Gastric Mucosal Damage Following Repeat Administration of Melanocortin Subtype-4 Receptor Ligands to Fischer 344 Rats

Abstract

Since the melanocortin system plays a role in the central control of feeding, melanocortin receptor ligands may be efficacious in treating human obesity. Ten structurally similar melanocortin subtype-4 receptor (MC4R) ligands from an aryl piperazine chemical platform were evaluated in female Fischer 344 rats to assess the toxicity of this class of compounds. Rats were orally gavaged with 100, 250, or 500 mg/kg of each compound in 10% acacia and purified water daily for 4 days. In treated rats, notable clinical observations included a dose-dependent decrease in mean body weight and food consumption. A morphologically unique compound-related histologic lesion occurred in the nonglandular gastric mucosa. The lesions consisted of multiple, raised, sometimes ulcerated, white foci which, microscopically, were discrete, intraepithelial vesicles containing dense accumulations of neutrophils continuous with inflammation in the submucosa. Ruptured vesicles resulted in ulcers and occasionally gastric perforation. The morphologic characteristics of this acute lesion were described and concluded to be a direct toxicity of the compounds unrelated to melanocortin-mediated pharmacology.

Introduction

The melanocortin subtype-4 receptor (MC4R) belongs to a family of seven transmembrane G protein-coupled receptors containing five melanocortin receptors, and plays a role in energy homeostasis. The biology of this family of receptors, which in addition to energy homeostasis includes pigmentation, inflammation, sexual function, and exocrine secretion, has been extensively reviewed (Schïoth, 2001; Gantz and Fong, 2003; Cone, 2005). Melanocortin subtype-4 receptors (MC4Rs) are primarily located throughout the central nervous system, at least in the rat, but also have been detected in some peripheral tissues in the developing rat fetus (Gantz et al., 1993; Mountjoy et al., 1994, 2003).

The MC4R is an attractive target for a human anti-obesity drug and is being aggressively explored by numerous pharmaceutical companies (Boyce and Duhl, 2004; Todorovic and Haskell-Luevano, 2005). Neuroanatomical mapping studies of the rat hypothalamus demonstrate MC4R in the parvicellular neurons of the paraventricular nucleus (PVN), an area involved in regulating energy balance and food intake. Substantial animal and human target validation supports its role in feeding and weight loss. In humans, mutations in the MC4R, the melanocortin precursor gene (proopiome-lanocortin; POMC), and prohormone convertase that processes the POMC precursor peptide result in obese humans (Jackson et al., 1997; Krude et al., 1998; Vaisse et al., 2000). In mice containing MC4R or POMC gene distruptions, obesity and hyperphagic phenotypes result. (Huszar et al., 1997; Coll et al., 2004). Numerous reviews, which provide additional evidence to support the role of melanocortins and feeding, are available (Vergoni and Bertolini, 2000; Williams et al., 2001; Barsh and Schwartz, 2002; Bays, 2004).

The aim of this study was to identify the potential toxicity of several selective MC4R agonists from an aryl piperazine chemical platform in the rat following repeat administration. Compounds used in this study were tested in either a luciferase or cAMP release functional activity assay in human embryonic kidney 293 cells stably transfected with human MC4R and were full agonists at the MC4R (data not shown). Full concentration-response curves for several of these compounds at each of the receptors tested have been previously published (Richardson et al., 2004). Our investigation here identified a morphologically unique gastric lesion in the nonglandular portion of the rat stomach. Lesions of irritation and toxicity, including inflammation, necrosis, erosion, ulceration, and proliferation, occur in the forestomach of rodents in association with a variety of chemical and pharmaceutical agents (Gopinath et al., 1987; Maronpot, 1999); however, the appearance and progression of this lesion make it particularly unique. The morphologic characteristics of this acute lesion are described and are concluded to be a direct toxicity of the compounds unrelated to melanocortin-mediated pharmacology.

Materials and Methods

Chemicals

Novel arylpiperazine-containing compounds (Figure 1) were synthesized from a common, previously described chemical platform (Richardson et al., 2004).

In Vitro Analysis

For the in vitro cytolethality assay, a rat skeletal muscle cell line (L6 cells, American Type Culture Collection, ATCC# CRL 1458) was used with stock cultures maintained in Medium M-199 with 5% fetal bovine serum and 25 mM HEPES buffer. Cultures were passed to maintain subconfluent densities and prevent myoblast fusion and were maintained in a humidified incubator (37°C, 5% CO₂). The test cultures were plated at 3 × 10⁴ cells per well in 96 well plates approximately 24 hours prior to the initiation of treatment with compounds. The compounds were prepared as 20 mM stock solutions in DMSO and diluted 1:100 into medium M-199 containing 3% fetal bovine serum and 25 mM HEPES buffer to yield a 200 μM incubation solution. Serial dilutions were performed to achieve final concentrations of 200, 80, 20, and 8 μM. Duplicate cultures were administered each dilution of the compounds. Following a 24-hour incubation, cultures were washed and cellular viability was determined using WST-1 (Roche Biosciences) as an indicator of mitochondrial metabolism. Absorbance of reduced WST-1 was measured at 440 nm using an absorbance microplate reader. The percent of control viability (WST-1 absorbance) for each concentration tested was used to calculate a Toxic Concentration 50% (TC₅₀) using a regression fit analysis.

Animals

Female Fischer 344 rats, 6–7 weeks old, were obtained from Harlan (Indianapolis, IN). Rats were acclimatized before use and housed individually in ventilated stainless steel cages with wire mesh floors and Lexan fronts. Rats had ad libitum access to Certified Rodent Diet 5002 (PMI Nutrition International, Inc.) and Greenfield city tap water. Rats were maintained in a continuously monitored environmentally controlled room on a 12-hour light:dark cycle with temperature and humidity maintained between 72 ± 3°F (22.2 ± 1.7°C) and 30–70%, respectively. All rats were acclimated for a minimum of one week prior to use. All experiments were conducted in compliance with the institutional guidelines for the care and use of laboratory animals.

Dose Administration

Rats were assigned to treatment groups (n = 3) by body weight stratification using a partitioning algorithm. Compounds were formulated daily in 10% acacia (Colloïdes Naturels International, Rouen Cedex, France) in purified water. Rats were administered 100, 250, and 500 mg/kg of each compound or vehicle control by oral gavage (10 ml/kg body weight) for 4 days. Clinical observations and food consumption were recorded daily and body weight was recorded at the beginning and the end of the study. Rats were fasted overnight prior to necropsy.

Toxicokinetics

For selected compounds, an additional 2 rats were added to each dose group to measure systemic exposure of the compounds. Rats were administered a single dose of compound and approximately 200 μl of blood was collected into heparinized syringes by retro-orbital bleeding. Blood was collected at various times between 0.5 and 24 hours postdose from isoflurane anesthetized rats. Each rat was used for 3 different collection time points. Plasma samples were analyzed using an electrospray LC/MS/MS on a Sciex API 3000 mass spectrometer using fast gradient elution on a 2 × 50 mm C18 HPLC column.

Clinical Pathology

Rats were anesthetized with isoflurane and approximately 0.6 ml of blood was collected by retro-orbital bleeding into tubes with EDTA. Plasma samples were prepared for standard hematology (erythrocyte count, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular hemoglobin, mean corpuscular hemoglobin concentration, total leukocyte count, leukocyte differential, platelet count, blood cell morphology, reticulocyte count).

Morphologic Pathology

Rats were euthanized by carbon dioxide asphyxiation. The necropsies included examinations of all external body surfaces and orifices; the thoracic, abdominal, pelvic, and cranial cavities and their viscera; cervical tissues and organs; and external surfaces of muscle, nerve, and spinal cord. The kidney liver, heart, lung, spleen, thymus, adrenal, stomach, duodenum, jejunum, ileum, bone marrow, cerebellum, cerebrum, brain stem, and diaphragm were fixed in 10% neutral buffered formalin. Preserved tissue specimens were trimmed, processed through graded alcohol and clearing agent, infiltrated and embedded in paraffin, microtomed, and stained with hematoxylin and eosin. Severity of lesions was classified using a 5-point grading scale of minimal, slight, moderate, marked, and severe.

Results

Not all rats survived until the scheduled study termination. Mortality was observed in the 500 mg/kg dose group with compounds 3, 4, and 9, and in the 250 and 500 mg/kg dose groups with compound 6. The cause of death could not be determined. In the surviving animals, soiling, rough and wet haircoat, chromodacryorrhea, lacrimation, and decreased and soft feces were common findings in compound-treated animals, and were considered signs of general debilitation. A dose-dependent decrease in body weight and food consumption was observed for all compounds tested; however, the severity of the decrease varied with compound (Table 1). For compounds in which plasma exposure was measured, an approximately linear, dose-proportional increase in exposure was present (Table 2). Exposure to compounds 4, 5, and 8 was approximately 100-fold lower than exposure for compounds 7 and 10.

Stomach lesions occurred in the nonglandular mucosa at doses of 100, 250 and/or 500 mg/kg with several test compounds (see Table 2). The incidence and severity were generally dose-progressive. At necropsy, the stomach of affected rats was dilated and fluid-filled, with multiple, small (1–2 mm), raised, sometimes ulcerated, white foci on the mucosa of the nonglandular portion of the stomach. Histologically, the lesions consisted of a range of acute necrotizing and inflammatory foci occurring randomly throughout the nonglandular region of the stomach only. The lesions included foci of acute submucosal edema and hemorrhage (Figure 2B) or areas of mucosal and submucosal necrosis and inflammation (Figure 2C). More advanced lesions consisted of microsabscessation in the submucosa and mucosa which often produced eruption of suppurative vesicles towards and into the epithelium (Figure 2D). Intraepithelial microabcesses were the most common findings, with continuity with the submucosa and macrovacuolar and ballooning degeneration of the overlying epithelium (Figure 2E). Rupture of the intraepithelial foci led to ulceration (Figure 2F) and when accompanied b y full thickness necrosis of the stomach wall produced gastric perforation in at least one rat given a dose of 500 mg/kg of a structurally similar MC4 agonist in addition to the 10 compounds in Figure 1, which was not disclosed.

Minimal changes in leukograms and hemograms of rats were characterized by decreases in lymphocytes, increases in neutrophils and monocytes, and decreases in reticulocyte counts and were considered typical of stress/endogenous corticosteroid release and/or inflammation (compounds 2, 4–10). These changes usually accompanied decreases in body weight and food consumption and histologic findings of bone marrow hypocellularity, lymphoid depletion in the thymus and spleen, and zymogen granule depletion in the pancreas, all of which were considered secondary to stress superimposed on inflammation.

Additional studies were conducted to assess the cytolethality of the MC4 agonists. Results from the cytolethality assay indicated that the compounds evaluated were moderately to very cytotoxic in vitro (Table 3).

Discussion

The toxicity of 10 structurally similar selective MC4R agonists was evaluated in the rat. As anticipated, a dose-responsive decreased body weight and food consumption were observed in these studies. Although mortality with a few compounds was observed, the remaining clinical observations were minor. Within the morphologic pathology findings was a lesion in the nonglandular mucosa of the stomach. The observed gastric lesion demonstrated unique morphology consisting of a mixed pattern of hemorrhage, necrosis and inflammation originating in the submucosa and causing necrosis of overlying epithelium, and/or individual, deep submucosal pustules erupting through the nonglandular mucosa forming cavitations and ulcers. Unlike typical gastric ulcerogens that produce erosive, inflammatory, and necrotic lesions that ulcerate and deepen from the surface mucosa into the underlying submucosa, the lesions presented here are clearly different due to the focal origin deep within the mucosa, the formation of pustules, and the eruption through the epithelium.

Chemical or pharmaceutical related pustular and/or vesicular lesions in the nonglandular mucosa of the rodent stomach are uncommon, but have been reported in the following toxicologic studies (Silver et al., 1982; Ghanayem et al., 1985; Wallig et al., 1988). The phytochemical 1-cyano-3,4-epithiobutane caused lesions very similar to what is reported here, but much less severe and lesions were seen only in male rats. Acrylnitrile, ethyl acrylate, and dibromoacetonitrile also cause gastric erosions and ulcerations. The lesions with these chemicals were similar to the lesions observed in our study with respect to involvement of the submucosa and overlying mucosal epithelium including macro-vacuolar change and ballooning degeneration, but dissimilar in terms of the magnitude of the inflammatory reaction. The acute lesions in our studies were more suppurative and formed microabscesses, however, temporal evaluation of the lesions caused by the MC4 compounds was not possible since studies greater than 4 days in duration were not performed. The 3 chemicals identified are glutathione depletors that are known to alter the sulfhydyl status of certain cholinergic receptors leading to an enhanced response to muscarinic agents in gastrointestinal toxicity (Ahmed et al., 1991). Other gastric/duodenal ulcerogens include propionitrile and non-nitrile cysteamine, which are known to increase gastric acidity and/or alter catecholamine and dopamine status (depletion) in the stomach (Szabo et al., 1987, 1997).

In our study, the etiology of the gastric lesions was not likely due to a surface caustic effect because the lesions were small, multifocal, randomly distributed and arose from within the submucosa. Although not measured in this study, the pH of these formulation was > 2 and within physiological range and therefore was not considered a factor in lesion development. Additionally, the moderate to severe cytotoxicity results did not align with a direct effect on the gastric mucosa resulting in the formation of the lesion. However, the in vitro cytolethality assay is a measurement of the direct toxicity of the parent compound on cells and does not take into consideration potential metabolites, systemic exposure, bioavailability, nor the effects from repeat administration of the compounds. This, combined with the observation of areas of overt epithelial necrosis, suggests that a caustic effect cannot be entirely eliminated. Focal inflammatory and necrotizing lesions in the submucosa may be attributed to vascular compromise and local ischemia; however, thrombosis, vasculitis, or fibrinoid necrosis were not apparent in any section.

The lesion resembled lesions in stratified epithelial tissues of other animal species in other disease processes such as toxic rumenitis in ruminants; however, the lesions in this study were not diffuse and were more pustular. The lesion did not appear iatrogenic (such as gavage trauma) and was probably not a manifestation or exacerbation of an infectious process, although based on the intense suppuration and microabscessation an infectious process could not be entirely eliminated. Stress-related gastric ulceration is a reported component of stress and food deprivation in which hemorrhagic lesions occur in the glandular mucosa (Iatropoulos and Davis, 1986) and ulcerative lesions occur in the nonglandular mucosa of the stomach of rats within 6 days (Nakagawa et al., 1985). While there were clinical pathology changes consistent with stress and there was decreased food consumption, the manifestation of the stomach lesion in these rats was unlike typical stress-related gastric lesions in morphology and severity, and the dose progressive incidence and severity suggest a likely compound-related effect.

Exposure data from several of the compounds was assessed to understand the relationship of the gastric lesions to systemic levels of the compounds. In rats with the lowest systemic exposure, which occurred with compounds 5 and 8, no evidence of gastric damage was observed. However, as exposure increased, both the incidence and severity of the lesion increased. It was unclear why the presence of the gastric lesion was not observed with compounds 7 and 10 at doses below 500 mg/kg, given that exposure at lower doses was associated with the lesion with other compounds. Since tissue distribution of the compounds was not assessed, it was difficult to determine a direct relationship between exposure and mucosal damage. Based on the limited exposure data and lack of correlation with the generation of the gastric lesion, no firm conclusions were drawn to connect systemic exposure to the lesion.

Melanocortin peptides, including adrenocorticotropin, and α-, β-, γ-melanocyte-stimulating hormones (MSH), mediate their effect through the melanocortin receptors (MC1-5). Evidence suggests that α-MSH is involved in gut cytoprotection, anti-inflammatory, and immunomodulating effects (Rajora et al., 1997; Konjevoda et al., 2001), which may help further support the conclusion that the gastric lesions observed in this study were not pharmacologically-mediated. α-MSH did not induce lesions in the normal rat gut mucosa and significantly reduced lesion formation in the ethanol-induced gastric lesion model (Konjevoda et al., 2001). Although not fully understood, evidence indicates the α-MSH mediates its anti-inflammatory effects through the MC1 and MC3 receptors and not the MC4R (Cantania et al., 2004). Therefore, any anti-inflammatory effect mediated by MC4R can be excluded.

Understanding MC4 receptor distribution and the potency of the tested compounds to this receptor was essential in understanding the potential link between the gastric lesion and MC4-mediated pharmacology. Using in situ hybridization, no expression of MC4 signal was measured in the developing rat fetus stomach (Mountjoy et al., 2003). The lack of MC4R in the rat stomach provides evidence to negate a direct MC4R-mediated gastric lesion. Additionally, the binding affinities (K_i) of each of the MC4 compounds in this study were determined by competitive inhibition of ¹²⁵I-radiolabeled-[Nle⁴,D-Phe⁷]-alpha-melanocyte stimulating hormone binding in human embryonic kidney 293 cells stably transfected with human MC4 receptors as previously described (Richardson et al., 2004). Since binding affinities to the human MC4 receptor are approximately 2-fold less potent than at the rat MC4 receptor while maintaining the same rank order of potency, measuring binding using the human receptor provides a reasonable estimate of the binding affinities in the rat (Husain et al., 2004). Compounds from this chemical platform were selective for the human MC4 receptor compared to other receptor subtypes including human MC1, MC3 and MC5 receptors (Richardson et al., 2004). No correlation between MC4 binding affinity and gastric lesion generation was present. Finally, no correlation between previously determined binding affinity (Husain et al., 2004) and cytolethality was present. Although the pathogenesis of this lesion was unknown, the evidence presented above supports the claim that a direct pharmacologically mediated effect via the MC4R in the rat stomach was unlikely.

In our study there was no toxicologically meaningful evidence of injury to the glandular portion of the stomach and small intestine. The lack of lesions in the glandular mucosa may be explained by the presence of a mucus coat and other protective features not present in the nonglandular mucosa. While the pathogenesis of this lesion probably did not involve the activation of MC4R, it was more likely related to the chemical structure of the compounds. Thus, exploration of other MC4R ligands with chemical structural diversity will likely provide a compound without the liability of the unique gastric lesion described in these studies.