TRAF6-dependent NF-kB transcriptional activity during mouse development



Nuclear factor-kappa B (NF-kB) transcriptional activity is induced by numerous stimuli. To identify tissues exhibiting NF-kB transcriptional activity during development, we analyzed transgenic reporter mice that express β-galactosidase from an NF-kB–responsive element. We report that NF-kB activation is widespread and present in numerous epithelial structures and within vasculature. Several regions of the developing central nervous system, including the roof plate and floor plate of the midbrain, show prominent NF-kB activation. To assess the role of the TRAF6 adaptor protein in developmental NF-kB activity, we analyzed NF-kB activation in reporter mice rendered null for TRAF6. Deletion of TRAF6 resulted in the loss of NF-kB activity in epithelia, in vasculature, and in roof and floor plate but had no effect on NF-kB activity developing telencephalon, choroid plexus, cochlear canal, and thymus. These data indicate that NF-kB transcriptional activity is present in a broad range of structures during development and that TRAF6 plays a critical role mediating developmental NF-kB activation in many but not all tissues. Developmental Dynamics 231:122–127, 2004. © 2004 Wiley-Liss, Inc.


Nuclear factor-kappa B (NF-kB) transcription factors are required during development and are necessary for inflammatory and immune responses in adults. The five mammalian NF-kB subunits (RelA (p65), NF-kB1 (p52/p100), NF-kB2 (p50/p105), RelB, and c-Rel) each contain a Rel homology domain that allows these factors to dimerize and bind DNA. NF-kB dimers are normally rendered inactive in the cytosol by virtue of their interaction with one of the inhibitory IkB proteins (IkBa, IkBb, IkBe, IkBg, Bcl-3). Transcription of NF-kB target genes typically occurs after stimuli-induced IkB protein degradation and translocation of NF-kB subunits to the nucleus (Gilmore, 1999; Karin and Ben-Neriah, 2000; Ghosh and Karin, 2002).

Several members of the TNF receptor family, interleukin receptor family, and Toll receptor family are potent inducers of NF-kB activity (Bradley and Pober, 2001; Chung et al., 2002). For each receptor class, ligand binding induces conformational changes that allow the receptors to complex with TRAF proteins, key intermediaries in the activation of NF-kB and Jun kinase signalling pathways (Bradley and Pober, 2001). Several receptors directly bind TRAF proteins, whereas others bind TRAFs through adaptors such as Myd88. Disruption of TRAF2, TRAF3, TRAF4, or TRAF5 alleles in mice results in specific defects in immune cell or respiratory system function (Chung et al., 2002), whereas loss of TRAF6 produces broader and more severe phenotypes that include osteopetrosis, lymph node agenesis, hypohidrotic ectodermal dysplasia, and exencephaly (Lomaga et al., 1999, 2000; Naito et al., 1999, 2002). The severe phenotype that results from TRAF6 deletion is consistent with its ability to bind multiple members of the TNF receptor, interleukin receptor, and Toll receptor families (O'Neill, 2002).

We previously produced a transgenic reporter mouse line in which the production of β-galactosidase (LacZ) is regulated by NF-kB activity in vivo (Bhakar et al., 2002). Our earlier work on this line examined NF-kB activity within the developing brain and demonstrated an important role for NF-kB in survival of developing cortical neurons. Here, we use this line to identify developing tissues that display TRAF6-dependent and -independent NF-kB activity.


