α1-antitrypsin mitigates NLRP3-inflammasome activation in amyloid β_1–42-stimulated murine astrocytes

Alzheimer’s disease (AD) is the most common form of dementia with more than 40 million patients affected worldwide [1]. By 2050, the number is expected to quadruple [2]. Age is the most important risk factor [3], because the incidence of AD doubles every 5 years after the age of 65 years [3]. There is no causal treatment so far. To date, almost all biologicals or secretase inhibitors have failed in clinical trials which emphasize the need for further research into novel therapeutic options. Treatment of patients, even with early symptoms, only starts when the disease pathology has progressed and neural tissue has irreversibly been damaged for years. Therefore, current trials focus on patients with prodromal disease signs [4‐11].

Following the amyloid hypothesis, accumulation of extracellular Aβ_1–42-oligomers is one of the earliest and driving factors for pathogenesis of AD [12, 13]. The majority of in vitro studies investigated the effect of Aβ_1–42 after an incubation time of 24–72 h [14‐16]. Current scientific literature reveals less data about a possible damaging effect of Aβ_1–42-stimulation on the central nervous system (CNS) after short-term stimulation of only a few hours [14].

Besides Aβ_1–42, AD is mainly characterized by hyperphosphorylation of tau and neuroinflammation mediated by microglia and astrocytes, causing neuronal cell death [17‐21]. A key component of the innate immune system is the NOD-like receptor family, pyrin domain-containing 3 (NLRP 3) [22, 23]. Though ubiquitously expressed in CNS, NLRP3 is found highly expressed in Alzheimer’s patients’ brains [22‐24]. Under physiological conditions, an inactive form of NLRP3 is located in the cytoplasm [14, 25]. However, in the absence of activating signals, the NLRP3-inflammasome is not able to oligomerize [25]. After NLRP3-receptors recognize danger signals released by damaged cells and pathogens [26], NLRP3, the adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD) and pro-caspase 1 form a subcellular multiprotein complex, known as NLRP3-inflammasome [22, 23, 27]. Subsequently, pro-caspase 1 is activated by autoproteolysis and catalyzes the cleavage of the precursors pro-IL-1β and pro-IL-18 [27]. Mostly induced by microglia and astrocytes, secretion of pro-inflammatory cytokines IL-1β and IL-18 drives inflammatory responses and causes neuronal damaging [27‐29].

Inflammasomes are linked to neurodegenerative diseases: activated NLRP3 was observed in Parkinson’s disease in the midbrain and cerebrospinal fluid [30‐33]. Furthermore, in an experimental ischemic stroke model, NLRP3-deficiency was protective against ischemic neuronal damage [34].

In a cellular model of Alzheimer’s disease, Halle et al. 2008 first described that activation of NLRP3-inflammasome is induced by Aβ_1–42 in microglia, leading to an overexpression of the pro-inflammatory cytokine IL-1β [14]. Moreover, Aβ_1–42 activates the NLRP3-inflammasome in astrocytes [35]. Alike microglia, as a part of the CNS immune response, reactive astrocytes surround amyloid deposits and perform phagocytosis [35‐37]. Most studies investigated Aβ_1–42-mediated inflammatory processes in microglia and little is known about inflammasome activation in Aβ_1–42-stimulated astrocytes [14, 15, 24, 38‐40]. Since inflammation occurs as one of the first cellular and molecular responses after cell stress, short-term effects of Aβ-stimulation in astrocytes need further characterization. Aside from cell culture models, also in human models of AD high expression of NLRP3-inflammasome was found. More precise, an upregulation of NLRP3 expression in peripheral monocytes from individuals with AD was identified [16]. Moreover, in frontal cortex and hippocampus lysates from AD patients increased amounts of cleaved caspase 1 were detected [24]. Interestingly, NLRP3 deficient mice carrying genes associated with familial AD were protected from spatial memory deficits [24].

Therefore, a potent inhibition of the NLRP3-inflammasome could be a new therapeutic approach. The protease inhibitor α1-antitrypsin (A1AT) is known for its anti-inflammatory and anti-apoptotic properties in both hepatic and lung cells [41‐44]. Conveniently, A1AT is therapeutically used in patients with A1AT-deficiency and therefore well-established as a pharmaceutical agent. Recently, we demonstrated that A1AT also protected neurons from glutamate-induced toxicity [45] and reduced Aβ_1–42-induced inflammation in microglial cells [15]. In addition, we found that A1AT inhibited calpain and stabilized calcium-homeostasis [15]. This study investigated the regulation of NLRP3-inflammasome by A1AT in Aβ_1–42-stimulated murine astrocytes. In order to elucidate a direct interaction between A1AT and the NLRP3-inflammasome, we have included an inhibitor of NLRP3. MCC950 is a highly potent and specific inhibitor of NLRP3, without affecting AIM2, NLRC4, or NLRP4 [46‐49]. Recent data revealed that MCC950 stimulated Aβ-phagocytosis in vitro, and reduced Aβ-accumulation in a mouse model of AD, which was associated with improved cognitive function [50].

After primary astrocytes were stimulated according to work flow (Fig. 1a), culture purity of 95% astrocytes on average was examined by ICC, less than 5% of the remaining cells were microglia (Fig. 1b, Additional file 2: Figure S1 and Additional file 1: Figure S2).

Activation of glia cells and overexpression of pro-inflammatory cytokines are regarded early events in Alzheimer’s disease [56]. In recent years, astrocytes have come into focus in neurodegenerative disorders such as AD and are seen in a new way. Astrocytes express a plethora of receptors and modulate cells and their function in their surroundings [57]. In brief, they are involved in excitotoxic glutamate release, secretion of pro-inflammatory cytokines, growth factor production, stabilization, and organization of the blood-brain barrier and Aβ_1–42 production [58]. In microglia cells, the activation of NLRP3 appears to be an essential step during AD. Following activation of NLRP3, inflammation is triggered by the activation of caspase 1 and generation of IL-1β. This hampers the phagocytic capability of microglia cells [39]. The NLRP3-inflammasome is important for the initiation and processing of neuroinflammatory processes, and especially in AD, NLRP3 is associated with age-related inflammation [59]. NLRP3 is known to be activated by Aβ_1–42-aggregates [14]. NLRP3 knock-out in transgenic animals carrying mutations associated with AD prevents AD pathology [24].

Our group has recently presented data in acute and chronic neurodegenerative disease models that different components of the NLRP3-inflammasome are allocated to astrocytes [60‐64]. With respect to AD, Couturier and co-workers were able to show that astrocytes produce and release IL-1β following Aβ_1–42-stimulation [18]. In this animal model, the downregulation of NLRP3-inflammasome activation leads to decreased amyloid plaques and a better memory performance [35]. Our data now show that stimulation of primary astrocytes with Aβ_1–42 induces a dose-dependent upregulation of NLRP3. This in turn is known to stimulate the activation of caspase 1 and IL-1β [39]. Our work further demonstrates that such an effect also occurs after short-term stimulation with Aβ_1–42. In contrast, the co-treatment of astrocytes with Aβ_1–42 and A1AT blocks the induction of NLRP3. The regulation of NLRP3 is complex and usually requires a two-step activation process [22]. Currently, it is well accepted that the first step includes a priming signal, usually provided by NF-ΚB signaling or secretion of endogenous cytokines such as IL-1a [65]. NLRP3 is then activated by a variety of cellular signals, amongst them are misfolded proteins. The large variety of possible regulatory signals and pathways suggest that the activation is rather the consequence of a disturbance of cellular equilibrium [22]. In 2012, Lee and co-workers have identified calcium signaling as essential in NLRP3 activation [66]. In this study, increased intracellular calcium levels are associated with NLRP3 activation. In astrocytes, Aβ_1–42 potentiates calcium signaling which is triggered by mGlu, α7nAChR, and purinergic substances [57].

Yet, there is little evidence on the regulation of the NLRP3-inflammasome by A1AT. Toldo et al. 2011 stated that A1AT inhibits caspase-1 [67]. Aggarwal et al. 2016 found that—by the presence of polyunsaturated fatty acids—A1AT downregulates NLRP3 and caspase 1 [68]. For astrocytes, no data exist with respect to mechanisms of action of A1AT. In previous studies, we have presented data which show that A1AT reduces inflammation in microglia cells mainly by controlling calcium signaling pathways [15]. A1AT has no effect on classic signaling pathways such as MAPK p38, p44/42, JNK, and cAMP-coupled mechanisms [15]. Using a fluorescent calcium dye, we have shown that A1AT reduces intracellular calcium concentrations in a microglial cell line [15]. A1AT had no direct effect on Aβ_1–42-oligomerization [15]. We therefore hypothesize that A1AT effects on NLRP3 upregulation in primary astrocytes are mainly triggered by an inhibition of calcium and calpain. This hypothesis needs further evaluation, since other reports also demonstrate that A1AT is able to reduce glutamate-induced toxicity in murine primary neurons [45]. Since astrocytes release glutamate in response to Aβ_1–42-stimulation, this could represent another way how A1AT prevents deleterious Aβ_1–42-induced inflammatory cascades in microglia, neurons, and astrocytes. However, our current study does not include research on how A1AT could regulate the NLRP3-inflammasome complex. Further, it must be remarked that microglia might have partially contributed to the observed results due to the slight contamination of approximately 5%.

Our data demonstrate that NLRP3-inflammasome components are upregulated time-dependently following Aβ_1–42-stimulation. This can be blocked by A1AT application. NLRP3 is the sensor protein of the NLRP3-inflammasome complex, which explains its upregulation after Aβ_1–42-stimulation. ASC, in contrast, has the caspase activating and recruitment domain. In our experiments, ASC concentration on the mRNA and protein level was unchanged, indicating that Aβ_1–42 and A1AT mainly regulate NLRP3, but not ASC. We assume the total amount of ASC to be sufficient to lead to the NLRP3-ASC-complex formation and caspase 1 binding.

To further investigate the direct effect of A1AT on inflammasome-dependent IL-1β maturation, we used a specific NLRP3-inhibitor called MCC950. As previously shown by Coll et al., pre-treatment with MCC950 prevents complex formation of apoptosis-associated speck-like protein containing a CARD (ASC) and blocks the release of IL-1β in immunological active cells, without affecting priming of NLRP3 [46]. Furthermore, MCC950 stimulated Aβ phagocytosis in vitro, and it reduced Aβ accumulation in a mouse model of AD, which was associated with improved cognitive function [50]. In our studies, MCC950-pretreatment in cells co-treated with A1AT and Aβ did not further drop IL-1β protein expression. Thus, A1AT had no effect on protein expression, when NLRP3 was selectively blocked. We therefore conclude that A1AT reduces IL-1β by inhibiting NLRP3-inflammasome.

These observations not only clearly highlight the importance of astroglial-related pro-inflammatory processes in the brain and in particular during AD, but also point at A1AT as a potent antagonist in astrocyte-dependent inflammatory signaling.

We demonstrate that Aβ_1–42-stimulation results in an upregulation of NLRP3, caspase 1, and its cleavage products in astrocytes. A1AT time-dependently hampers Aβ_1–42-triggered neuroinflammation by attenuating NLRP3-inflammasome expression. This suggests that A1AT offers a therapeutic opportunity for AD treatment.