Hypogonadotropic hypogonadism in nephrotic rats: Increased sensitivity to negative feedback effects of testosterone

Hypogonadotropic Hypogonadism in Nephrotic Rats: Increased Sensitivity to Negative Feedback Effects of Testosterone Allan R. Glass, Judith Beach, and Robert A. Vigersky Pituitary-testicular function was examined in adult male rats with aminonucleoside-induced nephrotic syndrome as a model for similar disease in humans. Nephrotic rats developed androgen deficiency, as manifested by decreased prostate and seminal vesicle weights, lower serum total and free testosterone levels, and reduced testosterone release from testes incubated in vitro. Despite hypoandrogenism, the weight and histologic appearance of the testes (light microscopy) were not affected in nephrotic rats. This androgen deficiency seemed to be a consequence of decreased gonadotropin output rather than primary testicular failure, since both pituitary gonadotropin content and serum gonadotropin levels (basally and after luteinizing hormone releasing factor: LHRH) were reduced in nephrotic rats. In addition, the percentage increase in testosterone release by testes incubated in vitro after addition of exogenous gonadotropin was similar in nephrotic and control groups. However, gonadotropin output in nephrotic rats was not impaired in the absence of testis, since no reduction was seen in either post-castration serum gonadotropin levels in vivo or gonadotropin release from pituitaries incubated in vitro. This presumed inhibitory effect of the testis on gonadotropin output in nephrotic rats was confirmed directly by demonstrating an increased sensitivity to testosterone-mediated suppression of gonadotropins in castrate animals in vivo. The presence or absence of albumin also seemed to modulate the suppressive effect of testosterone on gonadotropin output from normal pituitaries incubated in vitro. We conclude that nephrotic male rats develop hypogonadotropic hypogonadism secondary to an increase in sensitivity of the pituitary to the negative feedback effects of testosterone.

R

ENAL DISEASE, particularly chronic renal failure, has significant effects on the functioning of several endocrine systems, including the pituitarytesticular axis.’ For example, chronic renal failure in men routinely leads to primary hypogonadism. Likewise, nephrotic syndrome without renal failure leads to a variety of changes in endocrine organs of which the thyroid is perhaps the most intensively studied.*,-’ However, there is no published information concerning the effect of untreated nephrotic syndrome without uremia on gonadal function in animals or humans, and to explore this area we carried out the current study of pituitary-testicular function in rats with chemicallyinduced nephrotic syndrome. The results indicate that nephrotic rats without renal failure develop hypogonadotropic hypogonadism, which is related to an increase in pituitary sensitivity to the negative feedback effects of testosterone. MATERIALS AND METHODS Male Charles

Sprague-Dawley derived River Labs, Wilmington,

CD rats were obtained Mass, either as adults

from or as

From the Endocrine-Metabolic Service, Departments of Medicine and Clinical Investigation, Walter Reed Army Medical Center. Wash, DC, and from the Department of Medicine, Uniformed Services University of Health Sciences, Bethesda, Md. The opinions and assertions contained herein are the private views of the authors and are not to be construed as o$icial or representing the views of the Department of the Army or the Depariment of Defense. Address reprint requests to Allan R. Glass. MD, EndocrineMetabolic Service (70). Walter Reed Army Medical Center, WA, DC 20307-5001. This article is a US government work. There are no restrictions on its use.

