Attenuation of Diabetic Hyperphagia in Neuropeptide Y–Deficient Mice Dana K. Sindelar,1 Paul Mystkowski,1 Donald J. Marsh,2 Richard D. Palmiter,2 and Michael W. Schwartz1 The combined effects of increased hypothalamic signal- these responses when the stimulus is food deprivation. ing by neuropeptide Y (NPY) and decreased signaling by Moreover, fasting is a more potent stimulus to hypotha- melanocortins are hypothesized to stimulate food in- lamic AgRP gene expression than is STZ-diabetes. take when body fat stores are depleted. To investigate Therefore, central nervous system melanocortin signal- NPY’s role in the hyperphagic response to uncontrolled ing appears to be suppressed more effectively by fasting diabetes, streptozotocin (STZ) (200 mg/kg intraperito- than by uncontrolled diabetes, which provides a plausi- neally) or saline vehicle was given to NPY-deficient ble explanation for differences in the feeding response :Npy–/–) and wild-type (Npy؉/؉) mice. In Npy؉/؉ mice, to these two stimuli in mice lacking NPY. Diabetes 51) STZ-induced diabetes increased mean daily food intake 778–783, 2002 to plateau values 50% above baseline intake (؉2.0 ؎ 0.6 g/day; P < 0.05), an effect that was not seen in STZ- treated Npy–/– mice (؉0.8 ؎ 0.1 g/day; NS), despite comparably elevated levels of plasma glucose and com- daptive increases of food intake induced by parably decreased levels of body weight, fat content, depleted body energy stores are important for and plasma leptin. Unlike the impaired feeding response to uncontrolled diabetes, Npy–/– mice exhibit intact survival and appear to involve the coordinate hyperphagic responses to fasting (Erickson et al. [1], A regulation of multiple hypothalamic pathways Nature 381:415–418, 1996). To investigate whether dif- that can influence feeding behavior (2). For example, ferences in hypothalamic melanocortin signaling can hypothalamic neurons that contain neuropeptide Y (NPY), explain this discrepancy, we used in situ hybridization a potent stimulator of food intake, are activated when to compare the effects of STZ-diabetes and fasting on body fat content is reduced by energy restriction (3). pro-opiomelanocortin (POMC) and agouti-related pep- Conversely, the hypothalamic production of melanocort- tide (AgRP) mRNA levels in the hypothalamic arcuate ؉ ؉ ins (peptides with anorexic properties that are cleaved nucleus (ARC) of Npy–/– and Npy / mice. AgRP mRNA levels were increased by both fasting and STZ-diabetes, from pro-opiomelanocortin [POMC]) is reduced in this but the increase in STZ-diabetes was small (50–80%) setting (4). Acute or chonic energy deficits also increase compared with the effect of fasting (ϳ20-fold increase expression of the gene encoding agouti-related peptide of AgRP mRNA). STZ-diabetes also lowered POMC (AgRP), an endogenous antagonist of central nervous -mRNA levels by 65% in the ARC of Npy؉/؉ mice (P < system (CNS) melanocortin receptors (5) that is co-ex –/– 0.05) but by only 11% in Npy mice (NS); fasting pressed with NPY in neurons of the hypothalamic arcuate significantly lowered POMC mRNA levels in both geno- nucleus (ARC) (6). Fasting is an especially potent stimulus types. We conclude that NPY is required for both the to AgRP gene expression in mouse ARC, increasing AgRP increase of food intake and the decrease of hypotha- ϳ lamic POMC gene expression induced by uncontrolled mRNA by 20-fold (6). The combined effects of decreased diabetes. In contrast, NPY is not required for either of melanocortin signaling (due to both increased AgRP and decreased POMC biosynthesis) and increased NPY signal- ing, therefore, comprise an integrated mechanism to me- From the 1Department of Medicine, University of Washington, and Howard diate hyperphagia in response to a sustained energy Hughes Medical Institute, University of Washington, Seattle, Washington; and the 2Department of Biochemistry, University of Washington, Seattle, Wash- deficit. ington. The original description of mice with NPY deficiency Address correspondence and reprint requests to Michael W. Schwartz, –/– Harborview Medical Center, Division of Endocrinology, Box 359757, 325 9th (Npy ) due to targeted NPY gene disruption revealed Ave., Seattle, WA 98104. E-mail: [email protected]. them to have normal levels of daily food intake and body Received for publication 20 June 2001 and accepted in revised form 3 weight and to exhibit normal increases of food intake after December 2001. M.W.S. is on the Scientific Advisory Board of Millennium Pharmaceutical a fast (1). One possible explanation for this finding is that and has received consulting fees for work in this capacity. D.K.S. is currently neural control over feeding under these circumstances is employed as Senior Investigator at Eli Lilly, Inc., a position that he accepted sufficiently redundant that the loss of NPY is compensated after his work on this study was completed. D.J.M. is currently employed as Senior Investigator at Merck Pharmaceuticals, Inc., a position that he ac- for by other responses. Of potential importance in this cepted after his work on this study was completed. context is the marked reduction of melanocortin signaling Current address for D.K.S.: Eli Lilly & Co., Corporate Center, Drop 0545, –/– Indianapolis, IN 46285. Current address for D.J.M.: Merck & Co., Inc., induced by fasting, which may allow Npy mice to RY80T-126, P.O. Box 2000, Rahway, NJ 07065. increase food intake appropriately during refeeding. One AgRP, agouti-related peptide; ARC, hypothalamic arcuate nucleus; AUC, approach to test this hypothesis is to identify and study a area under the curve; CNS, central nervous system; Mcr, melanocortin receptor; MRS, magnetic resonance spectroscopy; NPY, neuropeptide Y; model in which hyperphagia depends more on increased POMC, pro-opiomelanocortin; STZ, streptozotocin. NPY signaling than on reduced melanocortin signaling. In 778 DIABETES, VOL. 51, MARCH 2002 D.K. SINDELAR AND ASSOCIATES such a model, the ability to increase food intake should be the same assay. Hybridization to NPY, POMC, or AgRP mRNA was performed 33 compromised in NPY-deficient mice. with antisense riboprobes transcribed from cDNA templates using [ P]UTP (5,16). The hybridization signal in the arcuate nucleus of each brain slice was Uncontrolled insulin-deficient diabetes induced by the determined from film autoradiograms using the MCID image analysis system ␤-cell toxin streptozotocin (STZ) is an established model (Imaging Research, St. Catherines, ON, Canada) as previously described (16). of sustained hyperphagia in rodents. Like the response to Both the film density and hybridization image area were measured, and the fasting, diabetic hyperphagia appears to involve increased product of these two measures was calculated as an index of mRNA content. Values for each animal reflect the mean of 6–12 measurements per animal. hypothalamic signaling by NPY (7–10) and is also accom- The mean value of each neuropeptide mRNA level obtained for wild-type panied by increased AgRP mRNA and reduced POMC controls (e.g., vehicle-treated Npyϩ/ϩ mice for the STZ-diabetes study or mRNA expression in ARC neurons (11–13). However, the Npyϩ/ϩ mice fed ad libitum for the fasting study) was considered to represent increase of AgRP mRNA levels induced by STZ-induced 100% of the control value for that study, and all individual hybridization values diabetes in mice appears to be much smaller (ϳ2-fold) from each experiment were normalized to this number. ϳ Body composition. Body fat content was determined post mortem in mice (13) than that induced by fasting ( 20-fold) (5). Diabetic from experiment 1 using magnetic resonance spectroscopy (MRS) as previ- hyperphagia may therefore depend to a greater extent on ously reported (17). Mice were placed within a custom-made radiofrequency NPY than on AgRP when compared to the hyperphagic coil used for both transmitting and receiving the resonant proton signal at response to fasting. 200.1 MHz. Body lipid mass was estimated from the area under the curve (AUC) of the lipid peak of the resultant MRS spectra, as previously described Based on this reasoning, we hypothesized that mice (17). For each genotype, the mean AUC for the lipid peak in the nondiabetic lacking NPY would manifest impaired feeding responses group was considered to represent 100% of the control value of relative body to uncontrolled diabetes, despite intact responses to fast- fat content, and individual AUC values for the diabetic groups were expressed ing (which induces a much larger increase of AgRP as a percent of this value. expression). Evidence that the orexigenic response to Plasma assays. Blood samples were collected on heparin and placed ϩ/ϩ immediately on ice in heparinized tubes. Plasma was separated by centrifu- centrally administered AgRP is heightened in Npy mice gation and stored at Ϫ20°C until determination of glucose and leptin concen- (14) provides additional support for this hypothesis. We trations. Plasma glucose was determined by the glucose oxidase method therefore measured food intake and hypothalamic levels (Beckman Instruments, Brea, CA). Plasma leptin levels were determined by of AgRP and POMC mRNA in Npy–/– and Npyϩ/ϩ mice radioimmunoassay (Linco Research, St. Louis, MO). Statistics. All values are reported as group means Ϯ SE. The level of made diabetic with STZ and compared these responses to significance was taken as P Յ 0.05, two-tailed. Data were analyzed for those induced by a 48-h fast in separate groups of each differences from baseline values using one-way ANOVA with repeated mea- genotype. sures, and Fisher’s protected least significance difference test was used for multiple comparisons when significant F ratios were obtained. Interactions between treatment and genotype were analyzed by group mean comparisons RESEARCH DESIGN AND METHODS using two-way ANOVA. Statistical comparisons were made using Statview Animals. Adult mice (C57Bl/6x129Sv/Ev mixed genetic background or pure (Calabasa, CA) software. 129Sv/Ev background) with targeted knockout of the Npy gene (Npy–/–) and wild-type littermate controls (Npyϩ/ϩ) were housed in individual cages in standard vivarium conditions (12 h:12 h light:dark cycle).