Evidence of Splanchnic-Brain Signaling in Inhibition of Ingestive Behavior by Middle Molecules

Isolation of Fractions from Plasma and Urine Ultrafiltrates

These fractions were collected from uremic patients with anorexia who came for hemodialysis treatment for the first time. The serum urea level was 34.4 ± 4.6 mmol/L and the serum creatinine level was 694.8 ± 48 μmol/L (mean ± SD) before dialysis started. The plasma ultrafiltrate was obtained by applying isolated ultrafiltration during the first 30 to 60 min of the treatment, using a hollow-fiber dialyzer with a cellulose acetate membrane (CA 170 G, Baxter) and a Gambro AK 10 dialysis monitor set in ultrafiltration mode, with no dialysis fluid passing through the dialyzer. The ultrafiltrate, usually 500 to 1000 ml depending on the degree of overhydration of the patient, was collected in a plastic bottle, which was then immediately frozen and stored at -20°C. After thawing, the ultrafiltrates from patients were pooled, frozen, and kept at -70°C, pending use in tests or further processing.

Plasma ultrafiltrate was also obtained from healthy volunteers. Heparinized venous blood (200 ml) was collected from each of them, and the plasma was separated by centrifugation and pooled. It was then ultrafiltered in vitro through the same type of dialyzer as used with the patient's ultrafiltrate. The pooled ultrafiltrate was frozen and kept at -70°C, pending use in the appetite test or further processing.

Morning midstream urine samples from volunteers with normal renal function were collected and pooled. The samples were filtered in vitro through the same type of dialyzer as used with the uremic plasma ultrafiltrate, then buffered with sodium hydroxide to a pH of 6.5. The samples were stored at -70°C before use.

The ultrafiltrates from normal and uremic plasma, as well as from normal urine, were fractionated according to molecular weight and size by the molecular filtration technique, and then concentrated, as described earlier (12), using a 400-ml stirred cell chamber and 76-mm disc membranes with molecular weight cutoff points of 10, 5, 1, 0.5, and 0.1 kD (Molecular/Por Type C, Spectrum, LA). The separations were performed in a cold room at 8°C to minimize bacterial growth.

The concentrated MM fractions from uremic and normal plasma and from normal urine, representing a molecular weight of 1 to 5 kD, were used in the experiments in this study. Precipitates, if any, were removed by centrifugation, pH was corrected to 6.5, and the fractions were stored at -70°C until use.

MM fractions from normal and uremic plasma ultrafiltrates were concentrated 25 times, and the corresponding fraction from normal urine ultrafiltrate was concentrated 10 times. The concentrations selected in the present study were based on dose-response curves in a previous study from our laboratory (12). These fractions were further concentrated 5 and 2 times, respectively, for use in the icv injection experiments, which required much smaller injection volumes.

Animals, Operations, and Injections

Male Wistar rats (Möllegård Breeding Laboratories, Ejby, Denmark), weighing 350 to 380 g, were maintained in an air-conditioned, temperature-controlled colony room in which the lights were off between noon and midnight. The rats had continuous access to food and water.

All surgery was performed under pentobarbital anesthesia (60 mg/kg ip; Mebumal, Nordvacc, Stockholm, Sweden). The intraoral (io) cannulas were implanted as described by Grill et al. (7). The animals were allowed to recover for 4 wk after the operation.

Cannulation of the right lateral brain ventricle was done while the rat was in a stereotaxic frame, using the following coordinates: 1.5 cm caudally, 1.5 laterally, and 3.5 cm ventrally from bregma. The concentrated MM fractions were injected slowly through a 28-gauge cannula (Plastic Products®, Inc., Roanoke), using a 25-μl Hamilton syringe.

Intravenous cannulas were prepared from silicone medical-grade tubing (inner diameter 0.51 mm; outer diameter 0.94 mm; Degania, Bethlehem, Israel) and implanted in the right jugular vein, as described by Remie et al. (13), except that the cannula was not anchored to the skull, but was exteriorized in the subcutaneous region of the neck, approximately 2 cm caudally from the ears. The cannulation was completed within 10 min and performed 10 h before behavioral testing. This cannulation procedure does not affect intake under the conditions of this study (10).

