Ubiquitination and Phosphorylation in the Regulation of NOD2 signaling and NOD2-mediated Disease

Abstract

The immune system is exquisitely balanced. It has the ability to effectively respond to and control infections while at the same time preventing inappropriate responses to self and environmental antigens. When this response goes awry, either through a failure to activate the immune response, or failure to terminate it, inflammatory pathology results. Posttranslational modifications (PTMs) such as ubiquitination and phosphorylation help ensure that the delicate balance underlying immune signal transduction is maintained. Ubiquitination and phosphorylation affect localization, activity, stability, and interactions of various components of the immune signal transduction machinery. Moreover, ubiquitination and phosphorylation are tightly linked, with one PTM affecting the other. Therefore, in order to find potential therapies for many immune-related pathologies, it is necessary to understand not only how the immune response is activated by ubiquitination and phosphorylation, but also how it is regulated by these PTMs at different stages of the response. An excellent system to study such activation and regulation is the NOD2 pathway. Dysregulation of NOD2 signaling is involved in the pathogenesis of a variety of inflammatory disorders including Crohn’s Disease, Early Onset Sarcoidosis, and Blau Syndrome. More recently NOD2 has been implicated in the development of autoimmune disease, allergy and asthma. This review will focus on what is currently known about how ubiquitination and phosphorylation regulate NOD2 signaling with particular emphasis on novel in vitro substrates which may serve as potential in vivo therapeutic targets for hyperactive NOD2 states.

Introduction

While sequencing of the human genome has given us the foundation for cellular processes, post-translational modifications allow for this foundation to be modified, altered and enriched in response to extracellular cues. Two key post-translational modifications that affect an array of proteins are phosphorylation and ubiquitination. During phosphorylation, kinases catalyze the transfer of a γ-phosphate group (usually from ATP) to the nucleophilic centers of acceptor molecules like the amino acids serine, threonine or tyrosine or to molecules like the phosphoinositides. In protein phosphorylation, the phosphorylation event can lead to a conformational change in the protein that then alters that protein’s activity, localization and/or stability [1]. Alternatively protein phosphorylation can serve to nucleate binding partners through phosphorylation-specific binding domains like SH2 domains [1]. Ubiquitination is equally important, regulating everything from cell cycle progression and cellular proliferation to antigen processing and immune signaling. Although initially believed to impart these functions solely through targeting proteins for proteasome-dependent degradation, it is now appreciated that ubiquitination also results in many non-degradative functions such as nucleation of signaling complexes, receptor endocytosis and even modulating kinase activity [2, 3].

Given their importance in multiple cellular processes, it is no surprise that there is considerable cross-talk between phosphorylation and ubiquitination. This interplay can be illustrated through a number of examples including the SCF E3 ubiquitin ligases in which substrate recognition and degradation is based on the presence of phosphorylated motifs within target proteins (phosphodegrons) [4] or the E3 ubiquitin ligase Smurf1 which recognizes GSK phosphorylation sites on Smad1 to cause Smad1 degradation [5]. This cross-talk has also recently been recognized to be a major component of both the activation and deactivation of the NOD2:RIP2 signaling pathway, a key innate immune pathway with a major role in granulomatous inflammatory disease.

1. The NOD2 signaling pathway

Nucleotide binding and oligomerization domain 2 (NOD2) is an intracellular protein that senses the presence of muramyl dipeptide (MDP), a breakdown product of peptidoglycan found in both gram-negative and gram-positive bacteria [6, 7]. Upon recognition of intracellular bacteria, the NOD2 pathway helps direct the secretion of antimicrobial peptides, cytokines and chemokines to induce innate immune clearance of the pathogen as well as to tailor the adaptive immune system to help fight the bacterial infection [8]. The NOD2 protein is comprised of two N-terminal CARD domains for interacting with other CARD containing proteins, a central NACHT domain for self-oligomerization and ten C-terminal leucine-rich repeats (LRRs) that are thought to be important for the intracellular recognition of MDP (Fig 1) [9]. Basally, NOD2 is thought to exist in an autoinhibited state with the LRRs folded back onto the NACHT domain. Exposure to MDP causes a conformational change in NOD2 allowing both NOD2 oligomerization through the NACHT domain and binding of the dual-specificity kinase RIP2 through homotypic CARD-CARD interactions [10]. Binding to NOD2 promotes RIP2 kinase activation and this active NOD2:RIP2 complex then activates a number of intracellular signaling pathways including the MAP Kinase pathways (JNK, ERK, p38), the NF-κB pathway (Fig2A), autophagy and antigen presentation [8, 11–16]. Of these, the best-studied effect of NOD2 activation is its effect on the transcription factor NF-κB (Fig 2A).

