The complementation of lymphotoxin deficiency with LIGHT, a newly discovered TNF family member, for the restoration of secondary lymphoid structure and function

0014-2980/02/0707-1969$17.50+.50/0© WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2002

The complementation of lymphotoxin deficiencywith LIGHT, a newly discovered TNF familymember, for the restoration of secondary lymphoidstructure and function

Jing Wang1, Amy Foster1, Robert Chin1, Ping Yu1, Yonglian Sun1, Yang Wang1, KlausPfeffer2 and Yang-Xin Fu1

1 Department of Pathology and Committee on Immunology, The University of Chicago, Chicago,USA

2 Institut of Medical Microbiology, Immunology, and Hygiene, Technical University of Munich,Munich, Germany

Highly organized lymphoid structures provide the intricate microenvironment essential forthe mediation of the effective immune responses. Compared with lymphotoxin g knockoutmice (LT g –/–), LT g receptor knockout (LT g R–/–) mice present with more severely disorganizedsplenic structures, suggesting the potential involvement of another ligand. LIGHT, a newlyidentified TNF family member, is a costimulatory molecule for T cells and binds to LT g R andherpes virus entry mediator (HVEM) in vitro. Here, we show that the complementation ofLT § –/– mice with a LIGHT transgene (LIGHT Tg/LT § –/–) leads to the restoration of secondarylymphoid tissue chemokine and T/B cell zone segregation. LIGHT Tg/LT § –/– mice also pre-serve dendritic cells, follicular dendritic cell networks, and germinal centers, though not themarginal zone. Consequently, IgG responses to soluble, but not particulate, antigens arerestored, confirming the role of primary follicle and marginal zone in the responses to solubleand particulate antigens. The failure of the LIGHT transgene to rescue the defective splenicstructures in LT g R–/– mice demonstrates that LIGHT can interact with LT g R in vivo. Moreseverely disorganized splenic structures developed after blockade of endogenous LIGHT inLT g –/– mice. These findings uncover the potential interaction between LIGHT and one of itsreceptors, LT g R, in supporting even in the absence of LT the development and maintenanceof lymphoid microenvironment.

Key words: Chemokine / Splenic structure / Antibody response / LIGHT / LT g receptor

Received 18/3/02Accepted 23/4/02

[I 22944]

Abbreviations: BLC: B lymphocyte chemoattractant(CXCL13) FDC: Follicular dendritic cell GC: Germinal cen-ter HVEM: Herpes virus entry mediator LIGHT: Homolo-gous to lymphotoxins, exhibits inducible expression, andcompetes with HSV glycoprotein D for HVEM on T cellsLIGHT Tg/LT § –/–: LIGHT transgene-complemented LT §knockout LT: Lymphotoxin LT I R: LT g receptor MZ: Mar-ginal zone PP: Peyer’s patches SLC: Secondary lymphoidtissue chemokine (CCL21) TD: T cell-dependent Wt: wild-type

1 Introduction

A proper immune response requires orchestrated in-teractions, both temporal and spatial, between rareantigen-specific T/B cells and antigen-presenting cells[1–3]. Lymphoid tissues provide an organized microenvi-ronment for these orchestrated interactions [4]. However,the cytokines and their receptors that are essential for

the development and maintenance of this organizedmicroenvironment have not been well-elucidated. Re-cent studies have indicated that two well-known proin-flammatory cytokines, tumor necrosis factor (TNF) andlymphotoxin (LT), play essential roles in lymphoid organ-ogenesis, in the establishment of the proper microenvi-ronment of secondary lymphoid tissues, and in the gen-eration of adaptive humoral immune responses [5–7].LT § exists as either a soluble homotrimer (LT § 3) or amembrane heterotrimer coupled with LT g , detectable onactivated lymphocytes [8–10]. TNFR-I and TNFR-II,expressed on a wide variety of cells, respond to LT § 3 andTNF § 3 whereas LT g R, which is expressed on cells ofnonhematopoietic origin, responds to LT § 1 g 2 heterotri-mers [9, 10].

Similar to LT § –/– mice, LT g R–/– mice lack lymph nodes(LN) and Peyer’s patches (PP) and show a severe disor-ganization of splenic architecture, characterized by theabsence of T/B cell segregation, marginal zones, follicu-lar dendritic cell (FDC) networks, and germinal centers

Eur. J. Immunol. 2002. 32: 1969–1979 Interaction between LIGHT and lymphotoxin g receptor for lymphoid tissues 1969

(GC) [5, 11, 12]. However, LT g –/– mice retain mesenteric(mLN) and cervical LN (cLN), and possess a reducedcapacity to form GC and FDC, suggesting a possible rolefor soluble LT § 3 [13, 14]. That LT g R–/– mice present with amore severe phenotype than LT g –/– mice suggests apotential interaction between LT g R and another ligand[15].

LIGHT, a newly identified TNF family member, is closelyrelated to LT and can bind to LT g R and HVEM in vitro[16]. LIGHT is predominantly expressed in lymphoid tis-sues, especially on activated T cells [16, 17]. In vitrostudies show that LIGHT is a costimulatory molecule forT cells but also induces apoptosis in certain tumor celllines [17–19]. Blockade of endogenous LIGHT effectivelyprevents the development of graft-versus-host disease(GVHD), whereas local overexpression of LIGHT en-hances tumor rejection [18]. Signaling via HVEM may beimportant for these LIGHT-mediated activities.

