Secondary Lymphoid Organs: Responding to Genetic and Environmental Cues in Ontogeny and the Immune Response¹

The mammalian immune system, a cooperative endeavor between the innate and acquired arms, provides an optimal environment for defense against the invasion of pathogens at any site in the body. The sites of organized lymphoid cell accumulations are termed primary and secondary lymphoid organs (SLOs)³. Diverse populations of functionally mature but naive lymphocytes are generated in the absence of foreign Ags in the primary lymphoid organs (thymus, fetal liver, and bone marrow). These cells seed the SLOs to optimally respond to foreign invaders.

In this article we review the structure, function, development, and maintenance of SLOs where T and B cells encounter Ag to generate effector cells or tolerance. SLOs include the spleen, lymph nodes (LNs), Peyer’s patches (PPs), tonsils, adenoids, and, in the mouse, rat, and rabbit, the nasal associated lymphoid tissue (NALT). During antigenic challenge, additional lymphoid tissues are apparent in the lung (bronchus- associated lymphoid tissue or BALTs) and intestine (isolated lymphoid follicles (ILFs)). The blood vessels of SLOs allow Ag access. LNs, which are also served by lymphatic vessels, are effective in mounting responses to Ags that are present in tissues. These Ags originate from foreign invaders that are transported by APCs or are derived from self-Ags. The capacity to discriminate between dangerous foreign Ags and benign self-Ags relies on the APCs, their state of activation, and the recognition capacity of the naive T and B cells. Dendritic cells (DCs) constitutively sample self-Ags and migrate to draining LNs even in the steady state. Because most self-Ag-bearing DCs in LNs are immature, they do not effectively activate naive cells. They regulate self-reactive T cells by inducing anergy and clonal deletion and/or by expanding regulatory T cells (1).

Tertiary lymphoid organs (TLOs) or, more accurately, tertiary lymphoid tissues are accumulations of lymphoid cells that arise in the adult. These ectopic lymphoid accumulations respond to environmental stimuli with chronic inflammation during microbial infection, graft rejection, or autoimmune disease. They are classified as lymphoid tissues because they resemble SLOs with regard to cellular composition and compartmentalization, chemokines, vasculature, and function (2).

Individual SLOs are similarly organized, although their vasculature, mode of Ag entrance, local environment, and the stimuli to which they are subjected may differ. All SLOs include B and T cells, APCs, stromal cells, and a vascular supply. We summarize similarities and differences in these organs as they impact upon our discussion of their developmental signals.

LNs are located at vascular junctions and are served by lymphatic vessels that deliver Ag and APCs. A capsule derived from lymphatic vessels surrounds the highly compartmentalized LN (Fig. 1 A). The collected lymph and cell contents enter the LN via several afferent lymphatic vessels and filter through the node through the lymphatic sinuses in the medulla (3). Factors from afferent lymph can be either transported deep into the LN cortex or move via the subcapsular sinus and leave through efferent lymphatic vessels (4). Cells and Ag also enter the LN via an arteriole, which branches into a capillary bed. The postcapillary venules called high endothelial venules (HEVs) have a distinctive appearance with a plump cuboidal endothelium. The adhesion molecules that slow the flow of naive lymphocytes include peripheral node addressin (PNAd) and mucosal-associated adhesion molecule 1 (MAdCAM-1). MAdCAM-1, the ligand for the integrin α₄β₇, is only apparent in early development on HEVs of peripheral LNs (PLNs). PNAd, the ligand for L-selectin, replaces MAdCAM-1 after birth in mouse peripheral LNs (5). MAdCAM-1 remains along with PNAd in mucosal LNs (MLNs).

The lymphoid chemokines CCL19 (ELC) and CCL21 (SLC), expressed on the HEVs (6), induce changes in the affinity of LFA-1 on lymphocytes, allowing their interaction with ICAM-1 on the HEVs and then the transmigration of T cells to the paracortical region. Both T cells and DCs are attracted to the paracortical region by their expression of CCR7, the receptor for CCL19 and CCL21. CXCL13 (BLC), produced by stromal cells in the B cell follicles, attracts CXCR5-expressing B cells (7). The cortical region consists of primary follicles of densely packed naive B cells and follicular dendritic cells (FDCs). Ag-activated B cells proliferate and secondary follicles and germinal centers develop. Plasma cells are concentrated in the medulla as they prepare to leave the LN and circulate to the bone marrow. A network composed of reticular fibers, fibrous extracellular matrix bundles, and fibroblastic reticular cells supports the entire LN (4). After naive T and B cells encounter Ag, they undergo extensive changes in expression of chemokine receptors and adhesion molecules that result in their movement to different areas of the LN or in leaving it altogether. A conduit system physically connects the lymphatic sinus with the walls of blood vessels and enables incoming factor(s) from lymph to move into the paracortical area (4, 8, 9).