Figure 1. NOD2 and disease.

Structure of the NOD2 protein showing location of disease-associated mutations and polymorphisms with respect to domain organization. Crohn’s Disease associated polymorphisms are indicated in boxes with the 3 most commonly occurring variants in bold. Blau Syndrome mutations are enclosed in circles above the schematic and Early Onset Sarcoidosis mutations are enclosed in circles below the schematic.

Figure 2. Activation and downregulation of NOD2 signaling.

A) Exposure to MDP liberates NOD2 from its autoinhibited state allowing it to bind and activate the kinase RIP2. RIP2 undergoes tyrosine autophosphorylation and is ubiquitinated by a number of E3 ligases. RIP2 also promotes ubiquitination of NEMO/IKK γand by doing so links together proximal and distal signaling components of the pathway. The TAB proteins of the TAK1/TAB2/TAB3 complex bind these ubiquitin chains allowing TAK1 to phosphorylate and activate the IKK complex. IKK activation leads to IκBα phosphorylation, marking IκB for degradation and liberating NF-κB for translocation to the nucleus. B) TRAF4 inhibits NOD2 signaling in a manner dependent on NOD2 binding and IKKα-mediated phosphorylation. The downstream events mediating such inhibition are unknown. A20 is phosphorylated and activated by IKKβ and together with ITCH, targets activated, tyrosine phosphorylated RIP2 for degradation. In a manner independent of its ubiquitin editing function, A20 also inhibits TAK1-mediated phosphorylation of IKK.

NF-κB activation must be tightly regulated to avoid inflammatory disease. Both the activation and the deactivation must be accurately coordinated such that only physiologic, beneficial inflammation results. Upon NOD2:RIP2 activation, RIP2 helps recruit the IKK signalosome via polyubiquitin networks. RIP2 promotes lysine (K) 63-linked polyubiquitination of NEMO and RIP2 itself is K63 ubiquitinated and bound by multiple E3 ligases - cIAP1, cIAP1, XIAP, TRAF2, TRAF5 and TRAF6 (Fig 2A) [17–21]. These K63 ubiquitin chains on both RIP2 and NEMO promote recruitment of TAK1 by serving as unique docking sites for TAB2 and TAB3 (TAK1-binding protein 2 and 3) which have ubiquitin-binding domains [19, 22]. Recruitment of TAK1 then leads to TAK1-mediated phosphorylation and activation of the IKK complex and subsequent activation of NF-κB (Fig2A) [23].

Downregulation of NOD2-induced NF-κB signaling is then accomplished through recruitment of multiple ubiquitin ligases and ubiquitin-editing enzymes. TRAF4 is an E3 ligase that has been shown to negatively regulate NOD2 signaling. Recent data from our laboratory has shown that TRAF4 binds directly to NOD2 in an agonist-dependent manner [24]. This recruitment brings TRAF4 in proximity such that it can serve as a substrate for IKKαMarinis et al., unpublishedresults. IKKα-mediated TRAF4 phosphorylation then allows dissociation of TRAF4 from NOD2 to mediate inhibition of NOD2 signaling (Fig2B). Both TRAF4 binding to NOD2 and IKKα–mediated TRAF4 phosphorylation are needed for the inhibitory effect of TRAF4.