574

I-day-old pups in litters. Beginning at age 21 days (weaning). all animals were given water and standard laboratory rat chow (Ralston Purina, St. Louis) ad libitum and were placed in a 14/10 light-dark environment. Nephrotic syndrome was induced by intravenous (IV) injection of 25 mg puromycin aminonucleoside (ICN Pharmaceutical, Cleveland) dissolved in saline into the surgically exposed femoral vein under ether anesthesia.4 Animals were killed I4 days after aminonucleoside injection, corresponding to the age of 98 days and age of 101 days in two separate studies. Some animals were given LHRH (30 or 300 ng/lOO g body weight ip) 30 minutes prior to their death. In addition, three separate experiments utilizing testosterone-replaced castrate animals were carried out. Testosterone replacement in these studies was effected via testosterone-filled silastic capsules’ of 5, IO, or I5 mm length (or empty capsules for castrate controls) implanted subcutaneously under ether anesthesia. In experiment one, castration (under ether anesthesia) was carried out at the age of 7 I days, at which time testosterone capsules were also inserted. Animals were killed at the age of 98 days with induction of nephrotic syndrome 2 weeks prior to their death. In experiments two and three, castration was carried out at weaning (the age of 21 days) with testosterone replacement delayed until the age of 85 days (experiment two. or the age of 87 days (experiment three). The killing occurred at the age of 99 days (experiment two) and the age of 101 days (experiment three) with induction of nephrotic syndrome 2 weeks prior to the killing. For all experimental regimens described, studies were carried out simultaneously in nephrotic rats and in nonnephrotic controls, with eight to ten animals in each group. Animals were killed by rapid aortic exsanguination under light ether anesthesia in order to obtain maximum blood. although animals receiving LHRH were killed by decapitation in order to meet time constraints. Ventral prostate, seminal vesicle, testis, and body weights were measured. Testes to be used for histologic examination were fixed in Bouin’s solution while testes to be incubated in vitro for measurement of testosterone output (vide infra) were kept on ice for short periods until incubation could be started. Anterior pituitary glands were removed for measurement of gonadotropin content (after homogenization) or were incubated in vitro to assess gonadotropin release (vide infra). Blood was allowed to clot at room temperature and serum was stored at -20 “C until assayed.

Metabolism, Vol 34, No 6 (June), 1985

HYPERGONADISM

575

IN NEPHROTIC RATS

Serum urea nitrogen or creatinine were measured by autoanalyzer in all specimens (Astra 8, Beckman Instruments, Brea, Calif), with subsequent exclusion of all animals showing uremia, as evidenced by serum urea nitrogen greater than 40 mg/dL and serum creatinine greater than I .8 mg/dL (approximately 1%of animals had to be excluded). Average serum urea nitrogen in nephrotic rats (21 f [SEMI mg/dL) was only slightly greater than in controls (18 + 1). although this difference was significant (P < 0.01) due to the large number of animals studied. Serum creatinine averaged 0.5 mg/dL in both nephrotic and control groups. Albumin was measured in all serum suecimens bv radial immunodiffusion6 using purified rat albumin’and antiseium against rat albumin obtained- from ICN Pharmaceutical, Cleveland. Animals not showing serum albumin less than 1.8g/dL were excluded from the nephrotic group (approximately 1% of the animals excluded). In addition, in ten control and ten nephrotic animals, 24-hour urine specimens were collected using metabolic cages and were similarly assayed for albumin. Luteinizing hormone (LH) and follicle-stimulating hormone (FSH) were measured in serum specimens, pituitary homogenates and media from in vitro pituitary incubations by radioimmunoassay (RIA) using the materials obtained from the National Pituitary Agency. Assay sensitivities were IO ng/mL for LH and 100 ng/mL for FSH, with intraassay variation of less than 9% and interassay variation of less than 15% for both assays. Testosterone was measured in serum specimens and in incubation media from both testis incubation by RIA after ether extraction and celite column chromatography (sensitivity IO ng/dL; intraassay variation 10%; interassay variation 14%). Free testosterone was measured using equilibrium dialysis at 37 OC.’ Samples from each nephrotic group and its corresponding control were always measured in the same assay run for these procedures. To assess testosterone release by testes in vitro, decapsulated testes from control and nephrotic animals were incubated at 37 “C in the Krebs-Ringer bicarbonate buffer for four hours, with (left testis) or without (right testis) added human chorionic gonadotropin (HCG; 400 mIU/mL), following which the incubation medium was aspirated and assayed for testosterone.’ Pituitary gonadotropin release in vitro was assessed by incubation of individual anterior pituitary halves from control and nephrotic rats (castrate or intact) in medium 199 for four hours at 37 “C, after an initial one hour preincubation period, with or without LHRH (20 ng/mL for noncastrate; 50 ng/mL for castrate) in the incubation medium.’ Incubation media were then aspirated and assayed for LH and FSH. Results are expressed per pituitary using the measured total pituitary weight and the weight of the incubated hemipituitary. To assess the effect of albumin on testosterone-mediated suppression of gonadotropin output in vitro, anterior hemipituitaries from normal adult male rats were incubated at 37 “C for four hours with LHRH (100 ng/mL) and either 50 or 100 rmol/L testosterone in the medium, either with or without added bovine albumin (4 g/dL; Sigma Chemical, St. Louis). After incubation, the medium was aspirated and assayed for LH and FSH, with the values expressed as a percentage of gonadotropin release by pituitaries similarly incubated in vitro in the absence of testosterone or albumin. Results were analyzed to assess a potential effect of addition of albumin by the Friedman nonparametric two-way analysis of variance.‘O All other statistical comparisons between control and nephrotic groups were done by Student’s t test, or if criteria for normality and equal variance could not be satisfied, by the nonparametric Wilcoxon rank sum test.

quent hypoalbuminemia (1.1 * 0.1 v 3.6 + 0.1 g/dL; P < 0.01). Following aminonucleoside injection, nephrotic rats do not gain (or lose) weight, in contrast to the normally growing nonnephrotic controls. Consequently, all groups of nephrotic rats weigh 8% to 15% less than corresponding controls at the killing. Nephrotic rats de;elopedhypoandrogenism, as man:fested by significant reductions in prostate and seminal vesicle weights and lower serum total and free testosterone levels (Table 1). In addition, testosterone release from testes incubated in vitro (without added HCG) was also significantly lower in the nephrotic group (Table 1). Despite this hypoandrogenism, the testis weight in nephrotic rats (1444 + 63 mg) was similar to that in controls (1491 rt 32 mg), and there was no difference between the two m-ouns in the histologic appearance of the testis unde; ligit microscopy. This relative lack of change in indices of spermatogenic function, despite the marked hypoandrogenism, may reflect the short time (2 weeks) between induction of nephrotic syndrome and the killing of the animals. To further evaluate androgen status in nephrotic syndrome, we measured in three separate experiments the prostate and seminal vesicle weights in castrate rats (control and nephrotic) after various doses of testosterone replacement (Table 2). Seminal vesicle weights were significantly different between the nephrotic and corresponding control groups in only one of nine comparisons, and prostate weights differed between the two groups in only two of nine comparisons (nephrotic prostates larger in one and smaller in the other). These few exceptions do not represent a consistent deviation from the general pattern of similar reproductive organ weights in control and nephrotic groups given the same testosterone replacement dose

Table ‘I. Pituitarv-Testicular

Function in Nmhrotic Rats* Control

Prostate weight lmg)

362 k 26$$

Seminal vesicle weight (mg)

504 t 19

383 + 22$$

Serum testosterone (ng/dL)

305 + 37

72 ? 8$$

18 + 2

6 + l$$

42 + 8

22 * 2$$

Serum free testosterone (ng/dL) Testosterone release by testis in vitro (ng/4h/testis)

LI I Serum lng/mLl Pituitary (fig/pituitary)

74 + 8 306 * 19

39 + 5$$ 249 + 14t

Serum (ng/mL)

366 t 17

317 +_ 7t

Pituitary (fig/pituitary)

149 * 17

FSH

RESULTS

*Mean + SEM (n = 8 to 10 per group).

Induction albuminuria

of nephrotic syndrome resulted in marked ([SEMI 259 * 20 mg/24 h) and conse-

Nephrotlc

489 r 24

t= $=

P < 0.05

vcontrol.

P < 0.01 vcontrol.

91 * 1;t

GLASS, BEACH, AND VIGERSKY

576

Table 2. Reproductive Organ Weights in Castrate, Testosterone-Replaced Rats’ Testosterone

Capsule Size (mm)

Experimentt

5

NO

10

15

377 % 27

403 + 21

359 ? 35

383 _c40

Prostate Weight (mg) Control

1

Nephrotic

297

k

19

193 + 1 l$

Control

80 + 5

147 f 11

181 + 11

Nephrotic

91 ? 7

124 + 9

176 + 10

Control

65 + 7

131 * 9

168 r 10

Nephrotic

96 + 5t

141 i 5

163 + 9

Seminal Vesicle Wt (mg) Control

296 zt 20

400 + 29

411 f 20

Nephrotic

273 ? 15

436 ? 30

392 t 24

Control

105 + 6

154i-7

198 + 4

Nephrotic

100 2 5

144 + 7

205 + 8

Control

106 r 8

160 ? 9

205 t 7

Nephrotic

132 + 7$

158 ? 7

195 + 7

*Mean + SEM (n = 8 to 10 per group). tExperiment one: castration and testosterone replacement age 7 1 days, sacrifice 4 weeks later; Experiments two and three: castratlon at age 2 1 days. testosterone replacement 2 weeks prior to sacrifice at age 100 days. $ = P i 0.05 Y control.

(ie, capsule size). If we thus assume that control and nephrotic rats with the same testosterone production rate (equivalent to the replacement dose in castrate rats) have similar prostate and seminal vesicle weights, then the previous finding that noncastrate nephrotic rats had smaller prostate and seminal vesicle weights than the corresponding noncastrate controls (Table 1) would suggest that the testosterone production rate in vivo, like that in vitro, must be lower in nephrotic rats. The hypoandrogenism in nephrotic rats was secondary to decreased gonadotropin output, as evidenced by lower serum and pituitary levels of LH and FSH (Table 1). In addition, the serum gonadotropin response to LHRH administration in vivo was impaired in nephrotic rats (Fig 1). Also suggesting that hypoandrogenism in nephrotic rats was a consequence of decreased gonadotropin output rather than an intrinsic testicular defect was the finding that, when exogenous hCG was added to testes incubated in vitro, the percentage increase in testosterone output was similar in control (635 * 161%) and nephrotic (581 k 53%) groups. The absolute levels of testosterone output in the presence of added hCG, however, were lower in the nephrotic group than in control (152 k 16 v 220 + 29 ng/4 h/testis; P < 0.05). The decreased gonadotropin output in nephrotic rats did not reflect an absolute inability to make gonadotropins. For example, the LH and FSH output from pituitary glands incubated in vitro, either with or without LHRH in the incubation medium, was similar in control and nephrotic rats (Table 3). Moreover, postcastration serum LH levels were actually higher in

SERUM FSH

SERUMLH

400 -

. = CONTROL A = NEPHROTIC * = p co.05

360 -

320 -

260-

660.

240-

600'

q200C

q520~ 5

160-

440'

120-

360'

60-

260~

40-

LHRHDOSE lng/l 00 g bodyWI) Serum LH (left) and FSH (right) 30 minutes after LHRH Fig 1. ip in control (solid circles) and nephrotic (open triangles) rats. Asterisk indicates P < 0.05 Y nephrotic group at the same time. Error bars indicate mean t SEM (n = 8 to 10).

HYPERGONADISM

577

IN NEPHROTIC RATS

Table 3. LH and FSH Release From Pituitary Glands Incubated LH Release lpg/4

in Vitro*

h)

FSH Release (fig/4

h)

LHRHt Added

Per Pituitary

Per mg Pituitary

Per Pituitary

Per mg Pituitary

Noncastrate Control

0

13.7 + 1.4

1.8 * 0.2

8.5 ? 0.9

1.1 * 0.1

Nephrotic

0

16.9 r 3.2

3.1 + 0.7

10.6 + 2.1

2.0 + 0.5

Control

+

26.1

+ 3.5

3.5 -t 0.6

15.8 f 1.8

2.1 * 0.2

Nephrotic

+

20.1

_+ 2.2

3.6 + 0.5

14.3 f 1.2

2.5 5 0.3

1.3 r 0.2

Castrate Control

0

54 + 5

5.0 + 0.8

14.2 ? 1.5

Nephrotic

0

58 + 7

5.7 + 0.7

12.2 + 1.5

1.2 + 0.1

Control

+

108 + 12

10.1 + 1.7

27.1

* 2.4

2.4 + 0.3

Nephrotic

+

121 t 27

13.1 + 1.9

28.8

r 8.1

2.6 ? 0.6

*All values are mean + SEM (n = 10 per group except n = 8 for noncastrate nephrotic). Values are fig gonadotropin released during four-hour incubation, expressed either per whole pituitary or per mg pituitary wet weight. None of the comparisons between control and nephrotic are statistically significant. TO = no LHRH added during incubation; + = LHRH addad during incubation (20 ng/mL for noncastrate; 50 ng/mL for castrate).

nephrotics than in controls in three separate studies, while postcastration serum FSH levels in nephrotics were slightly reduced in one study but not in two others (Table 4). Therefore, although gonadotropin output may be low in noncastrate nephrotic rats, it is not low if the testes are removed or if the pituitary gland is evaluated in vitro. These findings suggest that the low gonadotropin output in noncastrate nephrotic rats is secondary to an effect of the testis on the hypothalamic-pituitary unit rather than an intrinsic pituitary defect. To explore this possibility, the suppression of postcastration serum LH and FSH levels by various doses of testosterone replacement was assessed in three separate studies of control and nephrotic rats (Fig 2). In all three studies, testosterone-mediated suppression of gonadotropin output was greater in nephrotic than control groups. This increased gonadotropin suppression in nephrotic rats was not due to higher androgen levels attained after testosterone capsule implantation, since prostate and seminal vesicle weights were similar in nephrotic and control rats given the same size testosterone capsule (Table 2) and serum testosterone levels (measured in two studies) in nephrotic rats were the same or lower than those in controls implanted with the same size testosterone capsule (Table 5). To evaluate whether this increased testosterone sensitivity in nephrotic rats was due to hypoalbuminemia per se, the LHRH-stimulated gonadotropin output from normal pituitaries incubated in vitro was examined in the presence of testosterone, with or without albumin, in the incubation medium (Table 6). Addition of a commercial preparation of bovine serum albumin in physiologic concentration to the incubation medium reversed the testosterone-induced suppression of gonadotropin output, suggesting that the negative feedback effect of testosterone on gonadotropin release

may be enhanced by lowering the prevailing albumin concentrations. DISCUSSION

Adult male rats with hypoalbuminemia secondary to aminonucleoside-induced nephrotic syndrome were found to have hypoandrogenism, as evidenced by reduced prostate and seminal vesicle weight, lower serum total and free testosterone levels, decreased testosterone output from testes in vitro, and, by inference, decreased testosterone production rate in vivo. This hypoandrogenism seemed to be the result of decreased gonadotropin output, rather than primary testicular failure, since pituitary and serum gonadotropin levels were low and serum gonadotropin responses to LHRH administered in vivo were also decreased. Moreover, testosterone output from nephrotic testes incubated in vitro responded normally, expressed as percentage increase, to exogenous gonadotropin. This decreased gonadotropin output in nephrotic rats did not reflect an absolute inability to make gonadotropins, since postcastration rises in serum gonadotropin Table 4. Postcastration

Serum Gonadotropin

Serum LH (ng/mL)

Levels*

Serum FSH (ng/mL)

Experiment 1 Control Nephrotic

554 + 42 1,188

& 6lT

,470

i 38

,606 * 170

Experiment 2 Control Nephrotic

772 & 56 1,413

+ 78T

1 ,491

i 65

1 ,176

+ 62T

Experiment 3 Control

1,020

+ 77

2,059

+ 218

Nephrotic

1,878

+ 214t

2.802

? 283

*All values are mean + SEM tn = 8 to 10 per group). Castration occurred 4 weeks prior to killing in experiment one and 6 to 7 weeks prior to killing in experiments two and three. tP < 0.01 v control.

578

100

* c

. = CONTROL o = NEPHROTIC * = p co.05

80

2

Table 6. LHRH-Stimulated In Vitro

I”

* b\\*

C0ncentrat10n I”

Incubation

LH

Output

T

f 60 _, +!z 340

!I+ \ I \

4

\

\ \

I, B =

20

0

\

\

\

\

120

\

’ ‘4

_f PO

‘\\

pituitary

40-

5

*

levels were not impaired in nephrotic rats, while gonadotropin output measured in vitro, with or without LHRH, was normal. Rather, the low gonadotropin levels in nephrotic rats seemed to reflect increased sensitivity to the negative feedback effect of testosterone on gonadotropin output, as was directly demonstrated in vivo in castrate, testosterone-treated rats. However, our data do not exclude the possiblity that Table 5. Serum Testosterone

5

of in nephrotic

Rats

to decreased LH levels

in controls. of this

it

suggests

be a useful model of of the

to in nephrotic

testosterone’s be established to testosterone a variety of as starvation,” response

to a variety

or prepubera nonspecific In this a an effect of the or the

of

direct effect of to gain

15

10

in undernutrition.”

Experiment 1 69 * 5

96 + 7

68 t 6

68 f 4t

vitro

Experiment Control

72 t

137

140

Nephrotic

77 _t 9

125 + 16

109 _+

? SEM; values in

in LH levels,

of nephrotic it should be mentioned in nephrotic

Levels in Castrated

Testosterone-Treated

tP < 0.05 Yco”trol.

in our

to

15

Fig 2. Suppression of serum LH (top) and FSH (bottom) in castrate control (closed circles) and nephrotic (open triangles) rats by exogenous testosterone in three separate studies. Error bars indicate mean * SEM (n = 8 to 10) while asterisk indicates P < 0.05 v nephrotic. All values are expressed as a percentage of the level in castrate animals not receiving testosterone replacement.

44 t 3

(P < 0.01

in the

10

44 + 4’

mg

per group). Effect of

LH and

An interesting nephrotic

5 10 5 15 10 15 EXPERIMENT 3 EXPERIMENT 2 EXPERIMENT 1 TESTOSTERONE CAPSULE SIZElmml

Control

pg FSH

in testosterone-mediated of gonadotropins in nephrotic a greater production of in this

\

\ \

2L21-

r 21

or albumm m 5.0 wg LH

+ SEM (n = is significant

--_I

\ \ \

75 r 22

in vitro

*

\ \

74 I 110 x 14

55 * 123

the incubation medium (100%

* v

52 ‘- 5 119 -r 23

as percentage of gonadotropin

*

80-

0 4 0

*

looz g g 0 H

50 50 100

_-

*

I i z

VIGERSKY

*

*

~

GLASS, BEACH,

In contrast, of testosterone

be blocked by a commercial

testosterone

on gonadoin albumin to we cannot of the

HYPERGONADISM

579

IN NEPHROTIC RATS

nary in vitro experiment described reflect a contamination of the commercial albumin preparation used with sex-hormone-binding-globulin rather than an effect of the albumin itself. Whether this presumed in vitro effect of albumin can account quantitatively for the changes observed in hypoalbuminemic rats in vivo remains to be clarified, as does the mechanism by which albumin might exert this effect. Since Pardridge has proposed that albumin-bound hormone in circulation can be physiologically active,14 nephrosis-induced changes in bioavailability of various circulating androgen fractions might explain the apparent increase in testosterone sensitivity-although other mechanisms are also possible.

Also problematic is the relationship between our findings in nephrotic rats and potential abnormalities in humans with nephrotic syndrome. If nephrotic men, like rats, develop hypogonadotropic hypogonadism, then because nephrotic syndrome in humans can be prolonged without being fatal, adverse consequences (eg, delayed puberty in children, impotence, and androgen deficiency in men) might be anticipated. In view of our findings in rats, studies of pituitarygonadal function in humans with nephrotic syndrome would seem warranted. ACKNOWLEDGMENT We would like to thank Jeffrey Anderson for expert technical assistance

and Estelle Coleman

for proficient

secretarial

support.

REFERENCES 1. Morley JE, Melmed S: Gonadal dysfunction in systemic disorders. Metabolism 28:1051-1073, 1979 2. Gavin LA, McMahon FA, Castle JN, et al: Alterations in serum thyroid hormones and thyroxine-binding globulin in patients with nephrosis. J Clin Endocrinol Metab 46: 125-130, 1978 3. Afrasiabi MA, Vazir ND, Gwinup G, et al: Thyroid function studies in the nephrotic syndrome. Ann Intern Med 90: 335-338. 1979 4. Frenk S, Antonwicz I, Craig JM, et al: Experimental nephrotic syndrome induced in rats by aminonucleoside: Renal lesions and body electrolyte composition. Proc Sot Exp Biol Med 89:424-427, 1955 5. Legan SJ, Coon GA, Karsch FJ: Role of estrogen as initiator of daily LH surges in the ovariectomized rat. Endocrinology 96:5&56, 1975 6. Mancini G, Carbonara AO, Heremans JF: Immunochemical quantitation of antigens by single radial immunodiffusion. Immunochemistry 2:235-254, 1965 7. Vermeulen A, Stoica T, Verdonck L: The apparent free testosterone concentration, an index of androgenicity. J Clin Endocrinol Metab 33:759-767, 197 1

8. Payne AH, Kelch RP, Murono EP, et al: Hypothalamic, pituitary, and testicular function during sexual maturation in the male rat. J Endocrinol72: 17-26, 1977 9. O’Conner JL. Allen MB, Mahesh VB: Evaluation of in vitro incubation systems for the study of gonadotropin release. Endocr Res Commun 6: 15-36, 1979 IO. Conover WJ: Practical nonparametric statistics. New York, John Wiley, 197 1 p 273 11. Pirke KM, Spyra B: Influence of starvation on testosteroneluteinizing hormone feedback in the rat. Acta Endocrinol 96:413421, 1981 12. Glass AR, Steinberger A, Swerdloff R, et al: Pituitarytesticular function in protein-deficient rats. Follicle-stimulating hormone hyperresponse to castration and supersensitivity of gonadotropin secretion to androgen negative feedback. Endocrinology 110:1542-1546.1982 13. Smith ER, Damassa DA, Davidson JM: Feedback regulation and male puberty: Testosterone-luteinizing hormone relationships in the developing rat. Endocrinology 101:173-l 80, 1977 14. Pardridge WM: Transport of protein-bound hormones into tissues in viva. Endocr Rev 2: 103-I 23, 198 1