IκB Kinase Complex α Kinase Activity Controls Chemokine and High Endothelial Venule Gene Expression in Lymph Nodes and Nasal-Associated Lymphoid Tissue1

The lymphotoxin (LT) β receptor plays a critical role in secondary lymphoid organogenesis and the classical and alternative NF-κB pathways have been implicated in this process. IKKα is a key molecule for the activation of the alternative NF-κB pathway. However, its precise role and target genes in secondary lymphoid organogenesis remain unknown, particularly with regard to high endothelial venules (HEV). In this study, we show that IKKαAA mutant mice, who lack inducible kinase activity, have hypocellular lymph nodes (LN) and nasal-associated lymphoid (NALT) tissue characterized by marked defects in microarchitecture and HEV. In addition, IKKαAA LNs showed reduced lymphoid chemokine CCL19, CCL21, and CXCL13 expression. IKKαAA LN- and NALT-HEV were abnormal in appearance with reduced expression of peripheral node addressin (PNAd) explained by a severe reduction in the HEV-associated proteins, glycosylation-dependent cell adhesion molecule 1 (GlyCAM-1), and high endothelial cell sulfotransferase, a PNAd-generating enzyme that is a target of LTαβ. In this study, analysis of LTβ−/− mice identifies GlyCAM-1 as another LTβ-dependent gene. In contrast, TNFRI−/− mice, which lose classical NF-κB pathway activity but retain alternative NF-κB pathway activity, showed relatively normal GlyCAM-1 and HEC-6ST expression in LN-HEV. In addition, in this communication, it is demonstrated that LTβR is prominently expressed on LN- and NALT-HEV. Thus, these data reveal a critical role for IKKα in LN and NALT development, identify GlyCAM-1 and high endothelial cell sulfotransferase as new IKKα-dependent target genes, and suggest that LTβR signaling on HEV can regulate HEV-specific gene expression.

Given the relatively mild defects in TNF-␣ Ϫ/Ϫ mice, the phenotype of TNFRI Ϫ/Ϫ mice does not reflect signaling from TNF-␣ only but suggests an important role for LT␣:TNFRI interaction in secondary lymphoid organogenesis. LT␤ Ϫ/Ϫ mice lack PP and peripheral LNs (PLNs) but retain mesenteric LNs (MLN) and cervical LNs, and exhibit splenic defects that are less severe than those of LT␣ Ϫ/Ϫ mice (7).
Two NF-B signaling pathways, the canonical and alternative pathways that can be activated in response to TNFR1 or LT␤R engagement, have been implicated in differential regulation of secondary lymphoid organogenesis (8). Additional TNF family receptors, B cell-activating factor (BAFF)-R and CD40, also appear to participate in the alternative pathway (9). In mammals, the NF-B family of transcription factors includes five members that form various homo-and heterodimeric complexes: NF-B1 (p105 processed to p50), NF-B2 (p100 processed to p52), RelA (p65), RelB, and c-Rel (9). Activation of the canonical NF-B pathway can be induced by a wide variety of stimuli including proinflammatory cytokines, such as IL-1 and TNF-␣, bacterial endotoxins, and viral proteins (10). In most cell types, NF-B is sequestered in the cytoplasm by inhibitory IB proteins. Activation of NF-B depends on the IB kinase complex (IKK), which contains the IKK␥ regulatory subunit and two catalytic subunits, IKK␣ and IKK␤ (9). IKK␤ and IKK␥ are critical mediators of the canonical NF-B pathway required for phosphorylation of IB proteins, whereas IKK␣ is dispensable for activation of this pathway in response to most stimuli (10 -12). Following IKK activation, the IB proteins are phosphorylated and degraded through a ubiquitindependent process, thereby allowing nuclear translocation of NF-B dimers that activate expression of proinflammatory genes, including VCAM, MIP-1␤, and MIP-2 (8,9). LT␣␤ binding to LT␤R results in activation of the canonical NF-B pathway as well as the alternative NF-B pathway. The latter is based on phosphorylation-dependent proteolytic processing of cytosolic NF-B2 (p100) to p52 and nuclear translocation of p52:RelB heterodimers (8,9,(13)(14)(15). Activation of the alternative pathway requires the NF-B-inducing kinase (NIK) and IKK␣, but is independent of IKK␤ and IKK␥ (8,13,16). Early studies in embryonic fibroblasts from alymphoplasia mice (aly/aly), which lack functional NIK (17), and from