RIP2 is also actively downregulated. The HECT E3 ligase ITCH ubiquitinates RIP2 to downregulate NOD2-induced NF-κB responses [25]. Here the interplay of phosphorylation and ubiquitination is apparent. Phosphorylation is generally a low stoichiometry event. That is, signal amplification is generated starting from a low basal phosphorylation state towards a less low phosphorylation state. For instance, if 0.01% of the total molar amount of protein is basally phosphorylated and upon activation, 2–5% of the total molar amount becomes phosphorylated, this achieves a 200 to 500-fold enrichment in activation. By extension, if only 5% of a protein is phosphorylated, then enzymes such as ITCH that help degrade activated RIP2 must have mechanisms to distinguish activated RIP2 from inactivated RIP2. This distinction is made via the identification of tyrosine phosphorylated RIP2. Upon activation, RIP2 undergoes autophosphorylation on tyrosine 474 (Y474) [26]. ITCH recognizes this tyrosine phosphorylated, activated form of RIP2 to specifically target activated RIP2 for degradation (Fig 2B). An additional level of regulation lies with the protein A20. A20, which forms a complex with ITCH [27], has both ubiquitin ligase and ubiquitin-editing capability and is another potent negative regulator of NOD2 and NF-κB signaling [28, 29]. Primary macrophages deficient in A20 show prolonged NF-κB signaling in response to MDP stimulation and increased RIP2 ubiquitination [30]. Overexpression of WT but not the C103A A20 mutant (which is defective in its deubiquitinating activity) reduced RIP2 ubiquitination in overexpression systems [30]. Together, these suggest that A20’s deubiquitinase activity may contribute to the downregulation of NOD2 responses through deubiquitination of RIP2 (Fig 2B). This would represent a similar scenario to that shown for RIP1, where ITCH-mediated recruitment of A20 is needed for RIP1 deubiquitination and inhibition of NF-κB responses [27].

At the level of IKK complex activation, a number of ubiquitin ligases and ubiquitin-editing enzymes are also active. Aside from the well-documented IKKβ phosphorylation of the NF-κB inhibitor IκBα to promote its ubiquitination and degradation, our own laboratory has shown that IKKβ phosphorylates A20 on Ser381 and this phosphorylation event is important for inhibition of NF-κB (Fig 2B) [31]. Recently, there has also been emerging evidence to indicate that A20 can exert inhibitory effects independent of its deubiquitinase activity. Skaug et al. have shown that polyubiquitin binding by A20 allows it to bind NEMO and inhibit IKK phosphorylation by TAK1 (Fig 2B) [32]. This polyubiquitin binding occurs even in the absence of deubiquitinase activity. Most likely, A20 employs a combination of these mechanisms which renders it a very potent negative regulator of NOD2 signaling. Thus, while much attention has been focused on the NOD2:RIP2’s activation of NF-κB and its role in the coordination of cytokine release, equally important and perhaps more complicated are the mechanisms to downregulate NOD2:RIP2 signaling.

2. Disorders of Defective NOD2 Signaling

2.1 Crohn’s Disease

NOD2 has gained notoriety as a Crohn’s disease susceptibility gene when, in 2001, three independent laboratories identified polymorphisms within NOD2 which confer increased risk for developing Crohn’s disease, a type of inflammatory bowel disease (IBD) characterized by inflammation in the gastrointestinal tract which is transmural, discontinuous and granulomatous [33–35]. The CD-associated polymorphisms of NOD2 confer a 2–4 fold increased risk if heterozygous and a 20–40 fold increased risk if the individual is homozygous or compound heterozygous. These CD-associated polymorphisms are located within the LRR region of the protein and are associated with a loss-of-function phenotype (Fig 1, open bold boxes). That is, in vitro stimulation of cells harboring these CD-associated variants show loss of NF-κB activation [6, 36]. The finding that these loss-of-function polymorphisms precipitate inflammation seems, at first glance, counterintuitive and as such has been a topic of much debate. However, these results do not seem unreasonable if viewed as a setting of immunodeficiency: where the failure to respond to a pathogen ultimately leads to increased bacterial burden and a much more severe inflammatory response when the bacteria are detected through other pathways [37]. Alternatively, other groups have proposed that the 1007insC polymorphism actually exhibits a gain-of-function phenotype to actively repress IL-10 transcription, thereby allowing excessive inflammation [38]. While this finding is supported by the fact that IL-10 levels are decreased in cells from patients harboring the 1007insC polymorphism, the low penetrance of Crohn’s Disease in people with one or two Crohn’s Disease susceptibility alleles and the general autosomal recessive inheritance pattern suggests against a gain-of-function phenotype. This central paradox of NOD2 genetics elicits a number of potential explanations. However, to date, the exact mechanisms explaining the disparity between the effects of the loss-of-function polymorphisms on downstream NOD2 signaling and the pathology observed, have yet to be fully elucidated.

