INTRODUCTION

Top Abstract Introduction Methods Results Discussion Appendix References

IN ANIMAL STUDIES, satiety (a feeling of fullness or the desire to stop eating) can only be measured indirectly by behavioral outcomes such as the reduction or suppression of food intakes during experimental manipulation. In previous studies of intestinally perfused rats (24, 26), we confirmed and extended a variety of observations which indicated that intestinal sensors for specific nutrients signalled reductions of food intake. We found that such sensors were arrayed along the entire intestine, and several of our observations suggested 1) that satiety responses to nutrient-triggered signals increased sharply as the length of intestinal contact with nutrients extended; 2) studies in dogs and humans (21, 23) have indicated that free or emulsified dietary oils empty from the stomach initially at rates proportional to amounts ingested; and 3) that length of gut contacted by their lipolytic products, in turn, may vary with rates of duodenal entry of the fats (22). If these three ideas (ideas 1-3) are correct, then rats should progressively diminish intakes of food as increasing amounts of oil are ingested because satiety-triggering lipolytic products are spread farther along the gut with increasing intakes.

However, almost nothing is known about interrelationships between nutrient load, length of gut contacted by nutrients (or digestive products of nutrients), and reductions of food intake. The present experiments were undertaken in rats 1) to confirm that oil empties from the stomach at rates that vary with oil loads; 2) through the use of a potent lipase inhibitor (orlistat) to determine whether satiety was induced only by lipolytic products and not triglycerides; 3) to ascertain that increasing rates of oil entry into duodenum extend the length of gut contacted by high concentrations of lipolytic products; and 4) to judge whether length of gut contacted correlated in timing and magnitude with dose-responsive reductions of intake after intragastric instillations of oil. 5) Finally, with a variety of antagonists, we attempted to define how reductions were signalled by gut sensors sensitive to lipid, that is, whether signals were neurally transmitted or mediated by specific hormonal agents.

	METHODS

Top Abstract Introduction Methods Results Discussion Appendix References

Male Sprague-Dawley rats (300 g) were used throughout. Rats were kept in metabolic cages in a room with a natural light cycle and were trained to eat powdered rat chow (Richmond Standard, 5001 Laboratory Rodent Diet, 3.4 kcal/g; PMI Feeds, St. Louis, MO) from 0900-1600 daily. At 1600, food was removed until the next day's feeding. Water was provided ad libitum throughout the 24 h. The rats were also trained to accept brief orogastric intubation just before feeding. Formal studies began only after the animal's daily intakes became steady at 14 g or above despite these manipulations.

Animals were given intragastric instillates or "premeals" of constant volume. With premeals of corn oil (Mazola oil; Best Foods, Englewood Cliffs, NJ; specific gravity = 0.92; nutritional value = 8.6 kal/ml) or medium-chain triglyceride (MCT) oil (specific gravity = 0.93; nutritional value = 8.3 kcal/g; Mead Johnson, Evansville, IN), the volume of the premeals was always 4 ml made up of 0.15 M saline alone (control), saline plus 0.5, 1, or 2 ml of oil, or 4 ml (alone) of oil. The chosen volume of oil was injected first from one syringe, and then the remaining saline was injected through the gastric tube from a second syringe. To provide some contrasts to lipid, we also studied premeals of carbohydrate made from 73% (wt/wt) Polycose in water (nutritional value = 3.83 kcal/g of dry powder; Ross Laboratories, Columbus, OH), a mixture of glucose polymers. With Polycose, the premeals consisted of 6 ml of saline alone (control), 3 ml of saline plus 3 ml of Polycose, or 6 ml of Polycose alone. The 6 ml of Polycose (16.7 kcal) were approximately isocaloric with 2 ml of oil (16.0 kcal), but the 73% Polycose was hypertonic (1,100 mosmol/kgH₂O), whereas the oils were hypotonic.

We measured food consumed each hour during the 7-h feeding period. The feeding bowls were removed every hour until the last hour for <5 min, while spilled food was returned to the bowls, and the bowls were weighed. By following hourly food consumptions, we could estimate the magnitude and timing of reductions of food intakes after the various doses of nutrients when compared with control instillations of saline. Details of these calculations are given in the appropriate paragraphs in RESULTS.

To assess gastrointestinal transits, digestion, and absorption, we used [³H]glyceroltriether ([³H]GTE) as an indigestible, nonabsorbed fat marker, and [¹⁴C]triolein as a surrogate for the corn oil or long-chain triglycerides (LCT; see Ref. 22). A number of the oil premeals contained 0.6 µCi of [³H]GTE and 0.2 µCi of [¹⁴C]triolein. At 1, 2, 4, and 7 h after a 4-ml, radiolabeled oil premeal and the start of feeding, the rats were killed, and the gastrointestinal tract was immediately divided into five regions: stomach, proximal, middle, and distal thirds of small intestine, and colon (plus any feces expelled during the feeding period). We determined the net weight of gastric contents (food + secretions) and collected the liquid luminal contents from the small intestinal segments by flushing the segments with 30-40 ml of 10 mM sodium taurocholate. Solid gastric and colonic contents (plus expelled feces, if any) were separately collected, and these solid contents, as well as the washed wall segments of stomach and intestine, were liquefied by hydrolysis under reflux in boiling 10% KOH (vol/vol) in ethanol. The washed luminal contents and the acidified hydrolysates of solid samples were then extracted in 5:4 petroleum ether-ethanol, the ether evaporated, and the extracts were redissolved in liquid scintillation medium (Ecolume; ICN) for counting ³H and ¹⁴C. Corrections were made for quench and downscatter (22).

