This study shows that IL-1β inhibition during progressive systemic LPS-induced inflammation in near-term fetal sheep reduced microgliosis and apoptosis, and improved survival of oligodendrocytes in the large white matter tracts. The reduction in neuroinflammation was associated with reduced circulating pro- and anti-inflammatory cytokines and improved recovery of EEG power and fetal movement after LPS-exposure.
Clinically, perinatal infection/inflammation is associated with a high risk of neonatal mortality and morbidity. Moreover, in cases of perinatal infection/inflammation, systemic upregulation of IL-1β is associated with increased risk of short and long term neurodevelopmental impairment after birth [29, 30]. Increased IL-1β expression has been detected in the cerebrospinal fluid of term neonates with encephalopathy, and was strongly associated with impaired neurodevelopmental outcomes [31]. Furthermore, at post-mortem, neonates with white matter injury showed increased IL-1β expression localised to areas of white matter gliosis [15]. Similarly, increased circulating levels of IL-1 are associated with acute white matter injury and impaired neural metabolism [19, 20]. These data demonstrate a strong association between elevated systemic and central IL-1β production and perinatal brain injury. However, it remains unknown whether this association is causal. Critically, using a large animal translational model of perinatal infection/inflammation at term, this study shows that IL-1β plays an important role in the pathophysiology of white matter inflammation and injury, and suggests that targeted systemic inhibition may improve histological and functional outcomes.
Consistent with previous studies from our laboratory and others, LPS infusions were associated with a systemic inflammatory response as shown by elevated cytokine levels, systemic hypotension and tachycardia [7, 27, 32–34]. Repeated LPS infusions were associated with tolerance to subsequent doses indicating reprogramming of the innate immune system. In human and sheep monocytes, repeated LPS exposure is associated with decreased cytokine production and downregulation of the LPS receptor CD14[35–37]. Consistent with these findings, in vivo studies in fetal sheep have shown that repeated LPS exposure is associated with attenuation of systemic inflammation [27, 34, 38].
In the present study, we used the commercially available IL-1Ra, Anakinra, to inhibit IL-1 mediated systemic and central nervous system inflammation in near-term fetal sheep. Anakinra is a recombinant non-glycosylated form of the human IL-1Ra and has been FDA approved for treatment of chronic inflammatory conditions in adults and children. It exerts its physiological effects by binding to the IL-1 receptor and neutralising the effects of IL-1 to prevent downstream inflammatory signalling [39]. It has a molecular weight of 17 kDa and is capable of penetrating the blood brain barrier in humans and sheep [26, 40]. To the best of our knowledge, the temporal profile of circulating cytokines has not been assessed in the setting of IL-1Ra and systemic inflammation in the near-term fetus. Infusion of IL-1Ra starting 1 hour after LPS-induced inflammation led to a sustained reduction in circulating IL-6, from 6 h after the first LPS infusion, and reduced circulating IL-1β, TNF and IL-10 concentrations after the second LPS infusion. These data are consistent with in vitro and in vivo studies that reported inhibition of pro-and anti-inflammatory cytokines after IL-1Ra administration in adults with chronic inflammatory disease [41, 42], fetal sheep exposed to intra-amniotic LPS [43] and neonatal mice exposed to antenatal LPS and/or postnatal hyperoxia [44]. Collectively, these data demonstrate exogenous IL-1Ra can modulate systemic pro-and anti-inflammatory cytokine production in the fetus and neonate.
Elevated circulating levels of IL-1β are associated with impaired cerebral oxidative metabolism [18] and EEG suppression in neonates [45]. Similarly, in the present study we observed suppression of EEG power and nuchal EMG activity (reflecting reduced neural activity and fetal movement, respectively) after the first LPS infusion, and sustained reductions in neural activity and fetal movement during the recovery period. The suppression of EEG power and fetal movement may reflect inhibition of synaptic activity due to increased local cytokine production and/or hypoxia. Indeed, suppression of EEG activity and fetal movement after the first LPS infusion was associated with mild reductions in arterial PaO2 and SaO2. Inflammation and cerebral hypotension/hypoperfusion can trigger active EEG suppression through release of inhibitory neuromodulators and neurosteroids [46–48]. Although in the present study systemic hypotension and EEG suppression were not associated with reduced carotid artery perfusion, there was an increase in circulating lactate concentration in LPS + vehicle and LPS + IL1Ra treated groups after the first and second LPS infusions, suggesting impaired oxidative phosphorylation in response to LPS-induced inflammation at those times. These data are consistent with previous studies in preterm fetal and newborn sheep and raise the possibility that higher cerebral metabolic demand during fetal inflammation increases susceptibility to hypoxic-ischemic injury [7, 27, 49].
This concept is supported by studies in preterm and term neonates whereby antenatal/perinatal inflammation was linked to disturbances in cerebral oxidative metabolism, as shown by increased cerebral oxygen consumption on near infrared spectroscopy [50] and impaired cerebral oxidative metabolism on magnetic resonance spectroscopy [18]. By contrast, during the recovery period, when systemic oxygenation had normalised, suppression of EEG power and fetal movement in LPS exposed fetuses was associated with increased brain tissue IL-1β immunoreactivity. Thus, passive anoxic depolarization and inflammation-induced synaptic inhibition may have modulated EEG activity in LPS-exposed fetuses.
