- Email: [email protected]
Serum levels of soluble CD95 are not associated with amelioration of multiple sclerosis during pregnancy
Journal of the Neurological Sciences 252 (2007) 83 – 87 www.elsevier.com/locate/jns
Serum levels of soluble CD95 are not associated with amelioration of multiple sclerosis during pregnancy Stefan Ehrlich a , Judith Haas b , Frauke Zipp c , Carmen Infante-Duarte c,⁎ a
Department of Child and Adolescent Psychiatry, Charité - Universitaetsmedizin Berlin, Germany b Jewish Hospital Berlin, Germany c Institute of Neuroimmunology, Charité, Universitaetsmedizin Berlin, Germany
Received 30 December 2005; received in revised form 31 August 2006; accepted 13 October 2006
Abstract Multiple sclerosis (MS) is considered an autoimmune disease of the central nervous system. Like various other autoimmune disorders, MS normally improves during pregnancy. Pregnant MS patients experience a significant reduction in relapse rates and magnetic resonance (MR) disease activity. How sex steroid hormones affect disease course remains unclear. We hypothesized that hormonal changes during pregnancy might modulate the autoimmune response by enhancing apoptosis of autoreactive T lymphocytes. One of the most important effectors of apoptosis in T cells is the CD95/CD95L system. We have previously reported that the soluble form of CD95 (sCD95) can block CD95-mediated apoptosis and that MS patients show elevated levels of sCD95. Therefore, we considered whether gravidity might influence serum levels of sCD95 in patients, and analyzed the concentration of sCD95 in the sera of 61 patients with relapsing-remitting (RR) MS before, during and after pregnancy. We found no association between serum levels of sCD95 and pregnancy-related immune suppression in MS patients. Thus, sex steroid hormones do not seem to affect the production of anti-apoptotic sCD95. © 2006 Elsevier B.V. All rights reserved. Keywords: Multiple Sclerosis; Autoimmunity; Soluble CD95; T cell apoptosis; Pregnancy; Sex steroid hormones
1. Introduction Multiple sclerosis (MS) is a multifocal inflammatory disease of the brain and the spinal cord. The main cause of MS is unknown, although genetic and environmental factors have been shown to contribute to its etiology [1]. MS patients who become pregnant experience a significant reduction in relapse rates. The large, prospective European PRIMS (Pregnancy in Multiple Sclerosis) study showed an 87.5% decrease in the mean rate of relapse during the third trimester of pregnancy [2]. However, in the first three months postpartum relapse rate increases in 43% of the patients and returns to the pre-pregnancy level around the second trimester after delivery [3]. Furthermore, reduction in clinical relapse ⁎ Corresponding author. Institute of Neuroimmunology, Neuroscience Research Center, NWFZ 2680, Charité, 10098 Berlin, Germany. Tel.: +49 30 450 539028; fax: +49 30 450 539906. E-mail address: [email protected] (C. Infante-Duarte). 0022-510X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jns.2006.10.013
rate correlates with a decreased disease activity on magnetic resonance imaging (MRI) in pregnant MS patients [4]. This beneficial effect of pregnancy on the disease course has also been demonstrated in a rodent model of MS, experimental allergic encephalomyelitis (EAE) [5]. How pregnancy and sex hormones may influence MS is still unknown. MS is considered to be a T cell-mediated autoimmune disease [1] which might be promoted or even supported by the development of a pathogenic pro-inflammatory T cell response and/or defective mechanisms for eliminating activated autoreactive T cells [6,7]. Therefore, hormone-related modulation of the immune response might represent a plausible explanation for the powerful effect of pregnancy on MS course. The survival of activated T cells is regulated by different mechanisms, including TRAIL-and CD95-mediated apoptosis. CD95 is a cell surface receptor of the TNF/nerve growth factor receptor superfamily which signals apoptosis when bound by the CD95 ligand (CD95L). The CD95/CD95L
84
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87
system is of extraordinary importance in limiting immune responses. Natural mouse mutants which lack the genes encoding CD95 (lpr) and CD95L (gld) develop autoimmune diseases because activated T cells fail to undergo apoptosis in the periphery [8]. Human patients with autoimmune lymphoproliferative or Canale–Smith syndrome also carry mutations in the CD95 gene [9]. Alternative splicing of the CD95 mRNA leads to 5 different soluble forms of the CD95 receptor (sCD95) which all lack the intracellular death domain [10]. Serum levels of sCD95 are significantly elevated in systemic lupus erythematosus and in MS [11,12]. CD95 ligand- or antibodyinduced apoptosis of susceptible target cells was shown to be blocked by sCD95 and by serum of MS patients, dependent on the amount of sCD95 [11,13]. Here we examined the possible role of serum levels of sCD95 on the suppression of MS during pregnancy. 2. Methods 2.1. Patients Serum levels of sCD95 were assayed in sixty-one patients with clinically definite MS [14] and relapsing-remitting disease course plus planned pregnancy. Patients were ineligible if they suffered systemic infections, other autoimmune disorders or malignancies, or if they had received corticosteroid treatment less than one month prior to analysis. In the month before pregnancy one patient received interferon-β (IFN-β) and one patient glatiramer acetate (GA). In the postpartum period, these two patients and nine additional patients were treated with intravenous immunoglobulins (IVIGs), IFN-β or GA. Their disability scores, as defined by the expanded disability status scale (EDSS) [15], ranged from 0 to 7.5 (1.7 ± 0.1; Mean ± SEM). Serum samples were drawn at routine clinical appointments at multiple time points before, during, and after pregnancy, and frozen at − 70 °C. Serum samples were obtained with informed consent. The study was approved by the local Ethics Committee.
Fig. 1. Concentrations of sCD95 in the sera of 50 untreated MS patients before, during, and after pregnancy. The amount of sCD95 was measured by ELISA. Means were plotted with SEM.
covariance matrix”, “full maximum likelihood estimation”, and “type III sum of squares”. The “first order autoregressive repeated covariance matrix” is commonly used for longitudinal data sets and allows errors in the different time points to autocorrelate. Time (measured in “WEEKs”) was centered at the observed date of delivery. Growth curve models were conceptualized based on our preliminary data and on theoretical considerations. Our main hypothesis was that a change in sCD95 concentrations during pregnancy is associated with the suppression of MS activity. A linear “unconditional growth model” served as the control model (M0): Yij ¼ p0i þ p1i WEEKij þ eij ðbwithin personQ or blevel 1Q modelÞ ðbbetween personQ or blevel 2Q modelÞ p0i ¼ g00 þ f0i ðbbetween personQ or blevel 2Q modelÞ p1i ¼ g10 þ f1i
Yij π0i π1i εij γ ζ
sCD95 concentration of individual i at time point j individual i's true value of sCD95 concentration at time point 0 individual i's weekly rate of change in true sCD95 concentration random error of the data of individual i at time point j “level 2“ regression parameter “level 2“ residuals
2.2. Detection of sCD95 In contrast, model M1 was quadratic: sCD95 levels in the serum of patients were measured in dilutions of 1:10 by a sandwich enzyme-linked immunosorbent assay (ELISA) (Bender MedSystems, Vienna, Austria). The detection limit was confirmed to be 0.2 U/ml with an intraassay coefficient of variation of 2.3% and an interassay coefficient of variation of 3.5%. 2.3. Statistical analysis Differences in the sCD95 levels were compared by the Mann–Whitney U-test, ANOVA, and using individual growth curve modeling [16] with SPSS 11.0 (SPSS, Chicago, USA). The growth curve model was calculated using a mixed model procedure with a “first order autoregressive repeated
Yij ¼ p0i þ p1i WEEKij þ p2i WEEK2ij þ eij p0i ¼ g00 þ f0i p1i ¼ g10 þ f1i p2i ¼ g20 þ f2i Model M2 allowed for a discontinuous change of the slope due to an additional “level 1” term which was set to zero before and after pregnancy (PREG): Yij ¼ p0i þ p1i WEEKij þ p3i PREGij þ eij p0i ¼ g00 þ f0i p1i ¼ g10 þ f1i p3i ¼ g30 þ f3i
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87
85
Fig. 2. Change of sCD95 concentrations in pg/ml exemplified in 6 MS patients before, during and after pregnancy. The amount of sCD95 was measured by ELISA.
and Model M3 comprised interindividual differences in the change of trajectories depending on the change of EDSS scores (EDSS). This variable was set to one if the change of EDSS scores during pregnancy was b1:
before, during, and after pregnancy. In the group of untreated patients the mean sCD95 concentration was 582.4 ± 269.7 pg/ ml (Mean ± SEM) before pregnancy, 583.1 ± 71.5 pg/ml in the first trimester, 516.2 ± 42.4 pg/ml in the second trimester, 650.1 ± 45.9 pg/ml in the third, 657.0 ± 77.8 pg/ml in the first three postpartal months, and 523.8 ± 111.5 pg/ml three or more months after delivery (Fig. 1). One-way ANOVA analyses with the above-mentioned time periods showed no significant differences in sCD95 concentrations, both in the group of untreated patients (n = 50) and in the entire study population (F = 0.778, p = 0.57; F = 0.380, p = 0.86). Additional post hoc tests (Scheffé) confirmed these results. Further, we considered whether serum levels of sCD95 could be associated with EDSS during pregnancy in the
Yij ¼ p0i þ p1i WEEKij þ eij p0i ¼ g00 þ g01 EDSS þ f0i p1i ¼ g10 þ g11 EDSS þ f1i 3. Results The concentration of sCD95 in the serum of RRMS patients with planned pregnancy was analyzed using ELISA Table 1 Multilevel model for change fitted to sCD95 data of the untreated patients (n=50) Parameter
M0
M1
M2
M3
6.31877⁎⁎⁎ (.04756)
6.31956⁎⁎⁎ (.04766)
6.33826⁎⁎⁎ (.09578)
6.29681⁎⁎⁎ (.05768) .07544 (.10192) .00127 (.00329) − .00043 (.00476)
Fixed effects Initial status. π0i Rate of change. π1I Rate of change. π2i Rate of change. π3i
Intercept EDSS change Intercept EDSS change Intercept Intercept
γ00 γ01 γ10 γ11 γ20 γ30
0.00082 (.00237)
.00018 (.00339) − .00002 (.00007)
.00071 (.00242)
−.02532 (.10753)
Model statistics − 2 Log likelihood Akaike's Information Criterion (AIC) Schwarz's Bayesian Criterion (BIC) Difference to control model. −2 Log Likelihood
124.912 130.912 139.013
124.841 132.841 143.643 0.071 ns
124.856 132.856 143.658 0.056 ns
124.055 134.055 147.557 0.857 ns
Goodness of fit statistics were used for model comparison. This model predicts the natural logarithm of sCD95 as a function of weeks, pregnancy (at level-1) and EDSS change (at level-2). ⁎⁎⁎ b0.001.
86
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87
group of 50 untreated patients. As expected from previous reports [17], the study group tended to show a decreased EDSS during the third trimester of pregnancy (second trimester 1.94 ± 0.28; third trimester 1.71 ± 0.16, postpartum 1.91 ± 0.35). Nevertheless, our analyses showed no correlation between EDSS and sCD95 serum concentrations (Pearson Correlation: − 0.1, p = 0.4) In order to make an initial assessment of our data, we analyzed a subgroup of 30 randomly chosen MS patients and plotted their course of sCD95 concentrations before, during, and after pregnancy. The data obtained were heterogeneous and without apparent systematic pattern, as shown in the six examples represented in Fig. 2. To analyze such data, a repeated measurement ANOVA could have been applied. However, the assignment of exact time points to larger time periods, e.g. trimesters, would result in reduced statistical power. Hence, we used individual growth curve modeling to analyze our data (see Methods). The final models are presented in Table 1. Goodness of fit statistics indicated no significant difference between M1 and M0, M2 and M0 or M3 and M0. Furthermore, none of the predictor variables reached significance. Hence, we were unable to detect pregnancy- or EDSS-associated changes of sCD95 concentration in sera of MS patients over time.
the hormone-mediated improvement of regulatory immune mechanisms could have a beneficial effect on MS disease course. Moreover, estrogens were shown to inhibit the activation of microglial cells, which not only function as antigenpresenting cells but also produce pro-inflammatory cytokines such as TNF-α [22,23]. Another benefit of hormonal steroids might be a positive effect of progesterone on myelin formation and maintenance, as observed in animal studies [24]. In this context, a phase II prospective pilot trial in which 10 female MS patients received a 6-month treatment with high dosages of oral estriol showed a significant decease in the volume and the number of gadolinium-enhancing lesions on brain MRIs, although significant effects on relapse or EDSS score were not observed [25]. In summary, we were unable to find any implication of serum levels of sCD95 in pregnancy-related immune suppression in MS patients. The suppression of MS during pregnancy is more potent than the currently available treatments for this disease, but the mechanism of this effect is not yet understood. Thus, further studies on the immunoregulatory mechanisms involved in pregnancy could potentially result in the discovery of novel and safe MS treatments.
4. Discussion
We would like to thank all patients who participated in this study, as well as Alistair Noon and Andrew Mason for reading the manuscript as native English speakers. This study was supported by grants from the Charité (Rahel-Hirsch Stipend to CID) and from the German Ministry of Science (to FZ).
In this study we tested the hypothesis that the suppressive effect of pregnancy on MS activity might be mediated by falling sCD95 concentration in the serum of pregnant MS patients. In our previous work, we were able to show elevated sCD95 concentrations in MS patients compared to healthy controls. The results of a longitudinal study indicated that high sCD95 concentrations correlate with increasing disability measured with the expanded disability status scale (EDSS) [12]. Furthermore, we found that the sera of MS patients inhibit CD95 ligand-induced apoptosis of susceptible target cells in a concentration-dependent manner [13]. Thus, high sCD95 levels might allow for the survival of myelin-specific T cells and prolonged abnormal immune responses in MS patients. In this context it is highly interesting that Interferonβ, currently one of the most widely used treatments for RRMS, was shown to provoke an initial increase but subsequent decline of sCD95 levels in treated patients, indicating an enhanced T cell elimination [12]. In spite of these data and considerations, we were unable to show changes in the concentration of sCD95 during pregnancy or correlations between sCD95 and EDSS, although the sample size was adequate to reduce the beta error probability. Thus, mechanisms other than sCD95 might be responsible for the suppressive effect of pregnancy on MS. Pregnancy has long been considered to be a Th2-related occurrence [18,19]. However, it has recently been reported that gravidity is instead associated with an increased production of regulatory IL-10 [20] and an expansion of CD4+ CD25+ and IL-10+ regulatory T cells [21]. This indicates that
Acknowledgements
References [1] Sospedra M, Martin R. Immunology of multiple sclerosis. Annu Rev Immunol 2005;23:683–747. [2] Confavreux C, Hutchinson M, Hours MM, Cortinovis-Tourniaire P, Moreau T. Rate of pregnancy-related relapse in multiple sclerosis. Pregnancy in Multiple Sclerosis Group. N Engl J Med 1998;339:285–91. [3] Vukusic S, Hutchinson M, Hours M, Moreau T, Cortinovis-Tourniaire P, Adeleine P, et al. Pregnancy and multiple sclerosis (the PRIMS study): clinical predictors of post-partum relapse. Brain 2004;127:1353–60. [4] van Walderveen MA, Tas MW, Barkhof F, Polman CH, Frequin ST, Hommes OR, et al. Magnetic resonance evaluation of disease activity during pregnancy in multiple sclerosis. Neurology 1994;44:327–9. [5] Keith AB. Effect of pregnancy on experimental allergic encephalomyelitis in guinea pigs and rats. J Neurol Sci 1978;38:317–26. [6] Crawford MP, Yan SX, Ortega SB, Mehta RS, Hewitt RE, Price DA, et al. High prevalence of autoreactive, neuroantigen-specific CD8+ T cells in multiple sclerosis revealed by novel flow cytometric assay. Blood 2004;103:4222–31. [7] Zipp F, Krammer PH, Weller M. Immune (dys)regulation in multiple sclerosis: role of the CD95–CD95 ligand system. Immunol Today 1999;20:550–4 [00001545]. [8] Nagata S, Suda T. Fas and Fas ligand: lpr and gld mutations. Immunol Today 1995;16:39–43. [9] Nagata S. Human autoimmune lymphoproliferative syndrome, a defect in the apoptosis-inducing Fas receptor: a lesson from the mouse model. J Hum Genet 1998;43:2–8. [10] Papoff G, Cascino I, Eramo A, Starace G, Lynch DH, Ruberti G. An Nterminal domain shared by Fas/Apo-1 (CD95) soluble variants prevents cell death in vitro. J Immunol 1996;156:4622–30.
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87 [11] Cheng J, Zhou T, Liu C, Shapiro JP, Brauer MJ, Kiefer MC, et al. Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 1994;263:1759–62. [12] Zipp F, Weller M, Calabresi PA, Frank JA, Bash CN, Dichgans J, et al. Increased serum levels of soluble CD95 (APO-1/Fas) in relapsingremitting multiple sclerosis. Ann Neurol 1998;43:116–20. [13] Zipp F, Otzelberger K, Dichgans J, Martin R, Weller M. Serum CD95 of relapsing remitting multiple sclerosis patients protects from CD95mediated apoptosis. J Neuroimmunol 1998;86:151–4. [14] Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227–31. [15] Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:1444–52. [16] Singer JD, Willett JB. Applied longitudinal data analysis: modeling change and event occurrence. New York: Oxford Press; 2003. [17] Vukusic S, Confavreux C. Pregnancy and multiple sclerosis: the children of PRIMS. Clin Neurol Neurosurg 2006;108:266–70. [18] Wegmann TG, Lin H, Guilbert L, Mosmann TR. Bidirectional cytokine interactions in the maternal-fetal relationship: is successful pregnancy a TH2 phenomenon? Immunol Today 1993;14:353–6. [19] Correale J, Arias M, Gilmore W. Steroid hormone regulation of cytokine secretion by proteolipid protein-specific CD4+ T cell clones
[20]
[21] [22]
[23]
[24]
[25]
87
isolated from multiple sclerosis patients and normal control subjects. J Immunol 1998;161:3365–74. Gilmore W, Arias M, Stroud N, Stek A, McCarthy KA, Correale J. Preliminary studies of cytokine secretion patterns associated with pregnancy in MS patients. J Neurol Sci 2004;224:69–76. Zenclussen AC. CD4(+)CD25+ T regulatory cells in murine pregnancy. J Reprod Immunol 2005;65:101–10. Dimayuga FO, Reed JL, Carnero GA, Wang C, Dimayuga ER, Dimayuga VM, et al. Estrogen and brain inflammation: effects on microglial expression of MHC, costimulatory molecules and cytokines. J Neuroimmunol 2005;161:123–36. Bruce-Keller AJ, Keeling JL, Keller JN, Huang FF, Camondola S, Mattson MP. Antiinflammatory effects of estrogen on microglial activation. Endocrinology 2000;141:3646–56. Jung-Testas I, Do Thi A, Koenig H, Desarnaud F, Shazand K, Schumacher M, et al. Progesterone as a neurosteroid: synthesis and actions in rat glial cells. J Steroid Biochem Mol Biol 1999;69:97–107. Sicotte NL, Liva SM, Klutch R, Pfeiffer P, Bouvier S, Odesa S, et al. Treatment of multiple sclerosis with the pregnancy hormone estriol. Ann Neurol 2002;52:421–8.
Serum levels of soluble CD95 are not associated with amelioration of multiple sclerosis during pregnancy Stefan Ehrlich a , Judith Haas b , Frauke Zipp c , Carmen Infante-Duarte c,⁎ a
Department of Child and Adolescent Psychiatry, Charité - Universitaetsmedizin Berlin, Germany b Jewish Hospital Berlin, Germany c Institute of Neuroimmunology, Charité, Universitaetsmedizin Berlin, Germany
Received 30 December 2005; received in revised form 31 August 2006; accepted 13 October 2006
Abstract Multiple sclerosis (MS) is considered an autoimmune disease of the central nervous system. Like various other autoimmune disorders, MS normally improves during pregnancy. Pregnant MS patients experience a significant reduction in relapse rates and magnetic resonance (MR) disease activity. How sex steroid hormones affect disease course remains unclear. We hypothesized that hormonal changes during pregnancy might modulate the autoimmune response by enhancing apoptosis of autoreactive T lymphocytes. One of the most important effectors of apoptosis in T cells is the CD95/CD95L system. We have previously reported that the soluble form of CD95 (sCD95) can block CD95-mediated apoptosis and that MS patients show elevated levels of sCD95. Therefore, we considered whether gravidity might influence serum levels of sCD95 in patients, and analyzed the concentration of sCD95 in the sera of 61 patients with relapsing-remitting (RR) MS before, during and after pregnancy. We found no association between serum levels of sCD95 and pregnancy-related immune suppression in MS patients. Thus, sex steroid hormones do not seem to affect the production of anti-apoptotic sCD95. © 2006 Elsevier B.V. All rights reserved. Keywords: Multiple Sclerosis; Autoimmunity; Soluble CD95; T cell apoptosis; Pregnancy; Sex steroid hormones
1. Introduction Multiple sclerosis (MS) is a multifocal inflammatory disease of the brain and the spinal cord. The main cause of MS is unknown, although genetic and environmental factors have been shown to contribute to its etiology [1]. MS patients who become pregnant experience a significant reduction in relapse rates. The large, prospective European PRIMS (Pregnancy in Multiple Sclerosis) study showed an 87.5% decrease in the mean rate of relapse during the third trimester of pregnancy [2]. However, in the first three months postpartum relapse rate increases in 43% of the patients and returns to the pre-pregnancy level around the second trimester after delivery [3]. Furthermore, reduction in clinical relapse ⁎ Corresponding author. Institute of Neuroimmunology, Neuroscience Research Center, NWFZ 2680, Charité, 10098 Berlin, Germany. Tel.: +49 30 450 539028; fax: +49 30 450 539906. E-mail address: [email protected] (C. Infante-Duarte). 0022-510X/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jns.2006.10.013
rate correlates with a decreased disease activity on magnetic resonance imaging (MRI) in pregnant MS patients [4]. This beneficial effect of pregnancy on the disease course has also been demonstrated in a rodent model of MS, experimental allergic encephalomyelitis (EAE) [5]. How pregnancy and sex hormones may influence MS is still unknown. MS is considered to be a T cell-mediated autoimmune disease [1] which might be promoted or even supported by the development of a pathogenic pro-inflammatory T cell response and/or defective mechanisms for eliminating activated autoreactive T cells [6,7]. Therefore, hormone-related modulation of the immune response might represent a plausible explanation for the powerful effect of pregnancy on MS course. The survival of activated T cells is regulated by different mechanisms, including TRAIL-and CD95-mediated apoptosis. CD95 is a cell surface receptor of the TNF/nerve growth factor receptor superfamily which signals apoptosis when bound by the CD95 ligand (CD95L). The CD95/CD95L
84
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87
system is of extraordinary importance in limiting immune responses. Natural mouse mutants which lack the genes encoding CD95 (lpr) and CD95L (gld) develop autoimmune diseases because activated T cells fail to undergo apoptosis in the periphery [8]. Human patients with autoimmune lymphoproliferative or Canale–Smith syndrome also carry mutations in the CD95 gene [9]. Alternative splicing of the CD95 mRNA leads to 5 different soluble forms of the CD95 receptor (sCD95) which all lack the intracellular death domain [10]. Serum levels of sCD95 are significantly elevated in systemic lupus erythematosus and in MS [11,12]. CD95 ligand- or antibodyinduced apoptosis of susceptible target cells was shown to be blocked by sCD95 and by serum of MS patients, dependent on the amount of sCD95 [11,13]. Here we examined the possible role of serum levels of sCD95 on the suppression of MS during pregnancy. 2. Methods 2.1. Patients Serum levels of sCD95 were assayed in sixty-one patients with clinically definite MS [14] and relapsing-remitting disease course plus planned pregnancy. Patients were ineligible if they suffered systemic infections, other autoimmune disorders or malignancies, or if they had received corticosteroid treatment less than one month prior to analysis. In the month before pregnancy one patient received interferon-β (IFN-β) and one patient glatiramer acetate (GA). In the postpartum period, these two patients and nine additional patients were treated with intravenous immunoglobulins (IVIGs), IFN-β or GA. Their disability scores, as defined by the expanded disability status scale (EDSS) [15], ranged from 0 to 7.5 (1.7 ± 0.1; Mean ± SEM). Serum samples were drawn at routine clinical appointments at multiple time points before, during, and after pregnancy, and frozen at − 70 °C. Serum samples were obtained with informed consent. The study was approved by the local Ethics Committee.
Fig. 1. Concentrations of sCD95 in the sera of 50 untreated MS patients before, during, and after pregnancy. The amount of sCD95 was measured by ELISA. Means were plotted with SEM.
covariance matrix”, “full maximum likelihood estimation”, and “type III sum of squares”. The “first order autoregressive repeated covariance matrix” is commonly used for longitudinal data sets and allows errors in the different time points to autocorrelate. Time (measured in “WEEKs”) was centered at the observed date of delivery. Growth curve models were conceptualized based on our preliminary data and on theoretical considerations. Our main hypothesis was that a change in sCD95 concentrations during pregnancy is associated with the suppression of MS activity. A linear “unconditional growth model” served as the control model (M0): Yij ¼ p0i þ p1i WEEKij þ eij ðbwithin personQ or blevel 1Q modelÞ ðbbetween personQ or blevel 2Q modelÞ p0i ¼ g00 þ f0i ðbbetween personQ or blevel 2Q modelÞ p1i ¼ g10 þ f1i
Yij π0i π1i εij γ ζ
sCD95 concentration of individual i at time point j individual i's true value of sCD95 concentration at time point 0 individual i's weekly rate of change in true sCD95 concentration random error of the data of individual i at time point j “level 2“ regression parameter “level 2“ residuals
2.2. Detection of sCD95 In contrast, model M1 was quadratic: sCD95 levels in the serum of patients were measured in dilutions of 1:10 by a sandwich enzyme-linked immunosorbent assay (ELISA) (Bender MedSystems, Vienna, Austria). The detection limit was confirmed to be 0.2 U/ml with an intraassay coefficient of variation of 2.3% and an interassay coefficient of variation of 3.5%. 2.3. Statistical analysis Differences in the sCD95 levels were compared by the Mann–Whitney U-test, ANOVA, and using individual growth curve modeling [16] with SPSS 11.0 (SPSS, Chicago, USA). The growth curve model was calculated using a mixed model procedure with a “first order autoregressive repeated
Yij ¼ p0i þ p1i WEEKij þ p2i WEEK2ij þ eij p0i ¼ g00 þ f0i p1i ¼ g10 þ f1i p2i ¼ g20 þ f2i Model M2 allowed for a discontinuous change of the slope due to an additional “level 1” term which was set to zero before and after pregnancy (PREG): Yij ¼ p0i þ p1i WEEKij þ p3i PREGij þ eij p0i ¼ g00 þ f0i p1i ¼ g10 þ f1i p3i ¼ g30 þ f3i
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87
85
Fig. 2. Change of sCD95 concentrations in pg/ml exemplified in 6 MS patients before, during and after pregnancy. The amount of sCD95 was measured by ELISA.
and Model M3 comprised interindividual differences in the change of trajectories depending on the change of EDSS scores (EDSS). This variable was set to one if the change of EDSS scores during pregnancy was b1:
before, during, and after pregnancy. In the group of untreated patients the mean sCD95 concentration was 582.4 ± 269.7 pg/ ml (Mean ± SEM) before pregnancy, 583.1 ± 71.5 pg/ml in the first trimester, 516.2 ± 42.4 pg/ml in the second trimester, 650.1 ± 45.9 pg/ml in the third, 657.0 ± 77.8 pg/ml in the first three postpartal months, and 523.8 ± 111.5 pg/ml three or more months after delivery (Fig. 1). One-way ANOVA analyses with the above-mentioned time periods showed no significant differences in sCD95 concentrations, both in the group of untreated patients (n = 50) and in the entire study population (F = 0.778, p = 0.57; F = 0.380, p = 0.86). Additional post hoc tests (Scheffé) confirmed these results. Further, we considered whether serum levels of sCD95 could be associated with EDSS during pregnancy in the
Yij ¼ p0i þ p1i WEEKij þ eij p0i ¼ g00 þ g01 EDSS þ f0i p1i ¼ g10 þ g11 EDSS þ f1i 3. Results The concentration of sCD95 in the serum of RRMS patients with planned pregnancy was analyzed using ELISA Table 1 Multilevel model for change fitted to sCD95 data of the untreated patients (n=50) Parameter
M0
M1
M2
M3
6.31877⁎⁎⁎ (.04756)
6.31956⁎⁎⁎ (.04766)
6.33826⁎⁎⁎ (.09578)
6.29681⁎⁎⁎ (.05768) .07544 (.10192) .00127 (.00329) − .00043 (.00476)
Fixed effects Initial status. π0i Rate of change. π1I Rate of change. π2i Rate of change. π3i
Intercept EDSS change Intercept EDSS change Intercept Intercept
γ00 γ01 γ10 γ11 γ20 γ30
0.00082 (.00237)
.00018 (.00339) − .00002 (.00007)
.00071 (.00242)
−.02532 (.10753)
Model statistics − 2 Log likelihood Akaike's Information Criterion (AIC) Schwarz's Bayesian Criterion (BIC) Difference to control model. −2 Log Likelihood
124.912 130.912 139.013
124.841 132.841 143.643 0.071 ns
124.856 132.856 143.658 0.056 ns
124.055 134.055 147.557 0.857 ns
Goodness of fit statistics were used for model comparison. This model predicts the natural logarithm of sCD95 as a function of weeks, pregnancy (at level-1) and EDSS change (at level-2). ⁎⁎⁎ b0.001.
86
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87
group of 50 untreated patients. As expected from previous reports [17], the study group tended to show a decreased EDSS during the third trimester of pregnancy (second trimester 1.94 ± 0.28; third trimester 1.71 ± 0.16, postpartum 1.91 ± 0.35). Nevertheless, our analyses showed no correlation between EDSS and sCD95 serum concentrations (Pearson Correlation: − 0.1, p = 0.4) In order to make an initial assessment of our data, we analyzed a subgroup of 30 randomly chosen MS patients and plotted their course of sCD95 concentrations before, during, and after pregnancy. The data obtained were heterogeneous and without apparent systematic pattern, as shown in the six examples represented in Fig. 2. To analyze such data, a repeated measurement ANOVA could have been applied. However, the assignment of exact time points to larger time periods, e.g. trimesters, would result in reduced statistical power. Hence, we used individual growth curve modeling to analyze our data (see Methods). The final models are presented in Table 1. Goodness of fit statistics indicated no significant difference between M1 and M0, M2 and M0 or M3 and M0. Furthermore, none of the predictor variables reached significance. Hence, we were unable to detect pregnancy- or EDSS-associated changes of sCD95 concentration in sera of MS patients over time.
the hormone-mediated improvement of regulatory immune mechanisms could have a beneficial effect on MS disease course. Moreover, estrogens were shown to inhibit the activation of microglial cells, which not only function as antigenpresenting cells but also produce pro-inflammatory cytokines such as TNF-α [22,23]. Another benefit of hormonal steroids might be a positive effect of progesterone on myelin formation and maintenance, as observed in animal studies [24]. In this context, a phase II prospective pilot trial in which 10 female MS patients received a 6-month treatment with high dosages of oral estriol showed a significant decease in the volume and the number of gadolinium-enhancing lesions on brain MRIs, although significant effects on relapse or EDSS score were not observed [25]. In summary, we were unable to find any implication of serum levels of sCD95 in pregnancy-related immune suppression in MS patients. The suppression of MS during pregnancy is more potent than the currently available treatments for this disease, but the mechanism of this effect is not yet understood. Thus, further studies on the immunoregulatory mechanisms involved in pregnancy could potentially result in the discovery of novel and safe MS treatments.
4. Discussion
We would like to thank all patients who participated in this study, as well as Alistair Noon and Andrew Mason for reading the manuscript as native English speakers. This study was supported by grants from the Charité (Rahel-Hirsch Stipend to CID) and from the German Ministry of Science (to FZ).
In this study we tested the hypothesis that the suppressive effect of pregnancy on MS activity might be mediated by falling sCD95 concentration in the serum of pregnant MS patients. In our previous work, we were able to show elevated sCD95 concentrations in MS patients compared to healthy controls. The results of a longitudinal study indicated that high sCD95 concentrations correlate with increasing disability measured with the expanded disability status scale (EDSS) [12]. Furthermore, we found that the sera of MS patients inhibit CD95 ligand-induced apoptosis of susceptible target cells in a concentration-dependent manner [13]. Thus, high sCD95 levels might allow for the survival of myelin-specific T cells and prolonged abnormal immune responses in MS patients. In this context it is highly interesting that Interferonβ, currently one of the most widely used treatments for RRMS, was shown to provoke an initial increase but subsequent decline of sCD95 levels in treated patients, indicating an enhanced T cell elimination [12]. In spite of these data and considerations, we were unable to show changes in the concentration of sCD95 during pregnancy or correlations between sCD95 and EDSS, although the sample size was adequate to reduce the beta error probability. Thus, mechanisms other than sCD95 might be responsible for the suppressive effect of pregnancy on MS. Pregnancy has long been considered to be a Th2-related occurrence [18,19]. However, it has recently been reported that gravidity is instead associated with an increased production of regulatory IL-10 [20] and an expansion of CD4+ CD25+ and IL-10+ regulatory T cells [21]. This indicates that
Acknowledgements
References [1] Sospedra M, Martin R. Immunology of multiple sclerosis. Annu Rev Immunol 2005;23:683–747. [2] Confavreux C, Hutchinson M, Hours MM, Cortinovis-Tourniaire P, Moreau T. Rate of pregnancy-related relapse in multiple sclerosis. Pregnancy in Multiple Sclerosis Group. N Engl J Med 1998;339:285–91. [3] Vukusic S, Hutchinson M, Hours M, Moreau T, Cortinovis-Tourniaire P, Adeleine P, et al. Pregnancy and multiple sclerosis (the PRIMS study): clinical predictors of post-partum relapse. Brain 2004;127:1353–60. [4] van Walderveen MA, Tas MW, Barkhof F, Polman CH, Frequin ST, Hommes OR, et al. Magnetic resonance evaluation of disease activity during pregnancy in multiple sclerosis. Neurology 1994;44:327–9. [5] Keith AB. Effect of pregnancy on experimental allergic encephalomyelitis in guinea pigs and rats. J Neurol Sci 1978;38:317–26. [6] Crawford MP, Yan SX, Ortega SB, Mehta RS, Hewitt RE, Price DA, et al. High prevalence of autoreactive, neuroantigen-specific CD8+ T cells in multiple sclerosis revealed by novel flow cytometric assay. Blood 2004;103:4222–31. [7] Zipp F, Krammer PH, Weller M. Immune (dys)regulation in multiple sclerosis: role of the CD95–CD95 ligand system. Immunol Today 1999;20:550–4 [00001545]. [8] Nagata S, Suda T. Fas and Fas ligand: lpr and gld mutations. Immunol Today 1995;16:39–43. [9] Nagata S. Human autoimmune lymphoproliferative syndrome, a defect in the apoptosis-inducing Fas receptor: a lesson from the mouse model. J Hum Genet 1998;43:2–8. [10] Papoff G, Cascino I, Eramo A, Starace G, Lynch DH, Ruberti G. An Nterminal domain shared by Fas/Apo-1 (CD95) soluble variants prevents cell death in vitro. J Immunol 1996;156:4622–30.
S. Ehrlich et al. / Journal of the Neurological Sciences 252 (2007) 83–87 [11] Cheng J, Zhou T, Liu C, Shapiro JP, Brauer MJ, Kiefer MC, et al. Protection from Fas-mediated apoptosis by a soluble form of the Fas molecule. Science 1994;263:1759–62. [12] Zipp F, Weller M, Calabresi PA, Frank JA, Bash CN, Dichgans J, et al. Increased serum levels of soluble CD95 (APO-1/Fas) in relapsingremitting multiple sclerosis. Ann Neurol 1998;43:116–20. [13] Zipp F, Otzelberger K, Dichgans J, Martin R, Weller M. Serum CD95 of relapsing remitting multiple sclerosis patients protects from CD95mediated apoptosis. J Neuroimmunol 1998;86:151–4. [14] Poser CM, Paty DW, Scheinberg L, McDonald WI, Davis FA, Ebers GC, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227–31. [15] Kurtzke JF. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 1983;33:1444–52. [16] Singer JD, Willett JB. Applied longitudinal data analysis: modeling change and event occurrence. New York: Oxford Press; 2003. [17] Vukusic S, Confavreux C. Pregnancy and multiple sclerosis: the children of PRIMS. Clin Neurol Neurosurg 2006;108:266–70. [18] Wegmann TG, Lin H, Guilbert L, Mosmann TR. Bidirectional cytokine interactions in the maternal-fetal relationship: is successful pregnancy a TH2 phenomenon? Immunol Today 1993;14:353–6. [19] Correale J, Arias M, Gilmore W. Steroid hormone regulation of cytokine secretion by proteolipid protein-specific CD4+ T cell clones
[20]
[21] [22]
[23]
[24]
[25]
87
isolated from multiple sclerosis patients and normal control subjects. J Immunol 1998;161:3365–74. Gilmore W, Arias M, Stroud N, Stek A, McCarthy KA, Correale J. Preliminary studies of cytokine secretion patterns associated with pregnancy in MS patients. J Neurol Sci 2004;224:69–76. Zenclussen AC. CD4(+)CD25+ T regulatory cells in murine pregnancy. J Reprod Immunol 2005;65:101–10. Dimayuga FO, Reed JL, Carnero GA, Wang C, Dimayuga ER, Dimayuga VM, et al. Estrogen and brain inflammation: effects on microglial expression of MHC, costimulatory molecules and cytokines. J Neuroimmunol 2005;161:123–36. Bruce-Keller AJ, Keeling JL, Keller JN, Huang FF, Camondola S, Mattson MP. Antiinflammatory effects of estrogen on microglial activation. Endocrinology 2000;141:3646–56. Jung-Testas I, Do Thi A, Koenig H, Desarnaud F, Shazand K, Schumacher M, et al. Progesterone as a neurosteroid: synthesis and actions in rat glial cells. J Steroid Biochem Mol Biol 1999;69:97–107. Sicotte NL, Liva SM, Klutch R, Pfeiffer P, Bouvier S, Odesa S, et al. Treatment of multiple sclerosis with the pregnancy hormone estriol. Ann Neurol 2002;52:421–8.