Regulatory, Integrative and Comparative Physiology

Obesity-prone rats have normal blood-brain barrier transport but defective central leptin signaling before obesity onset

Barry E. Levin, Ambrose A. Dunn-Meynell, William A. Banks


Rats selectively bred to develop diet-induced obesity (DIO) were compared with those bred to be diet resistant (DR) on a 31% fat high-energy diet with regard to their central leptin signaling and blood-brain barrier (BBB) transport. Peripheral leptin injection (15 mg/kg ip) into lean 4- to 5-wk-old rats produced 54% less anorexia in DIO than DR rats. DIO rats also had 21, 63, and 64% less leptin-induced immunoreactive phosphorylated signal transducer and activator of transcription 3 (pSTAT3) expression in the hypothalamic arcuate, ventromedial, and dorsomedial nuclei, respectively. However, hindbrain leptin-induced nucleus tractus solitarius pSTAT3 and generalized sympathetic (24-h urine norepinephrine) activation were comparable. Reduced central leptin signaling was not due to defective BBB transport since transport did not differ between lean 4- to 5-wk-old DIO and DR rats. Conversely, DIO leptin BBB transport was reduced when they became obese at 23 wk of age on low-fat chow or after 6 wk on high-energy diet. In addition, leptin receptor mRNA expression was 23% lower in the arcuate nuclei of 4- to 5-wk-old DIO compared with DR rats. Thus a preexisting reduction in hypothalamic but not brain stem leptin signaling might contribute to the development of DIO when dietary fat and caloric density are increased. Defects in leptin transport appear to be an acquired defect associated with the development of obesity and possibly age.

  • signal transducer and activator of transcription 3
  • leptin receptor
  • sympathetic nervous system
  • leptin resistance

obesity in humans results from a combination of genetic and environmental influences (12). One possible cause of obesity in humans might be a resistance to the centrally mediated anorectic and catabolic effects of leptin. Such leptin resistance might have a dual source. There could be an impairment in the function of the saturable leptin transporter (5, 59) located at the blood-brain barrier (BBB). Additionally, there might be an impaired ability of leptin to induce responses at its receptor, a homologue of the IL-6 family of cytokine receptors (58). Inbred obese rodent strains with such specific genetic defects have been useful in understanding the leptin system. However, subsequent questions relating to how leptin resistance develops may be more profitably addressed in strains that, like humans (12), demonstrate the interaction of environmental influences with a polygenic propensity to develop obesity (33, 34). The Sprague-Dawley-derived diet-induced obese (DIO) and dietresistant (DR) rat strains offer such a unique model.

Outbred Sprague-Dawley rats have two populations with regard to the development of DIO. Approximately half the animals initially increase their caloric intake and develop DIO when the energy and fat density of the diet are modestly increased (29, 30, 32, 39, 41). The remaining rats are DR in that they gain no more weight or carcass fat than low-fat, chow-fed controls (29, 30, 32, 39, 41). These weight-gain patterns appear to be inherited as polygenic traits since selective breeding of outbred DIO and DR rats results in two substrains, which breed 100% true to their weight-gain phenotypes after only three to five breeding cycles (33, 36). Furthermore, the DIO phenotype persists in the offspring of matings performed by repeated back crosses between DIO and an obesity-resistant rat strain (34). Before the onset of obesity, outbred DIO-prone rats manifest a number of abnormalities of nervous system function, which might predispose them to develop obesity when offered a 31% fat, high-energy (HE) diet (40). Most prominent among these is a reduced sensitivity to the anorectic effects of centrally administered leptin (32).

The present DIO and DR substrains still breed true to their weight-gain phenotypes after more than 20 generations of selective breeding (33, 53). Although the DIO rats originally became obese, even on low-fat chow (33), the present generations of selectively bred DIO rats become heavier but only slightly more obese when fed chow from weaning (52). As with outbred DIO rats, selectively bred DIO rats now become significantly obese only when they are fed HE diet (34, 35, 53). Most of the accrued body fat on HE diet can be accounted for by increased caloric intake, which is not inhibited despite an early and marked increase in plasma leptin levels (34, 35, 53). This apparently reduced ability to sense and regulate rising leptin levels may well be a major predisposing cause of their obesity (32, 53).

