Epithelial NEMO links innate immunity to chronic
Arianna Nenci1,2*, Christoph Becker3*, Andy Wullaert1, Ralph Gareus1, Geert van Loo2, Silvio Danese4,
Marion Huth2, Alexei Nikolaev3, Clemens Neufert3, Blair Madison5, Deborah Gumucio5, Markus F. Neurath3*
& Manolis Pasparakis1,2
Deregulation of intestinal immune responses seems to have a
principal function in the pathogenesis of inflammatory bowel
disease1–4. The gut epithelium is critically involved in the main-
tenance of intestinal immune homeostasis—acting as a physical
barrier separating luminal bacteria and immune cells, and also
expressing antimicrobial peptides3,5,6. However, the molecular
mechanisms that control this function of gut epithelial cells are
in gut epithelial cells to control epithelial integrity and the inter-
action between the mucosal immune system and gut microflora.
Intestinal epithelial-cell-specific inhibition of NF-kB through
conditional ablation of NEMO (also called IkB kinase-c (IKKc))
or both IKK1 (IKKa) and IKK2 (IKKb)—IKK subunits essential
for NF-kB activation7–9—spontaneously caused severe chronic
of colonic epithelial cells, impaired expression of antimicrobial
peptides and translocation of bacteria into the mucosa. Con-
currently, this epithelial defect triggered a chronic inflammatory
response in the colon, initially dominated by innate immune cells
but later also involving T lymphocytes. Deficiency of the gene
encoding the adaptor protein MyD88 prevented the development
of intestinal inflammation, demonstrating that Toll-like receptor
activation by intestinal bacteria is essential for disease pathogene-
sis in this mouse model. Furthermore, NEMO deficiency sensi-
tized epithelial cells to tumour-necrosis factor (TNF)-induced
apoptosis, whereas TNF receptor-1 inactivation inhibited intest-
inal inflammation, demonstrating that TNF receptor-1 signalling
is crucial for disease induction. These findings demonstrate that a
primary NF-kB signalling defect in intestinal epithelial cells dis-
inflammatory-bowel-disease-like phenotype. Our results identify
NF-kB signalling in the gut epithelium as a critical regulator of
epithelial integrity and intestinal immune homeostasis, and have
important implications for understanding the mechanisms con-
trolling the pathogenesis of human inflammatory bowel disease.
To investigate the function of NF-kB signalling in the gut epithe-
lium in vivo, we generated mice lacking NEMO specifically in intest-
inal epithelial cells (NEMOIEC-KOmice) by crossing mice carrying
loxP-flanked (floxed, FL) NEMO (Ikbkg) alleles10with villin-Cre
transgenics11. Southern blot DNA analysis showed NEMO deletion
specifically in the small intestine and colon of NEMOIEC-KOmice
(Fig. 1b). Furthermore, immunostaining showed lack of NEMO
expression in most epithelial cells from the small intestine and colon
(Fig. 1a), indicating efficient NEMO ablation in intestinal epithelial
NEMOIEC-KOmice were born at the expected mendelian ratio;
however, at a young age they showed runting, diarrhoea and rectal
bleeding, indicating intestinal pathology. High-resolution mini-
endoscopy12revealed signs of severe colitis with thickening of the
bowel wall, a granular and bleeding mucosa, loss of regular vessel
architecture and diarrhoea in NEMOIEC-KOmice (Fig. 1c). Further-
more, macroscopic examination showed severe pancolitis affecting
all parts of the colon distal from the caecum but sparing the small
intestine (Fig. 1d). The colitis phenotype was observed in all
ther they were maintained in a specific pathogen-free or a conven-
tional animal facility. Macroscopic and histological examination of
6-week-old NEMOIEC-KOmice did not reveal any pathological find-
ings in the small intestine, whereas the colon was heavily inflamed
(Supplementary Fig. 1 and data not shown). Colitis in NEMOIEC-KO
mice was characterized by thickening of the mucosa, enlarged crypts
