Surface phenotype and function of tonsillar germinal center and mantle zone B cell subsets

Human peripheral lymphoid tissues contain distinct B cell populations that differ in their buoyant density, cell surface phenotype, and responsiveness to proliferative signals. Two major B cells subpopulations from reactive tonsils or lymph nodes have phenotypes that appear to correspond to mantle zone and germinal center B cells. These subpopulations were distinguished using two-color immunofluorescence to measure surface IgM, IgD, DR, DQ, and Bp35 (B1) in pairwise combinations. The two populations were separated on density gradients, and their proliferative responses to activation signals was studied. The dense B cells proliferated in response to anti-mu on beads or anti-Bp35 antibodies but not to T cell-derived growth factors. The dense B cells are almost all IgM+IgD+Bp35dull (mantle zone phenotype). In contract, more buoyant B cells proliferated in response to T cell factors alone but not to anti-mu on beads or anti-Bp35. These cells are IgD-, express little or no IgM and higher levels of Bp35 (germinal center phenotype). An additional minor subpopulation of dense Bp35dull IgD- B cells was detected in tonsils, suggesting that IgD may be lost prior to B cell entry into germinal centers. B cells in peripheral blood and spleen have surface phenotypes distinct from each other and from the tonsil subpopulations. In particular, the levels of IgM, Bp35, and DR all vary among different lymphoid locations. The mean surface density of DR is low in peripheral blood, intermediate in the spleen, and highest in reactive tonsils and lymph nodes. While Bp35 surface level does correlate with the stage of B cell activation, the surface level of DR, DQ, or DP does not since the surface density of DR, DQ, and DP are similar on both mantle zone and germinal center B cells. Both tonsillar populations express a wide range of class II antigen densities and display coordinate expression of DR and DQ antigen levels.     

Lymphocyte surface IgD and IgM in Macaca monkeys: ontogeny, tissue distribution and occurrence on individual lymphocytes.

A proportion of lymphocytes in blood, spleen and lymph nodes of nonhuman primates had immunoglobulin on their surfaces detectable by fluorescent antibody to human IgM and IgD. The majority of the individual lymphocytes having either IgM or IgD on their surfaces possessed both classes of immunoglobulin. Lymphocyte surface IgD was capped independently of surface IgM on the same cell when incubated at 37 degrees with anti-IgD. Lymphocytes with surface IgM and/or IgD were present in blood at birth and the percentages over the first 6 months of life were increased compared to older monkeys. A corona of cells faintly positive for both IgM and IgD was observed around germinal centres of both lymph nodes and spleen.     

Double labeling immunohistologic and flow cytometric analysis of human B cells with particular reference to Leu-8 expression

Previous results have shown that human lymphocyte subpopulations are heterogeneous as to Leu-8 expression. In the present study, we performed a heretofore unreported immunohistologic analysis and flow cytometric double labeling investigation focused on Leu-8+ and Leu-8- human B cells, with special reference to their expression of other B cell lineage antigens (Leu-14, B1, OKB7 or B2, OKB2, BA-1, BA-2, IgD, and IgM) or of a functional marker of cell proliferation (Ki-67). Immunohistologic analysis was performed on frozen sections of nine normal or reactive lymph node and tonsil biopsy specimens tested with either single or paired antibodies, the latter procedure (double labeling) being directed at revealing positively subtracted Leu-8+ or Leu-8- cells expressing a given marker depending on the antibody staining sequence used. Dual flow cytometry with paired antibodies was performed on cell suspensions from four normal or reactive lymph nodes and tonsils. As to the follicular district, immunohistologic analysis suggests that germinal center (proliferating) B lymphocytes, identifiable in cell suspension as Leu-8-B1+ and representing 80% of the B1+ cells, are Leu-8- and Leu-14+, BA-2+, OKB7+, and IgM+, being rarely IgD+. Most lymphocytes located in the mantle zone (resting cells), identifiable in cell suspension as Leu-8+B1+ and representing the remaining 20% of the B1+ cells, are Leu-8+, Leu-14+. OKB7+, OKB2+, BA-1+, IgD+, and IgM+. Furthermore, from our results, there is indirect evidence to support the existence of a Leu-8+BA-2+ lymphocyte subset located in the mantle zone. Immunohistologic study of four lymph nodes by Leu-8 and Ki-67 antibody (that recognizes a nuclear, cell proliferation-associated antigen absent only in the Go phase of the cell cycle) showed that these markers are mutually exclusive and added further evidence that Leu-8 antigen is a marker of resting lymphocytes.     

