CARD9 in the pathogenesis of axial spondyloarthritis

Abstract

Axial spondyloarthritis (axSpA) has been long classified as an autoimmune disease caused by a breakdown in the ability of the immune system to delineate self from foreign, resulting in self-reactive T cells. The strong genetic association of HLA-B27 supports this role for T cells. More recently, genetic and clinical studies indicate a prominent role of the environment in triggering axSpA, including an important role for microbes and the innate immune response. As an example, mutations in genes associated with innate immunity, including the anti-fungal signaling molecule Caspase recruitment domain-containing protein 9 (CARD9), have been linked to axSpA susceptibility. Thus, current thought classifies axSpA as a “mixed pattern condition” caused by both autoimmune and autoinflammatory mechanisms.

The goal of this review is to convey:

• Genetic/environmental mediating factors in axSpA

• Known roles for CARD9 in anti-fungal immunity versus sterile inflammation

• Previously characterized neutrophil-intrinsic roles for CARD9

• Studies supporting a role for CARD9^S12N mutation in promoting axSpA

1. Axial spondyloarthritis

Axial spondyloarthritis (axSpA) is a chronic inflammatory disease of the spine that can result in permanent fusion of the vertebrae. The prevalence of axSpA in the United States is 1%, with similar incidence rates between men and women [1]. This disease is comprised of an earlier or less severe form, known as non-radiographic axial spondyloarthritis, and a more severe form, known as radiographic axial spondyloarthritis (r-axSpA, previously referred to as ankylosing spondylitis [AS]) [2]. For many of the findings described in this review, it is unclear whether they apply to a particular sub-category of axSpA or the broader disease state. For this reason, we have elected to employ the more general term axSpA.

AxSpA is part of a genetically and clinically related group of rheumatologic diseases termed “spondyloarthritis” (SpA) [3]. While axSpA may be the most common form of SpA, other related disease types include a broad spectrum of clinical manifestations, namely, psoriatic arthritis (PsA), enteropathic arthritis (associated with inflammatory bowel disease [IBD]), and reactive arthritis, among others. AxSpA patients often present with extra-articular (non-joint) manifestations including 22–37% with anterior uveitis, 4–16% with IBD and 4–9% with psoriasis (Ps) [4]. Comorbidities associated with axSpA include but are not limited to osteoporosis, cardiovascular disease, hypertension, malignancy and infections [5]. Given the heterogeneity of disease between patients and overlapping clinical manifestations, axSpA has been historically challenging to diagnose. Consequently, the prevailing view is that individuals with this condition are likely underdiagnosed; a considerable quantity of evidence suggests that these patients also face prolonged delays in their diagnosis [6].

2. Complex interactions between the environment and genetics drive axSpA pathogenesis

The cause of axSpA remains unknown. Monogenetic twin studies indicate that heritable genetic factors including mutations contribute to up to 30% of axSpA disease origin [3]. Genome wide association studies (GWAS) indicate that 85–90% of AS patients express the human leukocyte antigen (HLA) class I B27 allele, making it arguably the most useful diagnostic biomarker available in the clinic [7]. HLA-B27 is a class of antigen presenting molecules expressed on the surface of most cell types in the human body that are critical for presentation of foreign antigen and subsequent induction of protective T cell responses. How HLA-B27 contributes to disease is unknown, but one hypothesis proposes that due to its protein structure, HLA-B27 can uniquely present self-peptides or microbial-peptides (including microbial antigens that are “molecular mimics” of self-proteins) and inadvertently activate self-targeting T cells that cause arthritis [8]. Thus, the strong connection between HLA-B27 and the risk of axSpA suggests self-reactive (i.e., autoreactive) T cells play a major role in pathogenesis.

GWAS have more recently described ~114 additional non-HLA genetic factors associated with axSpA risk, which are thought to contribute to 20% of genetic predisposition of the disease [9]. Several of these factors represent genetic mutations in genes that encode modulators of the innate immune response including cytokines and microbial signaling molecules (e.g., CARD9) [10]. The pro-inflammatory cytokines interleukin 17 (IL-17) and tumor necrosis factor (TNF) contribute to axSpA, but unlike TNF, which is a general pro-inflammatory cytokine that drives the pathogenesis of many diseases, IL-17 appears to be specific to axSpA and related SpA conditions. Human studies have identified a clear pathogenic role for CD4⁺ T helper cells that produce IL-17 (Th17 cells) in SpA. Elevated levels of IL-17 have been detected in the serum and synovial fluid of axSpA patients and Th17 cells are increased in blood compared to other rheumatic conditions, including rheumatoid arthritis [11]. Recent clinical trials have shown that IL-17 inhibition has substantial efficacy in ameliorating symptoms, proving IL-17 is key to the pathophysiology of axSpA [12].