However, it has been difficult to define the potential inter-action between LIGHT and LT g R in vivo due to a domi-nant role of membrane LT for LT g R, compound involve-ment of multiple ligands/receptors in the family, and lackof proper reagents. The role of T cell-derived LIGHT,where LIGHT is normally expressed, is also unknown. Tostudy if T cell-derived LIGHT and its interaction withLT g R are sufficient to support the formation of lymphoidtissues, LIGHT transgenic (LIGHT Tg) mice were gener-ated with the constitutive expression of LIGHT in T celllineage and crossed to either LT § –/– or LT g R–/– back-grounds. LIGHT-complemented LT § –/– mice (LIGHT Tg/LT § –/–) showed the intact T/B cell segregation accompa-nied by a substantial recovery of secondary lymphoidtissue chemokine (SLC) production and proper formationof dendritic cell (DC), FDC networks. LIGHT-mediatedrestoration of splenic architecture is dependent on thepresence of LT g R. Moreover, the blockade of endoge-nous LIGHT activity in LT g –/– mice led to a more severelydisrupted splenic architecture in LT g –/– mice, suggestinga potential interaction of LIGHT and LT g R for the optimalformation of lymphoid structures in wild-type (wt) mice.These results uncover the sufficiency of LIGHT, signalingvia LT g R in the absence of LT, in supporting the develop-ment and maintenance of the lymphoid microenviron-ment.

2 Results

2.1 Generation of LIGHT Tg/LT > –/– mice to studythe complementary role of LIGHT in theorganization of lymphoid structure

To determine whether the lymphoid abnormalities identi-fied in the LT § –/– mice can be rescued by a LIGHT-specific signal, we complemented LT § –/– mice with aLIGHT transgene (LIGHT Tg/LT § –/–). These transgenicmice were generated to express LIGHT protein under thecontrol of the proximal lck promoter and CD2 enhancer,which gives rise to constitutive expression of LIGHT inthe T cell lineage [20]. Two LIGHT Tg lines were crossedto LT § –/– mice and LIGHT Tg/LT § –/– mice were generatedto test whether constitutive expression of LIGHT canpotentially play a role in the organization of secondarylymphoid tissues in the absence of LT.

2.2 Intact splenic T/B cell segregation inLIGHT Tg/LT > –/– mice accompanied with thesubstantial recovery of SLC production in theT cell zones

To study the potential role of LIGHT in the complementa-tion of LT deficiency for lymphoid tissue structures,spleen sections of LIGHT Tg/LT § –/– mice were stainedwith anti-B220 (a marker for B cells) and anti-Thy1.2 (amarker for T cells) antibodies (Ab). Interestingly, theresults revealed a largely intact segregation of T/B cellzones with T cells mostly clustered around the centralarterioles and B cells mainly located peripheral to the Tcell zone (Fig. 1A). This structure is directly comparableto the T/B cell segregation in the spleens of wt mice(Fig. 1A). These data demonstrate that even in theabsence of LT, T cell-derived LIGHT is sufficient to main-tain focused T cell zones and the correct segregation ofT/B cells in the spleen.

We predicted that the restoration of T/B cell segregationmay be attributed to LIGHT-induced recovery of chemo-kine production. SLC is one of the chemokines attractingnaive T lymphocytes to the T cell zone in spleen, LN, andPP, and high endothelial venules (HEV) in LN and PP [21,22]. B lymphocyte chemoattractant (BLC) has been pro-posed to act as a B cell-homing chemokine [23]. Recentstudies have demonstrated that the levels of SLC andBLC are significantly reduced in LT § –/– mice [24]. Thus,we examined the expression of SLC in LIGHT Tg/LT § –/–

mice by immunostaining using Ab specific for SLC andB220. Double staining in wt mice revealed the specificexpression pattern of SLC in the T cell zone surroundedby B cells (Fig. 1B). However, the expression of SLC wascompletely undetectable in LT § –/– mice, consistent with

1970 J. Wang et al. Eur. J. Immunol. 2002. 32: 1969–1979

Fig. 1. The intact T/B cell segregation and full recovery of SLC production in LIGHT Tg/LT § –/– mice. (A) Spleen sections from wt,LT § –/–, and LIGHT Tg/LT § –/– mice were stained with anti-B220 for B cells (brown) and anti-Thy1.2 for T cells (blue) Ab. Originalmagnification: 100×. (B) Immunohistochemical analysis of SLC production in spleens from wt, LT § –/–, and LIGHT Tg/LT § –/– mice.Original magnification: 100×. Anti-B220 revealed the B cell zone (brown) and anti-SLC showed the specific T cell zone staining(blue).

previous studies. In contrast, LIGHT Tg/LT § –/– miceshowed full recovery of SLC expression to the levels pro-duced by wt mice (Fig. 1B). Such recovery is associatedwith the increased presence of DC in the spleen(Table 1). It is interesting to note that a recent studyshowed that transgenic expression of LT on B cells, on aLT § –/– background, was also sufficient for the restorationof SLC expression [25]. Whether the formation of the Tcell zone depends on B cell- or T cell-derived ligandsremains to be determined. We also examined the expres-sion of BLC, which remained undetectable in LIGHT Tg/LT § –/– mice (data not shown). Therefore, these resultsdemonstrate that T cell-derived LIGHT can rescue theexpression of SLC in a LT § –/– background. Furthermore,they raise the possibility that fully-recovered SLC, butnot BLC, production may serve as the underlying mech-anism for the proper development, in LIGHT Tg/LT § –/–

mice, of splenic microarchitecture, such as T/B cell seg-regation.

2.3 Restored formation of FDC and GC inLIGHT Tg/LT > –/– mice

Previous reports have suggested that B cell-derived LT isessential for the formation of FDC and GC [26, 27]. Toinvestigate whether T cell-derived LIGHT is sufficient torescue the deficient FDC networks in LT § –/– mice, spleensections from LIGHT Tg/LT § –/– mice were doubly stainedwith anti-B220 as a marker for B cells and an Ab to CR1

Table 1. Complementation of LT § –/– mice with LIGHT trans-gene in a LT g R-dependent mannera)

WT LT § –/– Tg LT § –/– Tg LT g R–/–

SLC ++ – ++ –

T/B cell segregation ++ – ++ –

CD11c+ DC ++ ± ++ ±

CR-Fc+ DC ++ – + –

FDC network ++ – + –

GC formation ++ – + –

Marginal zone ++ – – –

KLH IgG response ++ – ++ –

SRBC IgG response ++ – – –

a) ++: Strongly positive; +: positive; ±: weakly positive;–: undetectable.

