Using genes to triangulate the pathophysiology of granulomatous autoinflammatory disease: NOD2, PLCG2 and LACC1

Abstract

The intersection of granulomatosis and autoinflammatory disease is a rare occurrence that can be generally subdivided into purely granulomatous phenotypes and disease spectra that are inclusive of granulomatous features. NOD2 (nucleotide-binding oligomerization domain-containing protein 2)-related disease, which includes Blau syndrome and early-onset sarcoidosis, is the prototypic example of granulomatous inflammation in the context of monogenic autoinflammation. Granulomatous inflammation has also been observed in two related autoinflammatory diseases caused by mutations in PLCG2 (phospholipase Cγ2). More recently, mutations in LACC1 (laccase domain-containing protein 1) have been identified as the cause of a monogenic form of systemic juvenile idiopathic arthritis, which does not itself manifest granulomatous inflammation, but the same LACC1 mutations have also been shown to cause an early-onset, familial form of a well-known granulomatous condition, Crohn’s disease (CD). Rare genetic variants of PLCG2 have also been shown to cause a monogenic form of CD, and moreover common variants of all three of these genes have been implicated in polygenic forms of CD. Additionally, common variants of NOD2 and LACC1 have been implicated in susceptibility to leprosy, a granulomatous infection. Although no specific mechanistic link exists between these three genes, they form an intriguing web of susceptibility to both monogenic and polygenic autoinflammatory and granulomatous phenotypes.

Introduction

A granuloma is an inflammatory lesion that is marked by the aggregation of immune cells, including macrophages and lymphocytes, in response to non-degradable antigen (1, 2). Most granulomas form in response to infectious agents, where they develop as a physiological response to contain and prevent the spread of the causative organisms. Mycobacterial infection is the most common cause of granulomas and, in the case of tuberculosis, patients typically develop caseating granulomas that display features of necrosis. Granulomas may also develop in other circumstances, including in response to foreign bodies, in immunodeficiency disorders and in chronic inflammatory diseases, and in these cases the granulomas are not associated with infectious agents. In contrast to the granulomas of tuberculosis, most granulomas that develop in the context of non-infectious inflammatory diseases are non-caseating.

One group of inflammatory diseases that prominently exhibit granuloma formation is the granulomatous autoinflammatory diseases. The concept of autoinflammation emerged as a term to describe diseases that involved chronic, recurrent inflammation, in the absence of autoreactive lymphocytes or auto-antibodies that can occur without identifiable triggers (3). In practice, autoinflammatory diseases involve excessive inflammatory responses that arise in the context of activating molecular lesions of the innate immune system (4). Among the growing family of autoinflammatory diseases, there are several members in which granuloma formation is a prominent feature (1–3). Blau syndrome and early-onset sarcoidosis (EOS) both refer to granulomatous autoinflammation caused by mutations of nucleotide binding oligomerization domain containing protein 2 (NOD2), which encodes a pattern-recognition receptor. A second group of autoinflammatory diseases that display granulomas are phospholipase Cγ2 (PLCγ2)-associated antibody deficiency and immune dysregulation (PLAID) and autoinflammatory PLAID (APLAID), which are both caused by activating mutations of PLCG2 (which encodes PLCγ2). Finally, mutations in LACC1 (laccase domain-containing protein 1), which encodes the fatty acid metabolism-immune nexus (FAMIN) protein, have recently been recognized as a monogenic cause of both autoinflammation and granulomatous disease. In this review, we will discuss the clinical and genetic aspects of monogenic granulomatous autoinflammatory diseases and highlight the relationship of their causative genes with polygenic granulomatous conditions.

