Regulatory B-Cells Are Associated Negatively With Regulatory T-Cells and Positively With Cytokines in Peripheral Blood of Pregnant Women.

The analytic cohort in this study comprised pregnant women ( N = 29) who were, on average, 32 years old, had high school level of education (52%), were Mexico-born (55%), unmarried (52%), cohabitating with the baby’s father (59%), with 1.5 previous children, and 3.5 pregnancies inclusive of the focal pregnancy ( Table 1 ). Of the participants, N = 12 were primigravid and N = 17 were multigravid. The gestational ages at sample collection ranged from 5.1 to 15.7 weeks with an average of 12.1 weeks’ gestation ( Table 1 ). Flow cytometry analysis revealed that the median percentage of CD19 + B-cells in circulation of pregnant women was 11.9% (range: 4.37%–32.7%) ( Table S3 ). CD38 is upregulated with activation of B-cells and thus, we evaluated the percentage of CD38 + B-cells in circulation ( Figure 1A ). The median percent of total B-cells that expressed CD38 was 6.37% (range: 2.17%–15.8%) ( Table S3 ). We then quantified PD-L1 + B-cells, Bregs, and PD-L1 + Bregs by flow cytometry. A representative gating strategy for the measurement of Breg cells and PD-L1-expressing Bregs is shown in Figure S1 . The median percent of total B-cells that expressed PD-L1 was 3.43% (range: 1.45%–6.46%). We sub-gated on CD24 hi and CD38 hi B-cells to determine the percent of Bregs in peripheral blood circulation. Figure 1B shows the relative distribution in the percentage of CD19 + B-cells with Breg cell phenotype (CD24 hi CD38 hi ) (median: 7.55%, range: 1.90%–16.7%), and the percentage PD-L1 + Bregs (median: 4.41%; range: 0.78%–14.1%) ( Table S3 ). Moreover, we quantified T-cell and specific Treg cell subsets by flow cytometry. The median percentage of CD3 + T-cells from total viable cells analyzed was 48.1% (range: 21.9%–64.6%) ( Figure 1C and Table S3 ). The distribution of percent FoxP3 + cells within the CD3 + CD4 + CD25 hi CD127 lo population is shown in Figure 1D , which had a median percent value of 81.6% (range: 33.1%–93.2%). A representative gating strategy for the analysis of FoxP3 + Tregs and specific FoxP3 + Treg cell subsets expressing co-signaling molecules (Helios + , CD45RA + , CTLA-4 + , PD-1 + , TIGIT + , and TIM-3 + ) is shown in Figure S2 . The median composite score values for Th1, Th2, and Th17 cytokines were 1.02, 1.15, and 0.525, respectively ( Table S3 ). We first tested for associations between levels of Breg or PD-L1 + Bregs and cytokines, adjusting for gestational age at blood draw ( Figure 2 and Table S4 ). The overall Breg levels were positively associated with IL-6 (Beta = 2.563, CI [0.291, 4.835], p = 0.029) ( Table S4 ). PD-L1 + Breg levels were positively associated with IL-10 levels (Beta = 8.526, CI [1.625, 15.427], p = 0.017) ( Table S4 ). We also detected a positive trend between levels of PD-L1 + Bregs and IL-8 (Beta = 0.700, CI [0.015, 1.415], p = 0.054), although this fell above conventional statistical significance thresholds. Neither levels of Bregs or PD-L1 + Bregs exhibited significant associations with Th1, Th2, or Th17 cytokine composite scores ( Table S4 ). We then stratified the cohort by gravidity, with separate models for primigravid and multigravid pregnant women. Among primigravid women, results were generally similar to what we observed for the full cohort ( Table S5 ). Among multigravid women, the only statistically significant association was between levels of PD-L1 + Bregs and IL-8 ( Table S5 ). We observed no significant associations between Th1, Th2, or Th17 composite scores and Bregs or PD-L1 + Bregs of primigravid or multigravid pregnant women ( Table S5 ). To address Aim 2, we tested for associations between levels of Breg and PD-L1 + Bregs and FoxP3 + Tregs, and several subpopulations of Tregs, after adjusting for gestational age at sample collection ( Figure 3 and Table S6 ). Levels of the parent Breg population were not associated with Tregs or any of the subpopulations. Levels of PD-L1 + Bregs were negatively associated with the parent Treg population (Beta = −0.136, CI: −0.246 to −0.025, p value = 0.018), as well as Tregs expressing Helios (Beta = −0.160, CI: −0.245 to −0.076, p value = 0.001), CTLA-4 (Beta = −0.192, CI: −0.325 to −0.060, p value = 0.006), PD-1 (Beta = −0.156, CI: −0.271 to −0.040, p value = 0.010), TIGIT (B = −0.249, CI: −0.353 to −0.144, p value < 0.001), and TIM-3 (Beta = −0.679, CI: −1.301 to −0.058, p value = 0.033) ( Table S6 ). Figure 4 shows the correlation plots for significant regressions between the percentage of Bregs and PD-L1 + Bregs and cytokine levels (IL-6 and IL-10) and the percentage of Tregs and Treg subsets. We then conducted the analysis stratified by gravidity. The results were generally consistent with the analysis conducted on the whole cohort. Among both primigravid and multigravid pregnant women, only PD-L1 + Breg levels were associated with Treg levels ( Table S7 ). In primigravid women, levels of PD-L1 + Bregs were negatively associated with levels of Helios + Tregs (Beta = −0.210, CI: −0.376 to −0.044, p value = 0.019) and TIGIT + Tregs (Beta = −0.232, CI: −0.461 to −0.002, p value = 0.048) ( Table S7 ). For multigravid women, levels of PD-L1 + Bregs were negatively associated with Helios + Tregs (Beta = −0.136, CI: −0.250 to −0.022, p value = 0.023), CTLA-4 + Tregs (Beta = −0.223, CI: −0.434 to −0.013, p value = 0.039), PD-1 + Tregs (Beta = −0.240, CI: −0.407 to −0.073, p value = 0.008), and TIGIT + Tregs (Beta = −0.272, CI: −0.402 to −0.142, p value = 0.001) ( Table S7 ).

Participants for this study were recruited into Wave 2 of the Mothers’ Cultural Experiences (MCE) study, an NIH-funded study of how sociocultural stress and resilience relate to maternal-child psychobiology and development. The sociocultural and psychological assessments are not relevant for this study. Eligibility rules required women to be ≤ 16 weeks’ gestation at enrollment, age 18 or older, English or Spanish speaking, and self-identified as Latina, Hispanic, Chicana, Mexican, or Latin American for the purpose of other research questions related to cultural background and experiences. Although MCE Wave 2 is a longitudinal study, data for this manuscript were only derived from the first timepoint of MCE Wave 2; thus, these analyses are cross-sectional. First timepoint data collection (Timepoint 1) occurred from December 11, 2018, to March 3, 2020. Participants’ gestational ages ranged from 5.1 to 15.7 weeks (mean: 12.1 weeks). Informed, written consent was obtained from participants. This study was approved by the Institutional Review Boards of all participating institutions. The study adheres to the tenets of the Declaration of Helsinki. Treg and Breg cells and cytokine levels were measured from non-fasting morning blood samples drawn by antecubital venipuncture during morning prenatal visits by hospital or clinic phlebotomists. Isolated peripheral blood mononuclear cells (PBMCs) were used for Breg and Treg cell quantification by flow cytometry. Blood used to isolate PBMCs was drawn into sodium heparin vacutainers prior to other labs. Specimens were transported at room temperature to the UCLA Cousins Center Inflammatory Biology Core laboratory where PBMCs were isolated following standard procedures and frozen at −80°C. Plasma used for cytokine assays was extracted from blood collected in ethylene-diaminetetraacetic acid-treated tubes and kept at refrigerator temperature until plasma extraction within a few hours. Plasma aliquots were frozen at −80°C until cytokine measurement assay. The viability of PBMCs was assessed using a Zombie Aqua Fixable Viability Kit. For identification of Breg and PD-L1-expressing Breg cell populations, PBMCs were stained using the following panel of anti-human antibodies: CD14-Brilliant Violet 711, CD19-PerCP, CD24-PE, CD38-Brilliant Violet 605, and PD-L1-APC (BioLegend, San Diego, CA, USA). We included PD-L1 in this panel as it has been shown to modify the immunosuppressive capacity of B-cells [ 37 ]. For the identification and evaluation of Treg cells, a separate set of PBMCs was stained using the following panel of anti-human antibodies: CD3-Brilliant Violet 650, CD4-APC-Cyanine7 (APC-Cy7), CD8-Alexa Fluor 700, CD25-Brilliant Violet 421, CD127-Brilliant Violet 605, CD45RA-Brilliant Violet 785, CTLA-4-APC, FoxP3-PE, Helios-Alexa Fluor 488, PD-1-Brilliant Violet-711, TIM-3-PE/Dazzle CD366, and TIGIT-PE-Cy7 (BioLegend, San Diego, CA, USA). Stained cells were then analyzed by flow cytometry using a 4 laser AttuneNxT Accoustic Focusing cytometer (Invitrogen) with AttuneNxT software at the Janis V. Giorgi Flow Cytometry Core Laboratory at UCLA. Data were analyzed using the FlowJo software package (Tree Star, Ashland, OR, USA). We identified the positivity borderline from fluorescence minus one (FMO) control tubes. The gating strategies for Bregs and Tregs are shown in Figures S1 and S2 , respectively. Meso Scale Discovery V-PLEX multiplex assays were used to measure the concentrations of 13 cytokines in plasma of pregnant women: pro-inflammatory (IL-6, IL-1 β , TNF- α , IL-8, IL-17, IL-21, and IFN- γ ) and anti-inflammatory and/or immunoregulatory cytokines (IL-10, IL-12, IL-13, and IL-22) and cytokines mediating adaptive immunity (IL-2 and IL-4) (V-PLEX Pro-inflammatory Panel 1 and Th17 Panel 1 from Meso Scale Discovery, Meso Scale Diagnostics LLC, Rockville, MD). The average lower limit of detection, intra-, and inter-assay coefficients of variance (%CV) for each of the cytokines are presented in Table S1 . The average overall intra- and inter-assay %CV for the pro-inflammatory multiplex assay were 17.52% and 7.82%, and for the Th17 multiplex were 38.60% and 7.13%, respectively. For each individual cytokine, intra-assay %CV was considered acceptable if < 10 but higher values were considered acceptable if mean concentration was 10%. For multiple cytokines (IL-1 β , IL-4, IL-12, IL-13, and IL-17), high %CVs were driven by low or undetectable concentrations ( Table S1 ). Inter-assay %CV was considered acceptable if < 15, which all of them were. To limit the number of comparisons, we collapsed the 13 cytokines into 3 composite scores reflecting Th1, Th2, and Th17 type inflammation, and used these as dependent variables. We evaluated IL-10 independently in addition to these composite scores because of its direct relevance to Bregs. IL-6 and IL-8 were also used as dependent variables because they do not fit into only one of the Th1, Th2, or Th17 categories. IL-6 induces both Th2 and Th17 differentiation and IL-8 is a chemokine/chemoattractant molecule secreted by cells of the immune system (primarily macrophages), epithelial cells, and endothelial cells. To calculate composites that contained cytokines with high rates of undetectable values, we substituted the lowest limit of detection for participants with undetectable values. See Table S1 for further details. Cytokine composite scores were computed as the sum of the standardized value ( z -score) of each cytokine. The Th1 composite included IL-1 β , IL-2, IL-12, IFN- γ , and TNF- α ; the Th2 composite included IL-4, IL-10, and IL13; and the Th17 composite included IL-17, IL-21, and IL-22. Additionally, post-hoc, models were repeated using an alternative operationalization of the cytokine composite score, calculated as the raw sums of the unstandardized cytokine levels to confirm that results are not a spurious finding due to idiosyncratic operationalization technique ( Table S2 ). Multiple linear regression was used to estimate (1) associations between Breg cells and cytokines and (2) associations between Breg and Treg cells. Models were adjusted for gestational age, in weeks, at sample collection. p < 0.05 was considered statistically significant. We did not correct for multiple testing because this was an exploratory pilot study with a small sample size. All statistical analyses were conducted using R, version 4.2.2. To examine the relationships of interest in the cohort subsetted by gravidity, we defined primigravida as women who were pregnant for the first time based on self-report. Previous pregnancies could include childbearing, abortion (spontaneous or elective), or stillbirth. Additionally, five participants had blood draws and successful flow cytometry to measure both Bregs and Tregs at the second MCE assessment, which occurred at 23.12 weeks of gestation (range = 21.4–26.1 weeks). Although this sample size is not large enough for conventional statistical testing, intraindividual changes were documented from the first to the second assessment time points in Figure S3 to provide information for future studies.

