Both lymphotoxin-α (LTα)-deficient mice and alymphoplasia (aly) mice, a natural mutant strain, manifest a quite similar phenotype: lack of lymph nodes (LN) and Peyer’s patches (PP), with disturbed spleen architecture. The mechanisms underlying the defective lymphoid organogenesis in these mice were investigated by generating aggregation chimeras; ex vivo fused morulae were implanted into pseudo-pregnant host females and allowed to develop to term. Chimeric mice between LTα-deficient mice and wild-type mice restored LN and PP almost completely, suggesting that LTα expressed by circulating bone marrow-derived cells is essential for lymphoid organogenesis as well as for organization of spleen architecture. By contrast, chimeric mice between aly mice and wild-type mice showed only limited restoration of LN and PP. This suggests that the putative aly gene product does not act as a circulating ligand for lymphoid organogenesis, like LTα. Rather, abnormal development of lymphoid organs in aly mice seems most likely due to the defective development of the incipient stromal cells of the LN and PP. Supporting this hypothesis, up-regulation of VCAM-1 on aly mouse embryonic fibroblasts by signals through LTβR, which is exclusively expressed by nonlymphoid cells, was disturbed. These studies demonstrate that LTα and the putative aly gene product together control lymphoid organogenesis with a close mechanistic relationship in their biochemical pathways through governing the distinct cellular compartments, the former acting as a circulating ligand and the latter as a LTβR-signaling molecule expressed by the stroma of the lymphoid organs.
Lymphocytes continuously migrate from the blood to secondary lymphoid organs, where the systemic immune response against foreign Ags is initiated. Mechanisms for lymphocyte traffic have been well documented; homing to LN3 and Peyer’s patches (PP) occurs through the interaction between adhesion molecules expressed on high endothelial venules (HEV) of the lymphoid tissues and their counter-receptors on lymphocytes (1, 2). In contrast to the accumulated knowledge of the molecular basis for the lymphocyte traffic to the developed lymphoid organs, molecular basis for the ontogenic development of lymphoid organs themselves has been poorly understood. Morphological studies have demonstrated that LN in newborn mice are made up of richly cellular reticular tissue, undifferentiated postcapillary HEV, and nerve bundles. Lymphocytes begin to populate LN from the second day after birth, and lymphocyte diapedesis through the HEV becomes readily apparent from day 4 (3). Although recent studies have demonstrated that developing LN express mucosal addressin cell adhesion molecule-1 (MAdCAM-1) on HEV until 24 h after birth, and that a unique cell population of CD4+CD3−LTβ+ cells enters the LN using α4β7, a ligand for MAdCAM-1 (4, 5), mechanisms about how the incipient lymphoid organogenesis takes place still remain enigmatic.
Recent studies with gene-targeted mice manifesting abnormal development of the lymphoid organs have provided a new insight into these undefined processes. Lymphotoxin-α (LTα), originally discovered as a proinflammatory molecule (6), has turned out to be an essential factor that controls the genesis of the secondary lymphoid organs as well as for the organized spleen architecture (7, 8, 9). Lymphotoxin exists as two distinct forms: LTα3 as a soluble form and LTα/β heteromer (mainly as LTα1LTβ2) as a membrane-associated form, and each engages distinct receptor(s), the former with TNFR-I, TNFR-II, and herpesvirus entry mediator, and the latter with LTβR, respectively (10, 11, 12, 13, 14). Mice deficient for LTα were born with systemic absence of LN and PP (15, 16). Subsequent analyses of mice deficient for other LTα-related molecules have provided more detailed information on the action of LT for the lymphoid organogenesis. Although mice deficient for LTβ also lacked PP and most LN, they did possess mesenteric LN (17, 18). These results suggested that LTα1LTβ2 plays major roles of LT in the development of LN and PP, and that there also exists an LTα/β heteromer-independent pathway required for the development of mesenteric LN. This hypothesis was later proven by the in vivo administration of LTβR-Ig and TNFR-I-Ig fusion protein, which blocks the signals through LTβR and TNFR-I, respectively. Normal mice treated in utero with LTβR-Ig alone lacked systemic LN except for mesenteric LN, whereas concomitant administration of LTβR-Ig and TNFR-I-Ig or anti-TNF blocked the development of all LN, including mesenteric LN, indicating that TNF-TNFR-I axis also contributes to LN genesis (19). Consistent with this model, mice deficient for both LTβ and TNFR-I lacked all LN, including mesenteric LN (20). Conflicting data, however, also exist. It was demonstrated that mice deficient for LTβR lack all LN, including mesenteric LN, suggesting that LTβR is a primary receptor responsible for the development of all LN (21). Phenotypic difference between LTβ-deficient mice and LTβR-deficient mice also suggested that LTβR binds not only to LTα/β heteromer, but also to other ligand(s), such as LIGHT (14), which may mediate signals required for the development of mesenteric LN. Thus, the integration of the detailed phenotypic analyses of these gene-targeted mice with current perspectives of TNF/LT biology may illuminate many aspects of the lymphoid organogenesis.