(mAb 8C12) as a marker for FDC. Our data has revealedthe presence of organized B cell zones in the spleen ofLIGHT Tg/LT § –/– mice, located peripheral to T cell zones.Although, primary B cell follicles are not recovered ascompared with wt controls (Fig. 2). The organized B cellzones observed in LIGHT Tg/LT § –/– mice contained par-tially restored networks of CR1+ FDC (Fig. 2), confirmedby FDC-M1 staining (data not shown). Similar to wt FDC,


Fig. 2. The rescued formation of FDC and GC. Wt, LT § –/–, and LIGHT Tg/LT § –/– mice were immunized i.p. with KLH and SRBC,and 10 days later the spleens were collected. The sections of frozen spleen were stained with 8C12 (blue) to detect FDC, withPNA (blue) to detect GC, and with anti-B220 (brown) for B cells. LIGHT transgene could suboptimally rescue the formation ofFDC/GC.

restored FDC showed a typical network organization andlocalization in B cell area. Therefore, even in the absenceof LT, T cell-derived LIGHT has the capacity to support,though suboptimally, the development and maintenanceof T/B cell zones and FDC networks. However, the pri-mary B cell follicles are not recovered, suggesting T cell-derived LIGHT is not sufficient for the recovery of BLC,which may be required for the formation of primary B cellfollicles in LIGHT Tg/LT § –/– mice.

To determine whether the FDC clusters restored inLIGHT Tg/LT § –/– mice could sufficiently support the for-mation of secondary B cell follicles (GC) in the absenceof primary B cell follicles, the spleen of LIGHT Tg/LT § –/–

mice were harvested 10 days after i.p. immunization withthe T cell-dependent (TD) antigen KLH. Double immu-nostaining of the spleen sections was performed usinganti-B220 and peanut agglutinin (PNA) as a marker forGC B cells. Wt mice readily developed the characteristicPNA+ GC formed within B cell zones after immunization(Fig. 2). Consistent with previous data [28], LT § –/– micedid not form typical GC and only a few PNA+ cells formedthe aggregates around central arterioles (Fig. 2). Interest-ingly, as compared to wt mice, the spleen sections fromimmunized LIGHT Tg/LT § –/– mice showed the normalappearance of PNA+ cluster formed within B cell zones,although the number and size of these GC were reduced(Fig. 2). The localization of these GC and the finding thatthey overlapped with FDC networks demonstrate thatthe FDC networks rescued by LIGHT are functional forthe formation of GC upon immunization.

2.4 Rectified Ab responses in LIGHT Tg/LT > –/–

mice

Previous studies have shown that LT § –/– mice producehigh levels of IgM but very low levels of IgG Ab uponimmunization, indicating defective Ig class switching[29]. To examine whether the partially restored FDC andGC in the LIGHT Tg/LT § –/– mice could restore IgGresponses to a TD antigen, mice were immunized i.p.with KLH on day 0 and sera were collected on day 10.Anti-KLH IgG1 and IgG2a Ab responses were measuredusing an antigen-specific ELISA. LIGHT Tg/LT § –/– miceshowed significantly increased levels of IgG1 and IgG2aanti-KLH Ab compared to LT § –/– mice (Fig. 3). The levelsof IgG1 in LIGHT Tg/LT § –/– was comparable to wt mice,while IgG2a responses were lower than wt mice thoughsignificantly higher than LT § –/– mice (Fig. 3). Consistentwith the previous data, higher levels of IgM Ab weredetected in sera of LT § –/– mice (data not shown). Theseresults indicate that the constitutive expression of LIGHTtransgene on T cells is sufficient to restore the formationof functional GC in LT § –/– mice, which can effectivelymediate the antigen-specific IgG responses to a TD anti-gen.

2.5 LIGHT transgene fails to restore marginalzone (MZ) structures

The complete absence of MZ in LT § –/–, LT g –/– and LT g R–/–

mice has been shown by previous studies [5]. To exam-ine whether the splenic MZ in LIGHT Tg/LT § –/– mice wasrestored, we performed immunohistochemical staining


Fig. 3. The restored Ab response to a TD antigen. Wt, LT § –/–,and LIGHT Tg/LT § –/– mice were immunized i.p. with KLH(50 ? g/mouse) and 10 days later sera were collected.Antigen-specific IgG1 and IgG2a responses to KLH weredetermined by ELISA. Data are means ± SD of 4 mice pergroup.

of splenic sections using a marker specific for the mar-ginal sinus lining cells. Double immunostaining withMAdCAM-1 (MECA-367) and anti-B220 Ab revealed thecharacteristic organization of these MAdCAM-1+ endo-thelial cells peripheral to the white pulp in wt mice(Fig. 4A). However, similar to spleen sections from LT § –/–

mice, MAdCAM-1+ cells were completely undetectablein LIGHT Tg/LT § –/– mice (Fig. 4A). The absence of metal-lophilic macrophages was also observed using theMOMA-1 mAb. Sialoadhesin, another marker for mac-rophages in the MZ, was also absent in the spleen ofLIGHT Tg/LT § –/– mice (data not shown). Therefore,LIGHT transgenic expression cannot rescue the deficientMZ formation in LT § –/– mice. A previous study has shownthat B cell-derived LT is essential for the presence ofMAdCAM-1 in the MZ [5]. We speculate that T cell-derived LIGHT in the T cell zone in this case may not beable to substitute for the function of B cell-derived LT inthe development of MZ due to spatial or temporal effect.Alternatively, the presence of MZ may be developmen-tally fixed requiring ligands that signal via LT g R duringthe early life.