Although NOD2 polymorphisms are the most frequently altered defect in genetic CD, the majority of patients who develop Crohn’s Disease actually have wild-type NOD2. It is not unlikely, therefore, that in these cases the defect might lie in other components of the NOD2 signaling pathway which are faulty in either the propagation or the termination of this innate immune response. Additionally, as NOD2 and RIP2 are both NF-κB regulated genes, upon inflammatory stimulation, both NOD2 and RIP2 expression increase dramatically [39]. It has been postulated that inhibition of NOD2:RIP2 signaling in the background of WT NOD2 may be efficacious in Crohn’s disease [40, 41]. It appears that the mucosal immune system lies in such a delicate immunologic homeostasis that any perturbation, either positively or negatively, may result in inflammation.

2.2 Early Onset Sarcoidosis and Blau Syndrome

What is intriguing about the genetics of NOD2, is that a separate set of mutations all occurring within the NACHT domain result in a different type of granulomatous disease – Early Onset Sarcoidosis (EOS) and Blau Syndrome (Fig 1, open circles) [42, 43]. The EOS and Blau Syndrome mutations, unlike the CD-associated polymorphisms, are autosomal dominant, gain-of-function mutations that are associated with over-activation of the NOD2 pathway as evidenced by increased basal NF-κB activity even in the absence of agonist [43, 44]. EOS and Blau Syndrome are characterized by the triad of skin, joint and eye inflammation and are thought to be the sporadic and familial forms of the same genetic disease [43, 45]. The late-onset and adult forms of the disease often present with inflammatory involvement of the lungs and enlargement of the hilar lymph nodes [46, 47]. A recent study described what may be additional variants of NOD2 hyperactive disorders (so-called autoinflammatory NOD2 disorders) characterized by periodic fever, dermatitis, and polyarthritis [48]. A subset of these patients harbor one of the gain-of-function autosomal dominant mutations (R702W), while all were positive for the NOD2 IVS8 +158 mutation [48]. Whether this represents a distinct intronic mutation or is a marker for a co-segregating unidentified gain-of-function mutation, is yet to be discovered. Thus, as seen with both Crohn’s Disease and Early Onset Sarcoidosis, any perturbation of NOD2, either positive or negative, can lead to granulomatous inflammatory disease.

2.3 Allergy and Asthma

NOD2 has also been implicated in the development of allergic responses. Two independent genetic analysis of large German cohorts have found that some polymorphisms within NOD2 are associated with increased total serum IgE levels, atopic dermatitis, and atopic rhinitis but that other polymorphisms and haplotypes seem to be protective for these traits [49, 50]. In a separate study looking at Canadian and Australian populations, there was an association between NOD2 polymorphisms and airway hyperresponsiveness [51]. Studies in mice using the NOD2 agonist, MDP, similarly implicate NOD2 as being involved in the generation of Th2 immunity and the development of asthmatic and allergic responses. Immunization of mice using ovalbumin with MDP as an adjuvant showed increased secretion of Th2 related cytokines and elevation of antigen-specific IgG1 antibodies [52]. Other studies by the same group show that direct administration of MDP during intranasal tolerization protocols block respiratory tolerance and promote susceptibility to asthmatic disease [53]. NOD2 signaling in both stromal cells and DCs were determined to be crucial for development of optimal Th2 responses [54]. These studies are highly suggestive of overactive NOD2 signaling as a contributor to allergic and asthmatic disease.

2.4 Other Autoimmune Diseases

There is also emerging evidence for the involvement of NOD2 in autoimmune disease. Genetic analysis of NOD2 in individuals with multiple sclerosis (MS) and systemic lupus erythematosus (SLE) shows an increased susceptibility of developing SLE in individuals with harboring the 908R polymorphism [55]. Interestingly, occurrence of this polymorphism is also increased in a subset of IBD patients with spondylarthritis [56, 57]. Although initial studies of psoriatic arthritis have suggested an increase in the occurrence of this disease in patients harboring loss-of-function polymorphisms in NOD2, more recent studies have not been able to replicate these associations [58–61].