Additional agents contributing to the etiology of disease consist of a host of environmental factors, including microbes and the microbiome, smoking, and mechanical trauma [13,14]. It is now generally agreed that neither genetics nor the environment alone are sufficient to cause disease, but rather a synergistic interaction between the two causes axSpA (Fig. 1). Thus, axSpA is a complex polygenetic disease caused by an interplay of several different unique patient-specific genetic combinations and environmental triggers. These interactions create the “perfect storm” wherein a break in immune homeostasis leads to generation of self-reactive T cells, pro-inflammatory cytokine production, and hyperactive innate cellular responses—a phenomenon that propagates chronic sterile inflammation of the spine and peripheral tissues in axSpA (Fig. 1). Understanding how different genetic combinations and environmental factors result in axSpA may reveal novel diagnostic avenues and drug targets for future therapies. This review focuses on one such example of how new data in the field provides mechanistic insight that implicates the anti-microbial signaling molecule CARD9 as a putative contributor to axSpA pathogenesis.

Fig. 1. Overview of the pathogenesis of axSpA.

Genetic and clinical studies indicate a role for both the environment and genetic factors in axSpA pathogenesis. This interaction leads to the development of pathogenic T cells that recognize self-antigens (i.e., self-reactive T cells), pro-inflammatory cytokines, and hyper-innate immune responses that cumulatively contribute to inflammation and damage of the spine and other associated tissues.

3. CARD9 is a critical mediator of anti-fungal immunity

CARD9 is an intracellular adaptor protein expressed predominantly by myeloid cells that is critical to protection against fungal infection [15]. CARD9 functions as a major signaling hub downstream of microbial ligand engagement of a class of pattern recognition receptors called C-type lectin receptors (CLR). These receptors include but are not limited to Dectin-1, Dectin-2, and Mincle. Upon activation, CARD9 forms a complex with B-cell lymphoma/leukemia 10 (BCL10) and mucosa-associated lymphoid tissue 1 (MALT1) [15–17] that activates a downstream signaling cascade via the extracellular signal-regulated kinase (ERK) and the transcription factors nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and mitogen-activated protein kinase (MAPK). This then results in production of the pro-inflammatory cytokines TNF, interleukin 6 (IL-6), interleukin 12 (IL-12), interleukin 1 beta (IL-1β), and granulocyte-macrophage colony stimulating factor (GM-CSF) [15,17–19]. Although CARD9 functions downstream of several CLRs that interact with various fungal and bacterial ligands, patients with rare CARD9 loss-of-function (LoF) mutations are rendered immunocompromised with extreme susceptibility to invasive fungal infections caused by two specific opportunistic fungal pathogens, Candida albicans and Aspergillus fumigatus [15].

Neutrophils and Th17 cells are critical for antifungal immunity. Several studies on neutrophils from primary CARD9-deficient patients and Card9-deficient mice suggest that LoF of CARD9 results in significant defects in neutrophil trafficking to the central nervous system, neutrophil phagocytosis of fungi, pro-inflammatory cytokine production, and Th17 immunity [15]. These data present an unequivocal link between CARD9 and fungal immunity as well as CARD9 and neutrophil/Th17 responses. Combining these concepts leads to the intriguing notion of CARD9 as a pathogenic activator of axSpA, especially given that this disease involves a robust activation of innate immunity and Th17 cells reminiscent of fungal infections.

4. CARD9 is a genetic determinant of axSpA

Single nucleotide polymorphisms (SNPs) in CARD9 have been shown in several GWAS and clinical studies to confer significant risk for axSpA (Table 1) [10,20–25]. In particular, the mutation CARD9 rs4077515 that results in amino acid substitution of asparagine at position 12 for serine (CARD9^S12N) is possessed by 72% of axSpA patients and strongly linked to disease. CARD9 S12N is a proposed gain-of-function (GoF) mutation, as it coincides with increased expression of CARD9 mRNA in blood cells [26] and PBMCs [27] and has been linked to increased inflammation in IBD [28] and allergy [29]. The importance of CARD9 in driving inflammation is underscored by a compensatory CARD9 LoF mutation (CARD9^Δ11) shown to be protective against IBD [30]. Although the prevalence of this mutation is high, the penetrance of CARD9 S12N is likely much lower, as it is expressed by 20–40% of the general healthy population [31]. Importantly, unlike CARD9 LoF mutations associated with invasive fungal infections, the axSpA-associated CARD9 S12N mutation are not associated with increased fungal infections [32,33].

Table 1.

CARD9 associations in axial spondyloarthritis.

