Oestrogen modulates experimental autoimmune encephalomyelitis and interleukin-17 production via programmed death 1


H. Offner or C. Wang, Portland VA Medical Center R&D-31, 3710 SW US Veterans Hospital Rd., Portland, OR 97239, USA. Email: offnerva@ohsu.edu or wangch@ohsu.edu
Senior author: Halina Offner


The mechanism by which oestrogens suppress experimental autoimmune encephalomyelitis (EAE), an animal model of multiple sclerosis, is only partially understood. We here demonstrate that treatment with 17β-oestradiol (E2) in C57BL/6 mice boosted the expression of programmed death 1 (PD-1), a negative regulator of immune responses, in the CD4+ FoxP3+ regulatory T (Treg) cell compartment in a dose-dependent manner that correlated with the efficiency of EAE protection. Administration of E2 at pregnancy levels but not lower concentrations also enhanced the frequency of Treg cells. Additionally, E2 treatment drastically reduced the production of interleukin-17 (IL-17) in the periphery of immunized mice. However, E2 treatment did not protect against EAE or suppress IL-17 production in PD-1 gene-deficient mice. Finally, E2 failed to prevent Treg-deficient mice from developing spontaneous EAE. Taken together, our results suggest that E2-induced protection against EAE is mediated by upregulation of PD-1 expression within the Treg-cell compartment.


Multiple sclerosis (MS) is a progressive demyelinating neurological disease with an autoimmune component.1 Although the incidence of MS is higher in women, its relapse rates are decreased during late pregnancy,2,3 and treatment with pregnancy levels of oestriol reduces central nervous system lesions.4,5 We demonstrated previously that relatively low doses of 17β-oestradiol (E2) and oestriol confer potent protection against clinical and histological signs of experimental autoimmune encephalomyelitis (EAE),6,7 an animal model of MS. However, the molecular and cellular mechanisms by which oestrogens regulate MS and EAE have not been well characterized.

Several lines of evidence suggest that E2 is a potent regulator of the immune system and may act directly on oligodendrocytes and neurons. We reported that E2 treatment mitigated the production of proinflammatory cytokines, including tumour necrosis factor-α (TNF-α), release of chemokines and recruitment of inflammatory cells into the central nervous system,6–9 and boosted the frequency and function of regulatory T cells (Treg) through oestrogen receptor-α (ERα).10 On the other hand, by evaluating the E2 effect after transfer of encephalitogenic ERα+/+ or ERα−/− T cells into T-cell-deficient mice, we determined that E2 treatment does not directly act on pathogenic T cells.11 Similarly, anti-inflammatory cytokines including interleukin-10 (IL-10) and IL-4 were not required for E2 regulation because E2 was equally effective in wild-type mice and in IL-10 or IL-4 gene-deficient mice.7 Thus, E2 is a potent regulator of autoimmunity, but does not directly act on autoreactive T cells or through anti-inflammatory cytokines.

Programmed death 1 (PD-1) is a newly identified negative regulator of immune responses.10 PD-1 gene-deficient (PD-1KO) mice develop autoimmune disorders similar to lupus-like glomerulonephritis, arthritis or dilated cardiomyopathy as early as 5 weeks of age. In MS patients, a PD-1 polymorphism was associated with disease progression, possibly as the result of a partial defect in PD-1-mediated inhibition of T-cell activation.11 In the EAE model, it was shown that genetic disruption of PD-L1, an identified PD-1 ligand that can also bind to B7-1, converted an EAE-resistant mouse strain into a fully permissive one,12 suggesting that the PD-1 pathway may be critically involved in the process of EAE induction. Our recent studies in PD-1KO mice showed that PD-1 plays an essential role in the initiation of EAE (C. Wang, personal communication), and that E2-induced Treg-cell suppression involves PD-1.13 However, the precise role of PD-1 in E2-induced immunoregulation in EAE remains to be clarified.

Interleukin-17 is a critical proinflammatory cytokine and IL-17-producing helper T (Th17) cells may play a vital role in MS and EAE induction.14,15 IL-17 can synergize with TNF-α, which is downregulated by E27 and IL-1β to induce chemokine expression.13,16 Yet, it is not known whether or how E2 can regulate the production of IL-17 in mice with EAE.

The results presented below demonstrate that E2 treatment selectively upregulated the level of PD-1 in the CD4+ FoxP3+ Treg-cell compartment in immunized mice. The levels of PD-1 within the Treg-cell compartment, but not the frequency of Treg cells themselves, were closely linked to the disease suppressive activity of E2. Treatment with E2 also profoundly reduced the production of IL-17. The E2-mediated protection against EAE and reduction of IL-17 production were completely abrogated in PD-1KO mice. Therefore, we conclude that PD-1 is a critical mediator of the immunoregulatory effects of E2.