The connection between NOD2 and autoimmunity was further strengthened by the finding that stimulation of human dendritic cells with the NOD2 agonist MDP in combination with other TLR agonists (especially TLR2), promotes the development of T helper 17 (Th17) cells through the induction of IL-23 and IL-1 [62]. Th17 cells are a subset of T cells that secrete proinflammatory cytokines (most notably IL-17) and have been shown to be important in the control of certain bacterial infections but have also implicated in the pathogenesis of various autoimmune conditions [63]. Studies using genetically deficient mice also support the involvement of the NOD2 pathway. By subjecting mice to experimental autoimmune encephalomyelitis (EAE), a mouse model of MS, the authors show that mice deficient in NOD2 or RIP2 are more resistant to development of EAE [64]. Defective autoantigen-specific T cell development and expansion were not responsible for these results. Instead, a defect in RIP2 signaling in CNS-infiltrating dendritic cells (DCs) was suggested to drive the autoimmune pathology [64]. As with all studies using mouse models, it will be important to verify these findings in patient samples and through the use of human genetics. To date, however, there is no human genetic linkage of NOD2 to MS.

2.5 ITCH-deficiency

Not surprisingly, genetic loss of a potent negative regulator of the NOD2 signaling pathway, the E3 ubiquitin ligase ITCH, results in widespread inflammation in various mucosal compartments – the skin, the lungs and the intestinal tract [65]. The majority of patients found to be deficient in ITCH (harboring a truncation mutation rendering them essentially null) present with chronic lung disease, highlighting this as a major organ of involvement by ITCH and by proxy, NOD2 [66]. Our laboratory has shown that bone marrow-derived macrophages from ITCH-deficient mice (itchy) have increased basal and MDP-inducible NOD2 responses [25]. Furthermore, if we subject this genetically-susceptible strain of mice to a mouse sarcoidosis model (intraperitoneal immunization and intratracheal challenge with heat-killed Propionibacterium acnes), we are able to observe a much more pronounced granulomatous inflammatory lung response in the itchy mice compared to their wild-type counterparts (Tigno-Aranjuez et al., unpublished results). Therefore, ITCH-deficiency represents a unique setting of NOD2 overactivation, one where NOD2 status is genetically wild-type but where loss of negative regulation still precipitates disease.

Although loss-of-function CD-associated polymorphisms of NOD2 are most famous for being implicated in genetic Crohn’s disease, there are now multiple examples of NOD2 overactivation causing inflammatory disease (Table I). These now include: EOS, Blau Syndrome, autoinflammatory NOD2 disorders, allergy and asthma, EAE, Systemic Lupus Erythematosus, and ITCH-deficiency. Given that hyperactive NOD2:RIP2 signaling is implicated in the pathophysiology of these disorders when NOD2 is WT, in this genetic background, patients might benefit from pharmacologic downregulation of NOD2 signaling.

Table I.

How NOD2 Influences Inflammatory Disease.

Disease	Way in which NOD2 influences disease
Crohn’s Disease	Negatively: loss-of-function polymorphisms cause inflammatory granulomatous disease
Early Onset Sarcoidosis / Blau Syndrome	Positively: gain-of-function mutations cause inflammatory granulomatous disease
Allergy and Asthma	Positively: the NOD2 agonist MDP promotes Th2 immunity when used as an adjuvant; MDP administration during tolerization protocols blocks respiratory tolerance
Autoimmunity	Positively: increased susceptibility of developing SLE in patients with the 908R polymorphism; in murine models, NOD2 or RIP2 deficient mice showed more resistance to development of EAE
ITCH deficiency	Positively: loss of ITCH, a key negative regulator of NOD2 signaling, causes exacerbation of what should be normal NOD2 signaling

3. Novel NOD2 pathway Inhibitors

Until recently, it was believed that the kinase activity of RIP2 was dispensable for downstream activation of the NOD2 pathway. This conclusion was based on in vitro overexpression studies using kinase-dead versions of RIP2 which showed an equal or greater ability to activate NF-κB compared to wild-type RIP2 [12, 67]. However, more recent studies have begun to question these conclusions. One study using a kinase-dead (K47A) knock-in mouse showed that RIP2’s kinase activity was important for its expression [68]. A second study using an inhibitor of both RIP2 and p38 showed that the kinase activity of RIP2 was important for both RIP2 expression and for MDP-induced cytokine responses [69]. These, as well as our own observations, led us to more formally interrogate the importance of RIP2’s kinase domain in the activation of the NOD2 pathway. RIP2 was originally classified as a serine/threonine kinase [11, 70]. However, no RIP2 substrates had been clearly identified which could be used as potential targets or markers of activation of the NOD2 pathway. Previous use of RIP2 itself as a potential marker of NOD2 activation has also met with limited success as an anti-phospho Ser176 RIP2 antibody, which, although specific [71], does not seem to detect MDP-inducible phosphorylation of RIP2 (Abbott laboratory, unpublished results).

