Endogenous Ag requirement for induction and maintenance of T cell tolerance has been extensively investigated in mice that express a transgenic Ag and/or its cognate transgenic TCR. In contrast, studies on tolerance for physiologically expressed self Ag and normal T cells are limited. Herein, we showed that the murine ovarian-specific ZP3 Ag is detectable from birth. Tolerance to ZP3 is detected in female relative to male mice. In comparison to males, 100-fold more ovarian peptide (pZP3) is required to elicit a comparable pathogenic response in females. Female tolerance to pZP3 was dependent on the presence of endogenous ovarian Ag, because neonatal ovariectomy converted the female response to that of males. Moreover, in female mice that were ovariectomized from the ages of 1–6 wk, the pZP3 responses were enhanced to the male level if ovaries were removed up to 7 days, but not 3 days, before adult challenge with pZP3. Thus, the physiologically expressed ZP3 Ag induces tolerance to pZP3, and the maintenance of tolerance is critically dependent on the continuous presence of the endogenous ovarian Ag. In contrast, exposure to endogenous ovarian Ag confined to the neonatal period is insufficient for the induction and maintenance of tolerance to ZP3.
To elucidate the mechanism by which pathogenic T cells are rendered unresponsive to peripherally expressed self Ags is a central immunological issue. Ideally, one would address the role of self Ag in tolerance induction by studying immunocompetent T cells in their natural environment with or without the presence of the Ag in question. This has been approached experimentally in transgenic mice expressing genes encoding both an Ag in a defined tissue and a TCR of known specificity (1). The transgenic approach has been valuable in that it overcomes the low frequency of peptide-specific T cells of normal animals and provides a tissue-specific, cognate Ag that is absent in nontransgenic animals.
The induction of T cell unresponsiveness to various transgenic proteins has been readily demonstrated. However, the mechanisms of tolerance were found to vary. In the model in which self Ag expression is not associated with detectable changes in the cognate T cells, the expression of the transgenic Ag per se may provide insufficient stimulus to actively tolerize the Ag-specific T cells (2). In those models where there is active recognition of self Ag by autoreactive T cells, tolerance mechanisms have included the induction of anergy (3), deletion (4), and the down-regulation of T cell coreceptors (5). On the other hand, the induction of spontaneous autoimmune pathology in TCR transgenic mice and in TCR/neo-Ag double transgenic mouse models has also been reported (6, 7, 8). The many experimental outcomes based on transgenic models suggest that the induction and maintenance of tolerance are a multifaceted process, influenced by expression of the transgenic Ags in terms of ontogeny, location, level, and immunogenicity.
The time at which self Ag is first expressed is considered a key factor that determines whether the recognition of self Ag by autoreactive T cells leads to tolerance or autoimmunity. The paradigm proposes that self tolerance is induced predominantly within the neonatal period (9). Therefore, a popular approach to physiological tolerance has been to evaluate the impact of neonatal Ag exposure to subsequent Ag challenge. Most studies of neonatal tolerance have evaluated the influence of neonatal injection of allogeneic cells or foreign proteins or peptides delivered with adjuvant on the subsequent adult response to the same Ags (10, 11, 12). Recently, however, it has been shown that neonatal injection of Ags leads to an immune response rather than tolerance (13, 14, 15). Because both the transgenic approach and the neonatal tolerance model are contrived in nature, their results should be validated by a more physiological approach in which the T cell repertoire has not been skewed in favor of a particular TCR and where the self Ags under investigation are physiologically expressed.
Herein we report a study of the role of ovarian zona pellucida (ZP)5 in the induction and maintenance of T cell tolerance to the female-specific Ag, ZP3. Within the murine ZP3 sequence is an immunogenic 13-mer peptide (pZP3) with the capacity to induce T cell proliferative responses, T cell cytokine production, and ZP-specific Ab responses (16). The pZP3 also induces autoimmune ovarian disease (AOD) in endogenous ovaries of female mice and in ovarian grafts of both male and female mice (15, 16), and the disease is transferable to naive recipients by CD4+ T cells (16). In this study we determined the ontogeny of ZP3 autoantigen protein expression and investigated the requirement of endogenous ovarian Ag for induction and maintenance of tolerance to ZP3.