Memory B cells without somatic hypermutation are generated from Bcl6-deficient B cells

After immunization with T cell-dependent antigens, the high-affinity B cells selected in germinal centers differentiate into memory B cells or long-lived antibody-forming cells. However, a role for germinal centers in development of these B lineage cells is still controversial. We show here that Bcl6-deficient B cells, which cannot develop germinal centers, differentiated into IgM and IgG1 memory B cells in the spleen but barely differentiated into long-lived IgG1 antibody-forming cells in the bone marrow. Mutation in the V-heavy gene was null in these memory B cells. Therefore, Bcl6 and germinal center formation are essential for somatic hypermutation, and generation of memory B cells can occur independently of germinal center formation, somatic hypermutation, and Ig class switching.     

CD23 expression in aged rats.

CD23 is considered as an activation marker. Its expression on lymphocytes was studied during aging by using 1- to 142-week-old rats. Before analysis, all animals were injected with monoclonal IgE, in order to detect their full expression of CD23 molecules on their lymphocyte surface membranes. In elderly rats, a decrease was observed in the percentages of CD23+ lymphocytes from mesenteric lymph nodes, spleen, Peyer's patches and peripheral blood. In addition, these lymphocytes expressed significantly fewer CD23 molecules per cell. Almost all CD23+ cells belonged to the B lineage and were bearing both IgM and IgD on their membranes. It is suggested that a decrease in the number of both IgM+IgD+CD23+ lymphocytes and CD23 molecules per cell may contribute to immune deficiency in the elderly.     

IgD+IgM- B cells mount immune responses that exhibit altered antibody repertoire

IgM and IgD expression during B cell development and differentiation is strictly and developmentally controlled. Although studies have suggested subtle differences in B cell activation, tolerance, and affinity maturation when antigens ligate cell membrane IgM or IgD, the mechanisms that may explain these differences remain unknown and no drastic differences in immune responses have been reported in mice whose B cells selectively lack IgM or IgD. We now show that the antibody repertoire in IgM(-/-) mice is dramatically altered during the primary response to phosphorylcholine.In IgM(-/-) mice, B cells that are activated and differentiate into antibody-forming cells and germinal center B cells express VH genes other than the T15 genes that dominate in wild-type mice. The kinetics of the antigen-specific IgD primary antibody response in IgM(-/-) mice appears similar to that of IgG, but not to that of IgM in wild-type mice. Thus, our studies demonstrate that differences in the roles played by IgM and IgD in regulating the responsiveness and differentiation of B lymphocytes can have major biological consequences during adaptive immune responses.     

Phenotypic expression of B-lymphocytes. 2. Immunoglobulin expression of germinal center cells.

With a sensitive and specific immunohistochemical technique, immunoglobulin expression was examined in germinal centers with frozen and paraffin-embedded tissues. The majority of cells positive for surface immunoglobulin (IgM and IgD) were confined to the dark zone of the germinal center, while the cells positive for cytoplasmic immunoglobulin (IgM and IgG) were found in the light zone. There were considerable numbers of germinal center cells without identifiable immunoglobulin. The results, in conjunction with previous in vitro and in vivo B-cell activation studies, indicate that centroblasts (noncleaved cells) are formed during early stages of B-cell differentiation. Centroblasts further mature into centrocytes.     

B-cell production and differentiation in adult rats.

The B-cell development in a group of rats was suppressed for the first 45 days of life by serial administration of rabbit anti-rat IgM and IgD antibody. Total or near total suppression of B lymphopoiesis was achieved. At 45 days, suppression was stopped by injection of IgM and IgD rat paraproteins. The sequence of B-cell and plasma cell development following suppression was assessed by immunohistological analysis of spleen lymph nodes and small intestinal lamina propria. The main findings are listed below. Complete reconstitution of B-cell numbers occurs within 8 days, at which stage germinal centres are also present. B lymphopoiesis in the red pulp of the spleen differs from that reported for bone marrow. Cells develop expressing surface sIgM and sIgM with IgA, but not sIgD. sIgD-positive cells first appear in splenic follicles 2 days after stopping suppression, but their appearance in lymph nodes is delayed until after 3 days. At this stage, sIgD-positive cells become apparent in the splenic red pulp. IgM plasma cells appear from day 4. IgA plasma cells in the gut appear in small numbers at day 6, and gradually increase to normal numbers by day 14. sIgG2c expression in the splenic marginal zone did not approach normal levels, even 2 weeks after suppression was stopped.     