Studies with knockout mice of TNF/LT-related molecules have also unraveled essential actions of LT and TNF for the organization of lymphoid structure. In mice deficient for either LTα (22, 23), LTβ (17, 18), or TNF (24, 25), organized clusters of follicular dendritic cell (FDC) and germinal centers (GC) are absent from the spleen. In LTα−/− mice, formation of FDC clusters and GC was restored by transplantation of normal bone marrow (BM), indicating that the LTα-expressing cells required to establish these lymphoid structures are derived from BM (22). Subsequent analyses have pointed out B cells as essential source of LT required for this action (26, 27). In contrast to LTα−/− mice, when TNFR-I-deficient mice (28, 29) or LTβR-deficient mice (30) were reconstituted with wild-type BM cells, they showed no detectable FDC clusters or GC formation, suggesting that both TNFR-I and LTβR expression on non-BM-derived cells are necessary for the establishment of these structures. Thus, both BM-derived and non-BM-derived cells governed by the TNF/TNFR family members contribute to the organization of lymphoid structure.
In addition to these studies with gene-targeted mice, alymphoplasia (aly) mice, an autosomal recessive natural mutant strain, have provided a novel and unique model for the abnormal development of lymphoid organs (31). Like LTα−/− mice, aly mice lack all LN and PP, and spleen architecture such as development of GC and FDC clusters as well as marginal zone formation is disturbed (31, 32, 33). aly mice manifest additional immunodeficiencies, including disorganized thymic architecture, low serum Ig level, and impaired allogenic skin rejection, which are not observed in LTα−/− mice (15, 16, 31). Although a gene responsible for this mutant strain has been mapped to chromosome 11 by linkage analysis (34), little is known about how the putative aly gene product contributes to the lymphoid organogenesis. Because LTα−/− mice and aly mice manifest a quite similar phenotype (i.e., lack of systemic LN and PP, and disorganized spleen architecture), comparative studies on the mechanisms underlying the abnormal lymphoid organ development in these mice may provide clearer vision for the lymphoid organogenesis.
Although dominant roles of LTβR as well as supportive roles of TNFR-I in the development of secondary lymphoid organs have been demonstrated as described above, exact mechanisms about how LT control the development of lymphoid organs through these receptors are still largely unknown. Furthermore, we have no clues for the identity of the putative aly gene product that, like LT, is indispensable for lymphoid organ development. Our studies were undertaken to clarify the mechanisms underlying the defective lymphoid organogenesis in LTα−/− mice and aly mice. For this purpose, we have generated aggregation chimeras; ex vivo fused morulae were implanted into pseudo-pregnant host females and allowed to develop to term. Chimeric analyses demonstrate that LTα and the putative aly gene product control lymphoid organogenesis by governing distinct cellular compartments. LTα, expressed by the BM-derived cells, is essential not only for the organization of spleen architecture, but also for lymphoid organogenesis. By contrast, the putative aly gene product from BM-derived cells, if expressed, has no major roles in the development of secondary lymphoid organs. Rather, lack of LN and PP in aly mice may be caused by a defect of non-BM-derived cells, possibly through the defective development of the incipient stromal cells of the LN and PP. Finally, we demonstrate that signaling through LTβR, but not through TNFR-I, is impaired in embryonic fibroblasts (EF) isolated from aly mice. Because LTβR is exclusively expressed by nonlymphoid cells (11, 13), defective LTβR signaling in aly mice is consistent with the idea that lack of lymphoid organogenesis in this strain is caused by the defect of non-BM-derived cells.
Materials and Methods
LTα−/− mice were kindly provided by Dr. Chaplin (15). Transgenic mice expressing green fluorescence protein (GFP) under the control of a chicken β-actin promoter and CMV enhancer (hereafter GFP-Tg) were generated as previously described (35). aly mice (31) and C57BL/6J mice were purchased from CLEA Japan (Osaka, Japan). LTα−/− mice have agouti coat color, whereas aly mice and GFP-Tg are black coat color. Mice were maintained and bred under pathogen-free conditions. All animals used were handled in accordance with the Guide for Animal Experimentation of Ehime University, School of Medicine (Ehime, Japan), and experiments were initiated with 8- to 12-wk-old mice.
Generation of chimeric mice
To generate chimeras, the experimental morulae either from LTα−/− mouse matings or aly mouse matings were fused ex vivo with morulae from GFP-Tg matings, implanted together into pseudo-pregnant host females, and allowed to develop to term, as previously described (36). Chimeric mice between LTα−/− mice and aly mice were generated in the same way.
Suspensions of spleen and thymus were prepared by teasing the tissues between two frosted microscope slides. Spleen cell suspensions were depleted of RBC by osmotic lysis. For the assessment of GFP-expressing cells, cells were then subjected to the analysis with a FACScaliber flow cytometer (Becton Dickinson, San Jose, CA) with CELLQuest soft ware, as previously described (37). For the assessment of membrane-bound LT expression, spleen cells were stimulated with immobilized anti-CD3-ε mAb (clone 145-2C11; PharMingen, San Diego, CA) and Con A (5 μg/ml) for 18 h. After washing twice with PBS, cells were incubated with LTβR human IgG1 fusion protein (LTβR-Ig) (38) and then with PE-conjugated anti-human IgG (Calbiochem, La Jolla, CA). Cells were analyzed with a FACScaliber flow cytometer.
Assessment of signaling through LTβR and TNFR-I using EF
EF were established by the standard procedure (39). EF from both aly mice and C57BL/6J wild-type mice were cultured in RPMI 1640 medium (Life Technologies, Grand Island, NY) supplemented with 10% heat-inactivated FBS (Life Technologies), 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin at a density of 2 × 105 cells/well in a 6-well culture plate. Eighteen hours after incubation with either control mAb Ha4/8 (1 μg/ml), agonistic anti-LTβR mAb AC.H6 (1 μg/ml) (19), or human rTNF (100 U/ml) (Genzyme, Cambridge, MA), EF were harvested from the plate using 10 mM EDTA, and subjected to the flow-cytometric analysis using FITC-conjugated anti-VCAM-1 mAb (PharMingen), as described above. Expression of VCAM-1 on EF was assessed by the EPICS ELITE (Coulter, Hialeah, FL), according to the manufacturer’s instruction.
Membrane-associated LT expression is retained in aly mice
Although there is some phenotypic difference between LTα−/− mice and aly mice, these two strains possess an extremely similar phenotype: lack of LN and PP, and disorganized spleen architecture. Genetic analysis has demonstrated that LTα itself is not a gene responsible for the aly mutation; with the use of linkage analysis, aly gene has been mapped to mouse chromosome 11 in which neither LT genes nor LT-receptor genes are located. It is still possible, however, that the homozygous aly gene mutation may perturb the expression of membrane-associated LT to cause the LTα-deficient phenotype in aly mice. To test this, we first examined the expression of LT mRNA from aly spleen cells with Northern blot analysis. Both LTα expression after Con A stimulation and constitutive LTβ expression from aly homozygous mice (aly/aly) were indistinguishable from those seen in aly heterozygous mice (aly/+) (data not shown). Membrane-associated LT expression in aly mice was more directly tested by flow cytometry using LTβR-Ig, which binds to membrane-associated LTα1LTβ2. There was no binding of LTβR-Ig to unstimulated aly/+ spleen cells. After anti-CD3-ε plus Con A stimulation, aly/+ spleen cells expressed membrane-associated LT detected by this fusion protein (Fig. 1,A). Control polyclonal human IgG did not show any binding to aly/+ spleen cells under these conditions (data not shown). As expected, spleen cells from LTα−/− mice showed no binding to this fusion protein even after stimulation (Fig. 1,B). Although LTβR-Ig may also bind to LIGHT on activated lymphocytes, another ligand for LTβR, this did not occur under this experimental setting. LTβR-Ig bound to aly/aly spleen cells after anti-CD3-ε plus Con A stimulation (Fig. 1 C), demonstrating that membrane-associated LT expression is retained in aly mice with this in vitro system.