Materials and Methods
Mice and surgery
Adult C57BL/6 female and A/J male mice (6–8 wk), used for breeding to generate (C57BL/6 × A/J)F1 (B6AF1), were purchased from the National Cancer Institute (Frederick, MD). For surgeries, 24-h-old females were anesthetized by hypothermia, and 1-, 2-, 3-, 4-, 5-, and 6-wk-old females were anesthetized with 2,2,2-tribromoethanol (0.6 mg/25 g body weight; Aldrich, Milwaukee, WI). Via lower back incisions, the ovaries were removed, and the incision was closed with 8.0 or 9.0 silk for prejuvenile mice and 6.0 or 4.0 silk for postjuvenile mice (Surgical Specialties Group, Reading, PA). To implant ovarian grafts, an ovary from a normal 6- to 8-wk-old B6AF1 female was surgically inserted under the renal capsule of the anesthetized recipient through a small capsular incision. The peritoneal fascia was closed with 6.0 silk, and all skin wounds were closed with 4.0 silk (Surgical Specialties Group). All surgical procedures have been recently described (1). Treatment of mice was in accordance with the National Institutes of Health guidelines established at the University of Virginia.
The peptides, pZP3 (NSSSSQFQIHGPR), chimeric peptide 2 (CP2; NCAYKTTQANKQAQHGPRQ), and moth cytochrome c-derived peptide (pMCC; ELIAYLKQATK), were synthesized by F-moc chemistry on an AMS422 multipeptide synthesizer (Gilson, Middleton, WI). The peptides were analyzed and purified by HPLC on a Poros flow-through particle column (4.6 mmD/100 mml; POROS II R/H, PerSeptive Biosystems, Cambridge, MA). The purity of all peptides exceeded 90%.
Active induction of autoimmune ovarian disease
The detailed procedure for AOD induction has been published (1). Briefly, adult (6- to 8-wk-old) female and/or male mice were immunized with 0.1 ml of peptide and adjuvant emulsion, given in one footpad and at the base of the tail. The adjuvant emulsion consisted of equal volumes of CFA containing 1 mg/ml of Mycobacterium tuberculosis (Fisher, Fairlawn, NJ) and peptide in a 50-nmol dose (unless otherwise indicated) dissolved in deionized/distilled water. Male mice and ovariectomized females received a syngeneic ovarian graft at the time of adult immunization. Histologic evidence of autoimmune oophoritis was determined 14 days after immunization. Specificity for autoimmune oophoritis was previously determined for disease in the male ovarian implants (15).
Detection of ovarian antigenic ZP3 by circulating Abs and ZP3-specific T cells
For detection of ZP3 by Ab, neonatal B6AF1 female pups were administered i.p. 5 μg of a rat IgG2a mAb to mouse ZP3 (IE-10, a gift from Jurrien Dean, National Institutes of Health) or an isotype control Ab. At the time of sacrifice the ovaries were fixed in 4% paraformaldehyde, sectioned at 5-μm thickness, and stained with FITC-conjugated IgG goat Ab to rat IgG (Southern Biotechnology, Birmingham, AL) for 30 min at room temperature. The sections were given a final rinse with PBS before the application of a coverslip over Vectashield mounting medium (Vector, Burlingame, CA). Images were examined with a fluorescent microscope (Olympus, New Hyde Park, NY).
