Testosterone Effects on Spermatogenesis in the Gonadotropin-Releasing Hormone-Immunized Rat'

Transcription

1 BIOLOGY OF REPRODUCTION 5, (1994) Testosterone Effects on Spermatogenesis in the Gonadotropin-Releasing Hormone-Immunized Rat' ROBERT I. MCLACHLAN, 2'3 NIGEL G. WREFORD, 4 5 CON TSONIS, 6 DAVID M. DE KRETSER, 4 and DAVID M. ROBERTSON 3 Prince Henry's Institute of Medical Research, 3 Monash Medical Centre, and Institute of Reproduction and Development 4 and Department of Anatomy,5 Monash University, Clayton, Victoria, 3168, Australia Biotech Australia Pty Ltd, 6 Roseville, New South Wales, 263, Australia ABSTRACT Active immunization of adult rats with a GnRH fusion protein was used to inhibit gonadotropin secretion and to establish an in vivo model for studying the hormonal control of spermatogenesis. The model was characterized in terms of the efficacy of the immunogen as well as the time course and nature of the immunological and biological responses. To study the reinitiation of spermatogenesis, testosterone (T) was chosen for this initial study as it is known to restore spermatogenesis in gonadotropindeficient rats. Adult Sprague-Dawley rats were actively immunized with a proprietary GnRH immunogen (BA-11, 1 #Lg s.c.) every 4 wk. After 12 wk, 58% of animals showed markedly suppressed testicular size, as assessed by scrotal palpation, and were classed as responders. Serum LH, FSH, and T as well as the testicular elongated spermatid count (ESC) and epididymal sperm were undetectable in responding animals. Marked reductions in testicular (29% of control), prostatic (8% of control), and epididymal (32% of control) weights were seen. Spermatogenesis was severely disrupted with no evidence of progression beyond round spermatids. To study the action of T in GnRH-immunized animals, T (defined by lengths of s.c. silastic implant, T3-T24 cm) was given to responding animals. Animals were killed 2, 8, and 12 wk after T24 administration. In response to T24, serum T levels increased to 4 times control levels, serum FSH levels were restored to 65% of control levels by 2 wk, and serum LH remained undetectable. Testicular weight increased to 8% of control levels at 12 wk (p <.5 vs. control). Epididymal and prostatic weights were normalized by T. ESC increased to 82% of control values at 12 wk (11 ± 1 vs. control 134 ± 8 million/testis, p =.1). Spermatogenesis was histologically normal after 8 wk of T24 treatment. To study the time course and dose response of T action, animals were immunized with another GnRH immunogen (BA-17), which yielded an 87% response rate at 12 wk. Testicular weight increased by Day 5 of T24 treatment, and a clear dose-response effect was apparent. The restoration of ESC was delayed compared to that of testicular weight (no restoration at 2 wk) and required T6 treatment. Rats immunized for 2 wk and then given T24 treatment showed a similar pattern of restoration in testicular weight and ESC. Serum FSH was normalized by Day 2 of T treatment by doses - 3 cm. We conclude, first, that the GnRH immunization procedure using the BA-17 immunogen, by withdrawing gonadotropins, results in evidence of severe regression of spermatogenesis in -9% of rats within 12 wk and provides a convenient and highly efficient in vivo model for the study of spermatogenesis. The ability of the regressed testis to respond to T even after many weeks of profound regression indicates that a stable, albeit severely reduced, germ cell population is maintained and will respond to hormonal stimuli. Such a model of gonadotropin deficiency is ideal for the study of the action of hormones potentially regulating spermatogenesis. Our second conclusion is that T substantially, but not quantitatively, restores spermatogenesis either by the partial restoration of testicular T levels and/or by the concomitant restoration of serum FSH levels via stimulation of FSH secretion by GnRH-independent mechanisms. INTRODUCTION The relative contributions of FSH and testosterone (T) to the control of spermatogenesis remain controversial. In humans, both hormones are apparently required for quantitatively normal spermatogenesis [1]. On the other hand, there is evidence that passive immunization of rats against FSH does not interrupt spermatogenesis [2]. Gonadotropin deficiency can be induced by GnRH immunization [3-5], and in this setting T alone has been reported to restore spermatogenesis [3, 4]. Accepted September 2, Received May 26, 'Study supported by the National Health and Medical Research Council of Australia. R I. McLachlan is an R Douglas Wright Fellow of the Australian NH&MRC. 2 Correspondence: Dr. Robert I. McLachlan, Prince Henry's Institute of Medical Research, Monash Medical Centre, 246 Clayton Road, Clayton, 3168, Australia. Progress has been hampered by the lack of in vivo models that enable the specific manipulation of candidate regulators of spermatogenesis without affecting other hormonal systems. In vivo studies using hypophysectomy to examine the effects of FSH and T [6-11] are subject to the criticism that there is major disruption of other hormonal and metabolic systems of potential importance in spermatogenesis, e.g., thyroid-stimulating hormone (TSH) [12] and prolactin [13, 14]. Steroid treatment (low doses of T and estradiol [E]), while suppressing LH and T levels, only partially suppresses [9] or fails to affect serum FSH in rats [15, 16]. Blockade of GnRH action in vivo has been achieved by GnRH antagonist administration [17,18] and by active immunization with GnRH [3-5], these approaches inducing marked testicular regression. The administration of high doses of T to GnRH-immunized animals has been reported to maintain spermato- 271

