Abstract
The subcompartmentalization of the white pulp in the spleen is the result of interactions of specific resident stromal cells and migrating subtypes of lymphocytes. Because carbohydrate residues of cell membranes and extracellular matrices are involved in cell–cell and cell–matrix interactions, they were investigated in rat spleen by a broad panel of lectins. Splenic macrophages, which were also demonstrated by Perls' Prussian blue reaction, were labeled selectively by most mannose-specific lectins and gave the characteristic distribution patterns in all splenic (sub)compartments. One recently isolated lectin, Chelidonium majus agglutinin (CMA), visualized predominantly central arterioles, the reticular meshwork (RM) in the periarteriolar lymphatic sheaths (PALS), the circumferential reticulum cells limiting PALS and follicles, and some follicular dendritic cells (FDCs) in white pulp. The endothelial cells of venous sinuses in red pulp were also labeled by CMA and, if frozen sections were used, CMA also labeled the macrophages of the red pulp. Compared to CMA, the monoclonal antibody CD11, which can be used only in frozen sections, stained almost solely the fibrous (extracellular) component of the RM. Because CMA stains the reticulum cells in particular, it is better suited to visualize the stromal architecture of splenic white pulp than the monoclonal antibody. Because CMA can be applied to paraffin-embedded material, it is a particularly useful tool to study the splenic stromal architecture in archival material.
T
This functional subcompartmentalization of the splenic white pulp has been visualized by the use of monoclonal antibodies (MAbs) specific to various cell types: to T- and B-cells, their immune accessory elements, such as follicular dendritic cells (FDCs, B area-specific) or interdigitating dendritic cells (T area-specific), to the reticular cells, and to the extracellular matrix produced by them (Van den Berg et al. 1989; Yoshida et al. 1993; Wacker 1994). The reticulum cells and their matrix build up a reticular meshwork (RM) in the spleen, part of which embraces the lymphocytes in the white pulp. Lymphocyte traffic and homing in the white pulp are partially guided by this meshwork, which is therefore more than a bare connective tissue framework maintaining the structural integrity of this lymphatic tissue (Yoshida et al. 1991, 1993; Van den Berg et al. 1989).
In addition to MAbs, some cells of the immune system have been characterized by lectins (Strauchen 1984; Ree and Hsu 1983). Lectins are carbohydrate binding proteins, which recognize predominantly terminal carbohydrate residues of glycoconjugates. Lectins are therefore ideally suited to define cells and extracellular matrices by the composition of their glycoconjugates. This additional information is of particular interest in the immune system, in which carbohydrate residues serve as antigenic determinants and are also ligands for cell adhesion molecules that may direct lymphocyte traffic (Hughes 1992). Therefore, in this study we used various recently isolated lectins to determine the carbohydrate composition of the spleen.
Materials and Methods
Five male and female Wistar rats weighing 300–400 g were kept on a diet of Altromin 1314 (Altromin; Lage, Germany) which was supplemented with 200–220 mg Fe/kg. After sacrifice, the spleens were removed, cut into 1-mm slices, fixed in 4% paraformaldehyde buffered by 0.1 M phosphate buffer (pH 7.4), and routinely processed to paraffin. Sections were cut at 4 μm, deparaffinized in xylene, and hydrated through a series of graded alcohols. After incubation in Tris-buffered saline (TBS) containing 50 mM Tris, 150 mM NaCl, 1 mM MgCl2, 1 mM CaCl2, and HCl adjusted to pH 7.6, the sections were treated with 0.1 trypsin (Sigma; St Louis, MO) for 15 min at 37C. Sections were again transferred to TBS, washed three times, and then incubated with the biotinylated lectins at a concentration of 10 μg ml−1. All lectins were in-house preparations isolated according to previously described procedures (Van Damme et al. 1998). The origin of the lectins, their abbreviations, and nominal sugar specificities are listed in Table 1.
Cryostat sections (6 μm) were also cut, air-dried for 30 min, fixed in acetone for 10 min, and air-dried again. These sections were used for CMA staining and incubation with the MAb ED11, followed by a biotinylated anti-mouse antibody.
After a further wash in TBS, all sections incubated with the lectins or the biotinylated secondary antibody were treated with a biotin–streptavidin–alkaline phosphatase complex (Vectastain; Vector, Burlingame, CA), followed by a wash in TBS. Naphthol-AS-biphosphate was used as a substrate and hexatozized New Fuchsin was used for simultaneous coupling. Control sections were incubated the same way, omitting lectins or the primary antibody. Endogenous alkaline phosphatase activity in the frozen sections was blocked by levamisole. In the case of CMA, inhibition experiments using N, N′,N′′-triacetyl-chitotriose (Sigma) were carried out. CMA (10 μg ml−1) was incubated with chitotriose in concentrations of 25, 50, 75 and 100 mM for up to 24 hr. Counterstaining, when done, and mounting were performed by Mayers' hemalaun and in an aqueous medium, respectively.
Source, abbreviation, and nominal sugar specificity of the lectins used Lectin Source Abbreviation Nominal monosaccharide specificity a
aGalNAc, N-acetylgalactosamine; GlucNAc, N-acetylglucosamine.
In addition, the fibrous reticulum of the spleen was demonstrated by silver impregnation according to the Gomori method and storage iron was stained by Perls' Prussian blue reaction. Some of the silver-impregnated slides were used for additional CMA staining. Acid phosphatase was demonstrated according to the method of Leder (1967).
Results
Of the 16 lectins used, eight (i.e., UDA, BDA, IRA, CAA, GNA, AMA, LOA, HHA, and SNA-I) showed significant lectin binding to cells and one (CMA) also to extracellular matrix. The following lectins did not react or reacted only very weakly with a few mostly undiscernible cells: APA, DBA, SBA, ACA, and ASA. SNA-II stained only some granulocytes in the cords of red pulp.
CMA Binding in White Pulp
The reaction pattern of CMA in this splenic compartment in the paraformaldehyde-fixed, paraffin-embedded material showed no significant differences from that obtained from frozen sections. It consisted of strongly labeled endothelial and adventitial connective tissue cells of trabecular arteries and veins, central arteries, and arterioles, the RM of PALS, and scattered or groups of rounded cells in the MZ (Figures 2–4 and 6), which could be identified as macrophages by additional markers used. Even the smallest forming primary and secondary follicles were readily discernible in such preparations (Figures 1 and 2). Although reticular structures in the interior were moderately stained, they were quite sharply delineated against the MZ by the CMA reactivity of the circumferential reticulum cells which are adjacent to the inner surface of the marginal sinus (Figures 2–4 and 6). Follicles were either eccentrically connected with the PALS (Figures 1–3 and 6) or they were partially tied off and had their own demarcation against the MZ (Figure 4). Some of the more strongly labeled cells in primary follicles and germinal centers of secondary follicles in the paraffin-embedded material exhibited a dendritic morphology, thus indicating the presence of FDCs (Figure 2). The binding of CMA in white pulp was slightly diminished by preincubation with the trisaccharide chitotriose.
CMA labels central arterioles as well as a reticular meshwork in the periarteriolar lymphatic sheaths in white pulp and venous sinuses in red pulp of the rat spleen. Follicles (∗) correspond to bright spots eccentrically located in PALS or partially separated from them. The halos around the PALS and the bulging follicles represent the marginal zone. Bar = 1 mm.
CMA labels a central arteriole (ca), the reticular meshwork of the periarteriolar lymphatic sheath, the circumferential reticulum delimiting the white pulp, and venous sinuses of the red pulp (RP). Within a very small, probably forming primary follicle (F), a dendritic cell (arrow) is clearly marked, suggesting local transition of a reticular cell to a follicular dendritic cell. MZ, marginal zone. Bar = 200 μm.
CMA binding pattern. Periarteriolar lymphatic sheath with central arteriole (ca) and follicle (F). Both components of splenic white pulp are demarcated against the marginal zone (MZ) by a nearly continuous delicate band of lectin reactivity that corresponds to the circumferential reticulum. RP, red pulp. Bar = 200 μm.
CMA binding pattern. Central artery (ca) with the periarteriolar lymphatic sheath and a widely separated follicle (F). Both components of the white pulp are surrounded by the marginal zone (MZ). RP, red pulp. Bar =100 μm.
The RM within PALS visualized by CMA was much more prominent than the (classical) reticulum stained by silver impregnation or PAS reaction. CMA staining applied after silver impregnation suggests that the lectin is labeling reticular cells ensheathing reticular fibers that are impregnated by the silver salt (Figure 5).
CMA Binding in Red Pulp
CMA binding in this splenic compartment was dependent on the histological technique used. In both paraffin and frozen sections, the endothelial cells of venous sinuses in peripheral MZ and red pulp (Figures 1–4, 6, and 7) were very intensely labeled by CMA, but the reticulum of red pulp remained unstained. This binding pattern of CMA was completely inhibited by chitotriose. In frozen sections, the scattered macrophages of red pulp were intensely labeled by CMA (Figure 6).
ED11 Reactivity in the Spleen
The MAb ED11 labeled an RM within PALS that was more delicate and less continuous than that obtained by CMA binding (Figure 7a and 7b). The walls of the central arteries and arterioles, but not the endothelia, showed ED11 immunoreactivity. In the center of secondary follicles, web-like staining was present. The MZ and the red pulp remained unstained, but the splenic capsule and trabeculae were intensely labeled.
Lectin Binding of Splenic Macrophages
Most of the lectins with mannose specificity, such as GNA, AMA, LOA, and HHA, as well as the sialic acid-specific SNA-I and the GlcNAc-specific UDA, labeled splenic macrophages. Macrophages were present in three typical locations: (a) accumulated within the splenic cords of red pulp in large numbers; (b) scattered in the MZ, PALS, and lymphatic follicles, and (c) interfaced as an open ring between the PALS and follicles on one side and the MZ on the other (Figure 8). The scattered labeled cells in germinal centers of secondary follicles were typical “starry-sky” macrophages. In frozen sections, the macrophages of red pulp could be demonstrated by CMA binding (Figure 6).
Iron Staining of Splenic Macrophages
Because most of the CMA-positive macrophages in the splenic red pulp contained a yellow-brown pigment, which partially reacted with the macrophage-specific lectins, Perls' Prussian blue reaction was also used to visualize this special macrophage population. The cordal macrophages of the red pulp were very intensely iron-positive. The staining pattern ranged from a diffuse blue coloration of the cytoplasm to darker blue granules and polymorphous clumps that corresponded to the yellow-brown pigment of iron in unstained sections. Iron-positive macrophages were also found in the white pulp. These had fewer and smaller or no granules at all but contained iron, which was diffusely distributed throughout the cytoplasm. They were loosely scattered in the MZ, PALS, and follicles, but in the spleen of one animal formed an almost complete continuous band demarcating white pulp from the MZ (Figure 9).
Lectin Binding of Lymphocytes
Some of the lectins (i.e., BDA, IRA, AMA, LOA, HHA, CAA, and SNA-I) were diffusely and fine granularly reactive with large numbers of lymphocytes in the MZ and, to a lesser degree, with lymphocytes of the PALS and the follicles (BDA, CAA, SNA-I) marking these splenic compartments in their entirety. Therefore, these lectins indiscriminantly stained most if not all lymphocytes. In addition, centroblasts or centrocytes were never marked distinctively.
CMA binding pattern of PALS after silver impregnation. The CMA reactivity surrounds both the silver-impregnated reticular fibers and some elongated nuclei. Thus, reticular cells that ensheath the fibers are labeled by the lectin. ca, central artery. Bar = 20 μm.
CMA binding (frozen section) counterstained by Mayer's hemalaun. A peripheral central arteriole (ca) of the PALS branches and radiates through the marginal zone (MZ) into the red pulp (RP). In contrast to the paraffin-embedded material, the scattered macrophages of red pulp are CMA-positive. (
GNA-positive macrophages surround a follicle (F) and delineate it incompletely against the marginal zone. Nuclear counter staining by Mayer's hemalaum. Bar = 100 μm.
Visualization of iron-positive cells by Perls' Prussian blue reaction. Note the abundance of iron-storing macrophages in the splenic red pulp. These cordal macrophages are rich in polymorphous granular iron deposits (dark stained). Macrophages with weaker, predominantly diffuse iron staining correspond to marginal metallophils and are accumulated around PALS and follicles (F), thus producing an almost complete demarcation between white pulp and the marginal zone (MZ). Bar = 1 mm.
Discussion
Thus far, most knowledge of the histophysiology of the spleen has been derived from immunohistology with MAbs specific for immune cells, immune accessory cells, reticular cells, and macrophages. In the present study, carbohydrate determinants in the splenic red and white pulp were visualized by lectin binding. The results obtained with a recently discovered lectin, CMA, were particularly striking. CMA selectively outlined endothelial cells of central arteries/arterioles, an RM in the white pulp, and the lining cells of venous sinuses in the red pulp. In frozen sections, the macrophages of red pulp were also labeled by CMA binding.
The CMA-labeled reticulum was restricted to inner and outer PALS, including follicles. In the PALS, this reticulum was much more prominent than the (classical) reticulum stained by silver impregnation, PAS reaction, or labeling with MAb ED11. Silver impregnation and the PAS reaction label only components of the extracellular matrix such as glycoproteins of reticular fibers via their negative charges and their neutral carbohydrate content, respectively, and stain an extracellular reticulum in both the white and the red pulp. MAb ED11 coincides almost solely with the location of reticular fibers in PALS, follicles, trabeculae, and the capsule of the spleen, and stains FDCs (Van den Berg et al. 1989). The RM visualized by CMA in the PALS, however, is predominantly the result of labeling of the reticular cells and presumably also the fibers, which are produced and are ensheathed by these cells (Weiss 1964). Therefore, the RM identified by MAb ED11 was more delicate and restricted than that obtained by CMA.