To discriminate these two possibilities, we further inves-tigated the development of MZ using Ab blockingapproaches and bone marrow transfer (BMT). We firstdetermined whether the LIGHT transgene could rescuethe MZ structure in the absence of LT signals. Previousstudies suggested that the treatment of anti-LT g Ab

(BBF6 mAb), which binds solely to LT § 1 g 2 to block itssignaling via LT g R, can result in defective MZ formation[30]. The results obtained from anti-LT g Ab treatment inLIGHT Tg mice suggest that LIGHT transgene is suffi-cient to support the MZ structure in the absence of LTsignals in the adult mice (Fig. 4B, LIGHT Tg). At the sametime, the treatment with anti-LT g mAb dramaticallyreduced the intensity of MAdCAM-1+ staining in thespleen of wt mice, confirming the dominant effect of LT(Fig. 4B, wt). Thus, it appears that T cell-derived LIGHT issufficient to maintain the MZ structure in adult mice evenin the absence of LT signals from B cells.

We next investigated whether the formation of the MZ isdevelopmentally fixed. Wt and LT § –/– mice were reconsti-tuted with wt bone marrow (BM) and the presence of MZstructures were analyzed 7 weeks after BMT. MZ struc-tures were not restored in LT § –/– mice reconstituted withwt BM, even when the BMT was performed in recipientsas young as 9 days of age (data not shown). In contrast,such BMT rescued the formation of FDC and GC [26].Taken together, these results indicate that formation ofthe MZ requires the LT signal during the early life and isdevelopmentally fixed.

Due to its specific topographical location, the MZ isthought to be important in processing particulate anti-gens, and is involved in the defense against such anti-gens or bacterial invasion [31–33]. LIGHT Tg/LT § –/– miceprovide an interesting model to further test such ahypothesis. We therefore assayed the generation of totalIgG and IgG1 responses against SRBC, a typical particu-late antigen, in LIGHT Tg/LT § –/– mice (Fig. 4C). Very inter-estingly, SRBC-specific IgG response in LIGHT Tg/LT § –/–

mice was significantly reduced as compared to wt mice(Fig. 4C). SRBC-specific IgG1 production was alsoreduced in LIGHT Tg/LT § –/– mice (Fig. 4C). These resultssuggest that the MZ is actively involved in immuneresponses against blood-borne particulate antigens anda disrupted MZ in the spleen of LIGHT Tg/LT § –/– micemay directly result in impaired Ab responses to such par-ticulate antigens.

2.6 Restoration of LIGHT-mediatedmicroenvironment is dependent on LT I R

LIGHT can bind to LT g R and HVEM in vitro [16]. How-ever, it has not been shown that LIGHT can signal LT g Rin vivo. Thus, we next investigated whether LT g R isrequired for LIGHT-mediated restoration of splenicmicroarchitecture. LIGHT Tg mice were bred onto aLT g R–/– background. Immunostaining shows that thesplenic microarchitecture in LIGHT Tg/LT g R–/– mice isindistinguishable from LT g R–/– and LT § –/– mice, including


Fig. 4. Absent MZ structure in LIGHT Tg/LT § –/– mice. (A) Immunohistochemical analysis was performed on frozen sections ofspleens from wt, LT § –/–, and LIGHT Tg/LT § –/– mice immunized i.p. with SRBC. For the detection of MZ structure, the spleens werestained for MAdCAM-1 (blue) and B220 (brown). Original magnification: 100×. (B) Wt and LIGHT Tg mice were treated with con-trol Ab or anti-LT g Ab (100 ? g/mouse), and immunized i.p. with 108 SRBC. 10 days later spleens were harvested, and stained forMAdCAM-1 (blue) and B220 (brown). Original magnification: 100×. Anti-LT g Ab dramatically reduced the MAdCAM-1 staining inwt mice while the positive staining in LIGHT Tg mice was comparable to control Ab treated group. (C) Immune response to theparticulate antigen SRBC. Wt, LT § –/–, and LIGHT Tg/LT § –/– mice were immunized i.p. with 108 SRBC. Sera were collected 10 daysafter immunization and SRBC-specific IgG and IgG1 responses were determined by ELISA. Data are means ± SD of 4 mice pergroup.

the complete absence of T/B cell segregation and FDCnetworks (Fig. 5A). In addition, the recovery of SLC pro-duction in LIGHT Tg/LT § –/– mice appeared to be medi-ated by the interaction between LIGHT and LT g R sinceLIGHT Tg/LT g R–/– mice present with the complete ab-sence of SLC expression and reduced numbers of DC(Fig. 5B and Table 1). Consistent with the absence ofFDC networks, there is no GC formation in LIGHT Tg/LT g R–/– mice upon immunization (Fig. 5A). Anti-KLH IgG1and IgG2a responses were measured by ELISA, and theantigen-specific IgG production was also impaired inLIGHT Tg/LT g R–/– mice (Fig. 5C). Therefore, our datasuggested that the restoration of splenic structures inLIGHT Tg/LT § –/– mice is dependent on LT g R and thatLIGHT could interact with LT g R in vivo. This interactionsupports the formation of organized splenic microarchi-tecture, including T/B cell segregation and the devel-

opment of FDC/GC, for effective humoral immuneresponses to TD antigens.