2 272 MCLACHLAN ET AL. genesis quantitatively in the absence of measurable serum FSH and LH levels [3, 4]. In the present study we aimed to characterize an in vivo model of spermatogenic suppression in adult rats by active immunization against GnRH and then to apply this model in the evaluation of the hormonal control of spermatogenesis. In this report, the effects of T in this model are described. Animals MATERIALS AND METHODS Adult (9-11 days old, g) Sprague-Dawley rats were obtained from Monash Central Animal House and maintained at 22 C in a fixed 12L: 12D cycle with free access to food and water. The study was approved by Monash University Animal Ethics Committee. Animals were killed by decapitation, blood was allowed to clot at room temperature, and serum was collected after overnight storage at 4C. Testosterone Implants T implants were prepared using polydimethylsiloxane tubing (o.d mm; i.d mm; Dow-Corning, Midland, MI) and medical adhesive silicon Type A (Dow-Corning), as previously described [9]. T implants of varying length (T3-T24 cm) were implanted subcutaneously along the animals' dorsal surface. Elongated Spermatid Count The right testis was weighed and decapsulated into 2-4 ml.1 M phosphate buffer (ph 7.4) containing.154 M NaCl (PBS),.5% Triton-X-1, and.1% sodium azide and then homogenized for 3 sec in a tissue disperser (Polytron Kinematica, Lucerne, Switzerland). An aliquot was counted for elongated spermatids in a hemocytometer and the elongated spermatid count (ESC) per testis calculated [9]. The epididymides were dissected, weighed, finely minced prior to homogenization in the same buffer, and then counted in a hemocytometer. The prostates were also excised and weighed. Histology The left testis was fixed by immersion in Bouin's solution. After 1-2 h, the tunica albuginea was cut adjacent to the poles to give better fixation. After fixation for 24 h, the testis was placed in 7% alcohol prior to further processing. Testes were subsequently halved by being divided at right angles to the long axis, processed into paraffin, sectioned at 2 plm, and stained with hematoxylin and eosin. Sections were photographed using Agfa Copex Pan Rapid film via a Leitz photomicroscope with a 2x objective. The minimum diameter of tubule profile was measured as the tubule diameter. Hormone Assays Two rat FSH RIAs were performed using iodinated rat FSH (NIDDK rfsh 17 [Baltimore, MD]) as tracer and either an NIDDK rfsh antiserum (S11) or an in-house ovine human FSH antiserum (59/16) and NIDDK rfsh RP-2 as standard. All samples from an individual experiment were run in a single assay, and the within-assay coefficients of variation were 6.8% and 6.7%, respectively. The sensitivities of the assays were 1.8 ng/ml and 1.3 ng/ml, respectively. Data in Results were obtained through use of the S11 antiserum. The assay using 59/16 antiserum showed <1% crossreactivity with purified rat LH. A comparison of results from 4 selected samples analyzed by both assays yielded a correlation coefficient of r =.9 (data not shown). The LH RIA used iodinated rlh (NIDDK rlh 17) as tracer, NIDDK rlh S9 antiserum, and NIDDK rlh RP-2 as standard. All samples from an individual experiment were run in a single assay, and the within-assay coefficient of variation was 6.5%. The sensitivity of the assay was.14 ng/ml. The detection level of the gonadotropin assays was determined by the level of immunoreactivity in serum from hypophysectomized rats. Serum T was measured by RIA after extraction in hexane/chloroform as described previously [19]. The withinand between-assay coefficients of variation were 7.8 and 21.6%, respectively, and the sensitivity of the assay was.44 ng/ml. In the first immunization program, the T level in the testicular homogenates used for ESC determination could not be ascertained due to interference by Triton-X-1 in the RIA. In the second immunization program, an aliquot of the testicular homogenate was taken prior to the addition of Triton and was used, without extraction, in the RIA. All samples were run in the same assay and the sensitivity was.8 ng/g testis. Statistics Data are presented as the mean + standard deviation. Peritz's multiple range test [2] or Student's t-test was used to establish differences between treatment groups. Significance was taken to be p <.5, NS = not significant. GnRH Immunization Immunogen. Proprietary GnRH immunogen preparations (BA-11, BA-17), obtained from Biotech Australia Pty Ltd., (Roseville, Australia), incorporated an adjuvant free of mycobacterial components [21]. These immunogens were used in two immunization programs based on data in mice [21]. In both programs the GnRH immunogen was administered at a dose of 1 Lg protein at a single s.c. site every 4 wk until completion of the study. Control animals received injections of adjuvant only. Immunological Response. The antibody response to GnRH immunization using the first immunogen (BA-11) was assessed on the basis of the ability of serum both to bind

