Consequences of oxygen deprivation on myelination and sex-dependent alterations

Abstract

Oxygen deprivation is one of the main causes of morbidity and mortality in newborns, occurring with a higher prevalence in preterm infants, reaching 20 % to 50 % mortality in newborns in the perinatal period. When they survive, 25 % exhibit neuropsychological pathologies, such as learning difficulties, epilepsy, and cerebral palsy. White matter injury is one of the main features found in oxygen deprivation injury, which can lead to long-term functional impairments, including cognitive delay and motor deficits. The myelin sheath accounts for much of the white matter in the brain by surrounding axons and enabling the efficient conduction of action potentials. Mature oligodendrocytes, which synthesize and maintain myelination, also comprise a significant proportion of the brain's white matter. In recent years, oligodendrocytes and the myelination process have become potential therapeutic targets to minimize the effects of oxygen deprivation on the central nervous system. Moreover, evidence indicate that neuroinflammation and apoptotic pathways activated during oxygen deprivation may be influenced by sexual dimorphism. To summarize the most recent research about the impact of sexual dimorphism on the neuroinflammatory state and white matter injury after oxygen deprivation, this review presents an overview of the oligodendrocyte lineage development and myelination, the impact of oxygen deprivation and neuroinflammation on oligodendrocytes in neurodevelopmental disorders, and recent reports about sexual dimorphism regarding the neuroinflammation and white matter injury after neonatal oxygen deprivation.

1. Introduction

There is a critical window of neurodevelopment in the perinatal period, when multiple important events take place in brain that are sensitive to the brain tissue oxygenation (Cisneros-Franco et al., 2020; Ujhazy et al., 2013; Failor et al., 2010). Oxygen deprivation (OD) is among the main causes of neonatal morbidity, and affects 0.1 to 0.8 % of term-born babies and 60 % of preterm babies (MacDonald et al., 1980; Kurinczuk et al., 2010). Individuals who survive OD often develop permanent neurological sequelae, such as cognitive and motor deficits (Back and Miller, 2014; Vannucci, 2000; Wilson-Costello et al., 2005), and psychiatric disorders, such as autism spectrum disorder and schizophrenia (Giannopoulou et al., 2018; van Handel et al., 2007).

Among the processes disrupted by perinatal OD, myelination is one of the most prevalent, and white matter injury is a common finding in human post-mortem brains, as well as in live magnetic resonance images. Myelination is a prolonged neurodevelopmental process that ensures the efficiency and speed of action potentials by insulating axons with a specialized membrane compounded of 70 % lipids and 30 % proteins (van Tilborg et al., 2017). The myelination of axons is essential for the maturation of the central nervous system (CNS) and depends on glial cells called oligodendrocytes (OLs). Moreover, OLs support the maturation of neurons by the production of trophic factors, such as glial cell-derived neurotrophic factor (GDNF), brain-derived neurotrophic factor (BDNF), or insulin-like growth factor-1 (IGF-1) (Jang et al., 2019; Schmidt and Knosche, 2019; Wilkins, 2003; Zhao et al., 2001).

The development of the OL lineage is modulated by delicate interactions between CNS cells, the temporal production of trophic factors, signaling molecules, vesicular transport, energy input, as well as the control of intracellular transport, which shows a complex interaction between the axon, environment, and myelinating cell (Nishiyama et al., 2021; Santos et al., 2019). The impact of sexual dimorphism on the production and maturation of OLs has received less attention. However, some studies have observed sex-dependent differences in white matter as well as in the volume of white matter in specific structures, such as the corpus callosum (Cerghet et al., 2009; Fitch et al., 1990).

This review summarizes the latest research about the interaction between the neuroinflammatory status caused by OD in neonates and the associated white matter injury, with a special foccus on the sex-dependent consequences on myelination. To achieve this aim, we discuss (1) the physiological development of the oligodendrocyte lineage, (2) the myelination process, (3) how neonatal OD and neuroinflammation can interact at different stages of myelination, (4) the impact of neuroinflammation on myelination, (5) the common features of the neurodevelopmental disorders due to white matter injury and neonatal OD, and, lastly, (6) the recent reports about the sexual dimorphism in the myelination and neuroinflammation processes after neonatal OD events.

2. Development of oligodendrocytes

OLs arise from the oligodendrocyte progenitor cells (OPCs) that exhibit short processes emanating from opposite poles of the soma and bipolar morphological characteristics of neural precursor cells (Baldassarro et al., 2019; Sharma et al., 2020). OPCs emerge from three restricted periventricular germinative regions: the medial ganglionic eminence and entopeduncular area, the lateral ganglionic eminence, and the dorsal ventricular zone. The development of the OLs in the forebrain is characterized by three waves of OPC production. The first occurs at embryonic day (E) 12.5 in the ventricular zone of the ventral medial ganglionic eminence and entopeduncular area. The second wave occurs through precursors in the lateral and/or caudal ganglionic eminences at around E14.5. Finally, the third wave starts on postnatal day 0 (P0) with the migration of progenitors from the dorsal ventricular zone. In adulthood, OLs are evenly distributed throughout the adult CNS, so OPCs exhibit multidirectional migration from ventricular zones to distant locations (Kessaris et al., 2006).