In LPS + IL-1Ra treated fetuses we observed a reduction in the duration of EEG suppression and faster recovery of carotid artery perfusion after the first and second LPS infusions compared to the LPS + vehicle group, suggesting an improvement in cerebral metabolism. Furthermore, arterial lactate concentration did not differ from control in the LPS + IL-1Ra group but was higher in the LPS + vehicle group after the second LPS infusion, suggesting an intermediate improvement in oxidative phosphorylation with IL-1Ra-treatment. During the recovery period, EEG power and fetal movement were improved in the LPS + IL-1Ra treated group. Collectively these data suggest that IL-1 plays an important role in modulating EEG suppression during fetal inflammation.
LPS-induced fetal inflammation was associated with increased numbers of total and activated microglia, in addition to increased apoptosis (caspase-3) and IL-1β immunoreactivity, within the intragyral and periventricular white matter tracts. This was accompanied by reduced numbers of astrocytes and total (Olig-2+) oligodendrocytes in the intragyral and periventricular white matter. The reduction in astrocyte survival likely represents the acute phase of injury during neuroinflammation. For example, reduced astrocyte numbers were reported 48 h after hypoxia-ischemia in neonatal piglets and in mechanically ventilated newborn lambs [51, 52]. There was no apparent effect LPS-exposure or IL-1Ra treatment on myelination, as shown by no differences in numbers of immature and mature oligodendrocytes expressing the CNPase protein or area fraction of CNPase staining. These data are broadly consistent with previous studies in near-term fetal sheep that were exposed to LPS during a similar time-course [53]. This combination of reduced total (Olig-2+) oligodendrocytes with no change in the number of immature and mature (CNPase+) oligodendrocytes or myelination suggests selective loss of late oligodendrocyte precursors in LPS-exposed near-term fetuses. Similarly, we found no overt neuronal loss after LPS-exposure, as shown by no differences in cortical and striatal NeuN staining between the groups. The timing of white matter vulnerability in the near-term brain overlaps with late oligodendrocyte precursor cell proliferation [54]. These data suggest that late oligodendrocyte precursors in the near-term brain are vulnerable to LPS-induced fetal neuroinflammation. Pathologically, these data are highly consistent with clinical findings of diffuse white matter injury in cases of term neonatal encephalopathy [8, 55].The combination of diffuse white matter gliosis with selective oligodendrocyte loss and relative sparing of the gray matter suggests the pathological outcomes in the present study are comparable to a mild injury pattern, consistent with mild neonatal encephalopathy [55, 56].
IL-1Ra treatment during LPS-induced fetal inflammation reduced numbers of total and activated microglia, IL-1β immunoreactivity and apoptosis within the large white matter tracts, but had no effect on astrocyte survival. The reduction in microgliosis was associated with increased total numbers of oligodendrocytes, but had no effect on numbers of immature and mature oligodendrocytes (CNPase positive cells) or myelination (CNPase immunoreactivity), suggesting improved survival of oligodendrocyte precursors. These data are consistent with evidence that systemic and/or locally produced IL-1β plays an integral role in microglial infiltration and activation [57–59]. Furthermore, our data support a critical role for microglial activation in mediating acute oligodendrocyte loss during fetal inflammation [60] and indicate that targeted inhibition of IL-1β may be a viable therapeutic intervention. Consistent with these observations, administration of IL-1Ra to neonatal rats exposed to LPS or LPS and hypoxia-ischemia was associated with reduced brain tissue IL-1β expression, reduced gliosis, improved myelination and improved motor and cognitive function [16, 61]. In the intragyral white matter tracts, we observed increased numbers of total (Olig-2+) oligodendrocytes in LPS + IL-1Ra treated fetuses compared to vehicle controls. Oligodendrocytes and their progenitor cells express IL-1 receptors [62, 63]. Binding of the IL-1 receptor on oligodendrocytes is a key extracellular signal for initiating oligodendrocyte apoptosis [64]. Thus, we speculate the increased number of total oligodendrocytes in the LPS + IL-1Ra group compared to controls was mediated by anti-apoptotic effects of the IL-1Ra. Further studies are now required to examine the long-term effects of IL-1Ra on oligodendrocyte development and myelination.
In the LPS + IL-1Ra treated group we observed faster restoration of carotid artery perfusion to baseline levels after LPS infusions compared to the LPS + vehicle group. However, there was no effect of IL-1Ra on the severity or duration of hypotension or tachycardia after LPS infusions, suggesting the improved recovery of carotid artery perfusion in the IL-1Ra group was mediated by a local anti-inflammatory effect on the cerebral vasculature and/or tissue. For example, intracisternal injection of IL-1β in adult dogs was associated with a dose dependant increase in basilar artery perfusion and vasodilation, without affecting systemic blood pressure or heart rate. Furthermore, IL-1Ra-treatment reduced the IL-1β-mediated increase in cerebral artery vasodilation and perfusion, which was likely mediated by IL-1β-induced prostaglandin production [65]. By contrast, in preterm fetal sheep, systemic TNF blockade was associated with inhibition of systemic hypotension and tachycardia during LPS-induced inflammation [27]. Collectively, these data suggest that IL-1β has targeted effects on the cerebral vasculature whereas other cytokines, including TNF, play a greater role in modulating the cardiovascular adaptations to systemic inflammation in the fetus.
One of the key translational considerations for potential neuroprotectants is when to treat [66–68]. In the present study, we started IL-1Ra infusions 1 h after infusing LPS. The rationale was to establish proof-of-concept that systemic IL-1Ra started after early-onset fetal inflammation can alleviate neuroinflammation and injury. However, it is important to appreciate that in current practice it is unlikely that inflammation (infectious or sterile) can be detected and treated as soon as it begins. Thus, the present study supports further investigation to determine the window of opportunity of IL-1Ra for treating inflammation-induced brain injury.