Immunological pathways affected by SLE susceptibility variants

Current understanding of SLE pathogenesis can group gene products of identified SLE-associated gene variants into potentially influenced mechanistic pathways, such as: (1) DNA degradation, apoptosis and clearance of cellular debris, for example, TREX1 (three prime repair exonuclease 1) and DNASE1 (deoxyribonuclease 1); (2) defective clearance of immune complexes (IC) containing nuclear antigens, for example, complement components, and FCGR (Fc fragment of IgG, low affinity receptors); (3) Toll-like receptor (TLR) and type I interferon (IFN) pathway activation, for example, TLR7, IRF5 (IFN regulatory factor 5) and STAT4 (signal transducer and activator of transcription 4); (4) nuclear factor κB (NFκB) signalling, for example, TNFAIP3 (tumour necrosis factor (TNF) α-induced protein 3; (5) B cell function and signalling, for example, BANK1 (B-cell scaffold protein with ankyrin repeats 1) and BLK (B lymphoid tyrosine kinase); (6) T-cell signalling and function, for example, PTPN22 (protein tyrosine phosphatase, non-receptor type 22) and TNFSF4 (TNF superfamily, member 4); and (7) monocyte and neutrophil signalling and function, for example, ITGAM (integrin, α M); (see table 1 and the text below for a description of additional genes in each pathway).

Genetic variation in DNA degradation, apoptosis and clearance of cellular debris pathways

The proper disposal of intracellular constituents or infectious agents in a regulated manner may function inappropriately in SLE, leading to the abundance of self-antigens. Several variants of genes related to these pathways contribute to both monogenic and polygenic forms of SLE. Recessive mutations typically lead to Aicardi–Gutieres syndrome with deficiency of TREX1—an exonuclease involved in cell death, DNA degradation and cellular response to oxidative damage; deep sequencing of TREX1 identified a novel frameshift or missense mutations in patients with SLE but not controls.7 A large, multiethnic case–control study subsequently confirmed a TREX1 SLE risk haplotype associated with neurological manifestations in patients of European ancestry8 (see table 2 for candidate genes and associated subphenotypes). Similarly, mutations of ACP5 (acid phosphatase 5, tartrate resistant), which cause spondyloenchondrodysplasia due to deficiency of TRAP (tartate-resistant acid phosphatase), a protein that functions in lysosomal digestion, lead to elevated IFN-α activity and a spectrum of autoimmune diseases including SLE.9 Rare mutations of DNASE1, encoding the major nuclease present in serum, urine and secreta and its homolog, DNASE1L3 (deoxyribonuclease 1-like 3) have been identified in several patients with SLE from homogeneous and potentially consanguineous populations.19 ,31 In addition, SLE-associated variants have been described in European and Asian populations of ATG5 (autophagy related 5), encoding a protein that contributes to caspase-dependent apoptosis from FAS and TNF ligands, and degradation of cytoplasmic constituents.32

Genetic variants of IC clearance and phagocytosis pathways

Defective clearance of IC containing nuclear antigens in SLE leads to deposition in target organs. The incidence of SLE or lupus-like manifestations in individuals with a complete deficiency, due to a homozygous mutation, in one of the classic complement pathway genes ranges from 10% to 93% (C1Q and C1R/C1S, >90% penetrance; C4A and C4B, 75%; C2, 10–20%).24 C1Q and C4A, which had previously been described in monogenic forms of SLE, have also been implicated in polygenic SLE and with various clinical phenotypes.24 Genetic variants of CFHR1 and CFHR3 (complement factor H-related genes), which may contribute to alternative complement pathway regulation among other functions, have also been associated with SLE risk in multiple ancestral groups.33

FcγR gene variants with function in IC clearance are relevant in the development of several autoimmune diseases.34 The Fc receptor gene family region is complex and includes gene duplications and copy number variations, creating challenges to the investigation of gene structure. Inconsistencies between FcγR genetic studies in SLE have been attributed to ethnic differences and disease heterogeneity, as well as genotyping error. However, the role of FcγR variants in the risk of SLE is highlighted by several variants, including H131R of FCGR2A, F158V of FCGR3A and I187T of FCGR2B, which have been associated with SLE susceptibility in several ancestral populations, and with specific disease profiles.35 In addition, decreased copy numbers of FCGR3B, correlating with protein expression and IC clearance, is associated with SLE.36 FcγRII and FcγRIII, the low-affinity receptors for the IgG-Fc region, are important in phagocytosis, presentation of complexed antigen and cytokine response after receptor cross-linking. The FcγR are predominantly activating, except FcγRIIB, which can inhibit signalling through other FcγR and the B-cell receptor, neutrophils and macrophages;37 interestingly, a FCGRIIB functional SNP abrogates receptor function in SLE patients of both European and southeast Asian ancestries.38 Further investigation of the functional consequences of FcγR gene variants in SLE is warranted and will help to characterise their contributions to disease pathogenesis.

