Autoimmunity: homeostasis of innate immunity gone awry

Introduction

Autoimmunity is generally considered to represent an attack on self-tissues mediated by autoantibodies or self-reactive T cells. However, the recent identification of genetic variants in inflammatory diseases provides unequivocal examples of defects in innate immunity leading to autoimmunity. The complement system and DNases will be used as telling examples of how genetic variants in innate immune pathways produce injury directly or by triggering the adaptive immune system. The principles illustrated are likely representative of how other innate immune pathways mediate damage to self.

Direct tissue damage by dysregulated innate immunity

Atypical hemolytic uremic syndrome (aHUS): Hemolytic uremic syndrome (HUS) is a thrombotic microangiopathy (TMA) characterized by a triad of hemolytic anemia, thrombocytopenia and acute renal failure and is classified as enteropathic (diarrhea-associated) or non-enteropathic (atypical). Rare heterozygous mutations in a negative regulator (factor I, factor H or MCP) of the complement system’s ancient alternative pathway (AP) as well as gain of function mutations in an AP component (C3 or factor B) have been identified in over 50% of aHUS cases (Atkinson et al., 2005; Loirat & Fremeaux-Bacchi, 2011; Noris & Remuzzi, 2009; Richards et al., 2007; Rodriguez de Cordoba, 2010). AP features a feedback loop that can amplify complement responses initiated by classical or lectin pathway activation or if the AP “tickover” exceeds regulatory activity. Thus, individuals carrying a mutation are predisposed to developing a rapid and inappropriately regulated storm of AP mediated complement activity on damaged endothelial cells and exposed basement membranes culminating in a TMA that most severely affects the kidney.

aHUS exemplifies how targeted gene sequencing in a rare disease identified deleterious, unique variants in multiple components of an innate immune pathway. This newly acquired understanding of aHUS pathogenesis has led to the use of targeted therapy. Eculizumab, a monoclonal antibody to C5, is FDA-approved for aHUS albeit based on as yet unpublished data (FDA, 2011). A recent review discusses its use and efficacy (Westra et al., 2012). The concept of excessive AP response and the utility of eculizumab are currently being investigated in more common TMAs like enteropathic HUS (Menne et al., 2012) and preeclampsia (Salmon et al., 2011) as well as in a severe subset of preeclampsia known as hemolysis elevated liver enzymes low platelets (HELLP) syndrome (Fakhouri et al., 2008; Fang et al., 2008).

Age-related macular degeneration (AMD): AMD is a common, chronic organ specific example of AP driven tissue damage and is the leading cause of blindness in the industrialized world (Resnikoff et al., 2004). AMD causes progressive central vision loss and, at an early stage, is characterized by drusen, which are focal depositions of extracellular debris between the retinal pigment epithelium (RPE) and Bruch’s membrane. Drusen is a complex mixture of oxidized lipids, modified proteins, peptidoglycans and other cellular debris that stain positive for activation fragments and regulators of the AP (Johnson et al., 2001).

In 2005, four groups reported that a single nucleotide polymorphism (SNP) in factor H, the major regulator of the AP in blood, confers significant susceptibility to AMD. Y402H increases the risk of AMD two- to four-fold in heterozygotes (402_YH) and five- to seven-fold in homozygotes (402_HH) (Edwards et al., 2005; Hageman et al., 2005; Haines et al., 2005; Klein et al., 2005). Studies of purified factor H from genotyped individuals show that the wild type (402_YY) binds more efficiently to RPE, C-reactive protein (Anderson et al., 2010), and oxidized phospholipids than 402_YH or 402_HH (Shaw et al., 2012). Interestingly, a group A streptococcus (GAS) virulence factor binds less efficiently to 402_HH, leading to greater complement driven GAS opsonization and phagocytosis (Haapasalo et al., 2008) and therefore, providing an evolutionary reason for the high prevalence of this AMD risk allele.

A highly penetrant rare variant (R1210C), known to alter factor H binding to C3b and endothelial cells (Manuelian et al., 2003), was recently reported in AMD patients (Raychaudhuri et al., 2011). Other protective and risk associated SNPs have been identified in factor H (Hageman, 2008) and in other participants of AP (factor I, factor B and C3). The combined analyses of risk variants in members of the AP can account for nearly 75% of AMD cases (Gold et al., 2006; Seddon et al., 2009).

