If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Communications should be addressed to: Volpe; Department of Pediatric Newborn Medicine; Brigham and Women's Hospital; 221 Longwood Avenue, Room 343C; Boston, MA 02115 USA.
Department of Neurology, Harvard Medical School, Boston, MassachusettsDepartment of Pediatric Newborn Medicine, Harvard Medical School, Boston, Massachusetts
Prematurity, especially preterm birth (less than 32 weeks' gestation), is common and associated with high rates of both survival and neurodevelopmental disability, especially apparent in cognitive spheres. The neuropathological substrate of this disability is now recognized to be related to a variety of dysmaturational disturbances of the brain. These disturbances follow initial brain injury, particularly cerebral white matter injury, and involve many of the extraordinary array of developmental events active in cerebral white and gray matter structures during the premature period. This review delineates these developmental events and the dysmaturational disturbances that occur in premature infants. The cellular mechanisms involved in the genesis of the dysmaturation are emphasized, with particular focus on the preoligodendrocyte. A central role for the diffusely distributed activated microglia and reactive astrocytes in the dysmaturation is now apparent. As these dysmaturational cellular mechanisms appear to occur over a relatively long time window, interventions to prevent or ameliorate the dysmaturation, that is, neurorestorative interventions, seem possible. Such interventions include pharmacologic agents, especially erythropoietin, and particular attention has also been paid to such nutritional factors as quality and source of milk, breastfeeding, polyunsaturated fatty acids, iron, and zinc. Recent studies also suggest a potent role for interventions directed at various experiential factors in the neonatal period and infancy, i.e., provision of optimal auditory and visual exposures, minimization of pain and stress, and a variety of other means of environmental behavioral enrichment, in enhancing brain development.
Preterm birth (less than 37 weeks' gestation) is an enormous public health problem worldwide. According to the World Health Organization, approximately 15 million premature infants are born yearly and account for approximately one million deaths.
Global, regional, and national causes of under-5 mortality in 2000-15: an updated systematic analysis with implications for the Sustainable Development Goals.
The United States ranks sixth among countries in terms of the number of preterm births. According to the Centers for Disease Control and Prevention, from 2014 to 2017 the preterm birth rate rose in the United States to approximately 10%. Of the approximately four million births in the United States, 1.4%, or about 56,000, are of very low birth weight (less than 1500 g).
Survival rates vary markedly as a function of gestational age but are at least 95% at 32 weeks', 90% at 28 weeks', and 60% to 65% at 24 weeks' gestation.
The substantial survival rates, unfortunately, are accompanied by relatively high incidences of neurological disability, for example, cerebral palsy in 5% to 10%, other motor disturbances in 25% to 40%, and cognitive, attentional, behavioral, and socialization disturbances in 25% to 50%.
Developmental coordination disorder and its association with developmental comorbidities at 6.5 years in apparently healthy children born extremely preterm.
The neuropathological substrate of this disability in preterm infants, especially those very preterm (less than 32 weeks' gestation) and extremely preterm (less than 28 weeks' gestation), consists of a combination of cerebral white matter injury (WMI) and especially, subsequent dysmaturational events in both white matter and neuroaxonal structures (see later). This combination of WMI and disturbances of gray matter structures has been termed the encephalopathy of prematurity.
Subsequent work now suggests that although WMI is an important and likely initiating event, multiple subsequent dysmaturational events are most critical in determining the outcomes (see later). Moreover, because these dysmaturational events evolve over a very prolonged period (many months), a relatively long time window exists for interventions to prevent, counteract, or ameliorate the dysmaturation, i.e., neurorestorative interventions (see later).
In the following discussion, I will review the multiple maturational events occurring in infant brain during the premature period; the dysmaturational events observed in premature infants, including the importance of the initiating cerebral WMI; the dysmaturational events that may occur without WMI; and the potential neuroprotective and neurorestorative interventions.
Brain maturation during the premature period
The brain dysmaturation that occurs in premature infants (see later) involves the multiple active developmental events occurring in human cerebrum during the period of 20 to 40 weeks' gestation and beyond. The rapidity and complexity of these cellular events underlie, to a considerable degree, their vulnerability to perturbations. The principal components involved include the oligodendroglial (OL) lineage, especially the preoligodendrocyte (pre-OL), cerebral white matter axons, subplate neurons, cerebral cortex, thalamus, and basal ganglia (Fig 1). In addition, microglia and astrocytes, especially in the white matter, are involved importantly in both normal development and dysmaturation of these principal components. The major developmental events during this period have been summarized in detail elsewhere.
A brief review of the temporal aspects of these events is appropriate here (Table 1).
FIGURE 1Schematic of major components involved in brain maturation in the premature period. See text for details. GP, globus pallidus; Pre-OL, pre-oligodendrocyte; Put, putamen; SPN, subplate neuron; Thal, thalamus.
TABLE 1Major Developmental Events During the Premature Period
20-24 weeks
Proliferation of OL progenitors
Cerebral white matter axons (projection, commissural, and association) grow actively
Subplate neuronal layer well established
Thalamic afferent axons synapse abundantly on subplate neurons
24-32 weeks
OL progenitor differentiation leads to prominence of pre-OLs in cerebral white matter
Cerebral white matter axons continue active growth
Pre-OLs begin ensheathment of cerebral white matter axons
Subplate reaches maximum size (several times thicker than cortical plate at 27-30 weeks)
Thalamocortical afferent axons depart subplate neurons and enter cerebral cortex
Cerebral cortical dendritic development and synaptogenesis become prominent
Callosal (commissural) and association (corticocortical) axons enter subplate
GABAergic neurons migrate into cerebral white matter
32-40 weeks
Pre-OLs remain the predominant cell of OL lineage in cerebral white matter until approximately 40 weeks when they and the more differentiated “immature” OL each account for approximately 50% of the OL lineage
Subplate layer gradually decreases
Callosal and corticocortical axons depart subplate and enter cerebral cortex
GABAergic neurons migrate to cerebral cortex and populate upper cortical layers
Cerebral cortical dendritic development and synaptogenesis become marked
These cells are generated from OL progenitors and are the principal phase of the OL lineage during the premature period (Table 1, Fig 2). Pre-OLs account for 90% of the lineage during the peak period of WMI in premature infants. Even at term, pre-OLs account for 50% of the lineage in cerebral white matter, whereas approximately 50% of the lineage is the more differentiated “immature” OLs.
Mature, myelin-producing OLs do not develop in human cerebral white matter to an appreciable degree until post-term. The pre-OL begins ensheathment of white matter axons at approximately 30 weeks' gestation (Fig 3).
Arrested oligodendrocyte lineage progression during human cerebral white matter development: Dissociation between the timing of progenitor differentiation and myelinogenesis.
and, as a consequence, axonal function. The latter is the critical driving force for cerebral cortical development (see later), which evolves rapidly as an activity-dependent process during the third trimester of gestation.
FIGURE 2Major phases of the oligodendroglial lineage. The pre-OL (circled) is by far the predominant form during the premature period. OL, oligodendrocyte.