Materials and methods


Mice used for these experiments were age-matched females (7–11 weeks old) that were rested for at least 7 days before treatment or immunization. The PD-1KO mice, which were backcrossed with C57BL/6 (B6) mice for 11 generations, were from Dr Honjo at Kyoto University (Kyoto, Japan). B6 mice and RAG1−/− mice on the B10.PL background were purchased from the Jackson Laboratory (Bar Harbor, ME). MBP-TCRα/β double-transgenic mice were a gift from Dr Juan LaFaille of New York University. Animals were housed and cared for according to institutional guidelines in the animal resource facility at the Veterans Affairs Medical Center, Portland, OR.


Slow-releasing E2 and placebo pellets were purchased from Innovative Research of America (Sarasota, FL). These pellets (3-mm diameter) are designed to release their contents at a constant rate over 60 days. Mouse (m) MOG-35-55 peptide was synthesized at Beckman Institute, Stanford University (Palo Alto, CA). Pertussis toxin (PTX) was purchased from List Biological Laboratories (Campbell, CA) and intracellular staining kits and all antibodies were purchased from eBioscience (San Diego, CA). Luminex Bio-Plex mouse cytokine assay kits were purchased from Bio-Rad (Hercules, CA).

Induction of EAE and treatment

Mice were inoculated subcutaneously in the flanks with 0·2 ml of an emulsion containing 200 μg of mMOG-35-55 peptide and 200 μg of complete Freund’s adjuvant containing 200 μg heat-killed Mycobacterium tuberculosis H37RA. On days 0 and 2, mice were injected intraperitonealy with 75 and 200 ng PTX, respectively. The mice were assessed daily for clinical signs of EAE: 0 = normal, 1 = limp tail or mild hind limb weakness, 2 = moderate hind limb weakness or mild ataxia, 3 = moderately severe hind limb weakness, 4 = severe hind limb weakness or mild forelimb weakness or moderate ataxia, 5 = paraplegia with no more than moderate forelimb weakness, and 6 = paraplegia with severe forelimb weakness or severe ataxia, moribund condition or dead. Pellets containing various doses of E2 or placebo were implanted subcutaneously (dorsally) into mice 1 week before immunization. Serum levels of E2 were monitored by radioimmunoassay as described previously.6,17

Lymphocyte proliferation assay

Lymphocytes were harvested from spleens and lymph nodes and cultured in a 96-well flat-bottom tissue culture plate at 4 × 105 cells/well in stimulation medium in the presence of antigen-presenting cells (APC), irradiated (2500 rads) syngeneic thymocytes, at a ratio of 1 : 10 (T : APC), either with or without mMOG-35-55 peptide at varying concentrations. The cells were incubated for 3 days at 37° in 7% CO2, and pulsed with 0·5 μCi [3H]thymidine for the final 18 hr of culture. The cells were harvested onto glass fibre filters, and incorporated radioactivity was measured with a liquid scintillation counter. The counts per minute (c.p.m.; mean ± SD) were calculated from triplicate wells. The stimulation index was calculated by dividing the experimental c.p.m. by the control c.p.m.


Intact spinal columns were removed from experimental and control group mice. The spinal cords were dissected after fixation in 4% paraformaldehyde, dehydrated and embedded in paraffin before sectioning. To examine neuroinflammation, the sections were stained with haematoxylin and eosin (H & E). To examine demyelination the sections were stained with Luxol fast blue plus periodic acid Schiff. The sections were analysed by light microscopy after staining and were recorded with a digital camera.

Flow cytometry

For membrane staining (for CD4 and CD19), 1 million cells were stained at 4° in the dark with appropriate antibody dilutions in staining buffer (phosphate-buffered saline containing 0·5% bovine serum albumin and 0·02% sodium azide). Intracellular staining (for FoxP3 and PD-1) was performed following the protocol recommended by eBioscience. Briefly, 4 million cells were surface stained following standard procedures. After washing, the cells were fixed overnight and washed twice with 0·5 ml permeabilization buffer. The cells were costained for 15 min with FcBlock and immunoglobulin G–peridinin chlorophyll protein followed 30 min later with fluorescently labelled antibodies to FoxP3 or isotype control. The cells were then washed twice with 2 ml permeabilization buffer and once with 1 ml staining buffer, and resuspended in staining buffer. Flow cytometry data were collected on LSRII and FACSCalibur flow cytometers (BD Bioscience, San Jose, CA), and analysed using flowjo software (Tree Star, Ashland, OR). Data represent 50 000–100 000 events.

