Portal de Periódicos da CAPES

Multifaceted roles of neuroinflammation: the need to consider both sides of the coin

2016; Wiley; Volume: 136; Issue: S1 Linguagem: Inglês

10.1111/jnc.13530

1471-4159

Tammy Kielian,

Inflammasome and immune disorders

It is now widely accepted that neuroinflammation accompanies a wide range of central nervous system (CNS) diseases, including infection, trauma, and neurodegeneration. Historically, a focus has been on examining the deleterious aspects of neuroinflammation on CNS homeostasis, be it damaging effects of encephalitogenic Th1 and Th17 cells in multiple sclerosis (MS) and its animal model experimental autoimmune encephalomyelitis (EAE), negative effects of activated microglia and proinflammatory cytokines in Alzheimer's disease (AD), or peripheral leukocyte infiltrates in secondary damage following stroke. It is clear that if not tightly regulated, neuroinflammation has the potential to exert adverse effects on CNS integrity and function. Although the alternative viewpoint that many aspects of CNS immunity may in fact be protective has been proposed for some time, recent advances have begun to shed light on the complexity of neuroinflammatory outcomes during CNS insult/injury. This Special Issue on Neuroinflammation in the Journal of Neurochemistry showcases both the protective and destructive potential of neuroinflammatory responses in the CNS. The manuscripts in this Special Issue dovetail to illustrate the interrelatedness between neuroinflammatory outcomes in disparate CNS disorders and also point toward potential mechanisms of therapeutic intervention that could reverse neurodegeneration/neuropathology. The first manuscript in the Special Issue is an eloquent review of the neuroprotective attributes of microglia by Chen and Trapp (2015). As alluded to above, significant work has been done ascribing damaging effects of activated microglia, primarily via their ability to produce a wide range of proinflammatory cytokines and reactive oxygen/nitrogen species. However, it stands to reason that microglia also play crucial roles in maintaining CNS homeostasis as the sentinel innate immune cell in the CNS parenchyma. If microglia were only capable of exerting harmful effects, why would the CNS harbor a resident population that in essence would equate to a 'ticking time bomb'? In addition, if activated microglia were exclusively deleterious, the latency of neurodegenerative diseases would be accelerated because of rapid neuron destruction, which is clearly not the case for many chronic neurodegenerative disorders typified by reactive microglia (i.e. AD, Parkinson's disease (PD)). Chen and Trapp provide a comprehensive synopsis of recent evidence that supports neuroprotective effects of microglia via synaptic stripping, the ability of microglia to promote neurogenesis, the key role for microglial phagocytosis in CNS homeostasis, and the ability of microglia to dampen inflammation via anti-inflammatory cytokine production (i.e. TGF-β, IL-10) (Chen and Trapp 2015). One aspect of microglial activation that is worth highlighting is the classification of microglia as being polarized toward a 'M1' proinflammatory versus 'M2' anti-inflammatory state. This division was originally used to define the activation states of macrophages in vitro under highly controlled conditions (Gordon 2003) and it has become clear that there are many shades of gray as illustrated by recent transcriptional profiling of adult microglia as well as studies demonstrating that microglia can simultaneously exhibit both M1 and M2 characteristics (Hickman et al. 2013; Butovsky et al. 2014). In addition, a recent study has shown that microglia in a mouse AD model can express markers typically associated with 'M2' polarization, yet exert presumably neurotoxic effects by scavenging metabolites that are important for neuronal survival (Kan et al. 2015). This M1 versus M2 conundrum is nicely discussed by Chen and Trapp (2015). What remains to be determined, and represents a major question in the field, is what signals are critical for dictating whether microglia assume a protective versus destructive state. If we can understand the molecular triggers responsible for programming microglia they could potentially be harnessed to dictate desired outcomes in the setting of CNS disease/injury. The concept of microglia as sentinels of CNS surveillance is elegantly demonstrated by an innovative study by Hernandez et al. (2016). The authors utilize a murine model of unrestrained mild closed