sTLR9 on maternal Circulating immune cells as a negative regulatory phenotype during pregnancy

Abstract

TLR9 is an intracellular receptor that can also be localized to the cell surface, called sTLR9. sTLR9 is thought to have a negative immunomodulatory effect, which is conductive to the maintenance of immune tolerance. Since pregnancy is a physiological process accompanied with inflammation experienced by pregnant women while maintaining immune tolerance to the fetus, the change in sTLR9 of immune cells during pregnancy are worth studying. In this study, we first found that with the progress of pregnancy, the most significant change in PWBCs of pregnant women was the increasing percentage of neutrophils (Neu%) accompanied by the decreasing sTLR9⁺ Neu%. Then, we found that percentages and sTLR9 levels of sTLR9⁺ Neu were significantly higher in pregnant mice than those in non-pregnant mice, while the latter was obviously elevated in the first and second trimesters than that in third trimester and after delivery. In mice, the TLR9 agonist CpG ODN induced a proinflammatory environment characterized by a significant increase in Neu% and a decrease or no change in sTLR9⁺ Neu%. In this case, the delivery time of pregnant mice was not affected, but their newborn mice showed significant weight loss. These results link sTLR9 as an immune cell phenotype to immune tolerance status during pregnancy, providing a kind of new insights into the mechanisms by which pregnant mother maintain immune tolerance to the fetus.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-025-05858-8.

Highlights

Changes in sTLR9 of immune cells reflected the immune status of pregnant mothers.

sTLR9 as a negative regulatory phenotype of maternal immune cells during pregnancy.

Pregnant mother kept high ratio of sTLR9⁺ Neu in coping with acute immune stress.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-025-05858-8.

Introduction

Toll-like receptor (TLR9) is a pattern recognition receptor that can recognize and sense a variety of DNA molecules [1], such as microbial DNA rich in hypomethylated CpG motifs, artificial synthetic CpG oligodeoxynucleotide (CpG ODN), and self-DNA, like a mitochondrial DNA. TLR9 is mainly located in the endoplasmic reticulum (ER) and endosome of its expressing cells, which are intracellular receptors, and can also be localized to the cell surface, known as surface TLR9 (sTLR9). When DNA or CpG ODN as ligand enters the cell, TLR9 migrates from the ER to the endosome with the assistance of uncoordinated 93 homolog B1 (Unc93B1), becoming endosomal TLR9 (eTLR9) [2, 3]. After being cleaved by protease in the acidified environment of the endosome, TLR9 is activated and binds to the ligand to mediate the MyD88-dependent TLR9 signaling [4]. With the research progress, new discoveries have been made about the migration and distribution of TLR9. On the one hand, TLR9 exists in the ER and Golgi apparatus as a full-length inactive form, and then migrates with the assistance of Unc93B1 from the Golgi apparatus to either the endosome to be eTLR9 or the surface of cell membrane to become sTLR9 [5–7]. On the other hand, sTLR9 can interact with clathrin AP2 after internalization and migrate to the endosome with the assistance of Unc93B1, where it functions as eTLR9 [8, 9]. eTLR9 has been widely studied, but the study on the influencing factor for the sTLR9 on immune cells and its physiological significance is still lacking and inconclusive. From the perspective of TLR9 migration, the immune cell activation induced by TLR9 signaling should be accompanied by a decrease in their sTLR9 levels. We found in previous study that CpG ODN induces sTLR9 into endosomes while activating B cells [10]. Actually, sTLR9 has been found on almost all immune cells including neutrophils (Neu), B cells and monocytes (Mon)/macrophages (Mφ) [11], even on red blood cells [12]. One study suggested that sTLR9 may only be a reserve pool for eTLR9 because it does not bind to CpG ODN when located on B cells [13], hinting that it may be an inactive status on the cell surface. However, studies have shown that sTLR9 on Neu is also a DNA sensor [14]. Our previous studies revealed that sTLR9⁺ Neu played a negative immunoregulatory role in the systemic inflammatory response induced by TLR9 activation, and the increased number of sTLR9⁺ Neu contributed to the survival of mice from excessive systemic inflammatory reaction [15]. If sTLR9 on Neu is a negative immunomodulatory molecule, then it is worth considering whether sTLR9 on maternal immune cells including Neu is elevated as a negative regulator during pregnancy, so as to avoid premature initiation of an immune response conducive to parturition.