with loss of goblet cells, and a marked infiltration of mononuclear
cells into the mucosa and submucosa, similar to the histopathology
seen in patients with inflammatory bowel diseases (Fig. 1e and Sup-
plementary Fig. 2). Analysis of pro-inflammatory gene expression
showed upregulation of Il1b, Il6, Tnf and Ccl2 (also called MCP-1)
in the colon of NEMOIEC-KOmice already at 2weeks after birth
(Fig. 2a), which was even more pronounced in 6-week-old animals,
when Ccl5 (RANTES) was also induced. To investigate which cells
were responsible for the inflammatory response in the colon of
NEMOIEC-KOmice, we examined the cellular infiltrate within the
mucosa. In 2-week-old mice, inflammatory infiltrates were domi-
nated by large numbers of dendritic cells and granulocytes, and to
a much lesser extent by CD41T cells (Fig. 2b). However, at 12 and
36weeks, colonic tissue from NEMOIEC-KOmice showed numerous
lymphoid follicles and a massive infiltration with dendritic cells,
CD41T cells and granulocytes (Fig. 2b). Thus, intestinal epithelial-
cell-specific NEMO ablation causes the spontaneous development of
severe chronic colitis associated with increased expression of inflam-
matory cytokines and chemokines, an early accumulation of innate
immune cells, and the presence of T-cell infiltrates at later time
NEMO-deficient cells lack the ability to activate NF-kB and are
sensitive to apoptosis induced by cytokines such as TNF10,13,14.
Indeed, TdT-mediated dUTP nick end labelling (TUNEL) staining
of colon cross-sections from 2-week-old mice revealed extensive epi-
*These authors contributed equally to this work.
1Institute for Genetics, University of Cologne, Zu ¨lpicher Strasse 47, 50674 Cologne, Germany.2EMBL Mouse Biology Unit, I-00016 Monterotondo, Italy.3Laboratory of Clinical
Gastroenterology, Viale Manzoni 56, 20089 Rozzano, Milan, Italy.5Department of Cell & Developmental Biology, Center for Organogenesis, The University of Michigan, Ann Arbor,
Michigan 48109-0616, USA.
Vol 446|29 March 2007|doi:10.1038/nature05698
not in controls (Fig. 3a). Furthermore, histological analysis of colon
sections from these mice revealed the presence of crypts showing
extensive epithelial destruction with loss of epithelial cells in such
occurs in young NEMOIEC-KOmice, which probably allows for the
translocation of bacteria from the lumen into the mucosa. Con-
sistently, fluorescence in situ hybridization (FISH) using a probe
specific for the ribosomal RNA of all eubacteria showed the presence
of bacteria within the lamina propria of NEMOIEC-KOmice, in close
proximity to areas displaying defects in epithelial integrity (Fig. 3c).
presence of free bacterial rods in the mucosa, cells in these areas also
stained positive for bacterial rRNA, indicating phagocytosis of bac-
teria by immune cells (Fig. 3d). Immunofluorescence analysis
revealed an accumulation of CD11c1dendritic cells, neutrophils
and a few CD41T cells around areas showing local disruption to
epithelial integrity, suggesting that bacterial translocation into the
mucosa leads to immune cell recruitment and activation (Fig. 3e).
Quantification of apoptotic cells (TUNEL1), cells showing bacterial
phagocytosis (FISH1) and of neutrophils (MPO1) in colon sections
from newborn, 1-, 2- and 3-week-old mice showed that disruption
of epithelial integrity, bacterial translocation into the mucosa and
recruitment of neutrophils are first observed in 1-week-old mice and
the pathology develops progressively over time (Supplementary Fig.
3). Taken together, our data are consistent with a model in which
intestinal epithelial-cell-specific NEMO deficiency compromises
colonic epithelial integrity, resulting in translocation of luminal bac-
inflammatory response is initiated, characterized initially by activa-
tion of innate immune cells such as dendritic cells and neutrophils.