TRAF6 Deletion Reduces LacZ Expression in a Transgenic NF-kB Reporter Line

The NF-kB reporter line (NF-kBREP) used in these studies produces LacZ tagged with a nuclear localization sequence from an artificial promoter containing three tandem repeats of an NF-kB element derived from the HIV long terminal repeat. In vitro validation studies on mouse embryonic fibroblasts and neurons have demonstrated that LacZ expression in cells derived from these animals is NF-kB–dependent (Bhakar et al., 2002). To begin to identify signaling components that contribute to developmental NF-kB expression in vivo, we examined the expression pattern of LacZ in NF-kBREP animals that lacked the TRAF6 adaptor protein. Figure 1a shows whole-mount staining of a wild-type (TRAF6+/+) embryonic day (E) 13 NF-kBREP mouse and demonstrates that LacZ expression is prominent in the telencephalon, within vibrissae, in mammary buds, and in lacrimal glands. Figure 1b shows a TRAF6−/− NF-kBREP littermate that lacks both TRAF6 alleles. As previously reported, TRAF6 null embryos show profound exencephaly due to hyperplasia of the midbrain and hindbrain (Lomaga et al., 2000). Of interest, LacZ expression in the TRAF6−/− animals is lost in vibrissae, mammary glands, and lacrimal glands yet is maintained in the telencephalon. These data indicate a crucial role for TRAF6 in developmental NF-kB activity in many but not all developing tissues.

Figure 1.

Loss of TRAF6 expression ablates most but not all embryonic NF-kB reporter gene expression. Nuclear factor-kappa B (NF-kB)REP littermates that were TRAF6+/+ or TRAF6−/− were analyzed by LacZ whole-mount staining. a,b TRAF6+/+ animals (a) have prominent staining in vibrissae (V), lacrimal glands (L), and mammary glands (M), which is lost in the TRAF6-deficient background (b). The telencephalon (T) retains LacZ expression in the TRAF6−/− animals. Original magnification, ×2.5. Scale bar = 1 mm in b (applies to a,b).

NF-kB Activation in Developing Epidermal Tissues Is TRAF6-Dependent

To identify specific regions of TRAF6-dependent NF-kB activation in the developing mouse embryo, embryos were sectioned and assayed for LacZ activity. Consistent with the whole-mount staining shown in Figure 1, prominent LacZ activity was observed in sections of vibrissae, in mammary buds, and in lacrimal glands from TRAF6+/+ NF-kBREP embryos but was absent in these tissues in TRAF6−/− NF-kBREP littermates (Fig. 2b,e,h).

Figure 2.

a–o: Nuclear factor-kappa B (NF-kB) activity in the developing epithelial primordium requires TRAF6. a–l: LacZ expression in NF-kBREP littermates that were TRAF6+/− or TRAF6−/− was evaluated in vibrissae (a–c), mammary buds (d–f), lacrimal glands (g–i), and tooth buds (j–l). LacZ expression was prominent in epithelia-derived tissues in TRAF6+/− tissues, including vibrissae (a), mammary buds (d), lacrimal gland (g), and tooth bud (j), but was absent in these epithelia in TRAF6 null mice (b,e,h,k). m,n: LacZ activity was also apparent in the epithelium of the esophagus (m; arrow) and within the laryngeal additus (m; arrowhead) but was absent in TRAF6−/− animals (n). LacZ activity was also detected in the arytenoid swelling and was present in both TRAF+/− and TRAF6−/− animals (m and n; asterisks). Sections in c, f, i, l, o are adjacent section to those shown in b, e, h, k, and n but were stained with hematoxylin/eosin. Scale bars = 50 μm in a–o.

TRAF6 null animals show impairments both in tooth development and long bone formation (Lomaga et al., 1999; Naito et al., 1999; Armstrong et al., 2002). NF-kB activity was not detected in chondrocytes during prebone formation in the TRAF6+/+ NF-kBREP line (data not shown), but the tooth bud showed prominent LacZ expression in TRAF6+/+ NF-kBREP embryos (Fig. 2j). Reporter gene activity was completely absent in these tissues in Traf6 null animals (Fig. 2k). Esophageal and laryngeal epithelium also displayed high levels of LacZ expression in NF-kBREP embryos (Fig. 2m), which was dependent on the presence of TRAF6 (Fig. 2n).