Reference	CARD9 SNP	Main findings
[20]	rs10781499	Increased comorbidity rates due to pleiotropy not heterogeneity (axSpA, Crohn’s disease, Ps, primary sclerosing cholangitis, and ulcerative colitis).
[10]	rs4077515 and rs3812571 (tagged), rs11794847, rs10781505, rs10781499, rs3812570, rs3812571, rs10870149	Association between axSpA and IBD with CARD9; cis-acting effects on CARD9.
[21]	rs10781500	CARD9 strongly associated with disease pathogenesis, may point to mechanism for HLA-B27; supports β-glucan induced disease (IL-23-IL23R-IL-17 pathway).
[22]	rs11145763	Pediatric-age-of-onset autoimmune diseases.
[23]	rs1128905	CARD9 SNP located at exon-intron boundaries, predicted to influence splicing.
[24]	rs4077515, rs3812571	Chinese Han population shows minimal association between these CARD9 variants in axSpA, in contrast to European populations. Highlights importance of racial diversity in association studies.
[25]	rs4077515	Iranian population: CARD9 associated with axSpA in HLA-B27-positive patients, but protective in HLA-B27-negative patients.

5. Neutrophil-intrinsic CARD9 controls infection and sterile inflammation

Several GWAS that did not examine axSpA susceptibility have associated CARD9^S12N expression with increased neutrophil count and frequency in the blood [34–36], suggesting a positive relationship between CARD9 expression and neutrophilia. In support of this concept, CARD9 LoF in humans result in impaired neutrophil responses and invasive fungal infection. Card9 has been implicated as a neutrophil-intrinsic regulator of sterile inflammatory disease in experimental models, including collagen-induced arthritis [37] and dextran sulfate sodium (DSS)-induced colitis [38]. These data indicate that while the CARD9/neutrophil axis is necessary against fungal infection, dysregulated neutrophil responses induced by CARD9^S12N may contribute to sterile inflammation in axSpA.

Several studies have recently identified neutrophil-intrinsic functions for CARD9 (i.e., CARD9 expressed exclusively within neutrophils) in infection and sterile inflammation, including within human neutrophils and neutrophils from experimental mouse models, as reviewed in Fig. 2. These include neutrophil-intrinsic CARD9 as necessary for fungal killing via phagocytosis and fungal induced Th17 expansion [39–42]. Conversely, neutrophil-intrinsic Card9 promotes fungal β-glucan-induced Th17 responses, SpA in genetically predisposed SKG mice [42], arthritis in immune complex mediated sterile inflammation of the K/BxN serum transfer model and dermatitis in the skin blistering dermatitis mouse model [37].

Fig. 2. Neutrophil-intrinsic CARD9 in fungal infection and sterile inflammation.

Diagram summarizing the known effects of external environmental factors on the function of CARD9-sufficient and CARD9-deficient neutrophils. Immune functions downstream neutrophil-mediated immune responses are detailed following exposure to fungal pathogens (i.e., fungal infection [top panel]) or antibodies bound to antigen (i.e., immune complexes [middle panel]). The bottom panel shows how currently unknown environmental triggers may control neutrophils expressing the CARD9 S12N SNP to promote: disease and IL-17 production in axSpA, IBD, and experimental allergy.

6. The CARD9/neutrophil axis promotes Th17-mediated axSpA

Many studies have reported positive associations between axSpA and increased neutrophil responses. For example, axSpA patients may have dysregulated neutrophil responses, as clinically indicated by neutrophilia and increased neutrophil-to-lymphocyte ratios that positively associate with disease activity [43]. Several studies, including one published by our group using a cohort of U.S. Veterans, provide data suggesting that the neutrophil-to-lymphocyte ratio may provide information relevant to clinical decision-making and management of axSpA [44,45]. Neutrophils have been further implicated in the inflammatory milieu of axSpA, as evidenced by increased presence of the neutrophil activation byproduct, neutrophil extracellular traps (NETs), in the blood of axSpA patients [46]. Neutrophils are reported to have direct effects on basic CD4⁺ T cell function in vitro and in vivo, including a recent study indicating that NETs and associated histones are potent inducers of Th17 responses [47]. In support of a pathogenic role of neutrophils in axSpA initiation, studies conducted in human enthesis and experimental AS (SKG mice) suggest that neutrophils infiltrate the axial and peripheral entheseal sites early in disease and augment inflammation in curdlan-induced arthritis [48,49].

Mechanistically, we recently identified a neutrophil-intrinsic role for Card9 in promoting axSpA using the SKG mouse model [42]. SKG mice—like axSpA patients—have a genetic predisposition to axSpA in part due to the presence of self-reactive T cells. However, while in the case of axSpA patients wherein the environmental trigger of disease is largely unknown, arthritis in SKG mice precipitates following a single exposure to fungal β-glucans. Our data showed that within 5 days of fungal exposure, Card9 expression within neutrophils is required to induce arthritis in SKG mice. We demonstrated that Card9 functions within SKG neutrophils to expand pathogenic self-reactive T cells and that axSpA patient neutrophils were sufficient to expand autologous patient Th17 responses. Thus, fungal β-glucans activate Card9 signaling within neutrophils and neutrophil-intrinsic Card9 propagates pathogenic Th17 responses.

