- Email: [email protected]
Immunohistochemical labeling of androgen receptors in the brain of rat and monkey
Life Sciences, Vol. Printed in the U S A
50, pp. 409-417
Pergamon P r e s s
IMMUNOHISTOCHEMICAL LABELING OF ANDROGEN RECEPTORS IN T H E B R A I N O F R A T A N D M O N K E Y Andrew N. Clancy, Robert W. Bonsall and Richard P. Michael Department of Psychiatry, Emory University School of Medicine, Atlanta, GA 30322 and Georgia Mental Health Institute, Atlanta, GA 30306 (Received in final form December 3, 1991)
Summary Androgen receptor antibodies have recently been developed using fusion proteins containing fragments of human prostatic androgen receptor. We have used a polyclonal antibody raised in rabbits to label androgen receptors in brain sections from male and female rats and monkeys. Free-floating frozen sections were incubated in primary antibody, and processed by the j?eroxidaseavidin-biotin complex method using biotinylated anti-rabbit IgG. Nickel intensified diaminobenzidine was used as the chromagen, and neurons were labeled in the amygdala, hippocampus, bed nucleus of stria terminalis, septum, preoptic area, in several hypothalamic nuclei including the supraoptic and paraventricular nuclei, in several brain stem motor nuclei and in cerebral cortex. Staining was most intense in cell nuclei but also occurred in cytoplasm and in some neuronal processes. Labeling was more restricted in monkey than in rat brain. Omittin$ the primary antibody or pre-incubating the primary antibody with rat prostauc cytosol for control purposes demonstrated the specificity of stalmng. Conventional autoradiography has provided a great deal of information about steroidconcentrating systems in the brains of fishes (1), amphibia (2), reptiles (3), birds (4,5) and mammals (6-8). When the non-aromatizable androgen 3H-dihydrotestosterone, or a synthetic androgen such as 3H-methyltrienolone, is used to visualize target neurons in discrete brain regions, it is assumed that a somewhat specific steroid-receptor complex is formed in the cell nucleus. This is a plausible explanation for the accumulation of androgens because excess amounts of unlabeled androgens block labeling (8). Immunohistochemical techniques have the advantage of visualizing steroid receptors without the need to deplete endogenous steroids or the need for administering radioactive ligands. The immunohistochemical approach has provided valuable data on estrogen and progestin receptors in the brains of birds (4,9), guinea pigs (10) and rats (11,12), and also on androgen receptors in genital tract (13-15). However, immunohistochemical data on androgen receptors in brain are sparse (13,16), particularly in primates. Methods Animal~. Seven littermate Sprague-Dawley rats (5 male, 2 female, aged 6 mos), were used together with 1 adult male cynomolgus monkey (Macacafascicularis) (body weight 4.6 kg), 1 adult female cynomolgus monkey (b.w. 5.2 kg), 1 neonatal female cynomolgus monkey (b.w. 270 g), 1 adult male rhesus monkey (M. mulatta) (b.w. 9.2 kg) and 1 adult female rhesus Correspondence: Richard P. Michael, M.D., Department of Psychiatry, Emory University School of Medicine, Atlanta, GA 30322 Copyright
0024-3205/92 $5.00 + .00 o 1992 Pergamon Press plc All rights reserved.
410
Brain Androgen Receptor Immunoreactivity
Vol. 50, No. 6, 1992
monkey (b.w. 6.6 kg). All maintenance, surgical and experimental procedures were in accordance with institutional regulations and with the NIH Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1985). Reagents. The primary antibody was raised in a rabbit against a bacterially expressed fusion protein containing the N-terminal domain of the human androgen receptor (16,17) (donated by Affinity Bioreagents, PAl-110, and supplied as whole serum).