CMA is therefore very useful for overall structural analysis of the splenic white pulp by identifying all structures including early-forming primary follicles. This lectin might be of broader use than only in rat because it stained human spleen in a similar manner (unpublished observation). CMA binding was of moderate intensity in the internal reticulum of the follicles but was stronger in the circumferential reticular cells limiting the follicles against the MZ. When individual cells were labeled by CMA within follicles they had the typical dendritic shape of FDCs. At least in part, these FDCs appear to derive directly from local reticular cells when follicles are formed (Figure 2). These data corroborate earlier findings because transitions between both cell types had previously been revealed by electron microscopy in the native state (Heusermann et al. 1980) and by experimental antigen trapping (Dijkstra et al. 1984).
Serial frozen sections comparing CMA binding (
The staining patterns of CMA and MAb ED 11 in the splenic white pulp are differ not only in the structures they label (reticular cells and/or fibers) but also with respect to the determinants they recognize. These determinants must be different because the antibody failed to stain endothelial cells of central arteries/arterioles and sinuses in splenic red pulp that are intensively CMA-reactive.
Rat MAbs have revealed a remarkable heterogeneity of splenic white pulp RM in mice, and it has therefore been proposed that this heterogeneity reflects different functions, including maintenance of the compartment structure of the lymphoid organ and the segregated homing of T- and B-lymphocytes (Yoshida et al. 1991, 1993). The RM demonstrated by these antibodies is obviously the result of staining of reticular cells that can be explained functionally by the fact that reticular cells, but not the reticular fibers, contact the lymphocytes. The labeling pattern of some of these antibodies (e.g., F/R) extends beyond the RM components and also stains the venous sinuses of the red pulp in the mouse in a similar fashion as does CMA in the rat spleen (Yoshida et al. 1993).
Guidance of lymphocyte migration through the spleen may be functionally related to the determinants that are recognized by CMA. Recirculating B- and T-lymphocytes enter the spleen via the MZ and use anatomically determined pathways (Mitchell 1973; Nieuwenhuis and Ford 1976). They stream through the periphery of PALS to reach their final destinations, the central areas of PALS and the corona as well as the germinal centers of secondary follicles, respectively. It is of special interest that CMA recognizes (GlcNAc)n, which is one of the carbohydrate binding moieties of members of the selectin and S-lectin families (Hughes 1992), and therefore these carbohydrate residues may serve as a ligand for selectins that are involved in lymphocyte traffic.
GNA was previously described to label tingible body macrophages in the germinal centers of human Peyer's patches (Sharma et al. 1996) and in secondary follicles of lymph nodes in two seal species (Halichoerus grypus and Phoca vitulina) (Welsch et al. 1997). GNA is a lectin with only mannose specificity, as are the other macrophage-specific lectins AMA, LOA, and HHA, which also label these cells selectively. In contrast, ConA, which is also used to recognize cells of the monocyte/macrophage lineage (Strauchen 1984; Schumacher and Welsch 1987; Horst et al. 1992; Welsch et al. 1997), has a broader tissue binding pattern, which can be attributed to the fact that it recognizes both mannose and glucose.
Lectin binding and iron staining of macrophages demonstrated the abundance of macrophages in the cords of red pulp. The diffuse stainable iron of the cytoplasm and the polymorphous granular iron deposits in the macrophages correspond to free ferritin molecules of the cytosol and to ferritin- and hemosiderin-containing lysosomes, which are called siderosomes (Düllmann et al. 1992). At the border between the white pulp and the MZ, stored iron was a more sensitive macrophage marker than the mannose-specific lectins, which failed to produce an almost complete cellular demarcation between the white pulp and the MZ as seen by Perls' Prussian blue reaction in the spleen of one of the animals studied. The complete ring of iron-positive macrophages between PALS and follicles of white pulp and the MZ was adherent to the inner surface of the CMA-reactive circumferential reticulum, which on the outer side adjoins the marginal sinus of the white pulp. These macrophages correspond to the marginal metallophils (Snook 1964; Satodate et al. 1971; Kumagai et al. 1992). The metallophilic population of splenic macrophages was originally named for their argyrophilia (Marshall 1956). Insight into their functional significance was recently obtained when changes in the localization of iron and iron-related proteins under provoked immune responses were observed in splenic white pulp. After IV injection of a bacterial lipopolysaccharide, transferrin receptor-positive cells increased in the periphery of PALS and follicles as well as in the MZ. In addition, ferritin-positive cells increased in the white pulp and stainable iron accumulated in marginal metallophils (Kumagai et al. 1992). Therefore, the iron load of marginal metallophils may reflect prior immune challenges to the spleen. The role of iron in response to immunological challenges appears to be a regulatory one, because it has been shown in vitro that proliferation of rat splenic lymphocytes is inhibited by isoferritins proportional to their iron content (Cardier et al. 1995).
Footnotes
Acknowledgements
Acknowledgments
We wish to thank Prof Christine D. Dijkstra (Free University of Amsterdam) for the kind gift of the monoclonal antibody CD11.