Microglia and immune signaling as architects of the brain

Introduction

As the primary resident immune cells of the central nervous system (CNS), microglia perform a multitude of tasks both during development and throughout life in all vertebrate animals. Many of these vital contributions to developmental and homeostatic programs are highly conserved across species, with some notable differences observed in primates (Geirsdottir et al., 2019). In addition to protecting the nervous system against various sources of infection (Eyo and Wu, 2019), microglia secrete neurotrophic factors, participate in synaptic pruning and plasticity, and remove dead and apoptotic cells (Whitelaw et al., 2022). Moreover, dysregulation of microglia is associated with neurodevelopmental and neurodegenerative disorders (Mastenbroek et al., 2024; Solek et al., 2018; Smail and Lenz, 2024; Gao et al., 2023; Nakaso, 2024).

Microglia were first described as the “third glial cell” in 1919 by Spanish neuroscientist Pio del Rio-Hortega, which he proposed to be primarily mesoderm-derived scavengers of cellular debris [(Rio-Hortega, 1919a,b,c,d), see (Sierra et al., 2016) for translation]. For many decades thereafter, microglia were regarded as dormant or resting and ramified in the absence of immunological challenge (Barron, 1995). Upon injury or infection, they were thought to become “activated,” adopting a more amoeboid shape, increasing their motility and expressing a set of marker antigens also found in circulating monocytes. However, the advent of intravital imaging technology allowed a first observation of these cells in live animals and the demonstration that microglia are highly dynamic in their “resting” state (Nimmerjahn et al., 2005; Davalos et al., 2005). This led to an increased interest in the role of microglia in normal development and the adoption by many researchers of the more apt term “homeostatic” microglia. The following years of research established microglia as key players in successful brain development and wiring, with characterization of roles in neurogenesis, developmental neuronal apoptosis, and synaptic pruning (Paolicelli and Ferretti, 2017). In addition, findings from genome-wide association studies (GWAS) revealed that many immune genes expressed by microglia are implicated in psychiatric, neurodevelopmental (Network, 2015; Voineagu and Eapen, 2013), and neurodegenerative disorders (Holtman et al., 2024). This further spurred interest in a better characterization of microglial contributions to brain homeostasis and how these processes are dysregulated in various pathologies.

While microglia contribute the bulk of immune surveillance and signaling in the developing brain, other immune cells also play a role in CNS formation and refinement. For example, maternally-derived microchimeric microglia are found in the progeny's brain and help colonization of the brain by fetal microglia (Schepanski et al., 2025). The presence of these cells helps establish a stable microglial population and has a role in regulating synaptic pruning. Yolk sac-derived border associated macrophages are also present and populate the brain during development (Utz et al., 2020). Adaptive immune cells, such as B-1a lymphocytes and CD4+ T cells promote oligodendrogenesis and microglial maturation respectively (Pasciuto et al., 2020; Tanabe and Yamashita, 2018).

In this review, we will focus on the signaling molecules that are produced by microglia and those that recruit microglia to the sites of their activity. We will give an overview of the role of cytokines and chemokines, the complement system, membrane associated molecules and reactive oxygen species (ROS) in microglial function in developmental programs, including neurogenesis, synapse formation, maturation, and pruning.

Cytokines and chemokines

Cytokines are a broad category of small proteins used primarily by the immune system for cell signaling. Chemokines are a specialized set of cytokines that can induce a chemotactic response in cells expressing the corresponding receptors. Microglia secrete and respond to an array of cytokines and chemokines, many of which participate in neurodevelopment (Figure 1). One prominent class of these signaling molecules is pro-inflammatory cytokines, whose primary function is to respond to injury and infection by activating defense and repair mechanisms (Kany et al., 2019; O'Shea and Murray, 2008). Their functions have been extensively characterized in animal models of disease, and additional correlative evidence has been provided from human samples. In addition, several different avenues to assess cytokine function in human microglia are being pursued, including brain organoids (Yu et al., 2025), xenotransplantation of human microglia into mouse brain organoids (Popova et al., 2021), single-cell RNA sequencing (scRNA-seq) of isolated human fetal microglia (Kracht et al., 2020) as well as embryonic stem cells (ESCs), and induced pluripotent stem cell (iPSC)-derived microglia in culture or co-culture (Schepanski et al., 2025; Han C. Z. et al., 2023; Mordelt et al., 2026).

Fractalkine

Fractalkine (FKN, also known as C-X₃-C motif chemokine ligand 1-CX3CL1) is a membrane-bound protein that mainly mediates cellular adhesion, but which can be cleaved by proteases to produce its soluble form (sFKN) (Gunner et al., 2019). Both the membrane bound and soluble forms of FKN bind exclusively to C-X₃-C motif chemokine receptor 1 (CX3CR1) and modulate microglial activity via downstream signaling cascades (Harrison et al., 1998; Tarozzo et al., 2002, 2003; Maciejewski-Lenoir et al., 1999). FKN signaling is important for the regulation of neuroinflammation and neurological disease progression, synaptic plasticity, remyelination and maintaining neural homeostasis (Chen Y. et al., 2026). In physiological settings, neurons secrete FKN to attract microglia and the interaction of membrane bound FKN with CX3CR1 is thought to maintain brain homeostasis and ensure neuronal viability (De Felice et al., 2022; Goshi et al., 2020; Wood and Singer, 2025).

Correlative evidence of FKN roles in brain development comes from human studies and animal models of neurodevelopmental disorders (NDDs). A meta-analysis of gene expression profiles showed a significant down-regulation of CX3CR1 expression in schizophrenia patients compared to healthy controls (Bergon et al., 2015). FKN levels were also shown to be significantly lower in an autism spectrum disorder (ASD) patient group compared to control subjects (Kirsan et al., 2025). An analysis of human cohorts showed a genetic association of the Ala55Thr variant of CX3CR1 with ASD and schizophrenia phenotypes (Ishizuka et al., 2017).

In maternal immune activation (MIA) models in rodents, polyinosinic:polycytidilic acid [poly(I:C)] caused a decrease in microglial expression of CX3CR1 along with working memory and social behavior impairments (Mattei et al., 2017). Bacterial lipopolysaccharide (LPS)-induced maternal immune activation likewise showed a decrease in microglial CX3CR1 protein expression and a morphological phenotype in the hippocampus and dentate gyrus of male mice (de Cossio Fernandez et al., 2021).

Disruptions of FKN signaling suggest crucial roles in brain development. Indeed, in CX3CR1 knockout (KO) mice, microglial colonization of the hippocampus and somatosensory cortex in the postnatal brain is reduced compared to heterozygous or wildtype (WT) littermates (Paolicelli et al., 2014; Hoshiko et al., 2012). A greater number of apoptotic neurons was detected in the motor cortex of CX3CR1 KO mice (Ueno et al., 2013), suggesting a role for FKN signaling in the regulation of neuron number during normal development. CX3CR1 KO rats exhibit decreased hippocampal neurogenesis compared to WT animals, and interference with FKN signaling with CX3CR1 antibodies led to a decrease in the survival of immature neurons in young rats (Bachstetter et al., 2011). In FKN KO mice, cognitive impairments in long-term memory and spatial learning and memory were correlated with deficits in both hippocampal neurogenesis and long-term potentiation (LTP), which can be rescued by exogenous introduction of sFKN (Winter et al., 2020).

CX3CR1 signaling has also been shown to be important for synaptic pruning, the removal of excess or unused synapses during neural circuit refinement (Faust et al., 2021). In the mouse, significantly more PSD95 puncta were found in CX3CR1 KO hippocampal neurons, with a concurrent increase in dendritic spines and number of immature synapses at postnatal day 15 (P15) compared to WT littermates (Paolicelli et al., 2011). Downregulation of CX3CR1 in anterior cingulate cortex at P21 showed increased spine density and miniature excitatory post-synaptic current (mEPSC) frequency, suggestive of an impairment in pruning (Cao et al., 2021).

CX3CR1 KO mice also revealed a crucial role for FKN signaling in regulating the maturation of excitatory synapses during normal development. Hoshiko and colleagues found a delay in the maturation of N-methyl-D-aspartate (NMDA) receptor current kinetics at thalamocortical synapses during the second postnatal week in CX3CR1 KO mice compared with WT littermates, as well as reduced α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor content (Hoshiko et al., 2012). Conversely, exogenous application of FKN to hippocampal slices from 1 month old mice resulted in potentiation of NMDAR function due to the induction of D-serine release from microglia (Scianni et al., 2013). A study by Zhan et al. (2014) using whole-cell recording of Schaffer collateral synapses onto hippocampal CA1 pyramidal neurons, found that the ratio of spontaneous to miniature excitatory postsynaptic currents (EPSCs) was chronically reduced in the CX3CR1 KO mice at P38-P40, suggesting a failure to develop mature multisynapse boutons and therefore less efficient postsynaptic transmission when microglial function is compromised. CX3CR1 KO adult mice (10–12 weeks) were also shown to have impaired long-term potentiation in the hippocampus along with alterations in behavior, including a more anxious phenotype and memory encoding deficits (Mendez-Salcido et al., 2022; Rogers et al., 2011).

Hence, FKN has many demonstrated roles in brain development, from regulation of neuronal survival, to synaptic pruning, to modulation of synapse maturation. This highlights the crucial role of microglial signaling in healthy neurodevelopmental programs.

Interleukin-17A

Interleukin-17A (IL-17A) has long been viewed primarily as a key mediator of the neuroinflammatory pathways that cause or worsen neurodegenerative diseases (Chen et al., 2020). However, some roles of IL-17A in developmental programs are beginning to emerge. The expression levels of this cytokine are very low in the developing brain; however the IL-17A receptor is highly expressed in microglia (Das Sarma et al., 2009). Given that it shares protein structure elements with nerve growth factor (Hymowitz et al., 2001), it has been postulated to have neurotrophic properties. Studies examining serum levels of cytokines in human children have established a link between elevated IL-17A expression and ASD (Suzuki et al., 2011; Al-Ayadhi and Mostafa, 2012; Jones et al., 2017; Sallam et al., 2024; Barbosa et al., 2020; Hill et al., 2025), although one study reports a significant decrease in serum IL-17A levels in cases compared to cohort controls (Casey et al., 2022). Its levels are also reported to be reduced in children born following maternal Zika virus infection who exhibit microcephaly and intellectual disability (Fialho et al., 2023). This suggests that careful regulation of this cytokine is required for proper neuronal development. Indeed, a study using human iPSC-derived neural progenitor cells (NPCs) showed that application of IL-17A in culture increased expression of synaptic and neuronal polarity proteins in these cells (Gomes et al., 2022). In rodent model experiments, expression of synaptophysin and PSD95 proteins appeared to be decreased in neonatal (P6) IL-17A KO mice (Zhang Q. et al., 2025), further supporting a role for this cytokine in neurodevelopment. Increase in IL-17A levels during gestation, through prenatal stress (Amani et al., 2021), SARS-CoV-2 spike protein exposure (Erdogan et al., 2023), intraperitoneal injections of the viral mimic poly(I:C) (Garay et al., 2013) or of LPS (Erdogan et al., 2023; Yasumatsu et al., 2020), gestational hypothyroxinemia (Gonzalez-Madrid et al., 2024), maternal overexpression (Murakami et al., 2021) or recombinant IL-17 protein injection (Choi et al., 2016; Sasaki et al., 2020; Gumusoglu et al., 2020) were shown to correlate with a combination of changes in microglial morphology, cortical anomalies and/or rodent behavioral changes. Conversely, a decrease of IL-17A levels was detected in the hippocampus of P60 mice with autistic-like behavior following MIA (Dutra et al., 2023). Moreover, IL-17A helps to increase the inflammatory response by promoting the production of other cytokines, including IL-1β, TNF-α, and IL-6 (Choi et al., 2016), which provides further evidence for a potential role for this immunomodulator in neurodevelopmental programs. The precise mechanisms that are regulated by IL-17A during normal development remain to be fully characterized.

Interleukin-6

Interleukin-6 (IL-6) was first identified as a maturation factor for B-cells (Hirano et al., 1985), however, pleiotropic functions for this cytokine were rapidly characterized including on other immune cells, blood vessels, hepatocytes and neuronal cells (Kishimoto et al., 1995). Characterization of the effects of IL-6 on neuronal development are complicated by the intricate network of signaling molecules and receptors which converge on common intracellular pathways (Heinrich et al., 2003) and the multiple sources of IL-6 expression in the developing brain (Cornfield and Sills, 1991; Schobitz et al., 1993; Yan et al., 1992; Gadient and Otten, 1993, 1994a,b; Vallieres and Rivest, 1997).

Microglia express IL-6 in many inflammatory circumstances (Bobbo et al., 2019) but also express the IL-6 receptor-alpha chain (IL-6R), which binds IL-6, and the trans-membrane signaling subunit, glycoprotein 130 (gp130) (Burton et al., 2013; Couch et al., 2023). This places them in a position to mediate many of the different functions that IL-6 performs during neuronal development. Furthermore, maternally derived IL-6 was shown to influence embryonic microglial morphology in a mouse model of prenatal stress (Gumusoglu et al., 2020). In a murine MIA model, Pratt et al. (2013) showed that microglia at embryonic day 16.5 (E16.5) adopt an activated, inflammatory phenotype in maternal poly(I:C) injected WT but not IL-6 KO mice. Interestingly, CD11b-negative (non-microglial) fractions from IL-6 KO fetal mice, nonetheless, exhibited high basal expression of inflammatory cytokines even in saline-injected controls. This further supports the assertion that precisely controlled levels of IL-6 are necessary in the CNS for healthy development.

The specific roles of IL-6 in neurodevelopment are somewhat controversial but include regulation of neural stem cells (NSCs) both in vitro (Niwa et al., 1998) and in vivo (Bowen et al., 2011; Gallagher et al., 2013). IL-6 and its soluble receptor, sIL-6R, were shown to induce early neurogenesis and to initiate neuronal differentiation of human NSCs (Sulistio et al., 2018), and to promote neuronal differentiation in rat pheochromocytoma (PC12) cells in culture (Sterneck et al., 1996; Marz et al., 1997). On the other hand, IL-6 significantly reduces adult mouse hippocampal neuronal differentiation via the JAK/STAT3 pathway (Kong et al., 2019). IL-6 also has a potential neurotrophic role, as it was shown to promote the survival of postnatal rat neurons in culture (Hama et al., 1989, 1991). Moreover, inflammatory cytokines, including IL-6, can act as neuromodulators and influence neuronal excitability, in part by signaling through astrocytes (Vezzani and Viviani, 2015). Finally, increased expression of IL-6 either directly or through an infection model has been shown to cause abnormalities in neuroanatomical structural development and to impair neuronal plasticity [reviewed in (Wei et al., 2013)]. As one of the core pro-inflammatory cytokines expressed by microglia during development, IL-6 is a signaling factor that can influence multiple facets of neurodevelopment and neuromodulation.

Interleukin-1β

Interleukin-1β (IL-1β) is produced in the CNS by microglia (Giulian et al., 1986). As with most cytokines, the roles of IL-1β in pathology are much better characterized than its functions in healthy development. IL-1β is one of the main immune modulators linked to neuronal degeneration in Alzheimer's disease (Griffin and Mrak, 2002) and is at the center of signaling pathways for many additional cytokines (Kaneko et al., 2019). During neurodevelopment, elevated levels of IL-1β have been correlated with deficits in various rodent models of NDDs, from inflammatory activation of microglia to altered behavior indicative of pathology [reviewed in (Solek et al., 2018; Szabo et al., 2025)].

IL-1β is expressed at high levels in the brain during development, suggesting an essential role in neuronal development (Ma et al., 2014; Mousa et al., 1999; Giulian et al., 1988). In vitro studies showed that IL-1β influences the migration of embryonic (E14) mouse astrocyte progenitor cells in neurospheres by modulating calcium signaling (Striedinger and Scemes, 2008) and of juvenile (P15) rat cortical neurons in a dose-dependent manner (Ma et al., 2014). Further in vitro evidence suggests that IL-1β signaling contributes to the decrease in proliferation of E14 rat NPCs isolated from ventral mesencephalon (Crampton et al., 2012) and to their differentiation, likely through the Wnt5a/RhoA/JNK pathway (Park et al., 2018). IL-1β also has the potential to influence axonal growth cone turning of migrating neurons and is involved in the regulation of neurite outgrowth, as it acts synergistically with neurotrophin-3 to stimulate neurite density and length (Boato et al., 2011). In vivo inhibition of IL-1R1 in rat cortical neurons by RNAi resulted in a decrease in radial migration of these neurons at P5 (Ma et al., 2014), further supporting a role for IL-1β signaling in neuronal migration. Finally, IL-1β levels appear to need tight regulation in zebrafish axon arborization in larval stages, since both elevated levels and inhibition of expression of this cytokine led to alterations in retinal ganglion cell branching (Solek et al., 2021).

An instance of gene polymorphism in the IL-1β gene has been identified and is associated with increased incidence of ASD in humans (Pekkoc Uyanik et al., 2022), although it is unclear whether the genetic variants result in either gain- or loss-of-function phenotypes. Together, these data point to important roles for IL-1β in both healthy neuronal development and in pathological consequences resulting from either increased or decreased levels of expression.

Tumor necrosis factor alpha

Tumor necrosis factor alpha (TNF-α) is another pro-inflammatory cytokine that has been implicated in neuronal development and pathology. It is often upregulated alongside IL-1β and IL-6 in inflammatory environments (Solek et al., 2018; Zhang et al., 2026) with some exceptions (Chen Q. et al., 2026). As with many other signaling molecules involved in the CNS, the effects of TNF-α are not straightforward. For instance, high concentrations of TNF-α treatment have been shown to induce apoptosis of cultured neonatal mouse NPCs, while low concentrations of this cytokine promoted NPC proliferation and differentiation (Bernardino et al., 2008), suggesting a complex role for TNF-α in postnatal neurogenesis. Several roles for TNF-α secreted by microglia in healthy brain development have been postulated, although many are likely to be indirect. For instance, CXCR4 signaling in the presence of TNF-α was shown to increase intracellular Ca²⁺ levels and glutamate release in rat and human ESC-derived astrocytic cells in culture (Bezzi et al., 2001). Astrocytes have well-characterized and evolutionarily ancient roles in neuronal development, including in neuronal migration, axon specification and growth, neuronal differentiation and establishment and maintenance of proper synaptic connectivity (Allen and Eroglu, 2017; Chung et al., 2015; Allen and Lyons, 2018). TNF-α signaling acts as a permissive factor for the glutamate release step of astrocytic gliotransmission in the mouse dentate gyrus (Santello et al., 2011). Moreover, TNF-α can modulate excitability of all classes of voltage-gated channels, most notably voltage-gated sodium channels, a phenomenon extensively documented for peripheral dorsal root ganglion neurons in the context of neuropathic pain (Vezzani and Viviani, 2015).

