Maternal plasma proteomics in a rat model of pregnancy complications reveals immune and pro-coagulant gene pathway activation

Abstract

Problem:

The Brown Norway (BN) rat is a model of T-helper 2 immune diseases, and also a model of pregnancy disorders that include placental insufficiency, fetal loss, and preeclampsia-like symptoms. The aim of this study was to investigate the plasma proteomic/cytokine profile of pregnant BN rats in comparison to that of the Lewis (LEW) rat strain.

Method of Study:

Plasma proteomics differences were studied at day 13 of pregnancy in pooled plasma samples by differential in-gel electrophoresis, and protein identification was performed by mass spectrometry. Key protein findings and predicted cytokine differences were validated by ELISA using plasma from rats at various pregnancy stages. Proteomics data were used for Ingenuity Pathway analysis.

Results:

In-gel analysis revealed 74 proteins with differential expression between BN and LEW pregnant dams. ELISA studies confirmed increased maternal plasma levels of complement 4, prothrombin, and C-reactive protein in BN compared to LEW pregnancies. LEW pregnancies showed higher maternal plasma levels of transthyretin and haptoglobin than BN pregnancies. Ingenuity Pathway Analysis revealed that BN pregnancies are characterized by activation of pro-coagulant, reactive oxygen species, and immune-mediated chronic inflammation pathways, and suggested increased interleukin 6 and decreased transforming growth factor-β1 as potential upstream events. Plasma cytokine analysis revealed that pregnant BN dams have a switch from anti- to pro-inflammatory cytokines with the opposite switch observed in pregnant LEW dams.

Conclusion:

BN rats show a maternal pro-inflammatory response to pregnancy that likely contributes to the reproductive outcomes observed in this rat strain.

1. INTRODUCTION

Animal models of pregnancy aid researchers in understanding the complex maternal-fetal interactions, as well as physiological adaptations that are unique to the pregnancy state.^1–3 We and others have studied the reproductive phenotypes of several rat strains including the Brown Norway (BN) rat.^4–7 BN pregnancies are characterized by small litters and placental insufficiency due to decreased trophoblast invasion of maternal decidua and altered uterine natural killer cell marker expression.^4–7 Our lab has further characterized the reproductive phenotype of the BN rat as having decreased uteroplacental blood flow (placental insufficiency) with intrauterine fetal growth restriction that is corrected near term. Placental insufficiency is associated with increased labyrinth and mesometrial triangle angiogenesis.^4,7 Decreased uteroplacental and maternal renal blood flows are associated with late-stage maternal hypertension and proteinuria, which are pre-eclampsia like symptoms.⁷ Placental microarray studies suggested an upregulation of tumor necrosis factor alpha (TNFα)-mediated inflammatory pathways and increased renin-angiotensin system gene pathways in BN pregnancies.⁴ Systemically, we found that pregnant BN rats become severely vitamin D deficient at mid-gestation due to maternal renal and placental metabolic gene dysregulation. Vitamin D status returns to pre-pregnancy levels at the end of the lactation period.⁵

In contrast, another inbred rat strain, the Lewis (LEW) rat, shows pregnancy-induced increases in vitamin D status similar to healthy human pregnancies, a 10 mm Hg dip in blood pressure by mid-gestation, and decreased placental inflammatory gene expression.^5,7 LEW rat pregnancies have similar outcomes to those of the outbred Sprague-Dawley rat.⁵ Interestingly, studies on adult LEW rats have shown that this rat strain is predisposed to T-helper-1 (Th-1) inflammatory diseases such as acute transplant rejection and experimental acute encephalomyelitis.⁸ In contrast, BN rats are predisposed towards T-helper-2 (Th-2) inflammatory diseases such as allergen-driven asthma.⁸ Therefore, BN and LEW rats are useful models to study the genetic basis of immune-associated diseases.^8,9

Healthy pregnancies are characterized by maternal tolerance of the fetal implant that includes immune cell regulation at the fetal-maternal interface, as well as, maternal systemic immune adaptations.^10–16 Therefore, the aim of this study is to characterize the main changes that occur in the maternal circulatory proteome in BN and LEW rat strains as models of complicated and healthy pregnancy, respectively. Because BN rats model pregnancy complications such as preeclampsia, we hypothesize that BN rats would show systemic inflammatory protein profiles during pregnancy.

2. MATERIALS AND METHODS

2.1. Animals

Two-month old BN and LEW male and female rats were obtained from Charles River Laboratories (Cambridge, MA) and housed at the Animal Research Facility, Loma Linda University, Loma Linda, CA under conditions of 14h light, 10h darkness, room temperature of 20°C, and relative humidity of 30–60%. Rats were fed standard rat chow and tap water ad libitum. We chose to study the LEW inbred rat strain on the basis of previous publications demonstrating ‘opposite’ immune profiles to the inbred BN rat (pro-Th1 vs. pro-Th2, respectively) and similar reproductive outcomes to the outbred Sprague-Dawley rat.^6,7 All experimental protocols were in compliance with the Animal Welfare Act, and were approved by the Institutional Animal Care and Use Committee of Loma Linda University.

2.2. Study design, breeding, and tissue collection

We used a cross-sectional study design to study age-matched female rats at non-pregnancy (NP), and at 4 different pregnancy stages (day 9, day 13, day 17, and day 21). We used 4 rats for each of the 5 stages for a total of 20 rats for each strain (BN and LEW). Female rats were housed in single cages. Rats were bred by overnight monogamous pairing of a female with a strain-matched male. The following morning, we removed the male and that day was considered pregnancy day 0. Rats were weighed daily, and pregnancy was confirmed by examining vaginal plugs on day 0 and weight gain by day 10. To collect heparinized plasma for proteomic analysis, the rats were anesthetized with 2% isoflurane and blood was withdrawn directly from the beating heart using lithium heparin tubes (2 mL) , followed by heart/lung block removal. Blood was centrifuged at 5000 rpm × 5 min at 4°C to collect plasma. Plasma was stored at −70°C until further analysis. Rat pregnancy parameters such as litter size, fetal weight and placental weight were also obtained.

2.2. Plasma sample preparation

Differential in-gel electrophoresis (DIGE) and mass spectroscopy protein identification studies were performed by Applied Biomics, Inc. (Hayward, CA). Pooled maternal plasma samples at gestational day 13 from BN and LEW rats (n=3 rats per strain, 100 μl plasma from each rat) were used. The plasma samples were thawed and vortexed for 20 sec. The samples were spun for 30 min at 4 °C at 14,000 rpm and the supernatant was collected. Protein concentration was measured using Bio-Rad protein assay method. For each sample, 30μg of protein was mixed with 1.0 μl of diluted CyDye, and kept in dark on ice for 30 min. BN plasma proteins were labeled with Cy3 and LEW plasma proteins with Cy5, respectively. The labeling reaction was stopped by adding 1.0 μl of 10 mM lysine to each sample, and incubating in the dark on ice for an additional 15 min. The labeled samples were then mixed together. The 2X 2-D sample buffer (8 M urea, 4% CHAPS, 20 mg/ml DTT, 2% pharmalytes and trace amount of bromophenol blue), 100 μl DeStreak^™ solution and Rehydration buffer (7 M urea, 2 M thiourea, 4% CHAPS, 20 mg/ml DTT, 1% pharmalytes and trace amount of bromophenol blue) were added to the labeling mix to make a total volume of 250 μl.

2.3. Differential in-gel electrophoresis (DIGE)

After loading the labeled samples, isoelectric focusing (IEF, pH3–10, non-linear) was run following the protocol provided by GE Healthcare. Upon finishing the IEF, the immobilized pH gradient (IPG) strips were incubated in the freshly made equilibration buffer-1 (50 mM Tris-HCl, pH 8.8, containing 6 M urea, 30% glycerol, 2% SDS, trace amount of bromophenol blue and 10 mg/ml DTT) for 15 min with gentle shaking. Then the strips were rinsed in the freshly made equilibration buffer-2 (50 mM Tris-HCl, pH 8.8, containing 6 M urea, 30% glycerol, 2% SDS, trace amount of bromophenol blue and 45 mg/ml iodoacetamide) for 10 min with gentle shaking. Next the IPG strips were rinsed in the SDS-gel running buffer before transferring into 13.5% SDS-gels. The SDS-gels were run at 15 °C until the dye front ran out of the gels. Gel images were scanned immediately following the SDS-PAGE using Typhoon TRIO (GE Healthcare). The scanned images were then analyzed by Image Quant software (version 6.0, GE Healthcare), followed by in-gel analysis using DeCyder software (version 5.0, GE Healthcare). The fold change of the protein expression levels was obtained from in-gel DeCyder analysis.

2.4. Protein identification by mass spectrometry

The spots of interest (74 in total) were selected based on the in-gel analysis and spot picking design by DeCyder software and picked up by Ettan Spot Picker (Amersham BioSciences). The gel spots were washed twice and then digested in-gel with modified porcine trypsin protease (Trypsin Gold, Promega). The digested tryptic peptides were desalted by Zip-tip C18 (Millipore). Peptides were eluted from the Zip-tip with 0.5 μl of matrix solution ( α-cyano-4-hydroxycinnamic acid (5 mg/ml in 50% acetonitrile, 0.1% trifluoroacetic acid, 25 mM ammonium bicarbonate) and spotted on the AB SCIEX MALDI plate (Opti-TOFTM 384 Well Insert).

MALDI-TOF MS and TOF/TOF tandem MS/MS were performed on an AB SCIEX TOF/TOF^™ 5800 System (AB SCIEX, Framingham, MA). MALDI-TOF mass spectra were acquired in reflectron positive ion mode, averaging 4000 laser shots per spectrum. TOF/TOF tandem MS fragmentation spectra were acquired for each sample, averaging 4000 laser shots per fragmentation spectrum on each of the 10 most abundant ions present in each sample (excluding trypsin autolytic peptides and other known background ions).

Database search: Both of the resulting peptide mass and the associated fragmentation spectra were submitted to GPS Explorer workstation equipped with MASCOT search engine (Matrix science Inc., Boston, MA) to search the database of National Center for Biotechnology Information non-redundant (NCBInr) or Swiss-Prot database. Searches were performed without constraining protein molecular weight or isoelectric point, with variable carbamidomethylation of cysteine and oxidation of methionine residues, and with one missed cleavage also allowed in the search parameters.

2.5. Ingenuity pathway analysis (IPA)

The proteomics dataset provided by Applied Biomics consisted of 53 proteins with a protein or ion confidence interval (C.I) of greater than 95% that were differentially expressed in BN and LEW dams, and were input into IPA 2018 (Qiagen, Redwood City, CA) with a flexible format. The reference set used was the Ingenuity Knowledge Base with both direct and indirect relationships to identify all possible networks and upstream regulators. The rat accession number identifier type used was Uniprot/Swiss-Prot and the analysis was run with a stringent filter for all molecules and relationships. Interaction networks were analyzed in all species (rat, mouse, human) to observe all possible pathways. Briefly, a total of 18 proteins that were clearly upregulated or downregulated in BN compared to LEW rats were input into IPA; alpha-1B-glycoprotein (A1BG) and Ig kappa chain C were not included in IPA because of the ambiguity of the results.

The significance of a predicted pathway activation or inhibition is determined by a Z-score that is positive or negative, respectively. For individual proteins in a pathway, the color red indicated the protein was upregulated and green if it was downregulated in BN rats compared to LEW rats. If several proteins and observed results conform to the expected outcome within a pathway, then the Z-score algorithm applied an orange color for a likely activated pathway (positive Z-score) or a blue color for an inhibited pathway (negative Z-score).

2.6. Quantification of plasma proteomics data using ELISA

Based on IPA gene networks obtained with the proteomic dataset, 6 genes from the proteomic dataset were selected for ELISA quantification using commercially available kits. Plasma samples from non-pregnant and pregnant rats at days 9, 13, 17 and 21 of pregnancy (n=4 per group) were used to study the gestational variations in these selected proteins. The following proteins were validated by commercially available ELISA kits: total immunoglobulin G (IGG, catalog #41-IGGRT-E01, Alpco), C-reactive protein (CRP, catalog #ab108827, Abcam), haptoglobin (HP, catalog #ab108857, Abcam), transthyretin (TTR, catalog KA2137, Abnova), complement 4 (C4, catalog #LS-F21855, Lifespan Biosciences), and prothrombin (F2, catalog #LS-F13467, Lifespan Biosciences) according to the respective manufacturer’s protocol and using their controls and standard curves. All samples were run in duplicate.

2.7. Analysis of maternal circulating cytokine levels

Based on IPA predicted upstream regulators of the proteomic dataset, we quantified maternal plasma levels of selected cytokines. We used a multiplexed magnetic bead-based immunoassay kit (Milliplex catalog #RECYTMAG-65K, EMD Millipore Corp. Billerica, MA01821) to detect interleukin (IL)-2, IL4, IL5, IL10, interferon gamma (IFNγ), and TNFα following the manufacturer’s instructions. The plasma samples were centrifuged at 10,000 rpm for 4 min prior to use and then diluted 1:2 in assay buffer. Briefly, 25 μl of the diluted plasma were mixed with a cocktail of magnetic polystyrene microspheres coated with the relevant capture antibodies in wells of a 96-well plate and incubated at RT for 2h with agitation. The plate was washed and 25 μl of biotinylated detection antibodies were added to each well and incubated at RT for 1h with agitation. The plate was washed and the reaction mixtures were incubated with 25 μl of streptavidin-PE conjugate at RT for 30 min with agitation. Following incubation, the plate was washed and then read on MAGPIX (Luminex Corp., Austin, TX). The data was analyzed with Milliplex Analyst version 5.1 using a cubic spline curve fitting method (EMD Millipore Corp., Billerica, MA). The following cytokines were studied using commercially available ELISAs: IFNγ (Catalog ERIFNG, ThermoFisher Scientific), IL6 (Catalog 438204, Biolegend), TGFβ1 (BMS623, ThermoFisher Scientific), and IL17a (catalog BSM635, ThermoFisher Scientific). All samples were run in duplicate and analyzed according to the manufacturer’s instructions.

2.8. Statistical Analysis

All data are presented as mean ± standard error of the mean. We used univariate analysis to determine the differences for each variable using SPSS 22.0 software (IBM Corporation, Armonk, NY). Equal variance was determined by Levene’s test. Non-parametric data was log-transformed and reanalyzed. Two-way ANOVA was used to determine the significance of each of the two independent factors (rat strain and pregnancy stage) as well as the interaction between the two factors on the measured variable. To determine specific differences between each of the 10 groups (5 gestational stages, two rat strains), we used one-way ANOVA using a composite-independent factor followed by lest significant differences (LSD) posthoc analysis. Statistical significance was determined as p<0.05.

3. RESULTS

3.1. Maternal Plasma Proteomic studies on BN and LEW rats

The main fetal outcomes of the BN and LEW pregnancies used in this study are shown in Figure 1. There is a significant decrease in the litter size between early and late gestating BN pregnancies while LEW litter sizes remain stable throughout gestation (Figure 1A). BN pregnancies are also characterized by significantly lower fetal weights (Figure 1B) compared to LEW pregnancies. To uncover main differences in maternal responses to pregnancy, we performed a proteomic study in maternal plasma from gestational day 13 (mid-gestation). This day was chosen as significant fetal loss and increases in blood pressure occurs at this stage in pregnant BN dams. Plasma proteomic signatures showed some differentially expressed plasma proteins between pregnant BN and LEW rats (Figure 2A–C). From the proteomics dataset provided by Applied Biomics, we chose 53 proteins with a protein score confidence interval (CI) of more than 95% representing protein isoforms from 18 genes (Table 1). Several acute phase proteins were increased in BN compared to LEW rats including C-reactive protein (CRP), complement 4 (C4), prothrombin (F2), and alpha-1-acid glycoprotein. Interestingly, several protein spots corresponding to IGG and the corresponding light chain partners (immunoglobulin lambda and kappa) were significantly upregulated in BN compared to LEW pregnant dams (Table 1). Multiple spots corresponded to the same protein, such as alpha-1B-glycoprotein (A1BG), but these same proteins showed different isoelectric focusing migration patterns, with LEW A1BG proteins showing a more acidic isoelectric point (pI) compared to BN A1BG proteins (Figure 2, Table 1). LEW rats showed higher levels of plasma creatine kinase M (KCRM), apolipoprotein E (APOE), HP, and of TTR than pregnant BN rats (Table 1).

Figure 1.

Main fetal outcomes of BN and LEW pregnancies. Litter sizes (A), and fetal weight (B) are shown at pregnancy days 9, 13, 17, and 21. Bars represent the average of 4 pregnancies ± error. *p<0.05 BN vs. LEW (at the same pregnancy stage); †p<0.05 pregnancy day 9 (P9) vs. mid-to-late-stage pregnancy (P13, P17, P21) levels.

Figure 2.

Differential in-gel electrophoresis (DIGE) analysis of BN and LEW maternal plasma. Maternal plasma samples from BN and LEW rats were collected at pregnancy day 13 (mid-gestation) and pooled before DIGE analysis (n=3/rat strains). BN plasma proteins were labeled with Cy3 (green) and LEW plasma proteins were labeled with Cy5 (red), and 30 μg total protein per each rat strain were mixed and studied by DIGE as described in methods. DIGE images for A) BN, B) LEW, and C) merged BN and LEW plasma proteins are shown. Protein spots that yielded a protein score confidence of 95% or more are circled and numbered.

Table 1.

Protein spot identification by mass spectrometry

Spot #	Protein ID	Accession	BN/LEW Ratio†	Protein Score	Protein CI %
4, 5, 6	Alpha-1B-glycoprotein	A1BG	−8.7 to −16	>400	100
8, 9, 10	Alpha-1B-glycoprotein	A1BG	2.6 to 6.5	>400	100
2, 65	Prothrombin	THRB	3.7, 3.4	>200	100
3	Ig gamma-2A chain C region	IGG2A	4.5, 3.5	>200	100
11, 12	Ig gamma-2B chain C region	IGG2B	4.6, 9.8	160	100
15–19, 26–29	Haptoglobin	HP	−1.9 to −16	>200	100
23	Creatine kinase M-type	KCRM	−2.5	58	99
20–22	Alpha-1-acid glycoprotein	A1AG	1.9 to 3.0	>70	100
37	Apolipoprotein E	APOE	−1.9	587	100
36	Complement C4	C4	2.1	98	100
38–40, 42–44, 46, 47	Ig lambda-2 chain C region	LAC2	2.5 to 6.4	>100	100
41	C-reactive protein	CRP	3.2	148	100
45, 49–54, 56–59, 61–62	Ig kappa chain C region, B allele	KACB	2.1 to 7.8	>100	100
60	Ig kappa chain C region, B allele	KACB	−3.3	120	100
64	Galectin-5	LEG5	−8.2	145	100
68	Transthyretin	TTHY	−4.8	88	100
71	Hemoglobin subunit beta-1	HBB1	−2.1	259	100
72	Beta-2-microglobulin	B2MG	−1.7	82	100

^†A negative BN/LEW ratio of protein levels indicates LEW/BN values.

3.2. Ingenuity Pathway Analysis of maternal plasma proteomic data

IPA showed activation of pro-coagulant and pro-inflammatory gene networks in the maternal circulation of BN compared to LEW at midgestation (Figure 3). At day 13 of pregnancy, the release of F2, C4 and CRP proteins is associated with increased adhesion of blood cells and blood coagulation gene pathways in BN rats (Figure 3A). The increased maternal levels of CRP and F2 and the decreased levels of APOE, HP and Hemoglobin subunit beta-1 (HBB) in BN compared to LEW rats are associated with gene pathways involved in synthesis and production of reactive oxygen species (ROS) (Figure 3B). Also, the release of F2, C4, CRP, and IgG proteins is linked to the activation of immune cells especially leukocytes (Figure 3C). Finally, the increased plasma levels of these proteins along with CKM and ORM1, and the reduced release of APOE and beta-2-microglobulin (B2M) in pregnant BN rats, is associated with chronic inflammatory response towards pregnancy (Figure 3D). As upstream regulators of CRP and ORM1, the pro-inflammatory cytokines IL1 (not shown) and IL6 (Figure 3E) were predicted to be upregulated in the maternal circulation of pregnant BN compared to pregnant LEW rats. Furthermore, IL6 increases in pregnant BN dams would correlate with IL6-mediated downregulation of TTR and APOE (Figure 3E). Contrary to IL1 and IL6, transforming growth factor beta 1 (TGFβ1) is predicted to be downregulated in BN compared to LEW rats (Figure 3F), due to the known TGFβ1-mediated downregulation of F2, CRP, and CKM and upregulation of APOE.

Figure 3.

Maternal plasma proteomics Ingenuity Pathway Analysis. IPA was performed using the proteomics assay’s findings (Table 1). Shown are the main protein differences (BN compared to LEW, BN/LEW) and gene networks for coagulation (A), reactive oxygen species (B), immune cell-mediated (C), and chronic inflammation (D). Predictive upstream regulators include IL6 (E), and TGFβ1 (F).

3.3. Analysis of IPA key findings

Key proteins from the proteomic dataset involved in IPA gene networks were then quantified by ELISA in all rat groups (n=20 rats/strain) (Figure 4). The effect of the independent variables, pregnancy and rat strain, was determined by 2-way ANOVA. We found that both pregnancy and rat strain had a significant effect on F2, C4, CRP, total IGG, HP and TTR (Figure 4A–F). However, the interaction of pregnancy with rat strain (shown in the inserted text boxes as Inter.) was only significant for F2 (Figure 4A), indicating that the regulation of these proteins throughout pregnancy had a similar pattern in both BN and LEW rat strains. To determine if the differences between BN and LEW rat plasma protein were significant at pregnancy day 13, we performed a one-way ANOVA analysis using a composite (pregnancy*strain) independent variable and posthoc LSD analysis. We found that the levels of F2, C4, HP and TTR were significantly different between the rat strains at pregnancy day 13 validating the proteomic assay (Figs. 4A, B, E, F and Table 1). CRP and total IGG were not significantly different at pregnancy day 13 between rat strains, but were significantly higher in BN compared to LEW at pregnancy day 17 (Figs 4C, D and Table 1).

Figure 4.

Plasma proteomics assay validation. The changes of key proteins identified in the DIGE/MS and IPA analysis were studied in age-matched female rats starting at non-pregnancy (NP) and throughout pregnancy days 9, 13, 17, and 21 (n=4/ rat strain per gestational age; total 20/rat strain). A) Prothrombin (F2), B) complement 4 (C4), C) total immunoglobulin G (IgG), D) C-reactive protein (CRP), E) Haptoglobin (HP), and F) transthyretrin (TTR) protein levels were determined by ELISA. Graph points represent the average ± error per group (n=4). The 2-way ANOVA analysis shown in the text boxes indicates the significance for pregnancy (Preg), rat strain (strain) and the interaction of Preg × Strain (Inter). The 1-way ANOVA analysis using a composite-variable determined the significance between two specific data points. *p<0.05 BN vs. LEW (at the same pregnancy stage); †p<0.05 pregnancy (P) vs. pre-pregnancy (NP) levels.

To confirm the upstream cytokine differences predicted by IPA studies, the maternal plasma levels of IL6, TGFβ1, and other key cytokines were studied by multiplex or individual ELISAs (Figure 5). Prior to pregnancy, BN rats have lower basal levels of IL6, TNFα, IFNγ, and higher basal levels of TGFβ1 and IL10 than LEW rats (Figure 5A, B, C, D, and F). There were no differences between BN and LEW maternal plasma levels of IL6 at mid-gestation (P13, Figure 5A). However, the effect of pregnancy on IL6 is opposite in these two rat strains: IL6 levels increase 5-fold in BN rats with a reverse pattern shown in LEW rats (Figure 5A). 2-Way ANOVA showed a significant interaction between our two independent variables, pregnancy and rat strain (Figure 5A insert box). Similarly, the maternal plasma levels of TGFβ1 show an opposite pattern of regulation with pregnancy according to rat strain: TGFβ1 levels decrease throughout pregnancy in BN rats, but increase in LEW rats (Figure 5B). IFNγ shows a similar interaction between pregnancy and rat strain as IL6 (Figure 5A and D). There were significantly higher levels of TNFα at pregnancy day 9, and of IL2 at pregnancy day 13 in pregnant BN rats compared to LEW rats (Figures 5C and E). BN had significantly higher levels of IL10 at non-pregnancy and at pregnancy day 21 compared to LEW rats; and BN rats show a decrease in early (P9) to late (P17) gestation with a return to basal levels at P21 that was not observed in LEW dams (Figure 5F). There were no significant differences between rat strains or pregnancy stages in IL4 and IL5 levels (data not shown) and IL17A was not detectable in any sample.

Figure 5.

Plasma cytokine profile in BN and LEW dams. The plasma levels of IL6 (A), total TGFβ1 (B), TNFɑ (C), IFNγ (D), IL2 (E), and IL10 (F), were analyzed at non-pregnancy (NP) and through pregnancy for each rat strain by multiplex or individual ELISAs. Each point represents the average ± error (n=4). The 2-way ANOVA analysis shown in the text boxes indicates the significance for pregnancy (Preg), rat strain (Strain) and the interaction of Preg × Strain (Inter). The 1-way ANOVA analysis using a composite-variable determined the significance between two specific data points. *p<0.05 BN vs. LEW (at the same pregnancy stage); †p<0.05 pregnancy (P) vs. pre-pregnancy (NP) levels.

4. DISCUSSION

We have shown for the first time a novel circulating anti- to pro-inflammatory switch in a rat model of spontaneous placental insufficiency and preeclampsia-like symptoms. In contrast, we observed a significant decrease in the pro-inflammatory cytokine profile in LEW dams throughout pregnancy that is associated with better reproductive outcomes. The increased pro-inflammatory cytokine profile (i.e. increased IL6 and IFNγ, and decreased IL10) in pregnant BN dams correlates with increases in circulating levels of acute phase proteins such as CRP, F2, and C4 which are known to increase during chronic ‘cold’ (i.e. non-infectious) inflammatory diseases.^17–19 HP is an anti-oxidant protein that binds free hemoglobin and is therefore increased during inflammation, or decreased in the presence of hemolysis. Therefore, lower plasma HP in pregnant BN rats could correlate with increased reactive oxygen species or increased hemolysis. Of note, human studies have shown that plasma levels of HP are regulated differently during inflammation in lean versus obese subjects.^20–22 This is of great importance because obesity can induce additional physiological and gene dysregulation.^22,24 Other studies have observed a significant interaction of HP genotype with recurrent miscarriage and preeclampsia.^24,25 The plasticity of HP could explain the different results in various human proteomic studies where preeclampsia or recurrent miscarriage has been associated with either increased or decreased HP and HBB.^21,26–28 Finally, BN rats had lower levels of TTR, a negative acute phase protein whose function is to transport thyroid hormones (T4) and retinol binding protein. Some human studies have also observed decreased TTR or RBP4 (which is transported by TTR) in plasma from preeclamptic women.²⁹ Of note, many of the differentially expressed plasma proteins found in this study are produced by the liver under regulation of pregnancy specific hormones, placenta-secreted factors, in addition to immune cytokines. For these reasons, future research should include rat strain-specific hepatic adaptations to pregnancy and identification of differentially secreted placental factors.

Previous studies have shown that BN and LEW rats have genetically-determined immune phenotypes with non-pregnant BN rats showing propensity to Th2-mediated diseases, and non-pregnant LEW rats showing propensity to Th1-mediated diseases.^8,9 In this study, we confirmed that non-pregnant female BN rats have significantly higher basal levels of pro-Th2 cytokines (IL10, IL5 and TGFβ1), and significantly lower pro-Th1 cytokines (IL6, IFNγ, TNFα) than non-pregnant LEW rats, which is in accordance with studies from other labs.^8,9 Pregnancy significantly changes the systemic cytokine profiles in both BN and LEW rats. There is a significant increase in IL6, IFNγ, and TNFα, and a decrease in TGFβ1 and IL10 by mid-to-late gestation. In contrast, LEW rats show a significant decrease of pro-inflammatory cytokines IL2 and TNFα by midgestation and IL6 and IFNγ by late-gestation. Our data suggest that LEW rats show a trend towards a cytokine switch that is characteristic of healthy pregnancy in mammals.^10–16 The regulation of the systemic maternal immune system during pregnancy has been extensively reviewed.^10–16 It is well shown that maternal immune maladaptation is a key observation in multiple pregnancy disorders including recurrent miscarriage, preterm birth, intrauterine growth restriction, and preeclampsia.^10–16 A pro-inflammatory cytokine profile (in particular with high levels of IFNγ) has been consistently shown in women with recurrent miscarriage.^31–33 Of interest, LEW rats that show a pro-inflammatory cytokine profile at the start of pregnancy can lose their entire litter before day 10 of pregnancy (data not shown), an effect that could be related to their naturally higher pre-pregnancy pro-Th1 status. It is well known that pregnancy ameliorates the symptoms of women with pro-Th1 disease such as rheumatoid arthritis and multiple sclerosis.^34,35 Therefore, LEW rats could model women with a predisposition to Th1 disease. Future flow cytometry studies are needed to confirm a classic Th1 to Th2 switch in pregnant LEW rats.

In contrast to LEW pregnancies, this study found that BN pregnancies continue to present fetal loss, intrauterine growth restriction, and markers of ‘cold’ inflammation’ throughout the entire pregnancy. Increases in maternal IL6 and TNFɑ levels have been shown in women with preeclampsia, pro-inflammatory markers that are also upregulated in BN dams. Women with Th2-mediated inflammatory diseases such as systemic lupus erythematosus have higher frequency of preeclampsia and recurrent miscarriage in association with autoantibody production.^35,36 It is interesting to note that the current maternal proteomic study revealed elevated production of IGG in BN dams. Future assays will reveal whether BN dams produce antibodies that impair fetal growth or placental development. Maternal autoantibodies are important markers in the development of both recurrent miscarriage and preeclampsia.^36,37 Finally, BN rats are predisposed to develop allergen induced asthma.^8,9 This is of great interest because some studies show that asthmatic women have higher incidence of preeclampsia, and preeclampsia-born offspring have higher incidence of developing childhood asthma.³⁸ Altogether, this study suggests that the BN rat model can be useful in identifying the interactions between pregnancy-specific hormones and regulation of the maternal and fetal immune system.

This study has some important limitations. First, the proteomics methods used was not sensitive enough to detect proteins at sub-microgram level, thereby limiting the discovery of key proteins involved in the reproductive phenotypes of these rat strains. Another important limitation of the current study is that only maternal plasma was studied, therefore we cannot confirm pregnancy induced changes in major immune cells such as Th, Treg, and NK cells. Multiple reports demonstrate major changes of the maternal immune system that include an increase in tolerant T-regulatory cells which are recruited to the utero-placental milieu, an increase in pro-Th2 markers in CD4+ T-cells, and decreased expression of cytotoxic markers in natural killer cells (NK).^10–16 At the utero-placental site, multiple interactions between trophoblasts and maternal immune cells take place and are currently a topic of intense research.^{10–16,39,40} It has been shown that BN mesometrial triangles show decreased trophoblast invasion and uterine artery remodeling, compared to other rat strains including Sprague-Dawley and LEW rats.^4,6,7 These effects could be due in part to differences in the regulation of uterine NK cells (uNK), with BN uNK showing higher expression of cytotoxic markers such as granzyme B and perforin, and decreased tolerant markers such as the CD56 marker compared to uNK of other rat strains (4–5). In contrast, LEW and Sprague-Dawley show up to five fold higher uterine artery remodeling, trophoblast invasion of mesometrial triangles, and intense presence of CD56+ uNK cells compared to BN rats.^4,6 Therefore, future studies are needed to fully understand the molecular mechanisms that regulate the maternal-fetal immune interactions in BN and LEW pregnancies.

Finally, it is important to stress that the mechanisms that regulate the uteroplacental and maternal immune system are not fully understood. Hormonal changes, including increases in progesterone and estradiol, are key modulators of the immune system during pregnancy.^10–16 A previous study revealed progesterone receptor resistance in early BN gestation,⁴¹ although the mechanism leading to this observation remains unknown. Regarding differences in hormonal levels, we have previously shown that there are drastic differences in vitamin D homeostasis during pregnancy in BN dams compared to LEW and Sprague-Dawley dams.⁷ BN dams show a drastic five-fold drop in the levels of bioactive vitamin D, and supplementation with calcitriol significantly improved fetal/placental weights and fetal viability as shown by litter sizes at birth.⁷ This is of great interest, as maternal vitamin D deficiency has also been associated with preeclampsia and fetal-origins of postnatal asthma.^42,43 Altogether, we speculate that the BN rat could be a useful model to investigate the relationships between vitamin D deficiency during pregnancy, the increase in maternal pro-inflammatory cytokines, and the development of pregnancy disorders such as hypertension, proteinuria and placental insufficiency. In summary, BN rats are a useful model to investigate the cellular and molecular mechanisms responsible for a pregnancy-induced pro-inflammatory environment that occurs in certain pregnancy disorders.