Cachectic disease is characterized by decreased food intake and increased metabolic rate leading to body mass wasting. Mechanisms for appetite suppression have focused on leptin and IL-1β (Vellucci et al., 1995; Grunfeld et al., 1996; Sachot et al., 2004). Although leptin regulates food intake (Chilliard et al., 2005) and endocrine events in ruminants (Morrison et al., 2001; Daniel et al., 2002; Zieba et al., 2005), plasma leptin in sheep is unchanged after endotoxin (lipopolysaccharide, LPS; Soliman et al., 2001; Daniel et al., 2002), whereas leptin mRNA in renal fat was decreased (our unpublished data). Similarly, there was no recovery of feed intake in sheep given an IL-1β receptor antagonist (McMahon et al., 1999), leaving questions about the mechanisms for LPS-induced suppression of appetite in sheep.
One target for disease suppression of appetite is the melanocortin-4 receptor (MC4R). Proopiomelanocortin (POMC) gene expression and thus a-melanocyte-stimulating hormone (α-MSH) release is increased in the hypothalamus 4 h after LPS (Sergeyev et al., 2001). Moreover, IL-1β released α-MSH from hypothalamic slices (Scarlett et al., 2007). Evidence for a role for the MC4R and the melanocortin system is also based on studies that found that tumor-bearing MC4R knockout mice had no suppression of appetite compared with wild-type mice with tumors (Marks et al., 2003). In addition, injection of a MC4R antagonist into the lateral ventricle of the brain before intravenous injection of LPS prevented food intake inhibition (Huang et al., 1999). These studies argue that appetite suppression in chronic disease is mediated in part through α-MSH activation of the MC4R. Because agouti-related protein (AgRP) regulates appetite by antagonism of the MC4R in sheep (Wagner et al., 2004), this study was designed to address the hypothesis that AgRP may mediate some of the effects of LPS on appetite in sheep.
MATERIALS AND METHODS
All experiments were approved by the Auburn University Institutional Animal Care and Use Committee.
Animals and Maintenance
Castrated male sheep were housed indoors in individual pens with 2 sheep per room at 21°C and a 12-h photoperiod. The sheep were 1 yr of age and had a BW of 43.5 ± 7.8 kg. Sheep had ad libitum access to water and concentrate feed, which contained 12% crude protein on an as-fed basis and was calculated to meet 100% of their daily requirements.
Exp. 1. AgRP Antagonism of the MC4R in Sheep Treated with LPS
Following an overnight fast, sheep (n = 6) were anesthetized with triple drip (0.1 mg/mL of xylazine, 2 mg/mL of ketamine in 5% guaifenesin) at 1 mL/1.36 kg, placed in a sheep stereotaxic device (David Kopf Instruments, Tujunga, CA), and maintained under anesthesia with isoflourene. A guide cannula with luer stylette was placed into a lateral ventricle 15 mm caudal and 8 mm lateral to the bregma (McMahon et al., 1999; Wagner et al., 2004). Intracerebroventricular cannula placement was confirmed by taking a radiograph immediately after injecting 1 mL of radiopaque dye (Omnipaque 300; Sterling Drug Inc., New York, NY) into the cannulated ventricle.
Sheep were randomly placed in 1 of 4 treatment groups (saline, LPS, AgRP, AgRP + LPS), and a different treatment was applied each week until each sheep had received each treatment. However, sheep were not given LPS on consecutive weeks to prevent development of LPS tolerance (Elsasser et al., 2004). Sheep ad libitum fed were offered 1.8 kg of fresh feed at 0600 h, and 1 h later (0700 h), feed was removed and weighed. At 0730 h, sheep were injected intracerebroventricularly with either saline or AgRP (0.5 nmol/kg of BW; AgRP courtesy of Dan Marks, Oregon Health Science University, Portland, OR) through the lateral ventricle cannula 1 h before they were injected i.v. with either saline or LPS (0.6 μg/kg of BW) at time 0 and then again at 4 h. Feed was returned at 0900 h (90 min after AgRP injection and 30 min after LPS injection). Feed intake was monitored by weighing feed not consumed, calculating feed intake, and replenishing consumed feed at 1, 2, 4, 6, 8, 12, 24, and 48 h after LPS injections. Body temperature was monitored by rectal thermometer in all of the animals at the time of food intake measurements.