We used these counts for several different determinations. 1) The relative recoveries of ³H in each segment (luminal contents plus walls) at the four times of analyses were used as measures of transit over time. 2) Because the [³H]GTE was not absorbed but the [¹⁴C]triolein was digested and absorbed to varying degrees, the ratio of remaining ¹⁴C to ³H in luminal contents was used to indicate how much triolein remained unabsorbed by the time the meal traversed the segment (22). 3) To assess the degree of lipolysis, we divided samples in two to extract them at ambient pH (~6.5) and at pH >11. We had previously determined that all neutral lipids ([¹⁴C]triolein, [¹⁴C]monolein, [³H]GTE) were efficiently extracted from contents both at ambient pH and at pH 11, whereas 95% of [¹⁴C]oleic acid was extracted at ambient pH but 5% at pH 11 (22). The fraction of [¹⁴C]triolein remaining in luminal content that had been converted to [¹⁴C]oleate was so determined from these extracts at pH 6.5 and >11. These methods had been previously validated, and details of how we calculated outcomes are set forth in the APPENDIX. Nevertheless, [¹⁴C]triolein served as a surrogate for mixed LCTs that constituted corn oil and also as a surrogate for longer-chain MCTs. In previous studies (22), digestion and absorption of [¹⁴C]triolein was about two to three times faster than that of the mixed triglycerides in the corn oil, and once hydrolyzed, lipolytic products from the [¹⁴C]triolein were absorbed 1.6 times faster than products from corn oil. We found in the present study that fatty acids were released by pancreatic lipase from trilaurin three times faster and from MCT oil 15 times faster than from 95% pure triolein (Sigma Chemical).

To further determine where and how oil premeals signal satiety, we used four inhibitors that might interrupt such signals. The first was orlistat (formerly tetrahydrolipstatin; Hoffman LaRoche, Basel, Switzerland). This compound is a highly effective, nearly irreversible inhibitor of gastric and pancreatic lipases (21). Orlistat (800 mg) in 0.6 ml of ethanol was dissolved in 60 g of the corn oil from which portions were instilled as premeals. This concentration of orlistat in oil had previously been shown to provide >95% inhibition of pancreatic lipase (21). The second inhibitor was Pluronic acid L-81 (L-81), a hydrophobic detergent. This compound is known to block the export of absorbed fat from the enterocyte into lymph as chylomicra (30, 31). The L-81 was dissolved in the oil at a concentration of 1 g/36 g of oil from which portions were given as premeals. Because fatty acids in small intestine are potent releasers of cholecystokinin (CCK), we studied the effect of MK-329 (Merck), a potent and specific antagonist of CCK. The MK-329 was dissolved in dimethyl sulfoxide (DMSO) and then in sterile 0.15 M NaCl (50% as DMSO) and (or equal volumes of vehicle in control experiments) was given intraperitoneally 20 min before premeals and feeding. Finally, we used systemic capsaicin, a neurotoxin selective for unmyelinated, chemosensory nerves (10), to assess whether satiety signals were carried predominantly or exclusively by afferent nerves. Crystalline capsaicin (98% pure; Sigma) was dissolved completely with warming and stirring in a mixture of DMSO (10%), Tween 80 (10%), and 0.15 M NaCl (80%). The capsaicin was administered in three doses subcutaneously at two sites: at 0 h in a dose of 25 mg/kg and at 12 and again at 24 h in a dose of 50 mg/kg. The animals were anesthetized with halothane before every dose of capsaicin, and, before the first two doses, the rats were premedicated with intraperitoneal atropine (0.1 mg/kg). About 70% of animals fully recovered from the capsaicin treatment, and recovery was complete within about 3 days, by which time the animals were eating and behaving normally. Control animals were treated similarly using vehicle (DMSO + saline) in place of capsaicin.

We used two-way analyses of variance (ANOVAs) for unreplicated values to examine the effects of various treatments. The areas under the eating time courses (AUCs) were computed by trapezoidal rule and then normalized by taking the square roots (; see Refs. 21 and 23). Thus the two-way ANOVAs for were used to determine whether there was a significant treatment effect, and contrasts between treatments were used (after appropriate adjustments for numbers of comparisons) to determine whether food consumption after one treatment was significantly different from that after another. To determine whether there was a significant dose response, least-squares linear regressions of AUCs against dose were computed for each animal to determine individual slopes. We then computed the mean and SE of the slopes for all animals and determined by unpaired t-test whether the mean slope was significantly different from zero.

	RESULTS

Top Abstract Introduction Methods Results Discussion Appendix References

Character of Caloric Compensations

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Dose-responsive inhibitions of food intakes after premeals of corn oil, medium-chain triglyceride (MCT) oil, or 73% Polycose (Pc). Data are means from the numbers of rats indicated. SEs are displayed for all doses except 0.5, 1.0, and 2.0 ml of corn oil where they were omitted to avoid cluttering of the graph.