Inflammatory granuloma formation

Unlike infectious granulomas, the impetus for inflammatory granuloma formation is unclear. In the prototypical granulomatous diseases of sarcoidosis and Crohn’s disease (CD), cytokine stimulation leads to macrophage migration into a site of inflammation (5, 6). These macrophages produce tumor necrosis factor α (TNF-α), which recruits additional macrophages and lymphocytes to the area (6). It is hypothesized that M1 (pro-inflammatory) macrophages, activated by Toll-like receptor (TLR) ligands and interferon (IFN)-γ produced by T_h1 cells, predominate in the acute granulomatous process. Over time, the lesion undergoes tissue remodeling and becomes increasingly fibrotic, which is marked by a shift in macrophage polarity towards an M2 (remodeling/fibrosing) subtype (7, 8). Macrophages develop into epithelioid cells which eventually coalesce into multinucleated giant cells that secrete potent cytokines, including TNF-α, interleukin (IL)-1 and tumor growth factor-β (9, 10). These inflammatory factors attract CD4⁺ helper T cells, which help to further organize the granuloma. T_h1 cells are particularly responsive to stimulation by IFN-γ and IL-12, and once recruited to the site of inflammation, they secrete IL-2 to stimulate T-cell proliferation, as well as additional IFN-γ which perpetuates macrophage activation and amplifies macrophage TNF-α secretion (6). As the granuloma matures, T-cell polarity shifts towards T_h2 predominance, which is believed to contribute to increased fibrosis (9).

Recently, circumstantial evidence supporting a role for humoral immunity in granuloma formation has begun to accumulate. B lymphocytes are often seen at the periphery of both infectious and non-infectious granulomas, but sometimes B cells extend throughout the granuloma (9). Epithelioid cells of human sarcoid granulomas have been shown to express B-cell-activating factor (BAFF), a molecule that promotes survival and function of B lymphocytes (11). BAFF expression within granulomas suggests that B cells are involved in the pathogenesis of these lesions. Additionally, IgA-producing plasma cells are often found in or around granulomas (12). IgA-producing plasma cells are also known to secrete TNF-α, a cytokine closely linked to granuloma formation (13). Taken together, these observations suggest that altered B-cell behavior may contribute to granuloma formation and granulomatous inflammation.

Blau syndrome and EOS

Sarcoidosis is a systemic condition that is defined by non-caseating granulomatous inflammation of multiple organs with a predilection for pulmonary involvement. Although sarcoidosis is typically a polygenic condition, mutations in NOD2 cause a monogenic form of sarcoidosis. When present in families, this condition is called Blau syndrome, whereas sporadic cases are called EOS (14–16). Given that these conditions are clinically indistinguishable, we will refer to them henceforth as Blau/EOS. Blau/EOS is characterized by a core set of symptoms that include ocular, synovial and cutaneous granulomatosis (17, 18). Unlike adult and pediatric sarcoidosis, pulmonary involvement in NOD2-related disease is unusual (19).

Blau/EOS presents in early childhood, usually before 4 years of age, and its initial manifestation is often a rash that may range from fine papules to an ichthyosiform eruption. As the disease progresses, the rash may evolve into scaly, hyperpigmented patches. Biopsies of affected tissue characteristically reveal non-caseating epithelioid and giant-cell granulomas with a dense lymphocytic corona (20), and skin lesions are preferential sites for biopsy on the basis of their superior diagnostic yield (21). Articular involvement in Blau/EOS is usually symmetrical peripheral polyarthritis; however, tenosynovitis may also occur (19, 21). The majority of patients develop granulomatous uveitis, which can lead to significant visual impairments (21). Although the skin, synovium and eyes are the classically affected tissues, the granulomatous inflammation of Blau/EOS may manifest in many other ways, including fever, erythema nodosum, large-vessel vasculitis, pulmonary hypertension, systemic hypertension, cranial neuropathies, interstitial lung disease, granulomatous nephritis, hepatic granulomas and lymphadenopathy (18, 22–35).