IKK␣ Ϫ/Ϫ mice revealed that NIK and IKK␣ are dispensable for TNFRI-induced NF-B activation but are required for activation of the alternative NF-B pathway in response to LT␤R engagement (14). Other studies have demonstrated that LT␤R-induced NF-B2 processing and RelB nuclear translocation are NIK-and IKK␣-dependent (8,16). Early data implicated both the canonical and alternative NF-B pathways in control of lymphoid chemokine expression-CCL19 (EBV-induced molecule 1 ligand chemokine), CCL21 (secondary lymphoid chemokine), and CXCL13 (B lymphocyte chemoattractant; Ref. 18). More recent results derived from the activation of the LT␤R in vivo by treatment of mice with an LT␤R agonistic reagent suggests that the LT␤R3 NIK3 IKK␣ alternative pathway induces expression of these lymphoid chemokine genes (8). Furthermore, studies in LT␤-deficient and rat insulin promoter LT␣␤ transgenic mice indicate an involvement of LT␤R signaling in regulation of high endothelial venule (HEV)-specific gene expression (19). Though the classical NF-B pathway has been analyzed in endothelial cell lines in vitro, the roles of the classical and alternative NF-B pathways in HEV in vivo is completely unknown. This is the major focus of the present communication.
HEV are specialized postcapillary venules that facilitate L-selectin (CD62L) ϩ lymphocyte extravasation into LN and NALT through expression of adhesion molecules, peripheral node addressin (PNAd), and lymphoid chemokines. PNAd, a CD62L ligand detected by the MECA 79 Ab, is generated by high endothelial cell sulfotransferase (HEC-6ST)-mediated sulfation of various glycoproteins including CD34 and glycosylation-dependent cell adhesion molecule 1 (GlyCAM-1; Ref. 20). In the absence of HEC-6ST, the normal luminal MECA 79 staining pattern is lost, though abluminal staining remains (20). Although the LT␣: TNFRI system can regulate mucosal addressin cell adhesion molecule 1 expression (21)(22)(23), presumably through the canonical NF-B pathway, our previous data suggest that the LT␣␤:LT␤R system is a key regulator of PNAd through induction of HEC-6ST (19). One hypothesis to be tested in this study is that this occurs through the alternative NF-B pathway.
Analysis of aly/aly mice or mice with targeted disruptions of alternative NF-B pathway genes, including NIK, NF-B2, and RelB have revealed important roles for all of these molecules in the generation of normal splenic microarchitecture and PP development (24). Aly/aly and NIK Ϫ/Ϫ mice lack all LNs, while mice deficient in the downstream molecules, RelB and NF-B2, have a less profound defect in that they lack some, but not all LNs (24,25). A role for IKK␣ in LN organogenesis has not been established. IKK␣ Ϫ/Ϫ mice exhibit perinatal death associated with defects in skeletal morphogenesis and epidermal differentiation (26,27). Given the complete absence of LNs in NIK Ϫ/Ϫ mice, the relative paucity of LNs in RelB Ϫ/Ϫ and NF-B2 Ϫ/Ϫ mice, and IKK␣ perinatal lethality, studies addressing the role of these molecules in secondary lymphoid organogenesis have focused on PP and spleen development and have revealed the absence of PP and multiple splenic defects (24).
The NALT is an organized lymphoid tissue found in the rodent nasal tract that has structural similarities to PPs as well as the tonsils and adenoids in humans due to its common origin from the Waldeyer's ring (28). Previous studies have demonstrated that initiation of NALT organogenesis is independent of LT␣, LT␤, LT␤R, and NIK, NF-B2, and RelB. However, the NALT of mice deficient in these molecules exhibit reduced cellularity and diminished expression of lymphoid chemokines and vascular addressins (24,29). To address the contribution of the alternative NF-B pathway to LN and NALT organogenesis, particularly the role of IKK␣ in these processes, we have studied IKK␣ AA mutant mice, in which two amino acid substitutions were introduced into the activating phosphorylation sites of IKK␣, resulting in prevention of its activation by upstream stimuli (30). IKK␣ AA mice are viable and display only minor developmental defects (30). In regard to lymphoid organ development, IKK␣ AA mice were found to lack PP and exhibit defects in splenic organization similar to IKK␣ Ϫ/Ϫ mice (13).