Pregnancy is a physiological process accompanied by the inflammatory response, from the embryo implantation, the pregnancy maintenance to the parturition [16]. The parturition is the final inflammatory stage of pregnancy. In order to establish and maintain a successful pregnancy, the maternal immune system must keep an immune tolerance state to the semi-allogeneic fetus [17]. Although there is much research on the maintenance of immune tolerance in pregnant mother during pregnancy, most have focused on the interaction among various immune cells, decidual stromal cells and trophoblasts at the maternal-fetal interface in the first trimester [18–28]. There are few studies on how peripheral blood immune cells, including B cells and innate immune cells such as Neu and Mon, participate in the maintenance of immune tolerance in pregnant mother. In normal pregnancy, it is thought that the innate immune cells of pregnant mother cannot be activated to trigger a strong inflammatory response [29]. The negative regulation of inflammation may help explain the maintenance of immune non-response or low response during pregnancy. The elevated levels of sTLR9 on Neu and Mon/Mφ negatively regulated TLR9-activated excessive inflammatory responses [15, 30]. sTLR9-expressing Neu can produce large amounts of interleukin 10 (IL-10) at early stage of the inflammation induced by TLR9 agonist [30]. As IL-10 is an anti-inflammatory factor and IL-10-deficient mice are prone to fetal loss (absorption) [31], we wondered if sTLR9 would be expressed at elevated levels on maternal peripheral white blood cells (PWBCs) during maintenance period of pregnancy, thereby acting as a negative regulator to avoid excessive or premature inflammatory responses. There is an increasing amount of cell-free fetal DNA (cffDNA) from the placenta in the circulation of pregnant women. After 20 weeks of pregnancy, the amount of cffDNA entering the maternal circulation steadily increases at a rate of 1% per week, reaching a peak just before delivery [32–34]. Unlike maternal cell-free DNA (cfDNA) without hypomethylated CpG motif, cffDNA contains 0.06% hypomethylated CpG motif and can therefore act as a TLR9 agonist [35, 36]. In the absence of infection, cffDNA can activate the eTLR9 of immune cells, thereby triggering inflammatory responses and uterine contractions leading to delivery [37, 38]. However, during a normal pregnancy, immune cells in the circulation of pregnant mother are constantly stimulated by cffDNA, but not activated to launch a strong inflammatory response [29], until the end of pregnancy, when the concentration of cffDNA reaches its peak, thus triggering an inflammatory response through TLR9 signaling to initiate labor [39]. Why TLR9 is not activated by cffDNA and triggers the inherent immune inflammatory response during the maintenance of normal pregnancy, thus leading to abortion, premature birth and other adverse pregnancies. Perhaps there is a certain proportion of sTLR9⁺ PWBCs in the pregnant maternal circulation that is involved in controlling the intensity of the inflammatory response during pregnancy, until the last moments of delivery. To date, there have been no studies on the changes in sTLR9 of maternal PWBCs during pregnancy.

In this study, we investigated the changes in sTLR9 of circulating PWBCs in pregnant women and pregnant mice. It was proved that sTLR9 is a negative regulatory phenotype of maternal PWBCs during pregnancy. A certain proportion of sTLR9⁺ immune cells, especially sTLR9⁺ Neu, in pregnant mothers is conducive to maintaining a state of immune tolerance or low inflammatory response during normal pregnancy, and also effectively coping with the effects of acute immune stress to the pregnancy.

Materials and methods

Human sample collection from pregnant and non-pregnant women

Human peripheral blood samples in this study were from 51 cases of pregnant women and 9 cases of non-pregnant women, in The First Hospital of Jilin University (Jilin, China) from September 2019 to October 2019. The inclusion criteria for pregnant women were single birth, normal temperature, normal pressure, normal pre-pregnancy body mass index and gestational weight gain. The exclusion criteria for pregnant women were chronic diseases, infectious diseases, and pregnancy complications. This study has been carried out in accordance with the World Medical Association Declaration of Helsinki*, approved by the Ethics Committee of The First Hospital of Jilin University (Ethical Approved No. 2018 − 481), and all subjects provided written informed consent. The general information of pregnant and non-pregnant women is shown in Table 1.

Table 1.

Characteristics of blood samples of pregnant and non-pregnant women

Groups	Pregnant women					Non-pregnant women	p1	p2
	Antepartum				Pospartum(1-2 days after delivery)
	Total	1st trimester (<14 weeks)	2nd trimester (14-27+6 weeks)	3rd trimester(≥28 weeks)
Number of samples, n	51	16	18	17	13	9
Age, years, mean(±SD)	30.88±4.84	30.94±5.59	32.06±4.98	29.59±3.78	30.46±5.12	33.56±4.53
Gestational age, days, mean (±SD)		49.31±8.53	144.3±22.02	221±14.18
WBCs, mean % (±SD)
Neu%	0.71±0.07	0.64±0.06	0.74±0.04	0.74±0.05	0.81±0.05	0.62±0.08	<0.0001	<0.0001
Lym%	0.21±0.06	0.28±0.06	0.19±0.03	0.18±0.03	0.12±0.04	0.30±0.08	<0.0001	<0.0001
Mon%	0.06±0.01	0.06±0.01	0.05±0.01	0.06±0.02	0.06±0.02	0.05±0.01
WBCs, mean # (±SD)
Neu#	6.38±2.19	4.52±1.35	7.42±1.91	7.02±2.07	10.75±3.89	4.37±1.06	0.0001	<0.0001
Lym#	1.81±0.41	1.88±0.35	1.83±0.43	1.74±0.45	1.55±0.55	2.01±0.37
Mon#	0.52±0.17	0.44±0.12	0.53±0.13	0.59±0.22	0.83±0.25	0.38±0.06	0.0192	<0.0001

Data are presented as median ±SD. p1 Kruskal Wallis test between 1st trimester, 2nd trimester and 3rd trimester; p2 Kruskal Wallis test between Antepartum, Postpartum and Non-pregnancy;# Cell counts (×10⁹/L); WBC white blood cell; Neu neutrophil; Lym lymphocyte; Mon monocyte

Mice and animal experiments

Female and male (for mating purposes only) ICR mice and Balb/c mice were purchased from Liaoning Changsheng Biotechnology (Shenyang, China) or Yisi Laboratory Animal Technology (Changchun, China). All mice were maintained in microisolator cages under laminar flow condition and received human care in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals (NIH Publications No. 80 − 23) revised in 1996 and the guidelines of Jilin University. All mouse experiments were carried out in Animal center and the in vitro experiments using mouse splenocytes were performed in Department of Molecular Biology, College of Basic Medical Sciences, Jilin University. During experiments, mice were given free access to food and water. The experimental manipulation was undertaken with the approval of the Scientific Investigation Board of Science and Technology of Jilin Province, China. With the authorization of The College of Basic Medical Sciences of Jilin University ethics committee (2020–98 and 2024–292).

