Division of Neuroscience, Oregon Health and Science University, Ore-on National Primate Research Center, 505 NW 185th Avenue, Beaver-
on, OR 97006, USA
Department of Physiology and Pharmacology, Oregon Health and Sci-nce University, 3181 SW Sam Jackson Park Road, Portland, OR 97239,SA
bstract—In the rodent, arcuate nucleus of the hypothala-us (ARH)-derived neuropeptide Y (NPY) and proopiomela-ocortin (POMC) neurons have efferent projections through-ut the hypothalamus that do not fully mature until the sec-nd and third postnatal weeks. Since this process is likelyompleted by birth in primates we characterized the ontog-ny of NPY and melanocortin systems in the fetal Japaneseacaque during the late second (G100), early third (G130)
nd late third trimesters (G170). NPY mRNA was expressed inhe ARH, paraventricular nucleus (PVH), and dorsomedialucleus of the hypothalamus (DMH) as early as G100. ARH-erived NPY projections to the PVH were initiated at G100 butere limited and variable; however, there was a modest in-rease in density and number by G130. ARH-NPY/agouti-elated peptide (AgRP) fiber projections to efferent targetites were completely developed by G170, but the densityontinued to increase in the postnatal period. In contrast toPY/AgRP projections, �MSH fibers were minimal at G100nd G130 but were moderate at G170. This study also re-ealed several significant species differences between ro-ent and the nonhuman primate (NHP). There were few NPY/atecholamine projections to the PVH and ARH prior to birth,hile projections were increased in the adult. A substantialroportion of the catecholamine fibers did not coexpressPY. In addition, cocaine and amphetamine-related transcript
europeptide Y/agouti-related peptide (NPY/AgRP) androopiomelanocortin (POMC) neurons in the arcuate nu-leus of the hypothalamus (ARH) function as key centralegulators of food intake and homeostatic feedback controlf energy balance in both rodents and nonhuman primatesNHP) (Kagotani et al., 1989; Porte et al., 1998; Kalra etl., 1999; Larsen et al., 1999; Schwartz et al., 2000; Koe-ler et al., 2001; Grove et al., 2003b), in addition to playingrole in the reproductive, neuroendocrine and stress axes
Kalra, 1986; Hokfelt et al., 1998; Li et al., 1999, 2000; Bellt al., 2000; Conrad and McEwen, 2000; Campbell et al.,001). Although NPY neurons are present throughout therain in both the normal adult rodent and NHP, the majorrigin of hypothalamic NPY production is the ARH. Thenique feature of ARH-NPY neurons is that they are theole source of AgRP, which has become a useful marker ofRH-NPY neurons (Broberger et al., 1998; Hahn et al.,998; Chen et al., 1999; Grove et al., 2003a). ARH-NPY/gRP neurons have a broad range of projections through-ut the hypothalamus and forebrain; however, it is therojections to the paraventricular nucleus (PVH), dorsome-ial hypothalamic nucleus (DMH), perifornical regionPFR) and lateral hypothalamic area (LHA) that have re-eived the most study in regard to their role in the regula-ion of energy balance. In the rat, a low level of NPYeptide is discernible as early as embryonic day 14, and
mmunoreactive (-ir) fibers in the ARH are present twoays after birth. However, the projections of ARH neurons,
ncluding NPY/AgRP and POMC neurons, to downstreamarget sites do not start developing in the rodent until thend of the first postnatal week and do not fully develop untilhe third postnatal week (Grove et al., 2003a; Grove andmith, 2003; Bouret et al., 2004a). Recently, it was alsoemonstrated that ARH projections to the PVH do not fullyevelop in the leptin deficient ob/ob or the leptin receptorutant db/db mice, suggesting that leptin is a critical factor
or the development of ARH circuits (Bouret et al., 2004b).here is a natural leptin surge during the second postnataleek (Ahima et al., 1998) of rodents, supporting its phys-
ological role, and leptin replacement to ob/ob mice spe-ifically during the second to third postnatal week restoreshe development of ARH projections (Bouret et al., 2004b).
While these circuits have been extensively studied indult rodents, and more recently in the developing rodent,
ittle has been done to elucidate development of thesemportant energy homeostatic circuits in primates. In the
resent study, we report the presence of NPY mRNA in the
ved.
AWNtivapmiA
A
AaaPpIUau
iom3isfsvgwibscerfibNt
H
MCwssi
I
hacws1
o
hccBtt8gtwaRdvfide6cwtm
D
DE(SPu
I
sfp(cfe
tctaMwsb3aFtns(UcSv(itL
p
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986976
RH during late second and third trimesters of pregnancy.e characterized the development of ARH and brainstemPY efferent projections during the late second and third
rimester in the NHP, with a focus on the PVH as anmportant “feeding center.” We also characterized the de-elopment of � melanocyte stimulating hormone (�MSH)nd cocaine and amphetamine-related transcript (CART)rojections from the ARH. Finally, we measured fetal andaternal leptin during late pregnancy to determine if there
s a surge of leptin that correlates with the development ofRH circuits.
EXPERIMENTAL PROCEDURES
nimals
ll animals were maintained in outdoor/indoor group housing innatural light schedule. Animals were given water and food
d libitum consisting of high protein monkey chow biscuits (Ralstonurina Co., St Louis, MO, USA). All animal procedures were ap-roved by the Oregon National Primate Research Center (ONPRC)nstitutional Animal Care and Use Committee and conformed to.S. National Institutes of Health guidelines on the ethical use ofnimals. Every effort was made to minimize the number of animalssed and their suffering.
Pregnant Macaca fuscata, Japanese macaque monkeys weredentified by the Department of Animal Resources during routinebservations of the animal colony. Initial pregnancy was deter-ined by palpation and then confirmed by ultrasound examination0 and 60 days following initial knowledge of pregnancy. Animals
n the study were all multiparous and no twins were included in thetudy. Cesarean sections were performed at day 100, 130 or 170etal gestational age (full term is 175 days) by the trained surgicaltaff at ONPRC. The adult animals were allowed to recover undereterinary supervision before they were returned to the outdoorroup housing. The fetuses were taken directly to necropsy andere deeply anesthetized with sodium pentobarbital (�30 mg/kg
.v.). The chest cavity was opened and the ascending aorta wasisected and the cardiovascular system was flushed with 0.9%aline containing 5000 units of heparin/l to inhibit clotting. Trans-arotid perfusion of the head with 4% paraformaldehyde in buff-red sodium phosphate (pH 7.4) was performed. The brain wasemoved, blocked and post-fixed overnight at 4 °C in the samexative and subsequently transferred for 24 h to 10% glyceroluffered with NaPO4 and for 48 h into 20% glycerol buffered withaPO4. Tissue was frozen by submersion in ��50 °C methylbu-
ane and then stored in �80 °C until sectioning.
ormone assays
aternal and fetal leptin was determined by RIA (Linco, St.harles, MO, USA) as directed by assay protocol. Insulin levelsere measured by an electrochemiluminescence immunoas-ay using an automated instrument (Elecsys 2010 immunoas-ay analyzer, Roche Diagnostics, Palo Alto, CA, USA). The
nterassay CV was less than 5%.
n situ hybridization
Experiment 1: In situ hybridization for NPY mRNA. Coronalypothalamic sections (25 �m) were collected in 1:12 series usingfreezing microtome. Sections were stored in ethylene glycol
ryoprotectant at �20 °C until time of use. One series of sectionsas wet mounted in RNAse free KPBS on Fisher brand Superfrostlides. Animals were studied at approximate gestational days 100,30 and 170 and as adults (5–7 years of age) (n�3/group).
In situ hybridization for NPY mRNA was performed as previ-
usly described (Chen et al., 1999; Grove et al., 2001). Briefly, a i
uman prepro-NPY cRNA probe was transcribed from a 500-bpDNA in which 50% of the UTP was 35S labeled. The saturatingoncentration of the probe used in the assay was 0.3 �g/ml/kb.rain sections were fixed in 4% paraformaldehyde (pH 7.4),
reated with protein kinase at 37 °C to increase penetration, andhen with 0.25% acetic anhydride in 0.1 M triethanolamine (pH.0). Sections were then rinsed in 2� SSC, dehydrated throughraded series of alcohols, delipidated in chloroform, rehydratedhrough a second series of alcohols and air-dried. The sectionsere exposed to the labeled probes overnight in a moist chambert 55 °C. After incubation the slides were washed in 4� SSC, inNase A at 37 °C, and in 0.1� SSC at 60 °C. Slides were thenehydrated through a graded series of alcohols and dried. Forisualization of the probe, labeled sections were exposed to sheetlm (Biomax MR, Kodak) overnight. For histological analysis of theistribution of NPY mRNA, slides were dipped in Kodak NBT2mulsion (Eastman Kodak, Rochester, NY, USA) diluted 1:1 in00 mM ammonium acetate and then placed in light-tight boxesontaining desiccant and stored at 4 °C for 7 days. The slidesere then developed and counterstained with Cresyl Violet and
he distribution of silver grains was analyzed using darkfieldicroscopy.
arkfield microscopy
arkfield images of silver grains were captured using a Nikonclipse E800 Microscope coupled with a CoolSnap HQ camera
Photometrics, West Chester, PA, USA) along with Metamorphoftware (Universal Imaging Corp.). Objectives used were Nikonlan Fluor 4X (NA 0.13). The images were further processedsing Adobe Photoshop.
mmunofluorescence
Double-label immunofluorescence. Coronal hypothalamicections (25 �m) from the same animals as above were usedor immunofluorescence studies. Animals were studied at ap-roximately gestational days 100, 130 and 170 and as adults5–7 years of age) (n�3/group). Double-label immunofluores-ence coupled with confocal microscopic analysis was per-ormed as previously described (Grove et al., 2000; Campbellt al., 2001).
Experiment 2: ARH-NPY fiber projections. NPY fibers fromhe ARH were identified as those colocalized with AgRP. Toharacterize the development of ARH-NPY/AgRP efferent projec-ions, a cocktail of goat anti-NPY (a kind gift of P. Larsen; 1:5000),nd a guinea-pig anti-AgRP (produced by Antibodies Australia,elbourne, Australia; 1:5000) was used. Briefly, tissue sectionsere removed from cryoprotectant and washed in 0.05 M potas-ium phosphate-buffered saline (KPBS), preincubated in blockinguffer (KPBS�0.4% Triton X-100�2% normal donkey serum) for0 min at room temperature, followed by incubation with AgRPnd NPY-specific antibodies in blocking buffer for 48 h at 4 °C.ollowing washes in KPBS, tissue was incubated for 1 h in bio-
inylated donkey anti-guinea-pig antibody (1:600, Jackson Immu-oResearch Laboratories, Inc., West Grove, PA, USA), andubsequently washed and incubated in avidin–biotin solutionVectastain Elite ABC, Vector Laboratories, Burlingame, CA,SA) for 30 min. The AgRP-ir signal was further amplified using aommercial kit (tyramide signal amplification-indirect kit; NEN Lifeciences Products, Boston, MA, USA). AgRP-enhanced-ir wasisualized using Alexa 568 (red) conjugated to streptavidin1:1000, Molecular Probes, Eugene, OR, USA). For NPY colocal-zation, NPY-ir was visualized with Alexa 488 (green) conjugatedo a donkey anti-goat antibody (1:200, Jackson ImmunoResearchaboratories, West Grove, PA, USA).
Experiment 3: Brainstem NPY fiber projections. NPY fiberrojections from the brainstem were identified as those colocal-
zed with dopamine � hydroxylase (DBH)—a marker for catechol-
adrDd(
atapdv5(
C
ScPatClMdu
C
C1tCRllVscbd�fflri(cqn(
H
Fwea6ifkts
6m
mtttl1
N
AdlDpetptaa
Flm(
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986 977
mine neurons. To characterize the development of brainstemerived NPY, a cocktail of mouse DBH (Chemicon; 1:5000) andabbit anti-NPY (a kind gift of S. Brown, 1:75,000) was used.BH-ir and NPY-ir were visualized by direct labeling with TRITConkey anti-mouse and FITC donkey anti-rabbit, respectively1:200, Jackson ImmunoResearch Laboratories).
Experiment 4: �MSH and CART fiber projections. To char-cterize the development of POMC and CART efferent projec-ions, a cocktail of sheep �MSH (Chemicon; 1:5000) and a rabbitnti-CART (a kind gift of N. Vrang, 1:80,000) was used. Therotocol for signal enhancement for CART was similar to thatescribed above for AgRP. CART-ir was visualized with strepta-idin Alexa 488 (1:1000, Molecular Probes). For �MSH-ir, TRITC68 (red) was used conjugated to a donkey anti-sheep antibody1:200, Jackson ImmunoResearch Laboratories).
omputer-assisted line drawings
ingle-label immunohistochemistry for AgRP was done using DABhromagen enhancement. The Rhesus Macaque Brain Atlas byaxinos et al. (1999) was used to match hypothalamic levels in thedult and G130 animal at levels �5.0, �5.5, �6.0 and �6.75 caudalo bregma. Using a Nikon Eclipse E800 microscope coupled with aoolSnap HQ camera (Photometrics) adjacent images of single-
abeled AgRP-ir were snapped at 20� using Metamorph software.ontages were constructed using Adobe Illustrator. Lines wererawn over contiguous fiber projections and terminal axonal fieldssing the pencil tool. AgRP-ir was used as a marker of ARH-NPY.
onfocal laser microscopy
onfocal laser microscopy, as previously described (Li et al.,999, 2000; Grove et al., 2000; Campbell et al., 2001), was usedo analyze the double-label immunofluorescent images. The Leicaorporation (Germany) TSC SP confocal system consisted of aBE inverted microscope equipped with an Ar laser producing
ight at 488 nm (for visualization of FITC and Alexa 488) and a Kraser producing light at 568 nm (for visualization of Alexa 568).arious objectives (25�/NA 0.75 and 40�/NA 1.25) were used tocan the images. For each experiment, fluorophore signals werehecked individually for bleed through to the apposing detector. Allleed through was eliminated by adjusting the laser intensity andetector window width. A series of continuous optical sections, 0.5m intervals along the z axis of the tissue section, was scanned
or each fluorescent signal. The signals were obtained for eachuorophore on one series of optical sections and stored sepa-ately as a series of 512�512 or 1024�1024 pixel images. Themages were then processed with the MetaMorph Imaging SystemUniversal Imaging Corporation, West Chester, PA, USA). Theonfocal images are presented as projections of a stack of se-uential optical slices totaling 10 �m of the section. The bright-ess and contrast of the images were adjusted in PhotoshopAdobe Systems Inc., San Jose, CA, USA).
RESULTS
ormonal environment
etal leptin levels were low at G100 and G130 comparedith G170 (Fig. 1). However, there were dramatic differ-nces in fetal body weights between the gestationalges (G100, 172 g�12; G130, 414 g�37; G170,36 g�39). Since body fat levels were not measured, it
s difficult to determine if leptin levels correlate with bodyat as they do in adult animals. Furthermore, it is un-nown how much of the leptin is actually coming fromhe fetus or from the placenta. Fetal insulin levels did not
ignificantly change at the three gestational ages (G100, a
Maternal leptin was elevated throughout the third tri-ester (G130 and G170) compared with the second
rimester (G100), but was highest during the late thirdrimester (G170) (Fig. 1). However, like humans, most ofhis increased leptin secretion during the third trimester isikely due to placental leptin secretion (Schubring et al.,997; Sagawa et al., 2002).
PY mRNA expression in the hypothalamus
s previously shown (Grove et al., 2003b), adult animalsisplayed intense NPY expression in the ARH and low
evels in the PVH and scattered cells in the area of theMH (Fig. 2). Similar to the adult, NPY mRNA was ex-ressed within the ARH at all fetal ages investigated, andxhibited intense expression in both G100 and G130 fe-uses (Fig. 2). Furthermore, scattered NPY mRNA ex-ressing cells were present in the PVH at G100; however,hese neurons were not evident in the older gestationalges. The reason for this is unknown, but even in adultnimals the levels are low in the PVH in unfasted animals
ig. 1. Leptin levels during development. Fetal and maternal leptinevels increase with prolonged gestational age. Values represent the
ean�S.E.M. Asterisk indicates significant difference by ANOVA.P�0.05.)
nd, therefore, may be simply due to detection sensitivity.
A
Pi(a
Noaa
Foa ptured a
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986978
RH-NPY/AgRP cell bodies
reviously, our group, and others, have shown that AgRPs expressed in the vast majority of ARH-NPY neuronsHahn et al., 1998; Chen et al., 1999; Haskell-Luevano et
ig. 2. Postnatal development of NPY. Images are dark-field photomicf ARH NPY. NPY mRNA appears to be highly expressed at all age gccording to the Paxinos atlas (Paxinos et al., 1999). Images were ca
l., 1999). This has been shown for both rodents and t
HPs. In the present study, we used AgRP-ir as a markerf ARH-NPY projections (Broberger et al., 1998; Grove etl., 2003a). Similar to previous data (Haskell-Luevano etl., 1999), there was an abundance of NPY/AgRP fibers in
of silver grains representing human prepro-NPY cRNA probe labelingestigated. These images correspond to �6.0 with respect to bregmat 4� magnification.
rographsroups inv
he ARH of the adult monkey; however, NPY-ir cell bodies
wdmrlpwtbks
ctsu
ilFAl
F(dN2 hey corr
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986 979
ere not evident, while many AgRP-ir cell bodies could beetected with this antibody (Figs. 3 and 4). This is a com-on limitation of this technique in which the peptide is
apidly transported from the cells bodies and not accumu-ated. It is difficult in the normal adult animal to detecteptide within the cell bodies without colchicine treatment,hich could not be done in these NHP studies. But, while
here was an abundance of double-label NPY/AgRP-ir fi-ers (shown as yellow) within the ARH of the adult mon-ey, it is worth noting that there is also an abundance of
ig. 3. Development of NPY/AgRP neurons in the ARH. Figures repregreen) and AgRP (red), where colocalization is represented as yelloetectable at G100 but sparse NPY/AgRP-ir cell bodies were visualizote that NPY-ir is not readily detectable in ARH cell bodies of the ad5� oil objective (0.75 NA) and represent an area of 400�400 �m. T
ingle-label NPY-ir fibers as well (Fig. 3). These results m
ontrast with low levels of single-label NPY-ir in the ARH ofhe rat (Broberger et al., 1998; Grove et al., 2003a), anduggest that another source of NPY significantly contrib-tes to the peptide within the ARH of the NHP (see below).
Even though there was an abundance of NPY mRNAn the ARH at G100 (late second trimester), only lightlyabeled NPY/AgRP-ir cell bodies were detected at this age.urthermore, in contrast to older ages, only scattered NPY/gRP-ir fibers were present, as well as scattered single-
abel NPY-ir fibers. This indicates that the ARH at G100
r confocal digital images of double-label immunofluorescence for NPYimmunoreactivity was always colocalized with NPY. NPY/AgRP-ir is
ased density of fibers and soma appears with increased animal age.AGPR-ir in cell bodies is more evident. Images were captured with aespond to approximate level illustrated by area A in Fig. 5.
sent colow. AgRPed. Increult, while
ay not be fully developed, which is in agreement with the
da
Aretsatl
ha
A
INpPefi
Faio correspo
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986980
evelopment of the ARH in the human fetus (Koutcherov etl., 2003).
By G130, an abundance of AgRP-ir and some NPY/gRP-ir cells were evident throughout the ARH. The
eadily detectable cell bodies are likely due to increasedxpression and decreased transport of the peptide out ofhe ARH. There was also a moderate concentration ofingle-label NPY-ir fibers evident at this age, indicating thatnother source of NPY is innervating the ARH. By G170,he density of both double-label NPY/AgRP-ir and single-
ig. 4. Development of NPY/AgRP neurons in the PVH. Figures reprend AgRP. Single-label NPY fibers are shown in green, AgRP fibers ar
nnervation was present at G100 with increases at G130 and G170. Objective (0.75 NA) and represents an area of 400�400 �m. Images
abel NPY-ir fibers increased compared with earlier ages; (
owever, they still had not reached the levels seen in thedult.
RH-NPY projections to PVH
n the adult hypothalamus there was an abundance ofPY/AgRP-ir and single-label NPY-ir fibers within all hy-othalamic regions, with the highest concentration in theVH (Fig. 4) and DMH and lower levels in the LHA. How-ver, qualitatively the relative concentration of NPY/AgRPbers (originating from the ARH) to single-label NPY fibers
r confocal digital images of double-label immunofluorescence for NPYin red and double-label NPY/AgRP fibers are shown in yellow. Sparsevels were lower then the adult. Images were captured with a 25� oilnd to approximate level illustrated by area B in Fig. 5.
sent coloe shownverall le
originating from other sources, i.e. PVH, DMH or brain-
seif
ttaTGsisApie
B
AfitcttdbvlwTseriwNebpbctDpDw6wl
ts(Notfi
�
TsNnCiC(at�At�HdhioisAcwpga(Ahttec�A
Dsnsa(Sshdet1Ar
st
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986 981
tem) is low compared with that observed in the rat (Brob-rger et al., 1998; Grove et al., 2003a). This indicates that
n the NHP more of the NPY inputs into the PVH may comerom sources other than the ARH.
As early as G100, NPY/AgRP-ir fibers were present inhe PVH (Fig. 4) and other hypothalamic areas, althoughhey were relatively sparse. Single-label NPY fibers werelso sparsely distributed throughout the PVH at this age.he density of NPY/AgRP-ir fibers greatly increased at130 and G170. Again, the levels of both NPY/AgRP andingle-label NPY fibers at G170 were less than the densityn adults, particularly for NPY/AgRP. These data arehown schematically by computer-assisted line drawing ofgRP-ir indicating the extent of ARH-derived NPY to hy-othalamic nuclei at G130 and in the adult (Fig. 5). This
ndicates that ARH projections begin developing during thearly third trimester of pregnancy in the NHP.
rainstem NPY/catecholamine projections
s discussed above, an abundance of single-label NPY-irbers was observed within the ARH, PVH and other hypo-halamic regions at all ages. It is well known that NPY/atecholamine neurons from the brainstem projecthroughout the hypothalamus. To determine if this is true inhe NHP brainstem, we performed a preliminary study forouble-label immunofluorescence for NPY and DBH in therainstem. Within the solitary tract nucleus (NTS) and theentral lateral medulla (VLM) there were many double-abel fibers; however, single-label NPY-ir and DBH-ir fibersere readily evident throughout the brainstem (Fig. 6A).he source of the single-label NPY-ir fibers in the brain-tem is unknown and may originate from the ARH, PVH orven locally within the brainstem. NPY-ir neurons wereeadily detectable in the NTS and VLM; however, visual-zation of catecholamine cell bodies with this DBH antibodyas very limited, thus we were unable to confirm that allPY-ir cell bodies in the NTS and VLM were indeed cat-cholaminergic. These results indicate that it is not possi-le to completely characterize all of the brainstem NPYrojections simply by colocalization with DBH as is possi-le in the rodent. However, it remains true that NPY colo-alized with DBH marks brainstem-derived NPY. Keepinghis in mind, we characterized the development of NPY/BH projections into the PVH and ARH in the NHP. Sur-risingly, while there was a moderate scattering of NPY/BH-ir fibers observed in the PVH of the adult NHP, thereas an abundance of double-label fibers in the ARH (Fig.C and E). Furthermore, the majority of DBH-ir in the ARHas colocalized with NPY, while the presence of single-
abel DBH-ir fibers was more abundant in the PVH.Very few NPY/DBH fibers were observed in the hypo-
halamus at any gestational age, this included sparselycattered NPY/DBH fibers in the ARH or PVH at near termG170; Fig. 6B and D). This suggests that very few of thePY-ir present within the hypothalamus during fetal devel-pment likely come from the brainstem. However, as men-ioned above, it is possible that some brainstem NPY-ir
bers may not coexpress DBH. i
MSH and CART are not colocalized in the NHP
he primary source of �MSH is the ARH, while a smallubpopulation of �MSH (POMC) neurons, resides in theTS. In the rodent, nearly all �MSH neurons in the arcuateucleus co-express CART (Vrang et al., 1999b). AdditionalART neurons are present throughout the hypothalamus,
ncluding the PVH, LHA and DMH. Furthermore, whileART and �MSH neurons are both present in the NTS
Vrang et al., 1999b; Ellacott and Cone, 2004; Appleyard etl., 2005), these two neuropeptides are not colocalized inhis nucleus. Therefore, in the rodent, colocalization ofMSH and CART peptides in terminal fields is a marker ofRH-derived �MSH. In this study, we attempted to utilize
his technique to characterize the development of ARH-MSH projections through its colocalization with CART.owever, while in the adult NHP we observed an abun-ance of both �MSH-ir and CART-ir fibers throughout theypothalamus, these neuropeptides were never colocal-
zed in fibers or cell bodies, at least within the sensitivity ofur assay (Fig. 7). Furthermore, single-label in situ hybrid-
zation showed an abundance of CART mRNA in the PVH,upraoptic nucleus and LHA, with very low levels in theRH (data not shown). There was, however, evidence oflose appositions between �MSH-ir and CART-ir fibersithin the PVH of the adult NHP, suggesting that these twoopulations of neurons may be interacting at efferent tar-et sites. CART-ir fibers were sparse in the hypothalamust G100 (not shown), but were relatively abundant by G130Fig. 7). In contrast, �MSH-ir neurons were evident in theRH as early as G100 and G130 (Fig. 7 bottom panels);owever, �MSH-ir fibers did not show significant innerva-
ion of other hypothalamic sites, such as the PVH at theseimes (Fig. 7 top panels). By G170, �MSH-ir fibers werevident throughout the hypothalamus but were moderateompared with the adult. These findings suggest that ARH-MSH projections may not develop until after ARH-NPY/gRP projections.
DISCUSSION
uring the first 2 weeks in the developing rodent, there aretriking developmental changes in the NPY system. Mostotably, there is a transient expression of NPY mRNA ineveral hypothalamic areas, including the DMH, PFR, PVHnd LHA in addition to the normal expression in the ARHSinger et al., 2000; Grove et al., 2001, 2005; Grove andmith, 2003). Furthermore, we and others have demon-trated that while NPY-ir fibers are present throughout theypothalamus at birth, the ARH-NPY projections do notevelop until the second and third postnatal week (Grovet al., 2003a; Bouret et al., 2004a). Finally, a surge of leptinhat occurs in the early postnatal period (Ahima et al.,998) appears to be critical for the full development of theRH circuits, including both NPY/AgRP and MSH in the
odent (Bouret et al., 2004b).In contrast to the rodent, the present study demon-
trated that there is an abundance of NPY/AgRP fibershroughout the hypothalamus at birth in the NHP, indicat-
ng that these projections develop prior to birth. Indeed, as
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986982
eeNhi(tgpmTtAu
2advt(
w(ccwtf
FdcsFa
Facta ages wero n Fig. 5 b
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986 983
arly as the late second trimester (G100), NPY mRNAxpression (Fig. 2) was high in the ARH and scatteredPY/AgRP fibers (Figs. 3–5) were evident throughout theypothalamus. The density of these projections greatly
ncreased during the third trimester leading up to birthFigs. 3–5). However, the density of projections at nearerm (G170) is less than that observed in the adult, sug-esting that there is further development in the density ofrojections during the postnatal period, which is in agree-ent with the reports of others (El Majdoubi et al., 2000).hese findings indicate that the late second and early third
rimesters may be a critical period for the development ofRH circuits, which are key for homeostatic feedback reg-lation of body weight (Schwartz et al., 2000; Saper et al.,
ig. 5. Schematic drawings of ARH-derived NPY projections in the hyrawings illustrate the distribution of ARH-derived hypothalamic NPY fiorresponding approximately to �5.0, �5.5, �6.0 and �6.75 caudal tections were chosen approximating similar levels by anatomical marke
ig. 6. Presence of NPY/DBH fibers in the brainstem. Figures represnd DBH. Single-label NPY and D�H fibers are shown in green and rell bodies are present in the NTS with single- and double-label NPY/he PVH and ARH (B and D). Both single-label DBH and co-labeled N
10-�m thick collection of optical slices taken at 0.5-�m intervals. Imf 400�400 �m. Images correspond to approximate level illustrated i
iber projections are illustrated by lines. Darkened black shapes represent cron “x.” Boxes A and B in lower panel illustrate the approximate areas in color
002). This also indicates that maternal health (obesitynd diabetes) and diet could have a critical impact on theevelopment of these projections; this remains to be in-estigated. In contrast, the postnatal environment (postna-al nutrition) may have a greater impact in the rodentGrove and Smith, 2003).
In the current study we found that fetal leptin levelsere very low to undetectable during the late second
G100) and early third (G130) trimester and did not in-rease until right before birth (Fig. 1), this low leptin isonsistent with the near absence at G100 and G130 ofhite adipose tissue which develops during the 3rd trimes-
er (K. L. Grove et al., unpublished observations). There-ore, in the NHP, there was no evidence of a leptin surge
us of the G130 and adult Japanese macaque. Computer-assisted lineterminal fields using AgRP-ir as the marker at four levels in the adultaccording to the Paxinos atlas (Paxinos et al., 1999). Correspondingh exact level is unknown. Terminal fields are demarcated with specks.
confocal digital images of double-label immunofluorescence for NPYctively, and double-label NPY/DBH fibers are shown in yellow. NPYrs (A). The G170 animal expresses few catecholamine fibers in bothare present in the adult PVH and ARH (C and E). Images represente captured with a 25� oil objective (0.75 NA) and represent an areay area B in the PVH panel and area A in the ARH panel.
pothalambers and
o bregmars, thoug
ent colored, respeDBH fibePY/DBH
ss section of vasculature. Detectible cell bodies are demarcated withphotomicrographs of ARH (A) and PVH (B).
paeAtp
hcNan(tib1Gpicd
NjtcmtAwidedsfNPGcot
FiCpa ages wero
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986984
rior to the development of ARH projections. However,lternative explanations need to be considered. Foremost,ven as early as G100, there were already low levels ofRH-NPY/AgRP-ir fibers in the PVH and DMH; therefore,
he leptin surge that initiates the development of theserojections may occur prior to G100.
Similar to the rodent, not all of the NPY-ir within theypothalamus of the NHP comes from the ARH, as indi-ated by the significant concentration of single-labelPY-ir fibers (Figs. 3 and 4). In fact, in the NHP hypothal-mus, a qualitatively higher proportion of NPY fibers wasot colocalized with AgRP than typically seen in rodentsBroberger et al., 1998; Grove et al., 2003a). In the rodent,he other main source of NPY within the hypothalamus,ncluding the PVH and ARH, is NPY projections from therainstem, predominantly the A1/A2 regions (Everitt et al.,984; Chronwall et al., 1985; de Quidt and Emson, 1986;rove et al., 2003a). However, in the NHP, a limited pro-ortion of the NPY fibers within the PVH at any of the ages
nvestigated was colocalized with DBH (a marker ofatecholamine fibers). In contrast, there was an abun-
ig. 7. Postnatal development of CART and �MSH in the PVHmmunofluorescence for CART (green) and �MSH (red). Evidence ofART and �MSH do not colocalize in the NHP though close appositioresent in the PVH. Blue and white arrows indicate the presence of sin10-�m thick collection of optical slices taken at 0.5-�m intervals. Im
f 400�400 �m.
ance of NPY/DBH-ir fibers within the ARH in the adult l
HP. While it is likely that not all brainstem NPY pro-ections co-express DBH, their colocalization does iden-ify NPY/catecholamine projections. These findings indi-ate that NPY/catecholaminergic neurons in the brainstemay preferentially innervate the ARH in the NHP, while in
he rodent there is a much broader projection pattern.lternatively, in the NHP, it has not been determinedhether all NPY produced in the brainstem strictly colocal-
zes with catecholamines, as has been shown in the ro-ent. The physiological importance of this species differ-nce is unknown, but there are other examples of speciesifferences in brainstem projections. Our group has alsohown that glucagon-like protein (GLP) neuron projectionsrom the NTS also preferentially innervate the ARH of theHP, while in the rodent the primary projections are to theVH and DMH (Tang-Christensen et al., 2001; Vrang androve, 2004). Therefore, primates may have a unique
ommunication between brainstem neurons and ARH 1strder homeostatic sensory neurons, whereas in the rodenthis interaction may occur in efferent target sites at the
. Figures represent color confocal digital images of double-labelitially manifests as early as G130 in the ARH and G170 in the PVH.vident occasionally. Yellow arrows indicate close apposition of fibersd CART and �MSH soma respectively in the ARH. Images represente captured with a 25� oil objective (0.75 NA) and represent an area
and ARH�MSH inns are egly labele
evel of 2nd order neurons.
hmApHAGpEfirpsfivtmt
tNewesiatIppb
bocrsphphncsaipbitPAPBdt
dvd(t
IsNiftostbaCiNdts
ACDaTD
A
A
B
B
B
B
C
C
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986 985
In the rodent, brainstem efferent projections to theypothalamus develop in the early postnatal period (Rina-an, 2001; Rinaman, 2003), prior to the development ofRH projections; this includes the NPY/catecholaminerojections, which are intact by P2 (Grove et al., 2003a).owever, in the NHP, NPY/catecholamine fibers in theRH were rare to sparse at G100 and G130. It is not until170 (late third trimester), after the development of ARHrojections that brainstem NPY fibers become common.ven at this late gestational age, the concentration ofbers is minimal compared with the adult (Fig. 6). It isecognized that brainstem NPY/catecholamine neuronslay several physiological roles, including stress respon-ivity, regulation of reproductive neuroendocrine axis andood intake. Thus, it is difficult to predict the physiologicalmportance of the late gestational and early postnatal de-elopment of these projections. It is also recognized thathe catecholaminergic projections to the hypothalamusay actually be intact earlier in the third trimester, but that
hey do not express NPY.While there were few NPY/AgRP or NPY/DBH fibers in
he PVH at G100, there was an abundance of single-labelPY fibers. The source of these fibers is unknown but thexpression in PVH neurons at this fetal age is consistentith local production. To date, nothing is known about thefferent targets of the PVH-NPY neurons. Our group hashown that these neurons are activated by fasting, indicat-
ng that they are responsive to changes in metabolic bal-nce, or the stress of fasting, but there is no evidence that
hey express leptin receptors or directly respond to leptin.t is possible that these neurons are not responsive toeripheral metabolic signals, such at leptin, insulin and guteptides, and thus are not part of the homeostatic feed-ack circuit.
One of the other significant species differences thatecame evident from this study is the lack of colocalizationf �MSH with CART. In the rodent ARH-�MSH neuronsoexpress CART (Vrang et al., 1999a), while �MSH neu-ons in the brainstem (NTS) do not. Our goal for thesetudies was to utilize CART as a marker of ARH-�MSHrojections in the NHP (as was done with NPY/AgRP);owever, this was not possible since CART and �MSHeptides were never colocalized in fiber terminals in theypothalamus. We were able to detect scattered CART-ireurons in the ARH of the adult NHP and these were nevero-labeled with �MSH-ir (Fig. 7). In situ hybridizationhowed a relatively low level of CART mRNA in this regions well (K. L. Grove et al., unpublished observations). This
s in contrast to Elias et al. (2001), who reported a denseopulation of CART-ir neurons in the ARH of the humanrain. Unfortunately, this study did not investigate colocal-
zation of CART with �MSH. Therefore, it was not possibleo determine the relative sources of �MSH fibers in theVH during the different fetal ages. Interestingly, althoughRH-NPY/AgRP-ir fibers were readily detectable in theVH by G130, �MSH-ir fibers were very sparse at this age.y near term (G170), �MSH fibers in the PVH were abun-ant. Within the ARH, �MSH-ir neurons were easily de-
ected by G130, indicating that the peptide is indeed pro-
uced but that projections to the PVH may not have de-eloped yet. This raises the possibility that there may be aifferent ontogeny for the development of the orexigenicNPY/AgRP) and anorexigenic (�MSH) projections out ofhe ARH.
CONCLUSION
n summary, the studies presented in this report demon-trate two key new findings. 1) The development of ARH-PY occurs during the third trimester in the NHP, which is
n contrast to the postnatal development in rodents. There-ore, health and dietary issues that arise during the thirdrimester may have long-term consequences on the devel-pment of these circuits. 2) There are several importantpecies specific differences in neuronal projections be-ween the NHP and rodent. The ARH is the major target ofrainstem NPY/catecholaminergic projections in the NHPs opposed to the PVH and DMH in rodents. Furthermore,ART protein was undetectable by immunohistochemistry
n � MSH neurons or fibers in the hypothalamus of theHP, which is in contrast to the rodent. These speciesifferences may be important in understanding the poten-ial of therapeutics/interventions for the treatment of obe-ity in humans.
cknowledgments—The authors would like to thank Dr. Andaornea for her expert assistance in confocal microscopy and Dr.avid Hess for his assistance with the RIAs. The authors wouldlso like to thank the Division of Animal Resources at ONPRC.his work was supported by National Institutes of Health grantsK060685, DK060685-S2, HD14643, HD18185 and RR00163.
REFERENCES
hima RS, Prabakaran D, Flier JS (1998) Postnatal leptin surge andregulation of circadian rhythm of leptin by feeding. Implications forenergy homeostasis and neuroendocrine function. J Clin Invest101:1020–1027.
ppleyard SM, Bailey TW, Doyle MW, Jin YH, Smart JL, Low MJ,Andresen MC (2005) Proopiomelanocortin neurons in nucleustractus solitarius are activated by visceral afferents: regulation bycholecystokinin and opioids. J Neurosci 25:3578–3585.
ell ME, Bhatnagar S, Akana SF, Choi S, Dallman MF (2000) Disrup-tion of arcuate/paraventricular nucleus connections changes bodyenergy balance and response to acute stress. J Neurosci 20:6707– 6713.
ouret SG, Draper SJ, Simerly RB (2004a) Formation of projectionpathways from the arcuate nucleus of the hypothalamus to hypo-thalamic regions implicated in the neural control of feeding behav-ior in mice. J Neurosci 24:2797–2805.
ouret SG, Draper SJ, Simerly RB (2004b) Trophic action of leptin onhypothalamic neurons that regulate feeding. Science 304:108–110.
roberger C, Johansen J, Johansson C, Schalling M, Hokfelt T (1998)The neuropeptide Y/agouti gene-related protein (AGRP) brain cir-cuitry in normal, anorectic, and monosodium glutamate-treatedmice. Proc Natl Acad Sci U S A 95:15043–15048.
ampbell RE, Ffrench-Mullen JMH, Cowley MA, Smith MS, Grove KL(2001) Hypothalamic circuitry of neuropeptide Y regulation of neu-roendocrine function and food intake via the Y5 receptor subtype.Neuroendocrinology 74:106–119.
hen P, Li C, Haskell-Luevano C, Cone RD, Smith MS (1999) Altered
expression of agouti-related protein and its colocalization with
C
C
d
E
E
E
E
G
G
G
G
G
G
H
H
H
K
K
K
K
K
L
L
L
P
P
R
R
S
S
S
S
S
T
V
V
V
B. E. Grayson et al. / Neuroscience 143 (2006) 975–986986
neuropeptide Y in the arcuate nucleus of the hypothalamus duringlactation. Endocrinology 140:2645–2650.
hronwall BM, DiMaggio DA, Massari VJ, Pickel VM, Ruggiero DA,O’Donohue TL (1985) The anatomy of neuropeptide-Y-containingneurons in rat brain. Neuroscience 15:1159–1181.
onrad CD, McEwen BS (2000) Acute stress increases neuropeptideY mRNA within the arcuate nucleus and hilus of the dentate gyrus.Brain Res Mol Brain Res 79:102–109.
e Quidt ME, Emson PC (1986) Distribution of neuropeptide Y-likeimmunoreactivity in the rat central nervous system-II. Immunohis-tochemical analysis. Neuroscience 18:545–618.
l Majdoubi M, Sahu A, Ramaswamy S, Plant TM (2000) Neuropep-tide Y: A hypothalamic brake restraining the onset of puberty inprimates. Proc Natl Acad Sci U S A 97:6179–6184.
lias CF, Lee CE, Kelly JF, Ahima RS, Kuhar M, Saper CB, ElmquistJK (2001) Characterization of CART neurons in the rat and humanhypothalamus. J Comp Neurol 432:1–19.
llacott KL, Cone RD (2004) The central melanocortin system and theintegration of short- and long-term regulators of energy homeosta-sis. Recent Prog Horm Res 59:395–408.
veritt BJ, Hokfelt T, Terenius L, Tatemoto K, Mutt V, Goldstein M(1984) Differential co-existence of neuropeptide Y (NPY)-like im-munoreactivity with catecholamines in the central nervous systemof the rat. Neuroscience 11:443–462.
rove KL, Allen S, Grayson BE, Smith MS (2003a) Postnatal devel-opment of the hypothalamic neuropeptide Y system. Neuroscience116:393–406.
rove KL, Chen P, Koegler FH, Schiffmaker A, Smith MS, Cameron JL(2003b) Fasting activates neuropeptide Y neurons in the arcuatenucleus and the paraventricular nucleus in the rhesus macaque.Brain Res Mol Brain Res 113:133–138.
rove KL, Brogan RS, Smith MS (2001) Novel expression ofneuropeptide Y (NPY) mRNA in hypothalamic regions duringdevelopment: Region specific effects of nutritional deprivation onNPY and agouti related protein mRNA. Endocrinology 142:4771–4776.
rove KL, Campbell RE, Ffrench-Mullen JM, Cowley MA, Smith MS(2000) Neuropeptide Y Y5 receptor protein in the cortical/limbicsystem and brainstem of the rat: expression on gamma-aminobu-tyric acid and corticotropin-releasing hormone neurons. Neuro-science 100:731–740.
rove KL, Grayson BE, Glavas MM, Xiao XQ, Smith MS (2005)Development of metabolic systems. Physiol Behav 86:646–660.
rove KL, Smith MS (2003) Ontogeny of the hypothalamic neuropep-tide Y system. Physiol Behav 79:47–63.
ahn TM, Breininger JF, Baskin DG, Schwartz MW (1998) Coexpres-sion of Agrp and NPY in fasting-activated hypothalamic neurons.Nat Neurosci 1:271–272.
askell-Luevano C, Chen P, Li C, Chang K, Smith MS, Cameron JL,Cone RD (1999) Characterization of the neuroanatomical distribu-tion of agouti-related protein immunoreactivity in the rhesus mon-key and the rat. Endocrinology 140:1408–1415.
okfelt T, Broberger C, Zhang X, Diez M, Kopp J, Xu Z, Landry M, BaoL, Schalling M, Koistinaho J, DeArmond SJ, Prusiner S, Gong J,Walsh JH (1998) Neuropeptide Y: some viewpoints on a multifac-eted peptide in the normal and diseased nervous system. BrainRes Brain Res Rev 26:154–166.
agotani Y, Hashimoto T, Tsuruo Y, Kawano H, Daikoku S, Chihara K(1989) Development of the neuronal system containing neuropep-
tide Y in the rat hypothalamus. Int J Dev Neurosci 7:359–374.
alra SP (1986) Neural circuitry involved in the control of LHRHsecretion: a model for preovulatory LH release. Front Neuroendo-crinol 9:31–75.
alra SP, Dube MG, Pu S, Horvath TL, Kalra PS (1999) Interactingappetite-regulating pathways in the hypothalamic regulation ofbody weight. Endocr Rev 20:68–100.
oegler FH, Grove KL, Schiffmacher A, Smith MS, Cameron JL (2001)Central melanocortin receptors mediate changes in food intake inthe rhesus macaque. Endocrinology 142:2586–2592.
outcherov Y, Mai JK, Paxinos G (2003) Hypothalamus of the humanfetus. J Chem Neuroanat 26:253–270.
arsen PJ, Tang-Christensen M, Stidsen CE, Madsen K, Smith MS,Cameron JL (1999) Activation of central neuropeptide Y Y1 recep-tors potently stimulates food intake in male rhesus monkeys. J ClinEndocrinol Metab 84:3781–3791.
i C, Chen P, Smith MS (1999) Morphological evidence for directinteraction between arcuate nucleus neuropeptide Y (NPY) neu-rons and gonadotropin-releasing hormone neurons and the possi-ble involvement of NPY Y1 receptors. Endocrinology 140:5382–5390.
i C, Chen P, Smith MS (2000) Corticotropin releasing hormone (CRH)neurons in the paraventricular nucleus (PVH) are direct targets forneuropeptide Y (NPY) neurons in the arcuate nucleus (ARH): Ananterograde tracing study. Brain Res 854:122–129.
axinos G, Huang X-F, Toga AW (1999) The rhesus monkey brain instereotaxic coordinates. New York: Academic Press.
orte D Jr, Seeley RJ, Woods SC, Baskin DG, Figlewicz DP, SchwartzMW (1998) Obesity, diabetes and the central nervous system.Diabetologia 41:863–881.
inaman L (2001) Postnatal development of catecholamine inputs tothe paraventricular nucleus of the hypothalamus in rats. J CompNeurol 438:411–422.
inaman L (2003) Postnatal development of hypothalamic inputs tothe dorsal vagal complex in rats. Physiol Behav 79:65–70.
agawa N, Yura S, Itoh H, Kakui K, Takemura M, Nuamah MA, OgawaY, Masuzaki H, Nakao K, Fujii S (2002) Possible role of placentalleptin in pregnancy: a review. Endocrine 19:65–71.
aper C, Chou T, Elmquist J (2002) The need to feed. Homeostaticand hedonic control of eating. Neuron 36:199.
chubring C, Kiess W, Englaro P, Rascher W, Dotsch J, Hanitsch S,Attanasio A, Blum WF (1997) Levels of leptin in maternal serum,amniotic fluid, and arterial and venous cord blood: relation toneonatal and placental weight. J Clin Endocrinol Metab 82:1480–1483.
chwartz MW, Woods SC, Porte D Jr, Seeley RJ, Baskin DG (2000)Central nervous system control of food intake. Nature 404:661–671.
inger LK, Kuper J, Brogan RS, Smith MS, Grove KL (2000) Novelexpression of hypothalamic neuropeptide Y during postnatal de-velopment in the rat. Neuroreport 11:1075–1080.
ang-Christensen M, Vrang N, Larsen PJ (2001) Glucagon-like pep-tide containing pathways in the regulation of feeding behaviour. IntJ Obes Relat Metab Disord 25 (Suppl 5):S42–S47.
rang N, Grove KL (2004) The brainstem preproglucagon system in anon-human primate brain. Soc Neurosci Abstr 928.910.
rang N, Larsen PJ, Clausen JT, Kristensen P (1999a) Neurochemicalcharacterization of hypothalamic cocaine- amphetamine-regulatedtranscript neurons. J Neurosci 19:RC5.
rang N, Larsen PJ, Clausen JT, Kristensen P (1999b) Neurochemicalcharacterization of hypothalamic cocaine-amphetamine-regulated
transcript neurons. J Neurosci 19:1–8.
(Accepted 18 August 2006)(Available online 9 October 2006)