Blau syndrome was originally mapped to chromosome 16q12.1-13 (36), the same region to which CD had been mapped (37). From this locus, NOD2 was identified as the first susceptibility gene for CD (38, 39) and missense mutations of the nucleotide-binding domain of NOD2 were found in families with Blau syndrome (14, 21, 40). Blau/EOS is most frequently caused by missense mutations of position 334 (p.Arg334Trp or p.Arg334Gln), though many other causative mutations have been reported (14, 22, 27, 31, 35, 41–43). Most of these mutations demonstrate complete penetrance; however, one mutation, p.Glu383Lys, has been reported to produce Blau syndrome with incomplete penetrance in some families (44–46).

NOD2 is principally expressed in monocytes, granulocytes, dendritic cells and Paneth cells (47–50), where it encodes the NOD2 protein. NOD2 is composed of a caspase-recruitment domain (CARD), a central nucleotide-oligomerization domain (called NACHT as it contains the proteins NAIP, CIITA, HET-E and TEP1) and nine carboxy-terminal leucine-rich repeats (LRRs) (51). Blau syndrome mutations primarily affect the NACHT domain, whereas CD risk variants are found in the LRRs (52). NOD2 belongs to the family of NOD-like receptors, a family of pattern-recognition receptors that are critical components of innate immune surveillance. At rest, NOD2 is autoinhibited; however, upon exposure to bacterial components, namely muramyl dipeptide (MDP), it unfolds, oligomerizes and is activated. This leads to activation of a serine/threonine kinase, receptor-interacting protein kinase 2 (RIP2), which in turn activates NF-κB and MAPK, resulting in expression of pro-inflammatory genes (51, 53).

It has been proposed that, in Blau/EOS, gain-of-function mutations in the NACHT domain permit spontaneous oligomerization and activation of NOD2, promoting downstream signaling and inflammation in the absence of bacterial stimuli. In support of this hypothesis, over-expression of disease-causing mutations in cell lines has been shown to cause enhanced NF-κB activation (40, 54). Furthermore, a study of Blau/EOS patients found increased phosphorylation and activation of RIP2 in patients’ PBMCs, relative to the PBMCs of controls subjects (55).

In addition to the observations that support the ‘gain-of-function’ hypothesis, some evidence suggests that Blau/EOS may instead be caused by loss-of-function NOD2 mutations. Several studies of peripheral blood cells from Blau/EOS patients have failed to find increased production of inflammatory cytokines, either at rest or upon stimulation with MDP, calling into question the gain-of-function effect of Blau/EOS mutations (56–58). Additionally, the severity of disease has not been reliably predicted from in vitro studies of increased basal NF-κB activity (59). Further adding to the ambiguity is a murine model in which the murine equivalent (p.Arg314Gln) of the most common Blau/EOS mutation (p.Arg334Gln) was introduced by knock-in (60). Human and murine NOD2 share nearly 80% homology on the amino acid level, including conservation at this and most sites of Blau/EOS mutations (61). It is, therefore, surprising that unlike Blau/EOS patients who bear this mutation, the p.Arg314Gln mouse model does not develop a spontaneous phenotype. However, upon stimulation with MDP the mutant mice produced lower circulating levels of inflammatory cytokines than their wild-type littermates. Similarly, MDP stimulation of macrophages isolated from mutant mice revealed decreased RIP2, MAPK and NF-κB activation, as well as reduced cytokine production, relative to wild-type mice. Macrophages from Blau/EOS patients demonstrated a comparably reduced response to MDP stimulation, relative to macrophages from healthy subjects (58). These findings seem to suggest that Blau/EOS does not arise from over-activation of NOD2, but rather results from loss of NOD2 function, similar to CD. Importantly, the lack of a spontaneous inflammatory phenotype suggests that the p.Arg314Gln mouse model is an imperfect representation of human NOD2 function. Therefore, caution must be exercised when attempting to generalize observations made in this model to human subjects carrying analogous mutations. Ultimately, the mechanism by which NOD2 mutations produce granulomatous inflammation in Blau/EOS remains to be elucidated.