FIGURE 3Pre-OL ensheathment of axons at 30 weeks' gestational age. Pre-OL is immunostained green, and axon, red. OL, oligodendrocyte. (From Back SA, Luo NL, Borenstein NS, Volpe JJ, Kinney HC. Arrested oligodendrocyte lineage progression during human cerebral white matter development: Dissociation between the timing of progenitor differentiation and myelinogenesis. J Neuropathol Exp Neurol. 2002; 61:197-211, with permission).
The pre-OL is a highly vulnerable cell, with particular susceptibility to such insults as hypoxia, ischemia, and inflammation, which lead to death via excitotoxic and free-radical-mediated mechanisms.
Suffice it to say here, many experimental studies of acute pre-OL death produced by hypoxia, ischemia, and inflammation have shown protective benefit for such agents as antiexcitotoxic, antiinflammatory, and antioxidant compounds (see later). Notably, however, as will be discussed later, in the premature infant with WMI, pre-OLs are replenished in the subacute period but fail to differentiate over the ensuing weeks or months to later phases of the OL lineage. As a result, hypomyelination is a hallmark of the disease.
Axons
Axonal development is remarkably active in the cerebrum during the premature period (and the early postnatal period) (Table 1).
showed marked expression in cerebral white matter to at least 37 weeks' gestation. Growing white matter axons reach approximately the subplate region at 20 weeks, the deep layers of the cortical plate at 27 weeks, and the entire cortex by 37 weeks (Fig 4). Axonal growth occurs primarily within the cortex after 37 weeks and into the first year of life. Based on work by Kostovic and coworkers,
the likely anatomic correlates of this progression in cerebral white matter during the premature period are growth of axons from thalamus to subplate neurons at 20 weeks and from subplate neurons to the cerebral cortex at 27 weeks (Table 1). Also, at 27 weeks, commissural and corticocortical cerebral white matter axons are actively growing, especially in the posterior periventricular regions, the so-called crossroads area. The increase in cerebral cortical expression of GAP-43 at 37 weeks may reflect a sum of continued cortical penetration from the subplate of thalamic ascending fibers and of commissural and corticocortical fibers (Fig 4). Thus it is apparent that the premature period is one of extraordinarily rapid axonal development, especially in cerebral white matter. Axons during this rapid growth period are exquisitely vulnerable to multiple insults (see later).
FIGURE 4GAP-43 expression in developing human parietal white matter and cortex. Cortex is indicated by an asterisk. Note at 20 postconceptional (PC) weeks (A) there is strong expression in cerebral white matter to a region below the cortical plate, likely subplate neurons. At 27 PC weeks (B) the expression begins to enter the cerebral cortex and continues in white matter. By 37 weeks (C), diffuse expression in cortex as well as continued expression in white matter are apparent. At 144 PC weeks (approximately age two years) (D), expression is prominent in cortex but not in white matter. (From Haynes RL, Borenstein NS, DeSilva TM, Folkerth RD, Liu LG, Volpe JJ, et al. Axonal development in the cerebral white matter of the human fetus and infant. J Comp Neurol. 2005; 484:156-67, with permission).
The cerebral cortex undergoes dramatic changes during the premature period. These events include attainment of proper alignment, orientation and layering of cortical neurons (six layers apparent by 30 gestational weeks), arrival of late migrating GABAergic neurons (principally to upper cortical layers), elaboration of dendritic and axonal ramifications (neurite outgrowth), onset of synaptogenesis, and a marked increase in cortical surface area with gyral development (Table 1).
Importantly, in this context the progress of dendritic development depends on the establishment of afferent input from cerebral white matter and then presumably synaptic activity.
Thus axonal input from subplate neurons and then from thalamic, commissural, and corticocortical fibers are the principal driving forces underlying cortical dendritic development.
for review). In a seminal study, Sarnat and coworkers studied synaptic development in human cerebral cortex from 6 to 41 weeks' gestational age with the immunomarker synaptophysin, which identified maturation of synaptic vesicles in axonal terminals. Thalamocortical axons exhibited intense staining in frontal cortex at approximately 26 weeks' gestation, and diffuse and uniformly strong staining was apparent throughout the cortex from 34 weeks onward.
The findings integrate closely with measures of axonal development in the last trimester of gestation and with previous delineations of electrencephalographic maturation in premature infants. Functional synaptic activity via the axonal input to cortex is mediated principally through excitatory amino acid receptors, both the excitatory Ca++-permeable N-methyl-d-aspartate and GluR2-deficient α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors, which exhibit exuberant expression in developing human cortex during this period.
Developmental regulation of alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptor subunit expression in forebrain and relationship to regional susceptibility to hypoxic/ischemic injury. I. Rodent cerebral white matter and cortex.
Developmental regulation of alpha-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid receptor subunit expression in forebrain and relationship to regional susceptibility to hypoxic/ischemic injury. II. Human cerebral white matter and cortex.
This role of functional activity has implications for the effects of a variety of environmental stimuli on cortical development in the premature infant, and for potential neurorestorative roles for such stimuli in the context of brain injury and dysmaturation (see later).
FIGURE 5Cerebral cortical development from 15 to 35 weeks' gestation, Golgi-Cox preparations. Magnification is the same for each sample. Note the remarkable apical and basilar dendritic development, especially after 24 weeks' gestation. (From Marin-Padilla M: Ontogenesis of the pyramidal cell of the mammalian neocortex and developmental cytoarchitectonics: A unifying theory, J Comp Neurol 321:233-240, 1992, with permission).
During the important period of 24 to 32 weeks (the peak period for the occurrence of cerebral WMI) the subplate reaches maximum size (several times thicker than the cortical plate at 27 to 30 weeks) (Fig 1, Table 1). These neurons elaborate a dendritic arbor with spines, receive synaptic inputs from ascending afferents from thalamus and distant cortical sites,
and extend axon collaterals to the overlying cerebral cortex and to other cortical and subcortical sites (thalamus, other cortical regions, corpus callosum). The crucial organizational functions of these neurons include provision of a transient synaptic site for ascending afferents from thalamus and other cortical sites, i.e., these “waiting” afferents cannot synapse yet in cortex because their neuronal targets have not yet differentiated. These afferents would undergo degeneration if they did not have the subplate neurons as transient targets. Moreover, the subplate neurons extend axons to cortex to promote cortical differentiation and to guide the afferent axons to cortex when sufficient cortical differentiation has occurred. Subplate axon collaterals also descend to pioneer or guide the initial axonal projections from cerebral cortex toward subcortical sites (e.g., thalamus, corpus callosum, other cortical sites). The subplate neuronal layer gradually decreases after 36 to 40 weeks' gestation.
Late migrating GABAergic neurons
Particularly characteristic of human cerebral cortical development is the relatively late generation of GABAergic neurons from the dorsal telencephalic subventricular zone and from the ventral ganglionic eminence (Fig 6).
The origin of these late generated neurons is approximately 65% from the dorsal subventricular zone and 35% from the ventral ganglionic eminence. A substantial proportion of the ultimate population of GABAergic cortical neurons migrate through the cerebral white matter to the cortex in the third trimester. This migration peaks around term and then declines within the first six postnatal months.
FIGURE 6GABAergic neuron development during the premature period. Proliferation of GABAergic neurons occurs in the dorsal subventricular zone (SVZ) and ventral ganglionic eminence; migration proceeds radially and tangentially to cortex and thalamus, as shown.
Microglia and astrocytes are key players in the development of the white and gray matter structures just described. These glial elements also play a major causal role in the dysmaturational events that occur with cerebral WMI. Emphasis in this section is on the roles of microglia and astrocytes in normal development. Their role in dysmaturation is discussed later in the section on neuropathology.
Microglia
Microglia play important roles in such aspects of brain development such as axonal development, OL differentiation-myelination, vascularization, synaptogenesis, synaptic pruning, and neural circuit formation.
The roles in OL development involve microglial proteins that stimulate pre-OL proliferation, enhance pre-OL survival and provide iron for OL differentiation, and secrete cytokines that enhance differentiation.
These cells are also the principal neuroimmune cells involved in neuroinflammatory responses. As part of the neuroinflammatory responses microglia can be destructive to cellular elements, such as pre-OLs, principally by generating free radicals, secreting injurious cytokines, and enhancing excitotoxicity (see later).
Reversible inhibitory effects of interferon-γ- and tumour necrosis factor-α on oligodendroglial lineage cell proliferation and differentiation in vitro.
Neuronal injuries induced by perinatal hypoxic-ischemic insults are potentiated by prenatal exposure to lipopolysaccharide: animal model for perinatally acquired encephalopathy.
Microglia have been characterized generally as pro-inflammatory (activated) (MI) or anti-inflammatory (M2). However, this bimodal characterization appears now to be too simplistic. Thus a recent landmark study in the developing mouse, utilizing molecular characterization methods, identified at least nine distinct microglial subpopulations with unique molecular signatures that changed over the course of development and exhibited marked spatial differences.
One distinct population was highly concentrated in axon tracts of the premyelinated brain. The molecular signatures of the microglial subpopulations in early development identified pathways associated with cell metabolism, growth, motility, and proliferation, among others. Studies in developing human brain will be of great interest.
Microglia become prominent in the human forebrain at 16 to 22 weeks' gestation and migrate progressively through the white matter from 20 to 35 weeks, and then to the cerebral cortex.
The critical point is that the cerebral white matter of the human premature infant is heavily populated with microglia during a period when various maturational events are occurring and when a variety of proinflammatory insults can lead to “activation” to destructive microglial phenotypes and WMI (see later). Moreover, because of the important role of microglial subpopulations in such important developmental events as OL development, axonal guidance, synaptogenesis, sculpting of neural networks, and cerebral connectivity, diversion of these normal cells to microglial phenotypes with primarily proinflammatory functions could contribute to disturbances in these maturational events observed in the premature brain.
Fibrous astrocytes (generated from radial glial fibers) increasingly populate the cerebral white matter. During development astrocytes are important in axonal guidance, angiogenesis, formation of the blood-brain barrier, synaptogenesis, neuronal survival, and axonal and synaptic pruning.
The molecular characteristics of astrocytes involved in facilitation of these events underlie such functions as expression of extracellular matrix (ECM) proteins and axonal guidance molecules, secretion of angiogenic factors, secretion of synaptogenesis molecules, clearance of extracellular glutamate, and secretion of various neurotrophic molecules. As will be discussed later, in the context of various brain insults (e.g., inflammation, hypoxia-ischemia), astrocytes can become “reactive” and exhibit a variety of metabolic changes that are deleterious to other white matter components, including pre-OLs.
Dysmaturation in premature brain
Overview
The principal manifestations of dysmaturation in premature brain have been elucidated by advanced magnetic resonance imaging (MRI) techniques in living infants (Table 2). Briefly, the abnormalities have by volumetric MRI, diminished regional volumes, especially of cerebral cortex, white matter, thalamus, and basal ganglia; by diffusion-based imaging, in cerebral white matter, decreased fractional anisotropy (FA) with relatively greater involvement of radial diffusivity (consistent with impairment of pre-OL ensheathment of axons), and in cerebral cortex, blunting of the normal decline in FA (consistent with impaired dendritic development); by surface-based MRI measures, decreased cerebral cortical surface area and cortical folding or gyrification; and by functional MRI, impaired development of measures of connectivity, including especially thalamocortical connectivity. The abnormalities have been elucidated most commonly at term equivalent age, but generally persist, or may increase later in infancy, childhood, adolescence, or adulthood.
The most common accompaniment by MRI has been cerebral WMI (see later). The dysmaturational events, in general, appear to be secondary to WMI (see later discussion in Mechanics of dysmaturation with cerebral white matter injury). The constellation of WMI and the accompanying disturbances of neuronal or axonal structures is generally referred to as the encephalopathy of prematurity.
However, recent work suggests that some of the dysmaturational events documented in premature infants are not clearly related to WMI and perhaps are primary disturbances (see later). The emphasis in the following section is on the relation of cerebral WMI and dysmaturational events. Brief consideration of potentially primary dysmaturational events, perhaps independent of WMI, will then be presented.
TABLE 2Dysmaturational Features in Premature Brain Elucidated by Advanced MRI
Volumetric MRI
Decreased regional volumes, especially cerebral cortex, white matter, thalamus, basal ganglia
Diffusion imaging
In cerebral white matter, decreased FA, relatively increased radial diffusivity, variability altered axial diffusivity In cerebral cortex, blunting of the normal decline in FA
Surface-based MRI measures
Decreased cerebral cortical surface area and cortical folding/gyrification
Functional MRI
Impaired development of measures of connectivity, including especially thalamocortical connectivity
Cerebral WMI encompasses a spectrum of neuropathology that ranges from overt periventricular leukomalacia (PVL) to diffuse white matter gliosis (DWMG) (without focal necroses) (Fig 7A-C). The two fundamental characteristics of PVL are focal necroses with loss of all cellular elements in periventricular white matter and a more diffuse lesion in cerebral white matter, consisting initially of death of early differentiating pre-OLs, accompanied by vigorous and persistent astrogliosis and microgliosis (Fig 8).
The focal necroses are essentially infarcts. Temporally, in the more diffuse lesion the pre-OL disturbance consists acutely of cell death, followed subacutely and chronically initially by proliferation of pre-OLs but then critically, by a failure of maturation.
Nitrosative and oxidative injury to premyelinating oligodendrocytes is accompanied by microglial activation in periventricular leukomalacia in the human premature infant.
As noted earlier, this pre-OL dysmaturation underlies the subsequent hypomyelination, a unifying feature of PVL. The mildest form of WMI, i.e., DWMG without focal necroses, now is the most common form of WMI in premature infants and is also accompanied by the pre-OL dysmaturation.
