Microglial Phagocytosis in Epilepsy: Mechanisms and Impact

Abstract

Microglia are resident immune cells critical in maintaining brain homeostasis via their surveillance and phagocytosis function. Under disease contexts like seizures and epilepsy, microglial phagocytic signaling is activated in response to both inflammatory and non-inflammatory cell death. This process involves a range of well-characterized ‘find me’ and ‘eat me’ signals, phagocytic receptors, and less well-characterized intracellular signaling pathways. In addition, epigenetic and transcriptional regulators orchestrate microglial responses to seizures, including the integration of phagocytic and inflammatory pathways. Interestingly, while inhibiting phagocytosis has been shown to improve neuronal survival and cognitive performance after seizures, it paradoxically increases the risk of developing spontaneous recurrent seizures (SRS). Reconciling these dual effects requires a deeper understanding the spatiotemporal dynamics of microglial phagocytosis. The objective of this review is to examine the mechanisms and impact of microglial phagocytosis in the context of epilepsy and to highlight unresolved questions that warrant further investigation in this emerging field.

Graphical Abstract

Introduction

Epilepsy is a significant neurological disorder affecting approximately 50 million people worldwide, including 2.8 million in the United States (Beghi et al., 2019; Kobau et al., 2023). Among these individuals, 13-36% suffer from drug-resistant epilepsy, characterized by the inability to achieve sustained seizure freedom despite adequate trials of two or more anti-seizure medications (Sultana et al., 2021). One proposed explanation for this therapeutic failure is that current drugs may not adequately target key mechanisms, including those involving non-neuronal cells (Bazhanova et al., 2021). In this context, microglia have emerged as promising therapeutic targets for epilepsy. These cells are the resident immune cells of the central nervous system and function as professional phagocytes. Originating from the yolk sac during early embryonic development, microglia migrate to the CNS, where they are maintained through self-renewal (Saijo et al., 2011).

Microglia play a crucial role in regulating neuronal activity during acute seizures and orchestrating responses to tissue injury in the sub-acute and chronic phases (Bosco et al., 2018; Umpierre & Wu, 2021; Zhao et al., 2024). Neuronal damage following seizures activate microglia, which subsequently reshape neuronal organization through inflammatory and phagocytic signaling (Figure 1). These processes significantly impact cognition and future seizure risk, both of which are critical clinical outcomes (Wyatt-Johnson & Brewster, 2019; Novak et al., 2022; Yu et al., 2023). Aberrant microglial phagocytosis has also been implicated in epileptogenesis (Chu et al., 2010; Libbey et al., 2010; Ma et al., 2013). A deeper understanding of microglial phagocytosis in epilepsy may reveal novel therapeutic targets, particularly for patients with drug-resistant epilepsy. This review examines recent advances in understanding the mechanisms, regulatory elements, and effects of microglial phagocytic signaling in epilepsy. Additionally, we highlight major knowledge gaps, particularly regarding the spatiotemporal dynamics and regulatory mechanisms of microglial phagocytosis following seizures.

Figure 1: Seizure induced cellular injury activates microglia phagocytosis which in turn shapes tissue reorganization and behaviors.

Seizure induced cellular injury is associated with release of ‘find me’ signals and pro-inflammatory cytokines, as well as display of ‘eat me’ signals by apoptotic and stressed neurons. The ‘find me’ and ‘eat me’ signals are recognized by microglial phagocytic receptors whose downstream signalling initiates cytoskeletal modifications necessary for approaching and engulfing the target. These targets include dead and injured neurons, synapses, and aberrant newborn neurons which in turn affects clinically significant behavioral outcomes.

A wide array of phagocytic receptors regulates microglial responses to seizures

Seizures lead to extensive neuronal damage and death through a variety of mechanisms (Dingledine et al., 2014). These include inflammatory modes such as necrosis, necroptosis, and pyroptosis and non-inflammatory modes like apoptosis, autophagy, and phagoptosis. Apoptotic cells are cleared by phagocytes including microglia through a process termed efferocytosis and is extensively reviewed elsewhere (Doran et al., 2019). However, microglia can also phagocytose stressed-but-viable neurons through phagoptosis where unlike in the case of apoptotic cell clearance, phagocytosis itself is the primary mode of cell death (Brown & Neher, 2014; Abiega et al., 2016; Marchi & Brewster, 2024). It occurs following reversible exposure of eat me signals like phosphatidylserine (PS) such that blocking any part of PS signaling cascade rescues a substantial proportion of neurons from cell death (Neher et al., 2011; Fricker et al., 2012; Bosco et al., 2025). Thus, microglia contribute to seizure-induced neuronal loss through both efferocytosis and phagoptosis.