Spleen SkinSmall intestine
Endoscopic colitis score
Figure 1 | Intestinal epithelium-specific NEMO ablation causes severe
spontaneous colitis. a, Immunofluorescence with anti-NEMO antibodies
shows efficient NEMO ablation in the intestinal epithelium of NEMOIEC-KO
mice. WT, wild type. b, Southern blot DNA analysis shows NEMO deletion
(Del) specifically in the intestines of NEMOIEC-KOmice. c, NEMOIEC-KO
(n53) and wild-type (n55) mice were examined at the age of 32–36weeks
using mini-endoscopy. Murine endoscopic index of colitis severity
is thickened and shortened, indicating severe colitis. e, Haematoxylin-and-
eosin-stained colon cross-sections show severe inflammation and loss of
goblet cells in NEMOIEC-KOmice. All scale bars, 50mm.
36 weeks old
12 weeks old
2 weeks old
2 weeks old
6 weeks old
Figure 2 | Inflammation in the colon of NEMOIEC-KOmice. a, Expression of
pro-inflammatory cytokines and chemokines in the colons of NEMOIEC-KO
and wild-type mice (n54–6 mice per group). Data shown are mean values
relative to ubiquitin(6s.e.m.); double asterisk, P,0.01. b, Colon cross-
sections from NEMOIEC-KOand wild-type mice were immunostained with
antibodies for CD4, CD11c or myeloperoxidase (MPO) (red). Nuclei were
counterstained with Hoechst 3342 (blue). Scale bars, 100mm.
NATURE|Vol 446|29 March 2007
This pathology starts in a localized fashion in scattered areas of the
At later time points, the inflammation affects the entire colon and
also involves T lymphocytes. In mice with advanced-stage colitis,
we could detect lymphoid follicles surrounded by dendritic cells,
whereas CD11c/FISH double staining revealed bacterial rRNA
signals co-localized with dendritic cells in these areas (Fig. 3f),
indicating that the presence of bacteria drives the immune response
also at later stages.
Antimicrobial peptide expression by the intestinal epithelium is
believed to have an important function in controlling the number of
bacteria around epithelial cells, and has recently been suggested to
have a principal role in the pathogenesis of inflammatory bowel
disease6. Because the NF-kB signalling pathway regulates antimicro-
bial peptide expression6, we investigated whether NEMO deficiency
affected antimicrobial peptide expression from the colonic epithe-
lium. Using a primer set that detects simultaneously all 17 mouse
only mildly reduced in 2-week-old NEMOIEC-KOmice compared to
controls, and was unaffected at 6weeks of age (Fig. 3g). However,
beta-defensin-3, the expression of which is induced in mouse prim-
ary colonic epithelial cells by bacteria16, was significantly downregu-
lated in 2- and 6-week-old NEMOIEC-KOmice compared with
homologous to mouse beta-defensin-3, predisposes to colonic
Crohn’s disease in humans17. Thus, reduced expression of defen-
sin-3 may lead to impaired antibacterial defence in the colon, and
might be implicated in the translocation of commensal bacteria into
Consistent with a previous report18, mice withintestinal epithelial-
Cre transgenic mice with mice carrying loxP-flanked IKK2 (ikbkb)
alleles19, did not develop spontaneous intestinal inflammation
(Supplementary Fig. 5). The differential effect of NEMO versus
IKK2 ablation in the gut epithelium is probably due to the fact that
pro-inflammatory-signal-induced NF-kB activity is completely
blocked in NEMO-deficient cells but only partially affected in
IKK2-deficient cells10,20. Indeed, NEMO-deficient intestinal epithelial
cells displayed a complete lack of NF-kB activation, whereas IKK2-
deficient intestinal epithelial cells displayed low levels of residual
NF-kB activity (Fig. 4a and Supplementary Fig. 6). Furthermore,
NEMO-deficient intestinal epithelial cells failed to upregulate NF-
kB-dependent genes, such as IkBa (also called Nfkbia) and IP10
(Cxcl10), upon in vivo lipopolysaccharide stimulation, whereas
IKK2-deficient intestinal epithelial cells showed intermediate levels
of induction compared to wild-type cells (Fig. 4b). Thus, NEMO
deficiency completely blocks canonical NF-kB activity in intestinal
epithelial cells, whereas IKK2-deficient cells show residual NF-kB
activation, presumably through compensatory signalling by means
of IKK1. To test whether IKK1 signalling protects the colon of
IKK2IEC-KOmice from intestinal inflammation, we generated mice
lacking both IKK1 and IKK2 in intestinal epithelial cells (IKK1/
2IEC-KO) by crossing the IKK2IEC-KOmice with mice carrying IKK1
(Chuk) loxP-flanked alleles21. Whereas intestinal epithelial-cell-
ble IKK1/2IEC-KOmice displayed diarrhoea and rectal bleeding, as well
as macroscopic, endoscopic and histological evidence of colitis (Fig.
4c–e and Supplementary Fig. 7). Thus, although neither IKK1IEC-KO
nor IKK2IEC-KOmice show spontaneous intestinal inflammation,
IKK1/2IEC-KOmice develop severe colitis, similarly to NEMOIEC-KO
mice. These results demonstrate that complete inhibition of canonical
NF-kB activity in intestinal epithelial cells—achieved by ablation of
colitis, and reveal that canonical IKK/NF-kB signalling in intestinal
epithelial cells has an essential function in maintaining epithelial inte-
grity and immune homeostasis in the colon. Moreover, our findings
uncover a redundancy between the two IkB kinases in mediating this
Thepresenceofbacteriawithin thelaminapropria maytriggerthe
development of chronic inflammatory colitis in NEMOIEC-KOmice
by means of Toll-like receptor (TLR) stimulation5. To investigate
whether TLR activation contributes to the development of colitis
1 week old 2 weeks old
CD11c / FISH
FISH sense control
Figure 3 | Compromised epithelial integrity and bacterial translocation in
the colon of NEMOIEC-KOmice. a, TUNEL staining (green) on colon
cross-sections from 2-week-old NEMOIEC-KOand wild-type mice.
b, Haematoxylin-and-eosin-stained colon cross-sections from 3-week-old
NEMOIEC-KO(KO) and wild-type mice. Note the epithelial disintegration in
several crypts. The arrowhead depicts two remaining epithelial cells in a
crypt showing nearly complete loss of epithelium. c, Bacterial FISH staining
(green) was detected in the lamina propria of NEMOIEC-KOcolon
(arrowheads), but was restricted to the lumen in wild-type mice (arrows).
d, Bacterial FISH staining (red) co-localizes with cells in two representative
and dendritic cells in the lamina propria of young NEMOIEC-KOmice.
f, Double FISH/CD11c staining on colon cross-sections from 36-week-old
NEMOIEC-KOmice shows co-localization of dendritic cells with bacteria
around a lymphoid follicle. g, Expression of cryptdins and of defensin-3 in
are shown as mean values relative to ubiquitin (6s.e.m.); asterisk, P,0.05;
double asterisk, P,0.01. All scale bars, 50mm.
NATURE|Vol 446|29 March 2007
adaptor molecule required for signalling by most members of the
gical analysis of colons from NEMOIEC-KO/MyD88-deficient mice
of age) (Fig. 4f), demonstrating that MyD88-dependent signals are
essential for disease pathogenesis in this model. These results suggest
that TLR-mediated bacterial recognition has a critical role in indu-
cing intestinal inflammation in NEMOIEC-KOmice; however, IL-1-
mediated signals that also depend on MyD88 might also be involved
in the pathogenesis of colitis.