FIGURE 7Spectrum of white matter injury (WMI) in premature infants. (A) to (C) illustrate the neuropathological spectrum of WMI. In severe WMI (A) the focal necrotic component consists of macroscopic areas of necrosis that result in cysts (i.e., “cystic” WMI). In moderate WMI (B) the focal necrotic component consists of small areas of necrosis that result in glial scars (i.e., “noncystic” WMI). In mild WMI (C) the focal component may be microscopic (less than 1 mm) or absent. In (A) through (C) the diffuse component of WMI consists of pre- oligodendrocyte (OL) injury or death (followed by pre-OL proliferation and maturation failure) and diffuse white matter gliosis (DWMG) involving activated microglia and reactive astrocytes. In panels (D) through (F) the magnetic resonance imaging correlates are shown. In (D), severe WMI, periventricular cysts are apparent; in (E), moderate WMI, punctate white matter lesions (PWMLs) but not cysts are seen; and in (F), mild WMI, only diffuse signal abnormality in white matter is apparent.
FIGURE 8Diffuse reactive astrogliosis (GFAP immunostain) and diffuse activation of microglia (CD-68 immunostain) in cerebral white matter in all three varieties of white matter injury (WMI) illustrated in Fig 7. GFAP, glial fibrillary acidic protein.
The relative distribution of the spectrum of cerebral WMI in the modern era has been delineated best by neuropathological studies. The two largest, most recent series demonstrate that compared with earlier studies the areas of focal necrosis are smaller and, indeed, most cases have few or none.
In the series by Pierson et al. (n = 41), true PVL, i.e., with focal necroses, occurred in 17 (42%). Importantly, nearly all of these lesions were less than 1 mm in size. In an additional 17 (42%), only DWMG without focal necroses was observed. (Only seven of the 41 brains were free of white matter abnormality.) Critically, Busser and coworkers observed in association with DWMG the sequelae of pre-OL death, i.e., the excess of pre-OLs and failure of pre-OL maturation.
Thus the full spectrum of cerebral WMI can be illustrated as shown in Fig 7A-C.
Pathogenesis
The pathogenesis of the focal necroses characteristic of PVL relates primarily to decreases in cerebral blood flow, related to a variety of perinatal or neonatal events, and the presence in the periventricular area of vascular border zones and end zones.
The diffuse abnormality, DWMG, relates in considerable part to similar, albeit less severe clinical events (see later).
In the diffuse lesion the pathogenesis of the acute pre-OL injury or death likely relates in part to the acute insults, noted above, as well as the accompanying disturbances that may predispose the pre-OL to injury (e.g., intrauterine growth retardation, systemic infection, impaired nutrition) (see later). The stimulus for the subsequent proliferative response of OL progenitors to produce abundant pre-OLs remains unclear. The pathogenesis of the subacute and chronic failure of maturation of these newly generated pre-OLs appears to relate to deleterious effects of the abundant activated microglia and reactive astrocytes characteristic of the diffuse lesion. These components likely are involved in other dysmaturational events relating to axonal and neuronal structures, as described next.
Deleterious roles of microglia and reactive astrocytes
The pro-inflammatory microglia may impair pre-OL maturation by release of reactive oxygen or nitrogen species or cytokines (e.g., tumor necrosis factor-α, interleukin [IL]-1β) that then act on pre-OLs.
Whether such subpopulations are involved in pre-OL dysmaturation is unknown, but it is noteworthy that a large population of potentially activatable microglia are present in normal developing white matter during the premature period (see earlier). In addition, proinflammatory microglia have been shown recently to induce formation of neurotoxic reactive astrocytes.
As discussed next, such astrocytes are important in the pre-OL maturational failure. Finally, the shift in microglial phenotype from an antiinflammatory to a proinflammatory activated phenotype diverts the critical roles of “normal” microglia in OL development described earlier.
The abundant “reactive” astrocytes (A1) in DWMG also likely play critical roles in the failure of pre-OL maturation.
The likely sequence involves the generation by reactive astrocytes of high-molecular-weight forms of hyaluronic acid. Astrocyte-associated ECM is also involved in this generation. ECM is also a key source of hyaluronidases, which convert the high-molecular-weight forms of hyaluronic acid to lower-molecular-weight forms. The latter lead to failure of pre-OL maturation, probably by activating Toll-like receptor-2 receptors on pre-OLs.
The particular role of hyaluronan is supported by the observation that pharmacologic inhibition of hyaluronidases promotes pre-OL maturation and myelination (see later). Other products of reactive astrocytes may also be involved in the pre-OL dysmaturation. Thus, in human WMI reactive astrocytes express large amounts of interferon-γ, and pre-OLs express the interferon-γ receptor,
Other products of astrocytes may contribute to the inhibition of pre-OL differentiation, for example, bone morphogenetic proteins and Notch ligand Jagged 1, but data on human preterm WMI are not yet available.
Finally, as noted for activated microglia, the shift in astrocyte phenotype from normal fibrous astrocytes to the toxic reactive phenotype also diverts the critical roles of astrocytes in the development of OLs (see earlier).
In view of the apparent critical roles of activated microglia and reactive astrocytes in disturbing pre-OL development (and likely also, aspects of axonal development), the question of the duration of DWMG in cerebral WMI of the premature infant becomes critical. Thus available evidence by MRI in vivo suggests that dysmaturation continues for many months and likely longer. Not unexpectedly, neuropathological data in human infants concerning duration of DWMG are somewhat scanty. However, available information suggests that DWMG is present for at least many months after the premature period and likely longer.
Nitrosative and oxidative injury to premyelinating oligodendrocytes is accompanied by microglial activation in periventricular leukomalacia in the human premature infant.
In the latter setting, these cells are considered important in subsequent degeneration of axons and neurons years later and to play a role in the enhanced incidence of degenerative disorders, such as Alzheimer and Parkinson diseases.
Identification in vivo
Neuroradiological identification of the cerebral WMI spectrum in vivo is made best by MRI but is not entirely satisfactory. Thus the most severe end of the WMI spectrum, i.e., severe WMI, with large areas of necrosis and apparent cystic change, are readily identified as such (Fig 7D). However, such lesions are observed on MRI (and by neuropathology) in less than 5% of infants in modern neonatal intensive care facilities.
More common are small areas of necrosis (more than 1 mm) in periventricular and central cerebral white matter, seen at term equivalent age as (noncystic) punctate white matter lesions (PWMLs) in 15% to 25%, i.e., moderate WMI (Fig 7E).
Notably, this incidence of noncystic PVL (PWMLs) is appreciably higher if scans are performed early in the neonatal period—presumably the gliotic scars contract sufficiently to become invisible to MRI by term equivalent age. The least severe end of the WMI spectrum, i.e., mild WMI, is likely a heterogeneous group. Thus the large majority of focal necroses observed postmortem are less than 1 mm in size
and likely below the resolution of most conventional MRI scanners. In addition, the MRI correlate of the very common DWMG, without focal necroses, also is unknown. Importantly, as with the diffuse gliotic component of overt PVL, DWMG alone appears to lead to pre-OL death and subsequent dysmaturation,
and thus may be very important clinically. The frequent isolated finding of diffuse signal abnormality in cerebral white matter (Fig 7F) may be the MRI correlate of mild WMI, although both the reproducibility of this imaging finding and the relation to outcome remain unclear.
The few excellent studies that identify WMI without detectable focal necroses by the presence of diminished FA on diffusion-based MRI (see later) may be the best in vivo correlate of the admixture of the WMI spectrum that includes the two large groups of (1) focal necroses too small for identification (with DWMG) and (2) DWMG without any focal necroses, the two forms that we refer to as mild WMI.
Clinical importance
The clinical importance of the cerebral WMI spectrum relates to the motor and cognitive deficits associated with the lesion and the subsequent dysmaturation. The clinical phenomena associated with moderate and severe WMI have been described in detail elsewhere.
Identification of the neurodevelopmental sequelae of mild WMI is hindered by the difficulty in identifying the lesion by conventional MRI, as used in most large-scale studies. Large-scale MRI studies of premature infants show, as expected, worsening clinical outcomes as a function of severity of WMI. However, it is noteworthy that infants with either no or “mild” abnormality in cerebral white matter by conventional MRI still exhibit neurological disability subsequently. Although cognitive scales utilized among studies vary, cognitive scores for infants (less than 28 to 30 weeks' gestation) with no or “mild” WMI are approximately 85 to 93.
In a particularly well-characterized study of 480 extremely preterm infants (less than 28 weeks' gestation), 20% of infants with no apparent WMI by conventional MRI had cognitive scores less than 85.
The possibility that neuroaxonal dysmaturation with mild WMI (see mechanisms of dysmaturation with cerebral white matter injury) is important in determination of these outcomes is suggested by follow-up studies that included assessment of gray matter abnormalities (as well as WMI).
As will be discussed later, studies that assess WMI by highly sensitive diffusion MRI measures show a clear association between mild WMI, dysmaturation, and subsequent cognitive disturbances.
Mechanisms of dysmaturation with cerebral white matter injury
The mechanisms of the dysmaturational features identified by MRI in premature brains (Table 2), especially in the context of WMI, are likely multiple. The prevailing theme is a sequence whereby the initial insult (hypoxia, ischemia, inflammation, infection, etc.) leads to primary cellular injury or death, which in turn results in the subsequent replenishment of pre-OLs but secondary dysmaturation. The cellular elements injured likely depend on the severity of the WMI. Thus in moderate to severe WMI (Fig 7A,B), all the rapidly developing cellular elements, as outlined next, appear to be injured, whereas in mild WMI (Fig 7C), the pre-OL may be the principal or only cellular element undergoing primary injury.
Dysmaturation with moderate-severe white matter injury
Pre-OL injury
Primary injury or death of the pre-OL, which is exquisitely vulnerable to hypoxic-ischemic, inflammatory, or related insults, is a consistent early feature of all forms of WMI.
Nitrosative and oxidative injury to premyelinating oligodendrocytes is accompanied by microglial activation in periventricular leukomalacia in the human premature infant.
Subsequently, over the ensuing weeks replenishment of the pre-OL pool occurs but subsequent maturation to mature, myelin-producing OLs fails. The important role of reactive astrocytes and activated microglia in this dysmaturation was described earlier. The result of this pre-OL dysmaturation is hypomyelination (Fig 9A). Also, however, pre-OL dysmaturation likely leads to failure of pre-OL ensheathment of axons, and as a consequence, impaired development, i.e., dysmaturation, of axons. The important trophic role of pre-OLs for axonal development, survival, and function was noted earlier. Indeed, this process is likely crucial for the exuberant axonal growth in cerebral white matter illustrated earlier (Fig 4) and the activity-dependent development of cerebral cortex (Fig 5). The consequences of the axonal disturbance would be diminished volumes of cerebral cortex and thalamus or basal ganglia, secondary to retrograde and anterograde (trans-synaptic) effects, i.e., involving projection fibers to and from the cortex, thalamus, and basal ganglia, i.e., thalamocortical, corticospinal, corticostriatal, and commissural and association fibers to and from the cortex, i.e., corticocortical (Fig 9A).
FIGURE 9Mechanisms of dysmaturation following injury to pre- oligodendrocytes (OLs) (A), axons (B), thalamus (C), subplate neurons, (D) or migrating GABAergic neurons (E). For sequences (A–D), the initial injury leads to multiple dysmaturational events involving pre-OLs and axons. The principal outcomes are the disturbances of myelination and cortical and thalamic development, as shown in vivo by magnetic resonance imaging. See text for details.
Primary injury to the rapidly developing, vulnerable, premyelinating axons in cerebral white matter could be a primary event with WMI (Fig 9B). Although axonal injury is shown readily in the areas of focal necrosis, a more widespread degeneration of axons detected by the apoptotic marker, fractin, also has been identified.
The dysmaturational events subsequent to axonal injury (Fig 9B) by anterograde and retrograde effects would result in the impairments of cortical and thalamic development and related abnormalities detected by MRI (see Table 2). An impairment of pre-OL maturation would result from the loss of trophic axonal signals, with the ultimate consequence, hypomyelination. A contributory role for deleterious effects of activated microglia and reactive astrocytes (see earlier) also seems likely. Moreover, because of the role of both these glial types in normal axonal development, diversion to activated or reactive phenotypes may further impair axonal development.
Thalamic injury
Primary injury to thalamus is suggested by a neuropathological study of human infants with moderate to severe WMI and thalamic abnormalities (neuronal loss, gliosis, axonal degeneration) detected in approximately 60%.
Primary injury to thalamus could lead to degeneration of axons originating and terminating in the thalamus and, as a consequence, to pre-OL dysmaturation and hypomyelination (Fig 9C).
Subplate neuronal injury
Primary injury to subplate neurons would be expected to have major secondary dysmaturational effects on thalamus by retrograde degenerative effects on ascending thalamic axons (“waiting afferents”), as well as on cerebral cortex by anterograde effects via loss of subplate neuronal axons to cortex and on descending cortical axonal projections by loss of guidance from subplate axonal collaterals (Fig 9D). Considerable experimental data support these contentions.
With axonal degeneration, subsequent disturbances in pre-OL development would be expected (Fig 9D). Although data are not entirely consistent, experimental studies suggest that subplate neurons are particularly vulnerable to hypoxia-ischemia.
Primary injury to late migrating GABAergic neurons seems possible because the migratory path of these late generated cells is from the dorsal subventricular zone through cerebral white matter to the cerebral cortex (Fig 6). Two neuropathological studies of moderate to severe WMI show a deficit in central white matter neurons consistent with late migrating GABAergic neurons.
The result of a disturbance in these neurons would be a deficit in cerebral cortical neurons, especially the upper cortical layers (Fig 9E). The MRI result would be diminished cerebral cortical volume, surface area, gyrification, and connectivity, as noted in advanced MRI studies (Table 2).
Conclusions
Thus, in moderate to severe WMI, i.e., identified by neonatal MRI by PWMLs (relatively common) or by cystic lesions (rare), several potential sequences of primary injury leading to dysmaturation and developmental impairments detected by advanced MRI techniques seem likely. Although pre-OL death and subsequent replenishment of pre-OLs, which then fail to mature, appear most consistent (Fig 9A), the other sequences depicted in Fig 9 may also occur to varying degrees, dependent in part on such factors as the gestational age of the infant; the nature, severity, and timing of the initiating insult(s); and the presence of other potentiating factors, for example, intrauterine adversity, postnatal infection, undernutrition, etc.
Dysmaturation with mild white matter injury
Mild WMI, as discussed earlier, is characterized by focal necrotic lesions less than approximately 1 mm in size, and thus undetectable by conventional MRI, or DWMG without focal necroses (Fig 7C). The dysmaturational features apparent subsequently in vivo by advanced MRI are similar in many respects to those described earlier for moderate and severe WMI (Table 2) but are less pronounced.
Thus a series of careful studies of premature infants without major WMI and utilizing diffusion-based MRI determinations of FA and related measures in cerebral white matter as a means to detect mild WMI, not readily apparent on conventional MRI, show at term equivalent age disturbances in volumetric development of cerebral cortex, cerebral white matter, thalamus, basal ganglia, cortical folding, cortical and white matter microstructure, and thalamocortical connectivity.
In a particularly large, recent series (n = 491), Barnett et al. identified lower FA in cerebral white matter with particularly high radial (versus axial) diffusion (RD).
The findings suggest that impaired pre-OL maturation is the critical finding in mild cerebral WMI. The lower FA values were independently associated with increased number of days on ventilation, perhaps consistent with chronic hypoxia or related insults and with fetal growth restriction. The latter has been shown to be associated with a degree of hypoxia and in experimental studies to lead to delayed OL maturation
—recall that the pre-OL is exquisitely vulnerable to hypoxic and related insults (see earlier). The white matter findings also related to prolonged parenteral nutrition and suggest that impaired nutrition may lead to impaired pre-OL development (see later). Importantly, the abnormal FA values in the large study of Barnett et al. were associated with impaired neurodevelopmental performance at age 20 months.
As noted earlier, detection of this milder but prevalent form of WMI cannot be made consistently by conventional MRI. The recent work just described with diffusion-based MRI indicates promise for detection in vivo.
Although the mechanisms for dysmaturation with mild WMI may overlap with those just described for moderate to severe WMI, major differences are likely. Thus, with mild WMI clear evidence for primary injury to components other than the pre-OL is lacking. It is most likely that with mild WMI the deleterious effects of the abundant activated microglia and reactive astrocytes are the dominant mediators of dysmaturation, especially to the pre-OL, and perhaps also to axons.
Pre-OL injury
Primary injury or death of the pre-OL with subsequent replenishment of pre-OLs but failure of maturation, as described for moderate to severe WMI (see earlier), may be the major mechanism for the widespread dysmaturation just described. The important role of activated microglia and reactive astrocytes was discussed earlier concerning moderate to severe WMI. The scenario to widespread dysmaturation, thus, would be similar to that described for moderate to severe WMI (Fig 9A).
Axonal injury
Although evidence for primary injury to axons in mild WMI is lacking, the deleterious effects of the abundant activated microglia and reactive astrocytes may disturb axonal development, separate from any effects on pre-OLs. In addition, as noted earlier, during normal development these glia are critical for axonal guidance and growth, and phenotypic diversion to activated or reactive cells could lead to dysmaturation. Thus a scenario similar to that depicted in Fig 9B seems possible.
Thalamic, subplate, late migrating GABAergic neuron injury
The scenarios described earlier for primary injury to these neural structures leading to dysmaturational gray matter disturbances in the setting of moderate to severe WMI (Fig 9C-E) cannot be ruled out in mild WMI but do not seem highly likely. For example, in the careful neuropathological study of Pierson et al.,
in the 17 infants with DWMG (and no focal necroses), neuronal loss in cortex, thalamus, and basal ganglia was observed in none.
Conclusions
The dysmaturational disturbances of white matter and gray matter structures apparent by advanced MRI methods in infants with mild WMI do not appear to be related to widespread injury. Pre-OL injury and dysmaturation do seem apparent, and thus the possibility of the multiple secondary developmental disturbances of gray and white matter structures described earlier (Fig 9A) is real. The abundant reactive astrocytes and activated microglia in cerebral white matter, i.e., DWMG, are also likely important in the pre-OL injury or dysmaturation. In addition, axonal injury and dysmaturation are also a potential consequence of the deleterious actions of these two glial types (Fig 9B).
Primary dysmaturation of gray matter structures
The possibility that the gray matter structures shown to exhibit secondary impaired development with encephalopathy of prematurity, as outlined in the preceding discussion, may exhibit primary dysmaturation is suggested by recent clinical and experimental studies. If primary dysmaturation does occur, the approaches to neuroprotection and neurorestoration (see later) could be quite different from those directed at secondary dysmaturation in the context of cerebral WMI.
Clinical data
Primary dysmaturation of cerebral cortex, in the absence of evidence for WMI, is suggested by a study of 95 premature infants studied by MRI at two time points in the neonatal period (32 and 40 weeks post-conception).
The principal finding was evidence for delayed microstructural development of cerebral cortical gray matter at multiple sites. Diffusion-based measurements showed delayed microstructural development in cerebral cortex, but not cerebral white matter, in association with impaired somatic growth. The expected normal developmental decline in FA in cortex was blunted, whereas the expected increase in FA in white matter was not. Thus no evidence for WMI or impaired white matter development could be identified. As described earlier,
radial diffusion, and not axial diffusion in the cortex, was particularly affected, again most consistent with impaired dendritic development. The association with impaired somatic growth raises the possibility that undernutrition is particularly involved, although detailed data regarding nutrition, caloric intake, and feeding were not available. However, it is noteworthy that several studies of premature newborns with intrauterine growth retardation also show a particular involvement of cerebral cortical development, including reduced cortical volume, reduced cortical surface area, and impaired gyrification.
Nonetheless, on balance, it does appear that disturbances in growth, perhaps secondary to undernutrition, either in the premature infant postnatally or in utero may have a primary dysmaturational effect on cerebral cortex. More data clearly are needed.
Experimental data
Three recent studies in a well-characterized preterm large animal (fetal sheep) model of cerebral ischemia raise the possibility of primary dysmaturation of cerebral cortex, subplate neurons, and caudate neurons.
Thus utilizing elegant neurobiological methods, Back and coworkers have shown disturbances in cortex, in dendritic development and synapse formation; in subplate neurons, in dendritic arborization and synaptic activity; and in caudate, in dendritic arborization, synaptogenesis, and synaptic activity.
As the basic experimental paradigm was designed originally to replicate cerebral WMI of the premature infant, these examples of neuronal dysmaturation were accompanied by pre-OL degeneration and dysmaturation and diffuse gliosis with reactive astrocytes and reactive microglia. A reasonable question is whether the four-week period required for the evolution of the cortical, subplate, or caudate neuronal dysmaturation is necessary because the dysmaturation is secondary to the pre-OL degeneration and dysmaturation as described earlier (Fig 9A). In the absence of a definitive answer to this question, the possibility that the hypoxic-ischemic insult leads primarily and directly to neuronal dysmaturations is real. Coupled with the clinical study described earlier, the latter possibility demands further research.
Conclusions
The clinical studies of premature infants with impaired somatic growth and of those with intrauterine growth retardation raise the possibility that cerebral cortical development may be affected directly, i.e., primarily, perhaps by nutritional factors. In view of the rapid development of cortex during the premature period and therefore its likely vulnerability to neonatal insults, such a possibility seems reasonable. Experimental data also raise the possibility of a primary dysmaturational effect for hypoxia-ischemia on cortical, subplate, and caudate neurons. However, as discussed, the available data do not rule out a primary effect on pre-OLs with secondary neuronal dysmaturation.
Neuroprotective and neurorestorative interventions
As the pervasive theme in this review is that pre-OL death leads to subsequent dysmaturation of both white and gray matter structures, interventions are best considered (1) as preventative of the initial death (i.e., neuroprotection) or (2) as amelioration or prevention of the subsequent dysmaturation (i.e., neurorestorative). Although there is overlap in this categorization, the distinction best facilitates the discussion that follows.
Neuroprotective interventions
Neuroprotective interventions have focused on prevention of pre-OL injury or death. Many excellent recent reviews have addressed this issue, and therefore this will not be discussed further.
The principal neuroprotective interventions and the likely mechanism(s) affected in the cascade to pre-OL death are shown in Table 3. Most of the mechanisms are also relevant to those examples of WMI that are accompanied by direct injury to axons and neurons as well as to pre-OLs. Of the interventions shown in Table 3, only erythropoietin (EPO) has been studied in detail in human premature infants and will be discussed here.
TABLE 3Neuroprotective Interventions to Prevent Pre-OL Injury/Death
it is a prime candidate for the prevention of pre-OL injury or death, the critical initial event in genesis of preterm WMI. EPO has been shown to prevent or mitigate WMI in a variety of experimental models.
Although numerous studies of EPO in premature infants have been carried out, a recent meta-analysis of four randomized controlled trials (RCTs) comprising 1133 infants is especially useful.
Prophylactic EPO administration reduced the incidence of Mental Developmental Index scores of less than 70 (odds ratio 0.51 [0.31 to 0.81], P < .005) at 18 to 24 months. As the total numbers of infants with less than 28 weeks' gestational age were not large enough to assess adequately the outcome in this critical group, more data are needed. A large multicenter randomized controlled trial in the United States (Preterm Erythropoietin Neuroprotection Trial, NCT01378273) is focused on this critical group, and results should be available this year.
A closer assessment of the key EPO trials suggests that the timing of EPO administration may be critical in the likelihood of benefit. Thus, in one series of studies utilizing early, relatively brief administrations of EPO (atless than 3 hours, at 12 to 18 hours, and at 36 to 42 hours after birth), no significant differences in outcome at two years could be discerned
Effect of early prophylactic high-dose recombinant human erythropoietin in very preterm enfants on neurodevelopmental outcome at 2 years: a randomized clinical trial.
However, in a study utilizing EPO administration (as EPO or its higher glycosylated derivative darbepoetin) thrice weekly through 35 weeks' postconceptual age, the treated infants had better cognitive outcomes and less neurodevelopmental impairment at age 3.5 to four years, when compared with placebo-treated infants.
Thus the two different protocols with regard to the timing of EPO administration suggest that with the early, relatively brief approach, EPO was functioning only as a neuroprotective agent, whereas with the more prolonged approach the agent may have functioned both as a neuroprotective and a neurorestorative intervention. Perhaps consistent with this notion, the largest study to date randomized 800 infants of less than 32 weeks' gestation to placebo or EPO administered intravenously within 72 hours of birth and then once every other day for two weeks.
The rate of moderate or severe neurological disability at 18 months' corrected age was significantly lower in the EPO group (7.1%) versus the placebo group (18.8%) (odds ratio = 0.22, confidence interval, 0.19 to 0.55, P < .001). Dosing in the aforementioned Preterm Erythropoietin Neuroprotection Trial will be still more prolonged, i.e., initially, single doses intravenously, every other day, from day one to day 11, and subsequently, doses subcutaneously every other day until 32 weeks. The potential mechanisms for EPO's benefit concerning brain maturation, i.e., neurorestorative effects, will be discussed in the next section.
Neurorestorative Interventions
The emphasis of this review has been on the evolution of the widespread dysmaturational events that follow the initial insult(s) and injury or death, especially to pre-OLs. These events develop over many weeks to months, and perhaps longer. This relatively protracted period raises the possibility of a long time window for interventions potentially capable of ameliorating or preventing the dysmaturation. I will term these interventions neurorestorative. The principal such interventions, shown in Tables 4 and 5, are classified based on their study in experimental settings only (Table 4) or in clinical settings with human infants, principally preterm infants (Table 5).
Both epidermal growth factor (EGF) and insulinlike growth factor (IGF-1) have beneficial effects in experimental models of preterm WMI (Table 4). The agents appear to exhibit both neuroprotective and neurorestorative properties. In a mouse model of preterm WMI, Scafidi et al. showed that either selective overexpression of human EGF receptor in OL lineage cells or the intranasal administration of EGF immediately after injury led to decreased OL death, enhanced generation of new OLs from progenitor cells, and promoted functional recovery.
The benign mode of administration of the EGF suggests potential clinical applicability.
IGF-1 has shown protective effects versus WMI both in neonatal animal models (hypoxia-ischemia, lipopolysaccharide-induced inflammation) and in cultured pre-OLs.
The agent also showed restorative effects, i.e., rescue of pre-OLs and promotion of myelination. Two issues limit enthusiasm for IGF-1: first, the peptide must be administered intraventricularly, and second, its effects are dose-related, with lower doses being effective but higher doses being toxic.
Hyaluronidase inhibitors
Pre-OL dysmaturation in chronic WMI appears related at least in considerable part to the astrocytic component of the diffuse gliosis characteristic of the lesion. Thus Back and coworkers have shown that reactive astrocytes synthesize high-molecular-weight forms of hyaluronic acid, which are readily detectable in the human lesion.
As described earlier, hyaluronic acid digestion products, generated from hyaluronidases in the disrupted ECM of WMI, lead to a block in pre-OL maturation. This block could be prevented by pharmacologic inhibition of hyaluronidase in vitro and in an animal model.
Whether use of a hyaluronidase inhibitor has value in preventing pre-OL dysmaturation in the human infant requires further study.
Microglial or astrocytic phenotypic manipulation
Abundant microglia are important components of the diffuse gliotic component of WMI (see earlier). These cells are principally in an activated, pro-inflammatory state (M1 phenotype). Their role in acute pre-OL injury or death likely relates to the generation of reactive oxygen and nitrogen species and secretion of injurious cytokines.
However, a variety of studies, performed in in vitro and in vivo models, including adult human lesions with failure of OL differentiation and myelin development (e.g., multiple sclerosis), suggest involvement of activated microglia in the subsequent pre-OL dysmaturation in human preterm WMI.
The data raise the possibility that interventions capable of converting microglia from a pro-inflammatory phenotype (M1) to an anti-inflammatory phenotype (M2) would have major potential as a neurorestorative therapy. Such immunomodulatory agents that cross the blood-brain barrier have been identified (e.g., minocycline, melatonin, minozac, etanercept) and are under study in human adult demyelinating diseases.
These components are short noncoding RNAs (18 to 22 nucleotides), which are transcriptional regulators of gene expression. Several microRNAs have been shown to promote or inhibit inflammatory responses in microglia. One prominent microRNA of activated microglia is mir-155, which is elevated in microglia in multiple sclerosis lesions. When silenced in vivo by intracerebral injection of IL-17 in early stages of experimental allergic encephalomyelitis, the pathologic and clinical effects of the demyelinating disease are blunted.
Thus the possibility of such systemic therapy seems real. Indeed, recent research shows that intravenous delivery of another microRNA (miR-124) that promotes polarization of microglia from an inflammatory (M1) to an anti-inflammatory (M2) phenotype via miR-124-enriched exosomes improves hippocampal neurogenesis and neurological function over four weeks after traumatic brain injury
MiR-124 Enriched exosomes promoted the M2 polarization of microglia and enhanced hippocampus neurogenesis after traumatic brain injury by inhibiting TLR4 pathway.
(see also later discussion of exosomes). Notably, because the anti-inflammatory phenotype of microglia (M2) is important in the facilitation of many brain developmental events as described earlier, these avenues of research suggest a major neurorestorative possibility for in vivo manipulation of microglia phenotype.
Similar considerations concerning glial manipulation from a “harmful” to developmentally “helpful” phenotype apply to the reactive astrocytes in the diffuse gliotic component. Their involvement in pre-OL injury and dysmaturation and the potential value of hyaluronidase inhibitors were discussed earlier. Prevention of the microglial induction of harmful, “neurotoxic” astrocytes (A1 phenotype) is an area of active current research.
The valuable results would be inhibition of pre-OL death and preservation of the maturational effects of the developmentally beneficial astrocytic phenotype (A2). A variety of drugs and neutralizing antibodies (e.g., to tumor necrosis factor-α and IL-α from microglia) to prevent induction of harmful reactive astrocytes are under study in animal models and in adult human neurodegenerative disorders.
Experimental studies of stroke and related ischemic brain injuries in neonatal animals suggest that stem cell therapies may be effective for restoration, particularly of OLs.
Intranasal delivery of bone marrow mesenchymal stem cells improved neurovascular regeneration and rescued neuropsychiatric deficits after neonatal stroke in rats.
The major types of cells used thus far include neural, embryonic, mesenchymal, umbilical cord, and induced pluripotent cells. In vitro manipulation of neural precursor cells before transplantation can enhance their capacity to undergo OL differentiation and axonal remyelination.
A variety of routes of cell administration have been utilized, and intranasal administration may be the most efficient. Stem cells administered by this route appear to target the injury site after entering the brain via olfactory neural processes traversing the cribriform plate.
Studies of rodent models of preterm brain injury have shown that the intranasal route of administration can be effective not only for mitigating injury to myelin but also for improving behavioral outcome.
Of particular relevance to cerebral WMI and pre-OL dysmaturation or myelination failure in the premature infant is a recent study of such injury produced by hypoxia-ischemia in the 3-day-old rat.
OL progenitor cells produced from embryonic stem cells were transplanted into the injured cerebrum. The transplanted cells survived, underwent differentiation, formed myelin sheaths, and stimulated proliferation of endogenous neural stem cells. Supporting a neurorestorative effect was the demonstration of functional benefit after six weeks. It will be of particular interest if the results can be replicated after intranasal administration.
A relevant human study in this context involves the transplantation of human neural stem cells into the brains of four infants with connatal Pelizaeus-Merzbacher disease.
After one year, evidence of myelin ensheathment of axons was obtained by diffusion tensor imaging. Direct extrapolation to the human preterm infant with WMI is difficult because of the mode of administration used. Nevertheless, the findings suggest promise for stem cell therapy as a neurorestorative therapy in such infants.
Exosomes
The precise neuroprotective factors released by stem cells are not known with certainty and may vary as a function of the injury. Notably, however, effects on pre-OL and myelin development and on behavioral outcome in a rodent model of preterm WMI was achieved with intravenous administration of extracellular vesicles, i.e., exosomes, derived from stem cells, in lieu of stem cells per se.
Mesenchymal stem cell-derived exosomes provide neuroprotection and improve long-term neurologic outcomes in a swine model of traumatic brain injury and hemorrhagic shock.
The great therapeutic potential of exosomes, isolated from blood, has been recognized only recently, and the capacity to induce OL differentiation and myelination could serve a crucial neurorestorative function in the premature infant. Human studies will be of great interest.
Dendrimers
Dendrimers are unique nanoparticles synthesized for a variety of functions, including targeted delivery of therapeutic agents to brain.
Their small size and tailorable surface functional groups make them valuable for this role. Drugs, and perhaps ultimately, microRNAs or silencing RNAs can be attached to the dendrimer. Several recent models of ischemia- or inflammation-induced neonatal or fetal brain injury have shown marked beneficial effects of dendrimer–N-acetylcysteine conjugates.
Uptake of dendrimer-drug by different cell types in the hippocampus after hypoxic-ischemic insult in neonatal mice: Effects of injury, microglial activation and hypothermia.
N-acetylcysteine is an antioxidant, and after intravenous administration of the conjugate, uptake into activated microglia, reactive astrocytes, and differentiating OLs could be demonstrated. Sustained prevention of OL injury and improved myelination were shown.
The principal cellular target appeared to be inflammatory microglia. Further studies will be of great interest.
Clinical studies
A burgeoning clinical literature suggests the possibility that the dysmaturation of both pre-OLs and gray matter structures after premature brain injury can be counteracted to a considerable extent. These neurorestorative interventions include pharmacologic agents, i.e., EPO, and modifications of nutrition and other environmental factors (Table 5). Implementation of these interventions during the vulnerable neonatal period, when the remarkable array of developmental events described earlier are proceeding most rapidly, is of critical importance. However, the beneficial effects of these interventions likely continue beyond
30107-9&id=gr1.jpg){kind=link}
30107-9&id=gr2.jpg){kind=link}
30107-9&id=gr3.jpg){kind=link}
30107-9&id=gr4.jpg){kind=link}
30107-9&id=gr5.jpg){kind=link}
30107-9&id=gr6.jpg){kind=link}
30107-9&id=gr7.jpg){kind=link}
30107-9&id=gr8.jpg){kind=link}
30107-9&id=gr9.jpg){kind=link}