GnRH-prohormone-containing neurons in the primate brain: Immunostaining for the GnRH-Associated peptide

Peptides, Vol. 8, pp. 335-346. ©PergamonJournals Ltd., 1987. Printedin the U.S.A.

0196-9781/87$3.00 + .00

GnRH-Prohormone-Containing Neurons in the Primate Brain: Immunostaining for the GnRH-Associated Peptide TIANBAO SONG,* KAROLY NIKOLICS,t PETER H. SEEBURGt AND PAUL C. GOLDSMITH .2

*Reproductive Endocrinology Center, Department o f Obstetrics, Gynecology and Reproductive Sciences University of California San Francisco, San Francisco, CA 94143-0556 tDepartment of Developmental Biology, Genetech, Inc. 460 Point San Bruno Boulevard, South San Francisco, CA 94080 Received 4 August 1986 SONG, T., K. NIKOLICS, P. H. SEEBURG AND P. C. GOLDSMITH. GnRH-prohormone-containingneurons in the primate brain: Immunostainingfor the GnRH-associated peptide. PEPTIDES 8(2) 335-346, 1987.--The structure of the prohormone for mammalian gonadotropin releasing hormone (proGnRH) includes the GnRH decapeptide followed by a 56 amino acid GnRH-associated peptide (GAP). In this study, we compared immunostaining of brain neurons and fibers for GAP and GnRH in fetal rhesus monkeys and juvenile baboons. We used antisera against different portions of human and rat GAP (proGnRH 14-24, proGnRH 40--53, and proGnRH 52-66) or against GnRH and the PAP technique. Liquid phase absorption with GAP or GnRH confirmed the specificity of these antisera. Major accumulations of GAP immunoreactive (GAP+) perikarya occurred in the medial septal and preoptic areas and the nucleus of the diagonal band of Broca (44.6% in rhesus, 49.6% in baboon), supraoptic region including the area dorsal to the optic tract (21.9% in rhesus, 23.0% in baboon), and the medial basal hypothalamus (15.7% in rhesus, 16.4% in baboon), especially at the infundibularlip. Occasional cell bodies were scattered throughout the hypothalamic and forebrain regions studied. GAP+ fibers were widely distributed, but formed well-defined pathways such as the periventricular and ventral hypothalamic tract. In addition, GAP+ nerve terminals with various densities occurred in the lamina terminalis, the zona externa of the infundibulum, and behind the infundibular stalk. Fetal rhesus macaques had more GAP+ cell bodies, denser fiber networks, and more distinct pathways than juvenile baboons. However, fiber and terminal immunostaining was somewhat less intense for GAP than GnRH in comparable regions. These results indicate that proGnRH (GAP) is present in the same population of neurons as GnRH in the primate brain. They also suggest that post-translational products of proGnRH are present in perikarya, axons and terminals, and that GnRH and GAP and/or further cleavage products are consecreted into hypophysial portal blood in the primate. Gonadotropin releasing hormone prohormone ProGnRH GnRH-associated peptide Rhesus macaque Baboon Immunocytochemistry Peroxidase anti-peroxidase Colloidal gold immunostaining [27] and human and rat hypothalamic pre-proGnRH [1] were determined by isolating cloned genomic and cDNA coding sequences. The 92 amino acid precursor (M.W. 10 K daltons) contains a single copy of the GnRH decapeptide, which is preceded by a 23 amino acid signal peptide. GnRH is followed by a Gly-Lys-Arg amidation and enzymatic cleavage site, and a 56 amino acid residue, the GnRH-associated peptide or GAP (Fig. 1). Using in situ hybridization, Shivers and co-workers were able to identify rat forebrain neurons which contain mRNA that hybridizes with a radiolabeled, synthetic oligodeoxyribonucleotide (59-mer) complementary to human proGnRH mRNA over the region including the coding sequence for GnRH [29]. Using an antiserum raised against human proGnRH 40--53

THE hypothalamic decapeptide gonadotropin releasing hormone (GnRH), which controls pituitary gonadotropin secretion, is postulated to be biosynthesized via a conventional ribosomal route in the form of a larger GnRH prohormone (proGnRH) [12]. Higher molecular weight immunoreactive forms of GnRH have been reported in extracts of the hypothalamus and other tissues of several mammalian species including the rat [2,9], and sheep and swine [21,22]. Immunocytochemical results in the rat obtained with GnRH antisera which bind different sequences of the GnRH molecule also suggested that GnRH in neurons might be extended at either its C- or N-terminal, or at both of its terminal ends [13, 14, 34]. Recently, the structure of human placental pre-proGnRH

~Supported by NIH Grants HD 10907 (P.C.G.) and HD 11979, the Andrew W. Mellon Foundation, and a Grant from the Research Evaluation and Allocation Committee, UCSF. 2Requests for reprints should be addressed to Dr. Paul C. Goldsmith.

335

SONG ET AL.