Measurement of IL-17 production

Lymph node and spleen cells were cultured at 4 × 106 cells/well in a 24-well flat-bottom culture plate in stimulation medium (RPMI-1640, 1% sodium pyruvate, 1%l-glutamine, 0·4%β2-mercaptoethanol, 10% fetal bovine serum) with 25 μg/ml mMOG-35-55 peptide for 48 hr. Supernatants were then harvested and stored at −80° until tested for cytokines. Culture supernatants were assessed for IL-17 production levels using a Luminex Bio-Plex mouse cytokine assay kit following the manufacturer’s instructions.

Statistical analyses

Mean values from each experiment were compared statistically. Differences in daily clinical scores, peak scores and cumulative disease index (CDI) were evaluated by the Mann–Whitney test; incidence was evaluated by Fisher’s exact test; disease onset, cytokine secretion, T-cell proliferation and flow cytometric data were compared by the unpaired Student’s t-test. Data are represented as mean ± standard deviation of the mean (SD). All presented data represent one of two to four independent experiments.


E2 treatment upregulated PD-1 selectively in the CD4+ FoxP3+ Treg-cell compartment

The effects of E2 treatment were evaluated on the expression pattern of PD-1 in the periphery of B6 mice immunized with mMOG-35-55 peptide. The mice were treated with 2·5 mg/60 day release E2 pellets 1 week before immunization and killed 27 days after immunization. The pellets were demonstrated to consistently release pregnancy levels of E2 (1·5–2 ng/ml) into the serum of treated mice.18 As shown in Fig. 1(a), E2 treatment increased the PD-1 levels in CD4+ FoxP3+ regulatory T cells, but not in CD4+ FoxP3 or CD4 T-cell populations. At this dose, E2 also increased the frequency of CD4+ FoxP3+ Treg cells (Fig. 1b). Therefore, at pregnancy levels, E2 treatment affected both the frequency and intensity of PD-1-expression in CD4+ FoxP3+ Treg cells.

Figure 1.

 Oestradiol (E2) treatment selectively upregulated programmed death 1 (PD-1) expression in the CD4+ FoxP3+ regulatory T (Treg) cell compartment. B6 mice were treated with E2 (2·5 mg/60 day release) or placebo pellets 7 days before immunization and killed 27 days after immunization. The frequency of CD4+ FoxP3+ cells and PD-1 expression levels were measured by flow cytometry. (a) The effect of E2 treatment on the level of PD-1 expression in CD4+ FoxP3+, CD4+ FoxP3 and CD4 cells. Each histogram overlay of PD-1 fluorescence intensity represents an E2-treated and a placebo-treated mouse. The bar chart represents the average of five or six mice/group. (b) The percentage of CD4+ FoxP3+ Treg cells in total CD4+ T cells was boosted by E2 treatment at pregnancy levels. Each quadruple-log plot of fluorescence intensity shows the results from a representative mouse. The bar chart represents the average of five or six mice/group. ***P < 0·001 as indicated by Student’s t-test.

PD-1 level in Treg cells was correlated to EAE suppression

We subsequently investigated whether E2-induced upregulation of PD-1 in Treg cells was dose dependent. Indeed, E2 pellets administered 7 days before immunization protected the mice from EAE induction in a dose-dependent manner (Fig. 2a). Similarly, E2-induced upregulation of PD-1 within the Treg compartment was also dose-dependent. The level of PD-1 in Treg was linked closely to the reduction in disease severity induced by E2 (Fig. 2b,c). Interestingly, only the 2·5 mg/60 day release E2 pellets, which produce pregnancy levels of serum E2 (not shown), but not pellets ranging from 0·001 to 0·71 mg, significantly increased the frequency of CD4+ FoxP3+ Treg cells. This result is in sharp contrast to the sensitive nature of E2-induced EAE suppression and upregulation of PD-1 expression in Treg, which increased after treatment with as little as 0·025 mg/60 day release E2 pellets. In unimmunized mice, however, E2 failed to induce any change in PD-1 levels in Treg cells, even though the Treg-cell frequency was slightly, but not significantly, increased. Consequently, E2 may only increase PD-1 expression in Treg cells upon induction of EAE.

Figure 2.

