Epidemiological studies have shown a strong association between perinatal infection/inflammation and brain damage in preterm infants and/or neurological handicap in survivors. Experimental studies have shown a causal effect of infection/inflammation on perinatal brain damage. Infection including inflammatory factors can disrupt programmes of brain development and, in particular, induce death and/or blockade of oligodendrocyte maturation, leading to myelin defects. Alternatively, in the so-called multiple-hit hypothesis, infection/inflammation can act as predisposing factors, making the brain more susceptible to a second stress (sensitization process), such as hypoxic–ischaemic or excitotoxic insults. Epidemiological data also suggest that perinatal exposure to inflammatory factors could predispose to long-term diseases including psychiatric disorders.
- cerebral palsy
- perinatal brain damage
Background to perinatal brain injuries
Damage to the immature brain around the time of birth presents a serious public health problem across the world. The Global Burden of Disease Study estimated that, in 2010, premature birth was the commonest cause of death under the age of 5 years, causing more deaths than malaria or pneumonia, and the loss of 77 million [95% CI (confidence interval) 66–88 million] DALYs (disability-adjusted life years), predominantly due to neurological damage. Birth asphyxia caused a further 0.5 million deaths and 45 million (95% CI 402000–619000) DALYs .
In the developed world, prematurity is increasing, now accounting for 13% of all births in the U.S.A. and 7% in the U.K. Almost 10% of infants born before 33 weeks develop cerebral palsy and approximately 35% have persisting cognitive and neuropsychiatric deficits [2,3], and, although more extreme prematurity leads to the most severe problems, even slight reductions in the period of intrauterine life has significant adverse effects [4,5]. There is now a crisis in the provision of long-term education and healthcare for this increasing number of children . Equally, advanced obstetrics has reduced, but not abolished, birth asphyxia, which in the U.K. leads to death or severe neurological impairment in 1 in 2000 births .
The cost of this reproductive casualty to individuals, families and society is very high: lifetime costs for the care of one child affected by cerebral palsy is approximately 1 million U.S. dollars , highlighting the need to develop new treatments and preventive strategies. A significant breakthrough has been made recently by research showing that hypothermia in infants diagnosed with HIE (hypoxic–ischaemic encephalopathy) (based on low Apgar and blood gasses) reduces brain injury, doubling the chance of normal survival . As well as saving a large number of individuals from death or neurodevelopmental disability, this simple intervention provides proof of concept that therapeutic interventions to treat brain injury in this vulnerable population can be effective. As such, investment in the search for the next generation of brain-protecting treatments for both preterm and asphyxia brain damage is a healthcare, societal and economic priority.
Designing such neuroprotectants requires a detailed knowledge of the evolving clinical phenotype and pathophysiology in this vulnerable population. Critically, improvements in factors including antenatal care and fetal monitoring mean that major changes have recently been observed in the panorama of brain damage and neurological consequences observed in preterm infants.
Brain damage associated with preterm birth
Owing to the improvements in perinatal care, at the clinical level, severe motor deficits are less frequent , whereas fine motor deficits, cognitive and learning impairments, behavioural disturbances and sensory deficits have become more prominent over the last 20 years . At the structural level, both imaging and neuropathological studies have shown that focal destructive lesions including HPI (haemorrhagic parenchymal infarction) and cystic PVL (periventricular leukomalacia) are less frequent than subtle, but diffuse, white matter abnormalities [11–13]. White matter injury predominates with lower gestational age, but, as above, recent neuropathological studies have shown a switch over the last 20 years from a loss of pre-oligodendrocytes to an excess of immature oligodendrocytes combined with a deficit of myelinating oligodendrocytes, suggesting oligodendrocyte maturation blockade [14,15]. In addition, although research on the white matter has been the focus in premature infants, recent imaging studies have shown subtle grey matter abnormalities including reduced thalamocortical connectivity, reduced intracortical connectivity and altered cortical microstructure [16,17]. Supporting an involvement of grey matter, neuropathological studies have revealed axonopathy and loss of cortical interneurons in some preterm infants [11,18,19].