head injury (mCHI), where by virtue of the head not being immobilized allows for rotational injury to occur in more rostral regions removed from the cortical impact site. This model potentially recapitulates many human TBIs that involve injury to one side of the brain along with rotational aspects that do not produce distinctive MRI findings, which was supported by minimal changes in T2-weighted MRI scans in mCHI mice compared to sham animals (Hernandez et al. 2016). The authors utilized a clever approach to leverage microglia as reporters of acute brain injury at the impact site versus distant regions that experienced rotational stress. Namely, visualization of fine microglial processes with the use of CX3CR1-GFP mice revealed significant changes in microglial morphology at sites distant from the impact site compared to the injured area. Specifically, microglial morphology was similar in sham treated and mCHI mice at the site of impact; however, the length of microglial processes and soma volume were significantly reduced at the anterior region that was removed from the impact site by at least 2 mm, reminiscent of microglial activation. The authors put forward an intriguing hypothesis that the rostral area of the cortex (anterior to the impact site) that is subjected to rotational stress provides greater 'danger signals' to elicit microglial reactivity compared to the region that is directly damaged (Hernandez et al. 2016). As a consequence, the shortened process length would be expected to impair normal microglial surveillance, which relies on dynamic process extension into the surrounding parenchyma. In effect, this would leave these regions vulnerable to subsequent insults from a loss of effective sensing of danger/pathogenic signals and could set the stage for ongoing and long-term inflammation that has been reported in other TBI studies (Hernandez et al. 2016). What remains to be determined is how long these morphological changes in microglial process retraction persist, to account for the long-term clinical sequelae that are often associated with mTBI in humans. Importantly, this study by Hernandez and colleagues nicely reiterates points raised by Chen and Trapp (2015) where microglial activation did not follow a linear spectrum. In particular, Hernandez et al. found a disconnect between microglial activation as measured by flow cytometry and qPCR where no changes between sham and mCHI animals were evident in any brain region, whereas microglial morphological changes were seen selectively at rostral sites removed from the damaged region (Hernandez et al. 2016). This suggests that microglial marker expression and morphological changes are differentially regulated, suggesting distinct mechanisms of induction and/or possible activation subtypes, reiterating the complexity of microglial responsiveness and their potential adaptation to deal with specific types of insult/injury. In terms of pathogenic inflammation, several studies have implicated aberrant inflammasome activity in a wide array of CNS diseases. An elegant review in this Special Issue by Freeman and Ting highlights our current understanding of the pathogenic role of the inflammasome in MS, AD, and TBI (Freeman and Ting 2015). The inflammasome is a multi-subunit cytoplasmic complex formed by a NOD-like receptor (NLR) family member, the adaptor molecule apoptosis-associated speck-like protein (ASC), and pro-caspase-1, which upon assembly catalyzes the proteolytic processing of pro-IL-1β and pro-IL-18 into their mature forms via its effector enzyme caspase-1 and can also trigger a specialized form of inflammatory cell death termed pyroptosis (Freeman and Ting 2015). There are more than 20 NLR genes in humans and more than 30 in mice, with the signal specificity and functional roles for NLRP1, NLRC4, AIM2, and NLRP3 being the best characterized and pathogenic roles in AD, TBI, and MS recently described (Freeman and Ting 2015). In terms of AD, a critical role for the NLRP3 inflammasome has been identified for microglial production of IL-1β in response to fibrillar amyloid-β (Aβ) (Halle et al. 2008) and crossing NLRP3-deficient mice to the APP/PS1 mouse AD model resulted in fewer Aβ plaques and enhanced phagocytosis of Aβ (Heneka et al. 2013). In addition to NLRP3, NLRP1 has also been linked to AD, since aged APPswe/PS1d1 mice displayed increased NLRP1 expression and siRNA knockdown of NLRP1 reduced pyroptosis and improved cognitive performance in these animals (Tan et al. 2014). Although IL-1β has been implicated in