To quantitate differences in caloric compensations over the entire 7 h of feeding, we calculated the number of kilocalories reduced from the day's intake as 3.4 times the difference between grams of chow consumed by the end of 7 h after the saline premeal and grams consumed by the end of 7 h after each dose of premeal nutrient. In each animal, the kilocalories reduced over the 7 h (Y) were examined by linear regression against the kilocalories instilled with each premeal (X). We then computed the mean ± SE for the slopes of these regressions among the individual rats. These calculations indicated that caloric compensations to the three doses of MCT oil were accurate over the range given, that is, the mean slope of the regressions (1.106 ± 0.069) did not differ significantly from 1.000. Only two doses of Polycose were given, but within these limits, compensation was also fairly accurate because the slope (1.319 ± 0.230) also did not differ significantly from 1.000. By contrast, compensation to the corn oil over the same range as the MCT oil was inaccurate because the slope (0.622 ± 0.045) was significantly (P < 0.0005) less than 1.000. Moreover, a one-way ANOVA indicated that the slope from the Polycose premeals did not differ significantly from the slope from the MCT oil meals but that the slope from the corn oil premeals was significantly (P < 0.05) less than that from the MCT oil.

Time courses of eating and satiation. Inspection of Fig. 1 revealed two time-related phenomena. First, the fasted animals ate more in the first hour after the saline premeals than in subsequent hours. Second, dose-related reductions of food intakes by the nutrient premeals were more prominent in the first hour than subsequently. In fact, with all three premeals, there were steeply dose-related inhibitions (P < 0.0005) of first-hour intakes (Table 1), but in subsequent hours after all doses of corn oil (other than 4 ml) or after both doses of Polycose, the rate of food consumption seemed to parallel that after the control instillations of saline. To confirm this impression, we removed the first hour from the analyses by two methods.

In the first method for graphic display (Fig. 2) we calculated in each animal the difference between grams of chow consumed in the first hour after the saline premeal and grams consumed in the first hour after each dose of premeal nutrient. Thereafter, for each dose of nutrient, this first-hour difference, in turn, was subtracted from the difference at each subsequent hour between consumption after saline and consumption after the nutrient. We then plotted, after this removal of the first-hour difference, the cumulative difference with time of consumption between the saline control and the dose of nutrient (Fig. 2). For each animal, the cumulative difference was analyzed by linear regression against time. If inhibition continued over most of the feeding period after the first hour, then these cumulative differences should exhibit a positive slope when regressed against time.

View larger version (20K): [in this window] [in a new window]   Fig. 2.   In this figure, we have removed first-hour differences by subtraction of cumulating differences of g eaten each hour after the saline minus g eaten after the nutrient premeals. Significantly positive slope to the lines would indicate continuing suppression of food intake over most of the 7 h after the nutrient premeal. All of the MCT oil premeals but only the 4-ml corn oil premeal exhibited a significantly positive slope. Slope after the 6 ml of Polycose was of borderline significance (P  0.05).

These analyses (Fig. 2) confirmed the dominance of the first hour in caloric compensations after Polycose or the lower doses of corn oil. For example, after the three lowest doses of corn oil, the slopes of the regressions did not significantly differ from zero, indicating that caloric compensation from these three lowest doses arose mostly within the first hour. Only after the 4-ml dose of corn oil was the slope significantly different from zero. Similarly, the slope did not differ from zero after the 3-ml dose of Polycose. After the 6 ml of Polycose, the slope was marginally different (P  0.05) from zero. By contrast, after the MCT meals there was a considerably larger, continuing suppression of intake after each dose.

A second analysis was done to entirely eliminate cumulative effects across the 2- to 7-h period. First, we examined grams of food ingested each hour from the second through the seventh hours. During this period, hourly intakes fluctuated among individual animals, but linear regressions (g eaten each hour by each animal vs. time) produced mean slopes that seldom differed significantly from zero and did not show treatment effects across doses after any of the three nutrients (by ANOVAs for repeated measures). This initial result indicated that hourly intakes over hours 2-7 could be fairly summarized as an average after each dose of nutrient. For each nutrient, average hourly intakes over hours 2-7 (Table 2) were found to be still dose responsive (P < 0.005 for LCT and MCT oils; P  0.01 for Polycose), even though the much steeper dose responsiveness in the first postcibal hour (compare Table 1 with Table 2) had been removed from the analysis. Thus both methods of analysis indicated weaker but still dose-related effects that persisted over the entire period of analysis. The suppressions over hours 2-7 were, however, more prominent for MCT oil than for corn oil or Polycose.

Gastrointestinal transits of the oil premeals. We examined gastrointestinal transits of [³H]GTE in the corn oil premeals in five animals killed at each time point, at 1, 2, 4, and 7 h after the instillation of the premeal and the start of feeding for each dose of corn oil (4 doses × 4 time points × 5 rats = 80 rats in total). To reduce biases from time-dependent improvements in our facility with these analyses, we distributed the various doses and times evenly over the 5 mo of analyses.

The percentage of [³H]GTE that had emptied from the stomach at each time point did not vary with the dose of oil in the premeal (Fig. 3); there were points of overlaps in the time courses of all of the doses, and there were no correlations between percent emptied at each time point and dose. Conversely, linear regressions of grams of oil emptied (i.e., fraction emptied × grams fed) against grams fed always revealed highly significant (P < 0.0001) correlations. These analyses indicated that, throughout the 7 h of observation, the amount of oil that had entered small intestine varied with the mass of oil in the stomach. However, <70% of the oil in the premeals had emptied by the end of the 7 h, and most of that had emptied in the first postcibal hour.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Gastric emptying (A) and gastrointestinal transit (B) of the [3H]glyceroltriether (GTE) oil marker after varying doses of corn oil in the 4-ml premeals. Individual points on the graphs are the averages from 5 animals at each dose and time.

