Cutting Edge: Crohn’s Disease-Associated Nod2 Mutation Limits Production of Proinflammatory Cytokines To Protect the Host from Enterococcus faecalis-Induced Lethality

Nucleotide-binding oligomerization domain 2 (Nod2), a member of the Nod-like receptor family, is activated by muramyl dipeptide (MDP), a conserved structure from bacterial peptidoglycan (1, 2). Upon MDP recognition, Nod2 induces the activation of the transcription factor NF-κB and MAPKs, leading to production of proinflammatory and antimicrobial molecules (1–3). The importance of Nod2 is underscored by the observation that several genetic variants of Nod2 are associated with increased susceptibility to Crohn’s disease (CD) (4, 5). Three Nod2 variants, G908R, R702W, and a frameshift deletion mutation at L1007 (L1007fsinsC), have been linked to CD development (4, 5). Although the precise mechanisms by which Nod2 promotes disease remain unclear, CD-associated human Nod2 variants exhibit reduced capacity to activate NF-κB following MDP stimulation (2), suggesting that the loss of Nod2 activation promotes CD.

Population studies have revealed that the main CD-associated Nod2 variants are present in ∼20% of the healthy white individuals (6). Furthermore, there are marked differences in the allelic frequencies of the Nod2 variants among white populations and geographical regions in Europe (7). Because of these observations, it has been suggested that Nod2 variants linked to CD are under selection pressure possibly through bacterial infections (8). In this article, we show that Nod2 contributes to innate immune responses and regulates survival following infection with Enterococcus faecalis, an enteric commensal that is a common cause of systemic infection and lethality in humans (9). Using a newly developed mouse model in which the common CD-associated frameshift mutation was introduced into the Nod2 locus, we show that mice homozygous for the Nod2 mutation are impaired in MDP recognition and exhibit reduced susceptibility to systemic infection with E. faecalis.

C57BL/6 mice were raised in our animal facility from breeders originally purchased from The Jackson Laboratory (Bar Harbor, ME). Nod2^−/− mice on a C57BL/6 background have been reported (10). The construction of Nod2 Leu980fs-knock-in mice was described (11). Homozygous Nod2 Leu980fs-knock-in (Nod2^m/m) mice and heterozygous Nod2^+/m littermates on a mixed 129-C57BL/6 background were backcrossed six times to C57BL/6 mice before being used for in vivo experiments. Control wild-type mice were derived from crosses between heterozygous Nod2^+/m mice backcrossed six times to C57BL/6 mice. The animal studies were conducted under approved protocols by the University of Michigan Committee on Use and Care of Animals.

Bone marrow-derived macrophages (BMDMs) were prepared as previously described (12). Lipoteichoic acid (LTA) was from InvivoGen and MDP was from Bachem. E. faecalis strain OG1RF (ATCC 47077) was from the American Type Culture Collection. Macrophages were infected with bacteria for 60 min at 37°C and washed twice with PBS and IMDM containing 50 μg/ml gentamycin to limit the growth of extracellular bacteria.

Nod1 and Nod2 stimulation was determined as previously reported (2).

Mouse cytokines were measured using ELISA kits from R&D Systems.

Immunoblotting for mouse IκB-α, phospho–IκB-α, p38 and phospho-p38, phospho-ERK, and phospho-JNK (Cell Signaling Technology, Beverly, MA) was performed as described (10). To detect Nod2, protein extracts from mouse small intestines were immunoprecipitated using a rabbit polyclonal anti-Nod2 Ab (13), followed by immunoblotting with anti-Nod2 mAb (13). Binding was revealed by ECL (Pierce).

Total RNA extracted from BMDMs was reverse transcribed and the resulting cDNA was applied as a template for PCR with primers spanning Nod2 2939^insC mutation (forward, 5′-TTGAGTGTGCTCTTCGCTGT-3′; reverse, 5′-ACCAACCATCACGACTCCTC-3′). Amplified products were purified by QIAEXII (Qiagen) and their ends blunted with T4 DNA polymerase (Promega) and ligated into pCR-Blunt II-Topo (Invitrogen) for sequence analysis.

Adult mice (6–12 wk old) were administered MDP (300 μg/mouse) by i.p. injection and serum IL-6 levels were determined 3 h after injection. To induce endotoxin shock, mice were prestimulated with MDP (300 μg/mouse) i.p. and challenged with ultrapure LPS (250 μg/mouse) from Escherichia coli O55:B5 (Sigma-Aldrich) i.p. 24 h later. For infection, mice were given E. faecalis (1 × 10⁸ CFU/animal) i.p. and bacterial loads in liver and spleen were determined by plating. TNF-α blocking IgG was a gift of T. Moore (University of Michigan).

Statistically significant differences between the two groups were determined by the two-tailed t test with unequal variance (Aspin–Welch t test). Bacterial counts of infected mice were analyzed by a Mann–Whitney U test. The survival rate of infected mice was analyzed using the log-rank test. Differences were considered significant when p values were < 0.05.

The innate immune sensors that are involved in the recognition of E. faecalis, a common cause of infection and lethality in humans, are unknown. To assess a role for Nod2 in the recognition of E. faecalis, we first determined whether products from the bacterium can induce the activation of Nod2 using an NF-κB reporter assay. The culture supernatant from E. faecalis induced robust Nod2-dependent NF-κB activation whereas extracts from the bacterial cell pellet induced little or no stimulation (Supplemental Fig. 1A). These results suggest that the bulk of the Nod2 stimulatory activity is released by the bacterium during growth. Consistent with the lack of the critical γ-d-glutamyl-meso-diaminopimelic acid dipeptide motif in the peptidoglycan of E. faecalis, neither the bacterial pellet nor the culture supernatant induced Nod1-dependent NF-κB activation (Supplemental Fig. 1A). To determine whether Nod2 plays a role in the immune response to live E. faecalis, macrophages were infected with the bacterium at different bacterial/macrophage ratios and cytokine responses were measured. Infection of wild-type macrophages with E. faecalis induced the production of TNF-α and IL-6 and this response was reduced in macrophages lacking Nod2 (Fig. 1A, Supplemental Fig. 1B). Consistently, phosphorylation of JNK and Iκ-Bα induced by E. faecalis infection was delayed and/or reduced in Nod2^−/− macrophages, whereas little or no difference was detected in p38 and ERK phosphorylation (Fig. 1B, Supplemental Fig. 1C). These results indicate that Nod2 contributes to the cytokine response triggered by E. faecalis in macrophages.

To assess whether Nod2 regulates the susceptibility to systemic infection, wild-type and Nod2^−/− mice were injected i.p. and survival was monitored over time. Notably, ∼80% of wild-type mice succumbed to infection as compared with only ∼20% of the Nod2^−/− mice (Fig. 2A). The increased survival of Nod2^−/− mice was associated with reduced levels of IL-6 and TNF-α in the blood when assessed 3 h postinfection (Fig. 2B). The reduced cytokine production was not explained by differential bacterial clearance because we found similar bacterial burden in the spleen and liver of wild-type and Nod2^−/− mice (Fig. 2C, Supplemental Fig. 2A). These results indicate that Nod2 regulates susceptibility to systemic infection with E. faecalis and this is associated with reduced production of TNF-α and IL-6 without affecting bacterial clearance.

Given that increased survival of Nod2^−/− mice after E. faecalis infection was associated with reduced amounts of TNF-α, we asked whether inhibition of TNF-α protects mice from infection induced lethality. Infection of mice i.p. induced 100% mortality in mice treated with control IgG, whereas only 20% of the mice injected with anti–TNF-α blocking Ab succumbed to infection (Fig. 2A). Treatment with anti–TNF-α Ab did not inhibit the production of TNF-α or IL-6 induced by E. faecalis infection (Supplemental Fig. 2B). In contrast, administration of anti–IL-6 blocking Ab did not affect survival of mice infected with E. faecalis (Supplemental Fig. 2C, 2D). The results indicate that TNF-α, but not IL-6, is important for lethality associated with E. faecalis infection in our mouse model.

Monocytes from individuals homozygous for the common CD-associated L1007fsinsC Nod2 mutation exhibit impaired cytokine response after stimulation with MDP (14). However, mouse macrophages homozygous for the equivalent L1007fsinsC Nod2 mutation displayed increased NF-κB activation in response to MDP, which suggested that this genetic variant acts as a gain-of-function mutation in the mouse (16). To clarify these puzzling observations, we used a new strain of mice in which a Nod2 mutation homologous to the human L1007fsinsC was introduced into the Nod2 locus by homologous recombination (11). Sequencing of cDNA prepared from macrophages of homozygous Nod2^2939iCstop mice revealed the expected cytosine insertion followed by the TAG translation termination codon resulting in a truncated Nod2 protein lacking the last 33 aa residues (Supplemental Fig. 3A, 3B). Immunoblotting analysis of Nod2 immunoprecipitates showed comparable levels of endogenous Nod2 in the intestinal tissue of wild-type and homozygous Nod2^m/m mice (Supplemental Fig. 3C). To assess the function of the Nod2^2939iCstop mutation, we stimulated macrophages from wild-type, Nod2^−/−, and Nod2^m/m mice with MDP and monitored MAPK and NF-κB activation. Stimulation of wild-type macrophages with MDP induced phosphorylation of MAPKs and IκBα, which was greatly reduced or abrogated in Nod2-deficient and homozygous Nod2^2939iCstop macrophages (Fig. 3A). Consistently, costimulation of macrophages from wild-type mice with MDP and LTA increased the production of TNF-α and IL-6 when compared with the response observed with LTA alone, which was impaired in Nod2^−/− and Nod2^m/m macrophages (Fig. 3B, Supplemental Fig. 3D). Furthermore, i.p. administration of MDP induced the secretion of IL-6 in the serum of wild-type mice, but not in Nod2^m/m mice (Fig. 3C). To further assess the function of the Nod2^2939iCstop mutation in vivo, wild-type and Nod2^m/m mice were injected i.p. with 300 μg MDP and 24 h later with 250 μg LPS, a protocol that promotes lethality in mice (3). Nod2^m/m mice were protected from lethality induced by administration of LPS and MDP when compared with wild-type mice (Fig. 3D). These results indicate that the mouse Nod2^2939iCstop mutation is insensitive to MDP stimulation in vitro and in vivo.