AD pathogenesis, inflammasome involvement in the disease appears to extend beyond cytokine processing, since studies where IL-1β was chronically over-expressed did not result in enhanced neurodegeneration (Shaftel et al. 2007) and actually reduced Aβ plaques in another study (Ghosh et al. 2013), highlighting the potential involvement of the inflammasome in neurodegeneration that extends beyond pro-IL-1β processing. It is well known that caspase-1 can cleave many other proteins, some of which could impact AD progression; however, the identity of these potential inflammasome targets remains to be identified. In TBI, inflammasome activation likely originates from danger-associated molecular patterns (DAMPs) released following tissue damage. However, most of the experimental evidence to date implicating inflammasome involvement in TBI has been observed in severe injury models (Freeman and Ting 2015). It remains to be determined whether danger signals are sensed following mild TBI; however, the study by Hernandez et al. (2016) suggests that microglia distant from the injury site are responding to some trigger; whether this elicits inflammasome activation remains to be determined. Finally, the NLRP3 inflammasome has been implicated in MS neuropathology, which is thought to be mediated, in part, by the ability of IL-1β to promote Th17 differentiation and IL-18 to facilitate Th1 differentiation in conjunction with IL-12 (Freeman and Ting 2015). Interestingly, recent evidence suggests that IFN-β, which represents one of the first lines of treatment for MS, dampens NLRP3 and NLRP1 inflammasome activation and IL-1β production. This is supported by clinical studies where MS patients that were responsive to IFN-β had elevated NLRP3 and IL-1β expression initially, whereas non-responsive patients did not, suggesting that NLRP3 may be a target for IFN-β therapy (Malhotra et al. 2015). Despite these pathogenic roles for the inflammasome, it is clear that inflammasome activation and IL-1β and IL-18 production are needed for an effective immune response to clear CNS infections (Hanamsagar et al. 2011, 2014) and, as such, a delicate balance must be achieved to avoid exaggerated inflammasome activity. Negative regulators of the inflammasome are continuing to be identified and degradation of the macromolecular inflammasome complex by autophagy is another means to quell activity. There has been comparatively less investigation of negative regulators of inflammasome signaling in the context of CNS diseases, which could provide important insights into how to control aberrant inflammasome activity during neurodegenerative disorders. Clinical relevance for a pathogenic role of the NLRP1 inflammasome in TBI is provided in a compelling report by de Rivero Vaccari and colleagues where higher levels of NLRP1 were detected in the CSF of TBI patients that received a clinical diagnosis of a poor or unfavorable outcome (de Rivero Vaccari et al. 2015). In contrast, patients with a favorable outcome following injury displayed lower NRLP1 levels. Since NLRP1 is expressed in neurons, it could represent a potential biomarker for the severity of neuronal damage. Indeed, NLRP1 was detected in exosomes derived from the CSF of TBI and spinal cord injury patients (de Rivero Vaccari et al. 2015). These findings are in agreement with recent studies where NRLP3 and ASC have been shown to be released extracellularly and exist in a 'prion-like' state (Baroja-Mazo et al. 2014; Franklin et al. 2014). It has been proposed that the uptake of exosomes rich in inflammasome components (or other inflammatory molecules) serves as a means to endow neighboring cells with an immediate arsenal to exert proinflammatory activity; effectively propagating inflammation. The authors suggest that since exosomes reflect the intracellular content, they can provide a sort of 'liquid biopsy' that can be exploited for biomarker identification (de Rivero Vaccari et al. 2015). However, one outstanding question is to elucidate the cellular source of exosomes in the in vivo setting, which will likely prove challenging, since neurons, astrocytes, oligodendrocytes, and microglia have all been reported to secrete exosomes (de Rivero Vaccari et al. 2015). The authors also employed an innovative approach to utilize neuronal exosomes to deliver an ASC siRNA in vivo and found that this led to significant knockdown of ASC protein levels and attenuated caspase-1 activation and IL-1β processing after SCI (de Rivero Vaccari et al. 2015). This study suggests that