Postnatal androgenization induces premature aging of rat ovaries

Steroids 65 (2000) 190 –205

Postnatal androgenization induces premature aging of rat ovaries Antonin Bukovskya,*, Maria E. Ayalab, Roberto Dominguezb, Jeffrey A. Keenana, Jay Wimalasenaa, Pamela P. McKenziec, Michael R. Caudlea a

Department of Obstetrics and Gynecology, University of Tennessee Medical Center, Graduate School of Medicine, Knoxville, TN, USA b Laboratory of Biology of Reproduction, Facultad de Estudios Profesionales Zaragoza, UNAM, Mexico D.F., Mexico c Department of Molecular Pharmacology, St. Jude Children’s Research Hospital, Memphis, TN, USA Received 19 August 1999; received in revised form 30 September 1999; accepted 14 October 1999

Abstract In the present paper, we report that ovaries of adult rats treated with testosterone propionate (TP) on a critical postnatal Day 5 exhibit histologic and immunohistochemical findings which resemble those of the anovulatory ovaries in middle-aged female rats. The sterile rat model has been long known whereas ovarian failure seems to be a reason for anovulation with normal hypothalamo-pituitary-gonadotropin background. Appropriate function of ovarian steroidogenic cells is also regulated by mesenchymal cells. To characterize the ovarian failure, we studied the histology, luteinizing hormone receptor (LHR) expression, and characterized changes of vascular pericytes, T cells, and dendritic cells in ovarian steroidogenic compartments consisting of interstitial cells (ISC) of ovarian interstitial glands, and granulosa and theca interna cells of ovarian follicles. Normal adult ovaries contained 63% of mature interstitial glands. The mature ISC exhibited moderate cytoplasmic and strong surface LHR expression and fine (⬍5 ␮m) cytoplasmic vacuoles (ISC of ‘luteal type’). They originated from young ISC of ‘thecal type,’ which exhibited strong cytoplasmic LHR expression. Remaining 37% were aged interstitial glands, which consisted of aged ISC (increased cytoplasmic vacuolization, nuclear pyknosis, and reduced surface LHR expression) and regressing ISC (weak cytoplasmic and no surface LHR expression). However, no mature ISC of ‘luteal type’ were detected in anovulatory ovaries of adult rats (45- and 60-day-old) injected with TP (100 or 500 ␮g) on postnatal Day 5 (TP rats). Their ovaries contained 96% of aged interstitial glands with aged and regressing ISC. Remaining 4% were abnormal interstitial glands with direct transition of young ISC of ‘thecal type’ into aged ISC (young/aged glands). Lack of mature ISC, and similar amount of aged (96%) and young/aged interstitial glands (4%) was also detected in anovulatory ovaries of untreated persistently estrous middle-aged (10-month-old) females (aging PE rats). The aging process in TP and aging PE rats was accompanied by regression of vascular pericytes, T cells, and dendritic cells within the interstitial glands. In addition, anovulatory ovaries of TP rats and aging PE females contained mature follicles exhibiting LHR overexpression by granulosa cells, and aged (cystic) follicles with reduced layers of granulosa cells lacking LHR expression. In contrast, when the rats were injected with 500 ␮g of TP later, on postnatal Day 10, the adult females exhibited estrous cycles and normal ovaries with corpora lutea. These results show that injection of TP during the critical postnatal period causes a lack of mature and preponderance of aged ISC in adult ovaries, accompanied by degeneration of mesenchymal cells. We suggest that mesenchymal cells regulate qualitative aspects of tissue-specific cells, and this function of mesenchymal cells is programmed during the critical period of development. © 2000 Elsevier Science Inc. All rights reserved. Keywords: Androgens; Ovarian steroidogenic cells; Cellular aging; Mesenchymal cells; Ontogeny; Immune adaptation

1. Introduction When testosterone propionate (TP) is injected on postnatal Day 5 to female rats, the adult females (TP rats) exhibit anovulation characterized by a lack of corpora lutea, the development of cystic follicles, and persistent vaginal estrus [1]. Pituitary responsiveness to luteinizing hormone * Corresponding author. Tel.: ⫹1-423-544-8969; fax: ⫹1-423-5446863. E-mail address: [email protected] (A. Bukovsky)

(LH) releasing hormone and hypothalamic response to electrochemical stimulation in TP rats are normal [2,3], and the basal LH, FSH, testosterone, and estradiol serum concentrations are not different from control values [4]. It has also been shown that the preoptic-suprachiasmatic area of the hypothalamus, which plays a dominant role in the cyclic regulation of LH, did not lose its sensitivity to estradiol and testosterone [5]. However, capacity of the ovaries to secrete estradiol in the form of a surge is essential for the gonadotropin surge required for induction of ovulation, and this ability of the ovaries is impaired in TP females [6]. These

0039-128X/00/$ – see front matter © 2000 Elsevier Science Inc. All rights reserved. PII: S 0 0 3 9 - 1 2 8 X ( 0 0 ) 0 0 1 0 1 - 4

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

results indicate that the anovulation may be caused by inappropriate signaling of steroidogenic cells from the ovary. Secretion of estradiol surge by granulosa cells of large antral (dominant) follicles is dependent upon timely enhanced synthesis of androgens by ovarian theca-interstitial cells during preovulatory period [7]. Lack of appropriate (timely limited) secretion of testosterone and androstenedione by ovarian interstitial cells (ISC) causes anovulation [8]. Ovaries of adult TP rats exhibit decreased steroidogenesis [9], and, in contrast to control rats, the FSH therapy does not increase estradiol concentrations [4]. These observations indicate, that ovaries of TP rats are unable to exhibit estradiol surge required for induction of ovulation by hypothalamo-pituitary system. Although the ISC are considered to be of fundamental importance in the formation of ovarian steroid hormones [10 –23], we did not find any information on the ISC in anovulatory ovaries of TP rats. Interstitial cells originate from the hypertrophied theca interna cells of degenerating follicles, and surrounding stromal elements and vasculature contribute to the development of interstitial glands [10]. The ISC possess cytological, histochemical and biochemical features of well established steroid-secreting cells [10]. Formation of androgenic steroids seems to be the principal product of ISC, but in some mammalian species progesterone and estrogens are also formed [8,10,12,14 –20,24 –31]. The cell cultures used for the studies of androgen synthesis and other functions of ISC consist of theca interna cells and interstitial cells (theca-interstitial cells) [26,32]. During ovarian follicle growth, precise regulation of the onset of androgen production by theca-interstitial cells is necessary for maintaining follicle viability [30]. Androgens stimulate follicular atresia by enhancing apoptosis of granulosa cells [33]. Temporary suppression of androgen production by theca-interstitial cells has been shown in vitro to be induced by hepatocyte growth factor [30], the fibroblast/ pericyte-derived cytokine [34,35]. The LH-stimulated androgen production (androstenedione) by theca-interstitial cells is also inhibited by tumor necrosis factor-␣ [36], a monocyte/macrophage-derived cytokine [37], and apoptosis of theca-interstitial cells is induced by transforming growth factor-␣ and transforming growth factor-␤ [38]. Transforming growth factors are produced by monocyte-derived cells (MDC) and T lymphocytes [39,40]. These observations indicate that specialized mesenchymal cells, such as pericytes, MDC and T lymphocytes, may significantly influence the function and regression of the ISC. In the ovaries of regularly cycling individuals the proportion of ISC exhibits degenerative changes [10]. After persisting for some time, the ISC gradually lose their cytoplasm and lipids. This is accompanied by a loss of blood vascularity, and some ISC show degenerative cytoplasmic changes and become ultimately refractory to gonadotropic stimulation [10].

191

Cellular degeneration in various tissues is morphologically characterized by increased cytoplasmic vacuolization and nuclear pyknosis resulting in the apoptotic cell death [41– 44]. The increased vacuolization precedes and accompanies apoptotic degeneration of the cells [45,46]. This study was intended to analyze a character of interstitial glands in the anovulatory ovaries of TP rats, and compare the data with ovaries of aging persistently estrous rats (aging PE rats). We studied the morphology and LH receptor (LHR) expression by ISC, and character of vascular pericytes, T lymphocytes, and monocyte-derived dendritic cells in interstitial glands. The results indicate that postnatal androgenization causes premature aging of ovaries, and this is accompanied by changes in behavior of mesenchymal cells in interstitial glands.

