Pregnancy, cytokines, and disease activity in systemic lupus erythematosus

Authors


Abstract

Objective

To evaluate levels of selected cytokines and soluble receptors involved in the humoral immune response during pregnancy in systemic lupus erythematosus (SLE) patients.

Methods

Seventeen consecutive SLE patients and 8 matched healthy controls were prospectively studied during pregnancy. Sera were obtained within the last 3 months prior to pregnancy; at 9, 17, and 29 weeks of pregnancy; and at 1 month after delivery. Serum levels of interleukin-10 (IL-10), interleukin-6 (IL-6), and soluble tumor necrosis factor receptors p55 (sTNFR I) and p75 (sTNFR II) were evaluated. SLE activity was measured by the European Consensus Lupus Activity Measurement score modified for pregnancy.

Results

IL-10 serum levels were found to be higher (P < 0.0001) in patients than in controls before conception, and still higher (P < 0.0001) in SLE patients during gestation, without intertrimester changes. In SLE patients, IL-6 serum levels did not increase in the third trimester of pregnancy, as was observed in controls (P = 0.011). No significant differences between SLE patients and controls were found in either sTNFR I or II levels or profiles before and during pregnancy. IL-10 and sTNFR I levels were significantly higher during pregnancy and postpartum in SLE patients with active disease (P = 0.03 and P = 0.01, respectively).

Conclusion

The levels of some cytokines involved in the humoral immune response seem to be modified in the peripheral circulation of pregnant SLE patients. The most relevant modifications are the lower than expected increase of IL-6 in the third trimester of gestation and persistently high levels of IL-10 during pregnancy.

INTRODUCTION

Pregnancy is a physiologic condition in which several immunoendocrine changes occur to achieve immunosuppression and tolerance by the immune system to paternal and fetal antigens (1, 2). The most relevant immunologic effect in normal pregnancy seems to be a steroid-induced T helper type 2 (Th2) cytokine polarization (1). Changes in concentrations of cytokines and soluble cytokine receptors toward a switch from Th1 to Th2 cytokine production have already been described at the maternal-fetal interface and in the maternal circulation (3, 4). The consequence is that cellular immunity and Th1 cytokines are inhibited, whereas humoral immunity, antibody production, and Th2-type cytokines are enhanced (1, 5–7).

Systemic lupus erythematosus (SLE) is considered to be a Th2 cytokine-driven disease (1, 8) due to the overexpression of Th2-type cytokines (5, 8–10). Among these cytokines, interleukin 10 (IL-10) seems to play a central role in the pathogenesis of SLE as well as in disease flare induction (11–13). Interleukin 6 (IL-6) is a potent contributor to the differentiation of Th0 into Th2 cells (14) and it has also been reported that IL-6 production increases in patients with active disease (9). Other markers of cellular immune activation in SLE, which also correlate with disease activity, are soluble tumor necrosis factor α receptors p55 (sTNFR I) and p75 (sTNFR II) (15–18). Both sTNFRs are considered surrogate markers of TNFα production and inhibitors of TNFα activity in vivo. It is worthy to note that TNFα levels are increased in the blood of SLE patients (17, 18) and it has been suggested that it plays a role in the pathogenesis of the disease (19).

Up to now, the behavior of such antiinflammatory cytokines as an expression of immune activation and SLE modulation during pregnancy has been poorly investigated (20–23). Recent data provide evidence of significantly low mean serum concentrations of steroid hormones in the third trimester of pregnancy in SLE patients (23, 24), which could account for the surprisingly low percentage of SLE relapses observed by us (23) and others (25–28), but not all (29), in this period of gestation. Possible cause–effect relationships between sex hormones and immune response may exist and are a matter of investigation (22).

The aim of our study was to evaluate some selected markers of humoral immune response during pregnancy in the systemic circulation of SLE patients to identify which of them could account for the decrease in disease activity observed during the third trimester of pregnancy (23).

PATIENTS AND METHODS

Seventeen consecutive successful pregnancies in 17 patients with SLE were prospectively studied. All patients (mean age 30.5 years, range 21–38; mean disease duration 91 months, range 18–184) fulfilled the American College of Rheumatology classification criteria for SLE (30). Eight pregnancies in 8 healthy volunteers (mean age 29 years, range 20–35) served as controls. Gestational period was similar in SLE patients (mean 37 weeks, range 31–41) and controls (mean 38 weeks, range 35–41).

All patients were enrolled in our protocol for pregnancy planning and followup, which was detailed previously (23). Briefly, patients were evaluated monthly by the same rheumatologist and obstetrician during their entire pregnancy, including the 10–12 weeks of the postpartum period. Twelve of 17 pregnancies were fully planned and 5 pregnancies were unexpected. According to our protocol, the pregnancy was planned when disease was inactive for at least 6 months.

In each clinical evaluation, routine laboratory tests were performed, including white blood cell count; urinalysis; and glucose, blood urea nitrogen, serum creatinine, IgG, IgM, and IgA levels. Antinuclear antibodies and anti– double stranded DNA (anti-dsDNA) antibodies were detected by indirect immunofluorescence using as substrate HEp-2 cells and Crithidia luciliae, respectively. Anti-extractable nuclear antigen antibodies were detected by counterimmunoelectrophoresis; serum complement fractions C3 and C4 and C-reactive protein were detected by nephelometry; anticardiolipin antibodies were detected by enzyme-linked immunosorbent assay (ELISA); and lupus anticoagulant by Russell viper venom time assay.

SLE activity was measured by European Consensus Lupus Activity Measure (ECLAM) score (31), modified for pregnancy as previously indicated (23). The disease was arbitrarily considered active during pregnancy in those patients in whom the mean value of ECLAM scores, obtained at each clinical evaluation during gestation and 1 month after delivery, was ≥2.

The laboratory and clinical findings of our patients at conception and data regarding pregnancy outcome were previously reported (23). For the purpose of this study, we considered only those pregnancies that ended in live births.

Cytokine assays.

Serum samples for cytokine testing were collected in the following periods: within the last 3 months prior to pregnancy in the 12 planned pregnancies; at 9, 17, and 29 weeks of pregnancy; and at 1 month after delivery in all cases. All serum samples were kept frozen (–80°C) until the cytokine assays were performed.

Serum concentrations of cytokines were assayed by ELISA methods: IL-10, Amersham Pharmacia Biotech (Uppsala, Sweden), limit of detection 0.1 pg/ml; IL-6, Tell Sciences (Needham, MA), limit of detection 7 pg/ml; serum sTNFR I and sTNFR II, Bender MedSystems (Vienna, Austria), limits of detection 80 pg/ml and 0.15 ng/ml, respectively. These assays were performed according to the manufacturers' instructions.

Statistical analysis.

Programs from the BMDP (La Jolla, CA) statistical package were used for calculations (32). A two-tailed analysis was used. The analysis of variance (ANOVA) for repeated measures was performed on protocol variables, considering the periods of data collection as within-subject factors and clinical classification (SLE, controls, active and inactive disease) as grouping factors. Within-group contrasts were performed according to Bonferroni's method. Results of ANOVA are presented reporting 3 F test values, namely FG for grouping factor, FT for time, and FGXT for the interaction between time and grouping factors. The BMDP program 6D was used for linear correlation and regression analysis.

RESULTS

Data regarding clinical manifestations, laboratory findings, and treatment of our patients during pregnancy and postpartum were detailed previously (23).

Because most of the pregnancies (70%) were planned on the basis of inactive disease, the patient mean ECLAM score during pregnancy was low. However, the disease was considered active during pregnancy in 9 patients (53%) in whom the mean value of ECLAM scores, obtained at each clinical evaluation during gestation and 1 month after delivery, was ≥2. Patients with active disease were taking a prednisone (mean ± SD) dosage higher than that taken by patients with inactive disease: 15.8 ± 5.6 mg/day, range 10–28, versus 2.9 ± 2.4 mg/day, range 0–5 (P = 0.0001).

Cytokines.

IL-10, IL-6, and sTNFR mean values, before and after pregnancy, are reported in Table 1. IL-10 serum levels were found significantly and persistently higher in SLE patients than in healthy women, both before (Table 1) and during (Figure 1) the gestational period. IL-10 levels did not vary throughout pregnancy in SLE patients, whereas in controls they were significantly higher in the third trimester than before conception (P = 0.029) or the first trimester (P = 0.038) (Table 1), suggesting an influence of physiologic steroid hormone increase, at least in the last period of normal pregnancy. During pregnancy, IL-10 serum levels were significantly higher (FG 5.25; P = 0.03) in SLE patients with active disease than in those with inactive disease (Figure 1).

Table 1. Serum IL-10, IL-6, and sTNFR levels before, during, and after pregnancy*
 IL-10 (pg/ml)IL-6 (pg/ml)sTNFR I (ng/ml)sTNFR II (ng/ml)
SLEControlsSLEControlsSLEControlsSLEControls
  • *

    Data presented are mean ± SD. We considered the sera of all 17 patients during pregnancy and postpartum, whereas before pregnancy we considered only the sera of 12 patients in whom pregnancy was fully planned. IL-10 = interleukin 10; IL-6 = interleukin 6; sTNFR I = soluble tumor necrosis factor receptor p 55; sTNFR II = soluble tumor necrosis factor receptor p75; SLE = systemic lupus erythematosus.

  • IL-10 in prepregnancy: SLE versus controls P < 0.0001.

  • IL-10 serum levels in controls: third trimester versus first trimester P = 0.038; third trimester versus prepregnancy P = 0.029.

  • §

    sTNFRI in prepregnancy: SLE versus controls P = 0.043.

  • IL-6 in the third trimester of pregnancy: SLE versus controls P = 0.006.

  • #

    sTNFR I in postpartum: SLE versus controls P = 0.040.

Prepregnancy3.00 ± 0.900.91 ± 0.432.95 ± 3.551.29 ± 0.730.26 ± 0.08§0.19 ± 0.040.61 ± 0.280.65 ± 0.46
First trimester3.02 ± 1.060.92 ± 0.502.67 ± 2.911.55 ± 0.830.26 ± 0.100.27 ± 0.050.77 ± 0.410.65 ± 0.23
Second trimester2.87 ± 1.041.33 ± 0.641.86 ± 1.582.91 ± 2.890-29 ± 0.080.31 ± 0.030.93 ± 0.630.75 ± 0.29
Third trimester3.27 ± 1.381.46 ± 0.682.32 ± 1.9513.82 ± 15.830.34 ± 0.090.42 ± 0.110.82 ± 0.320.89 ± 0.42
postpartum3.21 ± 1.251.43 ± 0.781.72 ± 1.296.10 ± 13.710.26 ± 0.06#0.21 ± 0.030.75 ± 0.380.70 ± 0.57
Figure 1.

Profiles over time of mean values of interleukin-10 (IL-10) and interleukin-6 (IL-6) serum levels during pregnancy and postpartum in systemic lupus erythematosus (SLE) patients, healthy controls, SLE patients with active disease, and SLE patients with inactive disease, evaluated by analysis of variance for repeated measures. FG = grouping factor; FT = time factor; FGXT = interaction between time and grouping factors; n.s. = not significant; Trim. = trimester.

Interestingly, SLE patients did not show an increase in IL-6 in the third trimester, which is a characteristic finding of normal pregnancy (P = 0.006; Table 1 and Figure 1). IL-6 level profiles were similar in patients with active and inactive disease (Figure 1).

Soluble TNFR I serum levels were significantly higher in patients than in controls before conception (P = 0.043) and postpartum (P = 0.040; Table 1), but they did not significantly differ during pregnancy, even though significant differences in sTNFR I kinetics were observed between patients and controls (FT 22.4, P < 0,0001; FGXT 3.9, P = 0.02; Figure 2). No differences were noted in either sTNFR II serum concentration or profile before pregnancy, during gestation, or postpartum.

Figure 2.

Profiles over time of mean values of soluble tumor necrosis factor α p55 (sTNFR I) and p75 (sTNFR II) serum levels during pregnancy and postpartum in systemic lupus erythematosus (SLE) patients, healthy controls, SLE patients with active disease, and SLE patients with inactive disease, evaluated by the analysis of variance for repeated measures. FG = grouping factor; FT = time factor; FGXT = interaction between time and grouping factors; n.s. = not significant; Trim. = trimester.

As for IL-10, serum levels of sTNFR I were found significantly higher in patients with active compared with inactive disease (Figure 2). We did not observe any correlation between IL-10, IL-6, sTNFR I, or sTNFR II and ECLAM score, leukocyte and lymphocyte count, gammaglobulin, IgG, IgA, IgM, or ANA titer within any distinct gestational period.

DISCUSSION

In the present study, we analyzed the alterations of some cytokines involved in immune-inflammatory activation in pregnant SLE patients. IL-10, IL-6, and sTNFR are implicated in the pathogenesis of SLE (8–10, 19, 20, 33) inasmuch as they are considered markers of disease activity, even though by different and independent pathways (33).

Both pregnancy and SLE are conditions in which the cytokine environment is predominantly Th2 polarized, with a progressive inhibition of cell-mediated immunity and a consequent enhancement of antibody-mediated immune responses (1, 9, 34). The immunomodulatory effects of pregnancy on a Th2-cytokine driven disease such as SLE are not understood. Up to now, IL-10, IL-6, and sTNFR have been poorly studied in lupus as well as in healthy pregnancy.

Given the overexpression of Th2-type cytokines in SLE, in particular IL-10 (8, 10–12, 35), a disease flare in pregnant patients affected with SLE would be frequently expected, particularly in the third trimester. However, previous prospective studies report a low percentage of flares in the third trimester of SLE pregnancy (25–28); our previous article (23) is consistent with such reports.

IL-10 decreases macrophage activation and antigen presentation, thereby directly and indirectly inhibiting T-cell functions (36–38). At the same time, IL-10 is a potent stimulator of B lymphocytes (39) and anti-DNA antibody production in SLE (13). The IL-10 overproduction observed in lupus patients (8, 10–12, 35) may derive, at least in part, from genetic polymorphisms (40–42). The ability to secrete IL-10 can vary according to the genetic composition of the IL-10 locus (43).

IL-10 serum levels do not directly reflect cellular IL-10 production (44). However, the IL-10 serum level (12, 44), more than IL-10 production (10, 45), correlates with disease activity and serum anti-dsDNA antibody titers. Thus, IL-10 serum level might be considered a marker of immune activation in SLE.

In keeping with previous data on IL-10 production in normal pregnancy (5, 46), IL-10 levels progressively increased in our healthy pregnant women. Conversely, the abnormally and persistently high serum levels of IL-10 in SLE patients, either before or during pregnancy, seem to support the view that the IL-10 overproduction in SLE is mainly constitutive rather than modulated by gonadal (steroid) hormones. IL-10 serum levels were persistently higher in our patients with active disease compared with those with inactive disease. Alternatively, this difference could be related to the corticosteroid dosage, which was higher in the former group compared with the latter. However, it is more likely that patients with a genetic up-regulation of IL-10 tend to have more active disease compared with patients with a different genetic background. In fact, IL-10 serum levels did not change during pregnancy in patients with either active or inactive disease, further confirming the lack of hormonal modulation influence.