Kinetics of germinal center development in lymph nodes of young and aging immune mice

Recent findings imply that germinal center paucity in old mice, at least in part, results from a defect in the mechanisms responsible for the transport of antigens to lymphoid nodules (follicles) and the consequent impairment of the antigen retaining reticulum (ARR) of follicular dendritic cells (FDCs). The present objective was to observe the kinetics of lymph node germinal center development in old mice having antigen transport and ARR deficits. Germinal center development was monitored in popliteal (PLN) and axillary (AXLN) lymph nodes of 6-8 wk and 23-mo-old horseradish peroxidase (HRP) immune C57BL/6 mice. Using the selective binding of germinal center B cells for peanut agglutinin (PNA), germinal centers were identified in serial vibratome sections following histochemical labeling with PNA-peroxidase conjugates at times 0, 15 min, 1, 3, 5, and 10 days after footpad challenge with 8 micrograms HRP. To follow the fate of preexisting (environmental antigen-induced) germinal centers and the development of de novo (HRP-induced) germinal centers, it was essential to distinguish between these germinal centers. Accordingly, PNA positive germinal centers associated with HRP-retaining (peroxidase positive) ARR were identified as de novo germinal centers and germinal centers not associated with a peroxidase positive ARR were classified as preexisting germinal centers. Kinetic analysis of PNA positive germinal centers showed the following: 1) Preexisting, environmentally-induced germinal centers dissociated and disappeared by day 3 as indicated by a decline in their numbers after antigen injection: the process of germinal center dissociation remained unaffected by aging. 2) The latency of de novo germinal center appearance was approximately equal in duration (approximately 3 days) to the disappearance of pre-existing germinal centers. 3) The number and size of de novo HRP-induced germinal centers increased through the experimental period in young lymph nodes, but in old mice these parameters were depressed, resulting in a significant germinal center deficit. 4) The ratio of HRP-retaining ARR to de novo induced germinal centers was 1:1 in young and responder old mice. This ratio was not affected by aging. This finding favored the concept that antigen retention in ARR is a requirement of germinal center development. The observations supported our hypothesis that germinal center development, at least in part, depends on a normal antigen transport by showing that in aged mice with defective antigen transport-related ARR and iccosome deficits there is an impaired development of germinal centers.     

High CD99 expression in memory T and B cells in reactive lymph nodes

We investigated the expression of CD99 in 35 hyperplastic perigastric lymph nodes, which were resected for gastric carcinoma or chronic peptic ulcer. Essentially, all lymphocytes in lymph nodes expressed CD99, but there were two populations with respect to the intensity of CD99 expression--CD99high and CD99low cells. We showed CD99high cells were distributed in paracortical and medullary cords by immunohistochemical study while germinal center cells were CD99low. Using three-color flow cytometric analysis with CD3, CD4, CD8, CD19, CD23, CD45RA, CD45RO, CD69, CD138, IgM, IgD, and IgG, most of CD99high cells were shown to be activated/memory T cells. CD4+CD45RO+ T cells were the subset revealing the highest intensity of CD99 expression while CD4+CD45RA+ T cells were CD99low. Among B cells, IgG+ B cells revealed a higher level of CD99 molecules than IgM+ B cells. These results suggest that CD99 is one of activation-related molecules which are upregulated in recently activated lymphocytes.     

Expression of HIV in lymph node cells of LAS patients. Immunohistology, in situ hybridization, and identification of target cells.

Monoclonal antibodies (MAb) anti-HIV core and envelope proteins and in situ hybridization, using cDNA HIV probe, were employed to determine which lymph node cells in LAS patients express viral antigens and viral nucleic acids. The results have been correlated with the histologic phases of LAS and with the germinal center lysis detected using DRC-1 MAb directed against follicular dendritic reticulum cells (FDRC). Viral antigens occasionally were detected on high endothelial cells of paracortical venules and frequently on germinal center FDR accessory cells; this last finding correlates well with the extent of FDRC lysis and of CD4+ and CD8+ lymphocyte infiltration in germinal centers. HIV replication, detected by in situ hybridization, was observed in mononucleated cells present in T and B areas and, in one case, in flat endothelium.     

Differential expression of GL7 activation antigen on bone marrow B cell subpopulations and peripheral B cells.

