We have shown that Peyer’s patch (PP) first develops as a simple and even cell aggregation during embryogenesis. To investigate when and how such a simple cell aggregation forms the complex PP architecture, we analyzed the distribution of cells expressing IL-7Rα (PP inducer cells), VCAM-1 (mesenchymal cells), CD11c (dendritic cells), and mature lymphocytes by whole-mount immunostaining of 17.5 days postcoitus to 2 days postpartum mouse gut. Our results show that compartmentalization of PP anlagen commences at day 18.5 of gestation by clustering and subsequent follicle formation of IL-7Rα+, VCAM-1+, and CD11c+ cells. This process adds the primitive architecture of PP anlage with several follicles in which IL-7Rα+ cells localize in the center, while VCAM-1+ and CD11c+ cells localize at the fringe. This follicle formation is accompanied by the establishment of PP-specific vascular network expressing mucosal addressin cellular adhesion molecule-1. Mature B and T lymphocytes entering in the PP anlage are distributed promptly to their own target zones; B cells to the follicle and T cells to nonfollicular zones. Our analysis of scid/scid mouse indicate that the initial processes including formation of PP-specific vascular network occur in the absence of lymphocytes. These observations indicate that the basic architecture of PP is formed by a set of cell lineages assembled during the initial phase of induction of PP anlagen before entry of mature lymphocytes.
Peyer’s patch (PP)3 organogenesis proceeds through three distinct steps that have been identified immunohistologically (1). The appearance of discrete spots expressing VCAM-1 and ICAM-1 is the earliest identifiable step and occurs at 15 days postcoitus (dpc) in the mouse. One to 2 days later, a set of hemopoietic cells, which are identifiable by expression of surface markers such as CD45, IL-7Rα, CD11c, and c-Fms, accumulate, though no mature lymphocytes are found at this time. Entry of mature lymphocytes expressing CD3 or B220 occurs after 18.5 dpc. We have demonstrated that the initial step of PP organogenesis, represented by the appearance of VCAM-1+/ICAM-1+ spots, is induced by IL-7Rα+ cells that express lymphotoxin (LT) α1β2 upon stimulation of IL-7Rα/common γ/Jak3 signaling pathway (2). Analysis of mice bearing null mutations of genes involved in this induction phase concurs with the three-step model proposed above (3, 4, 5). How primitive PP anlagen develops into more complex structures as found in the fully developed PP is unknown.
The structure of the mature PP is similar to that of other peripheral lymphoid tissues in that it consists of a number of discrete lymphoid follicles in which T, B, and dendritic cells are located in an organized manner (6). By antigenic stimulation, primary follicles undergo structural changes, producing highly developed secondary architectures with germinal centers (7, 8). While this latter process generates extensive complexity in the peripheral lymphoid tissues, the basic tissue framework has to be prepared before Ag stimulation. Indeed, studies of germfree mice show that lymphoid follicles with segregated areas of T and B cells are formed in an Ag-independent manner (9). Furthermore, Noelle et al. clearly demonstrated that the cellular organization that directs T and B cells into segregated regions is established before the entry of lymphocytes (10). These results suggest that the primitive PP anlage identified initially as a simple structure in which VCAM-1/ICAM-1+ cells are scattered rather homogeneously becomes more complex via a series of compartmentalizing events before lymphocyte entry.
The first primary follicles, indicated by B cell aggregates, are seen 2 wk after birth in the rat (11, 12, 13) and 10 days after birth in the mouse (14, 15). In this study, we investigated the process by which the basic structure of PP with segregated compartments is established from the primitive PP anlagen. Our results show that the formation of basic architecture of PP with separate follicle-like structures commences from 18.5 dpc and is completed within a few days after birth. Moreover, this process is independent of mature lymphocytes, which have been shown to be the chief components inducing PP architecture in later life.
Materials and Methods
C57BL/6 strain and C.B17/Icr-SCID Jcl were purchased from Japan SLC (Shizuoka, Japan) and Japan CLEA (Tokyo, Japan), respectively. Female and male mice were allowed to mate overnight, and those with vaginal plugs were judged pregnant. Noon of the day when the vaginal plug was found was designated 0.5 dpc.
