NLRP3 Inflammasome Is Expressed and Functional in Mouse Brain Microglia but Not in Astrocytes

Neuroinflammation is the local reaction of the brain to infection, trauma, toxic molecules or protein aggregates. The brain resident macrophages, microglia, are able to trigger an appropriate response involving secretion of cytokines and chemokines, resulting in the activation of astrocytes and recruitment of peripheral immune cells. IL-1β plays an important role in this response; yet its production and mode of action in the brain are not fully understood and its precise implication in neurodegenerative diseases needs further characterization. Our results indicate that the capacity to form a functional NLRP3 inflammasome and secretion of IL-1β is limited to the microglial compartment in the mouse brain. We were not able to observe IL-1β secretion from astrocytes, nor do they express all NLRP3 inflammasome components. Microglia were able to produce IL-1β in response to different classical inflammasome activators, such as ATP, Nigericin or Alum. Similarly, microglia secreted IL-18 and IL-1α, two other inflammasome-linked pro-inflammatory factors. Cell stimulation with α-synuclein, a neurodegenerative disease-related peptide, did not result in the release of active IL-1β by microglia, despite a weak pro-inflammatory effect. Amyloid-β peptides were able to activate the NLRP3 inflammasome in microglia and IL-1β secretion occurred in a P2X7 receptor-independent manner. Thus microglia-dependent inflammasome activation can play an important role in the brain and especially in neuroinflammatory conditions.


Introduction
Inflammatory processes have been implicated in both acute and chronic neurodegenerative conditions. Pro-inflammatory cytokine production has been associated with neuroinflammation and different neurodegenerative diseases, such as Alzheimer's disease (AD), Parkinson's disease (PD) and Multiple Sclerosis (MS) [1][2][3]. Although inflammation may not typically represent an initiating factor in neurodegenerative diseases, there is emerging evidence in animal models that sustained inflammatory responses involving microglia and astrocytes contribute to disease progression [4]. Among these responses, IL-1β secretion has been shown to have important functions in diseases of the Central Nervous System (CNS), yet the precise mechanism of action is not clear since both neurotoxic and regenerative roles have been described [5][6][7].
The molecular steps leading to IL-1β maturation take place in an intracellular complex termed the inflammasome [8]. The inflammasome is a caspase-1 activating multiprotein platform that results from oligomerization of inactive monomeric proteins from the nucleotidebinding domain, leucine-rich repeat (NLR) protein family. Different complexes have been described and are usually defined by the core NLR proteins that are implicated, such as NLRP1, NLRP3, NLRC4 (also called IPAF) or the HIN200 family member AIM2. The most intensively studied is the NLRP3 inflammasome which is formed when NLRP3 associates with the adaptor protein ASC and procaspase-1 upon activation by different stimuli, such as pathogen associated molecular patterns (PAMPs) or endogenous danger signals [9,10]. Inflammasome assembly results in the cleavage of caspase-1 from its proform to its enzymatically active form. This active caspase-1 then processes the cleavage of several substrates, such as pro-IL-1β and pro-IL-18 into the mature cytokines IL-1β and IL-18. IL-1α, another IL-1 family member, and High-mobility group box 1 (HMGB1) do not require cleavage by caspase-1 but their release in response to certain stimuli is dependent on the inflammasome [11,12]. All these cytokines are shown to be important actors in different neurodegenerative diseases [13,14]. The precise mechanism of inflammasome activation is not fully understood. NLRP3 can sense a variety of molecular structures and it is thought that inflammasome activation occurs probably not through direct ligand-receptor interaction but rather the sensing of cellular homeostasis disruption [15,16].
The role of the NLRP3 inflammasome in CNS diseases has recently started to be investigated in more detail. Several studies suggest a general involvement of the NLRP3 inflammasome in MS, a demyelinating inflammatory disease of the CNS. In experimental autoimmune encephalomyelitis (EAE), a mouse model of MS, inflammasome-deficient mice are protected from disease progression (reviewed in [17]). In addition to MS, the implication of the NLRP3 inflammasome in Alzheimer's disease has been shown both in vitro by the ability of amyloidbeta (Aβ) peptides to activate the inflammasome [18] and in vivo in an AD mouse model where NLRP3 knock-out mice were protected from spatial memory impairment and showed decreased Aβ plaque burden [19].