GL7 was originally described as a 35-kDa late activation antigen on mouse T and B cells. GL7 expression has also been demonstrated on thymocytes, germinal center B cells and some neuronal cell types. Flow-cytometry and immunohistochemistry were used to follow changes in the expression of GL7 during B cell development, amongst B cell subpopulations and various anatomical locations. GL7 is expressed as early as the pro-B cell stage and increases up to the pre-B-I stadium. Expression remains high on pre-B-II and on immature B cells, although slightly decreases during maturation. GL7 is almost completely downregulated when IgD appears on the cell surface. On the periphery only a few B cells are positive and these cells are almost exclusively found in the sIgD- germinal center areas of lymph nodes and spleen. The staining pattern of GL7 is very similar to that of PNA in the lymph nodes but in the bone marrow we have found both B220+PNA+GL7- and B220+PNA+GL7+ populations, showing that GL7 and the antigen recognized by PNA are different. After in vitro stimulation, the GL7(hi) B cell population has also been found to be IgD negative. Functional comparison between in vitro activated and MACS sorted GL7(hi) and GL7(lo/-) spleen B cells of immunized mice showed significantly higher specific and total antibody production as well as antigen presenting capacity in the GL7(hi) population.     

Comparison of IgD+ and IgD- thoracic duct B lymphocytes as germinal center precursor cells in the rat.

Genetically marked thoracic duct B cell subpopulations rich in either IgD+ or IgD- B cells were transferred to non-irradiated, congenic rats in order to compare the capacities of IgD+ versus IgD- B cells to form germinal centers (GCs). This comparison was made quantitatively based on flow cytometric analyses of lymph node cells prepared from chimeric rats 7 days after s.c. immunization. Donor-origin and host-origin B cells were distinguished using anti-Igk antibodies, and GC B cells were distinguished from other B cells in suspension by their lack of labeling with the mAb HIS22. IgK+ HIS22- lymph node cells corresponded well to GC B cells: they contained many large cells, were IgM+ but mostly IgD-, expressed relatively lower levels of IgM than HIS22+ B cells, and increased in number and frequency in response to antigen. Results from flow cytometric analyses, corroborated by immunofluorescence histochemical studies, showed that cell-for-cell, IgD- B cells from GCs much more efficiently than IgD+ cells. B cell populations enriched for IgD- cells became relatively more distributed to GCs than to other lymph node B cell areas and gave rise to many more GC B cells of donor origin per transferred B cell than whole, unseparated thoracic duct B cells (for which greater than 97% were IgD+). IgD- B cells from rats primed deliberately with antigen also became relatively more distributed to GCs and gave rise to more GC B cells of donor origin than either IgD+ B cells from primed donors or IgD- B cells from unprimed donors.(ABSTRACT TRUNCATED AT 250 WORDS)     

Analysis of cell division among subpopulations of lymphoid cells in sheep. II. Peripheral lymphocytes

The number, distribution and surface phenotype of dividing cells in lymph nodes and blood and differences between the cell-cycle status of lymphocyte subpopulations were studied in lambs using double-labelling techniques. Dividing cells were labelled in vivo for various time periods with 5-bromo-2-deoxyuridine (BrdU). After removal of lymphoid tissues, the proportions of constituent lymphocyte subpopulations which had synthesized DNA during the labelling period were measured by flow cytometry or immunohistochemistry using a panel of monoclonal antibodies (mAbs) specific for sheep lymphocyte differentiation antigens and MHC class I and class II antigens in conjunction with an anti-BrdU mAb. There was a higher overall level of cell division in the ileocaecal lymph node than in either the prescapular or parathymic lymph nodes. In all three lymph nodes, the majority of lymphocytes which incorporated BrdU occurred in B-cell follicles or germinal centers. CD4+ and CD8+ T cells had a higher level of cell division (LI 5-14%) than those recognized by mAb 197 (CD4- CD8- subset) (LI less than or equal to 3%).     

Unusual biochemical features and follicular dendritic cell expression of human Fcalpha/mu receptor

The Fc receptor for IgA and IgM (Fcalpha/muR) is of particular interest because it can bind antibodies of both IgM and IgA isotypes and thus may play a pivotal role in systemic and mucosal immunity. Using IgM and IgA ligands and newly generated Fcalpha/muR specific monoclonal antibodies we have defined biochemical features and cellular distribution of the human Fcalpha/muR. Both recombinant and native forms of human Fcalpha/muR are expressed on the cell surface as remarkably stable homodimeric transmembrane glycoproteins that can bind specifically polymeric IgM or IgA. The only human B cells to express Fcalpha/muR, albeit at very low levels, are found in the pre-germinal center subpopulation defined by the IgD+/CD38+ phenotype. Hence the expression pattern differs from that of the mouse wherein Fcalpha/muR is expressed by both circulating and resident B cell populations. Significantly, the predominant cell type expressing the Fcalpha/muR in humans is the follicular dendritic cell of germinal centers. The Fcalpha/muR may thus function in antigen presentation and B cell selection in the germinal center response.     

