The goals of this study were to elucidate immune compositional and functional shifts associated with fetal brain injury in a mouse model and to correlate our findings to human cord blood samples. Flow cytometry and ELISA data are each compiled from a minimum of two experiments per time point with our mouse model and an N=4-6 mice/group/experiment. Human cord blood was obtained from an existing biorepository of cord blood serum collected for detection of biomarkers of fetal pathologies.

These studies utilize a well-established mouse model of intrauterine inflammation created by our lab and resulting in fetal and postnatal brain injury (16, 17). For these experiments, timed-pregnant CD-1 mice were purchased from Charles River Laboratories (Wilmington, MA) and arrived at our animal facility at embryonic day 11 (E11). On E15, pregnant mice underwent a mini-laparotomy, while under isoflurane anesthesia. Each mouse received an injection into their right uterine horn, between the first and second amniotic sacs proximal to the cervix, as previously described (21–23). Each injection was either 100µl sterile phosphate buffered saline (PBS; saline control group) or 50µg/100µl lipopolysaccharide (LPS)/sterile PBS (IUI-exposed group). LPS was from E. coli 055:B5 (Sigma, St. Louis, MO). Mice were then euthanized by CO₂ either 48 or 72 hours post-IUI, with tissues collected only from mice that were still pregnant. All animal care and use procedures are approved by the University of Pennsylvania IACUC.

At either 48 or 72 hours post-IUI, mice were euthanized and tissues were collected into and washed with 10% charcoal-stripped fetal bovine serum (FBS; Gemini Bio, Sacramento, CA) in Hank’s balanced salt solution (HBSS) on ice. Maternal spleen was collected. Amniotic fluid was removed from each gestational sac with a 1ml syringe. The uterus was bisected and gestational sacs with fetal membranes and placentae were removed. The decidua was then scraped from uterus with a glass microscope slide into HBSS + 10% FBS on ice. Fetal tissues, placentae, fetal liver and fetal brains, were collected from the four fetuses proximal to the injection site and washed in HBSS + 10% FBS on ice. Tissues were either flash frozen in liquid nitrogen and stored at -80°C or immediately processed into a single cell suspension for flow cytometry.

Preparation of tissues for flow cytometry follows a protocol established with the placenta (24). Placentae were minced into 2mm pieces, suspended in 5ml of digest solution (HBSS + 10% FBS + 1mg/ml Collagenase IV, Gibco, Gaithersburg, MD), and incubated in a 37°C water bath for 30 minutes. Placentae (post-digest treatment), maternal spleens, decidua, fetal livers, and fetal brains were mechanically pressed through a 70µm cell strainer. The strainer was rinsed with 10ml of HBSS and strained cells were passed through a 40µm cell strainer into a 50ml conical tube. Cells were centrifuged at 1500rpm, 4°C for 10 minutes. Red blood cells were eliminated from placentae, maternal spleens, decidua, and fetal livers by suspension in 3ml of ACK lysing buffer (Gibco, Gaithersburg, MD) and incubation at room temperature for 10 minutes. ACK-treated cells and amniotic fluid samples were rinsed in 10ml of HBSS and centrifuged at 1500rpm, 4°C for 5 minutes. Single cell suspensions from maternal spleen, decidua, placentae, fetal livers, and amniotic fluid were suspended in 1ml of FACS buffer (PBS + 2% FBS + 20µM EDTA) and transferred into 5ml round-bottom tubes through 35µm filter caps. Cells were stored on ice until ready for staining.

Fetal brain single cell suspension was suspended in 5ml 35% Percoll and was slowly layered on top of 5ml 70% Percoll in a 15ml conical tube. Cells in Percoll gradient were centrifuged at 1000xg, 4°C, for 20 minutes without centrifuge break. Myelin and glial cells were removed from the top of the gradient and then 1ml at the interface was collected, containing mononuclear cells. Mononuclear cells were washed in 10ml HBSS and centrifuged at 1500rpm, 4°C for 10 minutes. Cells were resuspended in 1ml FACS buffer and transferred into 5ml round-bottom tubes through 35µm filter caps.

Cells were transferred to a 96-well round-bottom plate for staining, washed with 200µl of PBS, and the plate was centrifuged at 1500rpm, 4°C for 5 minutes. Cells were resuspended in 100µl of Live/Dead Fixable Aqua (Invitrogen, Waltham, MA) diluted 1:1000 in PBS and incubated for 30 minutes at 4°C in the dark. Cells were washed twice with 200µl of FACS buffer, centrifuging the plate between washes at 1500rpm, 4°C for 5 minutes. Cells were then resuspended in an extracellular antibody mix and incubated for 30 minutes at 4°C in the dark. Extracellular antibody mixes included: anti-CD3e-BV785, anti-CD4-BV605, anti-CD11c-PE.Cy7, anti-CD19-APC, anti-CD45-BV711, anti-CD49b-PacificBlue, anti-CD80-PacificBlue, anti-F4/80-FITC, anti-Ly6G-AF700, anti-TCRγ/δ-PerCP.Cy5.5 (BioLegend, San Diego, CA); anti-CD8a-APC.efluor780, and anti-Gr1-AF700 (eBioscience, San Diego, CA). The antibody clone and dilution used are listed in Table 1 . Cells were washed twice with 200µl of FACS buffer, centrifuging between washes at 1500rpm, 4°C for 5 minutes. Cells were resuspended in 200µl fixation/permeabilization reagent (Transcription Factor Staining Buffer Set, eBioscience, San Diego, CA) and incubated for 30 minutes at 4°C in the dark. Cells were washed twice with 200µl of permeabilization buffer (Transcription Factor Staining Buffer Set, eBioscience, San Diego, CA), centrifuging at 2000rpm, 4°C for 5 minutes between washes. Cells were resuspended in anti-FoxP3-PE (eBioscience, San Diego, CA) diluted 1:300 in 100µl permeabilization buffer and incubated for 30 minutes at room temperature in the dark. Cells were washed twice with 200µl of permeabilization buffer, centrifuging at 2000rpm, 4°C for 5 minutes between washes. Cells were then resuspended in 200µl of FACS buffer and transferred into 5ml round-bottom tubes through 35µm filter caps and stored at 4°C in dark. Single-stained CompBeads (BD Biosciences, Franklin Lakes, NJ) were used as compensation controls to create a compensation matrix, and fluorescence minus one-stained splenocytes and placental cells (FMOs) were used as gating controls to draw gates. See gating scheme ( Figure S1 ). All samples were analyzed by an LSR-II flow cytometer (BD Biosciences, Franklin Lakes, NJ).

Tissues to be used for ELISA were flash frozen in liquid nitrogen at the time of tissue harvest and were stored at -80°C. Amniotic fluid samples to be used for ELISA were centrifuged at 3000rpm for 5 minutes and the supernatant was flash frozen for future use. Protein extracts were made from placentae or fetal liver tissue by submerging 50mg of frozen tissue in 1ml of RIPA buffer with cOmplete mini protease inhibitor cocktail (Roche, Basel, Switzerland) in a 2ml round bottom tube with a 5mm steel bead. Tissue was homogenized on Tissue Lyser II (Qiagen, Venlo, Netherlands) for 10 minutes at 30/second. Homogenate was rested on ice for 40 minutes and then centrifuged at 14000xg for 10 minutes. Total protein in supernatant was quantified by the BCA protein assay kit (Pierce, Rockford, IL) following the manufacturer’s protocol. Quantikine ELISA kits (R&D Systems, Minneapolis, MN) for IL-6, IL-10, IFNγ, CCL3, and CCL5 were used to measure cytokine levels, following the manufacturer’s protocol. Cytokine levels were normalized to total protein.

Fetal brains, decidua and maternal spleen were collected 72 hours post-intrauterine injection of LPS and processed into a single cell suspension as noted above. CD3+ cells were isolated from the single-cell suspension using mouse CD3e microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) following the manufacturer’s protocol. Isolated CD3+ cells were resuspended at 10⁶ cells/ml in T cell stimulation media: RPMI-1640 with L-glutamine (Corning Inc., Corning NY) supplemented with 50µM beta-mercaptoethanol (Sigma-Aldrich, St. Louis, MO), 1X non-essential amino acids, 1X sodium pyruvate, 1X HEPES buffer (Gibco, Gaithersburg, MD), 1X Cell Stimulation Cocktail, 1X Protein Transport Inhibitor Cocktail (eBioscience, San Diego, CA), and 10% FBS. 100µl of each sample was plated on a 96-well round-bottom plate and incubated for 4 hours at 37°C. Cells were centrifuged at 400xg, 4°C, for 5 minutes, washed with PBS, resuspended in Live/Dead Fixable Aqua 1:2000 in 100µl of PBS, and incubated for 30 minutes at 4°C. Cells were then washed twice with FACS buffer; stained with extracellular markers, anti-CD45-BV711, anti-CD3ε-BV785, anti-TCRγ/δ -PerCP-Cy5.5, anti-TCRβ-APC-Cy7 (BioLegend, San Diego, CA); and incubated for 30 minutes at 4°C. Cells were washed twice with FACS buffer and resuspended in Fixation/Permeabilization reagent (eBioscience, San Diego, CA) for 30 minutes at 4°C. Cells were washed twice with Permeabilization Buffer and then incubated for 30 minutes at 4°C with intracellular cytokine staining antibodies: anti-TNFα-FITC (BioLegend, Sand Diego, CA), anti-IFNγ-BV421, and anti-IL-17A-PE (BD Biosciences, San Jose, CA). Cells were washed twice with Permeabilization Buffer and then resuspended in FACS buffer and stored at 4°C in dark until run on flow cytometer.

Cells were analyzed on an LSR-II flow cytometer (BD Biosciences, San Jose, CA) running FACSDiva software (BD Biosciences, San Jose, CA) in the University of Pennsylvania Flow Core. Flow cytometry data was then analyzed using FlowJo (BD Life Sciences, Ashland, OR) to identify immune cell subsets. High-dimensional, unbiased cell clustering and visualization tools were used to further analyze flow cytometry data. Specifically, t-SNE and FlowSOM algorithms were downloaded from the FlowJo Exchange and implemented using R/Bioconductor (25).

Cord blood was obtained from a previously published cohort of singleton pregnancies (N=723) (26, 27) of which 77 specimens were utilized in this study. Case specimens were selected as term neonates (>37 weeks gestational age at delivery) with documented HCA (N=41). Control specimens were also from term neonates (>37 weeks) that had placental histological examination but no HCA and were frequency matched for maternal race (N=36). All participants provided informed consent and the study was approved by the Institutional Review Board at the University of Pennsylvania (IRB #807678). Cord blood serum was analyzed by high sensitivity Milliplex assay (Millipore Sigma, Burlington, MA) for IFNγ and TNFα.

Statistical analyses were performed using GraphPad Prism. Data for every analysis is compiled from at least two experiments with 4-6 samples/group/experiment. Flow cytometry data was analyzed for differences between IUI and control samples collected at the same time – not for differences over time. All data was assessed for normality by Shapiro-Wilk test. Non-normally distributed data was analyzed by Mann-Whitney test. Normally distributed data was analyzed by unpaired student’s t-test, with Welch’s correction applied if the two groups differed in variance by F test. ELISA data and intracellular cytokine staining were analyzed by two-way ANOVA, followed by Tukey’s multiple comparison test if initial ANOVA was significant. Human cord blood data was non-normally distributed by Shapiro-Wilk test and analyzed for differences by Mann-Whitney test. Cord blood data was also divided by fetal sex and analyzed by two-way ANOVA to determine if fetal sex was a significant factor contributing to variance.

Our previous work utilizing our model of intrauterine inflammation (IUI) has demonstrated minimal maternal immune activation after IUI, specifically noted by a lack of maternal serum cytokine elevation (17). To confirm the hypothesis that the maternal systemic immune system is not impacted by IUI, we compared maternal splenic immune populations by flow cytometry from mice with IUI compared to saline-treated controls. From each spleen sample, 10⁶ events were run on the flow cytometer, leading to the requisition of approximately 5x10⁵ CD45+ cells ( Figure S2A ). At both 48- and 72-hours post-uterine injection, none of the following immune populations examined were significantly different between the two groups: macrophages, CD4+ T cells, CD8+ T cells, regulatory T cells, γ/δ T cells, B cells, NK cells, neutrophils, and dendritic cells ( Figure S2B–J ).

Decidua, the specialized uterine lining that develops during pregnancy, was collected from these same mice to examine immune changes local to the uterus at 48- and 72-hours post-uterine injection. The presence of IUI increased numbers of CD45+ cells in decidual tissue ( Figures 1A, D ), primarily due to an influx of decidual neutrophils ( Figures 1B, C, E ). This neutrophil influx was evident as both an increase in absolute number of decidual neutrophils ( Figures 1B ) and an increase in neutrophil frequency of decidual leukocytes ( Figure 1C ). The decidual neutrophil increase persisted in the mice with IUI at 72 hours post-uterine injection ( Figures 1B, C ) and these mice also had a slight elevation in decidual CD8+ T cells at 72 hours post-uterine injection ( Figures 1F ). The two most prevalent decidual immune cells, NK cells and macrophages (28, 29) did not differ between the IUI and control groups ( Figures 1G, H ).

To test the hypothesis that IUI would alter immune responses both within and across the placenta, flow cytometry was used to identify immune cells in the placentae and amniotic fluid of mice with IUI and controls. At 48 hours post-uterine injection, both the placentae and amniotic fluid of mice exposed to IUI had more total leukocytes than controls, but these levels normalized by 72 hours post-exposure ( Figures 2A–D ). Neutrophil increases in both the placenta and amniotic fluid accounted for most of the leukocyte increase in mice with IUI at 48 hours post-uterine injection and neutrophil counts in the placentae and amniotic fluid of mice with IUI also had normalized by 72 hours post-uterine injection ( Figures 2E, F ). Placentae, but not amniotic fluid, from mice with IUI had a small increase in macrophages ( Figures 2G, H ), but no other immune cell differences from control animals were observed by 72 hours post-uterine injection.

Given the increase in CD45+ cells in the placentae and amniotic fluid of mice with IUI, we asked whether this was accompanied by an increase in chemokines that may recruit CD45+ cells to these tissues. Our previous work with this model of IUI had demonstrated significant elevations in CCL3, CCL5, and IL-6 in both the placenta and amniotic fluid at just 6 hours post-exposure (16, 17, 30), so we investigated whether these chemokines remained elevated at later time points. CCL5 was the most increased: 31-fold in the placenta and 8-fold in the amniotic fluid of mice with IUI at 48 hours post-uterine injection; but protein levels of this chemokine normalized to control levels by 72 hours post-uterine injection ( Figures 3B, E ). CCL3 was also elevated in the placenta and amniotic fluid in mice with IUI at 48 hours and normalized by 72 hours post-uterine injection ( Figures 3A, D ). IL-6 was elevated in the placenta but not the amniotic fluid of mice with IUI at 48 hours and was not significantly increased in either tissue at 72-hours post-injection ( Figures 3C, F ).

Given that immune alterations were detected in the amniotic fluid of IUI-exposed mice, we hypothesized that the fetal immune system would be perturbed. In the fetal livers of mice exposed to IUI, CCL3 and IFNγ were elevated at 48 hours post-uterine injection and had normalized by 72 hours ( Figures 4A, C ). There was no significant difference in CCL5 levels between IUI-exposed and control fetuses ( Figure 4B ). No cellular immune changes were noted at 48 hours in the fetal liver. However, following the observed cytokine elevation at 48 hours, at 72 hours post-uterine injection, fetuses exposed to IUI had more CD45+ cells ( Figures 4D, E ), neutrophils ( Figures 4F, G ), and macrophages ( Figures 4H, I ) than control fetuses in their fetal livers.

