Peptide YY_(3–36) Inhibits Morning, but Not Evening, Food Intake and Decreases Body Weight in Rhesus Macaques

Adult male rhesus macaques (Macaca mulatta) were singly housed at the Oregon National Primate Research Center (ONPRC) in an environment controlled for humidity, temperature, and lighting (illuminated between 7:00 a.m. and 7:00 p.m.). Nine animals between 10 and 18 years old and between 8 and 23 kg body wt were studied. All procedures were approved by the ONPRC Institutional Animal Care and Use Committee. Experimental group size was five or more monkeys unless otherwise noted.

Animals were fed low-fat monkey diet (#5047, jumbo size; Purina, St. Louis, MO) twice daily. Generally, animals were fed 12 biscuits at 10:00 a.m. and again at 3:00 p.m. Biscuits weighed ∼16.5 g and had an energy density of 3.11 kcal/g (25% protein, 5% fat, 6.5% fiber, 6% ash, and 57.5% carbohydrate). A high-fat diet consisting of semi-solid cubes weighing ∼60 g with an energy density of 3.9 kcal/g (20% protein, 35% fat, and 45% carbohydrate) was provided as noted during specific experiments. Macronutrient composition of the high-fat diet was designed to mimic a typical North American diet (21). Drinking water was continuously available via pressurized drinking spouts.

Two chronically implanted catheters placed into separate subclavian, femoral, or jugular veins of each animal allowed for simultaneous infusion of PYY_(3–36) and collection of blood samples. Catheters remained in place for the duration of the experiments and were remotely accessible in an adjacent sampling room. Monkeys wore fitted nylon vests that were connected to 36-inch flexible metal tethers, which were fastened to swivels mounted to the tops of cages (22). This catheter protection system allowed for blood sampling and intravenous infusions without sedating or disturbing animals and allowed animals free range of movement within their cages at all times. Physiological saline drip (5 ml/h; Baxter Healthcare, Deerfield, IL) containing heparin sodium (4 units/ml) maintained catheter patency. Catheters were checked daily for patency, and animals were sedated weekly with Ketaset (ketamine hydrochloride, 10 mg/kg i.m.) to record body weight and inspect catheter systems.

Blood collections were performed remotely via intravenous catheters. Blood was drawn into syringes prerinsed with heparin (heparin sodium, 10,000 units/ml), transferred to test tubes containing peptidase inhibitor (aprotinin, 0.6 trypsin inhibiting units/ml blood; Sigma Chemical), and centrifuged at 1,600g for 15 min at 4°C. Plasma was transferred into polypropylene freezer vials and stored frozen at −80°C until assayed.

All samples were assayed in duplicate. Total PYY immunoreactivity was measured with a sensitive radioimmunoassay that detected both the cleaved form (PYY_3–36) and the full-length hormone (PYY_1–36); as described previously, all the results from our assays reflect total plasma PYY (15). Briefly, the assay was validated for monkey using serial dilutions of a pool of samples of high PYY_(3–36) values and performed in a total volume of 350 μl of 0.06 mol/l phosphate buffer, pH 7.3, containing 0.3% BSA. Samples were incubated for 3 days at 4°C before separation of free and antibody-bound label by goat anti-rabbit IgG serum. One hundred microliters of unextracted heparinized plasma was assayed. The lowest detectable level was 3.2 pg/tube. Intra-assay variation was determined at concentrations of 100 and 400 pg/ml and was 6.3 and 6.6%, respectively. Interassay variation was 7.2 and 8.1% for the range of values measured.

Several different sampling schedules were used to determine endogenous release of total PYY [both PYY_(1–36) and PYY_(3–36)]. For one study, samples were drawn hourly before, during, and after animals ate normal meals of low-fat diet (n = 4), high-fat diet (n = 2), or after fasting (n = 2). In a subsequent study, blood samples were taken 15, 30, 45, and 60 min after a 3:00 p.m. meal and every 30 min thereafter until 9:00 p.m., and two baseline blood samples were taken 30 and 60 min before feeding of a low-fat (n = 6) or high-fat (n = 6) meal. A final study involved blood sampling either at baseline (3:00 p.m. after a morning fast) or at 8:00 p.m. after ad libitum consumption of high-fat diet (n = 6), low-fat diet (n = 6), or fasting (n = 6).