Once transported across the BBB, leptin acts on the long (signaling) form of the leptin receptor (OB-Rb) (7, 58) to produce anorexia (14) and increase sympathetic activity (20). This response is mediated through activation of the Janus kinase 2-signal transducer and activator of transcription-3 (STAT3) pathway (58). Leptin's occupation of OB-Rb leads to phosphorylation of STAT3 (pSTAT3) and subsequent gene transcription (61). Although gene transcription may not be the proximate cause of leptin-induced anorexia, the expression of pSTAT3 serves as a useful surrogate for the activation of OB-Rb (47, 61). In the present studies, we found that preobese, selectively bred DIO rats had reduced arcuate nucleus (Arc) OB-Rb expression in association with reduced leptin-induced anorexia and hypothalamic pSTAT3 expression at a time when leptin BBB transport was normal. Impaired leptin transport developed only after DIO rats became obese and/or aged.


Animals and diet. Animal usage was in compliance with the Animal Care Committee of the East Orange and St. Louis Veterans Affairs Medical Centers (East Orange, NJ, St. Louis, MO) and the guidelines of the American Physiological Society (1). Male and female rats from of our in-house colony of rats selectively bred to be DIO or DR were used at a variety of ages. These substrains were derived from the outbred Sprague-Dawley strain (Charles River Laboratories) and were bred for high (DIO) and low (DR) weight gain on a HE diet (33). In all cases, litters were adjusted to 10 pups per dam on the day of delivery, and rats were weaned at 3 wk of age. Rats were fed Purina rat chow (no. 5001) and water ad libitum from weaning and were housed at 22-23°C on a 12:12-h light-dark schedule with lights out at 1800 (except where specified). Purina rat chow contains 3.30 kcal/g with 23.4% as protein, 4.5% as fat, and 72.1% as carbohydrate, which is primarily in the form of complex polysaccharide (38). In some experiments, some rats were fed HE diet, which is composed of 8% corn oil, 44% sweetened condensed milk, and 48% Purina rat chow (Research Diets no. C11024F, New Brunswick, NJ). HE diet contains 4.47 kcal/g with 21% of the metabolizable energy content as protein, 31% as fat, and 48% as carbohydrate, 50% of which is sucrose (38).

Experiment I: leptin inhibition of food intake, stimulation of sympathetic activity and pSTAT3 expression. During the first week of lactation, selectively bred DR and DIO dams and pups were placed on a reversed light-dark schedule (lights on 2400; lights off 1200). Pups were fed chow from weaning. At 4 wk of age, nine DIO and eight DR male rats were placed in metabolic cages for collection of 24-h urine norepinephrine (NE) levels and to monitor 4- and 24-h baseline intakes after intraperitoneal injection of 0.5 ml of saline, beginning at lights off (1200). The next day, the rats were injected with murine leptin (10 mg/kg ip in 0.5 ml of saline; Calbiochem), 24-h urines were collected, and intake was monitored at 4 and 24 h. After an additional 5 days, the procedure was repeated again, except that rats were injected with leptin (15 mg/kg ip) on the second day. After an additional 5 days of rest, five DIO and four DR rats were intraperitoneally injected a second time with 15 mg/kg leptin, and four DIO and four DR rats were injected a second time with saline for assessment of brain pSTAT3 expression. After 45 min, they were anesthetized with chloropent and then rapidly perfused with ice-cold saline for 1 min, followed by ice-cold 4% paraformaldehyde in 0.1 M PBS, pH 7.4, at ∼70 ml/min for 6-7 min. The brains were removed from the skull and postfixed in the 0.1 M PBS/paraformaldehyde solution at 4°C for 24 h. The brains were then sunk in 30% sucrose in PBS overnight with agitation at 4°C and then frozen in 2 methylbutane cooled to -50°C with dry ice. Brains were then stored at -70°C until processed for pSTAT3 immunocytochemistry. Urine NE levels were assayed by high-performance liquid chromotography with electrochemical detection (29).