Genetic variants of the TLR and type I IFN pathway

Increased expression of type I IFN and type I IFN-inducible genes is observed commonly in patients with SLE, suggesting a major role in disease pathogenesis, and leading to the development of anti-IFN-α therapy.39 Likely candidates for triggers of type I IFN activation are the binding of pattern recognition membrane and cytosolic receptors by exogenous viral agents or endogenous nucleic acids. Variants have been associated with the risk of SLE; for example, a functional 3′-untranslated region SNP of the X-linked TLR7 that confers elevated TLR7 expression and an increased IFN response has been associated with SLE in east Asians,17 which was subsequently confirmed in European-American, African-American and Hispanic populations.40

Variation in genes coding for transcription factors downstream of TLR, including IRF5,41 ,42 IRF743 and IRF8,44 ,45 has been associated with SLE susceptibility. Robust associations of four IRF5 functional variants in multiple ancestries define haplotypes associated with increased, decreased, or neutral levels of risk for SLE, with functional consequences on the expression of IRF5, IFN-α and IFN-inducible chemokines.16 Similarly, a non-synonymous IRF7 SNP (Q412R) confers increased IRF7 and downstream IFN pathway activation in European, Asian and African-American patients with SLE;43 additional IRF7 risk alleles in patients with SLE are associated with anti-double-stranded DNA antibodies (European-American and Hispanic-American individuals), and anti-Sm antibodies (African-American and Japanese individuals).12 ,13 IRF8 and susceptibility to SLE in a large multiethnic cohort was recently described.45 IRF8 encodes a transcription factor that acts in the type I IFN pathway, and also plays a role in B-cell and macrophage development;46 however, the causal variant in IRF8 has not yet been identified.

Several additional genes within or downstream of the type I IFN pathway have been associated with the risk of SLE, including STAT4, IFIH1 (IFN induced with helicase C domain 1), TYK2 (tyrosine kinase 2) and PRDM1 (PR domain zinc finger protein 1).44 ,47 IFIH1 detects RNA before type I IFN pathway activation; an allele of IFIH1 was associated with anti-dsDNA antibodies among patients of multiple ancestries with SLE.14 PRDM1 encodes BLIMP-1 that acts as a repressor of IFN-β gene expression, is an essential regulator of T-cell homeostasis, and may regulate both B-cell and T-cell differentiation. TYK2 interacts directly with the type I IFN receptor on engagement with IFN-α or β and contributes to downstream phosphorylation of the STAT family and other transcription factors. STAT4, encoding a protein that promotes transcription after type I IFN receptor activation, has been associated with increased susceptibility to SLE in several ancestral backgrounds,32 ,48 ,49 several subphenotypes and an early age of diagnosis.25 ,15 The genetic control of IFN activity in SLE was recently expanded to include a functional SLE risk variant in MIR146A, encoding a negative regulator of the type I IFN pathway. Decreased levels of miR-146A seen in peripheral blood mononuclear cells from Han Chinese patients with SLE may be due to decreased binding of transcription factor Ets-1 at the MIR146A promoter variant location;50 genetic variation in ETS1 has also been associated with the risk of SLE (see B-cell section below).

Genetic variation of the NFκB pathway

Genes that play a role in the NFκB pathway downstream of TLR engagement have also been associated with increased SLE susceptibility in multiple ancestries. For example, both risk and protective haplotypes of IRAK1 (interleukin-1 receptor-associated kinase 1) have been associated with SLE.51 The X-linked IRAK1 gene encodes a kinase that acts as the Myd88 complex on/off switch for activation of the NFκB inflammatory pathway. TNFAIP3, also associated with SLE and subphenotypes including renal disease,26 ,52 encodes A20, a deubiquitinating enzyme that inhibits NFκB, leading to protein degradation and interactions that inhibit NFκB activity and TNF-mediated programmed death. A dinucleotide polymorphism just downstream of the TNFAIP3 promoter region was linked to the decreased expression of A20 in patients with SLE of Korean and European ancestry, and may be the risk haplotype functional variant.53 TNIP1 (TNFAIP3 interacting protein 1), encoding the A20-interacting protein, has also been associated with the risk of SLE.47 ,48 Additional genes within the NFκB pathway associated with SLE susceptibility include: SLC15A4 (solute carrier family 15, member 4) encoding a peptide transporter that participates in NOD1-dependent NFκB signalling;48 PRKCB (protein kinase C, β), which is involved in B-cell receptor-mediated NFκB activation;54 and UBE2L3 (ubiquitin-conjugating enzyme E2L 3), encoding the enzyme UBCH7, which participates in the ubiquitination of an NFκB precursor, and may play a role in cell proliferation.55 A risk haplotype of UBE2L3 confers increased UBCH7 expression in patients with SLE;56 a variant contained in this haplotype has been associated with the presence of anti-dsDNA antibodies.15

Genetic variation of B-cell signalling and function

DNA or RNA released from dying or damaged cells can be recognised by autoreactive B cells leading to the activation and production of autoantibodies that, together with additional autoantigens, form IC that drive other proinflammatory responses. This critical role of B cells in the development of autoimmunity has led to the development of several targeted therapies, including anti-BLyS (B lymphocyte stimulator) and anti-CD20. Several B-cell-related gene variants are involved in cell signalling and have been associated with SLE susceptibility in multiple ancestral backgrounds. For example, BLK, encoding a protein functioning in intracellular signalling and the regulation of proliferation, differentiation and tolerance of B cells;48 ,49 ,57 and BANK1, whose gene product facilitates the release of intracellular calcium, altering the B-cell activation threshold;49 ,58 ,59 and LYN, encoding a binding partner of BANK1 that mediates B-cell inhibition.29 Three functional variants of BANK1, including R61H, A383T and rs17266594 (affecting alternative splicing) have been identified that contribute to sustained B-cell receptor signalling and B-cell hyperactivity.58

Other genes with roles in B-cell function associated with SLE susceptibility in single ancestry groups include those that encode: ETS1 (Ets-1 protein, or p54), which negatively regulates B-cell and T helper type 17 cell differentiation;59 IKZF1 (IKAROS family zinc finger 1), which regulates lymphocyte differentiation, proliferation and B-cell receptor signalling;44 AFF1 (AF4/FMR2 family member 1), which functions in normal lymphocyte development;60 RASGRP3 (ras guanyl-releasing protein 3), which transmits B-cell signals via Ras-ERK after B-cell receptor ligation, with potential impact on immunoglobulin production and B-cell proliferation;48 interleukin (IL)-21, which sustains antibody production, mediates antibody class switching and promotes differentiation of T helper type 17 cells (association with SLE described in European and Hispanic ancestry);61 and IL-10, which inhibits T cells and antigen-presenting cells while enhancing B-cell survival and activity.47 Ongoing work from our laboratory has identified a functional SNP in the 5′ region of IL10, which is associated with higher messenger RNA and protein expression in SLE.62

Genetic variation of T-cell signalling and function

Hyperactive B cells, resulting from T-cell and antigen stimulation, increase the production of autoantibodies in SLE. T-cell-directed therapies, including CTLA4 (cytotoxic T lymphocyte antigen 4) fusion proteins, have been developed to modulate T-cell costimulation. The production of superoxide by nicotinamide adenine dinucleotide phosphate, reduced form, oxidase in leucocytes stimulated by autoantigens may be affected by an amino acid change (H389Q) in NCF2 (neutrophil cytosolic factor 2), a SLE susceptibility gene encoding a subunit of the oxidase enzyme, implicating a role for decreased reactive oxygen species in SLE pathogenesis.44 ,63 Two genes important to T-cell signalling, R620W of PTPN22 and TNFSF4, have been associated with SLE risk:64 PTPN22 is critical for T-cell signalling; TNFSF4 (or OX40L) induces co-stimulatory signals that induce the activation and differentiation of B cells and inhibit T regulatory cells. Variants of PTPN22 have been associated with both gain of function and loss of function in patients with SLE.65 In addition, CD44 encodes a cell-surface protein that regulates lymphocyte activation and apoptosis, among other functions; specific transcript isoforms in T cells from patients with SLE suggest a role for CD44 in SLE pathogenesis.66 ,67

Genetic variation in monocytes and neutrophil signalling and function

Aberrant activation of monocytes and neutrophils, now recognised as important participants in SLE pathogenesis, includes the function of NET (neutrophil extracellular traps) containing DNA and neutrophil-derived proteins that trigger for IFN-α release and can directly damage tissue.39 Genetic variants related to adhesion and endothelial migration of both cell types have been associated with SLE susceptibility in multiple ancestries, specifically R77H ITGAM and most recently the ICAM (intercellular adhesion molecule) locus.32 ,57 ,68 ,69 ITGAM encodes the α chain of α_Mβ₂ integrin, which regulates neutrophil and monocyte adhesion and migration from the bloodstream via interactions with a wide range of structurally unrelated ligands, including ICAM1 and ICAM2.

Ancestral differences among SLE susceptibility genes

Most SLE-associated loci have been confirmed among multiple ancestries, suggesting that common pathways play a role in disease pathogenesis. However, some susceptibility genes may be unique to particular populations: for instance, PTPN22 SLE risk alleles have been observed in European-derived but not Asian or African-American populations,32 ,48 ,49 whereas loci within ETS1, WDFY4 and AFF1 have been associated with SLE only in Asian studies.59 ,60 Variability in allelic frequency among populations explains the difference in effect of PTPN22 R620W (rs2476601; European 2–15%; Asian nearly absent) and potentially ETS1 rs1128334 (Chinese and Japanese 31–45%; European, Hispanic and African 6–11%). However, other differences in effect sizes are seen among populations, including an increased number of risk alleles for several SLE loci in Amerindian compared with European ancestry,70 and in southern European compared with central European individuals.71 Possible explanations for such diversity in results include: differences in allelic linkage, different gene–gene interactions, unique environmental exposures and limitations of study design.

Genotype–phenotype effects in SLE

The diversity of clinical manifestations in SLE has generated significant interest in the potential for genetic prediction of disease subphenotypes. Insights into genotype effects on clinical/laboratory phenotypes have led to burgeoning understanding of the potential impact on SLE variability, and we will highlight some examples in the following paragraph. As described above, ITGAM has recently been associated with SLE in multiple ancestral groups. On subphenotype analysis, the ITGAM rs1143679 risk allele was found consistently to confer a higher risk of lupus nephritis in SLE patients of European and Asian ancestry when compared to SLE patients without the risk allele.72 ,73 Another robust example is that of IRF5, whereby IRF5 haplotypes determine the risk of increased IFN-α activity in SLE, which is influenced by the presence of autoantibody subsets.16 There may, in fact, be baseline increases in IFN activity in healthy individuals based on the IRF5 haplotype, potentially leading to an increased risk for the development of autoantibodies and autoimmune disease.74 An association of ITGAM and STAT4 also with anti-dsDNA antibody positivity in human SLE among several population groups additionally suggests the possibility of ‘autoantibody propensity loci’ in SLE,15 which may influence disease subphenotype. Cumulative evidence also points to HLA-DR2 and DR3-containing haplotypes as respectively contributing to genetic risk for the presence of autoantibodies in SLE;75 in a murine model, lupus-associated autoantigen proteins interact with HLA-DR3 and lead to autoreactive T-cell activation, resulting in the production of autoantibodies.76 Overall, however, the modest effects of most loci to date account for a small proportion of the heritability of SLE. Indeed, there have been divergences in genotype–phenotype associations described by studies of different cohorts possibly due to differences in samples sizes, or a lack of complete clinical information. Learning of more loci to account for heritability will enhance the link between genotype/phenotype in SLE.

Gene–gene interaction in SLE

Recent descriptions of gene–gene interactions, or epistasis, may explain some of the missing heritability in SLE. Using novel analytical tools, potential epistasis has been identified between the HLA region and CTLA4, ITGAM and IRF5, STAT4 and IRF5, PDCD1 and IL21 and between BLK and BANK1 and TNFSF4 in patients with SLE.77–80 These early results, whereby the presence of certain risk alleles may influence other risk alleles at different loci, suggest additional complexity to the previously accepted model of additive heritability in SLE. These results represent early stages of epistasis study in SLE; we anticipate novel insights from further analyses of functional variant interactions derived from growing numbers of susceptibility loci.