Effects of altered prenatal hormonal environment on expression of autoimmune disease in NZB/NZW mice.

F1 hybrid New Zealand Black (NZB) x New Zealand White (NZM) (NZB/NZW) mice spontaneously develop an autoimmune disease analogous to systemic lupus erythematosus (SLE). Testosterone experts a powerful suppressive effect on this disorder in adult NZB/NZW mice. A series of experiments was designed to determine if disease would also be suppressed by exposing fetal NZB/NZW mice to increased testosterone. A model was developed in which NZB dams carrying NZB/NZW fetuses were treated with testosterone in a dose adequate to masculinize the external genitalia in female fetuses. NZB/NZW mice that were derived from testosterone-treated dams and control NZB/NZW offspring were followed in a longevity study and had serial assays to assess development of SLE. Additional experiments were carried out to measure lymphocyte subsets and responses to mitogens. Results were compared with F1 hybrid offspring of C57BL/6 dams crossed with DBA/2 males, which are not autoimmune and do not develop SLE. Spleen cells from these groups were tested for Thy 1.2, CD4, CD8, and IgM receptors, and for responses to the mitogens Concanavalin A (ConA) and lipopolysaccharide. Control male NZB/NZW fetuses had unexpectedly high serum estradiol, which decreased significantly with maternal testosterone treatment. The testosterone-exposed male NZB/NZW fetuses developed into adults that lived longer than male NZB/NZW controls. Testosterone treatment of the dam was associated with elevated terminal anti-DNA levels but did not alter markers of renal diseases in adult NZB/NZW mice of either sex. Testosterone-exposed NZB/NZW females had altered T-lymphocyte subsets and testosterone-exposed males had increased response to ConA compared to controls. In male NZB/NZW fetuses whose mothers were administered testosterone, the naturally high level of circulating estradiol observed in untreated male fetuses was decreased significantly. This decrease was associated with an increase in longevity. This unique observation has important implications for fetal exposure to endocrine disruptors in the environment.


Introduction
Systemic lupus erythematosus (SLE) is a The predilection of SLE for women of chronic and incurable disease charac-reproductive age (1), association between terized by formation of autoantibodies disease flares and pregnancy (2), aberrant and immune-mediated damage to key pattern of estrogen metabolism (3), target organs such as the kidney (1). and relatively low levels of circulating This paper is part of the Wingspread Work Session on Chemically-induced Alterations in the Developing Immune System: The Wildlife/Human Connection held 10-12 February 1995 in Racine, Wisconsin. Manuscript received 5 October 95; manuscript accepted 5 March 1996. This research was supported by The NIH-NICHHD through Contract NO1-HD-52909, the Department of Veterans' Affairs, and grant DCB 8518094 from the National Science Foundation. Dr. Walker was supported during the tenure of this work by a Clinical Investigator Award from the Department of Veterans' Affairs. Address  Matings between New Zealand Black (NZB) and New Zealand White (NZW) mice produce F1 NZB/NZW hybrids, which develop a disease analogous to SLE. These animals produce antibodies directed against double-stranded DNA (anti-DNA) and die prematurely with glomerulonephritis and renal failure (5). Sex hormones are important modifying factors in the disease of NZB/NZW mice. Female NZB/NZW mice have early-onset disease and accelerated mortality, with mean longevity of 40 weeks (6), and autoantibody formation is accelerated by treatment with exogenous 17,3-estradiol (7).
In contrast, androgens are beneficial in NZB/NZW mice. Males are expected to live longer than the females. The mean age of death in male NZB/NZW mice is 64 weeks (6), but early death occurs if male mice are castrated before puberty (7). Treatment with exogenous testosterone improves survival in male castrates (7). Testosterone therapy also improves survival in NZB/NZW females that are treated before or after the onset of renal disease, and ovariectomy is not necessary for female mice to receive therapeutic benefit (8,9).
Although testosterone modulates the severity of autoimmune disease in NZB/ NZW mice, the critical periods at which key interactions occur between endogenous androgens and the immune system have not been completely defined. It seems logical to assume that the protective effects of naturally secreted androgens are exerted in the postpubertal period in males, when serum testosterone is found in high concentrations. Experimental evidence has shown, however, that early exposure to testosterone also has an important role in modifying the course of SLE in NZB/NZW males. Castration leads to acceleration of disease in male NZB/NZW mice if the surgery is performed at 2 weeks of age, but postpubertal castration is relatively ineffective (8,10). A recent report from this laboratory (11) showed that NZB/NZW males had little androgen receptor-mediated protection from severe disease between the ages of 6 weeks and 1 year.
Colborn et al. (12) have reviewed evidence that maternal exposure to endocrinedisrupting chemicals in the environment can permanently alter the functioning of developing systems, such as the immune system, in offspring due to transport of chemicals from the mother during fetal life and lactation. A number of endocrinedisrupting chemicals typically encountered in water, air, and food and commonly found in human tissue are capable of permanently disrupting developing systems by binding to receptors for estrogen and androgen, which regulate the differentiation of many tissues; endocrine disruptors can either mimic or antagonize the actions of natural sex steroids (13)(14)(15). Thus, findings concerning the role of estrogen and androgen in the etiology of autoimmune disease are relevant to our understanding of the potential impact that exposure to endocrine disruptors during development can have on this disease. This paper is the third in a series of publications that report the effects of hormonal manipulations on the developing immune system of NZB/NZW mice during prenatal life (16,17). A model was created in which NZB dams, pregnant with NZB/NZW fetuses, were treated in late pregnancy with testosterone or with the androgen blocker, flutamide. Both maternal treatments failed to produce long-lasting influences on female offspring, but the treatments were associated with increased longevity in the NZB/NZW males (16).
This paper focuses on gestational testosterone therapy in NZB dams and the influences of this treatment on their NZB/NZW offspring. Although maternal exposure to exogenous testosterone did not affect the lifespan in the NZB/NZW females produced by the treated dams, maternal testosterone treatment was associated with an unusual effect in the males. Unexpectedly, NZB/NZW males from androgen-implanted mothers had prolonged survival and lived significantly longer than comparable control males from untreated NZB dams ( Figure 1) (16).
The hormonal status of the male NZB/NZW fetuses was unique. At the gestational age of 18 days, control males had very high serum estradiol that exceeded concentrations in littermate females by a factor of 2.6. Maternal treatment with testosterone significantly reduced (by onethird) serum estradiol concentrations in the male fetuses (Table 1) (17). Testosteronetreated NZB dams had significantly elevated serum testosterone (data not shown), but maternal treatment did -not alter fetal serum testosterone levels ( Table 1) (17). Age at death (weeks) Figure 1. Survival in female and male NZB/NZW mice. Female mice and male mice were delivered by cesarean section from NZB dams that had been treated from days 13-18 of pregnancy with implanted testosterone, 0.75 mg. Control NZB/NZW females and males were delivered by cesarean section from sham-treated NZB dams. The mean age at death in female offspring of testosterone-treated dams did not differ significantly from offspring of control dams. Mean longevity in males from the dams treated with testosterone was prolonged (61 weeks±4) compared to males from control dams (48±2); p<0.001. Figure 1 is adapted from Keisler et al. (16) and is used with permission of the publisher. offspring. Animals were purchased at 5 weeks of age and maintained with an automatic 12-hr on, 12-hr off light cycle in the Research Service of the Harry S. Truman Memorial Veterans' Hospital under conditions described in an earlier publication (17). At 6 weeks of age, NZB females were paired with NZW males, and C57BL/6 females were paired with DBA/2 males. Females were checked daily for vaginal plugs, and the appearance ofa plug was day 0. Testosterone Treatment Implants were made of 1-cm lengths (between the plugged ends) of Silastic tubing Environmental Health Perspectives * Vol 104, Supplement 4 * August 1996 (0.06-inch inner diameter, 0.125-inch outer diameter; Dow Corning, Midland, MI) filled with 0.75 mg testosterone (Sigma Chemical Co., St. Louis, MO) in 0.02 ml sesame oil. Fifteen NZB dams and eight C57BL/6 dams were anesthetized with Metofane (methoxyflurane) (Pitman-Moore, Inc., Mundelein, IL) and implanted on day 13 of pregnancy (16). The 0.75-mg dose of testosterone was chosen because it was determined in preliminary experiments that this was the lowest dose that significantly masculinized the external genitalia of female offspring. Masculinization was defined on the basis of increased anogenital space (the length between the anus and the genital papilla that becomes the scrotum in males). The dose was, however, low enough so that defeminization of female fetuses did not occur (i.e., these animals retained the capacity to ovulate) (16).
The vagina of the female mouse is usually open by 4 weeks of age, but it was anticipated that opening would be altered by prenatal exposure to exogenous testosterone (18). Vaginal opening did not occur in female F1 offspring of testosterone-treated NZB and C57BL/6 dams by 12 weeks of age, and all of these offspring had hysterectomies (16) to prevent endometritis that would have occurred because of lack of uterine drainage. Ovaries were carefully left intact. This surgery did not cause expression of disease to differ, and data from these mice were pooled with data from other female NZB/NZW offspring of sham-treated dams (16).

Sham Treatment
Twelve NZB dams and 5 C57BL/6 dams received implants containing 0.02 ml sesame oil on day 13 of gestation, and 8 NZB dams and 4 C57BL/6 dams received injections of sesame oil-alcohol vehicle on days 13,14,15,16,17, and 18 of gestation. Offspring of dams that received either sham treatment had similar anogenital spaces and comparable expression of disease, and the results were combined.

Derivation ofOffipring
Pups were delivered from all dams by cesarean section on day 18 of gestation, fostered to post-partum CF-1 mothers from the outbred colony of F. vom Saal, and weaned 21 days after birth (16).
Longevity Study of NZB/NZW Off'ring A longevity study was performed to determine if NZB dams exposed to testosterone in the last third of pregnancy produced NZB/NZW offspring with altered expression of autoimmune disease. The following groups of NZB/NZW offspring were studied: female (n=24) and male (n=22) offspring of testosterone-treated dams and control female (n= 34) and male (n= 29) offspring of sham-treated dams.

Assessment ofActive Autoimmune Disease
Mice were bled from the orbital plexus and urine was collected at 12, 24, 36, and 48 weeks of age. Mice were examined daily for signs of disease. They were bled, checked for albuminuria, sacrificed, and necropsied according to the protocol described in earlier publications when they developed neoplasms or appeared moribund (19)(20)(21) (25). Fluorescence signals were collected at 5 1 5 ± 10 nM (FITC) and 600 ± 15 nM (PE) in singleor dual-parameter histogram format from 10,000 cells with the lymphocyte light scatter gates, as defined by forward and perpendicular light scatter signals. Data were expressed as the percentage of cells bearing the antigen of interest. When necessary, nonspecific staining was subtracted from positive cells.

Lymphocyte Transformation
Mitogenic responses of splenocytes to Concanavalin A (ConA) (Pharmacia), 0.5 jig, or lipopolysaccharide (LPS) (E. coli 026:B6; Difco, Detroit, MI), 40 pg, were assayed by incubating the cells with either ConA or LPS to stimulate cell division, pulsing the dividing cells with 3H-thymidine, harvesting the cells on glass fiber filter paper, and using recovered CPM as a measure of cell proliferation. This assay was performed as described in an earlier publication (24), except that phenol red-free RPMI 1640 media (Gibco, Grand Island, NY) was used. Mean values of triplicate samples were determined, and results were expressed as differences in counts per minute (ACPM) between stimulated and unstimulated cultures.

Statistical Analyses
Analysis of variance procedures and mean comparison tests were used to determine differences in dependent variables in each experiment (General Linear Models, SAS Institutes, Cary, NC). Planned comparisons were made using least significant difference procedures (26).

Results
Longevity Study of NZB/NZW Offsprng Testosterone treatment of NZB dams in the last third of gestation affected longevity of the offspring in a sex-dependent manner. Figure 1 illustrates cumulative survival in NZB/NZW females and males that were delivered from testosterone-treated or sham-treated control dams. The major cause of death in mice of both sexes was SLE, manifested as renal disease leading to renal failure and vasculitis (16). Age at death in female offspring of testosteronetreated dams was 34 ± 1 week; longevity in this group did not differ from female offspring of control dams (37 ± 2 weeks). In contrast, outcome in NZB/NZW males was affected by the maternal testosterone therapy. Males from testosterone-treated dams had lifespans of 61 ±4 weeks versus male offspring of control dams (48 ± 2 weeks; p<0.001) (16).

Assessment ofActive
Autoi'mune Disease Serial assays of anti-DNA, an important marker of disease activity in NZB/NZW mice (27), are shown in Table 2. All female offspring had the expected agedependent increase in anti-DNA concentration. In females produced by the testosterone-treated NZB dams, anti-DNA appeared to exceed control values at 36 weeks of age; this increase was significant in serum that was collected at spontaneous death. Anti-DNA concentrations in male offspring also increased with age and were not altered by testosterone treatment of the dam.
Progression of kidney disease was not changed by perinatal exposure to testosterone in mice of either sex. Table 3 illustrates albuminuria, which appeared in female and male mice by 24 weeks of age.
Albuminuria was not accelerated in offspring of testosterone-treated dams. Terminal increases in BUN concentrations reflected the development of renal failure at the time of spontaneous death in NZB/NZW mice of both sexes (Table 3).    Mitogenic Responses in C57/DBA2 Mice In nonautoimmune C57/DBA2 hybrids, active responses were observed when spleen cells were cultured with either mitogen. Response were not influenced by prenatal treatment, sex, or age (Table 5).

Discussion
Gonadal hormones mediate severity of autoimmune disease in NZB/NZW mice, but the mechanisms and timing of the most important interactions with the immune system have not previously been defined.
Estrogens are capable of suppressing natural killer cell cytotoxic activity (28) and stimulating intense B-cell activity (28,29) in NZB/NZW mice. Androgenic hormones exert widespread influences on cellmediated immunity, sustaining interleukin (IL)-2 production (30) and increasing T-cell activity in castrated males (31). It may be reasoned that hormones exert modulating effects within the developing immune system of the fetus, whereas other important influences occur when concentrations of hormones change at puberty or after the animal has reached sexual maturity. We have investigated hormonal interactions in the last third of gestation, a period in which gonadal steroids regulate the differentiation of numerous organ systems. Several unique models have been developed in this laboratory to facilitate studies of altered prenatal hormone concentrations on the subsequent course of autoimmune disease in NZB/NZW mice. In the testosterone implant model, pregnant NZB dams are treated on days 13 to 18 of gestation with implanted testosterone in a dose that somewhat masculinizes genitalia of the female NZB/NZW offspring (16). This treatment results in long-lived male NZB/ NZW fetuses that survive for a significantly longer period compared to NZB/NZW males from control dams (16).
This present paper presents measurements of three parameters of active autoimmune disease in female and male NZB/NZW offspring of testosterone-treated and control dams. Anti-DNA, albumin-uria, and BUN were assayed serially during the first year of life. These indicators increased as expected as the animals grew older. At 36 weeks of age, female NZB/NZW mice from testosterone-treated dams had active production of anti-DNA (mean = 61%). Anti-DNA remained elevated in the terminal phase of SLE in this group (60% vs 45% in females from shamtreated dams; p < 0.05). The increased anti-DNA concentration, a marker of active disease, was not associated with accelerated mortality in the females from hormonetreated dams.
This analysis suggested that prenatal exposure to exogenous androgens was associated with accelerated autoantibody formation in the females. Paradoxically, their lifespans were not shortened compared to same-sex controls. In the corresponding male littermates, testosterone-exposed offspring did survive longer than expected in the face of active autoimmune disease and renal insufficiency. Their increased longevity was compatible with the existence of a protective factor or factors that prevented early mortality in the face of active anti-DNA formation. Death was delayed in these males even though they had longstanding, active disease that paralleled males from sham-treated dams.
Our understanding of the protective effects that the prenatal hormone milieu may exert on longevity is confounded by changes that occur as the mice become adults. Females produce cyclic estrogen and prolactin, which are immunostimulatory, and males produce immune-suppressing androgens. We are currently examining mechanisms by which the prenatal hormonal environment could interact with post-pubertal hormone surges in adult NZB/NZW mice to prolong survival in animals with active SLE.
Additional groups of NZB/NZW offspring of testosterone-treated and control dams were tested for percentages of cells bearing T-lymphocyte or B-lymphocyte markers and for mitogenic responses to ConA and LPS. Results were compared with a genetically distinct F1 hybrid (C57/DBA2) that does not develop autoimmune disease. Of interest, 8-week-old female NZB/NZW offspring of testosterone-treated dams had significant reductions in percentages of spleen cells bearing surface antigens that are characteristic of T lymphocytes (Thy 1.2), and the helper/ inducer T-lymphocyte subset (CD4). This apparent reduction in T lymphocytes, however, did not suppress responses to the T-lymphocyte mitogen, Con-A. The difference in T-lymphocyte subsets did not persist to 16 weeks of age, and was not noted in NZB/NZW males or C57/DBA2 hybrids. Male NZB/NZW mice from treated dams had no changes in lymphocyte subsets, but a significant increase in response to ConA was noted at 8 weeks of age.
Maternal testosterone treatment did have some longstanding effects on lymphoid cell subsets and responses to mitogens. In testosterone-exposed female NZB/NZW offspring, the reduced numbers of CD4+ T-helper cells at 8 weeks of age may have represented either a lag in production, or increased destruction of a key group of lymphocytes that is needed to sustain autoimmune disease in NZB/NZW mice (32). The relative paucity of CD4+ cells could have been protective, in that it might have affected expression of disease and resulted in some prolongation of life in this group of female mice with very high anti-DNA levels. On the other hand, the trend to increased responses to ConA and LPS in females and the significant increase in ConA responsiveness in males from testosterone-treated dams argues for a transient increase in activity of the immune system at an early age. The significance of this response is not clear; it could reflect an altered production of lymphocyte growth factors that may or may not be associated with the phenomenon of increased longevity in testosterone-exposed NZB/NZW males.
Our evaluations of the immune system in these animals did not identify a single protective factor that could have extended longevity in testosterone-exposed NZB/ NZW mice. Future investigations will focus on identifying changes in the prenatal and postnatal hormonal environments that may have affected the genesis of lymphocyte subsets or reset these cells to produce immune mediators that have protective as well as immune-stimulating properties. Additional studies are required to define lymphocyte profiles and function in juvenile animals, and to separate the effects of prenatal hormone exposure from immunosuppressive actions of endogenous androgens.
In the course of investigating the hormonal environment in which fetal NZB/NZW mice develop, we discovered that unmanipulated NZB/NZW males had unprecedented elevation of serum estradiol on day 18 of fetal life. For comparison, the observed concentration of serum estradiol in 18-day C57/DBA2 fetuses was (mean ± SEM) 158 pg/ml ± 48 for females and 183 pg/ml±38 for males. Values for CF-I mouse fetuses at 18 days of gestation were females, 113 pg/ml ±7 and males, 90 pg/ml ±4 (33). In the male NZB/NZW fetuses, the concentration of estradiol, (mean ± SEM) 327 pg/ml ± 42, greatly exceeded that of the female NZB/NZW littermates ( 125 pg/ml ± 44), p < 0.01. Testosterone treatment of the NZB dams decreased serum estradiol in male NZB/ NZW fetuses to (mean) 205 pg/mi (17).
The maternal treatment that decreased fetal estradiol in the NZB/NZW males did not alter their circulating testosterone, but resulted in significant prolongation of lifespans. These findings lead to the prediction that increased concentrations of estradiol during fetal life are related to early mortality from active autoimmune disease in NZB/NZW males.
The findings that NZB/NZW male fetuses have elevated serum estradiol compared to concentrations in male fetuses in nonautoimmune strains (16,17,34) and that a decrease in serum estradiol during fetal life in NZB/NZW males was associated with an increase in longevity, have important implications for studies of environmental influences on autoimmune disease. Our findings suggest that environmental endocrine disrupting chemicals that can bind to estrogen receptors and mimic the action of estradiol may affect the fetus and alter the subsequent course of autoimmune disease.
Numerous chemicals present in food, water, and air can bind to estrogen receptors and are referred to as environmental estrogens. Man-made environmental chemicals with estrogenic activity include pesticides, such as DDT and methoxychlor, components of commonly used products, such as plastics and soaps, and the epoxy lining of cans, such as octylphenol, nonylphenol and bisphenol-A (14,15). Environmental estrogens synthesized by man differ from naturally occurring plant estrogens (phytoestrogens) in that only man-made chemicals persist while stored in fat for prolonged periods after ingestion; phytoestrogens are rapidly cleared from the circulation (14). Thus, maternal exposure to persistent environmental estrogens before or during pregnancy and lactation can result in exposure of offspring to the chemicals during differentiation of critical organ systems.
While it has been recognized for some time that endogenous steroids modulate immune function and are involved in sex differences in immune function, relatively little attention has been paid to the effects of exposure to steroids during critical periods in the organization of the developing immune system on subsequent immune function. Steroids exert organizational effects on differentiating tissues, including tissues in the immune system, such that the functioning of the tissue throughout the remainder of life is permanently altered as a result of a change in hormone levels during differentiation (35).
The possibility that exposure of fetuses or newborns to endocrine-disrupting chemicals might be involved in a host of diseases associated with disfunction in the immune system needs to be examined. Information concerning parental exposure to endocrine disruptors can then be analyzed in relation to diseases that do not become manifest until the offspring are adults.