3 EFFECT OF GnRH IMMUNIZATION ON RAT SPERMATOGENESIS FIG. 1. Changes in serum GnRH binding activity following GnRH immunization in animals that prior to Week 12 were unselected (solid circles) and in subsequently immunized animals showing testicular atrophy only (open circles). Hatched box represents period of T treatment. Results from animals not showing testicular atrophy (non responders) are not included. Nonimmunized controls, inverted solid triangles. radiolabeled GnRH and to neutralize the effects of GnRH in a pituitary cell culture system. The GnRH binding activity of serum from immunized rats was assessed using synthetic GnRH, iodinated by the chloramine T procedure, as tracer. Serum (1 Il, 1:5 in PBS buffer), tracer (1 cpm, 1 l PBS), and PBS (2 A1) were incubated overnight at 2 C. Control (preimmune) serum was used to assesss for nonspecific binding. Separation of bound from free tracer was based on the addition of ovine anti-rat IgG (1 I) as precipitating second antibody (2 h at 4 C) and then the addition of 6% polyethylene glycol (1 ml) followed by centrifugation (25 x g, 2 min). The capacity of serum samples to neutralize GnRH in vitro was assessed in a male rat pituitary cell culture system [22]. Briefly, pituitary cells from 9-day Sprague-Dawley rats were dispersed after collagenase/trypsin digestion and cultured in Dulbecco's Modified Eagle's Medium (DMEM) containing.5% BSA for 5 days prior to addition of samples. Serum (2 Il) was preincubated with bacitracin (1 U/ml) for 15 min at 2 C prior to addition to medium (4 1I) containing 1 nm GnRH (approximately the ED 7 dose for LH release). On the fifth day of culture, the medium was replaced with 1 pli of this GnRH/serum mixture (i.e., 5 1I original serum diluted to 1 1) and 3 l of warm DMEM and incubated for 4 h at 37 C. Medium was then collected and stored at -2 C prior to LH assay. Control wells contained GnRH alone or each serum sample in the absence of GnRH. The GnRH standard curve was assayed in the presence of normal rat serum (5 Rxl) containing bacitracin. LH was measured by RIA and LH release expressed as a percentage of that in basal (i.e., non-gnrh stimulated = 1%) wells. Biological response. The assessment of biological response to immunization was made after 12 wk, when testicular volumes were estimated by palpation under anesthesia. Response was defined as an estimated testicular volume of <.55 ml (vs. control, -1.8 ml). The accuracy of this estimate was confirmed by comparison with testicular volumes determined at laparotomy in 39 animals using the formula: length x width 2 x rr/6. Experimental Design Immunization program 1 (BA-11 immunogen). To examine the degree of testicular regression induced by GnRH immunization, serum hormone levels, reproductive organ weights, ESC, and light microscopy of the testis were assessed in rats immunized for 12 wk with either vehicle or immunogen. To examine the effects of T, responders were then divided into groups (n = 5-6 rats/group), treated with or without T24 implants, and killed 2, 8, and 12 wk later. Immunization program 2 (BA-17 immunogen). To examine the time and dose-response effects of T on spermatogenesis, animals were immunized every 4 wk and classified at 12 wk as responders as described above. Animals (n = 5-6/group) were then implanted with 3-, 6-, 1-, and 24-cm T implants; they were killed 2, 5, and 1 days (T24 only) and 2, 8, and 16 wk (all T doses) later for measurement of serum gonadotropins, testis weight, and ESC. RESULTS Characterization of the GnRH-lmmunized Model Immunological response to immunization. The level of iodinated GnRH binding in serum increased with each successive injection of the BA-11 immunogen (Fig. 1). Control and immunized responding animals were examined to determine the relationship between either iodinated GnRH binding in serum or the ability of serum to neutralize GnRH action on gonadotropin secretion by pituitary cells in culture, and testicular volume (as an index of gonadotropin withdrawal). In immunized non-t-treated animals, killed between 12 and 24 wk, iodinated GnRH binding in serum was % (n = 17), while in control animals, binding activity was < 2.5%. Serum from immunized animals suppressed GnRH-induced LH release in vitro. With the basal release of LH in non-gnrh-treated wells defined as 1%, GnRH alone increased LH release over the course of incubation to 335 t 36% (n = 18) of basal release. Co-incubation of GnRH with immune serum impaired GnRH action with LH release being only % (n = 17) of

4 274 MCLACHLAN ET AL. GnRH BINDING 6* ( ) GNRH BINDING UI-! A u v : S (b)... D - '- ~ VIT II UTIKAAJT6AiU CONTROL 4 a *1 a..71 'a M. '.2' Eo :: VCL a. C e 1.2 ' e 3 2 oi * I S o *. * FIG. 3. Changes in body and reproductive organ weights in response to GnRH immunization and testosterone administration (hatched box). At 12 wk, animals were classified as responders on the basis of evidence of testicular atrophy (see Materials and Methods) and either continued as immunized controls (solid circles) or received testosterone (24-cm implants, open circles) for 12 wk. All immunized animals prior to Week 12 were unselected. Data expressed as mean SD, n = 5-6/group. Nonimmunized control, inverted solid triangles. Prostatic tissue not available at 24-wk time point. 