TNF blockade is an effective therapy for inflammatory bowel dis-
ease patients23. To investigate whether TNF signalling contributes to
the development of colitis in NEMOIEC-KOmice, we crossed them
onto a TNF receptor-1 (TNFRI)-deficient background. Double
NEMOIEC-KO/TNFRI-deficient mice did not show macroscopic or
of age), demonstrating that TNFRI signalling is crucial for disease
pathogenesis in this model (Fig. 4g). Because cells lacking NEMO
are more sensitive to TNF-induced apoptosis than cells lacking
IKK2 (refs 10, 24), TNF-mediated killing of NEMO-deficient intest-
inal epithelial cells probably contributes to the induction of colon
inflammation by compromising epithelial integrity in NEMOIEC-KO
mice. Indeed, TNF administration caused increased apoptosis of
by crossing with TNF-deficient mice25and used here to facilitate
assessment of apoptosis in a disease-free colon—compared with
TNF-deficient, IKK2IEC-KOand wild-type mice (Supplementary Fig.
8), demonstrating that NEMO-deficient intestinal epithelial cells are
more sensitive to TNF-induced killing in vivo than IKK2-deficient
intestinal epithelial cells. These results suggest that one key mech-
anism by which TNF signalling induces colon inflammation in
NEMOIEC-KOmice isbykilling theNF-kB-deficientintestinalepithe-
lial cells, resulting in disruption of epithelial integrity. However, TNF
Taken together, our findings support a model (Fig. 5) in which
impaired NF-kB signalling in the colonic epithelium results in TNF-
mediated epithelial cell apoptosis and diminished expression of
antimicrobial peptides, thus compromising epithelial integrity and
allowing bacterial translocation into the mucosa of NEMOIEC-KO
mice. The presence of bacteria within the colonic mucosa activates
innate immune cells through MyD88-dependent TLR signalling,
triggering the expression of pro-inflammatory cytokines such as
TNF and IL-1b, which cause further destruction of the NF-kB-
deficient colonic epithelium and induce the recruitment of addi-
tional immune cells including T-lymphocytes, finally resulting in
Impairment of intestinal epithelial integrity and reduced crypt-
din expression in humans have been suggested as important cau-
sative factors for the development of intestinal inflammation in
patients with inflammatory bowel disease5,6. Our results with the
NEMOIEC-KOmice demonstrate that a primary epithelial defect
resulting in impairment of both the integrity and antimicrobial
defence of the gut epithelium can give rise to chronic intestinal
inflammation. The rapid onset and high penetrance of the colitis
phenotype in NEMOIEC-KOmice suggest that a defect in epithelial
NF-kB signalling may be an important proximal event in the patho-
genesis of inflammatory bowel disease. Thus, epithelial NF-kB defi-
ciency may provide a new paradigm for the development of an
inflammatory-bowel-disease-like disease. Furthermore, NF-kB sig-
nalling in intestinal epithelial cells emerges as a crucial factor for
Chronic intestinal inflammation
Impaired defensin expression
Figure 5 | NEMO-mediated NF-kB activation in intestinal epithelial cells is
essential for the maintenance of epithelial barrier integrity and immune
homeostasis in the colon. DC, dendritic cell; MW, macrophage; PAMPS,
pathogen-associated molecular patterns.
LPS – + –
– + – +– +
– +– +– +
MPO MPOCD11c CD11c
Figure 4 | NF-kB inhibition in intestinal epithelial cells causes
inflammatory colitis that depends on TNFRI and MyD88 signalling. a,
b, Intestinal epithelial cells were isolated from mice with the indicated
10g body weight) or PBS. a, Immunoblot analysis of IkBa phosphorylation
and degradation, and NEMO, IKK2, villin and actin expression.
b, Quantitative RT–PCR analysis of IkBa and IP10 expression. Mean
examination (d) reveal severe colitis in 11-week-old double IKK1/2IEC-KO
mice but not in IKK1IEC-KOmice. e, Colon cross-sections from 11-week-old
IKK1/2IEC-KOmice were stained with haematoxylin and eosin or with
antibodies for CD4, CD11c or MPO (red). Nuclei were counterstained with
Hoechst 3342 (blue). f, g, Macroscopic and histological examination of
colons from 8-week-old NEMOIEC-KO/MyD88KO(f) and NEMOIEC-KO/
strains. All scale bars, 200mm.
NATURE|Vol 446|29 March 2007
maintaining homeostasis between commensal microflora and host
minant for prevention of gut inflammation.