Transit of the [³H]GTE into the colon mirrored the time course of gastric emptying after about a 2-h offset for transit through the small intestine. Thus, on average, only 25% of the [³H]GTE that had emptied from the stomach reached the colon by 2 h, but by 4 h nearly 100% of marker emptied in the first 2 h had entered the colon. Moreover, the shapes of the time courses were similar for gastric emptying and colonic filling (Fig. 3). The percentage of [³H]GTE having reached the colon by each time point had no correlation with the grams of oil fed, that is, fractional intestinal transits were unchanged over an eightfold variation in dose, just as with gastric emptying.

We did a similar, but much more limited, analysis with premeals of MCT (data not shown) by examining transits in five animals each at 1, 2, 4, and 7 h after the 4-ml oil dose but only at 1 h after the 1-ml dose (25 rats in total). As with corn oil premeals, there was rapid gastric emptying of the MCT oil in the first hour, followed by very slowed emptying thereafter. Similarly, the grams of MCT oil emptied in the first hour varied (P < 0.0001) with the grams fed. The percentage of corn oil emptied from the stomach in the first hour (all doses in 20 rats) did not differ significantly (t-test) from the percentage of MCT oil emptied (both doses, 10 rats). However, the limited analyses suggested that small intestinal transit was slower after the MCT oil than after the corn oil; for example, by 4 h only ~40% of the MCT oil that had left the stomach had entered the colon.

Satiety depended on formation of lipolytic products. Several nutrient-driven, intestinal signals [like stimulation of pancreatic secretion by fats (25) or inhibition of gastric emptying by fats (15, 21)] are driven, not by polymeric foods but instead by more chemically active, hydrolytic products. By inhibiting lipolysis with orlistat, we tested how much of the caloric reduction to premeals of corn oil and of MCT oil was signalled through the generation of lipolytic products. To maximize sensitivity so that smaller numbers of animals could be used, we examined only the 4-ml dose of oil, the one that produced the largest reduction of food intakes compared with the saline control. On three consecutive experimental days, rats were given 4-ml premeals of saline, corn or MCT oil, and corn or MCT oil plus orlistat in a Latin square design. The orlistat completely blocked the reduction in food intake by either oil (Fig. 4) so that the responses to oil plus orlistat were insignificantly different from the responses to the control premeals of saline yet significantly different (P < 0.001) from the suppressed intakes after the oil premeals without orlistat.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Effect of the lipase inhibitor orlistat on the satiety response to 4 ml of oil in 15 rats (means ± SE). A: corn oil vs. corn oil plus orlistat were tested in 15 rats. B: experiment was repeated with MCT oil vs. MCT oil plus orlistat in a different group of 14 animals. In both cases, the added orlistat completely abolished the suppressions of food intake by the oils compared with the saline control.

To ascertain the inhibition of lipolysis, we examined in six other rats the effects of the same concentrations of orlistat on transit, digestion, and absorption of corn oil at 4 h after a 4-ml premeal (Table 3) using [¹⁴C]triolein and [³H]GTE as outlined. The lipase inhibitor increased food intake, whereas it reduced lipolysis in small intestine by 85% and virtually eliminated absorption of the [¹⁴C]triolein. It also significantly sped the gastric emptying of the oil premeal, as well as the gastrointestinal transit of oil to the colon. Despite increased food intake, the grams of gastric content (food + secretions) were significantly reduced after the oil plus orlistat vs. the oil alone (Table 3). The 18.2 g of gastric content at 4 h after the corn oil plus orlistat were not significantly different from the 16.6 g of gastric content in five other rats (see Fig. 7) 4 h after a 4-ml saline premeal (no corn oil).

Lipolysis, absorption, and length of intestinal contact. In the first two hours of more rapid gastric emptying, the fraction of corn oil that entered the intestine was the same, regardless of the dose (Fig. 3), so that the milligrams of oil entering the duodenum increased sharply with dose. As previously observed in intestinally perfused dogs (22), lipolysis increased with rising rates of duodenal entry of fat but did not keep pace with the increasing loads of substrate, so that it became less complete within the proximal small intestine, as the dose of corn oil exceeded 1.0 ml (Fig. 5A). At all doses (or thus rates of duodenal entry) of oil, absorption took more intestinal length to complete than did lipolysis, but this difference was more marked at higher rates of inflow. For example, duodenal inflows of oil were much slower by 4 h than during the first hour (Fig. 3), and, correspondingly, absorption of oil for all doses was more complete by the time the oil reached the distal one-third of bowel at 4 h than at 1 h, even though lipolysis (at least for the 0.5- to 2.0-ml doses) had been about as extensive at both times of sampling (Fig. 5). As milligrams of oil inflow increased, both the longer length of bowel to complete lipolysis and, much more so, the longer length to complete absorption of lipolytic products resulted in an increasing of length of bowel exposed to fatty acids.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Lipolysis and absorption of [14C]triolein contained within the corn oil premeals of various volumes. Averages for 5 animals at each point were derived from analyses of 14C-to-3H ratios in luminal contents in proximal, middle, and distal thirds of small intestine (segments 1, 2, and 3, respectively). Values represent the percentage of [14C]triolein digested or absorbed by the time that portion of the oil meal passed into and through the small intestinal segment. A: 1 h; B: 4 h.