These two examples illustrate a concept of “innate autoimmunity”: an overactive AP in a genetically susceptible host mediating damage to self (figure 1). In aHUS, the injury is an acute event focused on the endothelium and the kidney basement membrane, while in AMD tissue damage accumulates over years and even decades in the retina. Both diseases evince how a variant in a regulatory protein or a gain of function mutation in an activator of a simple proinflammatory pathway of innate immunity can lead to autoimmunity.

Figure 1.

Genetic variants in innate immune pathways mediate autoimmunity. Variants with a regulatory defect (bold arrow indicating too much innate immune activity) in the setting of a predisposing injury, lead to an excessive innate immune mediated inflammatory response. Variants with a clearance defect (dashed arrow indicating too little innate immune activity) lead to “garbage” accumulation. Modified “self” then triggers an adaptive immune response. Consequently, both overactivity and underactivity of an innate immune pathway could lead to autoimmunity.

Deficiencies in innate immunity generating an adaptive immune response

Systemic lupus erythematosus (SLE): Lupus has served as a paradigm for autoimmunity since antinuclear antibodies were first detected in these patients’ sera over 70 years ago. In the 1960s, immune complexes with complement fragments were noted in SLE renal lesions (Koffler et al., 1969b) and anti-dsDNA antibodies were shown to be a risk factor for lupus nephritis (Koffler et al., 1969a). These observations led to a SLE pathogenesis model in which autoantibodies bind nuclear antigens, fix the classical pathway of complement activation and deposit in tissues (Koffler et al., 1971).

Surprisingly, a deficiency in any of the early complement components of the classical pathway (C1q, C1r, C1s, C4 or C2) was identified a few years later to be a major risk factor for developing SLE. Approximately 90% of C1q and 80% of C4 deficient individuals develop SLE (Pickering et al., 2000; Walport, 2001). The “lupus paradox” arose because, if SLE is mediated by autoantibodies activating the classical pathway, should not C1q or C4 deficiency be protective (Carroll, 1998)?

Since the 1980s, the concept of defective clearance of immune complexes containing nuclear antigens has been a centerpiece of lupus pathophysiology (Atkinson, 1989; Lachmann, 1984). Initially, it was thought to be the defective removal of foreign nuclear material derived from retroviruses that led to “autoantigen” development. Subsequently, studies in complement deficient mice have turned the focus to self-nuclear antigens as the immunogen. C1q deficient (C1qa^−/−) mice develop high titer antinuclear antibodies and 25% develop glomerulonephritis (Botto et al., 1998). C1qa^−/− mice accumulate apoptotic bodies in their glomeruli that correlated with disease severity. This observation and other similar data led to the hypothesis that C1q and early classical pathway partners are critical for the proper clearance of nuclear debris derived from apoptotic and necrotic cells (Carroll, 1998; Walport, 2001).

Deoxyribonuclease1 (DNase1): The initial report of SLE patients having decreased levels of DNase1, the major extracellular endonuclease that cleaves both single and double-stranded DNA was in 1981 (Chitrabamrung et al., 1981). Lupus prone mice were also reported to have reduced levels of “DNase” in both sera and urine (Macanovic & Lachmann, 1997). Mice deficient in DNase1 developed both antinuclear and anti-dsDNA antibodies and glomerulonephritis (Napirei et al., 2000).

Subsequently, two SLE patients were reported to have heterozygous DNase1 mutations and decreased DNase enzymatic activity (Yasutomo et al., 2001). In 2011, Al-Mayouf et al. sequenced DNASE1L3, one of three human homologs of DNase1, in 6 Saudi Arabian families affected with predominately early onset, severe, anti-dsDNA positive SLE in an autosomal recessive pattern. A fully penetrant homozygous 1 base pair deletion in DNASE1L3 segregated perfectly with SLE. The protein encoded by the mutant DNASE1L3 lacked enzymatic activity (Al-Mayouf et al., 2011). These studies have shifted the SLE pathogenesis model toward a self-antigen driven disease caused by an inadequate clearance of “altered self” nuclear material (Walport, 2000).