Immunocytochemistry for pSTAT3 was carried out by modifications of the methods of Peterson et. al. (51). Serial 35-μm brain sections were cut through the hypothalamic Arc, ventromedial (VMN) and dorsomedial (DMN) nuclei (centered on the compact portion of the DMN), and the caudal nucleus tractus solitarius (NTS; at the level of the area postrema). Sections were floated in K+-PBS (KPBS) buffer with gentle agitation and washed three times. They were then incubated for 20 min in freshly prepared buffer containing 1% NaOH and 1% H2O2 in KPBS, and then washed twice for 5 min each in KPBS. Next, sections were placed in 0.3% glycine in KPBS for 10 min followed by two washes of 5 min each with KPBS. Sections were then placed in freshly prepared KPBS containing 0.15% SDS for 10 min and then washed three times for 10 min each in KPBS. Blocking was done with 4% horse serum in KPBS containing 0.4% Triton X for 2 h, and sections were then incubated overnight at 4°C with 1:1,000 dilution of rabbit anti-pSTAT3 antibody (Cell Signaling Technologies) in KPBS containing 4% horse serum and 0.4% Triton X. The next day, the solutions were warmed to room temperature for 1 h, rinsed six times for 10 min each in KPBS containing 1% horse serum and 0.02% Triton X, and incubated for 1 h with 1:250 biotinylated goat anti-rabbit antibody in KPBS containing 1% horse serum. They were then washed three times for 10 min each in KPBS containing 1% horse serum and incubated for 1 h with ABC reagent (Vector Laboratories), which had been prepared 30 min before in KPBS containing 1% horse serum. After six washes in KPBS, the sections were reacted with diaminobenzidine for ∼20 min, washed three times, mounted on slides, and dehydrated by standard procedures. The number of cells expressing pSTAT3 for saline- and leptin-treated rats was counted in three sections per rat in the Arc, VMN, and DMN portion of the hypothalamus and the caudal NTS by an observer blinded to the experimental groups by using a Bioquant image analysis system (Nashville, TN). The average number of cells counted in the three sections per area in each rat was taken for statistical comparisons.

Experiment II: effects of genotype, diet, and age on BBB leptin transport. Pregnant female DIO and DR rats were transported from the parent colony in E. Orange (B. Levin) to the vivarium in St. Louis (W. A. Banks) where they were allowed to deliver and nurse their pups. These rats were kept at 22-23°C on a 12:12-h light-dark cycle (lights on at 0800) and were fed chow and water ad libitum from weaning. Leptin transport across the BBB was assessed in chow-fed male and female DR and DIO rats at 4 wk (n = 12 per group) and 11 wk of age (n = 8 per group). Transporter activity was also assessed in chow-fed male DR and DIO rats, which were either continued on chow or switched to HE diet for 6 wk beginning at 11 wk of age (n = 8 per group). Transport activity was assessed in final groups of 12 DIO and 14 DR male rats, which were fed chow from weaning to 23 wk of age. Thus transport was assessed in chow-fed DIO and DR females at 4 and 11 wk of age and in chow-fed DIO and DR males at 4, 11, 17, and 23 wk of age, as well as in males fed HE diet from 11 to 17 wk of age.

Assessment of leptin transport. Leptin (Amgen, Thousand Oaks, CA) was radioactively labeled by the lactoperoxidase method with 131I. Radiolabeled leptin (I-Lep) was separated from free iodine on a Sephadex G-10 column. The I-Lep had a specific activity of ∼100-125 Ci/g and was stored at 4°C until use. The unidirectional influx rate of leptin from blood to brain was measured by multiple-time regression analysis (10, 50). Rats were anesthetized with pentobarbital, and the left jugular vein and right carotid artery were exposed. Rats were given an injection into the jugular vein of 0.2 ml of lactated Ringer solution containing 1% bovine serum albumin and about 5 × 105 counts/million (cpm) of I-Lep. Blood from the carotid artery was obtained between 1 and 10 min after intravenous injection (typically at 1-min intervals), centrifuged for 10 min at 5,000 g at 4°C, and the radioactivity level in the serum determined. Immediately after collection of each arterial blood sample, the rat was decapitated and the level of the radioactivity in the whole brain (with the pituitary and the pineal gland removed) determined. The brain-to-serum ratios of I-Lep in units of microliters per gram [(cpm/g of brain)/(cpm/μl serum)] were plotted against exposure time (Expt; in minutes) Math1 where Cp is serum concentration, (τ)dτ is an integration time constant, and Cpt is the level of radioactivity in serum at time t (10, 50). The slope of the linear portion of the relation between the brain-to-serum ratios and Expt was calculated by the least squares method (see Statistics below) and measures the unidirectional influx rate from blood to brain and is reported with its standard deviation of the residuals.

Leptin assay. Serum obtained from 10-wk-old DR and DIO male rats was frozen at -70°C until assayed. Leptin was measured with the mouse ELISA (Linco, St. Charles, MO). This assay is 50% cross reactive with rat leptin, and so the reported values were multiplied by two to give “corrected rat” values.