In this study, we provide in vivo evidence that IKK␣ kinase activity is required for normal LN and NALT organogenesis through regulation of tissue microarchitecture, cellular composition, lymphoid chemokine expression, and HEV-specific gene expression. In addition to highlighting a role for IKK␣ kinase activity in LN and NALT organogenesis, these studies identify GlyCAM-1 and HEC-6ST as new IKK␣-dependent target genes.

Materials and Methods
Mice TNFRI Ϫ/Ϫ mice were purchased from The Jackson Laboratory (Bar Harbor, ME). LT␤ Ϫ/Ϫ mice have been previously described (31). TNFRI Ϫ/Ϫ and LT␤ Ϫ/Ϫ mice, on a C57BL/6 background, were studied under a protocol approved by the Yale University Institutional Animal Care and Use Committee. IKK␣ AA mice, on a 129/C57BL/6 genetic background, were previously described (30) and studied under a standard protocol approved by the University of California, San Diego, Office of Animal Care (La Jolla, CA) according to National Institutes of Health guidelines. All mice were studied between 6 and 10 wk of age.

LN visualization
Adult mice were injected in the hind footpads with 10 l of 1% Evans blue in PBS (Eastman Kodak, Rochester, NY) 24 h before sacrifice. The dye becomes concentrated within lymphoid tissue and facilitates macroscopic detection of even very small LNs.

Histologic analysis
For immunofluorescence, LN and NALT tissues were dissected and frozen in OCT compound (Tissue-Tek; Sakura Finetek, Torrance, CA) on dry ice. NALT tissue was isolated as previously described (32). Seven-micrometer tissue sections were cut (longitudinal sections of NALT tissue) onto poly-L-lysine-coated glass slides (Sigma-Aldrich, St. Louis, MO), fixed in 100% cold acetone for 10 min, and stored at Ϫ70°C. For staining, slides were air-dried at room temperature. Sections were double-stained with anti-HEC-6ST generated by our laboratory (19) and MECA 79 (anti-PNAd; BD Pharmingen, San Diego, CA) as previously described (19), or with MECA 79 and anti-LT␤R (BD Pharmingen) Abs. For MECA 79 and LT␤R double-staining, MECA 79 was detected using Cy3-conjugated goat anti-rat IgM Ab (Jackson ImmunoResearch Laboratories, West Grove, PA), and LT␤R was detected using biotinylated anti-hamster IgG (Jackson Immu-noResearch Laboratories) secondary Ab followed by incubation with Cy2conjugated streptavidin (Jackson ImmunoResearch Laboratories). Slides were analyzed by fluorescence microscopy using a Zeiss Axioskop microscope (Carl Zeiss Microimaging, Thornwood, NY).
For immunohistochemistry, frozen sections of LN tissue were prepared and stained as previously described (19). Primary Abs used were as follows: anti-CD35 (CR1) from BD Pharmingen, and rabbit-anti-mouse Gly-CAM-1 (CAM 02) was a generous gift from Dr. S. Rosen (University of California, San Francisco, CA).

In situ hybridization
The technique previously described by Drayton et al. (19) was used. Briefly, fixed LNs were cut onto poly-L-lysine-coated slides and analyzed with the following digoxigenin (DIG)-labeled riboprobes: CCL21, CCL19, and CXCL13 sense and antisense probes. Signal was detected by overnight incubation of sections with alkaline-phosphatase-conjugated sheep anti-DIG Ab (Roche Diagnostic Systems, Mannheim, Germany) and was developed with NBT/5-bromo-4-chloro-3-indolyl phosphate (Invitrogen Life Technologies, Gaithersburg, MD).