Three prime repair exonuclease 1 (TREX1 or DNase3): Aicardi-Goutières syndrome (AGS) is a rare autosomal recessive disease of infancy featuring a severe encephalitis with prominent lymphocytic infiltration and elevated CSF type 1 IFN. Loss of function homozygous mutations in TREX1, the major human intraceullar 3′ to 5′ exonuclease, cause AGS (Crow et al., 2006; Rice et al., 2007). However, unlike AGS patients, the TREX1 knockout mouse died of a noninfectious autoimmune inflammatory myocarditis (Morita et al., 2004) but also featured a type 1 IFN dependent production of anti-nuclear antibodies and glomerulonephritis with immune complex deposition (Stetson et al., 2008).

Due to several autoimmune characteristics exhibited by TREX1 knockout mice and AGS patients, Lee-Kirsch and colleagues sequenced the TREX1 gene in lupus patients. In German and United Kingdom cohorts, they identified TREX1 mutations in 2% (9/417) of SLE patients (Lee-Kirsch et al., 2007). No functional analyses were performed on the missense mutations but one of the 9 mutations described, R114H, had been previously reported in AGS as having decreased enzymatic activity (Orebaugh et al., 2011). In a subsequent large GWAS, Namjou et al., reported multiple TREX1 SNPs associated with SLE that were unique to a particular ethnic population (Namjou et al., 2011). Through GWAS, approximately 20 distinct lupus susceptibility loci (OR 1.1 to 2.0) have been described, but none has as strong of an effect as TREX1 (OR 25) (Harley et al., 2009).

Gall and colleagues recently elucidated how TREX1 mutations could lead to autoimmunity. The Interferon Stimulatory DNA pathway (ISD) is one of two recognized Toll-like Receptor (TLR) independent intracellular DNA recognition pathways (Karayel et al., 2009). Upon activation by primarily ssDNA, ISD induces the expression of TREX1 and a lymphocyte-independent type I IFN response. In turn, TREX1 metabolizes ISD activators to limit the IFN response (Stetson et al., 2008). TREX1^−/− mice manifest a dysregulated type 1 IFN response that requires an intact adaptive immune system to manifest an autoimmune phenotype (Gall et al., 2012). In the case of TREX1^−/− mouse, the IFN response drives the development of autoreactive T cells while in DNase1^−/− mouse an inability to properly clear extracellular DNA leads to TLR activation. Activation of TLRs on antigen presenting cells then result in autoreactive T and B cells and a type 1 IFN response (Elkon & Wiedeman, 2012; Liu & Davidson, 2012).

In sum, a deficiency in innate immunity such as an early component of the classical complement pathway or a nuclease can lead to inadequate clearance of DNA debris. By one of several mechanisms, accumulated nuclear materials trigger the adaptive immune response to form autoreactive T cells and anti-nuclear antibodies.

Conclusion

In most paradigms of autoimmunity, innate immunity was traditionally viewed as playing a modest role in etiopathogenesis. However, the examples described herein demonstrate how disruptions in homeostasis of the innate immune pathways mediate self-tissue injury via an excessive inflammatory response or by triggering adaptive immunity. Too little regulation may produce an acute thrombomicroangiopathic state like aHUS or a chronic, localized inflammatory condition of the retina like AMD. On the other hand, a defect in innate immunity (i.e. complement or nucleases) may impair nuclear debris clearance. Such “garbage” is modified and can serve as a self-antigen as well as creating a cytokine milieu that drives the adaptive immune system to break T-cell tolerance and stimulate B cells to produce autoantibodies.

List of abbreviations defined in text

aHUS

atypical hemolytic uremic syndrome

TMA

thrombotic microangiopathy

AP

alternative pathway

HUS

hemolytic uremic syndrome

HELLP syndrome

hemolysis elevated liver enzymes low platelets

AMD

age-related macular degeneration

RPE

retinal pigment epithelium

SNP

single nucleotide polymorphism

GAS

group A streptococcus

SLE

systemic lupus erythematosus

DNase1

deoxyribonuclease1

TREX1

three prime repair exonuclease 1

AGS

Aicardi-Goutières syndrome

ISD

Interferon Stimulatory DNA pathway

TLR

Toll-like Receptor

GWAS

genome wide association study

Not defined

CSF

cerebrospinal fluid

IFN

interferon