Abstract
Objective: To evaluate the use of estriol in the treatment of experimental autoimmune encephalomyelitis (EAE) and other cell mediated autoimmune diseases.
Background: Experimental autoimmune encephalomyelitis is a T helper 1 (Th1)-mediated autoimmune demyelinating disease that is a useful model for the study of immune responses in MS. Interestingly, both EAE and MS have been shown to be ameliorated during late pregnancy.
Methods: Estriol, progesterone, and placebo pellets were implanted in mice during the effector phase of adoptive EAE. Disease scores were compared between treatment groups, and autoantigen-specific humoral and cellular responses were examined.
Results: Estriol treatment reduced the severity of EAE significantly compared with placebo treatment whereas progesterone treatment had no effect. Estriol doses that induced serum estriol levels that approximated estriol levels during late pregnancy were capable of ameliorating disease. Estriol-treated EAE mice had significantly higher levels of serum antibodies of the immunoglobulin (Ig) G1 isotype specific for the autoantigen myelin basic protein (MBP). Further, MBP-specific T-lymphocyte responses from estriol-treated EAE mice were characterized by significantly increased production of the Th2 cytokine interleukin 10 (IL-10). T lymphocytes were shown to be the primary source of IL-10 within antigen-stimulated splenocyte populations.
Conclusions: Estriol as a hormone involved in immune changes during pregnancy may provide a basis for the novel therapeutic use of estriol for MS and other putative Th1-mediated autoimmune diseases that improve during late pregnancy.
During reproductive ages, there is a distinct female preponderance of autoimmune diseases including MS, rheumatoid arthritis (RA), systemic lupus erythematosus (SLE), MG, Sjögren’s syndrome, and Hashimoto’s thyroiditis. Sex hormones or sex-linked gene inheritance may be responsible for the enhanced susceptibility of women to these autoimmune diseases. A role for sex hormones in susceptibility to autoimmune disease is supported by observations of alterations in disease symptomatology with alterations in sex hormone levels during pregnancy, menopause, or exogenous hormone administration in the form of hormone replacement therapy or oral contraceptives. Women with MS and RA experience a remission of symptoms during gestation,1-8 whereas women with SLE experience exacerbation of symptoms with gestation.9,10 Because MS and RA are presumed to be cell-mediated autoimmune diseases and SLE is a putative antibody-mediated disease, it has been hypothesized that cell-mediated autoimmune disorders may be ameliorated during pregnancy whereas autoantibody-mediated disorders may be exacerbated during pregnancy.11-13
Cell-mediated immunity is due to T lymphocytes of the T helper 1 (Th1) type that secrete interferon gamma (IFN-γ) and tumor necrosis factor beta (TFN-β), whereas humoral immunity is due to T lymphocytes of the Th2 type that secrete interleukin (IL)-10, IL-4, IL-5, and IL-6. A shift toward humoral (Th2) immunity has been demonstrated during pregnancy.11-17 It is advantageous evolutionarily to shift toward humoral (Th2) immunity during pregnancy because rejection of the fetoplacental unit during allopregnancy is induced by cell-mediated (Th1) immunity.17,18 This shift toward Th2 has not, however, been limited to immune responses localized to the fetoplacental unit, but has been shown to involve systemic immune deviation. Lymphocytes from peripheral blood and lymphoid tissue have demonstrated decreased Th1 cytokine production associated with reduced, delayed-type hypersensitivity responses and increased Th2 cytokine production associated with elevated antibody levels.11-18 This shift toward Th2 during pregnancy has been shown to have an effect on systemic disease. Pregnant mice with immune responses biased toward Th2 had an impaired ability to generate Th1 responses that were needed for clearance of infection, ultimately resulting in a poorer clinical outcome.14 This systemic shift toward Th2 immunity during pregnancy may underlie observations that Th1-mediated autoimmune diseases improve during pregnancy whereas Th2-mediated autoimmune diseases are exacerbated during pregnancy.
Although a shift toward humoral immunity has been demonstrated during pregnancy, mechanisms that have evolved to induce this shift have remained unclear. Possibilities include the local production of Th2 cytokines by the placenta19 or the production of Th2 cytokines by immune cells consequent to increased circulating levels of sex hormones during pregnancy. Consistent with the latter possibility, in vitro studies have demonstrated the ability of progesterone to increase IL-4 production and 17β-estradiol to increase IL-10 production during T-lymphocyte responses.20,21 Whether such effects on cytokine expression would occur when hormones are present in vivo at levels physiologic with pregnancy has remained unknown.
Experimental autoimmune encephalomyelitis (EAE) is a Th1-mediated autoimmune disease that is useful for the study of immune responses in MS.22,23 Because EAE and MS have each been shown to be ameliorated during late pregnancy,1-4 we used the EAE model to identify a hormone of late pregnancy that, when administered in vivo, could ameliorate disease and induce immune changes similar to those observed during pregnancy.
Methods.
Animals.
Female SJL/J and male C57BL10/J mice (age, 6 to 8 weeks) were purchased from Jackson Laboratories (Bar Harbor, ME). All studies were performed in accordance with internal review board approval from the UCLA Office of Protection of Research Subjects (no. 98-001-01).
Hormone pellets.
Ninety-day-release pellets of estriol (5, 15, or 30 mg) and progesterone (15 and 50 mg), as well as placebo pellets that contain the carrier binder only (cholesterol:methyl cellulose:α-lactulose), were purchased from Innovative Research of America (Sarasota, FL). Pellets were implanted subcutaneously in the scapular area of the neck using a trocar after administration of methoxyflurane inhalation anesthesia.24 Mice were implanted with pellets either 3 weeks before or 10 days after the adoptive transfer of myelin basic protein (MBP)-specific T cells to induce EAE.
Immunization and adoptive EAE induction.
Whole 18.5-kDa MBP was prepared from adult guinea pig spinal cord.25 MBP-specific cells for adoptive transfer were generated as described and injected intraperitoneally into naive mice (5 × 107 cells/0.3 mL phosphate-buffered saline [PBS] per mouse). Clinical signs of EAE were graded on a scale of 0 to 5 with increasing disability26: 0, unaffected; 1, complete tail limpness with no evidence of limb weakness; 2, no paralysis on ambulation with limb weakness only as evidenced by inability to remain upright when examiner attempts to role subject over; 3, partial paralysis of at least one limb on ambulation; 4, complete paralysis of at least one limb on ambulation; and 5, moribund.
Pathologic assessment.
Subsequent to receiving anesthesia, mice were perfused intracardially with 10% buffered formalin. After spinal cord dissection, 1-mm blocks of the lumbosacral region were prepared and fixed secondarily in 1% osmium tetroxide dehydrated in methanol and imbedded in Eponate 12 resin. One-micrometer-thick transverse sections were stained with toluidine blue for light microscopic assessment of demyelination, perivascular inflammation, and other pathology. Semiquantitative assessment of pathology from each mouse was performed by examining two blocks of spinal cord separated spatially by 1 to 2 mm.
Pregnant mice with EAE.
Adoptive EAE was induced in naive SJL/J female mice. At the peak of the first episode of disease (clinical grade, 3 to 4; day 11 posttransfer) mice were bred with C57BL10/J male mice to induce allopregnancies. Mice were then sacrificed during late pregnancy (days 17 through 19) to obtain serum to determine estriol levels.
Serum hormone levels.
Blood was obtained by intracardiac puncture from either pregnant mice or estriol-treated mice. Serum hormone levels were determined by ELISA, in duplicate, according to the manufacturer’s directions (Oxford Biomedical Research, Oxford, MI).
MBP-specific antibody isotype ELISA.
Ninety-six-well plates were coated with 10 μg/mL MBP overnight at 4 °C. After washing, plates were blocked with 1% bovine serum albumin (BSA) in PBS overnight at 4 °C. After washing, sera (diluted 1:8, 1:16, and 1:32) were applied to duplicate wells for 2 hours at 37 °C. After further washing, horse radish peroxidase-conjugated goat affinity purified antibody (diluted 1:1,000) specific for mouse immunoglobulin (Ig)M, IgG1, or IgG2a (Southern Biotechnology Assoc. Inc., Birmingham, AL) was added with development of color reaction using 3,3′, 5,5′-tetramethylbenzidine liquid substrate system (Sigma, St. Louis, MO). Color reaction was stopped with stop reagent (Biosource, Camarillo, CA)) and plates were read at 450 nm with an ultraviolet max kinetic microplate reader (Molecular Devices, Menlo Park, CA). Positive control antibody consisted of an MBP-specific monoclonal antibody of the IgG1 isotype (Accurate, Westbury, NY). Negative control wells were coated with hen egg-white lysozyme (instead of MBP) at 10 μg/mL.
Cytokine profiles.
Splenocytes were cultured at 4 × 106 cells/mL with whole 18.5-kDa MBP or MBP 83–102 peptide at a previously optimized concentration (25 μg/mL) as well as with media alone. Wells were harvested for collection of supernatants at 48 and 72 hours. ELISAs (Biosource) were used to determine levels of IFN-γ, IL-10, IL-4, IL-5, IL-2, and IL-12 in supernatants according to manufacturer’s instructions as applied previously.24
Enrichment for splenocyte cell subpopulations.
Splenocytes isolated from estriol-treated EAE mice were cultured with MBP. After 1 day in culture, nonadherent cells were harvested to obtain T and B lymphocytes for further purification into subpopulations, whereas adherent cells were used as the macrophage-enriched subpopulation. Nonadherent cells were purified further into T- and B-cell populations by magnetic activated cell sorting (MACS) as described previously.27 Briefly, nonadherent cells were resuspended in MACS buffer (PBS containing 2 mM ethylenediaminetetraacetic acid, 0.5% BSA, and 0.01% sodium azide) and separated into three groups. The first group of cells was not manipulated further and constituted the “unseparated” nonadherent cell population. It consisted of 45% B lymphocytes, 45% T lymphocytes, and 2% macrophages. The second and third groups of cells were incubated for 15 minutes at 4 °C with magnetic bead conjugated antibodies specific for B220 to label B lymphocytes or Thy1.2 to label T lymphocytes (Miltenyi Biotec, Auburn, CA). After two washes in MACS buffer, cells were added to the column in the magnetic field (VarioMACS; Miltenyi Biotec). The column was washed four times with MACS buffer. Cells retained in the column were then eluted, and flow cytometry analysis showed that cells incubated with B220 antibody were 93% B lymphocytes, 5% T lymphocytes, and 2% macrophages, whereas cells incubated with Thy1.2 antibody were 75% T lymphocytes, 23% B lymphocytes, and 2% macrophages.
Reverse transcriptase (RT)-PCR for IL-10 within each enriched cell subpopulation.
Total RNA from each of the four splenocyte cell subpopulations (unseparated, B lymphocyte, T lymphocyte, and macrophage) was isolated using TRIzol reagent (Life Technologies, Gaithersburg, MD) according to manufacturer’s instructions. One microgram of messenger RNA was reverse transcribed using the GeneAmp RNA PCR kit from Perkin-Elmer (Branchburg, NJ). Complementary DNA samples were amplified for 30 cycles using murine primers for IL-10 and β-actin (Clontech, Palo Alto, CA). PCR products were electrophoresed through 1.2% agarose gel and visualized by ethidium bromide staining. PCR products were then transferred to MagnaGraph nylon membranes (Micron Separations Inc., Westborough, MA) and hybridized using a 32P-labeled murine IL-10 probe (Clontech). The radiolabeled Southern blot was exposed to a phosphor storage screen (Molecular Dynamics, Sunnyvale, CA) and read using the PhosphorImager (Molecular Dynamics) as described.28
Statistical analysis.
Mean cytokine levels and antibody titers from mice from each EAE group were compared using one sample t-test for comparisons of means. Mean clinical scores for each group were compared using repeated-measures analysis of variance and Tukey’s least significant difference criteria for post hoc t-tests as applied previously.24
Results.
Treatment with estriol (E3) ameliorates EAE.
It has been demonstrated previously (as reviewed by Abramsky3) that the clinical course of EAE is less severe during late pregnancy. The purpose of the current study was to identify a hormone of pregnancy that might be responsible for this decrease in disease severity during late pregnancy. Estriol (E3) was a leading candidate. Estriol is made by the fetoplacental unit, and levels are highest during late pregnancy. Estriol (E3) is distinct from 17β-estradiol (E2), which is the primary estrogen of ovulatory cycles. Treatment of naive, nonpregnant SJL/J female mice with 90-day-release estriol pellets 3 weeks before the adoptive transfer of MBP-specific T lymphocytes reduced the severity of clinical EAE significantly compared with naive, nonpregnant SJL/J female control mice implanted in parallel with placebo pellets (figure 1). Mice receiving estriol doses of 5, 15, or 30 mg each demonstrated a significantly less severe clinical course compared with placebo in each experiment when comparing mean scores from each treatment group at day of sacrifice, as well as mean scores from repeated measures over all days of disease (table 1). The mean maximal acute-phase scores were also significantly lower with estriol treatment. Considering data from all experiments, the mean maximal acute-phase score for EAE mice treated with 15 mg estriol was 2.5 (n = 24), whereas the mean maximal acute-phase score for EAE mice treated in parallel with placebo was 3.9 (n = 29; p < 0.0005). Regarding the lower estriol dose, the mean maximal acute-phase score for EAE mice treated with 5 mg estriol was 2.6 (n = 14), whereas the mean maximal acute-phase score for EAE mice treated in parallel with placebo was 4.0 (n = 16; p < 0.005). This reduction in disease from a score of 3.9 to 4.0 (complete paralysis) during placebo treatment to 2.5 to 2.6 (weakness to partial paralysis) during estriol treatment was clinically dramatic. Pathologic evidence of a dramatic reduction in demyelination and inflammation in the lumbosacral spinal cords of mice treated with estriol confirmed the clinical therapeutic effect (table 2). Thus, estriol treatment clearly reduced EAE disease severity and there were no significant differences between estriol doses in the ability to ameliorate disease.
Figure 1. Graphical demonstration of one representative experiment of a total of eight experiments. Estriol treatment ameliorates experimental autoimmune encephalomyelitis (EAE). Mean clinical scores for EAE were reduced in mice treated with 5 mg (⋄; n = 5) and 15 mg (○; n = 5) estriol compared with mice treated with placebo (□; n = 4) (p = 0.05 for each estriol group). Mice were treated with 90-day-release hormone implants 3 weeks before the adoptive transfer of myelin basic protein-specific cells to induce EAE.
Estriol treatment ameliorates experimental autoimmune encephalomyelitis
Histopathologic analysis of mice treated with estriol or placebo
Whether estriol treatment could ameliorate EAE after disease induction was then determined by adoptively transferring MBP-specific cells before treatment with estriol. As demonstrated in figure 2, when mice received estriol treatment implants on the day of disease onset (10 days postadoptive transfer), the subsequent disease course was significantly less severe in estriol-treated mice compared with placebo-treated mice (p < 0.05).
Figure 2. Estriol ameliorates experimental autoimmune encephalomyelitis (EAE) when treatment is administered after disease induction. Mean clinical scores for EAE were reduced in mice treated at disease onset (day 10 postadoptive transfer) with 15 mg estriol (⋄; n = 9) compared with mice treated with placebo (□; n = 8) (p = 0.05). Tx = treatment.
In addition to estriol, pregnancy is characterized by elevated levels of progesterone. Thus we determined whether progesterone treatment might also ameliorate EAE. In contrast to our results with estriol treatment, progesterone treatment had no effect on EAE compared with placebo treatment (see table 1). Together these data implicate a role for estriol in the less severe course of EAE during pregnancy.
Serum estriol levels during estriol treatment compared with during pregnancy.
Because treatment with 90-day-release pellets of 5 and 15 mg estriol ameliorated the clinical course of EAE repeatedly, serum estriol levels induced by each of these doses was determined. Sera from mice treated with estriol were tested in parallel with sera from mice during late pregnancy to determine directly whether estriol treatment had induced serum levels that were physiologic with pregnancy. The 5-mg dose of estriol induced serum levels of estriol (mean, 1.8 ng/mL; n = 9) that closely approximated levels observed during late pregnancy (mean, 1.5 ng/mL; n = 3), with each more than 10-fold higher than levels in nonpregnant, nonestriol-treated mice (mean, 0.05 ng/mL; n = 17). The higher dose of estriol (15 mg) induced levels that were supraphysiologic (mean, 12.1 ng/mL; n = 11)—eightfold higher compared with physiologic late pregnancy. Thus, in vivo treatment with estriol to induce serum levels of estriol that were equal to or greater than those observed during physiologic late pregnancy each resulted in disease amelioration. Furthermore, levels of progesterone and cortisol were no different between estriol-treated and placebo-treated mice (data not shown), thereby suggesting that the effect of in vivo treatment with estriol on disease was not an indirect effect mediated through the hypothalamic pituitary axis.
Estriol treatment increases MBP-specific antibody of the IgG1 isotype.
An increase in humoral immunity had been observed previously during pregnancy.11-13 To determine whether estriol might be responsible for this increase in humoral immunity during pregnancy, we assessed humoral immunity in nonpregnant EAE mice treated with estriol in vivo. Levels of MBP-specific IgG1 were increased significantly in estriol-treated EAE mice compared with placebo-treated EAE mice (p < 0.00005) whereas levels of IgM and IgG2a were not increased significantly (figure 3). The highly significant increase in antibody of the IgG1 isotype in estriol-treated EAE mice is important because production of antibody of the IgG1 isotype is characteristic of Th2 responses. The level of antibody production did not differ between EAE mice treated with 5 or 15 mg of estriol (data not shown). Levels of MBP-specific antibodies of all isotypes were undetectable in untreated healthy female mice and in estriol-treated healthy mice that had not received MBP-specific T lymphocytes by adoptive transfer. These data implicate a role for estriol in the increase in humoral immune responses during pregnancy.
Figure 3. Serum levels of myelin basic protein (MBP)-specific antibodies of the immunoglobulin (Ig)G1 isotype are increased in estriol-treated experimental autoimmune encephalomyelitis (EAE) mice (shaded bars; n = 9) compared with placebo-treated EAE mice (white bars; n = 9). Levels of MBP-specific IgM, IgG1, and IgG2a were determined in sera from EAE mice treated with estriol or placebo. In sera from estriol-treated mice with EAE there were significantly higher levels of IgG1 (*p < 0.00005) compared with sera from placebo-treated mice with EAE. At the same dilution (1:32), levels of IgM and IgG2a were lower than IgG1, and not significantly different between treatment groups. Levels in healthy untreated mice (hatched bars; n = 3) were determined as a negative control. Because there were no differences in antibody levels from mice treated with 5 mg estriol versus 15 mg estriol, data from nine estriol-treated mice (which included five mice treated with 5 mg estriol and four mice treated with 15 mg estriol) were pooled for comparison with data from nine placebo-treated mice. These mice were selected randomly from three separate EAE groups sacrificed at day 20 or day 24 postadoptive transfer.
Estriol treatment increases IL-10 production.
Because estriol caused an increase in MBP-specific antibody of the IgG1 isotype, we hypothesized that estriol might also cause an increase in Th2 cytokine production. MBP-specific T-lymphocyte responses from cultured splenocytes were assessed for production of IL-10 as an indicator of the Th2 phenotype, and IFN-γ as an indicator of the Th1 phenotype. IL-10 levels from estriol-treated female mice with EAE sacrificed at the peak of the first episode of disease (day 10 postadoptive transfer) were significantly higher than IL-10 levels from placebo-treated female mice with EAE (p < 0.05), whereas IFN-γ production was not significantly different (figure 4). IL-10 levels from estriol-treated female mice with EAE sacrificed at a later time point (day 24 postadoptive transfer) were again significantly higher than IL-10 levels from placebo-treated female mice (p < 0.05), whereas IFN-γ production was again not significantly different (data not shown). There were no differences in levels of IL-5, IL-2, or IL-12 at any time point examined. IL-4 levels were low and also no different between treatment groups. Because IL-10 enhances the viability and antibody production of B cells,29,30 an increase in IL-10 production during the MBP-specific response in estriol-treated EAE mice is consistent with our observation of increased serum MBP-specific antibody levels in these mice.
Figure 4. Interleukin 10 (IL-10; black bars) levels are higher during myelin basic protein (MBP)-specific responses within splenocytes derived from estriol-treated experimental autoimmune encephalomyelitis (EAE) mice compared with placebo-treated EAE mice. IL-10 production during MBP-specific T-lymphocyte responses was determined from splenocytes of mice treated with estriol or placebo. On postadoptive transfer day 10, levels of interferon gamma (IFN-γ; white bars) and IL-10 were determined by ELISA from supernatants collected at 48 hours after stimulation of splenocytes with MBP. Splenocytes from estriol-treated mice produced significantly more IL-10 than splenocytes from placebo-treated mice (p = 0.05), whereas IFN-γ production was not significantly different between treatment groups. Data are representative from ELISAs done in duplicate and repeated twice.
Splenic T lymphocytes are primarily responsible for IL-10 production in estriol-treated EAE mice.
To determine which cell type within spleen was producing IL-10 during estriol treatment, it was first demonstrated that antigenic stimulation elicited IL-10 production (figure 5). IL-10 was increased when spleen cells from estriol-treated mice were cultured with MBP compared with media. Stimulation with the immunodominant peptide within MBP (83–102) was also shown to elicit IL-10 production. Because T cells respond to peptide sequences with major histocompatibility complex molecules, and B cells respond to whole protein, these observations were consistent with the hypothesis that T-cell stimulation induced IL-10 production by splenocytes. To determine whether T cells were the source of IL-10, splenocytes from estriol-treated mice were stimulated with antigen and enriched for T cells, B cells, or macrophages. Each enriched cell subpopulation was then assessed for IL-10 production by RT-PCR. As demonstrated in figure 6, RT-PCR analysis for IL-10 within each enriched cell population confirmed an increase in IL-10 signal in the T-cell-enriched population. Although these data do not rule out a small degree of IL-10 production by B lymphocytes or macrophages within spleens of estriol-treated mice with EAE, the data together indicate that the cell type within spleen that is primarily responsible for IL-10 production is the T lymphocyte.
Figure 5. Interleukin 10 (IL-10) production is induced on stimulation with whole myelin basic protein (MBP) or MBP peptide. Splenocytes derived from experimental autoimmune encephalomyelitis mice treated with estriol were stimulated in vitro with whole 18.5-kDa MBP (black bars), MBP peptide 83–102 (gray bars), or media alone (white bars). Supernatants were collected at 48 hours, and levels of IL-10 were determined by ELISA, as in figure 4. Data are representative of experiments repeated four times.
Figure 6. Splenic T lymphocytes are primarily responsible for interleukin 10 (IL-10) production in estriol-treated experimental autoimmune encephalomyelitis (EAE) mice. Spleens were harvested from estriol-treated mice with EAE, and splenocytes were cultured with myelin basic protein. After 1 day of culture, splenocytes were enriched for B lymphocytes, T lymphocytes, or macrophages. Total RNA was isolated from each enriched cell subpopulation and 1.0 μg RNA was reverse transcribed with primer sets specific for IL-10 and β-actin. (A) Reverse transcription PCR products were run on agarose gels in the presence of ethidium bromide. The IL-10 signal was increased in the T-lymphocyte-enriched subpopulation. In contrast, there was no increase in the β-actin control signal in the T-lymphocyte-enriched subpopulation. (B) To confirm specific amplification of IL-10 by the IL-10 primers, PCR products were transferred to nylon membranes and hy- bridized using 32P-labeled IL-10 probes. The radiolabeled Southern blot was exposed to a phosphor storage screen and read using the PhosphorImager. The increase in IL-10 signal within the T-lymphocyte-enriched subpopulation was confirmed. Data are representative of experiments repeated twice. + = positive control amplification of IL-10 and β-actin complementary DNA (cDNA); unsep = amplification of cDNA from the nonadherent, unseparated splenocyte subpopulation; B-cell = from the B-lymphocyte-enriched subpopulation; T-cell = from the T-lymphocyte-enriched subpopulation; Mφ = from the adherent cell (macrophage-enriched) subpopulation.
Discussion.
Although it has been well established that EAE, MS, and other putative Th1 (cell-mediated) autoimmune diseases improve clinically during pregnancy,1-8 mechanisms underlying this clinical improvement have remained unknown. To determine whether a hormone of pregnancy might be responsible for the clinical improvement during pregnancy, we administered hormones of pregnancy to nonpregnant mice with EAE. When mice were treated with estriol, the severity of EAE was decreased significantly compared with placebo-treated mice. Doses of estriol that induced serum estriol levels that were equal to or greater than estriol levels that occur physiologically during late pregnancy were equally effective in ameliorating disease. These findings in adoptive EAE in SJL mice are consistent with a previous report31 in that estriol treatment, in doses that approximated levels of pregnancy, suppressed the induction of active EAE in B10.RIII mice immunized with MBP peptide 89–101, complete Freund’s adjuvant, and pertussis toxin. A notable difference in the two studies is that we found a decrease in the severity of disease with minimal delay in the onset of disease, whereas the other report described a delay in the onset of disease with no effect on the severity of disease. The decrease in severity in our study versus the delay in onset of disease in the other study is likely related to differences in the method of EAE induction. Despite these differences, the studies concur in their conclusion that when estriol is administered in doses that approximate those of late pregnancy, there is a suppressive effect on clinical EAE. A delay in the onset of EAE using the active EAE model could indicate either that estriol decreases the ability of the host to generate encephalitogenic MBP-specific T lymphocytes or it decreases the ability of encephalitogenic MBP-specific T lymphocytes to induce disease. Results in our study using adoptive EAE demonstrate clearly that estriol decreases the ability of transferred encephalitogenic MBP-specific T cells to induce disease by demonstrating a decrease in EAE severity when only naive recipients of encephalitogenic T cells are exposed to estriol.
Other candidate hormones that might be responsible for the amelioration of EAE during pregnancy include 17β-estradiol (E2) and progesterone because levels of these hormones are also increased during late pregnancy. One would hypothesize that treatment with low doses of 17β-estradiol might exacerbate EAE whereas treatment with high doses of 17β-estradiol might ameliorate EAE based on the observation of a biphasic dose effect during in vitro studies. When a low dose of 17β-estradiol was added to cultures of human proteolipid-specific T-cell lines, there was an increase in tumor necrosis factor alpha (TNF-α)—a proinflammatory cytokine for EAE. When higher doses of 17β-estradiol were added to cultures to achieve levels approximating those of pregnancy, TNF-α production decreased and there was an increase in IL-10—an anti-inflammatory cytokine for EAE.21 17β-estradiol has been administered previously in vivo in parallel experiments with estriol to compare directly the effects of the two estrogens on EAE. Doses of 17β-estradiol that approximated pregnancy levels had a mild effect on disease, clearly less than the effect observed when estriol was used.31 Only very high doses of 17β-estradiol—ones that were supraphysiologic—could delay the onset of disease to the extent that it was delayed with estriol. Why would treatment with pregnancy doses of estriol ameliorate EAE to a greater extent than treatment with 17β-estradiol? The answer will likely depend upon the relative level of expression of the classic estrogen receptor alpha (ER-α) and the recently identified estrogen receptor beta (ER-β) within lymphoid tissues. In thymus, for example, levels of message for ER-β have appeared to be higher than levels for ER-α.32 In addition, the affinity of estriol was found to be relatively greater for ER-β than for ER-α.32 The relative level of stimulation of the two receptors by a given ligand will be important in determining the ultimate biological effect because the ligand-induced transactivation behavior of ER-α and ER-β differ.33,34
Progesterone levels are also increased during late pregnancy. Based on in vitro studies, one would predict that progesterone treatment would ameliorate EAE. In vitro studies of human T-cell lines have demonstrated that progesterone caused increased IL-4 production20—an anti-inflammatory cytokine for EAE. Despite these in vitro data, we observed no effect of progesterone treatment on the severity of EAE. These data are consistent with a previous observation that 17β-estradiol, but not progesterone, suppressed collagen arthritis.35 During arthritis there was, however, a mild additive effect of progesterone when used in combination with 17β-estradiol. Together these data suggest that the in vivo effect of progesterone on EAE and arthritis is minimal or negligible.
Thus, regarding the effect of pregnancy hormones on EAE, the reports together indicate that estriol, as opposed to 17β-estradiol and progesterone, is primarily responsible for the amelioration of disease during late pregnancy.
Because immune responses during late pregnancy have been characterized by a shift toward humoral immunity,11-17 we investigated whether treatment with estriol could recapitulate such alterations in immunity. Similar to observations that antibody production is increased during pregnancy,11-13 we found that MBP-specific antibody of the IgG1 isotype was increased in EAE mice treated with estriol. In addition, MBP-specific responses from EAE mice treated with estriol were characterized by increased production of the Th2 cytokine IL-10. Because IL-10 promotes the viability and antibody production of B cells,29,30 this increase in IL-10 production is consistent with our observation of increased antibody levels during estriol treatment. Together these data implicate estriol in the shift toward humoral immunity during pregnancy.
Although estriol may ameliorate EAE through multiple mechanisms, the increase in IL-10 production during the MBP-specific response in estriol-treated EAE mice likely plays a role. EAE is a Th1-mediated disease, and IL-10 has been shown to downregulate Th1 responses (as reviewed by Wegmann et al.12). EAE has been ameliorated by IL-10 treatment in several36-38 but not all39 studies, and EAE has been exacerbated by anti-IL-10 treatment.37,39 In addition, upregulated expression of IL-10 has been observed during the remission phase of EAE.40 Finally, IL-10 knockout mice have demonstrated an increased incidence of EAE.41,42 Thus, our finding of increased IL-10 production during the autoantigen-specific response in mice treated with estriol would appear to be at least one mechanism through which estriol ameliorates EAE.
Acknowledgments
Supported by grant no. NS36680-01 from the National Institutes of Health (R.R.V.) and grant no. JF2094-A-2 from the National Multiple Sclerosis Society (R.R.V.). Dr. Voskuhl is a Harry Weaver Neuroscience Scholar of the National Multiple Sclerosis Society.
Acknowledgment
The authors thank Allan J. MacKenzie–Graham (University of California Los Angeles [UCLA], Department of Neurology) for isolating 18.5 kDa MBP and Jenny Yu (UCLA, Department of Biostatistics) for assisting with the statistical analysis.
- Received June 23, 1998.
- Accepted December 19, 1998.
References
Disputes & Debates: Rapid online correspondence
NOTE: All authors' disclosures must be entered and current in our database before comments can be posted. Enter and update disclosures at http://submit.neurology.org. Exception: replies to comments concerning an article you originally authored do not require updated disclosures.
- Stay timely. Submit only on articles published within the last 8 weeks.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- 200 words maximum.
- 5 references maximum. Reference 1 must be the article on which you are commenting.
- 5 authors maximum. Exception: replies can include all original authors of the article.
- Submitted comments are subject to editing and editor review prior to posting.