Depending on the MM volume used in each experiment, the iv injections were given slowly in the tail (1 and 2 ml) or external jugular vein (4- and 6-ml injections). There were no behavioral differences when different intravenous routes were used.

Diet Solution

A carbohydrate solution containing the carbohydrate moiety of a mixed nutritional solution (Fortimel®, Nutricia Nordica AB, Stockholm, Sweden) with 10.4 g of starch was used as the intraoral diet solution. The caloric density of the solution is 100 kcal/100 ml.

Ingestive Behavior

The pellets were removed at 7 a.m., and the animals were tested 6 h later, i.e., 1 h after the lights were switched off. The rats were placed in circular (35-cm diameter) Plexiglas arenas and had their intraoral cannulas connected to a peristaltic pump (Alitea XV, Ventur Alitea, Stockholm, Sweden), which delivers a solution at 1 ml/min. This activates ingestive behavior that ends by the time the animal either actively rejects the solution or passively lets it drip from its mouth. At that time, the infusion is interrupted for 30 s and then continued. If the animal ingests the diet solution during at least 1 additional minute, the infusion is again stopped for 30 s by the time the animal rejects the solution or lets it drip from its mouth within 1 min after the start of the infusion. Usually, rats let the solution drip from their mouth within the first minute after the start of the second infusion period (9). Active rejection of the solution is less common. The pellets were replaced after testing. After recovery from the trauma of catheter implantation, a training and adaptation period is required, during which the intake of the orally infused solution gradually increases until a stable level of intake is attained. Then the rats were studied once a day on consecutive days. Under these conditions, rats commonly stabilize their carbohydrate and pellet intakes within 5 to 7 d (9, 10).

In all experiments, the io infusion of the carbohydrate solution was started 10 min after the injections of the MM fractions. Only rats that had approximately equal intakes after the adaptation period were used in these studies.

Sexual Behavior

Male rats were tested individually for sexual behavior at the same time, i.e., 1 h after the light was turned off and in the same arenas as used for ingestive behavior. A sexually receptive female (injected with 2 μg of estradiol benzoate [Sigma, St. Louis, MO], dissolved in 0.1 ml of sesame oil and injected subcutaneously [sc] 48 and 24 h before, and 0.5 mg of progesterone [Sigma], dissolved and injected in the same way as estradiol 6 h before testing) was placed with the male. Records were kept on the number of mounts with intromission (IF) before ejaculation and the latency from the first intromission to ejaculation, ejaculation latency (EL). The other parameters of sexual behavior, i.e., latency to the first intromission (IL), the frequency of mounts without intromission (MF), and the time between ejaculation and the next intromission, the postejaculatory interval (PEI), were also recorded. Sexually experienced male rats intromit within a few seconds when a female is available, show very few mounts before ejaculation, and have very stable EL and PEI (14). The parameters most sensitive to manipulations are the number of intromissions and the EL and PEI. The males were used after showing a stable pattern of sexual behavior in five preliminary tests separated by 2 d.

Procedure

Effect of ip, iv, or icv Injection of MM Fractions on Carbohydrate Intake. A group of six rats was injected ip with 2 ml of concentrated uremic plasma ultrafiltrate MM fraction or normal plasma ultrafiltrate MM fraction in random order on 2 consecutive days, and the amount of carbohydrate intake was determined. A similar experiment was done by ip injection of 0.5 ml of urine ultrafiltrate or saline in a separate group of rats (n = 6).

A group of eight rats was injected iv with 2 and 4 ml of uremic plasma ultrafiltrate MM fraction in random order on 3 consecutive days, and the amount of carbohydrate intake was determined. In a separate group (n = 6), 6 ml of the uremic MM fraction was given iv. In another group (n = 8), 2, 4, and 6 ml of normal plasma ultrafiltrate MM fraction was injected iv. Ingestion without any injection served as controls.

In eight rats, 0.5 and 1 ml of urine ultrafiltrate MM fraction or saline was injected iv, and the amount of carbohydrate intake was determined. In another group (n = 6), 2 ml of urine ultrafiltrate MM fraction or saline were given iv.

Eight male rats were injected icv with 5 or 10 μl of uremic plasma ultrafiltrate MM fraction or the corresponding fraction of normal plasma ultrafiltrate in random order on 3 consecutive days, and the amount of carbohydrate intake was determined. The experiment was repeated using 5 or 10 μl of urine ultrafiltrate or saline in another group of rats (n = 8).

Effect of ip Injection of MM Fractions on Sexual Behavior. Male rats (n = 6) were injected ip with 2 ml of normal plasma MM fraction or 2 ml of uremic plasma ultrafiltrate MM fraction and tested for sexual behavior 10 min later. The injections were given in random order on 2 consecutive days.

The experiment was repeated on another group of rats (n = 8) injected ip with 0.5 ml of normal saline (0.9% NaCl) or 0.5 ml of normal urine ultrafiltrate (10:1) MM fraction.

Statistical Analyses

Data are presented as mean ± SEM. A P value <0.05 was considered significant. Paired t tests were used for statistical analysis of differences. In the dose-response studies, the data were tested by ANOVA, which, when significant, was followed by Dunn's test with Bonferroni correction for multiple comparison.

Effect of MM Fractions on Carbohydrate Intake

The ingestion of the carbohydrate solution was reduced by 45.9% after ip injection of 2 ml of uremic plasma ultrafiltrate MM fraction, compared with the ingestion after injection of the corresponding fraction isolated from normal plasma (Figure 1). Intravenous injection of 2 ml of uremic plasma ultrafiltrate MM fraction had no effect, injection of 4 ml inhibited ingestion by 10% (NS), and injection of 6 ml inhibited ingestion by 33% (Figure 1). Intravenous injections of up to 6 ml of MM fraction isolated from normal plasma had no effect on ingestion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Effect of intraperitoneal (ip) or intravenous (iv) injections of uremic plasma ultrafiltrate middle molecule (MM) fractions on ingestion of carbohydrate solution in each group of male rats. The ingestion after ip injection of uremic ultrafiltrate (left) was compared with the ingestion after injection of the same volume of normal plasma ultrafiltrate. The ingestion after iv injections of the various doses (right) was compared with the ingestion without injection. Data are means ± SEM.

Intraperitoneal injection of 0.5 ml of MM fraction isolated from normal urine reduced carbohydrate intake by 76.3%. Intravenous injection of 0.5 ml of urine ultrafiltrate MM fraction had no effect, injection of 1 ml inhibited ingestion by 8% (NS), and injection of 2 ml inhibited ingestion by 32% (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Effect of ip or iv injections of normal urine ultrafiltrate MM fraction on ingestion of carbohydrate solution in male rats. In each group, the ingestion after injection of 0.5 ml of saline served as control. Data are means ± SEM.

Injections of 5- or 10-μl MM fractions isolated from uremic plasma ultrafiltrate into the lateral ventricle of the brain reduced the ingestion of the carbohydrate solution by 22.6 and 49.5%, respectively, compared with injection of the corresponding fractions isolated from normal plasma ultrafiltrate (Figure 3). Intracerebroventricular injection of 5 or 10 μl of 1 to 5 kD fraction isolated from normal urine reduced carbohydrate intake by 13.5 and 41.6%, respectively (Figure 4).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Effect of intracerebroventricular (icv) injections of concentrated uremic ultrafiltrate (125:1) MM fractions on ingestion of a carbohydrate solution in male rats. The concentrated MM fraction elicited strong inhibition of the ingestion of the carbohydrate solution (two-way ANOVA), which was dose-dependent. Data are means ± SEM. At each concentration, the ingestion after ultrafiltrate injection was compared with the ingestion after injection of normal plasma ultrafiltrate.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Effect of icv injections of concentrated normal urine ultrafiltrate (20:1) MM fractions on ingestion of a carbohydrate solution in male rats. The concentrated MM fraction elicited strong inhibition of the ingestion of the carbohydrate solution (two-way ANOVA), which was dose-dependent. Data are means ± SEM. At each concentration, the ingestion after ultrafiltrate injection was compared with the ingestion after injection of saline.

Effect of MM Fractions on Sexual Behavior

Intraperitoneal injection of MM fractions isolated from either uremic plasma ultrafiltrate or normal urine had no effect on sexual behavior (Table 1). Since ip injection of the MM had no effect on sexual behavior, we did not expect an effect of icv injection, and therefore this experiment was omitted.