We next examined the role of the Nod2^2939iCstop mutation in the regulation of proinflammatory cytokine responses postinfection with E. faecalis. Infection of wild-type macrophages with E. faecalis induced secretion of IL-6 and TNF-α and these responses were significantly impaired in macrophages homozygous for the Nod2^2939iCstop mutation (Fig. 4A). As a comparison, Nod2^−/− and homozygous Nod2^2939iCstop macrophages produced comparable amounts of IL-6 and TNF-α in response to bacterial infection that was lower than in wild-type macrophages (Fig. 4A, Supplemental Fig. 4). These results indicate that the Nod2^2939iCstop mutation is associated with defective recognition of E. faecalis in macrophages.

We next examined the susceptibility of wild-type and Nod2^m/m mice to i.p. administration of E. faecalis. In these experiments, we crossed heterozygous Nod2^+/m mice with homozygous Nod2^m/m mice to generate littermate mice heterozygous and homozygous for theNod2^2939iCstop mutation. We then challenged wild-type as well as heterozygous Nod2^+/m and homozygous Nod2^m/m littermates i.p. with E. faecalis and assessed cytokine production and mouse survival over time. Nod2^m/m mice exhibited increased survival compared with wild-type mice (67 versus 25%; p = 0.05) after i.p. challenge with E. faecalis (Fig. 4B). In contrast, heterozygous Nod2^+/m and wild-type mice exhibited comparable survival after bacterial challenge (Fig. 4B). The increased survival of Nod2^m/m mice was associated with reduced serum levels of IL-6 and TNF-α (Fig. 4C). These results indicate that homozygocity for the Nod2^2939iCstop mutation promotes survival after systemic infection with E. faecalis.

Using a newly developed model of the L1007fsinsC Nod2 mutation, we show that the resulting Nod2 truncated protein is impaired in MDP recognition and thus behaves as a loss-of-function mutation similar to that found in human cells (2, 14). There is no a clear explanation to account for the differences in results between our model and the model published by Maeda et al. (15). In contrast to our animal model, a TAG termination codon was not introduced after the cytosine insertion at position 2939 in the model reported by Maeda et al. (15), which is predicted to result in a frameshift mutation leading to the addition of 42 aa not present in the CD-associated or normal Nod2 protein. Thus, it is possible that differences in the resulting Nod2 mutant protein between the two models could explain, at least in part, the discrepancy in results. We show in this study that mice homozygous for the CD-associated Nod2^2939iCstop mutation displayed impaired production of proinflammatory cytokines and improved survival in response to systemic infection with E. faecalis. These results suggest that the CD-associated Nod2^2939iCstop mutation exhibits a dual function. Namely, it promotes the susceptibility to CD, but it also protects the host by limiting the production of harmful cytokines induced by systemic infection with E. faecalis. Because MDP synergizes with TLR ligands in the production of proinflammatory cytokines, it is likely that the role for Nod2^2939iCstop mutation in limiting the production of harmful cytokines and lethality in response to bacterial infection is not limited to E. faecalis. However, most studies to date have shown increased susceptibility of Nod2^−/− mice to infections with microbial pathogens (16). Thus, the contribution of Nod2 mutations in evolutionarily selecting for individuals who are L1007fsinsC remains unclear.

We are grateful to Richard Flavell for Nod2^−/− mice, Joel Whitfield from the Cellular Immunology Core Facility of the University of Michigan Cancer Center for ELISA assays, Sharon Koonse for animal husbandry, and Mizuho Hasegawa for advice.

The authors have no financial conflicts of interest.