Synthetic human PYY_(3–36) was obtained from Phoenix Pharmaceuticals (Belmont, CA). The (3–36) fragment of PYY was used for all infusion tests. Vehicle for PYY_(3–36) infusions was physiological saline containing 1% human serum albumin (Albuminar-25; Aventis Behring, Kankakee, IL).

Preliminary infusion studies established appropriate doses for approximating normal total PYY in fasted animals. Three separate 60-min infusions of PYY_(3–36) in ascending concentrations of 0.8, 2.4, and 4.8 pmol · kg⁻¹ · min⁻¹ were separated by 30-min saline washout periods (n = 3). A second set of test infusions with lower doses (0.2, 0.3, and 0.4 pmol · kg⁻¹ · min⁻¹) allowed accurate approximation of total postprandial PYY (n = 3). Animals were closely monitored to detect any adverse reactions to PYY_(3–36) dosing. Intermittent blood sampling determined the achieved circulating total PYY levels and guided the selection of doses for initial PYY_(3–36) infusion experiments.

Three dosing protocols were used: once-daily, twice-daily, and continuous infusion (see Fig. 1). The once-daily dosing protocol (n = 5–6) was used to test the hypothesis that elevation of circulating PYY_(3–36) levels before and during the morning meal decreases food intake and body weight. Seven days of vehicle infusion were followed by 7 consecutive days of PYY_(3–36) infusion at 0.3 pmol · kg⁻¹ · min⁻¹ (because this dose mimics total postprandial PYY levels, as determined by assay as described above) and 2 days at 0.8 pmol · kg⁻¹ · min⁻¹. Vehicle and low-dose (0.3 pmol · kg⁻¹ · min⁻¹) PYY_(3–36) infusions began at 7:00 a.m. and ended at 8:30 a.m., and the higher PYY_(3–36) dose (0.8 pmol · kg⁻¹ · min⁻¹) began at 8:30 a.m. and ended at 10:30 a.m. Animals were fed at 10:00 a.m. throughout the duration of the study. Detailed feeding pattern measurements were made on the last 2 days of vehicle infusion and the first 2 days of 0.3 pmol · kg⁻¹ · min⁻¹ and 0.8 pmol · kg⁻¹ · min⁻¹ PYY_(3–36) infusions. Total 24-h food intake was measured daily. Two blood samples were taken 15 min before and 15 min after the start of each infusion with the exception of the 1st day of the low-dose PYY_(3–36) infusion, when additional blood samples were taken at 30, 60, 90, 120, and 240 min after the infusion. During the high-dose PYY_(3–36) infusions, blood samples were taken 15 min before and 20 min after meal presentation.

The second set of experiments tested the hypothesis that daily PYY_(3–36) infusion before and during morning and evening meal onset suppresses food intake and leads to body weight loss (n = 5–6). Infusions occurred between 8:30 and 10:30 a.m. and 1:30 and 3:30 p.m., daily. Seven days of vehicle infusions preceded and followed PYY_(3–36) infusions. PYY_(3–36) (0.8 pmol · kg⁻¹ · min⁻¹) was infused for 5 days, followed by 11 days of 1.6 pmol · kg⁻¹ · min⁻¹. Animals were weighed before and after the vehicle treatment week, after the 5th day of the low PYY_(3–36) dose (PYY1), and again after the higher PYY_(3–36) dose (PYY2). Total daily food intake was recorded throughout the twice-daily infusion protocol, and detailed feeding patterns were recorded on 7th day of initial vehicle infusion, the 1st day of PYY1, days 1 and 5 of PYY2, and days 1 and 7 of the after treatment vehicle period. Blood samples were taken 15 min before and after meals and 30 min after the end of PYY_(3–36) infusions.