Monoclonal Abs against VCAM-1 (429 MVCAM.A; BD PharMingen, San Diego, CA), CD11c (HL3; BD PharMingen), mucosal addressin cellular adhesion molecule (MAdCAM)-1 (MECA367; BD PharMingen), platelet endothelial cellular adhesion molecule (PECAM)-1 (MEC13.3; BD PharMingen), CD3 (Y65.372; Seikagaku, Tokyo, Japan; and 2C11; BD PharMingen) were purchased. Monoclonal Ab against IL-7Rα (A7R34) (16), B220 (RA3-6B2) and Flk-1(AVAS12) (17) were purified from hybridoma culture supernatant as described.
Whole-mount immunostaining was performed as previously described (18) with slight modifications. In brief, excised guts were incubated in fixing solution (4% paraformaldehyde in PBS) for 30 min at 4°C. After absorbing excessive paraformaldehyde with 4% glycine in PBS for 30 min at 4°C, specimens were dehydrated by incubating 30 min each with 50, 75, 100, 100% methanol in PBS at 4°C. To block endogenous peroxidase, the fixed specimens were bleached (methanol: 30% H2O2 20:1) for 30 min at room temperature. For staining, the dehydrated specimens were first blocked by incubating twice in PBSMT (1.5% skim milk and 0.1% Triton X-100 in PBS) for 1 h at room temperature, incubated with PBSMT containing biotinylated anti-CD11c mAb HL3 (1 μg/ml), anti-IL7Rα mAb A7R34 (2 μg/ml), anti-VCAM-1 mAb 429 (2 μg/ml), anti-MAdCAM-1 mAb MECA-367 (2 μg/ml), anti-PECAM-1 mAb MEC13.3 (1 μg/ml), biotinylated anti-CD3 mAb Y65.372 (1 μg/ml), biotinylated anti-CD3 mAb 2C11 (0.5 μg/ml), or anti-B220 mAb RA3-6B2 (2 μg/ml) overnight at 4°C. After washing five times in PBSMT each for 1 h at 4°C, the primary Ab was detected by incubating 1 mg/ml HRP-conjugated anti-rat Ig Ab (Biosource, Camarillo, CA) overnight at 4°C. To biotinylated mAbs, ABC kits (Vectastain Elite ABC Kit; Vector Laboratories, Burlingame, CA) were used at 1:20 dilution in PBS, incubated for 2 h, and washed three times for 30 min at 4°C. After extensive washing with more than five exchanges of PBSMT, including the final 20 min wash in PBST (0.1% Triton X-100 in PBS) at room temperature, specimens were soaked in PBST containing 0.05% NiCl2 and 250 mg/ml diaminobenzidine (DAB; Dojin Chem, Kumamoto, Japan) for 10–30 min, and H2O2 was added to 0.01%. The enzymatic reaction was allowed to proceed until the desired color intensity was reached, and the specimens were rinsed three or four times in PBST. Finally, the specimens were dehydrated and solution was exchanged to glycerol before being photographed. Then 14.5-dpc spleen was whole-mount immunostained by MAdCAM-1 and was embedded in polyester wax and sectioned by microtome. The specimen was further immunostained by a rat anti-mouse Flk-1 mAb (10 μg/ml), followed by the peroxidase-labeled anti-rat IgG. To obtain a brown color, the coloration reaction was performed in the absence of NiCl2. It was observed by differential-interference-contrast microscopy.
Whole-mount immunofluorescent staining was conducted in combination with the whole-mount immunostaining described above. FITC-conjugated HL3 (0.5 μg/ml) with Alexa594-conjugated A7R34 (1 μg/ml) in combination with biotinylated 429 (2 μg/ml) was used as a first step Ab in the above protocol. The location of each PP was identified by detecting 429 mAb binding by DAB/NiCl2 color-substrate reaction. Single PPs were then excised and mounted flat. Similarly, combinations of HL3-FITC/429-biotin followed by streptavidin (SA)-Alexa594 and anti-rat Ig HRP, 492-FITC/Alexa594-A7R/biotin-CD11c followed by SA-HRP, as well as A7R/2C11-biotin followed by SA-Texas Red (Molecular Probes, Eugene, OR) and anti-rat Ig HRP were also developed by DAB/NiCl2 and were excised and mounted flat. Fluorescent signals were detected by confocal microscopy (TCS-NT/Leica, Bensheim, Germany) or light microscopy (Axiophoto/Zeiss, München, Germany).
