Implications of fractalkine on glial function, ablation and glial proteins/receptors/markers—understanding its therapeutic usefulness in neurological settings: a narrative review

Abstract

Background

Fractalkine (CX3CL1) is a chemokine predominantly released by neurons. As a signaling molecule, CX3CL1 facilitates talk between neurons and glia. CX3CL1 is considered as a potential target which could alleviate neuroinflammation. However, certain controversial results and ambiguous role of CX3CL1 make it inexorable to decipher the overall effects of CX3CL1 on the physiopathology of glial cells.

Main body of the abstract

Implications of cross-talk between CX3CL1 and different glial proteins/receptors/markers will give a bird eye view of the therapeutic significance of CX3CL1. Keeping with the need, this review identifies the effects of CX3CL1 on glial physiopathology, glial ablation, and gives a wide coverage on the effects of CX3CL1 on certain glial proteins/receptors/markers.

Short conclusion

Pinpoint prediction of the therapeutic effect of CX3CL1 on neuroinflammation needs further research. This is owing to certain obscure roles and implications of CX3CL1 on different glial proteins/receptors/markers, which are crucial under neurological settings. Further challenges are imposed due to the dichotomous roles played by CX3CL1. The age-old chemokine shows many newer scopes of research in near future. Thus, overall assessment of the effect of CX3CL1 becomes crucial prior to its administration in neuroinflammation.

Background

Isotropic fractionator reveals that human brain contains less than 100 billion glial cells and the ratio of glia to neurons is less than 1:1, earlier which was believed to be 10:1 [1]. Glial cells respond to neurons and facilitate neuroimmune interactions [2]. Glial cells such as microglia communicate with each other, astrocytes and neurons through intracellular calcium signaling [3]. Two-way communication between neurons and glia ensures normal functioning of the nervous system by facilitating axonal conduction, synaptic transmission, and information processing [2]. The same glia when undergo reactive changes become the hallmark of neurodegenerative diseases, as seen in the case of reactive microgliosis and astrogliosis [4, 5]. In this regard, no research has completely shown the overall influence of fractalkine (CX3CL1) on the physiopathology of microglia and astroglia or on reactive microgliosis and astrogliosis. Thus, the effects of CX3CL1 should be considered with respect to the expressed proteins or markers which play a crucial role in glial activity. The review holds its significance owing to the expectation that CX3CL1 can alleviate neurodegenerative conditions and considered to be a potential target to counteract neuroinflammation in brain [6]. However, the role of CX3CL1 signaling in brain injury is controversial and is not completely understood [7]. Nevertheless, the significance of CX3CL1 in neuron-glia communication is evident from the fact that neurons in central nervous system (CNS) constitutively express CX3CL1 [8], whereas microglia express CX3CL1 receptor (CX3CR1) [8, 9], and CX3CR1 deficiency shows premature ageing of transcriptome in microglia [9]. Similarly, astroglia also express CX3CL1 [10] and CX3CR1 [11]. Thus, a review on the modulating effects of CX3CL1 on the functionalities of glial cells through receptors/proteins/markers expressed by the latter will give the overall view of the impact of CX3CL1 in neuroinflammation.

Objectives

The review aims to decipher the following:

• Role of microglia in physiological and pathological conditions with special emphasis on the implications of CX3CL1.

• Effect of microglia ablation on pathological condition and the implications of CX3CL1.

• Role of CX3CL1 in modulating different proteins/receptors/markers expressed by microglia.

• Role of astroglia in physiological and pathological conditions with special emphasis on the implications of CX3CL1.

• Effect of astroglia ablation on pathological condition and the implications of CX3CL1.

• Role of CX3CL1 in modulating different proteins/receptors/markers expressed by astroglia.

• Overall assessment of the therapeutic potential of CX3CL1 in neuroinflammation.

Main text

Microglia

Physiopathology of microglia, their subtypes and the role of CX3CL1

Microglia play significant role in synaptic pruning and synaptic maturation [12]. Absence of pruning leads to an increased density of excitatory synapses and subsequent synapse plasticity damage which leads to the loss of memory [13]. During early life, microglia eliminate unnecessary synapses by engulfing dendritic spines and lack of microglial activity leads to an increase in PSD95 and SHANK3 proteins. An increase in SHANK3 protein is correlated with autism [14]. Downregulation of Akt/mTOR pathway and downstream signalling molecule such as p70S6K and eIF4B were found in idiopathic autism [15]. Another claim correlates upregulation of PI3K/AKT/mTOR pathway in the pathogenesis of autism [16]. In this regard, CX3CL1-CX3CR1 interaction is found to activate PI3K/AKT signalling [17] but can prevent mTOR expression [18]. This may raise concern on the practical implications of CX3CL1 in autism. However, it is noteworthy that absence of CX3CL1 or complement component microglia signaling pathway leads to the development of excess immature synapses [19]. Previous study has already proved the significance of CX3CL1 signaling in developmental pruning of neurons [20] and its relation with autism [21]. This may raise inquisitiveness about the roles played by the isoforms of CX3CL1 in the pruning process. It is believed that soluble CX3CL1 plays a role in signalling microglia through CX3CR1 and leads to their migration or proliferation, whereas membrane bound CX3CL1 plays a role in microglial recognition of synapse for an effective pruning [12]. Similarly, it is reported that signaling from both soluble and membrane bound isoforms of CX3CL1 can alleviate neurological deficits. However, they are very specific about their roles. Each type modulates specific pathological events and suppresses distinct diseases [22].

Microglia have many physiological and pathological involvements depending on their subtypes. In the CNS, microglia will be either in the resting or in the active form. Active microglia are classified as pro-inflammatory (M1) type and anti-inflammatory (M2) type [23, 24]. M2 microglia are further classified as M2a (involved in repair and regeneration), M2b and M2c (immunoregulatory property, phagocytosis and removal of tissue debris) [25, 26]. Activation of these types depends on different stimuli. One such stimulus which plays a role in the pathological activation of microglia (M1 type) is hypoxia. Hypoxia exacerbates amyloid beta (Aβ) and tau pathologies, which in turn act as the key triggers for Alzheimer’s disease (AD) [27]. Activated M1 microglia release cytokines such as interleukin-α (IL-1α), IL-1β, IL-6, IL-12, IL-23, cyclooxygenase 2, tumor necrosis factor α (TNF-α), reactive oxygen species (ROS), and complement factors. Their continuous release initiates neuroinflammation and neurodegeneration [28,29,30]. This explains the relation between reactive microgliosis and neurodegenerative diseases, which are characterized by Aβ plaque formation, neurofibrillary tangle (tau deposits) and chronic neuroinflammation [4, 5]. All the above facts show the significance of hypoxia in neuropathology and underscore the need to examine the effects of CX3CL1 in this setting. In this regard, CX3CL1 plays a neuroprotective role in permanent middle cerebral artery occlusion-induced ischemic injury. CX3CL1 reduces caspase-3 activation [31]. Suppression of caspase-3 will prevent neuronal cell death, functional decline and Aβ plaque formation in AD brain [32]. Even tau protein-related neurotoxicity can be modulated by suppression of caspase-3 [33]. It was found that M1 microglia mediated neurodegenerative effect triggers an increase in CX3CR1 expressing microglia [34]. This facilitates internalization of extracellular tau within the microglia through CX3CR1 and subsequent clearance. However, the binding efficacy of phosphorylated tau is somewhat less than non-phosphorylated tau. Irrespective of the difference in binding affinity, CX3CR1 will reduce the burden of extracellular tau on neighbouring neurons and alleviate neurodegeneration and AD [35]. CX3CL1 may increase the expression of CX3CR1 [36]. Further, CX3CL1 can prevent tau-induced microgliosis by activating nuclear factor erythroid 2–related factor 2 (NRF2) and heme oxygenase 1 [37]. CX3CL1 also reduces the expression of IL-1β, IL-6 and TNF-α. Disruption of CX3CL1-CX3CR1 signaling triggered IL-1β mediated decreased survival and proliferation of neural progenitor cells [38]. A decrease in IL-1β alleviates tauopathy-induced cognition deficits [39]. Similarly, a decrease in IL-6 and TNF-α will have neuroprotection because IL-6 induces tauopathy [40] and TNF-α is known for producing Aβ plaque and hyperphosphorylation of tau [41]. Contrary to the negative role of TNF-α, prophylactic administration of IL-1α, IL-1β, IL-6 and TNF-α has neuroprotective effect against N-methyl-D-aspartate (NMDA)-induced neurotoxicity. In this regard, neurons can produce cytokines including TNF-α [42]. This proves the existence of a thin line of demarcation between neuroinflammatory and neuroprotective role of the mediators and the modulating effect of CX3CL1-CX3CR1 interaction on the expression and role of these mediators.

Transition of the M1 microglia into M2 type is the hall mark event in suppressing neuroinflammation and neurodegeneration. M2 microglia augment wound healing through high levels of arginase-1 and cluster of differentiation 206 (CD206) [24, 29]. In this context, CX3CL1-CX3CR1 interaction leads to the increased expression of arginase-1 [43]. Similarly, CX3CR1 expressing cells also express CD206 [44], indicating some kind of positive correlation between CX3CR1 and CD206. However, the beneficial effects of arginase-1 expression in M2-mediated neuroprotection are contradicted by a research where inhibition of arginase-1 and arginase-2 has a protective role against middle cerebral artery occlusion-induced ischemic stroke (Fig. 1a) [45]. This raises concern on the beneficial role of CX3CL1 in the context of ischemia. Contextually, other reports claim that neurons constitutively express and release CX3CL1 in response to excitotoxic and ischemic insult. CX3CL1 protects neurons from ischemic and excitotoxic insult by increasing the expression of CXCL16 mRNA in microglia and consequent increased secretion of CXCL16 [46]. Amidst the controversial role of CX3CL1 in ischemia/hypoxia-induced neuroinflammation, the alternative general approach to predict its effect in neuroinflammation is by observing the expression pattern of M1/M2 microglia. Neuronal CX3CL1 regulates the expression of M1 microglia by a negative feedback mechanism [47]. CX3CL1 suppresses the activation of M1 microglia by suppressing the pro-inflammatory glial gene. CX3CL1 suppresses the expression of genes favouring glycolytic pathway and hinder the expression of M1 microglia [48]. In this setting, CX3CR1 plays an important role. CX3CR1 expression can be seen in both M1 (low level) and M2 (high level) microglia. CX3CL1 converts M1 to M2 microglia by increasing the expression of CX3CR1 [36, 49] and subsequent binding with CX3CR1 [26]. Latest research also corroborates the role of CX3CL1 in transforming M1 to M2 by interacting with CX3CR1 [26]. The intricate relation between cytokines, CX3CL1, CX3CR1 and M2 microglia activation can be understood through different findings where both anti and pro-inflammatory cytokines such as IL-4 and TNF-α stimulate the expression of CX3CL1 [50] and anti-inflammatory cytokines such as IL-4, IL-10 or IL-13 initiate M2 activation [30, 51]. Contextually, IL-10 plays a significant role by increasing the expression of CX3CR1 in lipopolysaccharide (LPS) treated cell line [52]. All these previous results along with the recent ones strongly support the role of CX3CL1 in regulating the expression and activation of M2 microglia. This also shows the role of different mediators (pro and anti-inflammatory) in establishing communication between neurons and glia through the expression of CX3CL1 and CX3CR1, respectively. These processes are regulated by feedback loop, because CX3CL1 can reduce the expression of TNF-α [38, 53], at the same time can also increase the expression of TNF-α and other cytokines (Fig. 1b) [7]. This shows the complex bidirectional role of CX3CL1 in the expression of inflammatory mediators.

Fig. 1

a Neuronal CX3CL1 and microglial CX3CR1 interaction increases proinflammatory cytokines in ischemic brain leading to an increase in ROS and neurodegeneration. CX3CL1–CX3CR1 interaction can also decrease the expression of cytokines and neutrophils following activation of CREB and provide neuroprotection. By contrast, suppression of CREB provides neuroprotection. CX3CL1-CX3CR1 interaction increases arginase-1 and provide neuroprotection. By contrast, inhibition of arginase-1 and 2 prevents stroke in ischemic brain. b CX3CL1-CX3CR1 interaction can increase as well as decrease the expression of cytokines. Increase in cytokines causes neurodegeneration via activation of M1 microglia, whereas a decrease leads to neuroprotection. CX3CL1–CX3CR1 interaction provides neuroprotection via conversion of M1 to M2 microglia. c CX3CL1–CX3CR1 interaction may cause hyperphosphorylation of tau which may lead to neurodegeneration. Increased expression of CX3CL1 prevents tau-CX3CR1 complex formation and tau clearance. CX3CL1–CX3CR1 interaction decreases Aβ deposition. Decreased Aβ deposition occurs via decrease in caspase-3, as a result of an increased expression of CX3CL1. By contrast, increased level of CX3CL1 can decrease TLR4 expression, which leads to a subsequent decrease in the scavenging of Aβ. Deposited Aβ can bind with neuronal CX3CR1 causing neurotoxicity and neurodegeneration. d CX3CL1–CX3CR1 interaction can suppress C/EBPα mediated microglial activation and consequent neurodegeneration. C/EBPα may decrease the expression of proinflammatory cytokines and provide neuroprotection. C/EBPα may increase ROS and produce neurodegeneration. Decreased expression of C/EBPα activates TLR-4 and increases proinflammatory markers causing neurodegeneration. e CX3CL1–CX3CR1 interaction provides neuroprotection by increasing thrombospondins-1 which leads to a decrease in Aβ pathology. By contrast, thrombospondins-1 may increase the expression of post synaptically silent receptors which are linked with neurodegenerative disorders

Though CX3CL1 is known for its anti-inflammatory role in the CNS, yet there is a pathological concern associated with it. This is due to the fact that TNF-α liberated by M1 microglia, increases brain endothelial cell proliferation [54]. Stimulation of endothelial cell by TNF-α, interferon-gamma (INF-γ) and IL-1 liberates endothelial CX3CL1. Endothelial CX3CL1 mediates leukocyte recruitment and can produce atherosclerotic lesion [55, 56]. Intracranial atherosclerotic disease is again a common reason behind stroke [57], which eventually leads to neuroinflammation and secondary neuronal damage [58]. The detrimental role of CX3CL1-CX3CR1 interaction is proved in ischemic brain, which has shown elevated expression of TNF-α, IL-1β and IL-6 [7]. Similarly, another report related CX3CL1 signaling with neurotoxicity [59]. However, tagging CX3CL1 signaling with neurotoxicity is again contradictory with respect to certain other observation. This is partly owing to the relation between CX3CL1 and INF-γ. Unlike soluble CX3CL1, which has a negligible effect on INF-γ production, membrane bound CX3CL1 is a strong inducer of INF-γ by natural killer (NK) cells [55] and brain has clusters of NK cells [60]. It seems, CX3CL1 and INF-γ can induce expression of each other [55, 56]. This makes the pathological role of CX3CL1 debatable because knocking down of INF-γ receptor in the brain in comparison with periphery leads to higher incidence and severity of multiple sclerosis [61]. Contrary to the effect of CX3CL1-CX3CR1 interaction in ischemic brain [7], CX3CL1 can suppress the secretion of TNF-α, IL-1β, IL-6 and nitric oxide (NO) from M1 microglia [53]. Subsequent research has confirmed that CX3CR1 deficient microglia produce IL-1β and trigger inflammation [62]. This once again reflects the complex role of CX3CL1 in neuroinflammation. Further, the role of CX3CL1 in neuroprotection is evident from its function to impede demyelination of neurons [63]. It has a prominent role in suppressing α-synuclein mediated neurodegeneration and subsequent Parkinson’s disease [64]. This shows immense networking of CX3CL1 with different mediators.

Similarly, a decrease in CX3CR1 is related with an increased tauopathy [35]. As CX3CR1 binds with tau, hence, understanding the effects of tau-CX3CR1 binding is necessary owing to the report that such interaction disrupts the communication between neurons and microglia [65]. If this happens, it may impact synaptic pruning which is dependent on CX3CL1-CX3CR1 interaction [20]. The significance of synaptic pruning is already known from the fact that autism is characterized by an excessive synapse formation [21]. Interestingly, a report confirms that presence of CX3CL1 reduces the interaction of tau with CX3CR1, this is due to a higher binding affinity of CX3CL1 for CX3CR1 [35]. Thus, presence of tau in the vicinity of CX3CR1 should not affect CX3CL1-CX3CR1 interaction and communication between neuron and microglia. This raises concern on the long-term therapeutic administration of CX3CL1, because exogenous CX3CL1 may affect tau clearance through CX3CR1.

Effect of microglia ablation in neuropathophysiology and the impact of CX3CL1

The complex role of microglia in neurodegenerative disease progression could be understood from a target region specific ablation of microglia in AD model of adult mice. As per the report, microglial ablation in the target region led to a decrease in tauopathy and neurotoxicity [66] but had no impact on Aβ plaque formation [67]. Contextually, another report proves the involvement of reactive microglia in hyperphosphorylation of tau and its synaptic spreading, leading to tauopathy [68]. However, study involving occlusion of common carotid artery has rather shown an increase in infarct size (by 60%), dysregulated neuronal calcium response and neuronal death, following selective microglial ablation [69]. Thus, microglial ablation may have positive as well as negative consequences. In this regard, CX3CL1 can reduce cerebral infarct size and neurological deficits in rodents with permanent middle cerebral artery occlusion [31]. This underscores the significance of interaction between CX3CL1 and microglial CX3CR1 in the brain. Hypoxia activates M1 microglia [27] and they are associated with a larger infarct size [48] and glutamate toxicity [70], hence, depletion of M1 microglia should have positive outcome rather than negative one. This indicates pathology specific dichotomy of microglia. Similarly, depletion of CX3CR1 microglia in 6–12-week-old male mice has worsened acute seizures and increased mortality rate [71]. The result seems to be the consequence of M1 microglia deletion [71], but it may raise concern on the actual type of microglia (M1 or M2) whose deletion led to aggravated seizure outcome. This is owing to the expression of CX3CR1 by both M1 (low level) and M2 (high level) microglia and the role of CX3CL1 in converting M1 to M2 type [36, 49]. Thus, until the characterization of microglia and determining the rate at which CX3CL1 coverts M1 to M2, the concern remains unaddressed. It is noteworthy that healthy mice brains neither exclusively express M1 nor M2 phenotypic markers at any time [72] and under normal CNS functioning, the environment in CNS shows more skewing towards M2 microglia [73]. Thus, CX3CR1 microglia which are depleted in the study with seizure mice model [71] should be of M1 type. This once again proves the double edged role of M1 microglia and also raises question on the possible role of CX3CL1 in this particular setting.

In neuropathophysiology, the significance of M2 microglia is evident from the fact that synaptically released glutamate can initiate and spread seizure by acting on ionotropic and metabotropic receptors [74] and M2 microglia secrete brain-derived neurotrophic factor (BDNF) [75]. BDNF protects neurons against glutamate-induced neurotoxicity and death [76]. It is found that CX3CL1 plays a significant role by promoting the expression of BDNF [77]. This indicates that neuronal CX3CL1 can manipulate glial cells to secrete BDNF and provide neuroprotection.

The mysterious role of CX3CL1 can be observed where neither an increase nor a decrease in microglial count has affected Aβ plaque formation [67]. However, a decreased interaction between CX3CL1 and CX3CR1 is responsible for reduced Aβ deposition [78]. On the other hand, CX3CL1 reduces caspase-3 activation [31] and reduction of caspase-3 will prevent Aβ plaque formation in AD brain (Fig. 1c) [32]. This indicates the complex nature of CX3CL1 in neuroinflammation. Indiscriminate microglial ablation should increase deposition of tau and subsequent tauopathy as a result of ablation related decline in the population of CX3CR1. Concern to the matter is increased due to the fact that deposition of tau can promote microglial [79] and consequent astrocytic [80] activation, which in turn release cytokines [28, 81], this may lead to secondary tauopathy [82]. Thus, tauopathy and microglial/astroglial activation will help each other’s propagation, this will emerge as a vicious cycle of neuronal damage.

Above facts signify the interaction between neuronal CX3CL1 and microglial CX3CR1 in suppressing neuroinflammation and neurodegeneration. However, the relation between CX3CR1 and Aβ peptide remains ambiguous. The effects of Aβ peptide and tau on neuropathy are dependent on the membrane bound form of CX3CL1, rather than the soluble form of CX3CL1. As per a study, disrupted signaling from membrane bound CX3CL1 reduces the deposition of fibrillar Aβ, by contrast, favours hyperphosphorylation of tau [83]. Though membrane bound CX3CL1 is identified as a key player to modulate the effects of Aβ peptide and tau protein, yet other studies have proved that membrane bound CX3CL1 can partially restore spatial learning and memory [22]. The significance of membrane bound CX3CL1 can be further realized from the fact that soluble CX3CL1 originates following the cleavage of membrane bound CX3CL1 of neurons [84]. Soluble CX3CL1 partially corrects cognitive and motor function and is capable of restoring neurogenesis and long term potentiation. However, this property is not visible for the membrane bound form [22]. This emphasizes on the significance of both the forms of CX3CL1 in neuron-microglia communication and cognition. Lastly, the negative impact of microglial ablation can be established from the fact that neuronal CX3CR1 suppresses synaptic transmission and has been identified as a potential target of Aβ-42 to mediate neurotoxicity in the absence of microglia [59]. Thus, the effect of CX3CL1 differs based on the interaction with neuronal or microglial types of CX3CR1.

Role of other microglial markers in neuroinflammation and their modulation by CX3CL1

Other than CX3CR1, microglia express several other receptors/proteins such as cluster of differentiation receptors (CD11b, CD14, CD68), colony-stimulating factor 1 receptor (CSF1R), toll-like receptors (TLR), human leukocyte antigen (HLA)-DR, nucleotide-binding oligomerization domain (NOD) proteins, macrophage scavenger receptor (MSR)-A, C-type lectin receptors, Fcγ receptors [85,86,87,88,89], β2 receptors, α2A receptors [90]. Presence of myriad of receptors entrusts microglia with a complex role and triggers curiosity to look for any possible modulating effect of CX3CL1 on these receptors. Among different receptors, CSF1R (known as CD115) and TLR deserve special mention. CSF1R, predominantly expressed on microglial surface, plays a house keeping role including development and survival of microglia and a key player in neuroinflammation [91]. Previous study has proved that blockade of CSF1R can reduce proliferation of microglia/macrophages and gliosis, and this may suppress neuroinflammation and promote motor recovery [92]. CSF1 and IL-34 are the ligands for CSF1R [91], and both are responsible for many inflammatory conditions [93]. The role of CSF-1 in promoting neuroinflammation is evident [92] from the fact that antibodies to CSF-1 led to a decreased level of microglia in the white matter [94]. Growth and differentiation of microglia occur under the influence of secreted CSF-1 [95]. At the same time CSF-1 can increase microglia mediated neurotoxicity following Aβ deposition [96]. Though it raises concern over the role of CSF-1 in microglia mediated neuronal damage, yet the downstream events are opposite to the expectation. CSF-1 and CSF1R interaction activated cAMP response element-binding protein (CREB) and prevented apoptosis and inflammation. This occurs through a sequential activation of phospholipase C gamma 2 (PLCG2), protein kinase C epsilon (PKCε) and cAMP response element-binding protein (CREB). This interaction inhibited neutrophil infiltration and downregulated the expressions of IL-1β and TNF-α [95]. This draws attention towards CX3CL1-CX3CR1 interaction, because CX3CR1 signaling also activates CREB [48]. Further, a research has shown that CSF1R deletion-induced suppressed proliferation of microglia can be compensated by the infiltration of CX3CR1 positive cells, even though the blood brain barrier (BBB) is intact [97]. This perhaps indicates that CX3CL1-CX3CR1 signaling remains as a substitute for CSF1R signaling, and mediates activation of CREB and subsequent neuroprotection. However, CREB mediated neuroprotective role of CX3CL1 comes under question following the revelation that suppression of CREB prevents inflammation and provides neuroprotection [98].

Other than microglia, CSF-1 is expressed by neurons, astrocytes and oligodendrocytes, and its secretion promotes growth of microglia [95]. Thus, microglial proliferation may be regulated by a combined signaling from neurons, astrocytes and oligodendrocytes or by a predominant cell type. Though a study claims that CSF1R does not regulate microglial differentiation, yet the significance of CSF-1 with respect to regulating microglial density and distribution remains crucial [99]. Owing to the significance of CSF-1, it becomes necessary to pinpoint the most potential influencer of microglial density and distribution. As neurons generate earlier to glia in the developing mammalian cortex [100] hence, neuronal involvement in glial development through secreted CSF-1 is apparent. Even research has proved the pivotal role of neurons in manipulating microglia through CSF-1, during nerve injury [101]. It is noteworthy that CSF1R activation triggers tyrosine phosphorylation cascades which regulate many pathways involved in macrophage survival, differentiation and proliferation. One of these pathways such as janus kinase (JAK)/STAT signaling also depend on tyrosine kinase mediated activation of STAT1 and STAT3 [102]. Soluble CX3CL1-induced CX3CR1 internalization has decreased tyrosine phosphorylation of microglial STAT proteins [103]. This may affect microglial survival and proliferation. This finding is different from the observation where CX3CL1 is found to increase the population of M2 microglia via CX3CR1 [26, 36, 49].

Similarly, IL-34 (another ligand for CSF1R) promotes development of microglia [104]. IL-34 is constitutively expressed in the brain [105] by neurons [104]. Its role in neuroinflammation is evident from the fact that antibody to IL-34 led to a significant depletion of microglia in the gray matter [94]. However, previous study has proved the beneficial role of IL-34 in suppressing oligomeric Aβ neurotoxicity in AD [85]. It is worth mentioning that M1 microglia secrete many mediators including TNF-α, IL-1β [28,29,30] and IL-17 [106]. Eventually, TNF-α, IL-1β and IL-17 induce expression of IL-34 mRNA [107, 108]. However, among these cytokines, TNF-α has induced significant expression of IL-34 mRNA compared to IL-1β and IL-17 [108]. IL-34 so produced can prevent the production of TNF-α by downregulating the expression of Dectin-1 and TLR2 [109]. This represents a kind of neuronal feedback through IL-34 on microglial CSF1R to modulate neuroinflammation. In this setting CX3CL1 can suppress TNF-α mediated IL-34 expression by its ability to reduce expression and secretion of TNF-α [38, 53] and secretion of IL-1β from M1 microglia [53]. This not only reflects neuronal control on microglial cytokines via CX3CL1 but also represents the effect of CX3CL1 on neuronal IL-34. However, it is difficult to predict the actual impact of CX3CL1 on cytokine mediated expression and secretion of IL-34 because CX3CL1 can also increase the expression of TNF-α and other cytokines [7].

In CNS, the expression of TLR is limited to neurons, astrocytes and microglia [110]. Among different TLRs (TLR1 to TLR10) [110], only three are intracellular (TLR3, TLR7, TLR9) and remaining are present on the plasma membrane [111]. Among the TLRs, TLR4 deserves special mention because α-synuclein secreted by neurons leads to the activation of TLR4 on microglia, leading to its clearance [112]. The pathological aggregation is due to calpain-1 enzyme, which gives rise to C-terminal-truncated synuclein, which aggregates faster than its precursor α-synuclein [113, 114]. The truncated synuclein is the most potent activator of TLR4 receptors on microglia. Activated TLR4 mediates release of proinflammatory cytokine and ROS production by microglia [115]. Strangely, CX3CL1 can reduce the expression of TLR4 [116] but can suppress α-synuclein mediated neurodegeneration [64]. TLR4 has a scavenging effect on Aβ; however, prolonged expression of TLR4 leads to Aβ deposition in the brain [117]. Reduction of TLR4 by CX3CL1 may have dual effects on Aβ clearance.

Among different microglial receptors, β2 receptors mediate the surveillance function of resting microglia, whereas α2A receptors mediate the functions of active microglia, both receptors are activated by norepinephrine to facilitate microglia mediated surveillance [90]. Microglia mediated surveillance of brain parenchyma relies mainly upon two processes. One is lamellipodia (large processes with a bulbous tip) and the other is filopodia (bundles of parallel actin filaments) which extend from the tip of lamellipodia, as well as from the body of the microglia. A third process named uropod helps in the movement of the soma of microglia. Lamellipodia attached to the microglia actively monitor the brain parenchyma for any molecular signal of injury or harmful stimulus, even when microglia are under resting condition. Filopodia perform a faster screening of the surrounding compared to lamellipodia [89, 118]. Microglia mediated neuroprotection relies on fast nanoscale sensing of any kind of obnoxious stimulus with the help of filopodia [118, 119]. Rapid extension of filopodia requires an increase in intracellular cAMP concentration, initiated by the activation of metabotropic receptor coupled with G_s type of G-protein. In this aspect, β₂ adrenergic receptor activation by agonists (isoproterenol and norepinephrine) increases cAMP, leading to an increased turnover of filopodia [118]. This explains the role of norepinephrine and β₂ receptor in inducing filopodia and subsequent neuroprotection. Similarly, norepinephrine also induces release of soluble CX3CL1 by neurons and exerts anti-inflammatory effect on neurons via CX3CL1 [120]. The above statements perhaps give a hint regarding a correlation between CX3CL1 release and extension of filopodia. However, a study has put a question on such correlation because norepinephrine has been found to suppress CX3CL1 in the presence of lipopolysaccharide (LPS) [121], which produces neuroinflammation through IL-1β, TNF-α and IL-6 [122]. These cytokines increase production of ROS and a consequent ROS mediated signaling, which causes cellular pathology [123, 124]. M1 microglia are one among different sources of ROS. M1 microglia secrete nitrite [125], and nitrite can increase the amount of NO [126] responsible for neurodegeneration [125]. In this aspect, it is noteworthy that neuronal CX3CL1 can prevent the accumulation of nitrite in microglia [120], the release of NO from M1 microglia [48, 53] and suppress the secretion of TNF-α, IL-1β, IL-6 from M1 microglia [53]. Thus, suppression of neuronal CX3CL1 by norepinephrine in the presence of LPS raises question on the neuroprotective role of norepinephrine and its actual relation with CX3CL1. However, the significance of CX3CL1 remains undiminished due to its ability to inhibit the release of TNF-α from LPS stimulated microglia [127]. Additionally, study has confirmed the roles of CX3CL1-CX3CR1 signaling in facilitating faster screening of the environment by microglial processes as well as in migration of the microglia towards the source of injury [128]. Another report shows that CX3CL1 regulates macrophage migration in CNS and their conversion into microglia [47]. Irrespective of this fact, the role of CX3CL1 in the development of filopodia needs further exploration. In this aspect, it was found that cyclic guanosine monophosphate (cGMP) induced by NO can also induce expression of filopodia by increasing intracellular cAMP. The maintenance and motility of filopodia are again actin dependent but not microtubule dependent [118]. This further reveals the role of CX3CL1 in the development of filopodia, because interaction of CX3CL1 with CX3CR1 initiates the coupling of the latter with G-protein to initiate actin polymerization in microglia [129]. Actin polymerization and co-localization of ionized calcium-binding adaptor protein-1 (IBA-1) lead to filopodia development, microglial movement and phagocytosis of extra neuronal tau oligomers. This prevents the propagation of extracellular tau into surrounding neurons and further damage. This protects against neurodegenerative conditions such as AD [130]. The association between actin and CX3CL1 can also be assumed from the fact that disruption of actin facilitates the release of soluble CX3CL1 following an interaction between endothelial membrane bound CX3CL1 and metalloprotease [131].

Apart from β2 adrenergic receptor, CX3CR1 also facilitates the function of microglia as “sensome” by binding with CX3CL1. This enables microglia to perceive microenvironmental changes and perform a site specific motile response [89]. Soluble CX3CL1 has an overwhelming role in this regard, because it has more anti-inflammatory and neuroprotective affect compared to the membrane bound form [84]. However, previous research has proved that membrane bound form of CX3CL1 is a stronger inducer of microglial chemotaxis than the soluble forms of CX3CL1 [34]. This finding is different from subsequent claim which identifies soluble CX3CL1 as the inducer of microglial chemotaxis [12]. Membrane bound CX3CL1 rather enables microglia to identify the injured neurons and surround it [34]. This may have negative outcome because neuronal injury represents a site predominantly occupied by M1 microglia [132]. M1 microglia mediated massive release of glutamate leads to NMDA receptor mediated glutamate toxicity [70]. NMDA facilitates the disruptive effect of Aβ on dendritic spine [133]. Both soluble oligomer and fibrillar Aβ can activate microglia. However, oligomeric form is a stronger activator than the fibrillar type [134]. This way the vicious cycle of microglia mediated Aβ toxicity and Aβ mediated microglia activation continues to produce neuroinflammation and neurodegeneration. This gives the impression that CX3CL1 will play a negative role in the context of neuronal injury by signaling microglia. Conversely, after signaling microglial recruitment, CX3CL1 modulates their function and high level of CX3CL1 in damaged neurons protects against glutamate excitotoxicity and degeneration [29]. This indicates the significance of CX3CL1 mediated modulation of microglial function.

The significance of CX3CL1-CX3CR1 interaction is also apparent from the fact that a reduced interaction may lead to ageing of microglia [9]. Similarly, absence of CX3CR1 expression will decline neurogenesis in the hippocampus and olfactory bulb of adults, and will adversely affect integration of newborn neurons [9]. Further, lack of microglial CX3CR1 results in down regulation of NRF2 and consequent absence of heme oxygenase-1 gene expression. It also leads to a decrease in TAM family of receptor (Tyro-3, Axl and Mer), all together leads to neurological deficits [135]. On a similar note, CX3CL1 deficiency reduces the expression of class I multiple-synapse boutons, which are expressed under the influence of microglia-induced spine-head filopodia [19]. These multiple-synapse boutons represent synapse with more than one dendritic spine. Such arrangement is a prerequisite for hippocampus-dependent associative learning [136]. Similarly, spine-head filopodia formation under the influence of cholinergic stimulation indicates the significance of the former in cognition and in pathological condition like AD [137]. This proves the significance of CX3CL1 in cognition.

Microglia being the “main resident antigen-presenting cell of the brain” constitutively express HLA-DR molecule (MHC Class II). However, the expression level varies according to the location of microglia. It is reported that corpus callosum and the capsula interna of white matter present microglia with higher expression of HLA-DR compared to the microglia of the grey matter [138]. The fact regarding constitutive expression of HLA-DR on microglia is in dispute with previous research, which proves that HLA-DR is highly induced on microglia during AD and Parkinson’s disease [86]. Similarly, subsequent study claims that MHC Class II is undetectable on microglia, except for the activated form [6]. Recent study also recognizes MHC Class II as M1 microglia marker [139]. Its recognition as M1 microglia marker and role in neuroinflammation can be realized from the fact that IFN-γ induces the expression of HLA-DR, and TNF-α potentiates the role of IFN-γ in HLA-DR expression on microglia [140]. It is noteworthy that IFN-γ can induce release of TNF-α. TNF-α mediates maturation of antigen presenting cells, and has an indirect role in cellular apoptosis but has nothing to do with HLA-DR mediated cell death [141]. HLA-DR shows a significant positive association with diffuse plaque in population with or without dementia. Higher expression of HLA-DR is linked with poor cognition in AD [142]. Thus, if CX3CL1 has to prevent cell death then IFN-γ should be one of the targets of CX3CL1. It is noteworthy that CX3CL1 reduces expression of MHC Class II on microglia, and prevents neurotoxicity due to reactive microglia [6]. However, contradiction on the suppressive role of CX3CL1 on HLA-DR lies within the fact that CX3CL1 is a strong inducer of INF-γ by NK cells [55]. This calls for further research to simplify the understanding on the complex role of CX3CL1.

Primary microglia constitutively express NOD like receptor (NOD2) and the expression is upregulated following bacterial [87] and viral [143] invasion. The ligands for TLR also upregulate the expression of NOD2 in microglia. NOD2 activates nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB), which in turn lead to the transcription of inflammatory cytokines [87, 144] such as IL-1β, IL-6, TNF-α etc. [145].

CX3CL1-CX3CR1 signaling increases the expression of NF-kB, IL-1β and IL-6 by microglia [146]. This shows similarity between NOD2 and CX3CL1-CX3CR1 signaling, where both translate into the production of inflammatory cytokines. However, this conclusion is challenged by a quite opposite finding where CX3CL1-CX3CR1 signaling has suppressed the expression of IL-1β, IL-6 and TNF-α by microglia [38, 127].

Reactive microglia express MSR receptor in AD lesions [88]. IFN-γ and IL-1α induce the expression of MSR mRNA [147]. Expression of MSR-A receptor (also known as CD204, MSR-1, SCARA1) [148] plays a major role in the uptake of soluble and fibrillar amyloid beta-42 (Aβ-42) peptide [148, 149]. In spite of this effect, MSR-A strangely shows significant positive correlation with diffuse plaque, neuritic plaque, tangle formation in dementia population with AD pathology. MSR-A expression is related to poor cognition [142]. Activation of MSR-A with its ligand has increased the amplitude of LPS-induced inflammation through TLR4. TLR4 activates NF-kB and consequent expression of inflammatory cytokines [150]. As CX3CL1 can reduce the expression of TLR4 [116], hence, it is expected that CX3CL1 will suppress the aggravation of inflammation by MSR-A. This is in contrast with the finding which claims that CX3CL1-CX3CR1 signaling is responsible for fructose-induced neuroinflammation in hippocampal neurons, which have shown an increase in TLR4/NF-kB expression and signaling [151].

Microglia express CD14 but characterization of microglia based upon the expression seems to have disparity. This is because a recent publication identifies CD14 to be a general microglia marker, not the marker for the activated type [51]. However, previous publication claims that CD14 is expressed by M1 microglia [152]. CD14 has a prominent role in ischemia-induced neuronal damage. It is reported that ischemia leads to the production of NO by BV2 cells. NO activates CD14, which in turn induces TNF-α production. CD14/TLR4 receptor complex activates NF-κB pathway to produce proinflammatory cytokines [153]. This proves CD14 to be a proinflammatory marker for the M1 microglia. Next, question arises on the modulating effect of CX3CL1-CX3CR1 signaling on CD14. It is difficult to sketch the role of CX3CL1-CX3CR1 signaling, because one way it increases the expression of NF-kB, IL-1β and IL-6 by microglia [146] and by the other it suppresses the expression of IL-1β, IL-6 and TNF-α by microglia [38, 127].

Microglia express CD16 (Fcγ receptor). Characterization of microglia based on the expression of CD16 could not reach a consensus. Once CD16 was considered both a general and active state microglia marker [51], whereas another classification identifies it only as a proinflammatory marker for M1 microglia [154]. Irrespective of the classification, CD16 is known to induce inflammatory signal [51, 154]. This makes the understanding a little complicated because microglia with “triggering receptor expressed on myeloid cells 2” (TREM2) are characterized by the expression of CD16 receptor [155] and the same TREM2 induces M2 microglia and suppresses neuroinflammation [156]. It is reported that compared to TREM2− microglia, TREM2+ microglia provide better resolution of lesion after perforant pathway transection. TREM2 mediated clearance of myelin debris is accompanied by the secretion of anti-inflammatory cytokines. This is linked with proper axonal collateral sprouting and neuronal regeneration [155]. Deficiency of TREM2 also increases Aβ plaque [157]. Thus, CD16 seems to be both a proinflammatory and anti-inflammatory marker. Further, question arises on the modulating effect of CX3CL1-CX3CR1 signaling on CD16. Though there is no much evidence to prove the modulating effect of CX3CL1 signaling on CD16, yet a research on the effect of CNS insult shows higher expression of markers for both M1 (CD16) and M2 (CD206) microglia along with CX3CL1 and CX3CR1 [158]. It seems CX3CL1 has no direct effect on CD16 expression. However, in a different context it is reported that soluble CX3CL1 signaling can promote the proinflammatory effect of CD16 [159]. This again claims against the proven neuroprotective effect of CD16 expressing TREM2 microglia [155].

M1 microglia express CD32 (Fcγ receptor) [154, 160]. The role of CD32 as proinflammatory marker comes under question owing to the fact that TREM2 microglia also express CD32 receptor and have neuroprotective role [155]. TREM2 induce M2 microglia and suppress neuroinflammation [156]. Similarly, there is another classification where CD32 is considered as a marker for M2 microglia [152]. However, an increasing number of evidence suggests CD32 to be a M1 microglia marker with proinflammatory role [139, 160]. Though CD32 is recognized as a proinflammatory marker, yet it shares an inverse correlation with Aβ peptide load in the brain of AD patients [161]. This perhaps indicates Aβ peptide clearance role of CD32 microglia, because CD32 expressing microglia is predicted to have phagocytic activity in brain with excitotoxic injury [162]. Further, a question arises on the modulating effect of CX3CL1 signaling on CD32 microglia. So far there is no research available to prove the modulating effect. However, reduced Aβ deposition consequent to a decreased CX3CL1-CX3CR1 interaction [78] leads to the prediction that CX3CL1-CX3CR1 signaling shares an inverse relation with the expression of CD32 microglia.

Microglia constitutively express CD36 (class B scavenger receptor); however, pathological condition such as AD also shows their presence in the brain. CD36 increases the production of cytokines, chemokines and ROSby microglia when stimulated with fibrillar Aβ [163]. By contrast, CD36 expression on microglia leads to the engulfment of myelin debris and its clearance. This inhibits myelin debris-induced expression of TNF-α, IL-6 and nitric oxide synthase mRNAs in multiple sclerosis and can inhibit neuroinflammation [164]. Here, TREM2 deserves special mention because microglia express TREM2 receptor. TREM2 upregulates CCAAT/enhancer-binding protein alpha (C/EBPα), which in turn increases expression of CD36 and promotes phagocytosis of Aβ peptides and dead neural cells by microglia. TREM2 prevents loss of memory and learning in AD through C/EBPα [165]. However, opposing claim shows that CD36 is responsible for Aβ deposition in the microvessels of brain and its deficiency prevents neurovascular dysfunction irrespective of the elevated level of Aβ peptide in the brain [166]. The opposing claims pertinent to CD36 raises concern over the neuroprotective role of TREM2. Nevertheless, TREM2 overexpression prevents loss of neurons, synapse and hyperphosphorylation of tau protein. TREM2 prevents hyperphosphorylation by suppressing the activity of cyclin-dependent kinase 5 (CDK5) and glycogen synthase kinase 3β (GSK3β) [156]. It is reported that CX3CL1-CX3CR1 signaling promotes microRNA-124 (miR-124) delivery from neuron to microglia, which in turn targets C/EBPα to inhibit the activation of microglia. This leads to reduced neurodegeneration and improved neuronal function (short and long term) after subarachnoid haemorrhage [167]. However, absence of C/EBPα decreases the expression of CD11b on microglia [167]. CD11b deficient microglia activate TLR4 and produce more of IL-6, TNF-α and NF-kB compared to anti-inflammatory cytokines [168]. TLR4 also elevates the expression of CD86 [169] on microglia [170], and CD86 is a proinflammatory marker [171]. Association of CD86 with neuroinflammation can be observed in the context of ischemia/reperfusion injury. Ischemia/reperfusion injury deranges BBB and facilitates entry of activated T-cells into brain parenchyma. These T-cells secrete IFN-γ. IFN-γ induces the expression of MHCClass II and costimulatory molecule like CD80 and CD86 on CD11c⁺ microglia [172], and the role of CD11c⁺ microglia in neuroinflammation is already known [173]. Similarly, proofs regarding the proinflammatory role of TLR4 can be obtained from the finding where a decreased expression of TLR4 polarizes microglia to M2 type and inhibits neuroinflammation [170]. Thus, sketching the actual effect of CX3CL1-CX3CR1 signaling in this context is very complex. Contextually, the effect of CX3CL1-CX3CR1 signaling is further complicated by findings where C/EBPα prevents the expression of proinflammatory cytokines but increases ROSproduction (Fig. 1d) [167, 168].

Microglia express CD40 [152]. CD40 is a costimulatory molecule for TLR4. The role of CD40 in neuropathophysiology is a complex one. One way it suppresses ATP-TLR4 mediated NLR Family Pyrin Domain Containing 3 (NLRP3) inflammasome activation and secretion of IL-1β, and the other way CD40 stimulation not only strengthens LPS-induced upregulation of TLR4 but also enhances its own expression. Consequently, CD40 and TLR4 synergistically unleash microglial catastrophe through the production of proinflammatory cytokines such as IL-12, IL-6 and TNF-α, except for IL-1β [174]. It is noteworthy that LPS or IFN-γ induces cytosolic phospholipase A₂ alpha (cPLA₂α), which in turn induces CD40 but activation of cPLA₂α by IFN-γ was mediated by TNF-α [175]. Thus, once TNF-α is released following CD40 stimulation [174], it indirectly induces CD40 by facilitating IFN-γ mediated induction of cPLA₂α [175]. Thus, if CX3CL1 has to play as an anti-inflammatory agent, then it has to block the expression of TNF-α, other cytokines and cPLA₂α. It is observed that CX3CL1 shares an inverse correlation with the expression of TNF-α, IL-6, IL-1β by brain tissue, as proved by a research on ischemia-induced neuronal autophagy [176]. This is again quite opposite to the finding where CX3CL1-CX3CR1 signaling increased these markers in ischemic brain [7]. This perhaps demands for further research to understand the complexity of the role of CX3CL1. Further, it is reported that cPLA2α positive microglia are responsible for the production of significantly higher amount of NO and ROS as compared to cPLA2α negative microglia. LPS or IFN-γ depends on cPLA2α for the production of significant amount of NO and ROS by microglia to inflict neuroinflammation [177]. In this regard, cPLA2α is activated by phosphatidylinositol-4,5-bisphosphate (PIP₂) and ceramide-1-phosphate (C-1-P). PIP₂ activates cPLA2α independent of calcium but C-1-P causes calcium dependent activation of cPLA2α [178, 179]. Thus, to modulate the activity of cPLA2α, CX3CL1 has to target PIP₂ and C-1-P. It is noteworthy that CX3CL1 activates phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) [180]. PIK3CA in turn may convert PIP₂ to phosphatidylinositol 3,4,5-trisphosphate [181]. This conversion may affect the proinflammatory role of cPLA2α. Involvement of CX3CL1 in the expression of C-1-P may be predicted from the fact that CX3CL1 treatment makes primary cultures of murine microglia to release adenosine [182], most likely by degrading ATP [183]. ATP is a key player in the production of C-1-P [184]. CX3CL1-CX3CR1 signaling degrades ATP [182, 183], hence, energy released during the breakdown of ATP may enhance the production of C-1-P in the brain. Thus, further research is required to confirm the role of CX3CL1 in the production of C-1-P.

Microglia express CD64 (Fcγ receptor). CD64 is responsible for diffuse plaque in population with no dementia. CD64 expression is also significantly linked with diffuse and neuritic plaque in dementia with AD pathology [142]. It is reported that neuron derived immunoglobulin G (IgG) increases the expression of CD64 [185]. Though there is no direct evidence on the influence of CX3CL1 on CD64 expression, yet results from a different experimental setting have proved that absence of CX3CL1 or CX3CR1 led to impaired IgG production [186]. This perhaps indicates the significance of CX3CL1-CX3CR1 signaling on IgG mediated expression of CD64. Though this represents a negative perspective facilitating plaque formation and neuronal damage, yet a report proves the role of IgG-CX3CL1 or IgG-CX3CR1 infusion in alleviating neurodegeneration following status epilepticus [187]. Interestingly, increased plaque formation is associated with epilepsy [188] and decrease in CX3CL1-CX3CR1 interaction is associated with reduced plaque formation [78]. Reduced plaque formation will lead to a decrease in epilepsy and neurodegeneration. This represents two aspects. First aspect is the facilitating role of CX3CL1-CX3CR1 signaling on IgG expression [186] and consequent expression of CD64 [185] which leads to plaque formation [142], leading to epilepsy [188]. The second aspect is the post expression synergism between IgG and CX3CL1 or IgG and CX3CR1 to prevent neurodegeneration in status epilepticus [187]. This represents the complex role of CX3CL1-CX3CR1 signaling though IgG and CD64.

TREM2 microglia express CD68 receptor and TREM2 is linked with neuroprotective effect [155]. However, CD68 is considered a M1 microglia marker [160] and is expressed during AD [189]. CD68 is responsible for diffuse plaque in population with no dementia. It is significantly linked with neuritic plaque and tangle formation in dementia population with AD pathology. CD68 expression is related to poor cognition [142]. It is found that CX3CR1 shares a positive correlation with the expression of CD68, absence of CX3CR1 or CX3CL1-CX3CR1 signaling leads to a decreased expression of CD68 microglia [78].

Microglia express CD74 [152]. Expression of CD74 increases in neuroinflammatory condition like AD [190]. It is reported that therapeutic agent such as xamoterol alleviates manifestation of AD via PKA/cAMP pCREB signaling, which results in the decreased mRNA expression of many M1 microglial markers including CD74 [191]. CX3CL1 signaling may reduce the expression of CD74, because CX3CL1-CX3CR1 signaling activates CREB [48]. However, the effect of CX3CL1 on CD74 may be independent of cAMP activation, because CX3CL1 induces a significant decrease in cAMP signaling [192]. The correlation between CX3CL1 and CD74 can also be established through “migration inhibitory factor” (MIF), a chemokine produced by microglia and neurons in brain. MIF binds to CD74 and promotes expression of TNF-α, IL-6, IFN-γ, plaque formation and neuroinflammation in AD [193]. In a different context, it was found that IL-6 antagonist can decrease MIF levels [194]. As CX3CL1 suppresses IL-6 production [38], hence may suppress MIF mediated CD74 activation.

M2 Microglia express CD163 [152, 195]. CD163 microglia are predominant in the frontal and occipital cortices of AD, and associated with the phagocytosis of Aβ plaque [196]. Expression of CD163 indicates neuroprotection, however, it shares an inverse relation with the expression of CREB protein [98]. This makes the role of CX3CL1 somewhat complicated due to the fact that CX3CL1 activates CREB [48] as well as M2 microglia [26]. However, it is observed that a higher expression of CX3CR1 and moderate expression of CX3CL1 share an inverse relation with the expression of CD163 [197]. By contrast, CX3CL1 has been found to indirectly increase Cd163 gene expression [198]. Thus, further research is required to understand the complex role of CX3CL1 with respect to CD163.

Expression of ionized calcium-binding adapter molecule 1 (Iba1) by microglia is related with good cognition and absence of dementia [142]. In AD, Aβ-42 correlates inversely with Iba-1 [161]. However, the complex role of Iba1 microglia is evident from the fact that Iba1 positive microglia express NLRP3, which is responsible for caspase-1 dependent maturation of proinflammatory cytokines and cell death [199]. It is reported that decreased CX3CL1-CX3CR1 signaling leads to an increase in the expression of Iba1 positive microglia. Activation of Iba1 microglia is again responsible for postoperative cognitive dysfunction [77]. However, a research has contradicted the inverse relation between CX3CL1 and Iba1, where intrathecal injection of CX3CL1 led to an increase in Iba1-IR positive microglia in the spinal dorsal horn of C57BL/6J mice [200]. Thus, the actual role of CX3CL1 in the expression of Iba1 positive microglia remains elusive.

Microglia express lectins such as galectins, siglecs and C-type lectins. Among these lectins, galectin-3 deserves special mention because of its upregulation by inflammatory mediators [201]. Except for amyotrophic lateral sclerosis, galectin-3 upregulation causes neuroinflammation and neurodegeneration through NF-κB and NLRP3 [202]. Galectin-3 released by microglia is required for complete activation of microglial TLR4 and subsequent neuroinflammation [203]. Though there is no research to prove the effect of CX3CL1 on the expression of galectin-3, yet some different studies help to establish the link between. It was found that the concentrations of CX3CL1 and galectin-3 are much higher in systemic sclerosis patients and these proteins are thought to increase and exacerbate the fibrotic processes in systemic sclerosis [204]. This perhaps indicates that the expressions of CX3CL1 and galactin-3 are independent of each other or bear a positive correlation. This needs further research to reach into a conclusion. However, the need for establishing the relation between CX3CL1 and galactin-3 in systemic sclerosis can be justified by the finding that systemic sclerosis patients suffer from dysexecutive syndrome [205], which in most cases bear AD as the underlying aetiology [206].

CD206 and CD209 are the C-type lectins expressed by M2 microglia [195]. However, CD209 microglia are considered as intermediate microglia phenotype, because they are present in preactive lesions as well as in nonlesion white and gray matter of brain with multiple sclerosis [207]. CX3CL1 dependent expression of CD206 microglia have neuroprotective role. It was found that CX3CL1 facilitates the polarization of TNF-α microglia to CD206 microglia and can prevent ischemic stress associated disorders [208]. However, the established role of CX3CL1-CX3CR1 interaction to polarize M1 microglia to the anti-inflammatory M2 type [26] comes under question because M2a microglia markers CD206 and YM1 are expressed in the cortex of brain-injured CX3CR1^−/− mice [209]. In this context, the extent of cognitive dysfunction and neuronal death after the acute post-injury period were much greater in CX3CR1^−/− animals than the wild-type [209]. Contrary to the expression of CD206 in CX3CR1^−/− animals, the expression of CD209 is downregulated in TSC1^CX3CR1CKO microglia which are isolated from animals produced by cross breading TSC1^flox/flox mice with CX3CR1-cre mice [210], and CX3CR1-cre mouse does not produce endogenous CX3CR1 [211].

All the above evidences speak in favour of the differential effect of CX3CL1 on different mediators and markers of neuroinflammation. This makes the neuroprotective role of CX3CL1 a debatable matter.

Astroglia

Physiopathology of astroglia, their subtypes and the role of CX3CL1

Astrocytes provide structural support to CNS. They provide nutritional support to the neurons by storing glycogen and maintaining water and ionic balance. They form BBB and maintain its integrity, regulate permeability. They express ionotropic and metabotropic membrane receptors in the tripartite synapse and maintain homeostasis [119, 212,213,214]. A study with the BBB model in AD has revealed that expression level of CX3CL1 in the peripheral blood mononuclear cells (PBMCs) from AD BBB model was much lower compared to PBMCs from BBB of wild type model [215]. Contextually, AD is characterized by breakdown and early dysfunction of BBB [216]. This indicates the significance of CX3CL1 in maintaining the competence of BBB. However, another research with Alzheimer’s patients has made this understanding complicated owing to an increase and a decrease in the expression of CX3CL1 by PBMCs in BBB models of mild and moderate AD, respectively. Both expression values in mild and moderate AD were greater than the control group [217]. This indicates the need of further research on the role of CX3CL1 in BBB associated AD.

Astrocytes actively regulate cerebral blood flow. They release prostaglandin E2 (PGE2) and epoxyeicosatrienoic acids to produce vasodilation. They also release arachidonic acid which undergoes ω-hydroxylation to produce 20-hydroxyeicosatetraenoic acid in the blood vessel to produce vasoconstriction [218, 219]. However, results obtained from previous research are quite opposite, where PGE2 has produced vasoconstriction on the isolated parenchymal arterioles, which were preconstricted [220]. CX3CL1-CX3CR1 signaling has been found to increase the expression of PGE2 by satellite glial cells, indicating the complex role of CX3CL1 in cerebral blood flow [221].

Astrocytes are necessary for the proliferation of microglia under the influence of neuronal CX3CL1. Antibody to CX3CL1 or astroglial CX3CR1 abolishes the influence of CX3CL1 on microglial proliferation [222].

Astrocytes are the only cell types in the CNS which possess glutamine synthetase, and responsible for the glutamine content of the CNS [223]. Astrocytes convert glucose to glutamine. This glutamine eventually acts as the precursor for glutamate or gamma amino butyric acid (GABA) in the neuron [224]. Other than synthesizing glutamine from glucose, astrocytes can also synthesize it by taking up excess glutamate from the tripartite synapse and allowing astroglial glutamine synthetase to do the rest of the work [225, 226]. This will diminish the chances of glutamate excitotoxicity and neuroinflammation. The role of astrocytes in neuroprotection against glutamate excitotoxicity in rat brain is already proved [227]. In this regard, CX3CL1 protects against glutamate toxicity via adenosine receptors (mainly A₁R and in part A₃R). CX3CL1 induces release of D-serine (a co-agonist of NR2A type of synaptic NMDA receptor) by glial cells, CX3CL1 and D-serine induce the expression of adenosine receptor (A_2AR), which is required for phosphorylation of CREB, a prerequisite for neuroprotection against NMDA toxicity [228]. Initial astroglial atrophy in the blood vessels and synapses is prominent in AD. This attributes to the dysfunction of the neuro-vascular unit and glutamate toxicity in tripartite synapses [214, 229]. Prolonged glutamate toxicity is injurious to astrocytes and causes death of astrocytes [230]. This will exacerbate neurodegeneration because diseased astrocytes lead to secondary reactive astrogliosis [231]. Even previous reports show that astrocytes undergo transition into reactive astrocytes following injury-induced release of inflammatory mediators by M1 microglia [80, 232], oligodendrocytes, neurons, endothelial cells, leucocytes and other astrocytes [232]. However, the extent of neuronal damage due to reactive astrogliosis is difficult to predict. This is owing to the fact that one way reactive astrocytes expressing A_2AR and purinergic (P2Y1) receptors induce complement component C3, NF-kB and IL-1β, which are responsible for neurodegeneration [233, 234]. Other way, CX3CL1 through A_2AR provides neuroprotection against NMDA toxicity [228]. Astrocytes are recognized as intermediary/secondary targets of CX3CL1 activity, because absence of astrocytes completely disrupts the neuroprotective role of CX3CL1-stimulated microglia against excitotoxicity [235]. Astrocytes maintain the neurotransmitter pool around the synapse. Astrocyte processes have abundance of transporters which uptake GABA, glutamate [236], and glycine [237]. Thus, astrocytes are actively involved in the regulation of neuronal excitability. However, the actual process of regulation of neurotransmitter pool is more complex, because sequestration of glutamate and GABA by astrocytes is dependent on extracellular glycine release, which is mediated by Sodium-Neutral Amino Acid Transporter (SNAT3) [238]. The complexity of this process is realized from the report that extracellular glutamate triggers astroglial glutamate release to the adjacent neurons, by inducing Ca²⁺ release from the intracellular stores of astrocytes [225]. This puts a question on the role of CX3CL1 because CX3CL1-CX3CR1 interaction in astrocytes mobilizes astrocytic intracellular Ca²⁺ [222] and increased concentration of astrocytic Ca²⁺ leads to the release of gliotransmitters such as glutamate, ATP and D-serine [239]. This makes the role of CX3CL1 ambiguous in the context of excitotoxicity. This suggests for the presence of a delicate regulation between glutamate and glycine to maintain the glutamate gradient inside and outside astrocytes.

The significance of astrocytes can be realized from their ability to release different growth factors such as BDNF, glial cell line derived neurotrophic factor (GDNF), nerve growth factor (NGF), platelet derived growth factor (PDGF), and certain other substance such as heat shock proteins (HSP) [212, 240, 241]. The significance of CX3CL1 in expression of BDNF is already proved [77]. Similarly, CX3CL1 induces the expression of A_2AR [228] and A_2AR is required by GDNF to evoke the release of dopamine in the striatum and provide neuroprotection [242]. CX3CL1 is also observed to mediate the effect of NGF [243]. Research also shows positive correlation between expression of CX3CL1 and HSP-72 [244]. Another HSP, called heme oxygenase-1 is secreted by astrocytes and this HSP is associated with neuroprotection in intracerebral hemorrhage [245]. Soluble CX3CL1 has been found to induce the expression of heme oxygenase-1 [246]. All these evidences show enormous amount of crosstalk by CX3CL1 in favour of neuroprotection.

Astrocytes also synthesize L-lactate and cholesterol. Extracellular release of L-lactate provides energy to the neurons and the cholesterol forms the essential component of myelin sheath of neurons during synaptogenesis [224, 247, 248]. L-lactate can offer neuroprotection via mild oxidative burst leading to unfolded protein response and activation of NRF2. L-lactate does not induce severe oxidative stress, rather offers resistance against it [249]. It is found that CX3CL1 can activate NRF2 and provide neuroprotection [37].

In contrast to the claims about the negative role of astrogliosis, a study shows that astrogliosis shields the neurons from inflammatory lesion. During astrogliosis, astrocytes undergo hypertrophy and wrap the healthy neurons to protect from a brain lesion [250]. The role of astrocytes to counteract neurodegeneration is believed due to their ability to promote the clearance of interstitial Aβ peptide and encapsulate dendritic spines, dendritic shafts, axonal boutons. Astrocytes also modulate the stabilization and maturation of dendritic spines [214, 229]. In spite of the beneficial roles, astrogliosis is linked with neuroinflammation [4, 5]. This is due to the fact that cytokines released during pathological process, work synergistically to elevate the expression of β-secretase, amyloid precursor protein (APP) in astroglia and consequent massive production and secretion of Aβ peptide into the surrounding. This is a hallmark event in the progression of AD [251]. However, the effect of this secreted Aβ peptide on neurons may be neutralized by CX3CL1 C-terminal fragment, generated following the cleavage of membrane anchored CX3CL1 by metalloproteases and β-secretase (BACE1). This can significantly reduce Aβ deposition and neuronal loss [252]. Reduced Aβ deposition will also suppress α-synuclein toxicity. This is owing to the fact that Aβ may disrupt protein clearance, enhance phosphorylation and promote aggregation of α-synuclein [253]. Further cleavage of membrane anchored CX3CL1 C-terminal fragment by γ-secretase releases the intracellular domain of CX3CL1, which communicates with the nucleus and promotes neurogenesis [252]. Thus, CX3CL1 offers wide coverage against astrogliosis-related neurotoxicity. However, CX3CL1 may not protect against long-term assault because persistent reactive astrogliosis is maladaptive and the reason behind loss of neuronal plasticity and other regenerative process [254]. To decipher the bipolar role of astrocytes, knowledge of its location specific/functional classification is much desired. Based on location, human cortical astrocytes are classified as inter-laminar astrocytes (present in superficial cortical layers), polarized astrocytes or varicose projection astrocytes (present in the deep cortical layers) [224, 255]. Another location specific classification divides astrocytes as protoplasmic astrocytes (present in the gray matter) and fibrous astrocytes (present in the white matter) [240, 256]. In this regard, protoplasmic astrocytes deserve special mention, because they favour accumulation of α-synuclein during Parkinson’s disease, AD and epilepsy [240]. Such accumulation of α-synuclein is toxic to astrocytes [257] and can lead to the transformation of astrocytes into reactive astrocytes [258]. Reactive astrocytes with their larger size and increased expression of glial fibrillary acidic protein (GFAP) [259], induce neurotoxicity by producing ROS or certain pro-inflammatory cytokines [232]. The role of CX3CL1 in suppressing α-synuclein mediated neurodegeneration is already known [64]. This will suppress subsequent production of ROS and cytokines by reactive astrocytes and provide neuroprotection. Further, it is desirable to observe the effect of CX3CL1 on subtypes of reactive astrocytes which are classified based on upregulating stimuli. Based on upregulating stimuli and functional differences reactive astrocytes are classified as A1 astrocytes and A2 astrocytes. A1 are upregulated by inflammation and responsible for pathological outcomes (neurodegeneration), whereas A2 are upregulated by ischemia and have a neuroprotective role [256, 260]. Owing to the neurodegenerative role of A1 astrocytes, its suppression by CX3CL1 will alleviate diseases affecting cognitive functions. Hence, the impact of CX3CL1 on the upregulating stimulus of A1 astrocytes is a crucial predictor of the effect of CX3CL1 on neuroinflammation. Critical observation suggests that A1 astrocytes can also be upregulated by ischemia, because ischemia shows aggregates of RNA binding proteins such as TAR-DNA binding protein-43 (TDP43) and many others [261]. Among these proteins, expression of high levels of intracellular TDP43 is known to activate M1 microglia [30], which in turn activate A1 astrocytes by secreting TNF-α, IL-1α, and complement component 1, q subcomponent C1q [80, 259, 262]. Thus, accumulated RNA binding protein will involve A1 astrocytes, to negatively affect the survival of neurons and initiate neurodegeneration in amyotrophic lateral sclerosis and frontotemporal dementia [261]. Similarly, when astrocytes are activated with IL-1β and IFN-γ, they also liberate TNF-α [81]. This shows that astrocytes in turn can activate microglia, because M1 microglia are activated by TNF-α [263]. Thus, activation of astroglia and microglia seems to share a converse pathway which is TNF-α dependent. Previous report has claimed that astrocytes can express CX3CL1 under normal and pathological condition, and astroglial CX3CL1 signaling induce microglial chemotaxis [11]. Even subsequent research has confirmed that IL-1β, TNF-α and IFN-γ can stimulate the secretion of soluble CX3CL1 from activated astrocytes [10]. The ability of astroglia to activate microglia is corroborated by another finding where reactive astrocytes secrete significant amount of Aβ peptide in AD [232] and microglia start clustering around Aβ plaques [89]. These compelling evidences suggest that astrocytes have a strong influence on microglial recruitment. Astroglial CX3CL1 will impede the disruptive role of M1 microglia and prevent subsequent ischemia-induced neuronal damage because CX3CL1 is already known to be neuroprotective against ischemia-induced neuronal damage [31]. By contrast, CX3CR1 has been found to be a contributor to cerebral ischemia and postischemic inflammation [264]. This raises concern on the protective role of CX3CL1-CX3CR1 signaling via modulation of upregulating stimulus for A1 astrocytes.

In continuation to the aspects of microglial recruitment it is found that microglia may be recruited by many cell types including microglia, but through astrocytes. This is corroborated by the fact that microglia release TNF-α and IL-1β [28]. Both TNF-α and IL-1β can induce astroglial CX3CL1 [8, 11, 183] and CX3CL1 is responsible for microglial chemotaxis [11]. Other macrophage cell types which can release TNF-α include tanycytes, ependymocytes, and cerebrospinal fluid-contacting neurons [265]. As CX3CL1 reduces the expression of TNF-α and IL-1β [38], this will suppress chronic microglial recruitment by other cell types also. This represents a feedback loop, where astrocytes regulate microglial chemotaxis and neuroinflammation by other cell types also. However, the role of CX3CL1 in increasing TNF-α and other cytokines [7] imposes a big contradiction in this setting.

Further, it is noted that bradykinin-induced increase in astrocytic intracellular calcium leads to astrocytic ATP release in the surrounding. This ATP activates P2X7 receptors on microglia, leading to an increased microglial permeability and apoptotic death of microglia [266]. This perhaps indicates that activation of M1 microglia subsequently activates the degradative role of astrocytes on M1 subtype. This is due to fact that M1 microglia aggravate neurotoxic agent mediated demyelination of neurons by producing TNF-α and IFN-γ [267]. It is noteworthy that IFN-γ stimulates ATP mediated microglial death by astrocytes [266]. This represents a strange mechanism of microglia-induced microglial death through astrocytes. CX3CR1 deficient mononuclear phagocytes show upregulation of P2X7, consequent release of IL-1β and neurodegeneration [268]. This indicates the suppressive role of CX3CR1 in P2X7 expression. Next, it is desirable to observe the effect of CX3CL1 on TNF-α and IFN-γ because IFN-γ also upregulates the expression of HLA-DR in astrocytes. TNF-α does not have any direct role in HLA-DR upregulation but enhances the effect of IFN-γ on HLA-DR expression [269]. This imparts astrocytes with the function of antigen presenting cells [270]. Soluble CX3CL1 shows negligible effect but membrane bound CX3CL1 shows significant increase in the expression of IFN-γ by NK cells [55]. This may augment inflammatory process. However, suppression of TNF-α expression by CX3CL1 [38] may partially subdue the expression of HLA-DR.

The impact of astrocytic CX3CL1 on microglial CX3CR1 can be observed during spinal nerve injury, where an increased expression of astrocytic CX3CL1 leads to the upregulation of microglial CX3CR1 [271]. CX3CL1 and CX3CR1 are considered possible mediators of neuropathic pain [271]. Subsequent research also confirmed the role of activated microglia in the maintenance of chronic pain post spinal cord injury [272]. However, the role of astroglial CX3CL1 needs further exploration because CX3CL1 is also constitutively expressed in astrocytes, and reported to have physiological role other than pathological involvement. Even, CX3CR1 is also expressed on astrocytes [11]. Thus, the impacts of neuronal and astroglial CX3CL1 on CX3CR1 of microglia and astroglia need to be explored. Similarly, the role of neuronal CX3CR1 needs further exploration. Neuronal CX3CL1 has been proved to increase the neuroprotective role of astroglia in excitotoxicity and ischemia. However, this role is microglia dependent because CX3CL1-microglia interaction is the prerequisite for the expression of astroglial CCL2 in a co-culture of astroglia and microglia, which eventually provides neuroprotection [46]. When it comes to the role of astroglia in suppressing NF-kB mediated neuroinflammation, it is observed that CX3CL1-CX3CR1 signaling did not contribute to the process. Absence of CX3CR1 on astroglia did not show any increase in the expression of NF-kB as compared to CX3CR1 positive astroglia [146]. This reflects that NF-kB mediated release of inflammatory mediators [87] will not be checked by CX3CL1-CX3CR1 interaction on astroglia. However, astroglia may engage neuronal CX3CL1 to produce their effect on the mediators of inflammation. This is due to a prediction that astroglial CX3CL1 may promote expression of membrane bound CX3CL1 on neuron and a subsequent release of soluble CX3CL1 [10]. In this regard, the neurotoxic role of neuronal CX3CR1 in the absence of microglia is already discussed above [59].

Astrocytes liberate interleukin-1 (IL-1β), IL-6, chemokine C-X-C motif ligand-1 (CXCL1), IL-8 (CXCL8), INF-γ-induced protein 10 (IP-10)/CXCL10, monocyte chemoattractant protein 1 (MCP-1)/chemokine C–C motif ligand 2 (CCL2), macrophage migration inhibitory factor (MIF), macrophage inflammatory protein 1 alpha (MIP-1α)/CCL3, granulocyte colony stimulating factor (G-CSF) and granulocyte–macrophage colony stimulating factors (GM-CSF) in late stage AD. These massive cytokine and chemokine storms are reported to favour leukocyte infiltration into the brain and mount a chronic inflammatory process [214, 273,274,275]. However, the relation among astrogliosis, liberation of cytokines and chemokines, and chronic neuroinflammation in late stage AD is ambiguous. This is due to the report that, reactive astrogliosis is observed in late stage AD [276]. Such reactive astrogliosis leads to glial scar formation [277]. Glial scar prevents neurodegeneration, reduces demyelination, isolates damaged tissues and repairs injured nervous tissues, repairs BBB, impedes infiltration of inflammatory cells, and reduces loss of oligodendrocytes in white matter adjacent to the injury site [232, 259, 278]. Astroglial scar also produces chronic inactive lesion [11]. In fact, CX3CL1 induces the expression of CXCL16 in astrocytes [46]. Previous research has already proved that CXCL16 can be produced by reactive astrocytes. CXCL16 undergoes shedding to upregulate the expression of its receptor CXCR6 in glial precursor cells. CXCR6-positive glial precursor cells will be attracted by CXCL16. This will lead to their invasion and proliferation. This will favour subsequent astrogliosis and glial scar formation [279]. Irrespective of the effects of astrogliosis, AD represents a progressive degenerative brain disease [280]. This raises questions on the significance of CX3CL1-induced glial scar formation and its implication on neuroinflammation.

Effect of astroglia ablation in neuropathophysiology and the impact of CX3CL1

Ablation study has proved that microglia mediate tauopathy, however, it is reported that astrocytes can express lower level of tau proteins [281]. Astrocytic apoE4 has been proved as the key role player in tau-induced synaptic loss and phagocytosis of synaptic elements by microglia [282]. Thus, ablation of astrocytes seems to be beneficial in the context of tauopathy. The overall effect of astroglial ablation should be considered with respect to neuronal function. It is reported that ablation of astrocytes leads to motor discoordination and increases chances of excitotoxicity due to a reduced expression of glutamate transporters [283]. Similarly, reactive astrocytes in epileptic tissue can promote as well as oppose seizure development [284]. In this context, astrocytes play a dichotomous role. The significance of astrocytes in mediating the neuroprotective role of neuronal CX3CL1 and suppressing excitotoxicity through microglia is already discussed [235]. This represents the significance of a nexus of neuronal CX3CL1, astroglia and microglia in alleviating excitotoxicity. A recent study has put a question on the age old belief on the destructive role of reactive astrocytes in neuroinflammation. The study has revealed that amyloid pathology and memory loss in AD is aggravated after the loss of reactive astrocytes [285]. Similarly, another study has proved that astroglia play a neuroprotective role and helps in memory retention. Further, this study also corroborates that Aβ removal is dependent on astrocytes and pharmacological ablation of astrocytes leads to an increased expression of pro-inflammatory markers as well as loss of dendritic spine [286]. However, other study has proved that cholesterol derived from astrocytes is required by apolipoprotein E (apoE) to transfer neuronal APP inside lipid clusters, where APP is acted upon by β-secretase and γ-secretase to produce Aβ peptide. Both apoE3 and apoE4 isomers have similar actions on APP. Cholesterol also regulates the formation of lipid clusters. Inhibition of astrocyte cholesterol synthesis significantly reduces amyloid and tau burden. It is also proved that knocking down cholesterol synthesis or treatment with cholesterol-free apoE causes APP to come out of lipid clusters, which is then processed by α-secretase to produce neuroprotective metabolite [287]. Studies with multiple preclinical models reported that overexpression of astrocytic apoE4 increases hyperphosphorylation and misfolding of neuronal tau, which eventually leads to tauopathy [288]. It is reported that apoE4 astrocytes show dysfunctional astrocytic mitochondria with reduced fission and mitophagy. Dysfunctional astrocytic mitochondria are considered to play a role in AD [289]. Dysfunctional astrocytic mitochondria loose calcium handling ability, which in turn contributes to the progression of Parkinson’s disease [290]. All the above evidences show the beneficial role of astrocytes as well as the pathological significance of astrocytic cholesterol and apoE4. This revelation calls for the investigation of the influence of CX3CL1 on the negative role of astrocytes. The correlation between CX3CL1/CX3CR1 and cholesterol synthesis can be proved from previous studies where blockade of 3-hydroxy-3-methylglutaryl (HMG)-coenzyme A (CoA) reductase decreases the expression of CX3CL1 and CX3CR1. It is to be noted that CX3CL1-CX3CR1 interaction is found to play a role in atherogenesis [291,292,293]. The involvement of CX3CL1 in low density lipoprotein (LDL) cholesterol mediated atherosclerosis is proved by the fluorescence intensity study of CX3CL1 targeted nanofiber [294]. Thus, the role of CX3CL1 is evident in cholesterol synthesis. However, the role of CX3CL1 with respect to astroglial apoE4 has to be established. Surprisingly, no evidence could be retrieved regarding the role of CX3CL1 on astroglial apoE4. This opens newer avenue to research and establish the impact of CX3CL1 on astroglial apoE4. Nevertheless, the impact of CX3CL1/CX3CR1 on cholesterol synthesis seems to be a matter of concern in the event of astroglial cholesterol synthesis and subsequent neurodegeneration.

Role of other astroglial markers in neuroinflammation and their modulation by CX3CL1

Astrocytes in experimental autoimmune encephalomyelitis, produce macrophage inflammatory protein-3α (MIP-3α), a CC Chemokine, also known as CCL20. IL-1β and TNF-α can induce the expression and secretion of CCL20 by astrocytes. CCL20 is a chemoattractant for polarized T helper (Th) cells (both Th1 and Th2) [295]. High levels of Th cells (Th1 and others) along with chemokine and cytokines can damage BBB and activate resident astrocytes and microglia to produce neuroinflammation [296]. Thus, astrocytes attract Th cells to activate more and more astrocytes and microglia to aggravate neurodegenerative condition like autoimmune encephalomyelitis [295]. The release of Th1 and others are also associated with multiple sclerosis [296]. Similarly, Th1 and Th2 are responsible for the progression of AD [297]. This necessitates the establishment of a correlation between CX3CL1 and CCL20 as well as between CX3CL1 and Th cells. It was found that astrocytes show upregulation of CCL20 and CX3CL1 along with many other mediators under the influence of neuronal α-synuclein [298]. This perhaps indicates a positive correlation between CCL20 and CX3CL1. Under a different setting it is confirmed, that CX3CL1 indeed shares a positive correlation with CCL20 [299]. This again raises question on the neuroprotective role of CX3CL1.

Astrocytes constitutively express TLR4. LPS stimulated TLR4 requires soluble CD14 for the production of TNF-α, IL-6 and activation of NF-kβ [300]. This can activate M1 microglia via TNF-α [263]. TNF-α and other cytokines can stimulate the secretion of soluble CX3CL1 from activated astrocytes [10]. CX3CL1-CX3CR1 signaling, has been found to increase the expression of IL-1β, IL-6 and NF-kB by microglia [146] as well as decrease the expression of IL-1β, IL-6 and TNF-α by microglia [38, 127]. This dichotomous role of CX3CL1 on the release of cytokines makes its impact on astrocytic TLR4 and CD14 somewhat illusive. However, CX3CL1 has a protective role on CX3CR1-WT/WT CD14⁺ monocytes from serum starvation-induced death [301].

Astrocytes express CD36 and CD47. Both CD36 and CD47 engulf Aβ peptide in actin polymerization dependent manner [302]. CX3CL1-CX3CR1 interaction is already known for initiating actin polymerization [129]. This shows the role of CX3CL1 in engulfing Aβ through CD36 and CD47. CD36 is also required for astrogliosis and glial scar formation [303] and the role of CX3CL1 in glial scar formation by astrocytes is already proved [46, 279].

Astrocytes express aldehyde dehydrogenase family 1 member L1 (Aldh1L1). Aldh1L1 has many functions such as conversion of folate, nucleotide biosynthesis and cell division. Reactive astrocytes show upregulation of Aldh1L1 and GFAP in acute neural injury and chronic neurodegenerative conditions [304]. Expression of GFAP is the indication for astrocyte reactivity because astrocytes with ALDH1L1⁺GFAP⁻ are identified as resting astrocytes and ALDH1L1⁺GFAP⁺ as reactive astrocytes [305]. Aldh1L1 levels increase in response to focal demyelination injury [306]. However, the expression level of Aldh1L1 by astrocytes in response to saline or 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) treatment did not show a significant difference. This is similar in the case of groups treated with saline or trimethyltin (TMT) type of neurotoxins [307]. Thus, GFAP bears more significance compared to Aldh1L1 as a marker of astrocyte reactivity. An increase in neuronal CX3CL1 does not have any significant effect on promoting astrogenesis or astrocytosis, as confirmed by the expression level of GFAP [308].

Astrocytes express glycolytic enzyme aldolase C (AldoC), which can be found in astrocyte derived small extracellular vesicles (sEVs). Astrocyte can transfer AldoC containing sEVs to hippocampal neurons. Astrocytic sEVs also carry micro RNAs, which are present in astrocytes. One specific micro RNA is miR-26a-5p. It is observed that miR-26a-5p containing sEVs decrease dendritic complexity in neurons, and the content of AldoC influences the role of sEVs [309]. Decreased dendritic complexity may lead to neurodegeneration [310] and pathologies like AD. The relation with AD can be explained based on the functions of different spine types. Thin spines are apparently required for learning and new memory formation, whereas mushroom spines may represent sites of memory storage and stubby spines may give rise to thin or mushroom spines. AD reports a reduction in the density of mushroom spine. Thus, loss of dendritic complexity is a crucial determinant of AD [311]. CX3CL1-CX3CR1 signaling has been found to prevent dendritic loss and death in striatal neurons cultured with mixed glial population [312]. Though protective effect on dendrites seems to be the exclusive role of microglia rather than astrocytes [312], yet the involvement of astrocytes in microglia mediated neuroprotection cannot be denied [235]. Conversely, a report proves that CX3CR1 knockout microglia support increased dendritic spine density in neurons [12]. This depicts the complexity of CX3CL1 functioning.

Astrocytes secrete S100 calcium-binding protein B (S100B), a soluble protein which can induce neurite outgrowth [313]. Neurite elongation and branching are essential to ensure proper wiring of neuronal network, which is again required for brain development [314]. Within physiological range S100B has a trophic role on neurons, but a higher concentration of S100B is proinflammatory [313, 315]. Higher level of S100B is found in aged brain and in many neurodegenerative conditions including AD. Microglia express receptor for advanced glycation end products (RAGE) and S100B can engage microglial RAGE to facilitate microglial activation and movement. Such interaction underlies the pathogenesis of neuroinflammation [315]. S100B can also induce RAGE in cortical neurons [313]. Neuronal RAGE is a receptor for Aβ peptide [316] and amphoterin [317] and plays a dichotomous role. One way, amphoterin binds with RAGE to promote neurite outgrowth and the other way Aβ peptide-RAGE interaction results in oxidative stress and consequent neurodegeneration. Strangely, the same Aβ peptide may facilitate neurite outgrowth [317]. RAGE is also responsible for advanced glycation end products (AGE) mediated hyperphosphorylation of tau proteins and memory deficits, leading to AD like pathologies [318]. Suppressing excessive induction of S100B may suppress RAGE and provide neuroprotection in AD like pathologies [315]. This is concluded owing to the role of Aβ peptide-RAGE interaction in oxidative stress [317]. CX3CL1 shows inverse relation with the expression of S100B in ischemic stroke patients and a decreased S100B level indicates better recovery [319]. However, predicting the actual effect of CX3CL1 on S100B is not easy. This is owing to the fact that IL-1β induces secretion of S100B by astrocytes [39] and CX3CL1-CX3CR1 signaling may increase the expression of IL-1β by microglia [146] as well as decrease the expression of IL-1β by microglia [38, 127].

Astrocytes express RNA‐binding protein Fragile X Mental Retardation Protein (FMRP). FMRP loss from astrocytes reduces glutamate transporter GLT‐1 (EAAT2) expression and glutamate uptake by astrocytes. This may cause glutamate toxicity. Strangely, knocking out Fmr1 gene may lead to increased expression of tumor necrosis factor receptor 2 (TNFR2) (an anti-inflammatory receptor for TNF-α) and S100B by astrocytes. TNFR2 promotes myelination [320]. However, the overall benefits of increased expression of TNFR2 may be questionable due to the role of enhanced TNFR2 signaling in promoting atherosclerosis [321]. Atherosclerosis is again responsible for neuroinflammation and brain atrophy [322]. CX3CL1 increases mRNA expression and activity of GLT-1 on astrocytes and protects neurons against excitotoxicity [235]. At the same time impeding CX3CL1-CX3CR1 signaling alleviates severity of atherosclerosis [323]. This will reduce neuroinflammation and brain atrophy [322].

Astrocytes express aquaporin-4 (AQP4) protein, responsible for brain water and volume homeostasis. However, its involvement in activating microglia and inducing neuroinflammation in neurotoxin-induced Parkinson’s disease model has also been revealed [324]. Contrary to this pathological involvement, astrocytic AQP4 is involved in the clearance of Aβ from brain via lymphatic clearance. Hence, it seems to be a target to alleviate AD [325]. However, a significant increase in AQP4 expression is related to edema in AD brain and cerebral amyloid angiopathy [326]. AQP4 expression is dependent on the activation of p38 mitogen-activated protein kinase (p38-MAPK) pathway [327]. Suppression of p38-MAPK pathway leads to a decrease in the production of astrocytic soluble CX3CL1, and astrocytic CX3CL1 is neuroprotective. This reveals the beneficial effect of p38-MAPK pathway [10]. At the same time this pathway also plays a negative role in the context of ischemia which is linked to CX3CL1. The alleviating role of CX3CL1 against ischemia-induced neuronal damage is well documented [31]. Strangely, CX3CL1-CX3CR1 interaction on microglia had a detrimental effect on ischemic brain injury via p38-MAPK/PKC signaling [7]. Thus, CX3CL1 and p38-MAPK share a complicated relation in neuroinflammation.

Immature astrocytes express thrombospondins-1 and thrombospondins-2. Deficiency of both these proteins reduces synaptogenesis [328]. Though, thrombospondins-1 encoding gene THBS1 is associated with the risk for autism [329], yet thrombospondins-1 plays a neuroprotective role. Thrombospondins-1 impedes Aβ mediated synaptic pathology in AD [330]. CX3CL1-CX3CR1 interaction enhances expression of thrombospondins-1 [331] and will alleviate AD. However, thrombospondins are known to induce synapses which are presynaptically active (synapse with AMPA-NMDA receptor) but post synaptically silent (synapse with only NMDA receptor) [328]. Silent synapses are linked with brain trauma, addiction and neurodegenerative disorders [332]. This raises question on the overall neuronal effect of CX3CL1 through thrombospondins mediated induction of silent synapse (Fig. 1e).

Some of the contrasting issues on the neuroprotective and neurodegenerative roles of CX3CL1 are depicted (Fig. 1).

Discussion

The networking among neurons, microglia and astrocytes is enormous and the overall effect of the networking on neuropathophysiology is still elusive in the context of CX3CL1. One way, absence of astrocytes abolishes the neuroprotective role of CX3CL1-stimulated microglia against excitotoxicity [235] and the other way, neuronal CX3CL1 enhances neuroprotective role of astroglia in excitotoxicity and ischemia. However, this is dependent on CX3CL1-microglia interaction [46]. This indicates that both astrocytes and microglia are required to support each other to retain their neuroprotective role and augment the function of neurons. Irrespective of this understanding the therapeutic effect of exogenous CX3CL1 cannot be predicted accurately. This is due to two opposing claims with respect to a single disease. The first one shows that soluble CX3CL1 is effective in alleviating Parkinsonism [84], whereas the second one shows that injection of exogenous CX3CL1 in unilateral substantia nigra leads to microglial activation, depletion of dopaminergic neuron and motor dysfunction leading to Parkinsonism. Expectedly, these effects are abolished by blocking CX3CR1 [333].

There are gaps in knowledge with respect to the direct effect of CX3CL1 on the expression of NOD2, CD14, CD16 and CD64 on microglia. This indicates the need for further research to envisage the overall scope of targeting CX3CL1 to treat neuroinflammation. CD14 deserves special mention because of the dichotomous effects of CX3CL1 on the inflammatory mediators which are expressed as downstream events of CD14 activation [38, 127, 146].

Astrocytes synthesize glutamine and release in the tripartite synapse. This glutamine is taken up by neurons for synthesizing GABA or glutamate. Release of synthesized glutamate by the neurons triggers further glutamate release from the astrocyte [224, 225]. This represents a perspective that neuronally released glutamate controls the release of astroglial glutamate and any dysregulation will induce glutamate toxicity. This diminishes the role of astrocytes as the primary regulator of synaptic glutamate concentration and excitotoxicity to neurons. This is corroborated by the revelation that CX3CL1-induced release of glial D-serine along with CX3CL1 protects against NMDA toxicity [228]. However, contradiction to the neuroprotective role of CX3CL1 can be found where CX3CL1 cannot protect cortical neurons against glutamate-induced death unless neuronal culture contains 90% astrocytes and 10% microglia [235]. Thus, mere administration of exogenous CX3CL1 may not provide neuroprotection unless brain has an optimum population ratio of glial cells.

Protoplasmic astrocytes favour accumulation of α-synuclein during Parkinson’s disease, AD and epilepsy [240]. α-synuclein, which is predominantly expressed in neurons, ultimately finds its way inside astrocytes through tunnelling nanotubes. It is hypothesized that astrocytic lysosomes cause truncation of α-synuclein, leading to astrocytic dysfunction [334]. The C-terminal-truncation of α-synuclein exacerbates the aggregation and cytotoxicity of α-synuclein in Parkinson’s disease [335]. CX3CL1 impedes α-synuclein mediated neurodegeneration [64]. As CX3CL1 and α-synuclein originate from neurons, hence, it may be concluded that neurons are quite capable of preventing self-injury resulting from biochemical messenger originating within them. However, this conclusion is not applicable in every aspect, as can be seen in case of the differential effects of CX3CR1 on Aβ peptide and tau protein [83] and strange findings where acutely applied Aβ oligomers can increase dendritic complexity and spine density. This shows that Aβ may alleviate AD, because AD displays a reduction in the spine density [311]. Further, the benefits of CX3CL1 mediated reduced expression of IL-1β [38] and consequent alleviation of tauopathy-induced cognition deficits [39] come under question. This is due to the facilitating role of IL-1β in alpha-cleavage of APP and a consequent decrease in Aβ production [336] and the generation of neuroprotective metabolites [337].

The claims regarding the effects of CX3CL1 are in continuous contradiction with respect to TLR4 [116, 151], IL-1β, IL-6, TNF-α expression [38, 127, 146] and neuronal effect. This suggests that some effects of CX3CL1 are not completely understood in the context of neuroinflammation.

Further, the role of CX3CL1-CX3CR1 signaling to prevent dendritic loss [312] is controversial with respect to a report where CX3CR1 knockout microglia can increase dendritic spine density in neurons [12]. As dendritic spines grow from dendrites [338], hence, it may be predicted that CX3CR1 knockout increases dendritic spine on the leftover dendrites. This again raises question on the probability of occurrence of autism, this is owing to the positive correlation between increased dendritic spine density and autism [339]. At the same time, reduced dendritic spine formation due to over expression of CX3CR1 [12] and the probability of consequently occurring schizophrenia [340] need further consideration. This is owing to the fact that CX3CL1 may increase the expression of CX3CR1 [36]. Thus, there is no direct answer with respect to certain possibilities arising after an increased or decreased CX3CL1-CX3CR1 interaction. This calls for an overall assessment of the possible neuronal outcome before administering CX3CL1 in order to alleviate neurological deficits associated with any neurological disorder.

Conclusion

CX3CL1 shows immense networking with receptors/proteins/markers expressed by the glial cells and has a modulating effect on the physiopathology of glial cells. CX3CL1-CX3CR1 interaction seems to be beneficial in neurological deficits. However, certain contradictory observations and missing links suggest for an overall assessment of the implications of CX3CL1-CX3CR1 interaction in neurological settings.