Logo of immunologyLink to Publisher's site
Immunology. Dec 2005; 116(4): 429–442.
PMCID: PMC1802440

CD27+ B cells in human lymphatic organs: re-evaluating the splenic marginal zone


The marginal zone of human spleens is regarded as an organ-specific region harbouring sessile memory B cells. This opinion has arisen by extrapolating from results obtained in mice and rats. Detection of CD27+ B cells in situ now revealed similarities among the most superficial region of B-cell follicles in human spleens, reactive lymph nodes, inflamed appendices, tonsils and terminal ilea. The follicular surface in these organs consists of small naïve immunoglobulin D (IgD)+ CD27 B cells predominating in an inner area and larger IgD+/– CD27+ B cells prevailing in a more superficial position. CD27+ B cells may, however, also occupy the entire follicular periphery around the germinal centre. Together with additional peculiarities this distribution indicates a fundamental microanatomical difference among the human and rodent splenic white pulp. We hypothesize that the follicular periphery represents a recirculation compartment both for naïve and memory/natural reactive B cells in all human secondary lymphatic organs. This assumption implies a difference in recirculation behaviour among human and rodent B memory cells.

Keywords: marginal zone, CD27+ B cells, human spleen, human lymph node, human tonsil


The marginal zone (MZ) is a special B-cell compartment of the spleen, which delimits the white pulp from the red pulp. In mice and rats it is assumed to be a spleen-specific microanatomical region primarily occupied by memory B cells. At the same time the MZ is passed by recirculating naïve B and T cells entering into the splenic white pulp. MZ memory B cells are specific for T-dependent antigens, but B cells responding to T-independent antigens, especially to repetitive polysaccharide structures named T-independent type 2 (TI-2) antigens, also occur in the MZ.1,2 In addition, autoreactive and polyreactive B cells preferentially localize to the rodent MZ.3 Mouse spleens exhibit rather small MZs while in rats the MZs are broad well-delimited areas.4 In both species the MZ is separated from the innermost white pulp, i.e. from the T-cell zone and the follicles, by a so-called marginal sinus and a ring of marginal metallophilic macrophages.5 Special MZ macrophages are present among the MZ B cells. In rats and mice, the majority of MZ B cells are supposed to be sessile and to derive from recirculating precursors.6,7 In spite of the fact that rat MZ B cells appear to exert memory functions8 their antigen receptors were reported not to exhibit hypermutated immunoglobulin variable regions.9

The human splenic MZ has been difficult to define histologically, because a marginal sinus and marginal metallophilic macrophages are lacking in this species.1012 Microdissected human splenic MZ B cells were, however, demonstrated to carry hypermutated surface immunoglobulin.13 Human MZ B cells are supposed to express the phenotype CD27+ immunoglobulin M (IgM)+ IgD+ or +/– and to play a decisive role for immunoglobulin production against encapsulated bacteria. The immunoglobulin produced possesses hypermutated binding regions, although the bacterial antigens recognized are TI-2 antigens which are unable to elicit T-dependent B memory cells in rodents. This paradox was recently discussed.14,15 Potential explanations include the hypothesis that the antigenic polysaccharides are somehow coupled to proteins in vivo or that certain human B cells acquire hypermutated immunoglobulin in a T-cell independent fashion.

CD27, a member of the tumour necrosis factor-receptor family, has been characterized as an antigen on human T cells, natural killer cells, plasma cells and memory B cells.16 In human blood CD27 is present on B memory cells with hypermutated variable region genes, while naïve B cells are IgD+ CD27.17,18 Memory cells represent about 40% of blood B cells in humans17 but only 5% in mice.19 Human blood memory B cells form at least four subpopulations.17 Two large populations comprise ‘switched’ CD27+ cells and IgD+ IgM+ CD27+ cells, respectively. In addition, there is a smaller population of IgD IgM+ CD27+ cells, called ‘IgM only’ B memory cells, which seems to depend on the presence of the spleen.20 Finally, a fourth IgD+ IgM CD27+ human memory B cell population, called ‘IgD only’ cells, occurs at very low frequency and exhibits high immunoglobulin mutation rates.17 In human and rodent lymphatic organs recirculating naïve IgD+ B cells are present in primary follicles or in the mantle zone surrounding germinal centres (GCs) of secondary follicles. CD27+ cells have been detected in the supposed marginal zone of human spleen cryosections by immunohistology, but T and B cells were not differentiated in these studies.21,22

The regulation of human CD27 expression is complicated and differs in human B and T lymphocytes. While CD27 is absent in naïve recirculating B cells, it is present in recirculating T cells. B memory and plasma cells express CD27. Certain T memory cells are CD27+ in humans, but CD27 is down-regulated on CD8+ T effector cells.23 The same may also be true for mice.24 The function of CD27 on B cells also differs among species. Thus, ligation of CD27 on activated mouse B cells promotes B memory cell development and inhibits plasma cell formation, while plasma cell differentiation is obviously promoted in humans.25,26 In addition, CD27 does not represent a memory B cell antigen in mice.27 The role of CD27 on human centroblasts, centrocytes and plasmablasts has not been studied in detail. Engagement of CD27 on human T cells appears to support activation and proliferation.28

We wanted to investigate whether the area occupied by CD27+ B cells in human spleens forms a separate compartment and whether this compartment is spleen-specific. For this purpose we developed highly sensitive methods to differentiate CD27-positive B cells from T cells in paraffin sections by immunoenzymatic subtractive double-staining and to visualize coexpression of CD27, IgD and IgM by immunofluorescence. We then compared the staining patterns of spleens to those of reactive lymph nodes, tonsils, appendices and secondary follicles of the terminal ileum.

Materials and methods

Specimens (Table 1, Table 2)

Table 1
Spleen specimens
Table 2
Specimens of other lymphatic organs

Spleen specimens of patients aged 6–79 years (n = 10), metastasis-free lymph node specimens of patients aged 57–73 years (n = 6), specimens of resected appendices of patients aged 13–80 years (n = 6), specimens of resected terminal ilea of patients aged 26–78 years (n = 6) and tonsil specimens of patients aged 6–40 years (n = 6) were obtained from the archive of the Institute of Pathology of Marburg University. The tissues had been fixed in 3·7% formalin in water for an unknown duration before paraffin embedding.