The general mechanisms of microglia phagocytosis involve detection of apoptotic and damaged cells via a variety of find me signals, engulfment via eat me signals, intracellular signaling, phagolysosome formation and maturation (Figure 2). Engagement of phagocytic receptors lead to the activation of intracellular tyrosine kinases and cytoskeletal rearrangement. The specific mechanisms downstream of each receptor are not fully understood (Medina & Ravichandran, 2016). Additionally, phagocytic receptors also participate in signaling which influences microglia proliferation and the production or suppression of pro-inflammatory NFκB related cytokines (Wyatt-Johnson & Brewster, 2019). Below we discuss several of these phagocytic receptors, ligands, and their signaling mechanisms.

Figure 2: Ligand-receptor interactions regulating microglial phagocytosis in epilepsy.

Following a seizure, a variety of ‘find me’ and ‘eat me’ signals promote microglia phagocytosis of dead and stressed neurons, while ‘don’t eat me’ signals protect healthy cells. This wide array of receptor-ligand interactions facilitates fine tuning of microglia response based on the substrate type and extracellular milieu, as well as in coordinating phagocytic response with other signaling pathways such as inflammatory cytokine release and proliferation. Purines (ATP, UDP, UTP), S1P, and sCX3CR1 are common find me signals which bind to a variety of GPCRs (G protein-coupled receptors). Similarly, phosphatidylserine (PS) and activated complement components are common ‘eat me’ signals. Receptors like BAI1, TIM4, MerTK, and TREM2 interact with PS either directly or via bridging molecules like ProtS, Gas6, tubby, and TULP.

Nucleotides and purinergic signaling

ATP, UTP, and UDP are amongst the best known find me signals in the nucleotide category (Figure 2) (Koizumi et al., 2007; Elliott et al., 2009). During apoptosis, ATP and UDP are released via a variety of mechanisms including through Pannexin 1 (PANX1) channels activated by action of caspases (Elliott et al., 2009; Chekeni et al., 2010), lysosomal or vesicular secretion (Martins et al., 2013), passive release from necrotic cells, or via other channels such as connexins (Dosch et al., 2018). P2Y6 and P2Y2 receptors respond to these nucleotide signals and facilitate removal of apoptotic cells (Hide et al., 2023; Umpierre et al., 2024). UDP-P2Y6 signaling has been studied in epilepsy and known to affect neuronal phagocytosis by modulating microglial calcium activity, promoting lysosome biogenesis (calcium dependent) and release of NFκB associated cytokines (calcium independent) (Koizumi et al., 2007; Umpierre et al., 2024). There is also some evidence supporting a role for the microglial P2Y12 receptor in phagocytosis (Eyo et al., 2014; Inouye, 2018; Diaz-Aparicio et al., 2020; Chidambaram et al., 2022). P2Y12 is a Gi coupled GPCR, and we recently showed that prolonged activation of Gi signaling via chemogenetic manipulation arrested microglia in a homeostatic state and reduced phagocytosis of neurons post seizures (Dheer et al., 2024).

Phosphatidylserine, MerTK, and TREM2 signaling

Phosphatidylserine (PS) is an asymmetrically distributed membrane phospholipid maintained in the inner leaflet of the membrane through the action of flippase. During apoptosis, caspases activate ‘scramblase’ which exposes PS on the outer membrane (Vorselen, 2022). This exposed PS is bound directly by microglial receptors like BAI4 and TIM-4 (Vorselen, 2022), providing an initial tether followed by binding of soluble bridging molecules with their receptors to complete engulfment (Fricker et al., 2012; Vorselen, 2022). Protein S, Gas6, tubby, and tubby-like protein 1 are such bridging molecules mediating interaction between PS and the TAM family of receptors (TYRO3, Axl, and Mer receptor tyrosine kinase or MerTK) on microglia (Figure 2) (Shafit-Zagardo et al., 2018).

MerTK and Axl activation leads to autophosphorylation and downstream activation of phospholipase C gamma (PLC-γ), phosphatidylinositol 3-kinase (PI3K), and subsequent activation of Rac GTPase (Yadav et al., 2025). This activates cytoskeletal elements leading to pseudopod extension and formation of phagocytic cup (Botelho & Grinstein, 2011). MerTK plays a role in both tethering and engulfment of apoptotic cells (Dransfield et al., 2015). Additionally, both MerTK and Axl signaling suppresses the pro-inflammatory NFκB signaling (Wu et al., 2018; Vago et al., 2021) and promotes resolution of inflammation (Cai et al., 2018). Microglial loss of TAM receptors was associated with deficient phagocytosis of excitatory synapses and newborn dentate gyrus neurons and an increased incidence of seizure-related mortality in mouse model of Alzheimer’s disease (Huang & Lemke, 2022).

TREM2 is another important PS receptor (Bosco et al., 2025). Mutations in TREM2 and its adaptor protein TYROBP have been identified as the causes of Nasu-Hakola disease, a neurodegenerative disease commonly presenting with seizures (Paloneva et al., 2000; Paloneva et al., 2001; Paloneva et al., 2002). PI3K, SYK, and PLC-γ mediate phagocytic signaling downstream of TREM2 (Yao et al., 2019). Like MerTK, TREM2 activation also suppresses pro-inflammatory NFκB signaling (Li et al., 2019; Yao et al., 2019). TREM2 deficiency was associated with decreased neuron loss and increased spontaneous seizures in mice (Bosco et al., 2025). Furthermore, TREM2 promotes microglial proliferation (Wang et al., 2020; Bosco et al., 2025) which is essential to meet the increased phagocytic demand in the aftermath of seizure induced neuronal damage.

Complement signaling

Levels of complement components C1q and C3 are increased in both rodent models of epilepsy as well as epileptic patients including those with drug resistant seizures (Wei et al., 2021; Gruber et al., 2022). C1q and its downstream component activated C3 (C3b and iC3b) opsonize immature CNS synapses during a discrete window of postnatal development (Stevens et al., 2021). These tagged synapses are eliminated through C3b mediated activation of microglial complement receptor 3 (CR3; Figure 2). Spleen tyrosine kinase (Syk) and Rho GTPase mediate the downstream signaling from the complement receptor 3 with subsequent engagement of cytoskeletal elements (Uribe-Querol & Rosales, 2020). Deficiency of complement components C1q and C3 is associated with increased seizures in both rodents and humans secondary to deficits in synaptic phagocytosis (Chu et al., 2010; Libbey et al., 2010; Schaarenburg et al., 2016).

Fractalkine (CX3CL1-CX3CR1) signaling

CX3CL1 is the only known member of the CX3C chemokine family (Zhao et al., 2023). It exists in two forms- membrane bound and soluble (sCX3CR1), the latter forming from cleavage of membrane bound form by metalloproteinases ADAM10 and ADAM17 (Figure 2) (Zhao et al., 2023). It signals via the CX3CR1 receptor that is expressed widely by immune cells including microglia (Sokolowski et al., 2014; Zhao et al., 2023). This signaling is multifaceted such that both membrane bound and soluble CX3CL1 can act as a chemoattractant, membrane bound CX3CL1 can additionally function as an adhesion molecule, and the activation of CX3CR1 receptor raises microglial calcium and promotes chemotaxis (Harrison et al., 1998; Zhao et al., 2023). Increased expression of CX3CL1 and its association with neuronal loss has been observed in the brains of epileptic patients and rodent models (Xu et al., 2012). Blocking either CX3CL1 or its microglial receptor CX3CR1 was found to reduce hippocampal neuronal loss after status epilepticus in rodent models (Yeo et al., 2011; Ali et al., 2015 Feb), highlighting the significance of fractalkine signaling in microglia-neuron interaction in epilepsy.

Sphingosine-1-phosphate (S1P) signaling

Apoptotic cells have been shown to upregulate sphingosine kinase 1 (SphK1) which in turn increases S1P secretion (Xiao et al., 2023). S1P acts as a potent chemoattractant for macrophages, monocytes, and microglia (Zahiri et al., 2021). While microglia can detect S1P via S1P receptors, a recent study identified that S1P can also influence phagocytosis by binding to and activating the microglial TREM2 (Figure 2) (Xue et al., 2021). S1P signaling plays important roles in many epilepsy related mechanisms including neuronal apoptosis, blood-brain barrier stability, and inflammation (Wang et al., 2024). While there is some evidence for a therapeutic role of S1P immunomodulators in reducing microglial activation and inflammatory responses in epilepsy (Leo et al., 2017), direct investigation of microglia specific S1P signaling is lacking.

Other find me and eat me signals

Lysophosphatidylcholine (LPC) and its receptor G2A are other important receptor/ ligand systems involved in ‘find me’ signaling (Lauber et al., 2003; Peter et al., 2008; Inose et al., 2015). Similarly, calreticulin is upregulated on the surface of apoptotic cells and recognized by the LDL-receptor related protein 1 (LRP1) receptor with their interaction modulated by the ‘don’t eat me’ signal CD47 (Gardai et al., 2005). Blocking LRP1 signaling in microglia has been shown to reduce neuroinflammation in neuropathic pain (Brifault et al., 2019). In the context of epilepsy, LRP1 expression has been found to be increased in epileptic foci in multiple cell types including microglia (Wang et al., 2021). However, there are currently no studies that have examined either LPC-G2A or calreticulin-LRP1 signaling in epilepsy is any detail.

Don’t eat me signals protect healthy cells from phagocytosis

Don’t eat me signals serve to protect healthy cells through negative regulation of phagocytosis. Currently known ‘don’t eat me signals’ include- CD47, programmed cell death ligand 1 (PDL-1), beta-2-microglobulin, and CD24 (Gardai et al., 2005; Barkal et al., 2019; Khalaji et al., 2023). Neuronal CD47 and its microglial receptor SIRPα have been shown to be involved in synaptic pruning (Lehrman et al., 2018). However, there is paucity of direct evidence for a role of ‘don’t eat me’ signals in epileptogenesis and post-seizure neuronal loss (Figure 2).

Lysosomal function and inflammatory cytokine release is closely linked with phagocytic signaling

Lysosomal function

Factors that affect microglial lysosomal function and their fusion with phagosomes also influence phagocytic and autophagic activity. For example, presenelin-1 (PSN1) gene mutations are the most common cause of autosomal dominant Alzheimer’s disease with 20% patients also presenting with epilepsy (Larner, 2011). Mechanistically, PSN1 gene is involved in regulating lysosomal acidification and its fusion with autophagosome (Quick et al., 2023). Similarly, progranulin is essential for maintaining lysosomal acidification (Tanaka et al., 2017) and its loss-of-function mutations are associated with neuronal ceroid lipofuscinoses type 11 which presents with neurodegeneration and refractory seizures in children (Nóbrega et al., 2025). Recently, P2Y6 signaling was also shown to regulate lysosomal biogenesis in the context of seizures (Umpierre et al., 2024). Lastly, microglia phagocytosis maybe an important mechanism in ATP6V1A encephalopathy where over 80% of patients present with seizures (Guerrini et al., 2022). Interestingly, a recent study showed that microglial glycoprotein non-metastatic melanoma B (GPNMB) is upregulated in seizures and interacts with ATP6V1A to facilitate phagocytosis. Deficiency of GPNMB was associated with phagocytic defects and worsened seizures (Liu et al., 2025). Thus, proteins regulating lysosomal function serve as important links between phagocytosis and genetically predisposed epilepsy development.

NFκB signaling

NFκB signaling complex controls multiple aspects of innate and adaptive immunity by regulating inflammatory responses (Liu et al., 2017b). Phagocytic receptors that respond to damage associated molecular patterns (DAMPS) or pathogen associated molecular patterns (PAMPS) upregulate NFκB mediated pro-inflammatory signaling (Figure 1). These include P2Y6, toll like receptors (TLRs), complement receptor 3 (CR3), and Fcγ receptors (Cao et al., 2010; Moretti & Blander, 2014; Matsuda et al., 2015; Reis et al., 2019; Umpierre et al., 2024). Conversely, the phagocytic receptors that respond to non-DAMP, non-PAMP signals and maintain homeostatic clearance functions of microglia downregulate NFκB associated pro-inflammatory cytokine release (Figure 1). These include MerTK, TREM2/ DAP12, and Axl receptors (Hasanbasic et al., 2004; Tibrewal et al., 2008; Yao et al., 2019). NFκB signaling is thus an important effector determining the pro- or non-inflammatory nature of phagocytic signaling. Downregulating NFκB signaling has been shown to be beneficial in epilepsy due to reduced pro-inflammatory responses (Cai & Lin, 2022). In macrophages, blocking NFκB has been shown to reduce phagocytic responses to pathogens (Zou & Shankar, 2015; Liu et al., 2017a). However, whether the same holds true in epilepsy remains scantly studied (Wu et al., 2022).

Epigenetic regulators and transcription factors provide high level regulation of phagocytic machinery

Epigenetic regulators: PRC2, Kdm6a, and Kdm6b

Regional heterogeneity in the expression of phagocytic machinery and phagocytic ability of microglia is well established (Biase et al., 2017; Ayata et al., 2018). This heterogeneity is maintained by epigenetic regulators such as the Polycomb repressive complex 2 (PRC2) and functions in matching the microglial phagocytic activity with the burden of apoptotic cell clearance under homeostatic conditions (Ayata et al., 2018). For example, cerebellar microglia have higher phagocytic activity than striatal microglia, in line with higher apoptotic cell burden in cerebellum (Ayata et al., 2018). PRC2 functions by silencing phagocytosis related genes by methylating histone H3 at lysine 27 (H3K27me3). Whereas in cerebellum, induction of histone demethylases Kdm6a and Kdm6b relieve this suppression (Ayata et al., 2018). Disruption of this spatially heterogenous phagocytic activity by suppression of PRC2 was associated with increased phagocytic activity of striatal microglia, altered morphology of striatal medium spiny neurons, behavioral abnormalities, and seizures (Figure 3) (Ayata et al., 2018).

Figure 3: Aberrant phagocytosis contributes to epileptogenesis.

Phagocytic effectors, transcription factors, and epigenetic regulators maintain homeostatic control and regional heterogeneity of microglia phagocytosis. Both excessive as well as insufficient phagocytic activity can lead to epileptogenic remodeling of neuronal circuits. For example, elevated mTOR and PU.1 activity enhance microglial phagocytosis while deficiency of TREM2 and complements reduce phagocytosis, both results in increased seizures.

Transcription factors: PU.1, p53, mTOR, PPARs, and Nrf2

PU.1:

PU.1 (encoded by gene SPI1) is an essential transcription factor in microgliogenesis and defining microglia identity from erythromyeloid progenitors (Kierdorf et al., 2013). Silencing PU.1 downregulates phagocytic receptors and signaling partners like TREM2 and TYROBP/ DAP12, and reduces microglia phagocytosis, indicating its continued role in this process beyond development (Smith et al., 2013; Huang et al., 2017; Rustenhoven et al., 2018). Upregulation of SPI1/PU.1 is seen in tuberous sclerosis complex and focal cortical dysplasia which present with epilepsy (Figure 3) (Zimmer et al., 2021). There is some evidence to indicate that dysregulated phagocytosis may be involved in the pathogenesis of seizures in these cortical malformations (Zhao et al., 2018).

p53 tumor suppressor gene:

p53 is a tumor-suppressor gene and transcriptional regulator of cell cycle progression, DNA damage response, and apoptosis (Kruiswijk et al., 2015). Its activity is increased in microglia exposed to oxidative stress and DNA damage (Holtman et al., 2017). Knocking out or suppressing p53 in microglia has been shown to reduce pro-inflammatory cytokine release and increase phagocytic ability (Jayadev et al., 2011). Its effect on phagocytosis is mediated via the downstream regulation of Death Domain 1α (DD1α) receptor that engages in homophilic intermolecular interaction between apoptotic cell and macrophages (Yoon et al., 2015). Further, DD1α also promotes immune tolerance to apoptotic cell antigens through inhibitory co-stimulation of T cells (Yoon et al., 2015). It was observed that p53 general knockout mice experience more severe seizures with along with short but not long term neuroprotection in a kainate induced model of epilepsy (Engel et al., 2010). It is possible that the intricate balance between pro-inflammatory and phagocytic signaling, regulated by microglial p53, may influence these findings. Microglia-specific p53 deletion is needed to further elucidate its specific contributions.

mTOR:

mechanistic Target of Rapamycin (mTOR) signaling regulates fundamental cellular processes like growth, metabolism, and autophagy (Saxton & Sabatini, 2017). Selectively elevating mTOR signaling in microglia was associated with increased proliferation and phagocytic activity as well as development of spontaneous seizures, without a significant increase in pro-inflammatory cytokines (Figure 3) (Zhao et al., 2018). RNA-seq analysis revealed significant changes in many genes directly or indirectly involved in phagocytosis, thus highlighting that aberrant microglia phagocytosis may contribute to epileptogenesis (Zhao et al., 2018). However, microglia specific deletion of mTOR is also associated with worsened outcomes including increased neuronal loss and more severe spontaneous recurrent seizures, along with reduced proliferative responses to injury. (Zhao et al., 2020). These observations suggest that an optimal level of microglial mTOR activity is essential for a balanced response to seizure induced neuronal injury.

PPAR:

Peroxisome proliferator-activated receptors (PPARs) are a family of nuclear receptor transcription factors that function as lipid sensors and are involved in fatty acid metabolism (Tyagi et al., 2011). PPARγ agonists were found to increase amyloid-β phagocytosis through upregulation of phagocytic receptors Axl, MerTK, TREM2, and CD36 (Yamanaka et al., 2012; Savage et al., 2015). In the context of epilepsy, PPARγ agonists were found to rescue neuron loss, suppress the development of spontaneous recurrent seizures, reduce microglial activation, and inflammation by modulating the BDNF/TrkB signaling (Hong et al., 2013; Peng et al., 2019). However, the contribution of phagocytosis in these effects needs further evaluation.

Nrf2:

Nuclear factor erythroid 2-related factor 2 (Nrf2) is a master regulator of antioxidant response to oxidative stress and a proposed therapeutic target in epilepsy (Carmona-Aparicio et al., 2015). Activation of Nrf2 upregulates antioxidative defense and phagocytic ability of microglia, the latter via up-regulation of scavenger receptor CD36 (Zhao et al., 2015).

Phagocytosis shapes tissue reorganization after seizures

Post seizure pathology comprises widespread changes in tissue organization including loss of neurons and synapses, aberrant neurogenesis, and glial hypertrophy. Below we discuss the changes shaped by the phagocytic activity of microglia.

Neuronal loss

Epilepsy is associated with neuronal loss in multiple brain regions including hippocampus, amygdala, and entorhinal cortex . While some of the cell loss may be a direct consequence of seizure (Dingledine et al., 2014), microglial phagocytosis also plays an important role (Figure 1). This is evident from the neuroprotection observed when microglial phagocytic ability is compromised (Yeo et al., 2011; Umpierre et al., 2024; Bosco et al., 2025). For example, knocking out phagocytic receptors TREM2, P2Y6, C3R, or CX3CR1 as well as ‘eat me’ signals like complement 3 is associated with reduced neuronal loss after seizures (Yeo et al., 2011; Wei et al., 2021; Schartz et al., 2023; Umpierre et al., 2024; Bosco et al., 2025).

The mechanism of this neuroprotection likely involves a combination of reduced phagoptosis (Brown & Neher, 2014) and downregulation of pro-inflammatory NFκB associated cytokines downstream of phagocytic signaling (Umpierre et al., 2024). However, whether the stressed neurons that did not undergo phagoptosis function normally is unclear. In a kainate induced mice model of epilepsy, TREM2 knock out was associated with reduced phagocytic clearance of neurons as well as increased risk of spontaneous seizures (Bosco et al., 2025). The authors speculate that these ‘damaged neurons’ that survive may be pro-epileptogenic and contribute to the development of spontaneous recurrent seizures (Bosco et al., 2025). While others have suggested that neuronal death itself (Dudek et al., 2010) or the biochemical pathways causing programmed cell death may contribute to epileptogenesis (Dingledine et al., 2014).

Synapse reorganization

During brain development, microglia modulate synaptic connectivity in an activity dependent manner (Tremblay et al., 2010; Schafer et al., 2012; Nebeling et al., 2023). Although the mechanisms behind this, specially whether it involves a primarily microglia driven pruning of synapse has recently been debated (Pereira-Iglesias et al., 2024). Complement mediated signaling as described previously plays a key role in microglia-synapse interaction (Gasque, 2004; Stevens et al., 2007). Additionally, CX3CR1 (Paolicelli et al., 2011) and TREM2 (Filipello et al., 2018) signaling has also been shown to be important in microglial synaptic pruning. It has been observed that reduced microglial synaptic pruning during development such as in C1q knock out mice may predispose to recurrent seizures (Figure 3) (Chu et al., 2010; Ma et al., 2013). However, excessive synaptic pruning such as that observed with loss of Polycomb repressive complex 2 (PRC2), an epigenetic regulator of phagocytosis (Ayata et al., 2018) or with increased levels of C1q or C3 deposition in adult brain (Hong et al., 2016; Gomez-Arboledas et al., 2021) can also predispose to increased seizures (Figure 3). Conversely, a recent study showed that blocking C3 mediated inflammation and synaptic phagocytosis with captopril reduces risk of spontaneous seizures and cognitive deficits (Dong et al., 2022). Further, in the hours to days following seizures, microglia were shown to facilitate structural resolution of injured, beaded dendrites through direct physical interactions with their process pouches (Eyo et al., 2021). This may also play a role in synapse reorganization.

Aberrant neurogenesis

Aberrant neurogenesis is increased in the dentate gyrus following seizures (Victor & Tsirka, 2020; Lybrand et al., 2021). Non-inflammatory, homeostatic microglia keep excess neurogenesis in check through apoptosis coupled phagocytosis (Sierra et al., 2014). Further, following inflammatory challenge, microglia use various strategies to boost their phagocytic capacity to efficiently clear increased apoptotic cell load (Abiega et al., 2016). However, this apoptosis-phagocytosis coupling is lost in both patients as well as rodent model of mesial temporal lobe epilepsy (MTLE). These microglia show reduced surveillance and expression of apoptotic cell recognition receptors with resulting accumulation of apoptotic newborn neurons in the neurogenic niche (Abiega et al., 2016). However, microglial phagoptosis of viable newborn cells is increased around the same time, indicating a compensatory mechanism, specially keeping ectopic granule cells in check (Abiega et al., 2016; Luo et al., 2016; Eyo et al., 2017). It has been shown that aberrant hippocampal neurogenesis contributes to epileptogenesis and its associated cognitive decline (Cho et al., 2015; Lybrand et al., 2021). Interestingly, microglial ablation or P2Y12 knockout was found to be associated with reduced abberent neurogenesis and immature neuronal projections following seizures (Mo et al., 2019). Similarly, deficiency of either P2Y12 or MerTK/Axl phagocytic receptors was found to disrupt basal neurogenesis (Diaz-Aparicio et al., 2020). This could be due to disruption of the beneficial roles played by microglia in the proliferation and integration of neural progenitor cells, including through their phagocytic secretome (Matsuda et al., 2015; Sandvig et al., 2018; Diaz-Aparicio et al., 2020).

Microglia phagocytosis modulates epilepsy related behavioral outcomes

Cognitive function

Seizures are commonly associated with cognitive decline and behavioral abnormalities in both epilepsy patients and rodent models (Figure 1) (Elger et al., 2004; Novak et al., 2022). These abnormalities include deficits in memory and attention, low IQ, and poor scholastic performance (Lenck-Santini & Scott, 2015). The mechanisms of these deficits are multifactorial, involving a complex interplay between seizure etiology, the seizures themselves, interictal discharges, and anti-epileptic medications (Lenck-Santini & Scott, 2015). Neuronal loss in hippocampus is one such mechanism. Loss of hippocampal volume and neurons is seen in both patients as well as animal models of temporal lobe epilepsy and correlates with cognitive decline and psychiatric comorbidity (Cascino, 2003; Fuerst et al., 2003; Umpierre et al., 2024; Bosco et al., 2025). Reducing hippocampal neuronal loss by attenuating microglial phagocytosis and pro-inflammatory cytokine production is associated with improved cognitive performance (Tian et al., 2017; Schartz et al., 2023; Umpierre et al., 2024).

Thus, microglial phagocytic machinery maybe an attractive target for therapeutic interventions for improving cognition in epilepsy. However, a better understanding of the spatio-temporal dynamics of phagocytosis following acute seizures is required to selectively reduce detrimental effects and identify neuronal substrates underlying various cognitive and behavioral abnormalities. Also, the network effect of modulating phagocytosis, i.e. how it affects the interaction between different brain regions necessary for normal cognition has not been studied. Loss of homeostatic microglial function, such as regulating neuronal excitability (Badimon et al., 2020; Wu et al., 2020) following seizures may also be involved in cognitive deficits and need exploration.

Spontaneous recurrent seizures (SRS)

We have shown that microglia depletion increases the risk of SRS in mice (Wu et al., 2020). More recently we demonstrated that this increase in SRS is recapitulated in mice with TREM2 knockout associated phagocytic deficits (Bosco et al., 2025). The neuronal substrate that mediates the increased risk of seizures following status epilepticus and spared by microglial phagocytic defect is not well understood. However, aberrant neurogenesis in the dentate gyrus, ectopic granule cells, and immature neuronal projections after seizures are promising candidates.

While we have extensively discussed the effect of seizures on inducing phagocytosis, it is important to note that aberrant phagocytosis may itself be epileptogenic (Figure 3). C1q or C3 deficiency is associated with failure of synaptic pruning, enhanced excitatory synaptic connectivity and increased seizures in mice (Chu et al., 2010; Libbey et al., 2010; Ma et al., 2013). Seizures were also the most frequent neuropsychiatric symptoms reported in C1q deficient patients diagnosed with systemic lupus erythematosus (SLE) (Schaarenburg et al., 2016). Similarly, an abnormal increase in microglia phagocytosis can also increase seizure risk. Enhanced microglial mTOR signaling was associated with loss of cortical and hippocampal synapses, development of spontaneous recurrent seizures, and mortality (Zhao et al., 2018). Increased mTOR signaling is epileptogenic and observed in patients with tuberous sclerosis (Crino, 2016).

Lastly, it has been suggested that microglia may preferentially prune inhibitory synapses contributing to the excitation/ inhibition imbalance and epileptogenesis (Andoh et al., 2019; Fan et al., 2023). Interestingly, a recent study found that PS exposure may play a role in this. In mice with acute neuronal deletion of a flippase chaperone, a preferential exposure of PS in neuronal soma and a microglia MerTK dependent loss of inhibitory post-synapses were observed along with increased excitability and seizures (Park et al., 2021).

Concluding remarks and future directions

Over the past two decades, numerous groundbreaking discoveries have deepened our understanding of phagocytosis in epileptogenesis, post-seizure pathology, and the effects of seizures on phagocytic signaling (Wyatt-Johnson & Brewster, 2019). These advances include the identification of complementary ‘find me’ and ‘eat me’ signaling pathways, which facilitate the efficient clearance of damaged and stressed neurons (Medina & Ravichandran, 2016; Galloway et al., 2019; Wyatt-Johnson & Brewster, 2019). In parallel, a diverse array of phagocytic receptors (Figure 2) enables microglia to fine-tune their responses by recognizing DAMPS and apoptotic cells, leading to either the stimulation or suppression of pro-inflammatory cytokine release (Galloway et al., 2019; Wyatt-Johnson & Brewster, 2019). These responses are orchestrated by epigenetic regulators and transcription factors. However, significant gaps remain in our knowledge. The mechanisms underlying the dynamic switching between pro- and anti-inflammatory cytokine release downstream of phagocytic signaling are poorly understood (Paolicelli et al., 2022). Similarly, the role of ‘don’t eat me’ signals in epilepsy remains largely unexplored (Medina & Ravichandran, 2016). Furthermore, the interplay between phagocytic signaling and other pathways, such as those regulating proliferation and metabolism, which are concurrently altered in epilepsy, is yet to be fully elucidated.

Seizures exert a complex influence on microglial phagocytosis. Increased apoptotic and necrotic cell burden drives microglia to adopt compensatory mechanisms to meet heightened phagocytic demands (Abiega et al., 2016). However, in cases of recurrent or severe seizures, a prolonged pro-inflammatory environment can impair phagocytic capacity by downregulating phagocytic receptors (Cherry et al., 2014). This phenomenon underscores the growing interest in anti-inflammatory therapies for managing seizures and epilepsy, some of which target phagocytosis pathways (Dey et al., 2016; Dong et al., 2022).

Following status epilepticus, microglial phagocytosis has been implicated in cognitive deficits (Schartz et al., 2023; Umpierre et al., 2024). Current rodent studies predominantly focus on hippocampal-dependent memory assessments (Antunes & Biala, 2011). However, a more comprehensive evaluation is needed to characterize additional cognitive and psychological abnormalities, identify their underlying neuronal substrates, and investigate the role of microglial phagocytosis in shaping these behavioral outcomes.

Interestingly, microglial phagocytosis plays a pivotal role in modulating the risk of spontaneous recurrent seizures following status epilepticus. Emerging evidence suggests that aberrant neurogenesis in the dentate gyrus may serve as the primary substrate for this effect. Microglia exhibit a complex relationship with newborn neurons in the neurogenic niche: while they regulate aberrant neurogenesis by phagocytosing both apoptotic and viable newborn neurons, they also promote neurogenesis and support the growth of immature neuronal projections (Sierra et al., 2010; Abiega et al., 2016; Mo et al., 2019). The mechanisms through which microglia drive aberrant neurogenesis, however, remain poorly understood. Additionally, microglia contribute to epileptogenesis by mediating abnormal synaptic phagocytosis during synaptic development. Dysregulated microglial activity, whether excessive or insufficient, can have widespread effects on the central nervous system, as observed in conditions such as tuberous sclerosis or C1q deficiency-related seizures in systemic lupus erythematosus (Schaarenburg et al., 2016; Zhao et al., 2018). These insights underscore the potential of targeting microglial phagocytosis as a therapeutic strategy for such disorders.

Reconciling whether microglial phagocytosis is beneficial, or detrimental following status epilepticus may ultimately depend on contexts. Temporarily restricting phagocytosis could be advantageous, as it might enhance cognitive function without significantly increasing the risk of future seizures. This hypothesis could be tested in rodent models by attenuating microglial phagocytosis at various time points after an initial status epilepticus. However, prolonged attenuation of phagocytosis might elevate the risk of subsequent seizures, leading to cumulative cognitive decline over time. Another way to separate the beneficial from detrimental effect maybe by spatially restricting the attenuation of phagocytosis to the select brain regions. For instance, pharmacological agents that reduce phagocytosis could be delivered directly to the hippocampus following status epilepticus in rodent models. Subsequent assessments could evaluate both cognitive outcomes and the risk of developing spontaneous recurrent seizures (SRS). If these interventions show promise, future approaches could explore the use of small molecules targeting epigenetic regulators (Cheng et al., 2019; Singh et al., 2023) to modulate microglial phagocytosis. This strategy could enable precise, spatially restricted attenuation or enhancement of phagocytosis, optimizing therapeutic outcomes.