1 - o CONTROL GROUP REGRESSED PARTIAL NON REGRESSED GNRH IMMUNIZED GROUPS FIG. 2. a) GnRH binding activity in serum and (b) in vitro neutralization of GnRH by serum from nonimmunized control animals and GnRHimmunized animals that failed to show testicular regression at 12 wk but were killed at 24 wk. Animals were then classified as showing testicular regression (testis volumes <.55 ml, n = 14), partial regression (testis volumes.82,.84 ml), or no regression (testicular volumes > 1.38 ml, n = 7). Data are expressed as a percentage of basal LH release in control wells (see Materials and Methods). Dotted line, basal LH release; broken line, mean GnRH-induced LH release in the presence of control serum. Individual data and mean SD are shown. **p <.1 compared to the nonregressed group. basal. Serum from nonimmunized control animals showed no neutralizing activity. Animals that had failed to respond after 12 wk were immunized for a further 12 wk. In these 23 animals, 14 responded (giving an overall response rate by 24 wk of 85%), 2 showed a partial response (testicular volumes.82,.84 ml), and 7 showed little or no regression (testicular volumes > 1.38 ml). Serum GnRH binding and in vitro neutralizing activities were significantly higher in animals showing testicular regression (Fig. 2). In the nonresponding group (n = 7), no neutralizing activity was detected although GnRH binding was seen in all samples. Overall these data suggest that GnRH neutralizing activity correlates more closely than GnRH binding with testicular regression. Biological response to GnRH immunization. The effect of GnRH immunization in producing a biological response was assessed by the degree of testicular atrophy. After 12 wk of immunization, 58% (35 of 6 animals) and 87% (216 of 249 animals) of animals showed testicular atrophy in the programs using the BA-11 and BA-17 immunogens, respectively. After GnRH immunization with the BA- 11 immunogen, body and reproductive organ weights were unchanged until Week 1, when some animals showed major reductions in testis, epididymal, and prostatic weight (Fig. 3, a-d). At 12 wk, animals judged to have responded to immunization (testicular weight g vs g, Fig. 3b) showed evidence of T withdrawal, including the reduction in epididymal weight to 32% of the control value (.2 ±.3 vs g,p <.1, Fig. 3c) and in prostatic weight to 8% of the control value ( vs g,p <.1, Fig. 3d). The ESC and epididymal sperm content (Fig. 4) and serum T, FSH, and LH levels (Fig. 5) were undetectable in responding animals at 12 wk.

5 EFFECT OF GnRH IMMUNIZATION ON RAT SPERMATOGENESIS 275 (a) 2 (a) 15 C T12 9 i 6 33 ().6 (b) 25 2 rc.3 no c 15 E FIG. 4. Changes in (a) testicular elongated spermatid count and (b) epididymal sperm content following GnRH immunization and testosterone treatment (hatched box). Data are shown from the eighth week of immunization onwards. At 12 wk, animals were classified as responders on the basis of evidence of testicular atrophy (see Materials and Methods) and either continued as immunized controls (solid circles) or received testosterone (24-cm implants, open circles) for 12 wk. Data presented as mean - SD, n = 5-6/group, ***p =.1 compared with controls, inverted solid triangles. Histological examination showed that spermatogenesis was severely disrupted by GnRH immunization (Fig. 6). At 12 wk, seminiferous tubule diameter was reduced from im to 129 ± 12 jim, and the intertubular spaces contained atrophic Leydig cells. There was no evidence of progression of spermatogenesis beyond the round spermatid stage, and degenerating round spermatids were observed in the lumen. After 2 wk of immunization (Fig. 6b), a similar pattern was maintained (tubule diameter im), although degenerating pachytene spermatocytes were also observed. Effects of Testosterone Following GnRH Immunization In response to T24 treatment, testicular weight increased from 29% to 46% of values for nonimmunized controls at 2 wk and to 8% of control values by 12 wk, but remained. (C) FIG. 5. Changes in (a) serum testosterone, (b) serum LH, and (c) serum FSH levels following GnRH immunization and testosterone administration (hatched box). At 12 wk, animals were classified as responders on the basis of evidence of testicular atrophy (see Materials and Methods) and either continued as immunized controls (solid circles) or received testosterone (24- cm implants, open circles) for 12 wk. Data expressed as mean + SD, n = 5-6/group. Nonimmunized controls, inverted solid triangles. significantly less than values for nonimmunized controls ( g vs ±.22 g, p <.5, Fig. 3b). Epididymal and prostatic weights returned to nonimmunized control values after 12 and 2 wk of T treatment, respectively (Fig. 3, c and d). Body weight was significantly reduced after 12 wk in both T-treated animals and immunized controls (8% and 74% of values for nonimmunized controls, respectively, p <.1) (Fig. 3a). The ESC rose progressively during T treatment, to 13, 54, and 82% of nonimmunized control levels at 2, 8, and 12 wk of T24 treatment, respectively, (Fig. 4a); however, the 12-wk ESC remaining significantly reduced ( vs million, p =.1). ESC

6 276 MCLACHLAN ET AL. FIG. 6. Testicular sections from (a) control animals showing normal spermatogenesis; (b) animals immunized against GnRH for 2 wk, showing regressed tubules and absence of spermiogenesis; (c) animals immunized against GnRH for 12 wk followed by 2 wk of testosterone treatment (24-cm implants), showing partial restoration of spermiogenesis; (d) animals immunized against GnRH for 12 wk followed by 8 wk of testosterone treatment (24-cm implants), showing qualitative restoration of spermatogenesis and tubule diameter. Magnification x 2, scale bar is 5 ium. when expressed per gram of testicular tissue was, however, normalized in T-treated animals (data not shown). Epididymal sperm content remained undetectable at 2 wk but returned to control values by 8 wk of T24 treatment (Fig. 4b). Serum T levels remained undetectable in GnRH-immunized animals while T24 treatment increased levels to 11.9 ng/ml compared with 2.5 ng/ml in nonimmunized controls (Fig. 5a). Serum LH