There is no gold standard treatment for Blau/EOS, but most clinicians use glucocorticoids, immunosuppressants and agents that block TNF-α as the mainstays of treatment (30). Additionally, treatments directed against IL-1 have been employed. There are conflicting reports about the efficacy of recombinant IL-1-receptor antagonist (IL-1RA; anakinra) in Blau syndrome (22, 56); however, there is a report describing the successful use of the anti-IL-1β antibody, canakinumab (61). Because of the lack of a universally effective treatment strategy, patients often struggle with persistent arthritic and ocular complications and are at risk for multi-organ involvement.

PLCγ2-associated antibody deficiency and immune dysregulation

PLAID is an autosomal dominant condition with a broad constellation of clinical and laboratory features (62). It is characterized universally by cold urticaria, together with antibody deficiency and immune dysregulation manifesting as susceptibility to recurrent infection, atopic disease, autoimmunity and cutaneous granulomatous inflammation. Many clinical features of PLAID correspond to abnormalities of specific immune cells that signal through PLCγ2, such as cold urticaria (mast cells), antibody deficiency (B cells), autoimmunity (NK and B cells) and susceptibility to infection (NK and B cells). Signaling abnormalities in macrophages and neutrophils may also contribute to the pathogenesis of granulomatosis in PLAID (63).

The cold urticaria that develops in PLAID patients is provoked by evaporative cooling, leading to pruritic, localized erythematous skin lesions (62, 63). Patients frequently report that urticarial episodes develop during or following swimming or bathing, or upon exposure to cool air while perspiring. The cold urticaria of PLAID is not produced by direct cold stimulation or ice cube testing; however, some patients report that the consumption of frozen foods, such as ice cream, causes a burning sensation in the throat or chest. Histologic examination of biopsies of the cold-induced skin lesions of PLAID demonstrates the presence of mast cells and positive staining for tryptase, confirming its urticarial nature.

Immune deficiency with increased susceptibility to infection is commonly observed in PLAID, where antibody deficiency and B lymphocyte abnormalities are the most common features (62). Quantitatively, most patients have low serum IgG, IgM and/or IgA levels. Qualitatively, many PLAID patients demonstrate reduced antibody responses to specific stimuli, such as pneumococcal antigens. PLAID patients have normal numbers of T cells, but almost all have low numbers of circulating class-switched, CD27⁺ memory B lymphocytes. The majority of subjects also have decreased numbers of NK cells. Nearly half of PLAID patients develop recurrent sinopulmonary infections, while recurrent viral upper respiratory infections, viral pharyngitis, early-onset shingles and onychomycosis have also been observed in multiple patients.

Additionally, 3 of 27 PLAID patients had been diagnosed with common variable immune deficiency (CVID) and had been treated with intravenous immunoglobulin for severe, recurrent pneumoniae with bronchiectasis. This is important because of the striking phenotypic overlap between PLAID and CVID. Both conditions are marked by antibody deficiency, together with susceptibility to infection, autoimmunity and granulomatous disease (62, 64). Moreover, both conditions are linked to defects of the adaptive immune system (65). Collectively, this raises the possibility of shared pathophysiology between PLAID and other forms of CVID.

The immune dysregulation of PLAID extends beyond recurrent infections. Symptomatic allergic disease is present in most people with PLAID, including asthma, eczema, allergic rhinitis and food allergies. Autoimmunity is also commonly found, with two-thirds of patients having positive antinuclear antibody testing. A quarter of patients have symptomatic autoimmunity, including vitiligo and autoimmune thyroid disease, and cutaneous granulomatous lesions were also present in a quarter of patients. Among those with granulomatous skin lesions, many developed lesions of the nose, ears and fingers within the first few days of life. Although the skin lesions eventually resolve in most subjects, the rash sometimes persisted and worsened, resulting in soft-tissue damage and destruction of nasal and auricular cartilage. Additionally, in some cases granulomatous dermatitis developed later in life, primarily in areas that could be exposed to cold.