Experiment III: OB-Rb expression. Selectively bred male DIO and DR (n = 8 per group) rats were placed on a reversed 12-h light-dark schedule during the first week of suckling, with lights off at 1200 (see Experiment I). They were fed Purina rat chow (no. 5001) ad libitum from weaning. At 5 wk of age, they were decapitated without fasting during the 2 h before lights on (1000-1200). Their brains were quickly removed and frozen on dry ice. At this time, trunk blood was taken for leptin assay by radioimmunoassay using antibodies directed against rat leptin (Linco). Brains were stored at -70°C until assay. A 300-μm section was cut from frozen brains on a cryostat at -12°C such that the compact portion of the DMN was in the center of the slice. Cut sections were placed in RNA Later (Ambion). After at least 1 h, sections were placed under a dissecting microscope, and the Arc, VMN, and DMN were micropunched bilaterally by minor modifications (30) of the method of Palkovits (49). The punched samples were kept in RNA for <2 wk before being reversed transcribed to cDNA.

Micropunch samples were subsequently assayed by real-time (quantitative) PCR. They were first homogenized ultrasonically in a solution containing guanidinium thiocyanate and nucleotides. Then mRNA was purified by using silica columns (Ambion RNAqueous kit). After removal of genomic DNA with DNase, mRNA was reversed transcribed with random hexamer priming using Superscript 3 (Invitrogen). Samples were then treated with RNaseH (Ambion), and the resulting purified cDNA was aliquoted and frozen. Primer sets for cyclophilin and OB-Rb mRNA were designed by reference to published sequences, and their specificity was verified by using Genbank and by comparing the sequenced PCR product for both cyclophilin and OB-Rb to these references. For each mRNA species, a pair of conventional primers was used in combination with a sequence-specific 6-carboxyfluorescein (FAM)-labeled probe to allow real-time PCR quantitation by using an Applied Biosystems 7700 Sequence detector set for 40 PCR cycles. The primers for cyclophilin were Genbank (NM 017101) forward (253-272) AATGGCACTGGTGGCAAGTC; reverse (330-310) GCCAGGACCTGTATGCTTCAG; FAM-labeled probe (275-290) TCTACGGAGAGAAATT. The primers for OB-Rb were Genbank (AF287268) forward (2913-2937) AACCTGTGAGGATGAGTGTCAGAGT; reverse (3005-2984) CCTTGCTCTTCATCAGTTTCCA; FAM-labeled probe ATGCAACGCTGGTCAG (2954-2969). Reference standards for each hypothalamic nucleus were created for cyclophilin and OB-Rb from aliquots of samples from all samples of those areas used. These pooled samples were used to generate standard curves from which the respective quantitative data were read. Data were then expressed as the ratio of OB-Rb to cyclophilin.

Statistics. Food intake and urine NE levels after saline or leptin administration were assessed by one-way ANOVA for repeated measures with post hoc Bonferonni comparisons when significant differences were found. The number of cells expressing pSTAT3 and OB-Rb expression in each brain area were compared by two-way ANOVA (phenotype × treatment) with post hoc Bonferonni comparisons where appropriate.

For leptin transport, regression lines were calculated by the least squares method with the Prism 3.0 program (GraphPad, San Diego, CA) and the slope with its standard deviation of the mean reported. Regression lines were compared for statistical differences with the Prism 3.0 program, which first determines whether there are differences between slopes and, if not, whether there are differences between intercepts. When more than two slopes were compared, ANOVA was followed by Newman-Keuls range test with the standard deviation of the mean taken as the standard error term and, because two means (the slope and the intercept) were calculated from the data, n-1 was used as the value for n.


Experiment I: leptin inhibition of food intake, stimulation of sympathetic activity, and pSTAT3 expression. Male 4- to 5-wk- old DIO and DR rats were injected with saline to determine baseline intake and then with leptin (10 and 15 mg/kg ip) to compare the effects on food intake, sympathetic activity (urine NE levels) and activation of the leptin signaling pathway (pSTAT3 expression). At 4 wk of age, DIO rats in this study weighed 85 ± 2 g and DR rats weighed 95 ± 3 g (P = 0.01). At 5 wk of age, DIO rats weighed 145 ± 4 g and DR rats weighed 152 ± 5 g (P = 0.1). Neither DIO nor DR rats had significant reductions in food intake at either 4 or 24 h after 10 mg/kg ip leptin compared with saline injections (data not shown). After 15 mg/kg ip leptin, 4-h intake was inhibited by 13% in DIO and 24% in DR rats compared with their salinetreated baselines (Fig. 1; P = 0.016). By 24 h, leptin inhibition of intake did not differ between DIO and DR rats (Fig. 1). After saline injection, there was little constitutive expression of pSTAT3 immunoreactivity in either hypothalamic or NTS neurons (Fig. 2). On the other hand, there was significant expression of pSTAT3 immunoreactivity in neurons within these areas after leptin injections. In the Arc, expression was most prominent in the medial portion (Fig. 2A). Overall, the number of Arc cells expressing pSTAT3 in DIO rats was only 79% of DR rats (Fig. 2B; P = 0.039). In the VMN, leptin induced a gradient of decreasing pSTAT3-expressing neurons from the dorsomedial to central to ventrolateral portions (Fig. 2A). In each portion of the VMN, DIO rats had only 36-38% as many pSTAT3-expressing neurons (Fig. 2B; P = 0.013 for VMN overall). In the DMN, most leptin-induced pSTAT3 expression occurred in the ventral portion (Fig. 2A), and DIO rats had only 38% as many pSTAT3 immunoreactive DMN cells as DR rats (Fig. 2B). Finally, there was moderate expression of leptin-induced pSTAT3 in the NTS. However, in contrast to hypothalamic leptin-induced pSTAT3 induction, there was no significant difference between DIO and DR rats in pSTAT3 induction in the NTS (Fig. 2B).

Fig. 1.

Inhibition of food intake (as a percent of each rat's own saline-injected baseline intake) in 4- to 5-wk-old diet-resistant (DR) and diet-induced obesity (DIO) rats (n = 8 DR and 9 DIO/group) at 4 and 24 h after leptin administration (15 mg/kg ip). Data are means ± SE. **P = 0.01 when leptin-induced inhibition of intake at 4 h was compared between DR and DIO rats.

Fig. 2.

Leptin-induced expression of phosphorylated signal transducer and activator of transcription 3 (pSTAT3) in neurons of the hypothalamic arcuate (Arc), ventromedial (VMN), and dorsomedial nuclei (DMN), and the nucleus tractus solitarius (NTS). Chow-fed, 4- to 5-wk-old male DR (n = 8) and DIO rats (n = 9) were intraperitoneally injected with saline or 15 mg/kg leptin and killed 45 min later. A: coronal section through the ventrobasalar hypothalamus immunostained for pSTAT3. B: cells expressing immunoreactive pSTAT3 were counted in each brain area and expressed as means ± SE. III, third cerebral ventricle. *P ≤ 0.05 when the number of leptin-induced cells expressing pSTAT3 were compared between DR and DIO rats for each brain area.

At baseline, 4-wk-old DIO rats had 35% higher 24-h urine NE levels than DR rats (Fig. 3A; P = 0.05). As with its effect on food intake, injection of 10 mg/kg ip leptin did not alter 24-h urine NE levels in either genotype, although DIO rats no longer had higher levels than DR rats. After 15 mg/kg leptin, 24-h urine NE levels rose significantly above baseline in both DIO (127%; P = 0.006) and DR rats (110%; P = 0.05), and DIO levels were again higher (46%; P = 0.01) than DR levels. When the leptin-induced percent changes from baseline for 24-h urine NE and 4-h intake inhibition were plotted against each other (Fig. 3B), there was a significant correlation between these parameters in DR (r = 0.90; P = 0.01) but not in DIO rats. Curiously, half of the DR rats actually had a reduced 24-h urine NE response to leptin administration compared with saline-injected controls (Fig. 3B). There were no correlations between leptin-induced inhibition of 24-h intake and 24-h urine NE levels for either genotype.

Fig. 3.

Effects of leptin on 24-h urine norepinephrine (NE) levels and food intake were assessed in 4- to 5-wk-old chow-fed male DR (n = 8; ○) and DIO (n = 9; •) rats. A: 24-h urine NE levels (ng) at baseline (0) and after intraperitoneal administration of 10 and 15 mg/kg leptin. Data are means ± SE. *P ≤ 0.05 when leptin levels at a given dose were compared between DR and DIO rats. +P ≤ 0.05 when leptin levels for a given genotype after leptin were compared with their own baseline levels. B: leptin-induced (15 mg/kg ip) percent change in 24-h urine NE levels from baseline [Δ24-h urine NE (%Baseline)] compared with percent leptin-induced inhibition of 4-h food intake [intake inhibition (%Baseline)] in the same rat. There was a significant correlation for DR but not DIO rats.

Experiment II: Effects of genotype, gender, diet, and age on BBB leptin transport. The impaired anorectic and hypothalamic pSTAT3 expression in 4- to 5-wk-old DIO rats might have been due to reduced BBB transport. To address this issue, transport was compared between DIO and DR rats of differing ages, sexes, and degrees of obesity. At 4 wk of age, there were no differences in leptin transport across the BBB among lean, chow-fed DR or DIO males or females. All four groups had similar transport rates (Table 1; Fig. 4). Body weights differed among the four groups, and the weights of each group were different from those of the other three [Table 1: F(3,43) = 24.5; P < 0.001]. At this age, as in the 4-wk-old rats in Experiment I, DIO rats weighed 10% less than their DR counterparts. At 11 wk of age, there were further differences in body weight among the groups [Fig. 5; F(3,28) = 238; P < 0.001]. DIO males weighed ∼29% more than DR males (P < 0.001), and DIO females weighed ∼26% more than DR females (P < 0.001). The mean serum leptin level in seven DIO males (2.54 ± 0.31 ng/ml) was twice as high as in seven DR males (1.34 ± 0.39 ng/ml; P < 0.05). However, there were no significant differences in transport rate among the four groups (Fig. 5).

View this table:
Table 1.

Transport rates across the BBB (Ki) of I-Lep and body weights in 4-wk-old DR and DIO rats

Fig. 4.

Relation between brain-to-serum 131I-leptin (I-Lep) ratios and exposure time (Expt) for 4-wk-old DR and DIO males (n = 12 per genotype). The relation was statistically significant, demonstrating uptake of I-Lep from blood by brain. The slopes measure the unidirectional influx rate of entry (Ki) in units of μl/g-min (see Table 1 for values). There was no significant difference between the two groups.

Fig. 5.

Leptin transport rates (Ki; left) and body weights (right) in 11-wk-old male and female DR and DIO rats (n = 8/group). Data are means ± SE. There were no differences among Ki. Body weights of groups with differing superscripts (right) differed from each other at P ≤ 0.001 by post hoc t-test after significant intergroup differences were found by ANOVA.

When 11-wk-old male rats were either switched to HE diet or continued on chow for 6 wk, there were further significant intergroup differences in body weight [Fig. 6; F(3,28) = 2.5; P < 0.001]. DIO rats fed HE diet weighed 24% more than chow-fed DIO rats, 43% more than DR rats fed HE diet, and 54% more than chow-fed DR rats (P < 0.001 for all). The 7% difference in body weight between chow-fed DR and DR rats fed HE diet was not statistically different. The chow-fed DR and DIO rats and DR rats fed HE diet all transported I-Lep into the brain at similar rates of transport that did not differ from one another. However, only small amounts of labeled leptin were recovered from the brains of DIO rats fed HE diet and the ratios of brain to serum leptin were extremely variable among the animals. Thus the relationship between brain-to-serum ratios and time was not statistically significant for this group, i.e., no statistically significant transport had occurred in DIO rats fed HE diet. Nevertheless, a regression line was calculated for these animals to allow for intergroup statistical comparison. This comparison showed significant differences among the slopes of the regression lines for the four groups [F(3,23) = 4.12; P < 0.05]. Post hoc analysis showed that the slope of the line for DIO rats on HE diet was different from that of DR rats on HE diet (P < 0.05) and from that of chow-fed DIO rats (P < 0.05). The half-time disappearance rate of I-Lep from blood did not differ among these four groups, demonstrating that this saturable clearance mechanism was not affected.

Fig. 6.

Leptin transport rates (Ki; left) and body weights (right) in male DR and DIO rats fed chow or high-energy diet for 6 wk from 11 wk of age (n = 8/group). Data are means ± SE. Both Ki and body weight groups with differing superscripts differed from each other at P ≤ 0.05 by post hoc t-test after significant intergroup differences were found by ANOVA.

Final groups of DR and DIO rats were fed chow from weaning to 23 wk of age and were then evaluated for leptin transport (Fig. 7). Diet had no effect on transport rates in DR rats at this age, but the DIO group had lost the ability to transport leptin and was, therefore, significantly lower than time-matched chow-fed DR rats (P < 0.05). Figure 7 summarizes the changes in leptin transport rates for chow-fed DR and DIO male rats at 4, 11, 17, and 23 wk of age.

Fig. 7.

Leptin transport rates (Ki) for I-Lep in chow-fed DR and DIO male rats at 4 (n = 12/genotype), 11 (n = 8/genotype), 17 (n = 8/genotype), and 23 wk (n = 14 DR and 12 DIO) of age. Data are means ± SE. *The last transport rate value for DIO rats did not differ significantly from zero but did differ (P < 0.05) from the last value for DR rats.

Experiment III: DIO and DR OB-Rb mRNA expression. At 5 wk of age, when peripheral leptin levels were statistically comparable in chow-fed DIO (2.3 ± 0.2 ng/ml) and DR rats (1.8 ± 0.2 ng/ml), DIO Arc OB-Rb mRNA expression was 77% of that in DR rats (Table 2). Despite reduced leptininduced pSTAT3 expression in both the VMN and DMN of DIO rats, there were no differences in OB-Rb expression between DIO and DR rats in those hypothalamic nuclei (Table 2).

View this table:
Table 2.

Real-time PCR ratio of hypothalamic leptin receptor to cyclophilin mRNA


The rat model of DIO is a good surrogate for human obesity. As with human obesity (11), the DIO trait appears to be inherited in a polygenic fashion (33, 34), and DIO rats develop many features of the metabolic syndrome (19, 41, 42, 60, 63). In this study, we show that selectively bred DIO rats have a defect in central leptin signaling, which antedates the onset of obesity. At 4-5 wk of age, selectively bred DIO rats had reduced Arc OB-Rb expression in association with an attenuated anorectic response and activation of hypothalamic pSTAT3 after peripherally administered leptin. This defect was not attributable to deficient BBB transport of leptin since transport did not differ between DIO and DR rats at this age. DIO rats acquired defective leptin transport only as they aged and/or became obese. Their reduced OB-Rb mRNA and leptininduced pSTAT3 expression in the Arc, a nucleus known for its critical role in energy homeostasis (40, 57), suggests that impaired Arc leptin signaling might underlie their impaired anorectic response to leptin. On the other hand, the lack of impaired leptin-induced activation of pSTAT3 in the NTS, increased leptin-induced urine NE levels, and their known propensity to develop obesity-related hypertension (19) suggest that leptin signaling in this critical hindbrain autonomic nucleus (44) might not be altered in DIO rats.

The reduced Arc OB-Rb mRNA expression in DIO rats exists before the development of obesity or exposure to HE diet. Reduced mRNA expression does not guarantee that receptor protein expression or binding will be decreased. However, their attenuated leptin-induced Arc pSTAT3 expression suggests that DIO rats had a parallel or even greater reduction in functional Arc leptin signaling (21, 23, 27, 28, 47). It is unclear why there were no similar reductions in DIO VMN and DMN OB-Rb expression by quantitative PCR since their leptin-induced pSTAT3 expression was reduced. Also, our laboratory previously found (35) reduced OB-Rb mRNA expression in both the VMN and DMN of DIO rats by in situ hybridization. However, those reductions were small (∼10%) and might not have been detected here by the combination of micropunch and quantitative PCR. Although there is no conclusive evidence that leptin might affect food intake by a direct effect on the VMN, there is evidence that the DMN might play a role comparable to that ascribed to the Arc (8) and that this might be mediated, in part, by leptin (9). However, the selective decrease of both OB-Rb expression and leptin-induced STAT3 activation in the Arc of DIO rats, along with the well-established role of the Arc in energy homeostasis (40, 57), suggests that defective Arc leptin signaling might be a critical determinant of the development of obesity in DIO rats fed HE diet. One cautionary note is that the leptin signaling pathways involved in producing anorexia might not be dependent on STAT3 activation. But assessment of immunoreactive pSTAT3 activation is currently the best surrogate for shortterm postreceptor signaling available at the cellular level for leptin signaling.

Although the present studies suggest that 4- to 5-wk-old DIO rats have reduced leptin sensitivity compared with DR rats, they actually had lighter body weights at 4 wk of age. Because rats at this age have only 5-6% body fat (52), it is likely that their differences in body weight were due to relatively less lean body mass in DIO rats. Although there is no obvious explanation for this apparent reversal of expected relative body weights, it is unlikely that it is due to the lighting schedule used. DIO rats raised under both reversed (Experiment I) and conventional lighting schedules (Experiment II) weighed less than DR rats at 4 wk of age. Although it is unclear why there were differences in body weight at 4 wk of age, these differences disappeared at 5 wk of age.

The reduced responsiveness of 4- to 5-wk-old DIO rats to peripherally injected leptin cannot be explained by reduced access of leptin to central sites. DIO rats at this age had comparable BBB leptin transport to DR rats. Also, outbred DIO rats, from which the selectively bred DIO rats in the present studies are derived (33), have reduced sensitivity to the anorectic effects of centrally administered leptin (32). Moreover, no BBB transport would be required for leptin to reach some neurons in the Arc or NTS since both reside close to brain areas with no BBB (13, 22). Because a major portion of leptin's anorectic effect is mediated through Arc neurons (48, 57), peripherally administered leptin might reach this target, engage the OB-Rb on Arc neurons and activate STAT3, as well as producing anorexia without having to cross the BBB. Similarly, the fact that transport should not be required for STAT3 activation in the NTS suggests that differences between DIO and DR rats in the anorectic effects of peripherally administered leptin are not mediated through the NTS because there was no difference in leptin-induced pSTAT3 activation in this nucleus.

The dissociation between the anorectic effect of leptin and 24-h urine NE levels in DIO rats is in agreement with other studies showing a dissociation between the anorectic and sympathoexcitatory effects of leptin (17, 46, 52). In fact, this dissociation might explain why DIO rats have constitutively elevated urine NE levels (29, 30, 33) and develop obesityassociated hypertension (19) but do not downregulate their food intake or Arc neuropeptide Y expression until plasma leptin levels appreciably exceed those of DR levels (32, 37). Aside from its generally catabolic effect on energy homeostasis, leptin increases both general sympathetic activity and blood pressure (18, 26). Interestingly, the dissociation between leptin-induced anorexia and sympathoexcitation did not hold for DR rats, possibly because urinary NE was used as an index of sympathetic activity. Clearly, sympathetic activity varies by organ system. For example, although DIO-prone rats have constitutively elevated levels of urine NE excretion (29, 33), they actually have reduced cardiac and pancreatic, but not brown adipose, sympathetic activity (30). Similarly, the effect of leptin on sympathetic activity appears to be highly organ specific (25, 45). Another possibility is that altered urine NE levels might be due to an effect of leptin on the renal clearance of NE (16).

Leptin sensitivity appears to decrease as a function of age, increasing dietary fat content and obesity (3, 15, 21, 43, 54-56, 62, 64). In rats, these factors are generally not dissociable, since most rat strains become obese with age and on high-fat diets. Here, selectively bred DIO rats developed defective leptin transport only when fed HE diet for 6 wk from 11 wk of age or when fed chow from weaning to an age of 23 wk. Although we did not measure leptin levels or carcass fat in the 23-wk-old chow-fed rats, our laboratory previously showed (31) that chow-fed DIO rats of comparable age have fourfold higher leptin levels, suggesting that they are fatter than comparable DR rats. In the present study, same-age DR rats never developed defective leptin transport, although there was a tendency toward reduced transport with age in both DR and DIO rats. In light of these results and those of others (3, 21, 24), it seems likely that defective leptin transport in DIO rats is more a function of the development of obesity than of age alone (3, 4, 6). Thus, in support of data derived from inbred mice (4), defective leptin transport in selectively bred DIO rats appears to be an acquired rather than inherited defect (33).

In conclusion, our data support previous studies showing that reductions in central leptin signaling and defects in leptin BBB transport can develop independently and are likely to have different underlying mechanisms. In the DIO rat, reduced central leptin sensitivity appears to be an inherent trait, whereas defective transport is an acquired one. Reduced sensitivity to the anorectic effects of leptin would confer a distinct competitive advantage on feral animals where food availability was limited and only sporadically available. When food was readily available, rats could store more energy as fat before leptin exerted its inhibitory effect on intake and stimulatory effects on thermogenesis. In humans, reduced leptin signaling should confer a similar competitive advantage where energy sources were limited. However, in times of excess food availability, it would promote the development and maintenance of obesity.


This work was funded by the Research Service of the Veterans Administration (to B. E. Levin, A. A. Dunn-Meynell, and W. A. Banks) and National Institutes of Health Grants RO1 DK-30066 (to B. E. Levin), R01 NS-41863, and R01 AA12743 (to W. A. Banks).


The authors thank Charlie Salter, Eugenia Dziopa, Antoinette Moralishvilli, Odeal Gordon, Michael L. Niehoff, and Kathy M. Wolf for expert technical assistance. We thank Amgen for the kind gift of leptin.


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract