Defining the mechanisms of neuroinflammation and antepartum neuroprotection in human neural progenitor cells and microglia

Abstract

BACKGROUND

Corticosteroids and magnesium sulfate are neuroprotective agents that decrease disabilities in children born prematurely. The neuroprotective mechanisms of these drugs remain unclear, and despite widespread use, gestational age remains a major risk factor for fetal neurological injury. Towards full term, the abundance of neural progenitor cells declines, and microglia increase in number. The efficacy of neuroprotective agents may be cell-type specific at the time of prematurity. We hypothesize that the relationship between gestational age and neuroprotection is because of cell-specific responses to antepartum treatments.

OBJECTIVE

This study evaluates in vitro inflammatory cytokine production by human fetal microglia and neural progenitor cells exposed to magnesium sulfate and betamethasone to determine the neuroprotective mechanisms of these drugs.

STUDY DESIGN

Commercially available human fetal microglia (HMC3) and neural progenitor cells (ReNcells) were exposed to interleukin-1 beta, a known instigator of perinatal neuroinflammatory injury. Cells were also given magnesium sulfate, betamethasone, magnesium sulfate and betamethasone, or the vehicle (control) for 24 hours. Inflammatory responses were evaluated by gene expression and Milliplex immunology assay of the cell media to measure cytokine production and secretion.

RESULTS

With interleukin-1 beta exposure, inflammatory cytokines were upregulated in both cell types according to gene expression and were found at elevated concentrations within cell media. Substantial production of the pro-inflammatory cytokines interleukin-6 and interleukin-8 was observed by both cell types, but only microglia produced the anti-inflammatory cytokine interleukin-10. In interleukin-1 beta exposed neural progenitor cells, treatment with magnesium sulfate slightly lowered interleukin-6 and interleukin-8, but this did not occur with betamethasone or magnesium sulfate/betamethasone. In microglia, betamethasone alone and magnesium sulfate/betamethasone substantially reduced most cytokines upregulated by interleukin-1 beta. A discordance was observed between interleukin-8 expression within microglia and its concentration measured in cell media. Additional analysis revealed that betamethasone reduced the translation of interleukin-8 rather than its extracellular secretion.

CONCLUSION

Pro-inflammatory cytokine expression and production by microglia and neural progenitor cells were impacted differently by each drug. Microglial responses to interleukin-1 beta were substantially attenuated by betamethasone, whereas magnesium sulfate had a mild influence on neural progenitor cells. Furthermore, our results suggest that the neuroprotective mechanisms of betamethasone may be attributed to the reduction of cytokine translation within microglia. Additional investigation is warranted to elucidate whether neuroprotection is attributed to cell-specific drug interactions at the time of the inciting event.

AJOG Global Reports at a Glance.

Why was this study conducted?

Antepartum treatments reduce inflammatory injury, but the mechanisms of neuroprotection are unknown, and efficacy remains poor. A greater understanding of cellular responses is needed to optimize these treatments.

Key findings

Betamethasone reduced pro-inflammatory cytokines in human microglia. Magnesium sulfate mildly attenuated cytokine production by human neural progenitor cells.

What does this add to what is known?

The potential mechanisms of neuroprotection may involve anti-inflammatory effects of corticosteroids on microglia but not on neural progenitor cells.

Introduction

Although the exact etiology of poor developmental outcomes with prematurity is not always clear, neuroinflammatory injury can be instigated in the presence and absence of intraamniotic infection.^1,2 In utero neuroinflammation is mediated by several cytokines such as interleukin-1 beta (IL-1β), interleukin-6 (IL-6), interleukin-8 (IL-8), and tumor necrosis factor (TNF).3, 4, 5, 6, 7 These cytokines mediate perinatal neuroinflammatory injury that results in complex anatomical changes with reductions in brain volume, loss of white matter, and other effects at the cellular level.^1,8, 9, 10, 11, 12 Of note, IL-1β can invoke both preterm birth and systemic inflammatory responses, including the upregulation of cytokines (IL-6 and IL-8) associated with cerebral white matter injury.^4,13, 14, 15, 16

