Abstract
Mechanisms of androgen-induced thymic involution are largely undefined. We have found that significant decreases in thymic size occur 2–4 h after a dose of testosterone is administered to castrated male mice. This rapid rate of change suggests a role for androgen-induced apoptosis in modulating the size and composition of the thymus. Using thymic organ cultures to define these effects of androgens, we found that dihydrotestosterone treatment of thymus tissues from females or from castrated males results in enhancement of thymocyte apoptosis. Intact (androgen-replete) or testicular feminization, Tfm/Y (androgen-resistant) mice failed to show apoptotic change with androgen treatment, although the apoptotic response to glucocorticoids was present, suggesting a requirement for a functional androgen receptor. Acceleration of thymocyte apoptosis by androgens may mediate processes of thymocyte selection, with the potential to impart gender-specific characteristics on the peripheral T cell repertoire.
ANDROGENS exert considerable influence on the size and composition of the thymus. Removal of androgens by castration results in thymic enlargement, even in aged animals, whereas androgen replacement reverses these effects (1, 2). Mechanisms of androgen-induced involution of the thymus are incompletely understood but might be postulated to involve changes in cell proliferation, cell trafficking, and cell death. Previous studies have indicated that the first of these mechanisms, cell proliferation, contributes to thymic enlargement after castration (3). Increased rates of thymocyte proliferation are observed within the first few days after castration, subsiding after the thymus achieves a new steady-state size by 10 days after surgery (3). More recently, we have observed that marked decreases in thymus size occur in castrated mice within hours of a single testosterone injection. Such a rapid change seemed unlikely to be explained by decreased rates of cell proliferation but suggested that increased rates of thymocyte apoptosis might be induced.
It is well-recognized that apoptosis of thymocytes can be triggered by glucocorticoids (4, 5) through mechanisms that are dependent on calcium flux (6) and endogenous nuclease activation (7) and that require active gene transcription and protein synthesis (8). Limited previous studies with androgens have not shown increased rates of apoptosis in T cells or thymocytes. In the murine T cell line CTLL-2, no DNA fragmentation was observed when cells were cultured with relatively high concentrations of testosterone (10−5m) (9). Administration of testosterone to rats in vivo in other studies failed to result in demonstrable induction of thymic nuclease activity (10). One potential problem with in vivo studies is the rapid disappearance of apoptotic cells from tissues. Such cells are rarely observed in the thymus, despite significant rates of ongoing cell death, and it is likely that after apoptosis, dead cells are quickly removed (11).
The current studies, which used an in vitro organ culture approach, indicate that treatment of thymus tissue with dihydrotestosterone results in accelerated rates of thymocyte programmed cell death, as shown by DNA fragmentation. These findings suggest a role for androgens in mechanisms of thymocyte selection and in shaping of the peripheral T cell repertoire.
Materials and Methods
Mice and treatment protocols
Normal male and female C57 Bl/6 mice were obtained from Harlan Laboratories (Indianapolis, IN). Castration of the C57 males was performed at 3 weeks of age, and studies were done 10 days or more after surgery, as previously described (2). In vivo hormone treatment was achieved using a single sc injection of testosterone cypionate (1 mg; Upjohn, Kalamazoo, MI) or dexamethasone (0.1 mg; Sigma Chemical Co., St. Louis, MO). Control animals in these experiments received injections of the vehicle (cottonseed oil; Sigma) alone. Testicular feminization mice (Tfm/Y) were obtained from Jackson Laboratories (Bar Harbor, ME) at 6–8 weeks of age.
Thymus organ culture
Cultures of thymus tissues were performed in 12-well plates (CoStar, Cambridge, MA) using techniques similar to those in other reports (12–15). Each well was prepared by immersing a piece of Gelfoam (Upjohn) in 2 ml of culture medium consisting of RPMI 1640 (Gibco, Grand Island, NY) supplemented with 10% charcoal-stripped FCS (Hyclone, Logan, UT). After 2 h, the Gelfoam was overlaid with a sterile Nucleopore polycarbonate filter (CoStar), and pieces of thymus were placed on top of this filter. The hormones dexamethasone and dihydrotestosterone (Sigma Chemical Co.) were added at 10−6m; control wells received the ethanol vehicle alone (0.1%). Cultures were incubated for various times up to 20 h, as indicated in the individual experiments, and tissues were then used to prepare cell suspensions or were embedded in paraffin for histological analyses.
DNA fragmentation on agarose gels
Thymocyte cell suspensions were prepared from organ culture tissues after 5 h of incubation. The tissue was ground between glass microscope slides, producing a single cell suspension that showed more than 90% viability by trypan blue dye exclusion. Equal numbers of thymocytes (4 × 106) from each of the treated tissues (control and hormone) were lysed with TE/Triton-X buffer on ice for 15 min. DNA fragments resisting centrifugation at 13,000 × g were extracted with phenol and chloroform, precipitated with cold ethanol, and resuspended in TE buffer. Each sample was loaded into a separate lane on a 1.8% agarose gel and electrophoresed at 60 V. The lanes do not contain equal quantities of DNA, but the total amount of low-molecular-weight DNA derived from each of the cell samples was loaded. Bands were visualized by staining with ethidium bromide.
Cell death enzyme-linked immunosorbent assay (ELISA)
The Cell Death Detection ELISA (Boehringer Mannheim, Indianapolis, IN) was used to quantitate fragmented DNA. Cell suspensions were prepared from organ culture tissues after 5 h or 20 h of incubation. Cells from each culture (0.5 × 106) were lysed in the buffer provided at 4 C for 30 min. The lysate was centrifuged to pellet large pieces of unfragmented DNA, and the supernatant (containing small DNA fragments) was removed for assay. Samples were stored at −70 C if not used immediately. For assay, polystyrene microwells were coated with the antihistone antibody (provided with the kit) and then washed and blocked. Diluted supernatants were added in replicates of 3 or 4 and incubated for 90 min at room temperature. Control wells received buffer alone. After incubation, the wells were washed and then incubated with a peroxidase-conjugated anti-DNA antibody. Bound anti-DNA was detected using the substrate ABTS (2, 2′-azino-di-[3-ethylbenzthiazoline sulfonate)). Color was quantitated at 405 nm using an automated plate reader. Higher values in this assay indicate the presence of larger amounts of fragmented DNA. Results are expressed as the percent increase (or decrease) relative to the untreated control sample in each experiment.
Detection of apoptosis in situ
Apoptotic cells were visualized in situ in paraffin-embedded tissue sections from 20-h thymus organ cultures using the ApopTag Apoptosis Detection Kit (Oncor, Gaithersburg, MD). This method is based on identification of fragmented pieces of genomic DNA. Digoxigenin-conjugated nucleotides were added to the fragmented site catalytically by terminal deoxynucleotidyl transferase. Parallel control samples were done without added enzyme, according to the manufacturer’s instructions. The newly-added nucleotides were labeled with an fluorescein isothiocyanate-conjugated antidigoxigenin antibody. Counterstaining with propidium iodide allowed identification of nuclei. When viewed with a fluorescence microscope, apoptotic nuclei appear yellow-green, whereas intact nuclei appear red or orange.
This technique also was used to examine suspensions of thymocytes prepared from the organ culture tissues. Cells were fixed with 1% paraformaldehyde, layered onto microscope slides in 2% BSA/PBS, and air-dried. Numbers of positively stained cells, as indicated by yellow-green fluorescence, were scored out of a total of 200 cells. Each sample was analyzed in duplicate.
Statistical analyses
Data were summarized as mean ± sem. Group means were compared using a 2-tailed Student’s t test. For the ELISA, results were expressed as percent increases or decreases in the hormone-treated cultures, compared with the corresponding control. A P value less than 0.05 was considered significant.
Results
Thymic involution after a single injection of testosterone
Castrated C57 male mice, injected with a single 1-mg dose of testosterone, showed a significant reduction in thymic size 2–4 h later (Fig. 1). Thymus glands, removed from five testosterone-treated castrated mice 2 h after hormone injection, had a mean weight of 81 ± 9 (sem) mg, compared with a corresponding value of 96 ± 8 mg in the oil-treated animals, representing an average decrease of 16 ± 6% for the testosterone-treated group (P < 0.04). In a separate series of animals studied 4 h after injection, eight mice treated with testosterone had a mean thymus weight of 50 ± 5 mg, which was significantly less than the corresponding control value of 105 ± 10 mg (P = 0.004). A similar reduction in thymic weight was observed 4 h after an injection of dexamethasone given at a dose equivalent to approximately 10% of the dose of testosterone (Fig. 1).
Thymus weights in castrated male mice treated with testosterone (TEST) for 2 h (closed bar) or 4 h (hatched bar) or dexamethasone (DEX) for 4–5 h (open bar). Values represent percent of oil-treated controls in each experiment and are shown as mean ± sem. Significance between controls and hormone-treated animals was determined by t test. Numbers of animals per group: testosterone (2 h), N = 5; testosterone (4 h), N = 8; and dexamethasone, N = 3.
DNA fragmentation measured by gel electrophoresis
Organ cultures were carried out using thymus tissue from a castrated C57 male mouse. After a 5-h incubation period with hormones, cultures containing 10−6m of either dexamethasone or dihydrotestosterone showed greater amounts of fragmented DNA than controls (Fig. 2). In cultures that were continued for more than 20 h, high levels of DNA fragmentation were seen in both the controls and the hormone-treated samples (data not shown).
Agarose gel electrophoresis of DNA fragments. Thymocytes were prepared from a 5-h organ culture of thymus from a castrated animal. The samples were treated in vitro with the ethanol vehicle alone (left lane), dexamethasone (10−6m, middle), or testosterone (10−6m, right). DNA fragments were isolated from lysed cells by centrifugation, loaded and electrophoresed on a 1.8% agarose gel with added ethidium bromide for visualization of the bands.
Measurement of DNA fragmentation in organ culture tissues by ELISA
Thymus tissues from C57 mice, cultured for 5 or 20 h with dexamethasone, showed significantly elevated levels of DNA fragmentation by ELISA. Values for the dexamethasone-treated samples ranged from 9–98% greater than the corresponding controls, with an overall mean increase of 53 ± 10% (P = 0.0016) (Fig. 3). Dihydrotestosterone also induced DNA fragmentation. Females (N = 3) and castrated males (N = 4) showed higher levels of DNA fragmentation in dihydrotestosterone-treated cultures, compared with corresponding controls, with relative increases ranging from 6–97%. For all seven animals, considered together, the mean increase was 35 ± 11% (P = 0.0013). Thymus tissues from two intact (androgen-replete) C57 males and from three Tfm/Y (androgen-resistant) males did not produce more DNA fragments in response to dihydrotestosterone; OD values were less than corresponding control cultures. Thymus tissue samples from these mice, nevertheless, did show enhanced DNA fragmentation in response to dexamethasone.
DNA fragmentation in thymus organ cultures measured by ELISA. Tissues from the indicated animals were cultured for 5 or 20 h in vitro, and DNA fragments were isolated. Cultures contained either dexamethasone (open bars) or dihydrotestosterone (solid bars) at 10−6m. Values represent mean percent of increase or decrease, compared with the corresponding control culture, without added hormone (± sem). Groups are compared for statistically significant differences by t test.
Visualization of apoptotic nuclei in thymus tissues
Evidence for androgen-induced apoptosis also was observed using in situ labeling of DNA fragments. Thymus tissues, cultured for 20 h without added hormone, showed largely intact nuclei (Fig. 4; left panel). By contrast, the majority of cells exposed to dexamethasone (Fig. 4, center) or dihydrotestosterone (Fig. 4, right) showed evidence of apoptotic change. The ApopTag method also was used to evaluate dispersed thymocytes from these same cultured tissues. The average number of cells with fragmented DNA was 20% for the control, 59% for dihydrotestosterone, and 64% for dexamethasone. These findings confirmed the visual impressions from examination of the stained sections, indicating that androgens accelerate the process of thymocyte apoptosis.
Visualization of apoptotic nuclei in organ culture tissues. Control (left panel), dexamethasone-treated (center), and androgen-treated (right) sections from a 20-h organ culture were stained for DNA fragmentation using terminal deoxynucleotidyl transferase-mediated incorporation of digoxigenin-conjugated nucleotides into DNA fragments. Detection was with fluorescein isothiocyanate-conjugated antidigoxigenin antibody (yellow-green fluorescence). Sections are counterstained with propidium iodide to show intact nuclei (orange-red).
Discussion
The results of the present studies indicate that androgens accelerate thymocyte apoptosis. Apoptosis, induced by androgens, was qualitatively documented by agarose gel electrophoresis (DNA laddering) using organ culture tissues and a 5-h incubation period. Recognizing the limitations of this technique, two additional methods were used, namely the cell death ELISA for DNA fragments and in situ labeling of cells containing DNA fragments (ApopTag, Oncor). The greater precision of the ELISA approach may explain why elevated levels of apoptosis induced by androgens were shown, even in 20-h organ cultures, whereas by gel electrophoresis, qualitative differences between treated and untreated samples were not readily apparent.
Normal females and castrated males showed increases in thymocyte apoptosis after androgen administration, whereas intact (androgen-replete) males and Tfm/Y (androgen-resistant) males exhibited no increase in thymocyte apoptosis after such hormonal therapy. These data are consistent with a receptor-mediated effect of androgens. In previously reported studies, the mature murine T cell line CTLL-2 failed to show testosterone-induced apoptosis (9). Because peripheral T cells generally do not show androgen receptor (AR) positivity (16), this lack of response is most likely caused by the absence of functional ARs in these cells.
At least one previous study has reported using thymic organ cultures to examine effects of sex hormones (17). These investigators added hormones, including dihydrotestosterone to murine fetal thymus lobes for culture periods of 7–10 days and saw no effect on cell number or viability, although corticosteroids had the anticipated effect of diminishing the number of viable cells. This result was clarified by further experiments indicating that AR protein expression in the neonatal thymus does not reach significant levels until the animal is at least 4 weeks of age. This result, therefore, can be interpreted as further evidence that effects of androgens on thymocyte apoptosis are dependent on expression of a functional AR.
Previous examination of steroid hormone-mediated thymocyte apoptosis in rats revealed no evidence of androgen-induced apoptosis resulting from in vivo treatment with testosterone, although changes induced by glucocorticoids were readily detected (10). Several explanations are possible for the observed failure of androgens to initiate apoptosis in androgen-replete animals. First, androgens might regulate AR levels (i.e. down-regulate receptor and thereby reduce androgen sensitivity), but this seems unlikely. In previous experiments, we found no difference in immunoreactive AR levels between thymocytes from intact and androgen-deprived animals (18). Whether postreceptor signaling events might be modulated is unknown. Finally, whether a discrete subpopulation of androgen-sensitive thymocytes, susceptible to apoptosis, is depleted from the organ in intact male animals is a possibility, but such cells have not been definitively identified.
Thymocyte apoptosis is an important mechanism in the process of selection of cells that are to become mature peripheral T cells (19, 20). Many of the details of this selection process are only partially understood, but it seems that determination of whether positive or negative selection takes place is critically dependent on the surface density of relevant MHC molecules on thymic epithelial cells (21, 22). The MHC molecules bind self peptides that engage T cell receptors on immature thymocytes. Whether androgens might function to regulate the apoptotic pathways of these selection processes is unknown, but the present experiments suggest that such regulation is a possibility.
Mechanisms of androgen-induced thymocyte apoptosis remain undefined by the present studies. Other preliminary findings (which are not shown) suggest that the enlarged thymuses from castrated animals contain elevated levels of the BCL2 protein, an oncogene product that is generally, but not universally, associated with inhibition of apoptosis (23, 24). Whether this or other components of the programmed cell death pathway are altered by androgens remains to be elucidated.
The goal of our studies of the effects of androgens on the immune system is to understand mechanisms underlying the sexual dimorphism of immune responses and propensity to autoimmunity (25). Accumulated data suggest that androgens most likely exert effects on developing T cells within the thymus, rather than on mature T effector cells, because expression of ARs is not detected in the peripheral organs of the immune system (such as the spleen) (16, 26). Effects of androgen deprivation in the periphery include decreased numbers of splenic T cells, enhanced production of the Th1-type cytokines interleukin-2 and interferon-γ (27), and diminished ability of T cells to inhibit proliferation of other spleen cells (28). Effects of androgens on developing T cell subsets may thus result in the generation of a peripheral T cell repertoire that is skewed toward these functional characteristics. These studies leave open the question of how the hormonal milieu impacts on T cell selection in females. Estrogens may act to alter expression of thymocyte subsets that are distinct from those responding to androgens (29). The specific response to androgens and estrogens could be responsible for gender-specific differences in peripheral T cell populations (25).
Acknowledgements
The expert assistance of Maxine Turney is greatly appreciated. Steve Rivera, of Cytometry Associates, Brentwood, Tennessee, performed the flow cytometric analyses.
References
Author notes
This study was supported by the National Institutes of Health (DK-41053, to N.J.O.) and the Lupus Foundation of America (to W.J.K.), as well as by funds from the Nashville Chapter of the Lupus Foundation of America (to W.J.K. and N.J.O.).