Increased immunoglobulin E Fc receptor bearing cells in germinal centers of hyperimmunoglobulinemia E patients

Lymph nodes from 6 patients with hyperimmunoglobulinemia E (hyper-IgE; 4 with Kimura's disease, 1 with atopic dermatitis, and 1 with immunoblastic lymphadenopathy-like T cell lymphoma) and from 7 patients with normal IgE levels were studied to determine the localization of dendritic reticulum cells and of cells bearing Fc epsilon and C3d receptors and immunoglobulin E. The avidin-biotin-glucose oxidase method was used for unfixed biopsy specimens. To identify the above-mentioned cells, H107, a murine monoclonal antibody specific to the Fc epsilon receptor molecule, and corresponding antibodies specific to the other cell types were used. In 5 hyper-IgE patients (4 with Kimura's disease and 1 with atopic dermatitis) all germinal centers of the lymph nodes showed heavy reticular staining with H107, the dendritic reticulum cells being most intensely stained. In contrast, the germinal centers of the lymph nodes in the 7 patients with a normal IgE level only lightly or partially stained with H107. The staining pattern of anti-IgE was similar to that of H107 in the hyper-IgE cases. Likewise, C3d receptor and dendritic reticulum cell-related antigens were demonstrated very intensely in all germinal centers in lymph nodes of patients with hyper-IgE and normal IgE levels. These findings suggest that in a hyper-IgE state increased numbers of dendritic reticulum cells in a germinal center express the Fc epsilon receptor and that such cells may play a role in the differentiation of IgE-producing memory B cells.     

IgD Receptors of Lymphoid Cells

While present in normal human serum in very low amounts and undetectable in sera of non-human primates as well as of mice, IgD is found on the surface of the majority of B lymphocytes in all the above mentioned species.

Lymphocytes which carry IgD on their membrane also have IgM. The two molecules are present in relative amounts that can be very different in different cells.

Both IgM and IgD of a single cell are the actual product of the cell itself. They have the same light chain and, more importantly, the same combining site and idiotype.

IgD/IgM bearing lymphocytes are the majority of all B lymphocytes in spleen, lymph nodes and Peyer's patches, whereas in bone marrow they account for half of the immunoglobulin positive cells. Although the percentage of double IgD/IgM cells is very similar in different tissues, the total amount of IgD, as well as the relative amounts of IgD and IgM as detected by biochemical methods varies. In fact, lymph nodes, and even more Peyer's patches are much richer than spleen in cells having levels of IgD higher than those of ISM; conversely, in the bone marrow, all the positive cells have very low levels of IgD.

In ontogeny, as in evolution, IgD appears after IgM: in human foetuses IgD bearing cells are not detectable before 13 weeks of gestation, and in the muse they appear only after birth.

IgD receptors seem to disappear from B cells which undergo maturation to secretion, as indicated by the fact that only a proportion of IgM secreting plasma cells show membrane IgD.

IgD is never found on the membrane of IgG-containing cells, and also lymphocytes bearing simultaneously IgD and IgG are very rare, and it might well be that for these cells, the double expression for short periods of time does not actually correspond to simultaneous synthesis.     

IgD Receptors of Lymphoid Cells

While present in normal human serum in very low amounts and undetectable in sera of non-human primates as well as of mice, IgD is found on the surface of the majority of B lymphocytes in all the above mentioned species.

Lymphocytes which carry IgD on their membrane also have IgM. The two molecules are present in relative amounts that can be very different in different cells.

Both IgM and IgD of a single cell are the actual product of the cell itself. They have the same light chain and, more importantly, the same combining site and idiotype.

IgD/IgM bearing lymphocytes are the majority of all B lymphocytes in spleen, lymph nodes and Peyer's patches, whereas in bone marrow they account for half of the immunoglobulin positive cells. Although the percentage of double IgD/IgM cells is very similar in different tissues, the total amount of IgD, as well as the relative amounts of IgD and IgM as detected by biochemical methods varies. In fact, lymph nodes, and even more Peyer's patches are much richer than spleen in cells having levels of IgD higher than those of ISM; conversely, in the bone marrow, all the positive cells have very low levels of IgD.

In ontogeny, as in evolution, IgD appears after IgM: in human foetuses IgD bearing cells are not detectable before 13 weeks of gestation, and in the muse they appear only after birth.

IgD receptors seem to disappear from B cells which undergo maturation to secretion, as indicated by the fact that only a proportion of IgM secreting plasma cells show membrane IgD.

IgD is never found on the membrane of IgG-containing cells, and also lymphocytes bearing simultaneously IgD and IgG are very rare, and it might well be that for these cells, the double expression for short periods of time does not actually correspond to simultaneous synthesis.     

B lymphocyte differentiation in the rat: production and characterization of monoclonal antibodies to B lineage-associated antigens

Three mouse monoclonal antibodies (mAb) directed against rat B lineage antigens were produced. The mAb, designated HIS14 (IgG1), HIS22 (IgM) and HIS24 (IgG2b), were characterized for binding to lymphoid and nonlymphoid tissues by immunoperoxidase staining of frozen sections and by (double-) immunofluorescence staining of single cell suspensions from lymphoid organs. HIS14 recognized a pan B cell determinant: it reacted with virtually all cells of each anatomic B cell compartment and with about 95% of surface (s)Ig+ cells in thoracic duct lymph and in suspensions of spleen and lymph nodes. HIS22 and HIS24 detected B lineage-associated antigens expressed by major subpopulations of B cells. HIS22 predominantly stained the lymphocyte corona, but not (or weakly) the germinal centers and splenic marginal zones, whereas HIS24 reacted with both corona and germinal center and not (or weakly) with marginal zone. In accordance with this, substantial proportions of sIg+ cells in spleen cell suspensions did not express HIS22 or HIS24 determinants (20% and 27%, respectively). In bone marrow the vast majority of cytomplasmic mu+ pre-B cells were HIS14+ and HIS24+, and up to one third also HIS22+, indicating an appearance of the determinants early in B lymphocytopoiesis. The antigens recognized by HIS14, HIS22 and HIS24 are lost during the final stage of B cell differentiation: none of the mAb bound to plasma cells. As far as detectable, neither cells of myeloid and erythroid lineages in bone marrow nor thymocytes were stained by HIS14, HIS22, or HIS24. In suspensions of peripheral lymphoid organs (spleen and lymph nodes) but not in thoracic duct lymph, HIS14 and HIS24 labeled a small proportion (12% and 14%, respectively) of Ig- cells. HIS22 did not bind to Ig- peripheral lymphocytes. Reactivity of HIS14, HIS22 and HIS24 with nonlymphoid tissues was virtually absent; HIS22 stained the high endothelial venules in lymph nodes and Peyer's patches. As determined by immunoblotting, the antigenic determinants on lymph node cells recognized by HIS14, HIS22 and HIS24 were present on molecules with an apparent molecular mass of 205 kDa, 210 (and 175) kDa and 205 kDa, respectively, which is similar to the molecular mass of the B cell form of the rat leukocyte common antigen. In addition, the antigens recognized by HIS14, HIS22 and HIS24 co-capped with the leukocyte common antigen. This suggests that each of the three mAb recognize determinants present on the B cell form of the leukocyte common antigen.     

Kinetics of the tingible body macrophage response in mouse germinal center development and its depression with age

Although tingible body macrophages (TBM) have been recognized in germinal centers for over 100 years, their role in the germinal center response is not clear. In this study, the kinetics of the TBM response was quantitatively assessed and correlated with the kinetics of germinal center development in young mice. The TBM response in old mice (which have an age-related depression of germinal center development; Szakal et al., 1990) was analyzed for comparison. Young and old immune mice were challenged with human serum albumin and 0, 1, 3, 5, 7, 10, and 14 days later the popliteal and axillary lymph nodes were evaluated. Germinal centers were localized histochemically in alternate serial sections using horseradish peroxidase conjugated peanut agglutinin. TBM numbers were determined per germinal center on adjacent sections by the presence of tingible bodies or histochemically by using the monoclonal antibody Mac-2. Analysis of lymph nodes from young mice showed that TBM numbers decreased with the dissociation of preexisting germinal centers. TBM reappeared 5 days after challenge and the TBM kinetics paralleled the increase in size of de novo germinal centers. In fact, a constant ratio of one TBM to every 350-450 B cells was maintained from day 5 to day 10. In old lymph nodes, TBM were generally absent throughout germinal center development. The lack of TBM prior to germinal center development and their absence in aged mice are inconsistent with the concept that TBM are required for the induction of the germinal center reaction. However, the data are consistent with a role for TBM in regulating the magnitude of the germinal center reaction.