336 ABBREVIATIONS ADH AHA AP BSA DBB DOT GAP+ GnRH+ IGS IL INF ITP LH LPOA MBH ME

dorsal hypothalamic area anterior hypothalamic area ansa peduncularis bovine serum albumin diagonal band of Broca dorsal to the optic tract GnRH associated peptide immunoreactive gonadotropin releasing hormone immunoreactive immunogold staining infundibular lip infundibulum inferior thalamic peduncle lateral hypothalamus lateral preoptic area medial basal hypothalamus median eminence

MPOA MSN NGS OT OVLT PAP PLP PMA proGnRH+ PVZ SBTI SCN SEP SON TBS VHT

medial preoptic area medial septal nucleus normal goat serum optic tract organum vasculosum of the lamina terminalis peroxidase-antiperoxidase periodate-lysine-paraformaldehyde premammillary area GnRH prohormone immunoreactive periventricular zone soybean trypsin inhibitor suprachiasmatic nucleus septal area supraoptic nucleus tris buffered saline ventral hypothalamic tract

Gly-Lys-Arg GAP

I~signal peptide-I-GnRX~ # -23

1

10 14

LR1

24

24A

40

I---

I 52 53

66

69

39A '~1 I ' ~ KN1

FIG. 1. Schematic diagram of the pre-proGnRH molecule. The solid, vertical lines indicate boundaries of the portions of preproGnRH denoted by the names and amino acid numbers appearing above. The shaded regions represent the antigenic determinants against which the antisera indicated below were raised.

(GAP 27-40), G A P has been colocalized with GnRH in rat median eminence axon terminals, and is presumably secreted into portal blood simultaneously with GnRH [25]. GAP is a potent releaser of gonadotropins, even though no copy of GnRH is present [24], and there is some evidence of separation of neuroendocrine effects within the molecule [23]. Although whole GAP has been reported to inhibit prolactin secretion from cultured rat anterior pituitary cells more effectively than dopamine [24], it is not yet certain whether G A P also represents the long sought after peptidergic PRL inhibitory hormone [36,37]. So far, G A P has been localized only in the brain of rats using immunocytochemistry [25]. Differences in the distribution of GnRH-containing neurons between rats and primates has been confirmed [10, 15, 31]. Moreover, it has been suggested that the dynamics of GnRH biosynthesis, posttranslational processing, storage and secretion, are considerably different in these two species [14]. In view of these reports, we used antisera against different amino acid sequences (rGAP 1-11 and hGAP 27-40 or rGAP 3%53) to investigate the distribution of GAP immunoreactive ( G A P + ) elements in rhesus fetuses and juvenile baboons. In this study, we report these immunostaining results, and compare them with that for GnRH immunoreactive ( G n R H + ) elements in the hope of elucidating the nature of post-translational processing of proGnRH in different groups of GnRH

neurons. A preliminary report of these results has appeared [32]. METHOD

Animals Three fetal rhesus monkeys (Macaca mulatta, two females, one male), and two female juvenile baboons (Papio anubis) were used in this study. The fetuses were delivered by cesarean section at age 109-156 days gestation in the U C S F Reproductive Endocrinology Center. The baboons, shared with other experiments, were subject to infundibular stalk transection 24 hours before sacrifice.

Tissue Preparation The brains of rhesus fetuses were removed immediately after decapitation under deep ketamine anesthesia (10 mg/kg IM and more as needed). The entire brain was immersed in a fixative containing 3% paraformaldehyde, 0.12% glutaraldehyde, 5% sucrose and 0.9% NaC1 in 0.1 M sodium cacodylate buffer at pH 7.4 [16]. After about 30 rain, a tissue block including the forebrain and hypothalamus was dissected and immersed in the same fixative, which was injected into the third and lateral ventricles with a syringe with blunt needle to enhance penetration. Tissue blocks remained in the fixative solution for 24-48 hours at 4°C.

PROGnRH I M M U N O S T A I N I N G IN PRIMATE BRAIN TABLE 1 SPECIFICITY TEST OF ANTISERABY ABSORPTION WITH GnRH AND PEPTIDE ANTIGEN FRAGMENTSOF GAP Immunostaining Results Antibody + Peptide Peptide

LR 1

24A

39A

KN 1

+ + +

+ + +

+ + + -

337 dehyde. The antigenic determinants of LR1 are amino acids 3, 4, and 7-10 of the GnRH decapeptide [38]. Working dilutions of these primary antisera were 1:2,000 for 24A, 39A and KN1, and 1:20,000 for LR1 in TBS containing 1% NGS. Incubation for 68 hours at 4°C was followed by thorough washing in TBS.

Immunostaining Techniques GnRH GAP 1-11 GAP 27-40 GAP 3%53

-

+ + +

Juvenile baboons were deeply anesthetized with ketamine (l0 mg/kg and more as needed) and given somlethol (150 mg/kg to effect). Following bilateral thoracotomy, one baboon was perfused via the ascending aorta for 20 minutes with PLP fixative (3% paraformaldehyde, l0 mM sodium periodate in 75 mM lysine-phosphate buffer, pH 6.2) [16], and the other with the same fixative as that used in the fetuses. A peristaltic pump (Manostat, New York, NY) was used to maintain a steady flow rate of 30 ml/min, and the upper part of the abdominal aorta was clamped to promote perfusion of the brain. After the brain was removed from the skull, forebrain and hypothalamus were blocked out and immersed in the same fixative as the perfusing solution for 24-48 hours at 4°C. Tissue blocks were then thoroughly washed in buffer and embedded in 2.5% agar. Serial sections at a thickness of 40 microns were cut in the frontal plane on a Lancer Vibratome (Sherwood Medical), collected in groups of 10, and stored in the fixative at 4°C until immunostained.

Immunocytochemical Staining The unlabeled peroxidase-antiperoxidase (PAP) technique [33] and colloidal gold immunostaining (IGS) method [5] were performed as previously described in detail [34]. In brief, tissue sections were sequentially pretreated with 1% sodium borohydride, 20 mM lysine, 0.1% saponin plus 0.5% H202, and 3% normal goat serum (NGS) prepared in 0.2 M Tris buffered saline (TBS) at pH 7.6 to suppress non-specific staining and to promote penetration of subsequent immunoreagents.

Antisera The antigenic domains of the primary antisera used are presented schematically in Fig. 1. Antiserum 24A (3rd bleed) was produced in rabbit against the 11 amino acid N-terminal of rat G A P (rGAP 1-11 or proGnRH 14-24) conjugated to bovine serum albumin (BSA) and soybean trypsin inhibitor (SBTI). Antiserum 39A (3rd bleed) was produced against amino acids 40-53 from the mid-region of human proGnRH (hGAP 27-40) conjugated to BSA. Antisera KN1 was also generated in rabbit against a synthetic peptide containing residues 52-66 from the C-terminal of rat proGnRH (rGAP 3%53) conjugated to SBTI. Before use, antiserum 24A was routinely absorbed with 5% BSA and 1% SBTI in TBS, while antisera 39A and KN1 were absorbed with 5% BSA and 1% SBTI, respectively, to diminish non-specific staining. Antiserum LR1 (bleed 20-6-79) (generously provided by Dr. R. Benoit, The Salk Institute) was raised in rabbit against [D-Lys6]-GnRH conjugated to ovalbumin with glutaral-

For the PAP method, sections were incubated in goat anti-rabbit IgG and then rabbit PAP, both diluted 1:100 in TBS containing 1% NGS, for 30 minutes at 23°C, each followed by 3 washes in buffer. The peroxidase reaction product was developed with 0.05% DAB and 0.01% H202 in TBS for 10 minutes. For the IGS method, sections were first washed in 0.1% BSA-20 mM TBS, pH 8.2. They were then incubated for 2-4 hours at 23°C in 1:4 goat antirabbit IgG coupled to 15 nm colloidal gold particles (250 mg IgG/ml) in 1% BSA-20 mM TBS, pH 8.2. The development of pink to red cell bodies and fibers was monitored under the light microscope. When ideal staining was obtained, the sections were washed twice on 0.1% BSA-20 mM TBS and then twice with TBS.

Light Microscopy Sections stained with both the PAP and IGS methods were mounted on gelatin coated glass slides, dehydrated in ascending ethanols, cleared in xylene and coverslipped in Permount. Observations and photographs were made using a Leitz Ortholux II microscope equipped with a VarioOrthomat Camera. The number of G n R H + and G A P + neurons in the forebrain and hypothalamus were counted in comparable vibratome sections (every tenth 40 micron section from the septum to the anterior premamillary area). These represented 1/10 of the total immunopositive neurons per animal [26]. Statistical significance was determined using Student's t-test.

Immunostaining Controls Specificity of immunostaining was verified using liquid phase absorption and cross absorption of primary antisera with the synthetic peptide antigen sequences at 4°C for 24 hours before use. Antisera 24A, 39A and KN1 were incubated with 1 /zg GAP 1-11, 27-40, and 3%50 or 5 /~g LHRH/ml 1:2,000 dilution. Likewise, antiserum LR1 was absorbed with 1/xg L H R H or 5/xg of each G A P fragment/ml 1:20,000 dilution. Technical controls for each method included substitution of normal rabbit serum at equivalent concentration or TBS for the primary antisera, and of NGS or TBS for the second antibody. RESULTS All primary antibodies used in this study are highly specific for the peptide antigens against which they were raised. Immunostaining for GnRH or for the N-terminal, midregion, or C-terminal portions of GAP was blocked by liquid phase absorption with their respective peptide antigens (Table 1). However, immunostaining for GnRH with 1:20,000 antiserum LR1 was not affected by liquid phase absorption with 5/~g of any of the G A P fragments/ml, nor could 5 p,g GnRH/ml eliminate the reaction product produced with 1:2,000 of the GAP antisera 24A, 39A or KN1. Substitution

338

S O N G ET A L

FIG. 2. Morphology of proGnRH neurons stained with antisera 24A (a, c, e and h), KN1 (b, d and f) and 39A (g). (a) Fusiform bipolar cell with two thick processes in the ITP of the baboon. (PAP) x490. (b) Fusiform cell with a beaded axon in the MPOA of the fetal rhesus macaque. (PAP) x490. (c) Unipolar cell with a beaded fiber in the SON region of the fetal rhesus macaque. (1GS) x490. (d) Tripolar cell with a beaded axon emanating from the cell body in the DBB of the fetal rhesus macaque. (PAP) ×490. (e) Round unipolar cell in the SON re#on of the baboon. (PAP) x 490. (f) Round and oval bipolar cells with one thick and one thinner process in the MPOA of the fetal rhesus macaque. (PAP) x490. (g) A tripolar cell extends a process (arrow) contacting that of a bipolar cell in the IL of the baboon. (PAP) x 490. (h) Granular reaction product occurs in the perikaryon, but not in the nucleus of a cell in the MPOA of the fetal rhesus macaque. (IGS) x 1225.

