Crohn’s disease, a chronic inflammatory bowel disease, affects many people. For example, more than 0.3% of the populations of Canada and Germany have the condition, and its incidence is increasing worldwide1. Better therapies are needed, but progress in treating Crohn’s disease has been hampered by the lack of understanding of how it arises. Writing in Nature, Nayar et al.2 shed light on a long-standing mystery about one risk factor for Crohn’s disease, and their findings have important clinical implications.
Crohn’s disease can affect any part of the gut. Most commonly, it affects the ileum region, causing inflammation that frequently results in fibrosis (the deposition of fibrous connective tissue as an injury response). This leads to the narrowing (or stricture) of the lumen of the ileum, which often requires surgical intervention3. Crohn’s disease provides a useful model of illnesses that are mediated by genes and environmental interactions. In this case, genetic susceptibility underpins the disease-causing inflammatory responses to gut microorganisms.
Genetic variations, called polymorphisms, of the NOD2 gene are the strongest genetic risk association for Crohn’s disease; approximately 20% of all such risk of developing the disease is related to three single nucleotide polymorphisms of this gene4. Furthermore, NOD2 mutations are strong predictive factors for the development of ileum strictures and for the need for surgery in Crohn’s disease, which is a widely validated association between the genetic underpinnings of this condition and manifestations of the disease3.
However, connecting the NOD2 gene to disease susceptibility presented a paradox. NOD2 is an intracellular receptor (Fig. 1) that recognizes the molecule muramyl dipeptide (MDP) — a ubiquitous component of bacterial cell walls. Before NOD2 was described as a risk gene for Crohn’s disease, NOD2 function was best understood in immune cells that aid the innate branch of immune defences. NOD2 activation in these cells leads to the expression of inflammatory molecules called cytokines, and an abnormally intense inflammatory response can mediate intestinal damage in Crohn’s disease5,6. One might therefore have expected that NOD2 mutations known as loss-of-function mutations, which do not generate a fully functional version of the encoded protein, would protect against Crohn’s disease. Yet such loss-of-function mutations of NOD2 were identified as risk factors for the disease. Subsequent research therefore pivoted to focus on a different aspect of NOD2 biology in the intestine, investigating how functional NOD2 maintains homeostasis in the intestine, where the body’s largest biomass of immunologically active cells is constantly exposed to MDP from gut microbes, and how NOD2 mutations perturb this balance and lead to disease5.
The role of NOD2 mutations in the emergence of fibrosis of the ileum was unknown before the present study. The authors sought to understand what drives inflammation and fibrosis in Crohn’s disease, and linked these biological insights to NOD2 through research using human cells, human intestinal tissue and a zebrafish model.
First, the authors used single-cell sequencing of RNA from the inflamed tissue of ileum samples removed during surgery from people with Crohn’s disease. These cells revealed a gene-expression signature associated with activated macrophage and fibroblast cells. The authors also identified a key cell type that expresses markers of both myeloid and fibroblast cellular lineages. These discoveries suggest that a population of inflammatory macrophages in the ileum differentiates to become activated fibroblasts during the course of disease.
Strikingly, the authors demonstrate the evolutionary conservation of these cellular populations in an experimental model of intestinal inflammation — zebrafish treated with the molecule dextran sodium sulfate (DSS). This molecule has long been used to induce intestinal damage and inflammation in a standard rodent model. The in vivo modelling of human inflammatory bowel diseases has been dominated by mouse models. However, as Nayar and colleagues demonstrate, zebrafish offer a useful alternative for relatively high-throughput investigations and rapid assessment of correlations with human disease. Indeed, zebrafish and mammalian intestines have a similar form (morphology). Moreover, like humans, zebrafish have innate and adaptive branches of their immune-defence responses, and intestinal inflammation of zebrafish is also dependent on the community of gut microorganisms7. Gene-editing tools, such as CRISPR, aid the rapid modification of genes of interest in zebrafish.
The authors studied intestinal inflammation in zebrafish engineered to have nod2 deficiency. These fish, treated with DSS, had increased numbers of leukocyte immune cells in their intestines, a hallmark of inflammation, compared with zebrafish with normal nod2. But the zebrafish model is relevant only if a human correlation can be established. Accordingly, using data from children newly diagnosed with Crohn’s disease, the authors show that an increase in the number of copies of a NOD2 mutation (associated with the risk of Crohn’s disease) indeed correlated with an activated macrophage and fibroblast gene-expression signature in ileum tissue.
To understand NOD2 function in human cells that can differentiate in vitro, the authors used peripheral blood monocytes from healthy volunteers, and determined whether the cells had one, two or no copies of NOD2 mutations linked to susceptibility to Crohn’s disease. The cells were then differentiated in vitro with and without MDP. The authors observed a higher number of activated fibroblasts for cells with two copies of NOD2 mutations compared with cells with wild-type NOD2. Furthermore, an increase in the number of NOD2 mutations was associated with a corresponding enrichment in the number of fibroblasts with a gene-expression signature characteristic of activated cells. Interestingly, zebrafish with nod2 deficiencies, which were given MDP, had a gene-expression signature characteristic of activated fibroblasts that persisted even during recovery from injury mediated by DSS, compared with zebrafish that have wild-type nod2. These data suggest that nod2 deficits inhibit efficient recovery (resolution) from fibrosis and inflammation.
To further elucidate the molecular basis of the fibrosis-linked gene-expression signature associated with NOD2 risk mutations, the authors searched for upstream transcriptional regulators of this pathway. They identified the gene encoding STAT3 as being markedly upregulated in activated fibroblasts and macrophages. STAT3 is a transcriptional regulator of key components of inflammatory and fibrotic responses in inflammatory bowel diseases, and acts through the cytokine receptor gp130. Analyses of clinical data revealed upregulated expression of gp130-regulated genes encoding the proteins IL-6, oncostatin M and IL-11 in people with Crohn’s disease who did not respond to therapy targeting the tumour-necrosis factor (TNF) protein (anti-TNF antibodies are a common treatment for Crohn’s disease). The discovery supports a role for gp130 signalling in this group of therapy-resistant individuals.
The authors hypothesized that gp130 blockade might lessen the abnormalities that occur with NOD2 mutation. They tested this idea by using bazedoxifene, a gp130 inhibitor, on MDP-treated human cells with NOD2 mutations. Bazedoxifene indeed lessened the fibrotic-associated gene-expression signature and reversed the cellular shape changes that are characteristic of activated fibroblasts. This drug also reduced the intestinal damage found in nod2 mutant zebrafish treated with DSS.
Starting with the clinical characteristics of fibrosis in Crohn’s disease, this work describes a molecular pathway linked to NOD2 mutations associated with the disease, and concludes with a potential therapeutic insight to address the pressing clinical problems of fibrosis and anti-TNF drug resistance. By underpinning the genetics and the clinical outcomes to this cellular and molecular pathway, the study provides a road map to understanding present and future therapeutic approaches.
Many interesting avenues of investigation remain. NOD2 is the only described recognition pathway for MDP, yet this paper demonstrates MDP-induced cellular and molecular changes in the absence of nod2 in zebrafish. This implies that there are MDP signalling pathways that have not yet been described. Bazedoxifene was initially characterized as a selective inhibitor of the oestrogen receptor8, raising the concern that the drug might have adverse effects on other signalling pathways if used as a therapeutic for Crohn’s disease. The gp130 receptor has multiple ligand binding partners that influence a broad range of immune responses. Hence, understanding the specific gp130 ligands that orchestrate NOD2-mediated molecular events could lead to more selective, effective and safer therapeutic interventions than would globally inhibiting gp130 signalling, or targeting other clinically relevant signalling pathways, such as the Janus kinase enzymes (inhibitors for which are in late-stage clinical development for Crohn’s disease).
Not everyone with Crohn’s disease has NOD2 mutations associated with disease risk. Indeed, in individuals of certain ethnic groups, such as people of Chinese, Malay or Indian heritage, disease of the ileum is a prominent clinical feature of Crohn’s disease, yet NOD2 is not associated with disease risk in this population5,9. Perhaps the molecular signature of activated macrophages and fibroblasts is the relevant unifying signature for individuals with Crohn’s disease of the ileum. It is probable that different genetic landscapes might result in the same clinical and molecular outcomes. Hence, Nayar and colleagues’ results move the field a step closer to a molecular classification of Crohn’s disease that might clarify a complex condition that has approximately 200 genetic regions associated with disease risk10, and diverse clinical manifestations.
Nature 593, 201-203 (2021)