Our study investigates the associations between levels of Bregs and PD-L1 + Bregs with cytokines and Tregs in peripheral blood circulation of pregnant women. In multiple regression models, adjusted for gestational age at sample collection, we found a positive association between PD-L1 + Bregs and IL-10, which is consistent with the fact that Bregs secrete IL-10. In addition, we observed a positive association between levels of Bregs and IL-6. Other studies have shown that IL-6 can induce IL-10 expressing Bregs [ 38 ]. Recent studies have also shown that IL-10 + B-cells often secrete pro-inflammatory cytokines like IL-6 and TNF- α and that the generation of IL-10-producing B-cells often depends on cytokine levels in the environment and the activated state of cells [ 39 , 40 ]. Moreover, levels of PD-L1 + Bregs were negatively associated with FoxP3 + Tregs and various subsets of Tregs: PD-1 + , CTLA-4 + , Helios + , TIGIT + , and TIM-3 + Tregs. These markers are known to modify the immunosuppressive function of Treg cells and each of these subsets has been implicated in maternal tolerance of pregnancy in humans or murine models [ 41 – 43 ]. No significant associations between the parent population of Bregs and these Treg subsets were found. Collectively, these results suggest an inverse relationship between levels of PD-L1-expressing Bregs and Tregs. Bregs can suppress inflammation through elevated expression of PD-L1, which can engage with PD-1 expressed on T-cells, including Tregs, and exert negative signals to mediate T-cell suppression [ 15 , 34 ]. Our results may capture a temporal Breg-dominated stage of immunotolerance of pregnancy. It is possible that high levels of PD-L1 expressing Bregs in early pregnancy may facilitate the expansion of Tregs and a shift toward a Treg-dominated state of immunotolerance at later gestational weeks. It is also possible that different associations may be observed in decidual or uterine tissues, where FoxP3-expressing T-cells have been shown to colocalize with B-cells, potentially influencing the recruitment of Tregs in these tissues [ 44 ]. For models stratified by gravidity, we found positive associations of levels of Bregs with IL-6 and PD-L1 + Bregs with IL-10 among primigravid pregnant women. These findings suggest potentially unique cytokine activities associated with PD-L1 + Bregs in women experiencing a first pregnancy. Future studies can address how adaptive immunity differs between primigravid and multigravid women. In addition, we observed a positive association between levels of PD-L1 + Bregs and IL-8 in multigravid pregnant women. Although it did not reach the statistical significance threshold, we observed a similar trend among the full cohort, and primigravid pregnant women exhibited a non-significant positive association between Bregs and IL-8. The association of Bregs with enhanced IL-8 levels was counter to our prediction of lower pro-inflammatory cytokines. This may be due to the role of normative levels of IL-8 in healthy pregnancy. IL-8 is secreted by decidual stromal cells, the glandular epithelium, and macrophages, and plays a role in regulating the timing of parturition at term [ 45 – 47 ]. Given the crucial role of IL-8 in pregnancy, it is possible that Bregs may selectively permit its production. However, high levels of IL-8 are associated with obstetric pathologies, including pre-eclampsia [ 48 ] and endometriosis [ 49 ]. The stronger relation between levels of PD-L1 + Bregs and IL-8 among the multigravid subset of pregnant women could suggest that Bregs are recruited more vigorously in subsequent pregnancies that are prone to obstetric complications. Further research in a larger, high-risk cohort would be necessary to examine this possibility. Moreover, we observed significant negative associations between levels of PD-L1 + Bregs and FoxP3 + Tregs and several subpopulations of Tregs that express immune checkpoint markers, including Helios, CTLA-4, PD-1, TIGIT, and TIM-3. These relationships were consistent across nearly all Treg subtypes in the full cohort and the multigravid subset of cohort. Specifically, we found that levels of PD-L1 + Bregs were only negatively associated with Helios + and TIGIT + Treg cells in primigravid pregnant women, while PD-L1 + Bregs were negatively associated with Helios + , CTLA-4 + , PD-1 + , and TIGIT + Tregs in multigravid pregnant women. Collectively, these results suggest an inverse relationship between levels of PD-L1 + Bregs and Tregs in this cohort of pregnant women, which is largely irrespective of gravidity. Prior to statistical analysis, we did not predict a directionality of correlation because either a positive or negative correlation could have been predicted based on different justifications. The inverse correlation we observed is consistent with the possibility that only Bregs or Tregs need to be predominant to exert suppressive function in pregnancy. We speculate that, perhaps, if both were high simultaneously, it could exert too much immunosuppression. It is also possible that Bregs are important for the maintenance of maternal-fetal tolerance in early pregnancy. While longitudinal studies of Bregs beginning in the first trimester are lacking, pregnant women in the third trimester had lower levels of circulating Bregs compared to nonpregnant individuals [ 23 ]. This is consistent with longitudinal studies of the whole blood transcriptome, which show declines in B-cell mRNA across pregnancy [ 50 ]. In contrast, Treg levels have been shown to peak in the third trimester and into the early postpartum [ 35 , 51 ] Thus, our results may capture a temporal stage of pregnancy characterized by Breg-dominated immunosuppression. It is also possible that associations in the periphery may not be reflective of associations in decidual tissues. Tregs in the periphery have also been shown to be preferentially recruited to the decidua [ 52 ], which may underly some of these associations. To our knowledge, the only study documenting correlations between Bregs and Tregs in pregnancy reported a positive association between CD24 hi CD38 hi Bregs in blood samples collected in the third trimester [ 53 ]. Future studies are needed to elucidate the interaction between Breg and Treg recruitment and function in the context of pregnancy in longitudinal studies to assess changes across gestation. Pregnancy-associated hormones have also been suggested to modify B-cell function, therefore changes in these hormones may be related to our results. For example, estradiol, progesterone, and human chorionic gonadotropin (hCG) are thought to drive the expansion of IL-10-producing B-cells or IL-10-producing Bregs during pregnancy [ 54 ]. A study by Muzzio et al. found that estrogen exerts immune inhibition in murine pregnancy by inducing Breg cell maturation in gravid uterus [ 55 ]. In human pregnancy, hCG has been shown to promote IL-10 secretion of B-cells and promote the regulatory function of Bregs [ 7 , 9 ]. Moreover, the transition from Breg- to Treg-dominated maintenance of pregnancy tolerance may occur as hCG levels decline beginning around the 10th week of gestation as Bregs are recruited in pregnancy in an hCG-dependent mechanism [ 7 , 9 ]. Additional evidence suggests that Bregs are lowest in the third trimester such that circulating Breg levels are lower than those in non-pregnant and postpartum women [ 23 ]. Increase in the levels of progesterone during pregnancy may also promote a shift toward a Treg-dominated maintenance state. Progesterone may exert immunomodulatory effects on decidual DCs and uterine NK cells, keeping them in an immature state that may facilitate the expansion of Tregs [ 56 ]. Estrogen receptor positive B-cells may also contribute to this transition as they perform Breg-like functions, including the stimulation of Treg production and function [ 57 ]. Our results may reflect a Breg-dominated immunoregulatory state, and Treg levels may increase later in pregnancy as hCG levels decline and estrogen and progesterone levels increase. Thus, future studies should investigate associations for Breg cells with levels of pregnancy-associated hormones across different timepoints of gestation. Our observation of a more consistent negative relationship between PD-L1 + Bregs and Tregs among multigravid than primigravid pregnant women could reflect a scenario in which Tregs from a previous pregnancy inhibit the differentiation of Bregs in a subsequent pregnancy. This speculation is based on evidence that Treg levels are not only sustained but may continue to rise to 12 months postpartum [ 35 ]. Humoral immunity may play a more central role in primigravid pregnancy, whereas the elevation of Tregs postpartum may be recruited for future pregnancies. More longitudinal data would be needed to clarify this possibility. Our study contributes to efforts to characterize the maternal immune system during human pregnancy. It is necessary to understand how maternal-fetal tolerance is maintained during normal pregnancy and in the context of infection, immune intolerance, autoimmunity, and in the face of maternal health complications. We also contribute to the literature examining cell-surface molecules on B-cells or Bregs, such as PD-L1 [ 12 ]. Our results point to potential tradeoffs between Breg and Tregmediated immunity that might be different between primigravid and multigravid women. Future studies are needed to investigate the orchestrated response of Bregs and their crosstalk with other cells during pregnancy. Additionally, this study was conducted with a specific population of self-identified Latina, Hispanic, Chicana, Mexican, or Latin-American women. While race/ethnicity are social rather than biological categories, often biology can be influenced by our environments in ways that follow social categories [ 58 ]. Research focused on this population is important as this group is experiencing the largest temporal increase in the prevalence of pregnancy complications, such as hypertensive disorders of pregnancy, chronic hypertension, and diabetes mellitus, compared to other racial and ethnic groups in the United States [ 59 ]. Their results are likely due to increasing stressors specific to Latinos in the United States [ 60 ]. Future research should be conducted with other diverse cohorts. For example, in the United States, Black women have the highest rates of maternal morbidity and mortality out of all racial and ethnic groups [ 59 , 61 ], justifying the need for studies of gestational health and immune dynamics to understand systemic health disparities. This exploratory pilot study is limited by a small sample size. Future studies in larger cohorts with statistical correction for multiple testing are needed to confirm the replicability of our results. Also, we were unable to investigate the role of Bregs in pregnancy complications. Additionally, our sampling period had a relatively large window, ranging from 5.1 to 15.7 weeks of gestation. While we also collected blood samples from 21.4 to 26.1 weeks of gestation, we only had sufficient aliquots of viable cells from five participants and were not able to include these later samples in our analyses. Furthermore, this study focused on analyzing Bregs and CD4 + CD25 hi CD127 lo FoxP3 + Tregs but other subsets of Tregs have been described. CD4 + CD25 − FoxP3 − and CD8 + CD25 − FoxP3 − Tregs characterized by surface expression of HLA-G were first identified in the thymus and exhibit suppressive and regulatory properties [ 62 ]. These HLA-G + Treg cell subsets can suppress immune responses through secretion of IL-10, and they have been found to be elevated in pregnant women, but lower frequencies have been observed in peripheral blood of pre-eclamptic women [ 63 , 64 ]. Future studies are needed to investigate associations for Bregs with these Treg cell subsets and other Th2-like Treg cells that can secrete IL-17 and IL-4 [ 65 ] and that may be important for establishing tolerogenic responses in pregnancy. Finally, our study used biomarkers in peripheral blood circulation and did not access the functional capacity of isolated cells or cells from the maternal-fetal matrix that might be more relevant to pregnancy tolerance.

We found that circulating levels of PD-L1 + Bregs were associated with several cytokines and Treg cell populations in pregnant women. Levels of these PD-L1-expressing Bregs were negatively correlated with Tregs, suggesting that immunoregulatory capacity in pregnancy may be maintained through a balance of regulatory B- and T-cells. We also found differences between primigravid and multigravid pregnancies, where PD-L1 + Bregs levels were positively associated with IL-10 in primigravid and with IL-8 in multigravid women. Among both primigravida and multigravida, levels of Bregs were negatively associated with Tregs and specific Treg subsets. Further investigation into the interrelationships between distinct Breg and Treg subpopulations may help elucidate the mechanisms underlying immunological tolerance and control in pregnancy, and in women with pregnancy complications.

During pregnancy, the maternal immune system tolerates the semi-allogeneic fetus through a balance of regulatory and inflammatory immune responses. Regulatory T-cells (Tregs) play an important role in maintaining immunotolerance during gestation [ 1 , 2 ]. Healthy pregnancy is characterized by increased Tregs both in the uterine microenvironment as well as in maternal circulation [ 3 ]. Maternal peripheral blood comes into contact with fetal syncytiotrophoblasts, necessitating tolerogenic adaptation to sustain the pregnancy [ 4 ]. Typically in pregnancy, total effector Tregs and clonal effector Tregs are elevated at early gestation, and clonal effector Tregs markedly increase at late gestation [ 5 , 6 ]. However, less is known about how B-cells contribute to healthy and successful pregnancies. B-cells play a vital role in immunotolerance by preventing autoimmunity and the generation of autoantibodies. During pregnancy, B-cells maintain immune balance by exerting humoral activity and the generation of asymmetric antibodies against paternal antigens [ 7 – 9 ]. Apart from its humoral role, B-cells possess immunoregulatory functions. Regulatory B-cells (Bregs) are crucially involved in regulating the adaptive immune response and exert effector functions that include the secretion of anti-inflammatory cytokines [ 10 ]. Bregs were first characterized with the phenotype of CD19 + CD24 hi CD38 hi in peripheral blood [ 11 ]. Bregs comprise less than 10% of the total B-cell population in peripheral blood circulation, and one of their main functional characterizations is the capability to produce interleukin (IL)-10, IL-35, and TGF- β to reduce inflammation [ 10 , 12 – 14 ]. This suppressive function can occur in an IL-10 dependent and IL-10 independent manner [ 15 – 18 ]. In addition, Bregs maintain homeostasis by impairing antigen presenting cells and inducing the expansion and differentiation of Treg cells through secretion of IL-10 and TGF- β and by preventing the maturation of CD11c + CD80 + dendric cells (DCs), as shown in the spleens of abortion-prone mice [ 19 ]. The maintenance of immature DCs promotes the expansion of peripheral Tregs in mice [ 20 ]. There is also evidence suggesting that Bregs can exert safeguarding functions by protecting the mother from neonatal sepsis [ 21 ]. While well-studied in other contexts, such as cancer and autoimmunity, the literature on the role and function of Bregs in pregnancy is relatively small. There is evidence suggesting that Bregs promote immunological tolerance, which was first observed in mice [ 19 ]. Bregs have been shown to be elevated in the peripheral blood of women in the first trimester of pregnancy [ 9 ], as well as in first trimester decidual tissues [ 22 ], and decline in the third trimester and at delivery [ 23 ]. Levels in decidual tissue in early pregnancy are higher than in endometrial tissue from non-pregnant individuals and are positively correlated with the number of decidual CD123 − CD11c + myeloid dendritic cells (mDCs) and suppressive natural killer (NK) cells, suggesting that decidual Bregs may play a role in promoting a microenvironment that favors the establishment and maintenance of pregnancy [ 22 ]. The interactions between maternal-fetal components may educate naïve B-cells to become Bregs to allow the growth of the fetus [ 14 ]. It is thought that Bregs have the capability to restore fetal tolerance in mice with immune-mediated pregnancy complications [ 19 ]. Additionally, reduced levels and/or dysfunction of circulating Bregs may indicate poor pregnancy outcomes [ 24 ]. One study found that peripheral Breg levels were significantly lower in preterm delivery, while Treg levels were not different between term and preterm deliveries [ 25 ]. Moreover, significant inverse associations between Bregs and levels of pro-inflammatory cytokines were found in serum of women who delivered preterm, and elevated levels of maternal B-cells and the IL-6 cytokine were observed prior to those premature deliveries [ 4 , 25 ]. Further evidence justifies investigation of Bregs and IL-10 producing B-cells in pregnancy. Women with recurrent pregnancy loss (RPL) have lower proportions of CD19 + IL-10 + B-cells and reduced levels of IL-10 in circulation [ 26 ]. In addition, mRNA levels of IL-10 and PD-L1 were significantly lower in those RPL patients [ 26 ]. Others have shown that the expression of programmed death-ligand 1 (PD-L1) on peripheral and decidual immune cell populations plays an important role in the maintenance of pregnancy [ 27 ]. PD-L1 is an immune checkpoint molecule that binds to its receptor programmed cell death-1 (PD-1) to inhibit T-cell activation. PD-1 regulates immune responses during acute infection, autoimmunity, and cancer, and the interaction of PD-1/PD-L1 forms a negative costimulatory pathway to balance the immune system [ 28 ]. PD-L1 is expressed on the surface of T-cells, NKT-like cells, B-cells, plasmablasts, antigen-presenting cells, and cells from non-lymphoid tissues [ 29 ]. PD-L1 is also highly expressed on the surface of syncytiotrophoblasts in early to late stages of pregnancy [ 30 ]. Decidual Tregs can express CD25, CD45RA, CTLA-4, TIGIT, TIM-3, Helios, and PD-L1 [ 31 ]. Others have shown that PD-L1 expression on the surface of Tregs exerts suppressive effects to control the maternal immune response. In an in vivo mouse model system, PD-L1 plays a critical role in fetal-maternal tolerance when PD-L1 blockade or deficiency results in decreased allogeneic fetal survival rate [ 32 ]. In addition, deficiency of PD-L1 may influence effector T-cells toward fetal rejection in favoring T helper (Th) 1 and Th17 cell development and expansion [ 32 ]. In an alloantigen-specific model, PD-L1 was shown to be associated with Th17 cells and fetomaternal tolerance [ 33 ]. Moreover, others have shown that PD-L1 expressing B-cells can protect and reduce disease severity in experimental models of autoimmune encephalomyelitis by restricting the ability of helper T-cells to cause inflammation and inducing Treg cell activity [ 15 , 34 ]. The relationship between Breg and Treg cell levels is understudied, particularly in the context of pregnancy. In this study, we pursue two specific aims to assess associations of Bregs (CD19 + CD24 hi CD38 hi ) and PD-L1 + Bregs with (1) cytokines and (2) Tregs (CD3 + CD4 + CD25 hi CD127 lo FoxP3 + ) in peripheral blood of pregnant women. We predict that Bregs will be positively associated with circulating levels of the anti-inflammatory cytokine IL-10 and negatively associated with pro-inflammatory cytokines. We predict that Bregs interact with Tregs in pregnancy. Due to their seemingly overlapping functions, it could be hypothesized that Bregs should be associated with Tregs positively (together coordinating immunoregulation) or negatively (expansion of one or the other immunoregulatory compartment could be sufficient). Therefore, our second aim does not have a directional hypothesis and remains more exploratory. The expansion of the Breg or Treg compartments during pregnancy may have postpartum, long-lasting effects in women. For example, Treg levels have been shown to continue to rise 12 months postpartum [ 35 ], and immune phenotypes may differ between primigravid and multigravid women, at least in decidual tissue [ 36 ]. Therefore, to check that any observed patterns are the result of the biological events occurring in a single pregnancy (and not cumulatively across multiple pregnancies), we conduct secondary analyses of the same regression models separately on the cohort subsets of primigravida (first pregnancy) and multigravida.