We calculated (see APPENDIX) the milligrams of unhydrolyzed oil remaining as well as milligrams of free fatty acid (FFA) that remained in the lumens of proximal, middle, and distal thirds of intestine and colon (Fig. 6) with time after the various doses of oil. During the early period of rapid gastric emptying of oil (i.e., the first 2 h), all three segments of small bowel plus the colon were exposed to FFAs, but the intensity of exposure (i.e., mg/segment) varied directly with the oil dose (thus rate of oil entry, mg/h, into duodenum). As duodenal entry of oil slowed (in hours 4 and 7), both the length of intestinal contact and intensity of contact per segment diminished. For example, with the 0.5-ml dose of oil, the middle and distal thirds of small intestine were no longer contacted by FFAs in the fourth and seventh hours (Fig. 6), because absorption of lipolytic products had been completed more proximally (Fig. 5). However, with the 2.0- and 4.0-ml doses of oil, after which oil entry (mg/h) was proportionately still four to eight times faster than entry after the 0.5-ml dose, the distal small intestine and colon continued to be contacted by FFAs over hours 4-7, and the intensities of exposure were greater in distal small and large bowel after the 4-ml compared with the 2-ml doses. The profile after the 1.0-ml dose was intermediate between the 0.5- and 2.0-ml doses. Thus both the length (no. of segments) of bowel contacted and the intensities (mg/segment) of contact varied directly with the dose of oil given in the premeals.

View larger version (38K): [in this window] [in a new window]   Fig. 6.   Average (from 5 rats at each time and dose of oil) mg of free fatty acid (FFA) and mg of unhydrolyzed oil remaining unabsorbed in the segmental lumens at 4 times after each of the 4 premeals of corn oil as calculated from analyses of 3H and 14C (see APPENDIX). Prox, proximal; Mid, middle; Dist, distal.

There were no qualitative differences of digestion or absorption of [¹⁴C]triolein, whether the tracer was added to 4 ml corn oil or 4 ml MCT oil meals. For example, both lipolysis and absorption of the radiotriolein were incomplete by the time the oil reached the distal small intestine in the earlier hours after the MCT oil premeal; whereas, later, lipolysis was completed within the small intestinal length (data not shown). Moreover, the percentage of [¹⁴C]triolein entering the colon-feces over time was nearly identical whether the radiotriolein was given with the corn oil or with the MCT oil.

Mediation of the Satiety Responses

View larger version (34K): [in this window] [in a new window]   Fig. 7.   Average net weights of stomach contents (food + secretions) at times indicated after 4-ml premeals that contained 0-4 ml of corn oil or MCT oil. Rats were allowed unrestricted access to food after the gastric instillations of the premeals at time 0. Each point is the average from 5 rats. SEs are graphed for 0- and 2-ml doses of corn oil.

Predominance of long-chain fatty acids. In our previous study (24), we found that, on a molar basis, intestinally perfused dodecanoic acid (C-12) was as potent as oleic acid (C-18) at inhibiting sham feeding in rats, whereas neither decanoic (C-10) nor octanoic (C-8) fatty acids inhibited. The detergent L-81 is known to block the formation of chylomicra and thus the transport of long-chain fatty acids from the absorptive cell to the lymph (29, 30). In addition, the apolipoprotein A-IV (Apo A-IV) is synthesized and secreted into the lymph under the stimulus of absorption and lymphatic transport of long-chain fatty acids (1, 12). Moreover, Apo A-IV may cross the blood-brain-barrier and directly trigger receptors to Apo A-IV that signal satiety from the hypothalamus (5, 6). Lymphatic transport of absorbed fatty acids, which is inhibited by L-81, starts with chain lengths longer than 10 carbons (13), just as inhibition of sham feeding starts with chain lengths of intestinally perfused fatty acids >10 carbons. Therefore, an effect of L-81 on caloric compensations to MCT and LCT premeals might provide important insights into the mechanisms that induce satiety in response to each type of oil.

We studied the effect of 2.8% (vol/vol) of L-81 in 4 ml of corn oil vs. 4 ml of corn oil without L-81 vs. 4 ml of saline alone. These three treatments were given every other day in a Latin square randomization to 25 rats (Fig. 8). The L-81 completely abolished the satiety response to the corn oil for the first 4 h. Food intake after the 4 ml of corn oil without L-81 was significantly (P < 0.01) lower, but intake after corn oil plus L-81 was not significantly different from that after saline. The effect of L-81 was specific to fat, since emulsifying 2.8% L-81 in 6 ml of 73% Polycose did not alter the inhibition of the Polycose premeal (data not shown).

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Effect of Pluronic acid L-81 (L-81), an inhibitor of lymphatic transport of absorbed fatty acids longer than 10 carbons, on satiety after 4-ml premeals of corn oil (A) and of MCT oil (B) in 25 rats (means ± SE). L-81 eliminated the satiating effect of the long-chain triglycerides (LCTs) in corn oil and increased the g of chow eaten in the first 3 h after the MCT oil (which contains only 4% LCTs) by about the same amounts as after the corn oil.

The same 6-day protocol was then used in 25 rats to examine the effect of L-81 on satiety from MCT oil (Fig. 8). The three responses were each significantly different from each of the others (P < 0.005), but, clearly, the addition of L-81 to the 4 ml of MCT oil reduced the satiating effect of the MCT oil by ~60% in the first three postcibal hours. The absolute change in grams of food eaten in the first 3 h with the addition of L-81 to the oils was about as great after the MCT oil as after the LCT oil. For each of the first 3 h, we computed the differences in grams eaten by each rat between oil plus L-81 and oil, and we then summed the hourly differences. The sum of these hourly differences between MCT oil plus L-81 and MCT oil (6.6 ± 1.3) was not significantly different (by unpaired t-test) from the sum of hourly differences between LCT oil plus L-81 and LCT oil (7.3 ± 1.2).