Although deletion of TRAF6 resulted in loss of NF-kB activity in the epithelial tissues, the TRAF6 null tissues appear morphologically normal at this stage of development, indicating that loss of LacZ expression reflects reduced NF-kB activity and is not secondary to cell depletion in the null animals (compare adjacent panels in Fig. 2). These data indicate that NF-kB is prominent in developing epidermal structures and reveal that TRAF6 is required for NF-kB activation in these tissues during development.

Vascular NF-kB Expression Is Regulated by TRAF6

Assessment of NF-kB expression in the vasculature of the NF-kBREP line revealed that LacZ expression was present in essentially all arterial vessels within the TRAF6+/+ NF-kBREP embryos, including intercostal (Fig. 3a, arrow) and basilar (Fig. 3c) arteries. The descending aorta (Fig. 3a, arrowheads) also showed LacZ expression, but levels were reduced relative to the smaller vessels. Although arterial morphology is normal in TRAF6−/− NF-kBREP animals, TRAF6 deletion results in complete loss of LacZ expression (Fig. 3b,d), indicating that developmental activation of NF-kB in these vessels is TRAF6-dependent. Venous structures showed minimal or no LacZ expression, irrespective of TRAF6 genotype (data not shown). Thus, developmental NF-kB activation in the vasculature is limited to arterial vessels and dependent on TRAF6 expression.

Figure 3.

Nuclear factor-kappa B (NF-kB) expression is present in developing arterial vasculature and requires TRAF6. a,c: LacZ expression is shown in an intercostal artery (a, arrow), descending aorta (a, arrowheads) and in the basilar artery (c, arrow). b,d: The abundant LacZ activity in the basilar and intercostal arteries is lost in TRAF6 null animals (b,d). a,b: Lower but readily detectable levels of LacZ activity were present in the aorta of TRAF6+/+ animals (a) but were absent from TRAF6−/− animals (b). Scale bars = 100 μm in a–d

NF-kB Activity in the Developing Central Nervous System

The NF-kBREP mouse displays high levels of LacZ expression in the developing telencephalon, thalamus, and corpus striatum. We have previously established that neurons are the primary cell type expressing LacZ in the central nervous system and have shown that LacZ expression in neurons reflects bona fide NF-kB activity (Bhakar et al., 2002). Of interest, LacZ expression in these areas is reduced but not lost n NF-kBREP embryos that lack TRAF6 (Fig. 4a–c), indicating that activation of neuronal NF-kB activity occurs through TRAF6-dependent and TRAF6-independent mechanisms.

Figure 4.

Nuclear factor-kappa B (NF-kB) is regulated by TRAF6 in the roof plate and floor plates. a–c: LacZ expression in the telencephalon is present in TRAF6+/+ animals (a) and reduced but not lost in TRAF6−/− littermates (c). d–h: LacZ expression is also present in roof plate (d–f) and floor plate (g,h) of NF-kBREP littermates that were TRAF6+/+ (d,g, arrow) but is absent in sections from TRAF6−/− animals (f,h). g,h: NF-kBREP animals that were TRAF6+/+ also display LacZ staining in the pineal primordium (g, arrowhead), which is absent in TRAF6−/− animals (h). Scale bars = 50 μm in a–c, 100 μm in d–h.

Intriguingly, the roof and floor plates of the developing midbrain also show prominent LacZ expression in the NF-kBREP mouse (Fig. 4d–g). The LacZ expression is well delimited and shows a sharp border relative to the surrounding tissue, raising the possibility that NF-kB activity has a role in patterning the neural tube. Figure 4f and h show that embryos deficient in TRAF6 show complete loss of LacZ expression from both these locations and from pineal primordium.