2.7 The potential role of endogenous LIGHTin the establishment of splenicmicroarchitecture

Previous studies have shown that T/B cell segregation inthe spleen of LT g –/– mice is less severely impaired thanLT g R–/– mice [13–15]. To investigate the possible role ofLIGHT, another ligand for LT g R, in the organization ofmicroarchitecture in the secondary lymphoid tissues,LT g –/– mice were treated with LT g R-Ig to block the activ-ity of endogenous LIGHT in vivo. Immunofluorescent Tand B cell staining shows a disrupted splenic architec-ture in LT g –/– mice, but not as severe as that seen in the


Fig. 5. The interaction between LIGHT and LT g R was essential for the rescuing effects of LIGHT transgene. (A) LIGHT Tg micewere bred to LT g R–/– background (LIGHT Tg/LT g R–/–). Wt, LIGHT Tg/LT g R–/–, and LT g R–/– mice were immunized i.p. with KLH andSRBC, and 10 days later the spleens were collected. The sections of frozen spleen were stained with anti-Thy1.2 (blue) for Tcells, with anti-B220 (brown) for B cells, with 8C12 (blue) for FDC and with PNA (blue) for GC. Original magnification: 100×. (B)Immunohistochemical analysis of SLC production in spleen from wt, LIGHT Tg/LT g R–/–, and LT g R–/– mice. Original magnification:100×. Anti-B220 revealed the B cell zone (brown) and anti-SLC showed the specific T cell zone staining (blue). (C) Wt, LIGHT Tg/LT g R–/–, and LT g R–/– mice were immunized i.p. with KLH (50 ? g/mouse) and 10 days later sera were collected. Antigen-specificIgG1 and IgG2a responses to KLH were determined by ELISA. Data are means ± SD of 4 mice per group.

spleen of LT g R–/– mice (Fig. 6). In contrast, the blockadeof LIGHT activity by the administration of LT g R-Igresulted in a more disorganized splenic architecture withT and B cell completely intermixed in LT g –/– mice (Fig. 6).Moreover, our recent data show that constitutive expres-sion of LIGHT can lead to ectopic lymphoid-like tissue

formation in the intestine [20]. Therefore, like TNF and LT,up-regulated LIGHT may play a role in ectopic lymphoidtissue formation and chronic inflammation in the diseasemodels.


Fig. 6. The possible role of LIGHT in the splenic microarchi-tecture establishment. Spleen sections from 5–8 weeks oldnaive wt, LT g R–/–, LT g –/– mice treated with PBS (LT g –/–, con-trol) and LT g –/– mice treated with LT g R-Ig (LT g –/–, treated) areshown as digital overlays of T and B cell immunofluores-cence. Original magnification: 100×. B cells (green), T cells(red) and areas of T and B cells codistribution (yellow).

3 Discussion

LIGHT, a newly discovered TNF family member, was ini-tially identified as a costimulatory molecule for T cells,and capable of inducing cell death in some situations[16, 17]. Blockade of systemic LIGHT activity preventsthe development of GVHD. Local expression of LIGHT intumors leads to their rejection, probably due to theenhanced tumor immunity caused by the formation oflymphoid-like tissues inside tumor [18, 34]. In this study,we identify a potential role of LIGHT in the developmentand maintenance of splenic microarchitecture by itsinteraction with LT g R in the absence of membrane LT.Using the complementation approach, we show thatLIGHT transgene-complemented LT § –/– mice displaywell-segregated T/B cell zones and retain the capacity toform FDC networks and GC. These structures supportthe development of a TD humoral immune response.Furthermore, we reveal that LIGHT can interact withLT g R in vivo and trigger the downstream signaling path-way that mediates the restoration of splenic structures,humoral immune responses, and chemokine production(summarized in Table 1). Importantly, blocking endoge-nous LIGHT activity in LT g –/– mice leads to further dis-turbed T/B cell segregation, suggesting the potential role

of LIGHT in the maintenance of lymphoid structure andlymphoid organogenesis, especially in the absence ofLT g .

In LIGHT Tg/LT § –/– mice, the capacity to form GC/FDC issufficiently preserved to mediate humoral responses, asevidenced by the full recovery of IgG Ab production toKLH (Fig. 3). In contrast, the same set of mice failed togenerate IgG response to SRBC (Fig. 4C). The obviousdifference between these two experiments is the form ofthe antigens; SRBC is a typical particulate antigen whileKLH is a soluble antigen. It has been proposed that MZmay be required for the transportation of particulate anti-gens but this has been very difficult to confirm due toinability to specifically alter such structures and theinvolvement of multiple cellular components [31]. Recentstudies suggest that MZ B cells are required for an Abresponse to such antigens [32, 33]. In this study, weinvestigate the relationship between the structuraldefects and Ab responses with regards to the involve-ment of TNF/TNFR family members. LT § –/– and LT g R–/–

mice lack MZ but also show many other defects insplenic structure and fail to respond to both soluble andparticulate antigens in the absence of adjuvant [5, 15,35]. In contrast, LIGHT Tg/LT § –/– mice recovered severalstructural defects, but not the MZ. This provides an inter-esting model to address whether MZ is required for solu-ble or particulate antigens. Importantly, such mice couldrespond to soluble KLH but not particulate SRBC, dem-onstrating the essential role of MZ for the generation ofIgG responses to particulate antigens.

Similar to LN/PP, our data indicate that the presence ofMZ is developmentally fixed: 1) Lack of MZ and LN/PP inLT § –/– mice cannot be rescued by prolonged reconstitu-tion with LT-expressing bone marrow cells, even as earlyas a few days after birth (data not shown). 2) LIGHT Tg/LT § –/– mice provide strong signaling to LT g R but fail torestore MZ. Macroscopic examination of LIGHT Tg/LT § –/– mice (n G 16) showed that they still lack mLN, PLNand PP (data not shown). Since mature T cells thatexpress LIGHT can only migrate into the spleen a fewdays after birth, it is likely that LIGHT Tg/LT § –/– miceexpress a sufficient amount of LIGHT protein in the sec-ondary lymphoid tissue only postnatally, and miss thecritical window for the formation of LN and PP, which areprenatally determined.

LIGHT could play multiple roles for the generation ofeffective immune response: signaling via LT g R for theestablishment of the lymphoid structures and signalingvia HVEM for T cell activation. In addition, dysregulationof LIGHT in a site of local inflammation may facilitate theformation of tertiary lymphoid tissues. We have observedthat LIGHT Tg mice develop multiple lymphoid-like


structures in the gut and skin at the age of 5 months;moreover, the overexpression of LIGHT eventuallyinduces chronic inflammation and autoimmune diseases[20]. One striking feature of spontaneous autoimmunediabetes is the prototypic formation of lymphoid tissue-like structures within the pancreas [36, 37]. Our recentstudy demonstrates that the administration of eitherLT g R-Ig or HVEM-Ig to nonobese diabetic mice (NOD)prevents the onset of IDDM, suggesting that the ligandsfor LT g R play a critical role in the development of diabe-tes pathogenesis [20, 38]. Others also find that trans-genic mice expressing LT g R-Fc could prevent the devel-opment of diabetes in NOD mice [39]. Administration ofanti-LT g Ab fails to block the development of IDDM, sug-gesting that the up-regulation of LIGHT may play a domi-nant role in the pathogenesis of IDDM. The administra-tion of both LT g R-Ig and TNFR-Ig had a more profoundeffect on reversal of virus-induced infiltration in the lungthan single treatment [40] and prevention of inflamma-tory bowel diseases and arthritis [41]. It is likely that up-regulation of LIGHT or TNF in local immune responseand chronic inflammation may deliver a strong signal,even in the absence of LT, enough to form lymphoid-likestructures. Therefore, the combined treatment of solublereceptors and Ab may be required to block multipleligands necessary for the formation of lymphoid struc-tures or chemokine gradient during chronic inflamma-tion. Thus, our gene-modified mice may provide uniqueinsights into the potential role of LIGHT in various dis-ease models and its cooperation or complementationwith other TNF family members. T cell lineage-specificexpression of LIGHT will also allow us to further dissectmultiple roles of T cell-derived LIGHT in T cell activation,apoptosis, inflammation, and the development of lym-phoid tissues in the various situations.

4 Materials and methods

4.1 Mice

LT g R–/– mice [15] and LT § –/– mice [11] were backcrossed forfive and seven generations, respectively, to C57BL/6 miceand were maintained under specific pathogen-free condi-tions. Animal care and use were in accord with institutionalguidelines. Control Ab and anti-LT g Ab (a blocking Ab to themLT, kindly provided by Dr. Jeff Browning, Biogen, Cam-bridge, MA) were previously described [30]. LT g R-Ig wasgenerated as previously described [42]. LIGHT Tg mice weregenerated in the University of Chicago Cancer ResearchCenter Transgenic Mice Facility [20].

4.2 Immunohistochemical analysis of splenic structure

Spleens were harvested, treated, and stained as previouslydescribed [29, 35]. FDC-M1 was provided by Dr. Marie

Kosco-Vilbois (Geneva, Switzerland) and anti-SLC Ab waspurchased from Pepro Tech Inc (Rocky Hill, NJ). Anti-B220,Thy1.2, CR1, PNA, and MAdCAM-1 Ab were purchasedfrom PharMingen, San Diego, CA).

4.3 Immunization protocol

Mice were immunized intraperitoneally with 108 SRBC andKLH (50 ? g/mouse). In some experiments, anti-LT g Ab(100 ? g/mouse) was used on the day of immunization with108 SRBC. Sera were collected 10 days after the immuniza-tion for measurement of antigen-specific Ab.

4.4 ELISA for antigen-specific Ab

SRBC-specific Ab were measured and analyzed as previ-ously described [29]. The detection of serum KLH-specificAb was similar as described for SRBC-specific Ab.

4.5 In vivo treatment protocol

LT g –/– mice (5–8 weeks old) were purchased from Taconicand treated with LT g R-Ig (100 ? g/mouse) every 3 days for atotal of three times [42]. Spleens were harvested 3 days afterthe last treatment, embedded in OCT compound (Miles).Serial cryosections (8–10 ? m) were fixed in acetone andstained with PE-conjugated anti-Thy1.2 (for T cells) andFITC-conjugated anti-B220 (for B cells) Ab (PharMingen).Slides were mounted with Mowiol (Hoechst Celanese) con-taining 10% 1,4-diazobicyclo (2.2.2) octane. Samples wereanalyzed using a Zeiss Axioplan microscope and a Photo-metrics PXL CCD camera. Digital images were deconvolvedand overlaid using OpenLab.

Acknowledgements: We thank Shirley Bond for assis-tance with immunofluorescence microscopy and image pro-cessing. This research was supported in part by grants fromNIH (HD37104 and DK58897), and Juvenile Diabetes Foun-dation International (1–2000–875).

References

1 Zinkernagel, R. M., Localization dose and time of antigensdetermine immune reactivity. Semin. Immunol. 2000. 12:163–171.

2 Banchereau, J. and Steinman, R. M., Dendritic cells and thecontrol of immunity. Nature 1998. 392: 245–252.

3 Cyster, J. G., Chemokines and cell migration in secondary lym-phoid organs. Science 1999. 286: 2098–2102.

4 Ochsenbein, A. F., Sierro, S., Odermatt, B., Pericin, M., Karrer,U., Hermans, J., Hemmi, S., Hengartner, H. and Zinkernagel,R. M., Roles of tumour localization, second signals and crosspriming in cytotoxic T cell induction. Nature 2001. 411:1058–1064.


5 Fu, Y. X. and Chaplin, D. D., Development and maturation ofsecondary lymphoid tissues. Annu. Rev. Immunol. 1999. 17:399–433.

6 Ruddle, N. H., Lymphoid neo-organogenesis: lymphotoxin’s rolein inflammation and development. Immunol. Res. 1999. 19:119–125.

7 Ettinger, R., Mebius, R., Browning, J. L., Michie, S. A., vanTuijl, S., Kraal, G., van Ewijk, W. and McDevitt, H. O., Effects oftumor necrosis factor and lymphotoxin on peripheral lymphoidtissue development. Int. Immunol. 1998. 10: 727–741.

8 Browning, J. L., Ngam, E. A., Lawton, P., DeMarinis, J., Tizard,R., Chow, E. P., Hession, C., O’Brine, G. B., Foley, S. F. andWare, C. F., Lymphotoxin beta, a novel member of the TNF familythat forms a heteromeric complex with lymphotoxin on the cellsurface. Cell 1993. 72: 847–856.

9 Browning, J. L., Sizing, I. D., Lawton, P., Bourdon, P. R.,Rennert, P. D., Majeau, G. R., Ambrose, C. M., Hession, C.,Miatkowski, K., Griffiths, D. A., Ngamek, A., Meier, W., Ben-jamin, C. D. and Hochman, P. S., Characterization oflymphotoxin-alpha-beta complexes on the surface of mouselymphocytes. J. Immunol. 1997. 159: 3288–3298.

10 Ware, C. F., VanArsdale, T. L., Crowe, P. D. and Browning, J.L., The ligands and receptors of the lymphotoxin system. Curr.Top. Microbiol. Immunol. 1995. 198: 175–218.

11 De Togni, P., Goellner, J., Ruddle, N. H., Streeter, P. R., Fick,A., Mariathasan, S., Smith, S. C., Carlson, R., Shornick, L. P.,Strauss, S. J. et al., Abnormal development of peripheral lym-phoid organs in mice deficient in lymphotoxin. Science 1994.264: 703–707.

12 Banks, T. A., Rouse, B. T., Kerley, M. K., Blair, P. J., Godfrey, V.L., Kuklin, N. A., Bouley, D. M., Thomas, J., Kanangat, S. andMucenski, M. L., Lymphotoxin-alpha-deficient mice: effects onsecondary lymphoid organ development and humoral immuneresponsiveness. J. Immunol. 1995. 155: 1685–1693.

13 Alimzhanov, M. B., Kuprash, D. V., Koscovilbois, M. H., Luz,A., Turetskaya, R. L., Tarakhovsky, A., Rajewsky, K., Nedos-pasov, S. A. and Pfeffer, K., Abnormal development of second-ary lymphoid tissues in lymphotoxin beta-deficient mice. Proc.Natl. Acad. Sci. USA 1997. 94: 9302–9307.

14 Koni, P. A., Sacca, R., Lawton, P., Browning, J. L., Ruddle, N.H. and Flavell, R. A., Distinct roles in lymphoid organogenesis forlymphotoxin § and g revealed in lymphotoxin g -deficient mice.Immunity 1997. 6: 491–500.

15 Futterer, A., Mink, K., Luz, A., Kosco-Vilbois, M. H. and Pfef-fer, K., The lymphotoxin g receptor controls organogenesis andaffinity maturation in peripheral lymphoid tissues. Immunity 1998.9: 59–70.

16 Mauri, D. N., Ebner, R., Montgomery, R. I., Kochel, K. D.,Cheung, T. C., Yu, G.-L., Ruben, S., Murphy, M., Eisenberg, R.J., Cohen, G. H., Spear, P. G. and Ware, C. F., LIGHT, a newmember of the TNF superfamily, and lymphotoxin § are ligandsfor herpesvirus entry mediator. Immunity 1998. 8: 21–30.

17 Zhai, Y., Guo, R., Hsu, T. L., Yu, G. L., Ni, J., Kwon, B. S., Jiang,G. W., Lu, J., Tan, J., Ugustus, M., Carter, K., Rojas, L., Zhu, F.,Lincoln, C., Endress, G., Xing, L., Wang, S., Oh, K. O., Gentz,R., Ruben, S., Lippman, M. E., Hsieh, S. L. and Yang, D.,LIGHT, a novel ligand for lymphotoxin beta receptor and TR2/HVEM induces apoptosis and suppresses in vivo tumor formationvia gene transfer. J. Clin. Invest. 1998. 102: 1142–1151.

18 Tamada, K., Shimozaki, K., Chapoval, A. I., Zhu, G., Sica, G.,Flies, D., Boone, T., Hsu, H., Fu, Y. X., Nagata, S., Ni, J. andChen, L., Modulation of T cell-mediated immunity in tumor andgraft-versus-host disease models through the LIGHT co-stimulatory pathway. Nat. Med. 2000. 6: 283–289.

19 Rooney, I. A., Butrovich, K. D., Glass, A. A., Borboroglu, S.,Benedict, C. A., Whitbeck, J. C., Cohen, G. H., Eisenberg, R. J.and Ware, C. F., The lymphotoxin-beta receptor is necessary andsufficient for LIGHT-mediated apoptosis of tumor cells. J. Biol.Chem. 2000. 275: 14307–14315.

20 Wang, J., Lo, J., Foster, A., Yu, P., Chen, H. Y. W., Tamada, K.,Chen, L. and Fu, Y. X., The regulation of T cell homeostasis andautoimmunity by T cell derived LIGHT. J. Clin. Invest. 2001. 108:1771–1780.

21 Gunn, M. D., Tangemann, K., Tam, C., Cyster, J. G., Rosen, S.D. and Williams, L. T., A chemokine expressed in lymphoid highendothelial venules promotes the adhesion and chemotaxis ofnaive T lymphocytes. Proc. Natl. Acad. Sci. USA 1998. 95:258–263.

22 Tanabe, S., Lu, Z., Luo, Y., Quackenbush, E. J., Berman, M. A.,Collins-Racie, L. A., Mi, S., Reilly, C., Lo, D., Jacobs, K. A. andDorf, M. E., Identification of a new mouse beta-chemokine,thymus-derived chemotactic agent 4, with activity on T lympho-cytes and mesangial cells. J. Immunol. 1997. 159: 5671–5679.

23 Gunn, M. D., Ngo, V. N., Ansel, K. M., Ekland, E. H., Cyster, J.G. and Williams, L. T., A B cell-homing chemokine made in lym-phoid follicles activates Burkitt’s lymphoma receptor-1. Nature1998. 391: 799–803.

24 Ngo, V. N., Korner, H., Gunn, M. D., Schmidt, K. N., Riminton,D. S., Cooper, M. D., Browning, J. L., Sedgwick, J. D. and Cys-ter, J. G., Lymphotoxin alpha/beta and tumor necrosis factor arerequired for stromal cell expression of homing chemokines in Band T cell areas of the spleen. J. Exp. Med. 1999. 189: 403–412.

25 Ngo, V. N., Cornall, R. J. and Cyster, J. G., Splenic T zonedevelopment is B cell dependent. J. Exp. Med. 2001.194:1649–60.

26 Fu, Y. X., Huang, G., Wang, Y. and Chaplin, D. D., Lymphotoxin-expressing B cells induce formation of splenic clusters of follicu-lar dendritic cells. J. Exp. Med. 1998. 187: 1009–1018.

27 Wang, Y., Wang, J., Sun, Y., Wu, Q. and Fu, Y. X., Complemen-tary effects of TNF and lymphotoxin on the formation of germinalcenter and follicular dendritic cells. J. Immunol. 2001. 166:330–337.

28 Matsumoto, M., Mariathasan, S., Nahm, M., Baranyay, F.,Peschon, J. and Chaplin, D., The role of lymphotoxin and type ITNF receptor in the formation of Germinal centers. Science 1996.271: 1289–1291.

29 Fu, Y. X., Molina, H., Matsumoto, M., Huang, G. M., Min, J. J.and Chaplin, D. D., Lymphotoxin-alpha (Lt-Alpha) supportsdevelopment of splenic follicular structure that is required for IgGresponses. J. Exp. Med. 1997. 185: 2111–2120.

30 Mackay, F., Majeau, G. R., Lawton, P., Hochman, P. S. andBrowning, J. L., Lymphotoxin but not tumor necrosis factorfunctions to maintain splenic architecture and humoral respon-siveness in adult mice. Eur. J. Immunol. 1997. 27: 2033–2042.

31 Kraal, G., Cells in the marginal zone of the spleen. Int. Rev. Cytol.1992. 132: 31–74.

32 Martin, F. and Kearney, J. F., B cell subsets and the mature pre-immune repertoire. Marginal zone and B1 B cells as part of a“natural immune memory”. Immunol. Rev. 2000. 175: 70–79.

33 Guinamard, R., Okigaki, M., Schlessinger, J. and Ravetch, J.V., Absence of marginal zone B cells in Pyk-2-deficient micedefines their role in the humoral response. Nat. Immunol. 2000. 1:31–36.

34 Schrama, D., thor Straten, P., Fischer, W. H., McLellan, A. D.,Brocker, E. B., Reisfeld, R. A. and Becker, J. C., Targeting oflymphotoxin-alpha to the tumor elicits an efficient immune


response associated with induction of peripheral lymphoid-liketissue. Immunity 2001. 14: 111–121.

35 Wang, Y., Huang, G., Wang, J., Molina, H., Chaplin, D. D. andFu, Y. X., Antigen persistence is required for somatic mutationand affinity maturation of immunoglobulin. Eur. J. Immunol. 2000.30: 2226–2234.

36 Delovitch, T. L. and Singh, B., The nonobese diabetic mouse asa model of autoimmune diabetes: immune dysregulation gets theNOD [published erratum appears in Immunity 1998. 8: 531].Immunity 1997. 7: 727–738.

37 Tisch, R. and McDevitt, H., Insulin-dependent diabetes mellitus.Cell 1996. 85: 291–297.

38 Wu, Q., Salomon, B., Chen, M., Wang, Y., Hoffman, L. M.,Bluestone, J. A. and Fu, Y. X., Reversal of spontaneous autoim-mune insulitis in nonobese diabetic mice by soluble lymphotoxinreceptor. J. Exp. Med. 2001. 193: 1327–1332.

39 Ettinger, R., Munson, S. H., Chao, C. C., Vadeboncoeur, M.,Toma, J. and McDevitt, H. O., A critical role for lymphotoxin-beta receptor in the development of diabetes in nonobese dia-betic mice. J. Exp. Med. 2001. 193: 1333–1340.

40 Puglielli, M. T., Browning, J. L., Brewer, A. W., Schreiber, R. D.,Shieh, W. J., Altman, J. D., Oldstone, M. B., Zaki, S. R. andAhmed, R., Reversal of virus-induced systemic shock and respi-ratory failure by blockade of the lymphotoxin pathway. Nat. Med.1999. 5: 1370–1374.

41 Mackay, F., Browning, J. L., Lawton, P., Shah, S. A., Comiskey,M., Bhan, A. K., Mizoguchi, E., Terhorst, C. and Simpson, S.J., Both the lymphotoxin and tumor necrosis factor pathways areinvolved in experimental murine models of colitis. Gastroenterol-ogy 1998. 115: 1464–1475.

42 Wu, Q., Wang, Y., Wang, J., Hedgeman, E. O., Browning, J. L.and Fu, Y. X., The requirement of membrane lymphotoxin for thepresence of dendritic cells in lymphoid tissues. J. Exp. Med.1999. 190: 629–638.

Correspondence: Yang-Xin Fu, Department of Pathology,MC3083, 5841 S. Maryland, The University of Chicago, Chi-cago, IL 60637, USAFax: +1-773-834-8940e-mail: yfu — midway.uchicago.edu


Documents

The complementation of lymphotoxin deficiency with LIGHT, a newly discovered TNF family member, for the restoration of secondary lymphoid structure and function