The intraepithelial lymphocytes responded much more slowly and by day 5, had not yet returned to normal levels (King et al., 1997). Histologic evaluation of rats receiving TPN revealed markedly atrophied Peyer’s patches and restriction of intraepithelial lymphocytes to the basal third of the intestinal villi (some intraepithelial lymphocytes were noted in the distal two-thirds of the villi in control rats) (Tanaka et al., 1991).

MALT in Immunodeficient Mutant Mice and Rats

There are a number of immunodeficient mutant mice and rats, both naturally occurring and genetically engineered, in which the morphology of MALT is altered. In SCID mice (Figures 21 and 22), which are deficient in both B- and T-cells, the lymphoid structures of the gut and lung are very small (almost nonexistent) and disorganized with sparsely scattered lymphoid cells of variable size, often appearing immature (Custer et al., 1985). Effector sites are similarly affected, as exemplified by the relative lack of immune cells in the lamina propria of the intestine. This phenotype, however, may vary with background strain and age, as the SCID mutation is “leaky” and these mice may develop more B- and T-cells later in life. Athymic nude rats and mice (Figures 17–20), being T-cell deficient, have decreased numbers of intraepithelial lymphocytes and depleted T-cell regions (interfollicular/parafollicular regions) (Hanes, 2005; Rocha et al., 1994) which, in nude rats at least, have increased numbers of macrophages. Also, since the formation of germinal centers requires T-cell activity, nude rats and mice are typically devoid of germinal centers (Hanes, 2005). Other components of the lymphoid structures in these animals are variably present. For example, in both SCID and nude animals, the stromal cells of the lymphoid structures are comparable in number and appearance to those of immunocompetent strains (Custer et al., 1985; Hanes, 2005). RAG1−/− mice, a genetically engineered mutant deficient in B- and T-cells similar to SCID mice, have limited follicle associated epithelium and fewer M cells, but both are present and associated with small Peyer’s patch anlagen (MacDonald, 2003).