Signaling molecules such as sonic hedgehog, bone morphogenic protein, and Notch orchestrate the migration of the OPCs to their final locations (Hashimoto et al., 2018; Winkler and Franco, 2019). Once allocated, OPCs undergo local proliferation and differentiation, or apoptosis, in response to diffusible growth and transcription factors, such as Olig1, Olig2, SOX10, and Nkx2.2 (Liu et al., 2007; Qi et al., 2001; Zhou and Anderson, 2002; Zhou et al., 2001; Zhu et al., 2014). Also, the expression of tenascin-C - an extracellular matrix glycoprotein - by the OPCs may be associated to the PDGF signaling to regulate the maturation and survival of these cells (Garwood et al., 2004).

The OPCs can be identified by the expression of markers such as the platelet-derived growth factor-alpha receptors (PDGFRα), the GD3 ganglioside, the NG2, and the Nkx2.2 (Calver et al., 1998; Fruttiger et al., 1999; Gotoh et al., 2018; Liu et al., 2007). These markers play a continued role in the generation of OL throughout life, until these cells differentiate and mature into myelin-forming OL. Also, a recent study showed that Rnd3, a member of the subfamily of Rho GTPases, is expressed in a subset of OPCs and of mature OLs both in vivo and in vitro and is directly required for the differentiation of OLs (Madrigal et al., 2022). However, it is not clear if all OPCs undergo this same sequence of events. As OPCs differentiate they reduce their mitotic ability and gain the expression of sulfatides recognized by the O4 antibody and lose PDGFRα, GD3, and NG2 as they start expressing GalC (Pituch et al., 2015; Scott-Hewitt et al., 2017).

As OPCs mature and establish contact with target axons, OPCs exit the cell cycle and develop into pre-myelinating OLs, which begin to express myelin-related gene products, before axon ensheathment and myelination. CNPase or CNP is one of the earliest OL-specific proteins expressed, and its two isoforms are temporally regulated, with CNP2 being expressed by OPCs, and both CNP1 and CNP2 being expressed at the time of OL differentiation (Lee et al., 2006; Scherer et al., 1994).

This is followed by the expression of the major constituents of CNS myelin, the myelin basic protein (MBP) and proteolipid protein (PLP), which determines that pre-myelinating OLs reach the mature stage (Ray et al., 2003). Myelin-associated/oligodendrocyte basic proteins are expressed at the late stages of myelination, significantly later than the expression of MBP. Myelin-oligodendrocyte glycoprotein (MOG) is one of the last myelin proteins to be expressed and is often used as a marker for mature OLs (Scolding et al., 1989).

These multiple markers determine OLs functions at different stages of differentiation (Fig. 1). PDGFRα mediates the effects of platelet-derived growth factor with two A subunits (PDGF-AA), which is a potent mitogen and survival factor for OPCs and stimulates migration. The loss of PDGFRα culminates in a decrease in the migratory and proliferative capacity of OPCs and is dependent upon axon-derived factors for survival and differentiation (Deng et al., 2014). Fibroblast growth factor (FGF) signaling and the expression of FGF receptors (FGF-R) are developmentally regulated in OLs: both FGF-R1 and FGF-R2 are required for OPCs ventral generation in the forebrain; FGF-R2 acts as a late-stage regulatory mechanism of myelin growth and myelin sheath thickness; FGF-R3 is expressed transiently by pre-myelinating OLs and is important in the initiation of myelination (Furusho et al., 2012; Furusho et al., 2020; Furusho et al., 2011). Moreover, IGF-I is a survival factor for OLs and coordinates cell cycle progression of OPCs (Frederick and Wood, 2004; Masters et al., 1991; Palacios et al., 2007). However, the precise interactions of these trophic factors in vivo are unclear. When OPCs reach their final destinations, the contact with axons is sufficient for survival and differentiation, under the influence of neuregulin-1 (NRG-1) and Notch/Jagged signaling (Fernandez et al., 2000; Linying et al., 2014; Santhosh et al., 2017).

Fig. 1.

Markers of oligodendrocyte developmental stages. Oligodendrocyte progenitor cells mark the initial stages of differentiation. In this stage, cells can be identified by the expression of NG2, PDGFRα, NKx2.2, CNP2 and GD3. Pre-myelinating period is characterized by the initial expression of mature markers, such as MBP, MAG, CNP1, O4, O1 and GalC. Myelinating period is characterized by the expression of MOG. Olig 1 and 2 are expressed throughout the lineage and support the maturation process. NG2: neuron-glial antigen 2; PDGFαR: platelet-derived growth factor receptor A; Nkx2.2: NK2 homeobox 2; CNP2: CNPase 2; CNP1: CNPase 1; Olig 1: oligodendrocyte transcription factor 1; Olig 2: oligodendrocyte transcription factor 2; GD3: GD3 ganglioside; MBP: myelin basic protein; MAG: myelin associated glycoprotein; MOG: myelin-oligodendrocyte glycoprotein; GalC: galactocerebroside.

3. Myelination

Over the years, many models have been proposed to explain how OLs drastically change their plasma membranes to form internodes and myelinated axons. The “jelly roll or carpet crawler” model proposed for Bunge was inherited from the behavior of Schwann cells in the peripheral nervous system (PNS), According to this model, OLs project their membrane as an extension conformed with the shape of paranode of the axon to be enveloped and repeatedly encircle it (Bunge et al., 1961; Bunge et al., 1989). However, CNS myelination differs from the PNS by the multiple axons that an OL myelinates when compared to a single axon myelinated by a Schwann cell, and by the development of the internodes, of variable length conflicting with this model (Knobler et al., 1976). Subsequent models were proposed to reconcile these divergences, such as a “liquid croissant” (Sobottka et al., 2011).

However, with the advance of in vivo live-imaging techniques, a more suitable and current model of myelination in the CNS was proposed using zebrafish. This model argues that OLs project a growth cone toward the axon in a triangle-like shape, and the inner tongue repeatedly encircles the axon underneath the previous layer. In parallel, a process of lateral extension and compaction of the layers in the direction of the paranodes is coordinated by the interaction of the cytoplasmic channels and PI(3,4,5)P3 lipid levels (Snaidero et al., 2014).

Histological human post-mortem studies and recent in vivo MRI show that myelination follows a hierarchical sequence, in which brain regions dedicated to homeostatic and less complex functions are myelinated first, whereas those involved in higher cognitive functions, such as prefrontal cortex, are myelinated last (Brody et al., 1987; Deoni et al., 2011; Kinney et al., 1988). Moreover, myelination in the CNS is heterogeneous, not only between different areas but also in the same brain region. For example, pyramidal neurons of the somatosensory and visual cortices present different profiles of longitudinal myelinations along cortical layers (Tomassy et al., 2014). Until recently, it has been unclear how axons are selected to be enveloped, but physical, biophysical, and molecular cues seem to be involved in this complex process.

Considering physical aspects, CNS myelination is explained by axon geometries. The caliber of the axons seems to be determinant, since OLs do not wrap axons with a diameter smaller than 0.3 μm, and the thickness of the myelin sheath seems to be correlated to the axon caliber and internode length (Bechler et al., 2015). Cell surface molecules are also involved in myelination guidance. In neurons that should not be myelinated, inhibitory molecules as PSA-NCAM (Charles et al., 2000), JAM2 (Redmond et al., 2016), Lsamp, and Sema3 (Sharma et al., 2015; Syed et al., 2011) are expressed to repulse the neuron-OL interaction.

Myelination is enhanced by neuronal activity, which influences OPCs proliferation, differentiation, and myelin production in an adenosine-dependent mechanism (Gibson et al., 2014; Mitew et al., 2018; Stevens et al., 2002). Recently, a review about the activity-dependent myelination in the central nervous system showed how the advances in novel toolboxes, such as genetic, electrophysiological and the previously cited imaging approaches, led to a better understanding of how this neuron-oligodendrocyte cell interaction and myelination presents both precise temporal and spatial patterns, which can be regulated by extrinsic and intrinsic factors (Heflin and Sun, 2021).

Aligned to neuronal activity, it was demonstrated that dynorphin, released under high neuronal activity, also promotes developmental OPC differentiation and myelination, and this effect requires the expression of the dynorphin endogenous receptor, kappa opioid receptor, in OPCs (Osso et al., 2021). Also, Zonouzi and colleagues showed that certain OLs show a preference to myelinate axons from inhibitory or excitatory neurons and different classes of interneurons have different patterns of myelination in the neocortex (Zonouzi et al., 2019). This finding supports the hypothesis that different neuronal identities and electrophysiological features can influence the myelination process (Micheva et al., 2016).

Besides neurons, astrocytes and microglia are also important players in the myelination process. During myelination and early postnatal development, the astrocytic secretion of BDNF and ciliary neurotrophic factor (CTNF) potentiates myelination (Miyamoto et al., 2015; Stankoff et al., 2002). Also, the direct interaction between OLs and astrocyte connexins is important for the correct ensheathment of axons, since their deletion leads to myelination deficits, as vacuolization and a reduction in both the sheath thickness and OLs number (Tress et al., 2012). The importance of astrocytes in the adhesion of the OLs to the axons was shown using in vitro assays that manipulated PSA-NCAM in co-cultures of retinal ganglion cells and optic nerve (Meyer-Franke et al., 1999). Additionally, astrocytes support the activity of the oligodendrocytes and myelination in low-glucose conditions (Rinholm et al., 2011) as well as provide extracellular lipids that contribute to myelin membrane formation in normal development (Camargo et al., 2017). The contribution of the microglia to myelination, especially during development, is pivotal. Indeed, with the advance of transcriptome analyses, a novel population of microglia CD11c⁺ localized in the corpus callosum and cerebellum was described as the major source of IGF-1 – an important factor in the expression of myelin proteins, including PLP, MBP, MAG and MOG, and myelination (Wlodarczyk et al., 2017).

Fig. 2 illustrates the proposed myelination model for the CNS and the guidance factors involved in the myelination process.

Fig. 2.

CNS model of myelination and guidance factors of the myelination process. In the center, the current proposed model of myelination to the CNS is illustrated by an unmyelinated axon in parallel to an oligodendrocyte in three sequential steps of myelination and the respective cross-section illustration. In the first step, the triangle shape inner tongue is projected toward the axon, followed by the repeated encircle of the axon, lateral extension, and compaction of the myelin layers. Factors that guide the oligodendrocytes' decision to myelinate or not an axon are illustrated around the main image. Such factors include axon diameter, attractive or repulsive molecular cues, and neuronal activity. Moreover, the contribution of other glial cells is highlighted: astrocytes secrete BDNF and CTNF couple with oligodendrocytes by connexins; microglia CD11c + support the myelination process by the secretion of IGF-1, especially in the corpus callosum. IGF-1: insulin-like growth factor 1; BDNF: Brain-Derived Neurotrophic Factor; CTNF: ciliary neurotrophic factor.

In summary, the successful ensheathment of the axons by the membrane of OL projections is a multifactorial event, dependent on the axons' geometry, neuronal identity, adjacent glial cells, and molecules derived from the OLs and the surrounding tissue. However, there are still some gaps in the knowledge of the mechanisms underlying the process of oligodendrocyte development and myelination, which require further studies. In the next section, we discuss the impact of OD on the myelination process.

4. OD and myelination

Adequate myelination during brain development is essential for the appropriate performance of motor, cognitive, and sensory functions (Back and Miller, 2014; Branson, 2013; Volpe, 2012). Myelination begins during the gestational period and continues throughout life (Semple et al., 2013; Swire and Ffrench-Constant, 2018). Perinatal insults can result in myelin damage, with subsequent functional impairment.

OD during the perinatal period represents one of the major causes of mortality and morbidity in newborns. Among the risk factors for this condition are placental insufficiency, pre-eclampsia, anemia, cardiovascular disease, and maternal smoking. Neonates who survive OD can present permanent sequelae, such as seizures, cerebral palsy, and long-lasting cognitive and behavioral impairments (Rees and Inder, 2005). The OD causes extracellular accumulation of glutamate and other excitatory amino acids. The activation of excitatory receptors leads to a calcium (Ca²⁺) influx, which causes augmented levels of nitric oxide synthase, damage to the mitochondria, and subsequent cell death (Berger et al., 2016; Nyakas et al., 1996; Pearce, 1995). Thus, in the effort to understand the mechanisms underlying OD and to develop potential therapeutic approaches, there are different models developed in primates, sheep, piglets, and especially rodents that attempt to simulate the OD effects. Each model has particularities that are beyond the scope of this article, therefore, we will specify the OD models used throughout this section.

Although several regions of the developing brain are susceptible to hypoxic damage, OD in full-term newborns primarily affects the gray matter. On the other hand, the white matter is more affected in premature infants, and the resulting lesion is known as perinatal white matter injury (PWMI), which includes the periventricular leukomalacia (PVL) and diffuse non-cystic injury (Back, 2006).

It has already been reported the death of OPCs in the developing brain after OD in rodents and humans (Barradas et al., 2016; Kaur et al., 2013; Levison et al., 2001; Volpe, 2001). In addition to cell death, delayed differentiation in humans (Dunst et al., 2014; Wellmann et al., 2014), the interruption of OL maturation in the cortex, striatum, and hippocampus was also documented after OD (Back et al., 2001; Ness et al., 2001; Ziemka-Nalecz et al., 2018). These findings possibly corroborate with damages to myelination changes with MBP labeling, such as deformed myelin sheaths and a few myelinated axons observed in rats (Kessaris et al., 2006) and mice (Cai et al., 2011; Skoff et al., 2001),.

Among the factors involved in the aforementioned damages, the downregulation of several genes involved in OL differentiation after OD plays a fundamental role. IGF-1 is one of the main factors which stimulate the OLs proliferation, differentiation, and survival in rat brain cultures (Masters et al., 1991; McMorris and Dubois-Dalcq, 1988; McMorris et al., 1986; Mozell and McMorris, 1991) by the interaction with the PI3K/mTOR/Akt, MEK/ERK, and JAK/STAT pathways as observed in rat brains in vitro and in situ (Bibollet-Bahena and Almazan, 2009; Ness et al., 2001; Palacios et al., 2007; Subramaniam et al., 2005; Zheng et al., 2018). Changes in the expression of these genes impair myelinogenesis and result in the development of white matter diseases. For example, the reduction of IGF-1 mRNA levels observed after neonatal OD (Lee et al., 1996) and in newborns with hypoxic-ischemic encephalopathy (HIE) (Umran et al., 2016) may contribute to the vulnerability of the abnormal and deficient myelinogenesis and myelination (Ziemka-Nalecz et al., 2018). Recently, a decrease in IGF-1 levels was demonstrated in isolated cultures of OLs 72 h after oxygen and glucose deprivation. A significant decline in the number of dividing cells was seen, suggesting an autocrine effect of IGF-1 on cell division, which was impaired by hypoxia-ischemia (Janowska et al., 2020).

Furthermore, newborn brain's transcriptome analysis showed that hypoxia down-regulated OL genes encoding PLP, PDGFRα, cyclic nucleotide phosphodiesterase, and MAG, among these, MAG was the most altered. Thus, the loss of MAG expression after OD may be indicative of the loss of mature OLs (Curristin et al., 2002).

The development of OLs is also vulnerable to oxidative stress and glutamate accumulation triggered by the reperfusion after OD episodes. Nitrative and oxidative damage may be associated with a direct insult to pre-myelinating OLs (Haynes et al., 2003) by the release of free radicals induced by OD, which results in both decreased glutathione (cellular antioxidant) and promotes the accumulation of reactive oxygen species (ROS) (Haynes et al., 2003; Rathnasamy et al., 2011). Concomitantly, excess extracellular glutamate leads to an overactivation of AMPA and NMDA receptors, present in OLs, and leads to the death of these cells in the developing brain (Jensen, 2002; Salter and Fern, 2005; Sanchez-Gomez et al., 2003).

5. Neuroinflammation and myelination

Another reason for changes in the myelination process is the inflammatory response mediated by the release of cytokines, such as TNFα, IL-1β, IL-2, IL-6, IFNγ (Deguchi et al., 1997; Deng et al., 2008; Folkerth et al., 2004; Kadhim and Sebire, 2002; Kadhim et al., 2001; Kaur et al., 2013; Rathnasamy et al., 2011). Neuroinflammation presents two distinct phases (Fig. 3). The early phase is characterized by the direct activation of the microglia leading to the production of pro-inflammatory cytokines, which can also lead to reactive astrogliosis, thus contributing to the late inflammatory phase (Deguchi et al., 1997; Paschon et al., 2019).

Fig. 3.

Neuroinflammatory phases after neonatal oxygen deprivation. The schematic illustration of the early neuroinflammatory phase evidence the activation of the microglia and astrocytes especially by the interaction of the IL-18 secreted by the microglia with the IL-18R expressed by the astrocytes. Next, the interaction of IL-18, TNFα, and IL-1β released by microglia and astrocytes with their respective receptors expressed mainly in pre-myelinating OLs promotes death of the lineage or OPC maturation arrest, culminating in myelination deficits. In the latter phase, astrogliosis promotes the elevation of ROS and NOS levels; as immature OLs are susceptible to oxidative stress, later impairment of oligodendrocyte lineage continues.

The release of pro-inflammatory cytokines is known to reduce the differentiation and maturation of OPCs (Back et al., 2002; Ness et al., 2001; Paneth et al., 1990; Vela et al., 2002; Wellmann et al., 2014; Xie et al., 2016), to increase OL death (Li and Wong, 1998; Baerwald and Popko, 1998), and to suppress axon development (Han et al., 2017). All these changes result in decreased myelination, which was observed in studies by the reduction of MBP expression and the disruption of myelin sheaths in the white matter after hypoxia, with long-term functional deficits as a consequence (Kaur et al., 2006; Kurumatani et al., 1998; van Tilborg et al., 2018a; van Tilborg et al., 2018b; Wang et al., 2016).

Corroborating these findings, the increased expression of IL-18 by microglial and astrocytic cells have been described in the subcortical white matter in mice (Hedtjarn et al., 2005). Moreover, the expression of IL-18R in astrocytes, but not microglial cells, was reported in mice that were exposed to hypoxia, as well, patients with multiple sclerosis showed IL-18R expression in OLs and astrocytes (Cannella and Raine, 2004; Hedtjarn et al., 2005). Therefore, these results suggest that a hypoxic environment can cause neuroinflammation through the direct activation of microglial cells producing inflammatory mediators, such as IL-18, that amplify the reactive astrogliosis and damage to OLs.

Interleukin-specific receptors, such as IL-18R and IL-1R, and membrane receptors, including toll-like receptors (TLRs), mediate the upstream activation of JNK that enhances the transcription of genes involved in proliferation, cell survival and death, metabolism, and inflammation by the induction of the Activator protein 1 (AP-1) transcription factor. In rats subjected to OD, AP-1 showed to be elevated, and the inhibition of JNK promoted the reduction of brain damage and sensory-motor and cognitive behavioral improvements (Nijboer et al., 2010).

One of the inflammatory cytokines transcribed after the JNK activation is the TNFα that interacts with its receptors, TNFR1 and TNFR2. In human brains from early infantile deaths with PVL, an overexpression of TNFα and its receptors was identified in the affected areas, mainly in microglia, astrocytes, and OLs (Kadhim et al., 2006). These findings support the neurotoxic and pro-inflammatory actions, indicated by TNFα/TNRF1/JNK signaling in mice injected with LPS and exposed to hypoxia and ischemia. TNFR1-JNK, but not TNFR2 signaling, is a common pathway leading to apoptosis of pre-myelinating OLs (Wang et al., 2014).

In astrocyte cultures mRNA expression and protein levels of TNFα and IL-1β mRNA, proteins, their respective receptors TNF-R1 and IL-R1, and caspase-3 expression were upregulated significantly after hypoxic exposure (Deng et al., 2014). In addition, IL-1β may inhibit OL proliferation through its receptor (Vela et al., 2002). These results suggest that the release of TNFα and IL-1β by reactive astrogliosis causes the death of OLs and inhibition of OL proliferation.

Together, these data indicate that OD could activate astrocytes to produce inflammatory mediators, which would damage the OLs through their corresponding receptors. Therefore, not only microglia but also astrocytes may play an important role in white matter injury in the hypoxic neonatal brain. Astrogliosis boosts the potential for the production and release of pro-inflammatory cytokines and the overproduction of free radicals that are toxic to neurons and OLs (Kirkley et al., 2017). Cytokines can also be harmful to white matter, for example inhibiting the differentiation of OLs precursors (Favrais et al., 2011).

6. Myelination alterations and neurodevelopmental disorders (NDD)

Changes in myelination are found in different clinical conditions in which OD has occurred, just as in distinct animal models used to study the underlying mechanisms of the anoxic or hypoxic episodes (van Tilborg et al., 2017; Ziemka-Nalecz et al., 2018). The employment of these animal models recapitulates the OD episodes that occur in humans and allows the evaluation of lesions in the white matter, which are associated with the occurrence of cerebral palsy, epilepsy, schizophrenia, attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), learning and memory deficits, and cognitive impairments observed in humans (Cainelli et al., 2020; Juliano et al., 2015; van Tilborg et al., 2017).

Moreover, myelination disruption is not limited to the neonatal period. In a double-hit model in which neonatal hypoxia is performed after perinatal inflammation, van Tilborg and colleagues showed that Wistar rats that had experienced hypoxia after perinatal inflammation expressed lower levels of MBP 21 or 28 days after the stimulus and a delay in myelination up to the age of P30. These animals also presented impairment of OL maturation which was related to the increase in IBA1⁺ and GFAP⁺ cells, and also exhibited behavior similar to ASD, with decreased sociability, stereotyped behavior and increased anxiety behavior (van Tilborg et al., 2017). In another study, reduced myelination was accompanied by sensory and motor deficits in C57Bl/6 J mice after exposure to intermittent hypoxia in 7-day-old pups (Juliano et al., 2015). It might indicate that different models used to mimic neonatal OD may have different ways to impact the cellular mechanisms of OLs, influencing somatosensory development and cognitive impairment, evident in neurodevelopmental disorders.

These results agree with human clinical epidemiology where 25–50 % of individuals born prematurely and with extremely low birth weight present moderate cognitive, behavioral, and sensorimotor deficits throughout their lives (Volpe et al., 2011) whereas 5–10 % of surviving babies develop cerebral palsy (Doyle et al., 2010; Janowska et al., 2020).

Due to the loss of white matter integrity, pathologies such as schizophrenia and epilepsy are also associated with OD insults. In a study, 10-day-old C57/Bl6 mice exposed to hypoxia-ischemia show a pattern of neuronal activity similar to that seen in the EEG of human neonates with hypoxic-ischemic injury; the animals exhibited burst suppression interrupted by wave peaks during and after the hypoxia period in the hippocampus and cerebral cortex. Interestingly, in the context of the model used, HI with unilateral ischemia, the ischemic hemisphere showed greater activity than the contralateral hemisphere, with asymmetric activation occurring in the somatosensory cortex, hippocampal structures, and the olfactory system. Conversely, the anterior regions were activated bilaterally, especially in the motor cortex (Burnsed et al., 2019). Moreover, the expression of MBP was more prevalent in the anterior region of the corpus callosum, unlike the posterior region, which did not show MBP labelling and whose structures showed an asymmetric activation pattern (Burnsed et al., 2019). It is important to note that the corpus callosum is composed of axons that cross the midline to connect the cerebral hemispheres; however, the propagation of convulsive activity was shown bilaterally only in the frontal lobe, coinciding with the higher expression of MBP in the anterior part of the corpus callosum (Burnsed et al., 2019). These results suggests that, although the immaturity of myelination is important in limiting the propagation of seizure activity, the occurrence of the hypoxic-ischemic event triggers abnormalities in the activity of the neuronal circuitry and the present myelin is crucial in propagating such activity (Aberg et al., 2007; Nyakas et al., 1996).

Corroborating the hypothesis of the existence of myelination-associated long-term deficits, Wang and colleagues showed that, after neonatal hypoxia, 28-day-old C57Bl mice present neurons with disorganized sheaths and low-density of myelin (Wang et al., 2020). This is in line with van Tilborg's findings that showed disorganization of myelinated axons with fewer intersections and fiber length reduction in animal models of neonatal OD (Kurumatani et al., 1998). These results indicate a strong relationship between neonatal hypoxia events and the sequelae seen throughout life (Juliano et al., 2015), such as those related to attention, sociability, and task execution, which seem to be linked to disturbances in connectivity caused by white matter damage (Aberg et al., 2007). In schizophrenic patients, changes in myelin are common and include a reduced number of OLs, swollen processes, and axonal atrophy (Burnsed et al., 2019; Wang et al., 2018). Also, obstetric complications leading to preterm birth and neonatal hypoxia/anoxia are strong risk factors for the white matter damage observed in these patients (Berger et al., 2016; Takahashi et al., 2011). In an animal model using 1-day-old Wistar rats exposed to hypoxia-ischemia, the rats exhibited behaviors similar to those found in schizophrenia, such as impaired prepulse startle reflex inhibition and locomotor hyperactivity, 38 days after injury (Tejkalova et al., 2007). In another study in Sprague-Dawley rats, hyperactivity accompanied by memory deficits was associated with white matter and axonal damage (Thaker, 2008).

In conclusion, neurodevelopmental disorders share a close relationship with prematurity and neonatal OD. Moreover, OL development is vulnerable to inflammatory cytokines and oxidative stress, and the damage to this cell population ends up in impaired myelination and alterations of cell metabolism, which significantly impact the establishment of neuronal circuits and other important cellular processes through the different stages of neurodevelopment.

7. Myelination and sexual dimorphism in NDD

Studies have already noted that the incidence of many neurological disorders differs between the sexes and elucidating the underlying biological basis has become a high-priority challenge (Fabres et al., 2022; Netto et al., 2017; Sanches et al., 2013a). The observed physiological processes of myelination also appear to show differences depending on sex. Both the volume of white matter, the volume of myelinated nerve fibers and the volume of myelin sheaths in the white matter were significantly higher in male rats aged 6 to 8 months compared to female rats of the same age. However, these three evaluated parameters are shown to be higher in females than males when they reach around 18 months of age (Yang et al., 2008; Lai and Yang, 2011).

Previous studies suggest that female mice have smaller myelin volume and number of OLs compared to males (Cerghet et al., 2009; Schumacher et al., 2020; Swamydas et al., 2009). The difference in the number of OLs was most notable in younger rodents, with males showing a number of OLs 40 % higher than females. In rats, the proliferation of glial cells can be 2-fold higher in corpus callosum of females; however, cell death of glial cells (evaluated by active caspase-3, an apoptosis marker) was 50 % higher in female mice compared to males. Similar data were observed by Cerghet et al. when they observed that calpain increased 2-fold higher in females in the spinal cord and brain (Cerghet et al., 2009). Calpain, calcium-activated neutral proteinase, can be found in the myelin sheath and participates in the degradation of intracellular myelin protein with an increase in cytoplasmic calcium at pathophysiological levels (Shields et al., 1999). Therefore, as proliferation and cell death increase in females, OLs in females are likely to have a shorter life and faster turnover than males (Cerghet et al., 2009).

These developmental differences may be due to sexual steroids. The effects of sex steroids on myelin have been observed not only in the development of the white matter but also in the injured white matter. For example, estradiol has a dose-dependent protective effect against oxygen-deprivation-induced apoptotic cell death in primary OLs. However, estradiol and dihydrotestosterone (DHT) were shown to have similar effects in male and female mice. Estrogen had minor effects on OL numbers; and DHT reduced OL numbers in both sexes, but more so in females (Swamydas et al., 2009).

Also progesterone, one of the most widely studied neurosteroids, restored myelination and increased the density of OPC in a spinal cord injury model, and in combination with estradiol restored myelin protein expression in experimental autoimmune encephalitis (EAE), a multiple sclerosis model (Cerghet et al., 2009; Garay et al., 2008). However, administration of progesterone in a culture model of OLs can be seen that both sexes are protected, but females are slightly more protected than males. In other study, Marin-Hussteg and collaborators observed a higher number of oligodendrocytes in females compared to males using primary cultures of oligodendrocyte progenitors isolated from male and female neonatal rodent brains. This study also observed that estrogen receptor β are located mainly in the cytoplasm of oligodendrocytes, suggesting that estradiol can interact with this receptor and have some effect. Furthermore, they observed that treatment with progesterone increased cellular branching and the use of estradiol increased membrane sheet formation (Marin-Husstege et al., 2004). In these studies females always showed greater differences than males regarding changes in OLs numbers and signaling molecules.

While analyzing the adult brain model, progesterone therapy also promoted the repair of severe chronic demyelinating lesions induced by the OLs toxin cuprizone in female mice. Progesterone increased the density of OPC and mature OLs in the cuprizone-demyelinated corpus callosum and stimulated the formation of new myelin sheaths. Regarding myelination and remyelination during brain development, the proremyelinating effect of progesterone in adults was progesterone receptor (PR)-dependent (Schumacher et al., 2020).

Among the effects of progesterone, it was observed that in OL cultures the pAkt was more up-regulated in females than males, correlating with the effects of progesterone on OL numbers. The PI3K/Akt pathway is an important anti-apoptotic signaling and cell survival pathway. Akt activation stimulates substrates involved in cell survival and inhibits pro-apoptotic substrates such as caspase-3, inhibiting apoptosis pathways and contributing to cell survival by downstream mTOR activation via transcriptional regulation (Chong et al., 2007; Hay, 2005).

The administration of progesterone for treatment increases the expression of the mTOR protein which plays an important role in the Akt pathway which, in turn, acts on the survival of OLs. Using the PI3 kinase inhibitor whose effect is to block the Akt pathway, caused the reduction of the effects of progesterone, and thus reduced its influence on the regulation of OL survival, confirming that the Akt pathway is one of the keys to the action of progesterone in the central nervous system (Swamydas et al., 2009).

Additionally, some studies show that the striatum and white matter of females are more severely affected when exposed to a hypoxic environment (Arteni et al., 2010; Sanches et al., 2013a; Sanches et al., 2013b; Uluc et al., 2013), whereas others indicate that males are more severely affected (Zhu et al., 2003; Mayoral et al., 2009). Hypomyelination, resulting from the interruption of myelin synthesis after an injury, is commonly found in white matter injury (WMI) survivors, especially in diffuse WMI cases. Degeneration and dematuration of OLs may contribute to the pathogenesis of hypomyelination. In perinatal WMI, decreased expression of MBP was associated with long-term cognitive dysfunction (Liu et al., 2019).

Hypomyelination in the corpus callosum and cortex result in insufficient information processing and conduction speed between the cortical regions or between the cortex and the deep gray nuclei, such as the connections between the thalamus and the cortex. Liu and colleagues reported a reduction in numerical density and an increase in the ultrastructural defects of the synapses in the thalamus of rats with WMI, induced by HI. In addition, studies show that male mice would have greater expression of total MBP, summing the 4 isoforms, compared to females; however, the 21.5 kDa isoform was found to be more abundant in females by approximately 30–35 % (Benjamins and Morell, 1978; Cerghet et al., 2009). Therefore, if only female or male animals are used for white matter research under various experimental conditions, the conclusion may be affected by white matter sex differences.

In addition, sex differences in the corpus callosum were also identified in post-mortem studies in human brains, in this study it was observed that the width and surface area of the splenium were less in males when compared to females (DeLacoste-Utamsing and Holloway, 1982). In addition, increased fiber density was also observed in females compared to males in subregions of the corpus callosum (Highley et al., 1999).

An important question is whether the sex differences observed in myelin formation, particularly in the cells involved in myelination during the critical period of neurodevelopment, persist into adulthood. Daniel et al. 2015 observed an increase in the expression of MBP and PLP in adult female rats (around 85 days old) compared to male rats of the same age in the orbital frontal cortex (OFC) (Bayless and Daniel, 2015). In 2019, Daniel and collaborators verified whether the sexual difference previously observed in the OFC could be attributed to variations in neurodevelopment. For this, male and female animals were gonadectomized at two different ages, simulating adolescence and adulthood. Animals submitted to surgery in prior to the onset of puberty (adolescence) showed a reduction in MBP in OFC compared to animals without surgery and animals that underwent gonadectomization in adulthood. These results suggest a possible relationship between sex hormones and the neurodevelopment of myelinating cells and that this difference can persist into adulthood (Darling and Daniel, 2019).

Surprisingly, despite the evidence showing the importance of studying both sexual dimorphisms in NDDs, many studies still do not report the sex of animals used or use only males. Previous studies have observed that both inflammation and cell death in vitro and in vivo models of OD are sex-dependent. Therefore, our review aims to encourage further work that tries to understand the consequences of NDDs on myelination, death, proliferation, and differentiation of OLs and their sex-dependent effects.

Data availability

No data was used for the research described in the article.