P R O G n R H I M M U N O S T A I N I N G IN P R I M A T E B RA I N

FIG. 3. The same distribution of PAP-stained perikarya and fibers is seen with antiserum 24A (a) and with antiserum KN1 (b) in the septal-MPOA, and with antiserum 24A (c) and antiserum LR1 (d) in the DBB in the fetal rhesus macaque, x 123.

339

?IG. 4. Perikarya and fibers stained with antiserum 24A and PAP (a and b, left and right sides of the same section) and antiserum LR1 (c and d, left and ight sides of adjacent section) in the infundibular region of the fetal rhesus macaque. (a and b) Groups of neurons routinely occur at the infundibular lip IL). Only a few cells (asterisk) are within the arcuate nucleus (ARC) while a few are seen in the periventricular zone. (c and d) Two fiber pathways, the ~entral hypothalamic tract along the ventral brain surface, and the periventricular pathway along the third ventricle (III), project into the lateral and nedial parts of the infundibulum, respectively. Fewer proGnRH+ (a,b) cell bodies and fibers are found than those GnRH+ (c,d). Note the many fibers trcing through the ARC region (c,d). x 123.

PROGnRH

IMMUNOSTAINING

IN PRIMATE

BRAIN

FIG. 5. (a) Dense proGnRH+ fibers and fiber terminals stained with antiserum 24A and PAP occur in the lamina terminals (frontal section) of the fetal rhesus macaque. Some cell bodies (arrows) are also present in this structure. III, supraoptic recess of the third ventricle. x 196. (b) ProGnRH+ fiber terminals stained with antiserum KNl and PAP are located in the perimeter of the infundibulum of the fetal rhesus macaque (obliquely transected). x 140.

of the first or second antibodies with corresponding normal sera or TBS also completely abolished the immunocytochemical staining. Although specific immunolocalization of proGnRH occurs with all three GAP antisera, differences in their binding affinities was apparent. Antisera 24A and KNl gave better results, staining more GAP perikarya and fibers and with greater intensity than antiserum 39A. Immunostaining of cell bodies and fiber pathways was always better with the PAP

341

technique than with IGS method, due apparently to technical factors rather than the avidity of the primary antisera [16]. Using the IGS method, GAP+ perikarya and a few beaded fibers were best seen in fetal animals (Fig. 2c), and to a lesser extent in the baboon perfused with PLP fixative. In general, the best immunostaining was obtained in fetuses with antisera 24A and KNl, when compared at dilutions of 1:2,000 with the PAP technique. Granular immunoreaction products identifying GAP were seen throughout neuronal perikarya and processes, including both dendrites and axons, as well as in axon terminals, using both techniques (Fig. 2). GAP+ cells were small, and most were fusiform and bipolar in shape. Uusally, both bipolar processes were thick, or one somewhat thicker than the other, and non-beaded, suggesting they both were dendrites (Fig. 2a). Sometimes, a thin, beaded axon emanated from one pole (Fig. 2b). Both axons and dendrites could be traced for long distances. Rounded cell bodies were most frequently seen in the supraoptic nuclear (SON) region and medial preoptic area (MPOA) (Figs. 2c, e, f), which appeared unipolar or bipolar in the frontal plane. A few multipolar, especially tripolar, cells were observed in cell rich areas such as MPOA and infundibular lip (IL) (Fig. 2d,g), and apparent interneuronal contacts were occasionally found between them (Fig. 2g). GAP+ perikarya were found throughout the forebrain and hypothalamic regions studied. In the forebrain, most proGnRH cell bodies were present near midline structures such as the medial septal nucleus (MSN), anterior part of the MPOA (Fig. 3a,b), and organum vasculosum of the lamina terminalis (OVLT). They were mostly oriented dorsoventrally. Posteriorly, the cell bodies moved ventrolaterally from the MPOA to the nucleus of diagonal band of Broca (DBB) (Fig. 3c,d), with a few cells extending into the horizontal part o1 the DBB. GAP+ neurons in these areas accounted for 44.6% and 46.9% of the total GAP+ cells in the rhesus and baboon, respectively. A few GAP+ perikarya were also found in the lateral septal nucleus, the bed nucleus of the stria terminalis, and vertical part of the DBB. No GAP+ cell bodies were observed in the olfactory tubercle. In the hypothalamus, most GAP+ perikarya were observed in the SON region including the area dorsal to the optic tracts (DOT; rhesus 21.%, baboon 23.4%), and in the medial basal hypothalamus (MBH; rhesus 15.7%, baboon 16.4%), especially at the IL (Fig. 4). They occurred in clusters or groups of 2-4 cells, usually aligned with the fiber pathways. Around the SON, cell bodies had various orientations corresponding to their projections, and a few neurons were within the SON itself. More medially, numbers of GAP+ cell bodies increased, and most of them were located immediately above the lateral aspects of the OT, and oriented parallel to it. These neurons formed a continuum with those at the IL, where the neurons approached and ran parallel to the ventral hypothalamic surface (Fig. 4). GAP+ cell bodies were sparse midway between the DOT and the IL. A few perikarya were present in the arcuate (infundibular) nucleus (Fig. 4b), especially in its periventricular part (Fig. 4a,b). Some GAP+ perikarya were observed in the perventricular zone of the third ventricle throughout its rostrocaudal length, particularly in the preoptic (Fig. 7a) and infundibular (INF) regions (Fig. 4). These neurons usually paralleled the ventricular wall and were more numerous in the ventral portions. More laterally, some GAP+ neurons were seen in the lateral hypothalamic area. Most of them were in the area dorsal to the SON, and within the ansa