IL-6 promotes Th2 differentiation and simultaneously inhibits Th1 polarization (14). In SLE patients, an increased IL-6 production (47), as well as increased IL-6 levels in peripheral circulation (44, 47), have been reported. Moreover, IL-6 production seems to be correlated with SLE activity (9). In normal pregnancy, IL-6 levels rise significantly in the second and still more in the third trimester, and there is an additional rise with the onset of labor (48). Our data concur with those of Opsjln et al (48). IL-6 levels progressively increased during pregnancy in healthy controls, reaching the highest levels in the third trimester. In pregnant SLE patients, we observed a lower than expected IL-6 increase, particularly in the last trimester of gestation. The effects of estrogens on IL-6 production appear to be dose dependent: physiologic concentration of estrogen inhibits IL-6 production, whereas pharmacologic levels of estrogens, like those observed during late pregnancy in healthy subjects, seem to drive IL-6 production in monocytes/macrophages (49). Therefore, the lower than expected IL-6 serum levels observed in our patients in the third trimester could be due to the low serum levels of estrogen and/or progesterone (23, 24) observed in the same period of pregnancy in SLE (Figure 3) and might further explain the low percentage of SLE relapses reported in this gestational phase (25–28).

Figure 3.

Parallelism between interleukin 6 (IL-6) and 17-βestradiol A, or progesterone B, serum level profiles during pregnancy and postpartum in both systemic lupus erythematosus (SLE) patients and healthy subjects. Data on 17-βestradiol and progesterone serum levels have been previously reported in the same group of patients (23). Trim. = trimester.

Circulating forms of TNFRs are soluble cytokine binding proteins consisting of the extracellular domains of the TNF receptors, obtained by proteolytic digestion. These inhibitors act as TNFα antagonists and provide effective protection, in experimental animal models, against deleterious effects of TNFα (50). Their increased concentrations probably reflect either enhanced biosynthesis or cleavage stimulation. However, little is known of the signals and mechanisms that regulate sTNFR levels. One of the possible inducers is TNF itself (50).

sTNFR serum levels are increased in SLE patients compared with healthy subjects, and both receptors (sTNFR I and sTNFR II) may correlate with disease activity (15–18). It has been observed, although in a relatively small number of cases, that normal pregnancy was accompanied by increased production of sTNFR, which could impair somehow the biologic activity of TNFα (51). No previous data are available on sTNFR profiles during pregnancy and postpartum in SLE patients.

In keeping with previous reports (9, 14), prepregnancy sTNFR I levels were higher in our SLE patients than the controls. Neither receptor correlated with disease activity. Because sTNFRs appear to be influenced by renal impairment, it is noteworthy that all of our patients, but not those considered in the above-mentioned studies (9, 14), had normal renal function.

Notably, sTNFR I levels vary during pregnancy in SLE patients and in controls (51). However, sTNFR I concentrations were found to be higher in pregnant patients with active disease than in those with inactive disease (Figure 2); this is the same trend as seen in nonpregnant SLE patients.

In conclusion, the levels of some cytokines involved in the humoral immune response seem to be modified in the peripheral circulation of SLE pregnant patients. The most relevant modification seems to be the lower than expected increase of IL-6 in the third trimester of gestation. The blunting of IL-6 might reflect the lower levels of estrogens or progesterone in the last trimester of pregnancy (23, 24) and might account for the low percentage of SLE relapses reported in the third trimester of pregnancy by some authors (25–28).

The abnormally and persistently high levels of IL-10 observed in SLE even during pregnancy, seem to support the view that IL-10 overproduction in SLE is constitutive, even if some changes related to disease activity or treatment are detectable. Finally, the similar profile of both sTNFRs in patients and in controls during pregnancy could suggest that in such condition the TNF/sTNFR balance is not substantially modified although, as for IL-10, changes of sTNFR I levels related to disease activity or treatment are to be expected.

Ancillary