B Cell Development Undergoes Profound Modifications and Adaptations During Pregnancy in Mice1

Abstract

Pregnancy hides an immunological riddle combining two antagonistic characteristics of immunology: the existence of a tolerance that allows the gestation of a semiallogeneic fetus and proper protection against pathogens threatening the health of the immunocompromised mother. Despite the fundamental role that B cells play in orchestrating an immune response, their behavior in the context of pregnancy has been barely investigated. Here we demonstrate that numbers of pre/pro and immature B cells were progressively diminished in the bone marrow (BM) of pregnant mice, leading to a reduced influx of B cells in blood and spleen. Correspondingly, lower levels of B cell-activating factor of the TNF family were observed in serum of pregnant mice. In contrast to immature B cells, mature B cells were accumulated in the BM during pregnancy. Accordingly, higher numbers of mature B cells were observed in the lymph nodes draining the uterus as well as in the peritoneal cavity of pregnant mice, both tissues in close contact with the fetuses. Despite an increase in spleen size, pregnant mice showed lower numbers of splenic B cells, which was mirrored by lower numbers of immature and FO B cells. However, marginal zone B cells in the spleen increased during pregnancy. Additionally, serum IgM, IgA, and IgG3 titers were elevated in pregnant mice. Collectively, our data show how the B cell compartment adapts to the presence of the semiallogeneic fetus during gravidity.

Introduction

During mammalian pregnancies, the maternal immune system must be adapted in order to simultaneously “tolerate” the presence of the semiallogeneic fetus and to protect the mother against harmful pathogens. Failures in carrying on these adaptations may compromise not only the continuity of pregnancy but also the health of the mother. In this regard, it has been shown that pregnancy puts otherwise healthy women at an increased risk for serious complications from infections, such as influenza [1]. Indeed, the last pandemic influenza clearly showed that pregnant women are at particularly high risk of morbidity and mortality [2]. Furthermore, clinical course and severity of pre-existing autoimmune diseases may alter during pregnancy [3–5]. Surprisingly, despite the fundamental role that B cells play in orchestrating an immune response, their behavior in the context of gravidity has been barely addressed.

B cells can be divided into three different subsets: B1 B cells, follicular B cells (FO), and marginal zone (MZ) B cells. B1 B cells are further subdivided into B-1a and B-1b B cells [6]. B1 subsets are considered innate-like immune cells, and their role in the context of pregnancy has been previously investigated [7, 8]. Unlike B1 B cells, which are originated only during the embryonic life, FO and MZ B cells continuously originate from precursors present in the bone marrow (BM). In this compartment, the pro-B and pre-B cells represent the earliest B cell-linage stages. As soon as they express the cell surface marker IgM, they give rise to immature B cells, which leave the BM for further maturation in the periphery [9, 10]. Once in the spleen, immature B cells go through transitional developmental stages and originate either MZ or FO B cells. Each subset articulates a particular kind of immune response. Due to their preactivated phenotype, MZ B cells respond faster than FO B cells to blood-borne antigens [11]. Besides, MZ B cells maintain a basal polyreactive low-affinity antibody repertoire that acts as a first approach to address the threat of pathogens [12]. A more specific and powerful immune response is driven by the FO B cells, which generate a high-affinity antibody production through a T cell-dependent mechanism [6]. Hence, changes on the dynamic of B cell development during pregnancy may have an impact on the B cell repertoire, thus altering the capacity of the maternal immune system to fight pathogens.

Among the different molecules involved in the regulation of B cell development, the B cell-activating factor (BAFF) has been shown to have a central role [13–15].

This study aims to analyze how physiological changes occurring during gravidity alter or modify the development and behavior of B lymphocytes.

Materials and Methods

Animals

Six- to 8-wk-old female mice (C57B6/J) were purchased from Charles River. Mice were kept in our animal facility under optimal conditions in a 12L:12D cycle. Chow and water were supplied ad libitum. Animal experiments were carried out according to institutional guidelines after ministerial approval (institutional review board: Landesverwaltungsamt Sachsen-Anhalt [ID: FJ2-1019 to FJ] and Landesamt für Landwirtschaft, Lebensmittelsicherheit und Fischerei Mecklenburg-Vorpommern [7221.3-1-068/13 to F.J.]). The experiments were conducted in conformity with the European Communities Council Directive 86/609/EEC. One female mouse was placed with a BALB/c male in the same cage and checked daily for vaginal plugs. The day at which the vaginal plug was detected was considered Day 0 of pregnancy. Pregnant females were separated from the males and euthanized at Day 7 postplug (7 dpp), Day 14 postplug (14 dpp), or Day 18 postplug (18 dpp). As control, nonpregnant (np) age-matched females were included.