Mucosal-associated lymphoid tissues or MALTs, which quantitatively include the vast majority of lymphoid cells in the body, are major producers of secreted IgA and are responsible for inducing and maintaining tolerance to food Ags and commensal bacteria (10). Although the structures and locations of some MALTs are predetermined, all are somewhat plastic and prone to induction and remodeling due to their constant exposure to environmental Ags. The tonsils and adenoids in humans and NALTs in rodents are in fixed locations. So are PPs, but their number varies by species and Ag exposure. The BALTs and ILFs are located at predetermined sites in the lung and the small intestine but are even more plastic and subject to environmental influences. PPs, tonsils, adenoids, and NALTs exhibit the same basic organization as LNs with compartmentalized T and B areas, APCs, lymphoid chemokines, HEVs, and lymphatic vessels (Fig. 1 B). M cells, specialized APCs, also serve the NALTs, PPs, and BALTs. HEV vascular addressins differ between MALTs and LNs and between individual MALTs. PP HEVs express MAdCAM-1, but very low or undetectable PNAd. NALT HEVs express PNAd (11). BALTs are less organized but more environmentally regulated than PPs or NALTs. There is debate concerning whether BALTs should be considered as SLOs or as TLOs because they are not present in all species and are nearly entirely dependent on antigenic stimulation or age. Nevertheless, when they are present they exhibit organization similar to PPs, with T and B cell compartments and M cells. Their HEVs differ in that they express PNAd rather than MAdCAM-1 and high levels of VCAM-1 (12). Cryptopatches, composed of lin⁻c-kit⁺ cells, DCs, and VCAM-1⁺ stromal cells (13), with few or no mature T and B cells, give rise to ILFs (14, 15), which include B and T cells, resemble primitive LNs, and are found in the colon and small intestine.

The spleen is divided into two anatomically and functionally distinct areas: red pulp and white pulp. The red pulp, in its activities as a hematogenous organ, removes damaged cells and acts as a site for iron storage and turnover. The white pulp is an organized lymphoid structure. The spleen is highly vascularized but has no HEVs or afferent lymphatic vessels. Rather, the splenic artery, located immediately below the capsule, is the source of cells and Ags. The marginal sinus and marginal zone (MZ) demarcate the red and white pulp. The MZ, which surrounds the white pulp, represents a important transition between the innate and acquired immune systems; it is the first region after the red pulp encountered by blood-borne Ags and is richly supplied with specialized phagocytic cells, MZ macrophages and MZ metallophilic macrophages. The MZ also contains a specialized subset of B cells that differ phenotypically and functionally from follicular B cells and are considered a bridge between the innate and adaptive immune systems. The organization of the remainder of the white pulp of the spleen is similar to that of the LN, compartmentalized into B and T cell areas with the same lymphoid chemokines that are found in LNs and PPs. The white pulp consists of a central arteriole surrounded by T cells (the periarteriolar lymphoid sheath) which are surrounded by B cells. As in the LN, T cells interact with DCs, B cells migrate to the follicles where they interact with FDCs, and T cells interact with B cells at the border of the T and B cell areas, giving rise to germinal centers.

