Ovarian regulation of kisspeptin neurones in the arcuate nucleus of the rhesus monkey (Macaca mulatta)

Abstract

Tonic gonadotrophin secretion throughout the menstrual cycle is regulated by the negative feedback actions of ovarian oestradiol (E₂) and progesterone (P). While kisspeptin neurones in the arcuate nucleus (ARC) of the hypothalamus appear to play a major role in mediating these feedback actions of the steroids in non-primate species, this issue has been less well studied in the monkey. Here, we used immunohistochemistry (IHC) and in situ hybridization (ISH) to examine kisspeptin and KISS1 expression, respectively, in the mediobasal hypothalamus (MBH) of adult ovariectomised (OVX) rhesus monkeys. We also examined kisspeptin expression in the MBH of ovarian intact females, and the effect of E₂, P and E₂+P replacement on KISS1 expression in OVX animals. Kisspeptin or KISS1 expressing neurons and pronounced kisspeptin fibres were readily identified throughout the ARC of ovariectomised monkeys, but in intact animals on the other hand kisspeptin cell bodies were small in size and number and only fine fibers were observed. Replacement of OVX monkeys with physiologic levels of E₂, either alone or with luteal phase levels of P, abolished KISS1 expression in the ARC. Interestingly, P replacement alone for 14 days also resulted in a significant downregulation of KISS1 expression. These findings support the view that, in primates, as in rodents and sheep, kisspeptin signaling in ARC neurones appears to play an important role in mediating the negative feedback action of E₂ on gonadotrophin secretion, and indicate a need to further study their regulation by P.

Introduction

Tonic or basal gonadotrophin secretion throughout the greater part of the follicular and luteal phases of the menstrual cycle is regulated primarily by the negative feedback actions of ovarian oestradiol (E₂) and progesterone (P) (1-3). These ovarian steroids exert their actions either directly at the anterior pituitary or indirectly at a hypothalamic level to regulate pulsatile release of GnRH into the portal circulation. There is compelling evidence for a hypothalamic site of the negative feedback action of E₂ on LH secretion in the female primate. Administration of this steroid to adult ovariectomised (OVX) monkeys induces a suppression of GnRH release in the region of the pituitary stalk and median eminence with a latency of 1 to 2h (4), and an inhibition of multi-unit electrical activity (MUA) in the mediobasal hypothalamus (MBH) (5,6): a robust correlate of pulsatile LH release and presumably that of GnRH (7). Additionally, injections of E₂ directly into the hypothalamus or third cerebroventricle result in a decline of LH secretion in OVX rhesus monkeys (8,9).

In the case of P, an important site of action of this steroid in the female primate is also at the level of the hypothalamus. This action of P is indicated by the findings that LH pulse frequency, which is determined by the frequency of pulsatile GnRH release (7), is retarded during the luteal phase of the menstrual cycle when P secretion is elevated (10,11) and by exogenous P administration in the follicular phase (12).

Over the last decade, kisspeptin has come to be recognized as a major regulator of the GnRH neuronal network in mammalian species (13-18). Kisspeptin neurones are located in the rostral hypothalamus and MBH of the female brain in both primate and non-primate species: with those in the latter brain region located in the arcuate nucleus (ARC) (19-23). However, the distribution of kisspeptin peptide expressing neurons throughout the MBH has not been reported for the female monkey.

In non-primate species, the ARC kisspeptin neurons in females are considered to play a major role in mediating the negative feedback action of E₂ on LH secretion (16), and the same is likely the case for the monkey and human. The latter view is based primarily upon the observations of Rance and her colleagues (24) showing that expression in the ARC of the gene that encodes for kisspeptin (KISS1) is low in premenopausal women and monkeys, or in E₂ treated OVX cynomolgus monkeys, but elevated in postmenopausal women or untreated OVX macaques.