Cell Preparation and Flow Cytometry

Cells were isolated from the BM of femur and tibia, spleen, blood, peritoneal cavity (PerC), and para-aortic lymph nodes (lymph nodes draining the uterus). Erythrocytes lysis was performed for 5 min, and after washing, the cell count was determined using a Neubauer Chamber; 1 × 10⁶ cells were stained for 30 min at 4°C with specific labeled antibodies (see Table 1) or specific isotype controls. Data were acquired on FACSCalibur or FACS Canto (BD Biosciences) and analyzed by using FlowJo software (Tree Star Inc.). A detailed description of the gating strategies used to establish each subset according to a combination of their specific surface markers can be found in Supplementary Table S1 (all Supplemental Data are available online at www.biolreprod.org). To obtain the total cell number of each population, the percentage provided by the flow cytometry analysis (considering gating strategies) was referred to the cell count obtained with a Neubauer Chamber. In the case of blood samples, the cell count was divided by the total volume obtained on each blood collection procedure to obtain the results expressed as cells per milliliter of blood.

Table 1

Antibodies used in FACS analysis.

Molecule	Antibody clone	Label	Alternate nomenclature
CD23	B3B4	PE	FcεRII
CD5	53-7.3	APC
CD21	7G6	FITC
CD19	1d3	PE-Cy7
CD43	R2/60	FITC
CD45R	RA3-6B2	PE-Cy7	B220
IgM	II/41	APC
IgD	11-26c	PE
CD11c	M1/70	FITC	MAC-1
CD93	AA4.1	FITC	C1qRp

Analysis of Immunoglobulin in Serum

Serum levels of IgM, IgG1, IgG2a, IgG2b, IgG3, and IgA were assessed using the Milliplex Map mouse immunoglobulin kit (Millipore), and IgE in serum was evaluated by the Milliplex Map mouse IgE single-plex kit (Millipore) and subsequently analyzed on Luminex test equipment.

BAFF ELISA

Levels of BAFF in serum of pregnant and np animals were assayed using a commercially available ELISA kit from R&D Systems, following the supplier's recommendations.

Tissue Processing and Histopathological Morphometric Analysis

Spleens were formalin fixed, cut into two or three slices of 1 mm, and paraffin embedded; 2-μm sections were cut vertically and mounted on glass slides (clear glass microscope slides; Medite). Thereafter, the sections were deparaffinized with xylene and ethanol and stained with hematoxylin and eosin according to standard protocol. Slices were examined by two experienced histopathologists (K.U. and M.E.) using a Zeiss microscope (Axioplan2; Zeiss). Afterward, the hematoxylin and eosin-stained specimens were scanned (Leica SCN 400 slide scanner). Morphometric analysis was performed at a magnification level that still allowed a clear-cut separation of tissue compartments (i.e., red and white pulp) on the one hand with as much tissue area visible on the screen on the other. According to the shape and size of white pulp tissue, magnification varied between 120- and 200-fold. The relative volume of white and red pulp tissue was measured using Digital Image Hub (SlidePath; Leica Biosystems).

Statistical Analysis

Data were analyzed with PRISM software (ver. 5.01; GraphPad). Analysis of variance (ANOVA) followed by Tukey multiple post t-test or Kruskal-Wallis test as appropriate were applied to evaluate differences of means of multiple groups. Significant differences between groups are indicated with asterisks (*P < 0.05; **P < 0.01; ***P < 0.001). Nonparametric Spearman correlation was used to analyze the correlation between spleen weight and B220⁺ B cells in the spleen.