 Oestradiol (E2)-induced experimental autoimmune encephalomyelitis (EAE) suppression was correlated to the level of programmed death 1 (PD-1) expression in the regulatory T (Treg) cell compartment, but not to the frequency of Treg cells in CD4 T cells. Immunized mice were given pellets containing various doses of E2 or placebo 7 days before immunization. Unimmunized mice received the same amount of E2 and placebo as controls. (a) E2 treatment protected mice from EAE in a dose-dependent manner. (b) E2 treatment decreased the CDI and upregulated the PD-1 expression in Treg cells in a dose-dependent manner in immunized, but not unimmunized mice. (c) The CDIs of E2-treated mice negatively correlated to the level of PD-1 expression in the Treg compartment. *P < 0·05, **0·01 or ***0·001 as indicated by Mann–Whitney (for CDI) or Student’s t-test (for data other than CDI).

PD-1 is indispensible for E2-induced protection against EAE

The contribution of PD-1 to E2-induced protection against EAE was evaluated in age-matched wild-type (WT; B6) and PD-1KO female mice. Mice were implanted with E2 (2·5 mg/60 day release) or placebo pellets 1 week before immunization. As shown in Fig. 3(a), placebo-treated WT and PD-1KO mice developed severe EAE, whereas E2-treated WT mice developed no clinical signs of disease. In contrast, E2-treated PD-1KO mice had delayed onset but were not protected from EAE. T-cell responses to mMOG-35-55 peptide were more potent in PD-1KO mice compared with WT mice (Fig. 3b), which was consistent with the negative regulatory nature of PD-1. There was little difference in lymphocyte proliferation responses between placebo-treated and E2-treated mice. Pathologically, all placebo-treated mice and E2-treated PD-1KO mice had substantial central nervous system infiltration and demyelination, whereas no pathological signs of EAE were observed in E2-treated WT mice (Fig. 3c). Taken together, our results indicated a significant role for PD-1 in E2-stimulated protection against EAE.

Figure 3.

 Programmed death 1 (PD-1) is required for oestradiol (E2)-induced protection against clinical and histological experimental autoimmune encephalomyelitis (EAE). PD-1-deficient (PD-1KO) and wild-type (WT) mice were implanted with 2·5 mg/60 day release E2 pellets (or placebo) 1 week before immunization. Lymphocytes from spleen and lymph nodes were evaluated for T-cell proliferation and spinal cords were fixed for pathological examination. The experiment was repeated twice with a total of 5–10 mice in each group. (a) E2-induced protection against clinical EAE was abrogated in PD-1KO mice. (b) Lymphocytes from PD-1KO and WT mice proliferated in response to antigen stimulation. (c) E2-treatment prevented central nervous system infiltration and demyelination in WT but not in PD-1KO mice. Paraffin-embedded spinal cord sections were stained with Luxol fast blue plus periodic acid Schiff and recorded with a microscope equipped with a digital camera. **P < 0·01 as indicated by Student’s t-test.

E2 does not protect Treg-deficient mice against Spontaneous (Sp)-EAE

To study whether the PD-1 molecule expressed outside the Treg-cell compartment could mediate protection by E2, we tested the efficacy of E2 in blocking Sp-EAE in Treg-deficient mice generated by crossing MBP-TCRα/β double-transgenic mice with Rag1−/− mice on the B10.PL background.19 As shown in Fig. 4(a), no CD4+ FoxP3+ Treg cells were detectable in Treg-deficient mice, as expected. These mice typically develop Sp-EAE at about 10 weeks of age. As shown in Fig. 4(b), neither E2 nor placebo pellets administered when the Treg-deficient mice were ≥ 3 weeks old prevented them from developing Sp-EAE. The PD-1 molecules residing outside the Treg compartment did not, therefore, mediate the protective effects of E2.

Figure 4.

 Oestradiol (E2) did not protect regulatory T (Treg) cell-deficient mice from developing spontaneous experimental autoimmune encephalomyelitis (Sp-EAE). Treg-deficient mice were created by crossing MBP-specific TCRα/β double-transgenic mice with RAG1−/− mice. E2 (2·5 mg/60 day release) or placebo pellets were implanted 3 weeks after the mice were born. (a) CD4+ FoxP3+ Treg cells were not detected in Treg-deficient mice. (b) The severity of Sp-EAE in E2-treated and placebo-treated mice was comparable. Data represent mean daily disease scores. The experiments were repeated three times and four or five mice/group were included in the experiment shown.