In many cases, neurodegenerative diseases are related to an accumulation of protein aggregates such as Lewy bodies in PD, senile plaques in AD or PrP sc in prion related-disease. As shown in a recent study, PrP fibrils are able to induce inflammasome-dependent IL-1β secretion in microglia, suggesting a role for the NLRP3 inflammasome in prion diseases [20,21]. Besides, the capacity of similar aggregates, such as islet amyloid polypeptide (IAPP) in type II diabetes, to activate NLRP3 inflammasome via abnormal phagocytosis and lysosomal rupture mechanism has also been shown [22].
In view to understand the role played by IL-1β in neurodegenerative diseases, it is important to characterize in detail the inflammasome expression and activation in CNS inflammatory cells. Microglia are the resident macrophages of the brain and as such are described to secrete IL-1β. Astrocytes are the most abundant cell type in the brain where they maintain CNS homeostasis and provide neuronal support in healthy conditions. In the injured brain they play important roles during glial scar formation, a process referred to as reactive astrogliosis [23]. However it is not entirely clear if astrocytes are able to produce IL-1β. In addition there are only few studies about the role astrocytes can play during neuroinflammation with regard to inflammasome activation [24]. Therefore we aimed to investigate the relative contribution of microglia and astrocytes to inflammasome activation in the context of neurodegenerative disease-related peptide stimulation.

Materials and Methods
Mice C57BL/6JOlaHsd mice were from Harlan. Nlrp3 -/mice were provided by University of Lausanne [25]. Casp1 -/-/Casp11 -/-(sn#016621) (hereafter referred to as Casp1 -/-) and P2rx7 -/-(sn#005576) mice were obtained from The Jackson Laboratory (Bar Harbor, USA). All mouse strains were housed and bred at the animal-housing facility of the University of Luxembourg approved by the National Veterinary Inspection. All animal procedures were conducted in strict accordance with the FELASA Guide for the Care and Use of Laboratory Animals. The protocol was approved by the animal ethics and experimentation committee (AEEC) of the University of Luxembourg. All efforts were made to minimize suffering. Newborn mice were decapitated.

Cell cultures
Mixed glial cell cultures were prepared as previously described [26]. Briefly, after carefully removing meninges and large blood vessels, the brains from newborn mice were pooled and then minced in cold phosphate-buffered saline (PBS) solution. After mechanical dissociation cell suspensions were washed and plated in complete medium (Dulbecco's Modified Eagle Medium (DMEM, Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS, Gibco), 100 U/ml penicillin and 100 U/ml streptomycin, both from Lonza) at 37°C in a humidified atmosphere containing 5% CO 2 . The culture medium was changed after three days of culture. After 15 days, mixed glial cultures had reached confluence. Glial cells were separated by magnetic-activated cell sorting (MACS), as previously described [26]. Microglia were positively selected using an anti-CD11b antibody, following the manufacturer's instructions (Miltenyi), and plated in mixed glial culture conditioned medium to be treated twenty-four hours later. Astrocytes were negatively sorted during this process. The obtained "astrocyte-enriched cultures" were plated in 75 cm 2 flasks for seven days allowing the cultures to reach confluence. After one week, the MACS procedure was repeated in order to reduce microglial contamination of the astrocyte population.
Neurosphere-derived astrocytes were prepared as described in [27]. Briefly, neurospheres derived from embryonic neural stem cells were cultured and were differentiated to astrocytes by plating them on poly-L-ornithine coated flasks in complete DMEM.

Real-time PCR
Total RNA was isolated using the Invisorb Spin Cell RNA Mini Kit (Invitek) according to the manufacturer's protocol. Complementary DNA (cDNA) was synthesized from RNA samples using the ImProm-II Reverse Transcription System (Promega). Gene expression was analyzed using Bio-Rad iCycler (iQ5 Real-Time PCR Detection System, Bio-Rad Laboratories) with SYBR Green Supermix (Promega). Expression was normalized relative to the housekeeping gene L27. Primer sequences were designed using the Beacon Designer Software (Bio-Rad) ( Table 1).

Protein quantification assays
Cell culture supernatants were assayed, at the indicated times, for the presence of cytokines by using mouse IL-1β, TNFα, CXCL10 DuoSet ELISA and IL-1α DuoSet ELISA (all R&D) or IL-18 ELISA (MBL) according to the manufacturer's instructions. ATP was quantified in cell supernatants using ATP Determination Kit (Life technologies).

Statistical analysis
The significance of multiple treatments was analyzed by a one way analysis of variance (anova) followed by Tukey's or Dunnett's multiple comparisons tests. If multiple treatments were tested on different genotypes, a two-way anova was performed. The homoscedasticity and normality assumptions were assessed respectively by the Bartlett and Shapiro-Wilk tests. The statistical analysis of gene expression was performed on the delta-Ct values (Ct goi -Ct hkg ) and the analysis of protein quantification (Elisa) was performed on log-transformed concentrations (if present, zero values were removed by adding an offset to the dataset equal to half the smallest non-zero concentration). Non normal data was analyzed by a non-parametric analysis of variance (Kruskal-Wallis test) followed by pairwise comparisons with Mann-Whitney U-tests and Holm's correction. All statistical analyses were performed using R (http://www.r-project.org/) and differences with p-values less than 0.05 were considered significant. All experiments have been performed at least three times.

Brain microglia, but not astrocytes, express NLRP3 inflammasome components and substrates
In order to better understand inflammasome function in the brain, especially in the glial compartment, we aimed to analyze NLRP3, ASC and caspase-1 expression in the different cell types. Expression of inflammasome components was monitored at both the transcript and the protein level (Fig 1A and 1B). Microglia express inflammasome components NLRP3, ASC (Pycard) and caspase-1, similarly to bone marrow-derived macrophages that were used as control cells (Fig 1A and 1B). ASC and caspase-1 were constitutively present in microglia, whereas NLRP3 expression could be upregulated after priming with LPS. In addition to priming with LPS we used a complete cytokine mix (CCM: IL-1β, TNFα and IFNγ) that was also able to induce NLRP3 expression in microglia and BMDMs (Fig 1A and 1B).
In astrocytes however, although caspase-1 was expressed at low levels, only very weak levels of NLRP3 and ASC transcript and almost no NLRP3 and ASC protein could be detected, suggesting that these cells probably lack a functional NLRP3 inflammasome (Fig 1A and 1B). CCM was used in order to prime astrocytes as they do only weakly express TLR4 (Figure A in S1 File) and do not strongly respond to LPS stimulation [30]. However, we were still unable to robustly prime the inflammasome in astrocytes, as shown by the very moderate upregulation of NLRP3. Also priming of astrocytes with several other agents (e.g. P3C, as shown below), as well as combinations thereof (e.g. CCM and P3C, data not shown), did not result in significant expression of NLRP3 and/or IL-1β.
We further investigated expression of other inflammasome related proteins and found that both microglia and astrocytes expressed similar transcript levels of the intracellular sensors Aim2 and Nlrp1, whereas Nlrc4 was mainly expressed in microglia ( Fig 1C). Interestingly, Casp4 (caspase-11) mRNA could be upregulated in microglia and astrocytes after priming ( Fig  1C and Figure C in S1 File). In addition, classical inflammasome substrates, such as Il1b and Il18, were strongly induced by LPS priming in microglia. Il1b transcript remained virtually undetectable in astrocytes after CCM priming and Il18 basal expression levels remained unchanged after priming (Fig 1D). Il1a expression was inducible by LPS treatment in microglia and by CCM treatment in astrocytes. Hmgb1 was expressed by both microglia and astrocytes at mRNA and protein level (Fig 1D and 1E). Protein localization was monitored by immunocytochemistry in microglia, where LPS priming resulted in morphological changes monitoring cell activation, shown by both cell surface marker Iba1 and IL-1β expression (Fig 1E). The inflammasome components NLRP3, ASC and caspase-1, as well as HMGB1, were stained in control microglia and LPS-primed cells (Fig 1E).
In addition to the glial compartment, primary cortical neurons were analyzed by RT-PCR for the presence of inflammasome components and we could detect expression of Nlrp1, Nlrp2, Nlrp6 and Aim2, whereas classical NLRP3 inflammasome components were only weakly expressed ( Figure C in S1 File) These results thus indicate that microglia are competent to form a functional NLRP3 inflammasome, whereas for astrocytes this type of inflammasome is probably not formed and its activation seems compromised.
Microglia can secrete IL-1β and IL-18 in an NLRP3 inflammasomedependent manner As microglia are the brain resident macrophages and seem to express the minimal requirements for inflammasome activation, we investigated their ability to secrete IL-1β, IL-18 and IL-1α in response to classical inflammasome stimuli. When treated with widely used inflammasome activators, such as ATP and Nigericin, we observed robust IL-1β secretion and caspase-1 cleavage in LPS primed microglia (Fig 2A). In addition, particulate stimuli, such as monosodium urate (MSU) crystals or Alum, were also able to induce IL-1β secretion, although at lower levels (Fig 2A and 2B). Treatment with these different activators also resulted in IL-18 secretion by microglia (Fig 2C). Production of both IL-1β and IL-18 occurred in a NLRP3 inflammasome-dependent manner, as shown by the use of NLRP3-and caspase-1-deficient microglia, where the production of these cytokines was strongly reduced (Fig 2B and 2C). As both IL-1β and IL-1α are able to trigger the IL-1 receptor and their relative contributions in vivo are not well understood, we aimed to know if microglia could also produce IL-1α in response to inflammasome activation. We observed NLRP3 inflammasome-dependent production of IL-1α in response to ATP and Nigericin (Fig 2D and data not shown). In response to MSU and Alum, IL-1α was produced in a NLRP3-and caspase-1-independent manner, as also reported by others [31] (Fig 2D and data not shown). Production of all these cytokines required priming of microglia by LPS ( Figure A in S2 File). Microglia also produced TNFα in response to these treatments, but this occurred in a NLRP3-and caspase-1-independent manner ( Figure B in S2 File).
Inflammasome-dependent IL-1β production in microglia occurred through similar mechanisms than described in BMDMs. As expected, the P2X7 receptor was only required in response to stimulation with ATP, whereas Nigericin, MSU and Alum activated the inflammasome in a P2X7R-independent manner (Fig 2E). Potassium efflux was necessary for efficient inflammasome activation in microglia, as blocking it by addition of KCl to the extracellular medium strongly reduced IL-1β secretion (Fig 2F). This effect occurred in part at the priming step since we also observed reduced levels of pro-IL-1β (data not shown). Reactive oxygen species (ROS) were implicated in the microglial inflammasome activation process, since interfering with ROS production by N-acetyl cystein (NAC) impaired IL-1β secretion induced by ATP and Alum (Fig 2G and 2H). Lysosomal damage was implicated in inflammasome activation by Alum but not ATP, as shown by the use of the cathepsin B inhibitor Ca074-Me (Fig 2G and  2H). To determine if phagocytosis of Alum particles was required for inflammasome activation, we used the actin polymerization inhibitor cytochalasinD (cytoD). The presence of cytoD selectively reduced IL-1β secretion in response to Alum (Fig 2H), but not ATP (Fig 2G). Ca-074Me (10 μM) and cytochalasinD (cytoD, 2 μM)). Inhibitors were added for 30 min prior to inflammasome activation. Data shown are the mean ± SD of at least three independent experiments. *p<0.05, ns = not significant, compared to control (ctrl) (Fig 3B-3E) or compared to no treatment (Fig 3F-3H), #p<0.05, KO compared to WT.

Amyloid-β induced microglial inflammasome activation is P2X7Rindependent
Inflammasome function has been linked to a number of diseases that are characterized by peptide aggregates or particulate structures, such as gout or type 2 diabetes [32,33]. Neurodegenerative conditions, such as Alzheimer's or Parkinson's disease, are known to involve plaques formed by amyloid aggregates during the course of the disease [34]. In order to better understand their potential role in the development of the disease and more specifically in microglial activation, we treated cells with the amyloid-β fragment Aβ [25][26][27][28][29][30][31][32][33][34][35] . This Aβ peptide is unable to prime microglia, as shown by the lack of upregulation of different genes, such as Nlrp3, Il1b or Tnf upon treatment (Fig 3A). However, Aβ 25-35 is able to induce caspase-1 cleavage and IL-1β secretion in LPS primed microglia, whereas a control peptide (Aβ 35-25 ) is not (Fig 3B and 3C). IL-1β production is NLRP3-and caspase-1-dependent, as shown by the strong reduction of the cytokine in inflammasome-deficient microglia (Fig 3D). However inflammasome activation by Aβ is P2X7R-independent ( Fig 3E); there is no requirement for this receptor either directly or indirectly, as we could not detect ATP secretion in microglia treated with Aβ ( Fig 3F). Opposed to published results we were unable to detect significant IL-1β production by microglia treated with the Aβ 1-42 peptide whether we used oligomeric or fibrillar preparations of the peptide and different time points (Fig 3G and 3H) [18,35,36].

α-synuclein can modestly prime microglia, but does not result in significant IL-1β secretion
In addition to amyloid-β, we focused on the ability of α-synuclein (α-syn) to trigger inflammasome activation. α-syn is a component of Lewy bodies that are formed during Parkinson's disease and we were interested in its pro-inflammatory capacities as an aggregated peptide. Similarly to Aβ, α-syn can adopt different conformations such as oligomeric or fibrillar, and different degrees of toxicity are associated with each form. We therefore tested both forms for their ability to activate microglia. The oligomeric preparations showed slightly more transcriptional activation than the fibrillar forms concerning certain genes, such as Tnf or Nlrp3 ( Fig  4A). In addition to the native α-syn, we used a mutant peptide that occurs in familial forms of PD, the A53T mutant, and observed that this form has more pro-inflammatory potential than the WT form, but again only with the oligomeric conformation ( Fig 4A). The same results were observed for pro-IL-1β and NLRP3 at protein level, when both α-syn WT and A53T forms (oligomeric and fibrillar) were used to treat microglia for longer time points (Fig 4B). When using α-syn preparations for inflammasome activation, we did not observe significant IL-1β secretion in LPS primed microglia in response to all tested preparations (Fig 4C).

Astrocytes do not secrete IL-1β or IL-18 in response to inflammasome stimulation
As primary mixed glial cultures contain at least both glial cell types (astrocytes and microglia) and as microglia are extremely reactive immune cells, it is important to deplete them as much as possible from astrocyte cultures. Whereas microglia can be isolated to high purity by different techniques, astrocyte cultures mostly remain "contaminated" by microglia [37]. In order to analyze the purest possible astrocyte cultures in vitro, we performed double MACS isolation where astrocytes were negatively selected during the sorting [26]. In this way we were able to separate microglia and astrocytes efficiently starting from a mixed glial cell culture (MGC). We analyzed expression of different cell type specific markers (Iba1 and CD11b for microglia and GFAP for astrocytes) in the microglia fraction and in the astrocyte enriched culture-double MACS (AEC-M2), which were 98% and 99% pure respectively ( Figures A and B in S1 File).
In order to prime astrocytes for inflammasome activation we tested different products, such as LPS, P3C or different cytokines (IL-1β, TNFα and IFNγ) separately or in combination (CCM), and we observed that CCM was the most efficient in this cell type (Fig 5A). CCM treatment could induce transcriptional upregulation of Nlrp3, Casp-1, Il6, Nos2 and Cxcl10 (Fig 5A  and 5B), as well as the secretion of CXCL10 in the supernatant (Fig 5C). We were however unable to detect substantial IL-1β or IL-18 secretion by CCM-primed astrocytes treated with the NLRP3 inflammasome activators ATP, Nigericin, Aβ or the alternative activator α-synuclein (Fig 5D, 5E and 5G). In addition, treatment with the AIM2 inflammasome activator poly (dA:dT) also failed to induce IL-1β or IL-18 secretion in astrocytes (Fig 5F and 5G). The substantial amounts of CXCL10 secreted by astrocytes in response to CCM indicate that these cells can play important roles in the brain immune response, but are not able themselves to secrete inflammasome-dependent cytokines, such as IL-1β or IL-18.