Three days of PYY_(3–36) treatment (1.6 pmol · kg⁻¹ · min⁻¹; n = 3) were preceded and followed by 3 days of continuous vehicle infusion. Detailed measurements of feeding patterns were made on each day of treatment. Blood samples were drawn 15 min before both the morning and evening meals to assess total plasma PYY.

Animals (n = 5) were videotaped for 6 hours on a single day of vehicle and on a single day of PYY_(3–36) infusion (1.6 pmol · kg⁻¹ · min⁻¹ for 120 min) during the twice-daily infusion protocol to assess effects of PYY_(3–36) on general behavior and to determine whether appetite-suppressing effects of PYY_(3–36) were linked to sickness or malaise. Video cameras were placed on tripods in front of cages for several days before treatment to accustom animals to the presence of recording equipment. Behavioral analysis of focal observations was performed with a videotape player that was synchronized with a computer via The Observer 3.0 Software (Noldus Technologies). An observer blind to treatment scored the frequencies and durations of a battery of behaviors chosen to maximize detection of illness. Scored behaviors included drinking, eating food, touching food, standing, lying on side of body, locomotion, grooming, yawning, napping, salivation, vomiting, and stereotypical behavior. Frequency data are presented as mean occurrences per 6-h observation period ± SEM. Durations are presented in cumulative seconds of duration ± SEM during the 6-h observation period.

Intravenous glucose tolerance tests (IVGTTs) (n = 6) were performed by measuring blood glucose clearance after an intravenous bolus infusion of a sterile 50% dextrose solution (600 mg/kg). IVGTTs were performed before any vehicle testing, after 1 week of vehicle infusions, and again after twice-daily PYY_(3–36) treatment. Animals were not fed on the morning of IVGTTs. Blood glucose was measured immediately in whole blood with an Accu-chek Advantage blood glucose meter (Roche Diagnostics). Plasma was assayed for insulin using a Roche Diagnostics Elecysis 2010 analytical instrument by the ONPRC/OHSU Endocrine Services Laboratory. The interassay percentage of coefficient of variation for this assay was 3.5%. A subset of plasma samples from the IVGTT (n = 3) was assayed for total plasma PYY levels.

Statistical Software package SPSS 11.5 for Windows (SPSS) was used for all statistical calculations to evaluate the effect of PYY_(3–36) dose on feeding and body weight. To allow for combined analysis of multiple treatment days, doses, and time points in the same animals, the SPSS Linear Mixed Model was used for analyses unless specified otherwise. This model is often used for unbalanced datasets with nonuniform design and/or missing subjects because it estimates the marginal means for each dose and treatment regimen (we derived the degrees of freedom from these numbers). Like treatments (doses) from different test days and different times were collapsed within the model and used in the analysis to generate statistics for the overall effect of dose on the appropriate dependent variable. We did not collapse data across treatment days because of the small sample size inherent in monkey studies and the use of a linear mixed model. This increases the degrees of freedom but captures other variability inherent in the test protocols, and we believe it is the best method of analysis for this data. There was not enough statistical power to draw meaningful conclusions about the day-to-day repeatability and strength of the effect during the once- and twice-daily treatment protocols. Where appropriate, estimated marginal means generated by the statistical model were used to calculate an overall difference between doses for an experiment, and post hoc comparisons (Bonferroni adjusted) were used to compare effects between individual doses at specific time points. Body weight change during vehicle infusion was compared with body weight change during PYY_(3–36) treatment using paired t test. Graphs depict means of data ± 1 SEM. Data from days with the same treatment doses are displayed as means in graphs. The effect of PYY_(3–36) infusion on scored behaviors was statistically evaluated using paired sample t tests.

Total endogenous PYY ranged between 100 and 175 pg/ml and gradually declined overnight (Fig. 2A). Total PYY was positively correlated with intake of high-fat meal (r² = 0.71, P = 0.03); spontaneous intake of low-fat diet caused no significant change in total PYY; and fasting caused gradual decline with a nadir of ∼50 pg/ml overnight (Figs. 2A–C). The between-subjects variation in total PYY in response to spontaneous ingestion of low-fat diet ranged between 75 and 225 pg/ml (Fig. 2C).

Total PYY after test infusions (0.2, 0.3, and 0.4 pmol · kg⁻¹ · min⁻¹) was about one to three times greater than circulating endogenous postprandial levels (Fig. 3A). Infusions of 0.8, 2.4, and 4.8 pmol · kg⁻¹ · min⁻¹ produced peak total PYY ranging from 500 pg/ml to >5,000 pg/ml (assay upper limit, Fig. 3B). Careful observation of animals during test infusions did not reveal any indication of illness or altered behavior (see detailed behavioral analysis; Table 1). An assay for plasma cortisol also revealed no changes in cortisol concentrations during PYY_(3–36) infusions.

PYY_(3–36) infusion before and during the morning meal at doses that approximated endogenous postprandial levels, and at supraphysiological levels, suppressed acute food intake during that meal. Once-daily PYY_(3–36) infusion dose-dependently suppressed the rate of eating (Fig. 4A; dose by time interaction; F_{(14, 241.986)} = 2.255, P = 0.007), but total consumption for the morning meal was not affected (P ≥ 0.369). Morning PYY_(3–36) infusion did not affect rate of eating or total food consumption during the evening meal (Fig. 4B; dose-by-time interaction; F_{(8, 212.110)} = 0.721, P = 0.665) and had no effect of on total 24-h food intake (Fig. 4C; tests for effect of dose based on estimated marginal means: F_{(2, 26.218)} = 0.843, P = 0.442). The time course for total daily food intake is shown in Fig. 4D.

Infusions of PYY_(3–36) that increased total PYY levels two to five times physiological levels before and during the morning (Fig. 5A) and evening meals (Fig. 5B) reduced the initial rate of eating only during the morning meal (Fig. 5A; dose-by-time interaction, morning meal: F_{(21, 481)} = 22.642 P < 0.001) and produced a mild suppression in 24-h food intake that was not significant (F_{(2, 56.086)} = 2.997, P = 0.058; Fig. 5C). Multiple pairwise comparisons of the effect of PYY_(3–36) at individual time points during the morning meal showed significant reductions in intake with the 1.6 pmol · kg⁻¹ · min⁻¹ dose at 11:00 a.m. but not at the end of the morning meal or during the evening meal (Figs. 5A and B). Post hoc comparison of 1.6 pmol · kg⁻¹ · min⁻¹ versus vehicle doses revealed marginal significance (P = 0.052). Daily food intake was generally inhibited during the entire twice-daily PYY_(3–36) infusion period (Fig. 5D). There may be diminished effectiveness of the drug at the end of the PYY_(3–36) infusion period, although we did not have the statistical power to analyze the data for day to day differences in effect of PYY_(3–36).

The twice-daily PYY_(3–36) infusion caused a 1.9 ± 0.7% body wt loss (Fig. 5E; P = 0.017) that persisted through the second PYY_(3–36) dosing period (Fig. 5F). There was no additional weight loss, and body weight returned to baseline after cessation of treatment.

Analysis of videotapes revealed no nonspecific effects of the 1.6 pmol · kg⁻¹ · min⁻¹ PYY_(3–36) infusion on behavior (Table 1). There was a significant increase in the frequency (P = 0.028), but not duration, of touching biscuits. There were no statistical differences in grooming, locomotion, or napping and in no cases did animals display salivation or drooling, nor did animals exhibit general malaise. One animal appeared to regurgitate into its cheek pouches several times during the PYY_(3–36) infusion experiment (such regurgitation is a normal occurrence in some animals), but no overt vomiting was observed. No regurgitation or vomiting was observed in any other animals.

Repeated twice-daily PYY_(3–36) infusion had no effect on glucose clearance or insulin response profiles during IVGTTs. There was no effect of supraphysiological elevation of blood glucose levels during IVGTT on total circulating PYY (data not shown).

Continuous 24-h PYY_(3–36) infusion reduced food intake during the morning but not the evening meal (Figs. 6A and B; F_{(1, 85.772)} = 18.356, P < 0.001, F_{(1, 86.183)} = 1.077, P = 0.302, respectively; n = 3). Multiple comparisons of the effect of PYY_(3–36) dose on intake at specific time points during the morning meal revealed significant differences at 11:00, 12:00, and 2:30 p.m. (Fig. 6A). Continuous PYY_(3–36) infusion suppressed morning but not evening food intake (Fig. 6C). The decrease in morning food intake caused by PYY_(3–36) infusion also decreased 24-h food intake in all animals (Fig. 6D; F_{(1, 22)} = 9.746, P = 0.005). Continuous PYY_(3–36) infusion appeared to inhibit food intake each day of PYY_(3–36) infusion, but because of the small sample size, we were unable to make comparisons across time (Fig. 6E).

Total immunoreactive plasma PYY increased during PYY_(3–36) infusion. Furthermore, total PYY rose in proportion to increasing PYY_(3–36) infusion doses. At the end of the 0.3 pmol · kg⁻¹ · min⁻¹ s.i.d. infusions, average total PYY reached levels of ∼250 pg/ml—close to maximum observed endogenous levels (Fig. 7A). Total PYY levels increased to ∼450 pg/ml after 0.8 pmol · kg⁻¹ · min⁻¹ s.i.d. infusions. Total PYY during meal time in the twice-daily infusion protocol are shown in Fig. 7B. Because of an infusion problem, the 1st day of twice-daily PYY_(3–36) dosing was 0.6 pmol · kg⁻¹ · min⁻¹ (instead of 0.8 pmol · kg⁻¹ · min⁻¹), and this was reflected in total PYY levels_. Total PYY achieved after infusion of 0.8 pmol · kg⁻¹ · min⁻¹ was ∼500–800 pg/ml and increased to between 1,000 and 2,000 pg/ml for the 1.6 pmol · kg⁻¹ · min⁻¹ b.i.d. infusions. Total PYY decreased to about one-half of peak values 30 min after the infusions ended and quickly returned to baseline. Continuous infusion of PYY_(3–36) at 1.6 pmol · kg⁻¹ · min⁻¹ caused total PYY levels to peak in excess of 2,000 pg/ml (Fig. 7C). This result indicates that our assay detected increases in total PYY that were likely due to elevated plasma PYY_(3–36) from infusion, and our behavioral data are also primarily caused by action of PYY_(3–36).

This study sheds light on previous challenges in reproducing findings that PYY_(3–36) significantly decreases food intake (17–19). The data presented here demonstrate that, in this model, low doses of PYY_(3–36) that mimic physiological levels do not significantly reduce food intake—only sustained supraphysiological PYY levels effectively decreased morning but not evening food intake regardless of infusion protocol. This is in contrast to the long-lasting effect of PYY_(3–36) infusion in humans consuming a buffet meal in the middle of the day, who reported reduced appetite long after infusion cessation (15). In the case of once-daily infusions, the lack of an inhibitory effect on evening food intake is not surprising because elevated total morning PYY quickly returns to baseline after infusion. However, the finding that twice-daily and continuous infusions were only effective at suppressing morning meal intake is perplexing and requires further analysis.

One explanation is that decreased food intake in the morning increased hunger during the evening meal, superseding the minor suppressive effect of PYY_(3–36). However, this idea does not account for the fact that animals in the continuous infusion protocol consistently reduced morning intake but did not compensate for the caloric deficiency by increasing their evening meal intake. Alternatively, Y2 receptors might desensitize after activation by PYY_(3–36) in the morning, attenuating the effect of equal PYY_(3–36) concentrations in the evening. However, this possibility is also inconsistent with our results from the continuous PYY_(3–36) infusion protocol, in which animals consistently decreased their morning food intake, despite high total PYY throughout the entire day and night. It is also possible that priming by other signals (like cortisol) regulates sensitivity to PYY_(3–36). Because circulating cortisol is higher in the morning, this hormone might prime the ARH or other brain areas to the anorexigenic effects of PYY_(3–36), although much more extensive studies would be needed to explore this possibility. We feel that a fourth, and very likely, explanation is that because monkeys generally show a greater orexigenic drive in the afternoon compared with the morning (23), detection of a small anorexigenic effect in the afternoon is more difficult. Therefore, a mild anorexic signal, such as PYY_(3–36), would be more easily detected in the morning when orexigenic drive is lower.