Lack of dominant role of CCK. Among digestive products of macronutrients, lipolytic products in small intestine are believed to be the most potent releasers of CCK, and released CCK is regarded by many scientists as an important mediator of satiety. Therefore, we used the specific CCK antagonist MK-329 (Merck) to assess the role of CCK in the satiety responses to the corn and MCT oils. We gave MK-329 (vs. vehicle controls) intraperitoneally 20 min before the premeal at the start of the feeding period, in a dose of 0.1 mg/kg [known to effectively block peripheral CCK receptors and to inhibit satiety responses in rats under some circumstances (27)] and also in doses of 1.0 mg/kg.

In the first experiment, the effect of 1.0 mg/kg of intraperitoneal MK-329 was compared with the effect of intraperitoneal vehicle on food intake after 4-ml saline premeals. One-half of the rats received vehicle, and one-half received MK-329 on the first of two consecutive days. The CCK antagonist did not alter food intakes in this control experiment in which the animals consumed rat chow in the absence of corn oil. In the same 15 animals, MK-329 did not suppress the caloric compensations to 4-ml premeals of corn or MCT oils whether given in doses of 0.1 or 1.0 mg/kg (data not shown).

Failure to confirm neural dominance. Much indirect evidence suggests an important role of chemosensory, afferent nerves in the transmission of information that cues satiety from small intestine (20, 32). We therefore sought to demonstrate neural mediation through the use of systemic capsaicin, a selective neurotoxin for unmyelinated, afferent nerves that include chemosensory nerves from small intestine (10). The procedure for this selective denervation has been outlined (see METHODS). Before the procedure was undertaken, all rats underwent testing with premeals of corn oil (1.0-4.0 ml) and Polycose (3 and 6 ml). Equal numbers of rats were then assigned to treatment with capsaicin or vehicle but otherwise were treated identically with regard to anesthesia and recovery. Seventy percent of rats treated with systemic capsaicin survived and recovered completely, for a total of 16 surviving animals. After 1 wk of return to normal food intakes, the capsaicin-treated animals again underwent testing with premeals of corn oil and Polycose. This retesting was finished within 3 wk of the capsaicin treatment, during which weekly NH₄OH eye wipe tests elicited a positive wiping response in vehicle-treated animals but no reaction in capsaicin-treated rats, indicating sustained denervation of chemosensory afferents to the conjuctiva (10).

Both vehicle-treated animals (not shown) and capsaicin-treated rats continued to exhibit dose-responsive reductions of food intakes to the various doses of corn oil and of Polycose (data not shown). There was no evident change in the responses after capsaicin treatment; for example, the slopes of the dose responses were not significantly different within the same animals before compared with after the capsaicin (paired t-test).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion Appendix References

Dose-Responsive Quantitations

The results with inhibition of lipolysis by orlistat (Fig. 4 and Table 3) establish that only lipolytic products, not triglycerides, suppressed food intakes. Similarly, Maggio and Vasselli (17) observed that inhibition of intestinal saccharidases nearly completely abolished the reductions of intakes after premeals of corn starch in previously fasted rats. Taken together, these observations indicate that more chemically reactive, hydrolytic products of macronutrients trigger intestinal satiety, not the polymeric foods themselves (24). This idea is analogous to intestinal stimulation in other species of pancreatic secretion by fat (25) or intestinal inhibition of gastric emptying by fat (21, 23).

In our previous study (24), fatty acids intestinally perfused at 240 µmol/h inhibited sham or free feeding by 10-15%, at 480 µmol/h by 30%, and at 960 µmol/h by 63% in rats of similar size. In the present study, the rats consumed 26.5 kcal in the first hour after the saline (control) premeals. Allowing for the two times faster hydrolysis of the radiotriolein (see METHODS), we estimated (Table 4) that 220 µmol of FFAs entered small intestine in the first hour after 0.5 ml of oil and that this load reduced food intake by 14%. Similar calculations indicated that 480 µmol were released in the first hour after the 2.0-ml oil meals to give a 28% reduction, and 995 µmol released after the 4.0-ml corn oil inhibited intakes by 40%. Similarly, the reduction ratios (the ratio of kcal of reduced intake of chow to kcal of lipolytic products released in small intestine) were 2.7-4.3 for long-chain fatty acids in both studies (for the first hour, only, in the present experiments). The very close correspondences indicate that the observations in the former perfusion study and the present natural feeding study were expressions of the same biological processes.

Our paradigm utilized fasted animals, but others have investigated the effects of intragastric instillations on intakes in continuously feeding rats. The reductions of subsequent intakes after instilled glucose were dose dependent but amounted to 25-50% of the inhibition we observed in the first hours after ~4 g of carbohydrate (2). Similarly, Geliebter (7) observed 25-50% of the inhibition that we observed after 2.0-ml loads of corn oil. In contrast to our animals, which started feeding with empty stomachs, the freely and continuously feeding rats undoubtedly had stomach and small intestine already filled to some degree with food when the oil or glucose was instilled into the stomach. Under these circumstances, feedback inhibition of gastric emptying was already operating so that there would most likely have been a much reduced rate of entry of oil into small intestine, one more comparable to the fractional rates that we observed from the second hour and beyond. The much smaller satiations after the second hour (Tables 1 and 2) in our studies were similar in magnitude to those observed in continuously feeding rats in the first hours after gastric instillations.

Temporality and Intestinal vs. Postabsorptive Signals

The 4 ml of corn oil suppressed food intake most markedly in the first postcibal hour (Figs. 1 and 2), the brief period during which dose-related inhibitions arose after all doses of corn oil. Only the 4-ml dose of the corn oil produced a suppression of food intake for most of the 7 h of observation. Similar temporality was noted by Rolls et al. (28), who observed that oil or carbohydrate (and other constituents) in "premeal" snacks in fasted humans had satiating effects that rapidly diminished after ~90 min. However, the MCT oil produced sustained inhibitions after the entire range of doses.

It takes ~35 min from the start of intestinal perfusion of LCTs to the appearance of significant quantities of absorbed fatty acids 12 carbons in the lymph of rats (3, 8); transfer of absorbed fatty acids into lymph peaks 2 h thereafter (3, 8). Moreover, lymphatic transport from the whole small intestine of rats is rate limited to 200 µmol/h (3, 8, 31), a limitation that would have severely truncated any dose response based solely on molar absorption of LCTs into the systemic circulation. About 40% of the corn oil premeals emptied from the stomach in the first hour, and ~60% emptied over the 7 h of observation (Fig. 3). From the product of micromoles entering the duodenum times fraction absorbed by terminal ileum, we estimated that ~205-629 µmol of triglyceride were absorbed from the 0.5- to 4.0-ml loads of oil entering the small intestine in the first hour. These figures indicate that it must have taken >1.5 h for complete transfer of even the lowest first-hour load of absorbed corn oil to the systemic circulation and 3.5 h for complete transfer of the highest load absorbed in the first hour alone. Therefore, it is not reasonable to attribute these dose-responsive satiations (which were mostly limited to the first hour) to postabsorptive, systemic signals. However, it is less certain how much of the satiety response that lasted 7 h after the 4-ml corn oil loads arose from gut vs. the systemic circulation.

In contrast to the corn oil, absorbed MCTs are transferred predominantly to the portal vein (13), and this process is rapid. Thus, with the MCT oil, a substantial portion of the signal may have arisen from catabolism of fat absorbed into the systemic circulation. We found that MCT oil was hydrolyzed 30 times faster than corn oil; shorter-chain fatty acids released from MCT oil much more rapidly traverse unstirred water layers, and fatty acids 10 carbons are rapidly transferred directly to portal blood. Therefore, it can be reasonably assumed that the lower-molecular-weight triglycerides in the MCT oil were completely hydrolyzed and absorbed so that ~2,800 µmol would have entered the systemic circulation in the first hour after the 4.0-ml MCT oil meal. Most likely, the greater and more prolonged effects of even the lowest dose of MCT oil, when compared with corn oil, arose from continuing portal transfer of the fat into the systemic circulation and metabolism throughout the 7 h (16).

L-81 blocks only lymphatic transfer of fatty acids 12 carbons, the very same fatty acids that inhibited sham or free feeding by activating intestinal sensors in intestinally perfused rats (24). The L-81 blocked 100% of suppression of food intake by the corn oil and 50-60% of the effect of the MCT oil in the first 3 h. In fact, the L-81 plus oil increased the intakes of rat chow over oil without the L-81 by the same amounts after both corn oil and LCT oil. These observations indicate that one-half of the early suppression from the MCT oil was mediated by longer-chain fatty acids (principally C-12) within the MCT oil through the triggering of intestinal sensors. As suggested above, the remaining L-81-insensitive suppression by MCT oil probably resulted from systemic effects of absorbed medium-chain fatty acids.

Quantitation Through Length of Intestinal Contact

A study in rats (26) indicated that length of intestinal contact was a dominant factor in suppression of food intakes by intestinally perfused digestive products, more exactly, that suppression increased severalfold as length of contact extended from proximal into distal bowel. Even when length of contact is held constant, the amount of nutrient within a segment may also directly affect the response (26). The dose-related length and intensity of gut contact in the present study could well account for the observed suppression of food intakes. The changing spatiotemporality of spread and intensity of gut contact with FFAs after varied doses of oil premeals (Fig. 6) correlated better with the timing and duration of suppression of food intakes (Figs. 1 and 2) than did the stretch of the stomach, i.e., total gastric content (Fig. 7).

Because the [¹⁴C]triolein was hydrolyzed two to three times faster than the corn oil and its lipolytic products were absorbed 1.6 times faster than those from corn oil, it is a guess as to exactly how the majority of lipolytic products from the corn oil were distributed along the gut. Because the slower rate of absorption of [¹⁴C]oleate determined the extent of spread more than the speed of hydrolysis of [¹⁴C]triolein (Fig. 5), it is likely that the more slowly absorbed FFAs from the corn oil were spilled as much, if not more, into distal bowel than the [¹⁴C]oleate.