Immunoenzymatic techniques

Antigen retrieval and inactivation of endogenous peroxidase

Paraffin sections were prepared on silanized slides. The specimens were deparaffinized, autoclaved in citrate buffer pH 6·0 for 20 min, treated with 0·4 U/ml glucose oxidase/10 mm glucose/1 mm NaN3 in phosphate-buffered saline (PBS) and washed. Cryosections were fixed for 10 min at 4° in 100% isopropanol before glucose oxidase treatment.

Single staining

Anti-CD27 monoclonal antibodies (mAbs) 1A4 (Immunotech No. 2034, kindly donated by S. Weller, Hôpital Necker, Paris, France) and MT271 (Pharmingen no. 555439 via Becton-Dickinson, Heidelberg, Germany) were used on cryosections at a dilution of 1 : 100–1 : 300; mAbs 137B4 (Quartett, Berlin, Germany, No. 030410901) and CD27n (Hypromatrix Inc. Worcester, MA, No. HM 1073, via BioCat GmbH, Heidelberg, Germany) were used on cryosections and paraffin sections at 1 : 20–1 : 50. All anti-CD27 reagents were detected using the Vectastain anti-mouse ABC peroxidase system and tyramide amplification as described below.

To demonstrate strong expression of surface and intracellular IgA, polyclonal rabbit anti-human IgA (DAKO No. A 0262) was diluted 1 : 10 000–1 : 15 000 and visualized by the Vectastain anti-rabbit ABC system with tyramide amplification as described below.

Polyclonal rabbit anti-human IgD (DAKO No. A 0093) and anti-human IgM (DAKO No. A 425) were used at 1 : 400 and 1 : 1000, respectively, for intracellular and surface staining. Anti-human IgG (DAKO No. A 0423) was applied at 1 : 40 000 for demonstration of intracellular immunoglobulin. All anti-immunoglobulin reagents were detected with the Vectastain anti-rabbit ABC peroxidase system.

Double staining

Because CD27 is expressed by T and B cells, a subtractive staining method was applied. CD3+ T cells are first stained with a blue chromogen after applying an avidin-biotinylated alkaline phosphatase complex (ABC-AP) procedure. Deposition of the blue chromogen prevents all further reactivity of the stained CD3+ T cells with the second primary antibody. Thus, only CD27+ B cells are left to be revealed in brown colour by anti-CD27 and a consecutive avidin-biotinylated peroxidase complex (ABC-Peroxidase) procedure amplified by tyramide. Coexpression of two antigens is not detected by this method.

In detail, the sections were first incubated with polyclonal rabbit anti-human CD3 (DAKO No. A 0452) diluted 1 : 400 in PBS/1% bovine serum albumin (BSA)/0·1% NaN3 containing 3 µg/ml avidin for 24 hr at 4°. After washing, the Vectastain anti-rabbit ABC-AP (Vector No. AK-5000) reagents were applied according to the manufacturer's instructions, by first incubating the sections with biotinylated anti-rabbit IgG diluted 1 : 200 in PBS with 0·5% normal serum and 0·02 mg/ml biotin for 30 min at room temperature (RT) followed by the Vectastain AB-AP reagent at 1 : 100 in PBS for 30 min at RT. The presence of CD3 was then revealed in dark blue colour by visualizing alkaline phosphatase activity with a nitroblue tetrazolium/bromo-chloro-indolyl phosphate (NBT/BCIP) solution purchased from DAKO (No. K 0598) applied for 10 min at RT.

After washing, the second primary antibody, mAb 137B4 against human CD27, was diluted 1 : 20 in the buffer mentioned above and put on the sections for 24 hr at 4°. For detection of CD27 the Vectastain anti-mouse ABC-Peroxidase system was used acccording to the manufacturer's instructions applying biotinylated anti-mouse IgG at 1 : 200 for 30 min at RT in PBS with 0·5% normal serum and 0·02 mg/ml biotin followed by the Vectastain AB-Peroxidase reagent at 1 : 50 in PBS for 30 min at RT. Tyramide amplification was then carried out, using peroxidase-catalysed binding of biotinylated tyramide to antigen-containing areas. Biotinylated tyramide had been prepared before by incubating 50 mg NHS-LC-biotin (Pierce No. 21335) in 20 ml 0·025 m borate buffer pH 8·5 with 15 mg tyramine-HCl (Sigma No. T 2879) on a stirrer overnight. After filtration this solution was diluted 1 : 1000 in Tris-buffered saline pH 7·6 containing 0·09 mm H2O2 and used for incubating the sections for 10 min at RT. Finally, after washing, the avidin-biotinylated peroxidase complex of the Vectastain kit was applied again as described above and peroxidase reactivity was revealed in brown colour by a diaminobenzidine reaction. Formol pigment was removed by incubating the sections in 1% ammonia solution in 70% ethanol for 5 min at room temperature.

For investigating whether the putative plasmablasts in tonsil GCs were Ki67-positive, tonsil sections were prepared as described above. Rabbit anti-CD3 (final dilution 1 : 400) and mouse mAb MIB-5 (kindly donated by J. Gerdes, Borstel, Germany) against Ki67 (final dilution 1: 1500) were mixed and revealed in dark blue by the ABC-AP system as described above with the only alteration that in the second step biotinylated anti-rabbit IgG and biotinylated anti-mouse IgG were mixed at a final dilution of 1 : 200. After revealing CD3 plus Ki67 in blue colour with the NBT/BCIP reagent (see above), mAb 134B4 against CD27 was applied at 1 : 20 and demonstrated in brown colour by the ABC-Peroxidase system and tyramide amplification as mentioned above.

In all double-staining procedures performed dark blue or brown staining did not occur when either the first or, respectively, the second primary antibody was omitted while all other components of the system were present.

Immunofluorescence techniques

For immunofluorescence detection of coexpression, antigen retrieval was performed as described in below. CD27 was first visualized in green colour by applying the ABC-tyramide method as described above. Instead of the avidin-biotinylated peroxidase complex, streptavidin conjugated to the Alexa 488 chromogen (Molecular Probes, Eugene, OR, No. S 11223 obtained via Invitrogen, Karlsruhe, Germany) was used for the final incubation step. Subsequently, rabbit anti-human IgD (DAKO No. A 0093) or anti-human IgM (DAKO No. A 425) was applied at a dilution of 1 : 50 overnight. Then the TSA kit 41 (Molecular Probes, No. T-30954 obtained via Invitrogen) containing peroxidase-conjugated anti-rabbit immunoglobulin and Alexa 555-conjugated tyramide for red fluorescence was applied according to the manufacturer's instructions.

Similar to the light microscopic techniques, omission of first or second primary antibody was used to control the method.