For detection of ZP3 by T cells, peptide-specific T cell lines were derived. To obtain a T cell line, lymphocytes were obtained from the draining lymph nodes of mice that had been immunized with 50 μg of peptide in CFA 14 days previously. The cells were suspended at 106 cells/ml in complete DMEM (BioWhittaker, Walkersville, MD) supplemented with 10% heat-inactivated FCS (HyClone, Logan, UT), 1% sodium pyruvate, 1% nonessential amino acids, 1% glutamax, 100 U of penicillin, 100 μg/ml streptomycin (Life Technologies, Grand Island, MI), and 50 μM 2-ME (Sigma, St. Louis, MO). The cells were stimulated in the presence of 10 μM peptide and 3 × 106 cells/ml of irradiated (2000 rad) syngeneic spleen cells. Four days later T cell blasts were harvested on a Ficoll gradient and rested for 7–10 days. The stimulation and resting cycles were repeated until a stable cell line was obtained. The T cell lines used in this study were 1) a pZP3-specific line that produced typical Th1 T cell cytokines (IL-2 and IFN-γ), and 2) a moth cytochrome c peptide (ELIAYLKQATK; pMCC)-specific T cell line that produced typical Th1 T cell cytokines. For neonatal T cell transfers, B6AF1 pups were injected i.p. with 1 × 106 in vitro activated T cells. Ovaries as well as other organs were examined histologically on different days after cell transfer. The ovaries of the cell recipients were evaluated for the presence of T cells (CD3) and activated macrophages (F4/80 and MHC II) by immunohistochemistry as described previously (17).
Histological assessment for autoimmune oophoritis
Ovaries from immunized mice were fixed in Bouin’s fixative and embedded in paraffin. Thirty serial step sections per ovary, 5 μm thick, were stained with hematoxylin and eosin. Histopathology was evaluated for coded unknown specimens, and oophoritis severity was graded from 1 to 4 (1, focal inflammation in interstitial space; 2 and 3, increasing multifocal inflammatory foci and/or granuloma between and within ovarian follicles; and 4, loss of ovarian follicles and ovarian atrophy) (16).
Ag-specific lymphocyte proliferation assay
The popliteal, inguinal, and para-aortic lymph nodes of immunized mice were isolated and dispersed into single-cell suspensions in DMEM complete medium. Cells were cultured in flat-bottom 96-well plates (Costar, Cambridge, MA) at 3 × 105 cells/well in a total volume of 200 μl in the presence of increasing peptide concentrations. Cells were incubated for 72 h at 37°C in 5% CO2, at which time 0.5 μCi of [3H]thymidine was added. The cells were harvested 18 h later, and cell-associated radioactivity was determined on a beta spectrometer.
Lymph node cells from peptide-immunized mice were stimulated in vitro with 30 μM peptide at 3 × 106 cell/ml. On day 3, IFN-γ in supernatants was detected by indicator cells that produce NO2− upon stimulation with IFN-γ and LPS (18). Briefly, RAW 264.7 cells were allowed to adhere overnight in 96-well round-bottom plates at a concentration of 3 × 104 cells/well. Fifty microliters of sample supernatant was added to wells in triplicate together with 50 μl of LPS (80 ng/ml in complete DMEM). After 24 h at 37°C the supernatants (70 μl) were transferred to microtiter plates. Equal volumes of solutions A (0.1% naphthylenediamine dihydrochloride in distilled water) and B (1% sulfanilamide in 5% concentrated H3PO4; Sigma) were premixed, and 70 μl of the mixture was added. The absorbance at 540 nm was determined. IFN-γ levels were calculated against murine recombinant IFN-γ (PharMingen, San Diego, CA).
Ontogeny of ZP3 autoantigen expression
To study the role of physiologically expressed endogenous Ag ZP3 in tolerance induction, we first defined the ontogeny of expression of ZP3 in mouse ovary. Previous studies have demonstrated that ZP3 expression is exclusive to female ovaries. First, Northern blot analysis localized ZP3 mRNA transcripts to the ovaries and clearly demonstrated that ZP3 mRNA was not detectable in testis (19). Second, experiments using a luciferase reporter gene, whose expression was under the control of the ZP3 promoter, demonstrated that ZP3 is also not expressed in other organs, including the thymus (20). In addition to demonstrating the tissue and gender specificity of ZP3 expression, these studies demonstrated that oocyte-specific expression of murine ZP3 messenger RNA was detectable in the first week of life. To follow up on these studies, we determined the ontogeny of expression of the ovarian ZP3 autoepitopes recognized by ZP3-specific T cells and ZP3 Ab in vivo.