We did not expect to find that one-half of the satiation from MCT would be blocked by L-81, because this compound affects glycerides with fatty acids 12 carbons, which constituted only 4% of the fatty acids in MCT oil. In addition, it could not have been predicted that the MCT oil inhibited gastric emptying about as much as did the corn oil because only FFAs 12 carbons are potent (11). One explanation for 1) the unexpected potency of MCT oil at inhibiting gastric emptying and 2) the much higher than expected L-81-sensitive suppression of food intake by MCT oil might be that the 4% of fatty acids 12 carbons were spread along the entire length of small intestine to amplify the inhibitory responses out of proportion to their molar composition in the oil. Most likely, the much more abundant shorter-chain glycerides mechanically swept the 12 carbon glycerides dissolved in them along the gut and competed with the longer-chain glycerides for pancreatic lipolysis; both processes would slow absorption and spread the length of contact despite the very low amounts of longer glycerides initially present in the MCT oil. We had no [¹⁴C]trilaurin with which to examine this hypothesis, but the idea is supported by the observation that digestion and absorption along small intestine and length of spread of FFAs from traces of [¹⁴C]triolein added to the 4-ml doses of MCT oil showed quite similar temporal patterns (data not shown) to those observed when the [¹⁴C]triolein had been added to corn oil.

Mediators of Satiety

We did not observe any reversal of the satiety to 4-ml oil premeals by the specific CCK antagonist MK-329 after either dose. In many studies in a variety of species, including rat, this or similarly specific CCK antagonists have reversed a portion of satiety, that is, the injection of the antagonist increased the amount of diet consumed. However, these increases have not always been observed but have been most easily demonstrated during continuous feeding (as opposed to resumption of feeding after a prolonged fast) or after liquid instead of solid nutrients (27). Nevertheless, even under these more limited conditions, the increases in food consumed have always been modest. In humans, the CCK antagonist loxiglumide very modestly increased the intake of a test meal when the duodenum was perfused with fat at 6 g/h (14) but not when the duodenum was perfused at 36 g/h (4). CCK is stored in proximal small intestine while other hormonal peptides, like peptide YY, are stored in distal small bowel. It is possible that CCK prominently signals satiety only when nutrient contact is limited to proximal bowel but that, with longer lengths of contact, the prominence of CCK diminishes as additional mechanisms are activated from ileum. Our physiologically high, 4-ml dose of corn oil in these rats may have precluded a demonstration of the effect of CCK, just as a relatively high dose of fat did in humans (4, 14), possibly as the result of recruitment of other satiety signals from more distal small bowel.

There is a variety of circumstantial evidence that neural feedback is important for satiety. Thus nutrients (including fatty acids) in the intestinal lumen excited impulses along mesenteric afferent nerves (20), local anesthetic perfused with fatty acids markedly reduced inhibition of sham feeding in rats (9), and perivagal capsaicin (32) reduced inhibitions of sham feeding by nutrients instilled into rat small intestine. Nevertheless, we were unable to demonstrate that satiety was altered by selective denervation with capsai- cin of chemosensory, afferent nerves. It was unlikely that the capsaicin did not sufficiently denervate, as conjunctival reflexes were lost.

Unlike either the MK-329 or the capsaicin, the L-81 completely reversed the satiety response to the 4-ml dose of corn oil. It is now known that the synthesis and secretion of Apo A-IV is induced by absorption of LCTs (1); that L-81 blocks the synthesis and export of Apo A-IV from the absorptive cell (30, 31); that in the absence of L-81, release of Apo A-IV begins within 30 min of absorption of lipolytic products; that Apo A-IV can cross the blood-brain barrier (5); and that the hypothalamus has receptors to Apo A-IV that trigger satiety (6). In addition, recent studies indicate that contact of lipolytic products limited to ileum releases Apo A-IV from both ileal and jejunal mucosa (12), but jejunal exposure to lipids does not stimulate ileal export of Apo A-IV. This last finding indicates that secretion of Apo A-IV from the enterocyte can be dissociated from export of chylomicra themselves and that there may be an amplification of output of Apo A-IV when the entire small intestinal length is contacted by lipolytic products. Thus Apo A-IV itself could serve as a quantitative signal of the amounts of lipolytic products being absorbed along the entire small intestinal length. Although it is possible that all of the satiety response is mediated by direct contact of Apo A-IV with the hypothalamic satiety center, another possibility is that this same protein serves as a primary intermediary signal between absorptive cell and adjacent endocrine cells or absorptive cell and chemosensory nerve terminals in the neighboring intercellular space to trigger, respectively, hormone release and nerve impulses whenever the gut takes up lipolytic products.

Rats feeding after an overnight fast reduced their caloric consumptions in a dose-related fashion in response to intragastric instillations of corn or MCT oils, and tests established that intakes were suppressed only by lipolytic products. Both dose responsiveness and the timing of suppression corresponded to the extent over time of the distribution of lipolytic products along the full sensory surface of small and large intestine. These changing spatiotemporal distributions were primarily the outcome of dose-dependent speeds of gastric emptying, characterized by more rapid emptying in the first 1-2 h and slow, inhibited emptying thereafter. Aqueous premeals of carbohydrate exhibited similar temporal patterns of suppression of intakes in our rats. It is known that aqueous carbohydrates empty initially rapidly from rat stomach, slowing later, and that, during rapid emptying, carbohydrate products spread to colon but later recede to proximal bowel. Thus it is likely that spatiotemporal distributions of carbohydrate products along bowel after carbohydrate premeals similarly corresponded to timing of suppression of food intakes.

Normal humans have a similar intestinal distribution of sensors to lipid products and exhibit the same temporalities of gastric emptying of oil and satiation after various doses of oil. It is therefore likely 1) that the oil-induced satiety in normal humans derives from similar spatiotemporal distributions of lipolytic products and 2) that known alterations of these spatiotemporal patterns by gastric surgeries to treat peptic ulcers, by gastric bypass to treat obesity, or by diseases of digestion or absorption account for chronically low intakes of food frequent in these iatrogenic or pathological conditions.