Imm~lnohistochemical localization of androgen receptors. Animals were killed with a lethal injection of sodium pentobarbital and perfused transcardially, first with 0.9% saline and then with 2% paraformaldehyde, 0.15% picric acid in 0.1 M phosphate buffer (pH 7.4). Tissues were post-fixed for 60 min in the same fixative and stored overnight at 4°C in phosphatebuffered 30% sucrose. Free-floating coronal sections (25-40/~m) were collected on a freezing microtome and incubated for 60 min at 4°C with blocking solution containing normal goat serum, and for 36-40 hr at 4°C in 1:1000 to 1:5000 dilutions of the polyclonal antiserum. After three 5-min washes in phosphate buffer, sections were incubated for 2 hr at room temperature with the biotinylated secondary antibody (Vector Laboratories Vectastain rabbit ABC peroxidase kit or Vectastain Elite rabbit ABC peroxidase kit). For the peroxidase reaction, sections were washed and then incubated with avidin-biotinylated peroxidase complex for 60 min at room temperature followed by nickel intensified diaminobenzidine (Vector Laboratories DAB substrate kit). For controls, alternate sections were processed as above but without the primary antibody or were incubated with primary antibody preabsorbed with cytosol from rat prostate. Cytosols were prepared from the prostates of 4 adult male rats castrated 2 days earlier. Tissue was homogenized in 2 ml 0.1M phosphate buffer and normal goat serum (pH 7.4), centrifuged at 1000g for 10 min and then 100,000 g for 45 min at 2°C. The cytosolwas used to dilute the stock primary antiserum (1:10) to 1:1000 or 1:5000. After 18 hr, the cytosol-diluted antiserum was centrifuged at 5000g for 10 min and the supernatant was used as above. Results Antibody specificity
FIG 1 Rat brain stem: A. neurons labeled with rabbit polyclonal androgen receptor antibody; B. adjacent section when primary antibody was omitted; C. adjacent section using antibody pre-absorbed with rat prostatic cytosol. In B and C, labeling was eliminated. Scale = 25 ~m Widespread labeling occurred in rat brain using the rabbit polyclonal androgen receptor antibody. An example is illustrated in adjacent sections of brain stem motor neurons in Fig 1. A shows intense androgen receptor labeling (1:5000 dilution) of cell nuclei in neurons
Vol. 50, No. 6, 1992
Brain Androgen Receptor Immunoreactivity
411
of the motor nucleus of VII; B shows that omitting the androgen receptor antibody prevented staining in this brain region; and C shows that preabsorption of the androgen receptor antibody (1:5000 dilution) with rat prostatic cytosol, a tissue rich in androgen receptors, eliminated labeling in the same area. Labeling in monkey brain was also eliminated by omitting the primary antibody and much reduced by preabsorption with rat prostate cytosol: an example from the amygdala of an adult male is shown in Fig. 2. Findings such as these established the specificity o f this antibody for androgen receptors in the rat and monkey brain and also showed that the reagents employed to visualize the reaction did not themselves cause nonspecific staining.
FIG 2 Monkey amygdala: A. neurons labeled with rabbit polyclonal androgen receptor antibody; B. adjacent section when primary antibody was omitted; C. adjacent section using antibody pre-absorbed with rat prostatic cytosol. Scale = 25 #m Distribution of androgen receptor immunoreactivity in brain Rat. Rather than using a diagrammatic outline of the immunohistochemical staining of the brain with rabbit polyclonai androgen receptor antibody, we have compiled a photographic montage of a representative, but somewhat oblique, frontal section of the adult rat brain at about AP + 6.5 m its ventral aspect (Fig. 3). The locations of six regions, A though F, where neurons were heavily labeled are shown by the boxes, and a portion of each of these sites is shown at higher magnification in Fig. 4. In A, pyramidal cells in what may be layer V of the cerebral cortex were stained, and labeling extended throughout this layer. In some cases, apical dendrites appeared to have been lightly stained also. A second, much deeper layer of cortex superficial to the sub-cortical white matter, containing smaller fusiform neurons, was also labeled (not illustrated). In B, a tightly packed layer of what appeared to be pyramidal cells in Arnmon's horn was densely labeled, whereas neurons in the dentate gyrus were unlabeled. In C, a group of neurons in the dorsal thalamic region were labeled. In D, note a well-localized cluster of labeled cells in the anterior hypothalamic area. In E, many neurons were labeled in the supraoptic nucleus. In F, a region in the center of the anterior amygdaloid area showed densely packed labeled neurons. Rostral to the regions depicted in Fig. 3, labeled neurons were observed in the medial preoptic nucleus (Fig. 5A), bed nucleus of stria terminalis and lateral septum. Caudally, labeled neurons were found in the ventromedial hypothalamus (Fig. 5B), arcuate nucleus and brain stem (Fig. 1A). ~ y
In both adult male and female M. mulatta and M. fascicularis, the 3 regions most labeled using the rabbit polyclonal androgen receptor antibody were the medial preoptic nucleus (Fig. 6A), the supraoptic nucleus (Fig. 6B) and the paraventricular nucleus of the hypothalamus (not illustrated). Stained neurons were also observed in other regions of hypothalamus (Table 1) and in amygdala (Fig. 2A), but labeling was both less dense and
412
Brain Androgen Receptor Immunoreactivity
Vol. 50, No. 6, 1992
FIGS. 3 and 4 wiAbove) Montage of photographs of a frontal section of an adult rat brain stained th rabbit polyclonal androgen receptor antibody. Portions of the areas in marked boxes are shown at much higher magnification in Fig. 4 (right). A. Neurons in cerebrai cortex, probably layer V; B. Pyramidal cell layer of the hippocampus; C. Neurons in the dorsal thalamic region, sm = stria medullaris; D. Neurons in the anterior hypothalamic area, v= third ventricle; E. Neurons in the supraoptic nucleus, oc = optic chiasm; F. Neurons in the anterior amygdaloid area. Scales: Fig. 3 = 1 mm, Fig. 4 = 50/~m.
