Predominant suppression of follicle-stimulating hormone β-immunoreactivity after long-term treatment of intact and castrate adult male rats with the gonadotrophin-releasing hormone agonist deslorelin.

Continuous delivery of gonadotropin-releasing hormone (GnRH) agonists suppresses sex steroid synthesis in most male mammals (e.g., mouse ( 1 ), rat ( 2 ), dog ( 3 ), monkey ( 4 ), sheep ( 5 , 6 )), although there are a few notable exceptions (possum ( 7 ), red deer ( 8 ), wallaby ( 9 ) and especially the bull ( 10 , 11 )). As GnRH agonists are extremely effective in humans, they offer a powerful alternative to surgical orchidectomy for contraception and gonadal steroid-dependent disorders, particularly prostate cancer ( 12 - 14 ); approximately 600 000 men received GnRH agonist therapy in 2007 ( 15 ). GnRH agonist therapy is also used in women (endometriosis ( 16 ), uterine fibroids ( 17 ), central precocious puberty ( 18 )). The number of people receiving agonist treatment may increase significantly as this therapy is also under investigation for Alzheimer’s disease ( 19 ), bowel disease ( 20 ), polycystic ovarian syndrome ( 21 ) and preservation of ovarian function during chemotherapy ( 22 ). Thus, GnRH agonist therapy may become even more commonplace. The mechanism by which GnRH agonists suppresses the pituitary component of the reproductive axis is not fully understood. There are three aspects that may be compromised and these are not mutually exclusive. First, there is compelling evidence that the GnRH receptor is affected. After initially stimulating the reproductive axis, GnRH agonists induce internalization and down-regulation of the GnRH receptor ( 23 , 24 ). However, it has been suggested for many years that GnRH receptor modification fails to explain all changes induced by GnRH agonists ( 25 - 30 ). Second, indirect evidence suggests that chronic exposure to GnRH agonists may uncouple the associated intracellular signaling systems ( 31 - 33 ). Third, there may be a change in the cellular composition of the pituitary gland. This last component, which has not been investigated, was hypothesized to explain some of the effects of chronic GnRH agonist exposure in men. Specifically, if GnRH agonist therapy lasts for several years, recovery of testosterone levels takes 18 months or more after the cessation of treatment ( 34 - 39 ). Experimental studies have shown that chronic GnRH agonist exposure decreases the pituitary content of gonadotropins ( 40 - 43 ) and in vitro evidence suggests that this may also occur in humans ( 44 ). Such depletion may involve an impaired ability to synthesize the gonadotropins and/or the loss of gonadotropes entirely. Identifying potential cellular changes may help further elucidate the mechanisms through which GnRH agonists inhibit reproductive function and provide initial insight into why sustained suppression of the reproductive axis may occur after treatment ceases. Accordingly, this study investigated the effect of long-term exposure to the GnRH agonist deslorelin on the density of FSHβ-, LHβ- and αGSU-immunoreactive (-ir) cells within the pituitary gland of the adult male rat. To determine whether any pituitary changes were through a direct action of deslorelin or through the accompanying testosterone loss, a some rats were testosterone replaced. In a second experiment on castrate rats, we determined whether potential pituitary changes were due, at least in part, on the loss of some testicular factor other than testosterone, (e.g. inhibin).

FSHβ-ir cells were most abundant (mono-FSHβ and bi-hormonal cells) and deslorelin significantly suppressed immunoreactivity in both populations ( Figs 1 , 2A ; Supplementary data Table 1 ). Treatments did not affect αGSU-ir cell number, although deslorelin significantly decreased the percentage of β subunit-immunoreactive gonadotropes ( Fig. 2A ). Deslorelin did not suppress the percentage of LHβ-ir cells, although the proportion of mono-hormonal LHβ-ir cells increased as a result of the loss of FSHβ-ir in the bi-hormonal cells ( Figs 1 , 2A ). The percentage of bi-hormonal gonadotropes was significantly reduced by deslorelin. Although testosterone replacement did not reduce the suppression in the overall population of β subunit-immunoreactive gonadotropes, there was an increase in FSHβ-ir. Specifically, testosterone replacement did not reverse the suppression in mono-FSHβ-ir cells, but it increased (or prevented the loss of) FSHβ immunoreactivity in bi-hormonal cells. These FSH changes were also evident in the plasma: deslorelin reduced plasma FSH (88%; p<0.001; Fig. 2B ) and testosterone reduced the extent of suppression (44%; p<0.01). Deslorelin also suppressed plasma LH (47%; p<0.01) but to a smaller extent than FSH. In contrast to FSH, when testosterone was co-administered with deslorelin, plasma LH suppression was enhanced (70%; p<0.05 vs deslorelin alone). Deslorelin suppressed plasma testosterone. The percentage of αGSU-ir cells was not affected by castration, testosterone or deslorelin but was reduced in castrate rats receiving both testosterone and deslorelin. Castration did not increase the percentage of gonadotropes ( Figs 3 , 4A ; Supplementary data Table 1 ). However, the proportion of bi-hormonal and mono-LHβ-ir cells increased significantly with a resultant rise in the percentage of total LHβ-ir cells ( Fig. 4A ). In contrast, the proportion of mono-FSHβ-ir decreased, leading to no difference in the total FSHβ-ir cell population ( Fig. 4A ). These changes in gonadotropin immunoreactivity were reversed by testosterone replacement. Deslorelin suppressed FSHβ- and LHβ-ir below intact controls. As in the intact group, testosterone replacement in deslorelin-treated castrated rats did not reduce the suppression in the percentage of total β subunit-immunoreactive gonadotropes, but there was an increase in FSHβ-ir. Together, these data suggest that the real gonadotrope population remains robust in the face of endocrine changes, although the ability to synthesize the β subunits is dramatically perturbed. Castration increased plasma FSH 6-fold and LH 12-fold (p<0.0001 for both) and these increases were restored to the intact control rat range by testosterone ( Fig. 4B ). Deslorelin treatment also prevented the castration-induced increases and suppressed FSH below intact control rat levels. Testis weight was suppressed by 67% ( Fig. 5 ). Epididymis (control: 169.3±6.3; D: 54.1±3.5 mg/100g BW) weights were suppressed by deslorelin to a third of that in the untreated rat (p<0.001). Testosterone replacement partially alleviated this effect in the epididymis (D+T: 93.6±7.7 mg/100g BW) but not in the testis (D+T: 263.3±31.2 mg/100g BW). Spermatozoa were absent from the testes of the deslorelin-treated animals at the end of the study. Deslorelin reduced (control: 3.3±0.1; D: 2.8±0.1 mg/100g BW), whereas castration increased (4.0±0.2 mg/100g BW), pituitary weight (p<0.05 for both). Testosterone replacement reversed these effects in deslorelin-treated intact (3.1±0.2 mg/100g BW) and castrate (2.8±0.2 mg/100g BW), as well as deslorelin-treated castrate (3.4±0.1 mg/100g BW), rats. Deslorelin also prevented the castration-induced increase (3.3±0.1 mg/100g BW) in pituitary weight.