In addition to satiety, stress and illness are powerful regulators of food intake. We performed a radioimmunoassay for plasma cortisol (enhanced release of this hormone is associated with heightened stress) and found no change in plasma cortisol levels between vehicle and PYY_(3–36) infusions that ranged from 0.2 to 4.8 pmol · kg⁻¹ · min⁻¹ (data not shown). So although it is not likely that PYY_(3–36) reduces food intake by increasing stress, PYY_(3–36) might decrease appetite by inducing illness. For example, PYY_(3–36) could reduce food intake by activating chemosensitive hindbrain regions that mediate emesis (Y2 receptors are present in the rodent hindbrain [24,25]), although no direct evidence supports this claim. Our behavioral analysis showed that increased frequency of touching food after PYY_(3–36) treatment correlated with a nonsignificant decrease in the duration of eating. Although these data suggest an interest in food that is not actualized by consumption, the possibility that this behavior is an indication of simultaneous hunger and illness is inconsistent with the notable absence of other behavioral indicators of illness. Formal behavioral analysis for several indicators of sickness in monkeys found no differences between drug and vehicle in all monkeys but one, even when total PYY was 20 times normal physiological levels. More notably, clinical studies report no sensations of illness in humans after physiologically relevant PYY_(3–36) doses that inhibited 24-h food intake by 33% (15), indicating that illness is not a likely rationale for decreased appetite caused by PYY_(3–36) infusion.

Although we are unable to make statistical comparisons on the day-to-day efficacy of PYY_(3–36) to inhibit food intake, it generally inhibited food intake every day. The effect of PYY_(3–36) on food intake may have diminished toward the end of the infusion, and further longer-term studies will be needed to test this. There is tremendous variability in the effects of any drug on primate ingestive behavior; seemingly trivial changes in their environment can cause monkeys to alter their normal behavior. Therefore, small deviations from a general effect should not be overinterpreted. However, the weight loss persisted even if the effect on food intake faded, as is seen with many weight loss pharmacotherapies.

Two weeks of twice-daily PYY_(3–36) dosing had no effect on glucose clearance or insulin response to intravenous glucose during acute tests (results not shown). Only after spontaneous intake of a high-fat meal did total plasma PYY increase in proportion to calories. These results are consistent with the proposed release of PYY in response to ingested fat (26,27). If PYY release is mediated by an enteral, nutrient-dependent mechanism (26,28,29), it is possible that the effects of PYY_(3–36) infusion on appetite and body weight may be greater in animals consuming a high-fat diet.

The results presented here demonstrate that PYY_(3–36) transiently suppresses food intake and reduces body weight in monkeys, suggesting that targeting the PYY/Y2 system in humans may yield similar positive results. Because no physiological PYY_(3–36) doses significantly reduced total meal consumption, it is unlikely that PYY_(3–36) is a primary regulator of feeding behavior. Rather, our data suggest that postprandial PYY_(3–36) is part of an anthology of neurohormonal signals that regulate appetite. Pharmacological activation of the NPY/Y2 receptor will probably be more efficacious in individuals with deficiencies in endogenous Y2 ligands [improper PYY_(1–36) release or impaired enzymatic cleavage to PYY_(3–36)] or obese humans with lower PYY tone (14). This type of therapy would be in accordance with the precedent set by the high efficacy of leptin replacement therapy for weight loss in leptin-deficient humans (30,31). Nevertheless, our data do not completely dismiss PYY_(3–36)-based therapies. NPY-Y2 receptor activation may enhance the efficacy of other weight-loss treatments in overweight/obese patients with normally functioning PYY/Y2 systems and in patients who consume high-fat foods.

M.A.C. has received support from National Institutes of Health (NIH) Grant 5 P51-RR-00163, DK-62202. K.L.G. has received support from NIH Grant DK-060685. J.L.C. and F.H.K. have received support from NIH Grant DK-55819. P.J.E. has received support from Fogarty International Center Grant TW/HD-00668 (to P. Michael Conn).