The gender gap in autoimmunity and the contribution of sex hormones (estrogens, progestogens, androgens, and prolactin) to the pathogenesis of systemic lupus erythematosus (SLE) is the frequent subject of hypothesis and review (1–3). Few studies have, however, taken advantage of dramatic hormonal shifts during pregnancy and the postpartum period to examine potential relationships of these hormones to modulation of autoimmune disease. Recently, this physiologic model of dramatic immunoendocrine change has been revisited for rheumatoid arthritis(4) and multiple sclerosis(5), with the demonstration of clinically significant modulation of autoimmune disease by pregnancy or breast-feeding. In this issue of Arthritis Care & Research, Doria et al, in a systematic and illuminating investigation, confirm and extend observations of hormonal and immune changes occurring during pregnancy in patients with SLE(6). Although the issue of clinical SLE flare during pregnancy may be debated(7, 8) and confounded by therapeutic intervention, an incontrovertible fact is that dramatic hormonal shifts and altered immune and autoimmune parameters can occur. Of particular intrigue in the study by Doria et al(6) was the relatively unexpected finding of extraordinarily low mean serum concentrations of estrogen, progesterone, and testosterone in lupus pregnancies, compared with pregnancies in healthy individuals. This study confirms the earlier, similar findings of Jara-Quezada et al that demonstrated low serum estrogen concentrations in pregnant lupus patients with active disease(9). Furthermore, disease activity was low in the investigation by Doria et al, and worsened progressively throughout pregnancy in the presence of low mean serum dehydroepiandrosterone sulfate (DHEAS), testosterone, estrogen, and progesterone, with the exception of third-trimester disease activity. During the third trimester, the mean disease activity score partially regressed coincident with rising steroid hormone concentrations. Interestingly, corticosteroid treatment raised serum concentrations slightly towards more normal levels, but did not significantly alter the “abnormally” low levels. It would have been of interest to compare hormone concentrations in patients with active versus inactive disease, or high versus low SLE disease activity, particularly in relation to corticosteroid treatment.

Several aspects of this study (6) deserve attention and further investigation. First, the pattern of steroid hormones throughout pregnancy in a patient with lupus mirrors that of normal pregnancies, albeit at a much lower level, raising the question as to why gonadal steroids are suppressed in lupus pregnancies. Second, testosterone, 17β-estradiol, and progesterone moved in tandem throughout lupus pregnancy, suggesting a common regulatory process. Third, the highest levels of estrogen and progesterone occurred during the third trimester, a period of relative immunosuppression as documented by a reduction in mean SLE disease activity score and immunoglobulin concentration. This “snapshot” finding precipitates a basic question: does estrogen or progesterone suppress SLE, or does a drop in SLE disease activity lead to increased sex steroid concentrations? Fourth, in the postpartum period, when steroid levels fall again, SLE disease activity and immunoglobulins rise. This inverse relationship challenges the traditional thinking that increased concentrations of female gonadal steroids (i.e., estrogen or progesterone) directly stimulate SLE autoimmune responses. Conversely, most of the results in the study by Doria and colleagues remain consistent with the hypothesis that androgens (DHEAS and testosterone) suppress SLE disease activity. In the study by Doria et al, neither cytokines nor prolactin, another potent immunoendocrine hormone markedly altered by pregnancy and the postpartum, were measured. Although studies with innumerable variables must be focused, it must also be realized that cytokines impact the pituitary–gonadal axis and are altered during pregnancy and the postpartum(10). Moreover, in Jara-Quezada et al's study(9), higher prolactin concentrations occurred in pregnant active patients with SLE compared to control pregnancies. Prolactin has been shown to correlate with pregnant lupus disease activity(9, 11), and is known to suppress sex steroid concentrations(3, 10). As with most novel research findings, the results of Doria et al raise more questions than they answer. The reason for low levels of sex steroids in pregnant patients with SLE is unknown. Moreover, the possibilities must now be considered that estrogen or progesterone may be immunosuppressive to lupus disease activity, may normally stimulate an immunosuppressive cell or humoral factor, may be suppressed by inflammatory cytokines or prolactin, or may be directly suppressed by disease activity.

A multitude of complex issues pertinent to immunoendocrine control of lupus and lupus pregnancy remain to be examined. These complexities include, but are not limited to, the following:

  • 1
    roles of estrogen congeners and metabolites (i.e., estrone, estriol, methoxyestradiol, and 16α to 2α hydroxyestradiol ratios);
  • 2
    interconversions, or lack thereof, between steroids during lupus pregnancy (progesterone/DHEAS/testosterone can all be converted to 17β−estradiol);
  • 3
    potential alterations/abnormalities of metabolic enzymes involved in hormone metabolism (such as increased oxidation of testosterone);
  • 4
    pituitary–gonadal and pituitary–cytokine axes aberrations in pregnant lupus patients;
  • 5
    potential roles for human chorionic gonadotropin or human placental lactogen during lupus pregnancies;
  • 6
    temporal relationships of any cause-and-effect associations (as sex steroids normally act over days, weeks, months);
  • 7
    hormone- and cell-specific effects (3) that may independently influence final endpoint read-outs such as clinical disease activity compared to serologic manifestations.

Although non-hormonal explanations are possible, the study by Doria et al (6), and recent studies in rheumatoid arthritis(4) and multiple sclerosis(5), clearly demonstrate significant immunoendocrine changes in pregnant and postpartum patients with autoimmune disease, making conclusions regarding cause-and-effect hormonal immunomodulation nearly irresistible. However, others have suggested that there is no association between estrogen, progesterone, and immune responses (as assessed by cytokines) in patients with lupus(12). These descriptive investigations cry out for integration of endocrinologic, cellular, and molecular explanations for these phenomena. Doria et al provide a foundation from which to launch investigations to determine if sex hormones control lupus disease activity. Unfortunately, biologic explanations for clinical observations are almost always more complex than they appear.


  1. Top of page