Journal of Autoimmunity (1990) 3,257-270
Estrogen Induced Suppression of Collagen Arthritis V: Physiological Level of Estrogen in DBA/l Mice is Therapeutic on Established Arthritis, Suppresses Anti-type II Collagen T-cell Dependent Immunity and Stimulates Polyclonal B-cell Activity
Liselotte Jansson, Anita Mattsson, Ragnar Mattsson and Rikard Holmdahl Department
of Medical
Zoophysiology,
and Physiological
Chemistry
Uppsala University,
and Department
Uppsala,
of
Sweden
Immunization of castrated female DBA/l mice with rat type II collagen (CII) induces severe polyarthritis with an onset 3-5 weeks after immunization and with SO-lOO”/oincidence. Estrogen treatment, inducing physiological 17jl-estradiol (E2) levels, during a limited period before and after the immunization, or during another period before the expected onset of arthritis, delayed the arthritic onset by approximately 10 days but did not affect the incidence of severity of arthritis. Treatment with physiological doses of E2 after onset of arthritis decreased severity and duration of disease. The T-cell dependent anti-C11 autoantibody response was suppressed if the E2 treatment was given immediately before and after CII immunization and was not significantly affected if E2 treatment was given after CII immunization. Neither the total anti-C11 Ig levels nor the anti-C11 IgG2a levels correlated with development of arthritis. We also titrated the serum levels of estrogen and recorded the vaginal smear response after injections of various doses of E2. This enabled us to work in a physiological range of estrogen levels, spanning the levels found at the end of pregnancy and those found during the normal estrous cycle. These levels were found to suppress antigen-specific T-cell functions but enhance certain B-cell activities since the delayed type hypersensitivity (DTH) reaction against CII was suppressed while the total number of splenic Ig-secreting cells increased. These findings suggest that estrogen in physiological doses is therapeutic for the development of collageninduced arthritis and that estrogen exerts dualistic effects on the immune system by suppressing T-cell functions and stimulating certain B-cell activities. The suppressive effect on arthritis could not be explained by Correspondence to: Liselotte Jansson, Department of Medical and Physiological Chemistry, Box 575 Uppsala University, S-75123 Uppsala, Sweden. 257 0896-8411/90/030257+
14 $03.00/O
0 1990 Academic Press Limited
258
L. Jansson et al. suppression of anti-C11 autoantibody response and must therefore depend on other T-cell-mediated functions.
Introduction
It is well established that estrogen plays an important modulatory role on the immune system in mammals [ 1,2]. Estrogen also has modulatory effects on autoimmune diseases but has a disease-promoting effect on the course of human systemic lupus erythematosus (SLE) and may be an important contributing factor for the female preponderance and for the increased risk during pregnancy of SLE development [3]. Similarly, it is well established that the NZB/W mouse, which spontaneously develops lupus disease, is also sensitive to the action of estrogen. These findings have led to the suggestion that estrogen may be an accelerating factor for autoimmune diseases [4]. However, it has been increasingly evident that certain diseases also believed to be of autoimmune nature, such as rheumatoid arthritis (RA), are ameliorated under the influence of estrogen. Estrogen-containing contraceptive pill consumption and menopausal substitution treatment have been reported by several investigators to reduce the incidence of RA. These findings suggest that it is important to clarify the role of estrogen in the modulation of autoimmune arthritis. We have suggested the use of the experimental collagen-induced arthritis model (CIA) as a tool for these investigations since it both displays certain characteristics in common with RA, such as T-cell infiltration in afflicted joints [Sj, and a chronic disease process [6]. Moreover, the CIA model shows gender-dependent differences in the susceptibility to arthritis; in rats and apes there is a female preponderance and in mice a male preponderance for susceptibility to arthritis ([7,8] Larsson, Klareskog, Holmdahl, in preparation). Most interestingly, in both rats and mice, oophorectomy enhances susceptibility to arthritis, suggesting that estrogen may play an important regulatory role in the disease. We therefore have developed a program for investigation of the estrogen-mediated effects on CIA. We earlier reported that castration of female DBA/l mice increases the susceptibility [7] to CIA and that long-term estrogen treatment 193, but not progesterone treatment [lo], abrogates this increased susceptibility. We have now investigated the impact of physiological levels of estrogen on CT1 autoimmunity and on established CIA. Materials
and methods Mice
DBA/l mice, originally obtained from Jackson Laboratories, Inc., USA, were kept and bred in the animal unit at the Biomedical Center, Uppsala, Sweden. Female mice were used at the age of 8-10 weeks and were age-matched in all experiments. Operative
techniques and hormone treatment
The mice were castrated under chloral hydrate anesthesia; the ovaries were removed after a single incision through the back skin and a bilateral flank incision through the
Estrogen-induced
suppression of collagen arthritis
259
peritoneum. After surgery the mice were allowed to rest for 2 weeks before immunization. Estrogen therapy was given twice a week as subcutaneous injections of 17P-estradiol-benzoat (Sigma, St Louis, MO, USA) in a volume of 0.1 ml olive oil. Controls were given 0.1 ml olive oil. The efficiency of castration and sex hormone treatment were checked by inspection of vaginal smears [ 1 l] and determination of sex hormone serum levels. The mice were bled by retro-orbital puncture, and the sera collected individually and stored at 70°C until assayed. Collagens
Mouse Type II collagen (CII), from xiphoid cartilage, and rat CII, from the Swarm rat chondrosarcoma, were prepared by pepsin digestion and subsequent purification, as described by Miller (1972) [12]. The collagens were stored lyophilized and dissolved in 0.1 M acetic acid before use. Induction
of arthritis
Native CII dissolved in 0.1 M acetic acid at a concentration of I-2 mg/ml was emulsified in an equal volume of Freund’s complete adjuvant (Difco, Detroit, Michigan, USA) at 4”C, and 50 ~1 of this emulsion was injected intradermally around the root of the tail. The clinical severity of arthritis was quantified according to a graded scale as previously described [6]; 1 point = detectable swelling in one joint, 2 points = swelling in more than one but not all joints, 3 points = severe swelling of the entire paw and/or ankylosis. Each paw was graded; the maximum score per mouse was thus 12 points. Enzyme-linked
immunosorbent
assay (ELISA)
Quantification of anti-C11 reactive antibodies in sera was performed as previously described [ 131. Briefly, micro-ELISA plates (Dynatech, Plochingen, FRG) were coated with native mouse CII. Purified anti-C11 reactive antibodies were used as standards and titrated in lo-fold dilution steps in parallel with the unknown serum samples in ELISA plates. All tests were carried out in duplicate. The amount of bound antibody was estimated after incubation with goat anti-mouse IgG-Fc and goat anti-mouse IgM-Fc coupled to alkaline phosphatase (Jackson Lab, Avondale, Pennsylvania, USA) or with goat anti-IgG2a specific polyclonal antibodies coupled with horseradish peroxidase (Nordic Labs, Tilburg, the Netherlands). For each serum sample, the dilution that gave 50% absorbance as compared with maximum value obtained for the standard antibodies was related to the concentration (in mg/ml) of standard antibodies to give the same 50% absorbance value. By this method all calculations were performed with data from the steep portion of the slope, where the sample titration curves were parallel to the standard titration curves. The purified standard antibodies were obtained from rat CII immunized DBA/l mice. Assays of spleen functions
Cell suspensions were prepared from spleens, and the number of erythrocytes as well as the total cell number were estimated in a Coulter counter. The frequency of
260 L. Iansson et al. antibody-secreting cells was quantified in a protein plaque-forming cell assay as previously described in detail [ 141. Briefly, 30 ~1 spleen cell suspension (1 x lo6 cells per ml), 15 ~1 protein A (Pharmacia, Uppsala, Sweden) coupled sheep red blood cells (SRBC) (30% v/v), 10 ~1 guinea pig complement (1:2), 10 ul antiserum (1:401:80) and 100 nlO.5% agar (45°C) were mixed and plated on a 6-cm Petri dish (Nunc, Denmark). The dishes were then incubated overnight in humid air (37°C) and the plaques formed were counted. The antisera, rabbit anti-mouse IgG and rabbit anti-mouse IgM were produced in our laboratory.
Delayed
type hypersensitivity
(D TH)
assay
Rat CII-immunized DBA/l mice were challenged with 10 pg native rat CII, dissolved in 10 l.tl phosphate buffered saline (PBS), injected in the right ears at day 12 after immunization. The left ears were injected with 10 ul PBS. The ear swellings were measured 2 days later using a caliper.
