Dual roles of interleukin-33 in cognitive function by regulating central nervous system inflammation

IL-33 is localized in the nucleus as a transcription factor and extracellularly as a cytokine [9]. IL-33, ST2L, and IL-1 receptor accessory protein (IL-1RAcP) are highly expressed in brain cells, in particularly microglia, astrocytes, and neurons [5, 6, 25]. As a cytokine, IL-33 is upregulated and released upon inflammatory stimulation after injury and is thought to play roles in neuroinflammatory diseases via its binding to the ST2 receptor [6, 7, 31, 32]. The IL-33/ST2 signaling pathway participates in a complex network of multicellular interactions between the immune and nervous systems in the CNS [33] and is also central for infection-induced cognitive impairments [3]. However, it is worth noting that the IL-33/ST2 signaling pathway has dual roles in a range of infectious and inflammatory diseases [34], mediating both pathological immune responses and anti-inflammation neuronal responses [10, 31, 35].

In this part, the signaling pathways through which IL-33 exerts its protective effects on cognition are summarized. In IL-33^−/− mice (IL-33 gene deleted mice) and myeloid differentiation factor 88 (MyD88)-deficient mice, the phenomenon of long-term potentiation (LTP; one of the most commonly used models of measuring hippocampal memory) was abolished, and both the acquisition and retention latencies in the Morris water maze test (MWM; one of the most common behavioral methods for assessing spatial learning and memory) were increased [36]. This indicates that the IL-33/ST2 pathway may regulate spatial learning and memory by targeting MyD88 [36]. As a member of the IL-1 family, IL-33 acts through a signaling pathway similar to other cytokines in the family [5]. ST2 has been described as a negative regulator of Toll-like receptor-IL-1 receptor signaling, and functions as an important effector molecule of T helper type 2 responses [5, 37, 38]. In detail, IL-33 binds a cell membrane receptor composed of ST2L and IL-1RAcP, inducing the recruitment and activation of MyD88 to the receptor complex. Subsequently, the stimulated receptor-associated MyD88, as an adaptor protein, facilitates the recruitment and activation of interleukin-1 receptor-associated kinase 1 (IRAK1), interleukin-1 receptor-associated kinase 4 (IRAK4), and tumor necrosis factor receptor-associated factor 6 (TRAF6) [5, 39]. IRAKs further induce the downstream phosphorylation of IκB-α and IκB kinase (IKK), triggering the nuclear translocation of nuclear transcription factor-κB (NF-κB) and activation of C-Jun N-terminal kinase (JNK). In addition, the activated MyD88 induces the phosphorylation of extracellular signal-regulated kinases (ERK) and p38 mitogen-activated protein kinase (MAPK) [40], which constitute the MAPK signaling pathways along with JNK, leading to the triggering of activator protein-1 (AP-1). Finally, NF-κB and AP-1 transcription factors may induce the production of Th2-associated cytokines, including IL-4, IL-5, and IL-13 [41, 42] (Fig. 1). IL-4 and IL-13 induce arginase I expression via signal transducer and activator of transcription 6 (STAT6), and arginase I efficiently competes with inducible nitric oxide synthase (iNOS) for the substrate L-arginine, causing a decrease of NO generation in astrocytes and microglia [43] (Fig. 2). These Th2-associated downstream responses might balance the inflammatory Th1 responses to protect neurons from NO damage. The poor performance of IL-4- and IL-13-deficient mice in the MWM is accompanied by a decrease of brain-derived neurotrophic factor (BDNF) in hippocampal astrocytes. BDNF can promote learning-dependent synapse formation, which contributes to synaptic longevity and plasticity as well as the formation of neuronal circuits [44]. This indicates that IL-13 and IL-4 may also enable astrocytes to produce BDNF to aid spatial learning and memory consolidation [45]. In addition, the binding of intracellular IL-33 to NF-κB transcription factors was also found to contribute to the down-regulation of NF-κB and anti-inflammatory signaling in human HEK293RI cells (human embryonal kidney cells) in vitro. IL-33 interferes with the binding of p65 to κB consensus binding sites and reduces NF-κB/p65-triggered gene expression to dampen this pro-inflammatory signaling pathway [46]. Similarly, recombinant IL-33 was also observed to significantly reduce the nuclear translocation of NF-κB and disrupt its downstream pro-inflammatory mechanisms in monocytes derived from patients with AD and mild cognitive impairment (MCI) in vitro [47] (Fig. 2).

However, when released in excess of the physiological level, IL-33 may act as a pro-inflammatory cytokine to activate inflammatory cells, including astrocytes, microglia, and mast cells, in the CNS. IL-33 may activate the NF-κB signaling pathway, which upregulates numerous pro-inflammatory genes [48] and was implicated in the up-regulation of pro-inflammatory cytokines, chemokines, and neurotoxic mediators, such as IL-1β, IL-6, IL-8, IFN-γ, TNF-α, chemokine ligand 2 (CCL2), glia maturation factor (GMF), NO and ROS, resulting in inflammatory responses and cognitive dysfunction in neurodegenerative diseases [34, 49‐52] (Fig. 2). Thus, we speculate that IL-33 binding to ST2 may affect cognition through pro- or anti-inflammatory effects, which are mediated by the NF-KB signaling pathways in ST2-expressing cells, especially microglia and astrocytes.

The cytokines derived from T helper 1 cells (Th1) and T helper 17 cells (Th17), most prominently IFN-γ and IL-17, promote inflammation and subsequent secondary brain damage [96], while Th2-derived cytokines IL-4 and IL-13 have been shown to preserve spatial learning ability and memory by reducing inflammation and promoting neurogenesis [45]. A positive association between the serum levels of IL-17 and the montreal-cognitive assessment scores was noted in patients with schizophrenia, particularly in visuospatial and executive functioning, as well as in language functioning and delayed recall [97]. Regulatory T cells (Tregs) may also alleviate cognitive decline in mouse models of AD, possibly by constraining neuroinflammation through the upregulation of IL-10 [98]. The imbalance between Th1/Th17 and Th2/Tregs in the CNS was thought to be crucial for the neuroinflammatory and cognitive impairment in multiple neurological brain disorders, such as MS, brain injury, and ECM [79, 99‐101]. IL-33 is a potent inducer of type 2 immunity [102]. Th2 [103], mast cells, and ILC2 [104] express ST2 at significant levels, linking IL-33 to type 2 immune responses [105]. Mast cells and Th2 can respond to IL-33 stimulation, which leads to a significant increase in the production of Th2-associated anti-inflammatory cytokines, including IL-4, IL-5, and IL-13 [41, 42, 106, 107]. IL-33 can also induce Tregs to produce anti-inflammatory cytokines, including IL-10, IL-35, and TGF-β [108‐110]. In addition, IL-33 was found to decrease the secretion of Th17/Th1-associated pro-inflammatory cytokines, including IL-17, IL-6, IL-12, IFN-γ, IL-1β, and TNF-α [106, 108]. Hence, IL-33 may promote the transformation of Th cells into anti-inflammatory Th2 and Tregs, while also inhibiting their polarization into Th1 and Th17 (Fig. 4). This promotes the secretion of anti-inflammatory factors while at the same time inhibiting the secretion of pro-inflammatory factors, which may reduce neuroinflammation and neuronal damage.

The effect of IL-33 on the polarization of Th cells has also been demonstrated in several disease models, which are summarized in this review. IL-33 or ST2 knockout exacerbated spinal cord demyelination and EAE progression by promoting the development of pathogenic effector Th cells, including Th1 and Th17, while suppressing Th2 cells and Treg cells in the CNS of EAE mice [78, 111‐113]. Conversely, IL-33-treated EAE mice exhibited significantly reduced amounts of serum IL-17 and IFN-γ, along with increased levels of IL-5 and IL-13 compared with control mice [84]. Hence, IL-33 may play a protective role in EAE by promoting Th2 and inhibiting Th1/Th17. Mice with induced cerebrovascular occlusions, such as middle cerebral artery occlusion (MCAO) and unilateral common carotid artery occlusion, are commonly employed as experimental models of ischemic stroke [100, 114]. In mouse models of ischemic stroke, memory deficits in novel object recognition tasks after acute brain ischemia were associated with increased brain levels of the pro-inflammatory cytokines TNF-α and IL-1β [114]. In MCAO mice, administration of recombinant IL-33 promoted Th2-type effects following focal ischemic stroke, resulting in increased plasma levels of IL-4, IL-5, and IL-13, as well as attenuating the stroke-induced increases of the pro-inflammatory cytokines IFN-γ, IL-17, IL-1β and IL-6 [100, 115, 116]. In addition, IL-33 treatment increased the number of Tregs in the ischemic brain as well as the levels of IL-10, IL-35, and TGF-β in serum and brain tissues of MCAO mice [109, 110]. Thus, IL-33 treatment attenuated the severity of brain injury caused by dysregulated inflammation [100] and may mitigate cognitive impairment after focal cerebral ischemia. Similarly, in the oxygen–glucose deprivation neural cell model, IL-33 treatment effectively increased the concentration of IL-4 and TGF-β, while significantly reducing the level of IL-17 [117]. Additionally, in a murine model of ECM, IL-33 induced the release of Th2 cytokines IL-4, IL-5, and IL-13, as well as the expansion of Tregs, while suppressing the pro-inflammatory response, which alleviated cognitive impairment and prevented ECM development [3, 79]. These studies indicate that IL-33 may be at least partly responsible for an induced immune-shift from Th1 to Th2 response and suppressing the Th17 immune response to exert anti-inflammatory effects and reduce cognitive impairment after CNS injury.

However, some studies also showed increased concentrations of IL-33 in patients with MS and mouse models of EAE [113, 118]. In EAE mice, IL-33 treatment enhanced IFN-γ and IL-17 production by Th cells, while anti-IL-33 treatment reversed this increased and suppressed the development of EAE [113]. IL-33 may be involved in the pathogenesis of the EAE and MS through the reinforcement of Th17 and Th1 cell functions [113, 118], possibly correlated with excessive IL-33.

This review explores the possibility that IL-33 affects cognitive function by regulating cellular autophagy and apoptosis. Autophagy is a lysosome-dependent cellular degradation program. Moderate autophagy is an essential pathway for maintaining cellular homeostasis, but deregulated autophagy causes neuronal cell death, hippocampal shrinkage, and ultimately leads to a loss of synaptic plasticity [119]. Similarly, apoptosis is an evolutionarily conserved cell death pathway. Moderate apoptosis is necessary for normal development and maintenance of tissue homeostasis [120], while excessive neuronal apoptosis caused by excessive levels of pro-inflammatory molecules will accelerate neurodegeneration and cognitive impairment [14, 121, 122]. Endoplasmic reticulum stress (ERS) also leads to calcium dysregulation, which induces cell death pathways associated with autophagy and apoptosis, including activation of initiator caspases and the modulation of Bcl-2 family members [123]. Abnormal neuroinflammation, autophagy, and apoptosis are crucial factors of neuronal cell damage, which leads to the onset and progression of cerebrovascular disorders and cognitive dysfunction [124]. High levels of autophagy-related molecules such as Beclin-1, upregulation of pro-apoptotic proteins such as caspase-3, cleaved-caspase-3 (CC-3), Drp1, and Bax, as well as the downregulation of anti-apoptotic proteins such as Bcl-2 in the brain are associated with the decline of learning and memory function in a rat model of iron overloaded vascular dementia [125].