Fluorescence was investigated with a Leica TCS SP2 AOBS confocal microscope with Leica LSM software. For calculating the percentage of the CD27+ area occupied by IgD+ CD27+ or by IgM+CD27+ cells, the surface of the GC was defined as the inner border of the area evaluated. The surface of the follicle, i.e. the region where the density of B cells did no longer differ from that found in the red pulp, was chosen as the outer border. A preliminary study was performed in two spleens. Measurements always included three different follicles of the same section.


All primary antibodies were tested on paraffin sections and cryosections. In most cases they reacted identically with both types of specimens. Intracellular immunoglobulin was, however, only detected in paraffin sections. Monoclonal antibodies (mAbs) 1A4 and MT271 directed against CD27 were only applicable in cryosections, while anti-CD27 mAbs 137B4 and CD27n visualized T and B cells both in cryosections and paraffin sections. In contrast to the other two anti-CD27 reagents, mAbs 137B4 and CD27n additionally stained platelets and venous endothelium in cryosections, but not in paraffin sections. MAb 137B4 was used throughout the investigation. In paraffin sections it revealed three different intensities of B cell staining designated as CD27+/– CD27+ and CD27++.


CD27+ B cells in small secondary follicles

After staining with haemalum-eosin (HE) six of the 10 specimens investigated (Table 1) exhibited follicles with degenerating GCs lacking clearcut dark or light zones inside. This stage is typical of the majority of adult spleens with GCs. As described previously11,12 the GCs were surrounded by small lymphocytes with dark nuclei and scanty cytoplasm which were imperceptibly replaced by larger lymphocytes with paler nuclei and more cytoplasm towards the follicular surface.

After immunoenzymatic double-staining, CD3+ T cells appeared in dark blue colour, while CD3 CD27+ B cells were visualized in brown. The most conspicuous finding in all spleens with small secondary follicles was a brown-coloured ring-like or crescentic CD27+ area at the surface of the secondary follicles (Fig. 1a). This area was delimited from the GC by an almost unstained region. Such an arrangement was also found in one of the spleen specimens with large secondary follicles described below.

Figure 1
CD27+ B cells in small secondary follicles of human spleens. (a) Subtractive double-staining for CD3 (blue-black) and CD27 (brown). CD27+ B cells are present at the follicular surface, but do not occur at the periphery of the PALS. Seventeen-year-old ...

In the interior of most follicles, the surface of GCs was indicated by an accumulation of dark blue CD3+ GC T cells (Fig. 1a, c). There were, however, some degenerating GCs where T cells were so few that an unequivocal identification of the GC border was impossible. In spleens with degenerating GCs the majority of GC B cells were negative for CD27 or only faintly positive (Fig. 1a). However, single large CD3 CD27++ cells, which were tentatively interpreted as plasmablasts or plasma cells, occurred inside GCs (Fig. 1c). A predominantly CD27 space containing only single CD27+ B cells separated the surface of the GC from the ring of CD27+ B cells at the follicular surface (Figs 1a, c). This CD27 area was not clearly delimited, but the density of CD27+ B cells slowly increased from the GC in direction towards the follicular surface (Fig. 1a, c). In sections of the same specimen it was shown both by immunoenzymatic and by immunofluorescence techniques that B cells in GCs were IgD, while the CD27 areas surrounding them contained IgD+ B cells (Fig. 1b). Thus, the CD27 space corresponded to the mantle zone or corona, which represents the compartment of small recirculating B cells in secondary follicles. Scattered CD3+ T cells were present in the CD27 and in the CD27+ area (Fig. 1a, c).

Within the brown crescentic CD27+ area at the follicular surface there was a broad inner layer containing B cells more intensely stained for CD27 and a thin more superficial region adjacent to the red pulp with CD27+/– B cells (Fig. 1a). At the transition between both layers often a row of blue-staining CD3+ T cells together with unstained fibroblasts and connective tissue fibres was evident. Immunoenzymatic labelling revealed that the CD27+/– B cells at the surface of the CD27+ brown-staining ring continued only for a short distance along the T-cell region, the periarteriolar lymphatic sheath (PALS). Neither CD27+ nor CD27+/– B cells occurred at the surface of the PALS distant from the follicles.

Comparison to sections stained for IgD revealed that the majority of the B cells in the inner layer of the CD27+ ring structure expressed IgD only weakly in some individuals or did not stain for IgD at all in others (Fig. 1b). Interestingly, the most strongly surface IgA-positive B cells, which appeared to represent IgA memory cells, also occurred in the CD27+ area (Fig. 1e). Cells with strong intracellular staining for IgM, IgD, IgG or IgA were, however, more or less absent in this region indicating that the CD27+ B cells did not represent plasmablasts. IgM was present on the surface of B cells in the CD27 and in the CD27+ areas, while in GCs only the reticular pattern of follicular dendritic cells was visualized. Detection of coexpression by immunofluorescence revealed in a selected specimen that IgM+ CD27+ B cells were primarily located in the inner part of the CD27+ area, while IgM CD27+ cells, which most likely represented ‘switched’ memory B cells, were located more superficially (Fig. 1d). The same arrangement was also found with the few IgD+CD27+ B cells (Fig. 1b). The data obtained from Fig. 1(b, d) showed that about 3% of the CD27+ area were occupied by IgD+ CD27+ B cells and about 17% by IgM+ CD27+ B cells, respectively (IgD+ CD27+: 2·75 ± 1·50%, IgM+ CD27+: 17·36 ± 8·22%). By immunofluorescence the scattered IgM+ and IgD+ B cells in the most superficial area of the follicle were IgM+ CD27 and IgD+ CD27. The IgM+ CD27 and the IgD+ CD27 B cells were also present at the entire surface of the PALS (Fig. 2c). They obviously represented naïve recirculating and not MZ-type B cells. Thus, – in contrast to rats and mice – no trace of a MZ harbouring memory-type B cells was found at the surface of the human PALS. Interestingly, in the specimen investigated for coexpression, IgM+ CD27+ B cells were found in the interior of the entire T-cell region and at the follicular surface, while IgD+ CD27+ B cells were restricted to the follicular surface (Fig. 1f). In the PALS IgM+ CD27+ B cells could only be demonstrated by immunofluorescence, because the intense blue colour of the CD27+ T cells totally prevented detection of immunoglobulin in directly adjacent B cells by immunoenzymatic double-staining.

Figure 2
CD27+ B cells in large secondary follicles and in apparent primary follicles of human spleens. (a) Subtractive double-staining for CD3 (blue-black) and CD27 (brown). The entire follicular periphery outside the GC contains CD27+ B cells. Seventy-nine-year-old ...

A more detailed analysis of B cells with a switched immunoglobulin phenotype was not possible in paraffin sections because – in contrast to the other immunoglobulins analysed – the strong reactivity of interstitial IgG prevented the immunoenzymatic detection of surface IgG on B cells.

CD27+ B cells in large secondary follicles

Two of the spleen specimens investigated revealed secondary follicles with large GCs. The first case was a 79-year-old patient whose spleen was resected because of a gastric carcinoma. In this organ CD27+ B cells were found directly surrounding the GCs, which were well-demarcated by a superficial accumulation of GC T cells (Fig. 2a, b), but did not exhibit dark or light zones. Thus, the entire follicular surface consisted of only one single CD27+ area. Within this area CD27+ B cells tended to be somewhat less densely arranged directly adjacent to the GC and their number increased towards the follicular surface (Fig. 2b). The density of CD27+ B cells slightly varied among different follicles (Fig. 2a, b). IgD+ B cells occupied a broad area in this specimen, which appeared to overlap with the entire CD27+ region (Fig. 2c).

In the second case, a 6-year-old child was splenectomized because of a traumatic splenic rupture. The secondary follicles in this organ contained GCs with fully developed dark and light zones. Double staining revealed that centroblasts were CD27+, while the centrocytes in the light zone expressed CD27 only weakly (Fig. 2d). Large CD3 CD27++ cells resembling the scattered cells in degenerating GCs described above occurred preferentially between the dark and the light zone of the GCs. By immunofluorescence these cells were negative for IgD and IgM. Similar to the majority of spleens investigated, a small CD27 area surrounded the GC followed by the CD27+ follicular surface. The CD27 and the CD27+ area were asymmetrically arranged and orientated away from the PALS (Figs 2d, e). The CD27 region strongly reacted with anti-IgD while the CD27+ area stained only weakly (Fig. 2e). An immunofluorescence double staining experiment (Fig. 2f) showed that a minimum of about 20% of the CD27+ follicular surface area was occupied by IgD+/– CD27+ and by IgM+ CD27+ B cells (IgD+/– CD27+: 20·21 ± 6·79%, IgM+ CD27+: 20·81 ± 6·70%). Similar to the other spleens investigated, IgD+ CD27 and IgM+ CD27 B cells occurred at the follicular surface and along the periphery of the PALS.

Both cases with large GCs indicated that the amount of IgD expressed in CD27+ B cells at the follicular surface was individually variable potentially depending on the stage of GC development or on the age of the patient.

CD27+ B cells in apparent primary follicles

After HE staining two spleen specimens appeared to be entirely composed of primary follicles without GCs. As all follicles looked similar, tangential sectioning did not seem to cause this phenomenon. Staining for IgD revealed the follicles as uniform accumulations of B cells in one specimen and indicated tiny IgD GCs in the other. As we had found CD27+ B cells in the mantle zones of secondary follicles we wondered whether splenic primary follicles might entirely consist of CD27+ B cells. Subtractive double staining for CD3 and CD27 revealed that the entire interior of the follicles in one specimen was indeed populated by CD27+ B cells and a few T cells. However, the density of the CD27+ B cells was much less in the centre compared to the surface of the follicles (Fig. 2g). IgD+ B cells were uniformly and densely distributed in the follicular interior, while the follicular surface also revealed IgD+/– B cells (data not shown). In the second spleen specimen the apparent primary follicles had IgD CD27 centres with irregular contours and accumulations of CD3+ T cells inside. Thus, these follicles obviously represented highly degenerated secondary follicles with remnants of GCs and adjacent broad IgD+ mantle zones, which had not been recognized after HE staining. In the inner part of the follicular periphery IgD+ CD27 B cells prevailed, while at the outermost surface a large number of IgD+/– and CD27+ B cells occurred. However, a substantial number of CD27+ B cells was also present directly adjacent to the GC remnant (Fig. 2h). This staining pattern was similar to that observed in non-reactive parotid lymph nodes (data not shown) which only had small GCs in their follicles. Thus, it remains debatable whether true primary follicles do at all occur in adult lymphatic organs.

CD27+ B cells in the splenic red pulp

Two types of CD3 CD27+ cells occurred in the splenic red pulp. The first population consisted of evenly distributed cells of medium staining intensity and the second of large intensely positive cells which tended to form clusters (data not shown). These latter cells were interpreted to be plasma cells and plasmablasts. Scattered IgD+ cells were found everywhere in the red pulp. In addition, accumulations of IgD+ cells occurred around small arterioles. Cells strongly positive for IgM or for IgA were situated in the red pulp at low density and were classified as plasma cells. In three specimens with degenerated GCs a row of cells strongly positive for intracellular IgM extended along the surface of the PALS.

Lymph nodes, tonsils, appendices and follicles of the terminal ileum

Six specimens each of reactive lymph nodes, tonsils, appendices and follicles of the terminal ileum were investigated (Table 2). Most of these organs had been removed because of acute or chronic inflammation. Two appendices and the majority of the lymph nodes came from areas adjacent to malignant tumours. Thus, it has to be considered that the immunological state of these organs may have been fundamentally different from that of most spleens investigated before. Given this fact, it is not surprising that most organs exhibited full-blown secondary follicles with GCs composed of dark and light zones. The GCs were surrounded by asymmetrical mantle zones orientated towards the lymph node surface or towards the epithelium. The follicles of the terminal ileum were located in the lamina propria as single structures. In two cases they also reached the tela submucosa. They were relatively small and did not represent typical Peyer's patches. The reduced size of the follicles correlated with the rather atrophic state of the intestinal mucosa. In spite of this finding, ileal secondary follicles always had GCs with dark and light zones.

Neither conventional staining nor immunohistology for CD3, CD27 or IgD demonstrated any major differences in the overall composition of secondary follicles in lymph nodes, tonsils, appendices and ileal lamina propria. For this reason, the invariant findings are described below without differentiating between the single organs.

CD27+ B cells at follicular surfaces

Double staining for CD3 and CD27 revealed an area with CD3 CD27+ cells around the GCs in all organs. In five of six lymph nodes, five of six appendices, one of six tonsils and in all follicles of terminal ilea CD27+ B cells were the predominant cell type in this area (Fig. 3a–c, e). In contrast, in one lymph node, one appendix and in the majority of tonsil specimens only few CD27+ B cells occurred around the GCs. In general, the density of CD27+ B cells was lower in the direct vicinity of GCs and increased towards the follicular surface (Fig. 3a, c). The number of CD27+ B cells in the area bordering the GC did, however, differ among individuals and also among single follicles in the same specimen. In an example taken from an appendix, demonstration of IgD in adjacent sections revealed that the area occupied by CD27+ B cells contained IgD+ and also IgM+ B cells and thus corresponded to the mantle zone (Fig. 3b, d). A small number of CD3+ T cells was evenly distributed in this location (Fig. 3a–c). Structures resembling the fibroblast layer or the T cell ring present at the surface of splenic follicles were not prominent in the other lymphatic organs investigated. However, as the follicles were more deeply embedded in the T-cell zone in comparison to spleens, recognition of T-cell rings at the follicular surface might have been precluded.

Figure 3
CD27+ B cells in reactive lymph nodes, appendices, follicles of terminal ileum and tonsils. (a–c, e): Subtractive double-staining for CD3 (blue-black) and CD27 (brown). (a) CD27+ B cells in the mantle zones of cortical follicles outside the GCs. ...

CD27+ B cells in GCs

GCs were well delimited from the mantle zone by a dense accumulation of CD3+ T cells in the most apical part of the light zone orientated towards the lymph node surface or towards the epithelium (Fig. 3a–c). It was evident, that most of the B cells, i.e. the centroblasts, in the dark zone expressed CD27 (Fig. 3b). Tingible body macrophages often stood out as unstained patches in this area. The staining intensity of centroblasts was somewhat variable among the specimens and among different follicles. CD27 positivity diminished towards the light zone and most centrocytes were only faintly stained. Between dark and light zone a prominent accumulation of large CD3 CD27++ B cells was present in almost all GCs (Fig. 3b, e). As mentioned earlier, these cells were tentatively classified as plasmablasts or plasma cells. They were especially prominent in tonsil GCs (Fig. 3e). Double staining for Ki67 plus CD3 versus CD27 in selected tonsil specimens revealed that a substantial number of the CD27++ B cells in the GCs was Ki67 (Fig. 3g), although many Ki67+ B and T cells occurred in the light zone. In contrast to the other organs investigated, certain tonsil GCs exhibited scattered intensely IgD+ B cells (Fig. 3f). Single appendix and tonsil specimens also contained large strongly IgA+ cells in GCs. The distribution of the CD27++ B cells in the GCs did not correspond to that of cells strongly positive for IgD, IgM or IgA. Cells positive for intracellular IgG were, however, frequently found in tonsil GCs. It is thus most likely that the CD27++ B cells represented IgG+ plasma cells or their precursors. In tonsils and other organs large CD27++ B cells also occurred at the outer circumference of the dark and the light zone of many, but not all, GCs (Fig. 3e). The phenomenon was especially evident in tonsils because centroblasts were often less intensely CD27+ in these organs compared to reactive lymph nodes, appendices or ileal follicles.

CD27+ B cells in extrafollicular regions

In human tonsils the subepithelial region has been described as an equivalent of the splenic MZ29 and it was thus analyzed in more detail. In all mucosa-associated lymphatic organs investigated high numbers of CD27++ B cells occurred in a subepithelial position where they tended to form clusters. In the tonsil crypts these cells also occurred within the epithelium. They were interpreted as plasmablasts or plasma cells. In tonsils smaller CD27+ B cells were additionally present beneath the epithelium among the CD27++ cells. This phenomenon was, however, not observed in appendices and terminal ilea, although it cannot be excluded that some CD27+ B cells escaped detection among the many CD27++ cells in the lamina propria. The entire extrafollicular distribution of CD27+ B cells could, however, not be visualized, because – as mentioned above – these cells were not detectable in T-cell zones by immunoenzymatic double staining. Staining for IgA and IgM also revealed many strongly positive cells in the lamina propria below the epithelium with more cells expressing IgA than IgM. In lymph nodes CD27++ B cells occurred in the cortex below the marginal sinus and in the medulla.


Among mature human B lymphocytes CD27+ cells are supposed to represent cells with hypermutated surface immunoglobulin, which comprise immunoglobulin-switched memory cells and non-switched cells with potential antibacterial reactivity (tentatively called natural reactive B cells), as well as plasma cells/plasmablasts. The latter cell type can be differentiated from the former by conventional morphology based on its distribution and size as well as by immunohistological detection of large amounts of intracellular immunoglobulin. Our findings show that CD27+ B cells lacking intracellular IgM, IgG or IgA occur in the periphery of secondary follicles in all human lymphatic organs. This may indicate that the follicular periphery represents a potential recirculation compartment both for naïve and for memory/natural reactive B lymphocytes. Such a hypothesis is in accordance with a high proportion of CD27+ B cells found in human blood.17

We deliberately selected spleens with GCs from patients of different age to be able to accurately define the follicular compartments. Only two of the organs investigated corresponded to the majority of adult human spleens, which do not show GCs by conventional staining. Our findings indicate that the individual state of the immune system is more important for the morphology of splenic GCs than age per se. If adult spleens contain GCs at all, these structures are mostly found in a degenerated state without dark and light zones. This may reveal that – in contrast to mucosa-associated lymphatic organs – the spleen is not permanently confronted with new antigens in adults. We cannot exactly define the state of the GCs present in most of the spleens investigated. Remnants of GCs with a few proliferating B cells inside may persist for rather prolonged periods.

It may be argued that emergency treatment and short-term stress reactions cause the GC morphology observed. However, in specimens obtained during elective surgery the GCs in adult spleens also appeared in a state of decay, while infant spleens electively removed because of spherocytosis or autoimmune haemolytic anaemia (data not shown) and all adult gut-associated follicles always contained full-blown GCs. Thus, short-term effects of surgery or intensive care treatment are not very plausible and an effect related to the individual state of the immune system is much more likely to influence GC morphology. Interestingly, the expression of IgD by CD27+ B cells at the follicular surface seems to correlate with the stage of GC development in the spleen. Among the more than 80 adult spleens investigated in a previous publication12 most spleens with degenerating GCs showed almost no staining for IgD in the area of the CD27+ B cells. Spleens with more developed GCs did, however, often, but not always, exhibit a large number of IgD+/– CD27+ B cells at the follicular surface.

Klein et al.17 found that about 40% of human blood CD27+ memory B cells carry IgG, IgE or IgA and are IgD IgM, 40% only express IgD and IgM and 20% are IgD IgM+. The latter population does not switch to other immunoglobulins and is called ‘IgM-only’ B memory cells. Recent results indicate that ‘IgM only’ B cells expressing hypermutated immunoglobulin may develop independent of T cells and thus do not represent memory cells at all, but special natural reactive B cells.18 Our preliminary immunofluorescence data from two spleen specimens stained for CD27 and IgD or IgM do not entirely correspond to the figures of Klein et al.17 Apart from the fact that more organs need to be investigated, the discrepancies may show that recirculating B memory cells do not quantitatively pass through the spleen or that the composition of the follicular periphery and especially the expression of IgD by CD27+ B cells depends on local factors. In addition, IgD expression was at the lower detection limit of the highly amplifying immunofluorescence method used and may have been underestimated.

To our knowledge, there is only one study analysing human splenic B cells for coexpression of CD27 and IgD by flow cytometry.18 In this study about 17% of the splenic B cells in four children with spherocytosis were CD27+ IgD+ and no individual differences in IgD expression were reported. These results are, however, difficult to compare to ours, because we primarily investigated adults without splenic pathologies, where IgD expression may be less. For quantitative comparison of both methods spleen specimens need to be simultaneously analysed by flow cytometry and triple immunofluorescence on sections to detect B cells, T cells and CD27. This was, however, beyond the scope of our investigation.

The distribution of CD27+ cells in human GCs has not been described in detail so far. We found that centroblasts were CD27+, but centrocytes had a reduced expression. CD27+ centroblasts are also found in mice.27 Isolated human GC B cells up-regulate CD27 under the influence of IL-10.30 The large CD27++ B cells found in all full-blown GCs and also present in degenerating GCs may represent IgG-expressing plasmablasts or plasma cells, because they did not stain for intracellular IgM, IgD or IgA. In hyperplastic tonsils a large proportion of the CD27++ GC B cells were Ki67-negative, favouring their classification as plasma cells. This result corresponds to the experimental finding that plasma cells may arise in GCs after hyperimmunization. It is open to speculation whether these cells differentiate locally from centroblasts or immigrate as precursors. In mice, MZ B memory cells may enter GCs.31 On the other hand, antibody-secreting B cells circulate in the blood of mice and humans after immunization and respond to the chemokine CXCL13.3234 In addition, IgG-secreting cells have been localized in the light zone of human tonsillar GCs.32

CD27+ B cells may be widely distributed in extrafollicular locations. It was described that MZ-type B cells predominantly occur in the subepithelial lamina propria of human tonsils.29 We also found CD27+ B cells in this area, while at the same time the number of CD27+ B cells at the follicular surface of tonsils was less than in the other organs. The subtractive double-staining technique does not permit to visualize B cells in areas of high T-cell density. Thus, our study is not informative with respect to extrafollicular B cells, which need to be investigated with immunofluorescence techniques. Preliminary results of such studies in spleen sections indicate that IgM+ CD27+ B cells but almost no IgD+ CD27+ cells occur in the human PALS. Whether this phenomenon depends on age and immune state of the patients remains to be clarified.

Traditionally, the most superficial part of the human splenic white pulp has been named the MZ in analogy to rodents. The major B-cell population located in rat and mouse MZs is a mixed population of memory or polyreactive cells. These cells tend to stay in the MZ for a prolonged time because of increased expression and binding of integrins.35,36 In addition, mouse MZ B cells are retained in the MZ by interaction with the scavenger receptor MARCO on MZ macrophages.37 The interaction among macrophages and MZ B cells is reciprocal in mice, because, on the other hand, marginal metallophilic macrophages and MZ macrophages only exist when B cells are also present.38 Our previous publications already indicated that the rodent MZ is not equivalent to the area harbouring IgD+/– IgM+ B cells at the surface of human splenic follicles. The microanatomy of the superficial splenic white pulp in rats and mice differs fundamentally from that of humans. We found that humans lack a marginal sinus, marginal metallophilic macrophages and sialoadhesin-positive MZ macrophages.1012 Thus, the border between mantle zone and putative MZ is absent preventing a clearcut distinction between both compartments. In addition, the apparent human MZ is a follicular compartment and does not occur around the T-cell zones. We reported previously that the area appearing as the MZ in humans is subdivided into an inner and an outer part by specialized fibroblasts and CD4+ T cells.12

Considering these facts, the established names of splenic white pulp compartments in rodents cannot be used to describe our findings concerning CD27+ B cells in humans. We thus name the entire region outside the GC the follicular periphery, implying that this location comprises a preferentially IgD+ CD27 inner and a IgD+/– CD27+ outer area. The outermost follicular region also contains some IgD+ CD27 recirculating naïve B cells which also occur at the surface of the T-cell region. We assume that in human spleens migrating CD27+ memory B cells interact at the follicular surface with specialized fibroblasts, which express mucosal addressin cell adhesion molecule-1 (MAdCAM-1), vascular cell adhesion molecule-1, certain cytokeratins, smooth muscle alpha actin, thrombomodulin and CCL21.12 Expression of MAdCAM-1 is species-specific in humans and contrasts with the absence of this molecule in rat spleens.39 MAdCAM-1+ fibroblasts might retard the migration of CD27+ B cells so that they appear to aggregate in a superficial position around the follicles. This may be especially relevant for integrin alpha 4/beta 7-positive B memory cells arriving from the gut. We now show that B cells with a high surface expression of IgA, which most likely represent IgA memory cells, lodge near the fibroblasts in the outer periphery of human splenic follicles. This may indicate that IgA+ memory B cells preferentially recirculate between mucosae and spleen. It remains to be investigated whether these cells are also present in the follicular periphery of mesenteric lymph nodes. In the periphery of splenic follicles CD3+ CD4+ T cells accompany the special fibroblasts, so that an extension of the T-cell zone appears to be formed. Three-dimensional reconstruction revealed that this arrangement is not caused by a sectioning artefact.40

In its fundamental structure the surface of splenic follicles corresponds to that of follicles in other human lymphatic organs. The identity is, however, not total. Up to now we have not found specialized fibroblasts and rings of CD4+ T cells in follicles outside the spleen. In addition, CD27+ B cells occur in the periphery of splenic follicles more constantly and in larger numbers than in the other organs. It may be argued that the presence of CD27+ B cells in the mantle zone of certain spleens and other organs is entirely due to generalized or local immune reactions that promote the entry of these cells into the recirculation compartment. Thus, in entirely normal organs CD27+ B cells would never occur close to GCs. The interindividual variations observed indeed indicate a dynamically changing distribution of CD27+ B cells. However, ongoing immune reactions cannot be the only cause of this phenomenon, because spleens with apparent primary follicles (Fig. 2h) and non-reactive lymph nodes removed together with benign parotid tumours (data not shown) also exhibit a substantial number of CD27+ B cells in the mantle zone.

Thus, the distribution of CD27+ B cells observed may indicate species differences in the migration behaviour of B memory cells. It may be that a large number of mouse and rat B memory cells are more or less sessile in the splenic MZ, while these cells constantly recirculate in humans. The fact that memory cells with hypermutated immunoglobulin variable regions amount to 40% of blood B cells in humans, but only to 5% in mice19 is compatible with this assumption. However, memory cell accumulation depending on the longevity of the respective species may also explain the different percentages.

Several authors have described that special B cells are present at the surface of follicles in human lymphatic organs other than the spleen. Thus, an extensive MZ was observed in reactive lymph nodes after conventional staining.41 This indicates that under certain conditions memory B cells may occupy even larger follicular areas in lymph nodes than observed in our study. Spencer et al.42 already proposed that memory B cells are located superficially in human Peyer's patch follicles. Microdissection from human Peyer's patches43 and from an IgD CD23 region at the follicular surface of human lymph nodes44 revealed memory B cells with hypermutated immunoglobulin variable regions. In these studies B cells with non-mutated immunoglobulin V-regions were also found among the hypermutated cells showing that both cell types occur in the same microanatomical region.

Our study shows that the microanatomical differences among the white pulp surface in rodent and human spleens are so extensive that the term marginal zone is difficult to apply in humans. If CD27 is used as an antigen to demonstrate marginal zone-type B cells it is necessary to caution researchers that the entire follicular surface, which is traditionally called the mantle zone, may be occupied by a mixture of IgD+ CD27 and IgD+/– CD27+ B cells in several human lymphatic organs, including the spleen. In the spleen the tendency of naïve IgD+ CD27 B cells to predominate in the inner area of the follicular surface and of potentially recirculating IgD+/– CD27+ memory-type B cells to sojourn more superficially is most conspicuous. CD27+ B cells may occur in mantle zones because of ongoing immune reactions, but they are also found in this location without morphological or clinical evidence of immunological involvement.


This work was supported by grant Ste 360/10-1 of the Deutsche Forschungsgemeinschaft. Our thanks are due to K. Lampp and A. Seiler for expert technical assistance. The support by V. Stachniss, Department of Dentistry of Marburg University Hospital was fundamental for setting up digital photography. E. Weihe and M. Bette, Department of Neurobiology, Institute of Anatomy of Marburg University, kindly provided reagents and access to a microscope. I. Dammshäuser, Department of Reproductive Biology, Institute of Anatomy of Marburg University, donated biotinylated tyramide. The images were skilfully and empathically processed for publication by S. Horn, Institute of Molecular Biology and Tumour Research, University of Marburg. J. Jäkel, Institute of Pathology of Marburg University, contributed to this task.