Complementation of abnormal LN genesis in aly mice with LTα−/− mice
If membrane-associated LT is present and functional per se in aly mice, as suggested above, cells from aly/aly should reverse the phenotype of LTα−/− mice. This was tested by generating a chimeric mouse between LTα−/− mice and aly mice, in which cells from both strains coexist and interact with each other throughout development. LTα−/− mouse morulae were fused ex vivo with morulae obtained from aly/aly matings. One chimeric mouse was born, in which the contribution from both strains was verified by the coat color; both LTα−/− mouse-derived agouti hairs and aly/aly-derived black hairs were observed in this animal (data not shown). Upon detailed inspection, mesenteric LN (Fig. 2,A) and one lumbar LN (Fig. 2 B) were observed in this chimera, demonstrating that lack of LN in LTα−/− mice was restored by the cells from aly mice. Alternatively, the restoration of LN may represent the compensation for aly phenotype by the putative aly gene product-sufficient cells from LTα−/− mice. In either case, membrane-associated LT from aly mouse cells needs to be functional in order for this complementation to take place. Taken together, these results clearly demonstrate that lack of LN genesis in aly mice cannot be attributed to the lack of functional membrane-associated LT.
Spleen architecture demonstrated by immunohistochemistry manifested restoration of organized T cell/B cell segregation and GC formation, but not FDC cluster formation, in this chimera (data not shown).
Defective lymphoid organ development in LTα−/− mice is almost completely restored by the generation of chimeras with normal animals
We have demonstrated previously that lack of FDC clusters as well as the defective GC formation in the spleen from LTα−/− mice were restored by the transfer of wild-type BM cells, indicating that LTα-expressing cells required to establish these lymphoid structures are derived from BM (22, 28). In contrast to the spleen architecture, development of LN and PP was not restored in the same animals (37). Assuming that LTα-expressing cells required to generate LN and PP are also BM derived, we speculated that lack of LN and PP in LTα−/− mice is developmentally fixed. These hypotheses, however, have not been formally proven, because BM cells were transferred only into the adult mice in these experiments. We have approached this issue by generating aggregation chimeras between LTα−/− mouse morulae and LTα+/+ mouse morulae (summarized in Table I); both LTα-deficient BM-derived cells and LTα-sufficient BM-derived cells are expected to circulate in the body and to interact with the incipient stromal cells of the lymphoid organs throughout the development. Because we used GFP-Tg as LTα+/+ mice, chimeric contribution from each strain can be readily monitored by the detection of GFP. In particular, detection of GFP from the thymocytes and/or splenocytes by flow cytometry enabled us to focus on the chimerism of BM-derived cells, which are the major source of membrane-associated LT.
|Mouse .||Sexb .||Coat Colorc .||GFPd .||GFP-Positive Thymocytese .||Inguinal LN .||Mesenteric LN .||PPf .||Spleen Architecture .||.|
|.||.||.||.||.||.||.||.||GC .||FDC .|
|5||M||Black||−||13.1||R+: L−||+||2, small||+||+|
|Mouse .||Sexb .||Coat Colorc .||GFPd .||GFP-Positive Thymocytese .||Inguinal LN .||Mesenteric LN .||PPf .||Spleen Architecture .||.|
|.||.||.||.||.||.||.||.||GC .||FDC .|
|5||M||Black||−||13.1||R+: L−||+||2, small||+||+|
Note that two mice, LTα−/−/GFP-7 and aly/GFP-6, are apparently derived only from LTα−/− mice and aly mice, respectively (see text).
M, male; F, female.
Mixed coat color of agouti (derived from LTα−/− mice) and black (derived from GFP-Tg) is designated as mixed. aly mice have black coat color.
GFP expression from the body such as muscles, fat tissues and peritoneum is shown as + or −.
Percentages of GFP-positive thymocytes determined by the flow-cytometric analysis.
Mice with more than four PP are expressed as +, and mice lacking PP are expressed as −. In other cases, numbers of PP observed are shown.