exosomes may offer a novel therapeutic approach to deliver RNA-based compounds to block inflammation after CNS injury and, in particular, inflammasome activation that has been implicated in several chronic neurodegenerative diseases. Another interesting finding by de Rivero Vaccari and colleagues was that NLRP1 immunoreactivity extended well beyond the penumbra region of damaged tissue in clinical specimens from SCI patients (de Rivero Vaccari et al. 2015), which is reminiscent of the distant activation of microglia rostral to the injury site by Hernandez and colleagues in the mCHI model (Hernandez et al. 2016). The contribution from Su and colleagues represents an important advance toward better understanding the molecular determinants of microglial activation and function (Su et al. 2015). The authors have developed a novel method for isolating microglia from the adult brain, a task that has proved challenging based on the limited cell yield using conventional Percoll gradient centrifugation methods that can result in 'contamination' with other glial cell types. This new technique utilizes magnetic bead isolation technology to prepare adult microglia that can be maintained for downstream mechanistic experiments (Su et al. 2015). Historically, the majority of studies examining microglial activation and function in vitro have utilized neonatal microglia, since they are easier to purify and are capable of proliferation. However, evidence from a number of groups has demonstrated differences in the responsiveness of neonatal versus adult microglia (Floden and Combs 2006) and importantly, when considering chronic neurodegenerative diseases, microglia that are recovered from the brain of advanced disease have experienced a history of exposure to DAMPs and other regulatory cascades that likely alter their activation set point. This 'experience' cannot be easily recapitulated by neonatal microglia in vitro and should be considered when translating the significance of in vitro studies to in vivo disease conditions. Another significant advance provided by Su and colleagues is the use of adeno-associated virus (AAV) as a gene delivery system to interrogate mechanistic and therapeutic questions associated with microglial activation (Su et al. 2015). This is a critical issue, since it has long been recognized that primary microglia are notoriously difficult to transfect using traditional chemical or electroporation approaches that often induce toxicity and/or trigger microglial activation. The authors employed an ambitious approach by comparing the efficiency of four commercially available transfection reagents to AAV transduction using five different viral serotypes (AAV2, AAV5, AAV6, AAV8, and AAV9). They report superior gene delivery with AAV constructs compared to transfection reagents that displayed little/no efficacy. In addition, AAV serotypes did not induce any evidence of microglial activation, indicating that the virus itself does not elicit any changes that could influence phenotypes obtained following gene targeting. They also demonstrated that an AAV2-Cre construct could be used to knock down floxed miR-155 in microglia, providing proof-of-concept that viral-mediated gene targeting in microglia is feasible (Su et al. 2015). Given the interest in better understanding the role of microglia during neurodegenerative diseases, where age is a critical factor, these advances in combining adult microglial culture with AAV approaches to interrogate gene function could provide novel insights into microglial biology. Another emerging theme in neuroimmunology is the ability of prior experiences to shape the intensity and nature of subsequent CNS damage. One example is preconditioning, where prior exposure to a stimulus can modify subsequent inflammation. Preconditioning has been described for cerebral ischemia (Wang et al. 2015) and in this Special Issue, Bernardes and colleagues report the beneficial effects of preconditioning exercise (forced swimming) on minimizing disease severity in a mouse EAE model concomitant with reduced inflammation, demyelination, and axonal damage (Bernardes et al. 2015). Impressively, the beneficial effects of this prior exercise regimen were observed up to 28 days after the exercise period ended. Interestingly, exercise preconditioning significantly attenuated T and B cell infiltrates into the spinal cord; however, macrophage recruitment was not affected. Stereological quantification of mature oligodendrocytes versus oligodendrocyte progenitor cells revealed