Maternal and Fetal CD4+CD25+CD 127low/-Regulatory T Cells in Pregnancies with Gestational Diabetes, Preeclampsia, and Premature Rupture of Membranes

Background Regulatory T cells (Tregs) play a crucial role in maternal-fetal tolerance, but little is known about the characteristic of Tregs in peripheral blood (PB), maternal-fetal interface and cord blood (CB) in normal and complicated pregnancies with preeclampsia (PE), gestational diabetes mellitus (GDM) and premature rupture of membranes (PROM). Methods PB, retro-placental blood (RPB), and CB were collected immediately after delivery in women with normal full-term pregnancy (NP), PE, GDM, and PROM. The proportion of CD4 + CD25 + CD127 low/- T cells (Tregs) and the expression of PD-1, GITR, HLA-G and CTLA-4 on T cell subsets were investigated by ow cytometric analysis. The data were analyzed based on sample origins (PB, RPB, and CB) and the obstetrical study groups (NP, PE, GDM, and PROM).


Background
The maternal immune system has various mechanisms to tolerate semi-allogenic pregnancy. Tregs have been reported to play an essential role in the maintenance of pregnancy. Tregs, identi ed by the surface markers CD4 + , CD25 + , and CD127 dim/- [1], are increased in early pregnancy, peaked in mid-pregnancy and decreased in late pregnancy [2]. There are four main processes by which Tregs suppress immune responses: (1) modulation of dendritic cell (DC) function or maturation, (2) release of inhibitory cytokines, PE, which complicates approximately 10% of pregnancies, is a leading cause of maternal-fetal morbidity and mortality worldwide. There is strong evidence that disturbance in the immune response during early pregnancy appears to be the central cause of later placental pathology and secondary systemic reaction, such as hypertensive disorder [9][10][11][12]. There is substantial evidence that many pregnant women with PE have fewer and less functionally competent Tregs [13,14]. Therefore, it is conceivable that defective Tregs may be causatively involved in the development of PE. GDM is de ned as carbohydrate intolerance withonset or rst recognition during pregnancy. Recent evidence indicates that GDM is characterized not only by increased insulin resistance and glucose intolerance but by a state of low-grade systemic in ammation and dysregulation of the immune system, which induces an imbalance of Tregs [15,16].
Like PE and GDM, the etiology of PROM is multifactorial. However, many studies have focused on the detection of evidence of in ammation during pregnancy [17]. It is generally believed that preterm PROM is causally linked to intra-amniotic in ammation and intra-amniotic infection [18]. Studies have shown that women with intra-amniotic infection had a higher number of total T and CD4 + T cells in amniotic uid [19]. There are phenotypic changes in granulocytes and Tregs, which are consistent with the presence of intravascular in ammation in preterm PROM and preterm labor patients [20,21]. The in ammatory response is crucial to the onset of labor [22]. Hence, abnormal expression of Tregs could explain the activation of maternal in ammatory cascade, leading to PROM and labor. Since immunosuppressive Tregs were shown to control excessive in ammation, and immoderate immune responses are known to be of vital importance for the successful course of pregnancy, it could be possible that functional de ciencies of such cells are involved in the pathogenesis of PROM.
Diverse expression of surface markers on Tregs may re ect the changes in their function and capacity to expand during pregnancy. There are several critical surface markers on Tregs, such as programmed cell death receptor-1 (PD-1/CD279), glucocorticoid-inducible TNF receptor family-related protein (GITR/CD357), human leucocyte antigen-G (HLA-G) and cytotoxic T lymphocyte antigen 4 (CTLA-4/CD152). PD-1 is a cell surface receptor belonging to the CD28 family. PD-1 cross-linking with PD-L1 results in the promotion of Tregs development and function [23]. The proportion of PD-1 + T lymphocytes were elevated in a normal healthy pregnancy, while a de ciency of PD-1 expression might cause the overreaction of T cells, which occurred in abnormal pregnancies [24][25][26]. It is reasonable to assume that activation. CTLA-4 is constitutively expressed on Tregs, and its ligation positively reinforces the immunosuppressive functions of Tregs.
Tregs from different origins may have unique phenotypes and various expressions of surface markers.
The purpose of this study was to investigate the proportion of Tregs and their phenotypic characteristics in PB, RPB, and CB of normal pregnant women and women with PE, GDM, and PROM.

Study population
The study was approved by the ethics committee of Yuhuangding hospital, and all study subjects signed informed consent and permission form prior to entering the study. PB was drawn from 94 pregnant women, including 15 PE, 28 GDM, 23 PROM, and 28 NP. 15 CB and RPB were taken from each group separately. Immediately after the cord was clamped, CB and RPB were collected into EDTA-containing tubes for Tregs evaluation. All women with NP delivered vaginally without any complication. None had multiple gestations, pregnancies with chromosomal or fetal abnormalities, diabetes, pregnancyrelated hypertension, autoimmune or chronic disease. PE was de ned as the occurrence of hypertension; systolic blood pressure of ≥ 140 mm Hg or as the diastolic blood pressure of ≥ 90 mm Hg, plus albuminuria; presence of ≥ 300 mg of protein in 24h urine sample after 20 weeks of gestation in a woman who was normotensive before 20 weeks gestation [11]. The diagnosis of GDM was made between 24 and 28 weeks of gestation by a positive 2-h 75 g oral glucose tolerance test (OGTT) with the following criteria: a fasting plasma glucose ≥ 5.1 mmol/l (92 mg/dl), or a 1-h plasma glucose level of ≥ 10.0 mmol/l (180 mg/dl), or a 2-h plasma glucose of ≥ 8.5 mmol/l (153 mg/dl) [38]. The diagnosis of PROM was con rmed by the clinical ndings of a posterior vaginal pool and the vaginal PH or ferning test. The clinical characteristics of women with normal and complicated pregnancies are presented in Table 1.
Cells were analyzed on a FACS Canto II ow cytometer (BD Biosciences, USA); 200,000 events were recorded. The gating strategy for the detection of cells was as follows: The population of Tregs was characterized by the gating of CD3 + , CD4 + , CD25 + , and CD127 low/subsets. Then, within these subsets, PD-1, GITR, HLA-G, or CTLA-4 cells were gated, respectively. The collected data were exported to FlowJo software for analysis (Tree Star, Ashland, OR).

Statistical analysis
Statistical analysis were performed using statistical software GraphPad Prism, version 5 (GraphPad, San Diego, CA, USA). Data were described as mean ± standard error (SEM). Normality tests were used first to determine if all the data sets are in normal distribution. Student t-test was applied to test the differences between the two groups. ANOVA was used to test the differences between multiple groups. The correlations of the proportion of Tregs with various surface markers between different groups were analyzed by linear correlation analysis. Statistical signi cance was set at P < 0.05.

Clinical characteristics of the study population
The clinical characteristics of the obstetrical study populations, including age, obstetrical history, and outcome of the index pregnancies are listed in Table 1.
The proportions of CD3 + , CD4 + , and CD8 + T cells in PB, RPB, and CB from NP, PE, GDM, and PROM The proportion of CD3 + T cells in PB was higher compared to RPB and CB in each obstetrical group (P < 0.01, respectively). CD3 + T cells in each sample origin (PB, RPB, and CB) were the same among obstetrical study groups (P = NS, respectively).
The proportion of CD4 + T cells in RPB was lower compared to those of PB and CB in each obstetrical group (P < 0.01, respectively). The proportion of CD4 + T cells in each sample origin (PB, RPB, and CB) was the same among obstetrical study groups (P = NS, respectively).
The proportions of CD8 + T cells in PB, RPB, and CB have no differences in each obstetrical group (P = NS, respectively). The proportions of CD8 + T cells in each sample origin (PB, RPB, and CB) were the same among obstetrical study groups (P = NS, respectively) ( Table 2) The proportions of Tregs in PB, RPB, and CB from NP, PE, GDM, and PROM The proportions of Tregs in PB, RPB, and CB from NP were signi cantly higher than those of other obstetrical study groups (P < 0.01, respectively). In NP, the proportion of Tregs in RPB (5.25% ± 0.33%) was signi cantly lower than those from CB (6.85% ± 0.35%) and PB (7.85% ± 0.53%) (P < 0.01, respectively).
There were no differences in Tregs proportions of each sample origin among PE, GMD, and PROM groups (P = NS, respectively) (Fig. 1A).
Expression of PD-1, GITR, HLA-G, and CTLA-4 on CD4 + T cells in PB, RPB and CB from NP, PE, GDM and PROM The proportions of PD-1 + /CD4 + T cells in RPB and PB of NP were signi cantly higher than those of other obstetrical study groups (P < 0.01, respectively). In all obstetrical study groups, the proportions of PD-1 + /CD4 + T cells from RPB were higher than those of CB (P < 0.01, respectively) while the proportions of PD-1 + /CD4 + T cells from PB were higher than those of CB (P < 0.01, respectively).
In all obstetrical study groups, the proportions of HLA-G + /CD4 + T cells in PB were signi cantly lower than those of CB (P < 0.05, respectively) and RPB (P < 0.01, respectively) while the proportions of HLA-G + /CD4 + T cells in RPB were signi cantly lower than those of CB (P < 0.05, respectively).
The proportions of CTLA-4 + /CD4 + T cells had no statistical differences among the sample origin or obstetrical study groups (P = NS, respectively) ( Table 3).

Discussion
Our experimental results showed that the proportions of CD3 + T cells from NP, PE, GDM and PROM were signi cantly higher in PB than in RPB or CB. The proportions of CD4 + T cells in RPB from NP, GDM, PE, and PROM were signi cantly lower than those in PB or CB. Interestingly, there were no differences in CD8 + T cells among the sample origins or the obstetrical study groups. The RPB taken immediately after delivery may re ect maternal-fetal interface blood with the potential mixture of maternal blood [39]. We named it as "retro placental blood" in this study. During pregnancy, maternal placental circulation allows bidirectional passage of soluble antigens and nucleated blood cells through the maternal-fetal interface [40]. At the maternal-fetal junction, the fetal blood cells are proximity to the maternal blood, however the mixture of fetal and maternal blood cells does not occur because of the uteroplacental interface [41][42][43].
In this study, we report that CB retains a lower proportion of CD3 + T cells but the same proportion of CD4 + T cells with maternal PB. This nding was inconsistent with previous report which demonstrated that CD3 + and CD4 + T cells were increased in CB compared to PB [44]. More interestingly, our ndings showed that CD3 + and CD4 + T cells in RPB were signi cantly lower than those in PB, which suggested that T cellrelated immunity was suppressed at the maternal-fetal junction at the time of parturition. CD3 + T cells account for 10% of maternal immune cells in the rst-trimester human decidua, and about 30-45% of CD3 + T cells are CD4 + T cells. In contrast, CD3 + T cells represent about 80% of lymphocytes in peripheral blood, and approximately 65% of these CD3 + T cells are CD4 + T cells [45,46]. These data indicate that CD3 + and CD4 + T cells are rare in the decidua during early pregnancy, possibly due to a reduced capacity for T cell accumulation [47]. Our ndings showed that these cells were also decreased at the maternal-fetal junction during late pregnancy compared to those in PB regardless of obstetrical study groups.
Tregs play a role in modifying the maternal immune response to the fetoplacental 'allograft' within the uterus [41,43,48,49]. Previously, we reported that the proportions of Tregs were decreased in PB and decidua of women with unexplained RPL when compared to normal early pregnant women [5,38]. Contrarily, the temporary elevation of Tregs on the day of embryo transfer was associated with the higher embryo implantation rate [50]. The adoptive transfer of Tregs rescued pregnancy in abortion-prone mice model [51]. Few studies have investigated Tregs at the maternal-fetal junction and CB at the time of delivery [12, 52,53]. In this study, we found that the proportions of Tregs were signi cantly decreased in women with obstetrical complications as compared to NP regardless of sample origins. In addition, the proportions of Tregs in RPB of each obstetrical study group were lower than those in PB in all obstetrical study groups, including NP, and the reduction rates of Tregs were the same among the obstetrical study groups. Hence, the proportion of Tregs in RPB is determined by the proportion of PB Tregs, and the reduction rate of Tregs in RPB was not affected by the presence of obstetrical complications.
Data suggested that fetus-speci c Tregs are speci cally recruited from PB to the maternal-fetal interface [54]. Recruitment failure of CD4 + CD25 bright Tregs to the maternal-fetal interface may play a role in the development of obstetrical complications. Both maternal factors and fetal antigenicity were reported to play a role in determining maternal Treg accumulation at the uteroplacental interface [55]. The postpartum decline in Tregs was consistent with the withdrawal of the immunological stimulus of the allograft [2]. In this study, we found that proportions of Tregs in PB, RPB, and CB from PE, GDM and PROM were lower compared to NP, which was consistent with the previous study [56]. It may suggest that immune suppression in the maternal side, as well as the maternal-fetal interface, was less effective in obstetrical complications than those of normal pregnancies.
The characteristics of Tregs, determined by their surface expression of various markers, were different between PB, RPB, and CB in this study. Previous studies showed that PD-1 provided multiple and possibly redundant mechanisms to enhance the suppressive function and stability of Tregs and promote Tregs differentiation [57][58][59]. While the engagement of GITR on Tregs inhibited their suppressive function and in vivo GITR stimulation increased the numbers of Tregs [60][61][62]. Our experimental results showed that the proportion of PD-1 + and GITR + Tregs were higher in PB and RPB than in CB from NP. Placental or trophoblast-derived soluble factors, including M-CSF, IL-10, TGF-β, and TRAIL may expand and activate Tregs in maternal side but not in fetal side [63]. Maternal PB and RPB contain possibly induced Tregs resulting from an encounter with foreign antigens, while Tregs in CB are functionally mature immuneregulatory population with naive phenotype [64]. Considering the naive phenotype and incomplete post thymic expansion of CB Tregs, the regulatory mechanisms engaged in the augmentation (PID-1 and GITR) of their suppressive activity might be acquired during the postnatal maturation process [65][66][67]. In this study, we found that the proportion of PD-1 + /CD4 + and GITR + /CD4 + T cells as well as PD-1 + and GITR + Tregs were signi cantly lower in CB compared to PB and RPB. Besides, the high proportion of PD-1 + and GITR + Tregs from PB and RPB suggest that it may create a tolerogenic immune environment that allows the development of allogeneic placental and fetal tissues.
Studies have shown that PD-1 + T cells seem to be increased in healthy pregnant women compared to non-pregnant women [68], while decreased PD-1 + T cells were observed in PE and GDM [69][70][71]. In our studies, we found that PD-1 + Tregs in CB, PD-1 + Tregs and GITR + Tregs in PB from PE, GDM, and PROM were lower compared to those of NP. These ndings underscored the critical roles of PD-1 and GITR in regulating Treg function and maintaining immune suppression during pregnancy, whereas insu cient PD-1 and GITR expression might cause unsustainability in maternal-fetal immune tolerance and consequently, lead to pregnancy complications.
The higher proportions of HLA-G + Tregs from RPB and CB in NP, PE, GDM, and PROM are other signi cant ndings of the present study. The expression of HLA-G is con ned to immune-privileged sites, although the highest expression is shown in the placenta, where its primary role involves the protection of the fetus from maternal immunity, thus critically contributing to maternal-fetal tolerance [72]. Previous studies demonstrated that the HLA-G + /CD4 + T cells were expanded in the decidua compared to the periphery [35,73,74], which was in line with our results. In this study, we found a much higher proportion of HLA-G + Tregs at the maternal-fetal interface compared to PB.The high density of Tregs might be necessary for fetal acceptance at the site of the highest allogeneic stimulation inevitably presented by the fetus.
However, it remains to be clari ed whether the HLA-G + Tregs are recruited from the periphery or be induced at the maternal-fetal interface to promote tolerance. Considering the functions of HLA-G in pregnancy-associated immune tolerance, the increased expression of HLA-G on Tregs from RPB and CB may, at least in part, contribute to the restraint of undesirable maternal alloreactivity. HLA-G also stimulates the production of angiogenic factors and cytokines that favor embryo implantation, placental vascularization, and maternal-fetal tolerance [75]. Low blood concentrations of soluble HLA-G (sHLA-G) have been associated with increased risks of PE, GDM, and intrauterine growth restriction [76][77][78].In this study, the proportion of HLA-G + Tregs in all sample origins of NP tends to be higher than those of PE, GDM and PROM, however the differences did not reach a statistically signi cant level. It may be due to small sample size, and further study is needed.

CTLA-4 is accepted as a crucial negative regulator of T-cell responses. Upon the stimulation of the TCR,
CTLA-4 becomes stabilized on the surface of T cells, thus competing with CD28 for B7 binding and impairs the activation of T cells. It is reported that the expression of CTLA-4 on the cell surface is a key for Tregs to exert suppression by a critical mechanism [79][80][81][82]. In animal studies, treatment with CTLA-4blocking antibody caused greater susceptibility to fetal loss with altered cytokine pro les by decidual CD4 + T (dCD4 + T) cells [83]. It is also reported that percentages of CTLA-4 + cells were signi cantly increased in the endometrium of women with recurrent implantation failure (RIF) and RPL than those in controls [84]. This study showed that there was no difference in the proportions of CTLA-4 + Tregs between the obstetrical study groups, which may indicate the decreased contribution of CTLA-4 in Tregs suppression during pregnancy.

Availability of data and materials
The primary data for this study is available from the authors on direct request.   Values are mean ± SEM. NS:no significance Values are mean ± SEM. NS:no significance