**3. Discussion**

O ffspring born to preeclampsia patients have an increased risk of developing several neurological complications, including neurobehavioral abnormalities, cerebral palsy, cognitive impairment, and perinatal stroke [32–34]—the underlying mechanisms of which are not fully known. Here, we tested the hypothesis that utero-placental ischemia leads to *in utero* cerebrovascular changes, including micro-hemorrhages, which may contribute to future neurological complications. We found that indeed, placental ischemia in the pregnan<sup>t</sup> female rat induces an increased number of cerebral micro-bleeds, a more pro-inflammatory cerebral-tissue environment, and decreased microglial density in the sub-ventricular zone of fetal brains *in utero*.

Like previous findings, the RUPP procedure led to decreased maternal body weight and increased fetal demise, demonstrating successful induction of placental ischemia [19,35]. Because we did not see an increase in blood pressure in this small cohort of rats, our findings are attributable to placental ischemia, independent of hypertension. We hypothesized that as a result of reduced blood flow to the fetus, surviving fetuses would have evidence of hypoxia. We found significant increases in the hematocrits of fetuses exposed to placental ischemia, suggesting systemic hypoxia. Whether the increased fetal hematocrits were associated with increased angiogenesis through vascular endothelial growth factor (VEGF) in our model is not known. While this possibility cannot be ruled out, our finding of no di fference in VEGF levels in the brain homogenates of exposed fetuses suggests that cerebral angiogenesis may not be di fferent between the groups. VEGF also increases vascular permeability and could increase vessel leakage, causing extravasation of plasma proteins. Thus, as a measure of vascular dysfunction, we quantified micro-bleeds within tissue slices.

The incidence of cerebral micro-bleeds was significantly higher in the placental ischemia-exposed group compared to the sham control group. Micro-bleeds are thought to occur when blood vessels are structurally damaged [36]. Thus, an increased incidence of micro-bleeds is consistent with ongoing vascular damage in the brains of exposed fetuses. Cerebral micro-bleeds can lead to long-term neurological damage, including cognitive and motor deficits [36]. Vascular damage and increased blood-brain barrier permeability can result in increases in brain water content; therefore, we hypothesized that fetuses exposed to placental ischemia would have increased cerebral water content, a crude marker of cerebral edema. Contrary to our hypothesis, we found no di fference in brain water content between fetuses exposed to normal pregnancy and those exposed to placental ischemia. Even more interesting is the finding of a negative association between the number of cerebral micro-bleeds and brain water content. A recent study reported that micro-bleeds are observed in high-altitude-induced injury long after cerebral edema has resolved [31]. Thus, our finding of no di fference in brain water content even in the presence of micro-bleeds could be an indication of a resolution of edema. We do not know whether di fferent fetal or postnatal time-points would yield similar results and this is an area for future investigation. Additionally, we assessed water content in the entire fetal brain and could have missed regional changes in water content as a result.

Maternal inflammation and cerebral micro-bleeds are associated with increases in neuroinflammation. Therefore, we assessed the levels of di fferent cytokines/chemokines in brain homogenates of exposed fetuses. We found increases in the pro-inflammatory cytokines, IL-1β, IL-6, and IL-18, and a trend for increased levels of TNF α and IL-17 in brains of fetuses exposed to placental ischemia. Additionally, IL-10 and IL-4 levels tended to increase in brains of fetuses exposed to placental ischemia. Those data sugges<sup>t</sup> two possibilities. The first possibility is that there is an increased transfer of maternal inflammatory factors across the placental barrier into the fetal circulation. This is plausible since preeclampsia patients have increased circulating levels of tumor necrosis factor alpha (TNF α) [20], interleukin (IL)-2 [37,38], IL-6 [21], and IL-17 [39–41]. Placental ischemic rats also have increased levels of inflammatory cytokines in the circulation and cerebrospinal fluid [20,21,27,42,43]. We have not assessed changes in placental barrier permeability in the RUPP model of placental ischemia, and this is an area of ongoing investigation. The second possibility is that there is increased local production of these cytokines in the fetal brains following exposure to placental ischemia.

A key finding in this study is that exposure to placental ischemia leads to increased cerebral levels of the pro-inflammatory cytokine, IL-1β in utero. There is evidence that placentas from preeclampsia patients secrete more IL-1β compared to normotensive placentas [44]. Additionally, infusion of IL-1β into the brains of young rats induces blood-brain barrier breakdown [45]. Taken together, increased tissue levels of IL-1β may play a deleterious role at the blood-brain barrier, contributing to structural damage and subsequent cerebral micro-bleeds. Whether increased IL-1β has a causal role in increased cerebral micro-bleeds in placental ischemia-exposed fetuses will be assessed in future studies.

Our multi-plex cytokine analysis also revealed increased levels of IL-6 and IL-18 in the brains of exposed fetuses. Not only was cerebral IL-6 increased in response to placental ischemia, but fetal brain IL-6 levels positively correlated to the number of cerebral micro-bleeds. Our findings are consistent with reports that serum levels of both IL-6 and IL-18 are higher in patients with cerebral micro-bleeds compared to those without [46]. Another study found that in patients with ischemic cerebrovascular disease, IL-6 was associated with an increased risk for cerebral micro-bleeds in an elderly, community-based cohort [47]. Furthermore, in the developing brain, micro-bleeds were observed in pups exposed to lipopolysaccharide, coupled with intrauterine ischemia, mainly if they were vaginally delivered [48], suggesting that *in utero* insults make blood vessels more susceptible to injury. Taken together, those studies and our current findings support the hypothesis that maternal inflammation, induced by placental ischemia, contributes to weakened cerebral vessels, making them more susceptible to blood-brain barrier (BBB) disruption and cerebral micro-bleeds.

Previously, we reported that eotaxin (CCL11) is increased in the cerebrospinal fluid of placental ischemic dams [42,49]. In the current study, we found increased levels of eotaxin in the brains of fetuses exposed to placental ischemia. The consequences and source of fetal brain eotaxin levels were not directly investigated in this study, although eotaxin has been shown to promote glutamate-induced neurotoxicity [50]. Importantly, CCL11 has been shown to directly regulate neurogenesis [51], such that increased circulating eotaxin was associated with reduced neurogenesis in aged mice and infusion of CCL11 into young mice resulted in decreased neurogenesis and cognitive impairment. Thus, increased eotaxin levels in the brains of fetuses exposed to placental ischemia sugges<sup>t</sup> that neurogenesis may be impaired.

Microglia are essential during normal brain development and have important roles in pruning synapses, phagocytizing excess neuronal progenitor cells, and regulating the number of neurons at each stage of development [52]. Thus, changes in microglial numbers at late gestation could predict later neurological function. Microglia are also involved in secreting pro-inflammatory cytokines; thus, increased density of microglia could contribute to increases in local tissue inflammation. Following strokes, immune cells (microglia and macrophages) are activated by cytokines and chemokines, and migrate to damaged areas to remove dead neural cells. Those immune cells, induced by a chronic inflammatory environment, may become over-activated and produce large amounts of pro-inflammatory cytokines, disrupting neurogenesis and the BBB [53]. The presence of a pro-inflammatory environment and vascular damage in the brains of fetuses exposed to placental ischemia sugges<sup>t</sup> that microglia may be activated. We therefore hypothesized that brains from exposed fetuses would have an increased density of microglia. Contrary to our hypothesis, we found that placental ischemia exposed fetuses had a decreased density of microglia in the SVZ of the third ventricle. Microglia in that region are vital for neurogenesis and oligodendrogenesis during fetal development and throughout life [52]. Reduced microglial density in the SVZ of placental ischemia-exposed fetuses may predict deficits in neurogenesis and oligodendrogenesis a ffecting the central nervous system long term. This possibility is a subject of ongoing investigation.

There is evidence that in both rats and mice, microglia protect neonatal brains from injury after ischemic stroke, and that depletion of microglia led to worse outcomes [54]. Thus, increased cerebral micro-hemorrhages, as occurred in exposed fetuses, coupled with fewer microglia, may indicate more severe damage/outcomes. During development, microglia can be disturbed by cytokines and chemokines released under conditions of maternal inflammation [55] seen in placental ischemia.

To our knowledge, this is the first study reporting the impact of placental ischemia on cerebral micro-bleed incidence and neuroinflammation in the o ffspring in utero. The inclusion of male and female o ffspring in all analyses is a strength of the current study; however, we were unable to assess sex di fferences in our endpoints. Ongoing and future studies are now utilizing genotyping for sex to specifically assess sex di fferences. Additionally, this study utilized only one time-point (E19) and we are therefore unable to extend our findings to the postnatal period. Because we used brain homogenates to assess the inflammatory status, we were unable to address regional changes in expression of the di fferent cytokines, chemokines, and growth factors assessed. We limited our microglial analysis to the SVZ associated with the third ventricle for this study; however, analysis of microglial changes in other brain regions will be important.

In conclusion, the rat placental ischemia model induces increased cerebral micro-hemorrhages in the developing brain in utero and may be a good model to assess the etiology of cerebral micro-hemorrhages associated with preeclampsia-complicated pregnancies. Our data sugges<sup>t</sup> that a hypoxic environment in the fetus, induced by reducing maternal utero-placental perfusion, induces abnormalities in cerebrovascular structure, increased blood-brain barrier disruption and micro-bleeds in the developing fetal brain. This is associated with a pro-inflammatory environment and reduced microglial density in proliferative brain regions, and may underlie the increased neurodevelopmental abnormalities observed in o ffspring born to preeclampsia patients. Additional studies are required to elucidate the underlying causes of these observations and to establish whether there are time-dependent di fferences in the outcome measures.
