and present address: Dr D.M. Sansom, Division of Immunity and Infection, University of Birmingham, Birmingham, B15 2TT, UK. E-mail: D.M.Sansom@bham.ac.uk
CD4+ CD25+ T regulatory cells (TReg), suppress antigen-specific immune responses and are important for allograft tolerance. During pregnancy the mother tolerates an allograft expressing paternal antigens (the fetus) requiring substantial changes in immune regulation over a programmed period of time. We analysed whether immune-suppressive TReg cells were altered during pregnancy and therefore might play a part in this tolerant state. The presence of TReg cells was assessed in the blood of 25 non-pregnant, 63 pregnant and seven postnatal healthy women by flow cytometry. We observed an increase in circulating TReg cells during early pregnancy, peaking during the second trimester and then a decline postpartum. Isolated CD25+ CD4+ cells expressed FoxP3 messenger RNA, a marker of TReg cells, and suppressed proliferative responses of autologous CD4+ CD25– T cells to allogeneic dendritic cells. These data support the concept that normal pregnancy is associated with an elevation in the number of TReg cells which may be important in maintaining materno-fetal tolerance.
CD25+ CD4+ regulatory T cells (TReg) are a specialized subset of T cells involved in preventing autoimmunity.1–3 Studies have also demonstrated that TReg can prevent graft rejection,4,5 indicating that these cells are potent suppressors of T-cell responses. TReg arise in the thymus as a result of altered negative selection and are present from fetal life onwards. However, there are suggestions that TReg may also develop in response to antigenic stimulation under non-immunogenic circumstances.6
The mechanism by which TReg suppress responses is uncertain. Both cell contact and cytokine-based mechanisms have been proposed and roles for interleukin-10 and transforming growth factor-β (TGF-β) have been suggested.7–12 Interestingly, recent data indicate that TGF-β treatment of CD4+ CD25– cells may lead to the induction of the transcription factor FoxP3 and the generation of a regulatory phenotype. Furthermore, although classically described as being non-proliferative, recent evidence indicates that TReg can proliferate in vivo and are therefore capable of expanding in response to stimulating antigens.13
Whilst identification of TReg is associated with expression of CD25, this receptor is also expressed on normal activated T cells. However, in contrast to T cells that express CD25 during activation, TReg appear non-blastic and do not express other markers associated with recent activation. Recently, Foxp3, the gene encoding the transcription factor Scurfin, has been associated with the generation and function of murine TReg cells and does not appear to be expressed by normally activated T cells. Accordingly, Scurfin-deficient (sf) mice lack TReg cells and suffer autoimmunity and conversely, transgenic mice that over-express Foxp3 have increased numbers of TReg cells and display increased immunosuppressive activity when compared to wild-type mice.14,15
During pregnancy, the mother's immune system tolerates the fetus, which expresses paternal major histocompatibility complex antigens. Current evidence suggests that the maternal immune system in humans is aware of the presence of this ‘allograft’, but fails to mount a classical immune response;16 the mechanisms behind this tolerance are still being sought. One important mechanism that has emerged in materno-fetal tolerance is tryptophan catabolism by dendritic cells (DCs). The breakdown of tryptophan by the enzyme indoleamine 2,3 dioxygenase (IDO) leads to suppression of T-cell responses. Strikingly, inhibition of IDO has been shown to lead to abortion of allogeneic, but not syngeneic, concepti in mice.17,18 This strongly suggests that IDO activity is a key mechanism involved in materno-fetal tolerance in mice. Whilst studies in humans are limited, IDO has been detected in human decidual macrophages19 and in placental villous explants, and has been shown capable of inhibiting human T-cell responses.19 Intriguingly, IDO expression may also be influenced by cytotoxic T-lymphocyte antigen-4 (CTLA-4), a molecule expressed by TReg cells.20,21 We therefore hypothesized that CTLA-4-expressing TReg may play a role in inducing maternal tolerance to alloantigens during pregnancy. To test this hypothesis, we have evaluated changes in TReg numbers during the course of normal human pregnancy. Our data show that TReg cells are significantly increased during pregnancy and return to lower levels post partum.