Antepartum magnesium sulfate (MgSO₄) and corticosteroids mitigate some effects of early delivery.¹⁷ MgSO₄ decreases the rates of cerebral palsy if administered around the time of preterm birth at gestational ages less than 32 weeks.18, 19, 20, 21 Current dosing is mainly extrapolated from the use of MgSO₄ to prevent maternal seizures during preeclampsia.²² Corticosteroids such as betamethasone (BMZ) have traditionally been administered to accelerate fetal lung maturity, and their use has expanded as studies have revealed that these drugs decrease the rates of intraventricular hemorrhage and periventricular leukomalacia.23, 24, 25, 26, 27 The molecular mechanisms by which brain matter or development is “protected” remain unknown for these drugs in humans. Studies in animal models of in utero inflammation have demonstrated that BMZ and MgSO₄ reduce cytokines, microglial activation, and apoptosis.28, 29, 30, 31

Cellular growth in the human fetal brain increases exponentially from 13 to 20 weeks’ gestation and transitions to a linear phase from 22 weeks to term.³² A significant portion of the preterm fetal brain is composed of neural progenitor cells (NPCs) that give rise to neurons and glial cells.³³ Human NPCs can produce pro-inflammatory cytokines but are not considered major contributors.^34,35 Microglia, the brain’s resident macrophages derived from yolk-sac progenitors, are a major source of pro-inflammatory cytokines.36, 37, 38, 39, 40 In normal development, the number of NPCs decrease towards full term while the number of microglia increase.^33,39,41,42 Cell-specific inflammatory responses within the human fetal brain may contribute to greater vulnerability at earlier stages of neurodevelopment. Gestational age is a major risk factor for white matter injury, particularly for neonates born before 32 weeks.⁴³ The efficacy of neuroprotective agents may depend on the ability of these drugs to mitigate inflammatory responses by NPCs and microglia. We hypothesized that cell-specific responses underlie the efficacy of antepartum treatments and aimed to characterize the neuroprotective mechanisms of these drugs in human fetal NPCs and microglia.

Materials and methods

The commercially available human fetal neural progenitor cell line, ReNcell VM (Millipore Sigma, Temecula, CA), and human fetal microglial HMC3 cell line (ATCC, Manassas, VA), were thawed and cultured per the manufacturer’s guidelines.⁴⁴ First, we compared responses to cytokines associated with neuroinflammation by exposing HMC3 and ReNcells to 100 ng/mL recombinant human IL-1β, IL-6, and TNF (FUJIFILM Wako, Richmond, VA). Because of the robust response of both cell types to IL-1β, all subsequent comparisons were conducted with or without exposure to 100 ng/mL of only this cytokine. Drug treatments were conducted by supplementing the culture media with 1.29 mmol/L of MgSO₄ (Millipore Sigma, Temecula, CA) and 1 ng/mL pharmaceutical-grade BMZ (Celestone Soluspan; Organon, Jersey City, NJ). These concentrations reflect amounts observed in human umbilical cord blood in patients given standard doses of these drugs.45, 46, 47 For experimental control, cells were given an equal volume of the vehicle, which consisted of water for IL-1β and MgSO₄, and dimethyl sulfoxide for BMZ. Each condition was performed as 4–6 technical replicates within independent experimental runs for both cell types. After experimental exposure, cells were harvested for RNA extraction to compare gene expression. Media was collected for cytokine analysis at the intervals depicted within the figures. Cells were also harvested for protein extraction for Western blot analysis. Specific experimental and statistical methods are detailed in the supplemental materials and methods.