Vol. 50, NO. 6, 1992
Brain Androgen Receptor Immunoreactivity
FIG. 4
413
414
Brain Androgen Receptor Immunoreactivity
Vol. 50, No. 6, 1992
FIG 5 A. neurons in the rat medial preoptic nucleus and B. neurons in the ventromedial hypothalamic nucleus labeled with rabbit polyclonal androgen receptor antibody. Scale = 25/~m more restricted than in the rat. Neurons in the primate cerebral cortex and hippocampus were not consistently labeled, but full analysis of the material is not complete. TABLE 1 Comparison of Autoradiographic and Immunohistochemical Labeling in the Hypothalamus and Limbic System of Adult Rat and Monkey 3H-T n. septi lateralis n. preopticus medialis bed n. stria terminalis anterior hypothalamic area n. paraventricularis n. supraopticus n. ventromedialis n. arcuatus area amygdala anterior n. amygdaloideus basalis n. amygdaloideus centralis n. amygdaloideus corticalis n. amygdaloideus medialis hippocampus: pyramidal cell layer
Rat anti-AR
Monkey 3H-DHT anti-AR
+ + + + + + +
+ + + + + + + +
+ + + + + + +
+ + + + + + + +
+ + +
+ +
NA + NA
NA NA NA
+ +
+ +
+ +
+ +
+
+
+
NA
3H-T = autoradiography with 3H-testosterone, see ref 7; anti-AR = androgen receptor immunoreactivity; 3H-DHT = autoradiography with 3H-dihydrotestosterone, see ref 18.; NA = not available. Labeling in a single female M. fascicularis neonate (270 g), born prematurely, was more widespread than in the adult, and labeling was clearly observed in ventromedial and arcuate hypothalamic nuclei, septum, medial preoptic area, supraoptic nucleus, anterior hypothalamic area, paraventricular nucleus, bednucleus of stria terminalis and throughout the amygdala.
Vol. 50, No. 6, 1992
Brain Androgen Receptor Immunoreactivity
415
i
FIG 6 A. Neurons in the primate medial preoptic nucleus and B. neurons in the primate supraoptic nucleus labeled with rabbit polyclonal androgen receptor antibody. Scale = 100/~m High magnification revealed that neuronal nuclei were stained more densely in these 25-40 t~m brain sections than was the cytoplasm (Fig. 7), and this occurred in all brain areas studied. Cytoplasmic staining occurred only in neurons whose nuclei were densely stained. Combined immunohistochemistry and autoradiography in the same brain sections appeared to be a practical possibility. Fig. 8 shows neurons in the bed nucleus of stria terminalis of a castrated adult rat given 1 mCi/kg 3H-testosterone 60 min before death. Silver grains in the autoradiogram after 13 weeks were clustered over cell nuclei, indicating uptake of testosterone or its metabolites, and the section was incubated on the slide with rabbit polyclonal androgen receptor antibody, but without nickel intensification: this facilitated visualization of the silver grains in the photographic emulsion. Cell a, which contained androgen receptor immunoreactivity, also accumulated tritiated steroid from the sysFIG 7 temic circulation, and we interpret this to One of numerous neurons in the male rat brain be mainly androgen bound to androgen labeled with rabbit polyelonal androgen receptor receptor. Cell b, although not containing antibody. Nuclei stained densely while the eytoandrogen receptor immunoreactivity, did plasm stained more lightly. Scale --- 10 t~m. accumulate tritiated steroid/n vivo. We interpret this to be an estrogenic metabolite binding to the estrogen receptor. Discussion The availability of a specific antibody has made it possible to visualize androgen-receptor immunoreactivity in the brain of rat and monkey. In rat and in both species of primate, there was little or no staining in sections incubated in the absence of primary antibody or in sections using antibody preincubated with cytosol rich in androgen receptors. Thisprovided good evidence that the neuronal labeling was not an artifact but a consequence of the presence of androgen receptor immunoreactivity. It was clear that most of this im-
416
Brain Androgen Receptor In~munoreactivity
munoreactivity was located in the nucleus, but in neurons with stained nuclei, some staining of the cytoplasm often occurred. This raised the possibility that, contrary to a current view on the subcellular localization of steroid receptors, androgen receptors are not confined to the nucleus: a similar situation has been observed in immunohistochemical studies on estrogen receptors in brain (10,19). In an earlier report using 6-/~m sections, androgen receptor lmmunoreactivity appeared to be restricted to the nucleus (13), but in the current work, sections were 25-40 #m which would tend to enhance the density of cytoplasmic staining and reduce any loss of soluble cytoplasmic receptors. Cytoplasmic staining, like nuclear staining, in our thicker sections was eliminated by omitting the primary antibody and reduced by preabsorption with rat prostate cytosol.
Vol. 50, No. 6, 1992
FIG 8 Combined 3H-testosterone autoradiography and androgen receptor immunohistochemistry of adjacent neurons (a and b) in a 4/~m cryostat section of the bed nucleus of stria terminalis of an adult male rat. Cell a is labeled by both methods and cell b is labeled by autoradiography alone. Scale = 10/~m.
It was noteworthy that, after 13 weeks exposure, 4-/~m brain sections processed for autoradiosraphy retained androgen receptor immunoreactivity in cell nuclei. This facilitated dual labehng of neurons with 3H-testosterone and with the polyclonal antibody (Fig. 8). The observation that many neurons were labeled both autoradiographically and immunohistochemically provided further evidence that the rabbit polyclonal antibody reacted with androen receptors in brain. We hypothesize that neurons labeled autoradiographically but not beled lmmunohistochemically had accumulated metabolites of 3H-testosterone, probably 3H-estradiol formed by aromatization (20), binding to estrogen rather than androgen receptors. We do not favor the alternative explanation that some neurons lost immunoreactivity during processing because it is unlikel~y that closely adjacent neurons would have been affected so differently. There was considerable congruence in the adult rat brain between the distribution of androgen receptor immunoreactivity and the distribution of neurons labeled autoradiographically after administering 3H-testosterone (Table 1), but we have yet to make a comprehensive, quantitative analysis of this. An exception was the supraoptic nucleus, not generally thou~ght to be a target region for androgens, which contained many neurons labeled intensely with the polyclonal antibody in both rats and monkeys. Labeling in the supraoptic nucleus was greatly reduced by preincubating the antibody with rat prostatic cytosol, but this does not eliminate the possibility that the antibody might, under some circumstances, recognize non-receptor proteins or non-functional receptors such as those thought to be present in the fetal rat prostate (14). It was puzzling that labeling in adult monkey was more restricted than that m neonatal monkey and adult rat because androgentarget neurons are widely distributed in the adult primate brain (18). Autoradiography with labeled steroids, combined with immunohistochemistry, should help to increase our understanding of the nature of these brain-hormone interacuons and how they influence reproductive behavior. Acknowledgments This work was supported by USPHSgrants MH 19506 and MH 40420, by the Emory University Research Committee and by the Georgia Department of Human Resources.
Vol. 50, No. 6, 1992
Brain Androgen Receptor Immunoreactlvity
417
References 1. M.L FINE, D.A. KEEFER and H. RUSSEL-MERGENTHAL, Brain Res. 536 207-219 (1990). 2. D.B. KELLEY, J. Comp. Neurol. 199 221-231 (1981). 3. M.C. MARTINEZ-VARGAS, D.A. KEEFER and W.E. STUMPF, J. Exp. Zool. 205 141-147 (1978). 4. M. GAHR, Proc. Natl. Acad. Sci. 87 9445-9448 (1990). 5. J.L LUBISCHER and A.P. ARNOLD, J. Comp. Neurol. 291 450-456 (1990). 6. R.P. MICHAEL Br. Med. Bull., 21 87-90 (1965). 7. M. SAR and W.E. STUMPF, Anatomical Neuroendocrinology, W.E. Stumpf and LD. Grant (eds), 120-133, S. Karger, Basel (1975). 8. R.P. MICHAEL H.D. REES and R.W. BONSALL, Brain Res. 502 11-20 (1989). 9. J. BALTHAZART, A. FOIDART, C. SURLEMONT C. and N. HARADA, J. Neurobiol. 22 143-157 (1991). 10. J.D. BLAUSTEIN and J.C. TURCOTFE, Neuroendocrinology 49 454-461 (1989). 11. A. CINTRA, K. FUXE, A. HARFSTRAND, L.F. AGNATI, L.S. MILLER, J.L GREENE and J.A. GUSTAFSSON, Neurochem. Int. 8 587-595 (1986). 12. M. SAR and I. PARIKH, J. Steroid Biochem. 24 497-503 (1986). 13. M. SAR, D.B. LUBHAHN, F.S. FRENCH and E.M. WILSON, Endocrinology 127 31803186 (1990). 14. H. TAKEDA and C. CHANG, J. Endocrinol. 129 83-89 (1991). 15. S. HILD-PETITO, N.B. WEST, R.M. BRENNER and R.L. STOUFFER, Biol. Reprod. 44 561-568 (1991). 16. H. TAKEDA, G. CHODAK, S. MUTCHNIK, T. NAKAMOTO, and C. CHANG, J. Endocrinol. 126 17-25 (1990). 17. C. CHANG, C. WHELAN, T. POPOVICH, J. KOKONTIS and S. LIAO, Endocrinology 125 1097-1099 (1989). 18. R.W. BONSALL, H.D. REES and R.P. MICHAEL~ J. Steroid Biochem. 23 389-398 (1985). 19. I. PARIKH, K.G. RAJENDRAN, J.-L. SU, T. LOPEZ and M. SAR, J. Steroid Biochem. 27 185-192 (1987). 20. I. LIEBERBURG and B.S. MCEWEN, Endocrinology 100 588-597 (1977).
50, pp. 409-417
Pergamon P r e s s
IMMUNOHISTOCHEMICAL LABELING OF ANDROGEN RECEPTORS IN T H E B R A I N O F R A T A N D M O N K E Y Andrew N. Clancy, Robert W. Bonsall and Richard P. Michael Department of Psychiatry, Emory University School of Medicine, Atlanta, GA 30322 and Georgia Mental Health Institute, Atlanta, GA 30306 (Received in final form December 3, 1991)
Summary Androgen receptor antibodies have recently been developed using fusion proteins containing fragments of human prostatic androgen receptor. We have used a polyclonal antibody raised in rabbits to label androgen receptors in brain sections from male and female rats and monkeys. Free-floating frozen sections were incubated in primary antibody, and processed by the j?eroxidaseavidin-biotin complex method using biotinylated anti-rabbit IgG. Nickel intensified diaminobenzidine was used as the chromagen, and neurons were labeled in the amygdala, hippocampus, bed nucleus of stria terminalis, septum, preoptic area, in several hypothalamic nuclei including the supraoptic and paraventricular nuclei, in several brain stem motor nuclei and in cerebral cortex. Staining was most intense in cell nuclei but also occurred in cytoplasm and in some neuronal processes. Labeling was more restricted in monkey than in rat brain. Omittin$ the primary antibody or pre-incubating the primary antibody with rat prostauc cytosol for control purposes demonstrated the specificity of stalmng. Conventional autoradiography has provided a great deal of information about steroidconcentrating systems in the brains of fishes (1), amphibia (2), reptiles (3), birds (4,5) and mammals (6-8). When the non-aromatizable androgen 3H-dihydrotestosterone, or a synthetic androgen such as 3H-methyltrienolone, is used to visualize target neurons in discrete brain regions, it is assumed that a somewhat specific steroid-receptor complex is formed in the cell nucleus. This is a plausible explanation for the accumulation of androgens because excess amounts of unlabeled androgens block labeling (8). Immunohistochemical techniques have the advantage of visualizing steroid receptors without the need to deplete endogenous steroids or the need for administering radioactive ligands. The immunohistochemical approach has provided valuable data on estrogen and progestin receptors in the brains of birds (4,9), guinea pigs (10) and rats (11,12), and also on androgen receptors in genital tract (13-15). However, immunohistochemical data on androgen receptors in brain are sparse (13,16), particularly in primates. Methods Animal~. Seven littermate Sprague-Dawley rats (5 male, 2 female, aged 6 mos), were used together with 1 adult male cynomolgus monkey (Macacafascicularis) (body weight 4.6 kg), 1 adult female cynomolgus monkey (b.w. 5.2 kg), 1 neonatal female cynomolgus monkey (b.w. 270 g), 1 adult male rhesus monkey (M. mulatta) (b.w. 9.2 kg) and 1 adult female rhesus Correspondence: Richard P. Michael, M.D., Department of Psychiatry, Emory University School of Medicine, Atlanta, GA 30322 Copyright
0024-3205/92 $5.00 + .00 o 1992 Pergamon Press plc All rights reserved.
410
Brain Androgen Receptor Immunoreactivity
Vol. 50, No. 6, 1992
monkey (b.w. 6.6 kg). All maintenance, surgical and experimental procedures were in accordance with institutional regulations and with the NIH Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85-23, revised 1985). Reagents. The primary antibody was raised in a rabbit against a bacterially expressed fusion protein containing the N-terminal domain of the human androgen receptor (16,17) (donated by Affinity Bioreagents, PAl-110, and supplied as whole serum).
Imm~lnohistochemical localization of androgen receptors. Animals were killed with a lethal injection of sodium pentobarbital and perfused transcardially, first with 0.9% saline and then with 2% paraformaldehyde, 0.15% picric acid in 0.1 M phosphate buffer (pH 7.4). Tissues were post-fixed for 60 min in the same fixative and stored overnight at 4°C in phosphatebuffered 30% sucrose. Free-floating coronal sections (25-40/~m) were collected on a freezing microtome and incubated for 60 min at 4°C with blocking solution containing normal goat serum, and for 36-40 hr at 4°C in 1:1000 to 1:5000 dilutions of the polyclonal antiserum. After three 5-min washes in phosphate buffer, sections were incubated for 2 hr at room temperature with the biotinylated secondary antibody (Vector Laboratories Vectastain rabbit ABC peroxidase kit or Vectastain Elite rabbit ABC peroxidase kit). For the peroxidase reaction, sections were washed and then incubated with avidin-biotinylated peroxidase complex for 60 min at room temperature followed by nickel intensified diaminobenzidine (Vector Laboratories DAB substrate kit). For controls, alternate sections were processed as above but without the primary antibody or were incubated with primary antibody preabsorbed with cytosol from rat prostate. Cytosols were prepared from the prostates of 4 adult male rats castrated 2 days earlier. Tissue was homogenized in 2 ml 0.1M phosphate buffer and normal goat serum (pH 7.4), centrifuged at 1000g for 10 min and then 100,000 g for 45 min at 2°C. The cytosolwas used to dilute the stock primary antiserum (1:10) to 1:1000 or 1:5000. After 18 hr, the cytosol-diluted antiserum was centrifuged at 5000g for 10 min and the supernatant was used as above. Results Antibody specificity
FIG 1 Rat brain stem: A. neurons labeled with rabbit polyclonal androgen receptor antibody; B. adjacent section when primary antibody was omitted; C. adjacent section using antibody pre-absorbed with rat prostatic cytosol. In B and C, labeling was eliminated. Scale = 25 ~m Widespread labeling occurred in rat brain using the rabbit polyclonal androgen receptor antibody. An example is illustrated in adjacent sections of brain stem motor neurons in Fig 1. A shows intense androgen receptor labeling (1:5000 dilution) of cell nuclei in neurons
Vol. 50, No. 6, 1992
Brain Androgen Receptor Immunoreactivity
411
of the motor nucleus of VII; B shows that omitting the androgen receptor antibody prevented staining in this brain region; and C shows that preabsorption of the androgen receptor antibody (1:5000 dilution) with rat prostatic cytosol, a tissue rich in androgen receptors, eliminated labeling in the same area. Labeling in monkey brain was also eliminated by omitting the primary antibody and much reduced by preabsorption with rat prostate cytosol: an example from the amygdala of an adult male is shown in Fig. 2. Findings such as these established the specificity o f this antibody for androgen receptors in the rat and monkey brain and also showed that the reagents employed to visualize the reaction did not themselves cause nonspecific staining.
FIG 2 Monkey amygdala: A. neurons labeled with rabbit polyclonal androgen receptor antibody; B. adjacent section when primary antibody was omitted; C. adjacent section using antibody pre-absorbed with rat prostatic cytosol. Scale = 25 #m Distribution of androgen receptor immunoreactivity in brain Rat. Rather than using a diagrammatic outline of the immunohistochemical staining of the brain with rabbit polyclonai androgen receptor antibody, we have compiled a photographic montage of a representative, but somewhat oblique, frontal section of the adult rat brain at about AP + 6.5 m its ventral aspect (Fig. 3). The locations of six regions, A though F, where neurons were heavily labeled are shown by the boxes, and a portion of each of these sites is shown at higher magnification in Fig. 4. In A, pyramidal cells in what may be layer V of the cerebral cortex were stained, and labeling extended throughout this layer. In some cases, apical dendrites appeared to have been lightly stained also. A second, much deeper layer of cortex superficial to the sub-cortical white matter, containing smaller fusiform neurons, was also labeled (not illustrated). In B, a tightly packed layer of what appeared to be pyramidal cells in Arnmon's horn was densely labeled, whereas neurons in the dentate gyrus were unlabeled. In C, a group of neurons in the dorsal thalamic region were labeled. In D, note a well-localized cluster of labeled cells in the anterior hypothalamic area. In E, many neurons were labeled in the supraoptic nucleus. In F, a region in the center of the anterior amygdaloid area showed densely packed labeled neurons. Rostral to the regions depicted in Fig. 3, labeled neurons were observed in the medial preoptic nucleus (Fig. 5A), bed nucleus of stria terminalis and lateral septum. Caudally, labeled neurons were found in the ventromedial hypothalamus (Fig. 5B), arcuate nucleus and brain stem (Fig. 1A). ~ y
In both adult male and female M. mulatta and M. fascicularis, the 3 regions most labeled using the rabbit polyclonal androgen receptor antibody were the medial preoptic nucleus (Fig. 6A), the supraoptic nucleus (Fig. 6B) and the paraventricular nucleus of the hypothalamus (not illustrated). Stained neurons were also observed in other regions of hypothalamus (Table 1) and in amygdala (Fig. 2A), but labeling was both less dense and
412
Brain Androgen Receptor Immunoreactivity
Vol. 50, No. 6, 1992
FIGS. 3 and 4 wiAbove) Montage of photographs of a frontal section of an adult rat brain stained th rabbit polyclonal androgen receptor antibody. Portions of the areas in marked boxes are shown at much higher magnification in Fig. 4 (right). A. Neurons in cerebrai cortex, probably layer V; B. Pyramidal cell layer of the hippocampus; C. Neurons in the dorsal thalamic region, sm = stria medullaris; D. Neurons in the anterior hypothalamic area, v= third ventricle; E. Neurons in the supraoptic nucleus, oc = optic chiasm; F. Neurons in the anterior amygdaloid area. Scales: Fig. 3 = 1 mm, Fig. 4 = 50/~m.
Vol. 50, NO. 6, 1992
Brain Androgen Receptor Immunoreactivity
FIG. 4
413
414
Brain Androgen Receptor Immunoreactivity
Vol. 50, No. 6, 1992
FIG 5 A. neurons in the rat medial preoptic nucleus and B. neurons in the ventromedial hypothalamic nucleus labeled with rabbit polyclonal androgen receptor antibody. Scale = 25/~m more restricted than in the rat. Neurons in the primate cerebral cortex and hippocampus were not consistently labeled, but full analysis of the material is not complete. TABLE 1 Comparison of Autoradiographic and Immunohistochemical Labeling in the Hypothalamus and Limbic System of Adult Rat and Monkey 3H-T n. septi lateralis n. preopticus medialis bed n. stria terminalis anterior hypothalamic area n. paraventricularis n. supraopticus n. ventromedialis n. arcuatus area amygdala anterior n. amygdaloideus basalis n. amygdaloideus centralis n. amygdaloideus corticalis n. amygdaloideus medialis hippocampus: pyramidal cell layer
Rat anti-AR
Monkey 3H-DHT anti-AR
+ + + + + + +
+ + + + + + + +
+ + + + + + +
+ + + + + + + +
+ + +
+ +
NA + NA
NA NA NA
+ +
+ +
+ +
+ +
+
+
+
NA
3H-T = autoradiography with 3H-testosterone, see ref 7; anti-AR = androgen receptor immunoreactivity; 3H-DHT = autoradiography with 3H-dihydrotestosterone, see ref 18.; NA = not available. Labeling in a single female M. fascicularis neonate (270 g), born prematurely, was more widespread than in the adult, and labeling was clearly observed in ventromedial and arcuate hypothalamic nuclei, septum, medial preoptic area, supraoptic nucleus, anterior hypothalamic area, paraventricular nucleus, bednucleus of stria terminalis and throughout the amygdala.
Vol. 50, No. 6, 1992
Brain Androgen Receptor Immunoreactivity
415
i
FIG 6 A. Neurons in the primate medial preoptic nucleus and B. neurons in the primate supraoptic nucleus labeled with rabbit polyclonal androgen receptor antibody. Scale = 100/~m High magnification revealed that neuronal nuclei were stained more densely in these 25-40 t~m brain sections than was the cytoplasm (Fig. 7), and this occurred in all brain areas studied. Cytoplasmic staining occurred only in neurons whose nuclei were densely stained. Combined immunohistochemistry and autoradiography in the same brain sections appeared to be a practical possibility. Fig. 8 shows neurons in the bed nucleus of stria terminalis of a castrated adult rat given 1 mCi/kg 3H-testosterone 60 min before death. Silver grains in the autoradiogram after 13 weeks were clustered over cell nuclei, indicating uptake of testosterone or its metabolites, and the section was incubated on the slide with rabbit polyclonal androgen receptor antibody, but without nickel intensification: this facilitated visualization of the silver grains in the photographic emulsion. Cell a, which contained androgen receptor immunoreactivity, also accumulated tritiated steroid from the sysFIG 7 temic circulation, and we interpret this to One of numerous neurons in the male rat brain be mainly androgen bound to androgen labeled with rabbit polyelonal androgen receptor receptor. Cell b, although not containing antibody. Nuclei stained densely while the eytoandrogen receptor immunoreactivity, did plasm stained more lightly. Scale --- 10 t~m. accumulate tritiated steroid/n vivo. We interpret this to be an estrogenic metabolite binding to the estrogen receptor. Discussion The availability of a specific antibody has made it possible to visualize androgen-receptor immunoreactivity in the brain of rat and monkey. In rat and in both species of primate, there was little or no staining in sections incubated in the absence of primary antibody or in sections using antibody preincubated with cytosol rich in androgen receptors. Thisprovided good evidence that the neuronal labeling was not an artifact but a consequence of the presence of androgen receptor immunoreactivity. It was clear that most of this im-
416
Brain Androgen Receptor In~munoreactivity
munoreactivity was located in the nucleus, but in neurons with stained nuclei, some staining of the cytoplasm often occurred. This raised the possibility that, contrary to a current view on the subcellular localization of steroid receptors, androgen receptors are not confined to the nucleus: a similar situation has been observed in immunohistochemical studies on estrogen receptors in brain (10,19). In an earlier report using 6-/~m sections, androgen receptor lmmunoreactivity appeared to be restricted to the nucleus (13), but in the current work, sections were 25-40 #m which would tend to enhance the density of cytoplasmic staining and reduce any loss of soluble cytoplasmic receptors. Cytoplasmic staining, like nuclear staining, in our thicker sections was eliminated by omitting the primary antibody and reduced by preabsorption with rat prostate cytosol.
Vol. 50, No. 6, 1992
FIG 8 Combined 3H-testosterone autoradiography and androgen receptor immunohistochemistry of adjacent neurons (a and b) in a 4/~m cryostat section of the bed nucleus of stria terminalis of an adult male rat. Cell a is labeled by both methods and cell b is labeled by autoradiography alone. Scale = 10/~m.
It was noteworthy that, after 13 weeks exposure, 4-/~m brain sections processed for autoradiosraphy retained androgen receptor immunoreactivity in cell nuclei. This facilitated dual labehng of neurons with 3H-testosterone and with the polyclonal antibody (Fig. 8). The observation that many neurons were labeled both autoradiographically and immunohistochemically provided further evidence that the rabbit polyclonal antibody reacted with androen receptors in brain. We hypothesize that neurons labeled autoradiographically but not beled lmmunohistochemically had accumulated metabolites of 3H-testosterone, probably 3H-estradiol formed by aromatization (20), binding to estrogen rather than androgen receptors. We do not favor the alternative explanation that some neurons lost immunoreactivity during processing because it is unlikel~y that closely adjacent neurons would have been affected so differently. There was considerable congruence in the adult rat brain between the distribution of androgen receptor immunoreactivity and the distribution of neurons labeled autoradiographically after administering 3H-testosterone (Table 1), but we have yet to make a comprehensive, quantitative analysis of this. An exception was the supraoptic nucleus, not generally thou~ght to be a target region for androgens, which contained many neurons labeled intensely with the polyclonal antibody in both rats and monkeys. Labeling in the supraoptic nucleus was greatly reduced by preincubating the antibody with rat prostatic cytosol, but this does not eliminate the possibility that the antibody might, under some circumstances, recognize non-receptor proteins or non-functional receptors such as those thought to be present in the fetal rat prostate (14). It was puzzling that labeling in adult monkey was more restricted than that m neonatal monkey and adult rat because androgentarget neurons are widely distributed in the adult primate brain (18). Autoradiography with labeled steroids, combined with immunohistochemistry, should help to increase our understanding of the nature of these brain-hormone interacuons and how they influence reproductive behavior. Acknowledgments This work was supported by USPHSgrants MH 19506 and MH 40420, by the Emory University Research Committee and by the Georgia Department of Human Resources.
Vol. 50, No. 6, 1992
Brain Androgen Receptor Immunoreactlvity
417
References 1. M.L FINE, D.A. KEEFER and H. RUSSEL-MERGENTHAL, Brain Res. 536 207-219 (1990). 2. D.B. KELLEY, J. Comp. Neurol. 199 221-231 (1981). 3. M.C. MARTINEZ-VARGAS, D.A. KEEFER and W.E. STUMPF, J. Exp. Zool. 205 141-147 (1978). 4. M. GAHR, Proc. Natl. Acad. Sci. 87 9445-9448 (1990). 5. J.L LUBISCHER and A.P. ARNOLD, J. Comp. Neurol. 291 450-456 (1990). 6. R.P. MICHAEL Br. Med. Bull., 21 87-90 (1965). 7. M. SAR and W.E. STUMPF, Anatomical Neuroendocrinology, W.E. Stumpf and LD. Grant (eds), 120-133, S. Karger, Basel (1975). 8. R.P. MICHAEL H.D. REES and R.W. BONSALL, Brain Res. 502 11-20 (1989). 9. J. BALTHAZART, A. FOIDART, C. SURLEMONT C. and N. HARADA, J. Neurobiol. 22 143-157 (1991). 10. J.D. BLAUSTEIN and J.C. TURCOTFE, Neuroendocrinology 49 454-461 (1989). 11. A. CINTRA, K. FUXE, A. HARFSTRAND, L.F. AGNATI, L.S. MILLER, J.L GREENE and J.A. GUSTAFSSON, Neurochem. Int. 8 587-595 (1986). 12. M. SAR and I. PARIKH, J. Steroid Biochem. 24 497-503 (1986). 13. M. SAR, D.B. LUBHAHN, F.S. FRENCH and E.M. WILSON, Endocrinology 127 31803186 (1990). 14. H. TAKEDA and C. CHANG, J. Endocrinol. 129 83-89 (1991). 15. S. HILD-PETITO, N.B. WEST, R.M. BRENNER and R.L. STOUFFER, Biol. Reprod. 44 561-568 (1991). 16. H. TAKEDA, G. CHODAK, S. MUTCHNIK, T. NAKAMOTO, and C. CHANG, J. Endocrinol. 126 17-25 (1990). 17. C. CHANG, C. WHELAN, T. POPOVICH, J. KOKONTIS and S. LIAO, Endocrinology 125 1097-1099 (1989). 18. R.W. BONSALL, H.D. REES and R.P. MICHAEL~ J. Steroid Biochem. 23 389-398 (1985). 19. I. PARIKH, K.G. RAJENDRAN, J.-L. SU, T. LOPEZ and M. SAR, J. Steroid Biochem. 27 185-192 (1987). 20. I. LIEBERBURG and B.S. MCEWEN, Endocrinology 100 588-597 (1977).