decreased numbers of the latter in the white matter of exercised mice with EAE, which may be reflective of increased differentiation into mature oligodendrocytes that affords protection against demyelination and axonal damage associated with prior exercise (Bernardes et al. 2015). It is important to note that physical exercise programs have been shown to attenuate symptoms in MS patients, such as walking disability, pain, and depression (Motl et al. 2008; Feinstein et al. 2014). Collectively, these findings underscore the possibility that physical exercise may have a prophylactic effect toward curbing neuroinflammation, although the mechanism(s) of action remain to be completely elucidated. Neuroinflammation is also a hallmark of metabolic disorders and Platt and colleagues provide a comprehensive synopsis of inflammatory perturbations associated with Niemann–Pick disease type C (NPC) (Platt et al. 2015). NPC is a lysosomal storage disease caused by an inherited mutation in one of two lysosomal proteins, namely NPC1 or NPC2. These mutations result in lysosomal dysfunction and the accumulation of lipid storage material, primarily composed of cholesterol, sphingosine, sphingomyelin, and glycosphingolipids. Neuroinflammation is a common feature of lysosomal storage diseases that cause neurodegeneration, and NPC represents one example (Platt et al. 2015). The disease is associated with early microglial and astrocyte activation that predates the onset of clinical symptoms and neuronal loss. Interestingly, although NPC is typified by neurodegeneration, evidence exists to support a non-cell autonomous influence on neuronal death, as revealed by the ability of targeting Npc1 expression in astrocytes to increase the life expectancy of NPC1-deficient mice by greater than 3-fold (Zhang et al. 2008). The ability of non-steroidal anti-inflammatory drugs to significantly prolong lifespan, reduce microglial reactivity and Purkinje cell loss, and delay disease onset in NPC1-deficient mice provides strong evidence to implicate a pathogenic role for inflammation in the disease process (Smith et al. 2009). In addition to CNS inflammation, NPC is also typified by peripheral immune defects. In particular, deficiencies in natural killer (NK) and invariant natural killer T (iNKT) cells have been described in NPC and could conceivably alter the ability to thwart peripheral immune challenges in affected patients (Platt et al. 2015). Importantly, as highlighted by other contributions in this Special Issue, crosstalk between CNS and peripheral immune responses can influence outcomes, and changes in both compartments in NPC represent another such example. Collectively, this Special Issue highlights several conceptual issues that are at the forefront of the neuroimmunology field, namely a need to better understand the neuroprotective aspects of microglia that have historically been viewed as a key contributor to neurodegenerative processes when homeostatic mechanisms have passed the 'point of no return' (Chen and Trapp 2015), molecular mechanisms that bridge neurodegeneration across a wide range of diseases (i.e. inflammasome activation) (Freeman and Ting 2015), as well as the identification of novel therapeutic delivery approaches (exosomes) (de Rivero Vaccari et al. 2015) and research methods to study adult microglia and probe genetic mechanisms of action via AAV-mediated transduction (Su et al. 2015). In addition, a better understanding of the events that elicit neuroinflammation in the context of lysosomal storage diseases (Platt et al. 2015) and how prior experiences can shape the nature, intensity, and duration of CNS inflammatory responses (Bernardes et al. 2015) will be important to identify potential therapeutic targets that may afford protection across a wide range of CNS disorders. Finally, taking an 'outside the box' approach to leverage the exquisite surveillance capabilities of microglia as a means to flag areas of subtle dysfunction that are too diminutive to be detected with currently available technologies (i.e. MRI) (Hernandez et al. 2016) may provide a unique window to identify aspects of CNS pathology that until now have remained elusive. There is no doubt that the complexity of neuroinflammation and the cadre of disorders that are typified by inflammatory responses will continue to provide new and exciting areas of discovery for many years to come. Tammy Kielian is the Deputy-Chief Editor for the area 'Neuroinflammation & Neuroimmunology' and solicited the articles in this special issue.