The composition of the granulomatous plaques from PLAID patients is analogous to that of granulomas in sarcoidosis, CVID and other inflammatory conditions (63). Histologically, PLAID granulomas have a core of CD68⁺ histiocytes and multinucleated giant cells encased by a mild lymphocytic infiltrate (62, 63). It has been proposed that dysregulation of macrophages and/or neutrophils may contribute to granuloma formation in inflammatory diseases (63). In PLAID, neutrophils and monocytes from patients become spontaneously activated by cold exposure (63). Given that the granulomas of PLAID occur predominantly in the coldest parts of the body, it is hypothesized that chronic cold-induced activation of macrophages and/or neutrophils in the skin underlies the cutaneous granulomas of PLAID (63).

PLAID has been identified in nearly 50 members of three large, unrelated multi-generational families (62), and within these families it is fully penetrant. PLAID is caused by heterozygous genomic deletions of PLCG2, which encodes PLCγ2, an enzyme predominantly expressed in lymphoid and myeloid cells. In addition to these three families, other families with features clinically consistent with PLAID but without deletions of PLCG2 have been identified (63, 66). Among the families with PLCG2 deletions, each one was different, including two distinct deletions of exon 19 and a deletion that spans from exons 20 to 22. Each of the deletions produced a PLCG2 transcript with an in-frame deletion that alters the carboxy-terminal Src homology 2 domain (cSH2), a critical component of the PLCγ2 autoinhibitory domain (67).

PLCγ2 catalyzes the hydrolysis of phosphatidylinositol bisphosphate (PIP2) to inositol triphosphate (IP3) and diacylglycerol (DAG), which leads to the mobilization of calcium stores from the endoplasmic reticulum and facilitates downstream cellular activation. PLAID-causing mutations of PLCγ2 disrupt its autoinhibitory domain. In wild-type PLCγ2, a series of phosphorylation events produces an allosteric change that unveils the catalytic site and the phospholipase enzyme becomes active (67). In the presence of PLAID-causing deletions, phospholipase activity is decoupled from the activating phosphorylation events, producing constitutive enzymatic activity, in vitro (62, 63). Paradoxically, B cells, NK cells and mast cells bearing these mutations each demonstrate an anergic phenotype in response to conventional stimuli at physiologic temperatures. However, when exposed to subphysiologic temperatures, cells bearing PLCG2 deletions demonstrate spontaneous activation, which produces their unifying phenotype, cold urticaria (62, 63).

There is no effective targeted pharmacotherapy for PLAID, and therefore the approach to its management is directed towards specific symptoms. Lifestyle modification with cold avoidance is the mainstay of treatment for the cold-associated symptoms of PLAID. Antihistamines may provide symptomatic relief for atopic features. Similarly, immunoglobulin replacement may be used to treat immune deficiency in patients with antibody deficiency and recurrent infection.

Autoinflammatory PLCγ2-associated antibody deficiency and immune dysregulation

In the wake of the discovery of PLAID, a point mutation in PLCG2 was identified as the cause of another granulomatous autoinflammatory disease (68). This disease is referred to as APLAID and affected individuals share several important clinical features with PLAID patients, including antibody deficiency with reduced levels of IgM and IgA, striking reductions in circulating class-switched CD27⁺ memory B lymphocytes, reduced antibody responses to specific stimuli (i.e. pneumococcal antigens), recurrent bacterial and viral infections and early-onset granulomatous inflammation of the skin. Additionally, APLAID patients develop episodes of inflammation marked by elevations of the C-reactive protein and erythrocyte sedimentation rate, early-onset colitis, interstitial pneumonitis and posterior uveitis. Unlike PLAID, these patients did not manifest cold urticaria or clinically relevant atopic or autoimmune disease.

APLAID was identified by exome sequencing of an affected father and daughter, both of whom carried the novel p.Ser707Tyr mutation of PLCG2. Even though this mutation is in the cSH2 domain of PLCγ2, the same domain affected by the mutations that cause PLAID, it differs from the PLAID mutants in that it does not lead to constitutive enzymatic activity and the enzyme requires upstream signaling events for its activation (68). Cells bearing the p.Ser707Tyr mutation do not become spontaneously activated when exposed to cold environments, but upon physiologic stimulation, this mutation leads to hypermorphic enzymatic activity in vitro (68, 69). Unlike the anergic responses caused by PLAID mutations, cells bearing p.Ser707Tyr display hyper-responsiveness to stimulation in vitro and ex vivo. Despite this functional difference, APLAID patients manifest many of the same B-cell abnormalities as people with PLAID, including quantitative and specific antibody deficiency and low class-switched CD27⁺ memory B cells.

Like PLAID, there is not an effective treatment for APLAID. Immunoglobulin replacement has been shown to reduce the frequency of infections; however, patients continued to develop new inflammatory features despite this treatment. Moreover, the inflammatory features of APLAID have not responded to either TNF-α blockers or IL-1-directed agents.

Systemic juvenile idiopathic arthritis

Systemic juvenile idiopathic arthritis (sJIA) is a genetically complex inflammatory disease of childhood that is characterized by the presence of chronic arthritis and prolonged episodes of fever, together with an evanescent salmon pink skin rash, lymphadenopathy, hepatosplenomegaly and serositis (70).

Recent studies have demonstrated that sJIA risk is influenced by a variety of genetic factors (71–73); however, sJIA does not typically display Mendelian inheritance. Despite this fact, an autosomal recessive form of sJIA was recently described in five consanguineous families from the southern province of Saudi Arabia (74). Using a combination of linkage analysis and whole-exome sequencing, investigators discovered homozygous p.Cys284Arg mutations in the protein encoded by LACC1 in every affected subject. Since that report, two additional LACC1 mutations (p.Cys43Tyr fs*6 and p.Thr276Lys fs*2) have been identified as causes of recessively inherited juvenile idiopathic arthritis, including not only sJIA, but also rheumatoid factor negative polyarthritis (75) and extended oligoarthritis (76). All affected children with LACC1 mutations and JIA displayed significant biochemical evidence of inflammation (74–76). Beyond rare mutations of LACC1, a recent study also reported an association between common genetic variants of LACC1 and non-systemic forms of juvenile idiopathic arthritis, including p.Ile254Val (77).

The function of the protein encoded by LACC1 was unknown at the time of the initial report; however, subsequent studies have provided some insight into its function. Investigators have shown that the protein, which they called FAMIN, is a peroxisomal protein expressed most highly in neutrophils, monocytes, macrophages and dendritic cells (78–80). FAMIN critically regulates mitochondrial energetics (78), potentially through its influence on peroxisome proliferator-activated receptors (PPAR) signaling (79). It is closely tied with the NOD2 pathway, and its production helps to generate intracellular signaling and cytokine secretion (80). Loss of FAMIN function negatively impacts both fatty acid metabolism and glycolysis, resulting in impairment of reactive oxygen species generation, inflammasome activation and bacterial clearance by macrophages (78). Through this metabolic dysregulation, FAMIN loss-of-function mutants lead to detrimental effects on the function of M0, M1 and M2 macrophages. The p.Cys284Arg mutation (identified in the autosomal recessive form of sJIA) and the p.Ile254Val mutation (implicated in non-systemic forms of JIA) have both been evaluated to determine their effects on FAMIN function (78–80). The p.Cys284Arg mutation leads to a near complete loss of function, while the p.IleI254Val polymorphism results in diminished function of FAMIN (78–80). Additional studies looking at the p.Ile254Val mutation found that this substitution led to decreased downstream products of NOD2 activation, including reduced NF-κB and AP-1 transcriptional activity (80). Together, these results begin to demonstrate how mutations in LACC1 contribute to the pathophysiology of sJIA and integrate fatty acid metabolism and innate immunity.