Preparation of single-cell suspensions for flow cytometry and cell sorting
The whole gut of embryonic and neonatal mice was chopped to small pieces with dissecting scissors and incubated in 2 U/ml(2.60 mg/ml) dispase (Life Technologies, Rockville, MD) in PBS for 30 min at 37°C. In some experiments, two to three PP in the upper intestine were dissected from intestines using a stereomicroscope and treated with the dispase solution. After pipetting 20 times, dissociated cells were washed with HBSS containing 2% BSA (HBSS/BSA; Sigma, St. Louis, MO). Cells were filtered through nylon mesh to remove large clumps, washed with HBSS/BSA, and centrifuged for 10 min at 1000 rpm at 4°C. Nonspecific binding of mAb was blocked by suspending cells (1 × 106) in 50 μl of normal mouse serum for 20 min at 4°C, stained with FITC-conjugated 429 (10 μg/ml) or Alexa488-conjugated A7R34 (5 μg/ml) with PE-conjugated HL3 (10 μg/ml) or biotin-conjugated A7R34 (5 μg/ml) added with PE-conjugated SA (5 μg/ml). Stained cells were washed twice in HBSS/BSA and analyzed using FACSVantage (Becton Dickinson, Mountain View, CA).
Cell preparations for RT-PCR analysis
For RT-PCR analysis, cells were labeled with A7R34-Alexa488 and HL3-PE. After living cells were selected by staining with propidium iodide (Sigma), IL-7Rα single-positive cells and CD11c single-positive cells was sorted using FACSVantage (Becton Dickinson). To obtain pure cell populations, each population was subjected to two cycles of sorting. The purity of each cell population ranged from 89 to 96%.
RT-PCR detection of chemokine receptors in sorted mouse gut cell populations
Total RNA was extracted from 70,000 cells directly sorted into vials containing ISOGEN LS (Nippon Gene, Tokyo, Japan), and total RNA was isolated according to the manufacturer’s protocol. First-strand cDNA was prepared from total RNA by reverse transcriptase using oligo(dT) primers (Superscript2; Life Technologies). In each PCR for detection of CXCR5, CCR7, and GAPDH gene expression, cDNA corresponding to an equivalent of 3500 cells was incubated with 100 pg of the following primer sets; B lymphocyte chemoattractant (BLC), sense 5′-TCACCTTGTTGGGTACCCCAGCAA-3′, antisense 5′-ATACACAGACTTCTGCGCAC-3′; CXCR5, sense 5′-TTGTTGGCAGTGCCTATCACTGTCC-3′, antisense 5′-CTCGTGTACCATAACGACCCGTAC-3′; EBI1-ligand chemokine (ELC), sense 5′-CTGCCTCAGATTATCTGCCAT, antisense 5′-GCCAGAGTGATTCACATCTCT; secondary lymphoid-tissue chemokine (SLC), sense 5′-ATGGCTCAGATGATGACTCTG, antisense 5′-GTGTCTGTTCAGTTCTCTTGC; CCR7, sense 5′-ACAGCGGCCTCCAGAAGAACAGCGG, antisense 5′-TGACGTCATAGGCAATGTTGAGCTG; GAPDH, sense 5′-ATGGTGAAGGTCGGTGTGAACGGATTTGGC-3′, antisense 5′-GC ATCGAAGGTGGAAGAGTGGGAGTTGCTG-3′.
Semiquantitative RT-PCR analysis was performed as follows. The amount of the template cDNA from each sample was equalized by comparing the concentration of RT-PCR products with GAPDH primer sets at various amplifying cycles. The minimum number of amplifying cycles that detected the PCR products were determined for each specimen by each primer set.
Cell migration into separate compartments during PP organogenesis
Our previous studies defined the primitive PP anlage as a discrete region expressing VCAM-1/ICAM-1 (1). Judging from the homogeneous staining pattern of VCAM-1 (Fig. 1,A) and IL-7Rα (Fig. 1,C) within the primitive anlagen, the structure appears relatively simple with evenly distributed VCAM-1+ and IL-7Rα+ cells. However, in the adult PP, VCAM-1+ (Fig. 1,B) and IL-7Rα+ (Fig. 1 D) cells have separated into distinct units in which T and B lymphocytes and dendritic cells are segregated within different compartments. Thus, the precisely regulated process of cell segregation in the PP must take place between the formation of the primitive anlagen and mature PP. While Ag-induced reorganization of the architecture of peripheral lymphoid tissues has been defined to some extent, earlier processes remains completely obscure. Therefore, we first attempted to define the process in which a simple PP anlage develops higher order structures.