Exp. 2. RNA Isolation and Real-Time PCR
This experiment was performed to determine whether AgRP mRNA was altered by LPS treatment. Sheep (n = 20) were injected with either saline or LPS (0.6 μg/kg of BW), and sheep were killed at 3 or 6 h after treatment by i.v. injection of Beuthanasia D (1 mL/4.5 kg; Schering-Plough Animal Health Corp., Union, NJ). Hypothalami were removed within 3 min of euthanasia and stored at −70°C until extracted. Total RNA was extracted using Trizol reagent and purified using the RNeasy kit (Qiagen, Valencia, CA). Quality of RNA was determined spectrophotometrically and by electrophoresis. The total RNA (1 μg) was reverse-transcribed (Iscript cDNA synthesis kit, BioRad, Hercules, CA), and the cDNA was subjected to reverse transcription-PCR using primers designed from ovine AgRP sequences (GenBank number AY310396) and primers designed for control housekeeping genes (bovine glyceraldehyde phosphate dehydrogenase AF106860, human acidic ribosomal phosphoprotein NM_001002, human cyclophilin A AY00846, porcine ubiquitin U72496, porcine β-actin AJ312193), and Eppendorf Taq DNA Polymerase (Fisher Scientific, Pittsburgh, PA). Products from the PCR were cloned into pCRII (Invitrogen, Carlsbad, CA), and recombinant plasmids were transformed in Escherichia coli INVαF′-competent cells. Plasmids were isolated and quantified by optical measurements at 260 nm. The PCR product sequences were determined (Auburn University Genomics and Sequencing Laboratory), and sequences were verified using GenBank.
Temperature gradients were performed for all primer sets to determine optimum annealing temperature and primer quality for duplicate blank, standards, and reverse transcriptase product for each temperature. In addition, standards or PCR products were run in 10-fold dilution series to determine an optimum range for the standard curves. Dilutions of plasmids with the gene of interest were used as standards. Sample dilutions were linear across the range of the assay.
Real-time PCR was performed using a BioRad MyIQ thermocycler with a total reaction volume of 30 μL (15 μL of SYBR Green, 0.6 μL of each primer, and 13.8 μL of reverse transcriptase product plus water) as described previously (Whitlock et al., 2005; Elsasser et al., 2007). A denaturation cycle of 95°C for 3 min was run followed by amplification and quantification at 95°C for 30 s, 55°C for 1 min, 72°C for 30 s, and repeated for 40 cycles, except AgRP, which was at 59°C and 50 cycles. This was followed by 95°C for 1 min and 55°C for 1 min for 1 cycle each. Melt curves were performed as follows: 55°C for 10 s for 100 cycles with a temperature increase of 0.4°C each cycle. Primer efficiencies were greater than 88.5%. Primers for the 6 housekeeping genes were evaluated for use by the geNorm program (Ghent University Hospital Center for Molecular Genetics, Ghent, Belgium), and cyclophilin A, β-actin, and acidic ribosomal phosphoprotein were found to be acceptable primer pairs for use as a control and were all used in the calculations. Negative controls lacking cDNA or reverse transcriptase were also utilized. Results for AgRP were analyzed using the BioRad Excel gene expression macro and the ΔΔCT data graphed and utilized for statistical analysis.
Exp. 3. Dual-Label Immunohistochemistry
Sheep (randomly placed in treatment groups) were injected with either saline or LPS (0.6 μg/kg of BW) at time 0 and then once again at 4 h. At 6 h, the animals were killed as described in Exp. 2, and their hypothalami were fixed by phosphate-buffered formalin infusion in situ (Wagner et al., 2004). Hypothalami were removed and postfixed in fresh 4% paraformaldehyde with addition of 2.5% acrolein (Polysciences Inc., Warrington, PA) for 24 h at 4°C and sectioned on a freezing microtome at 40-μm intervals and stored in a cryoprotectant solution until immunohistochemistry could be performed on free-floating sections. In the first reaction, c-Fos was detected using a rabbit polyclonal antibody at a dilution of 1:80,000 (Ab-5; Oncogene Research Products, Cambridge, MA), and the reaction was visualized with the chromogen diaminobenzedine with the addition of nickel sulfate to produce a black precipitate in cell nuclei. In the second reaction, either AgRP was detected using a rabbit polyclonal antibody at a dilution of 1:10,000 or α-MSH was detected using a rabbit polyclonal antibody at a dilution of 1:30,000 (Phoenix Pharmaceuticals Inc., Belmont, CA), and the reactions were visualized with the chromogen diaminobenzedine alone to produce a brown precipitate in cell cytoplasm.