Mice. The mouse lines used in this study have been described previously:
NEMOFL(ref. 10), IKK2FL(ref. 19), villin-Cre11, TNFRI-deficient26, MyD88-
deficient27, TNF-deficient25, IKK1FL(ref. 21). The experiments were performed
on mice backcrossed into the C57BL/6 genetic background for at least five gen-
erations. In all experiments, littermates carrying the loxP-flanked alleles but not
expressing Cre recombinase were used as wild-type controls. The mice used in
this study were housed in individually ventilated cage systems at the EMBL
Mouse Biology Unit or at the Institute for Genetics of the University of
Cologne, in either specific pathogen-free or conventional animal facilities.
Immunohistochemistry, FISH and apoptosis. Immunofluorescence of cryo-
sections was performed usingthe TSA Cy3 system (PerkinElmer) and a fluores-
cence microscope (Olympus fluorescence microscope) as described28. In situ
hybridization of bacterial rRNA on glass slides was performed as previously
described28. Apoptosis was analysed using the in situ cell death detection kit
Endoscopic procedures. For endoscopic colitis monitoring, a high-resolution
mouse video endoscope, denoted Coloview, was used, as previously described29.
Southern blot analysis. Southern blot analysis was performed as described
Isolation of intestinal epithelial cells. Intestinal epithelial cells were isolated
from small intestines or colons (see Supplementary Information).
Immunoblot analysis. Total cell extracts were electrophoresed by SDS–
branes (Millipore). Membranes were probed with the antibodies listed in the
Quantitative real-time PCR. Total RNA from colon, small intestine and intest-
inal epithelial cells was isolated using Trizol (Invitrogen) and analysed by real-
Biosystems). Individual primer sequences are listed in the Supplementary
Information. All values were normalized to the level of ubiquitin messenger
Received 8 December 2006; accepted 23 February 2007.
Published online 14 March 2007.
1.Bouma, G. & Strober, W. The immunological and genetic basis of inflammatory
bowel disease. Nature Rev. Immunol. 3, 521–533 (2003).
Hanauer, S. B. Inflammatory bowel disease: epidemiology, pathogenesis, and
therapeutic opportunities. Inflamm. Bowel Dis. 12 (suppl. 1), S3–S9 (2006).
Macdonald, T. T. & Monteleone, G. Immunity, inflammation, and allergy in the
gut. Science 307, 1920–1925 (2005).
Neurath, M. F., Finotto, S. & Glimcher, L. H. The role of Th1/Th2 polarization in
mucosal immunity. Nature Med. 8, 567–573 (2002).
J. Immunol. 174, 4453–4460 (2005).
Wehkamp, J., Fellermann, K., Herrlinger, K. R., Bevins, C. L. & Stange, E. F.
Mechanisms of disease: defensins in gastrointestinal diseases. Nature Clin. Pract.
Gastroenterol. Hepatol. 2, 406–415 (2005).
Karin, M. & Greten, F. R. NF-kB: linking inflammation and immunity to cancer
development and progression. Nature Rev. Immunol. 5, 749–759 (2005).
2, 725–734 (2002).
Ghosh, S. & Karin, M. Missing pieces in the NF-kB puzzle. Cell 109 (suppl.)
10. Schmidt-Supprian, M. et al. NEMO/IKKc-deficient mice model incontinentia
pigmenti. Mol. Cell 5, 981–992 (2000).
11.Madison, B.B.etal. Ciselementsofthevillingenecontrolexpressioninrestricted
domains of the vertical (crypt) and horizontal (duodenum, cecum) axes of the
intestine. J. Biol. Chem. 277, 33275–33283 (2002).
of IL-6 trans-signaling. Immunity 21, 491–501 (2004).
13. Rudolph, D. et al. Severe liver degeneration and lack of NF-kB activation in
NEMO/IKKc-deficient mice. Genes Dev. 14, 854–862 (2000).