The role of ARC kisspeptin neurons in mediating the negative feedback action by P has been less extensively studied than that of E₂, and reports are inconsistent. In OVX rats, P treatment in combination with E₂ suppressed hypothalamic Kiss1 mRNA levels when assessed by semi-quantitative RT-PCR, but treatment with P alone was without affect (25). In one study of OVX sheep employing in situ hybridization (ISH), P was reported to inhibit Kiss1 expression in ARC (26), but in a second and currently unpublished study by the same group of investigators Kiss1 mRNA levels in the ARC and median eminence as determined by qRT-PCR were not effected by this steroid (Robert L. Goodman, personal communication). The action of P on KISS1 expression in the ARC of the primate brain had not been studied directly.

For the foregoing reasons, we performed two separate but related studies. One systematically compared the distribution of immunohistochemically identified kisspeptin neurones in the hypothalamus of intact and OVX female rhesus monkeys. The other examined the effects of physiological replacement with P, alone, or in combination with E₂, to adult OVX rhesus monkeys on KISS1 mRNA expression in the ARC as determined by ISH, using hypothalamic tissue that had been generated during a previous study describing the effect of ovarian steroids on expression of monoamine oxidase mRNAs in the brain of this macaque (27).

Materials and methods

Tissue

The tissue used for the ISH study was generated for an earlier experiment that was conducted at the Oregon National Primate Research Center (ONPRC), and which has been previously described in detail (27). Briefly, adult hysterectomised/OVX rhesus monkeys were divided into four experimental groups, each comprised of five animals and implanted with either empty (Control) or E₂-containing Silastic capsules for 28 days (OVX+E₂), or with empty Silastic capsules for 28 days and P-containing capsules for the last 14 days (OVX+P) or with E₂-filled capsules for 28 days and P-containing capsules for the last 14 days (OVX+E₂+P). Mean serum concentrations of E₂ and P as measured by RIA on the last day of treatments were reported at the time of the original publication (27) as 95 ± 26 pg/ml and 9.6 ± 1.1 ng/ml for E₂ and P treated animals, respectively. Corresponding values for animals implanted with only empty Silastic capsules were 6 pg/ml and 0.2 ng/ml (E₂ and P, respectively). The brains of these animals had been fixed by transcardial perfusion with 4% paraformaldehyde at the end of treatment. Coronal 25 μm hypothalamic sections (collected at approximately every 250 μm) had been cut on a sliding microtome, mounted on Superfrost Plus slides, and stored at −80°C. They were then shipped on dry ice to the Magee-Womens Research Institute, where they were stored at −80°C until analyzed.

For assessing the effect of the steroid treatments on gonadotrophin secretion, additional OVX rhesus monkeys (N=8 per group) that had received steroid replacements at ONPRC identical to those administered to the animals used for ISH were studied. This approach was necessary because serum samples from the animals used for ISH were expended. Accordingly, serum samples collected on the last day of treatment were assayed for LH using a homologous (macaque) RIA previously described (28). Recombinant monkey LH (NHPP/NICHD-rec.mo.LH-RP-1, AFP6936A) was employed as standard. The intra-assay coefficient of variation and the sensitivity of the one assay employed were <6% and 0.3 ng/ml, respectively. Circulating steroid levels in this second group of animals were determined by the Endocrine Technology and Support Core at ONPRC using an Immulite 2000 (a chemiluminescence-based automatic clinical platform (Siemens Healthcare Diagnostics, Deerfield, IL) or a Roche Elecsys 2010 analyzer (also a chemiluminescence-based clinical platform by F. Hoffmann-La Roche Ltd, Basel Switzerland). The sensitivities of the E₂ determinations by the 2 platforms were 20 and 5 pg/ml, respectively, and for P 0.2 and 0.04 ng/ml, respectively. The intra- and inter-assay variations with the Immulite 2000 and Elecsys platforms were less than 15% for both assays, and the coefficient of correlations between the two platforms for both assays were >0.93. Circulating steroid levels in the second group of animals were as follows: E₂ in E₂-treated animals, 103±10 pg/ml; P in P-treated animals, 4.9±1.5 ng/ml; E₂ and P in animals implanted with empty capsules, 18±5 pg/ml and 0.13±0.02 ng/ml, respectively. For samples with undetectable levels of steroid, the sensitivity of the respective assay was assigned.