TRAF6 and Developmental NF-kB Expression

Although LacZ expression within exocrine and epithelial tissues was invariably lost in mice lacking TRAF6, LacZ expression remained in several tissues in TRAF6 null embryos. These areas include the choroid plexus of the fourth ventricle (Fig. 5a,b), the cochlear canals (Fig. 5c,d), and the thymus (Fig. 5e,f; Table 1). Together, these data demonstrate that NF-kB transcriptional activation occurs in a wide variety of structures during development and establishes that TRAF6 plays a crucial role in NF-kB activation in many but not all locations.

Figure 5.

TRAF6 independent nuclear factor-kappa B (NF-kB) expression. a–f: LacZ expression was detected in TRAF6+/+ (a,c,e) and TRAF6−/− (b,d,f) animals in the choroid plexus of the 4th ventricle (a,b, arrows), in the cochlear canal (c,d, arrows), and in the thymus (e,f, arrows) Scale bars = 100 μm in a–f.

Table 1. LacZ Expression in NF-kBREP Mice vs. NF-kBREP Mice Rendered Null For TRAF6 Expression
StructureNF-kBREPNF-kBREP/TRAF6 (−/−)
Floor plate++
Roof plate++
Corpus striatum++++
Choroid plexus (lateral ventricle)
Choroid plexus (4th ventricle)++++
Base of diencephalon (thalamus)++++
Basilar artery++
Spinal cord
Lacrimal glands++
Jacob's organ
Nasal serous glands++
Submandibular glands+
Submandibular ducts+
Circumvalate papillae+++
Entrance of esophagus++
Entrance of trachea
Tooth buds++
Mammary glands+++
Cochlear canal++++
Arytenoid swelling++++
Lung parenchyma
Liver parenchyma
Kidney parenchyma
Intestine parenchyma
Bladder parenchyma


Transgenic Mouse Maintenance and Genotyping

All mice were maintained on a pure C57/BL6 background. TRAF6 animals (Lomaga et al., 1999) were a kind gift of Dr. Tak Mak (University of Toronto). NF-kB transgenic reporter animals were genotyped as described (Bhakar et al., 2002) and TRAF6 mice were polymerase chain reaction genotyped by using TRAF6 primers (ACG GAA GCA AGC CTC TGT TCA TAC CG; CTG CAG TGA AAG ATG ACA GCG TGA GT; 400-bp product) and a primer directed against the neomycin resistance gene (CCA AGT GCC CAG CGG GGC TGC TAA AG; 350-bp product). PCR cycle parameters were 94°C for 1 min, followed by 40 cycles of 94°C for 30 sec, 65°C for 1 min, 72°C for 1 min, and a final 10-min extension at 72°C. For all studies, NF-kB transgenic reporter animals containing a single transgenic allele were used.

Whole-Mount and X-gal Staining

E13- to E15-day-old embryos were dissected, and tails were removed for genotyping. Whole-mount staining was performed as previously described (Bhakar et al., 2002). For X-gal detection within sections, embryos were fixed overnight in 2% PLP (2% paraformaldehyde, 74 mM L-lysine monohydrochloride, 10 mM sodium periodate in 0.1 M phosphate buffer, pH 7.4) and cryopreserved in 30% sucrose overnight. The 20-μm serial frozen sections were assayed for β-galactosidase activity by overnight incubation at 37°C in 80 mM dibasic sodium phosphate, 20 mM monobasic sodium phosphate, 2 mM MgCl2, 0.2% NP40, 1 mg/ml sodium deoxycholate, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, and 1 mg/ml X-gal (5-bromo-4-chloro-3-indolyl- β-D-galactoside; Invitrogen). Slides were then rinsed in PBS and counterstained in 20% eosin for 20 sec. After counterstaining, sections were dehydrated, cover-slipped in Permount, and visualized by using a Zeiss Axioskop 40 microscope under brightfield emission.


We thank Tak Mak and the Amgen Corporation for providing the TRAF6 null mice. A.B. was supported by a studentship from the CIHR and by a Tomlinson Fellowship. P.A.B. is Scientist of the Canadian Institute of Health Research and was funded by a grant from the Canadian Institute of Health Research.