14. Makris, C. et al. Female mice heterozygous for IKKc/NEMO deficiencies develop
a dermatopathy similar to the human X-linked disorder incontinentia pigmenti.
Mol. Cell 5, 969–979 (2000).
15. Darmoul, D. & Ouellette, A. J. Positional specificity of defensin gene expression
reveals Paneth cell heterogeneity in mouse small intestine. Am. J. Physiol. 271,
16. Lan, J. G. et al. Different cytokine response of primary colonic epithelial cells to
commensal bacteria. World J. Gastroenterol. 11, 3375–3384 (2005).
b-defensin 2 gene copy number predisposes to Crohn disease of thecolon. Am. J.
Hum. Genet. 79, 439–448 (2006).
colitis-associated cancer. Cell 118, 285–296 (2004).
19. Pasparakis, M. et al. TNF-mediated inflammatory skin disease in mice with
epidermis-specific deletion of IKK2. Nature 417, 861–866 (2002).
20. Li, Q., Estepa, G., Memet, S., Israel, A. & Verma, I. M. Complete lack of NF-kB
activity in IKK1 and IKK2 double-deficient mice: additional defect in neurulation.
Genes Dev. 14, 1729–1733 (2000).
21. Gareus, R. et al. Normal epidermal differentiation but impaired skin barrier
formation upon keratinocyte-restricted IKK1 ablation. Nature Cell Biol.
doi:10.1038/ncb1560 (in the press).
22. Kawai, T. & Akira, S. TLR signaling. Cell Death Differ. 13, 816–825 (2006).
23. Targan, S. R. et al. A short-term study of chimeric monoclonal antibody cA2 to
tumor necrosis factor a for Crohn’s disease. Crohn’s Disease cA2 Study Group.
N. Engl. J. Med. 337, 1029–1035 (1997).
24. Luedde, T. et al. Deletion of NEMO/IKKc in liver parenchymal cells causes
steatohepatitis and hepatocellular carcinoma. Cancer Cell 11, 119–132 (2007).
25. Pasparakis, M., Alexopoulou, L., Episkopou, V. & Kollias, G. Immune and
inflammatory responses in TNFa-deficient mice: a critical requirement for TNFa
in the formation of primary B cell follicles, follicular dendritic cell networks and
germinal centers, and in the maturation of the humoral immune response. J. Exp.
Med. 184, 1397–1411 (1996).
26. Pfeffer, K. et al. Mice deficient for the 55 kd tumor necrosis factor receptor are
resistant to endotoxic shock, yet succumb to L. monocytogenes infection. Cell 73,
27. Kawai, T., Adachi, O., Ogawa, T., Takeda, K. & Akira, S. Unresponsiveness of
MyD88-deficient mice to endotoxin. Immunity 11, 115–122 (1999).
28. Becker, C. et al. Constitutive p40 promoter activation and IL-23 production in the
terminal ileum mediated by dendritic cells. J. Clin. Invest. 112, 693–706 (2003).
29. Becker, C. et al. In vivo imaging of colitis and colon cancer development in mice
using high resolution chromoendoscopy. Gut 54, 950–954 (2005).
Supplementary Information is linked to the online version of the paper at
and MyD88-deficient mice, respectively. This work was supported by grants from
the German Research Council to C.B. and M.F.N., and by EU-FP6 grants MUGEN
and IMDEMI to M.P.; A.W. received a research fellowship from the Alexander von
Author Contributions A. Nenci, C.B., A.W., R.G, G.v.L, M.F.N and M.P. designed
the research. A. Nenci, C.B, A.W, R.G, G.v.L, S.D., M.H., A. Nikolaev and C.N.
performed the research. B.M. and D.G. contributed new reagents. A. Nenci, C.B.,
A.W., R.G, M.F.N and M.P. analysed the data and wrote the paper. M.F.N. and M.P
share senior authorship.
Author Information Reprints and permissions information is available at
www.nature.com/reprints. The authors declare no competing financial interests.
Correspondence and requests for materials should be addressed to M.P.
NATURE|Vol 446|29 March 2007