S O N G ET A L

342

FIG. 6. (a) ProGnRH+ (antiserum 24A) fibers in the retrochiasmatic region, just before the infundibulum. The fibers mainly radiate from the periventricular pathway, x 123. (b and c) ProGnRH+ fibers stained with antisera KN1 (b) and LR1 (c) and PAP from the terminations of both the ventral hypothalamic and periventricular tracts join and form a dense band in the midline at the ventral surface at the premamillary level. They have the same general appearance, except that fibers and the fiber band stained for GnRH are denser than those stained for proGnRH. III, third ventricle. × 196.

PROGnRH IMMUNOSTAINING

IN PRIMATE BRAIN

FIG. 7. ProGnRH fibers and fiber networks stained with antiserum 24A and PAP in (a) the preoptic periventricular zone, (b) paraventricular nucleus, (c) lateral premamillary area, and (d) the small round (unnamed) area just dorsal and slightly lateral to the top of the third ventricle at the premamiUary level. Arrows in b and c show occasional cell bodies in these areas. III, third ventricle. (a, b, d) x196. (c) ×123.

343

344

SONG ET AL.

100

.= 40

7=

20 0 SEP OVLT MPOA DBB SCN LPOA AHA SON DOT LH ADH PVZ MBH PMA BRAIN REGIONS

t~

--

b

SEP

OVLT

MPOA DBB

'"

SCN

~c

-

AHA

-

z -

< ,.r m

-

-

- - S O N

--

PVZ DOT

ADH PMA

MBH

80-

.

50 •

(/)

-

51

40"

Z

30-

ELI Z

t0,1

O

2

/

, - .," ~'A# 44

40

40 •

28 "A.

12

4.4

DISCUSSION

-a, 16

4

3.6

3.2

2.8

2.4

2

1,6

1.2

in the septo-MPOA region. Anterior to the I N F , these fibers projected downward through the retrochaismatic area (Fig. 6a). Behind the I N F , the fibers joined with those of the caudal VHT and formed a dense fiber band in the midline at the ventral brain surface (Fig. 6b,c). Less dense, parallel fibers were also seen within the AP at anterior levels and the ITP at posterior levels. Other accumulations of G A P + fibers with various orientations were also seen in the paraventricular nucleus (Fig. 7b), anterior premamillary area (PMA) near the brain surface (Fig. 7c), and the small round areas just dorsal and slightly lateral to the top of the third ventricle at the level of the PMA (Fig. 7d). No fiber network in this latter area could be found in the juvenile baboon. A few fibers were detected in some nearby extrahypothalamic areas including the medial amygdaloid nucleus, bed nucleus of the stria terminalis, the area around the bottom of the lateral ventricle at the preoptic level, and the periventricular nucleus of the thalamus. The distribution of G A P + perikarya and fibers was the same as that of G n R H + immunostained elements on adjacent or comparable sections (Fig. 3c and d; Fig. 4a,b and c,d; Fig. 6b and c). Although the number of cell bodies which were G n R H + was slightly greater than that for G A P (452 vs. 413 for 12 comparable 40 micron sections per fetal rhesus, n=3) (Fig. 8a,b), the difference is not statistically significant (p>0.80). In general, the density and staining intensity of G A P + fibers was less than those of G n R H + fibers in comparable brain regions. G n R H + axon terminals in the I N F and ME were also denser and more intensely stained than G A P + terminals.

0.8

DISTANCE ANTERIOR FROM MAMMiLLARY BODY

~

0.4

3

0

(ram.}

FIG. 8. (a) Numbers of GAP (proGnRH, 24A) and GnRH (LR1) immunoreactive neurons in various brain regions of the fetal rhesus macaque. (b) Distribution of GAP (proGnRH, 24A) and GnRH (LR1) immunopositive neurons/40 micron section in adjacent frontal sections from different levels of the fetal rhesus macaque brain (156 days gestation). The extent of corresponding brain regions is indicated above.

peduncularis (AP), inferior thalamic peduncle (ITP), or medial forebrain bundle. G A P + fibers were also present all over the forebrain and hypothalamus, but they were more numerous in cell-rich areas, and formed several well-defined pathways. Rostrally, dense beaded fibers ran dorsoventrally in the midline, on both sides of the MSN and MPOA, and in the preoptic periventricular zone (Fig. 3a,b). Some of these fibers extended into and ended in the lamina terminalis (Fig. 5a). Others extended ventrolaterally to the DBB, where various fiber orientations were seen (Fig. 3c). A distinct fiber pathway, which we previously identified as the ventral hypothalamic tract (VHT) by GnRH immunostaining [10], was clearly seen extending ventromedially from the SON region to the I L (Fig. 4a,b). VHT fibers entered the perimeter of the INF, and terminated in the zona externa or radiated toward the center of the stalk (Fig. 5b). Another G A P + fiber pathway entering the I N F was the periventircular pathway. Its fiber density was less than that of the VHT and less than that

This study demonstrates that the C-terminal portion of proGnRH, the GnRH-associated peptide or GAP, is present in a substantial number of forebrain and hypothalamic neurons of the fetal rhesus monkey and baboon. G A P + labeling was evident in all parts of reactive neurons, especially using either antiserum 24A (anti-N-terminal sequence of GAP) or KN1 (anti-C-terminal sequence of GAP). Specific immunostaining was completely eliminated by substitution of normal rabbit serum or TBS for the primary antisera. G A P + staining was blocked by immunoabsorption of the antisera with synthetic GAP, but not with GnRH. Moreover, no sequence homologies have been found between G A P and the other known hypothalamic precursors to neuroendocrine peptides [27]. Therefore, it is clear that the immunostained antigen represents GAP or proGnRH itself, but not the decapeptide GnRH. The morphology and distribution of G A P + neurons are indistinguishable from those of GnRH as shown in the present study. There is also good agreement with previous results for GnRH immunostaining obtained in the baboon [10,20] and the rhesus macaque [14,30]. As for GnRH, most G A P + cell bodies occurred in the septo-MPOA-DBB, SON, and I L regions, although occasional perikarya were seen in almost all forebrain and hypothalamic areas studied, as well as in some nearby extrahypothalamic sites. Furthermore, the presence of many more G A P + perikarya in the baboon and rhesus MBH than in the rat [25] confirms this fundamental species difference in the distribution of GnRH-containing neurons. In the individual animals studied here, totals of G n R H + perikarya were very close to those of G A P + perikarya (Fig. 8b). Slight variations may have resulted from a somewhat

PROGnRH

IMMUNOSTAINING

IN PRIMATE

345

BRAIN

heterogenous distribution of proGnRH/GnRH neurons in brain tissue, from antibody avidities, or from dissimilar tolerance of antigens to fixatives and the immunostaining procedures. Somewhat larger regional differences in numbers of GnRH+ versus GAP+ neurons were noted in the DBB and septal areas (Fig. 8a). The reasons for these immunostaining results are not clear, but they may be related to the neuroendocrine versus neuroregulatory nature of the neurons involved [20]. Overall, however, there is very good coincidence in the distribution between GAP+ and GnRH+ neurons (see Fig. 3c,d; Fig. 4; Fig. 6b,c; Fig. 8). Therefore, these data indicate that GAP is present in the same population of neurons as GnRH. It has been suggested that hypothalamic and placental proGnRH may be the same [ 11, and that human and rat share a single gene coding for this precursor protein [25]. Our results in non-human primates showing immunostaining of hypothalamic neurons with polyclonal antibodies raised against synthetic peptide fragments of human and especially rat proGnRH support and extend this suggestion to non-human primates as well. GAP+ fiber staining was observed in the septal, periventricular and VHT fiber pathways, and in the OVLT and ME terminal fields of primates. These results extend those in the rat showing that GAP and GnRH coexist in ME [251. Our observations also indicate that GAP is delivered via the same system as GnRH, and is coreleased with GnRH into portal blood. In addition, as we postulated for GnRH [20], the presence of GAP perikarya and fibers in extrahypothalamic regions and GAP+ staining in rat cerebral cortex (unpublished observation) suggests other functions of GAP as a neurotransmitter or neuromodulator [25]. It is well established that the magnocellular hypothalamic peptides, vasopressin and oxytocin, are synthesized with their associated neurophysins in common precursor molecules, respectively [3]. The nucleotide sequence of cloned cDNA encoding the bovine vasopressin/neurophysin precursor (propressophysin) has been determined [ 181. The structure of the oxytocin/neurophysin precursor is similar to that of propressophysin [17]. The pre-prohormone is produced in the rough endoplasmic reticulum and packaged into secretory granules in the Golgi complex. Post-translational processing of the prohormone then occurs in the neurosecretory granules, and the cleaved peptide products are co-released [4,8]. This mechanism of biosynthesis of peptidergic hormones is also apparently shared by the parvocellular hypothalamic peptides, as determined from the structure of the precursors for somatostatin [28], corticotropin-releasing factor [7], growth hormone-releasing factor [ll], and GnRH [27]. Prosomatostatin [19] and proGnRH [25] have both been shown to be present in rat brain with a distribution indistinguishable from that of somatostatin-14 and GnRH, respectively. Results of this study in primates, together with other immunostaining data, suggest that enzymatic cleavage of proGnRH occurs primarily in the neuronal cell body. It has

been reported that antisera which recognize cyclized N- and amidated C-terminals of GnRH do not immunostain the hormone in perikaryal RER or the Golgi apparatus [14]. It was therefore also suggested that LRl, with antigenic determinants of amino acids 3,4 and 7-10, might not recognize GnRH in its precursor form with the extension of GAP at its C-terminal [38]. However, antiserum LRl produces dense staining of GnRH neurosecretory granules throughout immunoreactive neurons. This implies that proGnRH processing occurs soon after synthesis in cell bodies. Intense perikaryal labeling with antiserum 24A against the GAP N-terminal provides little further information in this regard, since the binding requirements of 24A have not been rigorously characterized. However, recent results with an antiserum against a synthetic peptide fragment consisting of proGnRH 6-16 (including the GnRH-GAP cleavage site) have been very informative. Immunostaining with this antiserum, which does not recognize either GnRH or the N-terminal of GAP, is restricted to neuronal perikarya and proximal processes (0. K. Ronnekliev, personal communication). Since axons and terminals do not stain, enzymatic cleavage appears essentially complete prior to anterograde transport. Comparison of immunostaining results also indicated several differences between the GAP antisera. Perikaryal immunostaining was most intense with 24A, comparable to that with antiserum LRI against GnRH. Staining with KNl was less intense than 24A, but 39A gave weak cell body staining, and even poorer fiber staining in primates. Since antiserum 39A provided good staining in rat brain [25], this may reflect differential sensitivity of this antibody against the mid-region of GAP in brain tissue from different animal species. Further comparison of the GAP+ results indicated that the antisera stained different portions of neurons to different degrees. While GAP+ staining intensity in perikarya was generally greater with 24A than KNl, the difference was even more marked in nerve fibers (e.g., the VHT), and especially axon terminals (e.g., the ME). This suggests the possibility that GAP might undergo conformational change or further processing during axonal transport. Examples of such types of peptide processing, even at single basic amino acids, have been reported for several other peptide precursors [6]. However, since all three GAP antisera used in this study immunostain all parts of reactive neurons, each of their antigenic sequences, and indeed the entire GnRHassociated peptide, may be secreted into primate hypophysial portal blood.

ACKNOWLEDGEMENTS

The authors thank Claudia Schumann for her help in the preparation of this manuscript. The authors are also grateful to the Reproductive Endocrinology Center for supplying some of the animals used in this study.

REFERENCES 1. Adelman, J. P., A. J. Mason, J. S. Hayflick and P. H. Seeburg. Isolation of the gene and hypothalamic cDNA for the common precursor of gonadotropic-releasing hormone and prolactin release-inhibiting factor in human and rat. Proc Nat/ Acad Sci USA 83: 379-183, 1986.

2. Bamea, A. and J. C. Porter. Demonstration of a macromolecule cross-reacting with antibodies to luteinizing hormone releasing hormone and its tissue distribution. Biochem Biophys Res Commun 67: 1346-1352. 1975.

346

3. Brownstein, M. J. Biosynthesis of vasopressin and oxytocin. Annu Rev Physiol 45: 12%135, 1983. 4. Brownstein, M. J., J. T. Russell and H. Gainer. Synthesis, transport, and release of posterior pituitary hormones. Science 207: 373-378, 1980. 5. De May, J. Colloidal gold probes in immunocytochemistry. In: lmmunocytochemistry. Practical Applications in Pathology and Biology, edited by J. M. Polak and S. V. Noorden. Bristol: Wright.PSG, 1983, pp. 82-112. 6. Douglass, J., O. Civelli and E. Herbert. Polyprotein gene expression: Generation of diversity of neuroendocrine peptides. Annu Rev Biochem 53: 665-715, 1984. 7. Furutani, Y., Y. Morimoto, S. Shibahara, M. Noda, H. Takahashi, T. Hirose, M. Asai, S. Inayama, H. Hayashida, T. Miyata and S. Numa. Cloning and sequence analysis of cDNA for ovine corticotropin-releasing factor precursor. Nature 301: 537-540, 1983. 8. Gainer, H., Y. P. Loh and E. A. Neale. The organization of post-translational precursor processing in peptidergic neurosecretory cells. Prog Clin Biol Res 79: 131-145, 1982. 9. Gautron, J. P., E. Pattou and C. Kordon. Occurrence of higher molecular forms of LHRH in fractionated extracts from rat hypothalamus, cortex and placenta. Mol Cell Endocrino! 24: 1-15, 1981. 10. Goldsmith, P. C., R. Lamberts and L. R. Brezina. Gonadotropin-releasing hormone neurons and pathways in the primate hypothalamus and forebrain. In: Neuroendocrine Aspects o f Reproduction, vol editor R. L. Norman, vol 2, ORPRC Symposia on Reproductive Biology, series editors R. M. Brenner and C. H. Phoenix. New York: Academic Press, 1983, pp. 7-45. 11. Gubler, U., J. J. Monahan, P. T. Lomedico, R. S. Bhatt, K. J. Collier, B. J. Hoffman, P. Bohlen, F. Esch, N. Ling. F. Zeytin, P. Brazeau, M. S. Posnian and L. P. Gage. Cloning and sequence analysis of cDNA for the precursor of human growth hormone-releasing factor, somatocrinin. Proc Natl Acad Sci USA 80: 4311~,314, 1983. 12. Jutisz, M., R. Connis and M. Corbani. Biosynthesis of gonadotropin releasing hormone (GnRH): Present status. Psychoneuroendocrinology 8: 251-258, 1983. 13. King, J. C. and E. L. P. Anthony. Biosynthesis of LHRH: Inferences from immunocytochemical studies. Peptides 4: 963-970, 1983. 14. King, J. C., E. L. P. Anthony, D. M. Fitzgerald and E. G. Stopa. Luteinizing hormone-releasing hormone neurons in human preoptic/hypothalmus: Differential intraneuronal localization of immunoreactive forms. J Clin Endocrinol Metab 60: 88-97, 1985. 15. Kozlowski, G. P. and W. L. Dees. lmmunocytochemistry for LHRH neurons in the arcuate nucleus area of the rat: Fact or artifact? J Histochem Cytochem 32: 83-91, 1984. 16. Lamberts, R. and P. C. Goldsmith. Fixation, fine structure, and immunostaining for neuropeptides: Perfusion versus immersion of the neurendocrine hypothalamus. J Histochem Cytochem 34: 389-398, 1986. 17. Land, H., M. Grez, S. Rupport, H. Schmale, M. Rehbein, D. Richter and G. Schultz. Deduced aminoacid sequence from the bovine oxytocin-neurophysin I precursor cDNA. Nature 302: 342-344, 1983. 18. Land, H., G. Schutz, H. Schmale and D. Richter. Nucleotide sequence of cloned cDNA encoding bovine arginine vasopressin-neurophysin II precursor. Nature 295: 299-303, 1982. 19. Lechan, R., R. H. Goodman, M. Rosenblatt, S. Reichlin and J. F. Habener. Prosomatostatin-specific antigen in rat brain: Localization by immunocytochemical staining with an antiserum to a synthetic sequence of preprosomatostatin. Proc Natl Acad Sci USA 80: 2780-2784, 1983. 20. Marshall, P. E. and P. C. Goldsmith. Neuroregulatory and neuroendocrine GnRH pathways in the hypothalamus and forebrain of the baboon. Brain Res 193: 353-372, 1980.

SONG ET AL. 21. Millar, R. P., C. Aehnelt and G. Rossier. Higher molecular weight immunoreactive species of luteinizing hormone releasing hormone: Possible precursor of the hormone. Biochem Biophys Res Commun 74: 720-731, 1977. 22. Millar, R. P., P. Denniss, C. Tobler, J. C. King, A. V. Schally and A. Arimura. Presumptive prohormonal forms of hypothalamic peptide hormones. In: Cell Biology o f Hypothalamic Neurosecretion. Colloque International du CNRS No. 280, edited by J. D. Vincent and C. Kordon. Paris: CNRS, 1978, pp. 487510. 23. Millar, R. P., P. J. Wormald and R. C. de L. Milton. Stimulation of gonadotropin release by a non-GnRH peptide sequence of the GnRH precursor. Science 232: 68--70, 1986. 24. Nikolics, K., A. J. Mason, E. Szonyi, J. Ramachandran and P. H. Seeburg. A prolactin-inhibitory factor within the precursor for human gonadotropin-releasing hormone. Nature 316" 511517, 1985. 25. Phillips, H. S., K. Nikolics, D. Branton and P. H. Seeburg. Immunocytochemical localization in rat brain of a prolactin releasing-inhibiting sequence of gonadotropin-releasing hormone prohormone. Nature 316: 542-545, 1985. 26. Ronnekliev, O. K. and M. J. Kelly. Luteinizing hormone releasing hormone neuronal system during the estrous cycle of the female rat: Effects of surgically-induced persistent estrus. Neuroendocrinology 43: 564-576, 1986. 27. Seeburg, P. H. and J. P. Adelman. Characterization of cDNA for precursor of human luteinizing hormone releasing hormone. Nature 311: 666-668, 1984. 28. Shen, L.-P., R. L. Pictet and W. J. Rutter. Human somatostatin I: Sequence of the cDNA. Proc Natl Acad Sci USA 70: 45754579, 1982. 29. Shivers, B. D., R. E. Harlan, J. F. Hijtmancik, P. M. Conn and D. W. Pfaff. Localization of cell containing LHRH-like mRNA in rat forebrain using in situ hybridization. Endocrinology 118: 883-885, 1986. 30. Silverman, A. J., J. L. Antunes, G. M. Abrams, G. Nilaver, R. Thau, J. A. Robinson, M. Ferin and L. C. Krey. The luteinizing hormone-releasing hormone pathways in rhesus (Macacca mulatta) and pigtailed (Macacca nemestrina) monkeys: New observations on thick, unembedded sections. J Comp Neurol 211: 30%317, 1982. 31. Silverman, A. J., L. C. Krey and E. Z. Zimmerman. A comparative study of the luteinizing hormone releasing hormone (LHRH) neuronal networks in mammals. Biol Reprod 20: 98110, 1979. 32. Song, T., K. Nikolics, P. H. Seeburg and P. C. Goldsmith. GnRH prohormone immunoreactivity in the primate forebrain and hypothalamus. First International Congress of Neuroendocrinology, p. 89, 1986 (A-256). 33. Sternberger, L. A. lmmunocytochemistry. Englewood Cliffs, N J: Prentice-Hall, 1974. 34. Sternberger, L. A. and G. E. Hoffman. Immunocytology of luteinizing hormone-releasing hormone. Neuroendocrinology 25: 111-128, 1978. 35. Thind, K. K. and P. C. Goldsmith. Ultrastructural analysis of synapses involving tyrosine hydroxylase-containing neurons in the ventral periventricular hypothalamus of the macaque. Brain Res 366: 37-52, 1986. 36. Vijayan, E. and S. M. McCann. The effect of intraventricular injection of gamma-aminobutyric acid (GABA) on prolactin and gonadotropin release in conscious female rats. Brain Res 155: 35-42, 1978. 37. Weiner, R. I. and W. F. Ganong. Role of brain monoamines and histamines in regulation of anterior pituitary secretion. Physiol Rev 58: 905-976, 1978. 38. Witkin, J. W. and A. J. Silverman. Synaptology of luteinizing hormone-releasing hormone neurons in rat preoptic area. Peptides 6: 263-271, 1985.