E2 treatment regulates IL-17 in a PD-1-dependent manner

Lastly, we investigated the effect of E2 treatment on the production of IL-17. As shown in Fig. 5, E2 treatment greatly reduced the secretion of MOG-35-55 peptide-activated IL-17 in lymphocytes from WT mice. In contrast, E2 treatment slightly increased a lower baseline level of IL-17 secretion in MOG peptide-activated lymphocytes from PD-1KO mice. Therefore, E2 treatment regulated IL-17 production through a PD-1-dependent mechanism.

Figure 5.

 Oestradiol (E2) downregulated interleukin-17 (IL-17) production in a programmed death 1 (PD-1)-dependent manner. IL-17 secretion was evaluated in 48 hr supernatants from MOG-35-55 peptide-stimulated lymphocytes (25 μg/ml) from placebo- and E2-treated wild-type (WT) and PD-1-deficient (PD-1KO) mice. *P < 0·05 or ***0·001 compared to placebo control. The experiment was repeated twice with a total of 7–10 mice in each group.


The precise mechanisms by which oestrogens protect mice from EAE induction have not been completely defined. Our previous studies ruled out the possibility of direct E2 regulation of encephalitogenic T cells by E2,20 as well as the requirement for the anti-inflammatory cytokines, IL-4 and IL-10,7 but suggested a link between PD-1 and Treg-cell suppression in the E2 regulatory mechanism.13 We therefore evaluated critically the possible roles played by PD-1. We first demonstrated that E2-induced upregulation of PD-1 occurred selectively in the CD4+ FoxP3+ Treg-cell compartment. Indeed, the level of PD-1 expression in Treg cells correlated better with the disease-suppressing activity of E2 than the percentage of PD-1+ Treg cells, which increased only when pregnancy levels of E2 were given.10 We then showed that the suppressive effects of E2 on both EAE and IL-17 production were abrogated in PD-1KO mice. Finally, we demonstrated that E2 could not inhibit Sp-EAE in Treg-deficient mice, so ruling out any participation of PD-1 expressed outside the Treg-cell compartment. Taken together, our results indicate that PD-1 expressed within the Treg-cell compartment serves as a critical mediator of E2-induced protection against EAE.

Treatment with E2 is a powerful regulator of cytokines; E2 decreases the production of proinflammatory cytokines, such as TNF-α, and increases the production of anti-inflammatory cytokines, such as IL-4 and IL-10.7,8 Although there is no doubt that these cytokines are important immunoregulators, their role in E2-induced protection against EAE is not clear. First, the anti-inflammatory cytokines, IL-4 and IL-10, are dispensable because E2 could protect against EAE in IL-4 and IL-10 gene-deficient mice, even though their production was obviously enhanced in E2-treated WT mice.7 On the other hand, E2 treatment of WT mice decreased the production of the proinflammatory cytokine TNF-α.7,8 However, TNF-α gene-deficient mice on both the B6 and 129 backgrounds developed a more severe form of EAE characterized by increased inflammation and demyelination, which could be ameliorated by administration of TNF-α.21 Under these circumstances, TNF-α appeared to have an anti-inflammatory role in EAE, so it is difficult to argue that the decrease of TNF-α production by E2 could account for the complete blockade of EAE induction by E2. The finding that E2 drastically decreases the production of IL-17 in WT mice is significant because IL-17 and Th17 cells are known to be critically involved in the pathogenesis of MS and EAE.15 It was reported previously that induction of EAE was prevented in IL-17 gene-deficient mice.22 Our result in PD-1KO mice suggested that PD-1 expressed in Treg cells, which respond to low doses of E2, is a critical regulator of IL-17 production.

In summary, we demonstrated in this study that E2 protected against EAE and downregulated IL-17 through a PD-1-dependent pathway that selectively involved Treg cells. Our results therefore shed new light on the mechanism through which oestrogens may regulate autoimmunity.


This work was supported by National Institutes of Health grants NS45445, NS23221 and NS49210; National Multiple Sclerosis Society grants RG3405 and PP1295; The Nancy Davis Center without Walls; and the Biomedical Laboratory R&D Service, Department of Veterans Affairs. The authors thank Paul Bui for providing excellent breeding service and animal care, Dr Honjo of Kyoto University for providing PD-1KO mice, and Eva Niehaus for assistance in manuscript preparation.


Oregon Health & Science University, the Department of Veterans’ Affairs and Drs Offner and Vandenbark have a significant financial interest in Pipex, a company that may have a commercial interest in the results of this research and technology. This potential conflict was reviewed and managed by the Oregon Health & Science University and Portland VA Medical Center Conflict of Interest in Research Committees.