We then hypothesized that if Card9-deficient neutrophils had decreased capacity to induce pathogenic Th17 responses, then a CARD9 GoF mutation (i.e., CARD9 S12N) may enhance Th17/IL-17 responses in patients. Consistent with this hypothesis, we observed that serum from CARD9 S12N expressing axSpA patients had increased serum IL-17A compared to axSpA patients with two non-mutated copies of CARD9 (Fig. 2, bottom panel). These data support a hypothetical model of axSpA pathogenesis wherein multiple genetic factors (HLA-B27 and CARD9 S12N) might predispose patients to development of self-reactive T cells and simultaneous heightened neutrophil responses and that when coupled with a fungal exposure, synergize to drive Th17/IL-17 mediated axSpA (Fig. 3).

Fig. 3. Hypothetical role for CARD9^S12N SNP in axSpA pathogenesis.

We hypothesize that fungal exposure of patients with underlying genetic predisposition to axSpA (e.g., co-expression of HLA-B27 and CARD9^S12N SNP) may lead to the generation of pathogenic immune responses including self-reactive T cells and CARD9-activated inflammatory neutrophils. Downstream of these affects we further propose that an interaction between the self-reactive T cells and CARD9-activated neutrophils would result in development of pathogenic Th17 cells that produce IL-17A and target the enthesis and spine to promote disease in axSpA.

7. Summary

Due to the genetic and clinical heterogeneity of disease, axSpA is notoriously difficult to diagnose and treat. AxSpA is a polygenetic disease and it is likely that many combinations of genetic factors and mutations may work together in the context of an environmental stimuli to provoke chronic sterile inflammation. CARD9 plays a critical function in protection against invasive fungal infection and conversely in promoting sterile inflammation in axSpA. CARD9 is positively associated with inflammatory neutrophil responses and Th17 immunity and CARD9^S12N may enhance these functions. These data support a model whereby additive effects of co-expressing HLA-B27 and CARD9^S12N result in pathogenic Th17 cells that cause axSpA. These findings further support the rising paradigm in the field wherein axSpA is a “mixed pattern” disease caused by both self-reactive T cells (i.e., autoimmune T cellular responses) and pathogenic neutrophil responses (i.e., autoinflammatory innate cellular responses). Understanding how different non-HLA-B27 alleles may function to control disease may lead to robust precision diagnostics for patients with SpA diagnoses and reveal novel therapeutic targets for axSpA.

Practice points

• HLA-B27 has limited clinical utility with a growing number of HLA-B27 negative axSpA patients, so it is important to identify other genetic risk factors, such as CARD9.

• The genetic associations of CARD9, while intriguing and interesting, are not yet applicable in the clinical setting.

Research agenda

• axSpA polygenetic risk scores have shown favorable results for distinguishing axSpA from controls in the research setting but it is currently economically and technically challenging for clinicians.

• More mechanistic research studies aimed at understanding the cellular mechanism for each axSpA -associated mutation will provide insight into future diagnostics and treatment strategies.

• Future studies are needed to understand how neutrophil:T cell interactions contribute to disease in axSpA.

• Neutrophils are key to the pathogenesis of axSpA and more mechanistic studies are urgently needed to understand how to harness these cells for therapeutic reasons.

Abbreviations

AS

ankylosing spondylitis

axSpA

axial spondyloarthritis

BCL10

B-cell lymphoma/leukemia 10

CARD9

caspase recruitment domain-containing protein 9

DSS

dextran sulfate sodium

ERK

extracellular signal-regulated kinase

GM-CSF

granulocyte-macrophage colony-stimulating factor

GoF

gain-of-function

GWAS

genome wide association studies

HLA-B27

human leukocyte antigen class I B27 allele

IBD

inflammatory bowel disease

IL-1β

interleukin 1 beta

IL-6

interleukin 6

IL-12

Interleukin 12

IL-17

interleukin 17

IL-17i

IL-17 inhibitors

K/BxN

Mice expressing both the T cell receptor (TCR) transgene KRN and the MHC class II molecule A(g7)

LoF

loss-of-function

MALT1

mucosa-associated lymphoid tissue 1

MAPK

mitogen-activated protein kinase

NETS

neutrophil extracellular traps

NF-κB

nuclear factor kappa-light-chain-enhancer of activated B cells

Ps

psoriasis

PsA

psoriatic arthritis

PSC

primary sclerosing cholangitis

SNP

single nucleotide polymorphisms

SpA

spondyloarthritis

Th17

T helper 17 cells

TNF

tumor necrosis factor