Introduction
Liver cancer was the fourth leading cause of cancer-related death worldwide in 2018 [
1]. Hepatocellular carcinoma (HCC) constitutes 75–85% of all liver cancers and is closely associated with chronic liver injury, which leads to repetitive cell death and regeneration, as well as chronic liver inflammation and fibrosis, which causes DNA damage in hepatocytes. In response to DNA damage, the hepatocytes activate the DNA damage response (DDR) signal to repair DNA damage [
2]. Disruption of the damage repair pathways may cause DNA mutations and transformation of normal cells into tumor cells [
2], leading to genomic instability, and eventually, hepatocarcinogenesis.
Due to its close anatomical connection to the intestine via the portal vein, the liver is the first organ exposed to the gut microbiota and their metabolites, which contain pathogen-associated molecular patterns (PAMPs). There is increasing evidence showing that dysbiosis in the gut-liver axis may play an important role in hepatocarcinogenesis [
3‐
5]. For instance, in HCC patients, and diethylnitrosamine (DEN)-induced HCC models, alteration of the gut microbiota, increases in intestinal permeability and translocation of bacterial metabolites from the gut to the liver have been reported [
3‐
5]. Consequently, the gut-derived PAMPs may trigger persistent inflammation and hepatocellular injury in the liver, promoting HCC development. Furthermore, bacteria clearance using antibiotics has been shown to reduce HCC development in the DEN/CCl
4-induced HCC model [
4], suggesting that gut-derived PAMPs may participate in hepatocarcinogenesis. However, the mediators whereby the gut microbiota and their metabolites modulate liver tumorigenesis are still not well characterized.
We and others have previously shown that nucleotide-binding oligomerization domain 2 (NOD2), a member of pattern recognition receptors (PRRs), propagates the inflammatory response via RIP2 signaling in the circulatory system during bacterial invasion [
6,
7]. Once gut-derived PAMPs translocate to the liver via circulation, the liver senses them mainly through the PRRs expressed on the hepatocytes [
8]. NOD2 is a well-characterized intracellular PRR of the NOD-like receptor (NLR) family, consisting of a C-terminal leucine-rich repeat (LRR), a central nucleotide-binding (NATCH) domain and two N-terminal caspase activation and recruitment (CARD) domains [
9]. Unlike Toll-like receptor 4 (TLR4), which recognizes the LPS found mainly in gram (−) bacteria [
10], NOD2 senses muramyl dipeptide (MDP), which is present in both gram (−) and gram (+) bacteria, making it a broader sensor of bacterial invasion compared to other PRRs [
11]. Therefore, certain genetic variants of
Nod2 gene are associated with an increased susceptibility to digestive disease such as Crohn’s disease [
12,
13]. NOD2 has also been implicated in the liver inflammatory diseases, as evidenced by its role in the promotion of experimental cholestatic liver fibrosis and hepatitis through inflammatory cytokines production [
14‐
16]. However, the role of hepatic NOD2 in HCC remains elusive.
In this study, we further reported that NOD2 is an important hepatic sensor of gut microbial metabolites that promotes hepatocarcinogenesis. We demonstrate that NOD2 promotes hepatocarcinogenesis through two different mechanisms, including a previously unrecognized role, whereby NOD2 initiates a nuclear autophagy pathway, leading to increased DNA damage and genomic instability.
Discussion
Genomic instability induced by chronic inflammation can be detected in 90% of early HCCs and dysplastic nodules in cirrhotic liver. For instance, telomerase reverse transcriptase (TERT) promoter mutations are presented in 25% of dysplastic nodules, increased in early HCC (61%) and remained stable in progressed and advanced HCC [
30], which suggests that genetic instability might be an early hallmark in hepatocarcinogenesis. However, the mediators responsible for genetic instability in early HCC remain poorly understood. Our study reveals NOD2 as a novel hepatic mediator, sensing gut dysbiosis and linking gut-derived bacterial PAMPs with chronic inflammation and genomic instability in liver (summarized in Fig.
7h). We show that, clinically, NOD2 is overexpressed in HCC samples and closely correlates with poor prognosis of HCC patients. In DEN/CCl
4-treated HCC model, hepatic NOD2 is activated and NOD2 deletion attenuates the tumorigenesis of HCC. Mechanically, in a RIP2-dependent manner, NOD2 activation triggers a pro-inflammatory response through activation of the NF-κB, JAK2/STAT3 and MAPK pathways. Remarkably, in a RIP2-independent way, activated NOD2 undergoes nuclear translocation and increases lamin A/C protein degradation via the nuclear autophagy pathway, consequently increasing genomic instability.
Usually, bacteria and their metabolites can damage host cell DNA directly by their genotoxins or indirectly by triggering inflammatory response [
31]. Here, we found another approach whereby microbiota potentiates DNA damage through the disruption of the damage repair pathways. When DNA damage occurs, cells respond by activating DDR signal to recognize damaged DNA and recruiting DNA repair proteins to DNA damage sites to facilitate damage repair [
2]. The disruption of the damage repair pathways may cause mutations or inactivation of certain oncogenes or tumor-suppressor genes, which can lead to the death of normal cells or transformation into tumor cells [
2]. Consistence with this, we showed that persistent activation of hepatocyte NOD2 signaling by bacteria impairs DNA repair and increases carcinogenesis during the early stage of HCC development.
Notably, we discovered a novel mechanism linking NOD2 signaling to DDR through lamin A/C degradation. Lamin A/C, encoded by
LMNA gene, is a nuclear matrix protein key in the maintenance of genome stability [
32]. Mutations in
LMNA are associated with premature aging syndromes and muscular dystrophies, collectively termed as laminopathy [
33]. Patients with severe laminopathy exhibit nuclear abnormalities including genome instability, epigenetic dysregulation and telomere shortening, and die in their teens. Consistently, we observed that NOD2 activation RIP2 independently promotes lamin A/C protein degradation and consequently inhibits the repair of DEN-induced DNA damage, by decreasing the recruitment of 53BP1 to DNA damage sites, reducing SIRT6 enrichment to chromatin and diminishing the activity of NHEJ pathway. Supporting our finding, previous studies have reported that loss of lamin A/C increases DNA damage by suppression of these pathways [
25,
26,
34]. Recently, a study by Liu et al. [
35] showed that LMNA may be an oncogene in HCC development. Indeed, previous studies have indicated that LMNA expression may be a double-edged sword in tumor initiation and progression: increased lamin A/C levels could facilitate tumor cell proliferation and migration in HCC progression, while decreased lamin A/C could induce genomic instability in tumor initiation [
32]. Thus, it is worth noting that the role of LMNA in cancer is probably context dependent and vary with stage of disease.
We found that NOD2 promotes lamin A/C protein degradation through nuclear autophagy pathway and identified NOD2 as a previously unknown initiator of nuclear autophagy. Previous studies have reported that NOD2 recruits ATG16L1 to the cell membrane at the bacterial entry site to kill the invading bacteria in human colon epithelial cells, macrophages and dendritic cells [
36‐
38]. This antibacterial xenophage occurs in the cytoplasm, as a host defense machinery against pathogenic infections. We expanded these studies to hepatocytes and, for the first time, found that NOD2 activation triggers nuclear autophagy. To date, little is known about the role of nuclear autophagy in degradation of nuclear materials. A recent study reported that oncogenic stress-initiated nuclear autophagy mediates nuclear protein lamin B1 degradation, causing cell proliferation impairment, chromatin alteration and nuclear membrane disruption, and, eventually, driving cells senescence [
23]. Our study shows that Gut-derived NOD2 agonist MDP could initiate nuclear autophagy and mediate lamin A/C degradation, which leads to nuclear matrix disruption, chromosomal aberration and accumulation of genomic mutations and instability, thereby promoting tumor formation. Thus, NOD2 appears to be an important PRR whereby microbial metabolites may cause nuclear DNA damage through the degradation of nuclear components.
Recent gut microbiome analyses showed that both gram (+) and gram (−) bacteria are enriched in cirrhosis patients and early HCC [
39,
40]. In addition, oral treatment with gram (+)-specific antibiotics (vancomycin or ampicillin) and gram (−)-specific antibiotics (metronidazole and neomycin) was shown to significantly reduce liver inflammation and hepatocarcinogenesis [
5]. Collectively, these studies indicated that gut-derived PAMPs from gram (+) and gram (−) bacteria could contribute to hepatocarcinogenesis. In this context, NOD2 may serve as a unique bridge between gut-derived microbiota and hepatocarcinogenesis, because it can sense MDP from both gram (+) and gram (−) bacteria. Supporting this hypothesis, our findings in clinic samples, DEN/CCl
4-induced HCC mice models and primary hepatocytes treated with DEN, demonstrate that hepatic NOD2 links gut-derived microbiota with liver inflammation and DNA damage.
Recent studies demonstrated that several NOD2 inhibitors including the initial hit compound (GSK669) and its related analogues, GSK400 and GSK717, possess sub-micromolar in vitro activity [
41]. These inhibitors appear to exhibit significant selectivity for inhibition of NOD2-mediated responses since they do not block any of the other IL-8/NF-κB inducing pathways including NOD1, TNFR1 and TLR2 [
41]. Although NOD2 inhibitors have not yet been tested in the clinical trials, previous studies have reported that NOD2 inhibitors exhibited favorable anticancer activity in vivo [
42,
43]. Therefore, NOD2 inhibitors may be promising chemical tools to treat HCC patients, when in combination with other therapies such as chemotherapeutics or immunotherapy.
Conclusion
Here we provide evidence showing that NOD2, a general sensor of gram (+) and gram (−) bacteria, acts as a hepatic mediator linking the gut-derived bacterial PAMPs with chronic inflammation and the genomic instability in hepatocarcinogenesis. Clinically, NOD2 was overexpressed in HCC samples and closely correlated with poor prognosis of HCC patients. In our animal model, hepatic NOD2 was activated in the process of DEN/CCl
4-induced HCC, while NOD2 deletion attenuated the tumorigenesis of DEN/CCl
4-induced HCC. Mechanically, in a RIP2-dependent manner, NOD2 activation triggered a pro-inflammatory response through the activation of NF-κB, JAK2/STAT3 and MAPK pathways. Notably, we show that activated NOD2 acts as a previously unknown initiator of the nuclear autophagy pathway in a novel RIP2-independent manner, promoting the degradation of the nuclear component lamin A/C. A decrease in lamin A/C impairs the ability of hepatocytes to repair damaged DNA and, as a result, leads to increased genomic instability (summarized in Fig.
7h). Targeting hepatic NOD2 might represent a promising therapeutic strategy for HCC.
Limitations of study
Our study still has limitation that the mechanism for a higher expression of NOD2 in HCC is still unclear. Previous studies have demonstrated that NOD2 promotor contains two NF-κB binding sites that enable the transcriptional activation of NOD2, following stimulation with TNF-α, which resulted in increased NOD2 mRNA and protein level [
44‐
46].
Moreover, this up-regulation is augmented by IFN-γ [
44]. We estimated that up-regulation of NOD2 expression might be part of a positive regulatory loop induced by inflammatory cytokines or bacterial components, which needs the further investigation.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.