Objective
Magnesium sulfate is proposed to have neuroprotective effects in the offspring. We examined the effects of maternal magnesium sulfate administration on maternal and fetal inflammatory responses in a rat model of maternal infection.
Study Design
Pregnant rats were injected with saline, Gram-negative bacterial endotoxin lipopolysaccharide or lipopolysaccharide with magnesium sulfate (pre- and/or after lipopolysaccharide) to mimic infection. Maternal blood, amniotic fluid, fetal blood, and fetal brains were collected 4 hours after lipopolysaccharide and assayed for tumor necrosis factor, interleukin-6, monocyte chemoattractant protein-1, and growth-related oncogene-KC. In addition, the effect of magnesium sulfate on cytokine production by an astrocytoma cell line was assessed.
Results
Lipopolysaccharide administration induced tumor necrosis factor, interleukin-6, monocyte chemoattractant protein-1, and growth-related oncogene-KC expression in maternal and fetal compartments. Maternal magnesium sulfate treatment significantly attenuated lipopolysaccharide-induced multiple proinflammatory mediator levels in maternal and fetal compartments.
Conclusion
Antenatal magnesium sulfate administration significantly ameliorated maternal, fetal, and gestational tissue-associated inflammatory responses in an experimental model of maternal infection.
Preterm birth is a major cause of perinatal morbidity and accounts for approximately 70% of all perinatal mortality. Maternal infection, a major risk factor for preterm labor, is associated with significant neonatal morbidity, including neurologic injury. The maternal inflammatory response to infection is characterized by the production of proinflammatory mediators such as cytokines and chemokines, which can lead to tissue injury when produced in excess.
Lipopolysaccharide (LPS), the Gram negative bacterial endotoxin, has been used in laboratory animals to characterize the underlying pathogenesis of inflammation/infection-mediated preterm labor, to investigate infection-associated fetal neuronal damage, and to test various potential therapeutic interventions. Endotoxin administration to pregnant laboratory animals induces proinflammatory mediator production in the maternal and fetal compartments. This model mimics the elevated levels of cytokines and chemokines (eg, tumor necrosis factor-alpha [TNF], interleukin-6 [IL-6], monocyte chemoattractant protein [MCP-1, CCL2]) observed in maternal and fetal compartments during infection and/or preterm laboring humans. Excessive and sustained inflammatory mediator production is proposed to activate the prostaglandin cascade leading to preterm birth, a risk factor for neurologic damage. In addition, fetal exposure to inflammatory mediators is understood to be linked to neurologic insults in the offspring, including long-term neurocognitive dysfunction and cerebral palsy.
Several reports support the neuroprotective effect of magnesium sulfate (MgSO 4 ) in the fetus. A randomized clinical trial by Rouse and colleagues showed a significantly lower rate of cerebral palsy among infants born to women at 24-31 weeks’ gestation who received MgSO 4 tocolysis for preterm labor. A recent metaanalysis and the Cochrane Review also supported this association.
Although MgSO 4 has been used in obstetrics for suppressing preterm labor, its exact mechanism of action is unknown. Research in our laboratory revealing that MgSO 4 attenuated cytokine production by cultured endothelial cells highlights the antiinflammatory activity of MgSO 4 . Additional in vivo studies demonstrating exaggerated inflammatory responses in Mg-deficient animals in the absence and presence of LPS further support the antiinflammatory effects of MgSO 4 .
Given the emerging evidence for (1) the association between neuronal injury and cytokine-mediated damage in the fetus, (2) the potential protective role of antenatal MgSO 4 therapy on the development of cerebral palsy, and (3) the antiinflammatory activity of MgSO 4 , we sought to investigate the effects of MgSO 4 on cytokine and chemokine concentrations within maternal and fetal circulation, as well as in the amniotic fluid and fetal brain using a rodent model of LPS-induced maternal systemic infection, and to examine the effects of MgSO 4 on inflammatory mediator production by a human neuronal cell line.
Materials and Methods
We obtained approval from the Institutional Animal Care and Use Committee (IACUC) before conducting our study. Timed pregnant Sprague-Dawley rats (E9 and E11) (Charles River, Wilmington, MA) were housed individually under normal environmental conditions with free access to standard rat chow and water for at least 3 days before experimentation. LPS (strain Escherichia coli 055:B5) was purchased from Sigma-Aldrich (St. Louis, MO). MgSO 4 was purchased from APP Pharmaceuticals, LLC (Schaumburg, IL). Pregnant rats (gestational day 19) were randomly assigned by a technician blinded to the study design into 1 of 5 treatment groups (n = 6 per group): (1) saline (given subcutaneously [sc] q20 minutes for a period of 4 hours before and a period of 4 hours after saline intraperitoneally [ip]) (SSS)-negative controls; (2) LPS (1 mg/kg, ip) and saline (sc q20 minutes for a period of 4 hours before and a period of 4 hours after LPS) (SLS) to mimic infection-associated preterm labor; (3) LPS (1 mg/kg, ip) and MgSO 4 (270 mg/kg load and then 27 mg/kg sc q20 minutes for a period of 4 hours before and for a period of 4 hours after LPS) (MLM); (4) LPS (1 mg/kg, ip) and MgSO 4 (270 mg/kg load and then 27 mg/kg sc q20 minutes 4 hours before LPS only and saline given sc q20 minutes 4 hours after LPS) (MLS); or (5) LPS (1 mg/kg, ip) and saline given sc q20 minutes for a period of 4 hours before LPS and MgSO 4 (270 mg/kg load and then 27 mg/kg sc q20 minutes) for 4 hours after LPS (SLM). A similar MgSO 4 regimen was used by Hallak et al and the LPS administration protocol (route, dose, timing) was similar to Kumral et al. Maternal blood was drawn (retroorbitally) 1½ hours after ip administration of LPS or saline. Four hours post-LPS (or saline) injection, the dams were euthanized by CO 2 (1 group of rats were euthanized 90 minutes post-LPS for fetal brain messenger RNA [mRNA] studies). Maternal blood was collected via cardiac puncture, and amniotic fluid was collected from each gestational sac. Fetal blood (pooled from the pups of each dam) and brains were obtained at the time of decapitation. Blood specimens were collected in nonheparinized tubes, the blood was allowed to clot for 30 minutes, and then specimens were centrifuged to collect serum. Fetal brains were flash frozen in liquid nitrogen, and both brain and serum samples were stored at −80° C.
Magnesium (Mg) concentrations of maternal blood and amniotic fluid were assessed by the North Shore-LIJ Health System Core Laboratories. For cytokine assessment, fetal rat brains were homogenized in 4 volumes of phosphate-buffered saline (PBS) containing 0.1% NP-40 and protease inhibitors. Cytokines/chemokines (TNF, IL-6, MCP-1, and growth-related oncogene-KC [GRO-KC, CXCL1]) in the rat serum, amniotic fluid, and brain homogenates were analyzed using Luminex XMAP technology (Millipore, St. Louis, MO). Fetal brain cytokine/chemokine concentrations were adjusted for protein concentrations (Bio-Rad protein method, Hercules, CA). Cytokine data are shown as individual data points with geometric means. The sensitivities (pg/mL) of the assays were as follows: TNF: 4.88-24.4; IL-6: 9.8-24.4; MCP-1: 3.81-4.88; and GRO-KC: 2.06-18.44 and the precisions (intraassay/interassay, %CV) were as follows: TNF: 9.16/11.1; IL-6: 10.37/14.3; MCP-1: 3.81/16.5; and GRO-KC: 7.39/14.1.
The RNA was isolated from frozen fetal brains (at 90 minutes and 4 hours post-LPS) using the RNeasy kit (Qiagen, Valencia, CA) and RNA preparations with A260:280 ratios >1.9 were analyzed. The relative expression of GRO-KC, MCP-1, and CXCR2 mRNAs were assessed by quantitative real-time polymerase chain reaction (Q-PCR) using Roche Universal Probe Library technology. Reactions (performed in duplicate) were completed using 200 ng RNA, Eurogentec onestep qRT-PCR master mix, and 7900HT Fast Real-Time PCR system (Applied Biosystems, Foster City, CA). Conditions used for the RT-PCR reaction were: 48° C (30 minutes) and 95° C (10 minutes), then 45 cycles of 95° C (15 seconds) and 60° C (1 minute) using specific primer sequences: GRO-KC: F-cacactccaacagagcacca, R-tgacagcgcagctcattg; MCP-1: F-agcatccacgtgctgtctc R-gatcatcttgccagtgaatgagt; CXCR-2: F-atctttgctgtggtcctcgt, R-tgaacaggacaatgttgtaggg. Relative gene expression levels were calculated using the ΔΔCt method and expression levels were normalized to the housekeeping gene GAPDH (Universal Probe Library Rat GAPDH Gene Assay, Roche). Data are presented as mean fold change over saline control (mean ± SEM).
To investigate the effects of MgSO 4 on brain cytokine responses ex vivo, the U87-MG human neuronal gliobalstoma astrocytoma cell line (ATCC, Manassas, VA) was cultured in 96-well plates in DMEM containing 10% fetal bovine serum and penicillin-streptomycin, and then treated with vehicle or MgSO 4 (0.1-25 mM) for 3 hours before TNF treatment (50 ng/mL) or IL-1β (1 ng/mL) (n ≥4 wells per condition). After an overnight incubation, cell-free culture supernatants were collected and assayed for MCP-1 and IL-8 by ELISA (R&D Systems, Minneapolis, MN). All assays were repeated 2 times and data are shown as mean (± SEM).
Statistical analyses
The rat cytokine data (normalized by a log transformation) was analyzed in 2 steps. First, 2-sample t tests, using a log transformation of the cytokine levels, were used to compare the SSS (negative control) group with the SLS (positive control) group to confirm induction by LPS ( P < .05). In step 2, separate 1-way analysis of variances (ANOVAs) were used to analyze the cytokines concentrations in the maternal and fetal compartments. On significant ANOVA findings ( P < .05), each group (MLS, SLM, and MLM) was compared with the positive control group (SLS) using the Dunnett’s test. The standard assumptions of Gaussian residuals and equality of variance were tested. Q-PCR data were analyzed with 1-way ANOVAs with significance set at P < .05.
Results
At 90 minutes after maternal LPS administration (SLS) (when serum TNF levels peak), we observed a significant induction of TNF ( P < .0001), IL-6 ( P < .0292), MCP-1 ( P < .0007), and GRO-KC ( P < .0001) levels in the maternal serum when compared with control animals (SSS) ( Figure 1 , A-D). Treatment of pregnant animals with MgSO 4 pre- and post-LPS (MLM) significantly decreased LPS-induced TNF ( P < .0002), IL-6 ( P < .0289), MCP-1 ( P < .0234), and GRO-KC ( P < .0026) levels in the maternal blood at 1½ hrs post-LPS compared with saline-LPS treated animals (SLS) ( Figure 1 , A-D). No significant reductions in serum inflammatory concentrations were observed when MgSO 4 was given only before (MLS) or only after LPS treatment (SLM).
As expected, the levels of inflammatory mediators (TNF, P < .0002; IL-6, P < .0001; MCP-1, P < .0019; and GRO-KC, P < .0001) remained significantly elevated in the maternal serum when assessed 4 hours post-LPS (SLS) compared with controls (SSS) ( Figure 2 , A-D). Consistent with the results at 90 minutes post-LPS, we found that maternal treatment with MgSO 4 (pre- and post-LPS, MLM) significantly decreased TNF ( P < .0002), IL-6 ( P < .0013), and MCP-1 ( P < .0044) levels in the maternal serum when compared with saline-LPS treated animals (SLS) ( Figure 2 , A-C). Treatment with MgSO 4 pre- and post-LPS (MLM) reduced GRO-KC levels in the maternal blood compared with saline-LPS treated dams (SLS) at 4 hours post-LPS; however, this was not significant ( Figure 2 , D). MgSO 4 treatment when given only before LPS administration (MLS) significantly decreased TNF in maternal serum at 4 hours post-LPS by 3.1 fold ( P < .0296) compared with saline-LPS-treated animals (SLS) ( Figure 2 , A). Maternal serum Mg concentrations (mg/dL) at 4 hours after saline or LPS treatment were as follows: SSS: 2.6 ± 0.52; SLS: 3.0 ± 0.84; MLM: 6.067 ± 1.89 ( P < .05 vs SLS); MLS: 4.03 ± 1.32; SLM: 10.5 ± 2.85 ( P < .001 vs SLS). (MgSO 4 was first given as a bolus, followed by serial injections once every 20 minutes, as described in the Materials and Methods section.)
Consistent with previous studies demonstrating that maternal infections trigger fetal inflammatory responses, we observed a significant induction of TNF ( P < .0098), IL-6 ( P < .0001), MCP-1 ( P < .0032), and GRO-KC ( P < .0001) in the amniotic fluid 4 hours after maternal LPS administration ( Figure 3 , A-D). Amniotic fluid levels of TNF, IL-6, and MCP-1 ( Figure 3 ) were approximately 8- to 10-fold lower than those observed in maternal blood ( P < .01, P < .01, P < .001 for TNF, IL-6, and MCP-1, respectively, Figure 2 ), whereas amniotic GRO-KC levels were similar to maternal blood concentrations. TNF ( P < .0076), IL-6 ( P < .0047), MCP-1 ( P < .0036), and GRO-KC ( P < .006) levels in the amniotic fluid significantly declined in response to antenatal MgSO 4 therapy when administered pre- and post-LPS (MLM) ( Figure 3 , A, B, D) and only amniotic GRO-KC levels were significantly lower when MgSO 4 was given before LPS injection only (MLS) ( P < .0420) compared with the saline-LPS-treated animals (SLS) ( Figure 3 , D). Amniotic fluid Mg concentrations (mg/dL) at 4 hours after saline or LPS treatment were as follows: SSS: 2.86 ± 0.23; SLS: 3.13 ± 0.18; MLM: 3.87 ± 0.28 ( P < .002 vs SLS); MLS: 3.48 ± 0.2414; SLM: 3.73 ± 0.16.
Similar to our observations with cytokines in the amniotic fluid, we found that LPS administration (SLS) significantly increased fetal serum TNF, IL-6, MCP-1, and GRO-KC levels (assessed 4 hours post-LPS) when compared with saline-treated control animals (SSS) ( Figure 4 , A-D). At 4 hours post-LPS maternal administration, TNF, IL-6, and MCP-1 levels were significantly lower in the fetal blood ( Figure 4 ) compared with (1) maternal blood levels ( P < .01, P < .01, P < .001 for TNF, IL-6, and MCP-1, respectively, Figure 2 ) and (2) amniotic fluid levels of IL-6 ( P < .001) and MCP-1 ( P < .02) ( Figure 3 ). Although fetal serum concentrations of IL-6 and GRO-KC were reduced in response to MgSO 4 treatment (MLM), only fetal serum TNF ( P < .0281) levels were significantly decreased by MgSO 4 when administered pre- and post-LPS (MLM) when compared with saline-LPS treated animals (SLS) ( Figure 4 , A). Fetal blood volumes were too low to assess Mg concentrations.