remained undetectable in GnRHimmunized animals, with or without T treatment (Fig. 5b). Serum FSH levels were partially restored by T24 treatment at Week 2 and were at control levels at 8 and 12 wk of treatment (Fig. 5c). In response to T24 implants, elongating spermatids were apparent by 14 days (Fig. 6c) and tubule diameter increased to 183 ± 6 plm. By 8 wk of T24 treatment, a qualitative restoration of spermatogenesis had occurred, but tubule diameter remained significantly below control values (257 ± 3 vs im, p <.5). Testosterone Time and Dose-Response Effects Following testicular regression induced by immunization with the BA-17 immunogen, the effect of various doses of T on testis weight and ESC is shown in Figure 7. Testis weight increased by Day 1 in response to T24, and in all T-treated animals by 2 wk, with a clear dose-response effect of T being seen at 8 and 16 wk of T treatment. In comparison to restoration of testicular weight, the restoration of ESC was delayed and the T dose-response effect was less

7 EFFECT OF GnRH IMMUNIZATION ON RAT SPERMATOGENESIS ' U) * 1.- T. I Control T V ized controls. Serum T levels were in the control range with T3 and were increased 4.5-fold by T24 implants. Serum LH levels were undetectable throughout T treatment (data not shown). Testicular T levels in T3-treated animals were not significantly higher than in GnRH-immunized animals (3.5 ±.6 vs ng/g testis), while testicular T levels were partially restored by T1 and T24 treatment to and ng/g testis (13% and 27% of control, ng/g testis)..5 -., v = 2 L a t Control 16 2 FIG. 7. Time course of T action on testicular weight (top panel) and elongated spermatid count (ESC, bottom panel) in GnRH-immunized rats. Data shown are from the T-induced recovery phase associated with T3-, T6-, T1-, and T24-cm implants. All immunized animals had been selected previously as "responders" to 12 wk of GnRH immunization based on the degree of testicular regression, as described in Materials and Methods. Animals continued to be immunized every 4 wk throughout T treatment. A separate group was immunized for a total of 2 wk (corresponding to the 8-wk point above prior to receiving T24 treatment. Data are mean SD, n = 5-6/group. clear. With the T3 dose, the ESC remained suppressed despite a significant (p <.5) rise in testis weight. ESC was restored only by T doses T6 cm. Despite substantial restoration of testis weight and ESC to 92% and 75% of control values by T24 at 16 wk, both parameters remained significantly (p <.5) less than those in controls. In the group suppressed for 2 wk before T-induced recovery of spermatogenesis, the response of testis weight and ESC to T24 was similar to that seen in the immunized group for 12 wk prior to T treatment. Serum FSH levels were restored to normal by Day 5 of T24 treatment (Fig. 8). At T doses - T3, FSH levels at 2 wk were not significantly different from those in nonimmun- DISCUSSION This study presents a detailed characterization of an in vivo model system using GnRH immunization in the study of the hormonal control of spermatogenesis. Active immunization with the GnRH immunogen BA-17 produced undetectable gonadotropin and testosterone levels and profoundly suppressed testicular and accessory gland weight in -9% adult rats by 12 wk. Previous studies have shown that active immunization against GnRH can disrupt spermatogenesis [3, 5]. We have extended these studies by assessing additional parameters including the efficacy of immunization procedures, the time course of biological and immunological responses to immunization, and the utility of GnRH neutralization in vitro in following the immune response. Spermatogenesis was grossly disrupted by GnRH immunization: elongated spermatid numbers were less than 1% those of controls and qualitative histology showed a markedly reduced tubule diameter, depletion of spermatids, and some degeneration of pachytene spermatocytes. The ability of the regressed testis to respond to T even after many weeks of profound regression indicates that a stable, albeit severely reduced, germ cell population is maintained and will respond to hormonal stimuli. In contrast to hypophysectomy, the GnRH-immunized model of gonadotropin deficiency has the advantage of not disrupting other hormonal systems that may impact upon spermatogenesis. The GnRH-immunized rat model is therefore suited to study of the action of potential hormonal regulators of spermatogenesis in vivo. We chose to initially study the effects of T because it is well known to reinitiate spermatogenesis in a number of gonadotropin-deficient systems. Administration of T to normal adult rats, at doses giving slighty supra-physiological serum levels, suppresses LH secretion, thereby decreasing Leydig cell T secretion and causing intratesticular T levels to fall to 3-5% of control levels [3, 9,1, 16, 23]. As serum FSH levels are only partially suppressed, this model is predominantly one of LH/T deficiency. Originally we aimed to use the GnRH-immunized model to study the processes involved in the restoration of spermatogenesis by high-dose T in the absence of FSH and LH. Testicular T levels were 6.9% of control levels in the GnRH immunized animals; this result was comparable to that reported by Awoniyi et al. in a similar model (2 vs. 64 ng/ml, 3.1%, in seminiferous tu-

8 278 MCLACHLAN ET AL. C I == h) IL C V1 U-L a physiological serum T level, as is provided by a T3 implant, restored FSH levels in the gonadotropin-deficient animal. This effect was apparent by Day 5 of T treatment and persisted at all subsequent time points. There is extensive evidence that rat pituitary cells in culture continue to release FSH for long periods of time (up to 28 days) in the absense of GnRH, while LH secretion is rapidly lost in this setting [24]. Such data indicate that gonadotropes in vitro release FSH in a constituitive nongnrh-dependent fashion. Reports on the extent to which this GnRH-independent release occurs in vivo are inconsistent, some reports showing undetectable serum FSH levels (this study; [3,4,17,18]) and others showing only partial suppression [25,26]. On the other hand, a direct GnRH-independent action of T to stimulate pituitary FSH secretion has been found in many previous itudie unin ripnt in vitrn nititarv cll crlture- [27--1 A recent study [31] using a rat pituitary cell perifusion sys- T RESPONSE (CM) tem suggested a dual action of T to cause a monotrophic (b) increase in basal FSH secretion in the absence of GnRH while attenuating the GnRH-induced release of FSH in the presence of pulsatile GnRH. Finally, several studies have shown an action of T to increase serum FSH levels in GnRH antagonist-treated rats with otherwise undetectable gonadotropin levels [17,18]. The mechanisms by which andro- Control gens induce this monotrophic rise in FSH may involve the {?a~ stimulation ]t~~ of FSH3 subunit mrna levels [31, 32]. T treat- T24 _ ~ment presumably permits the increased release of FSH in vivo by mechanisms similar to those operating in vitro. In addition, as GnRH-immunized animals have grossly im- +, * paired spermatogenesis, it is also possible that a reduction immune in inhibin secretion may also contribute to the non-gnrh- Control dependent release of FSH. In contrast to our findings, in previous studies using a GnRH-immunized model to study T action on rat spermatogenesis, FSH levels remained un- ~~~~ 8 4 l~2 l detectable during T treatment [3,4]. We are unable to ex ~2 plain the difference with our results, although the immu- nogen used may have resulted in antibodies that colocalized FIG. a) Dose 8. response of T on serum FSH levels after 2 wk of T with GnR on the gonadotropes and mediated cytotoxicity. treatment (O)) of GnRH-immunized animals. b) The time course of T24-in- The mechanism of the T-induced restoration of sper- of serum FSH in GnRH-immunized rats. GnRH-immu- matogenesis is therefore unclear, as T may be acting di- duced restoration nized controlis, re solid mean circles. + Nonimmunized SD, controls, n inverted = 5-6/group. solid trian- rectly on the testis and/or by a concomitant restoration of gles. Dat ar e mean 5-6/group. ± SD, n FSH secretion, presumably by a direct GnRH-independent action on the pituitary. FSH would then promote the resic). n turn these T levels are similar to those covery of spermatogenesis by the stimulation of Sertoli cell bule fluid reported iri hypophysectoized rats (2-5% control [7,9,1). function. For example, FSH may enhance T action by prosuggest that low levels of androgen biosynthesis moting the accumulation of androgen-binding protein in These dat persist in the functional absence of gonadotropins. None- Sertoli cells [33 or may indirectly enhance Leydig cell T theless, the profound effect of this degree of T withdrawal secretion by paracrine factors of Sertoli cell origin [34]. The on testicul lar function in the GnRH-immunized model is ap- stimulation of FSH secretion by T in rodents limits the useparent fro*m the range of parameters described. fulness of these species in the identification of specific roles Unfort inately, with respect to use of this model to dis- for T and FSH in the control of spermatogenesis via stratic actions of T on spermatogenesis, we have shown egies that only block GnRH action, such as GnRH antagonist cern speci that T trea tment is associated with restoration of FSH levels, treatment or GnR immunization. this findin.g having been confirmed in two RA systems. The In response to T treatment, testicular weight increased dose-resp onse curve for T stimulation of FSH showed that prior to the ESC such that after 2 wk, ESC remained un-

9 EFFECT OF GnRH IMMUNIZATION ON RAT SPERMATOGENESIS 279 detectable despite a doubling of testicular weight. We attribute this increase in testicular weight to enhanced fluid production and/or to an increase in germ cell types prior to the acquisition of Triton resistance (-step 15 of spermiogenesis). The observation that elongated spermatids did not appear at 2 wk of T treatment suggests that few round spermatids were present at the onset of restoration and that some synchronization of the testis may occur after prolonged gonadotropin withdrawal. Quantitative histological studies are needed to clarify these issues. The increase in testicular weight seen with T3 implants occurred with serum T levels within the physiological range. As the ESC remained suppressed at this T dose, the data suggest that there is a differential threshold for T action on the testis with a higher level of T being required for the stimulation of spermatogenesis than for fluid production. Testicular weight and ESC were substantially restored by T but remained significantly below control levels even after wk (equivalent to seminiferous cycles). Nonetheless, spermatogenesis was normalized in terms of both the ESC/g of testis and qualitative histology. Whether the failure to quantitatively restore testis weight or ESC is due to partial deficiencies of testicular T and/or serum FSH levels, or to an irreversible effect of prolonged gonadotropin deficiency, is unclear. Importantly, T was equally effective in restoring testicular function in animals subjected to an additional 8 wk of gonadotropin withdrawal prior to highdose T, suggesting that the seminiferous epithelium maintains a stable, albeit severely reduced, germ cell population that is capable of briskly,responding to hormonal stimulation. Although the restoration of FSH levels may indirectly promote the function of the Leydig cell via stimulation of paracrine factors from the Sertoli cell [34], it seems likely that Leydig cell factors other than T remain deficient in GnRHimmunized animals perhaps accounting for the failure to quantitatively restore testicular weight. We have previously reported that sperm production was markedly reduced after LH and testicular T withdrawal induced by physiological doses of T in adult rats while at high doses of T, spermatogenesis was maintained at essentially normal levels [9, 23]. Stereological analysis suggested that T acts to promote maturation of round to elongated spermatids [23]. Serum FSH and intratesticular T levels in these T-treated animals were similar to those in GnRH-immunized animals receiving high-dose T in the current study. Although the initial degree of spermatogenic suppression was more severe in the GnRH-immunized animals, after a prolonged period of T-induced recovery (i.e., one cycle, 8 wk), the two models have essentially similar gonadotropin levels and spermatogenic development. Neither system allows the study of T action in the absence of FSH. In summary, we have characterised a GnRH-immunized model that is convenient and is effective in profoundly suppressing gonadotropin levels and testicular function. Although severe testicular regression is produced, restoration of spermatogenesis can be induced, even after many weeks. The model is ideal for the study of the hormonal control of spermatogenesis in vivo as it allows for the assessment of the action of candidate regulators in the absence of gonadotropins, provided there is no direct action of the hormones on the pitiutary such as that observed here with testosterone or such as that which may occur with proteins such as activin [35]. In this initial study we have shown that T treatment restores serum FSH levels and therefore that the recovery of spermatogenesis may be due to the independent or synergistic actions of either hormone. Additional strategies, e.g., immunization against both GnRH and FSH, are required in order to define any unique actions of T on spermatogenesis in rats. ACKNOWLEDGMENTS We thank Ms Sarah Meacham, Liza O'Donnell, and Anne O'Connor for their excellent technical assistance. We thank the NIDDK for the provision of the gonadotropin assay reagents. REFERENCES 1. Matsumoto AM, Karpas AE, Bremner WJ. Chronic human chorionic gonadotrophin administration in normal men: evidence that follicle-stimulating hormone is necessary for the maintenance of quantitatively normal spermatogenesis in man. J Clin Endocrinol & Metab 1986; 62: Dym M, Raj HGM, Lin YC, Chemes HE, Kotite NZ, Nayfeh SN, French FS. Is FSH required for maintenance of spermatogenesis in adult rats? J Reprod Fertil Suppl 1979; 26: Awoniyi CA, Santulli R, Chandrashekar V, Schanbacher BC, Zirkin BR. Quantitative restoration of advanced spermatogenic cells in adult male rats made azoospermic by active immunization against luteinizing hormone or gonadotrophinreleasing hormone. Endocrinology 1989; 125: Awoniyi CA, Zirkin BR, Chandrashekar V, Schlaff WD. Exogenously administered testosterone maintains spermatogenesis quantitatively in adult rats actively immunized against gonadotrophin-releasing hormone. Endocrinology 1992; 13: Rovan E, Fiebiger E, Kalla NR, Talwar GP, Aulitzky W, Frick J. Effect of active immunization to luteinizing-hormone-releasing hormone on the fertility and histoarchitecture of the reproductive organs of male rats. Urol Res 1992; 2: Elkington JSH, Blackshaw AW. Studies in testicular function. I. Quantitative effects of FSH, LH, testosterone and dihydrotestosterone on restoration and maintenance of spermatogenesis in the hypophysectomized rat. Aust J Biol Sci 1974: 27: Bartlett JMS, Weinbauer GF, Nieschlag E. Differential effects of FSH and testosterone on the maintenance of spermatogenesis in the adult hypophysectomized rat. J Endocrinol 1989; 121: Santulli R, Sprando RL, Awoniyi CA, Ewing LL, Zirkin BR To what extent can spermatogenesis be maintained in the hypophysectomized adult rat testis with exogenously administered testosterone? Endocrinology 199; 126: Sun Y-T, Irby DC, Robertson DM, de Kretser DM. The effects of exogenously administered testosterone on spermatogenesis in intact and hypophysectomized rats. Endocrinology 1989; 125: Awoniyi CA, Sprando RL, Santulli R, Chandrashakar V, Ewing LL, Zirkin BR. Restoration of spermatogenesis by exogenously administered testosterone in rats made azoospermic by hypophysectomy or withdrawal of luteinizing hormone alone. Endocrinology 199; 127: Huang HFS, Marshall GR, Rosenberg R, Nieschlag E. Restoration of spermatogenesis by high levels of testosterone in hypophysectomized rats after long-term regression. Acta Endocrinol Copenh 116: Ando S, Panno ML, Beraldi E, Tarantino G, Salerno M, Palermo S. Prati M, Fugassa E. Influence of hypothyroidism on testicular steroidogenesis in adult rats. Exp Clin Endocrinol 199; 96: Dombrowicz D, Sente B, Closset J, Hennen G. Dose-dependent effects of human prolactin on the immature hypophysectomized rat testis Endocrinology 1992; 13:695-7.

10 28 MCLACHLAN ET AL. 14. Takase M, Tsutsui K, Kawashima S. Effects of PRL and FSH on LH binding and number of Leydig cells in hypophysectomized mice Endocrinol Jpn 199; 37: Awoniyi CA, Santulli R, Sprando RL, Ewing LL, Zirkin BR Restoration of advanced spermatogenic cells in the experimentally regressed rat testis: quantitative relationship to testosterone concentration within the testis. Endocrinology 1989; 124: Zirkin BR, Santulli R, Awoniyi CA, Ewing LL. Maintenance of advanced spermatogenic cells in the adult rat testis: quantitative relationship to testosterone concentration within the testis, Endocrinology 1989; 124: Rea MA, Marshall GR, Weinbauer GF, Nieschlag E. Testosterone maintains pituitary and serum FSH and spermatogenesis in gonadotrophin-releasing hormone antagonist-suppressed rats. J Endocrinol 1986; 18: Bhasin S, Fielder UT, Swerdloff RS. Testosterone selectively increases serum follicle-stimulating hormone (FSH) but not luteinizing hormone (LH) in gonadotrophin-releasing hormone antagonist-treated male rats: evidence for differential regulation of LH and FSH secretion. Biol Reprod 1987; 37: Risbridger GP, Kerr JB, Peake RA, de Kretser DM. An assessment of Leydig cell function after bilateral or unilateral efferent duct ligation: further evidence for local control of Leydig cell function. Endocrinology 1981; 19: Harper JF. Peritz F test: basic program of a robust multiple comparison test for statistical analysis of all differences among group means. Comp Biol Med 1984; 14: Stewart A, Valentine J, Tsonis C, Borchers C, Russell-Jones G, Headon D, Worral M, Clifton G, McInerney B, Forage R. Immunization using recombinant TraT- LHRH fusion proteins. In: Brown F, Chaneck RM, Ginsberg H, Lerner RD (eds.), Vaccines 92. Cold Spring Harbor, NY: Cold Spring Harbor Lab Press; 1992: Scott RS, Burger HG, Quigg H. A simple and rapid in vitro bioassay for inhibin. Endocrinology 198; 17: Sun Y-T, Wreford NG, Robertson DM, de Kretser DM. Quantitative cytological studies of spermatogenesis in intact and hypophysectomized rats: identification of androgen-dependent stages. Endocrinology 199; 127: Farnworth PG, Robertson DM, dekretser DM, Burger HG. Effects of 31 kilodalton inhibin on follicle-stimulating hormone and luteinizing hormone in rat pituitary cells in vitro: actions under basal conditions Endocrinology 1988; 122: Hasegawa Y, Miyamoto K, Yazaki C, Igarashi M. Regulation of the second surge of follicle-stimulating hormone; effects of antiluteinizing hormone-releasing hormone serum and phenobarbitone Endocrinology 1981; 19: Grady RR, Shin L, Charlesworth MC, Cohenbecker IR, Smith M, Rivier C, Rivier J, Vale W, Schwartz NB. Differential suppression of FSH and LH in vivo by a GnRH antagonist. Neuroendocrinology 1985; 4: Drouin J, Labrie F. Selective effect of androgens on LH and FSH release in anterior pituitary cells in culture. Endocrinology 1976; 98: Denef C, Hautekeete E, Dewals R, de Wolf A Differential control of luteinizing hormone and follicle-stimulating hormone secretion by androgens in rat pituitary cells in culture: functional diversity of subpopulations separated by unit gravity sedimentation. Endocrinology 198; 16: Leveque NW, Grotjan HE Jr. Interaction of progesterone with testosterone and dihydrotestosterone on follicle-stimulating hormone release by cultures of rat anterior pituitary cells. Biol Reprod 1982; 27: Kitahara S, Winters SJ, Oshima H, Troen P. Effects of gonadal steroids on folliclestimulating hormone and luteinizing hormone secretion by pituitary cells from castrated and intact male rats. Biol Reprod 1991; 44: Winters SJ, Ishizaka K, Kitahara S, Troen P, Attardi B. Effects of testosterone on gonadotrophin subunit messenger ribonucleic acids in the presence or absence of gonadotrophin-releasing hormone. Endocrinology 1992; 13: Wierman ME, Wang C. Androgen selectively stimulates follicle-stimulating hormone-3 mrna levels after gonadotrophin-releasing hormone antagonist administration. Biol Reprod 199; 42: Huang HF, Pogach LM, Nathan E, Giglio W, Seebode l. Synergistic effects of follicle-stimulating hormone and testosterone on the maintenance of spermiogenesis in hypophysectomized rats: relationship with the androgen-binding protein status. Endocrinology 1991; 128: Verhoeven G, Cailleau J. A factor in spent media from Sertoli cell-enriched cultures that stimulates steroidogenesis in Leydig cells Mol Cell Endocrinol 1985; 4: Mather J, Attie K, Woodruff T, Rice G, Phillips D. Activin stimulates spermatogonial proliferation in germ-sertoli cell cocultures from immature rats. Endocrinology 199; 127: