Systemic lupus erythematosus (SLE) is a chronic multisystem autoimmune disorder that predominantly affects women of reproductive age. As clinical outcomes improve, pregnancy in these women is becoming more common. Although epidemiological data have documented an improvement in the prognosis of pregnancy in these women over recent years, they are still at significantly increased risk of pregnancy complications, such as miscarriage, stillbirth, pre-eclampsia and impaired foetal growth. The pathogenesis of SLE involves marked immune dysfunction, and in particular, the function of immunosuppressive elements of the immune system is impaired, including regulatory T-cell function. Because regulatory T cells are likely to be the key cell-modulating feto-maternal tolerance, this review overviews the possibility that regulatory T-cell impairments contribute to pregnancy pathology in women with SLE and contribute to the clinical challenge of managing these women during pregnancy.
Systemic lupus erythematosus (SLE) is a chronic multisystem autoimmune disorder that predominantly affects women of reproductive age in a male to female ratio of 1:9. Virtually, any system within the body can be affected, with skin and joint manifestations being commonest, but other clinical presentations include renal, haematological and more rarely pulmonary, cardiac and neurological complications. For many years, it was associated with uniformly poor pregnancy outcomes, and women were generally dissuaded from pregnancy. With improvements in the clinical management of SLE, and possibly the diagnosis of milder forms of the disease, pregnancy outcomes have significantly improved. SLE is characterized by periods of remission and flare that require close clinical follow-up and treatment with immunosuppressant drugs, including steroids and in some cases, the newer biological agents such as B-cell-targeted therapies. The clinical challenge now is to understand the disease process better to enable improved prediction of pregnancy outcomes, with earlier interventions when problems arise, and therefore tailored management of pregnancy.
Pathogenesis of SLE
Systemic lupus erythematosus is a typical autoimmune disorder in that antibodies are generated against self-antigens, and in the case of SLE, self-antigens include DNA and other nucleosomic elements. The presence of these autoantibodies forms part of the classification criteria and aid in the diagnosis. Autoantibodies, especially anti-double-stranded DNA, may in some cases reflect disease activity, whilst others can reflect or predict organ involvement and clinical manifestations, such as the association between anti-Ro (anti-SSA) and photosensitivity, and antiphospholipid antibodies with clotting risk. Deranged function of both the innate and adaptive immune system contributes to this lack of immune tolerance. The clear pathogenesis leading to this loss of tolerance has not been clearly elucidated but is likely the result of both a genetic predisposition and environmental influences. As a result, abnormalities in many immune cell and cytokine pathways have been described, with T-cell-dependant autoantibody production by clonal B cells being a central feature (Fig. 1). It has been suggested that under the influence of increased levels of type 1 interferons, monocytes in the serum of patients with SLE differentiate into myeloid dendritic cells, which function as professional antigen-presenting cells. These cells can take up circulating apoptotic fragments, including nuclear fragments and nucleosomes, and present them to CD4+ T cells. This then leads to T-cell activation and subsequent clonal B-cell expansion leading to production of autoantibodies against the nucleic fragments. These autoantibodies subsequently cause tissue damage through deposition of antigen-antibody complexes within basement membranes. Whether this is the direct cause of cellular damage (e.g. the complex is directly pathogenic) or whether damage occurs secondary to complement and cytokine activation is as yet unclear.
The role of T cell ‘help’ in the generation of autoantibodies and subsequent tissue damage is critical. Autoantibodies are, however, also known to be present in healthy individuals. For example, anti-nuclear antibodies have been reported to be detectable in up to 25% of some populations without SLE, with 2.5% having significantly high ‘positive’ titres. The comparative rarity of SLE suggests that not all these individuals go on to develop clinically significant disease, although some will be at risk of doing so. It has been proposed that it is T-cell activity and function that determines how an individual responds to the generation of these autoantibodies. Therefore, it is likely that B cells expressing autoreactive antibodies are subject to negative selection during differentiation in healthy individuals. T cells are involved in interactions with antigen-presenting cells and with B cells, via specific signalling pathways and cytokine production. Thus, T cells determine the generation of specific high-affinity autoantibodies and also regulate cytokine production, which contributes to tissue damage. Both T-cell signalling and cytokine levels are deranged in SLE. In particular, serum levels of interleukin-10 and interferon-α are raised in SLE and levels correlate with disease activity.[9, 10]
The immune system's capability to self-regulate via T cells is also compromised in SLE. A key cell type involved in this self-regulation is regulatory T cells (Treg cells). Treg cells were initially described in the late 1990s and are usually identified by positivity for CD4, CD25 and FOXP3. In an intact immune system, Treg cells provide an immunosuppressive function, via inhibition of IL-2 production, inhibition of CD4+ and CD8+ T-cell proliferation and suppression of the function of antigen-presenting cells (Fig. 1). The likely role of Treg cells in autoimmune disease is evident from rodent models in which CD4+ CD25+ cell numbers are reduced. These mice develop severe autoimmune pathology that is improved when these cells are replaced.[13, 14] These cells have also been investigated in lupus-prone mouse strains. BWF1 lupus-prone mice have been shown to have Treg cell numbers 40–50% lower than their healthy counterparts. Work in the Scurfy mouse demonstrated that FOXP3 was vital for immune regulation, as mice carrying a mutation in this gene suffered a fatal overproliferation of CD4+ CD8− T lymphocytes, a phenotype similar to mice lacking transforming growth factor-β1 (TGF β1). Furthermore, it has been shown that transfer of ex vivo generated Treg cells into lupus-prone mouse models has the ability to both reduce disease progression and improve survival.
Whilst the role of Treg cells in murine models of lupus is beyond dispute, the evidence for disruption of Treg cells numbers and function in human SLE has been conflicting. Variation in these reports reflects a combination of different clinical disease phenotypes and treatments, together with differing laboratory protocols for the identification of this cell subtype. This topic has been reviewed elsewhere, and overall, most studies point to a reduction in Treg cell number and function in SLE.[19-24] T helper-17 cells (Th17 cells) are a further class of T cell, defined by generation of IL-17, with important actions in SLE and likely to play a role in the pathogenesis. Th17 are considered pro-inflammatory, as IL-17 is a cytokine capable of stimulating the inflammatory response through chemokine and cytokine production, and proliferation and recruitment of neutrophils, macrophages and lymphocytes. Several studies have reported an increase in Th17 cells and IL17 in SLE, and in particular with disease flare.[27-30] It therefore seems that SLE is associated with a reduction in the levels and function of immunosuppressive Treg cells together with an increase in the pro-inflammatory Th17 cells (Fig. 1).
The differentiation of T cells into Treg cells or Th17 cells is also linked, as both require the presence of TGFβ1 (Fig. 1). TGFβ1 is known to be critical for the activation of Treg cells and the expression of FOXP3. However, in the copresence of IL-6 predominant differentiation into the Th17 subtype occurs. Furthermore, there is also evidence that Treg cells are capable of differentiating into Th17 cells. Sharma and coauthors demonstrated that blockade of indolamine 2,3, dehydrogenase resulted in conversion of Tregs into cells with the Th17 phenotype. It is well documented that low levels of TGFβ1 correlate with increased SLE disease activity and with cardiovascular complications of the disease. Furthermore, IL-6 activity is also likely to be increased in SLE. Thus, it is probable that SLE is associated with a reduction in immunosuppressive capability via Tregs and TGFβ1 and a concomitant increase in proinflammatory factors such as Th17 cells and their cytokines.
Immunological changes during pregnancy
A successful pregnancy in the face of the immunological challenge conferred by the implantation of the semi-allogenic foetus requires the presence of an intact and functioning immune system. The maternal immune system must adapt locally within the uterus to facilitate placentation and but also systemic changes are required as the maternal immune system must display tolerance, whilst maintaining competence against, for example, infection. Given the considerable immunological derangements observed in women with SLE, it is not surprising that SLE is associated with complications during pregnancy.
It is becoming apparent that Treg cells are critical in the immune tolerance required to facilitate a healthy pregnancy, both locally in the uterus, and in the systemic circulation, and this topic has been reviewed in detail by others. In mice, pregnancy is associated with an increase in Treg cells both in the uterus and in the systemic circulation, and there is now evidence that this Treg response is in part driven by paternal antigens. Treg differentiation can occur via two different pathways; the ‘thymic’, which generates tTreg or the peripheral route (or extrathymic), which generates pTreg (sometimes referred to as inducible or iTreg). In mice, these pathways may generate subclasses of Treg cells with different functions, with tTregs functioning to suppress autoimmunity, whilst pTregs act to control immune responses to traditional antigens including bacteria and allergens. Generation of Tregs seems to be controlled by a FOXP3 enhancer element, which is conserved in mammalian species. Experiments have shown that pTregs generated in response to pregnancy recognize the paternal antigens and also suppress the maternal immune effector cells. In mice lacking the enhancer element, no pTreg cell induction occurred and activated T cells invade the placenta with subsequent pregnancy loss.
Assessment of Treg cells during human pregnancy has been subject to the similar variability as seen in the SLE studies of Treg cells. Initial work suggested that Treg cells increase in the first trimester, peak in the second trimester before falling post-partum. We and others have shown a similar increase in Treg cell numbers with pregnancy.[44-46] In contrast, others have suggested no change in Tregs, or a fall with gestation.[47, 48] Several authors have also documented lower Treg cells and higher Th17 activity, both circulating and in the uterine decidua, in women suffering complications of pregnancy, such as miscarriage.[45, 49-53] Of particular interest is an early clinical trial in which patients with a history of recurrent miscarriage and low Treg cells were administered granulocyte colony-stimulating factor. Women randomized to receive this treatment had an increase in Treg cells and, importantly, an increased live birth rate. This study provides further evidence that immune tolerance mediated by Tregs is vital in allowing the development of a healthy pregnancy, and clearly, larger studies are now needed to examine this further. Altered Treg and Th17 cells have also been reported in women with pre-eclampsia. Pre-eclampsia is a third trimester complication presenting with hypertension and proteinuria, strongly associated with faulty placental development and a likely maladaptation to the necessary cardiovascular vasodilation of pregnancy. Our own as yet unpublished longitudinal observations have shown that women with SLE have lower Treg cell numbers than healthy women, and that in pregnancy, this pattern persists. This coincides with a lower capacity to activate TGFβ1, which is the key cytokine required for FOXP3 activity. Taken together, evidence suggests that women with SLE have a dysfunctional tolerogenic capability, which compromises their adaptation to pregnancy. This may underpin the increased risks of pregnancy in this cohort and offers clinical possibilities for intervention in the future.[56, 57]
Pregnancy complications and SLE – the clinical challenge
As survival improves in SLE and as better disease control is possible through more effective therapy, pregnancy in women with SLE is increasingly contemplated and, indeed, is no longer actively discouraged. It is currently estimated that SLE affects approximately 2/1000 pregnancies. Whilst early studies suggested pregnancy was associated with a very high rate of complications with a live-term birth rate in only 28% and a pre-term delivery rate approaching 50%, more recent studies have suggested a much improved prognosis, with live-term birth rates in the region of 70%. Although prognosis has improved, pregnancy in women with SLE remains high risk, as demonstrated by a recent systematic review of the literature, which reported that 25% experience a lupus flare, 16% develop hypertension and 7% develop pre-eclampsia. The risks of foetal complications were similarly high, with women with SLE having a 16% risk of miscarriage, 3% risk of stillbirth, 2% risk of neonatal death and a risk of intrauterine growth restriction of 12%. Patients with active disease at conception, those with nephritis and those with the antiphospholipid syndrome (APS) are at the highest risk of complications developing.[61, 62]
Antiphospholipid syndrome was first described in a group of patients with SLE with raised anticardiolipin antibody levels and clinical features of recurrent thrombosis in the 1980s. APS is associated with arterial or venous thrombosis and pregnancy complications such as miscarriage, preterm delivery, stillbirth and pre-eclampsia. Approximately 30–40% of women with SLE will also have antiphospholipid antibodies (APA).[64, 65] However, not all these women will have a history of thrombosis or pregnancy complications. Autoantibodies are generated against negatively charged phospholipids, a component of the cell membrane. Functionally, such antibodies cause prolongation of phospholipid-dependant coagulation assays, the so-called ‘lupus anticoagulant’ (LA). The association with thrombosis is stronger with LA than with anticardiolipin antibodies (aCL), and this may reflect the more ‘functional’ nature of the assay systems. In a meta-analysis of 25 studies involving more than 7000 patients with APA, the mean odds ratio for thrombosis was 1.6 for aCL and 11.0 for LA. A more recent study in SLE has also suggested that LA positivity is particularly associated with poor pregnancy outcomes. The PROMISSE study was a prospective study that examined pregnancy outcomes in 144 women with APA and/or SLE. This study documented inconsistency between laboratories in the measurement of these antibodies, emphasizing the difficulty in diagnosis. Nevertheless, the presence of the LA was the primary predictor of pregnancy complications, with 43% of women in this group suffering an adverse pregnancy outcome. Both a diagnosis of SLE and a history of previous thrombosis were also predictors in the multivariate analysis.
Systemic lupus erythematosus is therefore undoubtedly associated with adverse pregnancy outcomes. However, more extensive assessment and measurement of autoantibodies has probably resulted in the diagnosis of milder and much earlier disease, often associated with milder symptoms without deep organ involvement. It seems unlikely that many of these women are at substantial risk of pregnancy complications, but our evidence base in this population is insufficient. Nevertheless, the diagnosis produces considerable anxiety for the clinician and patient, and there is a lack of evidence base to inform management. The current clinical challenge is evidenced by the two recent cases seen on the same day in our service (Box 1). Case 1 is a typical case of SLE complicated by lupus nephritis in remission following Rituximab therapy (anti-B-cell agent). Several clinical features raise concerns regarding the prognosis of pregnancy, namely, previous nephritis, obesity, increased maternal age and continuing need for maintenance immunosuppression. However, the absence of hypertension and aCL/LA would be good prognostic indicators. The difficulties with attempting clinical predictions are also highlighted by case 2 (Box 2). Case 2 also had a history of lupus nephritis, but also had the presence of APS with a previous thrombosis. These would all be poor prognostic indicators for pregnancy. In spite of this, the pregnancy progressed without complication on subcutaneous low-molecular-weight heparin and aspirin together with regular antenatal follow-up, including assessment of placental function with ultrasound, and monitoring for blood pressure and proteinuria. The pregnancy had a successful outcome, and she had a normal delivery of a baby girl with a normal birth weight centile for gestation and ethnic group. These two cases illustrate that in current clinical practice, we have limited strategies to be able to accurately predict outcomes and therefore target surveillance and treatments. A clearer understanding of the immunology of SLE in the context of pregnancy may help the development of algorithms and clinical tools to facilitate better stratification of women either pre-conceptionally and/or in the early antenatal period.
Box 1. Case 1
This 37 years old attended the clinic for preconception counselling. She had a 15-year history of SLE including a history of biopsy proven class IV lupus nephritis, which was in remission following treatment with Rituximab (anti-B-cell agent) and maintenance Cyclosporine and Azathioprine (immunosuppressive agents). She also had clinical features of mixed connective tissue disorder in addition to her diagnosis of SLE, including severe Raynaud's Phenomenon, facial telangiectasia and oesophageal dysmotility. Her disease had been stable for about 12 months. She had no previous pregnancies, no history of hypertension or previous thrombosis. All previous tests for aCL and LA were negative. She was also negative for anti-SSA/B (associated with foetal heart block). Current treatment included prostacyclin infusions to treat Raynaud's syndrome, prednisolone 10 mg/day, azathioprine 100 mg/day, cyclosporin 75 mg twice/day, ramipril for renal protection, aspirin and hydroxychloroquine. She also had a body mass index (BMI) of 38 kg/m2 and markers for disease activity demonstrated quiescence, and she had no proteinuria. She wanted to know whether pregnancy was possible and likely to be successful.
Box 2. Case 2
This 35 years old was initially referred in from her local hospital at 12-week gestation. She had been diagnosed with SLE 10 years previously. It was her first pregnancy and her BMI was 28 kg/m2. She had a diagnosis of biopsy proven class IV lupus nephritis, which was in remission following a combination of steroids and azathioprine. She also had APS with positivity for the LA and a previous deep vein thrombosis and pulmonary embolus. She was therefore on lifelong anticoagulation with warfarin, which had been converted to low-molecular-weight heparin when pregnancy was confirmed. She was positive for anti-SSA. She was managed throughout the pregnancy on aspirin and low-molecular-weight heparin. The baby demonstrated a normal heart rate throughout and the growth of the fetus was in the normal range at her last ultrasound scan at 36 weeks. She was remained normotensive throughout with no development of proteinuria. She went into labour spontaneously at 38-week gestation and delivered a healthy girl weighing 2.65 kg.
This research was supported by the NIHR Wellcome Trust Manchester Clinical Research Facility, Lupus UK and Greater Manchester Collaboration for Leadership in Applied Health Research and Care (CLAHRC) and facilitated by the Manchester Biomedical Research Centre and the Greater Manchester Comprehensive Local Research Network. Professor Bruce is supported by Arthritis Research UK, The Manchester Academic Health Science Centre, the National Institute for Health Research (NIHR) Biomedical Research Unit Funding Scheme and the NIHR Manchester Biomedical Research Centre. The views expressed in this publication are those of the author(s) and not necessarily those of the NHS, the National Institute for Health Research or the Department of Health.