Historically, the treatment of sJIA has involved the early use of glucocorticoids to extinguish the inflammation, together with one or more disease-modifying anti-rheumatic drugs (81). However, aside from glucocorticoids, there is a lack of evidence that any of these treatments were effective in treating children with sJIA. More recently, anti-cytokine agents directed against IL-1 and IL-6 have been shown to abort inflammation and improve arthritis in some children with sJIA, importantly doing so without the need for co-administration of glucocorticoids (82–86). Moreover, biomarkers capable of prospectively guiding the personalized treatment of sJIA are beginning to emerge. Despite the apparent effectiveness of these agents and approaches, as many as half of children with sJIA will still develop chronic arthritis, highlighting the importance of ongoing efforts to elucidate the pathophysiologic mechanisms of sJIA (87). The form of sJIA caused by LACC1 mutations is particularly recalcitrant to treatment. Among the cohort with p.Cys284Arg mutations, all affected children received treatment with NSAIDS, steroids, methotrexate and etanercept, but no subject responded adequately. They were subsequently treated with traditional and biologic DMARDs, including adalimumab, rituximab and tocilizumab, but again no patient achieved remission. Instead they each progressively developed erosive, destructive arthritis (74). The sJIA patient with the p.Thr276Lys fs*2 mutation was similarly difficult to treat (76).

Although it is not a granulomatous autoinflammatory disease, sJIA is germane to this discussion because the same LACC1 mutation that caused the recessive form of sJIA was previously found to cause a recessively inherited form of early-onset CD—which is a granulomatous disease (88). It is interesting that the same mutation in LACC1/FAMIN, p.Cys284Arg, has been found to cause two immune-mediated diseases with apparently disparate phenotypes. Intriguingly, there are two reports that collectively describe 16 children in whom sJIA and CD co-occur (89, 90). The observation that sJIA and CD are coexistent in some patients, together with the fact that monogenic forms of both diseases are caused by the same recessive mutation of LACC1, suggests that despite their phenotypic differences, sJIA and CD may share elements of pathophysiology. In fact, the possibility of shared pathophysiology among this group of genes and diseases extends beyond sJIA and CD.

An unexpected triad: autoinflammation, inflammatory bowel disease and susceptibility to leprosy

In addition to causing monogenic granulomatous/autoinflammatory diseases, variants of NOD2, LACC1 and PLCG2 each influence susceptibility to CD, a polygenic granulomatous inflammatory disease (Table 1). NOD2 variants are the strongest genetic risk factors for CD, and the relationship of both common and rare NOD2 variants with CD has been extensively studied (39, 91, 92). LACC1 variation has also been implicated in CD susceptibility, with a rare LACC1 mutation causing monogenic CD (88) and several common variants influencing susceptibility to the more typical, polygenic disease (93–95). Similarly, common variants of PLCG2 are associated with polygenic inflammatory bowel disease (96) and a rare PLCG2 mutation causes a monogenic form of CD (68).

Table 1.

Mutations in NOD2, PLCG2 and LACC1 that are associated with granulomatous/autoinflammatory diseases, Crohn’s disease and leprosy

Genes	Granulomatous/autoinflammatory disease	Crohn’s disease	Leprosy (97–99)
NOD2	Blau syndrome (AD) (14, 22, 27, 31, 35, 41–43)	Polygenic Crohn’s disease (39, 91, 92)	rs9302752
	p.Arg334Trp	rs2066844 (p.Arg702Trp)	rs7194886
	p.Arg334Gln	rs2066845 (p.Gly908Arg)	rs1981760
	p.Glu383Lys	rs2066847 (p.Leu1007Pro fs)
	Many other variants
PLCG2	PLAID (AD) (62)	Monogenic Crohn’s disease (AD) (68)
	∆19 PLCG2, ∆20–22 PLCG2	p.Ser707Tyr
	APLAID (AD) (68)	Polygenic Crohn’s disease (96)
	p.Ser707Tyr	rs141548656 (p.His244Arg)
		rs17537869 (p.Arg268Trp)
LACC1	Systemic juvenile idiopathic arthritis (AR) (74, 76)	Monogenic Crohn’s disease (AR) (88)	rs3764147 (p.Ile254Val)
	p.Cys284Arg	p.Cys284Arg
	p.Thr276Lys fs*2	Polygenic Crohn’s disease (95)
	Non-systemic juvenile idiopathic arthritis (AR) (75, 76)	rs3764147 (p.Ile254Val)
	p.Cys43Tyr fs*6
	p.Thr276Lys fs*2
	Non-systemic juvenile idiopathic arthritis (polygenic) (77)
	rs3764147 (p.Ile254Val)
	rs3816311
	rs4942255
	rs9533685

AD, autosomal dominant; AR, autosomal recessive.

Each of these genes is also linked to leprosy, a chronic granulomatous infection caused by Mycobacterium leprae. In leprosy, one’s ability to contain the mycobacterial infection through granuloma formation is a critical determinant of the disease course, with paucibacillary (localized) disease resulting from successful containment and multibacillary (disseminated) disease resulting from what may amount to failed granuloma formation. Common genetic variants of both NOD2 and LACC1 are also associated with leprosy susceptibility (97–99), conferring greater risk of developing the disseminated form (99). Additionally, a study comparing localized and disseminated leprosy found that PLCG2 expression is much higher in localized leprosy than in the disseminated form (100), suggesting a role for PLCγ2 in granuloma formation and successful anti-mycobacterial host response.

The triangular relationship of each of these genes with monogenic autoinflammation, CD and leprosy is fascinating. Despite an absence of specific pathways or mechanisms that unify these three genes, their collective involvement in these granulomatous conditions suggests an element of shared pathophysiology. Indeed, each of these genes is involved in the regulation of innate immunity. NOD2 is a pattern-recognition receptor that recognizes components of mycobacterial cell walls, and its activation leads to pro-inflammatory cytokine and chemokine production. FAMIN, encoded by LACC1, is another important regulator of innate immunity through its recently recognized metabolic and energetic control of myeloid cells. Similarly, PLCγ2 regulates innate immune signaling in several ways, including the transmission of signals from extracellular receptors like dectin-1 and -2 in myeloid cells (101, 102).

One may hypothesize that each of these gene products represents a ‘molecular switch’ that controls some aspect of the myeloid responses that leads to granuloma formation. For example, the variants associated with disseminated leprosy may hinder the host’s ability to effectively form granulomas, thereby allowing mycobacterial dissemination. In contrast, the variants associated with inflammatory diseases may act to promote excessive granuloma formation. In the case of CD, which is thought to require both genetic and environmental contributions, risk variants may lead to exaggerated granuloma formation in response to otherwise normal exogenous triggers (69, 80, 103).

Finally, in monogenic granulomatous autoinflammatory diseases, the granulomas may develop spontaneously in the absence of the normal exogenous triggers. If this is indeed the case, then perhaps the key to effectively treating granulomatous autoinflammation lies in the ability to therapeutically target the aberrant signaling events upstream of granuloma formation, instead of attempting to block the ubiquitous downstream inflammatory pathways. Perhaps the unified investigation of these seemingly unrelated conditions could provide novel insight into the (patho)physiology of granuloma formation and, in turn, produce targeted therapeutic strategies for the full spectrum of granulomatous disease.

Conclusions

Granuloma formation is a pathologic phenomenon frequently identified in both infectious and autoinflammatory conditions. In this review, we discussed the phenotypic features of monogenic granulomatous autoinflammatory diseases and reviewed their relationship with mutations in NOD2, PLCG2 and LACC1. We also critically examined the genetic overlap of these genes with sJIA, CD and leprosy. The involvement of these three genes in seemingly disparate presentations of granulomatous disease is intriguing, and further study is needed to determine if a common pathophysiology exists. Such studies could lead to improved treatment options for patients and may help to prevent disability related to uncontrolled granuloma formation.

Conflicts of interest statement: the authors declared no conflicts of interest.