Results

Cytokines instigate inflammatory responses in both microglia and neural progenitor cells

Initially, we evaluated responses by NPCs and microglia to key inflammatory cytokines IL-1B, IL-6, and TNF for 24 hours and analyzed changes in gene expression (Figure 1, A). Microglia responded to IL-1β and TNF by upregulating IL1B, IL6, and TNF (Figure 1, B). The expression of inflammatory receptors was not different in microglia, with the exception of a slight downregulation of CD14 with IL-1β when compared with IL-6 exposure (Figure 1, C). In contrast, IL-6 did not impact microglial gene expression for the cytokines and receptors examined. For NPCs, IL-1β exposure stimulated the upregulation of IL1B, IL6, and TNF (Figure 2, A) and the receptors IL1R1 and IL6R (Figure 2, B). Exposure to TNF upregulated IL1B and TNF (Figure 2, A) but not IL6 or inflammatory receptor expression. Both IL-1β and IL-6 downregulated CD14, but there were no other differences with IL-6 exposure. Cytokine exposure did not change NPC expression of related genes (Figure 2, C). Based on these results, we conducted subsequent comparisons with IL-1β.

Figure 1.

Experimental design and cytokine responses by microglia

A, Representation of cell exposure and treatment duration. Box-and-whisker plots represent expression for cytokines (B) and inflammatory receptors (C) in the x-axis with respective exposures to the vehicle, IL-1β, IL-6 and TNF. X’s within boxes denote the means of n=4–6 technical replicates. *P<.05, *** P<.0005 by Tukey’s post hoc test.

CD14, cluster of differentiation 14; IL1B, interleukin-1 beta; IL6, interleukin-6; IL6R, interleukin-6 receptor; ILR1, interleukin-1 receptor 1; TLR4, Toll-like receptor 4; TNF, tumor necrosis factor; TNFRSF1, tumor necrosis factor receptor superfamily member 1.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Figure 2.

Cytokine responses by neural progenitor cells

Box-and-whisker plots represent expression for cytokines (A), inflammatory receptors (B), and neural associated genes (C). Genes are represented in the x-axis for each respective exposures to the vehicle, IL-1β, IL-6 and TNF. X’s within boxes denote the means of n=4–6 technical replicates. * P<.05, *** P<.0005 by Tukey’s post hoc test.

CD14, cluster of differentiation 14; IL1B, interleukin-1 beta; IL6, interleukin-6; IL6R, interleukin-6 receptor; ILR1, interleukin-1 receptor 1; MBP, myelin basic protein; NESTIN, nestin; SOX 2, SRY-box transcription factor; TLR4, Toll-like receptor 4; TNF, tumor necrosis factor; TNFRSF1, tumor necrosis factor receptor superfamily member 1; TUBB3, tubulin beta 3 class III.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Interleukin-1 beta exposure invokes longer term changes in neural progenitor cells

Given the robust response by NPCs to IL-1β, we next determined whether these cells sustained inflammatory responses. For this, NPCs were exposed to IL-1β for 24 hours, then washed and replenished with fresh media free of cytokines. These cells were then harvested for gene expression analysis 24 hours or 7 days after the removal of IL-1β (Figure 3, A). Cytokines remained upregulated 24 hours after the withdrawal of IL-1β (48 hours post-treatment) but were not different compared with controls after 7 days of IL-1β withdrawal (8 days total) (Figure 1, B). Most inflammatory receptors were higher at 48 hours, but all receptors examined were lower at 8 days compared with controls (Figure 2, C). Neural gene expression was different only for MBP at 48 hours, whereas NESTIN and SOX2 were lower at 8 days (Figure 3, D).

Figure 3.

Experimental outline and results from longer-term analysis of IL-1β exposure

A, representation timeline for 24 hours of exposure to the vehicle or IL-1β. Cells were then washed and replenished with media free of IL-1β to assess whether responses persist after its withdrawal. Box-and-whisker plots represent expression for cytokines (B), inflammatory receptors (C), and neural associated genes (D). Genes are represented in the x-axis for exposure to the vehicle or IL-1β at 24 or 48 hours. X’s within boxes denote the means of n=4–6 technical replicates. *P<.05, **P<.005, ***P<.0005 by Tukey’s post hoc test.

