Models of Fetal Brain Injury, Intrauterine Inflammation, and Preterm Birth

Authors


Correspondence

Irina Burd, MD, PhD, Maternal-Fetal Medicine, Department of Gynecology and Obstetrics, Johns Hopkins University, 600 North Wolfe Street, Baltimore, MD 20905, USA.

E-mail: iburd1@jhmi.edu

Abstract

Intrauterine infection and inflammation are known risk factors for brain damage in the neonate irrespective of the gestational age. Infection-induced maternal immune activation leads to a fetal inflammatory response mediated by cytokines that has been implicated in the development of not only periventricular leukomalacia and cerebral palsy but also a spectrum of neurodevelopmental disorders such as autism and schizophrenia (Behav Brain Res 2009; 204:313, Ann Neurol 2005; 57:67, Am J Obstet Gynecol 2000; 182:675). A common link among the neurobehavioral disorders associated with intrauterine inflammation appears to be the evidence for immune dysregulation in the developing brain (Behav Brain Res 2009; 204:313). The timing of the immune challenge with respect to the gestational age and neurologic development of the fetus may be crucial in the elicited response (J Neurosci 2006; 26:4752). Studies involving animal models of maternal inflammation serve a key role in elucidation of mechanisms involved in fetal brain injury associated with exposure to the maternal milieu. These animal models have been shown to result in fetal microglial activation, neurotoxicity as well motor deficits and behavioral abnormalities in the offspring (J Neurosci 2006; 26:4752, J Neurosci Res 2010; 88:172, Am J Obstet Gynecol 2009; 201:279, Am J Obstet Gynecol 2008; 199:651). A better understanding of the mechanisms of perinatal brain injury will allow discoveries of novel neuroprotective agents, better outcomes following preterm birth and stratification of fetuses and neonates for therapies in cases of preterm birth, preterm premature rupture of membranes, and chorioamnionitis.

Maternal and fetal immune dysregulation and impaired fetal brain development

The maternal immune system is regulated and maintained in an optimal balance during pregnancy such that it does not trigger rejection of the fetus because of a robust immune response, while adequately protecting the mother and fetus from pathogens and other environmental stimuli.[1, 2] Cytokines are expressed constitutively by the normal developing brain that suggests a role during development.[3, 4] Maternal cytokine-associated inflammatory response appears to be the link in the relationship between infections during pregnancy, and the development of cerebral palsy (CP) and other neuropsychiatric conditions such as schizophrenia and autism.[3, 5-7] Multiple animal experiments have shown that irrespective of the type of pathogen, and even in the absence of a pathogen, cytokine releasing treatment during pregnancy is probably the link in producing fetal brain injury.[8-12]

Any dysregulation of the normal expression of cytokines in the fetal brain may affect neurodevelopmental processes. This immune alteration may occur by a transplacental passage of cytokines produced by the maternal peripheral immune system, through a placental production and secretion of cytokines, or by fetal production of cytokines.[13] In the presence of infection or any inflammatory stimulus, the glial cells in the fetal brain can also produce pro-inflammatory cytokines.

The fetal immune system is known to be relatively poorly developed in the early-/mid-gestation period and becomes functionally mature later in gestation. This correlates with an ‘inverse developmental regulation’ of serum immune-inflammatory protein levels that decrease with increasing gestational age. The increased concentration of inflammatory proteins earlier in gestation is hypothesized to be protective because at this age, the fetus is at greater risk of infections because of the deficits in innate and adaptive immunity. Any alteration/imbalance in this balance of inflammatory proteins may make the developing brain more prone for injury.[14]