To address whether an inflammatory insult in the uterus could ultimately alter immune cell composition and function in the fetal brain, we isolated mononuclear cells from fetal brains and identified the immune cell types by flow cytometry (gating scheme in Figure 5A ). At 48 hours post-uterine injection, the only detectable difference in the fetal brain immune cells was a 5-fold increase in granulocytes in the IUI-exposed group ( Figure 5F ). However, by 72 hours post-uterine injection, fetal brains from IUI-exposed dams had more overall CD45+ cells ( Figure 5B ) and specifically non-microglial CD45+ cells ( Figure 5E ), including 7-fold more granulocytes ( Figure 5F ). IUI-exposure increased the numbers of activated microglia and macrophages but did not affect the numbers of resting microglia ( Figures 5C, D, G ).

Given limited studies on what immune cells can be found in the fetal brain, we relied on unbiased clustering algorithms to determine the immune population with the greatest difference between fetal brains from IUI-exposed and control dams. tSNE analysis creates a visual map of the different CD45+ cells found in the fetal brain, with a clear population enriched in the brains of fetuses exposed to IUI (black box, Figure 6A ). The flowSOM algorithm generated ten different clusters of immune cells in the fetal brain (different colored clusters in Figure 6A ). The absolute difference between the frequency of each flowSOM cluster in the fetal brains of IUI-exposed versus control mice was calculated and the population with the greatest difference was identified (purple population, Figures 6A, B ). This population accounted for 7% of all CD45+ cells in the brains from IUI-exposed fetuses, but less than 1% of CD45+ cells in brains of control fetuses. This immune cell population was Gr-1+ CD11b+ γ/δ T cells (histograms in Figure 6B ). Total γ/δ T cells were elevated in the fetal brains from IUI-exposed dams at both 48 and 72 hours post-uterine injection ( Figures 6C, D ). Total γ/δ T cells were also increased in the decidua of IUI-exposed dams at 72 hours post-uterine injection ( Figure S3A ), but were not changed in the spleen ( Figure S2F ), placenta, amniotic fluid or fetal liver ( Figures S3B–D ) by IUI.

To elucidate the function of the γ/δ T cells in the IUI-exposed fetal brains, T cells were isolated from pregnant mice 72-hours after exposure to IUI from the maternal spleen, decidua, and fetal brain. T cells were stimulated in vitro for four hours and stained for intracellular cytokines. Cells were also stained for TCRβ and TCRγδ to differentiate between α/β and γ/δ T cells. Cytokine production by total T cells ( Figures 7A–C ) and specifically γ/δ T cells ( Figures 7D–F ) varied immensely by anatomic location. Approximately 10% of all splenic T cells produced at least one measurable cytokines, 33.7% of decidual T cells produced TNFα, and nearly 100% of T cells in the fetal brain produced at least one cytokine ( Figures 7A, B ). Of fetal brain γ/δ T cells, 9.1% produced IL-17A and 7.9% produced TNFα; while 89.4% of fetal brain γ/δ T cells produced IFNγ, with some γ/δ T cells producing multiple cytokines ( Figures 7D, E ). Whether examining total T cells or only γ/δ T cells, only those isolated from the fetal brain were making significant IFNγ ( Figures 7C, F ).

Cord blood was collected from term pregnancies (delivery at greater than 37 weeks) and placentae from these pregnancies were examined for histological chorioamnionitis (HCA) as part of clinical care. Term HCA was used as a human correlate to our mouse model of low dose intrauterine inflammation without preterm birth, as HCA is defined by placental neutrophil infiltrates (7). A case control study was performed. Cases were defined by the finding of HCA on placental pathology (N=36) compared to controls who did not have HCA (N=41). Cases and controls were frequency matched by self-reported race ( Table 2 ). IFNγ and TNFα were measured in cord blood by a high-sensitivity Luminex assay. IFNγ levels were 1.8-fold higher in cord blood from neonates with HCA compared to controls (p<0.0001), but TNFα was equivalent between cases and controls ( Figures 8A, B ), indicating an IFNγ-specific inflammatory signature in neonates with HCA. Fetal sex did not contribute to cytokine level variance.