CD14, cluster of differentiation 14; IL1B, interleukin-1 beta; IL6, interleukin-6; IL6R, interleukin-6 receptor; ILR1, interleukin-1 receptor 1; MBP, myelin basic protein; NESTIN, nestin; SOX 2, SRY-box transcription factor; TLR4, Toll-like receptor 4; TNF, tumor necrosis factor; TNFRSF1, tumor necrosis factor receptor superfamily member 1; TUBB3, tubulin beta 3 class III.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Neural progenitor cells produce pro-inflammatory cytokines but not to the extent as microglia

We subsequently examined cytokine translation and secretion between NPCs and microglia by analyzing cytokine concentrations in the media. For this, we compared media collected 24 hours post IL-1β exposure to media collected at 48 hours following replenishment with cytokine-free media at 24 hours post IL-1β exposure (Figure 3, A). The concentrations of granulocyte-macrophage colony-stimulating factor (GM-CSF), TNF, IL-1β, IL-6, and IL-8 were greater in media collected from NPCs exposed to IL-1β at 24 hours compared with the control (Figure 4, A). Only IL-1β and IL-8 remained higher at 48 hours, after 24 hours of IL-1β withdrawal. Microglia also had higher levels of GM-CSF, TNF, IL-1β, IL-6, and IL-8 after 24 hours of IL-1β exposure, and these remained elevated at 48 hours (Figure 4, B).

Figure 4.

Cytokines are produced by neural progenitor cells and microglia

Bio-Plex analysis of cell media collected at 24 and 48 hours post exposure to the vehicle or IL-1β, represented in the experimental design depicted in Figure 3, A. Note the 48-hour timepoint represents media collected 24 hours after washout of exogenous IL-1β, whereas the 24 hour timepoints represent media that contains exogenous IL-1β. Box-and-whisker plots represent cytokine concentrations detected for NPCs (A) and microglia (B). Cytokines are listed in the x-axis for exposure to the vehicle or IL-1β. X’s within boxes denote the means of n=4–6 technical replicates. *P<.05, **P<.005, ***P<.0005 represent False Discovery Rate adjusted P values.

GM-CSF, granulocyte-macrophage colony-stimulating factor; IL-1β, interleukin-1 beta; IL-2, interleukin-2; IL-3, interleukin-3; IL-4, interleukin-4; IL-5, interleukin-5; IL-6, interleukin-6; IL-7, interleukin-7; IL-8, interleukin-8; IL-9, interleukin-9; IL-10, interleukin-10; IL-12, interleukin-12; IL-13, interleukin-13; NPCs, neural progenitor cells; TNF-α, tumor necrosis factor alpha.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Magnesium sulfate and betamethasone treatments do not substantially alter gene expression in neural progenitor cells and microglia

Based on the high levels of IL-8 produced by NPCs and microglia, we examined the expression of this cytokine rather than TNF in subsequent experiments. Here, we again exposed cells to IL-1β, but in combination with MgSO₄ and BMZ or the vehicle. Gene expression and cytokine production were examined after 24 hours of exposure to IL-1β and the drugs (Figure 5, A). Based on the pattern of expression in NPCs, NESTIN expression was lower with both drugs, both without IL-1β or with IL-1β and MgSO₄. The combination of drugs or IL-1β also altered MBP expression compared with the vehicle control (Figure 5, B). For both NPCs and microglia, the expression of certain cytokines was lower in cells exposed to both drugs in the absence of IL-1β (Figure 5, Figure 5, Figure 5).

Figure 5.

Experimental outline and comparisons of MgSO₄ and BMZ treatments

A, representation outline for 24 hours of exposure to the vehicle or IL-1β in combination with or without MgSO₄ and BMZ. Box-and-whisker plots represent expression for neural related genes and cytokines by NPCs (B) and cytokines by microglia (C). Treatment combinations are represented by +’s within grids along the x-axis. X’s within boxes denote the means of n=4–6 technical replicates. *P<.05, **P<.005, ***P<.0005 by Tukey’s post hoc test.

BMZ, betamethasone; IL-1β, interleukin-1 beta; IL6, interleukin-6; IL8, interleukin-8; MgSO4, magnesium sulfate; MBP, myelin basic protein; NESTIN, nestin; NPCs, neural progenitor cells; TUBB3, tubulin beta 3 class III.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Magnesium sulfate and betamethasone differently attenuate cytokine production by neural progenitor cells and microglia

Concentrations of GM-CSF, TNF, IL-1β, IL-6, and IL-8 measured in the media of NPCs were significantly greater with IL-1β exposure (Figure 6). However, levels of GM-CSF, TNF, IL-6, and IL-8 were lower with exposure to IL-1β and the combination of drugs but not with MgSO₄ or BMZ alone. With IL-1β exposure, microglia upregulated the production of cytokines and chemokines (Figure 7). In comparison with cells exposed only to IL-1β, the introduction of BMZ significantly reduced concentrations of GM-CSF, TNF, IL-5, IL-6, IL-7, IL-8, and IL-10. Alone, MgSO₄ did not show the same effect, but in combination with BMZ, certain factors were reduced.

Figure 6.

Cytokines production by neural progenitor cells with MgSO₄ and BMZ treatments

Bio-Plex analysis of cell media collected from NPCs following 24 hours of exposure to the vehicle or IL-1β in combination with or without MgSO₄ and BMZ. Box and whisker plots represent cytokine concentrations. Treatment combinations are represented by +’s within grids along the x-axis. X’s within boxes denote the means of n=4–6 technical replicates. *P<.05, **P<.005, ***P<.0005 by Tukey’s post hoc test.

BMZ, betamethasone; GM-CSF, granulocyte-macrophage colony-stimulating factor; IFN-λ, interferon lambda; IL-1β, interleukin-1 beta; IL-2, interleukin-2; IL-4, interleukin-4; IL-5, interleukin-5; IL-6, interleukin-6; IL-7, interleukin-7; IL-8, interleukin-8; IL-10, interleukin-10; IL-13, interleukin-13; MgSO4, magnesium sulfate; NPCs, neural progenitor cells; TNF, tumor necrosis factor.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Figure 7.

Cytokines production by microglia with MgSO₄ and BMZ treatments

Bio-Plex analysis of cell media collected from microglia following 24 hours of exposure to the vehicle or IL-1β in combination with or without MgSO₄ and/or BMZ. Box and whisker plots represent cytokine concentrations. Treatment combinations are represented by +’s within grids along the x-axis. X’s within boxes denote the means of n=4–6 technical replicates. *P<.05, **P<.005, ***P<.0005 by Tukey’s post hoc test.

BMZ, betamethasone; GM-CSF, granulocyte-macrophage colony-stimulating factor; IFN-λ, interferon lambda; IL-1β, interleukin-1 beta; IL-2, interleukin-2; IL-4, interleukin-4; IL-5, interleukin-5; IL-6, interleukin-6; IL-7, interleukin-7; IL-8, interleukin-8; IL-10, interleukin-10; IL-13, interleukin-13; MgSO4, magnesium sulfate; TNF, tumor necrosis factor.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Betamethasone alters interleukin-8 translation by microglia

To gain mechanistic insight, we examined whether the actions of BMZ observed in microglia relate to the translation of IL-8. For this, cells exposed to IL-1β for 24 hours were washed and harvested to measure intracellular levels of IL-8 by Western blot analysis. With BMZ, cells showed reduced levels of IL-8 compared with IL-1β alone (Figure 8, A). The interpretation of these results and the proposed model are depicted in Figure 8, B.

Figure 8.

Intracellular levels of IL-8 and proposed mechanism of BMZ

A, Western blot analysis of microglia cells exposed to vehicle (control), IL-1β or IL-1β with BMZ. The right image represents IL-8 at ∼8.5kDA and to the right the loading control Histone H3 at ∼15kDA. The ladder is annotated to the right of each image. The box and whisker plot to the right represent the amount of IL-8 normalized to the loading control. X’s within boxes denote the means of n=3 technical replicates. *P<.05 by student t test. B, Proposed molecular mechanisms by which BMZ attenuates cytokine translation or invokes degradation (Ub denotes ubiquitination) in microglia but not NPCs.

BMZ, betamethasone; H3, histone H3; IL-1β, interleukin-1 beta; IL-6, interleukin-6; IL-8, interleukin-8; kDa, kilodalton; NPCs, neural progenitor cells; TNF, tumor necrosis factor; Ub, ubiquitin.