Neonatal female B6AF1 mice were injected i.p. with a rat mAb specific for the ZP3335–342 B cell epitope. The ovaries were obtained at different time points within the first week and assessed for bound Ab by direct immunofluorescence. Ovarian-specific ZP3 Ag was found to react in vivo with the rat mAb as early as the day of birth (Fig. 1⇓A) and rapidly increased in expression level thereafter (Fig. 1⇓B). Thus, ZP3 is present and accessible to circulating Ab from the day of birth. To determine whether neonatal ovarian ZP3 Ag is processed and presented by APC for immune recognition, pZP3-specific T cells were transferred to neonatal mice. When the T cells were transferred on day 1, ovarian-specific inflammation associated with infiltration of T cells (Table I⇓ and Fig. 1⇓E) and MHC II-positive macrophages (data not shown) was detected 2 days later. Thus, ovarian ZP3 protein is processed and presented to proinflammatory CD4+ T cells by APC in the mouse ovary within the first 2–3 days of life, and the presented ZP3 peptides include the ZP3330–342 epitope. We conclude that murine ovarian ZP3 is an organ-specific and female-specific Ag, detectable and accessible to circulating T cells and Abs within the first 3 days of life.
Comparing adult male and female responses to pZP3
The early and organ-specific expression of ZP3 in female, but not male, mice may result in physiological tolerance in females. We therefore compared the in vivo sensitivities of male and female mice to adult challenge with pZP3. Male mice and female mice ovariectomized at 6 wk of age were implanted with an ovary and were challenged with different doses of pZP3 emulsified in CFA. Compared with male mice, a 100-fold greater immunizing dose of pZP3 was required to induce comparable disease in the ovarian grafts of female mice (Fig. 2⇓, A and B). To further compare the pZP3-specific responses between males and females, mice were immunized s.c. with 50 nmol of pZP3 in CFA, the pZP3 dose that elicited comparable ovarian pathology in the two sexes. Fourteen days later, the draining lymph node cells were assayed for proliferation and IFN-γ production in response to pZP3. Despite comparable incidences and severity in AOD, male mice consistently exhibited a significantly stronger proliferative response and produced significantly higher levels of IFN-γ compared with female mice (Fig. 2⇓, C and D).
The gender difference in immune responses to pZP3 most likely represents a relative state of self tolerance in female mice for their ovarian-specific self Ag. However, it is possible that the differences were due to female vs male sex hormones. Yet, our findings are contrary to the known gender differences seen in humans and other experimental models. Typically, female subjects are generally more responsive to autoantigens and/or are more susceptible to autoimmune disease (21, 22, 23). To control for possible sex hormone effects, adult male and female mice were challenged with two foreign peptide Ags, CP2 (consists of bovine RNase94–104 and QAQIHGPR, a modified native B cell epitope pZP3335–342) and pMCC, emulsified in CFA. The equivalent T cell proliferative response and IFN-γ production induced in male and female mice when challenged with CP2 (Fig. 2⇑, E and F) or pMCC (data not shown) indicates that the dominant male response to pZP3 is unlikely to be due to a hormone effect.
Effect of neonatal endogenous ovarian Ag depletion on female pZP3 responses
If endogenous ovarian ZP3 was responsible for the reduced female responses to pZP3, one might expect enhancement of the adult female response following ablation of the endogenous ZP3. We therefore compared the pZP3 responses of adult female mice that were either untreated or ovariectomized on day 0. At 6 wk of age the mice were injected with 50 nmol of pZP3 in CFA, a dose known to induce gender differences in proliferative and IFN-γ responses (Fig. 2⇑, C and D). As shown in Fig. 3⇓A, the proliferative responses of neonatally ovariectomized females were significantly greater than the responses of the females with intact ovaries, and their IFN-γ production by T cells was also increased to the level produced in male mice (Fig. 3⇓B). In addition, the ovarian graft pathology of neonatally ovariectomized females now approximated that of male mice (Fig. 4⇓A). As controls, mice with neonatal ovariectomy and normal females were challenged with foreign peptides and assessed for in vitro recall responses to the immunizing peptides. Both groups of animals responded equally to CP2 (Fig. 3⇓, C and D) and pMCC (data not shown). Thus, neither the stress of the neonatal ovariectomy nor the loss of ovarian hormones was responsible for the increase in responses to pZP3 in the neonatally ovariectomized mice. Together, the data demonstrate that endogenous ovarian-specific ZP3, present from birth, is responsible for the relatively tolerant state of adult females to pZP3.
Continuous endogenous Ag presence is required for tolerance to ZP3
A current view of self tolerance emphasizes the importance of the murine neonatal period as a time window for tolerance induction to self Ags that are expressed early in life. We therefore examined the requirement of endogenous ovarian Ag for tolerance induction beyond the neonatal time window. Cohorts of female mice were ovariectomized weekly from birth to 5.5 wk of age. At 6 wk of age, the mice were implanted with syngeneic, age-matched ovaries and were immunized with 5 nmol of pZP3 in CFA, a dose known to induce gender differences in AOD and IFN-γ responses. As controls, normal female mice ovariectomized at 6 wk and normal 6-wk-old males were implanted with syngeneic ovaries and immediately challenged with pZP3 in CFA. Fourteen days later, the ovarian implants of all the animals were assessed for pathology, and their regional lymph node cells were assayed for IFN-γ production in response to pZP3.
AOD in the ovarian grafts of female mice that had been ovariectomized at 0, 1, 2, 3, 4, or 5 wk were similar in severity as the grafts in male mice; they were significantly more severe than the ovarian grafts in normal female mice that had not been ovariectomized until 6 wk (Fig. 4⇑A). Similarly, IFN-γ production by lymph node cells of females ovariectomized at 0, 1, 2, 3, and 4 wk in response to pZP3 was elevated and approximated the levels in normal male mice (Fig. 4⇑B). These findings indicate that the relative tolerance to pZP3 in females does not depend on the exposure of ZP3 to the confined period of the neonatal age, but, rather, it is critically dependent on the presence of the ovary at all ages. Therefore, the continual presence of ZP3 is required for the maintenance of self tolerance to pZP3. The IFN-γ responses would indicate that the lack of endogenous ovaries for at least 2 wk results in the loss of tolerance (Fig. 4⇑B). However, the findings from AOD induction, the more sensitive in vivo indicator of T cell responses, indicate that the absence of ovarian Ag for 7 days, but not 3, can result in the loss of tolerance to pZP3 (Fig. 4⇑A).
To control for the effect of ovariectomy on the subsequent pZP3 responses, females ovariectomized at 2 and 3 wk of age, the ages at which ovariectomy demonstrated the largest increase in pZP3 responses, were challenged at 6 wk with the third-party Ag, CP2. As shown in Fig. 4⇑, production of IFN-γ, in response to CP2 by the ovariectomized females, was not altered by the surgery (Fig. 4⇑A), nor were CP2-specific proliferative responses (Fig. 4⇑D). Thus, the loss of ovaries did not induce a general enhancement of T cell responses in the mice. To further investigate the possible effect of ovarian hormone depletion, the level of serum estradiol in the 6-wk-old female mice that were ovariectomized at 4 wk of age was determined and was similar to that in age-matched, untreated female mice (data not shown). The finding of comparable circulating estradiol levels in animals with significantly different immunopathologic responses to pZP3 provides additional evidence that the variance in pathologic change in ovariectomized mice is due to the loss of endogenous ZP3.
Effect of animal age versus the duration of ovarian Ag absence as the basis of altered pZP3 responses
In the previous experiment mice ovariectomized at birth or between the ages of 1–4 or 5 wk had enhanced responses to pZP3. This may have occurred because of the age at which the ovaries were removed or the time elapsed between ovariectomy and the time of pZP3 immunization. To differentiate between these alternatives, we studied the responses of mice ovariectomized at an older age. The ovaries of 6-wk-old females were removed, and 4 wk later (10 wk of age) the mice received ovarian implants and were challenged with pZP3 in CFA. The immune responses to pZP3 were studied 2 wk later, at 12 wk of age. As shown in Fig. 5⇓A, induction of AOD in the adult ovariectomized female was comparable to AOD in age-matched males and exceeded that in normal age-matched females. Similar findings were made for IFN-γ responses (Fig. 5⇓B). Thus, ovariectomy of adult mice also resulted in the reversal of tolerance if sufficient time lapsed between surgery and challenge with pZP3. Therefore, the termination of tolerance to ZP3 was dependent on the duration of endogenous ZP3 absence and not on the age of the animals.
Endogenous Ag requirement for induction and maintenance of T cell tolerance has been extensively investigated in mice with a transgenic self Ag and/or a transgenic TCR. In contrast, there are only limited studies on tolerance for physiologically expressed self Ag. Herein, we investigated the induction and maintenance of tolerance to ovarian ZP3 Ag by comparing male and female responses to a ZP3 immunogenic peptide, pZP3. The murine ovarian ZP3 is organ specific and is expressed in female, but not male, mice. Therefore, if physiological expression of ZP3 were required for tolerance induction, only female mice would be tolerant to ZP3. Indeed, tolerance to ZP3 in the female mice was demonstrated by the requirement of 100-fold more pZP3 immunogen to induce disease in their syngeneic ovarian grafts compared with male mice. In addition, at a pZP3 dose that elicited comparable disease in the two genders, T cells of the female mice exhibited a significantly lower pZP3-specific proliferative response and produced less IFN-γ than male T cells. It is of interest to note that the tolerant state to ZP3 is not an all or none phenomenon, and it can be overcome when sufficient antigenic stimulation is provided.
To further demonstrate endogenous Ag dependency of female tolerance to pZP3, neonatal ovariectomy of female mice was found to enhance their pZP3 responses to those of male mice. The most significant finding in this study is the demonstration that tolerance to pZP3 was lost even when female mice were ovariectomized as adults. Remarkably, the pZP3 responses were enhanced to the level of male responses when the endogenous ZP3 was eliminated up to 7 (but not 3) days before challenge with pZP3. This finding on self tolerance is novel and indicates that the continuous presence of endogenous, physiologically expressed ZP3 is required for the maintenance of tolerance to ZP3. In contrast, exposure to endogenous ovarian Ag limited to the neonatal period is not sufficient for the induction of self tolerance.
Sex hormone differences between male and female mice and between ovariectomized and normal females could partially account for the differences observed in this study. Male and female animals and humans are known to differ in susceptibility to autoimmune diseases, and the difference may be explicable by the disparity in estradiol and testosterone levels (24). Ovariectomy has also been shown to affect the development of organ-specific autoimmune disease (25). However, the gender differences in the pZP3-specific responses observed are probably not dependent on hormone differences for the following reasons. First, the superior male response to pZP3 is completely contrary to most of the known female-dominant autoimmune responses and autoimmune diseases (24). Secondly, the male and female responses to third-party foreign peptides were equal, and the female responses to third-party peptides were unaffected by ovarian depletion. Finally, ovariectomy of 4-wk-old females converted the pZP3 response to that of male mice, yet the circulating estradiol levels were similar to those in normal females when both were evaluated at 6 wk of age. Nevertheless, to completely investigate the hormone effect will require a future study on female mice deficient in murine ZP3 (26).
This study may represent the first demonstration of continuous endogenous Ag requirement for maintenance of self tolerance, particularly in the context of a pathogenic autoimmune response to a physiologically expressed self Ag. However, research on tolerance to foreign Ags has reported similar results. For example, the seminal research of Billingham et al. (9) on neonatal tolerance to allogeneic lymphocytes initially suggested a requirement for the neonatal period for tolerance induction. Ag administered before or immediately after birth induced tolerance, whereas administration of Ag later in life induced immunity (27, 28). However, later studies detected persistence of the injected allogeneic cells (29, 30), and depletion of donor cells abrogated tolerance (31). Thus, persistent alloantigen was required for maintenance of the experimental adult tolerance. A similar requirement for Ag persistence has been demonstrated for maintaining T cell tolerance to foreign proteins (32, 33), peptide (34), and virus (35). Additional studies reached a similar conclusion by “parking” tolerized T cells in an immunodeficient host that lacked the tolerogen and demonstrating that the T cells regained the ability to respond to the nominal Ag (36, 37, 38).
Previous studies on tolerance to physiologically expressed self Ags also support the requirement for self Ag in tolerance induction. However, unlike the present study, they did not investigate whether continuous endogenous Ag expression was required to maintain tolerance. Triplett (39) removed the hypophysis of the tree frog larvae and implanted it under the skin of 2-wk-old larvae where the hypophysis survived and differentiated. When the hypophysectomized larvae developed into a frog and received its explanted hypophysis, the hypophysis graft was rejected. Thus, in the absence of the self Ag in the larvae, the adult tree frog failed to develop tolerance to its hypophysis. Similar results were obtained for thyroid Ags in the fetal lambs (40, 41). These findings have been interpreted to indicate the requirement of self Ag early in life for self tolerance induction. However, because they did not address the requirement for endogenous Ag in adults, these studies did not clarify whether persistence of endogenous Ag is required for the maintenance of self tolerance. In addition, these studies neither measure autoimmune responses nor provide evidence for autoimmune response as the basis for the tissue inflammation and organ destruction.
The need for the continuous presence of ZP3 for the maintenance of physiological tolerance suggests that the neonatal period may not be a uniquely tolerogenic developmental time window. Classical experiments on transplantation tolerance established the idea that a prenatal or neonatal Ag encounter was critical for the establishment of peripheral tolerance. Grafting studies, in which recipients were primed with donor cells before grafting, indicated that Ag first administered prenatally or postnatally induced tolerance, whereas administration of Ag later in life induced immunity (9, 27, 28). Similar findings were later made in transgenic models, where embryonic expression of a transgene induced tolerance and postnatal expression of the same transgene induced immunity (42, 43, 44). Recent studies have demonstrated, however, that neonates are fully capable of mounting immune responses (13, 45, 46). Our current results demonstrate that Ag exposure in the neonatal period may be required but is not sufficient for tolerance induction to pZP3. Removal of endogenous ZP3 by ovariectomy at any age (up to 7 days before challenge) reversed pZP3-specific tolerance. This is not due to late ontogeny of ZP3 expression. In vivo evidence demonstrated that ovarian ZP3 was recognized by Abs by day 0 and was processed by ovarian APC and presented to pZP3-specific T cells within the first 2–3 days of life. In addition, the ovarian ZP3 Ags were not sequestered; they were accessible to both circulating Ab and T cells in the neonatal period. Therefore, although ZP3 is expressed during the neonatal period, ZP3-specific tolerance is not solely induced during this developmental window.
The capacity of Ab and T cells to recognize and respond to endogenous ZP3 in the neonatal period is consistent with our prediction that endogenous ovarian Ags are perceived by the neonatal immune system and that this might induce physiologic tolerance to ovarian ZP3 (15). Our earlier study compared the responses of male and female mice injected neonatally with pZP3 in IFA. The male mice were tolerized to pZP3, and the tolerant state was associated with a concomitant Th2-biased, nonpathogenic memory T cell response to pZP3. Female mice, on the other hand, mounted a pathogenic response following neonatal injection of pZP3 in IFA. However, the pathogenic female response was abrogated when their ovaries were removed within the first week of life. This unanticipated result suggested that endogenous ZP3 had stimulated a pathogenic neonatal response in female mice injected with pZP3 in IFA. We reasoned that in normal females, the early stimulation by ovarian ZP3 might initiate the tolerant state for ovarian ZP3. Indeed, tolerance studies based on transgenic Ags and TCR have described T cell activation as a prelude to tolerance (47, 48). Thus, our previous and current data would suggest that Ag stimulation during and beyond the neonatal period is required for the induction and maintenance of tolerance to ZP3.
This study has not, however, addressed the mechanism by which endogenous ZP3 elicits and maintains self tolerance to ZP3. We did not obtain experimental evidence for differences in frequencies of responding T cells, in counter-regulation, or in anergy induction. The frequency of T cells in male mice that responded to pZP3 by IL-2 secretion was only slightly higher than that in female mice (data not shown). In addition, comparable levels of IL-4, IL-5, and IL-13 were produced by male and female T cells in response to pZP3 in vitro (data not shown). Finally, the low proliferative response of female T cells to pZP3 in culture was not altered by the addition of exogenous IL-2 or IFN-γ (data not shown). Thus, preliminary data suggest that female hyporesponsiveness was not readily explained by pZP3 T cell deletion, Th cell deviation, or anergy induction. However, the potential roles of these mechanisms in the AOD model require further investigation.
We also have not fully investigated the role of endogenous Ags in the maintenance of regulatory T cells (49, 50), a mechanism known to prevent murine ovarian autoimmunity induced by thymectomy on day 3 of life (51). CD4+ regulatory T cells that express CD25 or high levels of CD5 from normal adult B6AF1 or A/J mice inhibited the spontaneous development of AOD induced by thymectomy on day 3 of life (52, 53, 54, 55). To determine whether endogenous ovarian Ags are required for the generation of the CD25+ regulatory T cells, T cells from normal male, normal female, and neonatally ovariectomized female mice were compared for capacity to inhibit AOD in day 3 thymectomized female mice (53, 56, 57). In two of three studies a higher number of T cells from donors without ovaries was required to suppress AOD (56, 58). The possible requirement for endogenous ovarian Ag for the induction of regulatory cells is further supported by studies showing an endogenous Ag requirement for the generation of suppressor T cells in autoimmune thyroiditis (50) and murine autoimmune prostatitis (49). Thus these studies suggest that endogenous ovarian Ags might sustain self tolerance by maintaining the regulatory T cell population in normal females. Our day 3 thymectomy study, however, showed that T cells from male, female, and neonatally ovariectomized female mice suppressed AOD in mice equally (52). This is at variance with the studies described above. Although the discrepancy is not yet explicable, it may be related to a difference in environmental Ags, because regulatory T cells might respond to both endogenous peptides and environmental cross-reactive peptides (59), which suggests that the role of ZP3-specific regulatory T cells in AOD also requires further study.
In summary, by comparing the responses of male and female mice to the female-specific ovarian ZP3, we have demonstrated that endogenous ZP3 is required for induction and maintenance of female tolerance to ZP3. Most importantly, the Ag requirement is not confined to the neonatal period; instead, persistent, physiologically expressed self Ag is required for the maintenance of self tolerance.
We thank Jason Borillo and Joyce Nash for excellent technical assistance. Histology support was provided by the Cell Science Core of the U54 Center for Reproductive Research and by the Research Histology Core of the University of Virginia (Charlottesville, VA).
↵1 This work was supported by National Institutes of Health Grants AI41236 and HD29099, and an NIGMSMARC predoctoral fellowship (to K.M.G.). Peptides were supplied by Multiple Peptide Systems under National Institutes of Health Contract NO1HD02906.
↵2 K.M.G. and S.S.A. contributed equally to the work.
↵3 Current address: Ontario Cancer Institute Princess Margaret Hospital, 610 University Avenue, Toronto, Ontario, Canada M5G 2 M9.
↵4 Address correspondence and reprint requests to Dr. Kenneth S. K. Tung, Department of Pathology, Box 214, University of Virginia, Charlottesville, VA 22908. E-mail address:
↵5 Abbreviations used in this paper: ZP, zona pellucida; AOD, autoimmune ovarian disease; CP2, chimeric peptide 2; pMCC, moth cytochrome c-derived peptide; pZP3, ZP3 peptide.
- Received November 24, 1999.
- Accepted February 1, 2000.
- Copyright © 2000 by The American Association of Immunologists