Cutting Edge: Committed Th1 CD4⁺ T Cell Differentiation Blocks Pregnancy-Induced Foxp3 Expression with Antigen-Specific Fetal Loss

Pregnancy requires expanded tolerance to encompass foreign paternal Ags expressed by the developing fetus (1, 2). The accumulation of maternal Foxp3⁺ regulatory CD4⁺ T cells (Tregs) occurs in healthy human pregnancy, whereas complications, such as spontaneous abortion or preeclampsia, which likely stem from disrupted fetal tolerance, have been linked with blunted maternal Treg expansion (3–8). In particular, in uncomplicated human pregnancy, the natural heterogeneity between maternal and paternal HLA-C alloantigens was shown to recruit Tregs to the maternal–fetal interface, which is associated with silencing effector T cell inflammatory responses (9–12). In turn, complementary animal studies allowing for experimental Treg manipulation established that maternal Tregs begin accumulating within the uterine-draining lymph nodes shortly after conception in response to seminal fluid and their necessity for sustaining fetal tolerance during allogeneic pregnancy (13–17). Thus, expanded maternal Tregs protect immunologically foreign fetal tissue from rejection.

With increasingly recognized heterogeneity among Foxp3⁺ cells, the necessity for unique maternal Treg subsets based on origin and specificity has been proposed (18–20). For example, the accumulation of Foxp3⁺ CD4⁺ T cells with specificity to fetal-expressed Ag and fetal resorption induced by prior stimulation with surrogate fetal Ags each suggests that maternal Tregs with fetal specificity play important protective roles (18–21). Induced Foxp3 expression is also likely essential because a majority of maternal Tregs with fetal specificity arise from Foxp3⁻ CD4⁺ T cells during primary pregnancy, and fetal resorption occurs when peripheral Treg conversion is circumvented in mice with disruption of the foxp3 enhancer conserved noncoding sequence-1 (18, 19). However, despite accumulation of maternal Tregs with fetal specificity, their role in sustaining pregnancy remains uncertain, given the lack of tools for manipulating Tregs in an Ag-specific fashion. To investigate the necessity for maternal Tregs with fetal specificity, pregnancy outcomes were evaluated in mice containing CD4⁺ T cells, with surrogate fetal specificity stably differentiated into non-Treg effectors prior to mating. Collectively, these studies show that committed Th1 CD4⁺ T cell differentiation blocks pregnancy-induced Foxp3 expression, causing Ag-specific fetal loss.

C57BL/6, congenic CD45.1⁺, and CD90.1⁺ mice (all H-2^b), as well as mice expressing 2W1S_55–68 peptide behind the ubiquitously active β-actin promoter backcrossed to BALB/c (H-2^d) or C57BL/6 mice, were described (19, 22). Expression of the 2W1S transgene was screened using 2W1S primers: 5′-CCAATCTGTCTGGCATCTCC-3′ and 5′-ATGATGGCCATAGCTCCAAG-3′ (22). For infection, Listeria monocytogenes were grown to early log-phase (OD₆₀₀ 0.1), washed, suspended in PBS, and inoculated i.v. at the following dosages: delta actin assembly–inducing protein (ΔACTA) L. monocytogenes–expressing 2W1S peptide (Lm-2W1S) (10⁶ CFU), delta listeriolysin O delta phospholipase C (ΔLLOΔPLC) Lm-2W1S (10⁷ CFU), or nonrecombinant ∆ACTA L. monocytogenes (10⁶ CFU) (23–25). For adoptive transfer, CD4⁺ T cells from the spleen and lymph nodes were purified by negative selection, and one mouse equivalent of CD45.1⁺ and CD90.1⁺ cells at a 1:1 ratio was inoculated i.v. into CD45.2⁺ CD90.2⁺ recipient mice before mating. For depletion, anti-CD4 (GK1.5) or anti–IFN-γ (XMG1.2) Abs were administered i.p. 1 d prior to mating and weekly thereafter (500 μg/dose). All experiments were performed in accordance with Institutional Animal Care and Use Committee–approved protocols.

Mononuclear cells from the spleen and axillary, brachial, cervical, inguinal, mesenteric, pancreatic, and para-aortic/uterine lymph nodes were collected and enriched with PE-conjugated I-A^b 2W1S_55-68 tetramer (19, 26), followed by cell surface (CD4, CD44, CD25, CD8, CD11b, CD11c, B220, F4/80), intracellular (IFN-γ, IL-17), or intranuclear (Foxp3, T-bet) staining. For stimulation, PMA (100 ng/ml) and ionomycin (1 μg/ml) were added for 5 h in media supplemented with brefeldin A (22).

For Treg differentiation, purified CD4⁺ T cells were stimulated with syngeneic APCs, 2W1S_55–68 peptide (10 μM), IL-2 (20 ng/ml), and TGF-β (up to 1.6 ng/ml). For Th17 polarization, CD4⁺ T cells were stimulated with syngeneic APCs, 2W1S_55–68 peptide (10 μM), IL-6 (20 ng/ml), IL-23 (10 ng/ml), and TGF-β (1 ng/ml) in media supplemented with anti–IFN-γ and anti–IL-4 Abs (10 μg/ml each). Five days after stimulation, Foxp3 expression and cytokine production were analyzed by intranuclear and intracellular staining.

Differences in the percentage of cells, resorption frequency, and number of pups were analyzed using an unpaired Student t test (two groups) or ANOVA with Dunnett correction for multiple comparisons (more than two groups), with p < 0.05 taken as statistical significance.

Recombinant strains of the intracellular bacterium L. monocytogenes have been widely used to characterize the T cell response to infection. We reported previously that L. monocytogenes cytoplasmic entry primes Th1 CD4⁺ T cell polarization and differentiation stability for cells responsive to Lm-expressed Ag (23). However, the use of monoclonal cells from TCR-transgenic mice with fixed specificity may limit their applicability for Treg differentiation, given the discordance in affinity and TCR repertoires between Foxp3⁺ and Foxp3⁻ CD4⁺ T cells (27). These drawbacks are bypassed by using MHC tetramers to identify endogenous CD4⁺ T cells, allowing a more comprehensive analysis of the cumulative polyclonal response (26). Using this approach, we (24) showed that the recombinant wild-type or attenuated ΔACTA Lm mutant that retains cytoplasmic entry each primes selective expansion of Foxp3⁻ CD4⁺ T cells with I-A^b:2W1S_55–68 specificity (24). Reciprocally, pregnancy sired by 2W1S-expressing males stimulates induced Foxp3 expression and Treg accumulation among CD4⁺ T cells with the same I-A^b:2W1S_55–68 specificity (19). Given the sharp discordance in Treg differentiation between L. monocytogenes infection and fetal stimulation, we reasoned that committed Th1 differentiation by recombinant L. monocytogenes could be exploited to investigate the necessity for Foxp3 induction among maternal CD4⁺ T cells during pregnancy. As controls for infection and 2W1S_55–68 stimulation, CD4⁺ T cells primed by L. monocytogenes restricted from the cell cytoplasm (ΔLLOΔPLC L. monocytogenes), with increased plasticity for differentiation into other effector lineages, were evaluated in parallel (23, 25).

We found that ΔACTA L. monocytogenes and ΔLLOΔPLC L. monocytogenes, each engineered to express 2W1S_55–68 as a secreted recombinant protein, primed robust accumulation of Foxp3⁻ CD4⁺ T cells with 2W1S specificity (Fig. 1A). In turn, the percentage of Foxp3⁺ Tregs among 2W1S⁺ CD4⁺ T cells declined precipitously postinfection with each recombinant L. monocytogenes compared with naive controls. These shifts were restricted to CD4⁺ T cells with L. monocytogenes specificity, because Foxp3 expression among 2W1S⁻ CD4⁺ T cells remained similar in L. monocytogenes–infected and control mice (Fig. 1A). Furthermore, 2W1S⁺ CD4⁺ T cells primed by ΔACTA L. monocytogenes selectively upregulated T-bet and acquired the capacity to produce IFN-γ, whereas these shifts were markedly diminished for 2W1S⁺ cells stimulated by ΔLLOΔPLC L. monocytogenes (Fig. 1B). Thus, L. monocytogenes primes the selective expansion of Ag-specific non-Treg CD4⁺ T cells, regardless of cytoplasmic entry, whereas Th1 differentiation requires L. monocytogenes cytoplasmic entry.

Next, the capacity for Treg differentiation among 2W1S⁺ CD4⁺ T cells primed by each L. monocytogenes strain was addressed. We found that TGF-β stimulation, along with 2W1S_55–68 peptide and IL-2, induced Foxp3 expression in a dose-dependent fashion among 2W1S⁺ CD4⁺ T cells recovered from both groups of L. monocytogenes–infected mice. However, at each TGF-β concentration, Foxp3 expression was significantly reduced for cells from ΔACTA Lm-2W1S–infected mice compared with ΔLLOΔPLC Lm-2W1S–infected mice (Supplemental Fig. 1A). Similarly, after stimulation under Th17-polarizing conditions, 2W1S⁺ CD4⁺ T cells from ΔACTA Lm-2W1S–infected mice compared with ΔLLOΔPLC Lm-2W1S–infected mice produced sharply reduced amounts of IL-17 while retaining robust IFN-γ levels (Supplemental Fig. 1B). Thus, CD4⁺ T cells primed by ΔACTA L. monocytogenes maintain more stable Th1 commitment, with resiliency against differentiation into Tregs or other effector lineages, whereas CD4⁺ T cells primed by ΔLLOΔPLC L. monocytogenes have more differentiation plasticity.

To investigate how this discordance dictates Foxp3 expression in vivo, pregnancy-induced Treg differentiation among 2W1S⁺ CD4⁺ T cells primed by ΔACTA Lm-2W1S compared with ΔLLOΔPLC Lm-2W1S was evaluated. Sixty days postinfection with each L. monocytogenes strain, virgin female mice were mated with allogeneic 2W1S-expressing males, which transforms 2W1S_55–68 into a surrogate fetal Ag (19, 22). Consistent with our recent studies using this mating scheme, ∼30% of maternal 2W1S⁺ CD4⁺ T cells in control mice without prior infection became Foxp3⁺ by embryonic day 15.5 (Fig. 2A) (19). In contrast, only ∼2% of maternal 2W1S⁺ CD4⁺ T cells from mice previously infected with ΔACTA Lm-2W1S were Foxp3⁺, whereas a majority retained T-bet expression and IFN-γ production (Fig. 2). Comparatively, Foxp3 became more readily induced among 2W1S⁺ CD4⁺ T cells from mice previously infected with ΔLLOΔPLC Lm-2W1S, albeit at reduced levels compared with naive mice (Fig. 2). Thus, resiliency against Treg differentiation shown in vitro for CD4⁺ T cells primed by ΔACTA L. monocytogenes is maintained and becomes even more pronounced with pregnancy-induced 2W1S stimulation.

Given the unique activation of immune components after ΔACTA L. monocytogenes infection compared with ΔLLOΔPLC L. monocytogenes infection (23, 25), resiliency against subsequent Foxp3 induction among cells from ΔACTA Lm-2W1S–infected mice could also reflect CD4⁺ T cell–extrinsic differences, in addition to cell-intrinsic shifts in differentiation stability. To discriminate between these possibilities, Foxp3 expression among CD4⁺ T cells from ΔACTA Lm-2W1S–infected mice or nonrecombinant ΔACTA L. monocytogenes–infected control mice was evaluated after adoptive transfer into naive recipients that subsequently were mated with 2W1S-expressing males. Mice with discordant expression of the CD45.1/2 and CD90.1/2 congenic markers were used so that each donor cell subset could be discriminated from each other and endogenous recipient cells (Fig. 3). We found that resistance against pregnancy-induced Foxp3 expression among CD4⁺ T cells from ΔACTA Lm-2W1S–infected mice was sustained after adoptive transfer, whereas induced Foxp3 expression to levels indistinguishable from endogenous naive CD4⁺ T cells were found among donor cells from mice previously infected with nonrecombinant ΔACTA L. monocytogenes (Fig. 3). Thus, despite considerable active debate on the relative stability of Foxp3⁺ Tregs (28), resiliency against pregnancy-induced Foxp3 expression among Th1 effector CD4⁺ T cells primed by ΔACTA Lm-2W1S reflects differentiation stability that is both cell intrinsic and Ag specific.

Having established resiliency against Foxp3 induction among 2W1S⁺ CD4⁺ T cells after ΔACTA Lm-2W1S infection and the efficiency with which pregnancy-induced 2W1S stimulation primes accumulation of maternal Tregs with the same specificity, these conditions were combined sequentially to address the necessity for Treg differentiation among maternal CD4⁺ T cells with fetal specificity during pregnancy. Using males heterozygous for the 2W1S transgene for mating, a reduction in the expected 1:1 ratio of 2W1S⁺/2W1S⁻ pups would illustrate Ag-specific fetal loss. Remarkably, the percentage and number of 2W1S⁺ pups by embryonic day 15.5 were significantly reduced in mice previously infected with ΔACTA Lm-2W1S compared with ΔLLOΔPLC Lm-2W1S or no-infection controls (Fig. 4A). In particular, although the expected 50% 2W1S⁺ pups did not deviate in pregnancies among naive mice (49% [43–55%]; mean [95% confidence interval]) or mice with prior ΔLLOΔPLC Lm-2W1S infection (50% [44–56%]), 2W1S⁺ pups were significantly reduced in pregnancies among mice with prior ΔACTA Lm-2W1S infection (38% [30–46%]); and this decline paralleled a 26% reduction in the number of live 2W1S⁺ pups (3.3/litter in ΔACTA Lm-2W1S–infected mice compared with 4.6/litter in ΔLLOΔPLC Lm-2W1S–infected mice or 4.6/litter in naive control mice) (Fig. 4A). Additionally, no differences in fetal resorption (each group’s background level < 10% for each group) were found at these later pregnancy time points, consistent with the necessity of maternal Tregs for implantation or beginning earlier during allogeneic pregnancy (17).

To investigate whether the selective loss of 2W1S⁺ offspring in ΔACTA Lm-2W1S–infected mice was caused by impaired CD4⁺ T cell differentiation, we evaluated the impact of CD4 cell depletion prior to mating on subsequent pregnancy. We found that the reduction in 2W1S⁺ pups among ΔACTA Lm-2W1S–infected mice became overturned with CD4 cell depletion; the percentage and number of 2W1S⁺ offspring (52% [46–58%]; 4.8 pups/litter) each rebounded to levels indistinguishable from naive controls (Fig. 4A). Likewise, neutralizing the Th1 effector cytokine IFN-γ prior to mating also abolished the loss of 2W1S⁺ offspring in ΔACTA Lm-2W1S–infected mice (54% [48–60%]; 4.6 2W1S⁺ pups/litter) (Fig. 4A). Thus, CD4⁺ T cells and IFN-γ are each essential for the protective contribution of maternal Tregs with fetal specificity; these findings in mice reinforce the protective role of human decidual Tregs that suppress effector T cell IFN-γ production (9–12). Together, these results suggest that maternal Tregs with fetal specificity confer protection by mitigating activation of Th1 effector cells that are harmful to pregnancy. In this regard, detrimental effector CD4⁺ T cells appear to be less constrained by decidual chemokine gene silencing compared with CD8⁺ T cells (29).

Comparatively, the degree of fetal loss triggered by prior ΔACTA Lm-2W1S infection also can be viewed as being somewhat marginal because 38% of pups remained 2W1S⁺, and the reduction in 2W1S⁺ offspring was, on average, only slightly more than one pup per litter, despite overwhelming resiliency against Foxp3 induction among maternal 2W1S⁺ CD4⁺ T cells with this specificity. It is important to highlight that this significant loss of offspring reflects circumvented Foxp3 induction among maternal CD4⁺ T cells of only a single surrogate fetal specificity that may be compensated by Tregs with specificity to other maternal–paternal mismatch Ags in allogeneic pregnancy or, alternatively, Tregs with self-specificity (20). To discriminate between these possibilities, outcomes after syngeneic pregnancy in mice previously infected with ΔACTA Lm-2W1S and sired by 2W1S-expressing males on the C57BL/6 background were evaluated. Interestingly, the loss of 2W1S⁺ pups observed in allogeneic pregnancy did not occur in syngeneic pregnancy, where the percentage and number of 2W1S⁺ pups were indistinguishable between ΔACTA Lm-2W1S–infected mice and control mice (49% [42–55%] and 4.4 2W1S⁺ pups/litter after ΔACTA Lm-2W1S infection compared with 51% [44–57%] and 4.5 2W1S⁺ pups/litter for naive no-infection controls) (Fig. 4B). Considering the muted protective role of maternal Tregs in maintaining syngeneic pregnancy compared with allogeneic pregnancy (13, 17, 19), these findings are perhaps not unexpected and may be explained by Treg-independent immune-silencing mechanisms recently described during syngeneic pregnancy (30). However, for allogeneic pregnancy that more closely recapitulates the mismatch between MHC haplotype Ags in human pregnancy, the immune-activating properties of fetal MHC alloantigens appear to be needed for uncovering the protective contribution of maternal Tregs with specificity for the surrogate fetal-2W1S minor histocompatibility Ag (1, 2). Applied to the reproductive process in humans and other outbred species in which there is considerably more variability in MHC haplotype and minor histocompatibility Ags between individuals, maternal tolerance likely needs to expand even more to accommodate the enhanced repertoire of foreign paternal–fetal Ags. However, increased antigenic mismatch that stimulates maternal Treg expansion for a broader assortment of foreign fetal Ags may also offset fetal loss shown in this study with resistance against Treg differentiation among maternal CD4⁺ T cells of a single specificity. These additional questions will require more comprehensive immunological tools capable of tracking CD4⁺ T cells with specificity for a significantly broader array of fetal and nonfetal specificities. Nevertheless, by experimentally establishing overlap between a single L. monocytogenes–expressed and surrogate fetal Ag, the protective benefits of Treg differentiation among maternal CD4⁺ T cells with fetal specificity in optimal pregnancy outcomes is revealed.

The authors have no financial conflicts of interest.