Ligon. Neuroprotection in human fetal microglia and neural stem cells. Am J Obstet Gynecol Glob Rep 2026.

Comment

Principal findings

Treatment with BMZ and MgSO had little effect on pro-inflammatory responses by NPCs. In contrast, BMZ significantly reduced the effects of IL-1β on microglia. Further evidence suggests that the actions of BMZ in microglia relate to the translation of pro-inflammatory cytokines (Figure 8).

Results in the context of what is known

Neuroinflammation is triggered by cytokines released in settings of infection, trauma, or disease.48, 49, 50 Cytokines, including IL-1β, IL-6, and TNF, can cross the blood-brain barrier to evoke neuroinflammatory injury.50, 51, 52 Although etiologies and severity vary, elevated levels of pro-inflammatory cytokines are associated with white matter injury within the developing and adult brain.^48,53, 54, 55 Microglia are major contributors to cytokine-mediated neuroinflammatory injury.^56,57 Pro-inflammatory microglia release cytokines that mediate tissue injury and negatively impact neurogenesis by altering the differentiation potential of NPCs.58, 59, 60, 61 In response to cytokines such as IL-1β, NPC may lose their potential to differentiate.⁶⁰ The expression of genes essential for NPC maintenance and potential (NESTIN and SOX2) declines days after IL-1β exposure (Figure 3).^62,63 Although we did not test differentiation, these results support human NPC reprogramming under inflammatory conditions that may impair self-renewal and their ability to generate neurons and glia.^62,64 The upregulation of MBP at 48 hours suggests that NPC differentiation to oligodendrocytes may have been initiated by IL-1β exposure (Figure 3, D).⁶⁵ However, this was not consistent across experiments and may represent a subset of cells within the 48-hour time point.

Few studies have addressed the possibility that NPCs also promote neuroinflammation.^66,67 We observed that NPCs express pro-inflammatory receptors that are maintained after exposure to cytokines (Figure 1). The exception is the downregulation of CD14, which may alter TLR4 responses to bacterial endotoxins present with chorioamnionitis.⁶⁸ Our results also indicate that NPCs upregulate the expression of inflammatory cytokines to almost the same degree as microglia when exposed to IL-1β and TNF but not IL-6 (Figure 1, Figure 2). Although NPCs did not sustain cytokine production at the same concentrations as microglia, both cell types upregulated and secreted significant amounts of GM-CSF, IL-6, IL-8, and TNF when exposed to IL-1β (Figure 4).^69,70 Coincidentally, the expression of IL-1B was also substantially elevated in both cell types. Unfortunately, immunoassays, including the one we used, cannot distinguish between the recombinant human IL-1β added to the media and what the cells produce. Therefore, we can rely only on the detection of endogenous IL-1β in samples collected at 48 hours post-exposure that were washed to remove recombinant IL-1β after the initial 24 hours of exposure (Figure 4).

In addition to IL-1β, TNF exposure also invoked significant upregulation of cytokines in NPCs and microglia. With TNF, microglia upregulated IL-6⁷¹; however, this did not occur in NPCs, suggesting that responses within this pathway are modified by cell-specific epigenetics or signaling transduction mechanisms.⁷² In contrast, exogenous IL-6 did not upregulate cytokine expression in either cell type (Figure 1, Figure 2). Sarieva et al⁷³ demonstrated that similar NPCs (ReNcell CX) respond to IL-6; however, they did not examine cytokine expression. The microglia (HMC3) we used express the IL-6 receptor; however, to our knowledge, no published study has reported exogenous IL-6 administration for these cells. Contrary to our results, several reports have demonstrated the upregulation of cytokines by other human cell types exposed to IL-6 in vitro.74, 75, 76 These studies administered IL-6 in conjunction with soluble IL-6 receptor or other stimuli, suggesting that additional factors may be needed for IL-6-mediated cytokine upregulation. Further investigation is needed to discern the cascades of IL-6 inflammatory signaling between various cell types and conditions.

Higher concentrations of pro-inflammatory cytokines have been described in maternal serum and fetal tissue affected by preterm labor. These cytokines are present in the preterm fetal brain and create a feedback system that further aggravates inflammatory responses.^2,77 IL-1β is associated with both chronic and acute inflammation and has been implicated in parturition by directly inducing smooth muscle contraction in the uterus.⁷⁸ Regarding neuroinflammation, IL-1β, IL-6, and IL-8 are commonly elevated cytokines and predictors of brain injury in preterm neonates and in adults.^8,79,80 These cytokines affect differentiated neural cells by activating microglia and astrocytes, arresting maturation, and leading to neurotoxicity with subsequent death or damage. The effects of inflammation on NPC proliferation and migration is also reflected in adult neurodegenerative disorders.⁸¹ These pro-inflammatory cytokines are also associated with the fetal white matter lesions implicated in cerebral palsy.^11,12,48 Immune effector cells, including microglia, are activated with neuroinflammation, although some studies suggest they may initially aggravate injury further contributing to white matter disruption before assuming an anti-inflammatory role.¹ Anti-inflammatory IL-10 increases survival of NPCs and is associated with neurogenesis and oligodendrogenesis.⁸² However, it remains unclear whether NPCs produce IL-10 in vivo and our results show low levels of this cytokine that did not change under the tested conditions.

Clinical implications

Antepartum interventions currently used in the setting of preterm birth, MgSO₄ and corticosteroids, appear to be neuroprotective.^18,83 The benefits of corticosteroids on lung and vascular maturation may relate to neuroprotective mechanisms and the normalization of fetal development after premature birth. However, the efficacy of these drugs is poor, with the numbers needed to treat for the prevention of cerebral palsy being 1:46–74 and 1:80 for MgSO₄ and BMZ, respectively.^22,84 The underlying neuroprotective mechanisms of these drugs remain unclear, but understanding these mechanims is essential to address their poor efficacy and optimize their use. Evidence suggests MgSO₄ may inhibit N-methyl-D-aspartate (NMDA) channels, protecting the brain from NMDA receptor-mediated injury, while also acting on fetal neurovascular vasodilation, fetal convulsion suppression, and decreased release of nitric oxide.^85,86 In human umbilical cord blood-derived monocytes and human placentas, MgSO₄ appears to reduce levels of IL-6.^87,88 Studies examining BMZ suggest that endothelial cell stability and maturation of vascular structures may decrease intracranial hemorrhage and periventricular leukomalacia.^27,89,90 The anti-inflammatory and neuroprotective effects of these drugs are also supported by several animal studies.^{30,31,83,91,92} To our knowledge, no studies have evaluated the anti-inflammatory potential of MgSO₄ and BMZ in human neural cells.

The pattern of our results supports the concept that neuroinflammatory injury and the efficacy of these medications may be related to cell-specific responses by NPCs and microglia at the onset of preterm birth. A recent clinical trial suggests that BMZ given later in gestation (ie, at 34–36 weeks) may not provide neuroprotection.⁹³ One explanation for this observation is the lack of anti-inflammatory effects BMZ may have on human NPCs compared with microglia within this timeframe. Human microglia are at their lowest density between 30–34 weeks’ gestation and begin to increase in number at 36 weeks.³⁹ Temporal changes in the number of microglia may relate to the efficacy of BMZ at earlier gestational ages and the lack of effect between 34–36 weeks’ gestation. Correspondingly, the number of NPCs declines between 32–36 weeks’ gestation. Given that pro-inflammatory responses by human NPC were mildly attenuated by MgSO4 and not at all by BMZ, the overall poor efficacy of these drugs in preventing cerebral palsy may be related to this observation.

Research implications

These neuroprotective agents selectively attenuated inflammatory responses to IL-1β, with no obvious synergy when applied in combination. NPCs were largely unaffected by BMZ, whereas MgSO₄ showed a mild impact in reducing the concentration of GM-CSF, IL6, IL-8, and TNF (Figure 6). Microglia, unaffected by MgSO_4, were more broadly and dramatically affected by BMZ compared with the differences observed in NPCs (Figure 7). This was particularly profound for IL-8, which increased substantially in microglia exposed to IL-1β and was dramatically lower with BMZ. The average concentration of IL-8 was 4608 ng/mL with IL-1β but was 10-fold lower (430 ng/mL) with the addition of BMZ. For NPCs, the average concentration of IL-8 was 477 ng/mL with IL-1β and was 2-fold lower (282 ng/mL) with the inclusion of MgSO₄. In both cell types, the transcription of IL-8 did not change, suggesting that anti-inflammatory mechanisms relate to protein translation, degradation, or extracellular secretion. Given that IL-8 levels in NPCs were not as high or as impacted by MgSO₄, we focused on microglia to investigate the discrepancy between IL-8 expression and protein concentration. Our results from microglial intracellular analysis suggest that BMZ reduces the translation of IL-8. However, we cannot exclude the possibility that secretion was also impacted or that other posttranslational mechanisms may contribute to the decline in protein (Figure 8, A). Further research is necessitated to decipher the exact molecular interactions by which BMZ alters cytokine production and whether corticosteroids also inhibit their extracellular release (Figure 8, B).

Strengths and limitations

The greatest strength of our study is our comprehensive evaluation of inflammatory responses in conjunction with clinically relevant neuroprotective agents in human fetal brain cells. To provide clinical and translational relevance, our dosing resembles concentrations observed in human fetal cord blood. Finally, our investigation provides novel mechanistic insight into how BMZ may invoke neuroprotection by attenuating microglial translation of pro-inflammatory cytokines. This novel finding adds to the understanding of the molecular interactions by which corticosteroids alleviate neuroinflammation.

Study limitations primarily lie with the generalizability of our results and reliance on 2-dimensional culture, which does not account for cell interactions and the microenvironment present in vivo. Because of inherent limitations in studying pregnancy, even sophisticated animal or cell line models may not reflect the complexities of human neurodevelopment. The administration of MgSO₄ and BMZ, with regard to route, bioavailability, metabolized forms, and timing, are all factors that cannot be perfectly simulated in vitro. Furthermore, dexamethasone, which is also commonly administered for prematurity, or other formulations of betamethasone were not compared and require further investigation. Another limitation is our reliance on cell lines, as human fetal neural cells are inherently inaccessible. We believe that RenCells and HMC3 lines are the most suitable cells for this study given that they are the only commercially available fetal brain derived cells and that they faithfully model human NPCs and microglia.^67,94 Lastly, our study is limited to the investigation of antepartum drugs with IL-1β administered at supraphysiological levels relative to neonatal cerebral spinal fluid.⁹⁵ Neuroinflammation involves several cytokines that warrant further investigation alongside antepartum treatments.

Conclusion

The mitigation of adverse neurological outcomes from premature birth is a vital healthcare priority.⁹⁶ Prematurity-related neuroinflammation can affect movement, cognitive, and sensory skills that, in turn, require significant resources and may impair quality of life.^97,98 Our findings highlight cell-specific responses that may underlie neuroprotective mechanisms and the efficacy of antepartum medications: specifically, that BMZ dramatically attenuates microglial inflammatory cytokine production, whereas MgSO4 marginally impacts this process in NPCs. Further research is needed to translate our findings and broaden our understanding of the neuroprotective mechanisms within the developing brain.

CRediT authorship contribution statement

Sarah E. Ligon: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Katherine M. Leonard: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Jennifer R. Damicis: Writing – review & editing, Methodology, Investigation, Formal analysis. Marit K. Bastian: Writing – review & editing, Methodology, Investigation. Katherine E. Free: Writing – review & editing, Methodology, Investigation. Hillary J. Kinsman: Writing – review & editing, Methodology, Investigation, Formal analysis. Irina Burd: Writing – original draft, Formal analysis, Conceptualization. Peter G. Napolitano: Writing – review & editing, Writing – original draft, Supervision, Formal analysis, Conceptualization. Nicholas Ieronimakis: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization.