| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
Graduate School Neurosciences Amsterdam (A.P.G., U.A.U., D.F.S.), Netherlands Institute for Brain Research, 1105 AZ Amsterdam ZO, The Netherlands; Endocrine Unit (A.P.G., S.R.B.), Imperial College School of Medicine, Hammersmith Hospital, London W12 0NN, United Kingdom
Address all correspondence and requests for reprints to: D. F. Swaab, Netherlands Institute for Brain Research, Meibergdreef 33, 1105 AZ Amsterdam ZO, The Netherlands. E-mail: tgoldstone{at}yahoo.com
Abstract
Animal studies have demonstrated the importance of orexigenic NPY and agouti-related protein (AGRP) hypothalamic neurons, which are inhibited by the adipocyte hormone leptin, in the regulation of body weight and neuroendocrine secretion. We have examined NPY and AGRP neurons in postmortem human hypothalami from controls, Prader-Willi syndrome and other obese subjects, using quantitative immunocytochemistry (ICC) and in situ hybridization, to identify causes of leptin resistance in human obesity. Using combined ICC and in situ hybridization, AGRP, but not POMC, was colocalized with NPY in infundibular nucleus neurons. Infundibular nucleus (including median eminence) NPY ICC staining or mRNA expression, and AGRP ICC staining, increased with premorbid illness duration. NPY ICC staining and mRNA expression were reduced in obese subjects, but AGRP ICC staining was unchanged, correcting for illness duration. This suggests normal responses of NPY and AGRP neurons to peripheral signals, such as leptin and insulin, in human illness and obesity. The pathophysiology of obesity and illness-associated anorexia appear to lie in downstream or separate neuronal circuits, but the infundibular neurons may mediate neuroendocrine responses to illness. The implications for pharmacological treatment of human obesity are discussed.
ANIMAL STUDIES HAVE identified hypothalamic neuropeptide pathways involved in regulating feeding and metabolism (1). Major stimulatory feeding neuropeptides are NPY and AGRP (2, 3). The latter is an endogenous 
-MSH antagonist at the melanocortin-4 receptor (MC4R), colocalized with NPY in arcuate nucleus (ARC) neurons (4), called the infundibular nucleus (INF) in humans. NPY/AGRP neurons, activated by the fall in plasma leptin during starvation (5), form part of the homeostatic loop to increase food intake, reduce energy expenditure, and alter pituitary hormone secretion (6, 7, 8). They interact with -MSH, produced by ARC POMC-expressing neurons, which inhibit feeding via the MC4R, particularly in the paraventricular nucleus (PVN) and lateral hypothalamic area (LHA) (9, 10, 11).
Although similar effects on the hypothalamic-pituitary (HP) axis are seen during human starvation and illness (12, 13), disadvantageous anorexia and increased metabolic rate are often features of severe illness, contributing to cachexia, increased morbidity, and mortality (14). These conditions may be mediated by disturbances in the normal physiological response of hypothalamic pathways, perhaps via circulating cytokines or endotoxins (15).
Hypothalamic NPY and AGRP overactivity in rodents leads to obesity and neuroendocrine abnormalities (3, 16), as with leptin deficiency or receptor defects in rodents and humans (17). Identification of neuropeptides altered in human illness and obesity may provide pharmacological strategies for treatment. The importance of melanocortin pathways has been suggested by the childhood-onset obesity associated with mutations in POMC and MC4R genes (17). The role of hypothalamic NPY and AGRP in human body weight and neuroendocrine regulation is unknown (18).
Prader-Willi syndrome (PWS) is the commonest genetic syndromal cause of human obesity, characterized by insatiable hunger from childhood, mental retardation, hypogonadism, and GH deficiency (19). The molecular genetic cause is nonexpression of paternal genes in the PWS region on chromosome 15q11–13 (20). Many features are thought to arise from developmental abnormalities in the hypothalamus (21). Regulation of leptin secretion is normal in PWS (22, 23), with elevated leptin levels in proportion to their obesity, as seen in non-PWS obese adults (24). Available mouse models of PWS do not display hyperphagia or obesity (25, 26, 27), and so investigation of their cause must currently use postmortem human material (28, 29, 30).
We hypothesized that hypothalamic NPY/AGRP neurons mediate behavioral, metabolic, and neuroendocrine abnormalities or responses in human illness and obesity, with their activation in both PWS and non-PWS obesity, as a part of leptin resistance, contributing to energy storage. We therefore examined: 1) localization of AGRP and NPY in hypothalamic neurons using combined immunocytochemistry (ICC) and in situ hybridization (ISH), and 2) differences in hypothalamic NPY and AGRP peptide and NPY mRNA expression between control, PWS and other obese subjects, with quantitative ICC and ISH, using postmortem human material.
Materials and Methods
Hypothalamic material
Hypothalamic material was used from 22 control adults (4 premenopausal females, 5 postmenopausal females, 9 intact males, and 4 hypogonadal males who had undergone bilateral orchidectomy), 6 PWS obese adults, and 4 non-PWS obese adults (body mass index >30 kg/m2). Clinicopathological details are given in Table 1
.
|
). Examination of the clinical records was performed to calculate the postmortem delay (PMD), fixation time, tissue storage time (from date of autopsy to study), underlying medical problems, cause of death and the duration of the premorbid illness. Hypothalami were fixed in 4% formaldehyde, dehydrated, paraffin-embedded and serially sectioned at 6 µm. Every 100th section was stained with thionine for orientation to determine the level of the INF and median eminence (ME). NPY and AGRP immunocytochemistry
Every 100th section from the ME to the mamillary bodies was mounted for NPY ICC and adjacent sections for AGRP ICC. The NPY (Niepke 26/11/1988, Netherlands Institute for Brain Research) and AGRP (Phoenix Pharmaceuticals, Inc., Belmont, CA; catalog no. H-003-53) polyclonal rabbit antibodies were raised, respectively, against porcine NPY(1–36) (Sigma, St. Louis, MO) coupled to thyroglobulin (Sigma) with glutaraldehyde (Merck, Rahwah, NJ), and human AGRP(83–132). Mounted sections were dried at 37 C, deparaffinized, and rehydrated. Antigen retrieval was used for AGRP, but not NPY, ICC staining, by placing sections in preheated TBS (0.15 M NaCl, 0.05 M Tris, pH 7.6), in a 90 C water bath for 30 min. For both NPY and AGRP ICC staining, sections were then preincubated in TBS-milk [5% milk (Elk, Campina Melkunie, Eindhoven, The Netherlands) in TBS] for 1 h at room temperature (RT). Primary antiserum diluted in Supermix-milk (a solution of 0.5%Triton X-100, 0.25% gelatin, 5% milk powder in TBS, pH 7.6) was added at a concentration of 1:1000 (NPY) or 1:1500 (AGRP) for 1 h at RT, followed by an overnight incubation at 4 C. After rinsing in TBS-milk and TBS, sections were incubated in goat antirabbit IgG (NPY: Betsy, AGRP: Akon) 1:100 in TBS-Triton at RT, rinsed in TBS, incubated in peroxidase-antiperoxidase 1:1000 in TBS-Triton at RT, rinsed in TBS, incubated in DAB-nickel solution (0.5 mg/ml 3,3-diaminobenzidine, 0.01% H2O2, 2.33 mg/ml ammonium nickel sulfate in TBS) at RT, dehydrated, cleared in xylene and coverslipped.
Antibody specificity
Specificity of the NPY antibody has been previously examined (31) and was further confirmed by the absence of ICC staining in the human INF using preimmune serum or antiserum preadsorbed in the solid-phase with porcine NPY (pNPY) peptide (Peninsula Laboratories, Inc., Merseyside, UK) (fixed overnight with 4% formaldehyde onto gelatin-coated nitrocellulose filter paper, 0.1 µm), and the presence of staining when pre-adsorbed with AGRP(83–132) peptide (Phoenix Pharmaceuticals, Inc.), which did not differ from unadsorbed serum. For the AGRP antiserum, no staining was seen after omission of the primary antiserum or preadsorption of the antiserum with AGRP(83–132) peptide, whereas no reduction in staining was seen after preadsorption with pNPY peptide.
Immunocytochemistry quantification
An estimate of the volume of DAB-nickel precipitate throughout the INF/ME as a measure of total NPY- and AGRP-immunoreactivity (perikarya and fibers) was made by an image analysis system (IBAS, Kontron Elektronik, Munich, Germany) connected to a camera and plain objective microscope. The ICC stained region of the INF/ME for one half of the hypothalamus was outlined for each section at low magnification (Fig. 1, A and B), while blinded to the group of origin, and the area retrieved as a series of 10x objective images loaded for mask extraction (32). The ICC stained region of the INF/ME was identified by 1) a medial border defined by the midline and third ventricle, 2) a ventral border defined by the ventral hypothalamic surface, and lack of fiber staining in the outer layer of the ME, and 3) a distinct dorsolateral border, abutting a region of less dense fiber staining, coinciding with a region containing blood vessels (Fig. 1, A and B). This border of the INF was therefore easily distinguishable from the denser fiber staining of the VMH and DMH more dorsolaterally. The selected masks covered stained neuronal cell bodies and densely stained fibers. In each section, the masking area of NPY or AGRP ICC staining within the outlined area was calculated. For each subject, volume estimates of ICC staining were performed by a conversion program, based upon multiplication of the masking area by sample frequency over the whole length of the INF/ME that contained positive cell bodies. Mean (± SD) number of sections quantified per subject were 8 ± 2 for NPY and 7 ± 2 for AGRP. To determine the variability in the ICC protocol, adjacent sections throughout the whole length of the INF/ME for 2 control subjects (no. 93–139 and 98–089) were included in identical and separate runs for both NPY and AGRP ICC staining. For NPY the within run coefficient of variation was 8.9% and the between run coefficient of variation was 6.8%, and for AGRP 12.1% and 18.6%, respectively. Quality controls using adjacent sections were also included in separate ICC runs to ensure similar staining was achieved.
|
Mounting, deparaffinization, deproteination, and delipidation of the paraffin-embedded formalin-fixed sections for ISH was performed as described previously (33). An additional pretreatment step was added between deparaffinization and deproteination, with sections placed in preheated TBS, pH 7.6, in a 90 C water bath for 30 min. Antisense and sense 48-bp oligonucleotides (Eurogentec, Maastricht, The Netherlands), corresponding to bp 99–146 of the human NPY mRNA sequence (34), were 35S-deoxy-ATP end-labeled, dissolved in hybridization buffer to a concentration of 1 x 106 cpm/70 µl (33), and incubated with sections overnight at 45 C. After posthybridization washing (2 x SSC at 55 C, 1 x SSC at 55 C, twice in 0.1 x SSC at 55 C and twice in 0.1 x SSC at RT), sections were dehydrated and directly applied to Hyperßmax film (Amersham Pharmacia Biotech, Roosendaal, The Netherlands) for 8 d for autoradiography.
In situ hybridization quantification
For NPY ISH, three sections were taken from each subject, each being adjacent to the three INF/ME sections with the greatest masking area for NPY ICC. The average NPY ICC staining masking area for these sections was significantly correlated with the overall NPY ICC staining masking volume for the whole INF/ME determined from every 100th section (r = +0.95, P < 0.001), indicating that these sections gave a good representation of NPY ICC staining for the whole nucleus. The densitometric quantification of ISH signal by autoradiography using radioactive standards was performed as described previously (12). For each section, the background corrected mean density was multiplied by the area of the outlined region, and the average value of the sections was obtained for each subject. All sections were included in a single experiment. Two control subjects (nos. 93–139 and 93–025) were excluded from the analysis because of damage to sections.
Immunocytochemistry/ISH colocalization
For colocalization studies, sequential AGRP ICC, using DAB as a chromogen, followed by NPY ISH was performed, using similar techniques to those described previously (35). No additional prehybridization water bath treatment was performed during the ISH protocol. For emulsion autoradiography following ISH, slides were dipped in Kodak NTB2 emulsion and incubated for 5–8 d at 4 C, before developing and fixing. Sections were then counterstained with thionine, dehydrated, cleared in xylene, and coverslipped. Pilot studies revealed no overall effect of the ICC protocol on the ISH signal obtained and no nonspecific ISH signal was produced using sense probes after ICC (Fig. 1D). Colocalization of NPY in POMC neurons was also examined by sequential POMC ICC, using a sheep polyclonal -MSH antibody (Chemicon Intl. Inc., Temecula, CA) (9), followed by NPY ISH. The POMC ICC protocol used antigen retrieval for 30 min in 90 C Tris, pH 9, incubation in primary antibody (1:100,000, 1 h at RT, overnight at 4 C), biotinylated rabbit antisheep IgG (Vector Laboratories, Inc., Burlingame, CA; 1:400), ABC elite (Vector Laboratories, Inc.; 1:800), followed by DAB. The -MSH antibody was found to cross-react with -MSH, desacetyl -MSH and ACTH(1–39), but not ß-MSH, 
-MSH, ß-lipotropic hormone(61–77), neuropeptide EI, pNPY, or AGRP(83–132), using peptide spots fixed onto gelatin-coated nitrocellulose filter paper with 4% formaldehyde, stained with an identical ICC procedure to that used for the hypothalamic sections.
Statistics
Simple and multiple linear regression analysis was performed to assess any effect of age, PMD, fixation time, storage time, and duration of premorbid illness (independent variables) on ICC staining masking volumes or mRNA expression (dependent variable). Differences between groups were analyzed using Mann-Whitney test, analysis of covariance (ANCOVA) and multiple linear regression analysis. ICC staining volumes, mRNA expression data, and premorbid illness duration were log10 transformed for ANCOVA and regression analysis, as they were not normally distributed. To calculate the degree to which dependent variables were reduced in obesity, when adjusting for differences in significant covariates, the obesity regression coefficient (ß) was calculated by multiple regression using a group variable (0 = control, 1 = obese) as an independent variable, and converted into a figure for obese as a % of control, equal to 100/10-ß. A P value of less than 0.05 was considered to be significant.
Results
Colocalization and coregulation of NPY and AGRP
Cell bodies for both NPY and AGRP were found by ICC in overlapping areas of the INF and adjacent ME (inner layer only)/pituitary stalk in control, PWS and non-PWS obese subjects (Fig. 1, A and B). NPY mRNA was expressed in identical areas on emulsion autoradiography. No signal was detected using the sense NPY probe. The average NPY ICC staining area was positively correlated to the average NPY mRNA expression for adjacent sections (r = +0.42, P = 0.02). NPY/AGRP colocalization was examined in four control males (no. 94–118, no. 98–081, no. 93–025, no. 98–033), 1 PWS male (no. 95–104) and 1 non-PWS obese female (no. 97–065). Almost all AGRP ICC staining positive cells in the INF, displayed NPY mRNA expression in all subjects (Fig. 1, C, E, and G). There was a positive correlation between NPY and AGRP ICC staining volumes in the INF/ME (r = +0.44, P = 0.01). By contrast no colocalization was seen between NPY mRNA and POMC neurons (Fig. 1, F and H).
Effects of postmortem delay, fixation time, or storage time
Neuropeptides appear to be very stable in postmortem human material allowing quantitative ICC and ISH studies, and through appropriate matching of controls, detailed image analysis, and appreciation and adjustment for confounding variables, the results appear robust and reproducible (36). There was no significant correlation between the NPY ICC staining volume or mRNA expression, or AGRP ICC staining volume, in the INF/ME and PMD or fixation time in controls (r = -0.17 to +0.15, P = 0.5–1.0). There was also no significant influence of these covariates in ANCOVA between control and PWS (PMD: P = 0.4–0.9, fixation time: P = 0.2–0.9) or non-PWS obese (PMD: P = 0.7–0.8, fixation time: P = 0.5–1.0) groups.
There was, however, a significant negative relationship between storage time and NPY mRNA expression (r = -0.53, P = 0.02), but not NPY or AGRP ICC staining volume (r = +0.10, P = 0.6 and r = +0.21, P = 0.4, respectively) in controls. A negative effect of storage time on NPY mRNA expression (but not NPY or AGRP ICC staining) remained when including premorbid illness duration (see below) in multiple regression analysis, in both control (P = 0.008) and obese (PWS and non-PWS, P = 0.05) subjects. Adjustment was therefore made for differences in storage time in comparison of NPY mRNA expression between groups (Fig. 2B).
|
The positive relationship between NPY ICC staining and premorbid illness duration did not reach statistical significance in controls (r = +0.30, P = 0.18) but was significant in obese PWS subjects (r = +0.92, P = 0.01) (Figs. 3A and 4). NPY mRNA expression significantly increased with premorbid illness duration in controls (r = +0.71, P < 0.001) and obese PWS subjects (r = +0.66, P = 0.01) (Figs. 3B and 5), when including storage time in multiple regression analysis (overall r2 = 0.67 and 0.92, respectively). There was a significant positive relationship between AGRP ICC staining and premorbid illness duration in both controls (r = +0.82, P < 0.001) and obese PWS subjects (r = +0.91, P = 0.01) (Figs. 3C and 6). Adjustment was therefore made for differences in premorbid illness duration, in comparison of NPY ICC staining or mRNA expression and AGRP ICC staining between groups. The significance of these relationships in controls was unchanged, when including fixation time or postmortem delay as an independent variable in multiple regression analysis, which was not a significant covariate by itself (P = 0.5–0.9 for fixation time, P = 0.3–0.9 for PMD). Furthermore, in controls the positive relationship of NPY ICC, NPY mRNA and AGRP ICC with premorbid illness duration persisted with exclusion of the six subjects with long fixation time (>65 d) and PMD (>41 h) (r = + 0.443, P = 0.09; r = + 0.65, P = 0.01; r = + 0.83, P < 0.001).
|
|
|
|
NPY and AGRP in human obesity
There was no significant difference in NPY ICC staining volume, NPY mRNA expression or AGRP ICC staining volume, in PWS vs. non-PWS obese subjects (P = 0.8, 0.4, 0.7, respectively) (Fig. 2), and adjusting for premorbid illness duration and, for NPY mRNA expression, storage time (P = 0.9, 0.9, 0.7, respectively).
There was a significant reduction in NPY ICC staining volume in all obese subjects (PWS and non-PWS) compared with controls (median obese 0.149 mm3 vs. controls 0.550 mm3, P < 0.001, Fig. 2A). When correcting for differences in premorbid illness duration by multiple regression analysis, NPY ICC staining volume in all obese subjects was 27.7% (SE range 21.9–35.0%, P < 0.001) of that in control subjects (Figs. 2A and 4). This reduction was significant in both PWS and non-PWS obese subjects, without or with adjustment for differences in premorbid illness duration, which by itself was a significant covariate (both P < 0.001) (Figs. 2A, 3A, and 4). The reduction in all obese subjects remained significant (all P < 0.001) when adjusting for differences in fixation time or PMD, which by themselves were not significant covariates (P = 0.3 and 0.5), or when excluding the six control subjects and two obese subjects with long fixation time (>65 d) and PMD (>41 h).
There was a tendency for a reduction in NPY mRNA expression in all obese subjects (PWS and non-PWS) compared with controls (median obese 3.93 vs. controls 7.18, P = 0.07, Fig. 2B). When correcting for differences in premorbid illness duration and storage time by multiple regression analysis, NPY mRNA expression in all obese subjects was significantly reduced, being 45.2% (SE range 32.6–62.7%, P = 0.02) of that in control subjects (Figs. 3B and 6). Premorbid illness duration and storage time were significant covariates (both P < 0.001). The reduction remained significant (P = 0.04 and P = 0.02) with adjustment for differences in fixation time or PMD, which by themselves were not significant covariates (P = 0.7 and 0.11). However, due to the smaller number of subjects, only a trend remained when excluding those control and obese subjects with long fixation times and PMD (P = 0.08). The reduction in NPY mRNA expression in obese PWS subjects alone compared with controls, correcting for premorbid illness duration and storage time, did not quite reach statistical significance (P = 0.07) (Figs. 3B, 4B, and 6). The biological significance of this tendency for a reduction in NPY mRNA expression in obesity may therefore be questionable.
There was no reduction in NPY seen in hypogonadal male or female controls, as no significant differences in NPY ICC staining or mRNA expression were found between noncastrated and castrated male controls (median NPY ICC staining volume 0.439 vs. 0.691 mm3, P = 0.1; median NPY in situ signal (arbitrary units) 7.2 vs. 29.7, P = 0.6), or pre- and postmenopausal female controls (median NPY ICC staining 0.673 vs. 0.349 mm3, P = 0.3; median NPY in-situ signal 8.5 vs. 4.7, P = 1.0) (Fig. 2, A and B), and when correcting for differences in premorbid illness duration, and for NPY mRNA expression, storage time (P = 0.1, 0.7, 0.2, 0.2, respectively).
There was a tendency for a reduction in AGRP ICC staining volume in all obese subjects (median obese 0.047 mm3 vs. controls 0.116 mm3, P = 0.05), PWS (P = 0.15) or non-PWS obese subjects (P = 0.13) alone, compared with controls. However this was no longer apparent when adjusting for premorbid illness duration (P = 0.7, 0.9, 0.7, respectively), which by itself was a significant covariate (all P < 0.001) (Figs. 2C, 3C, and 6). AGRP ICC staining volume in all obese subjects was 99.8% (SE range 68.9–144.6%, P = 1.0) of that in control subjects, when adjusting for premorbid illness duration in multiple regression analysis. This remained nonsignificant (P = 0.7 and 1.0) with adjustment for differences in fixation time or PMD, which by themselves were not significant covariates (P = 0.7 and 0.5), or when excluding those control and obese subjects with long fixation times and PMD (P = 0.6).
Discussion
Examining hypothalamic NPY and AGRP neurons in postmortem material, using quantitative ICC and ISH, we provide the first evidence to support physiological changes in their activity in human obesity and illness. NPY or AGRP peptide ICC staining and NPY mRNA expression were increased in the INF/ME in prolonged illness. NPY ICC staining and mRNA expression were decreased in PWS and non-PWS obese adults, while AGRP staining was unchanged, excluding activation of these neurons in the pathophysiology of their obesity.
NPY and AGRP neuronal colocalization
NPY and AGRP were colocalized in INF/ME cell bodies, as seen previously (4, 37). No abnormal neuroanatomy or colocalization was seen in PWS or non-PWS obese subjects. By contrast, no colocalization was seen between NPY mRNA and POMC peptide in control or obese subjects. NPY peptide staining was positively correlated with NPY mRNA expression, suggesting similar changes in peptide synthesis.
NPY and AGRP in human illness
Activity of NPY/AGRP neurons, as measured by NPY ICC staining or mRNA expression, and AGRP ICC staining, was positively correlated with the duration of the premorbid illness, in both controls and obese PWS subjects. This agrees with a previous study that showed, by RIA, increases in NPY in the INF, and terminal fields in the PVN or dorsomedial hypothalamus (DMN), in humans who died after prolonged respiratory illness vs. those with more acute death (38). This suggests that the anorexia seen in illness is not due to a reduction in the activity of NPY/AGRP neurons. Similarly, no reduction in ARC NPY is seen in animal models of anorexia associated with carcinoma, endotoxin, or cytokines (39, 40).
Diminished food intake and weight loss during illness-associated anorexia, leading to reductions in plasma leptin, may be responsible for the increased NPY/AGRP neuronal activity (5, 41, 42, 43). This would be an appropriate response to stimulate food intake and reduce energy expenditure, as part of a homeostatic loop. Additional explanations for increased NPY activity may be HP-adrenal axis stimulation (13) because glucocorticoids increase hypothalamic NPY (44), or altered insulin secretion or resistance, as insulin decreases hypothalamic NPY (45), and perhaps AGRP (46, 47).
Our results suggest that anorexic factors in illness act either distally to NPY/AGRP or in separate neuronal circuits. Such pathways could be suitable drug targets to improve survival in such circumstances. Increased MC4R signaling, via increased POMC neuronal activity, might counteract increased AGRP activity, although rodent studies have been inconsistent (39, 48). However, the increase in both endogenous hypothalamic NPY and AGRP in human illness, raises doubts about the success of exogenous NPY agonists or -MSH antagonists as such treatments.
NPY and AGRP also mediate effects of starvation on the HP axis in rodents, that occurs, at least partially, in response to falls in leptin (49). This includes suppression of the HP-thyroid and -gonadal axes (6, 7), and stimulation of the HP-adrenal axis (16), via NPY or AGRP innervation of CRH, TRH, and GnRH neurons (37, 50, 51). Our results support the hypothesis that, as in starved rodents, activation of NPY/AGRP neurons may mediate similar neuroendocrine changes that occur during human illness, with increased cortisol secretion, reduced TSH and thyroxine secretion (nonthyroidal illness), and hypogonadism (8, 12, 13).
NPY and AGRP in PWS and obesity
PWS is associated with life-threatening hyperphagia, as well as GH deficiency and hypothalamic hypogonadism. Given the effects of chronic NPY agonism (6, 16) or melanocortin antagonism (3, 52), and reduced leptin or MC4R signaling through mutations in both rodents and humans (17, 53, 54, 55), increased activity of hypothalamic NPY/AGRP neurons might be responsible for hyperphagia in PWS, and NPY hyperactivity for hypogonadism and GH deficiency, perhaps through defects in leptin signaling.
However, there is no evidence from our study for increased activity of these neurons in obese PWS adults. On the contrary, our results suggest decreased activity of NPY neurons because NPY ICC staining was significantly decreased, and there was a statistical tendency for a similar decrease in NPY mRNA expression. This reduction in NPY did not appear to be a primary abnormality because it was not seen in 2 PWS infants (aged 6 and 8 months, no. 98–168 and 99–079), before the onset of hyperphagia and obesity (unpublished results). The decrease in NPY in obese PWS adults did not appear to be due to hypogonadism because it was not seen in castrated male controls or postmenopausal women, but a similar reduction was seen in non-PWS obesity. This is consistent with inhibition of INF NPY neurons in obesity, through elevations in peripheral signals, such as plasma leptin (5) and insulin (45), though others factors may be involved. One non-PWS obese subject was on prednisolone, but this is unlikely to have adversely affected the results because glucocorticoids increase hypothalamic NPY (44).
Similar reductions in hypothalamic NPY are seen in diet-induced obesity in rodents (1), in response to hyperleptinaemia, though this may only be transient (56). Mouse strains more resistant, but not those most susceptible, to chronic high-fat feeding, maintain reduced NPY mRNA expression (57). We found no evidence for activation of a separate NPY neuron population in the DMN of PWS and non-PWS obese humans, as seen in rodent models of obesity (58, 59). It is not yet clear if this has any aetiological role in their obesity. There may be a species difference because no staining for neuropeptides or neurotransmitters has yet specifically identified neuronal cell bodies in the human, unlike the rodent, DMN (60). A similar augmentation of NPY expression in the DMN is seen in VMH-lesioned rats, which display hyperphagia, obesity, and leptin resistance (1). However, in this model, their reduction in ARC NPY mRNA, and PVN NPY release, is also associated with an increased sensitivity to the orexigenic actions of NPY, related to a concomitant increase in NPY Y1 receptor expression and perhaps also the reduction in ARC POMC mRNA (61, 62, 63). Study of NPY receptor and POMC expression will therefore also be important in interpreting the consequences of the INF NPY reduction in PWS and non-PWS obesity seen in our study.
By contrast, AGRP peptide ICC staining was unchanged in PWS and non-PWS obese adults. This difference between NPY and AGRP changes in obesity has also been seen in rodents, where dietary-induced obesity led to a reduction in ARC NPY but no change in AGRP mRNA in one study (59). This may represent a different dose response of target neuropeptides to peripheral signals such as leptin (64). The neuronal response to falling leptin during starvation and illness (increasing both NPY and AGRP) may be distinct from the response to hyperleptinaemia seen in obesity (decreasing only NPY). Alternatively, there may be adaptation of AGRP responses to long-term changes in nutritional state. In support of this, we have found very high levels of NPY ICC staining (INF/ME masking volume 0.924 mm3) and mRNA expression (26.0 arbitrary units) in a 34-yr-old female with anorexia nervosa (no. 99–058), but average levels of AGRP ICC staining (INF/ME masking volume 0.071 mm3) (unpublished results). However, both AGRP and NPY mRNA were only transiently reduced after exposure of mice to a high-fat diet in a longitudinal study (56). The relatively small number of subjects might also explain this apparent difference between NPY and AGRP in our study.
Prolonged illness in obese PWS adults was, however, associated with significant increases in NPY and AGRP ICC staining, and NPY mRNA expression. As in controls, this increase may be caused by reductions in food intake, body weight, and leptin. Activation of similar body weight defense mechanisms may contribute to the failure of dietary attempts to reduce body fat in obese humans, by increasing appetite and reducing energy expenditure. In this case, NPY antagonists or MC4R agonists might help to maintain or promote weight loss during dieting.
These results suggest an appropriate response of NPY/AGRP neurons in human obesity and support the conclusion that NPY/AGRP neurons in the INF/ME are not defective in PWS. Leptin resistance in PWS thus appears to be independent of the leptin transport/receptor/signaling pathway, at least in NPY/AGRP neurons, because obesity occurs without increases in these orexigenic hypothalamic neuropeptides, but in fact a decrease in NPY peptide or mRNA. This suggests normal leptin receptor expression in the PWS INF/ME, supported by our demonstration of long isoform leptin receptor expression in PWS lymphocytes (23). The small number of other obese subjects in our study means that general application of this conclusion to the heterogeneous nature of non-PWS obesity must, however, be cautious.
Conclusion
This study has demonstrated appropriate changes in hypothalamic NPY and AGRP, colocalized in human INF/ME neurons, in illness and obesity (Table 2). The pathophysiology of illness-associated anorexia, and obesity in PWS and non-PWS subjects, appears to lie in downstream or separate circuits. Particular candidates include anorexigenic neurons in the PVN, containing oxytocin [total and oxytocin PVN cell number are reduced in PWS (29)], CRH (1), IL-1ß (39, 65), in the INF expressing POMC and CART (1), serotoninergic pathways (66), and orexigenic MCH and orexin neurons in the LHA (9, 10). However, our results support a physiological role for hypothalamic NPY and AGRP in human energy homeostasis and neuroendocrine regulation.
|
Acknowledgments
R. S. Williams (Shriver Center, Walton, MA), A. Schulze (Odense University Hospital, Denmark) and M. Bojsen-Møller (The John Kennedy Institute, Glostrup, Denmark), A. Holland and J. Xuereb (University of Cambridge, Cambridge, UK), M. E. J. Schipper (OLVG, Amsterdam, The Netherlands), H. M. Evenhuis (Mental Retardation Institute De Hooge Burch, Zwammerdam, The Netherlands), and R. A. C. Roos (AZL, Leiden, The Netherlands), P. T. Botha and L. Thornton (Hastings Memorial Hospital, Hastings, New Zealand), U. Eiholzer (Foundation for Growth Puberty Adolescence, Zürich, Switzerland), and C. Markwalder (Kantonspital, St. Gallen, Switzerland) for provision of brain material and clinical information from PWS subjects; Netherlands Brain Bank staff (R. Ravid, W. Kamphorst, F. C. Stam, P. van der Valk, B. Fisser, A. Holtrop, J. Wouda, M. Kooreman, M. Kahlmann and A. C. E. de Nijs) for the provision and processing of hypothalamic material and clinical information; patients, their families and other clinicians for tissue donation; J. J. van Heerikhuize and C. W. Pool for their technical assistance; M. A. Hofman for statistical comments; A. Holland, L. van der Ploeg, and X. Guan for helpful comments.
Footnotes
This work received financial support from Merck Research Laboratories (Rahway, NJ), the UK Medical Research Council, the Royal Society of London, and the Royal College of Physicians (London), PAD 9607.
1 Present address: Department of Endocrinology, St. Bartholomew’s Hospital, West Smithfield, London EC1A 7BE, United Kingdom. 
Abbreviations: AGRP, Agouti-related protein; ANCOVA, analysis of covariance; ARC, arcuate nucleus; DAB, 3,3-diaminobenzidine; HP, hypothalamic-pituitary; ICC, immunocytochemistry; INF, infundibular nucleus; ISH, in situ hybridization; LHA, lateral hypothalamic area; MC4R, melanocortin-4 receptor; ME, median eminence; PMD, postmortem delay; pNPY, porcine NPY; PVN, paraventricular nucleus; PWS, Prader-Willi syndrome; RT, room temperature.
Received July 23, 2001.
Accepted October 30, 2001.
References
-melanocyte-stimulating hormone. Diabetes 49:177–182[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  A. P. Goldstone, A. J. Holland, B. P. Hauffa, A. C. Hokken-Koelega, M. Tauber, and on behalf of speakers contributors at the Second E Recommendations for the Diagnosis and Management of Prader-Willi Syndrome J. Clin. Endocrinol. Metab., November 1, 2008; 93(11): 4183 - 4197. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Fronczek, S. Overeem, S. Y. Y. Lee, Ingrid. M. Hegeman, J. van Pelt, Sjoerd. G. van Duinen, G. J. Lammers, and D. F. Swaab Hypocretin (orexin) loss in Parkinson's disease Brain, June 1, 2007; 130(6): 1577 - 1585. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Fronczek, G. J. Lammers, R. Balesar, U. A. Unmehopa, and D. F. Swaab The Number of Hypothalamic Hypocretin (Orexin) Neurons Is Not Affected in Prader-Willi Syndrome J. Clin. Endocrinol. Metab., September 1, 2005; 90(9): 5466 - 5470. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. Goldstone, M. Patterson, N. Kalingag, M. A. Ghatei, A. E. Brynes, S. R. Bloom, A. B. Grossman, and M. Korbonits Fasting and Postprandial Hyperghrelinemia in Prader-Willi Syndrome Is Partially Explained by Hypoinsulinemia, and Is Not Due to Peptide YY3-36 Deficiency or Seen in Hypothalamic Obesity Due to Craniopharyngioma J. Clin. Endocrinol. Metab., May 1, 2005; 90(5): 2681 - 2690. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
U. A. Unmehopa, J. J. van Heerikhuize, W. Spijkstra, J. W. Woods, A. D. Howard, E. Zycband, S. D. Feighner, D. L. Hreniuk, O. C. Palyha, X.-M. Guan, et al. Increased Melanin Concentrating Hormone Receptor Type I in the Human Hypothalamic Infundibular Nucleus in Cachexia J. Clin. Endocrinol. Metab., April 1, 2005; 90(4): 2412 - 2419. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Alkemade, C. L. Vuijst, U. A. Unmehopa, O. Bakker, B. Vennstrom, W. M. Wiersinga, D. F. Swaab, and E. Fliers Thyroid Hormone Receptor Expression in the Human Hypothalamus and Anterior Pituitary J. Clin. Endocrinol. Metab., February 1, 2005; 90(2): 904 - 912. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Gavrila, J. L. Chan, L. C. Miller, K. Heist, N. Yiannakouris, and C. S. Mantzoros Circulating Melanin-Concentrating Hormone, Agouti-Related Protein, and {alpha}-Melanocyte-Stimulating Hormone Levels in Relation to Body Composition: Alterations in Response to Food Deprivation and Recombinant Human Leptin Administration J. Clin. Endocrinol. Metab., February 1, 2005; 90(2): 1047 - 1054. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. M. Escobar, S. J. Krajewski, T. Sandoval-Guzman, M. L. Voytko, and N. E. Rance Neuropeptide Y Gene Expression Is Increased in the Hypothalamus of Older Women J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2338 - 2343. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. P. Goldstone, E. L. Thomas, A. E. Brynes, G. Castroman, R. Edwards, M. A. Ghatei, G. Frost, A. J. Holland, A. B. Grossman, M. Korbonits, et al. Elevated Fasting Plasma Ghrelin in Prader-Willi Syndrome Adults Is Not Solely Explained by Their Reduced Visceral Adiposity and Insulin Resistance J. Clin. Endocrinol. Metab., April 1, 2004; 89(4): 1718 - 1726. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Hestiantoro and D. F. Swaab Changes in Estrogen Receptor-{alpha} and -{beta} in the Infundibular Nucleus of the Human Hypothalamus Are Related to the Occurrence of Alzheimer's Disease Neuropathology J. Clin. Endocrinol. Metab., April 1, 2004; 89(4): 1912 - 1925. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Horsthemke, H. Nazlican, J. Husing, L. Klein-Hitpass, U. Claussen, S. Michel, C. Lich, G. Gillessen-Kaesbach, and K. Buiting Somatic mosaicism for maternal uniparental disomy 15 in a girl with Prader-Willi syndrome: confirmation by cell cloning and identification of candidate downstream genes Hum. Mol. Genet., October 16, 2003; 12(20): 2723 - 2732. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
hypogonadal females (postmenopausal or PWS), 
intact males, 
hypogonadal males (castrated controls or PWS). Line represents median for each group. Note that in obese subjects (PWS and non-PWS) compared with controls, there is a significant reduction in NPY ICC staining and mRNA, but no difference in AGRP ICC staining, when adjusting for significant covariates. P values: a, Mann-Whitney test; b, adjusting for differences in premorbid illness duration by ANCOVA; c, adjusting for differences in premorbid illness duration and storage time by ANCOVA.