Quantitative RT-PCR
Total RNA was prepared from WT and IKK␣ AA LNs with an RNeasy kit (Qiagen, Valencia, CA) and treated with RNase-Free DNase set to remove residual genomic DNA (Qiagen). First strand cDNA was prepared using oligo(dT) primers and SuperScript II reverse transcriptase (Invitrogen Life Technologies). Samples were analyzed in triplicate with SYBR Green Master Mix (Qiagen). GAPDH was used as an endogenous control to determine relative expression of target genes. Data were collected using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA). CCL21 and CXCL13 oligonucleotide primers have been previously described (8).

Rudimentary LNs from IKK␣ AA mice exhibit structural and cellular defects
Previous studies have established a role for IKK␣ kinase activity in PP and splenic development through analysis of IKK␣ AA mutant mice (30). However, LN development has not been studied in detail. To further determine the contribution of IKK␣ to lymphoid organs, we investigated the presence of LNs in IKK␣ AA mice by injection of Evans blue in the hind footpad. Upon initial macroscopic examination of untreated IKK␣ AA mice, inguinal LNs appeared to be absent while the presence of other LNs, including the MLNs, was apparent. However, closer examination of Evans blueinjected animals revealed the presence of several other LNs, including inguinal LNs (Fig. 1A). Although all LNs analyzed (axillary, brachial, iliac, lumbar, inguinal, popliteal, cervical, and mesenteric) could be detected in IKK␣ AA mice, they were smaller than those of wild-type (WT) mice (Fig. 1A) and exhibited reduced cellularity (Table I). Of the LNs examined, inguinal, axillary, and popliteal LNs exhibited the most striking reduction in size (Fig.  1A, and data not shown). Given the reduction in the size of IKK␣ AA LNs, we next examined the cellular composition of these tissues by flow cytometric analysis (FACS). Quantitation of absolute cell number and FACS analysis of IKK␣ AA LN cells revealed a reduction in total cellularity. All populations were affected ( Fig.  1 and Table I) and, in fact, there was a large proportion of cells that did not react with standard phenotypic markers. Most notably, the proportion and number of mature B220 ϩ CD19 ϩ B cells was markedly reduced ( Fig. 1B and Table I). Although B220 ϩ CD19 ϩ B cells comprised 40% of the total LN cells in WT controls, this population was reduced to 4% in IKK␣ AA LNs. No significant differences in the proportion of CD4 ϩ and CD8 ϩ T cells were observed in WT and IKK␣ AA LNs (Fig. 1B) or the thymus (data not shown). In addition to B cell defects, we also observed a reduction in the proportion and number of CD11c ϩ dendritic cells and CD11b ϩ macrophages in IKK␣ AA LNs ( Fig. 1B and Table I). This impaired LN cellular composition prompted us to investigate the contribution of IKK␣ AA kinase activity to other hallmarks of normal LNs. Two-color immunofluorescence analysis of WT and IKK␣ AA LNs revealed numerous defects in IKK␣ AA LN microarchitecture. WT LNs exhibited T and B cell compartmentalization and prominent B cell follicles (Fig. 2). Consistent with the paucity of B cells revealed by flow cytometric analysis, IKK␣ AA LNs were characterized by only a thin rim of B cells located in the LN cortex (Fig. 2). Further immunohistochemical analysis of the B cell compartment in these LNs with anti-CD35 revealed fewer and smaller follicular dendritic cell (FDC) networks than in WT controls (Fig. 2).  a Cells were isolated from WT and IKK␣ AA PLNs. PLNs included pooled brachial and axillary LNs. Cells were first counted by trypan blue exclusion on a hemacytometer, the proportion of the indicated cell populations was determined by flow cytometry, and the absolute cell number was determined; (n ϭ 3).

Diminished stromal and HEV-specific lymphoid chemokine expression in IKK␣ AA LNs
Having noted impaired follicle development and cellular recruitment to IKK␣ AA LNs, we sought to identify a mechanism for this defect. In situ hybridization analysis for lymphoid chemokine expression (CCL19, CCL21, and CXCL13) revealed a substantial reduction in the accumulation of all three chemokine transcripts in IKK␣ AA LNs when compared with WT LNs (Fig. 3A). Consistent with fewer B cells in IKK␣ AA LNs, CXCL13 transcripts were significantly reduced and in some cases, undetectable by in situ hybridization analysis (Fig. 3A). Stromal expression of CCL21 and CCL19 was also reduced in IKK␣ AA LNs (Fig. 3A). Moreover, HEV-specific CCL21 expression was greatly diminished (Fig. 3A,  inset). Quantitative RT-PCR analysis revealed a respective 10-and 27-fold reduction in CCL21 and CXCL13 expression compared with WT controls (Fig. 3B).
To further examine the contribution of IKK␣ kinase activity to the HEV phenotype, we investigated PNAd, the HEV-borne CD62L ligand, and HEC-6ST expression by two-color immunofluorescence on WT, TNFRI Ϫ/Ϫ , LT␤ Ϫ/Ϫ , and IKK␣ AA LNs. WT LN-HEV exhibited concomitant pericellular (i.e., luminal and abluminal) PNAd expression and high levels of HEC-6ST (Fig. 4, E and I). Analysis of TNFRI Ϫ/Ϫ LNs revealed a pattern of PNAd and HEC-6ST expression similar to that seen in WT controls, although a few small vessels displayed only abluminal PNAd expression  with no detectable HEC-6ST (Fig. 4, F and J). The HEV phenotype in TNFRI Ϫ/Ϫ mice is the reciprocal of that seen in LT␤ Ϫ/Ϫ mice. Whereas TNFRI Ϫ/Ϫ mice exhibited predominately pericellular PNAd expression on HEV, we have previously shown that LT␤ Ϫ/Ϫ mice exhibit a higher proportion of HEV displaying only abluminal PNAd and a corresponding absence of HEC-6ST on these vessels (Fig. 4, H and L; Ref. 19). In contrast to WT and TNFRI Ϫ/Ϫ LN-HEV, IKK␣ AA LN-HEV displayed an abnormal pattern of PNAd staining characterized by a lack of distinct pericellular PNAd expression (Fig. 4G) presumably due to the modification of other glycoproteins, such as CD34, by a sulfotransferase other than HEC-6ST as previously discussed (19,20). Furthermore, HEC-6ST expression was markedly reduced compared with WT controls and TNFRI Ϫ/Ϫ mice (Fig. 4K). Consistent with diminished HEC-6ST expression on IKK␣ AA LN-HEV, quantitative RT-PCR revealed a 4-fold reduction in this transcript compared with WT (data not shown).
Given diminished HEC-6ST expression and nearly undetectable GlyCAM-1 expression in IKK␣ AA LN-HEV, we next examined CD62L high cell recruitment to these LNs as an indicator of HEV function. We have previously reported that CD62L ϩ cell accumulation is reduced in LT␤ Ϫ/Ϫ MLN (19). Although the proportion of CD62L ϩ cells was similar in both WT, TNFRI Ϫ/Ϫ , and IKK␣ AA MLNs, the intensity of CD62L staining was reduced on only those cells isolated from IKK␣ AA MLNs pointing to a reduction in functional CD62L on HEV (Fig. 5A, and data not shown). Whereas nearly 70% of WT MLN cells exhibited high CD62L expression, this population was dramatically reduced to 24% in IKK␣ AA MLNs (Fig. 5B)

IKK␣ kinase activity is important for normal NALT organogenesis
We also investigated the involvement of IKK␣ in NALT development. Gross histology revealed the presence of this organ in both WT and IKK␣ AA mice (Fig. 6). However, when compared with WT controls, IKK␣ AA NALTs were reduced in size (Fig. 6,  A and E). Whereas the WT NALT is characterized by a large proportion of B cells surrounded by a smaller ring of T cells (Fig.  6B), IKK␣ AA NALTs exhibited marked structural disruptions including comingling of T and B cells (Fig. 6F). As was seen in the LN, both the number of T and B cells in the NALT appeared reduced but the effect was more pronounced with regard to B cells.   (19). Original objective, ϫ40.

LT␤R is expressed on LN-and NALT-HEV
We have previously shown that LT␤, likely signaling through the LT␤R, is important for the optimal expression of HEC-6ST and PNAd on HEV (19), and shown here with analysis of LT␤ Ϫ/Ϫ mice that GlyCAM-1 is also an LT␤ target. Given that IKK␣ is a crucial component of LT␤R signaling and that there were HEV defects in IKK␣ AA LN and NALT tissue, we sought to determine whether LT␤R was expressed on HEV to determine whether LT␤R activity could contribute directly to HEV-specific gene expression. C57BL/6 LN and NALT tissue were analyzed by twocolor immunofluorescence with MECA 79 (anti-PNAd) and anti-LT␤R Abs (Fig. 7). Examination of LN and NALT tissue revealed prominent coexpression of both PNAd and LT␤R on HEV. Interestingly, while PNAd ϩ LT␤R ϩ HEV were readily detected in both LN and NALT tissue, there are also some HEV that expressed only the LT␤R, particularly in the NALT.

Discussion
In this paper, we have studied the contribution of IKK␣, an essential component of the alternative NF-B pathway, to LN and NALT organogenesis. By using IKK␣ AA knockin mice, we provide evidence that IKK␣ activation and kinase activity contribute in a crucial manner to both LN and NALT organogenesis through regulation of tissue cellularity, organization, chemokine expression, and HEV development. IKK␣ AA mice were characterized by the presence of small, hypocellular LN and NALT tissue exhibiting multiple structural defects. Although IKK␣ AA kinase activity is not critical for T and B cell compartmentalization in the LN, it plays an important role in B cell follicle development and is also crucial for compartmentalization of the NALT. Furthermore, consistent with the hypothesis that signaling via the alternative NF-B pathway regulates lymphoid chemokine expression (8), we find that IKK␣ AA mice have severe defects in LN expression of CCL21, CCL19, and CXCL13. The reduced expression of these chemokines involved in the recruitment of lymphoid and myeloid cells to secondary lymphoid organs (33) may account, in part, for the reduced numbers of cells. Further analysis also revealed a previously unidentified role for IKK␣ in HEV development through regulation of HEV-specific gene expression. It is likely that impaired expression of HEV genes also contributes to the significant reduction in cell recruitment to IKK␣ AA LN and NALT organs. It is of interest that CD62L expression on B cells is only half that of T cells, and that when CD62L expression is limited, reduced rolling is seen on HEV where CD62L ligand expression is also reduced (34). In previous reports, it has been noted that additional CD62L ligands probably exist (35,36) and these ligands could bind cells expressing different levels of CD62L and could account for the entry of CD62L low cells into IKK␣ AA LNs. It is also possible that a different ligand/receptor pair could be used by these cells.
Numerous reports addressing the cellular and molecular interactions that govern secondary lymphoid organogenesis have demonstrated that LN, PP, and NALT each require a different set of signals for their development (7). For example, LT␤Rand NIKdeficient mice lack all LNs and PP, yet retain NALT formation, suggesting that signaling via the alternative NF-B pathway is not crucial for initiation of NALT development (29). In this study, we report that IKK␣ kinase activity is also dispensable for NALT formation but plays an important role in NALT maturation. Unlike LNs, which complete development during embryogenesis or shortly thereafter, the NALT continues to develop up to 6 wks after birth. Similar to LNs, IKK␣ kinase activity plays a crucial role in NALT architecture, cellularity, and HEV development.
We have previously described a role for LT␣␤ in HEV development through the regulation of PNAd and HEC-6ST expression and proposed that this regulation occurs through the LT␤R3 NIK3 IKK␣ alternative NF-B pathway (19). In support of this hypothesis, in this study, we show that IKK␣ is necessary for optimal PNAd and HEC-6ST expression on LN-and NALT-HEV. In addition, we demonstrate that a mechanism by which IKK␣ regulates the HEV phenotype is through regulation of Gly-CAM-1, a core glycoprotein in the PNAd-generating pathway, and regulation of the lymphoid chemokine, CCL21. We also demonstrate that similar to IKK␣, LT␤ also regulates GlyCAM-1 expression in HEV. The observation that TNFRI Ϫ/Ϫ LNs have a slightly reduced expression of both HEC-6ST and GlyCAM-1 indicates that while the canonical NF-B pathway may not play a primary role in HEV-specific gene expression, it does cooperate with the alternative NF-B pathway in this process.
