Regulatory T Cells: New Keys for Further Unlocking the Enigma of Fetal Tolerance and Pregnancy Complications

Reproductive success in eutherian placental mammals allowing prolonged in utero fetal maturation and protection presents a conundrum whereby the mother must tolerate, as well as provide nourishment to, the immunologically foreign fetus. More than 60 years ago, Sir Peter Medawar (1) posited theories to explain how fetal tolerance may occur, including physical separation between maternal and fetal tissues, fetal antigenic immaturity, and maternal immune suppression. Since then, evidence supporting the partial validity of these explanations has been shown, with more comprehensive molecular and cellular characterization of maternal reproductive and fetal tissues. For example, entrapment of APCs within the decidua, local exclusion of effector T cells through chemokine gene silencing, and reduced complement deposition together create a formidable immunological barrier (2–4). Similarly, the diminished or skewed MHC expression by trophoblastic cells illustrates the additional contribution of antigenic immaturity in sustaining fetal tolerance (5, 6). However, considering the increasingly established transfer of cellular vesicles or intact cells between mother and fetus, as well as systemic recognition of the fetus by maternal immune cells (7–9), these processes that work locally at the maternal–fetal interface are likely incomplete and implicate the necessity for complementary systemic immunological shifts.

Earlier studies showed that healthy pregnancies were associated with reduced IFN-γ and reciprocally increased Th2 polarization of maternal PBMCs, whereas complications, such as preeclampsia and spontaneous abortion, were each linked with more IFN-γ production (10). Although these findings initially were interpreted to imply a requirement for maternal Th2 CD4⁺ T cell polarization in sustaining fetal tolerance, normal pregnancy outcomes in mice, despite individual or combined defects in Th2 cytokines (e.g., IL-4, IL-5, IL-9, and IL-13), refined these interpretations to instead implicate a necessity for maintaining anti-Th1 responses (11, 12). More recently, as additional CD4⁺ T cell subsets and non-Th1–differentiation programs have been identified, our conceptual understanding of the protective immunological changes that occur during pregnancy has shifted in parallel.

In particular, the identification of regulatory T cells (Tregs) as a distinct CD4⁺ T cell lineage dedicated to silencing activation of other immune components has rekindled the consideration of active immune suppression in many physiological and disease processes. Most peripheral Tregs acquire Foxp3 expression within the thymus based on specificity for self-Ag and suppress self-reactive immune components that escape central tolerance (13). However, the additional capacity for peripheral conversion of naive CD4⁺ T cells with specificity for a near infinite array of immunologically foreign Ags into Tregs establishes an adaptable arsenal of suppressive cells capable of responding to fluctuating environmental cues. In this regard, Tregs have already been shown to restrain activation of immune cells with commensal specificity that protects against autoimmunity and refines immunity to pathogens (14–16). In this review, accumulating evidence implicating extended protective roles for Tregs in accommodating the expanded repertoire of immunologically foreign Ags expressed by the developing fetus during pregnancy is summarized (Fig. 1).

The importance of Tregs likely begins during the menstrual cycle to render female reproductive tissues receptive to foreign paternal–fetal Ags. Cyclic reproductive hormones leading up to pregnancy have potent quantitative and qualitative impacts on Tregs (17–21). For example, the magnitude of peripheral maternal Treg expansion is tightly correlated with serum estradiol levels that progressively rise just prior to ovulation (17). In turn, progesterone that accumulates in the luteal phase after ovulation acts synergistically with TGF-β and IL-2 to induce Foxp3 expression and augment Treg suppression (18, 19, 21). Thus, female reproductive hormones work together with immune cells in a fine-tuned fashion for pregnancy preparation.

Maternal Treg expansion during pregnancy was first described in a series of observations comparing CD25 expression among maternal CD4⁺ T cells. Mahmoud et al. (22, 23) reported a 1.6-fold increase in CD25⁺ “activated” CD4⁺ T cells in the peripheral blood of women during the third trimester compared with nonpregnant controls. With awareness that CD25 expression more likely identifies immune-suppressive cells, greater expansion of maternal CD25⁺ CD4⁺ Tregs that peak in the first and second trimesters of pregnancy were subsequently described (24, 25). Moreover, in parallel with accumulation of immunologically foreign fetal tissue at the maternal–fetal interface, CD25⁺ CD4⁺ Tregs with increased CTLA-4 expression were found within the decidua (25, 26). Since these initial reports, a wealth of human epidemiological studies have replicated these key findings. Collectively, they provide a consensus that maternal Tregs expand from a baseline of 1–10% of peripheral CD4⁺ T cells to 3–25% by midgestation, with subsequent contraction to near background levels later in pregnancy (Table I) (22–38).

Importantly, however, there have also been some notable inconsistencies in Treg shifts during pregnancy. Mjösberg et al. (34) found that FOXP3 expression among maternal peripheral T cells was primarily restricted to the CD4^dim CD25^high subset, and these cells declined by ∼50% in the second trimester of pregnancy. Interestingly, using other permutations of CD25 or FOXP3 coexpression among CD4⁺ T cells, other investigators (36, 37) linked diminishing peripheral Tregs with migration into the decidua. However, several caveats related to the analysis of human Tregs need to be considered when evaluating these results in aggregate. The first is the high degree of natural variation among individuals illustrated by the wide range of peripheral Treg frequencies, even among nonpregnant controls. The second relates to discordant markers used for identifying human Tregs. Although Foxp3 is expressed exclusively by Tregs in mice (39, 40), FOXP3 expression in humans occurs for both immune-suppressive Tregs and activated effector T cells (41, 42). To address this limitation, expression of CD25 and/or downregulation of IL-7R (CD127), along with Foxp3, have been used (Table I). However, although high-level CD25 expression has consistently identified cells with suppressive properties in vitro, their ability to restrain effector T cell activation in vivo remains incompletely defined, and diminished CD127 expression that more uniformly identifies suppressive cells has not been consistently used (43, 44). Nevertheless, despite the inherent heterogeneity among human individuals combined with unique permutations of FOXP3, CD25, and expression of other CD4⁺ T cell–intrinsic molecules used to identify Tregs, the physiological accumulation of maternal Tregs during pregnancy has been widely replicated (Table I). Moreover, given the increasingly recognized functional specialization of Treg subsets with regard to specificity or molecules used to mediate context-specific immune suppression (39, 40, 45), individual studies in which bulk maternal Tregs do not expand significantly does not necessarily negate their importance in pregnancy. Instead, uncovering the most critical protective features of maternal Tregs will likely require analysis of distinct Treg subsets based on specificity and local accumulation at the maternal–fetal interface.

Analysis of maternal Tregs in human pregnancy complications provides additional evidence supporting their protective necessity. Blunted expansion and/or functional decline has been described in many seemingly unrelated complications, such as spontaneous abortion, preeclampsia, and prematurity, which potentially share underlying defects in fetal tolerance (Table II). Among women with preeclampsia compared with gestational aged-matched controls, an ∼33% reduction in peripheral Treg expansion has been consistently described using various combinations of molecular markers (CD25^high CD4⁺, CD127^low CD25^high CD4⁺, FOXP3⁺ CD4⁺) (17, 23, 26, 27, 29–33, 36, 46–58). Furthermore, a representative case series of 43 women with preeclampsia showed that impaired Treg expansion was further exacerbated by increased levels of IL-17–producing CD4⁺ T cells (31). These findings are consistent with the common developmental origins but dichotomous differentiation programming for Tregs and Th17 CD4⁺ T cells that may explain pathologically elevated levels of proinflammatory cytokines, such as IL-6 and TNF-α, during preeclampsia (59–61). Moreover, systemic inflammation that occurs during preeclampsia is likely fueled by defects in fetal tolerance, considering that delivery of the fetus and placenta remains the most definitive and only effective therapy (62). Thus, active suppression of immune components targeting foreign fetal tissue work in tandem with reproductive hormones and fetal cells that induce Foxp3 expression to silence activation of pathogenic effector T cells in healthy pregnancies (17–21, 63).

Quantitative defects in maternal Tregs also have been described in cases of spontaneous abortion (Table II). Sasaki et al. (26) reported an ∼33% decline in peripheral blood CD25^bright CD4⁺ T cells among women with spontaneous abortion compared with uncomplicated pregnancy. This decline was even more pronounced for decidual Tregs: CD25^bright CD4⁺ T cells declined from 22% in cases of elective abortion compared with 7% in spontaneous abortion (26, 30). Similarly, women with recurrent spontaneous abortion consistently had 40–90% reductions in circulating and decidual Tregs (Table II). Interestingly, along with these quantitative Treg reductions, diminished suppressive potency for maternal Tregs also likely contributes to recurrent pregnancy loss, because production of the suppressive cytokines IL-10 and TGF-β by CD127^dim CD25⁺ CD4⁺ T cells recovered from women with recurrent abortions were reduced to 2 and 1%, respectively, compared with 9 and 14% among gestational aged–matched controls (54). Reductions in Treg-suppressive potency also were described in preterm labor, which further illustrates the importance of maintaining qualitative shifts in suppressive function, along with quantitative expansion of these cells, throughout pregnancy (64). Collectively, these associations between maternal Treg expansion in uncomplicated human pregnancies and blunted accumulation in pregnancy complications establish the groundwork implicating a protective role for these cells in sustaining fetal tolerance.

Human studies, although instrumental in describing molecular and cellular changes associated with pregnancy, are inherently limited by ethical considerations that preclude experimental manipulation and analysis of reproductive tissue required for establishing the cause and effect relationship between maternal Tregs and pregnancy outcomes. Fortunately, animal studies have bypassed many of these roadblocks and, together, more definitively demonstrate the protective necessity of maternal Tregs. In turn, redirecting the wealth of transgenic mouse tools to investigate fetal tolerance provides additional mechanistic insights about how Tregs mediate immune tolerance in other contexts besides pregnancy.

First, maternal Tregs accumulate with similar magnitude and tempo in mice compared with humans during pregnancy (Table III) (35, 65–73). Using inbred mouse strains with discordant MHC haplotypes for mating that recapitulates the natural heterogeneity between analogous maternal–paternal Ags in human pregnancy, circulating maternal CD25⁺ or Foxp3⁺ CD4⁺ T cells expand significantly above background levels within 2 d after pregnancy and reach peak ∼2-fold–enriched levels by midgestation (65, 69, 71, 74). At early pregnancy time points, immune-suppressive cytokines present in seminal fluid alone may foster local Treg expansion (75). Reciprocally for abortion-prone matings among defined strains of inbred mice (e.g., DBA/2J [H2^d] ♂ × CBA/J [H2^k] ♀), maternal Tregs are sharply reduced to levels comparable to or below those found in nonpregnant controls (67). Additional comparisons using inbred mice with identical or discordant MHC haplotypes illustrate a direct correlation between the magnitude of maternal Treg expansion and degree of mismatch between maternal–paternal alloantigens. For example, maternal Treg expansion is eliminated or diminished to nonsignificant differences during syngeneic pregnancy among genetically homogenous mice in which the only potential source of antigenic mismatch are those encoded by the Y chromosome, whereas it is consistently more robust in allogeneic pregnancies among MHC haplotype–discordant mice (35, 70, 71, 73). Thus, although female reproductive hormones and seminal fluid participate in pregnancy preparation and early pregnancy (17–21, 75), the more robust Treg taccumulation during allogeneic pregnancy compared with syngeneic pregnancy highlights the additive or potentially synergistic contribution of maternal–fetal antigenic discordance.

Applying Ag-specific tools to track CD4⁺ T cells demonstrates even more pronounced accumulation of maternal Tregs with fetal specificity (72, 73, 76, 77). Using female mice with fixed TCR specificity or repopulated with donor monoclonal T cells for breeding showed that pregnancy stimulates induced Foxp3 expression among CD4⁺ T cells with specificity to naturally occurring fetal alloantigens or OVA expressed as a surrogate fetal Ag (73, 76). Similarly, when male mice that ubiquitously express the I-A^b 2W1S_55–68 peptide were used for breeding with nontransgenic females, endogenous maternal Tregs with surrogate fetal-2W1S specificity, recovered systemically from the spleen and peripheral lymph nodes, expanded 72-fold compared with bulk maternal Tregs that expanded <2-fold (71, 77). Although the exact molecular and cellular modifications that drive induced Foxp3 expression and expansion of Tregs with fetal specificity remain incompletely defined, essential clues likely reside within decidual maternal APCs stimulated by fetal components that prime tolerogenic, as opposed to activated, effector T cell phenotypes (4, 78). A recent study showed thymic stromal lymphopoietin produced by trophoblast cells markedly augments the capacity of dendritic cells to induce FOXP3 and CD25 expression among CD25⁻ CD4⁺ human decidual cells (63). These protective benefits of qualitative shifts in maternal Tregs are consistent with the reinstatement of healthy pregnancy by donor Tregs from pregnant, but not virgin, control mice in abortion-prone matings between DBA/2J ♂ × CBA/J ♀ mouse strains (67). Thus, Tregs expand with comparable tempo and magnitude during uncomplicated human and allogeneic mouse pregnancy, whereas naturally occurring complications in both species are associated with blunted accumulation of these cells. In turn, mouse pregnancy studies further uncovered that Treg expansion likely represents a more focused response toward immunologically foreign fetal Ags.

The protective necessity for maternal Tregs is reinforced by evaluating pregnancy outcomes after experimental Treg manipulation. In pioneering studies by Aluvihare et al. (65), depletion of CD25⁺ cells among donor lymphocytes was shown to induce fetal wastage after transfer into T cell–deficient female recipients prior to pregnancy. Interestingly, even in these initial experiments, the necessity for maternal Tregs linked with immunologically distinct fetal Ag was shown, because resorption observed in allogeneic pregnancies was eliminated in syngeneic matings (65). Directly manipulating Tregs with anti-CD25 Ab administered to pregnant mice showed that Treg depletion in early pregnancy (embryonic day [E]2.5) caused implantation failure, whereas depletion later (E4.5 or E7.5) induced fetal wastage (69, 74). Similar results were obtained using transgenic mice that coexpress the high-affinity human diphtheria toxin receptor with Foxp3, which bypasses potential limitations associated with manipulating Tregs based on CD25 expression. Targeted ablation of maternal Foxp3⁺ cells was shown to induce near complete fetal resorption in allogeneic pregnancy that also became significantly reduced in syngeneic matings (71, 77). Therefore, although diphtheria toxin–induced apoptotic cell death may cause inflammation that contributes to these detrimental outcomes (79–81), the consensus that selectively eliminating Tregs, regardless of experimental approach, triggers fetal wastage illustrates an essential protective role for these cells (65, 69, 74, 80).

Exploiting the X-linked inheritance of foxp3 and random inactivation of this chromosome that renders only one-half of Tregs susceptible to ablation in Foxp3^DTR/WT heterozygous females showed that even partial transient Treg depletion causes fetal resorption associated with expansion and IFN-γ production by fetal-specific maternal effector T cells (71, 77). Considering that the maternal Treg nadir reaches prepregnancy levels after depletion in Foxp3^DTR/WT mice, which directly parallels their blunted expansion in naturally occurring human pregnancy complications (Table II), it is likely that the sustained accumulation of maternal Tregs is required for maintaining allogeneic pregnancy. Furthermore, the selective loss of male pups with partial Treg depletion in syngeneic pregnancy indicates that the protective benefits of maternal Tregs are likely not limited to MHC haplotype Ags and may extend to minor maternal–fetal discordant alloantigens as well (70, 82).

More recent studies with secondary fetal Ag stimulation further highlight protection by maternal Tregs with fetal specificity. Using MHC class II tetramers to track CD4⁺ T cells with the aforementioned I-A^b 2W1S_55–68 surrogate fetal specificity, maternal Tregs with this specificity were retained at enriched levels postpartum and re-expanded with accelerated tempo following fetal-2W1S restimulation in subsequent pregnancy (77). Remarkably, with this enriched retained pool of fetal-specific memory maternal Tregs, secondary pregnancy becomes markedly more refractory to partial ablation of bulk Foxp3⁺ cells, suggesting enhanced protection by Tregs with fetal specificity. These findings provide critical immunological clues to explain the partner-specific protective benefits conferred by prior pregnancy against complications, such as preeclampsia, in subsequent pregnancies (77, 83). However, considering that human epidemiological analyses also showed that these protective benefits wane as the interpregnancy interval becomes significantly extended (84, 85), a gradual decline in Treg memory that parallels diminishing numbers of effector CD4⁺ T cells with time after Ag elimination is predicted (86–88). Thus, establishing the durability and potential for amplifying the protective properties of maternal Tregs, analogous to booster vaccines for enhancing immunity against pathogens, represents an important area for translational application of memory Tregs.

Using a complementary approach, other recent studies also showed that stimulation, under conditions that block pregnancy-induced Treg expansion among maternal CD4⁺ T cells with fetal-2W1S specificity, causes a selective loss of 2W1S⁺ offspring in pregnancies sired by 2W1S-expressing males (89). This reduction in offspring that occurs in an Ag-specific fashion parallels similar results with preconceptual immunization with other surrogate fetal Ags (72). Together, these findings strongly suggest that Tregs with fetal specificity play distinct protective roles in sustaining and reinforcing fetal tolerance. Hence, further application of Ag-specific tools for investigating Tregs has enormous potential for uncovering other critical facets of how these cells work in pregnancy and other biological contexts in which the importance of Treg Ag specificity has been implicated (e.g., tolerance to commensal or pathogenic microbes, autoimmunity) (14–16, 39, 40, 90).

Application of Ag-specific tools may be especially illuminating for investigating the molecular basis by which Tregs mediate immune suppression. Even with identification of unique and context-specific suppressive properties for many individual Foxp3⁺ cell-intrinsic molecules, none has been shown to individually reproduce all aspects of Treg suppression (40, 45). For example, although Treg-intrinsic CTLA-4 is essential for averting systemic autoimmunity (91), targeted deletion of this molecule in Foxp3⁺ cells throughout development in mice does not recapitulate the more rapid onset of fatal disease observed when Tregs are completely lacking. Similarly, despite protection against inflammation at mucosal surfaces by Treg production of IL-10, neither Foxp3⁺ cell-intrinsic nor complete deficiency in IL-10 causes systemic autoimmunity (92). With regard to pregnancy, neither of these molecules best implicated to mediate Treg protection from autoimmunity appears to be essential. Mice with targeted IL-10 deficiency or treated with IL-10R–neutralizing Ab showed no significant loss of live offspring after allogeneic pregnancy (11, 71). Similarly, CTLA-4 neutralization neither induces fetal resorption nor overrides protection conferred by donor Tregs in abortion-prone pregnancies (93). Interestingly, however, Treg-intrinsic defects in PDL-1 or neutralization of PD-1 prior to adoptive transfer each fails to rescue protection against fetal wastage in abortion-prone matings (93, 94). Nevertheless, it is more likely that multiple Treg-associated molecules play functionally redundant roles in sustaining fetal tolerance, because only inconsistent results are observed after allogeneic pregnancy in PD-1– or PDL-1–deficient mice (95). Considering our newfound appreciation of Treg specificity, establishing pregnancy-induced shifts in the expression of immune-suppressive molecules among maternal Tregs with fetal specificities compared with other specificities may be essential for uncovering the molecular basis by which Tregs sustain fetal tolerance.

Although pregnancy represents a ubiquitous model of immune tolerance, the molecular and cellular alterations required for sustaining fetal tolerance have remained enigmatic. With the seminal discovery of Tregs as a unique CD4⁺ T cell lineage dedicated to immune suppression, their necessity in averting autoimmunity and maintaining tolerance to self and “extended-self” commensal Ags swiftly ensued. In parallel, a compelling complement of human and animal data have collectively illustrated that the necessity of immune suppression by Tregs extends to maintaining fetal tolerance. Given the importance of reproductive fitness in species survival, it also can be argued that the advantages gained from more prolonged in utero fetal development were a dominant factor in positive selection that endowed CD4⁺ T cells with induced Foxp3 expression to encompass immunologically foreign fetal Ags. This more prominent consideration of reproductive success driving stepwise gains in Treg suppression is supported by comparative genomic analyses across species, illustrating conservation within foxp3 compared with associated enhancer elements (76, 96). For example, although foxp3 is stringently conserved in mammals, the conserved noncoding sequence 1 enhancer required for induced Foxp3 expression is present only in placental mammals and is distinctively absent in egg-laying monotremes and most marsupial species (76). This capacity for enhanced accumulation of maternal Tregs with fetal specificity among eutherian placental mammals protects against maternal–fetal conflict, allowing more prolonged in utero fetal development; therefore, it likely played preeminent roles in species selection and survival (76). Based on this assertion, establishing how Tregs work during pregnancy has both direct translational implications for improving maternal–fetal health and enormous potential for uncovering the fundamental biology for how immunity against extended self-Ags and foreign Ags is regulated. Considering that shifts in maternal Tregs that establish their underlying importance in the reproductive process was first described only 10 years ago, we are enthusiastically optimistic that further investigation in this exciting area will soon unveil the keys to more comprehensively unlock the enigma of fetal tolerance and pregnancy complications.

The authors have no financial conflicts of interest.