To investigate the initial process of compartmentalization, the localization of IL-7Rα+, VCAM-1+, and CD11c+ cells was analyzed after the formation of primitive PP anlage. We have proposed that IL-7Rα+ and VCAM-1+ cells form a mutually interacting unit for making an organizing center of the PP anlage (19). Whole-mount preparations of 17.5-dpc to 2-days postpartum (dpp) embryonic intestines were stained by anti-IL-7Rα, anti-VCAM-1, and anti-CD11c mAbs. In Fig. 2, two types of staining pattern were observed. IL-7Rα+ or VCAM-1+ cells, which are distributed diffusely in the PP anlagen (Fig. 2, A and F), first formed irregularly segregated subregions (Fig. 2, B, C, G, and H), which subsequently split into several apparent compartments of similar size (Fig. 2, D, E, I, and J). CD11c+ cells first appear in the periphery of PP anlagen (Fig. 2,K) but eventually fused to the IL-7Rα+ area after birth (Fig. 2, L–O). These observations indicated that the three markers that we have used to define PP inducers, PP organizers (which we defined as a population activated by the inducer), and dendritic cells are sufficient to detect the initial process toward compartmentalization of PP from 18.5 dpc. Within 2 days of birth, positioning of each component is complete, thereby forming several follicle like structures (Fig. 2, E, J, and O).
Analysis of immunofluorescent staining of each follicle-like structure shows distinct distribution patterns for each cell population whereby IL-7Rα+ cells are concentrated in the center (Fig. 3, B and H), while VCAM-1+(Fig. 3, E and G) and CD11c+ (Fig. 3, A and D) components accumulate at the fringe. However, these regions are not demarcated sharply.
To show that CD11c+, VCAM-1+, and IL-7Rα+ cells are distinct populations, we prepared single-cell suspensions from two to three PPs from upper intestines of 2-day-old neonates. As the size of PP at this stage is too small to dissect specifically, the cell preparation may contain surrounding tissues. Thus, the percentage of each population in Fig. 4 may not represent the actual proportion in the PP. Nevertheless, Fig. 4 demonstrated clearly that expression of VCAM-1, IL-7Rα, and CD11c are completely segregated into distinct cell populations. These results suggest that initial compartmentalization of the PP anlagen is established through coordinated migration of distinct cell types that assembled in the primitive PP anlagen.
Expression of chemokine receptors in the initial phase of PP compartmentalization
Chemokines and their receptors play an essential role in regulating the compartmentalization of hemopoietic cells in the secondary lymphoid organs (20, 21, 22, 23, 24). Among the three cellular components, CD11c+ and IL-7Rα+ cells are CD45+ hemopoietic cells (25). To gain insight into the mechanisms underlying the distinct movement of the two hemopoietic cell populations during the initial phase of compartmentalization, CD11c+ and IL-7Rα+ cells were sorted to a purity of 89–96%, and the expression of chemokine receptors was analyzed by RT-PCR. As shown in Fig. 4 B, IL-7Rα+ cells expressed both CXCR5 and CCR7, whereas CD11c+ cells expressed only CCR7. While expression level may decrease during the process of compartmentalization, expression of CXCR5 and CCR7 in each population is maintained throughout the process from 17.5 dpc to 2 dpp. As BLC, ELC, and SLC are expressed in the VCAM-1+ stromal cell component in the PP anlagen (Ref. 38 and K. Honda, unpublished observation), IL-7Rα+ and CD11c+ cells are likely to have different affinities for regions expressing these chemokines.
Development of vascular structures
High endothelial venules (HEVs) are a hallmark of peripheral lymphoid organs. In adult PP, MAdCAM-1 is specifically expressed on HEV and functions as an adhesion molecule during lymphocyte homing (26, 27). Mebius et al. demonstrated that MAdCAM-1 is expressed by HEV in lymph nodes from the neonatal stage (28), suggesting that this specialized vascular structure may also be established in parallel with the compartmentalization process of PP described above. Thus, we investigated the temporal relationship between compartmentalization of PP and development of vascular architecture.
While we initially expected MAdCAM-1 staining to reveal the process of HEV development in PP, we found that MAdCAM-1 is detectable in the intestine before the formation of PP anlagen. As shown in Fig. 5,A, MAdCAM-1 expression was detectable in the embryo as early as 9.5 dpc. As compared with PECAM-1, which is expressed by all vascular endothelial cells (ECs) (29) (Fig. 5,D), MAdCAM-1 is expressed specifically by venous ECs. The vein-specific expression of MAdCAM-1 is most clearly displayed in the mesenteric veins that run parallel to arteries and lymphatics (Fig. 5,B) that express PECAM-1 but not MAdCAM-1 (Fig. 5, E and F). Moreover, analyses of other regions demonstrated that venous specific expression of MAdCAM-1 is a general phenomenon during development of vascular system (data not shown). While most venous EC express MAdCAM-1 during embryogenesis, the expression disappears by birth except in PP (Fig. 5,C), spleen, and mesenteric lymph nodes (MLN). As detected by expression of nkx-2.5, the primordium of the spleen is first formed at 10.5 dpc (30). Expression of MAdCAM-1 is detected in the vascular structures of the developing spleen from 12.5 dpc. Double staining of 14.5-dpc spleen by anti-Flk-1 and anti-MAdCAM-1 (Fig. 5 G) shows that all vascular structures are Flk-1+, whereas only a portion are stained with MAdCAM-1. The primordium of MLN can also be detected as aggregates of IL-7Rα+ cells at 12.5 dpc (data not shown). MAdCAM-1 expression on vascular structures is seen from 15.5 dpc in MLNs (data not shown). These results suggests that venous-specific expression of MAdCAM-1 is a general phenomenon during organogenesis of peripheral lymphoid tissues.
In the developing PP anlagen, MAdCAM-1 was first detected on 16.5 dpc (Fig. 2,P). While MAdCAM-1 expression in the PP anlagen starts from a homogeneous pattern, a strong net-like expression emerges on 17.5 dpc (Fig. 2,P) and eventually becomes restricted to HEV (Fig. 2,T). Expression in surrounding tissues disappears gradually. From 18.5 dpc (Fig. 2,Q) to 0 dpp (Fig. 2,R), when entry of mature lymphocytes is observed, MAdCAM-1 is strongly expressed on vessels in the PP anlagen, though compartmentalization is not yet observed. From 1 dpp, the distribution pattern of MAdCAM-1+ vessels changes in parallel with the compartmentalization of VCAM-1+, IL-7Rα+, and CD11c+ cells. MAdCAM-1+ vessels become associated with the developing follicle-like structure by 1 dpp (Fig. 2,S) and start to surround them from 2 dpp (Fig. 2 T), while the compartmentalization appeared to occur earlier. This time course indicates that the process of cell compartmentalization is followed by reorganization of vascular architecture, thereby each newly formed compartment is supplied with discrete vascular units.
Entry of mature lymphocytes
Our previous study demonstrated that entry of mature lymphocytes to PP anlagen is detected primarily at around 18.5 dpc (1). B220+ cells (Fig. 6, A and B) and CD3+ (Fig. 6, E and F) cells were distributed evenly over the PP anlagen from 18.5 dpc to 0 dpp. This diffuse pattern is consistent with the architecture of vascular network before compartmentalization. Along with formation of follicle-like structure with MAdCAM-1+ HEV, B220+ cells accumulate in the central region of each follicle-like structure (Fig. 6,C) and eventually show a similar distribution pattern as that of IL-7Rα+ cells (Fig. 6,D). In contrast, T cells remain over the outside space of follicle-like structures (Fig. 6, G and H). Note that this pattern is different from those of IL-7Rα+, CD11c+, and VCAM-1+ cells. However, in the fully mature PP, T cell distribution became similar to that of CD11c, indicating that the segregation pattern of these two cell types found in early neonates is transitional.
Early compartmentalization is independent of mature lymphocyte entry
Mature lymphocytes play an important role in the formation of lymphoid follicles (10). Thus, it is likely that lymphocyte entry into the PP anlagen provides momentum to induce cell segregation and subsequent outward movement of the follicle-like structures within the PP. To test this possibility, we stained the gut of scid/scid mice in which mature lymphocytes are rarely observed (31). As shown in Fig. 7, all processes of compartmentalization described in the preceding section occur normally in the gut of scid/scid mouse, although the size of scid/scid PP were much smaller. This conclusion was confirmed by staining >10 PPs from each scid/scid mouse. As Croitoru et al and Falk et al. have shown that adult scid/scid (32) and rag1 or rag2 null mutant mice have CD3+ intraepithelial lymphocytes (33), it is possible that these mature lymphocytes are involved in PP organogenesis in scid/scid mice. However, we could not detect any CD3- nor B220-positive cells by whole-mount immunostaining of guts of 17.5-dpc to 2-dpp scid/scid mice (Fig. 8; data not shown). Defective lymphogenesis in each scid/scid mouse was also confirmed by staining the spleen by anti-CD3 and anti-B220. This result indicates that the early compartmentalization process with concomitant formation of the vascular system is independent of entry of mature lymphocytes.
Our previous studies analyzed the initial steps of PP organogenesis by which VCAM-1/ICAM-1+ regions are induced on the antimesenteric side of embryonic intestine (1). Induced PP anlagen, appear as homogeneous regions as revealed by immunostaining with various mAbs. In contrast, the architecture of mature PP consists of several discrete follicles, each being further divided into segregated compartments (34). Thus, a series of compartmentalization processes must be initiated after the formation of the PP anlagen, histological dissection of which was the major aim of this study.
Using whole-mount immunostaining, we succeeded to detect the first processes of compartmentalization as early as 18.5 dpc. The first event detected in this study is outward spreading of IL-7Rα+ cells that rapidly develop into several segregated compartments. As this population is also present in scid/scid mice, IL-7Rα+ cells may represent LTα1β2-producing PP inducers that we have characterized in previous studies (2). Almost concomitant with this event, diffusely stained VCAM-1+ regions segregate to form highly complex structures by 2 dpp. It should be noted that IL-7Rα+ and VCAM-1+ cells behave similarly during the initial step of the compartmentalization. This is consistent with our hypothesis that two populations interact reciprocally to form an organizing center for PP organogenesis.
CD11c+ dendritic cells are also one of the major cell populations within the secondary lymphoid organs and are found from an early phase of PP organogenesis. As compared with IL-7Rα+ cells, which are distributed homogeneously in the PP anlage, CD11c+ cells distributed more densely in the outer rim of PP anlagen at 17.5 dpc. At 18.5 dpc, when segregation of IL-7Rα+ cells starts, they tend to fuse to the developing follicle-like structures where IL-7Rα+ cells are concentrated. In the follicle-like structures of 2-dpp mouse, IL-7Rα+ cells localize more preferentially in the central region of the follicle-like structure, whereas CD11c+ cells display a tendency to localize at the fringe. This distribution pattern of IL-7Rα+ and CD11c+ cells suggests that each follicle-like structure can be divided to the core and fringe region in the next step, although the two regions are not sharply demarcated. Nevertheless, the behaviors of IL-7Rα+ and CD11c+ cells imply that the complexity of PP architecture is induced in a stepwise manner by creating distinct, novel compartments within previously formed compartment. In this study, we analyzed expression of CXCR5 and CCR7 in these two hemopoietic populations, as they were shown to be essential to establishing an organized PP architecture (19, 23). Interestingly, IL-7Rα+ cells express both CXCR5 and CCR7, while CD11c+ cells express only CCR7, with these expression patterns remaining unchanged during the initial phase of compartmentalization from 18.5 dpc to 2 dpp. This difference in the chemokine receptor expression may correlate with the distinct distribution patterns of the two populations. In fact, about a half of cxcr5 −/− mice were reported to lack PP (20), implicating its role in the induction phase of PP organogenesis. However, induction of PP anlage and subsequent compartmentalization of IL-7Rα+ and CD11c+ cells are normal in plt mutant mice which have a defect in SLC and ELC expression (H. Hashi et al., unpublished observation). Consistent with this, ccr7−/− mice have been shown to develop normal numbers of PP, though the ultimate PP architecture is disturbed (24). Thus, the significance of the differential expression pattern of these chemokine receptors remains to be elucidated in future studies. Nevertheless, comparison of mice with mutations of various chemokine and chemokine receptor genes in terms of the processes defined here will be important to fully understand the molecular mechanisms regulating the compartmentalization process. Moreover, further investigations of the expression of different chemokines in each newly emerging compartment is needed.
The mature PP architecture should be equipped with a specific vascular structure including HEV and identified by expression of MAdCAM-1. In the intestine, expression of MAdCAM-1 becomes detectable in venous ECs over the intestine at 13.5 dpc. Thus, MAdCAM-1 expression is not specific to HEV and is induced at restricted stages of development of venous ECs in a spatio-temporally regulated manner. This notion is consistent with the observation by Iizuka et al. that in rat embryo MAdCAM-1 is also expressed on nonmucosal tissues such as in blood vascular ECs in the fetal skin and neonatal thymus (35).
While PP organogenesis commences from 15.5 dpc, clear MAdCAM-1 expression in this region is first detected at 16.5 dpc. Judging from the staining pattern, possibly all newly developing venous ECs in the PP anlagen are MAdCAM-1+. However, unlike other regions, high expression is maintained in PP. This is likely to be due to expression of LTα1β2 in the IL-7Rα+ PP inducers, as LTα1β2 was shown to be a potent inducer of MAdCAM-1 (36) as are IL-1 and TNF-α (37). Along with compartmentalization of IL-7Rα+ cells, MAdCAM-1 expression became restricted to the ECs surrounding each follicle-like structure. It appears that MAdCAM-1+ vessels are pushed away to the rim of the follicle-like structure by increasing entry of hemopoietic cell lineages. Correspondingly, in the scid/scid mouse in which entry of mature lymphocytes is defective, the average diameter of each follicle-like structures surrounded by MAdCAM-1+ vessels appears to be smaller than that of the wild type.
Interestingly, even before the formation of discrete follicle-like structures with HEV, mature lymphocytes enter the PP anlage from 18.5 dpc. The first mature lymphocytes detected in PP are distributed rather evenly with a tendency to concentrate in the central part of PP, but B and T cells are not segregated yet. As soon as the follicle-like structures are formed, B220+ cells migrate to the these area, while most CD3+ cells are distributed within the space between these follicle-like structures. Noelle et al. has shown that the segregation of T and B cells is regulated by LTα produced by nonlymphoid cells (10). Thus, it is likely that LTα1β2 produced in IL-7Rα+ PP inducer cells also play a role in this process by forming a B cell target zone before lymphocyte entry.
It is still unclear how segregation is initiated in the apparently simple and homogeneous primitive PP anlagen. We speculate that CD11c+ cells play a role in this initial segregation. Unlike VCAM-1+ or IL-7Rα+ cells, which are distributed evenly over the PP anlagen, CD11c+ cells first accumulated in the periphery of the PP anlagen at 17.5 dpc. Thus entry of CD11c+ cells from the periphery of the PP anlage potentially adds a different influence by disturbing the interactions between IL-7Rα+ and VCAM-1+ cells.
In conclusion, we defined the initial process in formation of follicle-like structures in the PP. To our knowledge, this is the first description of how the rudiments of primary follicles are established. Thus, our present study, though remaining largely phenomenological, succeeded to specify the new process that should be examined in future investigation of PP organogenesis.
We gratefully thank Reina E. Mebius for the generous gift of mAb MECA-367.
This work was supported by Japanese Ministry of Education, Science, and Culture Grants 07CE2005, 07457085, 12219209, and 06277102 (to S.N.), a grant for Pediatric Research (9C-5) from the Ministry of Health and Welfare of Japan, and a Japanese Government Science and Technology Agency grant (to H.Y.).
Abbreviations used in this paper: PP, Peyer’s patch; dpc, days postcoitus; dpp, days postpartum; MAdCAM-1, mucosal addressin cell adhesion molecule; LT, lymphotoxin; PECAM-1, platelet endothelial cell adhesion molecule-1; HEV, high endothelial venule; EC, endothelial cell; BLC, B lymphocyte chemoattractant; SLC,secondary lymphoid-tissue chemokine; ELC, EBI1-ligand chemokine; DAB, diaminobenzidine; SA, streptavidin; MLN, mesenteric lymph node.