Although IKK␣ is crucial for LT␤R signaling, we note it has also been implicated in signaling downstream of other receptors including CD40 (37), BAFF-R (38), and RANK (30). Therefore, it is conceivable that these other receptors also contribute to the secondary lymphoid organ defects observed in IKK␣ AA mice. Because CD40 is necessary for the clonal expansion of B cells (39), and BAFF contributes to their maturation (40,41), it is possible that the reduction in mature B cell cellularity in IKK␣ AA LNs is due, in part, to a greater proportion of immature B cells. Nonetheless, the fact that LT␤ Ϫ/Ϫ and IKK␣ AA mice have overlapping defects in HEV further supports the notion that the LT␤R3 NIK3 IKK␣ alternative NF-B pathway regulates HEVspecific gene expression and plays a key role in LN and NALT organogenesis.
An interesting question is whether the HEV defects in IKK␣ AA mice are a result of diminished IKK␣ activity in high endothelial cells and/or defective LN stromal cell signaling whose products affect the HEV phenotype. The most direct model for LT␣␤ and IKK␣ involvement in HEV-specific gene expression is that LT␣␤ binds to LT␤R expressed on endothelial cells, engages the alternative NF-B pathway, triggers IKK␣ activation and subsequent transcription of HEC-6ST, GlyCAM-1, and CCL21. Although LT␤R expression has been detected on HUVEC cells (42), its expression on HEV had not been described previously. In this paper, we establish that LT␤R is highly expressed on HEV in both the LN and NALT; a finding in support of the hypothesis that LT␣␤ can exert its effects directly on high endothelial cells. Future studies will address whether direct signaling via the LT␤R on high endothelial cells in vivo regulates gene expression.
Targeted disruptions of alternative NF-B pathway components including LT␤, LT␤R, NIK, NF-B2, and RelB, result in extensive defects in secondary lymphoid organogenesis (7,24). The defects in IKK␣ AA mice are not as severe as those seen in mice deficient in the upstream signaling molecule, NIK, in that IKK␣ AA mice retain LN and NALT albeit these organs are defective. One possibility for the retention of small, rudimentary LNs and NALT tissue in IKK␣ AA mice is that the IKK␣ AA mutation only prevents inducible IKK␣ kinase activity but permits basal activity (13). However, as the phenotype of IKK␣ AA mice is similar to that of mice deficient in downstream molecules of the alternative NF-B pathway including NF-B2 and RelB, the basal activity of the IKK␣ AA variant must be very low. Most notably, NF-B2 Ϫ/Ϫ and IKK␣ AA mice have strikingly similar LN defects-small, undeveloped LNs, defective follicle formation, and a marked reduction in B cell accumulation (24). Therefore, while IKK␣ kinase activity is not essential for initiation of LN and NALT organogenesis, it is crucial for later stages of development including HEV formation, lymphoid chemokine expression, and proper tissue microarchitecture.
Given the gross similarities between IKK␣ AA , NF-B2 Ϫ/Ϫ , and RelB Ϫ/Ϫ mice, a previously identified role for IKK␣ in LT␤R activation of the alternative NF-B pathway, and the normal activation of the canonical NF-B pathway by TNF-␣ in IKK␣ AA mice (8,30), it is likely that the secondary lymphoid organ defects in IKK␣ AA mice are attributable to abrogated activation of the alternative NF-B pathway. However, the fact that LNs are not completely eliminated in IKK␣ AA , NF-B2 Ϫ/Ϫ , and RelB Ϫ/Ϫ mice suggests that components of the canonical pathway are also involved in this process. For example, molecules such as VCAM-1 and mucosal addressin cell adhesion molecule 1, induced by both TNFRI and LT␤R through the canonical pathway, are also crucial for normal LN development (7). In addition, both the LT␣:TNFRI and LT␣␤:LT␤R ligand:receptor pairs regulate lymphoid chemokine expression (8,18). Therefore, while both NF-B pathways differentially regulate gene expression, their activities likely cooperate in secondary lymphoid organogenesis and HEV development.