To get pregnant mice, mice were acclimated for 1–2 weeks and their blood pressure was measured using Tail-cuff SBP (BP-2010 A, Beijing Soft Biotechnology Co., LTD) before mating. Two or three virgin female mice were mated with one male mouse overnight. Mating success was assessed the following morning by looking at the vaginal plugs of the female mice. The day the vaginal plug was seen was recorded as day 0.5 (E0.5) of pregnancy. In the early (E9.5), middle (E14.5) and third (E19.5) trimesters of pregnancy, and 2nd/5th day after delivery or −1st day before mating, peripheral blood was collected through the tail vein of pregnant mice. The blood samples were treated with anticoagulation for subsequent testing.

In vivo experiments, pregnant or non-pregnant female mice were injected with CpG ODN 1826 (CpG1826, TLR9 agonist), CCT ODN (SAT05f, TLR9 inhibitor) [40] or NaCl through tail vein and then bled for collecting peripheral blood samples at different time points after injection. The injection dose of CpG ODN was determined based on our previous study and comprehensive literature review. While 5 µg CpG ODN suffices as a vaccine adjuvant to promote antibody production [41], intravenous administration of 100 µg CpG ODN achieves a robust stress response without inducing pharmacological toxicity [42]. The blood samples were treated with anticoagulation for subsequent testing. In addition, the natural delivery time of the pregnant mice and the appearance and weight of newborn mice were recorded.

In vitro experiment using mouse splenocytes

Mouse splenocytes were prepared according to the method reported previously [10] and cultured in complete RPMI 1640 medium containing 10% heat-inactivated FBS and antibiotics (100 U/mL penicillin and 100 mg/mL streptomycin) in a 5% CO₂ humidified incubator at 37 ℃. For in vitro experiment, splenocytes (3 × 10⁶ cells/mL) were stimulated using CpG ODN 5805 (CpG5805, a TLR9 agonist) [10] at 3 µg/mL alone or combined with chloroquine (2 µg/mL) (Sigma-Aldrich, St. Louis, MI, USA) or Ap2-siRNA (GenePharma, Shanghai, China) at 37 ℃, 5%CO₂ for 48 h. Then, cells were harvested for surface staining with fluorescence-labeled antibodies followed by detection using flow cytometry. The CpG ODN was synthesized by fully phosphorothioate-modification at Sangon Biotech (Shanghai, China) and diluted in PBS followed by testing with Limulus amebocyte lysate assay (Associates of Cape Cod, Inc., MA, USA) to confirm that there was no detectable endotoxin.

Flow cytometry for human and mouse PWBCs

For surface staining human peripheral white blood cells (PWBCs), 100 µL of human whole peripheral blood was centrifuged (3000 rpm, 5 min, 4 ℃) (Eppendorf, Hamburg, Germany, Cat. No: 5424R) and the supernatant was discarded. Each tube was added with fluorescence-labeled mAbs of anti-CD45 (PerCP, clone 2D1, #652803, BD Biosciences, NJ, USA) and anti-TLR9 (PE, clone eB72-1665, #560425, BD Biosciences, NJ, USA).

For surface staining mouse PWBCs, 50 µL of mouse whole peripheral blood was added with fluorescence-labeled mAbs of anti-CD19 (PE, clone 1D3, #557399, BD Biosciences, NJ, USA), anti-Ly6G (FITC, clone 1A8, #551450, BD Biosciences, NJ, USA) and anti-TLR9 (Alexa Fluor 647, clone 1138D, #FAB7960R, R&D Systems, MN, USA). The flow cytometry gating strategy [43–45] sequentially identified: (i) Alive cells (P1), (ii) Singlets (P2), (iii) cell subsets from P2, and (iv) sTLR9⁺ populations within each subset.

All stained samples were incubated on ice in dark for 30 min. Then, the mouse sample was treated with 0.5 mL of ACK buffer (1/9 V/V) (NH4Cl 8024 mg/L, KHCO3 1001 mg/L, Na2EDTA 3.7 mg/L, pH 7.2–7.4) (Sinopharm Chemical Reagent Co. Ltd., Shanghai, China) on ice for 2 min, and the human sample was treated with 1.0 mL of 1× lysing solution (#349202, BD Biosciences, NJ, USA) at room temperature for 15 min to lyse red blood cells (RBCs). Thereafter, all samples were centrifuged at 2500–3000 rpm (5 min, 4 ^◦C), and the supernatants were discarded, which was repeated once. The remaining PWBCs were washed twice with PBS, resuspended with PBS/2% FBS, and then detected with an Accuri C6 flow cytometer (BD Biosciences, NJ, USA).

Flow cytometry for cultured mouse splenocytes

The mouse splenocytes cultured with different reagents were harvested for surface staining with fluorescence-labeled mAbs of anti-CD19 (PE, clone 1D3, #553786, BD Biosciences, NJ, USA), anti-CD40 (FITC, clone 3/23, #124608, BioLegend, San Diego, California), anti-TLR9 (Alexa Fluor 647, clone #1138D, #FAB7960R, R&D Systems, MN, USA) and corresponding isotype controls (BD Biosciences, NJ, USA) on ice in dark for 30 min. After staining, the cells were washed twice in PBS/2% FBS before detection. All stained cells were detected on an Accuri C6 flow cytometer.

RNA interference

Splenocytes were transfected with siRNAs targeting Ap2 mRNA (siRNA-Ap2) or with control siRNA (siRNA-NC) using Lipofectamine™ Reagent 3000 (L3000-001, Invitrogen). For each RNAi assay, the isolated splenocytes were cultured in Opti-MEM at 3 × 10⁶ cells/well in a 24-well plate. After 4 h of the transfection, the culture supernatant in the well of the plate was discarded and 1mL of RPMI 1640 containing 10% FBS was added for further study. All siRNAs were designed and synthesized in Shanghai GenePharma Co., Ltd (Shanghai, China). The interference of siRNA-Ap2 was detected by RT-qPCR and Western blotting. The sequence of siRNA-Ap2 selected for subsequent studies is as follows: 5’-GCACUGAAAACCUUCAUCATT-3’ and 5’-UGAUGAAGGUUUUCAGUGCTT-3’. Quantitative real-time polymerase chain reaction (RT-qPCR) was performed using two-step SYBR green qPCR assays. The data were acquired using the Step One™ real-time PCR system (Applied Biosystems, Foster City, CA, USA). The mRNA level of Ap2 was normalized with the mRNA level of Gapdh and analyzed with 2^−△△Ct method. The protein level of AP2 was detected by Western blotting and normalized with the level of GAPDH by the ratio of gray values.

Statistical analysis

Data are shown as mean ± SD. All calculations and statistical analysis were performed using GraphPad Prism 8.0 software (GraphPad, San Diego, CA) for Windows. Comparisons between groups were conducted using analysis of unpaired t tests. p < 0.05 was regarded as statistically significant.

Results

Changes in sTLR9 of peripheral white blood cells (PWBCs) of pregnant women

To study the relationship between sTLR9 and pregnancy, we collected peripheral blood samples from normal pregnant women at different pregnant stages and women with non-pregnancy (Table 1), and detected the number and percentage of PWBCs and their subsets, as well as the expression of sTLR9 on these cells in the samples by flow cytometry. After gating based on CD45/SSC, we analyzed the expression of sTLR9 on neutrophils (Neu), monocytes (Mon) and lymphocytes (Lym) in PWBCs (Fig. 1A). In the analysis of PWBC subsets in pregnant women, we found the percentage of Neu was significantly higher in peripheral blood of both antepartum (AP) and postpartum (PP) pregnant women than that of non-pregnant women, and higher in the second and third trimester than in the first trimester (Fig. 1B). This finding is consistent with literature reports [46, 47]. When analyzing the expression of sTLR9 on different subsets of PWBCs in pregnant and non-pregnant women, we found in addition to the decrease of percentage of sTLR9⁺ Mon in AP and PP pregnant women compared with that in non-pregnant women, the percentages of sTLR9⁺ Neu/Lym and the levels of sTLR9 on Neu/Mon/Lym were higher or tended to be higher in AP pregnant women than in non-pregnant women or PP women (Fig. 1C). We then analyzed the changes of sTLR9 on subsets of PWBCs in pregnant women during different gestational stages. It was found that in pregnant women, the percentages of sTLR9⁺ Neu and sTLR9⁺ Mon were higher in the 1 st trimester than in the 2nd and 3rd trimesters, and the percentage of sTLR9⁺ Lym showed the same trend. The level of sTLR9 was significantly higher on Mon of pregnant women in the early pregnancy than that in the late pregnancy, and showed a similar trend for Neu and Lym (Fig. 1D). This result suggests that the change of sTLR9 on PWBCs may be a phenotypic characteristic reflecting the immune status in pregnant women, and its increase indicates that the immune status is tolerable.

Fig. 1.

Analysis of sTLR9 expression on peripheral white blood cells of pregnant women by flow cytometry. (A) The experiment procedure for surface staining of peripheral white blood cells (PWBCs) with fluorescently labeled anti-CD45 and anti-TLR9 mAbs and the gating strategy of flow cytometry. (B) Percentage of neutrophils (Neu), monocytes (Mon) and lymphocytes (Lym) in CD45⁺ PWBCs of peripheral blood of pregnant women or non-pregnant women. (C) The percentage of sTLR9⁺ cells in Neu, Mon and Lym of CD45⁺ PWBCs in peripheral blood of pregnant women or non-pregnant women, and sTLR9 levels on them. (D) The percentage of sTLR9⁺ cells in Neu, Mon and Lym of CD45⁺ PWBCs in peripheral blood of pregnant women with different trimesters, and sTLR9 levels on them

Changes in sTLR9 of Neu and B cells in peripheral blood of pregnant mice

To confirm the above findings on PWBCs of pregnant women, we detected the sTLR9 on Ly6G⁺ Neu and CD19⁺ B cells in PWBCs of pregnant mice by flow cytometry (Fig. 2A). The results showed that the percentages of Ly6G⁺ and CD19⁺ cells in peripheral blood had no statistical difference between pregnant and non-pregnant mice, or between mice at different trimesters and postpartum mice (Fig. 2B). The percentage of sTLR9⁺Ly6G⁺ cells and their sTLR9 levels in either antepartum (AP) or postpartum (PP) pregnant mice were all significantly higher than those of non-pregnant mice. Similarly, sTLR9 levels on CD19⁺ B cells in AP and PP pregnant mice were also significantly higher than those of non-pregnant mice (Fig. 2C). This result suggests that the peripheral Neu and B cells of pregnant mice seem to be in a state of low response. Then, we analyzed the expression of sTLR9 on Neu and B cells in pregnant mice with different gestation stages. The result showed that the percentage of sTLR9⁺Ly6G⁺ cells did not differ during the whole pregnancy and the day 2 postpartum, but increased significantly on the day 5 postpartum. There was also no difference in the percentage of sTLR9⁺CD19⁺ cells from the first trimester to five days postpartum. However, sTLR9 levels on Neu and B cells differed significantly among pregnancy periods in pregnant mice. sTLR9 levels on Neu and B cells are higher during pregnancy at both the first and second trimesters, especially that at the second trimester. By the third trimester and postpartum, sTLR9 levels on both types of cells were significantly reduced compared to that of the pregnancy at first and second trimesters (Fig. 2D). This suggests that the number of sTLR9-expressing cells remained relatively stable, while the expression level of sTLR9 per cell fluctuated during pregnancy to adapt to physiological changes. The same result can be obtained from the dynamic change curve of sTLR9 levels on Neu and B cells in peripheral blood of 7 pregnant mice (Fig. 2E). These results indicate that the changes in sTLR9 of Neu and B cells in peripheral blood of pregnant mice are different at various stages of pregnancy, and the decrease in sTLR9 levels in the 3rd trimester may be related to the pro-inflammatory environment before delivery, which also explains why the sTLR9 levels on these two cells remain low for a certain period of time after delivery. Thus, elevated sTLR9 may indeed be a phenotype of peripheral white blood cells, representing a low maternal immune response during pregnancy.

Fig. 2.

Expression of sTLR9 on Neu and B cells in peripheral white blood cells of pregnant mice detected by flow cytometry. (A) Experimental procedures of surface staining for sTLR9, Ly6G and CD19 of peripheral white blood cells (PWBCs), and gating strategies for flow cytometry. (B) Percentage of Neu and B cells in PWBCs of pregnant mice. (C) Percentage of sTLR9⁺ Neu and sTLR9⁺ B cells and their sTLR9 expression level in PWBCs of pregnant mice. (D) Percentage of sTLR9⁺ Neu and sTLR9⁺ B cells and their sTLR9 expression level in PWBCs of pregnant mice with different trimester. (E) Dynamic change curves of sTLR9 levels on Neu and B cells in peripheral blood of pregnant mice during pregnancy and after delivery. NP, non-pregnant; AP, antepartum; PP, postpartum

The ascended sTLR9 of B cells is a phenotypic change in their low response to TLR9 agonist stimulation

To demonstrate that elevated sTLR9 represents a low response or inactivation state of immune cells, we selected mouse splenic B cells as our study subjects because B cells are known to constitutively express TLR9 in both humans and mice and can be activated by the TLR9 agonist CpG ODN. On the activated B cells, the expression of co-stimulatory molecules such as CD40 is increased, and the expression of surface TLR9 (sTLR9) is decreased (Fig. 3A) [10]. We stimulated mouse spleen cells with CpG ODN to establish an experimental platform for the activation of splenic B cells accompanied by a decrease in their sTLR9 levels. In CpG ODN-stimulated splenocytes, CD40⁺ B cells increased from 45 to 55% while sTLR9⁺ B cells decreased from 16 to 11%. Meanwhile, CpG ODN stimulation also increased the level of CD40 by about 0.2 to 0.5 times and reduced the level of sTLR9 by 0.4-fold on those B cells (Fig. 3B). Then, according to the literature that sTLR9 can interact with clathrin AP2 after internalization and migrate to the endosome with the assistance of Unc93B1 [8, 9], we applied siRNA targeting Ap2 mRNA (siRNA-Ap2) to mouse splenocytes to see what effect it had on splenic B cell activation and sTLR9 levels (Fig. 3C). After siRNA-Ap2, which could effectively inhibit the expression of AP2, was identified at the mRNA level and protein level by RT-qPCR and Western blotting, respectively (Fig. 3D), we used flow cytometry to detect the effect of siRNA-Ap2 on the expression of sTLR9 and CD40 on CpG ODN-induced splenic B cells. We found that siRNA-Ap2 not only allowed sTLR9 to reside on the surface of CpG ODN-induced B cells, but also down-regulated the expression of CD40 on these B cells (Fig. 3E). This result suggests that the elevation of sTLR9 represents non-response or low response of immune cells. To confirm the infer, based on the characteristics of TLR9 activation dependent on endosomal acidification environment [48], we used chloroquine to block CpG ODN-induced B cell activation (Fig. 3F). We found that chloroquine reduced the expression of CD40 on CpG ODN-induced splenic B cells, but increased the expression of sTLR9 on these B cells (Fig. 3G), indicating that increased sTLR9 levels could act as a phenotypic change in low or non-responsive immune cells.

Fig. 3.

CpG ODN induced the change of sTLR9 on the splenic B cells and its influencing factors. (A) Schematic diagram of TLR9 trafficking route and B cell activation induced by CpG ODN. (B) Changes in sTLR9 and CD40 of splenic B cells in mouse splenocytes stimulated for 24 h by CpG ODN in vitro. (C) Schematic diagram of Ap2 siRNA affecting AP2 function. (D) Screening of siRNA targeting Ap2 mRNA. Three siRNAs targeting Ap2 mRNA and one control siRNA-NC. (E) Effect of siRNA-Ap2 on CpG ODN-induced mouse splenic B cell activation and sTLR9 levels. (F) Diagram of chloroquine acting on B cell endosomes. (G) Effect of chloroquine on the change in sTLR9 and CD40 of spleen B cells induced by CpG ODN. Each point represents the average of the data obtained from two or three repeated wells of one murine-derived splenocytes

Effect of CpG ODN on percentage of circulating sTLR9-expressing immune cells in pregnant and non-pregnant mice

Since increased cffDNA in maternal circulation during pregnancy has been shown to activate TLR9 and participate in the initiation of delivery-related inflammatory responses, we injected mice with the TLR9 agonist CpG ODN through the tail vein to observe whether it altered the immune environment of the mice. We first injected 5 µg of CpG ODN intravenously into non-pregnant mice, and collected peripheral white blood cells (PWBCs) of the mice at 1 h, 4 h, 6 h and 24 h after injection, respectively. After detecting PWBCs in these mice by flow cytometry, we found that the percentage of Neu increased significantly at 4 h and 6 h after injection of CpG ODN, about twice that of mice injected with NaCl. Meanwhile, the percentage of sTLR9⁺ Neu decreased significantly at 4 h after CpG ODN injection. There was no significant difference in B cell changes between CpG ODN and NaCl groups (Fig. 4A). We then injected the same dose of CpG ODN into pregnant mice at E17.5 and examined the changes of B cells and Neu in PWBCs of these mice at 4 h and 6 h after injection. The results showed that, compared with the NaCl group, the percentage of B cells in CpG ODN group at both time points was decreased, in which the percentage of B cells at 4 h was significantly decreased, while the percentage of Neu only showed an increasing trend. In addition, the percentage of sTLR9⁺ B cells and sTLR9⁺ Neu showed no difference between the CpG ODN and NaCl groups (Fig. 4B). This suggests that 5 µg of CpG ODN may not be sufficient to affect the immune response of pregnant mice. Therefore, we increased the dose of CpG ODN to 100 µg per mouse in the following experiment to see its effect on Neu in PWBCs of pregnant and non-pregnant mice. We also first measured changes in circulating Neu and sTLR9⁺ Neu in non-pregnant mice at 1 h and 6 h after injection. After seeing that the change at 1 h was not significant, we analyzed the data obtained at 6 h. The results showed that CpG ODN increased Neu from 15 to 60% and reduced sTLR9⁺ Neu from 5 to 3.5% in mice compared with NaCl (Fig. 4C). It is suggested that the decrease in percentage of sTLR9⁺ Neu accompanied with the increase in percentage of overall Neu represent the formation of proinflammatory environment in mice under the agitation of CpG ODN. We then measured the percentage of Neu and sTLR9⁺ Neu in PWBCs of E17.5 pregnant mice treated with CpG ODN for 6 h. Meanwhile, in addition to NaCl as a control, we also selected a TLR9 inhibitory ODN (SAT05f) as a control. We found that CpG ODN significantly increased circulating Neu from 15 to 60% in pregnant mice compared to NaCl, while SAT05f did not. Neither CpG ODN nor NaCl nor SAT05f reduced the percentage of circulating sTLR9⁺ Neu in pregnant mice, which is inconsistent with what is found in non-pregnant mice (Fig. 4D). Considering that sTLR9 levels peaked at E14.5 in pregnant ICR mice as shown in Fig. 2E, whereas we examined sTLR9 changes at E17.5, suggesting that the E17.5 time point may not be optimal for observing sTLR9 level variations. Therefore, we measured the changes in circulating Neu and the percentage of sTLR9⁺Neu in pregnant mice at E14.5 after 6 h treatment with either CpG ODN or SAT05f. The results showed that CpG ODN significantly increased the percentage of circulating Neu in pregnant mice but markedly reduced the percentage of sTLR9⁺Neu, whereas SAT05f exhibited no regulatory effects (Fig. 4E). These findings suggest that CpG ODN injection indeed induced a heightened pro-inflammatory environment in pregnant mice.

Fig. 4.

Changes of percentages of sTLR9-expressing immune cells in peripheral blood of pregnant or non-pregnant mice treated by CpG ODN. (A) Dynamic changes of percentages of circulating B cells and Neu, and their sTLR9⁺ B/Neu in non-pregnant mice treated with CpG ODN (5 μg per mouse) or NaCl for various times. (B) Percentages of circulating B cells and Neu and their sTLR9⁺ B/Neu in E17.5 pregnant mice treated with CpG ODN or NaCl for 4 h and 6 h. (C) Percentages of circulating Neu and sTLR9⁺ Neu in non-pregnant mice treated 1 h and 6 h with CpG ODN (100 μg per mouse) or NaCl. (D) The changes of percentage of circulating Neu and sTLR9⁺ Neu in E17.5 pregnant mice treated with 100 μg CpG ODN, 100 μg SAT05f or NaCl for 6 h. (E) The changes of percentage of circulating Neu and sTLR9⁺ Neu in E14.5 pregnant mice treated with 100 μg CpG ODN, 100 μg SAT05f or NaCl for 6 h. There were three pregnant mice in each group

Effect of intravenous administration of CpG ODN in pregnant mothers on their delivery time and neonatal mice

Considering that the increased inflammatory response in pregnant mice may promote labor or affect the newborn, we recorded the labor timing of mother mice and the weight of newborn mice. We first analyzed the body weight of all newborn mice delivered from mother mice in each group, and found that the weight of neonatal mice in CpG ODN group was significantly lower than that in NaCl group and SAT05f group (Fig. 5A). The result indicates that pregnant mice given intravenous CpG ODN may have been born prematurely, because their babies lost weight. To support our infer, we conducted the following study. We randomly analyzed the delivery time of three pregnant mice in each group and the weight of their babies. The results showed that the delivery time of pregnant mice injected intravenously with CpG ODN at E17.5 was not different from that of other groups, all among E19.5-E21.5. However, neonatal mice born from those mother mice in CpG ODN group had significantly lower body weight, while which didn’t in SAT05f group (Fig. 5B, left). The result suggests that CpG ODN did affect neonatal weight, but did not cause preterm birth in pregnant mice. In addition, we found that the rate of postnatal weight gain of neonatal mice in the CpG ODN group was not significantly different from that in the other groups, and even faster (Fig. 5B. right). This result may suggest that although the birth weight of newborn mice in the CpG ODN group was low, the mice themselves were not harmed by the mother’s body changes. We then injected CpG ODN or NaCl intravenously to pregnant mice for three consecutive days starting from E15.5, once a day, and found that the delivery time of pregnant mice in the two groups still remained similar, but the weight of newborn mice from the CpG ODN group was significantly reduced (Fig. 5C). This result, combined with the previous fact that CpG ODN significantly increased the percentage of Neu, suggests that TLR9 agonists did create a pro-inflammatory environment in the pregnant mother, but the mother stabilized her immune tolerance status by an unknown mechanism, thereby preventing premature labor. However, during this period of resistance to inflammation, the maternal supply of nutrients to the fetus is insufficient.

Fig. 5.

Effect of intravenous administration of CpG ODN in pregnant mothers on their delivery time and neonatal mice. (A) Photos and weights of neonatal mice produced by pregnant mothers treated with CpG ODN (100 μg per mouse), SAT05f (100 μg per mouse) or NaCl once at E17.5. The photos of neonatal mice were taken from a litter of neonatal mice delivered by one of mother mice in each group. The weight results were derived from the weight of all neonatal mice delivered by all mother mice in each group. There were 81, 88, and 76 newborn mice delivered by mother mice of NaCl, CpG ODN, and SAT05f groups, respectively. (B) The delivery time of three mother mice in each group and the weight of their newborn mice. All mother mice were given the injection of NaCl, CpG ODN or SAT05f once at E17.5. (C) Effects of three consecutive injections with NaCl or CpG ODN on delivery time of mother mice and the weight of their newborn mice. In (C) & (D), "+" means giving birth once.

Discussion

In this study, we detected the changes of sTLR9 on circulating PWBCs of pregnant women and pregnant mice, and found that it may be an immune cell phenotype reflecting the maternal immune status during pregnancy, and its elevation is conducive to the maintenance of immune tolerance.

TLR9, as an intracellular pattern recognition receptor, is considered to be a negative regulatory mode to avoid overresponse to TLR9 agonists if it is localized to the surface of the cell membrane. Our previous research found that sTLR9⁺ Neu has a negative immunomodulatory effect in acute inflammatory responses. In this study, we found that there is a certain proportion of sTLR9⁺ PWBCs, especially sTLR9⁺ Neu, in the circulation of pregnant women, and the proportion of such cells is higher in the first and second trimesters. We also found significant increases in percentages of sTLR9⁺ Neu and sTLR9⁺ B cells in pregnant mice. Since pregnancy is a process accompanied by gradually increasing inflammatory response, and the immune system of pregnant mother needs to maintain immune tolerance or low response to the allogeneic fetus, the negative immune regulation is particularly important. In the PWBCs of pregnant women, we mainly observed an increase in the number and percentage of Neu with the progression of pregnancy and an increase in the percentage of sTLR9⁺ Neu in the first and second trimesters. In the PWBCs of pregnant mice, we did not observe a significant increase in the number and percentage of Neu, but did observe a significant increase in the percentage of sTLR9⁺ Neu and sTLR9⁺ B cells. In addition, the more significant change was the increased level of sTLR9 on Neu and B cells in pregnant mice during the pregnancy maintenance, especially in the second trimester. This may be related to the following factors. (1) Neu is the largest proportion of immune cells in human PWBCs, but not the largest proportion in mouse PWBCs [49–51]. However, in both human and mice, Neu is the immune cell that responds fastest to various stimuli, such as infections. Therefore, changes in Neu may have the greatest impact on the immune status of pregnant mothers. (2) The cffDNA in the circulation of pregnant women began to gradually increase during pregnancy [32–34]. Since cffDNA has been shown to be an agonist of TLR9, the elevated sTLR9 on PWBCs may be a self-regulatory mechanism to avoid overactivation of TLR9 signaling. For this purpose, fine-tuning sTLR9 expression per cell could offer greater sensitivity and efficacy compared to decreasing sTLR9⁺ cell populations. Our in vitro study of mouse splenic B cells also demonstrated that the increase of sTLR9 levels is accompanied by the decrease of cell activity. (3) Since the inflammatory state of pregnancy is a slow worsening process, the proportion of sTLR9⁺ cells in PWBCs of pregnant mothers is also a slow reducing process, which should be a result of self-adaptation and adjustment, so that the body can adapt to the emergence of acute and severe inflammation required at the end of pregnancy. A recent study has shown that the expression of immune checkpoint protein Programmed death-ligand 1 (PD-L1) on Neu across maternal peripheral blood, vascular endothelium and maternal-fetal interface tissue of C57BL/6 pregnant mice showed a downward trend from days 10.5 to 18.5 of embryo. Moreover, PD-L1 levels on maternal-fetal interface tissue mononuclear phagocytes also decreased significantly with the prolongation of gestation time. These results suggest that less need for immune suppression in third trimester, as parturition nears [52]. This seems to support our conjecture, as our previous work has shown that sTLR9⁺ Neu play a negative regulatory role in excessive inflammatory responses [15, 30], suggesting that sTLR9 acts somewhat like a checkpoint molecule.