Insight into lymphoid organ development has come from analyses of mice that are deficient in particular genes, transgenic for lymphoid chemokines or cytokines, or have been treated with cytokine inhibitors (2). The action of multiple transcription factors, cells, cytokines, chemokines, and signaling molecules is critical for the development of a fully functional lymphoid system. The results of deletion of individual cytokines, chemokines, and transcription factors have been catalogued in several recent reviews (16, 17, 18); the reader is referred to tables in these publications. In this article we emphasize the importance of cooperating chemokines and cytokines in individual SLOs. Lymphotoxin (LT) (LTα, TNF-β) was initially described (19, 20) as a cytotoxic factor made by activated “lymphocytes” (T cells) that correlated positively with delayed type hypersensitivity (21). LT is also crucial for lymphoid organ development. Secreted LTα₃, signaling through TNFRI p55, and membrane-bound LTα₁β₂, signaling through the LTβ receptor (LTβR) and TRAF (TNF receptor-associated factor) 6, are absolutely crucial for SLO development. Lta^−/− mice lack all LNs and PPs and exhibit a severely disorganized NALT (11, 22, 23, 24). Their splenic defects include a disorganized white pulp with loss of T and B cell compartmentalization and loss of MZ macrophages, metallophilic macrophages, MZ B cells, MAdCAM-1 sinus lining cells, and germinal centers. Ltb^−/− mice lack PLNs but retain mesenteric, sacral, and cervical LNs (25, 26, 27). The splenic disorganization of Ltb^−/− mice is somewhat less pronounced than that of Lta^−/− mice. The phenotype of Ltbr^−/− mice is similar to that of Lta^−/− mice (28). LIGHT (lymphotoxin-related inducible ligand) is also recognized by the LTβR, but no defect in lymphoid organ development is observed in Light^−/− mice (29). However, mice doubly deficient in LIGHT and LTβ have fewer mesenteric LNs than those deficient in LTβ alone, suggesting an additive effect of the two LTβR ligands. Treatment of pregnant mice with LTβR-Ig inhibits most LNs in the developing embryos, depending on the time of administration. However, mesenteric LNs are not inhibited by this treatment. These studies indicate that individual LNs differ in their kinetics and cytokine requirements during ontogeny (30). TNF-α participates in lymphoid organ development, although its role is not as crucial as that of LT. Tnfa^−/− mice lack those splenic populations that are also absent in Lta^−/− mice (31); some reports indicate that they also lack PPs (32). Tnfr1^−/− mice show splenic but not LN defects similar to those of Lta^−/− mice with the exception of MZ B cells, and they do not give rise to mature ILFs (14). Cooperation between the various cytokines is evident from studies of mice doubly deficient in LTβR and TNFRp55. These mice lack MLNs and exhibit more severe splenic defects than do Ltb^−/− mice (33), suggesting that LTα₃ signaling through TNFRp55 is necessary for MLNs. Similarly, evidence for cooperation between LTαβ and TNF in splenic organization is apparent from studies of mice doubly deficient in LTβ and TNF (33).

Receptor activator for NF-κB (RANK) ligand (RANKL), another member of the larger TNF family, also participates in LN development. Mice deficient in RANK (also called TRANCER or OPG) (34) or RANKL (also called TRANCE or OPGL) (35, 36) exhibit defects in LNs but not in PPs. Mice deficient in the common γ-chain of several cytokine receptors, including IL-7, lack LNs (37), and Il7ra^−/− mice lack PPs (38) and some LNs (39). Thus, the statement that IL-7, signaling through the IL-7R and the JAK3 kinase, is required for PP but not LN development (40) is incorrect, because Jak3^−/− mice lack both PLNs (41) and PPs (38). LNs of Il7r^−/− mice are poorly populated (42), indicating that IL-7 is crucial for their maintenance. Furthermore, transgenic overexpression of IL-7 directed by a class II Eα promoter resulted not only in increased numbers of PPs but also in the generation of additional LNs (43).

CXCR5- or CXCL13-deficient mice (44, 45) lack some LNs and almost all PPs (44, 45). Mice doubly deficient in CXCL13 and IL-7Rα lack all LNs, including MLNs, indicating cooperation between these factors (39). plt/plt mice lack CCL19 and CCL21 but retain LNs and PPs, although they do exhibit defects in T cell homing and NALT maturation (46).

Transcription factors crucial for lymphoid organogenesis include the helix-loop-helix transcription factor inhibitor Id2 (47) and the retinoid acid-related orphan receptors (RORs) RORγ and RORγt (48, 49). Mice that lack RORγt lack LNs and PPs (48), although their NALTs are normal in size and cellularity (24). Factors that uniquely affect splenic development include homeobox genes and transcription factors that are expressed in the splenopancreatic mesenchyme at embryonic day (E)10.5; for example, Hox11 mice are asplenic (summarized in Ref. 50). Both the canonical and alternative NF-κB signaling pathways contribute to SLO development (51). The alternative pathway, characterized by the NF-κB-inducing kinase NIK and IKKα, is particularly important. aly/aly mice, which have a point mutation in Nik, lack all LNs and PPs (52, 53). Mice with a mutated form of the Ikka gene have defective HEVs, further confirming that the LTβR signal regulates HEVs and lymphoid chemokines through the alternative NF-κB pathway (54).

