Appetite Suppression and Weight Reduction by a Centrally Active Aminosterol

Experiments were performed in accordance with guidelines and regulations of the National Institutes of Health and Institutional Animal Care and Use Committee of the University of Pennsylvania.

Male 12-week-old C57Bl/6J mice (Jackson Laboratories, Bar Harbor, ME) were housed individually in a 12:12 h light-dark cycle (lights on at 0600; temperature 22°C) and allowed normal laboratory diet and water ad libitum. Synthetic MSI-1436 and squalamine (MSI-1256) were provided by Genaera (Plymouth Meeting, PA). Our preliminary studies confirmed that squalamine did not affect food intake or body weight (5). In contrast, intraperitoneal (IP) injection of MSI-1436 during the light or dark cycle reduced food intake and body weight in a dose-dependent manner (data not shown). We selected a lower dose of MSI-1436 that did not suppress drinking as previously described (5). MSI-1436 (5 mg/kg IP × three doses) was administered at 0900–1000 at 3-day intervals to a group of mice (n = 6). Control mice (n = 6) were treated with vehicle (100 μl endotoxin-free H₂O IP). A third group of mice (n = 6) was pair-fed to the daily intake of MSI-1436 mice. Indirect calorimetry was performed after the final treatment. The mice were acclimatized to the test cage for 2 days, and energy expenditure was measured at 15-min intervals for 24 h on the third day (Oxymax Equalflow System; Columbus Instruments, Columbus, OH) (6). The following settings were used per cycle: air flow 500 ml/min, sample flow 400 ml/min, settle time 120 s, measuring time 60 s, temperature 22°C, respiratory exchange ratio (RER) = volume of carbon dioxide generated (Vco₂) divided by oxygen consumption (Vo₂). Heat (kcal/h) = 3.815 + 1.232 × RER. Total and ambulatory activity were measured simultaneously with photodetectors (Optovarimex System; Columbus Instruments). Rectal temperature was measured with a thermistor (Physitemp Instruments, Clifton, NJ).

The mice were killed by CO₂ inhalation, and blood was obtained by cardiac puncture. Glucose and triglycerides (Sigma, St. Louis, MO) and nonesterified fatty acids (NEFAs) (Wako Chemicals, Richmond, VA) were measured with enzyme assays. Plasma insulin, leptin, corticosterone, and thyroxine were measured by radioimmunoassay as previously described (7,8). Uncoupling protein (UCP)-1 mRNA was measured in brown adipose tissue (BAT) by Northern blot analysis using a cDNA probe (provided by Mitch Lazar, University of Pennsylvania). The rest of the carcass was dried to a constant weight at 60°C to determine water content and digested in ethanol-KOH, and fat (triglyceride) content was measured with a colorimetric assay (Sigma) (8).

We determined whether MSI-1436 could prevent the hyperphagia normally associated with fasting. Male 12-week-old C57Bl/6J mice were fasted for 48 h and received a single injection of MSI-1436 (5 mg/kg IP) or vehicle (n = 6/group) after the fast. They were housed in individual cages and allowed ad libitum access to normal diet and water. Food intake and body weight were measured daily. Feeding frequency and duration were monitored using infrared detectors (Vitalview System; Mini Mitter, Bend, OR). The data on feeding frequency and duration were collated in 1-h bins and analyzed using Actiview software (Mini Mitter).

Male Sprague-Dawley rats (250–300 g) were obtained from Taconic Farms (Germantown, NY), housed in a 12:12 h light-dark cycle (lights on at 0600; temperature 22°C), and allowed normal laboratory food and water ad libitum. The animals were anesthetized with sodium pentobarbital, and a 22-gauge stainless steel guide cannula with obturator (Plastics One, Roanoke, VA) was implanted unilaterally in the lateral cerebral ventricle using the following coordinates: 0.8 mm posterior to bregma, 1.5 mm lateral to the midline, and 4 mm below the skull. The intracerebroventricular (ICV) cannula was attached to the cranium with stainless steel screws and dental cement. Cannula placement was verified with the drinking response to ICV angiotensin II injection and histologically after the experiment. Only data from rats with correctly positioned cannulas were included in the analysis.

The animals were housed individually after surgery and handled daily to habituate, and two sham injections were performed at 3-day intervals. Experiments were performed after restoration of body weight (∼1 week). MSI-1436 (2, 10, or 30 μg) or vehicle (5 μl endotoxin-free H₂O) was administered ICV over 1 min to unanesthetized rats (n = 6) with an injector protruding 1 mm below the guide cannula into the cerebral ventricle. Other rats (n = 6/group) received a single intraperitoneal injection of MSI-1436 (5 mg/kg) or vehicle (200 μl endotoxin-free H₂O). The animals were presented with a preweighed amount of pellet food, and food consumption and body weight were measured daily. The effect of MSI-1436 on conditioned taste aversion was investigated in other rats using the two-bottle paradigm (9). The animals (n = 4 per group) were treated with a single injection of MSI-1436 (30 μg ICV), MSI-1436 (10 mg/kg IP), vehicle (ICV or IP), or LiCl (100 mg/kg). Saccharin preference ratio was determined as the volume of 0.1% saccharin divided by water consumed in 24 h (9).

Expression of the immediate-early gene c-fos has been used to map neuronal pathways for feeding regulation (10–13). Fos immunostaining was performed on coronal brain sections from pathogen-free male Sprague-Dawley rats (Taconic Farms) as previously described (10). The animals were treated with MSI-1436 (5 mg/kg in 200 μl endotoxin-free H₂O IP or 20 μg in 5 μl endotoxin-free H₂O ICV; n = 4/group). Control rats (n = 4) were treated with vehicle (IP or ICV). The animals were anesthetized with sodium pentobarbital 2 h later (1000–1200) and perfused transcardially with PBS followed by 10% neutral-buffered formalin. Brains were cryoprotected in 20% sucrose/PBS, and free-floating coronal sections (40 μm) were cut on a sliding microtome and processed for Fos immunoreactivity (10). The sections were examined with a Nikon E600 microscope equipped with a digital camera and image analysis system (SPOT Diagnostic Instruments, Sterling Heights, MI). Fos-positive cells/ region/hemisphere with distinct nuclear staining were counted by an observer blinded to the experimental protocol using software provided by the manufacturer (Phase 3 Imaging, Glenmills, PA).

Based on the induction of strong Fos-immunoreactivity in the paraventricular hypothalamic nucleus (PVN) (see results), we determined whether injection of MSI-1436 into this hypothalamic region would recapitulate the response to ICV administration. A 26-gauge stainless steel cannula with obturator (Plastics One) was implanted into the PVN in rats as described (13). One week later, MSI-1436 (0.5 or 1 μg) or vehicle (0.5 μl H₂O) was microinjected into the PVN. The animals were housed singly, and food intake and body weight were measured daily. The injection site was verified histologically, and only data from confirmed PVN injection (n = 4/group) were included in the analysis.

Male 12-week-old C57Bl/6J mice were treated with three injections of MSI-1436 (5 mg/kg IP) or vehicle (100 μl endotoxin-free H₂O) at 3-day intervals. A third group was pair-fed to MSI-1436 treatment. Three days after the last injection, the mice were killed, hypothalami were dissected, and quantitative RT-PCR analysis was performed using specific primers for agouti-related peptide (AGRP), neuropeptide Y (NPY), pro-opiomelanocortin (POMC), cocaine- and amphetamine-regulated transcript (CART), melanin-concentrating hormone, and β-actin (14). For corticotropin-releasing hormone, the following primers were used: upstream, 5′-gcatcctgagagaagtccctctg-3′; downstream, 5′-aagttagccgcagcgtggtc-3′, corresponding to nucleotides 437–459 and 1476–1495 (15). The signal was measured by Phosphoimager (Molecular Dynamics, Sunnyvale, CA), and densities corresponding to mRNA expression of various peptides were normalized to β-actin (14).

The effects of MSI-1436 on food intake, body weight, and other parameters were compared by ANOVA. Pairwise differences between groups were determined using Fisher’s protected least significant differences test; P < 0.05 was considered significant.

To ascertain whether weight reduction by MSI-1436 was mediated entirely by inhibition of feeding, the daily intake of a group of pair-fed mice was matched with that of MSI-1436-treated mice. Cumulative food intake was decreased by ∼20% in MSI-1436-treated and pair-fed mice (Fig. 1A); however, body weight (Fig. 1B) and body fat (Fig. 1C) were significantly lower in MSI-1436-treated mice than in pair-fed mice, indicating that inhibition of food intake alone could not account for the weight loss in MSI-1436-treated mice. Oxygen consumption (Vo₂) (Fig. 1D) and heat (data not shown) were maintained at high levels in MSI-1436-treated mice despite weight reduction. RER, an index of metabolic fuel use, was not different between MSI-1436-treated (0.8 ± 0.01) and vehicle-treated (0.84 ± 0.02) mice. In contrast, Vo₂ was lower in pair-fed mice (Fig. 1D). RER was significantly lower in pair-fed mice (0.77 ± 0.02) than in vehicle-treated mice (0.84 ± 0.02) (P < 0.05). Body temperature, UCP-1 mRNA expression in BAT, and locomotor activity (beam breaks) were not affected by MSI-1436 (data not shown).

Serum triglycerides, insulin, and leptin were reduced in both MSI-1436-treated and pair-fed mice (Table 1). NEFAs and the NEFA-to-triglyceride ratio were twofold greater in pair-fed mice than in MSI-1436-treated mice, consistent with increased lipolysis (Table 1). Thyroxine was decreased and corticosterone was increased in pair-fed mice but remained normal after MSI-1436 treatment (Table 1). Feeding frequency and food consumption increased rapidly in vehicle-treated fasted mice, leading to restoration of body weight within 48 h (Figs. 2A–C). In contrast, MSI-1436 blunted the postfast hyperphagia and weight gain (Figs. 2A–C). The duration of feeding bouts was not significantly different between MSI-1436 treatment and pair-feeding (data not shown).

Cerebral ventricular injection of MSI-1436 resulted in a dose-related decrease in food intake and body weight (Figs. 3A and B). Water intake was reduced slightly by the highest ICV dose (30 μg MSI-1436) but was not statistically significant (33.7 ± 2.6 vs. 30 ± 4.6 ml/24 h; P = 0.21). The effective ICV dose of MSI-1436 was <1,000 of the peripheral dose (Figs. 3A and B). In all cases, body weight was restored several days after MSI-1436 treatment (Fig. 3B). MSI-1436 treatment did not cause taste aversion, as evidenced by a normal saccharin preference ratio of 0.8 in vehicle, ICV, and IP MSI-1436-treated rats. In contrast, saccharin preference ratio was markedly suppressed to 0.38 (P < 0.001) after LiCl administration, consistent with aversion behavior.

The distribution of potential MSI-1436 targets in the brain was evaluated using Fos immunohistochemistry. A robust induction of Fos-immunoreactive cells was detected in the PVN after central or peripheral MSI-1436 administration (Figs. 4A and B). Fewer numbers of Fos-positive cells were observed in other forebrain regions involved in ingestive behavior, energy balance, and glucose homeostasis, i.e., arcuate nucleus, perifornical region, zona incerta, lateral hypothalamic area, dorsolateral and ventromedial hypothalamic nuclei, and central amygdala (Fig. 4B). Injection of MSI-1436 into the PVN caused a dose-dependent reduction in food intake (Fig. 4C) and body weight (Fig. 4D). As with ICV treatment, the response to a single PVN injection persisted for several days (Figs. 4C and D).

The expression of hypothalamic neuropeptides was analyzed after MS-1436 treatment or pair-feeding (caloric restriction) (Fig. 5). MSI-1436 decreased NPY mRNA expression in the hypothalamus by 10% (P = 0.1) and AGRP mRNA by 60% (P < 0.001) (Fig. 5). In contrast, NPY mRNA increased by 12% (P = 0.08) and AGRP mRNA by 27% (P < 0.01) in pair-fed mice (Fig. 5). The expression of POMC, corticotropin-releasing hormone, CART, and melanin-concentrating hormone was not affected by MSI-1436 or caloric restriction (Fig. 5).

MSI-1436 and other aminosterols that affect feeding were discovered serendipitously during a search for antimicrobial compounds in the dogfish shark (4,5). Initial studies showed that MSI-1436 decreased body weight over long periods. However, there was a lack of understanding as to whether the sustained reduction in body weight was due solely to appetite suppression. Our studies provide evidence in support of a specific suppression of feeding as well as increased energy expenditure after MSI-1436 treatment. Importantly, MSI-1436 reduced body weight (specifically fat) without provoking the well-known response to caloric depletion (16,17). In the physiological model of energy homeostasis, the amount of energy stored in adipose tissue reflects the balance between energy intake and expenditure (16). A decrease in energy stores from fasting triggers various responses, e.g., reduced energy expenditure and overfeeding in an attempt to restore body weight (16). Other typical responses to food deprivation include decreased thyroid hormone, increased glucocorticoids, and increased lipolysis (evidenced by an increased NEFA-to-triglyceride ratio and reduced RER). These responses, which are mediated at least in part by reduced leptin level, are designed to defend body weight and lean mass by reducing energy expenditure and stimulating feeding (16,17). The counter-regulatory mechanisms may be responsible for the failure to sustain weight loss during forced caloric restriction (16). MSI-1436 was effective in reducing food intake and body weight in ad libitum-fed animals for several days and also prevented the hyperphagia and rapid weight gain normally associated with fasting, indicating that its pharmacological effects can override the physiological tendency to gain weight. The ability of MSI-1436 to maintain high-energy expenditure despite weight loss is not likely to be mediated through UCP-1 or locomotor activity because these factors were not altered.

Based on a previous study (5) and our own study (data not shown) showing that MSI-1436 reaches the brain after peripheral injection, we compared the dose-effect of ICV treatment versus systemic MSI-1436 treatment. Administration of MSI-1436 ICV was >1,000 times more potent than intraperitoneal treatment. Central or peripheral MSI-1436 treatment did not cause taste aversion, thus providing evidence against visceral illness as a cause of the reduced food intake and body weight. We mapped potential targets of MSI-1436 in the brain using Fos immunostaining. The induction of Fos immunoreactivity in the PVN is consistent with neuronal activation and might explain the effects of MSI-1436 on feeding and energy expenditure because this hypothalamic region mediates feeding behavior and autonomic function (10,16,17). Fos immunoreactivity in the PVN has been shown to be regulated by a variety of factors, such as leptin, urocortin, corticotropin-releasing factor, and AGRP (10–13). We observed fewer Fos-positive cells in other hypothalamic and forebrain regions implicated in energy homeostasis. However, activation of Fos within the PVN does not necessarily prove that MSI-1436 acts at this site. Thus, we determined whether direct administration of MSI-1436 into the PVN would produce an effect similar to that of ICV or peripheral MSI-1436 injection. Administration of one-thirtieth the ICV dose of MSI-1436 into the PVN resulted in a sustained reduction in food intake and body weight, suggesting that the PVN is a specific target of MSI-1436.

The PVN receives neuronal input from the arcuate nucleus, other hypothalamic regions, and the brainstem (16,17). The orexigenic neuropeptides NPY and AGRP are produced by the same neurons in the arcuate nucleus and increase body weight by stimulating appetite and increasing energy expenditure (16,17). Conversely, α-melanocyte-stimulating hormone (α-MSH) is produced by POMC neurons in the arcuate nucleus and decreases body weight by inhibiting appetite and increasing expenditure (16,17). The cellular action of α-MSH is mediated through antagonism of AGRP at melanocortin receptor-3 (MCR3) and MCR4. Hypothalamic neuronal circuits expressing these neuropeptides respond to leptin, insulin, glucocorticoids, and other endogenous factors (16,17) and can be pharmacologically influenced by other molecules such as ciliary neurotropic factor (CNTF) and the fatty acid synthase inhibitor C75 (18–21). MSI-1436 significantly inhibited AGRP mRNA and, to a lesser extent, NPY mRNA expression in the hypothalamus. This response was distinct from caloric deprivation, in which AGRP and NPY mRNA expression increased, consistent with their roles as stimulators of feeding (16,17).

We speculate that the central action of MSI-1436 is mediated at least in part through inhibition of AGRP/NPY projection to the PVN (16,17). The suppression of AGRP may explain the prolonged inhibition of feeding by MSI-1436 because AGRP has the unique ability to regulate food intake and body weight for several days after a single central injection (21). The ability of MSI-1436 to profoundly inhibit appetite while increasing energy expenditure for long periods distinguishes it from other weight-reducing compounds. In case the pharmacology of MSI-1436 is shown to be similar in humans, these properties could prove to be advantageous by counteracting the homeostatic metabolic responses that resist weight loss.

This work was supported by National Institutes of Health Grant P30 DK50306.

The authors thank Genaera Corporation (Plymouth Meeting, PA) for providing MSI-1436 and squalamine, Qinglan Jiang for initial studies, Mitchell A. Lazar for helpful discussions, and the Mouse Phenotyping, Physiology & Metabolism and Radioimmunoassay Cores of the Penn Diabetes Center and the Morphology Core of the Penn Center for the Molecular Study of Digestive Diseases for technical assistance.