Role of inflammation in perinatal brain damage
Together with the changing panorama of injury, on the pathophysiological level over recent years, the hypothesis of a purely hypoxic–ischaemic cause of injury has been replaced by a multifactorial hypothesis where systemic inflammation appears to play a key role [20,21]. Most epidemiological studies suggest a strong association between fetal infection/inflammation (chorioamnionitis) and brain damage, especially white matter injury, in the premature new born and neurological disability in survivors. In particular, studies are demonstrating not only that inflammation is present at birth, but also that the persistence of this process (for at least 14 days) is associated with poorer outcome . A wealth of experimental studies have demonstrated a causal effect of infection/inflammation in perinatal brain damage due to exposure alone or as part of a multiple-hit paradigm (discussed below and in [23–28]).
In the so-called multiple-hit hypothesis (Figure 1), infection/inflammation can act as predisposing factor (sensitizer), making the brain more susceptible to a second stress [29–31]. Indeed, injection of low-dose LPS (lipopolysaccharide) (that does not induce apparent brain damage by itself) to developing rats makes the newborn brain extremely susceptible to a very mild hypoxic–ischaemic insult (that does not induce apparent brain damage by itself)  (Figure 2). Similarly, injection of IL-1β (interleukin 1β) to newborn mice or rats makes the brain much more sensitive to an excitotoxic insult . Studies have evaluated the time window during which sensitization of the brain to hypoxia–ischaemia persists after exposure to inflammatory factors in the perinatal period [33,34]. These studies and others from our group illustrate that the presence of a systemic inflammatory challenge before hypoxia–ischaemia can strongly sensitize in the rodent . These effects do not appear to be due to differences in temperature or cerebral blood flow [32,34]. The mechanisms by which sensitization is working are not yet fully understood, but seem to include COX2 (cyclo-oxygenase 2) activation and excess production of PGE2 (prostaglandin E2) leading to neuroinflammation, increased metabotropic glutamatergic receptor activity linked to decreased expression of GRK2 (G-protein-coupled receptor kinase 2) and changes in gene transcription [33,35,36]. Furthermore, there are data suggesting that Toll-like receptor activation of the MyD88 (myeloid differentiation factor 88)-dependent pathway, leading to cytokine production may be involved [37,38].
Recent evidence also strongly supports the hypothesis that infection/inflammatory factors can induce brain damage by themselves, in the absence of a secondary hit. Accordingly, injection of Escherichia coli into pregnant rabbits induces periventricular white matter cysts and widespread white matter cell death, mimicking brain damage observed in preterm infants . In addition, injection of Ureaplasma parvum into pregnant mice induces myelin defects and loss of interneurons in the offspring . Similarly, injection of LPS to pregnant rats or mice induces transient central inflammation, white matter cell death, myelination defects and behavioural deficits in the offspring [41,42].
These latter studies provide the proof of concept that infection/inflammation can alter programmes of brain development. However, these studies rely on relatively severe infectious or inflammatory stimuli that probably do not reflect the levels of systemic inflammation observed in many human preterm infants. In this context, we have aimed to test the hypothesis that moderate systemic inflammation is sufficient to alter white matter development. Consequently, we developed a model where newborn mice received twice-daily intraperitoneal injections of IL-1β over 5 days and were studied for myelination, oligodendrogenesis, behaviour and MRI, in order to match this model with human data [23,43] (Figure 3). Mice exposed to IL-1β had a long-lasting myelination defect that was characterized by an increased number of non-myelinated axons. They also displayed a reduction in the diameter of the myelinated axons. In addition, IL-1β induced a significant reduction in the density of myelinating oligodendrocytes accompanied by an increased density of oligodendrocyte progenitors, suggesting a partial blockade in the oligodendrocyte maturation process. Accordingly, IL-1β disrupted the co-ordinated expression of several transcription factors known to control oligodendrocyte maturation. These cellular and molecular abnormalities were correlated with a reduced white matter fractional anisotropy on diffusion tensor imaging and with memory deficits. This original and clinically relevant model shows that moderate perinatal systemic inflammation during a period approximating 28–35 weeks gestation in the human alters the developmental programmes of the white matter. Additional data support the hypothesis that systemic inflammation impairs oligodendrocyte maturation through neuroinflammatory processes, including microglial activation [11,44,45], as understanding of microglial activation, i.e. phenotypes, and their roles in brain injuries is increasing [21,46–48]. Of note, a link between microglia phenotypes and myelination in adult CNS (central nervous system) suggest that microglia of a specific regenerative phenotype (characterized by CD206 and arginase 1 expression) regulate myelination . Ongoing studies address the molecular cross-talk between activated microglia and developing oligodendrocytes.