In order to differently address the question whether astrocytes are able to activate the classical NLRP3 inflammasome, we used neurosphere-derived astrocyte cultures [27]. When these cells were treated with LPS or CCM, we observed very low or no expression of Asc and Il1b, similarly to astrocyte-enriched cultures ( Figure A in S2 File), suggesting again that astrocytes do not express a functional NLRP3 inflammasome under these conditions.

Discussion
IL-1β production has been linked to neuroinflammatory conditions and neurodegenerative diseases for a long time with both beneficial and detrimental effects described. However the potential role of the NLRP3 inflammasome in the local production of IL-1β in the brain is not very well understood. Recently several reports started to investigate inflammasome activation in neuroinflammatory and neurodegenerative conditions, in spite of this fact a lot of basic questions remain. We therefore aimed to characterize in more detail the in vitro capacities of microglia and astrocytes separately to respond to inflammasome activation, since this issue has not clearly been addressed. Especially concerning astrocytes, it is not clear if they are able to express TLR4 and/or produce IL-1β as some conflicting reports exist [24,38]. Analysis of this phenomenon has often been complicated by the fact that, according to purification protocols used, remaining microglia most probably "contaminate" astrocyte cultures. We observed that less than 5% remaining microglia could result in detectable IL-1β production and therefore we purified our astrocyte cultures by two successive MACS passages in order to carefully deplete microglia. Recently, Facci and coworkers demonstrated that rat microglia-free astrocytes are unable to secrete IL-1β, by depleting microglia with the lysosomotropic agent Leu-Leu-OMe, a completely different approach than our MACS purification [39]. We were unable to observe any substantial IL-1β and IL-18 secretion in astrocyte cultures in response to classical NLRP3 inflammasome stimuli. This could be due to the lack of ASC and weak NLRP3 expression in these cells. In addition we could not induce pro-IL-1β expression in astrocytes, although we tested several potential inducers, such as LPS, P3C or CCM, and longer time points. These observations were confirmed in neurosphere-derived astrocyte cultures that are completely devoid of microglia. It has also been shown that in human astrocytes IL-1β mRNA was not translated to protein, probably by an active repression mechanism [40].
Concerning the role of other NLR proteins in the CNS, there is one report describing NLRP2 inflammasome expression in human astrocytes [41]. We did however not detect NLRP2 expression in our mouse astrocyte cultures, suggesting species-specific differences in inflammasome components expression. Nlrp2 expression in the mouse seems to be restricted to ovaries and plays a role in early embryonic development as a maternal effect gene [42]. We were however able to detect NLRP1 expression in astrocytes, as also described before in several reports that link NLRP1 inflammasome activation in astrocytes and neurons to in vivo models of spinal cord injury and thromboembolic stroke [43,44]. In addition to classical inflammasome activators, such as ATP or Nigericin, we also tested AIM2 activation by the use of poly (dA:dT) in astrocytes. In all cases, these potent microglial inflammasome activators did not result in detectable IL-1β or IL-18 production in astrocytes. As both NLRP3 and AIM2 are dependent on ASC in order to activate caspase-1, the lack of ASC in astrocytes could explain this observation. ASC-independent inflammasomes, such as NLRP1 or NLRC4 inflammasomes that can directly engage caspase-1 could lead to induction of pyroptosis and/or cleavage and secretion of IL-18. In addition the non-canonical inflammasome activation pathway could be functional in astrocytes, as we could show that caspase-11 expression can be upregulated upon priming.
A neuronal NLRP1 inflammasome has been described as a preassembled complex consisting of NLRP1, caspase-1, ASC, caspase-11 and XIAP [44]. We could demonstrate expression of ASC, caspase-1 and caspase-11 in cortical neurons, although they were weakly expressed. In contrast, NLRP1, NLRP2 and Aim2 were relatively abundant in cortical neurons. Functional analysis of these complexes has recently started and needs further characterization in order to better understand the importance of inflammasomes in neurons and neuronal disorders [45].
Microglia share several features of peripheral macrophages when they are isolated in vitro. They respond to classical NLRP3 inflammasome activation by secretion of IL-1β and other cytokines, such as IL-18, IL-1α and TNFα. IL-1β and IL-18 production in response to particulate stimuli, such as MSU or Alum, is weaker than after activation with ATP or Nigericin. We observed that microglial inflammasome activation occurs through similar mechanisms than in BMDMs. We could show that potassium efflux, ROS production and cathepsin B function are also required for efficient inflammasome activation in microglia.
A recent report showed that microglial IL-18 secretion in response to stimulation with S. aureus is NLRP3-and ASC-independent [46]. In microglia, activation by S.aureus seems to trigger additional pathways that can lead to mature IL-18, whereas in macrophages S.aureus induced IL-18 production has been described to be NLRP3 inflammasome-dependent [47]. We show that IL-18 production in microglia in response to classical activators, such as ATP, is NLRP3-and caspase-1-dependent as the corresponding knock-out microglia are no longer able to secrete IL-18. As IL-18 is emerging as an important player in neuroinflammation, this observation is clearly opening interesting questions concerning the differential production of IL-18 in the brain [13].
We were interested in the potential implication of the NLRP3 inflammasome in neurodegenerative diseases and focused on in vitro activation of microglia and astrocytes by peptide aggregates related to these conditions. Several reports already establish a link between amyloidβ and inflammasome-dependent IL-1β secretion in macrophages and microglia [18,35,36]. Despite published reports we were unsuccessful in triggering inflammasome activation by the full length peptide Aβ 1-42 . We tested fibrillar and oligomeric forms from different providers and were unable to detect IL-1β production in primary microglia. However, when we tested the peptide fragment Aβ 25-35 , we could observe inflammasome activation consistent with another report [35], as monitored by IL-1β secretion in microglia, but in a P2X7R-independent way (Fig 3E). The in vivo relevance of the NLRP3 inflammasome in Alzheimer's disease has recently been demonstrated in the APP/PS1 mouse model where NLRP3-deficient mice show less severe disease symptoms [19]. The conformation of Aβ peptides seems to be of great importance for immune activation in vitro and may explain the different results obtained [48]. In addition, several reports describe the ability of Aβ to directly prime microglia, resulting in proIL-1β production. We could not observe this with the Aβ 25-35 peptide we used ( Fig 3A); the Aβ 1-42 may have more pro-inflammatory capacities. However, priming microglia or macrophages with LPS before activating the inflammasome with Aβ may be required for a robust activation and detection of IL-1β in vitro, as suggested by several published studies [18,35,36,49,50].
In correlation to Aβ, we tested the capacity of α-synuclein, a PD-related peptide aggregate, to activate the inflammasome in microglia. α-syn treatment did not result in mature IL-1β secretion in our hands, although we tested different kinetics and concentrations. Opposed to this are recent reports describing substantial triggering of the inflammasome by aggregated αsyn in human monocytes and THP1 cells [51,52]. One explanation might be the different aggregation forms used in our study, in addition inflammasome activation in mouse macrophages and in human monocytes occurs through different mechanisms [53]. Astrocytes treated with Aβ [25][26][27][28][29][30][31][32][33][34][35] or α-syn, did not secrete any IL-1β or IL-18 ( Fig 5E) and only very modest levels of CXCL10 (data not shown), indicating that astrocytes do not strongly react to these peptides, at least in vitro. During the course of neurodegenerative diseases, it may be that microglial inflammasomes are preferentially activated in response to neuronal death and associated ATP release rather than by the presence of senile plaques and aggregated peptides. A recent study also underlines the fact that adult microglia considerably differ from neonatal microglia in their gene signature and this could also explain functional differences in inflammasome activation in the adult diseased brain [54].
It remains crucial to analyze microglia-astrocyte crosstalk in vivo to fully understand the importance of NLRP3 inflammasome function in the brain and the potential therapeutic interests in neurodegenerative diseases, yet our results suggest that microglia are the main cell type in the brain responsible for IL-1β and IL-18 secretion.
Supporting Information S1 File. Characterization of cell expression markers and inflammasome components in primary microglia, astrocytes and neurosphere-derived astrocytes.