Although RIP2 was originally described as a serine-threonine kinase, we have recently found that RIP2 is actually a dual-specificity kinase, capable of phosphorylating not only serines and threonines but tyrosines as well [26]. Our laboratory has demonstrated that RIP2 is not only a bona fide tyrosine kinase through in vitro kinase assays, but has shown that RIP2 is inducibly tyrosine autophosphorylated at tyrosine 474 (Y474) in response to NOD2 activation [26]. This finding opened up new avenues for identifying potential inhibitors of the NOD2 pathway. Using a small molecule tyrosine kinase inhibitor screen, we discovered two additional FDA-approved drugs - Tarceva (Erlotinib) and Iressa (Gefitinib), which are capable of inhibiting RIP2 tyrosine activity in the nanomolar range [26]. We showed that these drugs can inhibit downstream MDP-induced cytokine responses and are effective, at least in vitroin alleviating the exacerbated cytokine responses in NOD2 hyperactive states such as in primary macrophages which lack ITCH, or in macrophage cell lines overexpressing the Blau Syndrome-associated mutation R334Q [26]. Whether these drugs can also inhibit these inflammatory states in vivo still remains to be tested. It would also be interesting to determine whether lung cancer patients who have been receiving or who had received Tarceva and Iressa show symptoms correlating with modulation of NOD2 signaling.

Tarceva and Iressa were designed as inhibitors of EGFR, a receptor tyrosine kinase, and are used most commonly for treatment of lung cancers [72]. These drugs work as competitive inhibitors of ATP by binding to the ATP-binding pocket of their target kinase. Because the RIP2-inhibitory effect of Tarceva and Iressa was found as an off-target effect of these drugs, caution must be used in applying these inhibitors in situations where EGFR signaling may be important. Although we have shown that Tarceva and Iressa are specific for RIP2 relative to EGFR in macrophages, we cannot rule out the possibility that there are additional off-target effects of these agents. For this reason, and given that RIP2 remains an attractive pharmacologic target in inflammatory disease, it will be important to develop specific RIP2 antagonists.

Additionally, a number of ubiquitin ligases have been associated with a positive influence on the NOD2 signaling pathway (cIAP1, cIAP2, XIAP) [20, 21]. Numerous pharmacological inhibitors for these ligases are available and are currently at various stages of clinical trials (Table II). Although originally designed for inhibition of IAPs in association with cancer, it is possible that the same inhibitors could be repurposed for the downregulation of NOD2 signaling as well, albeit in a less-specific manner than NOD2 or RIP2 inhibitors. As of the writing of this review, there has been successful development of selective and potent NOD1 inhibitors which show inhibition of NOD1-mediated activation of NF-κB and NF-κB mediated cytokine responses [73–75]. The same studies which screened and tested for NOD1 inhibitors also led to the identification of NOD2 specific inhibitors. However, the structural, chemical, and functional information for the compounds which resulted from this screen have yet to be published. There is one report which describes arene chromium-complexes (AKS-01 being the most potent) as specific inhibitors of NOD2-mediated NF-κB signaling [76, 77]. These compounds seem to be active in a variety of cell lines (HEK 293, THP, primary murine DCs) and show no effects on reduction of NF-κB signaling mediated through other pathways such as TNFα, TLR2 and TLR4 and only a slight reduction in NOD1-mediated NF-κB signaling [76, 77]. However, as with the Erlotinib and Gefitinib, further studies are needed to determine in vivo efficacy and safety.

Table II.

Drugs which inhibit NOD2 signaling

Drug	Manufacturer	Target	NOD2 Pathway off-target	Mode of Action	Phase	Reference
AT-406	Ascenta Therapeutics	IAPs		Orally active Smac mimetic; binds to the BIR3 regions of XIAP, cIAP1 and cIAP2 ; promotes caspase-9 activation for XIAP and promotes cIAP1 degradation	I	[77]
TL-32711	TetraLogic Pharmaceuticals	IAPs		Smac mimetic	II	[78]
LBW242, LCL161	Novartis	IAPs		Smac mimetic	Preclinical, I	[79]
GDC-0152, GDC-0917	Genentech	IAPs		Smac mimetic	I	[78]
AEG35156, AEG40826-2HCl	Aegera	XIAP		XIAP antisense oligonucleotide	I/II I	[80]
SB 203580	GlaxoSmithKline	P38	RIP2	competitive inhibitor of ATP binding	Preclinical	[81]
SB 203580 SB 220025 PD 169316	GlaxoSmithKline	P38	RIP2	competitive inhibitor of ATP binding	Preclinical	[82]
Iressa (Gefitinib), Tarceva (Erlotinib)	AstraZeneca, Genentech	EGFR	RIP2	competitive inhibitor of ATP binding	FDA approved	[26]
arene—Cr(CO)3 complexes such as AKS-01	University of Cologne	NOD2	NOD2	proposed competitive inhibitor of MDP	Preclinical	[83]

4. Potential Avenues for Research

There are still many unanswered questions regarding NOD2 signaling. For example, although RIP2 undergoes inducible tyrosine autophosphorylation in response to MDP stimulation, the other downstream targets of RIP2 kinase activity have yet to be discovered. Can these be exploited for clinical gain? Second, RIP2 is modified by multiple ubiquitin ligases. How is this affecting its kinase activity and if so, how is this modulation achieved? Does regulation between the different ubiquitin ligases occur? Is this affected by RIP2 kinase activity? Third, we have mentioned numerous ubiquitin ligases and the substrates involved in the modulation of NOD2 signaling. If we understood the exact nature of ubiquitin ligase substrate recognition in these cases, would it be feasible to design small molecule inhibitors (decoys) to disrupt the complexes which are formed specifically in NOD2 signaling? What are the other NOD2 pathway-specific substrates for these ligases (for example, no substrates have been described for TRAF4)? Fourth, it is clear that dysregulated NOD2 signaling results in a variety of disorders. What molecular events are common among certain disease states? Can these be used as diagnostic markers and utilized in a personalized approach to treatment?

Although high throughput screens (for kinase and ligase substrates) will probably be needed to answer many of these questions, movement toward a systems biology approach to understanding NOD2 signaling is most likely in order. The downstream effects of NOD2 and RIP2 activation outside of NF-κB are still not clearly delineated. Although it is well documented that engagement of NOD2 activates the MAPK pathways, the exact mechanism of how this occurs is still unknown. It is also well known in the field that MDP is a very weak agonist, at least compared to other ligands for TLRs. Are we missing something crucial in downstream NOD2-mediated effects? Furthermore, there is clearly cross-talk between the different microbe associated pattern receptors. For instance, NOD2 has repeatedly been associated with a number of antiviral recognition receptors and pathways [78, 79]. How is RIP2 kinase activity affecting activation of these pathways and which ubiquitin ligases and substrates are recruited to mediate specific antiviral responses? The vast developments in the fields of genetics and the decrease in the costs for sequencing analysis open up new avenues for further advancement of the NOD2 field. For example, one could employ whole exome sequencing to identify new and rare mutations or polymorphisms in NOD2 itself or in components of the NOD2 pathway. One level above studying the genetic differences which underlie many of the aforementioned diseases is the study of epigenetic changes which occur in NOD2-associated disorders. ChIP-Seq can be used to determine not only protein-DNA interaction events specific to NOD2 activation but DNA methylation and histone modification events as well. RNA sequencing technology now offers greater advantages from traditional microarray technology, removing the bias of focusing on only selected known genetic sequences (hybrid-based technology) but also allowing quantitation of expression levels especially for very low or very high mRNA expression. In addition, RNA-seq allows one to determine differential splicing patterns and RNA-editing events. These technologies combined with the appropriate questions, proper controls, and relevant study design, are bound to yield valuable information to further our current understanding of NOD2 signaling and provide insight on how we might be able to arrive at targeted therapies for NOD2-related disorders.

Highlights.

• The NOD2 pathway is regulated by phosphorylation and ubiquitination

• A number of disease settings may benefit from inhibition of NOD2 signaling

• We describe potential therapeutic targets in the NOD2 pathway

• We describe currently available inhibitors of NOD2 signaling