Chorioamnionitis Induces a Unique Time Course of Inflammatory Changes and Immune Reponses in the Brain and Spleen

Abstract

Introduction:

Inflammatory oxidative micro-environments can alter heme oxygenase-1 (HO-1) homeostasis. Dysregulation of HO-1 favors proinflammatory signals while transferrin receptor 1 (TfR1) regulates anti-inflammatory signal transduction. Previously, we have shown that chorioamnionitis (CHORIO) induces sustained elevations in HO-1/TfR1 at postnatal day (P)2 and mononuclear cell-driven inflammation at term age equivalent P7. Here, we hypothesized that an altered HO-1/TfR1 developmental time course would coincide with inflammatory/immune signal changes in the brain.

Methods:

To induce CHORIO in rats, we performed a laparotomy followed by bilateral transient uterine artery occlusion and intra-amniotic injection of lipopolysaccharide (LPS) at E18. The control group received laparotomy only with equivalent duration of anesthesia. We used realtime polymerase chain reaction (RT-PCR), multiplex electrochemiluminescent immunoassay (MECI), and flow cytometry (FC) to study changes in pro- and anti-inflammatory gene expression, immune cell secretome, and immune cells at critical and clinically relevant timepoints following CHORIO.

Results:

We found an acute reduction in an anti-inflammatory signals in the cortex on embryonic day (E)19. This was followed by an increased pro-inflammatory signature on postnatal day (P)2. There were also significant alterations in the phenotypic distribution of splenic T cells on P7, a key organ in immune function. Furthermore, we showed that the micro-environment of the cortex at P21 was skewed towards a pro-inflammatory state by significant increases in IL-6. The prominence of T-helper cells (Th) in brain concomitant with a proinflammatory state at P21 suggests emerging inflammation and potential for neural injury.

Conclusions:

Defining how these and other inflammatory/immune signatures contribute to perinatal brain injury and investigating distinct immune signatures at more developmental time courses will be beneficial for targeting emerging therapies.

Introduction

Chorioamnionitis (CHORIO), inflammation, oxidative stress, hypoperfusion of the placenta and brain, are common risk factors for brain injury in preterm neonates [1–6]. Specific associations of CHORIO with lasting brain injury and adverse neurodevelopmental outcomes in vulnerable preterm neonates have been documented in preclinical and clinical studies [1, 7–10]. CHORIO is defined as the inflammation of the chorioamnionic plate of the placenta and can lead to fetal inflammatory response syndrome (FIRS),[11], an independent risk factor for neonatal morbidities [12]. Inflammation in preterm neonates can alter brain cortical development by affecting oligodendrocytes, neurogenesis, neuronal migration, maturation, differentiation, or by activating microglia and astrocytes and decreasing cortical myelin [13–18]. Furthermore, CHORIO induces an immunotoxic exposure that alters the fetomaternal interface within critical neural and immune developmental windows and can have lifelong consequences by reprogramming the neonatal immune system, leading to neuroimmune and peripheral immune alterations [6, 19–25].

Dysregulation of heme oxygenase-1 (HO-1) is associated with pregnancy complications and many diseases in the neonate[26–31]. Inflammatory/oxidative micro-environments can affect HO-1 homeostasis. In the absence of an infection or inflammation, HO-1 is developmentally regulated [32, 33]. HO-1 and transferrin receptor-1 (TfR1) play important roles in regulating cytokines, chemokines, differentiation of innate and adaptive immune cells, and immune function [34, 35]. Perturbed HO-1 regulation promotes pro-inflammatory pathways. In contrast, TfR1 controls anti-inflammatory signal transduction [34, 35].

Previously, we demonstrated elevations in cerebral cortex HO-1 mRNA in utero and in the early perinatal period concomitant with sustained increase in HO-1/TfR1 ratios after chorioamnionitis [34]. Concurrently, CD45⁺CD11b/c⁺ mononuclear cells at P7, term-equivalent age, were increased [34]. These data support that CHORIO disrupts immune system maturation and alters the developmental regulation of HO-1 and TfR1. Given the importance of HO-1 and TfR1 in oxidoreductive balance, iron metabolism, and immune regulation [32, 36–42], we extended our studies to determine how CHORIO affects cerebral inflammation and central and peripheral immunoreactivity through P21. We hypothesized CHORIO induces dynamic pro- and anti-inflammatory and immune signature changes in the developing cerebral cortex. Furthermore, we posited a pro-inflammatory reprogramming of T-cell populations in the periphery and central nervous system (CNS).

Materials and Methods

Animals

Timed-pregnant adult female Sprague-Dawley rats were purchased from Charles River Laboratories (Wilmington, MA, USA) and were acclimated. All experimental procedures were approved by the Institutional Care and Use Committee (IACUC) at Johns Hopkins University School of Medicine. ARRIVE guidelines were followed [43].

CHORIO and Sham (Control) Groups

CHORIO was induced as described previously [1, 10, 23, 24, 34, 44]. Briefly, pregnant dams were first randomized to sham or CHORIO groups. On embryonic day (E)18, laparotomies were performed. For the CHORIO group, uterine arteries were transiently occluded for 60 min, and lipopolysaccharide (LPS, Escherichia coli, serotype O111:B4, Sigma, St Louis, MO; 4 μg/amniotic sac) was injected intra-amniotically. For the sham group, uteri were externalized; however, uterine arteries were not occluded nor was LPS injected. All groups received equivalent duration of isoflurane anesthesia. After laparotomy closure, dams were allowed to recover in their respective cages with adequate pain control. At E18.5 and E19 (perinatal timepoints), a subset of dams was euthanized for collection of fetal tissues. For postnatal studies (P2 to P21), dams were allowed to deliver their litters at E22. Offspring were housed with their birth mother. Both sexes of fetuses or offspring were used in all experiments. A total of 22 sham and 33 CHORIO litters were used for experiments at E18.5, E19, P2, P7 and P21 timepoints.

Tissue Collection

At E18.5 or E19, a subset of dams was euthanized by decapitation and fetal brains and spleens were immediately collected and snap frozen. For postnatal timepoints (P2 to P21), pups were sedated briefly with an intraperitoneal injection of ketamine plus xylazine. Forebrains at E18.5 and E19, whole brain cortices at P2, P7, and P21 were immediately flash frozen and stored at −80°C for quantitative polymerase chain reaction (qPCR). For flow cytometry (FC) studies, the pups were not perfused but we first collected their blood by intracardiac access and then utilized their spleens at P7, and brains at P21, to assess immune status of the pups at term-equivalent age and toddler age, respectively. In addition, at P21, blood was collected by intracardiac blood aspiration from the left ventricle and then centrifuged at 6000xg for 15 min at 4°C to separate serum. Serum was carefully pipetted into 1.5-mL cryotubes and stored at −80°C. Whole spleens and micro dissected brains were stored at −80°C until use for qPCR and multiplex electro chemiluminescent immunoassays (MECI).

RNA Extraction and Quantitative Real-Time Polymerase Chain Reaction (RT-PCR)

RNA was isolated from E18.5 (n = 7 chorio and 7 sham), E19 (n = 6 and 6), P2 (n = 7–15 and 7–24), P7 (n = 6 and 7), and P21 (n = 7 and 7) sham and CHORIO brains, respectively. Total RNA was extracted using the Direct-zol RNA Miniprep Quick protocol (Zymo Research, Irvine, CA, USA). RNA concentration and purity were measured using NanoDrop. cDNA reaction was completed on 900 ng of total RNA using Bio-Rad 5x iSCRIPT reverse transcription SuperMix for RT-PCR on a MiniAmp thermal cycler (ThermoFisher Scientific, Waltham, MA, USA). We compared △△CT values to naïve pooled E19 whole brains at every age, pooled naïve E19 comparator internal control was chosen for developmental comparison, and normalized to 18S endogenous control as previously published [1, 45]. Primer sequences for tumor necrosis factor-alpha (TNF-α), cluster of differentiation (CD) 86 and Arginase-1 (Arg1) are listed in Table 1. Primer sequences were verified using the Basic Local Alignment Search Tool (BLAST) for Nucleotides on National Center for Biotechnology Information (NCBI) website. RT-PCR data are presented as fold change. We employed strict a priori criteria for experimental/technical replicates. All samples were run in triplicate. When triplicates exceeded 0.25 SD of the mean CT value, we repeated that evaluation. To standardize transcripts between experiments, Ct from samples of interest were compared to Ct values from pooled samples from naïve E19 brains, with gene-of-interest transcription normalized to a 18s ribosomal endogenous control[24, 34].

Table 1:

Inflammatory-Immune Gene Primer Sequences.

Target	Primer Sequence
TNF-α	(F)5’GTG ATC GGT CCC AAC AAG GA3’ (R)5’CTC CCA CCC TAC TTT GCT TGT G3’
CD86	(F)5’TAGGGATAACCAGGCTCTAC3’ (R)5’CGTGGGTGTCTTTTGCTGTA3’
Arg1	(F) 5’CCACAGTCTGGCAGTTGGAAG3’ (R) 5’GGTTGTCAGGGGAGTGTTGATG3’
18S	(F)5’TCCCTAGTGATCCCCGAGAAGT3’ (R)5’CCCTTAATGGCAGTGATAGCGA3’

Isolation of Single Cells

Spleens from sham (n = 7) and CHORIO pups at P7 (n = 6) were mechanically dissociated using a Gentle MACS Dissociator (Miltenyi Biotec, Auburn, CA, USA). Cells were filtered through a BD Falcon 70-μm cell strainer and washed with ice-cold PBS followed by filtering through a 35-μm cell strainer prior to staining.

Single cell suspensions were prepared from sham (n = 7) and CHORIO brains at P21 (n = 6) using a Neural Tissue Dissociation Kit (T) (Miltenyi Biotec) following manufacturer’s protocol. Briefly, after removing the frontal poles and cerebellum, 400 mg of brain tissue from bilateral hemispheres were placed in an enzyme T and buffer X mix at 37°C. Then, serial mechanical and enzymatic digestions steps were completed utilizing Miltenyi Gentle MACS Dissociator and MACS Mix Tube rotator, respectively. Single cells were filtered using a BD Falcon cell strainer (70 μm) and washed with Hank’s Balanced Salt Solution (HBSS). Single cells were then resuspended in PBS and then layered onto Ficoll Paque Plus (GE Healthcare, Chicago, IL, USA) to separate out leukocyte populations.

Multiparameter FC

Antibodies for FC were purchased from BD Bioscience (San Jose, CA, USA), BioLegend (San Diego, CA, USA), and Novus Biologicals (Littleton, CO, USA) for anti-CD45 PerCp-Cy5.5, purified mouse anti-rat CD32 Fc block, anti-CD3 APC, anti-CD4 APC Cy7, anti-CD8 Alexa Fluor 594, anti-CD25 BV421, anti-FoxP3 FITC, and Live/Dead Fixable Aqua Dead Cell Stain Kit, for 405 nm from Thermo Fisher. Single cells were counted using a Countess^™ II FL Automated Cell Counter (ThermoFisher Scientific). 1 × 10⁶ live single mononuclear spleen and brain cells were incubated with viability dye in PBS on ice for 30 min. After viability staining, cells were incubated with anti-CD32 Fc Block and stained with T-cell antibody cocktail in MACS buffer. Fluorescent minus one (FMO) gating controls, unstained similarly-treated single cells were used for detecting autofluorescence and OneComp eBeads were used for compensation controls, as previously described [34]. Multiparameter FC was performed using a Cytek Aurora and data analyzed using FCS Express Plus 7 software (De Novo Software, Pasadena, CA, USA) as previously described [34].

Immune Phenotyping

We gated on forward scatter height versus forward scatter area to exclude debris and doublets. Then, we sequentially gated on live cells, CD45⁺ leukocytes, CD3⁺CD4⁺ T-helper (Th) cells, CD3⁻CD4⁺ immature T cells, CD3⁺CD4⁺CD25⁺Foxp3⁺ T-regulatory cells (Tregs), Median fluorescent intensity (MFI) Tregs, CD3⁺CD8⁺ cytotoxic T cells, CD3⁺CD4⁻CD8⁻ double-negative T cells and CD3⁻CD4⁻ cells.

MECI

We used a V-plex pro-inflammatory panel 2 (rat) kit to assay the following cytokines and chemokines: interferon gamma (IFN-γ), interleukin-1 beta (IL-1β), IL-4, IL-5, IL-6, chemokine (C-X-C motif) ligand 1 (CXCL-1), IL-10, IL-13, and TNF-α. Briefly, sera, micro dissected cortices and spleens were collected at P21, and immediately flash frozen. Cortices and spleens were homogenized in the presence of protease and phosphatase inhibitors in a buffered sucrose solution, centrifuged at 4200xg for 10 min. Protein concentrations in supernatants were determined using the Bradford assay (Bio-Rad, Hercules, CA, USA). 100 μg of protein per tissue (cortex, spleen) and 1:3 dilutions of serum were measured in duplicate on multi-spot plates per manufacturer’s protocol (Mesoscale Discovery, Gaithersburg, MD, USA). Plates were read on a Quickplex SQ 120 Imager (Mesoscale Discovery) [23–25, 44]. MECI was performed in sham and CHORIO pups at P21 on spleens (n = 7 and 7, respectively), sera (n = 7 and 7, respectively), and cortices (n = 7 and 7, respectively).

Statistical Analyses

Prism Software Version 8.3.0 (GraphPad, San Diego, CA, USA) was used for all data analyses. Normality was tested by Shapiro-Wilk test. For normally distributed data, we used Student’s t-test with Welch’s correction or ordinary 2-way ANOVA with Bonferroni post-hoc correction for multiple comparisons. For non-parametric data, Mann-Whitney or Kruskal-Wallis with Dunn’s multiple comparisons test were used. Data are presented as mean ± standard error of the mean (SEM). Minimum 6 samples from different litters were used in all experiments [34, 46, 47].

Results

Ontogeny of Inflammatory/Immune Gene Expression in the Sham Cortex

Baseline inflammatory/immune gene expression was first determined in sham control cortex and expressed as relative expression (RQ) after normalization to 18S levels. Figure 1a shows that TNF-α expression was significantly higher at E18.5 (8.0 ± 1.5) compared with that at E19 (0.13 ± 0.06, p = 0.001), at P2 (0.33 ± 0.19, p < 0.01), at P7 (0.39 ± 0.17, p < 0.05) and at P21 (0.34 ± 0.16, p < 0.01). Cortical CD86 and Arg1 expressions (Fig. 1b and Fig1c), however, was not significantly different at E18.5 (1.6 ± 0.20), E19 (3.7 ± 0.99), P2 (2.8 ± 0.35), P7 (2.9 ± 0.55), or P21 (2.07 ± 0.69).

Fig. 1.

Inflammatory/immune gene expression in the sham and CHORIO cortex E18.5 through P21. a. TNF-α; b. CD86; and c. Arg1. Data are presented as mean ± SEM fold change as normalized to 18S levels, ordinary two-way ANOVA with Bonferroni correction for multiple comparisons. E18.5 (n = 7), E19 (n = 6), P2 (n = 7 to 15), P7 (n = 3), and P21 (n = 7). RQ defines relative quantification. *, p < 0.05; **, p < 0.009; ***

CHORIO Induces Cortical Pro- and Anti-Inflammatory Gene Expression in the Perinatal Period

After we established the ontogeny of classical whole cortical pro- (TNF-α, CD86) and anti-inflammatory (Arg1) gene expression from the perinatal period (E18.5) to toddler-equivalent age (P21) in controls, we determined the impact of CHORIO during this developmental window. This timeline marks the peak brain growth, neurogenesis, synaptogenesis, and neural-immune maturation. At P2, we found that TNF-α gene expression in CHORIO brains were significantly higher (p = 0.008) when compared with that of sham controls. At all other timepoints, TNF-α expression was not statistically different (Fig. 1a). Similarly, CD86 gene expression in P2 CHORIO increased significantly by 8-fold (p = 0.01) compared with that of sham controls (Fig. 1b). At all other timepoints, CD86 expression was not significantly different (Fig. 1b). For Arg1, expression significantly decreased by 43% (p = 0.03) at E19 in CHORIO compared with sham brains (Fig. 1c). Arg 1, however, was similar at E18.5, P2, P7, and P21 (Fig 1c). Taken together, these results affirm the importance of the dysregulated anti- and pro-inflammatory signatures in CHORIO cortex, after a single intrauterine insult, in the prenatal and early postnatal life. Accordingly, we decided to study brain and the peripheral inflammatory-immune signatures at term-and toddler-equivalent ages in the next study.

CHORIO Alters Key Splenic Immune Cell Populations at P7

Because the spleen is a key secondary lymphoid organ that has central functions in the maintenance of a healthy immune system [48], we studied how the distribution of splenic immune cells are affected at term-equivalent age (P7) following CHORIO. We demonstrated that CHORIO significantly alters the percentages of key T-cell populations, including Th, immature T cells and cytotoxic T cells. Specifically, CHORIO increased the percentages of CD3⁺CD4⁺ Th and immature T cells while decreasing cytotoxic T cells. The percentage of gated CD45⁺ leukocytes was not different between sham (54.9 ± 3.8%) and CHORIO (44.8 ± 3.9%) spleens (Fig. 2a). We detected a significant increase in the percentage of CD3⁺CD4⁺ Th cells in CHORIO (3.6 ± 0.4%, p < 0.005) compared with sham (2.1 ± 0.1%) spleens (Fig. 2b). In CHORIO spleens, a significant concomitant increase in the percentage of CD3⁻CD4⁺ immature T cells (23.3 ± 1.0%, p = 0.001) was also found compared with sham spleens (10.6 ± 2.3%, Fig. 2c). The total percentage of CD3⁺CD4⁺CD25⁺FoxP3⁺ Tregs (sham: 7.2 ± 3.3% vs CHORIO: 15.6 ± 0.8%) (Fig. 2d) and median fluorescent intensity (MFI) of gated CD3⁺CD4⁺CD25⁺FoxP3⁺ Tregs (18880 ± 1175 vs. 18800 ± 923.1) were similar at P7 when compared with Sham (21957 ± 288) at P7 (Fig. 2e). In contrast, the percentage of CD3⁺CD8⁺ cytotoxic T cells were decreased in CHORIO spleens (0.9 ± 0.2, p = 0.02) compared to sham (2.1 ± 0.4) Fig. 2f) and CD3⁺CD4⁻CD8⁻ double-negative T cells (Fig. 2g) were not different. CD3⁻CD4⁻ cells significantly decreased by 17.2% in CHORIO (68.6 ± 1.0%, p = 0.0001) compared with sham (82.9 ± 1.9%) spleens (Fig. 2h). Our FC data showed altered percentages of inflammatory T cells, which suggests immune reprogramming following a single intrauterine CHORIO insult. Therefore, we asked whether the cytokine and chemokine profiles of these cells between CHORIO and sham spleens would be different.

Fig. 2.

CHORIO-induced phenotypic alterations at P7 in splenic T-cell populations. a. Leukocytes (CD45⁺); b. Th cells (CD3⁺CD4⁺); c. Immature T-cells (CD3⁻CD4⁺); d. Tregs (CD25⁺FoxP3⁺); and. e. FoxP3 Tregs (CD25⁺FoxP3⁺) MFI; f. CD3⁺CD8⁺ cells; g. CD3⁺CD4⁻CD8⁻ cells; and h. CD3⁻CD4⁻ cells. Data are presented as mean ± SEM % live cells or MFI, Welch’s t-test. Sham n = 7 and CHORIO n = 6. *, p < 0.02; **, p < 0.005; ***, p < 0.001; ****, p < 0.0001

Spleen Inflammatory Secretome at P7 and P21

We further interrogated immune responses at term equivalent age by studying cytokine and chemokines in spleen, peripheral blood and cortices at P7 and at toddler equivalence at P21. At P7, CXCL-1 was significantly increased in CHORIO spleens (34.1 ± 3.16 pg/100 mcg protein, p = 0.02) compared to sham controls (24.49 ± 1.64 pg/100 mcg protein) and TNF-α was higher in CHORIO spleens (2.34 ± 0.18 pg/100 mcg protein, p = 0.0047) compared to sham controls (1.57 ± 0.13 pg/100 mcg protein) Table 2.

Table 2:

Cytokine and Chemokine (C/C) Profiles

C/C	Spleen		Cortex		Serum		p value
	Sham	CHORIO	Sham	CHORIO	Sham	CHORIO
IFN-γ
P7	0.59 ± 0.19	0.23 ± 0.08s	3.84 ± 0.16	4.3 ± 0.43	11.54 ± 1.69	13.76 ± 1.49	ns *0.01
P21	0 ± 0	0 ± 0	3.23 ± 0.77	5.46 ± 2.1	5.3 ± 1.3	10.3 ± 1.7*
IL-1β
P7	75.27 ± 3.29	83.56 ± 6.35	19.05 ± 0.49	21.78 ± 1.77	1.50 ± 0.36	3.26 ± 0.93	ns **0.04
P21	117.7 ± 20.39	105.4 ± 8.71	14.24 ± 2.94	14.81 ± 6.08	5.2 ± 2.3	25.1 ± 7.7
IL-4
P7	0.05 ± 0.02	0.07 ± 0.02	0.3 ± 0.05	0.4 ± 0.1	2.93 ± 0.34	3.34 ± 0.28	ns ns
P21	0.06 ± 0.03	0.14 ± 0.05	0.04 ± 0.04	0.10 ± 0.06	2.14 ± 0.35	2.15 ± 0.43
IL-5
P7	3.22 ± 1.06	3.18 ± 1.94	8.51 ± 0.94	9.54 ± 1.85	0 ± 0	6.23 ± 1.76**	***0.021 ns
P21	0.13 ± 0.13	0 ± 0	0 ± 0	1.34 ± 0.93	1 ± 1	0 ± 0
IL-6
P7	2.63 ± 0.38	3.16 ± 0.44	6.36 ± 1.06	8.38 ± 2.24	36.68 ± 5.52	38.78 ± 6.15	ns Ψ0.048 ω0.0006
P21	5.64 ± 2.78	2.31 ± 1.66	10.04 ± 3.83	23.57 ± 4.76 Ψ	76.56 ± 21.3	321.7 ± 76.8ω
CXCL-1
P7	24.49 ± 1.64	34.1 ± 3.16ε	3.46 ± 0.54	2.99 ± 0.66	213.8 ± 29.86	281.3 ± 29.46	ε0.0247 ns
P21	64.91 ± 7.38	54.54 ± 5.32	10.73 ± 1.53	9.77 ± 0.75	77.50 ± 12.71	104.5 ± 41.76
IL-10
P7	1.76 ± 0.37	1.84 ± 0.19	5.21 ± 1.01	6.62 ± 1.83	70.07 ± 6.9	78.05 ± 6.36	ns Λ0.0344 κ0.02
P21	0.19 ± 0.13	0.34 ± 0.34	0.94 ± 0.65	3.01 ± 0.92Λ	41.5 ± 6.2	60.2 ± 7.4κ
IL-13
P7	0.09 ± 0.08	0.17 ± 0.08	1.23 ± 0.29	1.60 ± 0.58	4.72 ± 0.72	4.73 ± 0.9	ns ns
P21	0 ± 0	0 ± 0	0 ± 0	0.29 ± 0.20	8.99 ± 1.22	8.63 ± 1.66
TNF-α
P7	1.57 ± 0.13	2.34 ± 0.18β	0.59 ± 0.09	0.75 ± 0.19	7.26 ± 0.39	7.26 ± 0.43	β0.0047 ns
P21	4.79 ± 0.63	4.99 ± 0.56	0.24 ± 0.04	0.31 ± 0.098s	2.44 ± 0.28	2.33 ± 0.20

Data presented as Mean ± SEM.

At P21, CHORIO did not alter splenic cytokine and chemokine responses (expressed as pg/100 μg protein) compared with sham controls for all studied targets including: IL-6 (2.31 ± 1.65 vs 5.64 ± 2.77); TNF-α (4.98 ± 0.56 vs 4.78 ± 0.62); INF-γ (0 ± 0 vs 0 ± 0); IL-10 (0.34 ± 0.34 vs 0.18 ± 0.13); IL-1β (105.4 ± 8.7 vs 117.7 ± 20.4); CXCL1 (54.5 ± 5.32 vs 64.9 ± 7.38); IL-4 (0.14 ± 0.05 vs 0.05 ± 0.02); and IL-5 (0 ± 0 vs 0.12 ± 0.12) pg/100 mcg protein (Table 2). CHORIO is a major contributor to perinatal brain injury. Hence, after establishing the dysregulated pro- and anti-inflammatory signatures in the CHORIO cortex and identifying T-cell immune alterations in the key peripheral lymphoid organ spleen, we investigated peripheral blood and CNS inflammatory-immune changes in CHORIO in toddler-equivalent age (P21).

CNS and Systemic Peripheral Blood Inflammatory Secretome at P7 and P21

Next, we sought to identify whether the CNS and systemic chemokine and cytokine patterns were altered at term- (P7) and toddler-equivalent age (P21). No differences in brain cortical cytokine and chemokine levels were identified at P7 (Table 2). We detected a significant increase in IL-6 (23.6 ± 4.8 pg/100 μg protein, p < 0.05) in cortices of CHORIO pups when compared with those of sham controls (10.0 ± 3.8 pg/100 μg protein, Fig. 3a) at P21. Additionally, there was a significant increase in IL-10 (3.0 ± 0.9 pg/100 μg protein, p = 0.03) in cortices of CHORIO pups compared with those of sham controls (0.9 ± 0.7 pg/100 μg protein, Fig. 3b) at P21. Other cytokines and chemokines were not significantly different in CHORIO compared with sham control cortex at P21 (IFN-γ and IL-1β are shown in Figs. 3c and 3d, respectively).

Fig. 3.

Pro-inflammatory cytokine response at P21 in CHORIO pup brains (n = 7) compared with sham controls (n = 7). a. IL-6; b. IL-10; c. IFN-γ; and d. IL-1β. Data are presented as mean ± SEM pg/100 μg protein, Welch’s t-test for normally distributed data and Mann Whitney test for non-normal distributed data. *, p < 0.05

Notably, the above changes in brain cytokines paralleled the increases in serum cytokines at P21. Of note, at P7, IL-5 in CHORIO serum was significantly higher than sham controls (6.23 ± 1.758 vs. 0 ± 0 pg/mL, p = 0.02, Figs 4a), while similar at P21 (Figs 4b). Specifically, IL-6 in CHORIO serum were higher than that of sham controls (321.7 ± 76.8 and 76.6 ± 21.3 pg/mL, respectively, p = 0.0006, Fig. 4c); increased IL-10 in CHORIO serum compared with sham (60.2 ± 7.4 and 41.5 ± 6.2 pg/mL, respectively, p = 0.02, Fig. 4d); increased IFN-γ in CHORIO serum compared with sham ((10.3 ± 1.7 and 5.3 ± 1.3 pg/mL, respectively, p = 0.04, Fig. 4e); and increased IL-1β in CHORIO serum (compared with sham (25.1 ± 7.7 and 5.2 ± 2.2 pg/mL, respectively, p = 0.03, Fig. 4f). Thus, we showed that the acute systemic inflammatory response in CHORIO offspring at P7 is followed by a sustained inflammatory phenotype in serum and CNS at P21 when compared to sham controls.

Fig. 4.

Serum cytokines. P7: a. IL-5. P21: b. IL5; c. IL-6; d. IL-10; e. IFN-γ; and f. IL-1β. Data are presented as mean ± SEM pg/ml, Welch’s t-test for normally distributed data and Mann Whitney test for non-normal distributed data. Sham n = 6–7 and CHORIO n = 4–7. *, p < 0.05; ***, p < 0.0007

CHORIO Increases CD3⁺CD4⁺ Th in the Brain at P21

Given that we detected significant pro-inflammatory cytokine signatures in the CNS at toddler-equivalent age, we next determined whether immune T-cell distributions were altered at P21. In support of emerging sustained inflammatory/immune alterations, CHORIO increased CD3⁺CD4⁺ Th cells in the brains when compared with sham controls. The percentages of CD45⁺ leukocytes (Fig. 5a) and CD3⁺ T cells (Fig. 5b) were similar. Specifically, the percentage of CD3⁺ CD4⁺ Th cells significantly increased 2.5-fold in CHORIO (0.14 ± 0.03) over sham control (0.06 ± 0.03, p < 0.02, Fig. 5c) brains. However, the percentage of CD3⁺CD8⁺ cytotoxic T cells (Fig. 5d) in were similar in both groups. We previously showed that in CHORIO brains, Th cells are increased at P60. Our current data point to an earlier developmental time course for brain T-cell dysregulation in CHORIO.

Fig. 5.

CHORIO-induced phenotypic alterations in cortical T-cell populations at P21. a. Leukocytes (CD45⁺); b. T lymphocytes (CD3⁺); c. Th cells (CD3⁺CD4⁺); and d. cytotoxic T cells (CD3⁺CD8⁺). Data are presented as mean ± SEM % of gated cells, Welch’s t-test for normally distributed data and Mann Whitney test for non-normal distributed data. Sham n = 4 and CHORIO n = 6. *, p = 0.02

Discussion

Perinatal infection, inflammation, and/or CHORIO can disturb homeostasis during pregnancy [49]. Preterm neonates are at a disproportionately higher risk for inflammation/infection-related adverse events including perinatal brain injury (PBI) [50]. The effects of early alterations in inflammatory and immune pathways are multifold and can result in lifelong sequalae in many organs, manifesting as neurological and developmental deficits [50–52]. Cortical brain injury following inflammatory insult is heterogenous involving neurons, oligodendrocytes, microglia, and astrocytes [13–18]. Both gray and white matter is affected, including cortical myelin [13–15, 17]. Previously, we have shown that there is an increase in brain mononuclear cell populations compared with sham controls 2 weeks after chorioamnionitis induced in rats and a dysregulation of HO-1/TfR1 in the cortex of animals with CHORIO [34]. In our current study, we observed an acute decrease in anti-inflammatory genes in the cortex at E19 followed by an increase in pro-inflammatory genes at P2. Importantly, the timing of these anti- and pro-inflammatory changes coincide with our previously observed dysregulation of HO-1/TfR1 at these same critical developmental timepoints [34]. Furthermore, we have found that the pro-inflammatory changes in CHORIO was not limited to innate immune cells in the CNS, but also included the adaptive immune system, the spleen. Specifically, in CHORIO spleens at P7, Th and immature T cells were increased, while cytotoxic T cells were decreased. The prominence of Th cells at P21 suggests an emerging chronic neuroinflammation given that the Th predominance persists at P60 in this model[25]. Additionally, the micro-environment in brain cortex at P21 is pro-inflammatory coincident with significant increases in IL-6. IL-6 is well-known to orchestrate the transition from acute to chronic inflammation by altering inflammatory milieu, recruitment of mononuclear cells and altering T cell immune responses [53].

It is well known that TNF-α is developmentally regulated and it plays important roles in immunomodulation and organogenesis during embryonic life [54]. Additionally, TNF-α signaling is important in CNS development [54]. For instance, early in mouse embryo at E12.5, TNF-α promotes the survival, proliferation, and differentiation of neural progenitors [54, 55]. However, cellular responses to TNF-α switch from proliferation and differentiation to neuronal apoptosis later in gestation [54]. A known role of TNF-α is indeed in inflammation and inflammatory diseases, including pathogenesis of acute and chronic neurodegenerative disorders [54, 56, 57]. TNF-α signaling is complex which is driven partly by TNF-α’s relative affinity to bind to its receptors TNFR1 and TNFR2; In an acute insult TNF-α has higher affinity to bind to TNFR1 and favor pro-inflammatory responses, while relative abundance of and TNFR1/TNFR2 across different brain regions such as in cortex, could be variable and differ during acute and chronic phases of injury [58, 59]. Therefore, studying baseline TNF-α expression at critical perinatal and postnatal windows, in the absence of pathological or injury-induced inflammation, is crucial. Furthermore, brain cortical and amniotic fluid TNF-α and IL-6 levels can also increase after a prenatal transient hypoxic-ischemic insult alone, in the absence of amniotic LPS administration.[60] The complexity of inflammatory responses following fetal exposure to transient hypoxia and systemic infections require further investigation of cytokine-chemokine-immune interactions.

After determining the potential ontogeny of key pro- and anti-inflammatory/immune markers in the cortices in this critical window, we asked whether they would be altered by CHORIO. CD86 is an important co-stimulatory molecule that is present in antigen presenting cells, mononuclear cells, as well as microglia in the brain. Arg1 is a traditional anti-inflammatory mononuclear marker [61]. Mononuclear cell populations including microglia are dynamic [62, 63]. For instance, the transcriptomic signatures for microglia can differ in health and diseased states, through the lifespan and based on CNS micro-environment [62, 63]. We observed high CD86 expression in CHORIO cortices at P2 signifying the presence of an increased pro-inflammatory signature at P2. We also show an acute decrease in the anti-inflammatory marker Arg1 at E19 prior to the increase in CD86 at P2 and a prominent increase in IL-6 in the cortices at P21, which collectively suggest a pro-inflammatory micro-environment [61]. Our findings of increased CD86 at P2 are in alignment with our previous FC studies using this animal model and detecting increased mononuclear cells at P2 and P7 [24, 34]. Indeed, the trend in increased mononuclear cells persists until P120 [25]. These results collectively denote the importance of a loss of homeostasis in inflammatory-immune-induced perinatal brain injury. Microglia developmental window is tightly regulated at the transcriptional level [63]. Insults leading to maternal immune activation prematurely change microglia transcription to a profile representative of a later developmental stage and alter homeostasis [63–65]. Interestingly, in a 2015 study, CD86 expressed on microglia was found to be associated with a microglial phenotype that promotes neurorepair [66]. Upregulation of CD86 in rat microglia resulted in a phenotype switch from compact to a ramified type and led to axonal growth, thereby providing neuromodulation [66]. CD86 was absent in astrocytes in this study [66]. In contrast, in a mouse stroke model that carries a genetic deficiency in inflammatory stimuli induce immunosuppressive gene 1 (IRG1), showed increased CD86⁺ microglial inflammation and immune cell infiltration [67]. Notably, IRG1 was found to repress HO-1 expression, however upregulation of microglial HO-1 attenuated ischemic brain injury [67]. Hence, based on our current findings it is imperative to further our studies to the specific mononuclear cell populations and microglia in CHORIO brains to determine how they contribute to perinatal brain injury and neuroinflammation.

The spleen is an organ involved in local and systemic immune regulation and immunotolerance [48]. We observed disruptions in splenic immune signatures after CHORIO in the immediate postnatal period at term-equivalent age. Our data revealed increases in CD3⁺CD4⁺ Th and CD3⁻CD4⁺ immature T cells and decreases in cytotoxic T cells. Interestingly, Th cells migrate towards and colocalize with CD3⁻CD4⁺ cells within the spleen [68]. CD3⁻CD4⁺ immature or innate-like T cells are also known for their high constitutive TNF-α secretory capacity and their ability to alter tissue micro-environments in both humans and embryonic (E15), neonatal (P2) and adult cells of rodent spleens [69–72]. CD3⁻CD4⁺ immature or innate-like T cells were previously identified as originating from the fetal spleen [73]. Furthermore, in rodent secondary lymphoid organs, specifically the spleen, CD3⁻CD4⁺ T cells contribute to the structural organization of B and T cells and support memory T cells via survival signals [70–72]. Interestingly, despite lack of, or very low levels, of surface CD3 expression in CD3⁻CD4⁺ T cells, CD3 was indeed present intracellularly, committing these CD3⁻CD4⁺ cells to T cell lineage [69]. Furthermore, CD3⁻CD4⁺ immature or innate-like T cells have been shown to be increased in human chronic inflammatory diseases such as rheumatoid arthritis and psoriasis, which correlate with disease severity [69, 74]. Hence, our results suggest that CHORIO may reprogram splenic immune cells to a pro-inflammatory phenotype at P7. Splenic cytokine and chemokine responses in CHORIO and sham were similar at P21 and the splenic immune profile warrants further study.

We then asked whether this pro-inflammatory signature would be prominent in the cortices and peripheral blood of CHORIO pups at a later timepoint, at toddler equivalent age. Importantly, we demonstrated that pro-inflammatory CNS cytokine IL-6 response in P21 CHORIO pups exceeds the level of increase in anti-inflammatory CNS cytokine IL-10. Moreover, we showed presence of an inflammatory serum phenotype by a 4.2-fold increase in IL-6 and 4.8-fold increase in IL-1β, in CHORIO sera compared to Sham. These findings are important because, IL-6 can promote transition from an innate to adaptive immune response in CNS, promote increase in certain subsets of CD4⁺ Th cells, such as Th17, contribute to CD8⁺ cytotoxic T-cell differentiation while inhibiting Treg differentiation [75, 76]. Furthermore, in the CNS, IL-6 can be produced by a variety of cells, including neurons, astrocytes, microglia, and endothelial cells after an injury [76]. IL-6, in addition to supporting transition from an innate to adaptive immune response in the CNS, is a driver for chronic inflammation and has important roles in recovery and normal functions in the CNS [76]. The role of IL-6 in the switch from acute phase of inflammation to chronic phase of inflammation is a well-known factor [53]. In this role IL-6 supports mononuclear cells, macrophages, and T-cell responses [53]. Notably, our previous and current innate and adaptive immune results are in alignment with this existing literature [23, 25, 34]. Collectively, we show supporting inflammatory evidence in CHORIO brains at toddler equivalent age in humans.

Next, we investigated whether CHORIO would result in sustained T-cell immune changes in the brain, given serum IL-6 can promote T cells. We observed a Th-predominant cerebral inflammation at P21 in conjunction with a Th-predominant inflammation in the spleen at P7 and elevated serum IL-6 at P21. It has been reported that in an inflamed cerebral micro-environment, Th cells through their secretome can further activate resident microglia and resulting in pro-inflammatory phenotypes of infiltrating mononuclear cells and other T-cell subsets [77–80]. This links to the observed increase in mononuclear cells in CHORIO brains compared to Sham from early postnatal (P2-P7) through adulthood (P120) [25, 34]. Transient peripheral mononuclear cell infiltration following injury is well described in CNS [81]. This can lead to peripheral to central immune cell trafficking [81–83]. Taken together, this immune reprogramming may highlight the risk of emerging chronic neuroinflammation after exposure to CHORIO.

We acknowledge that there are limitations to our study. First, we were not powered to detect sex differences, although we did use both sexes in all experiments. Second, we studied the cortical gene expression from whole cortices and not from isolated microglia, or other neural or immune cell populations and could not comment on possible regional or subpopulation differences. We acknowledge the potential of interdependence of some inflammatory genes and markers and this necessitates future investigation. Third, western blot (WB) or immunohistochemistry (IHC) will be done in future experiments to establish the relationship between mRNA and protein expression and the immune cell population changes documented. Fourth, splenic immune cell phenotypes from P2 to P21 and beyond remain to be studied.

In conclusion, CHORIO yields a sustained inflammatory micro-environment defined by increases in cortical IL-6 in rats and the concurrent presence of Th-predominant T cells at P21, toddler-equivalent age in humans. These data support emerging Th-specific chronic inflammation. Notably, HO-1/TfR1 is known to regulate the balance of IL-6 and IL-10 that is crucial for T-cell differentiation. Further studies to elucidate the role of dysregulated HO-1/TfR1 homeostasis on sustained T-cell neuroinflammation are warranted. Defining how these inflammatory/immune signatures contribute overall to brain injury and investigating distinct immune signatures through developmental time course will be beneficial for targeting HO-1/TfR1 pathway-focused emerging therapies in the future.

Data Availability Statement

The data that support the findings of this study are not publicly available due to privacy reasons but are available from the corresponding author upon reasonable request.