TNF-α can also modulate the composition of GABA and glutamate receptor subunits at neuronal membranes; the cytokine-induced changes resulting in increased neuronal network excitability (Vezzani and Viviani, 2015). Interestingly, TNF-α released by astrocytes is responsible for the homeostatic “up-scaling” of AMPAR-mediated synaptic transmission in cortical neurons in response to chronic neural activity blockade and to monocular deprivation during the developmental critical period for ocular dominance plasticity (Kaneko et al., 2008; Heir et al., 2024). In the developing Xenopus tadpole visual system, chronic exposure to TNF-α accelerated the developmental unsilencing of immature NMDAR-only synapses via trafficking of AMPARs to retinotectal synapses, and led to expanded dendritic arborizations of optic tectal neurons (Lee et al., 2010). Therefore, TNF-α is poised to be a critical regulator of neuronal development, with closely controlled levels produced by both microglia and astrocytes being necessary for healthy circuit wiring and function.

Interleukin-33

Interleukin-33 (IL-33) is mainly secreted by astrocytes in development, and its expression decreases in the mature brain (Wicher et al., 2013), although it is also expressed by oligodendrocytes (Allan et al., 2016) and hippocampal neurons (Nguyen et al., 2020). This crucial cytokine, which functions as an alarmin and helps promote the immune response following cellular necrosis (Gadani et al., 2015), has demonstrated roles in neurodevelopment [reviewed in (Luo and Yang, 2026)]. It was shown to influence microglial synapse engulfment in the developing mouse thalamus and spinal cord, as animals with IL-33 conditionally deleted in astrocytes showed increased excitatory and inhibitory synapse numbers at P30 and fewer PSD95 puncta engulfed by microglia at P15 (Vainchtein et al., 2018). Conversely, exogenous introduction of excess IL-33 in the developing mouse brain at P15 resulted in reduced number of VGLUT-1 synapses in somatosensory thalamus and increased microglial engulfment of synaptic material (Han R. T. et al., 2023). Loss of neuronal expression of IL-33 or deletion of Il1rl1 (ST2), the receptor for IL-33 in microglia, resulted in impaired spine plasticity, reduced newborn neuron integration, and memory task deficits (Nguyen et al., 2020). IL-33 has also been shown to regulate microglial phagocytic activity in mouse cortical development in a process that involves mitochondrial regulation of bioenergetic metabolism in microglia, since disruption of the ST2-AKT pathway resulted in a reduction of IL-33 induced synaptic engulfment by microglial BV2 cells in vitro (He et al., 2022). Finally, IL-33 is involved in oligodendrocyte precursor cell maturation and myelination since both processes were impacted in IL-33 KO mice (Sung et al., 2019). Overall, signaling of IL-33 through microglia has many demonstrated roles in brain development but also in continued brain health.

Interferon-γ

Interferon-γ (IFN-γ) is another pro-inflammatory cytokine produced by microglia. Elevated gestational and perinatal levels of IFN-γ have been found in some mothers of children with NDDs (Jones et al., 2017; Allswede et al., 2016), and in the patients themselves (Masi et al., 2015), suggesting that dysregulation of this cytokine may contribute to neurodevelopmental abnormalities. IFN-γ has been shown to negatively regulate the neuronal stem cell niche in mice in vivo, and in vitro neurosphere formation from NPCs isolated from P2 and adult mice is reduced in the presence of excess IFN-γ (Li L. et al., 2010). This effect curiously was greater in the absence of microglia, suggesting they may release factors that protect against IFN-γ-mediated inhibition. IFN-γ was also shown to promote neuronal differentiation and neurite outgrowth in rat, mouse and human NSCs in vitro (Wong et al., 2004; Johansson et al., 2008; Leipzig et al., 2010; Song et al., 2005), and acts through microglia to enhance neurogenesis in mouse NPC co-cultures (Butovsky et al., 2006). Moreover, IFN-γ also promotes migration and reduces proliferation and cell survival of neural progenitors (Wong et al., 2004; Song et al., 2005; Ben-Hur et al., 2003). Conversely, IFN-γ was shown to inhibit neuronal differentiation of mouse NPCs in vitro and in the E14.5 mouse brain in vivo (Ahn et al., 2015). IFN-γ was also shown to confer neuroprotection against apoptosis via an Erk1/2-dependent mechanism (O'Donnell et al., 2015). Finally, studies in IFN-γ KO mice revealed behavioral impairments and defects in neuronal connectivity (Filiano et al., 2016), suggesting that some level of expression of this cytokine is necessary for proper neuronal development and that alterations in those levels are likely to have functional consequences.

Other cytokines and chemokines

Many other cytokines and chemokines have potential roles in neurodevelopment although their functions have mostly been investigated in the context of neuroinflammation [reviewed in (Zhang et al., 2026)]. A few studies have hinted at developmental functions for these signaling molecules in development. For instance, increased levels of Granulocyte Colony-Stimulating Factor (G-CSF) in mice led to increased dendritic spine density, a higher proportion of immature spines in the medial prefrontal cortex and altered social behavior in the offspring (Kirikae et al., 2026). The anti-inflammatory cytokine IL-10 is mostly investigated in the context where it mitigates the negative effects of neuroinflammation (Lobo-Silva et al., 2016). However, it has also been shown to regulate neurogenesis by delaying rat and mouse neural progenitor differentiation in vitro (Perez-Asensio et al., 2013), to promote neurite outgrowth in juvenile (P16-18) rat primary cortical neurons in culture (Chen et al., 2016) and to be important for microglia-induced synapse formation in cultured embryonic (E18) rat hippocampal neurons (Lim et al., 2013). IL-4, another anti-inflammatory cytokine, has been implicated in learning and memory, as IL-4 KO mice exhibited cognitive deficits which could be rescued by transferring WT T-cells (Derecki et al., 2010). Prematurely high levels of IL-4 in the mouse brain caused reduced microglia-mediated pruning of whole granule cells (GCs), excessive connectivity from parallel fibers onto Purkinje cells (PCs), and an attention-deficit/hyperactivity disorder (ADHD)-like phenotype (Guedes et al., 2023). IL-2 has long been known to have effects on neurons including promoting cortical, hippocampal and striatal neuron survival, extension of neurites and the activation of microglia (Hanisch and Quirion, 1995), although more detailed characterization of its contribution in neurodevelopment has been minimal (Petitto et al., 2012). And even though it was shown to be negatively correlated with ASD in GWAS datasets (Guedes et al., 2023; Liu et al., 2025), Il-2 is infrequently reported in human and animal model studies of NDDs. In addition, many chemokines besides FKN signal through microglia and have demonstrated roles not only in brain development but also in neuropathology [reviewed in (Watson et al., 2020)].

The cytokine and chemokine landscape in neurodevelopmental processes is vast and complex (Figure 1). Besides these signaling molecules having multiple cellular targets, many of them influence the expression of an array of other cytokines and chemokines, creating a network of inter-cellular interactions which can be tricky to untangle experimentally. Disease models and animal KOs offer a window into these processes, however, understanding how these crucial immune messengers influence neuronal development, circuit formation and synaptic refinement in the healthy brain requires further investigation.

Immune recognition and phagocytic signaling pathways

Among the immune-related pathways implicated in healthy brain development, innate immune receptors and other membrane associated signaling molecules have emerged as important regulators in neurodevelopment. Toll-like receptors (TLRs), Major Histocompatibility Class I (MHC-I) molecules, phosphatidylserine receptors and purinergic receptors are all required to sense pathogen presence and damage, and to launch a proper immune response to protect the host (Salvador, 2022; Qian et al., 2026; Liu et al., 2023; Eberhardt et al., 2022). However, studies in neurodevelopment and in neurodevelopmental disorders have also revealed their involvement in healthy brain development, indicating their roles to be more pleiotropic than initially thought.

Major histocompatibility class I molecules

MHC-I molecules are generally known as the antigen presenting receptors found on all nucleated cells (Liu et al., 2023). They present intracellular peptides to be detected by patrolling immune cells. In turn, non-self proteins, such as viral antigens, are detected on the surface of infected cells and tagged for degradation. However, numerous studies have noted a function of MHC-I proteins in neuronal function through the modulation of neurogenesis, synaptogenesis, and synaptic plasticity during development (Corriveau et al., 1998; Huh et al., 2000; Nelson et al., 2013). Their role in the nervous system was first discovered in 1998, when Corriveau et al. (1998) noted a positive spatial and temporal correlation of MHC-I expression in the dorsal lateral geniculate nucleus (dLGN) and hippocampus with levels of neuronal activity. In addition, inhibition of spontaneous activity led to lowered expression of MHC-I on the cell surface. A subsequent study noted that knock-out of MHC-I led to a decreased segregation of contralateral and ipsilateral retinogeniculate projections to the dLGN and enhanced LTP in the hippocampus (Huh et al., 2000). Loss of neuronal cell surface expression of MHC-I molecules also led to a defect in long-term depression (LTD) in mouse hippocampal slices. It was subsequently found that MHC-I KO mice exhibited an increase in the size of the presynaptic boutons relative to WT littermates and a concurrent increase in mEPSC frequencies in hippocampal neurons (Goddard et al., 2007). Conversely, elevated levels of MHC-I in transgenic mouse neurons led to significant reductions in the presynaptic markers synaptophysin and GAP43 in the dentate gyrus and CA3 area and a modest decrease in the number of pyramidal neurons in the hippocampal CA1 region compared to WT mice (Wu et al., 2011). As MHC-I molecules are known for recruitment of immune response components, their correlation with activity may indicate their roles in microglial recruitment and identification of specific synapses for refinement of the circuit during development. Moreover, MHC-I expression can be regulated by cytokines (Neumann et al., 1997), further strengthening their roles as developmental modulators of neuronal circuits.

Toll-like receptors

TLRs were named after their structural similarity to the Drosophila Toll receptors [reviewed in (Valanne et al., 2022, 2011)]. They were originally identified due to their effect on embryonic development, but also mediate immune responses in the fly. In mammals they are involved in innate immunity, sensing components of infectious pathogens, including bacterial wall and foreign DNA (Verma et al., 2026). Many studies now note that TLRs do not only function in immune defense but also participate in neurodevelopment.

Different TLR proteins are expressed in mammalian microglia, neurons, astrocytes and oligodendrocytes (Hanke and Kielian, 2011) and expression levels in neurons vary throughout development (Abarca-Merlin et al., 2024). Several TLRs have been implicated in neurogenesis (Barak et al., 2014). TLR2 KO mice exhibit a decrease in NPC proliferation in the dentate gyrus at P7 and a significant decrease in the number and length of neurites of TLR2 KO neurons in culture compared to WT (Rolls et al., 2007). TLR3 appears to have an opposite effect, with NPCs from TLR3 KO mice capable of forming a significantly higher number of neurospheres, and having more proliferating cells in the dorsal telencephalon at E13.5 than WT controls (Lathia et al., 2008). In addition, TLR3 KO mice exhibit increased neurogenesis in the hippocampus and deficits in learning and memory (Okun et al., 2010). Young TLR4 KO mice also show increased proliferation in the hippocampus, suggesting a role for this receptor as a negative regulator of neurogenesis (Connolly et al., 2020). In addition, TLR4 is a negative regulator of retinal progenitor cells (RPCs), since a deficiency in TLR4 results in increased proliferation and neuronal differentiation in the early postnatal mouse retina, whereas activation of TLR4 restricts RPC proliferation and differentiation in vitro (Shechter et al., 2008). TLR5 is able to enhance neural differentiation of mouse embryonic stem cells in vitro, and promotes differentiation of NSCs in the dentate gyrus (Seong et al., 2022).

TLR7 and TLR8 have overlapping but distinct roles in axon and dendritic morphology and synaptic organization (Hung et al., 2018). TLR8 activation induces dendritic pruning but has no obvious effect on axonal growth (Hung et al., 2018), whereas TLR7 influences arbor elaboration, since Tlr7 knockdown leads to longer dendrites in cultured mouse cortical neurons (Liu et al., 2013). TLR4 signaling through Hevin, a synaptogenic protein secreted by astrocytes, is also involved in the process of thalamocortical synapse pruning in young (P15) mice (Ramirez et al., 2026). TLR activation in glia, particularly TLR10 in microglia (Verma and Kim, 2016) and TLR3 in astrocytes (Hung et al., 2018; Farina et al., 2005), has also been shown to promote expression of a number of cytokines and chemokines, including IL-6, TNF-α and IL-10. All these studies contribute to the better understanding of the roles of TLRs in the mammalian brain, where they regulate the proliferation and differentiation of NPCs into neurons or glial cells. They also control cell death and survival, neurite growth, synaptogenesis, and the modulation of spine density and size through communication with glial cells, including microglia.

Purinergic receptors

P2Y12 receptors (P2Y12R) are purinergic receptors that have well-described roles in microglial function in neuroinflammation (Gomez Morillas et al., 2021). Their expression is a critical marker differentiating microglia from macrophages in the nervous system (Mildner et al., 2017) and is present throughout life, although expression levels vary over time (Mildner et al., 2017; Crain et al., 2009; Peng L. et al., 2019). They are activated by interaction with adenosine diphosphate (ADP), which causes microglial activation including extension of processes and phagocytic activity (Illes et al., 2021; Ohsawa et al., 2010). An extensive review of the many roles of P2Y12 receptors in microglia has been published previously (Lin et al., 2021), but here we highlight the main functions related to neurodevelopment.

Purinergic signaling, including via P2Y12, has been implicated in regulation of neurogenesis and brain development (Rodrigues et al., 2019). Microglial P2Y12 signaling has been shown to regulate neurogenesis and neuronal projections (Mo et al., 2019), as well as to promote proliferation of adult mouse subventricular zone cells (Suyama et al., 2012). KO of P2Y12 receptors in microglia leads to aberrant formation of cortical layers, suggesting a role for microglia in neuronal migration (Cserép et al., 2022). P2Y12 signaling is additionally involved in activity-dependent remodeling of neuronal networks, as microglia rapidly respond to experience-dependent changes in neuronal activity, such as monocular deprivation, and that response is abrogated in the absence of P2Y12R activity (Sipe et al., 2016).

P2Y12Rs also contribute to the microglial response to synaptic activity (Abbracchio et al., 2009). Microglia directly interact with the cell bodies of pyramidal neurons through P2Y12R-rich contacts called somatic microglial junctions (Cserép et al., 2020; Posfai et al., 2026). These junctions form and re-form in a dynamic way during development and have been detected in the ventricular zone/subventricular zone of E15 mice (Cserép et al., 2022). The loss of P2Y12R receptors in microglia leads to increased excitability of neurons in the hippocampus and increased anxiety-like behavior in mice (Peng L. et al., 2019). Both microglial depletion and genetic deletion of P2Y12R lead to increased baseline and evoked firing rates in narrow-spiking interneuron populations in the mouse barrel cortex (Kiraly et al., 2026). Moreover, mutations in the locus encoding P2Y12 receptors are observed in some epileptic patients (Wang et al., 2022), suggesting a neurodevelopmental role for these receptors in humans (Kiraly et al., 2026). Taken together, the current data point to a conserved role of P2Y12 receptors in microglial mobilization for neuronal survival, migration and refinement, where deficits lead to aberrant states caused by miswired circuits.

Phosphatidylserine

Phosphatidylserine (PS) is a type of anionic phospholipid that is normally constrained in healthy cells to the intracellular leaflet of the plasma membrane bilayer through the activity of specific lipid transporter enzymes called “flippases” (Peet et al., 2020). It can be “flopped” to the outer membrane of the cell during cellular injury or under conditions where flippase activity is reduced (e.g., in apoptosis). In addition, PS can be actively externalized by “floppases” and bidirectional “scramblases” such as Xk-related protein 8 (Xkr8) (Neniskyte et al., 2023). Exposed extracellular PS serves as a canonical “eat me” cue that engages engulfment or pruning by phagocytic cells, such as microglia, which express diverse PS receptors (Figure 2). Furthermore, because of the high energetic demands of flippase activity, ATP depletion resulting from reduction or impairment of mitochondrial activity can also result in externalization of PS and increase phagocytic pruning (Shacham-Silverberg et al., 2018). Therefore, failure of proper PS metabolism and maintenance has been associated with a broad range of neurodevelopmental disorders, including ASD (Ma et al., 2022; Li et al., 2024).

One example of a flippase implicated in NDD is the ATPase Phospholipid Transporting 8A1 (ATP8A1), a member of the P4-ATPase family. Elevated expression levels of ATP8A1 have been reported in adolescent ASD patients (Kerr et al., 2016). The gene encoding ATP8A1 is located in band 4p13, a locus associated with brain development and GABAergic synaptic function, and which is genetically linked to ASD and other NDDs (Polan et al., 2014). As ATP8A1 removes PS from the outer leaflet, it could potentially prevent PS signaling for synaptic pruning by glia (Raiders et al., 2021). Indeed, lentiviral overexpression of ATP8A1 in hippocampus does result in a reduction in synapse number and strength, however ATP8A1 KO mice appear to exhibit a similar phenotype (Kerr et al., 2016).

Conversely, flippase loss-of-function can result in excessive pruning of neurons due to increased exposure of PS in the outer membrane leaflet. CDC50A is a chaperone protein for phospholipid flippases such as ATP11A (Li et al., 2021). In the developing nervous system, its expression is positively regulated by neural activity, where it limits PS exposure at the cell surface. Knockdown of CDC50A results in increased engulfment of presynaptic axon terminals by microglia and induces synapse elimination in the retinogeniculate projection, consistent with PS acting to promote synaptic phagocytosis in an activity regulated manner. Interestingly, postsynaptic KO of CDC50A was found to selectively promote elimination of inhibitory synapses, an effect that was prevented in mice in which the microglial phagocytic receptor MERTK, a member of the TAM (Tyro-3, Axl, Mertk) family of receptor tyrosine kinases, was knocked out (Park et al., 2021).

Another well-characterized microglial PS receptor is the S4 splice variant of the adhesion G protein-coupled receptor GPR56, a PS-binding protein that is expressed at high levels in microglia during development (Li et al., 2020). GPR56 KO mice exhibit reduced microglial engulfment of PS, along with increased synapse densities in the dLGN and hippocampus. Furthermore, segregation of RGC inputs in the dLGN at P10 is impaired in GPR56 KO mice.

Initially identified through its association with neurodegenerative disorders, including Alzheimer's (Guerreiro and Hardy, 2013) and Parkinson's (Colonna and Butovsky, 2017) diseases, Triggering Receptor Expressed on Myeloid cells 2 (TREM2) is another PS receptor that contributes to developmental synapse elimination by microglia (Filipello et al., 2018; Scott-Hewitt et al., 2020). TREM2 KO mice reveal higher densities of excitatory synaptic markers, as well as increased mEPSC frequency, in the hippocampus at P18-20. WT hippocampal neurons co-cultured with microglia for 24 h results in a microglia-mediated reduction in synapse number. However, this reduction was abrogated when microglia from TREM2-lacking mice were used. In a mouse model of childhood adversity, it was found that deprivation by limited bedding reduced TREM2 levels which contributed to impaired microglial pruning in the hippocampus (Ahmed et al., 2026).

Other PS receptors may be required for PS recognition and participate in phagosome formation and stabilization. A study of microglial neuronal engulfment in embryonic zebrafish noted that the PS receptors BAI-1 and TIM-4, actively participate in the mechanics of phagocytosis (Mazaheri et al., 2014). Morpholino knockdown of BAI-1 led to a doubling of the time required for the formation of phagosomes by the microglia, while the KO of TIM-4 led to the formation of unstable phagosomes, prone to losing their cargo, in turn preventing effective responses to PS displayed on neurons.

Reactive oxygen species

Reactive oxygen is produced as a byproduct of oxidative phosphorylation, and is often regarded as a harmful, damaging radical molecule associated with pathological states. Many neurodegenerative disorders involve an increase in ROS, which in turn is associated with oxidative stress and cellular damage (Oswald et al., 2018a). ROS is often generated as a response to neuroinflammation. ROS can also activate the inflammasome, a cytoplasmic multiprotein complex that increases inflammatory response (Yang et al., 2013).

However, ROS is also an important physiological signaling molecule in neurodevelopment. NSCs are capable of maintaining high ROS levels intracellularly, which they use as a second messenger to promote proliferation over quiescence (Le Belle et al., 2011). Hippocampal neurons also have been found to express NADPH oxidases and NOX2, which generate ROS during development. Inhibition of ROS production leads to delayed growth and reduced axon lengths (Wilson et al., 2015). Concentrations of ROS appear to reflect activity levels in neurons (Bindokas et al., 1996), and studies noted that submerging hippocampal slices in superoxide scavengers can prevent NMDAR-mediated long-term potentiation by preventing appropriate modulation of Protein Kinase C activity (Klann et al., 1998). Conversely, an increase in ROS is sufficient to inhibit both evoked and spontaneous synaptic vesicle fusion and neurotransmitter release at the mouse and frog neuromuscular junctions through its impairment of SNAP25 (Giniatullin et al., 2006).

Activity-dependent synaptic refinement in the Drosophila model has also been demonstrated to depend on ROS generation. Excess excitation of motor neurons was found to promote mitochondrial metabolic ROS-activated stress signaling to alter neuromuscular junction architecture both pre- and postsynaptically (Oswald et al., 2018b; Peng J. J. et al., 2019).

Caspase-3

Caspases are a family of cysteine-dependent, aspartate-directed proteases that are responsible for regulating programmed cell death (apoptosis) and certain developmental processes. Many apoptotic pathways converge onto caspase-3 and its activation is often recognized as the “point of no return” for cell death (Porter et al., 1999). Caspase-3 function has also been implicated in neuronal development and synaptic refinement. In the developing retinotectal system of Xenopus tadpoles, it was demonstrated that plasticity-inducing visual stimulation could drive caspase activation, leading to the rapid degradation of MEF2, which in turn altered both visual response properties and dendritic remodeling by tectal neurons (Chen et al., 2012). This result is consistent with demonstrations in mouse hippocampus that NMDAR-dependent LTD (but not LTP) requires mitochondrial activation of caspase-3 (Li Z. et al., 2010). In the developing mouse retinogeniculate projection, caspase-3 levels were found to be upregulated in postsynaptic dLGN neurons when the retinal afferents from one eye were silenced by monocular viral expression of tetanus toxin, but not when both eyes were silenced (Yu et al., 2025). In Casp3 KO mice, activity-dependent anatomical segregation of retinal afferents into eye-specific laminae in the dLGN is disrupted, along with a failure to prune the number of functional inputs onto individual dLGN neurons measured by whole-cell recording. Moreover, microglial engulfment of axon terminals is significantly reduced in the Casp3 KO mice.

In another study microglia were co-culltured with hippocampal neurons expressing hM3Dq, a designer receptor exclusively activated by designer drugs (DREADD), which increases neural activity upon addition of clozapine-N-oxide (Andoh et al., 2025). Using immunocytochemistry for cleaved caspase-3 as well as a fluorescent reporter of caspase-3 activity, they observed increased caspase-3 activation in axons and increased microglial engulfment of presynaptic material. Interestingly, decreased activity via inhibitory DREADD expression did not affect either caspase-3 activation or microglial engulfment compared to controls, indicating that caspase activity is not perfectly coupled with neuronal activity. Furthermore, the study showed by fluorescent labeling of PS and complement C1q protein that elevating caspase-3 activity promoted the targeting of these “eat me” signals to presynaptic sites. This finding supports reports that caspase-mediated inactivation of flippases may underlie PS externalization at the plasma membrane (Mandal et al., 2002).

Complement system in neurodevelopment

The complement system consists of a series of extracellular proteins that aids in clearing pathogens and unwanted cellular debris as a part of innate immunity (Taylor et al., 1998). Beyond the well-characterized functions in innate immunity, roles for the complement system have been identified in neurodevelopment, including in neurogenesis, neuronal migration and circuit wiring refinement (Magdalon et al., 2020). Most of the protein components of the complement system are expressed locally in the brain during embryonic and early postnatal development (Zhang Y. et al., 2025). To date, complement C3, C4 and C1q have been most extensively associated with neurodevelopmental functions, with other proteins also implicated in neuronal circuit formation programs [reviewed in (Magdalon et al., 2020)].

The complement system has 3 different ways of activation, distinguished by the initiating stimulus: the classical pathway, the lectin pathway and the alternative pathway (Taylor et al., 1998; Sarma and Ward, 2011). In the classical pathway, C1q, as part of the C1 complex (C1qr₂s₂), recognizes its target molecule, which can include apoptotic bodies, cellular blebs and antibody-antigen complexes. Following tagging by C1q, the c1r and c1s serine proteases are activated. Activated C1s cleaves C4 and C2, in turn leading to the formation of the C3 convertase C4b2b which cleaves and activates C3, the central effector for complement clearance through all three pathways (Sarma and Ward, 2011). While the classical pathway of the complement system has been implicated in neurodevelopmental processes, for example by producing “eat me” signals for synaptic pruning (Taylor et al., 1998; Sarma and Ward, 2011) (Figure 2), the roles of the lectin and alternative pathways in brain development are less clearly defined and may not be involved.

C1q expression in the brain is punctate and localized to synapses during development (Stevens et al., 2007). Mice deficient for C1q show a significant impairment in synapse elimination, with a higher number of excitatory synaptic connections observed in C1q KO mice (Stevens et al., 2007; Chu et al., 2010). C1q was recently found to recognize and bind to PS, a component of the cellular membrane normally exposed on apoptotic cells that is also found at synapses, suggesting a synergistic mechanism by which synapses could be targeted for pruning by both the PS and complement pathways (Scott-Hewitt et al., 2020; Andoh et al., 2025).

The roles of C3 in neurodevelopment include NPC proliferation, neuronal migration and synaptic pruning (Magdalon et al., 2020). Genetic KO of C3 leads to an increased pool of NPCs in the subventricular zone in early development (Gorelik et al., 2017a). Pharmacological block and genetic deletion of the C3a receptor (C3aR) resulted in opposing phenotypes in the mouse ventricular zone, with a surprising decrease in proliferation in C3aR KO animals compared to WT mice (Coulthard et al., 2018). Therefore, the role of C3 signaling in neuronal proliferation requires further investigation. Deletion of C3 leads to deficits in neuronal migration in the developing mouse cortex, a phenotype that could be rescued by expression of C3a peptide mimics that bind to the C3a receptor (Gorelik et al., 2017a). The functions of C3 in neurodevelopment also involve the complement protein SERPING1, a C1 inhibitor, since its deletion leads to a decrease in NSC proliferation and impeded neuronal migration (Gorelik et al., 2017b).

Many studies have also noted a function of C3 in synaptic pruning, with its loss leading to aberrant pruning in diverse animal models (Lim and Ruthazer, 2021; Schafer et al., 2012). In the developing Xenopus retinotectal system, overexpression of C3 targeted to the presynaptic membrane of retinal ganglion cells resulted in significantly reduced axon arbor length and axon branch complexity compared to control axons (Lim and Ruthazer, 2021). Genetic KO of C3 has the interesting effect of improving place and reversal learning compared to control mice, while reducing probability of release at glutamatergic synapses in the hippocampus (Perez-Alcazar et al., 2014). Non-canonical roles for C3 and C1q include influencing the release of glutamate from astrocytes and neurons and regulating the function of glutamatergic receptor subtypes, including NMDA receptors, AMPA/kainate receptors, and metabotropic glutamate receptors in the plasma membrane (Pittaluga et al., 2025).

Further support for a role of complement pathway components in neurodevelopment comes from evidence of abnormal expression levels and genetic risk in NDDs and psychiatric disorders (Sierra et al., 2022). For instance, GWAS strongly links schizophrenia risk with the MHC locus, where structural variation and copy number diversity drive altered expression levels of C4 (Sekar et al., 2016). In a mouse KO of C4, the segregation of eye-specific laminae in the dLGN was significantly reduced, consistent with impaired synaptic pruning. Additionally, increased levels of C4a, a C4 protein fragment released upon C4 protein activation, have been observed in plasma from schizophrenia patients (Kalinowski et al., 2021). Moreover, iPSC-derived astrocytes from ASD patients showed decreased C4 mRNA and protein expression compared to healthy controls (Mansur et al., 2021). This aligns with the hypothesis that schizophrenia involves excessive synaptic elimination, whereas many forms of ASD, including Fragile X syndrome, are characterized by impaired or insufficient pruning (Faust et al., 2021).

“Don't eat me” signals

Developmental refinement of neural circuits by synaptic pruning involves the selective sparing of a population of appropriate contacts while targeting inappropriate synapses for elimination by microglia. In contrast to the cell-autonomous regulation of PS localization by membrane-localized flippase and scramblase activity, the pathways leading to microglial activation and target opsonization via complement signaling take place largely in the extracellular space. Thus, positive and negative cues that can distinguish between potential targets of complement-mediated phagocytosis should be present during development and be subject to activity-dependent regulation by neurons. One such candidate is CD47, which is expressed by developing neurons and which inhibits phagocytosis by binding its receptor on microglia, CD47-Signal Regulatory Protein α (SIRPα) receptor (Lehrman et al., 2018). Mice lacking CD47 exhibit increased engulfment of retinogeniculate axon terminals by microglia, as well as reduced synapse numbers in the dLGN. In the developing Xenopus retinotectal projection, complement signaling has been shown to mediate synaptic pruning and RGC arbor remodeling (Lim and Ruthazer, 2021). Xenopus RGCs express a complement C3-interacting protein called Amphibian Regulator of Complement Activation 3 (aRCA3), a putative homolog of mammalian CD46. Overexpression of aRCA3 in RGCs significantly reduces axonal engulfment by microglia and results in axonal arbors that grow to be larger and more extensively branched than controls.

Another promising complement regulatory protein expressed in neurons is CUB and Sushi Multiple Domains 1 (CSMD1) (Baum et al., 2024). CSMD1 is genetically associated with schizophrenia and human stem cell-derived neurons lacking CSMD1 are subject to increased deposition of complement C3. Synaptosomes derived from CSMD1 KO mice are preferentially phagocytosed by microglia in vitro compared to WT synaptosomes. Another sushi domain protein, sushi repeat protein X-linked 2 (SRPX2) is a neuronally expressed and secreted complement inhibitor (Cong et al., 2020). SRPX2 directly binds to C1q and blocks its activity, resulting in reduced synapse engulfment by microglia. SRPX2 KO mice have reduced numbers of synapses in the dLGN and somatosensory cortex, consistent with a role in protecting synapses from elimination.

Complement regulatory factors appear to play a key role in protecting certain synapses from pruning in development. As such, they also hold tremendous potential as therapeutic agents to prevent cellular degeneration and synapse loss in disease. Complement receptor 1-related protein Y (CRRY), a complement C3 inhibitor has been found to preserve visual function when expressed in the eye to limit RGC injury in a model of glaucoma (Bosco et al., 2018) and in experimental autoimmune encephalomyelitis (EAE), a model of multiple sclerosis (Werneburg et al., 2020).

Taken together, these findings indicate that complement proteins function in the proper formation and establishment of neuronal circuits by regulating neuronal precursor cell proliferation, neuronal migration, synaptic pruning and activity during development. A better understanding of the mechanisms by which complement signaling targets specific synapses to direct anatomical pruning and functional refinement of developing circuits is likely to provide valuable insights into the etiology of NDDs.

Conclusions

The formation, fine-tuning, and function of the central nervous system rely extensively on signaling and direct interactions from the innate immune system. This review has outlined just a fraction of the numerous molecular cues and structural interventions that immune mechanisms exert on the developing brain. Notably, pruning of synapses by microglia during circuit refinement has emerged as a fundamental process for healthy neurodevelopment. Genetic linkage studies further underscore this connection, strongly implicating aberrant immune signaling in a range of NDDs and psychiatric illnesses, including ASD and schizophrenia. Elucidating these precise neuroimmune pathways will be invaluable for uncovering new therapeutic targets and advancing our understanding of nervous system development.

Generative AI statement

The author(s) declared that Generative AI was used in the creation of this manuscript. Generative AI was used to proofread the final draft of the text for grammar and for verification of accuracy.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher's note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Footnotes

Edited by: Yael Mandelblat-Cerf, National Institute of Mental Health (NIH), United States

Reviewed by: Steven Ngandeu Schepanski, Charité University Medicine Berlin, Germany

Janay Vacharasin, Francis Marion University, United States

Daria Kveštak, University of Rijeka, Croatia