Long-term effects of perinatal systemic and brain inflammation: tertiary phase
Epidemiological data also suggest that perinatal exposure to inflammatory factors could predispose to long-term diseases including psychiatric disorders such as autism and schizophrenia . This could be particularly the case for preterm infants whose brains could be more sensitive to environmental factors when compared with full-term infants. In addition to developmental disruption associated with the initial insult to the immature brain, injury processes may persist for many months or years [51,52] (Figure 4). This tertiary phase has been demonstrated following adult brain injuries and models of injury [53–55]. We suggest that these tertiary mechanisms of damage might include persistent inflammation and epigenetic changes . We propose that these processes are implicit in prevention of endogenous repair and regeneration and predispose patients to development of future cognitive dysfunction and sensitization to further injury. If this hypothesis is true, treatment of tertiary mechanisms of damage might be possible by various means, including preventing the repressive effects of microglia over-activation and recapitulating developmentally permissive epigenetic conditions. Delayed therapy approaches have been effective against adult brain injuries and models [57–61], and increased study of a tertiary inflammatory process will open important avenues for therapeutic interventions following perinatal inflammatory brain damage.
Altogether, these clinical and experimental data strongly support the hypothesis that exposure to infection/inflammation during pregnancy or the perinatal period is deleterious for the developing brain. When combined with a second hit, this can lead to the appearance of perinatal brain damage that is accompanied by long-term neurological, cognitive and behavioural disabilities. In addition, some data also support the hypothesis that perinatal exposure to inflammatory factors can alter the programmes of brain development, leading to long-term deficits as well. Deciphering the underlying cellular and molecular mechanisms and the window for intervention should lead to the discovery of new therapeutic targets.
The research of the authors is supported by the Inserm (to O.B. and P.G.), Paris Diderot University (to O.B. and P.G.), the PremUP Foundation (to O.B. and P.G.), the Seventh Framework Programme of the European Union [grant number HEALTH-F2-2009-241778/Neurobid (to P.G.)], the Indo-French Centre for the Promotion of Advanced Research [CEFIPRA project number 4903-H (to P.G.)], the Leducq Foundation (to H.H. and P.G.), the Gueules Cassées Foundation (to V.D., S.S. and P.G.), the de Spoelberch Foundation (to P.G.), the Grace de Monaco Foundation (to P.G.), Medical Research Council (Sweden) [grant number 2012-2642 (to H.H.)], ALF-LUA [grant number ALFGBG2863 (to H.H.)], Wellcome Trust [programme grant number WT094823MA (to H.H. and P.G.)], SPARKS Foundation (to P.G. and H.H.), and the Assistance Publique-Hôpitaux de Paris (‘contrat hospitalier de recherche translationnelle’ to P.G.).
Brain Disorders Across the Lifespan: Translational Neuroscience from Molecule to Man: An Independent Meeting held at University College Cork, Ireland, 12–13 September 2013. Organized and Edited by Eoin Fleming (University College Cork, Ireland).
Abbreviations: CI, confidence interval; DALY, disability-adjusted life year; IL-1β, interleukin 1β; LPS, lipopolysaccharide; PVL, periventricular leukomalacia
- © The Authors Journal compilation © 2014 Biochemical Society