Hormone
assays
The content of 17P-estradiol (E2) in pooled mouse serum was determined by a previously described radioimmunoassay (Miles-Yeda, Rehovot, Israel) [ 151. The assays were blanked with serum from castrated DBA/l female mice.
Statistical
analysis
Continuous variables (i.e. antibody levels) were analysed in terms of their group means, by Student’s t-test, dichotomous variables (i.e. incidence of arthritis), by their proportionate group frequencies (chi-square test), and the significance of arthritic scores by the Mann-Whitney U-test.
Results Induction
of physiological
levels of estradiol
In the mouse, physiological levels of 178-estradiol (E2) during the peak of estrous cycle is non-detectable with all hitherto used methods [ 161 as well as with the currently used RIA method. To determine a treatment protocol which corresponds to the physiological state, we titrated treatment with E2 on castrated female DBA/l mice, and determined serum levels as well as vaginal smear responses (Table 1). These data show that an E2 level comparable to that found in DBA/ 1 pregnant state (115 pmole/l) [lo] was obtained one day after an injection of 3.2 pg E2 while the lowest recorded serum level of estrogen was obtained with an injection of 1.6 l.tg E2. However, the level induced with 1.6 ug E2 is probably much higher than the level during the estrogen peak in the estrous cycle since an 8 times lower dose still induces a positive vaginal smear. As shown in Table 1, we found the limit for development of a positive vaginal smear response was obtained 2 days after injection of 0.2 l.rgE2 and
Estrogen-induced
261
suuuression of collagen arthritis
Table 1. Estrogen levels and vaginal smear responses after injection of 17P-estradiol Levels of 17fl-estradiol Vaginal smear responses’ Dose
n
dl
d2
Olive oil 0.025 pg 0.05 pg
4 4 4 9 9 5 5 5 5 5 5
D D D D D D D D D D D
D D D D E,P E E E E E E
0.1 t% 0.2 Pg 0.4 I-G 0.8 pg 1.6 pg 3.2 pg 6.4 pg 12.8 pg
d3
D D D D E,P,D E E E E E E
(pmole/U d4
D D D D E,P,D nd nd nd E nd nd
dl
d2
d3
d4
< 12 nd
< 12 < 12
Estrogen Induced Suppression of Collagen Arthritis V: Physiological Level of Estrogen in DBA/l Mice is Therapeutic on Established Arthritis, Suppresses Anti-type II Collagen T-cell Dependent Immunity and Stimulates Polyclonal B-cell Activity
Liselotte Jansson, Anita Mattsson, Ragnar Mattsson and Rikard Holmdahl Department
of Medical
Zoophysiology,
and Physiological
Chemistry
Uppsala University,
and Department
Uppsala,
of
Sweden
Immunization of castrated female DBA/l mice with rat type II collagen (CII) induces severe polyarthritis with an onset 3-5 weeks after immunization and with SO-lOO”/oincidence. Estrogen treatment, inducing physiological 17jl-estradiol (E2) levels, during a limited period before and after the immunization, or during another period before the expected onset of arthritis, delayed the arthritic onset by approximately 10 days but did not affect the incidence of severity of arthritis. Treatment with physiological doses of E2 after onset of arthritis decreased severity and duration of disease. The T-cell dependent anti-C11 autoantibody response was suppressed if the E2 treatment was given immediately before and after CII immunization and was not significantly affected if E2 treatment was given after CII immunization. Neither the total anti-C11 Ig levels nor the anti-C11 IgG2a levels correlated with development of arthritis. We also titrated the serum levels of estrogen and recorded the vaginal smear response after injections of various doses of E2. This enabled us to work in a physiological range of estrogen levels, spanning the levels found at the end of pregnancy and those found during the normal estrous cycle. These levels were found to suppress antigen-specific T-cell functions but enhance certain B-cell activities since the delayed type hypersensitivity (DTH) reaction against CII was suppressed while the total number of splenic Ig-secreting cells increased. These findings suggest that estrogen in physiological doses is therapeutic for the development of collageninduced arthritis and that estrogen exerts dualistic effects on the immune system by suppressing T-cell functions and stimulating certain B-cell activities. The suppressive effect on arthritis could not be explained by Correspondence to: Liselotte Jansson, Department of Medical and Physiological Chemistry, Box 575 Uppsala University, S-75123 Uppsala, Sweden. 257 0896-8411/90/030257+
14 $03.00/O
0 1990 Academic Press Limited
258
L. Jansson et al. suppression of anti-C11 autoantibody response and must therefore depend on other T-cell-mediated functions.