The above explanation may only apply to normal pregnancy, as use of the TLR9 agonist CpG ODN did not significantly reduce the percentage of circulating sTLR9⁺ PWBCs in pregnant mice. May be, the presence of maternal circulating sTLR9⁺ PWBCs, especially sTLR9⁺ Neu, is a buffer mode to cope with sudden immune stress during pregnancy. In this study, we observed that percentage of Neu increased significantly and the percentage of sTLR9⁺ Neu decreased obviously in circulating PWBCs of non-pregnant mice treated with CpG ODN. However, in PWBCs of pregnant mice treated with CpG ODN in the same way as non-pregnant mice, the percentage of Neu also increased significantly, but the percentage of sTLR9⁺ Neu remained unchanged at E17.5. The significant increase in Neu should be a manifestation of the body’s inflammatory response, because Neu is the innate immune cell that responds the fastest to various infections, injuries and other stimuli. That is, we simulated infection stress with CpG ODN, causing both non-pregnant and pregnant mice to show a significantly increased percentage of Neu in their circulation, which should be a normal immune response to create a pro-inflammatory environment. As sTLR9⁺ Neu has negative immunomodulatory properties, its reduced percentage in the inflammatory response is also a relatively understandable result. This is consistent with what we have seen in normal pregnant women and pregnant mice, where the percentage of sTLR9⁺ Neu decreases as the pregnancy progresses. However, why did CpG ODN stimulation fail to induce a decrease in the percentage of circulating sTLR9⁺ Neu in pregnant mice? We do not yet know what is the mechanism behind this phenomenon. There are two possible reasons for this phenomenon. First, pregnant mice may be naturally more resistant to acute stress such as infection. Second, pregnant mice may have different response times to various stressors than non-pregnant mice, which may need to be determined by future dynamic observations. Considering that sTLR9 levels peaked at E14.5 in pregnant ICR mice as shown in Fig. 2E, whereas we examined sTLR9 changes at E17.5, suggesting that the E17.5 time point may not be optimal for observing sTLR9 level variations. Therefore, we measured the changes in circulating Neu and the percentage of sTLR9⁺Neu in pregnant mice at E14.5 after 6 h treatment with CpG ODN. We found that CpG ODN injection increased the percentage of circulating Neu while decreasing the proportion of sTLR9⁺Neu in E14.5 pregnant mice, indicating differential immune stress responses depending on gestational stages. According to our understanding of sTLR9⁺ Neu, the presence of a certain proportion of sTLR9⁺ Neu in maternal circulating PWBCs should be a way for pregnant mothers to stabilize the immune system when encountering acute excessive inflammatory stress. A recent study has shown that Neu play key roles in placental histopathological damage during Dengue virus infection, the unusual aggregation of granulocytes in the placenta resulted in abnormal expression of Matrix metalloproteinases that could affect placental vascularization, which may interfere with fetal-maternal exchange and ultimately lead to fetal intrauterine growth restriction [53]. In our study, CpG ODN was injected into the tail vein to stimulate pregnant mice and simulate the virus to induce immune inflammatory response. Whether the proliferation of Neu in peripheral blood also causes the accumulation of Neu in the placenta, causing fetal nutritional efficiency and hypoxia needs further experimental verification.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

TLR9

Toll-like receptor 9

sTLR9

surface Toll-like receptor 9

eTLR9

endosomal Toll-like receptor 9

PWBCs

peripheral white blood cells

CpG ODN

CpG oligodeoxynucleotide

cffDNA

cell-free fetal DNA

PBS

phosphate-buffered saline

FBS

fetal bovine serum

Authors’ contributions

Hong Wang: The overall idea of the topic, the specific implementation of the research, the collation and analysis of experimental data, and the writing of the draft article. Wenting Lu: Mainly participated in the in vitro experiment of mouse splenocytes. Mengru Zhu: Mainly participated in the in vivo experiment of pregnant mice. Yongli Yu: Guide the idea and direction of the project progress, and polish the article in English. Liying Wang: The basic idea of the topic, research guidance, article writing and revision.

Funding

This study is financially supported by the National Nature Scientific Foundation of China (31670937 and 81471888).

Data availability

The authors confirm that the data supporting the findings of the present study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval

This study was carried out in accordance with the World Medical Association Declaration of Helsinki*. Approval was granted by the Ethics Committee of The First Hospital of Jilin University approval to carry out the study (Ethical Approved No. 2018 − 481).

During the animal experiments, the mice were treated in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals and the guidelines of Jilin University, and with the approval of the Scientific Investigation Board of Science and Technology of Jilin Province, China. The Ethics Committee of The College of Basic Medical Sciences of Jilin University approved the mouse experiments in this article with the number of 2020-98 and 2024-292.

Consent for publication and consent to participate

Informed consent was obtained from all individual participants included in the study. Consent for publication was obtained from the participants.

Competing interests

The authors declare that there are no competing interests associated with the manuscript.