Some studies have shown that IL-33 may protect cognitive function by inhibiting excessive autophagy and apoptosis [85‐87, 126]. For example, IL-33 significantly inhibited the trauma-induced upregulation of the ERS-related protein glucose-regulated protein 78 (GRP-78; localized in the lumen of the ER, and involved in the folding and assembly of proteins in the ER), pro-apoptotic proteins caspase-3, CC-3, and Bax, as well as the autophagy-related molecule Beclin-1, while also downregulating the anti-apoptotic protein Bcl-2 in the hippocampus of TBI, ICH and MCAO model mice [86, 87, 127]. IL-33 pretreatment remarkably improved the motor function and MWM test scores in these mouse models. Notably, these positive effects of IL-33 can be blocked by anti-ST2 treatment [86, 87]. Thus, IL-33/ST2 signaling may mitigate TBI- and ICH-induced cognitive impairments through the inhibition of ERS, autophagy, and apoptosis [86, 87, 128]. Similarly, RNS induced significant increases in GRP-78, Drp1, and CC-3 expression and a decrease in Bcl-2 expression in the hippocampus, leading to neurobehavioral disorders, along with deficits of spatial learning and memory. Notably, these effects could be mitigated by IL-33 pretreatment in RNS mice [85].

Continuous synaptic remodeling in the hippocampus is essential for the encoding and retaining of learning and memory [129]. Synaptic plasticity, the ability of neurons to alter the structure and strength of synapses, is important for memory formation and consolidation [130]. Synaptic dysfunction is manifested in several ways, including loss of synapses, reduced synaptic function, and impairment of synaptic plasticity, which are the result of CNS insults and prolonged inflammation linked to cognitive deficits [14, 65, 131]. Synaptic impairment in its various forms is not only a hallmark of late-stage neurodegenerative diseases, such as Parkinson’s disease [132] and AD [133], but also a feature of the early progression of dementia [65]. Cytokines such as TNF-α, IL-1β, IL-18, and IL-6, as well as other molecules such as ROS, NO, and prostaglandin E2, are overexpressed during the neuroinflammatory process. These signals deeply influence neurotransmitter release and post-receptor signal transduction mechanisms, ultimately affecting synaptic function through a-calcium–calmodulin-dependent protein kinase II, MAPK, and ERK pathways or by directly influencing the synaptic memory processes [134]. At low physiological levels, these immune mediators might be essential for the induction and maintenance of neuroplasticity, but their overexpression during a neuroinflammatory process might result in neurodegeneration and the impairment of synaptic plasticity [134].

In this review, we summarize that the synaptic plasticity of neurons is also influenced by IL-33 in several studies [24, 135‐137]. Neuron-microglia communication through IL-33 is required for precision memory at remote time points, and IL-33 promotes microglial functions to optimize experience-dependent neural circuit plasticity, suggesting that IL-33 may play a role in memory consolidation [136]. IL-33 treatment significantly reversed the LTP impairment and contextual memory deficits of APP/PS1 mice, while also improving habituation to a new environment, which confirmed that IL-33 treatment reversed the reduction of synaptic plasticity in the hippocampus and cognitive deficits in APP/PS1 mice [2, 24]. The cornu ammonis 1 (CA1) sub-region constitutes the primary output node of the hippocampus, which is thought to be essential for most hippocampus-dependent memories [138]. IL-33 administration promoted the formation of functional excitatory synapses in hippocampal CA1 neurons in vivo, whereas conditional knockout of IL-33 in CA1 astrocytes decreased the number of excitatory synapses [137]. Importantly, blockade of IL-33 by intracerebroventricular administration of its decoy receptor inhibited homeostatic synaptic plasticity in CA1 pyramidal neurons in vivo, thereby impairing spatial memory formation in mice [137]. Thus, IL-33 may play a key role in learning and memory by enhancing synapse formation. The effect of IL-33 on synaptic plasticity modulation in HIV-infected human astrocyte cultures was indicated by an inverse correlation of IL-33 and myocyte enhancer factor 2C (MEF2C), a transcription factor that regulates synaptic function in the hippocampal region. A low concentration of IL-33 (120 ng/L) may exert a protective effect on synaptic plasticity with an increased MEF2C, while the higher dose of IL-33 (480 ng/L) led to a significant down-regulation of MEF2C and had a negative effect on synaptic plasticity. It was therefore speculated that the accumulation of IL-33 negatively impacted the synaptic plasticity, which may potentially be explained by the pro-inflammatory feedback effect of IL-33 [55, 135].