MRL/MpJ-Fas(lpr) mice show abnormalities in ovarian function and morphology with the progression of autoimmune disease.

The immune system is known to affect reproductive function, and maternal-fetal immune tolerance is essential for a successful pregnancy. To investigate the relationship between autoimmune disease and female reproductive function, we performed a comparative analysis of the ovarian phenotypes for C57BL/6 mice, autoimmune disease-prone MRL/MpJ (MRL/+) mice and congenic MRL/MpJ-Fas(lpr) (MRL/lpr) mice harboring a mutation in the Fas gene that speeds disease onset. Both MRL-background strains showed earlier vaginal opening than C57BL/6 mice. The estrous cycle became irregular by 6 and 12 months of age in MRL/lpr mice and mice of the other two strains, respectively. Histological analysis at 3 months revealed that the number of primordial follicles was smaller in MRL-background mice than in C57BL/6 mice after 3 months. In addition, MRL/lpr and MRL/+ mice displayed lower numbers of ovarian follicles and corpora lutea at 3 and 6 months, and 6 and 12 months, respectively, than that in age-matched C57BL/6 mice. MRL/lpr and MRL/+ mice developed ovarian interstitial glands after 3 and 6 months, respectively. In particular, MRL/lpr mice showed numerous infiltrating lymphocytes within the ovarian interstitia, and partially stratified ovarian surface epithelia with more developed microvilli than that observed in C57BL/6 mice at 6 months. No significant differences in serum hormone levels were observed between the strains. In conclusion, MRL/lpr mice display altered ovarian development, morphology and function consistent with the progression of severe autoimmune disease, as these findings are less severe in MRL/+ counterparts.


Introduction
In mammals, the female reproductive period is limited by the finite number of oocytes in the ovaries. In humans, the total number of oocytes gradually decreases with age due to physiological processes such as ovulation and follicular atresia. However, gonadal dysfunction, such as premature ovarian failure (POF), can cause an unexpected decrease in the number of oocytes stored in the ovaries. Patients with POF show amenorrhea, hypoestrogenism and elevated serum levels of follicle-stimulating hormone (FSH) by 40 years of age [1]. An estimated 1.5 million patients are diagnosed with POF each year, and $1% of women worldwide live with this disease [2,3]. Several reports have suggested that POF is caused by genetic factors, infectious agents or iatrogenic causes [1,2]. However, approximately one-third of POF cases are attributable to autoimmune diseases [2]. Autoimmuneassociated POF is characterized by lymphocytic oophoritis and/or by the presence of ovarian autoantibodies and is usually observed in conjunction with autoimmune diseases such as Addison's disease [1,2,4]. In most cases, it is difficult for patients with POF to reacquire fertility with ovulation of healthy oocytes [1,2].
In healthy individuals, many immune cells are present in the ovaries, suggesting that they play important roles in ovarian function in several species, such as rodents, cows, sheep, pigs and horses [5]. For example, many leukocytes, such as lymphocytes, granulocytes and macrophages, are found in the corpora lutea in rodents, cows and horses. Furthermore, the number and type of immune cells present in corpora lutea change with the estrous cycle, suggesting that these leukocytes participate in the development and regression of corpora lutea [6][7][8]. In addition to immune cells, a variety of cytokines and their receptors are expressed in the ovaries, indicating that they play important roles in the control of ovarian function [9]. For example, interleukin-6 (IL-6) is expressed in granulosa cells, where it regulates the expression of the luteinizing hormone (LH) receptor in rats [10]. Additionally, IL-1 signaling via IL-1b and its receptor are essential for ovarian folliculogenesis and ovulation in equine ovaries [11]. In pigs, tumor necrosis factor alpha (TNFa) acts as a survival factor for granulosa cells during follicular atresia [12]. Furthermore, several studies in clinical veterinary medicine also support the notion that altered immune function may affect reproductive function. Briefly, dogs suffering autoimmune-mediated thyroiditis are anestrous, suggesting a loss of reproductive function [13]. In laboratory animals, autoimmune regulator (Aire)-deficient mice, which show severe autoimmune phenotypes, are infertile, with follicular depletion and lymphocyte infiltration in their ovaries [3]. Furthermore, the pathogenesis of autoimmune disease is influenced by the production of sex steroid hormones in the gonads; in particular, estrogen aggravates systemic lupus erythematosus (SLE) in humans and rodent models [14,15]. These data suggest a close relationship between immune and reproductive function in mammals.
BXSB/MpJ-Yaa, NZBWF1 and MRL/MpJ-Fas lpr (MRL/ lpr) mice are used as autoimmune disease models, and their pathology resembles that of human SLE and rheumatoid arthritis [16,17]. MRL/lpr mice show a much more severe autoimmune phenotype than congenic MRL/MpJ (MRL/+) mice, which results from the lymphoproliferation (lpr) mutation of Fas that causes abnormal survival of autoreactive T lymphocytes [18]. In addition to the lpr mutation, the MRL background is prone to the development of autoimmune diseases, as it carries at least 13 disease susceptibility loci associated with the development of SLErelated diseases [19][20][21]. The onset of SLE is usually characterized by splenomegaly, autoantibody production and glomerulonephritis. Accordingly, spleen weights significantly correlated with the serum autoantibody levels in MRL background strain [22], supporting their use as indicators of disease severity. Interestingly, MRL mice also manifest several unique phenotypes in the female reproductive organs [23,24]. In particular, neonatal MRL/+ mice show accelerated folliculogenesis associated with the appearance of numerous mast cells in the ovary [24]. Further, adult MRL/+ mice frequently develop ovarian cysts, originating from the rete ovarii [23]. Although MRL/+ mice show several abnormalities in both immune and reproductive function, the pathological and etiological correlations remain unclear.
In this study, we used MRL-background mice to evaluate the relationship between autoimmune disease and female reproductive function. Female MRL/lpr mice showed abnormalities not only of ovarian function but also of morphology, characterized by early loss of the estrous cycle accompanied by a significant increase in spleen weight. These findings indicate that severe autoimmune disease affects female reproductive function, and provide novel insights into the treatment of infertility derived from immune system disruption

Animal and sample preparation
Female C57BL/6, MRL/+ and MRL/lpr mice were purchased from Japan SLC, Inc. (Hamamatsu, Japan). Mice were maintained according to The Guide for the Care and Use of Laboratory Animals of Hokkaido University, Graduate School of Veterinary Medicine (approved by the Association for Assessment and Accreditation of Laboratory Animal Care International). Virgin female mice at 3, 6, 9 and 12 months of age were used at the metestrus stage of the estrous cycle, as determined by monitoring vaginal smears. Body weights were measured, and then blood samples were collected by cutting vena cava under deep anesthesia (Avertin; 2,2,2-tribromoethanol dissolved in 2-methyl-2-butanol, 2.4 g/kg, administered intraperitoneally). After mice were euthanized by cervical dislocation, the ovaries and spleen were collected. Spleen weight was measured, and the spleen weight to body weight ratio was calculated as an index of autoimmune disease progression.

Examination of vaginal opening and estrous cycles
The vaginal opening status was monitored in all mice as an indicator of the onset of puberty. After vaginal opening, vaginal smears were collected for 10 consecutive days every month until 12 months of age to assess estrous cyclicity. Properties of epithelial cells, leukocytes and vaginal mucus were used to characterize each stage, as follows: ''proestrus'' was characterized by the appearance of round and nucleated epithelial cells without leukocytes; ''estrus'' was characterized by clustered and keratinized squamous epithelial cells; ''metestrus'' was characterized by a large number of leukocytes; and ''diestrus'' was characterized by low numbers of leukocytes and the presence of vaginal mucus. Metestrus was included in diestrus based on a previous report [25].

Histological analyses
Collected ovaries were fixed in 4% paraformaldehyde/0.1 M phosphate buffer (PB) overnight, embedded in paraffin and cut into sections (3 mm thick). Deparaffinized sections were stained with hematoxylin-eosin (HE) or periodic acid-Schiff (PAS) for evaluation of ovarian morphology. The number of primordial, primary, secondary and antral follicles was counted in five sections, and only follicles containing oocytes with an apparent nucleus were counted. The number of corpora lutea was also counted in five sections. Finally, the total number counted in all the five sections was used as the value for the respective ovary.
Some sections were immunostained using the following procedure: for antigen retrieval, sections were incubated in 20 mM Tris-HCl (pH 9.0) for 20 min at 105 C (CD3) or 0.05% trypsin/0.01 M phosphate-buffered saline (PBS, pH 7.4) for 5 min at 37 C (B220). The samples were then soaked in methanol containing 0.3% H 2 O 2 to block internal peroxidase activity. Sections blocked in 10% normal goat serum for 30 min at room temperature were incubated with rat anti-B220 (1:1600, Cedarlane, Ontario, Canada) or rabbit anti-CD3 (1:200, Nichirei, Tokyo, Japan) at 4 C overnight. After washing 3 times in PBS, sections were incubated with biotin-conjugated goat anti-rat IgG antibody (Caltag-Medsystems Limited, Buckingham, UK) for B220, or biotin-conjugated goat anti-rabbit IgG antibody (SABPO Kit, Nichirei, Tokyo, Japan), for 30 min at room temperature, washed again and incubated with streptavidin-biotin peroxidase complex (SABPO Kit, Nichirei) for 30 min. The sections were then incubated with 3,3 0 -diaminobenzidine tetrahydrochloride-H 2 O 2 solution. Finally, the sections were lightly counterstained with hematoxylin. The number of CD3-positive cells in each immunostained section was counted, and the total number in five sections was used as the value for the respective ovary.

Ultrastructural analysis
For scanning electron microscopy (SEM), halves of glutaraldehyde-fixed ovaries were placed in 2% tannic acid for 1 h at 4 C and post-fixed with 1% osmium tetroxide in 0.1 M PB for 1 h. Specimens were dehydrated through a series of graded alcohols, transferred into 3-methylbutyl acetate and dried using an HCP-2 critical point dryer (Hitachi, Tokyo, Japan). The dried specimens were sputter-coated using a Hitachi E-1030 ion sputter coater (Hitachi) and then examined on an S-4100 SEM (Hitachi) with an accelerating voltage of 5 kV.
For transmission electron microscopy (TEM), the ovaries of C57BL/6 and MRL/lpr mice at 6 months were fixed with 2.5% glutaraldehyde in 0.1 M PB for 4 h. Tissues were postfixed with 1% osmium tetroxide in 0.1 M PB for 2 h, dehydrated using a series of graded alcohols and embedded in epoxy resin (Quetol 812 Mixture; Nisshin EM, Tokyo, Japan). Semi-thin sections (0.5 mm thick) were stained with 1% toluidine blue and examined. Ultra-thin sections (70 nm thick) were double-stained with uranyl acetate and lead citrate and observed under a JEOL TEM (JEM-1210; JEOL, Tokyo, Japan).

Measurement of serum hormone levels
Collected blood was centrifuged at 885 Â g for 15 min. After clotted blood was removed, the samples were centrifuged at 22 136 Â g for 5 min, and the supernatant was immediately frozen at À80 C. Serum levels of endogenous mouse FSH and testosterone were measured using ELISA kits (Endocrine Technologies Inc., Newark, NJ) according to the manufacturer's instructions. All the samples were tested in duplicate.

Statistical analysis
The results are expressed as the mean ± SE and were analyzed using non-parametric statistical methods. The Kruskal-Wallis test was used to compare numerical results, and multiple comparisons were performed using Scheffe's method when significant differences were observed (p50.05).

Autoimmune disease symptoms of mice
The spleen weight to body weight ratio was used as an index of autoimmune disease in mice (Figure 1). At all the ages examined, MRL/lpr mice showed significantly higher values than the other two strains (p50.05). At 9 and 12 months, although MRL/+ mice showed slightly higher values than C57BL/6 mice, no significant differences were observed between C57BL/6 and MRL/+ mice.
Age-related changes in estrous cyclicity were examined next ( Figure 3). Figure 3(A) shows the representative pattern of mouse estrous cycle in each of the three strains at 3, 6, 9 and 12 months. C57BL/6 and MRL/+ mice showed regular cyclicity, with each cycle lasting 4-7 days until 9 months of age. These strains showed regular cyclicity until 11 months of age (data not shown). However, MRL/lpr mice lost cyclicity after 6 months, characterized by a prolonged diestrus period and a shortened estrus period. Figure 3(B) and (C) show the proportion of time spent in estrus and diestrus for each of the three strains. Estrus was significantly shorter in MRL/lpr mice than in the other strains at 6 months of age (p50.05). At 12 months, C57BL/6 and MRL/+ mice tended to show shorter and longer periods of estrus, respectively, compared to earlier months; however, no significant differences were observed among strains ( Figure 3B). In addition, diestrus was longer in MRL/lpr mice at 6 months of age than in the other strains (p50.05). At 12 months, C57BL/6 and MRL/lpr mice showed longer periods of diestrus than in earlier months, but no significant differences were observed among strains ( Figure 3C).
Age-related histological changes in mouse ovaries Figure 4 shows the ovarian histology of C57BL/6, MRL/+ and MRL/lpr mice at 3, 6 and 12 months. At 3 months, follicles at various stages of development were observed at the ovarian cortices in all three strains ( Figure 4A, D and G); the number tended to decrease with age ( Figure 4B, C, E, F, H and I). In particular, very few follicles were observed in MRL/ lpr mice at 12 months ( Figure 4I). At all the ages examined, a large number of corpora lutea were observed in C57BL/6 and MRL/ + mice ( Figure 4A-F); however, few were observed in MRL/lpr mice ( Figure 4G-I). Further, the ovarian interstitium was larger in MRL/lpr mice ( Figure 4G-I) than in the other strains.
Next, we quantified the number of follicles at each stage of follicular development (primordial, primary, secondary and antral) and the number of corpora lutea at 3, 6 and 12 months of age in all three strains. The total number of follicles decreased with age in all the strains ( Figure 4J-L), and the number of follicles observed at each stage differed among strains and at different ages of the mice. Briefly, MRL/+ and MRL/lpr mice showed a smaller number of primordial follicles than C57BL/6 mice at all the ages examined, and significant differences between strains were observed at 3 and 6 months ( Figure 4J and K) (p50.05). Further, MRL/+ mouse ovaries contained a significantly larger number of secondary follicles than the other two strains at 3 months ( Figure 4J) (p50.05). The total number of follicles was significantly lower in MRL/lpr mice than in the other two strains at 3 months, and in both MRL/+ and MRL/lpr mice than in C57BL6 mice at 6 months ( Figure 4J and K) (p50.05). In addition, the number of corpora lutea was lowest in MRL/lpr mice at all the ages examined; it was also significantly lower in MRL/+ than in C57BL/6 mice at 12 months ( Figure 4M ) (p50.05).
As shown in Figure 5, strain-related differences in ovarian morphology were observed at the ovarian interstitium and on the ovarian surface. The ovarian interstitium contains interstitial glands comprised of interstitial endocrine cells (IECs) ( Figure 5A-I). These IECs originate from internal theca or granulosa cells of atretic follicles and primarily produce testosterone in rodents [26]. IECs were clearly observed in C57BL/6 mice at 12 months ( Figure 5C) and in MRL/+ mice after 6 months ( Figure 5E and F). However, large numbers of IECs were clearly present in MRL/lpr mouse ovaries at all the ages examined ( Figure 5G-I). Finally, the ovarian surface epithelium (OSE) was a squamous monolayer in C57BL/6 and MRL/+ mice at all the ages examined ( Figure 5J and K), whereas a partially cuboidal, stratified epithelium was observed in MRL/lpr mice after 6 months ( Figure 5L).

Ultrastructure of the ovarian surface
The morphology of the ovarian surface was compared between C57BL/6 and MRL/lpr mice at 6 months by using  Figure 6A) and was mostly covered with oval-or spindle-shaped epithelial cells with microvilli ( Figure 6B). In contrast, the ovarian surface of MRL/lpr mice showed several dome-shaped bulges ( Figure 6C) covered by rounded epithelial cells that were fully covered in microvilli ( Figure 6D). The microvillus density of epithelial cells appeared to be higher in MRL/lpr mice than in C57BL/6 mice ( Figure 6B and D).
Next, the ovarian surface was observed using semi-thin and ultra-thin sections ( Figure 6E-H). In C57BL/6 mice, the OSE consisted of monolayered dark-or pale-colored cells ( Figure  6E). Under TEM, the dark-colored cells in semi-thin sections showed a highly electron-dense cytoplasm and few microvilli, whereas the pale-colored cells had a cytoplasm with low electron density and a large round nucleus similar to that found in oocytes ( Figure 6F). Likewise, two cell types were observed in the semi-thin OSE sections of MRL/lpr mice ( Figure 6G). However, the OSE cells and ovarian tunica albuginea (lowermost layer of the ovary capsule) were taller and thicker, respectively, in MRL/lpr mice than in C57BL/6 mice ( Figure 6E and G), which was also accompanied by an increase in the presence of collagen fibrils ( Figure 6H). Under TEM, the dark-colored cells in the semi-thin sections showed a highly electron-dense cytoplasm and abundant microvilli in MRL/lpr mice ( Figure 6H). The pale-colored cells in semithin sections had a cytoplasm with low electron density and a larger nucleus than the dark-colored cells ( Figure 6H). The microvillus density of cells with a highly electron-dense cytoplasm in MRL/lpr mice appeared to be higher than that in C57BL/6 mice ( Figure 6F and H), as well as in cells of MRL/lpr mice with a cytoplasm with low electron density ( Figure 6H).

Inflammatory cell infiltration in mouse ovaries
Immunohistochemical analysis was used to detect T and B cells in the ovaries of C57BL/6, MRL/+ and MRL/lpr mice (Figure 7). At 6 months, few CD3-positive T cells were observed in the ovaries of C57BL/6 and MRL/+ mice, which  tended to localize in or around the corpora lutea ( Figure 7A and B). However, numerous CD3-positive T cells were observed in the ovaries of MRL/lpr mice ( Figure 7C). These cells were widely distributed, and mainly localized to the ovarian interstitium ( Figure 7C). Quantification of CD3positive T cells in the ovaries showed that MRL/lpr mice showed significantly higher values than the other two strains at both 3 and 6 months of age ( Figure 7D) (p50.05). At 3 months, CD3-positive T cells were observed in or around the follicles of MRL/lpr mice ( Figure 7E and F). Further, few B220-positive B cells were observed in the ovaries of C57BL/ 6 and MRL/+ mice at any period examined ( Figure 7G and H), whereas some MRL/lpr mice showed numerous positive cells in the ovarian interstitium at 6 months ( Figure 7I).

Serum levels of FSH and testosterone in mice
We next compared the blood serum levels of FSH and testosterone -indicators of reproductive ability and IEC function, respectively -in C57BL/6, MRL/+ and MRL/lpr mice at 6 and 12 months (Figure 8) [26,27]. Although serum FSH levels tended to be slightly higher in MRL/lpr mice than in C57BL/6 and MRL/+ mice at 12 months, no significant differences among strains were observed at any age ( Figure 8A). Similar to FSH, serum testosterone levels were higher in MRL/lpr mice than in C57BL/6 and MRL/+ mice at 6 and 12 months, but the differences were not significant ( Figure 8B).

Discussion and conclusions
Several studies have indicated a close correlation between the immune and reproductive systems [5,9]. This study examined the effect of autoimmune disease on ovarian morphology using MRL-background mouse strains. From 3 months, MRL/ lpr mice showed autoimmune defects, indicated by significantly increased spleen weight compared to MRL/+ and C57BL/6 mice. At the same age, the total number of follicles and corpora lutea tended to be smaller in MRL/lpr mice than in mice of other strains. Further, numerous IEC and infiltrating lymphocytes were observed in MRL/lpr mouse ovaries. Importantly, lymphocytic oophoritis with loss of the follicle pool has also been reported in autoimmune-mediated POF [1,2,4]. Although the MRL background carries several susceptibility loci associated with the development of autoimmune defects [19,20], changes in ovarian morphology and development of autoimmune disease were less pronounced in MRL/+ mice than in MRL/lpr mice. Further, some MRL/+ and MRL/lpr mice developed ovarian cysts, as reported previously [23]; however, no correlation between cyst development and the parameters evaluated was found in this study (data not shown). Importantly, our data indicate that splenomegaly (3 months) precedes the onset of estrous cycle irregularity (6 months) in MRL/lpr mice. Therefore, the lpr mutation and/or its associated autoimmune defects may be crucial to the changes in ovarian morphology observed in MRL-background strains. It has been proposed that granulosa cells or oocytes may form from the mesodermal OSE [28,29]. In this study, both dark-and pale-colored cells were observed in semi-thin sections of the OSE. The pale cells were likely oocytes or their progenitors, considering their nuclear shape and previous reports [28]. The OSE showed a unique morphology in MRL/lpr mice but not in MRL/+ or C57BL/6 mice at 6 months. Ultrastructurally, the microvilli were more developed in MRL/lpr mice than in C57BL/6 mice, as shown by SEM examination of the OSE and TEM observation of darkcolored cells, and the OSE was partially stratified in MRL/lpr mice. Although no clear evidence elucidating the mechanism for altering OSE structures was found, altered morphology of the peritoneal mesothelium has been reported in ascitic conditions -such as intraperitoneal dialysis [30]. Further, patients with autoimmune diseases or ovarian cancer show altered amounts or components of ascitic fluid [31,32]. Therefore, the features of OSE in MRL/lpr mice may reflect an altered environment in or around the ovaries.
Further, only MRL/lpr mice showed a thickened ovarian tunica albuginea similar to that reported in patients with polycystic ovarian syndrome and in model mice [33,34]. Thickening of the tunica albuginea could contribute to infertility by disturbing ovulation. In this study, the ability of MRL/lpr mice to ovulate or become pregnant remains unclear; however, evaluation of the tunica albuginea may be important to elucidate the pathogenesis of ovulation disorders in autoimmune diseases, such as autoimmune-mediated POF.
MRL/lpr mice showed an irregular estrous cycle associated with the significant increase of spleen weights and morphological changes in the ovaries beginning at 6 months of age. The estrous cycle progresses through folliculogenesis,  H). Two cell types, pale-colored cells (black arrows) and black-colored cells (arrowheads) were observed in the superficial epithelium of C57BL/6 (E) and MRL/lpr (G) mice. The tunica albuginea (dual-headed arrows) were thicker in MRL/lpr mice than in C57BL/6 mice (E and G). In the ovarian surface epithelium of C57BL/6 mice (F), the round cells contained large nuclei, a few microvilli and a cytoplasm with a low electron density (arrowhead), whereas squamous cells had abundant microvilli and an electron-dense cytoplasm (arrows). In the ovarian surface epithelium of MRL/lpr mice (H), all the cells were taller than those of C57BL/6 mice (F and H), and cells with few microvilli and a cytoplasm with a low electron density bulged toward the peritoneal cavity (arrowheads). Thickened tunica albuginea in MRL/lpr mice contained numerous collagen fibers (H, asterisks). Similar to C57BL/6 mice, cells with abundant microvilli and an electron-dense cytoplasm (arrows) were observed, and its density appeared to be higher than in C57BL/6 mice (F and H). Bars ¼ 10 mm. ovulation and luteinization [35]. These processes are coordinated by oocytes, follicular epithelial cells/granulosa cells and lutein cells via the interaction of gonadotropins such as LH and FSH [35]. Furthermore, loss of the follicle pool and elevated serum FSH levels are pathological characteristics in human patients with POF showing an altered menstrual cycle [14]. In this study, no significant differences were observed in serum levels of FSH among strains, but MRL/lpr mice showed significantly decreased numbers of follicles and corpora lutea. Therefore, abnormal populations of ovarian functional units -such as follicles and corpora lutea -rather than serum gonadotropin levels, may have greater impact on the cyclicity of the estrous cycle in this murine autoimmune disease model. In this study, MRL/+ and MRL/lpr mice showed several common phenotypes, such as earlier vaginal opening and fewer primordial follicles than C57BL/6 mice, despite the significant increase of spleen weights observed in MRL/+ mice. A previous study suggested that MRL/+ mice showed enhanced progression through early folliculogenesis than C57BL/6 mice [22,36]. Therefore, the earlier onset of puberty and fewer primordial follicles compared to C57BL/6 mice are unique phenotypes in MRL-background strains.
However, the appearance of MRL-derived unique ovarian phenotypes differed between MRL/+ and MRL/lpr mice in several ways. Briefly, MRL/+ mouse ovaries contained a larger number of secondary follicles than those of MRL/lpr mice at 3 months. Further, compared to C57BL/6 mice, the total number of follicles was smaller from 3 to 6 months in MRL/lpr and MRL/+ mice, respectively. These results suggest that the unique phenotypes associated with folliculogenesis in MRL/+ mice were altered in MRL/lpr mice. Moreover, MRL/lpr mice exhibited an increased infiltration of T cells in or around follicles after 3 months of age, suggesting that these cells or their derived cytokines may affect folliculogenesis.
Testosterone-producing IECs originate from internal theca cells or granulosa cells of atretic follicles and develop in the ovarian interstitium of rodents [26]. In this study, although the number of corpora lutea was smaller, the number of IECs in the ovaries of MRL/lpr mice was increased after 3 months, and a similar tendency was observed in older MRL/+ mice. Therefore, the MRL background may contribute to the development of IECs in both MRL/+ and MRL/lpr mice. Further, the serum testosterone concentration tended to be higher in MRL/lpr mice than in mice of the other strains, possibly reflecting the development of IECs in MRL/lpr mice.
In mammalian ovaries, the development and regression of corpora lutea is regulated by cytokines produced by immune cells observed in or around corpora lutea via NFkB  signaling [5,9]. In the present study, few lymphocytes were observed in or around the corpora lutea of C57BL/6 and MRL/+ mice. In contrast, a pronounced infiltration of B and T cells into the ovarian interstitium was observed in MRL/lpr mice. Furthermore, previous reports have suggested that granulosa cells and theca cells express Fas, and that Fas-Fas ligand-mediated apoptosis is important for the regression of corpora lutea [37,38]. Based on these findings, we suggest that changes in the local immune microenvironment and/or cell death resulting from lpr mutation in the Fas gene may affect the differentiation of internal theca cells/granulosa cells into IECs.
In conclusion, we demonstrated a relationship between ovarian morphology and reproductive function in MRL/lpr autoimmune mice. These data may aid in clarifying the pathogenic mechanism of ovarian dysfunction in autoimmune disease. Briefly, intraovarian infiltration of immune cells, loss of the follicle pool and a disrupted estrous cycle are common pathological features of autoimmune disease between POF patients and MRL/lpr mice. While MRL/lpr mice did not clearly exhibit the increased levels of FSH characteristic of POF, they did display unique ovarian phenotypes including failures in follicle luteinization, increased IECs and altered ovarian surface morphology. Menstrual abnormalities, increased cycle length and decreased serum levels of hormones produced by corpora lutea were observed in patients with SLE [39,40]. Ovarian reserve markers such anti-Müllerian hormone and antral follicle count were decreased in adult patients with childhood-onset SLE [41]. In MRL/lpr mice, early onset puberty and altered follicular development were observed for the same time or earlier than the development of autoimmune defects indicated by increased spleen weight. Importantly, it was reported that early age at menarche is associated with an increased risk of SLE in human patients [42]. Taken collectively, we conclude that the Fas mutation and subsequent immune defects triggered in the MRL background likely leads to the abnormal reproductive phenotypes observed in MRL/lpr mice. This complexity causes the phenotypic similarities and differences observed between MRL/lpr mice and human clinical cases. This study provides a novel insight into the treatment of infertility resulting from disruption of the immune system.