Materials and methods
Ethics approval for this study was provided by the Local Research Ethics Committee and all women gave their informed, written consent. Venous blood from healthy women (aged 19–42 years) was subjected to an automated full blood count, to measure numbers of lymphocytes, and to flow cytometry to assess proportions of lymphocyte subsets. The whole blood fluorescence-activated cell sorter (FACS) lysis method was used with phycoerythrin- (PE), fluorescein isothiocyanate- (FITC) and peridinin chlorophyll protein (PerCP) fluorochrome-conjugated mouse monoclonal antibodies to CD3, CD4, CD8, CD11a, CD25, CD45RO, CD45RA, CD69, HLA-DR (Becton Dickinson, Oxford UK) and CXCR3 (R & D Systems, Oxford, UK). Briefly, whole blood (50 μl) was stained with 5 μl labelled antibody for 30 min. Red cells were then lysed by addition of FACS lysing solution (Becton Dickinson) according to the manufacturer's instructions. Cells were washed once in phosphate-buffered saline and analysed. For comparison, whole blood was diluted in RPMI-1640 and activated T cells were generated by incubation for 3 days with the mitogen phytohaemagglutinin at a concentration of 25 μg/ml.
Suppression of alloresponses by TReg
Functional lymphocyte studies were carried out using mononuclear cells isolated from blood by Ficoll density gradient centrifugation. CD4+ T cells were purified by immunomagnetic negative selection (Stem Cell Technologies, Meylan, France). CD25+ CD4+ cells were positively selected using anti-CD25 microbeads (Miltenyi Biotech, Bisley, UK). CD4+ CD25– cells were collected from the washing steps, and the CD4+ CD25+ cells were subsequently retrieved from the beads.
DCs were cultured from unrelated donors using monocytes purified from blood by negative immunomagnetic selection (Miltenyi Biotech) and cultured for 6 days in RPMI-1640 medium containing 10% fetal calf serum with granulocyte–macrophage colony-stimulating factor (800 U/ml), interleukin-4 (500 U/ml) and standard antibiotics.22
Alloantigen stimulation experiments were performed using the cultured DCs. CD25+-depleted CD4+ T cells (1 × 105) were stimulated with DCs at ratios of 1 : 10 and 1 : 100 DC : T cells in 200-μl final volumes. T-cell proliferation was measured by [3H]thymidine incorporation. The inhibitory effect of the CD25+ CD4+ T cells was assessed at the ratios shown.
Oligo-dT-primed first-strand cDNA was synthesized with a SuperScript™ First-strand Synthesis System (Invitrogen, Paisley, UK) as a template for real-time PCR. PCR primer pairs and internal TaqMan probes for the genes of interest were as follows:
Foxp3 forward 5′-GAGAAGCTGAGTGCCATGCA-3′ and reverse 5′-AGGAGCCCTTGTCGGATGAT-3′. FAM probe 5′-AAAATGGCACTGACCAAGGCTTCATCTGT-3′; and β-actin, forward 5′-CCTGGCACCCAGCACAAT-3′ and reverse 5′-GCCGATCCACACGGAGTACT-3′ VIC probe 5′-ATCAAGATCATTGCTCCTCCTGAGCGC-3′.
The FoxP3 primers were used at final concentrations of 300 nm and 400 nm, respectively, and the probe at a concentration of 200 nm. For β-actin, primers were used at final concentrations of 20 nm and 70 nm, respectively, and the probe at a final concentration of 100 nm. Other components of the PCR mix were from the TaqMan Universal Master Mix (PE Applied Biosystems, Warrington, UK). Forty-five cycles of 95° for 15 seconds and 60° for 1 min were carried out using an ABI Prism 7700 Sequence Detection system. The relative quantity of mRNA was calculated by plotting a curve of threshold cycle (CT) for FoxP3 and β-actin for each sample.
Data were non-parametrically distributed. The two-tailed Mann–Whitney test was used to compare pairs of groups, with significance defined as P < 0·05. The Kruskal–Wallis test was used to assess changes in the proportion of TReg cells with gestation.
Normal pregnancy is associated with an increase in CD4+ CD25+ TReg cells
Initially the percentage of TReg cells was evaluated (as a proportion of the total peripheral lymphocyte population) in pregnant individuals (n = 63) compared to non-pregnant controls (n = 25). This revealed (Fig. 1a) that the median percentage of CD4+ CD25+ T cells was doubled during pregnancy (8·9%, range 3·4–28·1%) compared to controls (4·4%, range 2·3–8·0%; P < 0·0001). Analysis by stage of pregnancy showed a significant increase from the non-pregnant controls to the first trimester [4·4% (n = 25) versus 6·7% (n = 5); P < 0·01] and from first trimester to second trimester [6·7% (n = 5) versus 10·9% (n = 13); P = 0·03]. However, there was no significant change from the second to third trimester [10·9% (n = 13) versus 8·9% (n = 45); P = 0·06]. The proportion of CD4+ CD25+ T cells was significantly lower 6–8 weeks following delivery than during the third trimester [8·9% (n = 45) versus 7·5%; P = 0·03 (n = 7)] but remained higher than pre-pregnancy (7·5% versus 4·4%; P < 0·01). A scatter plot of the distribution of results against gestational age is also shown (Fig. 1b). Analysis of variance using the Kruskal–Wallis test suggested that the variation in TReg cells with gestation was highly significant (P < 0·0001). Since the total lymphocyte counts were not significantly different between the pregnant and non-pregnant women, the increased proportion of CD4+ CD25+ T cells was also reflected in increased absolute numbers.
To characterize further the phenotype of the CD4+ CD25+ T cells, FACS analysis was carried out. Representative FACS plots are shown in Fig. 2. Results for TReg from a pregnant woman are shown in black and for comparison results from control blood stimulated for 3 days with phytohaemagglutinin are shown in grey. The levels of CD25 observed in TReg were considerably lower than those seen in T-cell blasts. In addition to small size (Fig. 2b), evidence that these CD4+ CD25+ T cells were TReg cells rather than recently activated T cells included a lack of expression of markers associated with activation (CD69 and HLA-DR). This contrasted with high levels of expression of CD69 and HLA-DR amongst the activated lymphocytes (Fig. 2d,e). CD45RO was present on 34–52%, indicating previous exposure to antigen, with a reciprocal proportion positive for CD45RA, demonstrating that CD25+ TReg were contained within both the CD45RO ‘memory’ population as well as the apparently naïve CD45RA cells. CXCR3, an important chemokine receptor for migration into allografts but also noted on regulatory T cells,23 was found on 38–50% of TReg cells in pregnancy and at higher levels than in blasts. Overall this analysis showed marked differences between CD4+ CD25+ cells purified from peripheral blood during pregnancy and CD4+ CD25+ cells generated as a consequence of immunogenic stimulation. This indicated that the CD25+ cells observed in pregnancy were not the consequence of classical stimulation in vivo.
TReg cells express FoxP3 and suppress alloresponses
FoxP3 is now an established marker for TReg cells in mice and humans.15 To provide further support for our hypothesis that CD4+ CD25+ T cells isolated during pregnancy represented TReg cells, the expression of FoxP3 within these cells was investigated using quantitative PCR (Fig. 3). This revealed that FoxP3 was highly enriched in CD4+ CD25+ T cells compared to CD25– T cells. These data were consistent with previous studies suggesting that FoxP3 is selectively expressed in TReg and supported our contention that the CD4+ CD25+ cells observed in pregnancy were indeed TReg cells.
To establish that the CD25+ T cells identified possessed regulatory function, their ability to suppress allo-responses in CD25– T cells was also analysed. Purified CD25– T cells were stimulated using allogeneic DCs and the effect of adding back CD25+ T cells was assessed. This demonstrated (Fig. 4) that the CD4+ CD25+ cells isolated from healthy pregnant women were capable of suppressing [3H]thymidine uptake in allogeneically stimulated lymphocytes by approximately 50%. Interestingly, no statistically significant difference in suppressive activity was found between five pregnant women and their non-pregnant controls. Overall these data suggest that the CD4+ CD25+ T cells isolated from pregnant women contain a regulatory population that on a per cell basis are of similar potency to those from non-pregnant controls, but which are present at a significantly elevated frequency.
We have investigated the prevalence of the CD4+ CD25+ T-cell subset in human pregnancy, to determine whether changes in this subset might be relevant to materno-fetal tolerance. We hypothesized that TReg cells may play a role in modifying the maternal immune response to the feto-placental ‘allograft’ within the uterus. The dramatic and rapid increase in circulating TReg cells during early pregnancy, peaking during the second trimester at a time when trophoblast invasion of the maternal decidua is maximal, is consistent with this.24 Equally, the decline in TReg cells postpartum is also consistent with the withdrawal of the immunological stimulus of the allograft. We have demonstrated that this population of CD25+ cells can inhibit the induction of lymphocyte proliferation by a third-party allogeneic stimulus and is therefore likely to be capable of this function in vivo. Since in vitro studies suggest that cell–cell contact is necessary for TReg-mediated immunosuppression, we are also extending studies of blood lymphocytes to an assessment of TReg cells within decidua.
Our functional studies demonstrate that these CD4+ CD25+ lymphocytes suppress [3H]thymidine incorporation in CD4+ CD25– lymphocytes stimulated with allogeneic DCs by around 50%. These results confirm that the CD4+ CD25+ lymphocytes studied have T-cell inhibitory activity, although, in our experiments so far, they do not appear any more potent than TReg cells from non-pregnant controls. However, given the increase in absolute numbers of TReg in pregnancy, one would predict that this should lead to a more immunosuppressed immune environment compatible with materno-fetal tolerance. An intriguing possibility is that the increase in regulatory cells reported here provides an explanation for the observation that a number of autoimmune conditions, such as systemic lupus erythematosus and rheumatoid arthritis, tend to remit during pregnancy.25,26 Since, once activated, TReg do not appear to act in an antigen-specific manner, allospecific TReg may be able to regulate T cells of other specificities, including those promoting autoimmunity.
The mechanism by which TReg could modulate responses during pregnancy is not yet established. However, recent studies in mice have demonstrated that tolerance to alloantigens is essential for successful pregnancy. Studies by Munn et al. demonstrated that allogeneic, but not syngeneic, concepti were aborted in a T-cell-dependent manner when the enzyme IDO was inhibited.18,27 Very recently studies have shown that soluble CTLA-4 or TReg expressing CTLA-4 can stimulate the up-regulation of IDO and the catabolism of tryptophan.21 Our studies could now provide a link between IDO activity and CTLA-4+ TReg cells in pregnancy. Thus increasing numbers of TReg could stimulate IDO expression in antigen-presenting cells and thereby decrease tryptophan levels. In support of this concept it has been shown that levels of tryptophan decline with each trimester during normal pregnancy.28 These data fit well with the increase in TReg numbers observed in the present study.
In conclusion, the results presented here support our hypothesis, and it seems plausible that during pregnancy an increase in the population of TReg cells plays a role in maternal immune tolerance of her allogeneic conceptus. Studies are now in progress to characterize the phenotype and function of TReg cells in both normal and compromised human pregnancies in more detail.
Whilst this manuscript was under revision, the paper by Aluvihare et al. reported findings highly consistent with our own, using a murine model.29 These data also show an increase in TReg in pregnant mice with gestation. However, they demonstrate that TReg are elevated irrespective of whether the fetus is allogeneic or syngeneic. Nonetheless, depletion of CD25+ T cells resulted in the selective abortion of allogeneic concepti. Overall these data are highly consistent with our findings in humans and suggest that elevation of regulatory T cells in normal pregnancy provides immune protection for the fetus.
The authors would like to thank the University of Birmingham Scientific Projects Committee for supporting this work. Y.Z. was funded by the BBSRC and the Wellcome Trust. D.M.S. was funded by the Arthritis Research Campaign and is an ARC Senior Research Fellow.