Exp. 4. Single-Label In Situ Hybridization
Sheep were randomly placed in 2 groups and treated with saline or LPS (0.6 μg/kg of BW), and sheep were killed at 6 h after treatment as described in Exp. 2. Hypothalami were removed and frozen in liquid nitrogen. Single-label in situ hybridization to identify cells expressing AgRP and POMC mRNA in sheep brains was performed as described previously (Marks et al., 1992). Coronal sections (20 μm) were cut on a cryostat and thaw-mounted onto Superfrost Plus slides (VWR Scientific, West Chester, PA). Hypothalamic sections were collected in a 1:4 series from the diagonal band of Broca caudally through the mammillary bodies. In the first study, antisense 33P-labeled sheep AgRP riboprobe (corresponding to bases 18 to 256; GenBank accession number AY310396) was denatured, dissolved in hybridization buffer at a concentration of 0.045 pmol/mL along with transfer RNA (1.7 mg/mL), and applied to slides. In the second study, antisense 33P-labeled sheep POMC riboprobe (corresponding to bases 66 to 392; GenBank accession number NM_001009266) was denatured, dissolved in hybridization buffer at a concentration of 0.05 pmol/mL along with transfer RNA (1.7 mg/mL), and applied to slides. Slides were covered with glass coverslips, placed in a humid chamber, and incubated overnight at 55°C. The following day, slides were treated with RNase A and washed under conditions of increasing stringency. The tissue was dehydrated in 100% ethanol, air-dried, and then dipped in NTB-2 liquid emulsion (Eastman Kodak Co., Rochester, NY) diluted 1:1 with distilled water. Slides were developed 5 d later and coverslipped.
Single-Label In Situ Hybridization Image Analysis.
All slides were assigned a random 3-letter code, alphabetized, and read unilaterally under dark-field illumination with custom software designed to count the total number of cells and the number of silver grains (corresponding to radiolabeled AgRP mRNA) over each cell. Cells with a signal-to-background ratio of at least 2 or greater were considered to express AgRP (first study) or POMC (second study) mRNA. Data are expressed as the total number of identifiable cells and grains per cell (a semiquantitative index of mRNA content/cell).
Exp. 5. Effects of Fasting on Recovery from LPS Appetite Suppression
This experiment was designed to physiologically increase endogenous orexigenic neurotransmitters and thus prevent LPS suppression of feed intake as observed with AgRP and neuropeptide Y (NPY) treatments. Six sheep were placed in an indoor facility as described in Exp. 1. To prevent sheep from becoming programmed to a feeding schedule, sheep were fed at different times of the day and at an amount of 3.5% of BW and provided water ad libitum. Sheep were randomly divided into 4 treatment groups (fed/saline, fasted/saline, fed/LPS, and fasted/LPS), and treatments were applied weekly such that each treatment was represented each week, each animal received each treatment, and the LPS was not given to any animal in consecutive weeks. In this protocol, fed sheep were provided feed at 1800 h on the day before an experiment and at 0600 h on the day of an experiment. Experiments began at 0800 h. Fasted sheep were provided feed at 1100 h the day before an experiment, and any remaining feed was removed at 1400 h. Jugular cannulas were inserted into each sheep the day before an experiment. On the day of the experiment, sheep were injected with either saline or with LPS (0.6 μg/kg of BW) at 0800 h. Feed (2.26 kg) was provided to all sheep 1 h after the saline or LPS injection (to allow the LPS time to affect feed intake). Feed not consumed was weighed at 2, 4, 6, 8, 10, and 24 h after saline and LPS injections, and feed consumption was calculated. Body temperature was sampled by rectal thermometer at the time of feed intake measurements.
Data for Exp. 1 were expressed as cumulative feed intake on a percentage of BW basis. Data were tested for effect of treatment (LPS, AgRP, saline or LPS and AgRP), time (1, 2, 4, 6, 8, 12, 24, or 48 h after treatment), and interactions on cumulative food intake and body temperature using a univariate split-plot model approach for repeated measures with JMP software (SAS Institute, Cary, NC). Subject within treatment was used as the error term for treatment. For mRNA and neurotransmitter immunohistochemistry, data were tested for effect of treatment by ANOVA. Comparisons of means for significant effects were performed using Student’s t-test. Data from Exp. 5 were analyzed using the same procedures as for Exp. 1, and specific differences among means were determined by Tukey’s honestly significant differences test. Values of P < 0.05 were considered significant.
Measurement of body temperature indicated that temperatures were increased in LPS-treated and AgRP + LPS-treated sheep compared with controls (P < 0.001), but there was no difference between the 2 groups (Figure 1). Interestingly, AgRP alone decreased body temperature at 6, 8, and 12 h after AgRP injection compared with saline-injected controls. There was a significant effect of treatment (P < 0.03), time (P < 0.001), and treatment × time interaction (P < 0.01) on cumulative feed intake (expressed as a percentage of BW to correct for variations in animal size effects on feed intake; Figure 2). Treatment with LPS resulted in decreased cumulative feed intake at 12, 24, and 48 h relative to sheep treated with AgRP + LPS (P < 0.05) and relative to sheep treated with saline or AgRP (P < 0.05). Feed intake for sheep treated with AgRP + LPS did not differ from sheep treated with saline alone or AgRP alone. Additionally, sheep treated with LPS had lower feed intake at 8 h posttreatment than sheep treated with AgRP (P < 0.05). Treatment with AgRP did not increase feed intake above controls.
To determine whether AgRP neurons were inhibited by LPS, AgRP mRNA was assessed at 3 and 6 h after LPS or saline injection. The expression of AgRP mRNA did not change at 3 h but was elevated in LPS compared with control sheep at 6 h (P < 0.03; Figure 3). The evaluation of POMC mRNA levels was performed, but there were no differences. This was not unexpected due to the inclusion of whole hypothalami, including the pituitary stalk, thus providing multiple sources of POMC mRNA that could potentially be regulated independently.
Based on the gene expression data, dual-label immunohistochemistry was designed to determine whether AgRP or α-MSH protein expression was altered in the hypothalamic arcuate nucleus (ARC) 6 h after injection of LPS in sheep (Figure 4). The number of neurons expressing AgRP in the ARC was not affected 6 h after LPS injection, whereas the percentage of AgRP neurons coexpressing c-Fos was elevated by LPS (P < 0.04), indicating an elevation in AgRP neuron activation (Figure 5). Similarly, the number of neurons positive for α-MSH did not differ with treatment, but the percentage of α-MSH neurons expressing c-Fos was decreased in LPS-treated sheep (P < 0.001), which indicates decreased α-MSH neuron activation 6 h after LPS (Figure 5).
Analysis of AgRP gene expression in the ARC by in situ hybridization revealed that the number of neurons positive for AgRP did not differ between groups (Figure 6). The number of grains per cell for AgRP was increased (P < 0.05) in the LPS-treated sheep (Figure 5), indicating an increase in AgRP gene expression (and in agreement with the real-time PCR data in Exp. 2). In addition, the number of POMC-expressing neurons in the ARC was not altered by treatment; however, POMC gene expression (based on the number of grains per cell) was decreased (P < 0.01) in the ARC of LPS-treated compared with saline-treated sheep (Figure 5).
In a study to physiologically increase orexigenic neurotransmitters, there was an effect of fed state (P < 0.01), LPS treatment (P < 0.001), time (P < 0.001), fed state × time interaction (P < 0.001), and LPS treatment × time interaction (P < 0.001) on body temperature (Figure 7). Mean body temperature was greater in fed-LPS than in fasted-LPS sheep. In contrast, mean temperature did not differ between fed-saline and fasted-saline sheep, and both had lower body temperatures than LPS-treated sheep. The LPS-treated sheep had greater body temperatures at 2, 4, and 6 h than saline-treated sheep at the same time periods (P < 0.05). Feed intake differences were also determined. There was an effect of nutritional state (P < 0.001), LPS treatment (P = 0.001), time (P < 0.001), nutritional state × time interaction (P < 0.002), and LPS treatment × time interaction (P < 0.0103). Using Tukey’s honestly significant differences test, saline-treated sheep that had previously been fasted had the greatest feed intake (P < 0.05), LPS-fasted and saline-fed sheep did not differ from each other, and LPS-fed sheep had the least intake (P < 0.05; Figure 7).
Appetite regulation is less studied in sheep than in other models, and although POMC neuron function has been a major focus of studies of LPS action, few studies have examined the effects of endotoxin on AgRP. In general, neural control of the appetite process in sheep follows the pattern of that seen in other species. For example, appetite neurotransmitter gene expression has the general patterns after fasting that would be expected (Henry et al., 2001a; Adam et al., 2002; Chaillou et al., 2002). Moreover, injection of selected neurotransmitters, orexin, AgRP, NPY, and melanin-concentrating hormone, into the lateral cerebroventricals of ad libitum-fed sheep resulted in an increase in food intake (Sartin et al., 2001; Wagner et al., 2004; Whitlock et al., 2005).
A consistent view has emerged from studies to date that supports a role for AgRP as an antagonist of the MC4R and leading to increased feed intake (Cone et al., 2001; Marks et al., 2001; Ramakrishnan et al., 2007). Moreover, studies in rodent models have demonstrated a role for α-MSH activation of the MC4R as a mechanism for disease-associated reductions in appetite (Huang et al., 1999; Marks et al., 2001; Sergeyev et al., 2001). Because studies in sheep have not supported an obligatory role for leptin or IL-1β in appetite suppression (McMahon et al., 1999; Daniel et al., 2002), the possibility of a central role for the MC4R in appetite control during disease in sheep was investigated. The intraventricular injection of AgRP before i.v. injection of LPS prevented the inhibition of feed intake in sheep, similar to the effects of a synthetic MC4R antagonist in rats (Huang et al., 1999). One possibility to consider is that the AgRP acted as a potent appetite stimulus and overwhelmed LPS actions to decrease feed intake. This is unlikely, because the dose of AgRP used did not increase feed intake in this study and was 25% of the dosage needed to increase feed intake in sheep (Wagner et al., 2004). Moreover, Obese et al. (2007) studied a potent feed intake stimulus, syndaphylin, and were unable to overcome the inhibition of feed intake induced by LPS. Thus, the AgRP effects in this study are likely mediated by actions on MC4R. The reduction in feed intake due to LPS is caused by α-MSH activation of MC4R (Huang et al., 1999; Marks and Cone, 2001; Scarlett et al., 2007). Because AgRP is an antagonist to the MC4R, these data indicate that blocking the MC4R prevents the inhibition of feed intake by LPS in sheep. These data support the hypothesis that the MC4R is a critical site for appetite suppression by LPS and is a potential target for pharmacologic intervention in disease suppression of appetite (Markison et al., 2005).
Gene expression of POMC is reported to be increased early in the onset of appetite suppression in rodents (Sergeyev et al., 2001), and IL-1β receptors are present on the POMC neuron that release α-MSH (Scarlett et al., 2007), all suggestive that activation of the MC4R is central to appetite suppression. Other studies in mice demonstrated that c-Fos expression was inhibited at 4 h but not 6 h in the ARC and at 6 h in the lateral hypothalamus, coinciding with the beginning of decreased feed intake at 6 h for this species (Becskei et al., 2008). However, we cannot directly compare intraperitoneal injections of high-dose LPS in rodents with the low-dose i.v. injections of LPS in sheep. Indeed, a more rapid onset and recovery from appetite inhibition would be expected in the sheep. Therefore, AgRP gene expression was examined at 3 h and at 6 h after LPS. The level of AgRP mRNA was unchanged at 3 h, suggesting that changes in POMC neurons are alone sufficient to decrease appetite. The levels of AgRP mRNA were increased at 6 h, and we have focused future studies on this time point. Both in situ hybridization and immunohistochemistry indicated an increase in gene expression for AgRP and increased activation of AgRP neurons at 6 h. In addition, there were fewer active α-MSH neurons and decreased expression of mRNA. Both the activation of the MC4R antagonist (AgRP) and the reduction of the MC4R agonist (α-MSH) should lead to improved appetite. Indeed, all LPS-treated sheep had begun to consume small amounts of feed by 8 h. Therefore, we hypothesized, based on the literature, that the suppression of appetite is caused by increased α-MSH activating the MC4R (Sergeyev et al., 2001; Scarlett et al., 2007), whereas the increased AgRP and decreased α-MSH in this study may represent an adaptive response to the LPS and the decreased feed intake. Elevated AgRP should block the MC4R and decreased α-MSH decrease the activation of the MC4R, which should provide a stimulus to feed intake as the LPS effects subside.
Infusion of pharmacologic levels of NPY or injection of AgRP in sheep (McMahon et al., 1999) and rodents (Huang et al., 1999) have been able to recover feed intake after LPS administration. In an effort to determine if this is a physiologically relevant response, we chose short-term fasting to elevate these critical appetite neurotransmitters. In short-term fasted sheep, increased AgRP expression is the primary appetite neurotransmitter response (McShane et al., 1993; Adam et al., 2002; Wagner et al., 2004), whereas there were no changes in gene expression for MC4R, melanin-concentrating hormone, or cocaine amphetamine-regulated transcript (Iqbal et al., 2001; Adam et al., 2002; Whitlock et al., 2005). In addition, the important POMC gene is minimally changed by fasting (McShane et al., 1993; Henty et al., 2001b; Adam et al., 2002). It should be noted that other factors, such as leptin and insulin, both appetite regulators, are also altered by fasting such that the effects of fasting are not limited to specific neurotransmitters (Altmann et al., 2006), although both hormones ultimately target these appetite neurotransmitters. Also, in rodent models, fasting did not modify the effects of LPS on hypothalamic cytokine gene expression (Gayle et al., 1999). Thus, IL-1β would still be present to activate α-MSH. Moreover, Li and Davidowa (2004) found that food restriction decreased the ability of melanocortins to activate the MC4R in the ventromedial nucleus of the hypothalamus, which was attributed to elevated AgRP expression. Therefore, we hypothesized that short-term fasting should elevate critical appetite stimulatory neurotransmitters (particularly AgRP) and should enhance recovery of appetite in sheep injected with LPS. Indeed, in this study, feed intake was normalized (to the level of fed controls) by fasting overnight. Unfortunately, feed intake in LPS-treated sheep was not restored to the level of the fasted-vehicle sheep, which suggested that fasting is not sufficient to completely prevent appetite suppression in sheep. The inability of fasting to completely restore appetite is also seen in mice fasted for 12 h before injection of LPS (Becskei et al., 2008). And in a study with IL-1β infusion in food-restricted rats (Mrosovsky et al., 1989), food intake was increased above controls during the first d 1 but on d 2 of IL-1 infusion were either slightly decreased (high-dose IL-1) or at the same intake level as controls (low IL-1 infusion). Nonetheless, fasted sheep given LPS consumed more feed than fed sheep given LPS, and this response is reminiscent of the normalization of feed intake in sheep treated with LPS and NPY (McMahon et al., 1999) or the effects of AgRP in this study. In addition to feed intake, body temperature was similarly decreased by AgRP and by fasting. These results confirm the hypothesis that a physiological elevation in orexigenic factors would have a positive effect on feed intake in this model, similar to our data with infusion of individual neurotransmitters (McMahon et al., 1999; Figure 2). Moreover, in agreement with Mrosovsky et al. (1989), the current data suggest that there is a resetting of appetite mechanisms depending on past nutritional history or perhaps a resetting of the threshold at which cytokines inhibit appetite mechanisms.
The data in this study provide evidence that AgRP treatment (antagonism of the MC4R) can prevent the effects of LPS to decrease feed intake. Moreover, as the appetite-suppressive effect of LPS progresses, there is a rise in AgRP and decreased α-MSH (leading to inactivation of the MC4R), which are changes consistent with increased appetite stimulation. These data suggest an important role for the MC4R in appetite control in disease in sheep. In view of the apparent role of AgRP, α-MSH, and the MC4R in the regulation of appetite, the development of MC4R antagonists capable of crossing the blood-brain barrier may have possible therapeutic value in ruminants to prevent or hasten the recovery from suppression of appetite (Sartin et al., 2005). In addition, because AgRP is a potent stimulus of appetite in sheep, these MC4R antagonists may be useful to activate appetite pathways in normal farm animals.