2. Experimental 2.1. Animals Rats of the CII-ZV strain, born and maintained in the Laboratory of Biology of Reproduction, Zaragoza, Mexico, were kept on an 14L-10D cycle (lights on 5:00 a.m.–7:00 p.m. h). Newborn rats were weaned at 21 days of age. Litters were randomly assigned to one of following groups: (1) females injected subcutaneously with 100 ␮g of TP in 0.05 ml of sesame oil (Sigma Chemical Co., St. Louis, MO, USA) on postnatal Day 5 (n ⫽ 12), (2) females injected with 500 ␮g of TP in 0.05 ml of oil on postnatal Day 5 (n ⫽ 16), (3) females injected with 500 ␮g of TP on postnatal Day 10 (n ⫽ 12), and (4) control females injected with 0.05 ml of oil on postnatal Day 5 (n ⫽ 13). For immunohistochemical studies both ovaries and segments of the uterus were removed at the age of 45 (six animals in each group) and 60 days (remaining animals). The 60-day-old rats were monitored by vaginal smears 10 to 12 days before sacrifice, and seven twomonth-old control females were sacrificed in proestrus. (5) Anovulatory ovaries of untreated middle-aged (280 – 300 days) PE rats were also investigated (n ⫽ 6). We included middle-aged rats in this study to investigate whether the anovulation with persistent estrus, characteristic for postnatally-androgenized and normal middle-aged rats, is associated with similar characteristics of steroidogenic structures in their ovaries. All animals were sacrificed in the early afternoon (2:00 p.m. h). The study was approved by the Institutional Animal Care and Use Committee. 2.2. Peroxidase immunohistochemistry Tissue samples from each animal (both ovaries and two 3-mm long segments of the uterus) were placed on the bottom of a single 10 ⫻ 10 ⫻ 5 mm plastic cryomold (Tissue-Tek, Miles, Naperville, IL, USA), embedded in O.C.T. compound (Miles), frozen in liquid nitrogen, and

192

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

stored at ⫺80°C. Frozen tissues were sliced into 22 sections (7 ␮m) through the central parts of the ovaries by using a cryostat microtome with specimen retraction during return travel (Carl Zeiss Microm HM 505 E; MICROM Laborgera¨te GmbH, Waldorf, Germany). The slides were dried overnight at room temperature, and stored at 4°C. In most instances, the immunohistochemical staining of cryostat sections was performed within one week. Two to four sections of both ovaries and uterine samples from each animal were stained for LHR (see bottom), one section was used for Papanicolaou staining. We also stained parallel sections for Thy-1 glycoprotein of rat pericytes (dilution 1:100; clone MRC OX-7, Serotec Ltd, Oxford, England) to analyze vascular pericytes and distinguish interstitial glands from regressing corpora lutea (CL). The regressing CL were of similar size, but exhibited significantly higher degree of vascularization when compared to interstitial glands. Additional sections were stained with MRC OX-8 antibody (Serotec) recognizing CD8 subset of rat T lymphocytes (dilution 1:100), MRC OX-62 (Serotec) to rat dendritic cells (1:10), and other markers not reported here. Unused sections were stored at ⫺20°C for additional staining, e.g. oil red O (see Section 3). For peroxidase immunohistochemistry, the labeled streptavidin-biotin method was employed. All steps were performed at room temperature. Universal DAKO LSAB™2 Peroxidase Kit K609 for use on rat specimens (DAKO Corporation, Carpinteria, CA, USA), reacting with mouse immunoglobulins, was used according to the instructions provided by vendor with some modifications. Briefly, cryostat sections were fixed five min with acetone, dried for 30 min, and rinsed in phosphate buffered saline (PBS). PBS was freshly prepared before staining from frozen 10⫻ stock solution, and pH adjusted to 7.22. Primary antibody, negative control reagent (DAKO), or PBS were applied for 15 min. Slides were washed twice in PBS, and link solution (biotinated anti-mouse immunoglobulin antibody) was applied for 15 min. After three washes in PBS, slides were incubated with streptavidin-peroxidase conjugate for 15 min and washed 3⫻ in PBS. Slides were incubated five min in substrate-chromogen solution, rinsed with PBS, washed in distilled water, counterstained with Harris hematoxylin diluted 1:20, dehydrated, and mounted. 2.3. Primary anti-LH receptor antibody The mouse anti-rat LHR monoclonal antibody 3B5 [47] was used in the form of tissue culture supernatant diluted 1:2 with PBS. This antibody has been developed by using affinity column purified rat LHR [48], and it has been shown to inhibit 125I-hCG and 125I-hLH binding to rat luteal cell membrane fractions [47]. It also exhibits strong cytoplasmic staining of the rat steroidogenic cells in theca interna, strong surface staining of mature ISC, moderate staining of granulosa cells in mature antral follicles, and strong surface staining of granulosa-lutein cells in the corpora lutea [47].

The 3B5 antibody has also been successfully used by other investigators to localize LH/hCG receptors in the rat and human ovaries [49,50]. 2.4. Video imaging Evaluation was performed on a Leitz DM RB (Leica Inc., Wetzlar, Germany) microscope equipped with differential interference contrast and a DEI-470 CCD Video Camera System (Optronics Engineering, Goleta, CA, USA) with detail enhancement. Video images were captured via a DT2255 MACII Frame Grabber (Data Translation, Marlboro, MA, USA) into the Apple Power Macintosh 8100/80 16/1000CD computer (Apple, Inc., Cupertino, CA, USA). The captured black and white images (256 shades of gray) were processed with NIH Image 1.58 public software (Wayne Rasband, NIH, Bethesda, MD, USA), and selected areas copied into the Microsoft™ PowerPoint™ 97 SR-2 (Microsoft Corporation, Redmont, WA, USA) software to allow multiple pictures per page. Each picture was assigned with descriptive letters and symbols, and the page printed by HP DeskJet 1120C Printer [Hewlett–Packard Singapore (Pte) Ltd., Singapore]. For prints and quantitative measurements (see below) magnifications 100⫻, 200⫻, 400⫻ or 1000⫻, and camera factor 1.5⫻ were used. 2.5. Quantitative immunohistochemistry Quantitative evaluation of LHR expression was performed in both ovaries of two-month-old control rats in proestrus, 2-month-old TP females (TP 500 ␮g, postnatal Day 5), and aging PE rats using NIH Image 1.58 software (Wayne Rasband). The illumination of empty field on slide (image background) was first adjusted to 40 in 0 to 255 scale, and blank field correction performed as described [47]. The density of peroxidase staining was evaluated in the 10, 000 square ␮m areas using the circle or rectangular tools for area selection. Each measurement represented a cumulative mean value obtained from 10 000 individual values (pixels). The cumulative value (crude measurement) was subtracted with tissue background (control staining) to obtain a net value of specific staining. Each column in Fig. 4A–C represents the mean ⫾ standard error of the 60 net values (measurements) obtained from 6 to 10 rats. 2.6. Quantitative evaluation of aged interstitial glands We used the Papanicolaou staining (polychrome EA60; Sigma) for evaluation of aged interstitial glands in control and anovulatory ovaries. In blank field corrected images (see above) the density slice was set to 0 to 40 (0 –255 scale) and the highlighted empty areas (cytoplasmic vacuoles) were calculated in the 10 000 sq ␮m rectangular or circled regions. The myometrium in uterine

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

193

Fig. 1. High power magnification of LHR expression in interstitial cells of control ovaries (A–D), theca interna cells (E), and mature CL (F). (A), young ISC of ‘thecal type’ show strong cytoplasmic staining (arrow) and LHR accumulation in the nuclear envelope (arrowhead)—see also (E). (B), mature ISC of ‘luteal type’ show strong surface staining (long arrowheads) and diminution of cytoplasmic LHR expression (arrows)—see also (F). Some cells show perinuclear LHR (short arrowhead). (C), aged ISC exhibit residual surface (arrowhead) and moderate cytoplasmic staining (arrow). (D), regressing ISC lack surface expression, and exhibit strong staining of nuclear envelope segments (arrowheads) and residual cytoplasmic staining (arrows). (E), theca interna cells show LHR expression similar to that of young ISC (for symbols, see A). (F) Mature granulosa lutein cells show LHR expression similar to that of mature ISC (for symbols see B). All immunohistochemical samples are weakly counterstained with hematoxylin.

samples was used to establish tissue background. The background value was 6.58 ⫾ 0.84/10 000 sq ␮m region. The interstitial glands with ISC lacking nuclear pyknosis and showing cytoplasmic vacuoles not exceeding 5 ␮m in diameter were considered as normal mature interstitial glands. The evaluation of LHR expression and Papanicolaou staining was also used to determine a percentage of mature and aged interstitial glands in the ovaries. The percentages were calculated from the total number of interstitial glands in tissue sections from 14 ovaries of cycling control rats in proestrus (371 glands evaluated), 16 ovaries of TP females (484 glands evaluated), and 12 anovulatory ovaries of aging PE rats (292 glands evaluated). The interstitial glands with majority of ISC exhibiting reduced LHR expression, enhanced cytoplasmic vacuolization and nuclear pyknosis, were considered as aged interstitial glands. In TP and aging PE ovaries no normal mature interstitial glands were detected, but these ovaries showed occasional abnormal glands exhibiting direct transformation of young into aged ISC (young/ aged interstitial glands).

2.7. Statistic Mean and SE were calculated using Microsoft™ Excel 97 SR-2 (Microsoft Corporation) software. ANOVA (Single Factor; Microsoft™ Excel 97 SR-2, Microsoft Corporation) was used to evaluate the significance of differences between columns (sets of values) obtained from both ovaries of rats in control, TP and aging PE rats. Values of P ⬍ 0.05 were considered significantly different.

3. Results 3.1. Control females All ovaries of control females contained corpora lutea and follicles in various stages of development. No nonspecific staining was evident with control immunohistochemical procedures (not shown). The LHR expression in the ovaries of cycling females was similar to that described previously in normal adult rats [47]. Based on the morphology and LHR expression, four

194

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

Fig. 2. LHR expression (A, C–G, and I) and Thy-1 pericytes (B, H and J), in ovaries of control (A–D), TP (E, G and H), and aging PE rats (F, I, and J). Dotted lines indicate borders of some interstitial glands in adjacent sections stained for Thy-1. (A), mature ISC and luteal cells show similar surface staining, which is depleted in aged ISC. Note strong cytoplasmic staining in theca interna and moderate in granulosa cells (g⫹). Dotted arrowhead indicates normal follicular vessels. (B), activated pericytes (arrowheads) accompany mature structures and periphery of regressing CL. Arrows indicate inactive pericytes. (C), developing interstitial gland shows transition of young into mature ISC (arched arrow). (D), in regressing gland, aged and regressing ISC are replaced by a new and undifferentiated population of stromal cells (s) lacking LHR. (E), young ISC are directly transformed into the aged ISC (arched arrow); no mature

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

types of ISC in control ovaries were distinguished: 1) Young ISC of small size with strong cytoplasmic LHR expression (Fig. 1A) similar to that of steroidogenic cells in theca interna (Fig. 1E). 2) Mature ISC of moderate size (smaller than luteal cells) with moderate cytoplasmic and strong surface LHR expression (Fig. 1B) similar to that of granulosa lutein cells in mature CL (Fig. 1F). 3) Aged ISC with partial depletion of LHR expression (Fig. 1C), large cytoplasmic vacuoles, and nuclear pyknosis (Fig. 3F). 4) Regressing ISC with residual cytoplasmic and perinuclear, and no surface LHR expression (Fig. 1D). Mature interstitial glands contained mature ISC only (m, Figs. 2A; 3A and B). In developing interstitial glands, the mature ISC originated from young precursors (Fig. 2C). In aged interstitial glands the mature ISC were converted into the aging and regressing ISC (Fig. 3F and G), and the space of regressing ISC was taken by a new and undifferentiated population of stromal cells (Fig. 2D). Fig. 2B shows that mature interstitial glands and follicles exhibited activated (wide) Thy-1 pericytes, and such pericytes were also associated with the periphery of regressing corpora lutea (arrowheads). However, regressing ISC and inactive ovarian stroma showed inactive (narrow) Thy-1 pericytes (arrows). 3.2. Ovaries of rats androgenized on postnatal Day 5 Ovaries of TP rats contained three types of ISC only: young, aged, and regressing; no mature ISC were detected. In developing interstitial glands the young ISC were directly transformed into the aged ISC (Fig. 2E). Majority of interstitial glands contained ISC of aged and regressing type (Fig. 2G). No mature interstitial glands were found, but occasional interstitial glands associated with large follicles contained young and aged ISC with strong cytoplasmic LHR expression (young/aged interstitial glands, Fig. 2G). Such abnormal interstitial glands were not detected in control ovaries. The changes in TP ovaries were not dependent on the dose of TP injected (100 or 500 ␮g) or the age of rats at which the ovarian samples were collected (45– 60 days). Staining for oil red O (mayonnaise smear used as a positive control) did not show presence of lipids in the cytoplasmic vacuoles (unpublished data). The average number of interstitial glands per central section in TP ovaries (30.3 ⫾ 2.1 standard error) did not show a significant difference when compared to controls (26.5 ⫾ 2.1; P ⫽ 0.22).

195

Fig. 2H shows that young/aged interstitial glands were associated with activated Thy-1 pericytes (arrowhead), areas of aged ISC (a) were associated with inactive pericytes (arrow), and areas of regressing (r) ISC showed regressing or no Thy-1 pericytes. Theca interna of mature follicles showed normal morphology and LHR expression, but granulosa cells overexpressed LHR. In addition, abnormal large antral follicles overexpressing LHR (g⫹⫹, Fig. 2G; compare with g⫹, Fig. 2A) and possessing thecal ‘pegs’ were found. These pegs (arrowheads, Fig. 2G) originated from the vascular layer under follicular basement membrane, and contained activated pericytes (arrowheads, Fig. 2H). To distinguish this type of large antral follicles from normal mature follicles we use the term ‘transitory’ follicles (tf, Fig. 2G)—see Section 4. The TP ovaries also contained cystic follicles, which were characterized by large antrum lined with reduced layer of granulosa cells lacking LHR expression. 3.3. Ovaries of rats androgenized on postnatal Day 10 Adult females injected with 500 ␮g TP on postnatal Day 10 exhibited cyclic changes in vaginal smears, their ovaries contained corpora lutea, and showed no abnormalities of LHR expression in ISC and granulosa cells (not shown). 3.4. Anovulatory ovaries of aging PE females The interstitial glands exhibited patterns similar to that found in TP rats. Developing interstitial glands often consisted of aging ISC only (a, Fig. 2F). Majority of interstitial glands exhibited aged and regressing ISC (Fig. 2I). The ovarian stroma contained many isolated cells exhibiting LHR expression (arrowhead, Fig. 2F). The average number of interstitial glands per central section of ovary was lower (24.3 ⫾ 4.0 versus 30.3 ⫾ 2.1 in TP rats), but the difference was not statistically significant (P ⫽ 0.18). Similarly to TP rats, the ovaries of aging PE females also contained mature and ‘transitory’ follicles with LHR overexpression, and cystic follicles (Fig. 2I). The Thy-1 pericytes in cystic follicles were inactive (Fig. 2J). 3.4. T lymphocytes and dendritic cells Fig. 3 demonstrates the size of cytoplasmic vacuoles (A, F, K, and P) in Papanicolaou staining (PAP), character of

ISC are present. (F), developing interstitial gland (top) exhibits aged ISC. Note normal preantral follicle at the bottom. Arrowhead indicates isolated ISC in the stroma. (G), interstitial glands contain ISC of aged and regressing type, but abnormal young/aged interstitial glands are occasionally present. Strong staining of granulosa cells (g⫹⫹). Arrowheads indicate vascular thecal pegs. (H), active pericytes (arrowheads) in young/aged interstitial gland and follicle. Arrow indicates inactive pericytes. (I), interstitial gland containing aged and regressing ISC. The ‘transitory’ follicle shows strong staining of granulosa cells. Arrowhead indicates thecal peg. Note a lack of staining of granulosa cells in cystic follicle. (J), aged region of interstitial gland shows regression of Thy-1 pericytes and regressing region shows no Thy-1 pericytes. Note inactive Thy-1 pericytes in cystic follicle. mf, mature follicle; g, granulosa cells; t, theca interna; e, theca externa; m, mature ISC; a, aged ISC; r, regressing ISC; mcl, mature CL; rcl, regressing CL; s, ovarian stroma; v, stromal vessels; y, young ISC; z, zona pellucida; o, normal oocyte; y/a, abnormal young/aged interstitial gland; tf, ‘transitory’ follicle; cf, cystic follicle.

196

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

Fig. 3. Papanicolaou staining (PAP), and localization of LHR, Thy-1 pericytes (PC), CD8 T cells (TC), and MRC OX-62 dendritic cells (DC) in semi-parallel sections through mature and aged interstitial glands of controls (C:) and young/aged and aged interstitial glands of TP rats (TP). PAP column: arrows indicate normal and arrowheads pyknotic nuclei. Note fine vacuoles in (A) and presence of normal and pyknotic nuclei in (K). LHR column: Arrow in (B) indicates strong surface staining of ‘luteal type’ in mature ISC, asterisks in (L) indicate strong cytoplasmic staining of ‘thecal type’ in young ISC. Arrowheads indicate aged cells. Note transitions of mature (m) into aged (a) and regressing (r) ISC in (G), and lack of LHR depletion in ‘early’ aged cells with large vacuoles in (L). PC column: plus signs indicate activated PC, arrows indicate inactive (narrow) PC. Arrowhead in (H) shows regressing PC. TC column: Arrows indicate normal and arrowheads regressing TC. DC column: Arrows indicate intact and arrowheads regressing DC, or their remnants (T). In (J) the DC regress among aged ISC— compare with (G). Note intact DC in adjacent stroma (s) in (O) and their regression (arrowheads) after entering (arched arrow) the young/aged interstitial gland. For additional details see text.

LHR expression (B, G, L, and Q), changes of Thy-1 pericytes (C, H, M, and R), and association of T lymphocytes (D, I, N, and S) and dendritic cells (E, J, O, and T) in semi-parallel sections of interstitial glands in control and TP ovaries. In controls, the mature interstitial glands exhibited fine cytoplasmic vacuoles and strong surface LHR expression in ISC, and occurrence of activated pericytes, many normal T

lymphocytes, and mature dendritic cells (DC). Aged interstitial glands showed ISC with large vacuoles and pyknotic nuclei, diminution of LHR expression, inactive and regressing pericytes, and degenerating T cells and DC. The degeneration of T lymphocytes and DC characteristically occurred in the inner aspects of aged interstitial glands (arrowheads, Fig. 3I and J), in connection with regressing ISC (r, Fig. 3G).

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

197

Fig. 4. Quantitative evaluation of LHR expression in interstitial cells (A), stroma (B), and granulosa cells of mature follicles (C) in the ovaries of control rats (C), TP females (TP), and aging PE rats (aPE). Net OD ⫽ net optical density. For further details see text. a, P ⬍ 0.001 versus controls; b, P ⬍ 0.001 versus control and TP rats.

The abnormal young/aged interstitial glands in ovaries of TP rats contained young ISC with fine vacuoles lacking nuclear pyknosis (top and right, Fig. 3K) and exhibiting strong cytoplasmic LHR expression (asterisks, Fig. 3L). These cells were directly transformed into ‘early’ aged ISC, showing large vacuoles and nuclear pyknosis, but no LHR depletion (arrowheads, Fig. 3K and L). Such interstitial glands exhibited inactive pericytes (Fig. 3M versus C) and many normal T cells. However, DC exhib-

ited degenerative changes (arrowheads, Fig. 3O). Aged interstitial glands in TP ovaries showed ISC with large vacuoles and pyknotic nuclei (Fig. 3P), diminution of LHR expression (Fig. 3Q), inactive pericytes (Fig. 3R), and degenerating T cells (Fig. 3S). Dendritic cells were virtually absent, except for the small fragments exhibiting weak staining (arrowhead, Fig. 3T). Similar features were detected in interstitial glands of aging PE rats (not shown).

198

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

Table 1 Percentage of mature and aged interstitial glands (ISG) and characteristics of large antral follicles in the ovaries of 2-month-old cycling control rats in proestrus (C), and anovulatory ovaries of TP rats (TP, 500 ␮g on postnatal Day 5), and middle-aged PE females (aPE) Rats

na

Mature ISG (%)b

Aged ISG (%)

LHR-E in GC (NMF)

NMF (%)

‘Transitory’ follicles (%)

Cystic follicles (%)

C TP aPE

14 20 12

63.29 ⫾ 4.29 4.25 ⫾ 0.89c 4.33 ⫾ 1.48c

36.71 ⫾ 4.29 95.75 ⫾ 0.89c 95.67 ⫾ 1.48c

61.86 ⫾ 1.94 102.49 ⫾ 1.09c 101.22 ⫾ 2.21c

100 55 33

0 80 67

0 85 100

Data represent the mean percent ⫾ standard error of mature and aged ISG in the ovaries, the mean net optical density ⫾ standard error of LHR expression (LHR-E) in granulosa cells (GC) of morphologically normal mature follicles (NMF), and percentage (%) of ovaries exhibiting particular type of large antral follicles. For details see Section 2 and text. a Total number of ovaries investigated. b Young/aged ISG in the ovaries of TP and aPE rats (see y/a in Fig. 1G and Section 3) c P ⬍ 0.001 versus controls.

3.5. Quantitative evaluation of LH receptor expression Quantitative data on LHR expression by ISC, ovarian stroma, and granulosa cells of mature antral follicles in the ovaries of control, TP, and aging PE rats are given in Fig. 4 and Table 1. The ISC (Fig. 4A) in the ovaries of TP and aging PE rats showed significantly lower LHR expression when compared to the ISC in control ovaries, apparently due to the lack of normal mature glands in TP and aging PE ovaries (Table 1). Note that LHR expression in ISC of aging PE females is also significantly lower when compared to that of TP rats. In ovarian stroma (Fig. 4B; note different scale versus Fig. 4A), the LHR expression was significantly higher in TP rats, and even more higher in aging PE rats.

When mature antral follicles with normal morphology were compared (Fig. 4C; the ‘transitory’ and cystic follicles excluded), the LHR expression by granulosa cells in TP and aging PE rats was significantly higher than that of control females. 3.6. Increased vacuolization of aged interstitial cells, and a percentage of aged interstitial glands Fig. 5A shows that mature interstitial glands in control ovaries (C/m) showed cytoplasmic vacuolization similar to that of granulosa lutein cells in the mature CL. Vacuolization of ISC in aged interstitial glands in control ovaries (C/a) and ovaries of TP rats was significantly higher when compared to normal mature interstitial glands. Ovaries of aging PE rats showed even more extensive vacuolization in aged interstitial glands. Fig. 5B and Table 1 show that aged interstitial glands represented 37% of all interstitial glands in control ovaries. However, 96% of interstitial glands in the ovaries of TP and aging PE females were aged interstitial glands. Table 1 also shows that no ‘transitory,’ and cystic follicles were found in control ovaries. Although large antral follicles with normal morphology were detected in the ovaries of 55% of TP and 33% of aged PE rats, the majority of anovulatory ovaries also exhibited ‘transitory’ and cystic follicles (Table 1).

4. Discussion 4.1. Similarities between anovulatory ovaries of androgenized and aging PE females

Fig. 5. Extent of vacuolization (area in sq ␮m/10 000 sq ␮m of tissue) in corpora lutea (CL), mature (C/m) and aged interstitial glands of control rats (C/a), and interstitial glands of TP females (TP), and aging PE rats (aPE). a, P ⬍ 0.001 versus previous tissue (column). B) Percent of aged from total number of glands evaluated in the ovaries of control rats (C), TP females (TP), and aging PE rats (aPE). For details see Section 2. a, P ⬍ 0.001 versus controls.

The key findings of this study are that the ovaries of TP and aging PE rats lacked normal mature ISC. The vast majority (96%) of interstitial glands were aged interstitial glands with diminution of LHR expression, contrasting 37% of aged interstitial glands in controls. While in control ovaries young ISC were transformed into the mature ISC, the ovaries of TP and aging PE rats showed direct transition of young into aged ISC. The aged and regressing ISC exhibited diminution of LHR

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

expression. It is now well documented that the steroidogenic function of ISC is LH-dependent [7,20,51]. Hence, low levels of LHR expression by ISC in the ovaries of TP and aging PE rats may indicate inability of these cells to secrete timely a sufficient amount of androgens, precursors of estrogens required for estradiol surge during preovulatory period [7]. In addition, depletion of LHR was associated with increased cytoplasmic vacuolization and nuclear pyknosis of ISC, suggesting that these cells are in the process of apoptotic cell death. These observations indicate that such ISC exhibit a morphofunctional senescence. Another abnormality found in anovulatory ovaries of TP rats and aging PE females was that mature follicles were not preserved in the preovulatory stage with moderate expression of LHR by granulosa cells characteristic for control ovaries. Instead, they showed significantly higher LHR expression by granulosa cells, and progression into the so called ‘transitory’ follicles and aged cystic follicles. Such abnormal follicles may result from the lack of ovulation. However, anovulatory ovaries of prepubertal and pubertal rats already contain normal mature follicles, and excessive follicles are eliminated by atresia [52,53]. Elimination of excessive follicles through atresia is ensured by the ovarian macrophages and T cells [47,53–56]. This suggests that the occurrence of ‘transitory’ and aged cystic follicles in the ovaries of TP and aged PE rats may result from a lack of recognition of excessive follicles by immune system effectors. Taken together, the TP and aging PE rats are for some reason unable to prevent accelerated aging of interstitial glands and transformation of large antral into cystic follicles. 4.2. Regression of interstitial glands A proportion of ISC in aged interstitial glands of control and anovulatory ovaries underwent regression. This was associated with regression of vascular pericytes. Activated pericytes accompany differentiating and mature ovarian and placental structures, and regression of pericytes is associated with aging and degeneration of tissue cells [43,55,57– 60]. Hence, the regression of vascular pericytes may also induce regression of interstitial glands. Why do pericytes in certain glands regress? The Thy-1 pericytes are accompanied by autonomic innervation, which was proposed to control their activity [57]. The autonomic innervation regulates the proportionality (quantity) of tissues within the body (reviewed in Ref. [59]). In the ovary, autonomic innervation is associated with microvasculature of interstitial glands [61– 63]. We speculate that when new interstitial glands are formed, the autonomic innervation ensures the regression of older interstitial glands, by causing regression of their pericytes. In that way, the relatively constant amount of interstitial tissue within normal and anovulatory ovaries can be maintained. Regression of T cells is associated with LHR depletion in interstitial glands of normal, TP, and aging PE ovaries. The regression of LHR expression in ISC may require

199

certain cytokines released from T cells undergoing suicide. In the stratified epithelium of ectocervix, the intraepithelial T cells exhibit apoptotic fragmentation after attaining intermediate layer of epithelial cells, and similar fragmentation of DC accompanies advanced aging of epithelial cells (unpublished data). Such regression of DC accompanies direct transition of young into aged ISC not yet exhibiting diminution of LHR expression (Fig. 3O). On the other hand, only normal mature glands exhibited intact mature DC. Hence, premature apoptotic fragmentation of DC in interstitial glands of TP ovaries may cause a lack of mature interstitial glands and premature aging of ISC. 4.3. Possible mechanisms causing premature aging of ovaries Injection of TP on postnatal Day 5 apparently induces the premature aging of adult ovaries. One possibility is that the TP induces permanent deterioration of tissue-specific cells in developing ovaries. However, the ovaries of adult TP rats resume ovulatory function when transplanted to normal rats, and ovaries of normal females do not ovulate when transplanted to TP rats [64]. Another possibility is that the postnatal injection of TP causes permanent deterioration (masculinization) to the developing hypothalamus [64]. However, it has been shown that the hypothalamus is not permanently affected, but the ovaries are unable to produce preovulatory surge of estradiol required for the hypothalamic stimulation of the LH surge [3,5,6]. Third possibility is that the TP causes some alteration in the patterns of pulsatile LH secretion [9], a feature also reported to occur in aging females [65]. However, the concept seems to be vague (both decrease and increase in LH pulse frequency are considered to contribute to the alteration of ovarian function—see response of Wise et al. to comments in Ref. [66]), and it has been admitted that altered hormonal signalization from the gonads of TP females may cause subsequent alteration of pulsatile LH secretion [9]. Fourth possibility is that the alteration of postnatal ovarian development by TP is encoded within the developing autonomic innervation, which participates in the regulation of differentiation of steroidogenic ovarian structures in adult ovaries [67,68]. However, alteration to the autonomic innervation is associated with quantitative (changes of compensatory hypertrophy and number of ovulations) [69 –71] rather than qualitative changes to the ovary (anovulation). Another possibility is that TP-induced alteration of ovarian development during immune adaptation is encoded within the developing immune system, which regulates differentiation, aging, and regression of ovarian structures in adult ovaries [72–79]. The immune system has been proposed to be a component of the ‘tissue control system’ (reviewed in Ref. [59]) cooperating with the endocrine and neural systems in the maintenance of tissue homeostasis, including the ovary [55].

200

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

Fig. 6. Simplified model on the role of the stop-effect (SE) of the TCS in the physiology and pathology of adult tissues. (A), no SE is exerted toward tissues that are either absent (corpus luteum) or differentiate into apoptotic cells (cornified layers in epidermis) during the immune adaptation. Specific cells in such ‘SE-independent’ tissues are stimulated to gradually differentiate and degenerate (thick arrow) through the four basic stages (1, 2, 3, and 4) and interfaces I, II, and III (dotted lines). (B), in ‘SE-dependent’ tissues, the differentiating tissue-specific cells are parked by the SE in the mature state. This is dependent on the presence of mature tissue cells during immune adaptation. (C), failure of tissue-specific cells to differentiate into the mature state during immune adaptation causes a ‘lower’ setting of the SE than required (shift to the left, versus B). Because of that, the cells in such adult tissues are parked by the SE in the immature state. (D), a ‘higher’ setting of the SE (shift to the right, versus B) during immune adaptation causes that cells in adult tissues are parked by the SE as aged cells. For the ovary, which differentiates later during immune adaptation, the regular SE is exerted for a shorter period during the lifetime when compared to the early differentiating tissue (liver). Consequently, the ovarian function physiologically expires sooner than function of other tissues—in normal middle-aged females. Dashed arrow indicates that the SE in aging females can be ‘shifted’ even more to the right, when compared to the TP females (solid thick arrow)—see text, and data in Figs. 4A and B, 5A.

4.4. The tissue control system (TCS) theory No population of somatic cells that multiplies in vitro expresses the full panoply of differentiated traits of which the cell type is capable [80]. To induce partial differentiation, the tissue culture should be supplemented with various growth factors and cytokines, i.e. substances in vivo produced by mesenchymal cells [34,81,82]. The TCS theory [57,59,83] deals with the role of vascular pericytes, MDC, and T and B lymphocytes in regulation of tissue function. It proposes that MDC stimulate differentiation of specific tissue cells. They also regulate expression of epitopes of specific tissue cells, and in that way control their recognition by circulating autoreactive T lymphocytes and autoantibodies. Such T cells and antibodies are suggested to participate in stimulation of tissue cell differentiation. This may ultimately result in the aging and degeneration of tissue cells. By the end of the immune adaptation in early ontogeny, the MDC wandering through various peripheral tissues are sup-

posed to encounter the most differentiated tissue cells in a tissue specific manner, bring this information into the developing lymphoid tissues, and program future populations of MDC for prevention of tissue cells to differentiate beyond the encoded state by so called ‘stop-effect’ of MDC. The dominant role of MDC in regulation of immune response toward foreign antigens has been proposed [84,85], and the same may apply for the function of MDC in regulation of self tissue function by TCS. The nature of the stop-effect is considered to reside in the inability of MDC to stimulate differentiation of tissue cells beyond the encoded stage. Fig. 6A–D indicates that either none or only one stop-effect can be exerted toward particular tissue-specific cell type. The stop-effect can be exerted at any stage of the MDC pathway (primitive monocyte-like cells accompanying immature tissue cells, DC precursors accompanying mature tissue cells, and mature DC accompanying aged/regressing tissue cells). In addition, each of the MDC phenotype can exhibit various types of the stop-

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

effect. The ‘direct’ stop-effect represents a situation when MDC by themselves do not promote continuation of differentiation of dependent tissue cells. The ‘next step’ stopeffect consists of the promotion of tissue cells into the next stage of differentiation, but expression of new epitopes, which can be recognized by tissue-specific T cells or natural autoantibodies, is not stimulated. The MDC may also act indirectly, by prevention of production of certain tissuespecific natural autoantibodies (‘short’ indirect stop-effect) or by inhibition of production, homing, and activation of tissue-committed autoreactive T cells (‘long’ indirect stopeffect). The last two events may contribute to the preservation of the functional CL during pregnancy [57]. Hence, in this way, the MDC are potentially capable to control entire differentiation pathway of specific cells in each tissue, and cause a preservation of tissue cells at any particular stage (Fig. 6A–D) and sub-stage (e.g. 3a, 3b and 3c—not shown in Fig. 6A) encoded during the immune adaptation. These considerations are important for the detailed analysis, e.g. subtle differences between the TP and aging PE ovaries (see below). The ability of MDC to preserve tissue cells in the functional state declines with age, and this is accompanied by functional decline of various tissues within the body, and results in menopause and increased incidence of degenerative diseases in human beings (Fig. 6D). 4.5. Immune adaptation and the rat ovarian model In large mammals, including primates, the immune adaptation is terminated during intrauterine life (circa end of the second trimester of intrauterine life in man), whereas in small laboratory rodents (rats and mice) the immune adaptation also encompasses several postnatal days [86]. Multiple postnatal injections of estradiol dipropionate completely abolish the occurrence of LHR expression in ISC and granulosa cells of developing rat ovaries, and such females exhibit retardation of follicular development and complete lack of LHR expression in adult ovaries [59], despite normal basal serum levels of gonadotropins [87,88]. This indicates that suppression of early ovarian development results in permanent ovarian immaturity. On the other hand, injection of TP causes a lack of postnatal differentiation of mature ISC, and stimulates direct transformation of young ISC with strong cytoplasmic LHR expression into the aged ISC lacking LHR expression in developing ovaries (unpublished data). This is exactly what is later found in the adult ovaries of TP rats (Fig. 2E). However, TP-induced anovulation can be prevented by neonatal injection of thymic cell suspension from immunocompetent prepubertal but not five-day-old normal female donors [89,90]. This suggests that certain thymic cells (thymocytes, or thymic MDC) of normal immunocompetent females carry an information on the appropriate differentiation of ovarian structures, and this information can be transferred to immunologically immature neonatal rats and

201

prevent ovarian dysfunction by TP. It is here shown that either TP- or age-induced anovulation is associated with a different behavior of the monocyte-derived DC to the ovarian ISC. This suggests that MDC may play an important role in the development of mature ISC in normal adult ovaries, and cause premature aging of ISC in TP and aging PE rats. 4.6. Immune adaptation and epigenetic programming of tissue function It is now well established that the function of the ovary and other tissues within the body is regulated by immune system components [57,60,75–77,79,91–94]. However, despite these accomplishments, an unifying theory is missing to ascertain the role of the immune system in tissue physiology. Instead, an idea on the inert relationship of the immune system toward ‘self’ still prevails [95,96]. The determination of the specific relationships between the immune system and other tissues is still poorly understood, and it may depend on the outcome of the immune adaptation. Immune adaptation is an epigenetic phenomenon [86,97–99] and its impact for the regulation of adult tissue function may be a consequence of recognition and memorization of peculiarly differentiated (immature, mature, or aged) tissue-specific cells by the developing immune system in a tissue-specific manner (Fig. 6A–D). The immune system is not genetically programmed for the recognition of self tissue-specific cells, and therefore it has to ‘learn’ the occurrence of tissue-specific cells and extent of their differentiation during immune adaptation. By other words, and in wide terms, the optimal fitness of tissue cells (optimal stage of differentiation) for their optimal function during childhood, adulthood, and aging is not known by the homeostatic mechanisms of a developing individual, and has to be established depending on the genetically programmed presence or absence, timing of appearance, and stage of differentiation of tissue-specific cells during early development. The optimal differentiation of tissue cells varies among tissues (liver versus epidermis) and the preservation of optimal function during the life-span also varies (liver versus ovary). Hence, programming of tissue function for the lifespan is, in principle, an open issue, and genetic or epigenetic (exogenous steroids) retardation or acceleration of certain tissue differentiation during early development may cause a rigid and permanent alteration of this tissue function (Fig. 6C and D). The TCS theory assumes that there are three mutually connected developmental aspects that pre-determine tissue function and aging in immunocompetent individuals. They consists of: 1. Presence or absence of tissue-specific cells during immune adaptation. If some tissue is absent, like the CL of menstruation, its life in adult individuals is extremely short. Due to the absence of the CL during immune adaptation no stop-effect is established (Fig.

202

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

6A), and regression of the entire structure is induced and maintained by the autoreactive T cells and antibodies [57]. Most importantly, such tissue is by itself unable to regenerate, versus epidermis, which also lacks the stop-effect (Fig. 6A) but differentiates in early ontogeny. 2. Timing of the tissue appearance before and during immune adaptation in early ontogeny. For instance liver, which is present from the earliest stages of ontogeny, can function in human beings over one hundred years, whereas the ovary, which differentiates relatively late, can function for about a half of that period. 3. Stage of tissue differentiation attained by the end of the immune adaptation. For instance, epidermis, which differentiates into apoptotic surface cells from early stage of development [100], shows the same extent of differentiation in adulthood. The liver, which lacks apoptotic cells during early development, does not exhibit apoptosis during subsequent period of life. The ISC, which normally differentiate into the mature state during immune adaptation, show the same differentiation in adult females. The TP rats, which exhibit aged ISC during immune adaptation, show prevalence of aged ISC in adult ovaries. Longlasting estrogenization causes inhibition of ovarian development and lack of mature structures in the adult ovary [59]. The potential mechanism(s) by which TP imprints its putative effects on the ovary during immune adaptation can be viewed as follows: Postnatal injection of TP causes acceleration of the ISC development toward aging. The occurrence of aged ISC is recognized and encoded by wandering MDC during immune adaptation. This causes that the aged ISC are interpreted by MDC to be the appropriate structures for adult ovaries. Because of that, the stop-effect of MDC is set higher then required for the normal function of ISC (Fig. 6D versus B). In addition, estrogens can act differently toward different estrogen-sensitive tissues, depending on their inhibitory or stimulatory effect. For instance long lasting estrogen treatment of postnatal rodents causes permanent inhibition of ovarian development [59]— due to the inhibition of postnatal serum levels of gonadotropins by estrogens [101]. On the other hand, a direct stimulatory effect of the estrogen on the developing vagina causes estrogen-independent cornification of the vagina in adult females [102], including ‘diethylstilbestrol daughters’ of women estrogenized during pregnancy [103]. Hence, the tissue-specific effect of steroids injected during immune adaptation appears to be a consequence of temporary alteration (inhibition or stimulation) of normal tissue development during immune adaptation, rather than a consequence of the direct effect of steroids on the developing immune system. Also, when a low dose of TP (5 ␮g) is injected on

postnatal Day 5, the rats exhibit so called delayed anovulatory syndrome with persistent estrus after a significant period (60 –100 days) of normal cycles [104]. This delayed manifestation of ovarian dysfunction, possibly caused by a shorter effect of TP on developing ovaries, resembles some human degenerative diseases with autoimmune character, which are also manifested after a shorter (juvenile diabetes mellitus) or longer period of normal tissue function (Alzheimer’s disease). On the other hand, permanent ovarian immaturity after postnatal estrogenization resembles a situation in certain muscular dystrophies, where the skeletal muscle development is inhibited during human fetal development by maternal serum factors, and muscle immaturity persists thereafter [105]. These dependencies are summarized in Fig. 6C and D. When compared to other tissues, the ovary differentiates relatively late in ontogeny. This may be why in normal females an appropriate setting of the stop-effect toward ovary is shifted to the ‘right’ (Fig. 6D versus 6B) at certain age [83], corresponding to the beginning of the immune senescence [43,106]. Subsequently, a preponderance of aged interstitial glands and development of aged cystic follicles is observed in aging PE rats. Although ovaries of TP and aging PE rats showed many similarities, the quantitative analysis indicated some differences. They were characterized by significantly higher LHR depletion in ISC (Fig. 4A), significantly higher LHR expression in the stroma (Fig. 4B), and significantly more extensive vacuolization of ISC (Fig. 5A) in aged PE rats, when compared to TP females. These data suggest that the stop-effect of MDC for ISC in aging PE rats is shifted more to the ‘right’ (dashed arrow, Fig. 6D) when compared to the TP females (solid thick arrow, Fig. 6D). This will allow the ISC in aging PE rats to persist in more advanced stage of degeneration (lower LHR expression and larger cytoplasmic vacuoles), and cause a delay of the final regression of some interstitial cells (isolated LHR⫹ cells in stroma).

5. Conclusion In conclusion, the TP injected within a narrow time window of early ovarian development causes in adult rat females histologic findings resembling those found in the anovulatory ovaries of the middle aged PE rats. This is associated with altered behavior of MDC toward interstitial cells. Better understanding of the role of early ovarian development for the function of adult ovaries may be helpful for a better understanding of similar mechanisms responsible for various epigenetically-induced diseases (Fig. 6C and D). Further studies are required to clarify a possible role of MDC in the TP-induced premature ovarian aging. They may include an adoptive transfer of defined subpopulations of thymic or spleen cells from immunocompetent TP rats to neonatal recipients, in order to induce the premature aging of the ovary by nonhormonal means.

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

Acknowledgments We thank the editors for the selection of appropriate and open-minded reviewers, and to the reviewers for highly professional and constructive criticism. The skillful technical assistance of Cathy Tulloch, H. T., of the Department of Pathology, The University of Tennessee Medical Center at Knoxville, is greatly appreciated.

References [1] Barraclough CA. Production of anovulatory, sterile rats by single injections of testosterone propionate. Endocrinology 1961;68:62–7. [2] Mennin SP, Kubo K, Gorski RA. Pituitary responsiveness to luteinizing hormone-releasing factor in normal and androgenized female rats. Endocrinology 1974;95:412– 6. [3] Kubo K, Mennin SP, Gorski RA. Similarity of plasma LH release in androgenized and normal rats following electrochemical stimulation of the basal forebrain. Endocrinology 1975;96:492–500. [4] Jones HM, Vernon MW, Rush ME. Systematic studies invalidate the neonatally androgenized rat as a model for polycystic ovary disease. Biol Reprod 1987;36:1253– 65. [5] Danguy A, Pasteels JL, Ectors F. Sensitivity of anterior hypothalamic areas to gonadal steroid implantation in androgenized female rats. J Endocrinol 1977;74:315–22. [6] Sarkar DK, Fink G. Mechanism of the first spontaneous gonadotrophin surge and that induced by pregnant mare serum and effects of neonatal androgen in rats. J Endocrinol 1979;83:339 –54. [7] Zachow RJ, Magoffin DA. Granulosa cell modulation of luteinizing hormone-dependent androgen production by ovarian theca-interstitial cells: a temporal switch from suppression to augmentation stimulated by follicle-stimulating hormone in vitro. Biol Reprod 1995; 53:758 – 65. [8] Abhilasha, Krishna A. High androgen production by ovarian thecal interstitial cells: a mechanism for delayed ovulation in a tropical vespertillionid bat, Scotophilus heathi. J Reprod Fertil 1996;106: 207–11. [9] Spinedi E, Mariani V, Bulfon M, Colombani–Vidal M, Scaglia H. Analysis of the hypothalamic-pituitary-ovary axis in the neonatallyandrogenized female rat. J Endocrinol Invest 1990;13:481– 8. [10] Guraya SS. Interstitial gland tissue of mammalian ovary. Acta Endocrinol Suppl (Copenh) 1973;171:5–27. [11] Guraya SS. Histochemical observations on the lipid changes in the ovarian interstitial gland tissue of rats. J Reprod Fertil 1975;45: 141–5. [12] Louvet JP, Harman SM, Schrieber JR, Ross GT. Evidence of a role of androgens in follicular maturation. Endocrinology 1975;97:366 – 72. [13] Guraya SS, Gupta SK. Morphology and histochemistry of ovarian changes in the field rat, Millardia meltada. Z Mikrosk Anat Forsch 1979;93:959 –73. [14] Hillier SG, Zeleznik AJ, Knazek RA, Ross GT. Hormonal regulation of preovulatory follicle maturation in the rat. J Reprod Fertil 1980; 60:219 –29. [15] Magoffin DA, Reynolds DS, Erickson GF. Direct inhibitory effect of GnRH on androgen secretion by ovarian interstitial cells. Endocrinology 1981;109:661–3. [16] Magoffin DA, Erickson GF. Direct inhibitory effect of estrogen on LH-stimulated androgen synthesis by ovarian cells cultured in defined medium. Mol Cell Endocrinol 1982;28:81–9. [17] Magoffin DA, Erickson GF. Mechanism by which GnRH inhibits androgen synthesis directly in ovarian interstitial cells. Mol Cell Endocrinol 1982;27:191– 8.

203

[18] Magoffin DA, Erickson GF. Prolactin inhibition of luteinizing hormone-stimulated androgen synthesis in ovarian interstitial cells cultured in defined medium: mechanism of action. Endocrinology 1982;111:2001–7. [19] Erickson GF, Magoffin DA, Jones KL. Theca function in polycystic ovaries of a patient with virilizing congenital adrenal hyperplasia. Fertil Steril 1989;51:173– 6. [20] Magoffin DA. Evidence that luteinizing hormone-stimulated differentiation of purified ovarian thecal-interstitial cells is mediated by both type I and type II adenosine 3⬘,5⬘-monophosphate-dependent protein kinases. Endocrinology 1989;125:1464 –73. [21] Nicosia M, Moger WH, Dyer CA, Prack MM, Williams DL. Apolipoprotein-E messenger RNA in rat ovary is expressed in theca and interstitial cells and presumptive macrophage, but not in granulosa cells. Mol Endocrinol 1992;6:978 – 88. [22] Magoffin DA, Weitsman SR. Differentiation of ovarian theca-interstitial cells in vitro: regulation of 17 alpha-hydroxylase messenger ribonucleic acid expression by luteinizing hormone and insulin-like growth factor-I. Endocrinology 1993;132:1945–51. [23] Magarelli PC, Zachow RJ, Magoffin DA. Developmental and hormonal regulation of rat theca-cell differentiation factor secretion in ovarian follicles. Biol Reprod 1996;55:416 –20. [24] Sridaran R, Gibori G. Intraovarian localization of luteinizing hormone/human chorionic gonadotropin stimulation of testosterone and estradiol synthesis in the pregnant rat. Endocrinology 1983;112: 1770 – 6. [25] Dyer CA, Erickson GF. Norepinephrine amplifies human chorionic gonadotropin-stimulated androgen biosynthesis by ovarian thecainterstitial cells. Endocrinology 1985;116:1645–52. [26] Magoffin DA, Erickson GF. An improved method for primary culture of ovarian androgen-producing cells in serum-free medium: effect of lipoproteins, insulin, and insulinlike growth factor-I. In Vitro Cell Dev Biol 1988;24:862–70. [27] Magoffin DA, Gancedo B, Erickson GF. Transforming growth factor-beta promotes differentiation of ovarian thecal-interstitial cells but inhibits androgen production. Endocrinology 1989;125:1951– 8. [28] Zachow RJ, Magoffin DA. Granulosa cell modulation of luteinizing hormone-dependent androgen production by ovarian theca-interstitial cells: a temporal switch from suppression to augmentation stimulated by follicle-stimulating hormone in vitro. Biol Reprod 1995; 53:758 – 65. [29] Gelety TJ, Magoffin DA. Ontogeny of steroidogenic enzyme gene expression in ovarian theca- interstitial cells in the rat: regulation by a paracrine theca-differentiating factor prior to achieving luteinizing hormone responsiveness. Biol Reprod 1997;56:938 – 45. [30] Zachow RJ, Weitsman SR, Magoffin DA. Hepatocyte growth factor regulates ovarian theca-interstitial cell differentiation and androgen production. Endocrinology 1997;138:691–7. [31] Zhang G, Curtiss LK, Wade RL, Dyer CA. An apolipoprotein E synthetic peptide selectively modulates the transcription of the gene for rat ovarian theca and interstitial cell P450 17alpha-hydroxylase, C17–20 lyase. J Lipid Res 1998;39:2406 –14. [32] Magoffin DA, Erickson GF. Purification of ovarian theca-interstitial cells by density gradient centrifugation. Endocrinology 1988;122: 2345–7. [33] Billig H, Furuta I, Hsueh AJ. Estrogens inhibit and androgens enhance ovarian granulosa cell apoptosis. Endocrinology 1993;133: 2204 –12. [34] Rosen EM, Grant DS, Kleinman HK, Goldberg ID, Bhargava MM, Nickoloff BJ, Kinsella JL, Polverini P. Scatter factor (hepatocyte growth factor) is a potent angiogenesis factor in vivo. Symp Soc Exp Biol 1993;47:227–34. [35] Diaz–Flores L, Gutierrez R, Varela H. Behavior of postcapillary venule pericytes during postnatal angiogenesis. J Morphol 1992; 213:33– 45.

204

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

[36] Zachow RJ, Tash JS, Terranova PF. Tumor necrosis factor-alpha induces clustering in ovarian theca-interstitial cells in vitro. Endocrinology 1992;131:2503–13. [37] Wang Y, Walsh SW. TNF alpha concentrations and mRNA expression are increased in preeclamptic placentas. J Reprod Immunol 1996;32:157– 69. [38] Foghi A, Teerds KJ, van der Donk H, Dorrington J. Induction of apoptosis in rat thecal/interstitial cells by transforming growth factor alpha plus transforming growth factor beta in vitro. J Endocrinol 1997;153:169 –78. [39] Guilbert L, Robertson SA, Wegmann TG. The trophoblast as an integral component of a macrophage-cytokine network. Immunol Cell Biol 1993;71:49 –57. [40] Kearsey JA, Stadnyk AW. Isolation and characterization of highly purified rat intestinal intraepithelial lymphocytes. J Immunol Methods 1996;194:35– 48. [41] Stanulis–Praeger BM. Cellular senescence revisited: a review. Mech Ageing Dev 1987;38:1– 48. [42] Eiben R, Liebich HG. Detection and evaluation of lesions in the retina, optic nerve, and ciliary muscle of old rats using a electron microscope. Lens Eye Toxic Res 1989;6:269 – 87. [43] Bukovsky A, Caudle MR, Keenan JA, Wimalasena J, Upadhyaya NB, Van Meter SE. Is irregular regression of corpora lutea in climacteric women caused by age-induced alterations in the “tissue control system”? Am J Reprod Immunol 1996;36:327– 41. [44] Sell DR, Primc M, Schafer IA, Kovach M, Weiss MA, Monnier VM. Cell-associated pentosidine as a marker of aging in human diploid cells in vitro and in vivo. Mech Ageing Dev 1998;105:221– 40. [45] Sloviter RS, Dean E, Sollas AL, Goodman JH. Apoptosis and necrosis induced in different hippocampal neuron populations by repetitive perforant path stimulation in the rat. J Comp Neurol 1996;366:516 –33. [46] Gomez–Cuetara C, Flores JM, Sanchez J, Rodriguez A, Sanchez MA. Histological changes in the uterus during postpartum in the mare. Anat Histol Embryol 1995;24:19 –23. [47] Bukovsky A, Chen TT, Wimalasena J, Caudle MR. Cellular localization of luteinizing hormone receptor immunoreactivity in the ovaries of immature, gonadotropin-primed and normal cycling rats. Biol Reprod 1993;48:1367– 82. [48] Indrapichate K, Meehan D, Lane TA, Chu S-Y, Rao CV, Johnson D, Chen TT, Wimalasena J. Biological actions of monoclonal luteinizing hormone/human chorionic gonadotropin receptor antibodies. Biol Reprod 1992;46:265–78. [49] Teerds KJ, Dorrington JH. Immunolocalization of transforming growth factor alpha and luteinizing hormone receptor in healthy and atretic follicles of the adult rat ovary. Biol Reprod 1995;52:500 – 8. [50] Takao Y, Honda T, Ueda M, Hattori N, Yamada S, Maeda M, Fujiwara H, Mori T, Wimalasena J. Immunohistochemical localization of the LH/HCG receptor in human ovary: HCG enhances cell surface expression of LH/HCG receptor on luteinizing granulosa cells in vitro. Mol Hum Reprod 1997;3:569 –78. [51] Gelety TJ, Magoffin DA. Ontogeny of steroidogenic enzyme gene expression in ovarian theca-interstitial cells in the rat: regulation by a paracrine theca-differentiating factor prior to achieving luteinizing hormone responsiveness. Biol Reprod 1997;56:938 – 45. [52] Peters H, McNatty KP. The Ovary. A Correlation of Structure and Function in Mammals. Los Angeles: University of California Press, 1980. [53] Gaytan F, Morales C, Bellido C, Aguilar E, Sanchez–Criado JE. Ovarian follicle macrophages: is follicular atresia in the immature rat a macrophage-mediated event? Biol Reprod 1998;58:52–9. [54] Bukovsky A, Trebichavsky I, Presl J, Zidovsky J. Submicroscopic evidence of lymphoid cells within the granulosa of the rat atretic follicle. IRCS Med Sci 1978;6:80.

[55] Bukovsky A, Presl J, Holub M. The ovarian follicle as a model for the cell-mediated control of tissue growth. Cell Tissue Res 1984; 236:717–24. [56] Greenwald GS, Terranova PF. Follicular selection and its control. In: Knobil E, Neill JD, editors. Physiology of Reproduction, Vol. 1. New York: Raven Press, 387– 445. [57] Bukovsky A, Caudle MR, Keenan JA, Wimalasena J, Upadhyaya NB, Van Meter SE. Is corpus luteum regression an immune-mediated event? Localization of immune system components, and luteinizing hormone receptor in human corpora lutea. Biol Reprod 1995; 53:1373– 84. [58] Bukovsky A, Caudle MR, Keenan JA, Wimalasena J, Foster JS, Van Meter SE. Quantitative evaluation of the cell cycle-related retinoblastoma protein and localization of Thy-1 differentiation protein and macrophages during follicular development and atresia, and in human corpora lutea. Biol Reprod 1995;52:776 –92. [59] Bukovsky A, Caudle MR, Keenan JA. Regulation of ovarian function by immune system components: the tissue control system (TCS). In: Motta PM, editor. Microscopy of Reproduction and Development: A Dynamic Approach. Rome: Antonio Delfino Editore, 1997. pp. 79 – 89. [60] Bukovsky A, Caudle MR, Keenan JA, Wimalasena J, McKenzie PP. Thy-1 differentiation protein and monocyte-derived cells during regeneration and aging of human placental villi. Am J Reprod Immunol 1999;42:135–52. [61] Ahmed CE, Dees WL, Ojeda SR. The immature rat ovary is innervated by vasoactive intestinal peptide (VIP)-containing fibers and responds to VIP with steroid secretion. Endocrinology 1986;118: 1682– 89. [62] Lawrence IEJ, Burden HW. The autonomic innervation of the interstitial gland of the rat ovary during pregnancy. Am J Anat 1976; 147:81–94. [63] Svensson KG, Owman Ch, Sjoberg NO, Sporring B, Walls B. Ultrastructural evidence for adrenergic innervation of the interstitial gland in the guinea pig ovary. Neuroendocrinology 1975;17:40 –7. [64] Gorski RA. Sexual differentiation of the hypothalamus. In: Mack HC, Sherman AI, editors, The Neuroendocrinology of Human Reproduction. Springfield: CC Thomas Publ., 1971. pp. 60 –90. [65] Wise PM, Krajnak KM, Kashon ML. Menopause: the aging of multiple pacemakers. Science 1996;273:67–70. [66] Knobil E, Yen SS. The advent of menopause. Science 1996;274: 18 –20. [67] Morley P, Armstrong DT, Calaresu FR. Ovarian nerve extracts influence androgen production by cultured ovarian thecal cells. Neuroendocrinology 1989;50:93–9. [68] Rosa, Prajiante TM, Almeida FH, Guimaraes MA, Lunardi LO. Ovarian granulosa and theca interstitial cells: a morphological and physiological analysis in guanethidine denervated rats at pre-puberty. Cell Mol Biol (Noisy-le-grand) 1997;43:409 –16. [69] Dominguez R. Differential ovulatory responses of the right and left ovaries of the adult rat to unilateral lesion and anaesthesia of the cervico-vaginal plexus. J Endocrinol 1990;124:43–5. [70] Nakamura Y, Kato H, Terranova PF. Abdominal vagotomy decreased the number of ova shed and serum progesterone levels on estrus in the cyclic hamster. Endocrinol Jpn 1992;39:141–5. [71] Chavez R, Dominguez R. Participation of the superior ovarian nerve in the regulation of compensatory ovarian hypertrophy: the effects of its section performed on each day of the oestrous cycle. J Endocrinol 1994;140:197–201. [72] Tamura K, Tamura H, Kumasaka K, Miyajima A, Suga T, Kogo H. Ovarian immune cells express granulocyte-macrophage colonystimulating factor (GM-CSF) during follicular growth and luteinization in gonadotropin-primed immature rodents. Mol Cell Endocrinol 1998;142:153– 63. [73] Vigano P, Gaffuri B, Ragni G, Di BA, Vignali M. Intercellular adhesion molecule-1 is expressed on human granulosa cells and

A. Bukovsky et al. / Steroids 65 (2000) 190 –205

[74]

[75]

[76]

[77] [78]

[79]

[80] [81]

[82]

[83] [84] [85] [86] [87]

[88]

[89]

mediates their binding to lymphoid cells. J Clin Endocrinol Metab 1997;82:101–5. Best CL, Pudney J, Anderson DJ, Hill JA. Modulation of human granulosa cell steroid production in vitro by tumor necrosis factor alpha: implications of white blood cells in culture. Obstet Gynecol 1994;84:121–7. Emi N, Kanzaki H, Yoshida M, Takakura K, Kariya M, Okamoto N, Imai K, Mori T. Lymphocytes stimulate progesterone production by cultured human granulosa luteal cells. Am J Obstet Gynecol 1991; 165:1469 –74. Hoek A, van Kasteren Y, de Haan–Meulman M, Schoemaker J, Drexhage HA. Dysfunction of monocytes and dendritic cells in patients with premature ovarian failure. Am J Reprod Immunol 1993;30:207–17. Chryssikopoulos A. The relationship between the immune and endocrine systems. Ann N Y Acad Sci 1997;816:83–93. Spanel–Borowski K, Ricken AM. Evidence for the maintenance of macrophage-like cells in long-term bovine granulosa cell cultures. Cell Tissue Res 1997;288:529 –38. Vinatier D, Dufour P, Tordjeman–Rizzi N, Prolongeau JF, Depret– Moser S, Monnier JC. Immunological aspects of ovarian function: role of the cytokines. Eur J Obstet Gynecol Reprod Biol 1995;63: 155– 68. Harris H. Unfinished business. In: Gordon S, editor, The Legacy of Cell Fusion. Oxford: Oxford University Press, 1994. pp. 286 –95. Hughes FM Jr, Lane TA, Chen TT, Gorospe WC. Effects of cytokines on porcine granulosa cell steroidogenesis in vitro. Biol Reprod 1990;43:812–7. Garcia–Lloret MI, Morrish DW, Wegmann TG, Honore L, Turner AR, Guilbert LJ. Demonstration of functional cytokine-placental interactions: CSF-1 and GM-CSF stimulate human cytotrophoblast differentiation and peptide hormone secretion. Exp Cell Res 1994; 214:46 –54. Bukovsky A, Michael SD, Presl J. Cell-mediated and neural control of morphostasis. Med Hypotheses 1991;36:261– 8. Pennisi E. Teetering on the brink of danger. Science 1996;271: 1665–7. Banchereau J, Steinman RM. Dendritic cells and the control of immunity. Nature 1998;392:245–52. Klein J. Immunology: the science of self-nonself discrimination. New York: John Wiley and Sons, Inc., 1982. Nagasawa H, Yanai R, Kikuyama S, Mori J. Pituitary secretion of prolactin, luteinizing hormone and follicle-stimulating hormone in adult female rats treated neonatally with oestrogen. J Endocrinol 1973;59:599 – 604. Matsumoto A, Asai T, Wakabayashi K. Effects of x-ray irradiation on the subsequent gonadotropin secretion in normal and neonatally estrogenized female rats. Endocrinol Jpn 1975;22:233– 41. Kincl FA. Permanent Atrophy of Gonadal Glands Induced by Steroid Hormones. Charles University, Prague. Ph.D. Thesis. 1965.

205

[90] Kincl FA, Oriol A, Folch Pi A, Maqueo M. Prevention of steroidinduced sterility in neonatal rats with thymic cell suspension. Proc Soc Exp Biol Med 1965;120:252–5. [91] Gaytan F, Bellido C, Morales C, Reymundo C, Aguilar E, van Rooijen N. Selective depletion of testicular macrophages and prevention of Leydig cell repopulation after treatment with ethylene dimethane sulfonate in rats. J Reprod Fertil 1994;101:175– 82. [92] Beagley KW, Husband AJ. Intraepithelial lymphocytes: origins, distribution, and function. Crit Rev Immunol 1998;18:237–54. [93] Kroemer G, Cuende E, Martinez C. Compartmentalization of the peripheral immune system. Adv Immunol 1993;53:157–216. [94] Avrameas S. Natural autoantibodies: from ‘horror autotoxicus’ to ‘gnothi seauton’. Immunol Today 1991;12:154 –9. [95] Huston DP. The biology of the immune system. J Am Med Assoc 1997;278:1804 –14. [96] Silverstein AM, Rose NR. On the mystique of the immunological self. Immunol Rev 1997;159:197–206. [97] Owen RD. Immunogenetic consequences of vascular anastomosis between bovine twins. Science 1945;102:400 –1. [98] Hasek M. Parabiosis of birds during embryonic development. Cesk Biol 1953;2:265–70. [99] Billingham RE, Brent L, Medawar PB. Actively acquired tolerance of foreign cells. Nature 1953;172:603– 6. [100] Holbrook KA, Odland GF. Regional development of the human epidermis in the first trimester embryo and the second trimester fetus. J Invest Dermatol 1980;74:161– 8. [101] Sharpe RM, Atanassova N, McKinnell C, Parte P, Turner KJ, Fisher JS, Kerr JB, Groome NP, Macpherson S, Millar MR, Saunders PT. Abnormalities in functional development of the Sertoli cells in rats treated neonatally with diethylstilbestrol: a possible role for estrogens in Sertoli cell development. Biol Reprod 1998;59:1084 –94. [102] Takasugi N. Development of permanently proliferated and cornified vaginal epithelium in mice treated neonatally with steroid hormones and the implication in tumorigenesis. Natl Cancer Inst Monogr 1979;51:57– 66. [103] Vingerhoets AJ, Assies J, Goodkin K, Van HG, Bekker MH. Prenatal diethylstilbestrol exposure and self-reported immune-related diseases. Eur J Obstet Gynecol Reprod Biol 1998;77:205–9. [104] Swanson HE, Werff ten Bosch JJ van. The “early-androgen” syndrome; differences in response to prenatal and postnatal administration of various doses of testosterone propionate in female and male rats. Acta Endocrinol (Copenh) 1964;47:37–50. [105] Farkas–Bargeton E, Barbet JP, Dancea S, Wehrle R, Checouri A, Dulac O. Immaturity of muscle fibers in the congenital form of myotonic dystrophy: its consequences and its origin. J Neurol Sci 1988;83:145–59. [106] Weksler ME. Immune senescence. Ann Neurol 1994;35(suppl): S35–7.