One mouse (LTα−/−/GFP-7) of 11 chimeras generated showed no evidence of chimeric contribution from GFP-Tg; coat color was agouti from LTα−/− mice, and no GFP expression from the body as well as from the thymocytes and splenocytes was detected. Upon detailed inspection, no LN and PP were observed in this mouse. Histological analysis demonstrated disturbed T cell/B cell organization without GC and FDC formation in the spleen (data not shown), suggesting that this mouse was apparently derived only from LTα−/− mice. Except for this mouse, the other 10 mice showed contribution from both LTα−/− mice and GFP-Tg. Coat color and GFP expression from the body (i.e., muscles, fat tissues, and peritoneum) as well as from the thymocytes and/or splenocytes showed wide variety of contribution from the donors in each chimera. Upon detailed inspection, all chimeric mice showed lymphoid organ development indistinguishable from that seen in wild-type mice, as exemplified in Fig. 2,C, except for one chimera; LTα−/−/GFP-6 showed only one PP with normal LN genesis. Percentages of the LTα-expressing cells evaluated by the detection of GFP from thymocytes varied among the chimeras, ranging from 8 to 66%, and this variation did not apparently affect the extent of restoration of lymphoid organogenesis in this range. Development of LN and PP occurred irrespective of the coat color, status of GFP expression from the body as long as LTα-expressing cells exist in the spleen and thymus. As expected, spleen architecture was also restored in these mice. Organized T cell/B cell segregation and FDC cluster formation were present in the spleens from all chimeric mice (Fig. 3, C and G). GC formation was also restored in all chimeras (data not shown). These results demonstrate that LTα-expressing BM-derived cells, if present throughout the development, can restore the lymphoid organogenesis in LTα−/− mice, supporting the idea that LTα-dependent interactions must occur during development in order for LN and PP to develop.
Defective lymphoid organ development in aly mice is only partially restored by the generation of chimeras with normal animals
To characterize the putative aly gene product, chimeric mice between aly mice and GFP-Tg were generated and evaluated for the restoration of lymphoid organ development. Because both aly and GFP-Tg have black coat color, chimerism cannot be assessed by the coat color. Chimerism could be evaluated, however, by the detection of GFP in tissues as well as from the thymocytes and/or splenocytes. Of 12 mice generated, one mouse (aly/GFP-6) did not show any evidence for the contribution from GFP-Tg: complete absence of GFP expression in tissues and from thymocytes and splenocytes, no LN and PP, and disturbed T cell/B cell segregation without GC and FDC formation in the spleen.
Of the other 11 chimeras, two mice (aly/GFP-10 and aly/GFP-11) showed no inguinal LN (Fig. 2,D), and one mouse (aly/GFP-5) had only right inguinal LN. Although four mice had more than four PP, seven mice had either less than three (aly/GFP-5, aly/GFP-7, and aly/GFP-12) or no PP at all (aly/GFP-8, aly/GFP-9, aly/GFP-10, and aly/GFP-11). Furthermore, one mouse that lacked inguinal LN and PP (aly/GFP-11) had only one small mesenteric LN (Table I). Development of other LN such as mandibular, iliac, lumbar, popliteal, and axillary LN was indistinguishable from those seen in wild-type mice. It is particularly important to note that there are many spleen cells derived from GFP-Tg even in the chimeric animals that had defective development of LN and/or PP; mice that lacked both inguinal LN and PP (aly/GFP-10 and aly/GFP-11) showed many GFP-expressing cells in the spleens (Fig. 4, C and D). Because BM-derived cells from GFP-Tg should have the normal putative aly gene product, if expressed, this suggests that lack of LN and/or PP in these chimeras was caused independent of the presence of putative aly gene product on the BM-derived cells. This limited restoration of lymphoid organ development sharply contrasts to that seen in the chimeras between LTα−/− mice and GFP-Tg in which relatively small numbers of the LTα-expressing BM-derived cells were sufficient to restore the lymphoid organogenesis.
Despite the partial restoration of LN and PP development in these chimeras, histological evaluation showed spleen architecture indistinguishable from that seen in wild-type mice; T cell/B cell organization was apparently normal with GC and FDC formation in all 11 chimeras, including aly/GFP-10 and aly/GFP-11 (Fig. 3, D and H). These results demonstrate that organized spleen architecture can be formed independent of the defective development of LN and/or PP.
