The study findings were published in the peer reviewed journal: Molecular Psychiatry.
https://www.nature.com/articles/s41380-022-01831-0
Several recent clinical studies have documented increased inflammasome activity in response to SARS-CoV-2 infection, leading to immune dysregulation that is associated with COVID-19 severity [55, 56]. In the periphery, it has been observed that monocytes from COVID-19 patients have increased inflammasome activation and undergo pyroptosis, which is associated with higher levels of plasma IL-1β in critically ill patients [53, 55], and the presence of ASC speck formation in lung macrophages [56].
In the CNS, a plurality of evidence has shown microglial activation in COVID-19 deceased patients, with enlargement of cell soma, and thickening of processes detected by staining for microglial markers [12, 13, 57]. The colocalization of SARS-CoV-2, ACE2, and NLRP3 within neurons, astrocytes, and microglia was also found in cerebral cortical tissue in three post-mortem COVID-19 cases [58].
Similarly, microglial activation has been confirmed in multiple preclinical in vivo models of COVID-19 [14,15,16]. Aligned with these results, here we show for the first time that SARS-CoV-2 invasion in the brain following intranasal infection of K18-hACE2 mice, leads to extensive microglial NLRP3 inflammasome activation.
While the ability of SARS-CoV-2 to enter the nervous system and to infect and replicate in CNS cells has been studied extensively, there are contradicting findings [59]. A remaining debate is whether the observed neurological manifestations in COVID-19 are attributed to virus invasion into the brain, or as a consequence driven by systemic factors or comorbidity.
In addition, SARS-CoV-2 genetic material and viral replication was recently reported in astrocytes in the brains of five individuals who died from COVID-19 presenting with brain damage [61]. Moreover, in vivo experiments performed in hamsters provide evidence that the basement membrane of the BBB is disrupted after SARS-CoV-2 infection, supporting that SARS-CoV-2 can cross the BBB in a transcellular pathway [62].
The capacity of SARS-CoV-2 to disrupt the blood-CSF barrier has also been demonstrated using human brain organoids [63], and human induced pluripotent stem cell-derived brain capillary endothelial-like cells [64], supporting that SARS-CoV-2 can use this route to enter the CNS.
Apart from the BBB, there is evidence demonstrating that SARS-CoV-2 can enter the CNS through cranial nerves that innervate the olfactory mucosa [65], as well through the vagus nerve to the brainstem, by immunohistochemical detection of SARS-CoV-2 in vagus nerve fibres [66].
SARS-CoV-2 entry into host cells has been thoroughly described and is mediated by the binding of viral spike protein to the human receptor ACE2 [67]. The expression of ACE2 receptor in the normal brain shows contradictory findings, with studies demonstrating no or low protein expression in the brain using tissue microarrays in a body-wide analysis [68].
However, more recently, using a collection of available datasets at the mRNA level, relatively high levels of ACE2 were found in the choroid plexus and paraventricular nuclei of the thalamus [43], which was confirmed at the protein level, through immunohistochemical staining for ACE2 in the choroid plexus and ependymal cells [69].
Interestingly, in COVID-19-affected brains, ACE2 expression is upregulated in endothelial cells in the white matter, with a correlation of higher expression in patients with more severe neurological symptoms [69].
The expression of ACE2 in the diseased brain has also been studied in the context of neurodegenerative diseases, with upregulated ACE2 expression observed in Alzheimer’s disease brains [46]. In the context of PD, angiotensin type-1 receptor (AT1) and ACE2 autoantibodies were found elevated in PD patients compared with age-matched controls in serum and in cerebrospinal fluid (CSF), suggesting that dysregulation of renin-angiotensin autoantibodies could contribute to PD progression [70].
Notably, in relation to the present study, the development of autoantibodies against ACE2 [71] and AT1 [72] has been also found after SARS-CoV-2 infection. Our results show that MDMi microglial cells express ACE2 receptor and, although the level is relatively low, SARS-CoV-2 is able to enter these cells but does not establish viral replication.
This has also been shown for human in vitro differentiated myeloid dendritic cells (mDC) as well as M1 and M2 macrophages, where in contrast to Vero E6 controls, no infectious virus production of SARS-CoV-2 is observed up to 48 h after inoculation [73]. One limitation of the MDMi cells used in our study, is the potential for residual characteristics to be carried over from their monocyte precursors [41].
Notably however, our data in MDMi cells aligns with the work of Yang et. al. where it was demonstrated that human pluripotent stem cell (hPSC)-derived microglia also express ACE2 receptor and are permissive to SARS-CoV-2-pseudo-virus entry [49].
Additionally, this study also found low or undetectable levels of viral RNA in hPSC-derived microglia exposed to infectious SARS-CoV-2, offering further evidence that while ACE2 mediated viral uptake is possible, hPSC-derived microglia do not support SARS-CoV-2 replication [49].
We also show that SARS-CoV-2 can activate the inflammasome in human microglia, through the read-out of cleaved IL-1β, cleaved caspase-1, and ASC speck formation in the supernatant. As it has been previously demonstrated that the interaction between ACE2 receptor and spike protein can induce the hyperactivation of NLRP3 in endothelial cells [74], we investigated the role of spike-ACE2 interaction relative to NLRP3 activation in microglia using a prefusion-stabilized SARS-CoV-2 spike protein (S-clamp) [37].
To confirm that NLRP3 activation on MDMi was ACE2 dependent we used a soluble human ACE2 receptor (hACE2-FcM), an ACE2 inhibitor (MLN-4760), and a well characterised monoclonal antibody (3E8) [52]. All three approaches confirmed that spike protein can activate NLRP3 in human microglia-like cells through ACE2.
Although we demonstrated that spike protein can activate the NLRP3 inflammasome in human microglia, it is worth noting that SARS-CoV-2 also encodes other viral proteins that could be involved in inflammasome activation. SARS-CoV-2 is comprised of a nucleocapsid protein (N), spike protein (S), membrane protein (M), and envelope protein (E), in addition to a series of accessory proteins (ORF3a, ORF6, ORF7a, ORF7b, ORF8, and ORF10).
Previous studies with the original SARS coronavirus, SARS-CoV have shown that protein E and ORF3a activate NLRP3, forming multimeric complexes that act as ion channels activating the NLRP3 inflammasome with IL-1β release, driven through NF-kB [75,76,77]. Moreover, recent evidence demonstrated that N-protein interacts directly with NLRP3, promoting the binding of NLRP3 with ASC, facilitating NLRP3 inflammasome assembly indicating another distinct mechanism of direct inflammasome activation through interaction of a viral protein with NLRP3 [78]. Our findings now provide further information that SARS-CoV-2 spike protein contributes directly to activating the NLRP3 inflammasome through ACE2.
Priming of the inflammasome in cells is a process necessary to induce transcriptional up-regulation of NLRP3 and pro-IL-1β [79]. Our initial observation that the SARS-CoV-2 virus itself can trigger inflammasome activation in MDMi without the need for priming supports a role for vigorous virus-mediated inflammasome activation in vivo.
We also confirmed that spike protein alone can prime the inflammasome through NF-kB in MDMi, allowing for NLRP3 activation with classical inflammasome activators ATP and nigericin, as has been previously reported in human monocytes, macrophages, and human lung epithelial cells [80, 81]. These findings support that human coronavirus spike protein can induce innate immune responses through NF-kB signalling.
We previously documented that activation of microglial NLRP3 inflammasomes through α-synuclein fibrils is a major driver of dopaminergic neuronal loss in experimental PD [4]. The accumulation of α-synuclein aggregates, as seen in Lewy bodies, and their spread throughout the brain is correlated with the stages of PD progression [82].
Of importance to the present study, there are increasing reports of significant neurological complications from SARS-CoV-2 infection in human patients [12, 13, 17, 23, 60, 61, 83, 84]. The correlation between viral infection and the manifestation of Parkinson-like symptoms has been described for a variety of viruses including influenza virus, JEV, and West Nile virus (WNV) infection resulting in tremor, myoclonus, rigidity, bradykinesia, and postural instability [85].
The activation of microglial NLRP3 inflammasome in the brain has also been demonstrated after viral infection with JEV [7] and WNV [86]. Indeed, we utilised WNV-infected C57BL6/J mice in this study as a control for our SARS-CoV-2 infected K18-hACE2 mice, which similarly demonstrated upregulated microglial NLRP3 inflammasome activation (Supplementary Fig. 4).
Moreover, post-mortem analysis performed on WNV-infected individuals showed an increased level of α-synuclein [87]. This finding prompted the hypothesis that α-synuclein is upregulated during infection as an antiviral factor in neurons, where it is proposed to act as a natural antimicrobial peptide to restrict viral infection in the brain [87, 88]. However, a recent study indicated that there were no alterations in α-synuclein levels in serum and CSF of COVID-19 patients with neurological symptoms [89].
These findings suggest that the reported cases of parkinsonism after SARS-CoV-2 infection could be a consequence of an increased proinflammatory environment, mediated by blood brain barrier (BBB) disruption [90], peripheral cell infiltration [91], and microglial activation [92]. These processes could be enhanced in the presence of ongoing synucleinopathies, or risk factors such as aging and poor health.
In addition, a recent study demonstrated that NLRP3 inflammasome genetic variants are associated with critical disease in severe COVID-19 patients, especially in elderly male individuals with reduced sickness symptom complex (SSC) and with increased body mass index (BMI), hypertension, and diabetes type 2 [93].
In conjunction, all these factors could lead to an accelerated neuronal loss, correlated with the reported parkinsonism symptoms and possible susceptibilities to developing PD post-SARS-CoV-2 infection.
Here we addressed the impact of SARS-CoV-2 on microglia in presence of α-synuclein. We showed that SARS-CoV-2 promotes α-synuclein mediated NLRP3 inflammasome activation by priming MDMi through spike protein, providing ex vivo support for the negative impact of SARS-CoV-2 on neurodegenerative diseases such as PD.
It is also worth noting that there are several lines of evidence in the literature indicating neurological complications resulting from SARS-CoV-2 infection.
These include:
(i) Possible cases of neuroinvasion by SARS-CoV-2 as observed in humans and preclinical COVID-19 models [12, 56, 60, 94, 95];
(ii) Extended microglial activation with pronounced neuroinflammation as reported in brain autopsies obtained from deceased SARS-CoV-2 patients [12, 13, 96];
(iii) Significant deterioration of motor performance and motor-related disability in PD patients recovering from COVID-19 [97, 98];
(iv) Accumulation of hyperphosphorylated Tau and α-synuclein occurring beyond viral clearance in SARS-CoV-2 infected hamsters [15] and Lewy body formation in SARS-CoV-2 infected macaques [99].
Thus, our finding complements the knowledge-gap in molecular mechanisms by which SARS-CoV-2 may activate microglia and lead to neurological manifestations. Our data suggest that the spike protein-mediated priming and/or activation of microglia through the ACE2-NF-kB axis may promote NLRP3 inflammasome activation leading to neuroinflammation and neurological phenotypes.
Further, this process may be enhanced in the presence of neurodegenerative disease triggers such as α-synuclein aggregates, supporting a possible role for COVID-19 in triggering brain diseases such as PD. Pharmacological inhibition of the NLRP3 inflammasome upon SARS-CoV-2 infection can suppress immune overactivation and alleviate COVID-19 in preclinical mice models [100].
Here, we further highlight the therapeutic potential of inhibiting NLRP3-driven microglial activation in the COVID-19 brain using an oral drug approach with the brain-penetrant small molecule MCC950 [4].
Since NLRP3 inhibitors are currently in clinical development for neurodegenerative diseases, including PD [4, 101], these findings also support a potential therapeutic avenue for treatment of SARS-CoV-2 driven neurological manifestations.