In the United States, approximately 12% of all live births are delivered preterm. Intrauterine inflammation in these cases has been demonstrated to be a major cause of adverse neurologic outcomes, including CP. Recent studies demonstrated that children born preterm also have significant cognitive and behavioral deficits. Corollary to that, it has been demonstrated that intrauterine inflammation results in fetal neuronal injury.[15, 16] In our studies comparing animal models of inflammation-induced preterm birth (lipopolysaccharide) with non-infectious induced preterm birth (RU486, an anti-progestational agent), we have demonstrated that it is not preterm birth per se that results in fetal neuronal aberrations but the exposure to intrauterine inflammation and the development of fetal neuroinflammation.[15]

Intrauterine inflammation associated with preterm birth, placental inflammation, or premature rupture of membrane may lead to a systemic inflammatory condition in the fetus called fetal inflammatory response syndrome (FIRS) through an alteration of pro-inflammatory cytokine levels.[17, 18] Maternal intrauterine infection/inflammation resulting in FIRS has been implicated in the development of periventricular leukomalacia (PVL) leading to motor and cognitive impairments in the infant. FIRS is associated with elevated levels of inflammatory cytokines such as interleukin (IL) 6, tumor necrosis factor (TNF) α receptors in the fetal plasma and high cerebrospinal fluid (CSF), and cord blood levels of IL-6 and IL-8.[17, 19, 20] Nelson et al.[21] reported a close association of elevated inflammatory cytokines in the neonatal blood and development of CP. Kadhim et al. [22] demonstrated that in PVL cases with infection, an increased infiltration of macrophage and high levels of IL-1β and TNF-α could be detected from early stages of pathologically identified PVL to later stages of cystic cavitation. The expression of TNF-α in PVL is particularly abundant in microglia surrounding necrotic foci in focal PVL[23] and further suggests that microglia (and astrocytes) may act as central candidates for the initiation of pathology and damage to the surrounding oligodendrocytes and neurons, via the production of cytokines.[24, 25]

Evidence of maternal infection/inflammation and brain injury

Epidemiologic studies indicate that maternal infections (viral, bacterial, or parasitic) are associated with an increased incidence of schizophrenia.[26, 27] Similarly, recent studies have demonstrated an increase in the incidence of autism spectrum disorders following exposure to prenatal infection.[1, 8, 28, 29] An increased pro-inflammatory cytokine profile (IFN-gamma and IL-6, respectively) was noted in the serum during mid-gestation in women who subsequently gave birth to a child with autism or developmental delay.[1]

Leviton and Gilles[30] first described the detrimental effect of bacterial endotoxin on myelinogenesis and white matter injury in autopsy samples of infants who died with perinatal telencephalic leukoencephalopathy. Elevated levels of cytokines have been observed in the neonatal blood and amniotic fluid of children with cerebral palsy.[21, 31] A meta-analysis study on the incidence of chorioamnionitis with CP reported that, in preterm infants, clinical chorioamnionitis was significantly associated with both CP [Relative risk (RR), 1.9; 95% CI 1.4–2.5] and cystic periventricular leukomalacia (cPVL) (RR, 3; 95% CI 2.2–4) while histological chorioamnionitis significantly associated with cPVL (RR, 2.1; 95% CI 1.5–2.9). In term infants, a positive association was found between clinical chorioamnionitis and CP (RR, 4.7; 95% CI 1.3–16.2). This study concluded that 28% of the preterm infants with chorioamnionitis may develop CP while the incidence of CP in term infants with a history of chorioamnionitis is 12%.[32]

Hypertrophic astrocytes and activated microglia are considered to be markers of white matter damage.[31, 33] Autopsy studies in patients with PVL have shown to express hypertrophic astrocytes, TNF α, and βAPP in the white matter region of the brain.[34] The presence of TNFα, IL1β, and IL6 was observed in 88% of the patients with PVL, and these cytokines were concentrated in the hypertrophic astrocytes and activated microglial cells.[31]

Timing of injury

The timing of maternal infection during pregnancy appears to play a crucial role in the neurodevelopmental response/outcome of the offspring. Distinct neurodevelopmental programs are affected at different stages, resulting in different responses.

An infection occurring in the second trimester may disrupt the migration of young post-mitotic neurons from the ventricular zone to the neocortex. Destruction of the existing neurons, or interference with migration or development of new neurons leads to an aberrant cortical development.[35] The presence of immature oligodendrocytes and ameboid microglia in the white matter region of the brain[36, 37] during the late second trimester may contribute to increased vulnerability of the brain to pathologic insult.[13, 38] Thus, the presence of inflammatory cytokines during the critical window may adversely affect the developing brain leading to abnormal brain development with cognitive and behavioral impairments.

Mayer et al. reported that the fetal brain has a differential effect to maternal infection/inflammation during the mid- and late-gestational period. Mid-gestational maternal infection resulted in suppressed exploratory behavior, while late gestational infection resulted in perseverative behavior that is implicated in schizophrenia, autistic spectrum disorders (ASD), addictive behavior as well as obsessive compulsive disorders (OCDs).[3] The type of infection along with timing may result in a different inflammatory response as reported recently where viral infections during the first trimester and bacterial infections during the second trimester were associated with the development of ASD in the offspring.[29]

Animal models

It is difficult to have an animal model that is clearly representative of perinatal brain injury in humans, as the developmental trajectories of glia, oligodendrocytes, and neurons show significant variability between species. Each species offers a unique window into a development of the human fetal brain, dependent on structural similarities and cell-type similarities in development. Dissecting human brain development by studying the insult across several species offers several advantages. Timing of injury relative to the development of various structures in the brain across various species would be crucial in predicting the outcome and in correlating with human development.

One of the first animal models of endotoxin-mediated white matter injury was demonstrated by Gilles in 1977. Injection of Escherichia coli endotoxin into the newborn kittens resulted in necrosis of white matter region followed by astrogliosis and macrophage infiltration. Cystic necrosis was also observed in thalamus and caudate region of the brain in kittens with severe damage.[39] An intense inflammatory response in the gray and white matter was observed in neonatal dogs exposed to the endotoxin on postnatal day 1–10. However, these neonatal dogs had necrotic lesions only in the forebrain white matter.[40]

Recently developed models aim to connect the impact of maternal inflammation and fetal exposure to the inflammatory milieu (animal models of maternal inflammation-induced fetal injury are depicted in Table 1). Yoon et al.[41] introduced E. coli into the cervix of pregnant rabbits on gestational day 21/22 and observed that all kits with evidence of intrauterine infection had white matter damage. Debillon et al.[42] injected E. coli into the uterine horns of pregnant rabbit on gestational day 27 and 28 and observed 100% placental inflammation and multifocal programmed cell death[42] and macrophage infiltration in the white matter of live fetuses after 48 hr of inoculation. In our animal model of maternal inflammation-mediated injury in the rabbit kits, we have observed microglial infiltration into the white matter region of the brain followed by decreased myelination and motor deficits on postnatal day 1.[43, 44]

Table 1. Models of Maternal Inflammation and Perinatal Brain Injury
SpeciesGestation (GA)Organism/agentMode of injectionHistology in offspringBehavioral outcome in offspringReferences
Mouse9 (GA) (corresponds to middle or late second trimester in human)Human influenza virus (H1N1)IntranasalDecrease in hippocampal and cortical thickness and decreased reelin Fatemi et al.[61]
Mouse9.5 (GA)Human influenza virusIntranasal Deficit in pre-pulse inhibition, acoustic reflex, exploratory novel object task and social interactionShi et al.[12]
Mouse9, 17 (GA)PolyI:CIntravenousGray matter injuryDeficit in open field and perseverative behaviorMeyer et al.[3, 6]
Mouse18 (GA)Human influenza virus (H1N1)IntranasalBrain atrophy, thinning of corpus callosum, decreased serotonin Fatemi et al.[62]
Mouse16 (GA) (Corresponds to middle second trimester in human, period of neurogenesis in hippocampus)Human influenza virus (H1N1)IntranasalDecreased hippocampal volume Fatemi et al.[63]
Mouse15 and 18.5 (GA)LPSUterine wallGray matter injury Burd et al.[15, 16, 45] Elovitz et al.[64]
Rat18 (GA)LPSIntraperitonealWhite matter injury Cai et al.[11]
Rat15 (GA) (corresponds to late onset human chorioamnionitis and critical period for glial development in rat brain)LPSIntracervicalWhite matter injury Bell et al.[9]
Rat15 (GA) (Corresponds to second trimester of human pregnancy)PolyI:CIntravenous Loss of latent inhibition, rapid reversal learning, and increased sensitivity to locomotor effect followingZuckerman et al.[65]
Rat17 (GA)LPSIntraperitonealGray matter injury Larouche et al.[66]
Rabbit21–22 (GA)Gardnerella vaginalisUterine hornsGray matter injury Field et al.[67]
Rabbit20–21 (GA)Escherichia coliIntracervicalWhite matter injury Yoon et al.[31, 41]
Rabbit27–28 (GA)E. coliUterine hornsWhite matter injury Debillon et al.[42]
Rabbit28 (GA) (Period of peak microglial activation and immature oligodendrocyte density)LPSUterine wallWhite and gray matter injuryMotor deficitsKannan et al.[44], Saadani et al.[43]

Chorioamnionitis at term has a higher incidence of CP compared with preterm infants.[32] A mouse model of intrauterine inflammation was created by intrauterine injection of lipopolysaccharide (LPS; an endotoxin and a component of a cell wall of gram-negative bacteria) on gestational day 18.5 (corresponding to the end of murine gestation).[45] Six hours later, the fetus had significantly high expression of pro-inflammatory cytokines such as IL1β and IL6 and decreased expression of microtubule-associated protein 2 (MAP2), a key cytoskeletal protein in developing neurons. A significant reduction in the dendritic arborization of MAP2 stained neurons was also observed in the cortical neuronal culture.[45] This study underlines the role of fetal neuronal injury in response to intrauterine inflammation and correlates the result to the adverse neurobehavioral outcomes such as cerebral palsy, ASD, schizophrenia, observed in children exposed to intrauterine inflammation.

Probable mechanisms of injury

Pro-inflammatory cytokines may cause (i) a direct injury to oligodendrocytes and neurons; (ii) a secondary injury through an activation of microglial cells in the white matter tracts that are normally present during development for remodeling and growth; once activated, microglia may secrete pro-inflammatory cytokines and free radicals resulting in injury to the surrounding cells; or (iii) a combination of both direct injury and secondary injury through activated microglia and astrocytes (Fig. 1). The microglial activation component may persist for a period of time resulting in ongoing neuroinflammation as seen in patients with ASD and PVL.[28, 46]

Figure 1.

Probable mechanisms of fetal brain injury with in utero exposure to maternal inflammation.

In recent years, a number of in vitro and in vivo studies have implicated microglial cells in the development of PVL. Haynes et al. (2003)[47] have shown the increased presence of activated microglia diffusely throughout the white matter in autopsy specimens of patients with PVL, indicating that activated microglia are involved in causing white matter damage by oxidative and nitrosative stress. Pro-inflammatory cytokines such as IL-1β has been shown to induce microglial activation in vitro.[48] LPS treatment results in oligodendrocyte damage only in the presence of microglial cells supporting the role of microglial cell activation in white matter injury.[49]

Activated microglial cells may induce injury by (i) producing a variety of pro-inflammatory cytokines, many of which are cytotoxic[50]; (ii) producing excitotoxic metabolites such as glutamate and quinolinic acid, which may cause glutamate receptor, or NMDA receptor mediated injury to oligodendrocytes,[51] and by (iii) releasing oxidative and nitrosative products.[52] The low resistance of oligodendrocytes to oxidative stress, presence of calcium permeable glutamate receptors, and presence of NMDA receptors in their myelinating processes make them highly susceptible for this injury.[50, 53, 54]

Exposure to maternal intrauterine endotoxin administration in rabbits results in a phenotype of CP along with a robust microgial activation in the periventricular white matter tracts along with a phenotype of cerebral palsy in the newborn rabbit. The extent of motor deficits appears to correlate with the extent of neuroinflammation as measured by PET imaging using (11C) PK11195. Endotoxin administration was also shown to be associated with decreased cortical serotonin in the neonate and impaired formation of serotonin-regulated projections to the sensory cortex. Impaired development of the sensory cortex and decreased serotonin are seen in patients with ASD.[44, 55-57] These studies indicate that ongoing activation of microglial cells may be responsible for the development of white matter and neuronal injury in infants exposed to intrauterine infections/inflammation (Fig. 2).

Figure 2.

Representative images of glia and neurons in the control and endotoxin group of rabbit kit on day 1 of life. Maternal intrauterine infection leads to activated microglia (lectin stain; arrows), astrogliosis (GFAP stain; arrows), and loss of oligodendrocytes (MBP stain) in the white matter region of the rabbit kit brain as observed on day 1 of life. Cresyl violet shows that white matter injury is also accompanied by increased neuronal death in the hippocampus (arrows).

Increased expression of cytokines most likely produced by activated microglia (namely TNF-α) was demonstrated in brains of developing fetuses after an inflammatory stimulus, and in autopsy specimens of children with PVL.[10] TNF-α was shown to decrease the number of oligodendrocyte progenitors, causing apoptosis of oligodendrocytes.[10] Incubation of hippocampal slices, from mature fetal guinea pig, with LPS increased their capacity to secrete TNF-α that may induce apoptotic processes in oligodendrocytes and their progenitors.[58] In addition, the combined application of TNF-α and interferon-γ severely reduced survival and inhibited differentiation of oligodendrocyte progenitors in primary culture prepared from neonatal rats, thus may underlie the disrupted myelination that characterizes PVL.[59] Higher levels of B-lymphocyte chemo-attractant, ciliary neurotrophic factor (CNTF), epidermal growth factors, IL-12, IL-15, monocyte chemo-attractant protein-3, and others were demonstrated in sera of children with CP.[59]

O'shea et al.[60] measured the concentration of 25 different inflammation-related protein on day 1, day 7, and day 14 from 939 infants with gestational age <28 weeks. They observed a significant association with increase in CRP on day 7 and an elevated levels of IL1β, IL6, TNFα, IL8, adhesion molecules such as ICAM-1, CRP, SAA, and MPO on day 14 with a Mental and Motor scale score of <55 at 2 years of age. This study support the hypothesis that systemic inflammation in the infants near term of birth can result in mental and motor disturbance at later stages in life. This study suggests that these biomarkers in neonatal blood may be used as predictors for inflammation and maybe useful for interventions that target inflammation mediated brain injury.

Future for therapy

Maternal infection/inflammation results in an increase in proinflammatory cytokines in the fetal brain with release of free radicals leading to a cascade of injury to developing oligodendrocytes and neurons through an activation of microglia. The presence of ongoing neuroinflammation in the postnatal period provides a window of opportunity for treatment in maternal inflammation-induced brain injury. Possible therapeutic intervention may include prenatal or postnatal therapy.

Therapies targeted to activated microglia and astrocytes may in the future attenuate the injury. Similarly, anti-cytokine treatments may help reduce the extent of the injury. Further research in this area is urgently needed to bring all of the available research discoveries into clinical practice. Better understanding of the mechanisms of maternal inflammatory response and fetal neuroinflammation, and discoveries of novel therapeutics will lead to better outcomes in neonatal intensive care units and quality of life for children exposed to intrauterine inflammation in utero.

Acknowledgements

This work was supported by the American Board of Obstetrics and Gynecology/American Association of Obstetricians and Gynecologists Foundation (ABOG/AAOGF) Grant (IB).

Ancillary