Lymphatic vessels are crucial components of the immune system and contribute to SLO development. The generation of embryonic lymphatic vessels from preexisting veins in pig embryos was first described in the early 1900s and has recently been molecularly defined (summarized in Ref. 55). At mouse E9, Sox18, a homeobox gene product, is expressed in several cell types including a subpopulation of endothelial cells in the cardinal vein (56). At E9.75 it directly activates Prox1, which is crucial for maintenance of the lymphatic phenotype (57). LYVE-1 is also expressed at this time in those lymphatic-biased polarized endothelial cells. Prox1 induces expression of a variety of genes, including integrin α₉ and VEGFR-3, allowing migration toward VEGF-C (58). LYVE-1⁺Prox1⁺VEGFR-3⁺ endothelial cells are committed toward a lymphatic pathway. Their further separation from venous endothelium requires a Syk/Slp-76 signal (summarized in Ref. 55).

Histologic studies in the rat revealed that popliteal and inguinal LN anlages originally appear in a limited mesenchymal area along the vein wall at E17. The next day, lymphatic vessels form a sac running parallel to the vein. The LN anlage develops into a bulb-shaped structure with lymphatic vessels and the subcapsular sinus originating from the remaining lymphatic vessel. At the next stage the LN divides into a primitive cortex, the basic network of reticular cells, and medulla, and lymphocytes scatter in the LN anlage. Blood vessels branch into the LN and later develop into HEVs (59). The primary follicles appear at day 18 after birth, indicating that B cell migration into LNs is a late event during lymphoid organogenesis Studies in the mouse confirm and extend these observations (summarized in Ref. 18).

Nishikawa and colleagues, based on their studies of PP development, articulated the concept that lymphoid organ development involves interactions between distinct cell types in several steps (60). The first step is the development of VCAM-1⁺ICAM-1⁺MAdCAM-1⁺“organizing clusters,” the second is the accumulation of cells expressing IL-7R and CD4, and the third is lymphocytes expressing CD3 or B220. The current model, which supports this concept, postulates interactions between initiator cells, inducer cells, and organizer cells. The various abbreviations for these cells have included LTin (lymphoid tissue initiator), LTi (lymphoid tissue inducer), and LTo (lymphoid tissue organizer) (18). Somewhat confusingly, LT in that terminology is an abbreviation for lymphoid tissue, not lymphotoxin. These cells have also been called PPin (Peyer’s patches initiator), PPi (PP inducer), and PPo (PP organizer) or LNi (lymph node inducer) and LNo (lymph node organizer) (42). To reduce confusion between lymphoid tissue and lymphotoxin and to emphasize the common mechanisms used in all SLOs, we call these cells ltini (lymphoid tissue initiator), ltind (lymphoid tissue inducer), and lto (lymphoid tissue organizer) cells. Although the relative importance of particular factors produced by these cells differs in individual lymphoid organs, the general scheme is similar. It has become clear that the development of a fully functional immune system involves cooperation not only between multiple cell types but also between multiple cytokines, chemokines, and their receptors.

The ltini cells have not been as extensively characterized as the ltind and lto cells, although at least in PPs the ltini cells are CD11c positive (60). Such CD45⁺CD4⁻CD3⁻IL-7R⁻CDllc⁺ cells in the mouse embryonic gut and adult spleen express RET, a tyrosine kinase receptor previously described as a neuroregulator. Because mice deficient in RET lack PPs (61), mesenchymal-derived PP lto cells produce artemin, a RET ligand (61), suggesting an interaction between ltini cells and lto cells in PP development. The role of RET in other SLOs has not been investigated.

Lto cells attract, activate, and respond to CD4⁺CD3⁻CD45⁺ ltind cells, which are derived from fetal liver progenitors (17, 49, 62, 63). These ltind cells are dependent on id2 (47) and RORγt (48), express the integrin α_4β7 (63), interact with VCAM-1 on resident lto cells, and accumulate in the developing LN or PP, forming clusters with lto cells to initiate a cascade of intracellular and intercellular events that lead to the maturation of the primordial SLOs. Although initiation of the LN anlages is LT independent, maintenance of lto cells is dependent on LTα (64). During this early step, depending on the organ, a positive feedback loop involves several signaling pathways between IL-7, CXCL13, LTβR, and TNFR expressed by lto cells and IL-7Rα, CXCR5, LTαβ and LTα, and TNF-α by ltind cells. At least two lto cell populations have been described. The proportion of cells that express high levels of VCAM, ICAM-1, and MAdCAM-1, compared with intermediate levels of those adhesion molecules, varies between neonatal MLNs and PLNs (65), providing further evidence of the somewhat subtle but nonetheless real differences in the regulation of these different LNs. Heterogeneity is also apparent in the expression levels of a variety of genes of the presumptive mesenteric LN lto cells compared with those from developing PPs (66). For example, although lto cells in both sites express RANKL, the levels differ, suggesting that the environment influences the relative importance of particular cytokines in individual organs in the course of development. Cryptopatch cells have some characteristics of other ltind cells in that they express RORγt and IL-7R, but they also express CCR6, and ILF development is dependent on IL-7, CCR6, and its receptor, CCL20 (67).

Although RANK and RANKL are required for LN development, there has been some confusion regarding the identity of ligand producers and responders. RANKL appears to be produced by both lti and lto cells. RANKL expression has been noted in lto stromal cells (40, 66, 68, 69), ltind cells (35, 70, 71), and both cell types (72). At least one publication describes the production of both RANK and RANKL by ltind cells (35). It is likely that the activity of this ligand-receptor pair is both autocrine and paracrine. Rossi et al. showed that RANKL production by thymic CD4⁺CD3⁻ cells affects stromal cells (71), whereas Yoshida et al. indicate that RANKL stimulates ltind cells to produce a factor that binds LTβR (presumably LTαβ), suggesting that ltind cells express the receptor (RANK) (40). Thus, the effect may be dependent on the microenvironment and developmental window of individual SLOs. As noted above, the relative importance of IL-7 for individual SLOs varies. IL-7/IL-7R signaling results in an increased number of CD4⁺CD3⁻ cells in the PPs, indicating that this cytokine is necessary for the maintenance of ltind cells in that organ. However, other data suggest that the ltind cells for PP and LNs are essentially similar; IL-7 administration to Traf-6^−/− LN-deficient embryos restores those organs (40). Initiation of NALT organogenesis occurs in the absence of IL-7R, LTβR, LTα, and the NF-κB-inducing kinase NIK (24, 73) but is dependent on CD3⁻CD4⁺CD45⁺ cells (73). However, LTα₃ and LTα₁β₂ are both crucial for postnatal NALT development with regard to the expression of lymphoid chemokines, HEV development, and response to Ag (11). The relative importance of the various cytokines in individual SLOs is illustrated in Fig. 2.

The prolonged interaction between ltind and lto cells promotes the development of HEVs. PNAd replaces MAdCAM-1 in the first few days in PLNs (5). LTα alone can induce MAdCAM-1, but PNAd requires LTαβ. PNAd expression is impaired in the remaining MLNs of Ltb^−/− mice, indicating that optimal LN HEV PNAd expression requires LTα₁β₂ signaling through the LTβR and the alternative NF-κB pathway (54, 74, 75). Because the maturation of HEVs is coincident with further development of the LN (5), the homing of LTαβ-expressing lymphocytes most likely also contributes to HEV maturation. NALT HEVs develop after birth and are regulated in a fashion similar to that of LNs, with LTαβ playing a crucial role in PNAd and the sulfotransferase that contributes to that functional addressin (11). Although an extensive literature exists on the development of lymphatic vessels from cardinal veins (55), less is known regarding their regulation in ontogeny in LNs. The roles of ltind and lto cells in this regard have not been extensively analyzed, although a recent publication indicates that even the defective LN anlages in Lta^−/− mice have LYVE-1⁺ vessels but no capsule (72).

Th17 cells that express IL-17A and IL-17F, IL-22, and IL-21 are implicated in several models of autoimmunity (summarized in Ref. 76). These cells have not previously been implicated in lymphoid organogenesis. However, the recent revelations that ltind cells and Th17 cells share cytokines and transcription factors buttress the concept that SLOs and TLOs represent the end products of similar mechanisms. RORγt, which as noted above is required for the development of most SLOs, is also crucial for differentiation of Th17 cells (77). Recent reports indicating that cells with the phenotypic characteristics of ltind cells in adult spleen (78) and human fetal lymphoid tissue (79) produce IL-17 further blur the difference between these cells and raise the intriguing possibility that this cytokine could contribute to SLO development. Although no abnormalities were reported in the cell populations of spleens, LNs, and thymi of Il17a^−/− (80) or Il17a^−/−Il17f^−/− mice (81), these organs were not analyzed histologically, leaving open the possibility that IL-17 could contribute to the microarchitecture or maintenance of SLOs.

SLOs located in anatomically distinct sites were formerly considered as static structures. Our thinking regarding this issue has evolved to an appreciation of the dynamism of lymphoid organs and the realization that SLOs are quite plastic and are influenced by their environments. HEV maturation, evident as a change from MAdCAM-1^high HEC6ST^low to PNAd^high HEC6ST^high (5), is coincident with the entrance of lymphocytes after birth, indicating that the mature HEV phenotype relies on the LN microenvironment (5, 59). Furthermore, interruption of lymphatic vessel flow drastically affects HEVs, reverting them to an immature phenotype (summarized in Ref. 82). Stromal cells with characteristics of lto cells have been described in mature LNs (68). CD45⁺CD4⁺CD3⁻ ltind cells are rare in adult SLOs (83), although they have been noted in adult spleens where they may contribute to the generation of memory (84). Furthermore, LT-producing cells, in addition to ltind cells such as T, B, and NK cells, presumably serve to maintain adult LNs. LTβR-Ig treatment of adult mice results in altered LNs with changes in cellular composition, HEV reversion to an immature state, inhibited FDC function, and disrupted immune responses to foreign Ags (82, 85, 86). Thus, in addition to its critical role in lymphoid organogenesis, LTβR is essential for maintaining LN function. The similarities between cells that induce inflammation (Th17) and those that induce lymphoid organs (ltind) add additional support to the notion that SLOs and TLOs share developmental programs and functions.

The most plastic and environmentally regulated lymphoid tissues are in the adult gut. The number and cellularity of PPs increase after immunization and decrease with aging (2, 87). After mice are exposed to microbes or during some forms of autoimmunity, cryptopatches give rise to ILFs (14, 49) and may even recapitulate the ontogenic program in response to these stimuli (88).

The spleen responds to its environment with changes in cellular compartmentalization. CMV infection results in disrupted white pulp compartmentalization of T and B cells in wild-type mice (89). The low level of CCL21 expression in Lta^−/− spleens (90) is reduced even further during infection with that virus, indicating that in the adult, LT-independent pathways can contribute to the maintenance of expression of lymphoid chemokines. Nevertheless, LT-producing CD4⁺CD3⁻ ltind cells are required to maintain T and B cell compartmentalization in the adult spleen (70). During infection with lymphocytic choriomeningitis virus, stromal cells and cellular organization are destroyed. Restoration is dependent on the proliferation of CD4⁺CD3⁻ cells and LTβR signaling (91), suggesting that the embryonic program can be reactivated under stress conditions.

TLOs are the most plastic and adaptable of the lymphoid tissues; lymphoid neogenesis can be induced by a variety of stimuli or inhibited by LTβR-Ig (summarized in Ref. 2). Their nimbleness in this regard suggests that they might represent the most primitive tissues in the immune system. Tertiary lymphoid tissues can be “turned off” (i.e., resolved) upon removal of the initial stimulus or after therapeutic intervention. For example, when β cells in the islets of Langerhans in type I diabetes mellitus are destroyed, the loss of antigenic stimulus is accompanied by TLO resolution. Similarly, LTβR-Ig treatment resolves established TLOs. These results recall those described above regarding the effects on established LNs of LTβR-Ig treatment (82, 85), again emphasizing the commonality of SLOs and TLOs.

In this communication, we have emphasized the similarities in structure and ontogeny of the various SLOs. Although a logical picture has emerged regarding the interaction of cytokines, chemokines, and the cells that produce them, much remains to be determined. What is the nature and origin of ltini cells? How are they activated? What governs the precise anatomical location of SLOs? Do homeobox genes contribute to LNs and PPs? Do TLOs recapitulate ontogeny? The plasticity of SLOs and TLOs, the persistence of ltind and lto cells, and the ability of other cells to assume their functions suggest that it may be possible to engineer lymphoid organs in individuals whose LNs are destroyed or nonfunctional due to surgical, genetic, or acquired immunodeficiencies.

The authors have no financial conflict of interest.