This is the first study reporting on specific changes in the immunocytochemically detectable gonadotrope population of the adult male rat pituitary gland following long-term treatment with a GnRH agonist. FSHβ-ir was compromised by deslorelin in intact rats and plasma FSH was suppressed as expected. The reduction in FSHβ-ir cells was evident in both the mono- and bi-hormonal gonadotrope populations. Such an FSH-dominant suppression has been observed previously in response to buserelin in the ewe ( 52 ). Testosterone replacement to deslorelin-treated intact rats caused a significant increase in FSH-ir, especially in bi-hormonal cells, and a commensurate increase in plasma FSH. This indicates that deslorelin’s effects are mediated, at least in part, through the loss of testosterone. Deslorelin did not affect the αGSU-ir population, suggesting that this agonist does not eliminate the gonadotrope population per se but that there is a reduced ability to synthesize the FSHβ subunit. In marked contrast to FSH, deslorelin did not affect the density of LHβ-ir cells in intact rats. LH secretion was slightly suppressed by deslorelin and this suppression was enhanced in the presence of testosterone, which is notably different to FSH. Deslorelin potently suppressed LHβ-ir and plasma LH after castration. Although GnRH agonists have been used clinically for decades as an alternative to surgical castration, their exact mechanism of action remains poorly understood. The ability of GnRH agonists to decrease pituitary gonadotropin release has been attributed to GnRH receptor internalization and down-regulation. Although receptor down-regulation has been extensively studied in short-term investigations, fewer studies have determined the effects of long-term GnRH agonist exposure. It is notable therefore that after 30 days of triptorelin treatment in rats, GnRH receptor mRNA rose from the initial short-term down-regulation, and GnRH receptor protein levels were back at pre-treatment levels ( 28 ). Also, although triptorelin suppressed GnRH-induced LH and FSH release in young girls treated chronically for precocious puberty, GnRH still stimulated αGSU secretion ( 53 , 54 ). αGSU secretion was also markedly increased in patients treated for pituitary tumors with GnRH agonists, although LH and FSH were suppressed ( 55 ). More recently, Hirsch et al ( 56 ) reported that αGSU release increased in response to long-term exposure to histrelin, an agonist that powerfully suppresses LH and FSH secretion. Elevated αGSU release in response to deslorelin may be mediated through increased PACAP ( 57 , 58 ), which stimulates αGSU ( 59 , 60 ). Our study concurs with studies on rats showing marked reduction in FSHβ mRNA and pituitary FSH content after 28 days of continuous buserelin ( 40 , 42 ). Together, these studies suggest extensive and coinciding decreases in FSHβ mRNA and FSH protein following treatment, leading to a substantial reduction in the number of gonadotropes synthesizing FSHβ. As GnRH stimulates PACAP ( 57 , 58 ), the effect of deslorelin on FSHβ may be transduced through PACAP and/or follistatin. Continuous PACAP suppressed FSHβ mRNA levels ( 61 ) and PACAP increased follistatin mRNA in mouse ( 62 , 63 ). There is preliminary evidence that follistatin, in addition to binding activin, may affect FSHβ mRNA in primate pituitary cultures ( 64 ) and LβT2 cells ( 65 ). The decrease in FSHβ-ir was consistent with the changes in plasma FSH; a ~90% reduction in circulating FSH levels in intact and castrate rats after deslorelin treatment. Continuous infusion of buserelin also suppressed plasma FSH in intact and castrate rats ( 42 ). Leuprorelin similarly reduced plasma FSH ( 66 ), whereas goserelin exerted a more modest reduction ( 67 ). Interestingly, plasma levels of FSH rose following daily injections of alarelin in rats even after 12 weeks of treatment ( 68 ), suggesting that not all agonists may work through a common mechanism. Different agonists clearly have different potencies, depending on the assay ( 69 , 70 ); deslorelin is considered to be 114 times more potent than GnRH, whereas leuprorelin is only 15 times more potent ( 48 ). Thus, it is noteworthy in red deer that although buserelin infusion stimulated reproductive activity ( 8 ), leuprorelin suppressed it ( 71 ). It is possible that these two GnRH agonists affect intracellular pathways differently, at least in the long term. Although plasma LH was diminished in response to deslorelin in intact rats, the decrease was about half that of FSH. This suppression, possibly coupled with the more dramatic loss in FSH, was clearly sufficient to suppress testosterone. A deslorelin-mediated suppression in LH was much more evident in the castrate rats (95%), which is similar to reports on bulls ( 72 ). Indeed, castration may be essential to consistently detect a GnRH agonist-mediated suppression of LH. Chronic exposure to goserelin ( 67 ), leuprorelin ( 66 ) and buserelin ( 43 ) had no effect on plasma LH in intact rats, whereas buserelin significantly reduced plasma LH in castrate rats ( 43 ). Some of the discrepancy may also originate from the inherent pulsatility of LH secretion, which can produce a large variability within a group with a single blood sample and changes in LH are best investigated by serial blood sampling. This is illustrated in a recent study on the effect of leuprorelin on LH in the pubertal human male, where no differences in LH release were apparent before or during treatment in the first 4 hours of the night, but clear suppression was evident thereafter ( 73 ). FSH suppression occurred throughout the study ( 73 ). Previous studies have reported reduced LHβ mRNA in the intact male rat in response to triptorelin ( 74 ) and buserelin ( 43 ). Thus, the absence of an effect of deslorelin on the number of cells immunocytochemically detectable for LHβ in intact rats was unexpected. Our result must be viewed within certain caveats. First, it should be noted that immunocytochemistry is not an accurate method to determine the amount of a target substance within a cell ( 75 , 76 ). Thus, if two gonadotropes differed significantly in LHβ content, they may be detected equally by immunocytochemistry. However, as determined by electron microscopy, long-term leuprorelin exposure in mice reduced the LHβ content of individual secretory granules ( 77 ). Second, FSH is largely secreted constitutively ( 78 ), and changes in synthesis will therefore be reflected accurately immunocytochemically. In contrast, LH is stored prior to secretion ( 78 ) and, thus, in the absence of stimulation, LH may still be immunocytochemically detectable. However, the retained presence of LHβ-ir in our study may also be the result of cellular changes within the gonadotrope. As noted above, GnRH induces PACAP and although continuous PACAP did not affect LHβ mRNA levels, it protected the LHβ mRNA from degradation and thus lengthened its half-life ( 61 ). Thus, it is possible that while LHβ stores in, and secretion from, individual gonadotropes are reduced in response to deslorelin, these LHβ-synthesizing cells remain immunocytochemically detectable. The specific mechanisms responsible for the deslorelin-mediated suppression of FSHβ-ir, but retention of LHβ-ir, warrant further attention. It is well known that castration increases the pituitary gland gonadotropin content (e.g. ( 79 )). Our study concurs with a recent report ( 80 ) that this effect is not through gonadotrope proliferation. The increased pituitary weight in castrated rats is likely to be the result of the increased size of gonadotropes, the so called “castration cells” (see Fig. 3 ), as well as increased mitosis in the non-gonadotrope pituitary cell population ( 80 ). Deslorelin suppressed circulating testosterone by 90%, with a corresponding 70% decrease in testicular weight, which can be used as a rough index of spermatogenesis suppression ( 81 ). In previous studies using different agonists, testosterone suppression was similar ( 82 , 83 ), but the impact on testicular weight was less pronounced ( 40 , 68 , 82 , 83 ). Unlike in men ( 84 ), fertility is compromised but not eliminated in male mice with disrupted FSH systems ( 85 , 86 ). Thus, it is probable that this “castration” of the testes by deslorelin is not mediated solely through the marked reduction in FSH, and the smaller LH reduction is also likely to be important. In this respect, although testosterone supplementation increased FSH, testicular weight did not increase. Also, at least part of the effect of GnRH agonists may be through a direct testicular effect. Testicular GnRH receptors have been reported in several species, including humans ( 87 , 88 ). In rats, GnRH affects Ca 2+ mobilization ( 89 ) as well as steroidogenesis in Leydig cells ( 90 ). This reported direct effect of GnRH on steroidogenesis is especially noteworthy for the current study, as although testosterone implants returned testosterone concentrations to the normal range in castrated rats, the testosterone elevation was lower in deslorelin-treated intact rats. Moreover, buserelin treatment for 5 months decreased testosterone but increased testicular levels of dihydrotestosterone and estradiol ( 91 ), suggesting increased activity of 5α-reductase and aromatase, respectively. The effects of testosterone replacement on plasma gonadotropins in deslorelin-treated rats were diametrically opposite: it enhanced the deslorelin-mediated suppression of LH but elevated FSH secretion. This finding concurs with studies on GnRH antagonist-treated intact and castrate male rats ( 92 , 93 ) and may have been even more pronounced in our study if testosterone concentrations had been restored to control levels. It is well established that testosterone consistently suppresses LH mRNA ( 94 ). The enhanced suppression of LH by testosterone in our deslorelin-treated rats suggests that the effects are additive, possibly synergistic. In contrast, testosterone increases FSH mRNA and secretion in rodents ( 95 ). Some of this effect of testosterone may be through increased FSHβ stability ( 93 ). It is also noteworthy that testosterone suppressed pituitary follistatin mRNA ( 95 , 96 ). Although testosterone regulation of PACAP in the brain has been demonstrated ( 97 , 98 ), we are unaware of studies in the pituitary gland. Our study suggests that increased activin secretion is unlikely to mediate the effects of testosterone on FSH. Activin stimulates both LHβ and FSHβ ( 99 ), but plasma LH and LHβ-ir were suppressed in the deslorelin-treated rats supplemented with testosterone. However, if the direct suppressive effect of testosterone on the LHβ promoter was dominant over the inductive effect of activin on the LHβ promoter, then increased activin coupled to reduced follistatin would increase FSHβ, but decrease LHβ, synthesis. Synergism between androgens and activin signaling on the FSHβ promoter has been observed in LβT2 cells ( 100 ) and androgen mediated suppression of TGF-β signaling is evident in prostate tissue ( 101 ). Clearly, more mechanistic research is indicated. In summary, our study demonstrates specific changes in the cellular composition of the adult male rat pituitary gland after long-term treatment with deslorelin. The deslorelin treatment we used was far more suppressive on testis weight than previously reported using other agonists in the rat. Deslorelin compromised the FSHβ-ir cell population in both intact and castrated rats and the deslorelin-induced change in the composition of gonadotropes was similar in both groups. At least some of the suppression is mediated through the loss of testosterone as the proportion of FSH-ir cells, as well as plasma FSH, increased with testosterone replacement in intact rats. As deslorelin-mediated changes on FSH were similar between intact and castrated rats, it suggests that testicular factors other than testosterone are not essential for suppression. LH may be less susceptible to the effect of deslorelin alone, although marked deslorelin-induced LH changes were evident following castration. In contrast to FSH, the addition of testosterone served to amplify the suppressive effects of deslorelin on plasma LH. We hypothesize that the strong reduction in FSH by deslorelin in intact animals may be pivotal to the efficacy of this GnRH agonist when used as a chemical castrant in the rat. Thus, from an applied perspective, it may be more important to analyze changes in FSH, than the more routinely analyzed LH, to gauge efficacy of a specific GnRH analog in steroid deprivation therapy in the male.

Adult male Sprague-Dawley rats (135.9±0.2 days old) were treated for 41.0±0.6 days with a sham implant (Control; n =13), a slow-release 1.1 mg deslorelin implant (DESL; n =13; Suprelorin ® ; Peptech Animal Health; Sydney, Australia) or deslorelin implant and a 15 mg testosterone pellet (DESL+T; n =7; 90-day release, Innovative Research of America, Sarasota, FL). As reported by others ( 45 ), and supported by preliminary studies (data not shown), this testosterone treatment restored total testosterone concentrations in castrated rats to the intact range. Free testosterone levels were also not different between intact (4.5±1.9pg/ml) and testosterone-treated castrate (3.3±1.2pg/ml) rats. Based on these preliminary investigations, the testosterone dose was not increased as we did not want confounding anabolic effects, if this steroid was elevated above the normal range. Deslorelin implants were produced by carefully cutting 4.7mg Suprelorin implants into quarters with a scalpel, with particular care taken to prevent implant fragmentation. Implants (n=14) weighed 23.1±0.2mg, equating to ~1.1mg deslorelin per implant. As the in vitro deslorelin release rate from an implant containing 5mg deslorelin is >1ug/day ( 3 ), we estimated the smaller implants would release ~250ng/day; this is likely to be an underestimate on account of the increased surface area. Plasma deslorelin was 92.9±9.3 pg/ml at the end of the study. This concentration is lower than reported for leuprorelin in men: levels of 190-430pg/ml and 930pg/ml leuprorelin were evident after 3 and 12 month depot Lupron implants, respectively ( 46 ). Similarly, a different 4 month implant produced mean serum leuprorelin levels from day 3 to day 112 of treatment of between 1ng/ml and 100pg/ml, with individuals ranging from 50pg/ml to 5ng/ml ( 47 ). Deslorelin is, however, considered to be a more potent GnRH agonist than leuprorelin ( 48 ). All implants were inserted sc aseptically in the scapular region under isoflurane anesthesia. Rats were placed in individual cages with nesting supplies and a heat source for 6 days and then returned to group housing (4-5 per cage). Food and water were provided ad libitum and rats were maintained on a 12L/12D cycle (06:00-18:00; 21°C). Rats were weighed twice a week for the duration of the study. All studies were performed under IACUC authorization from the University of Wyoming. A second group of rats, run concomitantly with Experiment 1, were bilaterally castrated and treated with a sham implant (castrate; n =8), a deslorelin implant (castrate+D; n =8), a deslorelin implant and testosterone (castrate+D+T; n =6), or only testosterone (castrate+T; n =8). Implants were inserted and animals maintained as described. Rats were weighed twice a week. At the end of treatment, rats were anesthetized with ip ketamine (100mg/kg)/xylazine (10mg/kg) and injected iv with heparin (200 units/kg). Blood samples were collected via cardiac puncture. Rats were perfused transcardially with 150 ml saline, 300 ml Zamboni’s fixative ( 49 ) and 125 ml 20% sucrose in 0.01M PBS. Retention of implants was confirmed and organs were extirpated and weighed. Pituitary glands were stored in 20% sucrose in PBS (4°C) and sagittal pituitary sections (20μm) cut on a cryostat, mounted on Silane-coated slides and stored (-80°C). For immunolabeling the FSHβ and LHβ subunits, slides were washed (all washes 3×5 min in 0.01M PBS), incubated (48h; 4°C) in 0.01M PBS containing 10% goat serum, 0.3% Triton and rabbit anti-rat FSHβ (AFP7798-1289; 1:56000; NIDDK) and guinea pig anti-rat LHβ (AFP-22238790GPOLHB; 1:62000; NIDDK). Slides were washed, incubated (90 min; room temp) with Alexa Fluor 488 F(ab’) 2 fragment of goat anti-rabbit IgG (1:8000; Invitrogen; Carlsbad, CA) combined with affinity purified Texas Red goat anti-guinea pig IgG F(ab’) 2 (1:500; Rockland; Gilbertsville, PA), washed and coverslipped with Vectashield mounting medium with 4’,6-diamidino-2-phenylindole (DAPI), a fluorescent stain that binds strongly to DNA (Vector; Burlingame, CA). For staining the αGSU, sections were incubated (48h; 4°C) in guinea pig anti-rat αGSU (AFP5191792; 1:2000; NIDDK), washed, incubated in the Texas Red goat anti-guinea pig IgG F(ab’) 2 (1:250, 90 min; room temp), washed and coverslipped with Vectashield and DAPI. For additional evidence of the specificity of the guinea-pig anti-rat αGSU antibody, we had previously performed a colocalization study with a rabbit anti-rat αGSU antibody (AFP66P9986; 1:4000; NIDDK). Labeled cells were visualized with the Alexa Fluor 488 F(ab’) 2 goat anti-rabbit IgG (1:8000; Invitrogen) combined with affinity purified Texas Red goat anti-guinea pig IgG F(ab’) 2 (1:500; Rockland) as described. There was 100% colocalization between these two antisera (data not shown). On a subset of pituitary sections from control rats (n=4), we also confirmed that all cells labeled with the rabbit anti-rat FSHβ antibody were also labeled with the guinea pig anti-rat αGSU antibody (data not shown). Gonadotropes are randomly distributed throughout the anterior pituitary, although there is a concentration along the anterior aspect of the pars distalis ( 50 ). For analysis, two mid-sagittal sections, at least 120 μm apart, were used from each animal. Images were taken of 5 randomly selected fields (220 × 295μm) from each section using a Nikon Eclipse E800 equipped with an X-Cite™ 120 Fluorescence Illumination System (EXFO; Mississauga, Ontario, Canada) and SPOT RTke digital camera (Diagnostic Instruments; Sterling Heights, MI). The illumination parameters (e.g. brightness, gamma) were fixed for each fluorescent tag. Using Image J software (NIH, Bethesda, MD), the number of FSHβ-, LHβ-, FSHβ+LHβ-, αGSU-ir cells, and DAPI-labeled nuclei, were marked and then counted by the software in each field. Only cells exhibiting cytoplasmic immunoreactivity with a nuclear exclusion, which corresponded to a DAPI-labeled nucleus, were marked. The mean value of the 10 fields from each pituitary was used to estimate the value for each rat. We have previously determined (data not shown) that this value falls within the 95% confidence interval of 30 fields. FSH was measured by EIA (#29-AE-R004; ALPCO Diagnostics; Salem, NH) with a sensitivity of 0.2 ng/ml and intra-assay coefficient of variation of 6%. The RIA for rLH used reagents from the NIDDK (rLH-I-8 for iodination, AFP-12066B; rLH-S-10 for antiserum, AFP-571487; rLH-RP-3 for standard, AFP-7187B) and rLH was iodinated by the chloramine-T method ( 51 ). Standards ranged from 0.4 to 6 ng/ml. Assay sensitivity (0.09 ng/ml) was calculated from the standard curve as the value that exceeded the 0ng/ml by 2SDs. The intra-assay coefficient of variation was <5%. Total testosterone was estimated by RIA (#TKTT2; Diagnostic Products Corporation; Los Angeles, CA) with a sensitivity of 0.05ng/ml and intra-assay coefficient of variation was 5%. Free testosterone was measured by EIA (#11-FTSHU-E01; ALPCO Diagnostics; Salem, NH) with a sensitivity 0.17pg/ml and an intra-assay coefficient of variation of <5%. For the extraction of deslorelin, 200μl plasma was diluted in 800μl saline and then acidified with an equal volume of 1% trifluoroacetic acid (TFA, HPLC grade). HF Bond Elut C18 OH columns (Varian, Palo Alto, CA) were equilibrated with HPLC grade acetonitrile (1ml) and 1% TFA (9ml). The acidified plasma was added to the columns and washed with 1% TFA (6ml). The peptide was eluted with 6ml of 60% acetonitrile in 1% TFA. The eluate was evaporated to about half volume in a Savant SpeedVac (Thermo Fisher Scientific, Waltham, MA), frozen, and evaporated to dryness in a freeze dryer (LabConco, Kansas City, MO). Dried extracts were stored at -20°C and then reconstituted in 150μl assay buffer at the time of assay. Deslorelin concentrations were determined by EIA (#S1175; Bachem, Torrance, CA) according to the manufacturer’s instructions. Sensitivity was 0.014ng/ml and the intra-assay coefficient of variation 12.6%. Recovery of deslorelin added to untreated control rat plasma averaged 93%. Each hormone was analyzed in a single assay for that hormone and if a concentration was below the threshold of detection then the value for that sample was set to the assay sensitivity. As there was no significant effect of any treatment on the density of DAPI-labeled cells in the pituitary gland (data not shown), the number of cells immunoreactive for FSHβ-, LHβ-, or αGSU-ir in a field was normalized to a percentage of total cells. This calculated percentage underestimates the true percentage of secretory cells as DAPI labeling does not discriminate between endocrine and non-endocrine (e.g. folliculostellate) cells. The number of mono-hormonal gonadotropes was calculated by subtracting bi-hormonal cells from the total of FSHβ-ir or LHβ-ir cells, respectively. A second level analysis was performed to determine the effect of treatment on the proportion of gonadotropes that were mono-FSHβ, mono-LHβ or bi-hormonal. The total gonadotrope population was estimated by adding the mono-hormonal FSHβ-ir, mono-hormonal LHβ-ir and bi-hormonal cells. Untreated intact (Control) rats in Experiment 1 were used in the statistical analyses of both experiments. Data were analyzed by ANOVA and Tukey’s post hoc test to compare groups. Plasma values were log transformed before ANOVA.