Introduction
It is well established that estrogen plays an important modulatory role on the immune system in mammals [ 1,2]. Estrogen also has modulatory effects on autoimmune diseases but has a disease-promoting effect on the course of human systemic lupus erythematosus (SLE) and may be an important contributing factor for the female preponderance and for the increased risk during pregnancy of SLE development [3]. Similarly, it is well established that the NZB/W mouse, which spontaneously develops lupus disease, is also sensitive to the action of estrogen. These findings have led to the suggestion that estrogen may be an accelerating factor for autoimmune diseases [4]. However, it has been increasingly evident that certain diseases also believed to be of autoimmune nature, such as rheumatoid arthritis (RA), are ameliorated under the influence of estrogen. Estrogen-containing contraceptive pill consumption and menopausal substitution treatment have been reported by several investigators to reduce the incidence of RA. These findings suggest that it is important to clarify the role of estrogen in the modulation of autoimmune arthritis. We have suggested the use of the experimental collagen-induced arthritis model (CIA) as a tool for these investigations since it both displays certain characteristics in common with RA, such as T-cell infiltration in afflicted joints [Sj, and a chronic disease process [6]. Moreover, the CIA model shows gender-dependent differences in the susceptibility to arthritis; in rats and apes there is a female preponderance and in mice a male preponderance for susceptibility to arthritis ([7,8] Larsson, Klareskog, Holmdahl, in preparation). Most interestingly, in both rats and mice, oophorectomy enhances susceptibility to arthritis, suggesting that estrogen may play an important regulatory role in the disease. We therefore have developed a program for investigation of the estrogen-mediated effects on CIA. We earlier reported that castration of female DBA/l mice increases the susceptibility [7] to CIA and that long-term estrogen treatment 193, but not progesterone treatment [lo], abrogates this increased susceptibility. We have now investigated the impact of physiological levels of estrogen on CT1 autoimmunity and on established CIA. Materials
and methods Mice
DBA/l mice, originally obtained from Jackson Laboratories, Inc., USA, were kept and bred in the animal unit at the Biomedical Center, Uppsala, Sweden. Female mice were used at the age of 8-10 weeks and were age-matched in all experiments. Operative
techniques and hormone treatment
The mice were castrated under chloral hydrate anesthesia; the ovaries were removed after a single incision through the back skin and a bilateral flank incision through the
Estrogen-induced
suppression of collagen arthritis
259
peritoneum. After surgery the mice were allowed to rest for 2 weeks before immunization. Estrogen therapy was given twice a week as subcutaneous injections of 17P-estradiol-benzoat (Sigma, St Louis, MO, USA) in a volume of 0.1 ml olive oil. Controls were given 0.1 ml olive oil. The efficiency of castration and sex hormone treatment were checked by inspection of vaginal smears [ 1 l] and determination of sex hormone serum levels. The mice were bled by retro-orbital puncture, and the sera collected individually and stored at 70°C until assayed. Collagens
Mouse Type II collagen (CII), from xiphoid cartilage, and rat CII, from the Swarm rat chondrosarcoma, were prepared by pepsin digestion and subsequent purification, as described by Miller (1972) [12]. The collagens were stored lyophilized and dissolved in 0.1 M acetic acid before use. Induction
of arthritis
Native CII dissolved in 0.1 M acetic acid at a concentration of I-2 mg/ml was emulsified in an equal volume of Freund’s complete adjuvant (Difco, Detroit, Michigan, USA) at 4”C, and 50 ~1 of this emulsion was injected intradermally around the root of the tail. The clinical severity of arthritis was quantified according to a graded scale as previously described [6]; 1 point = detectable swelling in one joint, 2 points = swelling in more than one but not all joints, 3 points = severe swelling of the entire paw and/or ankylosis. Each paw was graded; the maximum score per mouse was thus 12 points. Enzyme-linked
immunosorbent
assay (ELISA)
Quantification of anti-C11 reactive antibodies in sera was performed as previously described [ 131. Briefly, micro-ELISA plates (Dynatech, Plochingen, FRG) were coated with native mouse CII. Purified anti-C11 reactive antibodies were used as standards and titrated in lo-fold dilution steps in parallel with the unknown serum samples in ELISA plates. All tests were carried out in duplicate. The amount of bound antibody was estimated after incubation with goat anti-mouse IgG-Fc and goat anti-mouse IgM-Fc coupled to alkaline phosphatase (Jackson Lab, Avondale, Pennsylvania, USA) or with goat anti-IgG2a specific polyclonal antibodies coupled with horseradish peroxidase (Nordic Labs, Tilburg, the Netherlands). For each serum sample, the dilution that gave 50% absorbance as compared with maximum value obtained for the standard antibodies was related to the concentration (in mg/ml) of standard antibodies to give the same 50% absorbance value. By this method all calculations were performed with data from the steep portion of the slope, where the sample titration curves were parallel to the standard titration curves. The purified standard antibodies were obtained from rat CII immunized DBA/l mice. Assays of spleen functions
Cell suspensions were prepared from spleens, and the number of erythrocytes as well as the total cell number were estimated in a Coulter counter. The frequency of
260 L. Iansson et al. antibody-secreting cells was quantified in a protein plaque-forming cell assay as previously described in detail [ 141. Briefly, 30 ~1 spleen cell suspension (1 x lo6 cells per ml), 15 ~1 protein A (Pharmacia, Uppsala, Sweden) coupled sheep red blood cells (SRBC) (30% v/v), 10 ~1 guinea pig complement (1:2), 10 ul antiserum (1:401:80) and 100 nlO.5% agar (45°C) were mixed and plated on a 6-cm Petri dish (Nunc, Denmark). The dishes were then incubated overnight in humid air (37°C) and the plaques formed were counted. The antisera, rabbit anti-mouse IgG and rabbit anti-mouse IgM were produced in our laboratory.
Delayed
type hypersensitivity
(D TH)
assay
Rat CII-immunized DBA/l mice were challenged with 10 pg native rat CII, dissolved in 10 l.tl phosphate buffered saline (PBS), injected in the right ears at day 12 after immunization. The left ears were injected with 10 ul PBS. The ear swellings were measured 2 days later using a caliper.
Hormone
assays
The content of 17P-estradiol (E2) in pooled mouse serum was determined by a previously described radioimmunoassay (Miles-Yeda, Rehovot, Israel) [ 151. The assays were blanked with serum from castrated DBA/l female mice.
Statistical
analysis
Continuous variables (i.e. antibody levels) were analysed in terms of their group means, by Student’s t-test, dichotomous variables (i.e. incidence of arthritis), by their proportionate group frequencies (chi-square test), and the significance of arthritic scores by the Mann-Whitney U-test.
Results Induction
of physiological
levels of estradiol
In the mouse, physiological levels of 178-estradiol (E2) during the peak of estrous cycle is non-detectable with all hitherto used methods [ 161 as well as with the currently used RIA method. To determine a treatment protocol which corresponds to the physiological state, we titrated treatment with E2 on castrated female DBA/l mice, and determined serum levels as well as vaginal smear responses (Table 1). These data show that an E2 level comparable to that found in DBA/ 1 pregnant state (115 pmole/l) [lo] was obtained one day after an injection of 3.2 pg E2 while the lowest recorded serum level of estrogen was obtained with an injection of 1.6 l.tg E2. However, the level induced with 1.6 ug E2 is probably much higher than the level during the estrogen peak in the estrous cycle since an 8 times lower dose still induces a positive vaginal smear. As shown in Table 1, we found the limit for development of a positive vaginal smear response was obtained 2 days after injection of 0.2 l.rgE2 and
Estrogen-induced
261
suuuression of collagen arthritis
Table 1. Estrogen levels and vaginal smear responses after injection of 17P-estradiol Levels of 17fl-estradiol Vaginal smear responses’ Dose
n
dl
d2
Olive oil 0.025 pg 0.05 pg
4 4 4 9 9 5 5 5 5 5 5
D D D D D D D D D D D
D D D D E,P E E E E E E
0.1 t% 0.2 Pg 0.4 I-G 0.8 pg 1.6 pg 3.2 pg 6.4 pg 12.8 pg
d3
D D D D E,P,D E E E E E E
(pmole/U d4
D D D D E,P,D nd nd nd E nd nd
dl
d2
d3
d4
< 12 nd
< 12 < 12