Possible involvement of LTβR signaling in the abnormal lymphoid organ development in aly mice
The similar phenotypes of the LTα-deficient and aly mouse strains suggested that there might be a close mechanistic relationship in their affected biochemical pathways. The data described above showed no detectable alteration in the expression of the membrane-associated LT ligand in aly mice. We next examined downstream elements of the pathway at the level of LTβR signaling using EF isolated either from aly mice or from C57BL/6J wild-type mice. In the wild-type EF, VCAM-1 expression was up-regulated upon incubation with human TNF, which binds and signals exclusively through mouse TNFR-I (Fig. 5,A). Either agonistic anti-LTβR mAb AC.H6 (Fig. 5,C) or rLTα1LTβ2 (data not shown) also induced up-regulation of VCAM-1 from wild-type EF, although to a lesser extent compared with TNF stimulation. By contrast, up-regulation of VCAM-1 after stimulation either with agonistic anti-LTβR mAb (Fig. 5,D) or with rLTα1LTβ2 (data not shown) was absent from aly EF, although human TNF induced up-regulation of VCAM-1 on aly EF (Fig. 5 B). These results suggest that signaling through LTβR, but not through TNFR-I, is impaired in aly mice.
In the present study, the mechanisms underlying the defective development of secondary lymphoid organs in LTα−/− mice and aly mice, a natural mutant strain, were investigated by a novel approach. The results demonstrate that the phenotype of aly mice is not due to the lack of functional membrane-bound LT, but that defective signaling through LTβR seems likely. Studies using chimeric mice suggest that LT required for the genesis of LN and PP is provided by BM-derived cells, as shown for the organization of spleen architecture (22). By contrast, BM-derived cells do not account for the abnormal development of secondary lymphoid organs in aly mice. Thus, LTα and the putative aly gene product together control lymphoid organogenesis by governing the distinct cellular compartments with a close mechanistic relationship in their biochemical pathways.
Identification of the aly gene and/or aly gene product promises to facilitate our understanding of the mechanisms for lymphoid organogenesis as well as for immunodeficiency. The only information on aly gene to date, however, is its chromosomal location (34). Because LTα−/− mice and aly mice manifest a quite similar phenotype, we first investigated whether aly mice have membrane-associated LT. The results indicated that activated splenocytes from aly mice expressed membrane-associated LT whose function was proven by the complementation experiment; mesenteric LN and lumbar LN were restored in a chimera between LTα−/− mice and aly mice.
To clarify the roles of LTα and the putative aly gene product in lymphoid organogenesis, we have tested how the phenotype of mice deficient for those molecules can be rescued by generating chimeric mice with normal animals. In each chimeric mouse, cellular contribution from the donor strain is expected to be random. In fact, percentages of GFP-expressing BM-derived cells, coat color, and GFP expression in tissues varied among the mice. In the case of chimeric mice between LTα−/− mice and GFP-Tg, this variation did not apparently affect the extent of restoration of lymphoid organogenesis. Gene dosage effect of LT, however, has been reported. Mice heterozygous for both ltα and ltβ (LTα+/−LTβ+/− mice) showed complete lack of PP, and some of those mice also lacked inguinal LN (20). Although there was one chimera (LTα−/−/GFP-6) that had only one PP, the percentage of GFP-expressing (i.e., LTα-expressing) thymocytes in this mouse was not low (66%). The reason for the poor PP genesis despite sufficient numbers of LTα-expressing cells in this particular animal remains unknown.
Although transgenic expression of LTα under the control of rat insulin promoter in LTα−/− mice restored some LN such as mesenteric LN and cervical LN, development of inguinal LN, popliteal LN, and PP did not take place (40). Furthermore, reconstituted LN did not contain FDC clusters, and T cell/B cell segregation in the spleen remained disturbed in these mice. In contrast to these transgenic mice, chimeric mice between LTα−/− mice and LTα-sufficient mice, GFP-Tg, restored not only the development of all LN and PP, but also spleen architecture completely. We speculate that the difference between transgenic studies and the chimeric analyses in the present study is due to the differential location of the cells expressing LTα, the former mainly at ectopic sites such as pancreatic β cells, kidney, and skin, and the latter at more physiological sites, respectively.
Restoration of lymphoid organogenesis in chimeric mice between aly mice and GFP-Tg took place only partially. This limited restoration of lymphoid organ development sharply contrasts to that seen in the chimeras between LTα−/− mice and GFP-Tg; there were many aly/GFP-Tg chimeric mice that lacked LN and/or PP with many GFP-expressing BM-derived cells. This suggests that the putative aly gene product from BM-derived cells, if expressed, has no major role in the genesis of secondary lymphoid organs. Rather, lack of LN and PP in aly mice may be caused by the defect of non-BM-derived cells, possibly through the defective development of the incipient stromal cells of the lymphoid organs. We speculate that stroma of such missing LN and PP in chimeric mice were destined to derive and develop from cells contributed by aly mice. In this scenario, BM-derived cells from aly mice have no major responsibility for the lack of LN and PP in this strain, which is suggested by the complementation experiment with LTα−/− mice, as described above.
It may be informative to determine the stromal origin of the restored LN in these chimeras. Our efforts, however, to detect the GFP from the LN and spleen stroma under the fluorescent microscope were not successful; although we could detect strong GFP signals not only from BM-derived cells but also from some nonlymphoid tissues such as muscles and fat tissues, GFP expression from the lymphoid stroma did not give us a clear signal sufficient to determine the origin of the donor. Introduction of a detection marker whose expression is more ubiquitous may be required to solve this issue.
It is now clear that all LN are not generated equally. Differential requirement for TNF/LT receptor family members has been demonstrated in detail for the development of mesenteric LN (15, 16, 17, 18). Mice deficient in BLR1, a chemokine receptor expressed mainly on mature B cells, have defective development of inguinal LN and PP (41). Furthermore, LN at different anatomic sites are generated at a particular developmental stage in an ordered fashion (42). It may merit attention that inguinal LN and PP, which develop at a later gestational stage, were susceptible to the failure of restoration in our aly/GFP-Tg chimeric mice. This is also true for LTα−/− mice, in which LTα gene was introduced under the control of rat insulin promoter (40), and for LTα−/− mice in which agonistic anti-LTβR mAb was injected in utero to restore LN genesis (19). The reason that inguinal LN and PP become preferential targets for the developmental loss in these mice remains unknown.
To date, a number of classes of genes have been implicated for the generation of secondary lymphoid organs. Cytokine-related genes other than LT have been recognized to be involved in the lymphoid organogenesis; mice deficient in either IL-2Rγ chain (43), IL-7Rα, or Jak3 (44) have defective PP development. Chemokine receptor BLR1 has also been demonstrated to be essential for the development of inguinal LN and PP, as well as for the formation of B cell follicles in the spleen (41). Recently, it was also demonstrated that mice deficient in osteoprotegerin ligand have defective LN genesis (45). In addition, targeted deletion of transcription factors has caused abnormal development of lymphoid organs and/or lymphoid cells; Hox11-deficient mice lack spleen (46), Id2-deficient mice lack LN and PP (47), and Ikaros-deficient mice lack cells of all lymphoid lineages, LN, and PP (48). Mice deficient for both NF-κB1 (p50) and NF-κB2 (p52) also lack LN with disorganized spleen architecture (49). Thus, there are broader spectrum of factors than had been anticipated that influence the lymphoid organogenesis. Then, what kind of gene might aly be? Up-regulation of VCAM-1 after stimulation with agonistic anti-LTβR mAb was absent from aly EF, suggesting that abnormal LTβR signaling may account for the abnormal lymphoid organ development in aly mice. Consistent with this result, injection of agonistic anti-LTβR mAb into pregnant aly female, which could induce lymphoid organogenesis in LTα−/− mice (19), did not restore the development of LN in the progeny (our unpublished data). Furthermore, these results well reconcile the fact that LTβR is exclusively expressed by nonlymphoid cells, with the results demonstrating that lack of LN and PP genesis in aly mice is caused by the defect of non-BM-derived cells. Because LTβR gene itself is on mouse chromosome 6 (38), not on chromosome 11, and that LTβR mRNA expression from EF was indistinguishable between aly/+ mice and aly/aly mice by RT-PCR (unpublished data), the putative aly gene product should work downstream of the LTβR. In light of the fact that aly mice manifest some additional immunodeficient phenotypes to that seen in mice deficient for LTβR (21), it is reasonable to speculate that the putative aly gene product may be involved in the signaling more than through LTβR. Molecular cloning of aly gene is required to solve this issue.
aly mice also manifest some additional phenotypes of immunodeficiency to that seen in LTα−/− mice. LTα−/− mice, however, manifest more profound defects than aly mice in some aspects; aly mouse spleens contain a relatively well-formed T cell area, whereas LTα−/− mouse spleens lack organized T cell area completely (Fig. 3, A and B ). Recently, Alexopoulou et al. have demonstrated that spleen architecture, but not LN and PP genesis, in LTα−/− mice can be greatly improved by the transgenic expression of TNF, suggesting that defective signaling through TNFR-I in addition to the loss of LTβR signaling can be attributed to the disturbed spleen architecture in LTα−/− mice (50). In aly mice, signaling through TNFR-I was not apparently affected, as demonstrated by the up-regulation of VCAM-1 on EF as well as by the normal NF-κB induction from thymocytes after stimulation with human TNF (unpublished data). Thus, preserved TNFR-I signaling in aly mice may explain the less disturbed organized T cell area in aly mice than that from LTα−/− mice, although the putative aly gene product should be involved in other receptor signaling beyond LTβR.
The histological abnormality of the spleen in aly mice was not fully corrected by the BM transfer from wild-type mice (31, and our unpublished data). Based on this result, involvement of the stromal element in the abnormal lymphoid structure in aly mice was postulated. It was also demonstrated that no histogenesis of the LN or PP took place in adult aly mice that had received either BALB/c BM cells (31) or aly/+ BM cells (unpublished data). It is not clear, however, whether these results indicate the involvement of the stromal element in the abnormal lymphoid organogenesis in aly mice; failure of LN and PP genesis by wild-type BM cells could indicate that defective lymphoid organogenesis in aly mice is developmentally fixed. In fact, by manipulating LT-LTβR axis of wild-type mice or LTα−/− mice in utero, it was demonstrated that LN genesis is largely completed by day 17 of gestation (19, 42). It is our intention that we have introduced chimeric analyses that overcome the developmental barrier. Using chimeric analyses, we have demonstrated that lack of LN and PP genesis in aly mice seems most likely due to the defective stromal development of the lymphoid organs. It is important, however, to emphasize that aly mice do have some defect in the BM-derived cells as well. Abnormal function of the BM-derived cells in aly mice is exemplified by the fact that serum IgM and IgG in aly mice increased to normal levels after BM transplantation from wild-type mice (31). We speculate that the putative aly gene product plays important roles both in non-BM-derived cells and in BM-derived cells; abnormal development of LN and PP in aly mice is due to the defect in non-BM-derived cells, as demonstrated in the present study, and some of the impaired immune function may be caused by the lack of the putative aly gene product in the BM-derived cells, as suggested by the BM transfer experiments. The role of the putative aly gene product in immune regulation is currently under investigation.
There are several types of knockout mice that manifest abnormal GC and/or FDC development without gross defective lymphoid organogenesis (51). Conversely, mice that exhibit abnormal lymphoid organ development with intact GC and FDC development also exist; mice heterozygous for both ltα and ltβ had the ability to develop GC and FDC in the spleen in the absence of PP and/or inguinal LN (20). LTα−/− mice expressing transgenic TNF lacked all LN and PP, but retained a suboptimal capacity to develop primary B cell follicles that contained FDC clusters in the spleen, and these structures could support the formation of GC (50). Furthermore, Id2-deficient mice retain normal spleen architecture without LN and PP (47). In this study, we demonstrated that some of the aly-derived chimeras lacked some LN and/or PP with intact GC and FDC development in the spleen. Taken together, organization of the lymphoid structure takes place independently of lymphoid organogenesis, and vice versa. Thus, signals required for the development of lymphoid organs may be distinct from those for the organization of lymphoid architecture.
Note added in proof.
A gene responsible for the aly mutation has recently been identified by positional cloning (52).
We thank David D. Chaplin for encouraging this work and critical reading of the manuscript. We thank Jeffrey L. Browning for providing us with LTβR-Ig, agonistic anti-LTβR mAb, and rLTα1LTβ2. We also thank Masayuki Miyasaka and Hiroyasu Nakano for stimulating the discussion.
This work was supported in part by Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Culture, and Sports, Japan, and by the Mochida Memorial Foundation for Medical and Pharmaceutical Research.
Abbreviations used in this paper: LN, lymph node; BM, bone marrow; EF, embryonic fibroblast; FDC, follicular dendritic cell; GC, germinal center; GFP, green fluorescence protein; HEV, high endothelial venule; LT, lymphotoxin; PP, Peyer’s patch; Tg, transgenic mice.