Role of Lymphotoxin in Experimental Models of Infectious Diseases: Potential Benefits and Risks of a Therapeutic Inhibition of the Lymphotoxin-β Receptor Pathway

Expression and regulation of ligands and receptors.

Lymphotoxin is a TNF family cytokine. The seminal discovery of impaired secondary lymphoid organ formation in LTα gene-deficient (−/−) mice (11) has shed new light on the biological functions of LT, which was long considered to be a redundant cytokine for TNF-α. Figure 1, top left, describes the LT/LIGHT ligands and receptors. Soluble LTα₃ is a secreted protein that interacts with the TNF receptors I (55 kDa) and II (75 kDa) (TNFR-I and -II) (reviewed in reference 68). LTα is coexpressed with the membrane protein LTβ as LTαβ heterodimers, which are tethered to the cell membrane. LTα₁β₂ binds to a TNF family receptor known as LTβR. LIGHT is a second ligand interacting with the LTβR. LIGHT also binds to the TNF family receptors herpesvirus entry mediator (HVEM) and decoy receptor 3. Activated lymphocytes and a subset of resting B cells express LT. The LTβR is expressed mainly on nonhematopoietic and myeloid lineage cells (reviewed in reference 22). The expression of LTαβ and LIGHT is induced by activation of lymphoid cells and certain cytokines and chemokines, including interleukin 4 (IL-4), IL-7, CXC chemokine ligand 13 (CXCL13), and CCL19/CCL21 (22). While regulation of LTβR expression remains to be defined, HVEM expression is induced during T-cell activation (22). Figure 1, top right, depicts the factors, chemokines, and cytokines involved in LTαβ regulation and regulated by LTβR activation. Expression of LT on lymphocytes provides signals necessary for stromal cells to secrete CXCL13. CXC chemokine receptor 5⁺ (CXCR5⁺) B cells are attracted to such stromal cells. CCL21 attracts T cells and dendritic cells, which together with B cells and stromal cells form lymphoid follicles with separated T- and B-cell zones, high endothelial venules, and follicular dendritic cell (FDC) networks.

LT ligand/LT receptor LIGHT gene-deficient and transgenic mice.

The roles of the LT/LIGHT ligands and receptors have been characterized in gene-deficient mice. Gestational and postgestational inhibition of the LTβR allows us to distinguish between the specific functions of the LTαβ/LIGHT-LTβR pathway at different developmental time points. Table 1 summarizes the defects in intestinal lymphoid organ development observed in mice with disrupted LTβR- and TNFR-mediated signaling.

Deletion of the ltα gene in LTα^−/− mice is associated with the loss of all lymph nodes (LNs) and Peyer's patches (PPs) and changes to the lymphoid architecture of the spleen (3, 11). Targeted disruption of the ltβ gene results in loss of peripheral LNs, whereas the mucosa draining cervical and mesenteric LNs (MLNs) are retained (29). LTβR^−/− mice have a phenotype very similar to that of LTα^−/− mice (19). LIGHT, the second ligand of the LTβR, is predominantly expressed on T cells and serves as a costimulatory molecule (reviewed in reference 66). In contrast to LTα^−/− mice, LIGHT^−/− mice develop intact lymphoid organs, indicating a predominant role for the LTαβ-LTβR interaction in the formation of secondary lymphoid organs (61). Conversely, transgenic expression of LIGHT in LTα^−/− mice restores splenic T- and B-cell segregation, FDCs, and germinal-center (GC) formation but not marginal-zone (MZ) formation, suggesting that LIGHT can compensate for the loss of certain LTαβ functions in the presence of the LTβR (66). CD8⁺ cells from LIGHT^−/− mice show decreased proliferative responses, while the IL-2 secretion is decreased in CD4⁺ cells from these mice (61), indicating a role for LIGHT as a costimulatory molecule. Conversely, transgenic expression of LIGHT on T cells is associated with a hyperactivated enlarged T-cell compartment and spontaneous autoimmunity, including inflammatory bowel disease (67).

Gestational and postgestational inhibition of LTβR-mediated signaling.

Treatment with soluble LTβR-immunoglobulin G (IgG) fusion protein (LTβRIgG) similarly blocks the interaction of LTαβ and LIGHT with the LTβR. Therefore, the effects observed in LTβRIgG-treated animals are a result of the inhibition of both ligands. Gestational treatment with LTβRIgG prevents the formation of PPs and LNs (56). Postgestational LTβR signaling during the first 6 weeks after birth is critical for the development of intestinal lamina propria B cells, IgA secretion, and isolated lymphoid follicles of the intestine (38, 48).

Mode of action of LTβR inhibition in adult mice.

The lymphoid microenvironment is defined as the local interplay between mobile lymphocytes and the fixed reticular/stromal cells and includes cell adhesion, trafficking, chemokine function, and cellular positioning (22). Secondary lymphoid organs are structures with a high degree of plasticity. Inhibition of LTβR signaling in adult mice alters the lymphoid microenvironment. As shown in Fig. 1, bottom, FDC networks and GCs disintegrate and B-cell follicles disappear in the absence of LTβR signaling. Figure 2 depicts the lymphoid microarchitecture of a murine wild-type (wt) spleen and describes the changes observed in mice with blocked LTβR signaling. Table 2 and Fig. 3 summarize the changes to the lymphoid microenvironment observed in mice with disrupted LTα₁β₂-LTβR signaling. FDC networks consist of a scaffold of specialized reticular fibroblasts that retain and present intact antigen to B cells. Memory B cells develop during the GC reaction. Permanent LTα₁β₂-LTβR interaction is required to maintain a CXCL13 chemokine gradient, which attracts CXCR5⁺ B cells to the follicle and is also required to maintain the differentiation status of the recruited B cells and the FDCs in the network. Similarly, LTβR engagement is required for continued expression of the vascular-cell adhesion molecule 1 (VCAM1) by the FDC network (23, 25).

The expression of VCAM1 is similarly reduced in FDC networks and splenic MZs following LTβRIgG treatment. Splenic MZs are located between the red and white pulp of the spleen and consist of a reticular matrix harboring B cells, macrophages, and dendritic cells (DCs). Blood-borne antigen is captured in the MZ and presented to B cells. Marginal zones are critical for immunity against T-cell-independent bacterial antigens (24). MZ markers disappear following LTβRIgG treatment.

Inhibition of the LTβR also blocks the migration or maturation of the cysteine-rich domain of the mannose receptor (CR-Fc)-positive DCs (42, 71) Figure 3 describes three potential mechanisms by which inhibition of the LTβR alters autoimmunity and host defense.

PPs, isolated lymphoid follicles, cryptopatches, and colonic patches are organized lymphoid aggregates of the intestine. The number and cellular contents of these aggregates are reduced in adult mice undergoing anti-LTβR treatment (13).

The potential mode of action in anti-LTβR treatment is to impair immune function by preventing proper placement of T cells, B cells, and APCs in secondary lymphoid organs, thus preventing the induction of appropriate antigen-specific immune responses. In contrast, selective inhibition of LIGHT signaling results in a loss of LIGHT-mediated costimulatory stimuli. Soluble HVEM-Fc fusion protein selectively blocks LIGHT-HVEM interactions, inhibits CD3-induced T-cell proliferation, and reduces the frequency of spontaneous diabetes in nonobese diabetic mice (67).

Ectopic “tertiary” lymphoid organs and inflammatory diseases.

A number of human inflammatory and autoimmune disorders are associated with the formation of ectopic lymphoid structures at the site of the inflamed organ which resemble secondary lymphoid organs (69; reviewed in reference 60). It is likely that immune responses to self antigens expand in these de novo lymphoid organs, as they allow colocalization of antigen-specific T and B cells with APCs. In mice, ectopic lymphoid structures can be induced by transgenic overexpression of LTα (31), and they are called tertiary lymphoid tissues. Data from animal models of autoimmune diseases associated with the formation of ectopic lymphoid tissue indicate that autoimmunity is less severe, cured, or prevented if the LTβR is blocked in these conditions (reviewed in reference 22). A potential role for microbial pathogens in the pathogenesis of inflammatory disorders, such as rheumatoid arthritis or inflammatory bowel disease, has been discussed (18). Granulomas are common histological hallmark of Crohn's disease and mycobacterial infections and serve as sites where T cells, APCs, and antigens collocate (44). Granulomas resemble tertiary lymphoid tissues. The recruitment of T cells to granulomas was impaired in LTα^−/− (donor) → (transfer) wt (recipient) bone marrow chimeras infected with Mycobacterium tuberculosis (58). It is therefore possible that anti-LTβR targeted therapy might also shut down common pathways of host defense and inflammatory responses that might lead to autoimmunity in genetically predisposed persons.

Lymphotoxin and LIGHT contribute to central immune tolerance in mice.

While a stimulatory role for LTβR signaling in the induction of peripheral autoimmune disease has been demonstrated by effective treatment of such conditions by anti-LTβR therapy, the thymic expression of LTβR and LTβR ligands contributes to central tolerance. Lymphotoxin signaling is required for the expression of Aire, which is a key mediator of central tolerance for peripherally restricted antigens. Similarly, LTβR ligand expression on thymic epithelial cells is required for proper differentiation of thymic medullary epithelial cells. LTα^−/− and LTβR^−/− mice show infiltration of liver, lung, pancreatic islands, and kidney with activated lymphocytes, and autoantibodies can be detected in these mice (7, 10). Thus, LTβR signaling plays different roles in the peripheral and central control of immune tolerance.

Bacterial infections.

Both types of animal models predominantly requiring cellular immunity and T-cell-mediated humoral immunity for clearance of bacteria have been investigated. Host defense against intracellular bacteria eliminated by cellular immune mechanisms in mice with disrupted LTβR-mediated signaling was severely impaired in most studies. The immune response against bacterial pathogens inducing combined T- and B-cell immunity varied in mice with blocked LTβR activation with the different models tested, suggesting that certain bacterial antigens (Citrobacter rodentium) required LTβR-mediated signaling while others (Salmonella enterica) were cleared in an LTβR-independent fashion.

(i) Intracellular mycobacterial infections: BCG and M. tuberculosis.

While a central role for TNF-α in immunity against mycobacterial infections has been well characterized (1, 20, 27, 64), the contribution of soluble LTα₃ as the second TNFR-I ligand in antimycobacterial host defense was unknown. Therefore LTα^−/− mice and TNF-α and LTα double-gene-deficient (TNF/LTα^−/−) mice were infected with Mycobacterium bovis bacillus Calmette-Guérin (BCG) (8, 26, 51) or Mycobacterium tuberculosis (14). Studies investigating the role of LT in experimental mycobacterial disease are of clinical relevance considering the reactivation of tuberculosis observed in patients treated with TNF-α antagonists (21). Table 3 summarizes studies investigating bacterial infections in mice with blocked LTβR signaling. The course of mycobacterial infection was lethal in TNF/LTα^−/− and LTα^−/− mice. Survival was longer in BCG-infected LTα^−/− mice (182 days) and in TNF-α^−/− mice (56 days) than in TNF/LTα^−/− mice (35 days), indicating that the absence of TNF-α in this infection leads to broader immunodeficiency than the absence of LTα. The impaired antimycobacterial immune response in mice without TNF-α was aggravated by the simultaneous lack of LTα (8). Similarly, the bacterial loads of the lung on day 28 of the infection were 1,000-, 40-, and 1.4-fold higher in TNF/LTα^−/−, TNF-α^−/−, and LTα^−/− mice than in the respective wt mice. Introduction of an LTα transgene in TNF/LTα^−/− mice delayed disease onset but failed to restore resistance to BCG infection, suggesting a transient protective effect exerted by LTα in this disease model. Roach et al. generated LTα^−/− → wt bone marrow chimeras in order to investigate the role of soluble LTα₃ in antimycobacterial immunity (58). There was lethal disease in wt mice with lymph nodes and LTα^−/− bone marrow, indicating a critical role for soluble LTα in antimycobacterial immunity.

The proinflammatory cytokine TNF-α is secreted as a soluble TNF-α₃ molecule and is also tethered to the cell membrane. Olleros et al. investigated the role of membrane-bound noncleavable TNF-α by creating membrane TNF-α transgenic mice on the TNF/LTα^−/− background, thus specifically studying the role of membrane TNF-α in the absence of soluble TNF-α and LTα. Following inoculation with BCG, the infection was controlled in membrane TNF^+/+ transgenic TNF/LTα^−/− mice, though the bacterial load was higher in these mice than in wt animals. Thus, membrane TNF alone is capable of controlling BCG infections (51). As TNF-α^−/− mice also succumb to BCG infection, the expression of LTα in the absence of TNF is not sufficient to control this mycobacterial infection (8).

Causes for the impaired antimycobacterial immunity in LTα and TNF/LTα gene-deficient mice varied in the different models studied. The granulomatous responses to BCG infection were similarly delayed and impaired in TNF/LTα^−/− and TNF-α^−/− mice (8, 26). There were fewer macrophages with reduced inducible nitrite oxide synthase (iNOS) and acid phosphatase expression. Fewer T cells could be detected in these lesions. These observations indicate a central role for TNF in the recruitment of T cells and macrophages to granulomatous lesions, which cannot be compensated for by the presence of LTα. Conversely, transgenic expression of noncleavable membrane TNF-α in TNF/LTα^−/− mice resulted in a two- to fourfold increase in the number of hepatic granulomas, which were of smaller size and predominantly consisted of macrophages (51).

In LTα^−/− → wt bone marrow chimeras infected with M. tuberculosis, there was normal recruitment of T cells to the lungs (58). However, pulmonary T cells remained in the perivascular and peribronchial areas and failed to collocate with the macrophages in granulomas.

There are controversial findings generated in different systems regarding the role of LTα₁β₂-LTβR interaction in the control of mycobacterial infections (14, 39, 58). Wild-type mice infected with BCG and undergoing LTβRIgG treatment and LTβR^−/− mice infected with M. tuberculosis suffered a more severe course of disease (14, 39). Similarly, the bacterial loads in livers and lungs of LTβ^−/− mice infected with M. tuberculosis were elevated (14). LTβR^−/− → wt bone marrow chimeras failed to control M. tuberculosis infection (14). The impaired immune response against mycobacteria in mice with disrupted LTβR was associated with decreased iNOS activity in the lung and spleen (14, 39). In contrast, Roach reported normal clearance of mycobacterial infections in LTβ^−/− → wt bone marrow chimeras. The reasons for these discrepant observations are unknown and might be related to a different lymphoid microenvironment in LTβ^−/− → wt and LTβR^−/− → wt bone marrow chimeras. LIGHT, a second ligand of the LTβR, is not involved in the control of disease, as LIGHT^−/− mice cleared M. tuberculosis infections at the same rate as wt mice did (14).

The course of experimental murine listeriosis was more severe in TNF/LTα^−/− and LTβR^−/− mice, suggesting that in addition to TNF-α, LTα and engagement of the LTβR are critical for control of this intracellular pathogen (27, 53, 59).

Most studies indicate that interaction of LTα₃-TNFR and of LTβ with the LTβR is required for control of infections with the intracellular pathogens Mycobacterium and Listeria. The elimination of these pathogens depends strongly on cellular immunity. Only one study utilized LTβRIgG in adult mice (39), a situation comparable to human treatment of autoimmune diseases. This study showed a significant but moderate increase in the number of acid-fast bacilli (three- to fourfold) in LTβRIgG-treated mice compared to 10- to 1,000-fold increases observed in the studies utilizing gene-deficient mice. However, the biological relevance of this observation in terms of disease-related mortality was not investigated in this study, as all mice were sacrificed for in vitro analysis 4 weeks after infection, while most gene-deficient mice used in other studies died after day 30 following mycobacterial infection.

Experimental infectious colitis. (i) Salmonella enterica serovar Typhimurium and Salmonella enterica.

Infection of LT family gene-deficient mice with Salmonella enterica serovar Typhimurium has been utilized to investigate the roles of LTα and TNF-α in the regulation of anti-Salmonella immunity.

Oral infection of TNF/LTα^−/− mice with S. enterica serovar Typhimurium results in a lethal course of infection compared to mild disease in wt mice. This difference was most likely due to reduced recruitment of neutrophils to the site of infection, as well as reduced intracellular killing of S. enterica serovar Typhimurium by granulocytes (12).

Mice undergoing oral pretreatment with streptomycin develop infectious colitis, which closely resembles human S. enterica-induced colitis, following oral infection with S. enterica serovar Typhimurium. The development of S. enterica-induced colitis was not affected by the presence of PP, MLN, or the LTβR, as the courses of the infection in wt and LTβR^−/− mice without PP and MLN were similar. Infection of mice with S. enterica without antibiotic treatment induced a typhoid type of disease with bacterial expansion in PP and MLN. Interestingly, the typhoid type of S. enterica infection was also similar in wt and LTβR^−/− mice, indicating that while S. enterica might home to intestinal lymphoid organs, PP, MLN, and LTβR are not required for antibacterial immunity against this invasive pathogen (4).

(ii) Citrobacter rodentium.

We have recently investigated the role of LT α₁β₂-LTβR interactions in the course of infectious colitis induced by Citrobacter rodentium (65). Infection of mice with the gram-negative bacterium C. rodentium serves as an animal model of human infection with enteropathogenic and enterohemorrhagic Escherichia coli (36). In adult and immune-competent mice, there is only mild transient colitis with hyperplasia of infected colonic epithelial cells. The course of C. rodentium-induced colitis was more severe in LTβRIgG-treated mice, with increased disease-related mortality (65). Similarly, there was nearly 100% disease-related mortality in C. rodentium-infected LTα^−/−, LTβ^−/−, and LTβR^−/− mice, suggesting a critical role for LTα₁β₂-LTβR interactions in anti-Citrobacter immunity. In mice with disrupted LTβR signaling, there were fewer splenic CD11c⁺ dendritic cells following oral infection. FDCs were absent in the spleens of LTβRIgG-treated mice. Similarly, there were fewer colonic lymphoid follicles in LTβRIgG-treated mice and in the gene-deficient mice used. In LTβR^−/− mice, anti-Citrobacter IgG2a antibody titers were reduced while IgG1 titers were increased. Similarly, there was increased Citrobacter-induced secretion of IL-4 in LTβR^−/− mice. These observations indicate that the loss of local intestinal lymphoid organs and changes to antigen-presenting functions of the spleen are associated with impaired immunity against this noninvasive pathogen.

(iii) LPS-induced systemic shock.

A number of studies showed resistance of TNF/LTα^−/− mice against lethal endotoxemia induced by intravenous LPS injection (12, 17), depending on the bacterial origin of the LPS. Eugster described resistance to shock induced by coadministration of d-galactosamine and E. coli-derived LPS (17). Netea et al. demonstrated increased resistance of TNF/LTα^−/− mice to lethal endotoxemia induced by E. coli and K. pneumoniae LPS compared to S. enterica serovar Typhimurium LPS (46). These differences were associated with increased IL-1 and gamma interferon secretion following injection of the lethal S. enterica serovar Typhimurium LPS. BCG-sensitized TNF/LTα^−/− and TNF-α^−/− mice were completely resistant to E. coli LPS-induced shock, whereas LTα^−/− mice showed prolonged survival compared to wt mice (8). Thus, LTα contributes to septic shock, although TNF-α appears to be more potent in the induction of LPS shock than LTα.

Viral infections.

A number of studies have investigated the role of LT in viral infections, most of them studying influenza virus, herpesvirus, and lymphocytic choriomeningitis virus infections in gene-deficient mice with anatomical defects. Table 4 summarizes studies of experimental viral infections in mice. Except for two studies (37, 55), all of them utilized mice with genetic defects of the LT ligands. Similar to LPS-induced shock models, virus-induced systemic shock was less severe in mice with impaired LTβR, most likely due to a depletion of virus-specific CD8⁺ T cells following LTβRIgG treatment. Overall antiviral cytotoxic-T-cell immune responses were more or less impaired, and the clearance of the virus was slowed down or inhibited, leading to a lethal course in influenza A virus (40), murine cytomegalovirus (MCMV) (5), and Theiler's virus (37) infections. In the extensively studied lymphocytic choriomeningitis virus infection model, the defective antiviral immune response was secondary to the loss of the marginal zone in the spleen (6, 45) but not due to the absence of LTβ itself. Similarly, treatment of adult wt mice with LTβRIgG did not affect immunity against Theiler's virus infection, while LTα^−/− and LTβR^−/− mice failed to mount appropriate antiviral cytotoxic-T-cell responses (37), suggesting that changes to splenic and lymph node architecture, but not the presence of LTβ, were critical for clearing of the infection.

Parasite infections.

Studies investigating the role of LT in parasite infections are summarized in Table 5.

(i) Toxoplasma gondii.

Schlüter et al. compared the course of experimental toxoplasmosis in wt, TNF-α^−/−, LTα^−/−, and TNF/LTα^−/− mice in order to dissect the roles of both ligands of the TNF receptors in this infection (62). TNFR-I plays a predominant role in experimental toxoplasmosis. TNF-α induces toxoplasmastatic gamma interferon secretion in macrophages and microglial cells in the central nervous system (9, 34).

All gene-deficient mice tested in this study failed to control intracerebral T. gondii and succumbed to acute necrotizing Toxoplasma encephalitis. The lethal course of disease was associated with reduced intracerebral expression of iNOS and lower splenic NO levels. Experiments with bone marrow reconstitution chimeras demonstrated an exclusive role of TNF-α- and LTα-producing hematopoietic cells for surviving toxoplasmosis.

(ii) Leishmania.

Infection of LTβ^−/− mice with Leishmania major was associated with a fatal course of disease with visceral spread of parasites despite the resistant genetic background of the C57BL/6 mice used in this study (70). The impaired and delayed cellular and humoral anti-L. major immune response in LTβ^−/− mice was secondary to changes in the lymphoid architecture. Reconstitution of LTβ^−/− mice with wt bone marrow failed to restore effective antiparasite immunity, whereas wt mice receiving LTβ^−/− bone marrow were not immunocompromised.

Murine Leishmania donovani infection induces visceral leishmaniasis and is more severe in both TNF-α^−/− and LTα^−/− mice (15). Experiments with bone marrow radiation chimeras indicated a critical role for liver-generated LTα in the migration of leukocytes from periportal to sinusoidal areas, while T-cell-generated TNF-α and LTα were required for the control of parasite growth.

(iii) Trypanosoma brucei.

Infection of LTα^−/− and TNF/LTα^−/− double-gene-deficient mice with the extracellular parasite Trypanosoma brucei was associated with control of disease and slightly prolonged survival of LTα^−/− mice following infection (43). Trypanosoma-specific IgM and IgG2a serum antibody titers were increased in LTα^−/− mice, indicating that germinal centers and FDC networks were not required for this antiparasite humoral immune response.

(iv) Cerebral malaria.

Infection of mice with Plasmodium berghei serves as an animal model for human cerebral malaria. LTα^−/− mice were protected against cerebral malaria, as they did not develop perivascular cerebral hemorrhage. Bone marrow chimera experiments indicated that a radioresistant cerebral cell population is the source of the LTα required for extravasation of malaria-infected erythrocytes (16).

Prion disease/scrapie.

Transmissible spongiform encephalopathies (TSEs), or “prion diseases,” are chronic neurodegenerative diseases that affect humans and animals. Most TSEs, including human variant Creutzfeldt-Jakob disease and experimental prion disease in mice, are transmitted by peripheral exposure. TSE infection results in conversion of normal prion protein (PrP^c) to the disease-associated form, PrP^sc. Intracerebral or peripheral administration of prions to mice induces a rise of infectivity in the spleen and in other lymphoid organs long before the development of neuropathological changes. PrP^sc migrates from the lymphoid compartments to the central nervous system by neuronal transport. FDCs in the germinal centers of lymphoid organs have been implicated as initial sites of accumulation of PrP^sc. FDCs trap antigen-antibody complexes. Studies using intraperitoneal (i.p.) (41, 54) and oral (50) routes of scrapie infection provided different results regarding the role of FDCs and LT ligands/receptors in this infection. LTα^−/−, LTβ^−/−, TNF/LTα^−/−, and LTβR^−/− mice with disrupted LTα₁β₂-LTβR signaling undergoing i.p. inoculation resisted infection and contained no infectivity in spleens and lymph nodes (54). Similarly, pretreatment of wt mice with LTβRIgG prior to i.p. scrapie infection blocked early PrP^sc accumulation in the spleen and reduced disease susceptibility. In contrast, LTα^−/− mice orally infected with scrapie were susceptible to disease while LTβ^−/− mice were resistant (50). However, pretreatment of wt mice with LTβRIgG prior to oral infection with scrapie blocked PrP^sc in PP and MLN and prevented neuroinvasion (41).

As FDCs were similarly absent in TNF-α^−/−, TNFR-I^−/−, and LTα, LTβ^−/− mice but only LT gene-deficient mice were protected against experimental scrapie, FDCs are not required for the replication of scrapie in lymphoid tissue following i.p. infection. More likely, some other yet-undefined effect of impaired LTβR-mediated signaling is critical for control of the expansion of scrapie protein in lymphoid tissues. The susceptibility of LTα^−/− mice to oral scrapie and the resistance of LTβ^−/− mice and LTβRIgG-pretreated mice to oral infection are two controversial observations which require further investigation.