1. MacLennan ICM, Liu Y-J. Marginal zone B cells respond both to polysaccharide antigens and protein antigens. Res Immunol. 1991;142:346–51. [PubMed]
2. Martin F, Kearney JF. Marginal zone B cells. Nature Rev Immunol. 2002;2:323–35. [PubMed]
3. Kanayama N, Cascalho M, Ohmori H. Analysis of marginal zone B cell development in the mouse with limited B cell diversity. role of the antigen receptor signals in the recruitment of B cells to the marginal zone. J Immunol. 2005;174:1438–45. [PubMed]
4. Kumararatne DS, Bazin H, MacLennan ICM. Marginal zones: the major B cell compartment of rat spleens. Eur J Immunol. 1981;11:858–64. [PubMed]
5. Kraal G. Cells in the marginal zone of the spleen. Int Rev Cytol. 1992;132:31–74. [PubMed]
6. Gray D, MacLennan ICM, Bazin H, Khan M. Migrant µ+δ+ and static µ+δ B lymphocyte subsets. Eur J Immunol. 1982;12:564–9. [PubMed]
7. Lopes-Carvalho T, Kearney JF. Development and selection of marginal zone B cells. Immunol Rev. 2004;197:192–205. [PubMed]
8. Liu Y-J, Oldfield S, MacLennan ICM. Memory B cells in T cell-dependent antibody responses colonize the splenic marginal zones. Eur J Immunol. 1988;18:355–62. [PubMed]
9. Dammers PM, Visser A, Popa ER, Nieuwenhuis P, Kroese FGM. Most marginal zone B cells in rat express germline encoded IgVH genes and are ligand selected. J Immunol. 2000;165:6156–69. [PubMed]
10. Steiniger B, Barth P, Herbst B, Hartnell A, Crocker PR. The species-specific structure of microanatomical compartments in the human spleen: Strongly sialoadhesin-positive macrophages occur in the perifollicular zone, but not in the marginal zone. Immunology. 1997;92:307–16. [PMC free article] [PubMed]
11. Steiniger B, Barth P. Microanatomy and function of the spleen. Adv Anat Embryol. 2000;151:1–100. [PubMed]
12. Steiniger B, Hellinger A, Barth P. The perifollicular and marginal zones of the human splenic white pulp: do fibroblasts guide lymphocyte immigration? Am J Pathol. 2001;159:501–12. [PMC free article] [PubMed]
13. Dunn-Walters DK, Isaacson PG, Spencer J. Analysis of mutations in immunoglobulin heavy chain variable region genes of microdissected marginal zone (mgz) B cells suggests that the mgz of human spleen is a reservoir of memory B cells. J Exp Med. 1995;182:559–66. [PMC free article] [PubMed]
14. Weller S, Reynaud CA, Weill JC. Vaccination against encapsulated bacteria in humans: paradoxes. Trends Immunol. 2005;26:85–9. [PubMed]
15. Weill JC, Weller S, Reynaud CA. A bird's eye view on human B cells. Semin Immunol. 2004;16:277–81. [PubMed]
16. Agematsu K. Memory B cells and CD27. Histol Histopathol. 2000;15:573–6. [PubMed]
17. Klein U, Rajewsky K, Küppers R. Human immunoglobulin (Ig) M+IgD+ peripheral blood B cells expressing the CD27 cell surface antigen carry somatically mutated variable region genes: CD27 as a general marker for somatically mutated (memory) B cells. J Exp Med. 1998;188:1679–89. [PMC free article] [PubMed]
18. Weller S, Braun MC, Tan B, et al. Blood IgM ‘memory’ B cells are circulating splenic marginal zone B cells in humans. Blood. 2004;104:3647–54. [PMC free article] [PubMed]
19. Schittek B, Rajewsky K. Natural occurrence and origin of somatically mutated memory B cells in mice. J Exp Med. 1992;176:427–38. [PMC free article] [PubMed]
20. Kruetzmann Rosado MM, Weber H, et al. Human immunoglobulin M memory B cells controlling Streptococcus pneumoniae infections are generated in the spleen. J Exp Med. 2003;197:939–45. [PMC free article] [PubMed]
21. Tangye SG, Liu YJ, Aversa G, Phillips JH, de Vries JE. Identification of functional human splenic memory cells by expression of CD148 and CD27. J Exp Med. 1998;188:1691–703. [PMC free article] [PubMed]
22. Zandvoort A, Lodewijk ME, de Boer NK, Dammers PM, Kroese FGM, Timens W. CD27 expression in the human splenic marginal zone: the infant marginal zone is populated by naïve B cells. Tissue Antigens. 2001;58:234–42. [PubMed]
23. Tomiyama H, Takata H, Matsuda T, Takiguchi M. Phenotypic classification of human CD8+ T cells reflecting their function: inverse correlation between quantiative expression of CD27 and cytotoxic effector function. Eur J Immunol. 2004;34:999–1010. [PubMed]
24. Hendriks J, Gravestein LA, Tesselaar K, van Lier RAW, Schumacher TNM, Borst J. CD27 is required for generation and long-term maintenance of T cell immunity. Nat Immunol. 2000;1:433–40. [PubMed]
25. Raman VS, Akondy RS, Rath S, Bal V, George A. Ligation of CD27 on B cells in vivo during primary immunization enhances commitment to memory B cell responses. J Immunol. 2003;171:5876–81. [PubMed]
26. Agematsu K, Nagumo H, Oguchi Y, et al. Generation of plasma cells from peripheral blood memory B cells: synergistic effect of interleukin-10 and CD27/CD70 interaction. Blood. 1998;91:173–80. [PubMed]
27. Xiao Y, Hendriks J, Langerak P, Jacobs H, Borst J. CD27 is acquired by primed B cells at the centroblast stage and promotes germinal center formation. J Immunol. 2004;172:7432–41. [PubMed]
28. Hintzen RQ, Lens SMA, Lammers K, Kuiper H, Beckmann MP, van Lie RAW. Engagement of CD27 with its ligand CD70 provides a second signal for T cell activation. J Immunol. 1995;154:2612–23. [PubMed]
29. Dono M, Zupo S, Leanza N, Melioli G, Fogli M, Melagrana A, Chiorazzi N, Ferrarini M. Heterogeneity of tonsillar subepithelial B lymphocytes, the splenic marginal zone equivalents. J Immunol. 2000;164::5596–604. [PubMed]
30. Jung J, Choe J, Li L, Choi YS. Regulation of CD27 expression in the course of germinal center B cell differentiation: the pivotal role of IL-10. Eur J Immunol. 2000;30:2437–43. [PubMed]
31. Song H, Cerny J. Functional heterogeneity of marginal zone B cells revealed by their ability to generate both early antibody-forming cells and germinal centers with hypermutation and memory in response to a T-dependent antigen. J Exp Med. 2003;198:1923–35. [PMC free article] [PubMed]
32. Arce S, Luger E, Muehlinghaus G, et al. CD38 low IgG-secreting cells are precursors of various CD38 high-expressing plasma cell populations. J Leukoc Biol. 2004;75:1022–8. [PubMed]
33. Odendahl M, Mei H, Hoyer BF, et al. Generation of migratory antigen-specific plasma blasts and mobilization of resident plasma cells in a secondary immune response. Blood. 2005;105:1614–21. [PubMed]
34. Blink EJ, Light A, Kallies A, Nutt SL, Hodgkin PD, Tarlinton DM. Early appearance of germinal center-derived memory B cells and plasma cells in blood after primary immunization. J Exp Med. 2005;201:545–54. [PMC free article] [PubMed]
35. Lu TT, Cyster JG. Integrin-mediated long-term B cell retention in the splenic marginal zone. Science. 2002;297:409–12. [PubMed]
36. Gatto D, Ruedl C, Odermatt B, Bachmann MF. Rapid response of marginal zone B cells to viral particles. J Immunol. 2004;173:4308–16. [PubMed]
37. Karlsson MCI, Guinamard R, Bolland S, Sankala M, Steinmann RM, Ravetch JV. Macrophages control the retention and trafficking of B lymphocytes in the splenic marginal zone. J Exp Med. 2003;198:333–40. [PMC free article] [PubMed]
38. Nolte MA, Arens R, Kraus M, van Oers MHJ, Kraal G, van Lier RAW, Mebius RA. B cells are crucial for both development and maintenance of the splenic marginal zone. J Immunol. 2004;172:3620–7. [PubMed]
39. Iizuka T, Tanaka T, Suematsu M, et al. Stage-specific expression of mucosal addressin cell adhesion molecule-1 during embryogenesis in rats. J Immunol. 2000;164:2463–71. [PubMed]
40. Steiniger B, Rüttinger L, Barth PJ. The three-dimensional structure of human splenic white pulp compartments. J Histochem Cytochem. 2003;51:655–63. [PubMed]
41. Van den Oord JJ, de Wolf-Peeters C, Desmet VJ. The marginal zone in the human reactive lymph node. Am J Clin Pathol. 1986;86:475–9. [PubMed]
42. Spencer J, Finn T, Pulford KAF, Mason DY, Isaacson PG. The human gut contains a novel population of B lymphocytes which resemble marginal zone cells. Clin Exp Immunol. 1985;62:607–12. [PMC free article] [PubMed]
43. Dunn-Walters DK, Isaacson PG, Spencer J. Sequence analysis of rearranged IgVH genes from microdissected human Peyer's patch marginal zone B cells. Immunology. 1996;88:618–24. [PMC free article] [PubMed]
44. Tierens A, Delabie J, Michiels L, Vandenberghe P, De Wolf-Peeters C. Marginal-zone B cells in the human lymph node and spleen show somatic hypermutations and display clonal expansion. Blood. 1999;93:226–34. [PubMed]

Articles from Immunology are provided here courtesy of British Society for Immunology
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...