For immunohistochemistry (IHC), the hypothalami of 3 mid to late luteal phase ovarian intact (11.8±0.8 year of age) and 3 OVX rhesus monkeys (9.0±0.6 years of age) were used. The hypothalami from the 3 intact monkeys and two of the OVX monkeys were obtained from the Pathology Unit of the Wisconsin National Primate Research Center, and were immersion-fixed for ~24h with 4% paraformaldehyde and kept in 30% sucrose solution at 4°C for at least five days. The third OVX monkey hypothalamus was perfusion fixed and obtained from the ONPRC. Serial 50 μm coronal sections were cut with a sliding microtome and two successive sections per well were collected into a tris-buffer and stored at 4°C for up to four months until staining.

All tissue used in this study was generated from experiments that had been approved by the Institutional Animal Care and Use Committee at either ONPRC, the University of Wisconsin or the University of Pittsburgh.

Digoxigenin-labeled riboprobe synthesis

The digoxigenin (DIG)-labeled sense and antisense RNA probes were synthesised from a monkey KISS1 cDNA fragment (~300bp) subcloned into pGEM-T vector (29). For antisense and sense probe synthesis, the plasmid was linearised with EcoRI and BamHI, respectively. Linearised cDNA was transcribed for antisense and sense riboprobes using a Maxiscripts transcription kit (Ambion) and a DIG-RNA labeling mix (Roche Diagnostics) in the presence of T3 (for antisense) or T7 (for sense) RNA polymerases. DIG-labeled riboprobes were purified using NucAway Spin Columns (Ambion) and stored at −80°C.

Immunohistochemistry (IHC)

Hypothalamic sections taken at 300 to 400 μm intervals throughout the MBH from each animal were immunostained for kisspeptin with a polyclonal kisspeptin antibody raised in sheep (GQ2; a gift of Dr. Stephen Bloom, Imperial College, London, UK). Briefly, sections were washed 3 times with PBS and incubated with blocking solution consisting of 1 % BSA, 0.1 % gelatin, and 0.25 % triton × in PBS for 1 hr at room temperature. Sections were then incubated with GQ2 (30,000× dilution), in blocking solution for 72 h at 4°C. Subsequently, the sections were washed with PBS 3 times and exposed to a biotinylated 2^nd antibody with avidin-biotinylated peroxidase complex (ABC reagent, Vector Lab., Burlingame, CA). For visualization, 3,3′-diaminobenzidine (DAB), was used. Tissue sections were mounted with glycerine-jelly. As a negative control, sections were exposed to blocking solution in which the primary antibody was omitted: all other procedures were the same. Additional validation of the use of this antibody for IHC localization of kisspeptin in the monkey hypothalamus has been provided previously (30).

The distribution pattern, number and cell size of immunopositive kisspeptin neurons stained in sections from each hypothalamus were systematically analyzed as follows. Three evenly spaced immunostained sections taken from every 1000 μm interval were examined. After staining and before counting positive cells, each section from each monkey was aligned with respect to distance posterior to the caudal boundary of the optic chiasm. Cell body areas of 50 randomly selected kisspeptin-positive neurons per animal were measured using NIH ImageJ.

In Situ Hybridisation (ISH)

We followed with modifications the ISH procedure as described previously by Marks et al. (31) and employed by us earlier (32). Frozen sections were allowed to warm to 4°C and fixed in cold 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO) in diethylpyrocarbonate (DEPC)-treated 0.1 M PBS, pH 7.4, for 5 min. Slides were rinsed twice in cold 0.1 M PBS, and treated with proteinase K (1 μg/ml; Invitrogen) for 30 min at 37°C. Slides were then rinsed in DEPC treated water and triethanolamine (TEA; Sigma-Aldrich), and incubated with 0.25% acetic anhydride in 0.1M TEA for 10 min at room temperature (RT). Slides were soaked in 2× sodium chloride-sodium citrate (SSC, 1 × SSC = 150 mM NaCl and 15 mM sodium citrate, pH 7.4) for three min, dehydrated through graded alcohol and delipidated in chloroform (EMD chemicals) for five min followed by washes in alcohol and air-drying.

Pre-hybridisations and hybridisations were performed under HybriSlip (Research Products International Corp., Mt. Prospect, IL, USA). Slides were incubated with prehybridisation buffer containing 0.001% tRNA in hybridisation buffer (10% dextran sulfate; 50% formamide; 0.3 M NaCl; 9.2 mM Tris, pH 8.0; 0.92 mM EDTA, pH 8.0; 1 × Denhardt’s solution) at 58°C for 1 hr, rinsed in 2 × SSC, dehydrated through graded alcohol and air-dried. Slides were then hybridised with DIG-labeled antisense or sense probe diluted in hybridisation solution for 16-20 hr at 58°C in a humidified chamber. After hybridisation, coverslips were removed and the slides were washed in 4 × SSC for 15 min at 37°C and treated with RNase A (4 mg/ml, Sigma-Aldrich) for 30 min at 37°C and washed under conditions of increasing stringency (30 min in 2 × SSC at RT, twice 30 min washes in 0.1 × SSC at 63°C; 0.1 × SSC for 3 min at RT). Slides were then incubated in blocking buffer (1 × SSC, 2% normal sheep serum, 0.05%Triton X-100) for 1 hr followed by two washes in buffer 1 (0.15 M NaCl, 0.1 M Tris, pH 7.5) for 15 min at RT. Slides were incubated overnight in anti-digoxigenin-AP Conjugate (Roche Diagnostics, 1:1000 dilution in buffer 1 containing 2% NSS and 0.3% Triton X-100) for 16-20 hr at RT followed by two washes in buffer 1 for 15 min at RT. Slides were then equilibrated in buffer 2 (0.15 M NaCl, 0.1 M Tris, pH 9.5, 0.05 M MgCl₂) containing 48 mg/ml levamisole (Sigma-Aldrich) for 10 min at RT, and incubated in chromagen solution (75 mg/ml nitroblue tetrazolium, 50 mg/ml 5-bromo-4-chloro-3-indolyl phosphate, Sigma-Aldrich; 48 mg/ml levamisole in buffer 2) for 1-2 hr in the dark at RT. The chromagen reaction was stopped by washing the slides three times with buffer 3 (0.01M Tris, pH 8.0; 1mM EDTA, pH 8.0) for 10 min at RT. Slides were dehydrated through graded alcohol, cleared in histoclear (National Diagnostics) and coverslipped using Permount (Fisher Scientific, Pittsburgh, PA).

For initial standardisation of ISH with the DIG-labeled KISS1-cRNA probe, and as a positive control in all ISH assays, we used coronal hypothalamic sections (25 μm) from an adult castrated male monkey that earlier had been transcardially perfused with 4% paraformaldehyde (30). These sections exhibited consistent and intensely immunopositive kisspeptin neurones in ARC (30), and we reasoned therefore that KISS1 mRNA levels would likely be concomitantly high. For this purpose, sections were removed from the cryoprotectant in which they were stored at −20°C and washed in 0.1 M PBS, pH 7.4, eight times for 15 min at RT. Sections were then mounted on Superfrost Plus slides (Fisher Scientific, Pittsburgh, PA), dried at RT for 30 min and then placed on a slide warmer for 2 hr at 45°C and before storing at −80°C until ISH was conducted as described above. Figure 1 shows that neurones expressing KISS1 mRNA were readily detected in ARC of the castrated adult male monkey and that the hybridisation signal is absent when sense riboprobe was employed.

Figure 1.

KISS1 mRNA expressing neurones in the arcuate nucleus (ARC) of a castrated adult male rhesus monkey as detected by ISH. The left-hand panel shows hybridisation with the antisense riboprobe. Incubation with the sense probe resulted in loss of the hybridisation signal for KISS1 mRNA (right hand panel). The occasional black spot on the right hand sections incubated with sense riboprobe represents chromogen particles. V, 3rd cerebroventricle. Scale bars are 50 μm (top) and 10 μm (bottom).

Counterstaining of sections was performed using nuclear fast red (Sigma-Aldrich Inc., St. Louis, MO, USA).

Statistics

The significance of differences in number of kisspeptin neurones/section and immunopositive cell size between intact and OVX monkeys were examined by t-test. Immunopositive cell distribution throughout the MBH between ovarian intact and OVX monkeys was examined by two-way ANOVA with repeated measure, followed by Bonferroni’s post hoc tests. The Mann-Whitney test was used to determine the significance of differences in the mean number of KISS1 neurones in the OVX and OVX+P groups. ANOVA followed by Neuman-Keuls posthoc comparison was used to analyze LH suppression.

Results

Kisspeptin positive neurones in MBH of intact and OVX monkeys

A large number of darkly-stained kisspeptin positive neurones were seen throughout the ARC and periventricular region of the hypothalamus of OVX monkeys (Fig. 2A). These cells were most dense in the middle to caudal ARC. In contrast, a small number of lightly, albeit distinctly, stained kisspeptin positive neurones were observed in this region in intact monkeys (Fig. 2B). Importantly, the soma size (1160.7±73.2 μm², Table 1) of kisspeptin neurones in OVX monkeys (Fig. 2D) was significantly greater than that (631.7±14.0 μm², Table 1) of kisspeptin neurones in ovarian intact monkeys (Fig. 2E). Moreover, thick fibers with large varicosities were readily seen throughout the MBH and especially in the median eminence- stalk region of OVX monkeys (Fig. 2D), whereas less prominent fine fibres with small varicosities were found in the ARC and median eminence of ovarian intact animals (Fig. 2B and 2E). Systematic cell counts indicated that kisspeptin neurones in the middle to caudal portion of the ARC were affected by ovariectomy (Fig. 3). The kisspeptin cell number/section (442±93) of OVX monkeys was significantly greater than that (92±25) in ovarian intact monkeys (Table 1). No immunopositive soma or fibres were found in control stained tissues (Fig. 2C and 2F).

Figure 2.

Kisspeptin neurones in the arcuate nucleus (ARC) of female monkeys are up-regulated in the absence of the ovary. Photomicrographs show kisspeptin immunostaining with GQ2 in one ovariectomised (OVX, A and D) and one ovarian intact (B and E) female monkey. Kisspeptin neurones in the OVX hypothalamus were large in size and number compared to the intact situation. Similarly, thicker prominent fibers with large varicosities were seen in OVX females, whereas fine fibers with small varicosities were seen in intact counterparts. Panels C and F show minimal non-specific staining (primary antibody omitted). ME, median eminence; V, 3rd cerebroventricle. Scale bars in A/B/C and D/E/F are 100 μm and 25 μm, respectively.

Table 1.

Effects of ovariectomy on the number and size of immunopositive kisspeptin neurons in adult female monkeys

	Ovarian Intact	OVX
Total # of animals	3	3
Mean #KISS1 neurones/section	92±25	*442±93
Mean Cell size (μm2)	631.7±14.0	***1160.7±73.2

^*Significantly different from Intact at p<0.02

^***Significantly different from Intact at p<0.0001

Figure 3.

Antero-posterior distribution of kisspeptin immunopositive neurones throughout the MBH from ovariectomised (OVX, closed bars) and ovarian intact (open bars) adult monkeys (N=3 per group). The x-axis indicates distance from the caudal edge of the optic chiasm. The numbers 3, 6, 9, and 12 are 1000, 2000, 3000, and 4000 μm from the posterior boundary of the optic chiasm. Note that kisspeptin neurones in the middle to caudal portion of the ARC were affected by ovariectomy. *, ***, indicate P<0.05 and P<0.001, respectively.

KISS1 expression in hypothalamus of OVX monkeys and effect of E₂ and/or P replacement

In control OVX adult monkeys implanted with empty Silastic capsules, intense hybridisation with the antisense KISS1 probe was observed in neurones throughout much of the ARC (Fig. 4). Approximately 60% of the MBH sections examined in the control OVX animals were KISS1 positive and the mean number of KISS1 neurones in KISS1 positive sections was 38.2±7.0 (Table 2). In four of the five control OVX monkeys, KISS1 expression spanned a length of ARC ranging from 1250 to 2000 μm (i.e. 5 to 8 sequential sections collected at 250 μm intervals). In the remaining control animal, however, KISS1 positive neurones were observed in only one section. KISS1 expressing neurones were not observed in the anterior pole of ARC, but the number of these neurones progressively increased toward the midtuberal area and were maximal in the pre-mammillary region of the nucleus (Fig. 4 and 5). KISS1 expression was not observed elsewhere in the MBH. In striking contrast to the control castrate situation, E₂ replacement for 28 days, alone or in combination with P for the last 14 days of treatment, abolished the KISS1 signal at all levels of ARC (Fig. 4, Table 2). Interestingly, P alone, also resulted in a decrease in both the number of KISS1 expressing neurones in ARC and in the hybridisation signal of individual KISS1 positive neurones (Figs. 4 and 5, Table 2). The later parameter was assessed qualitatively.

Figure 4.

Effects of replacement with physiological levels of oestradiol (E₂) or progesterone (P) on KISS1 expression in the arcuate nucleus (ARC) of the ovariectomised (OVX) adult rhesus monkey. KISS1 mRNA in the ARC as revealed by ISH is shown for a control OVX monkey (implanted with empty Silastic capsule) in the left hand panels and for an OVX animal treated with E₂ (OVX+E₂) or P (OVX+P) in the center and right hand panels, respectively. Photomicrographs in the top and bottom rows were taken at 20× and 40×, respectively. Sections were counterstained with nuclear fast red. Scale bar: top, 50 μm; bottom, 10 μm.

Table 2.

Changes in the number of KISS1 expressing neurons produced by oestradiol (E2) and progesterone (P) replacement, either alone or in combination, in ovariectomised (OVX) monkeys. The number of slides per animal is provided to indicate that not each monkey was examined at each of the eleven 250 μm intervals throughout the anteroposterior extent of the MBH (see Fig 5).

	OVX	OVX+E2	OVX+P	OVX+E2+P
Number of animals	5	5	5	5
Number of slides examined/animal	6	5	8	5
% KISS1 positive slides/animal	61	0	25	0
Mean #KISS1 neurones/positive slide/animal (±SE)	38.2±7.0*	-	12.1±4.0	-

^*Significantly different (P<0.05) from OVX+P

Figure 5.

Antero-posterior distribution of KISS1 neurones throughout the MBH of OVX adult rhesus monkeys that were implanted with empty (Control, open data points) or progesterone-filled Silastic capsules (closed data points). Section numbers on the x-axis indicate relative anterior-posterior level with section 1 most anterior with respect to the optic chiasm. Distance between sections, 250 μm. Three to 5 animals (average 4) analyzed at levels 2-9.

Effects of E₂ and/or P replacement on serum LH levels in OVX monkeys

E₂ replacement alone or in combination with P resulted in a marked suppression in circulating LH levels (Fig 6). P replacement alone was without effect.

Figure 6.

Serum LH concentrations (mean± SE) in ovariectomised (OVX) adult rhesus monkeys that were implanted with empty (OVX), oestradiol -filled (OVX+E₂), progesterone-filled (OVX+P), or E₂ and P-filled (E₂+P) Silastic capsules. The number of capsules implanted and the duration of implantation was the same as that used for the animals in which hypothalamic KISS1 expression was determined, and the resulting circulating steroid levels were similar (see text). N=8. Groups with the different letters are significantly different from each other (p<0.05).

Discussion

In coronal sections of the MBH from OVX adult rhesus monkeys both kisspeptin and KISS1 mRNA expressing perikarya were found in the ARC. Although kisspeptin positive neurones were also observed in the periventricular region of the hypothalamus, KISS1 mRNA in the MBH was restricted to ARC. This distribution of ARC kisspeptin neurones is similar to that previously reported for kisspeptin positive neurones in the MBH of castrate adult male rhesus monkeys (30): ie few cell bodies in the retrochiasmatic aspect of the ARC with numbers increasing progressively as the nucleus extends caudally with maximal numbers in the middle to caudal ARC including the pre-mammillary region. Intensely stained kisspeptin immunopositive fibres with large varicosities were also observed in the MBH of the OVX monkeys, and these were particularly apparent in the ARC - median eminencestalk region. Thus, the distribution and intensity of kisspeptin immunopositive profiles in the MBH of OVX monkeys is similar to that previously described for postmenopausal women (33), but appears to differ from that in OVX mice where kisspeptin fibres in the ARC are weakly stained relatively to the ovarian intact situation (34). In the case of the monkey, however, it may now be proposed that loss of the negative feedback action of E₂ leads to a marked increase in the synthesis of kisspeptin in the ARC, as reflected in OVX animals by an upregulation of expression of both mRNA and peptide, together with increased release of kisspeptin in the region of the stalk-median eminence (35).

The number of kisspeptin neurons revealed by IHC was markedly greater than that revealed by ISH. Although a contributing factor may be related to section thickness (50 μm for IHC versus 25 μm for ISH), other possibilities should be considered. First, the relationship between cellular levels of kisspeptin and KISS1 mRNA is unknown; second, the relationship between the sensitivities of the IHC and ISH procedures is unknown; third, the ISH and IHC data were obtained in different laboratories from different animals that had been killed at different times and their brains processed according to different protocols. For all these reasons, the only quantitative comparisons that may be interpreted with confidence are those made within the analysis of protein (IHC) or mRNA (ISH).

In ovarian intact monkeys, the perikarya of kisspeptin immunopositive neurones were small, lightly stained and few in number, and in OVX monkeys replaced with E₂, either alone or in combination with P, KISS1 mRNA was not detected. These observations are consistent with those reported earlier. First, Rometo et al. (24) using ISH reported a hypertrophy of KISS1 neurones in the menopausal and OVX conditions in female cynomolgus monkeys and in women, and that KISS1 expression in the ARC was inversely related to the E₂ status of the subjects. Second, one of our laboratories had also demonstrated with qRT-PCR, significantly elevated levels of KISS1 expression in the MBH of young OVX and aged (menopausal) rhesus monkeys as compared to young eugonadal females (36). Third, a similar regimen of E₂ treatment to that used in the present study was found by qRT-PCR to reduce levels of KISS1 mRNA in tissue blocks comprised of ARC and median eminence from young and old OVX monkeys (37).

Although circulating LH concentrations were not determined in the animals used for the ISH study, suppression of LH in OVX rhesus monkeys by circulating E₂ concentrations in the range produced here have been reported (38,39). Moreover, LH secretion was dramatically suppressed in an additional group of OVX rhesus monkeys that were treated at the ONPRC with an identical E₂ replacement regimen to that employed in the present study. For the foregoing reasons, it seems reasonable to propose that the E₂ induced decrease in KISS1 expression in the ARC of OVX monkeys demonstrated here, and in the earlier study of Rometo et al. (24), led to a reduction in kisspeptin content and release that, in turn, led to a decrease in GnRH secretion and an associated suppression of LH secretion. The latter view is consistent with earlier findings that administration of E₂ to adult OVX monkeys led to a rapid suppression of hypothalamic GnRH release (4), and to an inhibition of MUA in the MBH (5,6).

E₂ negative feedback on LH secretion at the level of the anterior pituitary, on the hand, is considered by some to represent the major site of this steroid’s action in controlling tonic gonadotrophin secretion throughout the menstrual cycle (2,3,40). The latter view is based on the following findings. First ovulatory menstrual cycles may be restored in monkeys bearing hypothalamic lesions that abolish gonadotrophin secretion, and in women deficient in hypothalamic GnRH, by administration of an invariant pulsatile infusion of the synthetic decapeptide (2,3,40). In these experimental paradigms, the low gonadotrophin levels of the early follicular phase of a normal menstrual cycle are replicated indicating that the negative feedback action of E₂ responsible for these low follicular phase levels in the foregoing primate models is exerted directly at the pituitary. Second, E₂ administration to ARC lesioned OVX monkeys, in which gonadotrophin secretion is driven by an unchanging pulsatile GnRH replacement, leads to a profound suppression in LH secretion (41). Again in this paradigm, the negative feedback action of E₂ on gonadotrophin secretion must be exerted on the pituitary because the GnRH drive to the pituitary is clamped experimentally. A pituitary site of negative feedback action of E₂ has also been inferred from the finding that MUA in the MBH of the intact female monkey continues unabated during the follicular phase of the menstrual cycle (42) when LH secretion is controlled by the negative feedback action of E₂.

Overall, the foregoing findings provide compelling evidence for both a hypothalamic and pituitary site of negative feedback action of E₂ to regulate tonic gonadotrophin secretion in primates. The relative importance of these two negative feedback mechanisms of action and how they are integrated to achieve the physiological control of tonic LH secretion during the menstrual cycle remains to be determined.

The present finding that replacement of OVX animals with P alone resulted in a marked decrease in the number of KISS1 positive neurones in the ARC, is similar to that of a study from Clarke’s laboratory (26), in which OVX ewes also received P replacement for two weeks during the breeding season and Kiss1 expression was measured by ISH. In contrast, when Kiss1 mRNA levels in punches of the arcuate region were determined with qRTPCR the same group found P replacement in OVX ewes to be without effect (Dr. Robert L Goodman, personal communication). Similarly, in a study of OVX rats that utilized qRTPCR to measure mRNA levels in the MBH, P suppression of Kiss1 expression was not observed (25).

Although it is generally recognized that pretreatment with E₂ is required for upregulation of P receptor (PR) expression in the hypothalamus (43-45), oestrogen independent expression of PR in brain has been reported (46), as has conversion of P to E₂ by neural steroidogenic enzymes (47,48).

Regardless of the mechanism underlying the ability of P to suppress KISS1 expression in the OVX monkey, the significance of this action of the steroid is unclear. In this regard, an early study demonstrated that P replacement in OVX monkeys, which produced luteal phase levels of circulating P for a period of time similar to that employed here, had little impact on circulating gonadotrophin levels as measured in daily samples (38). Indeed, the inability of luteal phase levels of P to suppress LH secretion in the absence of E₂ was confirmed in the present study, albeit not in the monkeys used for ISH. We recognize that an important component of P negative feedback action on LH secretion is mediated by a deceleration in LH pulse frequency (10-12), and therefore subtle effects of P on LH may have been missed in the present study because only a single sample was analysed from each animal.

In contrast to the OVX situation, Smith et al (21) studying the ovarian intact monkey found that the numbers of KISS1 expressing neurons in ARC during the early-follicular and mid-luteal phases of the menstrual cycle (stages with comparable E2 levels but dramatically different P concentrations, ref 1) were similar, further suggesting that P regulation of ARC kisspeptin neurons in the OVX monkey may be unrelated to the control of tonic gonadotropin secretion. However the case may be, the apparent oestrogen independent action of P to regulate ARC kisspeptin gene expression, and its functional significance, merit further study.

In summary, the present study of the rhesus monkey extends previous work in the closely related cynomolgus macaque by Rometo et al. (24) by showing the location of kisspeptin peptide in the MBH of a female primate, and demonstrating that 1) the increase in KISS1 mRNA levels induced in ARC by ovariectomy (24) is paralleled in this region of the hypothalamus by an elevation in the expression of the peptide encoded by the gene, 2) the distribution pattern of KISS1 and kisspeptin throughout the MBH of the female monkey is similar, and 3) the action of E₂ replacement to downregulate the expression of KISS1 in ARC of the OVX monkey is mimicked to a certain extent by P replacement, alone. From a comparative perspective, our results provide further evidence to support the notion that, as in non-primate species, kisspeptin neurons in the ARC represent a key component of the negative feedback control system whereby the ovary regulates tonic gonadotrophin secretion in primates. Conceptually, therefore, it may be argued that in contrast to the fundamental species differences in the mechanisms utilized by mammalian species to trigger the surge mode of gonadotrophin secretion prior to ovulation (49), the control system governing tonic gonadotrophin secretion has not been subjected to such marked evolutionary pressures.