Microglial Self-Recognition STINGs in A-T Neurodegeneration

Abstract

Microglial inflammation is often seen as a secondary event in neurodegeneration. A recent study by Song et al. demonstrates that loss of ataxia telangiectasia mutated (ATM) activates microglia through the cytosolic DNA sensor STING. This highlights the ability of microglia to recognize and respond to self-DNA, with potentially neurotoxic consequences.

Ataxia telangiectasia (A-T) is a pleiotropic disease characterized by cerebellar Purkinje neuron degeneration, cancer, and immune dysfunction. A-T is caused by loss of function mutations in the ATM gene. The encoded ATM protein is a PI3K family kinase that is essential for the double-strand (ds) DNA break repair machinery and for cellular redox balance.

Cell-autonomous effects of ATM dysfunction have been well studied in neurons. However, recent studies indicate that the predominant Purkinje cell degeneration evidenced in A-T may not be solely explained by neuronal ATM loss. Initial evidence [1] demonstrated that ATM inhibition solely in glial cells was sufficient to activate an innate immune response and induce neurodegeneration. Interestingly, a later study showed that cortical and cerebellar neurons in ATM null mice respond differently to immune challenge [2], in a manner that seemed to correlate with microglial phenotype. Together with other reports of microglial activation, these results suggested that microglia might contribute to A-T pathogenesis and area-specific susceptibility [3]. However, the mechanisms underlying microglial activation in A-T remained unknown.

A recent study from the laboratory of Karl Herrup provides new insight into this issue [4]. Song et al. show that ATM-deficient microglia undergo cell-autonomous activation triggered by presence of single-stranded (ss) and dsDNA in the cytoplasm. In particular, they demonstrate that deficient DNA repair caused by genetic and pharmacological inhibition of ATM results in leaching of ss/dsDNA into the cytoplasmic compartment of the microglia. In microglia, the presence of this cytoplasmic self-DNA induces signaling via stimulator of inter-feron genes (STING), a crucial regulator of the innate immune response to cytosolic nucleic acids (Figure 1A). STING activation results in nuclear factor κB (NF-κB)-dependent transcriptional upregulation of interleukin 1 (IL-1b) and production of pro-IL-1β. Although loss of ATM function led to the accumulation of ss/ds self-DNA in the cytoplasm of neurons and fibroblasts, this alone did not seem to cause inflammation. This may be due to the absence of STING in neurons – single-cell RNA sequencing indicates that, in the brain, STING is predominantly expressed in microglia (Figure 1B) [5].

Figure 1. Microglia Are Key Responders to Cytoplasmic DNA Accumulation through the Stimulator of Interferon Genes (STING) Pathway.

(A) Proposed mechanism of the innate immune response of microglia in ataxia telangiectasia (A-T). Normally, ATM (ataxia telangiectasia mutated) responds to DNA double-strand (ds) breaks and aids in their repair, whereas in A-T there is a reduction in the function of ATM and thus there is an accumulation of DNA breaks. Song et al. show that the accumulation of ds breaks leads to activation of the innate immune response in microglia via the accumulation of single-stranded and dsDNA in the cytoplasm [4]. Cytoplasmic DNA then activates STING and the AIM2 inflammasome, leading to nuclear factor (NF)-κB activation and interleukin (IL)-1b release from microglia. This novel pathway for an overall increase in the activation state of microglia and subsequent neuroinflammation then contributes to, and exacerbates, cell-intrinsic degeneration of Purkinje neurons in the cerebellum.

(B) Brain RNA sequencing (RNA-seq) data indicate predominant expression of STING in microglia in both mice and humans [5]. Abbreviation: FPKM, fragments per kilobase million.

To be released and exert neurotoxic effects, pro-IL-1β must be processed by caspase 1 in active inflammasomes. Indeed, Song et al. show that pharmacological inhibition of ATM induces the formation of an inflammasome containing absent in melanoma (AIM2) and caspase 1. Subsequent IL-1β release resulted in synaptic degeneration and apoptosis in neuronal cultures because both were blunted by pretreatment with an IL-1 receptor antagonist. Of note, synaptic loss and neural death were only induced when neural cultures were either cocultured with microglia or treated with microglia-conditioned media, and not in response to pharmacological or genetic loss of ATM function in neurons. Therefore, the pathologies seen in vitro were strictly due to microglial activation.

Cytoplasmic nucleic acid recognition may be one of nature’s oldest protective strategies, whereby the detection of invading genetic material activates mechanisms to degrade the pathogen and destroy its ability to replicate. Although this ancient innate immunity is crucial for efficient clearance of viral and bacterial infections, it may also cause a neurotoxic autoimmune response when reacting to ‘self’ nucleic acids. In A-T, deficient DNA repair leads to the cytoplasmic accumulation of ss/ds self-DNA that is recognized by this innate immune system, resulting in microglial activation and the secretion of neurotoxic cytokines. From a therapeutic perspective, these results suggest that targeting innate immune activation may alleviate the neurodegenerative aspects of this disease. Indeed, early targeting of innate immune pathways with anti-inflammatory drugs effectively inhibits disease phenotype development in ATM null mice [6].

The association of this microglial autoimmunity pathway and Purkinje cell death may, in part, explain the regional specificity of A-T neuropathology. Although ubiquitous throughout the CNS, microglia are regionally diverse [7]. Cerebellar microglia are distinct from cerebral microglia owing to their hypervigilant state. This is evidenced by increased expression of immune-amplifying as well as immune-response/recognition genes, including genes involved in cytoplasmic DNA recognition such as Zbp1 [7]. The distinct ‘alert’ state of cerebellar microglia may prime the inflammatory response to their microenvironment, tipping the balance between beneficial and neurotoxic at the slightest viral or bacterial insult, with Purkinje cells in the crosshairs.

Innate autoimmune dysfunction may also occur in other neurodegenerative diseases in which mutated proteins play a role in DNA repair, including Huntington’s disease (HD) and spinocerebellar ataxia type 3 (SCA3) [8]. In HD and SCA3, polyglutamine expansion impedes transcription-coupled repair complex (TCR) function, resulting in the accumulation of DNA damage and cell toxicity [8]. Even with intact DNA repair, unusual self-nucleic acid may be recognized by this system, as recently shown in a Drosophila model of polyglutamine repeat neurodegenerative diseases, where abnormally expanded repeat dsRNA triggered an antiviral autoimmune response [9]. Mutant proteins can also trigger autoimmunity by directly affecting the function of key players in the nucleic acid immune response. For example, phospholipase D3 (PLD3), an endonuclease that regulates DNA sensing by endosomal Toll-like receptor 9 (TLR9), is implicated in Alzheimer’s disease and is regulated by ataxin 3 [10]. Therefore, it would be of interest to determine whether cytoplasmic nucleic acids and/or STING/NF-κB inflammatory pathways are present in microglia in these other neurodegenerative diseases.

In conclusion, the novel inflammatory cascade described by Song et al. may be a broadly applicable mechanism for neuroinflammatory induction in a variety of diseases. The pattern of inflammation induction and subsequent neural degeneration may be explained by the regional diversity of microglial alertness as well as of local mutant protein expression. Therefore, neuroinflammation, and especially microglial activation, if further validated, could be an effective and broadly applicable therapeutic target, but may need to be targeted early in disease.