Results

Early Stages of B Cell Development Are Progressively Reduced While Mature B Cells Are Increased in the BM During Murine Pregnancy

The analysis of different B cell subpopulations in the BM of pregnant and np females depicted profound changes concerning B lymphopoiesis. The total numbers of B220 intermediate and surface IgM negative B220^intsIgM⁻ pre/pro B cells [9] in the BM of pregnant mice at 7 dpp revealed no differences when compared to np mice (Fig. 1). However, as pregnancy advanced, the numbers of B220^intsIgM⁻ pre/pro B cells significantly dropped from 9.8 ± 0.8 × 10⁶ cells in np mice to 3.8 ± 0.5 × 10⁶ cells at 14 dpp and kept decreasing to 2.3 ± 0.4 × 10⁶ cells toward the end of pregnancy (18 dpp). The decrease in the numbers of pre/pro B cells in the BM of pregnant mice was partly accompanied by a reduction in the numbers of B220^intsIgM⁺ immature B cells at 14 dpp compared to np mice (1.4 ± 0.2 × 10⁶ vs. 3.3 ± 0.45 × 10⁶ cells; Fig. 1). However, when analyzed at late pregnancy (18 dpp), numbers of B220^intsIgM⁺ immature B cells were similar to those observed in np control females (Fig. 1). Unlike immature B cells, B220⁺sIgM⁺ mature/recirculating B cell numbers significantly increased in pregnant mice at late pregnancy as compared to np as well as to pregnant mice at 7 and 14 dpp (8.9 ± 1.2 × 10⁶ vs. 2.1 ± 0.3 × 10⁶, 1.6 ± 0.2 × 10⁶, and 2.3 ± 0.3 × 10⁶ cells, respectively; Fig. 1). Interestingly, when the ratio between immature and mature B cells was assayed, a clear preponderance of mature B cells over immature B cells was evidenced at 14 dpp and late pregnancy (Fig. 1). In summary, pregnancy was associated with a gradual reduction in the numbers of pre/pro and immature B cells in the BM, while an accumulation of mature/recirculating B cells was observed in this tissue.

Fig. 1

Pregnancy induces a reduction of B cell production from the BM. BM cells obtained and pooled from femur and tibia of pregnant as well asnp females were analyzed for the expression of B220 and sIgM. Representative dot plots from np as well as pregnant females (7, 14, and 18 dpp) depicting the gating strategy for B220^intsIgM⁻ (pre/pro B cells), B220^intsIgM⁺ (immature B cells), and B220⁺sIgM⁺ (mature/recirculating B cells). Graphs show total number of B220^intsIgM⁻ pre/pro B cell, B220^intsIgM⁺ immature B cells, and B220⁺sIgM⁺ mature/recirculating B cells in the BM of pregnant (7, 14, and 18 dpp) as well as np females. Data are expressed as single dots, and means are indicated. *P < 0.05, **P < 0.01, and ***P < 0.001 as analyzed by one-way ANOVA, followed by a Tukey multiple comparison test.

The Splenic Architecture As Well As Splenic B Cell Populations Are Strongly Modified During Murine Pregnancy

Macroscopically, the spleen during pregnancy depicted a strong splenomegaly, which was clearly evident at 14 dpp (Fig. 2A). However, at late pregnancy stages (18 dpp), the size of the spleen decreased to the size observed in np control mice (Fig. 2A). Deeper histological analysis depicted, in an otherwise regular spleen, a decrease of relative white pulp (WP) tissue volume (ratio of WP to total spleen parenchyma), occurring already at Day 7 of pregnancy and peaking at Day 14 (Supplemental Figure S1). At late pregnancy stages (18 dpp), the ratio increased significantly again compared to the one at Day 14 of pregnancy and was not significantly different from pregnant mice at Day 7 of pregnancy. Nevertheless, the relative WP tissue volume was still significantly lower in comparison with nonpregnancy stages (Supplemental Figure S1). However, when analyzed independently of the total parenchyma, volume of WP did not show significant differences among the groups (data not shown).

Fig. 2

Splenic B cell compartment is strongly modified during pregnancy. A) Spleen weight from np as well as pregnant (7, 14, and 18 dpp) females. B) Correlation analysis between spleen weight (g) and total number of B220⁺ B cells. Splenocytes isolated from np as well as pregnant mice at 7 and 14 dpp were stained for different B cell markers and analyzed by flow cytometry. C) Representative histograms showing B220⁺ splenocytes from np as well as pregnant (7, 14, and 18 dpp) mice. Bar graph depicts the total numbers of B220⁺ cells in the spleen. D) Representative pseudocolor dot plots showing the expression of CD21/35 and CD23 on B220⁺ gated splenocytes from np as well as pregnant (7, 14, and 18 dpp) females. FO cells identified as CD21/35^intCD23^hi and MZ cells identified as CD21/35^hiCD23^low. Bar graphs depict the total numbers of CD21/35^intCD23^hi (B220 gated) FO and CD21/35^hiCD23^low (B220 gated) MZ B cells. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 as analyzed by one-way ANOVA, followed by a Tukey multiple comparison test.

Remarkably, the increase in spleen size observed during pregnancy was negatively correlated with the numbers of total B220⁺ B cells in the spleen of pregnant mice (Fig. 2A). Numbers of B220⁺ B cells were similar in pregnant mice at Day 7 of pregnancy as compared to np mice (33.1 ± 0.6 × 10⁶ vs. 29.5 ± 1.5 × 10⁶ cells; Fig. 2C). However, at Day 14 of pregnancy, when the spleen size was at the highest, the numbers of B220⁺ B cells significantly dropped as compared to np (20.74 ± 0.9 × 10⁶ vs. 29.58 ± 1.5 × 10⁶) and pregnant mice at 7 dpp (20.74 ± 0.9 × 10⁶ vs. 33.16 ± 0.6 × 10⁶; Fig. 2C). At late pregnancy (18 dpp), when the size of the spleen decreased, the numbers of B220⁺ B cells significantly increased as compared to Day 14 of pregnancy (27.27 ± 0.8 × 10⁶ vs. 20.74 ± 0.9 × 10⁶; Fig. 2C). Alongside changes in spleen morphology as well as total splenic B cell numbers during gravidity, profound modifications in splenic B cell populations were also observed. Following egress from the BM, immature B cells traffic to the spleen [10, 16]. Newly arrived immature B cells express high levels of surface sIgM and very little sIgD (sIgM^hisIgD^lo) and undergo maturational progression through a sIgM^hisIgD^hi stage to the most mature sIgM^losIgD^hi stage that is reportedly equivalent to the FO B-cell subset [16]. Pregnant mice showed a significant decrease in the numbers of immature sIgM^hisIgD^lo (B220⁺gated) at Days 14 and 18 of pregnancy as compared to Day 7 of pregnancy and to np control mice (Supplemental Figure S2A). Levels of transitional sIgM^hisIgD^hi and mature sIgM^losIgD^hi B cell stages suffered a slight but not significant increase at Day 7 of pregnancy as compared to np mice (Supplemental Figure S2A). At Day 14 of pregnancy, numbers of both populations significantly decreased as compared to Day 7 of pregnancy (Supplemental Figure S2A). At late pregnancy, numbers of sIgM^losIgD^hi mature but not sIgM^hisIgD^hi transitional B cells increased again up to the levels observed in np and pregnant mice at Day 7 of pregnancy (Supplemental Figure S2A).

According to the surface expression of CD23 and IgM, B220⁺CD93⁺ immature B cells can be further subdivided into three transitional subsets: T1 (sIgM⁺CD23^lo), T2 (sIgM⁺CD23⁺), and T3 (sIgM^loCD23⁺) [6]. We observed that numbers of T1 and T2 B cell subsets significantly increased at Day 7 of pregnancy of pregnancy as compared to np mice and then declined again toward Days 14 and 18 of pregnancy (Supplemental Figure S2B). Unlike T1 and T2, numbers of the sIgM^loCD23⁺ T3 subset remained unchanged at Day 7 of pregnancy as compared to np mice but then decreased (Supplemental Figure S2B). We have additionally analyzed the numbers of B220⁺CD23^hiCD21^hisIgM^hi T2-MZP (transitional 2 marginal zone precursors), which give rise mainly to MZ B cells [17], and found that T2-MZP were slightly but not significantly increased in the spleen of pregnant mice (most evident at Day 7 of pregnancy) as compared to np mice (Supplemental Figure S3).

To further define the subsets of mature splenic B220⁺ cells, we analyzed the expression of CD23 and CD21/35 on B220 gated cells. FO B cells are defined as CD23^hiCD21/35^int and MZ B cells as CD23^loCD21^hi [18]. The numbers of CD23^hiCD21/35^int (B220 gated) FO B cell significantly dropped from 19.4 ± 1.2 × 10⁶ cells in np mice to 13.20 ± 0.6 × 10⁶ cells in pregnant mice at 14 dpp and then increased again toward the end of pregnancy (17.8 ± 0.6 × 10⁶ cells at 18 dpp; Fig. 2D). CD23^loCD21^hi (B220 gated) MZ B cell numbers slightly but not significantly increased during pregnancy as compared to np mice (Fig. 2). However, when using a more specific gating strategy to define MZ B cells [17, 19], a significant increase in the numbers of B220⁺CD23^−/loCD21^hisIgM^hi MZ B cells was observed at late pregnancy as compared to np as well as pregnant mice at Days 7 and 14 of pregnancy (2.4 ± 0.1 × 10⁶ vs. 1.0 ± 0.1 × 10⁶, 1.5 ± 0.1 × 10⁶, and 1.2 ± 0.1 × 10⁶, respectively; Supplemental Figure S4). Correspondingly, the ratio between FO and MZ B cells depicted a clear preponderance of MZ B cells toward FO B cells in the spleen of pregnant mice (Fig. 2D and Supplemental Figure S4).

Pregnant Mice Display Increased Levels of IgM, IgA, and IgG3 Immunoglobulins in Serum

The main function of the B cell compartment is the development and maintenance of a broad repertoire of antibody-producing cells that release different subtypes of immunoglobulin [20]. Even though pregnancy has been classically associated with a shift toward a humoral Th2 type immune response [21], a detailed characterization of the B cell repertoire induced during pregnancy is missing. We observed that levels of IgM were significantly higher in the serum of pregnant mice analyzed at 7 and 14 dpp as compared to np mice (Fig. 3). At late pregnancy (18 dpp), IgM levels were similar to those observed in np animals. Similarly, IgA levels were slightly but not significantly augmented at Day 7 of pregnancy as compared to np mice and decreased at Day 14 of pregnancy. At the end of pregnancy (18 dpp), IgA levels were significantly augmented as compared to np mice. Analysis of different IgGs isotypes depicted a slight but not significant increase of IgG1, IgG2a, and IgG2b at Day 7 of pregnancy as compared to np mice (Fig. 3). As pregnancy advanced, levels of IgG1, IgG2a, and IgG2b significantly dropped (Fig. 3). Unlike other analyzed IgGs, IgG3 levels were significantly augmented in the serum of pregnant mice at Day 7 of pregnancy as compared to np animals (Fig. 3). However, at Days 14 and 18 of pregnancy, serum levels of IgG3 were comparable to those observed in np mice (Fig. 3). Serum levels of IgE did not suffer significant alterations during pregnancy as compared to the ones in np mice (Fig. 3).

Fig. 3

Immunoglobulins levels in serum of pregnant and np mice. Levels of total IgM, IgA, and IgE as well as different IgG isotypes (IgG1, IgG2a, IgG2b, and IgG3) were assayed in serum of np (n = 10) as well as pregnant mice at 7 dpp (n = 11), 14 dpp (n = 11), and 18 dpp (n = 5) by a multiplex system. Data are expressed as box and whisker plots showing median. *P < 0.05, **P < 0.01, and ***P < 0.001 as analyzed by the nonparametric Kruskal-Wallis test.

BAFF Levels Are Decreased in the Serum of Pregnant Mice

Having observed profound changes in the composition of transitional as well as mature (FO and MZ) B cells in the spleen of pregnant mice, we then considered analyzing factors that may account for these modifications. The cytokine BAFF of the TNF family is essential for B cell generation and maintenance in the periphery [14, 15]. Indeed, elevated levels of BAFF result in increased numbers of FO and MZ B cells [22, 23]. In contrast, animals lacking BAFF receptor (BR3) have reduced numbers of mature B cells [14, 15]. We observed that serum levels of BAFF were slightly but not significantly increased at Day 7 of pregnancy as compared to np control mice (Fig. 4). However, levels of BAFF significantly dropped at Day 14 of pregnancy and remained very low toward the end of pregnancy (18 dpp) as compared to np or pregnant mice at Day 7 of pregnancy (Fig. 4).

Fig. 4

Levels of BAFF are diminished in serum of pregnant mice. Bar graph showing the concentration of BAFF (ng/ml) in serum from (np (n = 10) as well as pregnant (7 dpp, n = 10; 14 dpp, n = 10; and 18 dpp, n = 5) mice. Data are presented as mean ± SEM. **P < 0.01 and ***P < 0.001 as analyzed by one-way ANOVA, followed by a Tukey multiple comparison test.

Total B Cell Numbers Are Diminished in Blood During Pregnancy

In agreement with the lower production of B cells from the BM during pregnancy, the analysis of B cells in blood depicted a significant decrease in the total numbers of CD19⁺ B cells in pregnant mice at Days 7 and 14 of pregnancy as compared to np control mice (1.2 ± 0.05 × 10⁶ cells/ml and 0.9 ± 0.04 × 10⁶ cells/ml vs. 1.9 ± 0.1 × 10⁶ cells/ml, respectively; Fig. 5A). At late pregnancy, numbers of CD19⁺ B cells were similar to those observed in np control mice (Fig. 5A). Similarly, numbers of sIgM⁺sIgD^lo, sIgM⁺sIgD⁺, and sIgM^losIgD⁺ B cells were significantly decreased at Days 7 and 14 of pregnancy as compared to np mice (Fig. 5A). At late pregnancy, numbers of sIgM⁺sIgD⁺ and sIgM^losIgD⁺ B cells were comparable to those observed in np mice (Fig. 5B). However, numbers of sIgM⁺sIgD^lo B cells were still significantly lower as compared to np control mice (Fig. 5B).

Fig. 5

Pregnant mice depict a B cell lymphopenia in blood. A) Representative histograms showing CD19⁺ cells on PBMCs from np as well as pregnant (7, 14, and 18 dpp) mice. Bar graph depicts the total numbers of total CD19⁺ B cells per milliliter of blood. B) Representative contour plots depicting the expression of surface IgD and IgM on PBMCs from np as well as pregnant (7, 14, and 18 dpp) mice. Graphs depict the total numbers of sIgD^hisIgM^lo, sIgD^hisIgM^hi, and sIgD^losIgM^hi cells per milliliter of blood. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 as analyzed by one-way ANOVA, followed by a Tukey multiple comparison test.

Immature As Well As Mature B Cell Numbers Are Increased in the Lymph Nodes Draining the Uterus During Pregnancy

It has been previously shown by microscopy using F (ab)₂ staining that the percentages of B cells in murine uterus-draining para-aortic lymph nodes (PLN) were increased during pregnancy [24]. Here we partially confirmed and then extended these results. Unlike Newport and Carter [24], we observed that total B220⁺ B cell numbers significantly decreased in PLN of pregnant mice at Day 7 of pregnancy as compared to np mice (Fig. 6). However, similarly to what had previously been observed [24], the numbers of total B220⁺ B cells were 2.3- and 2-fold higher in PLN of pregnant mice at Days 14 and 18 of pregnancy, respectively, as compared to np mice (Fig. 6A). In addition, and in line with the kinetic of total B220⁺ B cells, the numbers of B220⁺CD93⁺ immature, as well as B220⁺CD93⁻ mature B cells, were significantly higher in pregnant mice analyzed at Days 14 and 18 of pregnancy compared to np mice (Fig. 6B). We further confirmed the augmentation of mature B cells in the PLN of pregnant mice by staining PLN cells with CD21 (Fig. 6C). Despite the similar kinetics displayed by mature and immature B cells, the ratio between both populations revealed a slight tendency favoring the mature phenotype as gravidity progressed (Fig. 6D). Moreover, pregnant mice at late pregnancy (18 dpp) showed a significantly higher mature/immature ratio compared to np mice (Fig. 6D).

Fig. 6

The lymph nodes draining the uterus are enriched in B cells during pregnancy. PLN cells were isolated from np as well as pregnant mice (7, 14, and 18 dpp) and analyzed for different B cell markers. A) Representative histograms depicting the analysis of total B220⁺ B cells in PLN. Bar graph shows the total numbers of B220⁺ B cells from np as well as pregnant (7, 14, and 18 dpp) mice. B) Representative contour plots showing the gating strategy for the analysis of total B220⁺CD93⁺ (immature) and B220⁺CD93⁻ (mature) B cells in PLN. Bar graph depicts the total numbers of B220⁺CD93⁺ and B220⁺CD93⁻ B cells from np as well as pregnant (7, 14, and 18 dpp) mice. C) Representative contour plots showing gating strategy for the analysis of B220⁺CD21⁺ mature B cells in the spleen from np as well as pregnant (7, 14, and 18 dpp) mice. Bar graph depicts total numbers of B220⁺CD21⁺ (mature) cells in PLN. D) Bar graph showing the ratio between mature and immature B cells in PLN. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 as analyzed by one-way ANOVA, followed by a Tukey multiple comparison test.

B1 and B2 B Cell Numbers Are Increased in the Peritoneal Cavity During Pregnancy

Due to the close proximity to the fetuses, we next concentrated on analyzing the B cell compartment in the peritoneal cavity (PerC) of pregnant mice and compared it to np control females. As shown in Figure 7A, total numbers of CD19⁺ B cells significantly increased already at Day 7 of pregnancy and remained very high at Days 14 and 18 of pregnancy as compared to np control mice (9.6 ± 0.4 × 10⁶ cells [7 dpp], 10.1 ± 0.3 × 10⁶ cells [14 dpp], and 8.1 ± 0.1 × 10⁶ cells [18 dpp] vs. 4.9 ± 0.3 × 10⁶ cells [np]). The augmentation in the total numbers of B cells in PerC during pregnancy was accompanied by a significant increase in the numbers of CD19⁺CD23⁺ B2 B cells at Days 7, 14, and 18 of pregnancy compared to np mice (Fig. 7B). Similarly, CD19⁺CD23⁻ B1 B cell numbers also increased at Days 7 and 14 of pregnancy but dropped toward the end of pregnancy (18 dpp; Fig. 7B). A deeper analysis of the B1 B cell compartment depicted a slight but not significant increase of CD19⁺CD23⁻CD11b⁺CD5⁻ B-1b B at Day 7 of pregnancy that reached a significant peak at Day 14 of pregnancy before decreasing again at Day 18 of pregnancy (Fig. 7C). Unlike B-1b, CD19⁺CD23⁻CD11b⁺CD5⁺ B-1a B cell numbers did not show major alterations during pregnancy.

Fig. 7

Peritoneal cavity of pregnant mice harbors elevated numbers of B1 and B2 B cells. A) Representative histograms showing the analysis of total CD19⁺ cells on peritoneal lymphocytes from np as well as pregnant (7, 14, and 18 dpp) mice. Bar graph depicts the total numbers of CD19⁺ B cells. B) Representative contour plots depicting the gating strategy used for the analysis of CD19⁺CD23⁺ B2 and CD19⁺CD23⁻ B1 B cells from np and pregnant (7, 14, and 18 dpp) mice. Bar graphs show the total number of B2 and B1 B cells. C) Representative contour plots depicting the gating strategy used for the analysis of CD11b⁺CD5⁺ (gated on CD19⁺CD23⁻ cells) B-1a and CD11b⁺CD5⁻ (gated on CD19⁺CD23⁻ cells) B-1b B cells from np and pregnant (7, 14, and 18 dpp) mice. Bar graphs show the total numbers of B-1a and B-1b b cells. Data are presented as mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001 as analyzed by one-way ANOVA, followed by a Tukey multiple comparison test.

Discussion

The present study provides clear evidence that B cell development undergoes profound adaptations during murine pregnancy. These adaptations begin early in B cell development as pregnant mice showed a gradual reduction in the numbers of pre/pro and, to a minor extent, immature B cells in their BM. Correspondingly, it has been previously demonstrated that murine pregnancy is associated with a suppression of B lymphopoiesis [25]. However, in their work, Medina et al. have pooled pregnant animals from Day 6.5 to Day 19.5 of pregnancy, which hinders the identification of the exact pregnancy time when the changes on B lymphopoiesis occur. Indeed, unlike Medina et al., we showed that B lymphopoiesis is not affected at Day 7 of pregnancy. Nevertheless, as pregnancy advanced, a clear reduction in the numbers of pre/pro and immature B cells was noted. Although we did not fully investigate the mechanisms driving B lymphopenia during pregnancy, pioneer works performed by Kincade's laboratory have pointed to the female sex hormone estradiol as one of the main forces [26–28]. Bosco et al. have extended these results and showed that the estradiol effect on B lymphopoiesis seems to be mediated by an extrinsic limitation in IL-7 availability in the BM microenvironment rather than by the direct effect of estradiol on B cells [29]. Interestingly, the authors have shown a dose-dependent effect of estradiol on IL-7 production by BM stromal cell lines [29]. This would explain why we did not observe differences in numbers of B cell precursors at Day 7 of pregnancy, when levels of estradiol are still low and comparable to those in np mice [30]. Remarkably, at Days 14 and 18 of pregnancy, when levels of estradiol significantly rise, a clear reduction of B cell precursor was observed. In addition to hormonal influence on B lymphopoiesis, antigen-induced deletion of immature B cells was also documented in the second half of pregnancy in an MHCI transgenic mouse model [31].

In contrast to immature B cells, numbers of mature B cells significantly increased in the BM of pregnant mice, suggesting that despite a reduction in B lymphopoiesis, B cell maturation is induced during gravidity. Accordingly, higher numbers of mature B cells were observed in the lymph node draining the uterus as well as in the peritoneal cavity of pregnant mice. This prompted us to further investigate B cell development in the periphery. Following egress from the BM, immature B cells traffic to the spleen, where they continue their development [10, 16]. Newly arrived immature B cells undergo a maturational process involving different transitional stages and eventually give rise to either mature FO or MZ B cells [17]. It has been shown that in case of the absence or reduction of B cell influx from the BM, the splenic MZ population, which possesses a preactivated phenotype, is preferentially maintained [32, 33]. It is believed that this phenomenon is triggered as a compensatory mechanism that ensures the development of the effector branch of the B cell compartment, thus maximizing the capacity for defense [34]. Interestingly, we observed a gradual reduction of B lymphopoiesis in the BM during pregnancy that was mirrored by lower numbers of newly emigrated as well as FO B cells in the spleen. Notably, MZ B cell numbers were either unaltered or increased in pregnant mice. More important, when the MZ/FO B cell ratio was analyzed, a clear preponderance of MZ B cells was observed, undoubtedly indicating a bias toward MZ B cell development during gravidity. After encountering an antigen, FO B cells produce T-dependent high-affinity immunoglobulin and memory B cells [6]. This process lasts 5–7 days, a time that can be fatal in the case of blood-borne pathogens with short period of replication [12] and even worse in the case of an immunocompromised pregnant woman. MZ B cells compensate this gap by activating a faster T-independent pathway that generates short-life plasmablasts that secrete low-affinity IgM but also IgA and IgG3 [12]. Hence, it is reasonable to argue that expansion of the MZ B cell compartment may represent a mechanism to compensate for the B cell lymphopenia observed during pregnancy, thus maximizing the capacity of the immune system to protect the mother. Our findings concerning the transient increase of IgM, IgA, and IgG3 immunoglobulins in serum reinforce this idea.

In an effort to identify a factor responsible for the variations in the composition of splenic B cells during pregnancy, we next concentrated on the TNF family ligand BAFF/Blys, which is an essential factor for B cell development [14, 15, 35]. Unexpectedly, we observed a significant reduction in the levels of BAFF in the serum of pregnant mice. Although MZ B cells are markedly reduced in numbers in BAFF^−/− mice, it has been proposed that this could be explained by the diminished survival of all follicular precursors and not by a direct effect of BAFF on MZ B cells [17]. Indeed, BAFF is an essential survival factor for transitional B cells. Apart from T1 B cells, BAFF receptor is expressed in all other splenic B cell populations [13]. In agreement with this, we found that T2 B cells were slightly and T3 significantly reduced in the spleen of pregnant mice. However, numbers of T2-MZP cell, which give raise mainly to MZ B cells [17], were either slightly increased or not modified during pregnancy.

It is also important to note that BAFF can support the survival and proliferation of autoreactive B cells, which have higher BAFF dependence [36]. Indeed, autoimmunity is often associated with elevated levels of BAFF [37]. Thus, the strong reduction in systemic BAFF levels in synergy with B cell lymphopenia observed during pregnancy may represent an acquired protective mechanism aimed at reducing the occurrence of autoreactive B cells. In fact, it has been demonstrated that elevated levels of BAFF in the presence of B cell lymphopenia may contribute to increased B cell response, leading to autoimmunity [38]. Remarkably, it is very well known that women suffering from systemic lupus erythematous, an autoimmune disease highly associated with elevated levels of BAFF, experience a worsening of their symptoms during pregnancy [3, 4]. Future works should address the link between physiological lymphopenia induced during gravidity and high levels of BAFF in pregnant women suffering from systemic lupus erythematous.

Overall, we demonstrated in this work that pregnancy was associated with a gradual reduction in the production of B cells from the BM as well as serum levels of BAFF, possibly representing a physiological adaptation to avoid the occurrence of autoreactive B cells. This affected the peripheral B cell compartment, as B lymphopenia in blood as well as in the spleen was observed. However, an expansion of a mature and preactivated MZ B cell compartment was depicted in pregnant mice, most likely representing a compensatory mechanisms tending to maximize the capacity of the maternal immune system to fight pathogens.

This novel piece of information helps us better understand how the process of pregnancy tolerance occurs and interpret why pregnancy may increase susceptibility to infections and worsen the symptoms of some autoimmune diseases.

Acknowledgment

We kindly thank Denice De Carlo for her helpful English language assistance and Oliver Nicolai for his technical support in with the Milliplex Map procedure and measurements.

1

Supported by grants from the Fritz Thyssen Foundation to F.J. (Az. 10.12.2.155) and intramural founding from Greifswald University.