Enterostatin deficiency increases serum cholesterol but does not influence growth and food intake in mice

Abstract

A pentapeptide released from procolipase, enterostatin, selectively attenuates dietary fat intake when administered peripherally or centrally. Enterostatin may act through the afferent vagus nerve and in the hypothalamus and amygdala, primarily in the central nucleus of the amygdala. To investigate the physiological role of endogenous enterostatin, we created an enterostatin-deficient, colipase-sufficient (Ent^−/−) mouse. Ent^−/− mice are viable, normally active, and fertile. They exhibit normal growth on low-fat and high-fat diets. Furthermore, Ent^−/− mice develop diet-induced obesity, as do Ent^+/+ mice, and have normal responses to a two-macronutrient choice diet and to a switch from a high-fat to a low-fat diet. Levels of total serum (P = 0.004) and non-HDL (P ≤ 0.001) cholesterol were higher and levels of HDL cholesterol (P = 0.01) were lower in Ent^−/− than in wild-type mice. To determine whether enterostatin contributed to the decreased survival or whether colipase deficiency was the sole contributor, we administered enterostatin to procolipase-deficient (Clps^−/−) mouse pups. Enterostatin significantly improved survival (P ≤ 0.001). Our results demonstrate that enterostatin is not critically required to regulate food intake or growth, suggesting that other pathways may compensate for the loss of enterostatin. Enterostatin has developmental effects on survival of newborns and alters cholesterol metabolism.

colipase is a 10-kDa protein secreted by the pancreas, where it interacts with pancreatic triglyceride lipase (PTL) to facilitate the digestion of dietary fats (5, 30). The pancreas synthesizes colipase as a proprotein, procolipase. The proform is converted to colipase in the duodenum by the proteolytic cleavage of a pentapeptide from procolipase. The propiece is present and highly conserved in all procolipase molecules, from dogfish to humans (9). Although the propiece was long considered an activation peptide, the observation that the propiece of colipase decreased the appetite of rabbits led to the alternative hypothesis that the pentapeptide is a hormone that controls fat intake (8).

The peptide, named enterostatin, reduced food intake in a general way when injected centrally or peripherally into a rat (10, 12, 40). Importantly, enterostatin also reduced food intake if given intraduodenally (32, 33). In animals given a choice of fat, carbohydrate, or protein, enterostatin specifically reduced fat intake (12). When presented with a choice between a low-fat and a high-fat diet, rats given enterostatin consumed the low-fat diet. These effects occurred in normal rats and in genetically obese Zucker or Osborne-Mendel rats (10, 35). Changing the amino acid sequence or removing the COOH-terminal amino acid abolished the inhibition of fat intake (23). Rats given enterostatin exhibited the normal sequence of behavior, grooming, exploration, and resting that is seen after a satiating meal (22). The duration and amount of water consumption were not affected by enterostatin. These findings suggest that enterostatin is a satiety factor that limits fat intake and that the effects on satiety promote body weight loss (11, 35).

The physiological role of enterostatin as a satiety factor has not been confirmed by characterizing animals deficient in enterostatin. A mouse model of procolipase deficiency has been reported (5). In procolipase-deficient (Clps^−/−) mice, decreased neonatal survival, fat malabsorption in newborns and in adults fed a high-fat diet, and decreased rate of weight gain until weaning, when the rate of weight gain was normal, were observed. Clps^−/− mice fed a high-fat diet maintained their weight, despite marked steatorrhea, by eating twice as much as controls. The phenotype of the mice indicated that procolipase has essential functions in postnatal development and in regulation of body weight set point. Since enterostatin and colipase were missing in these mice, no conclusions could be made about the role of enterostatin in any of the alterations.

To define the physiological role of enterostatin deficiency, we created a mouse with deletion of the nucleotides encoding enterostatin from the procolipase gene. That is, we made a mouse with enterostatin deficiency and colipase sufficiency (Ent^−/−). The mice were characterized for colipase function by a test of their ability to digest and absorb dietary fat. We then determined the effects of enterostatin deficiency on food intake, weight gain, and metabolism. In addition, we examined the effect of endogenous enterostatin on the neonatal phenotype in the colipase-deficient mice.

MATERIALS AND METHODS

DNA methods.

All manipulations of DNA were done by standard methods (39). After restriction digests, the desired fragments were separated by agarose gel electrophoresis and extracted from the gel using a Qiaquick gel extraction kit (Qiagen) following the manufacturer's instructions. Mutagenesis was accomplished with the QuickChange (Stratagene) method according to the manufacturer's instructions and under PCR conditions. A Robocycler (Stratagene) was used for the PCR. After the reaction, the products were transformed into Escherichia coli, and plasmid DNA was isolated from individual colonies using a Qiagen Miniprep kit. We screened the isolated DNA by digestion with Sma I, since the single site in the insert was removed by the mutagenesis. The products were separated by electrophoresis in a 0.9% agarose gel followed by ethidium bromide staining. Plasmids that produced the proper pattern were analyzed by dideoxynucleotide sequencing to confirm the deletion of the sequences encoding enterostatin.

Transgenic mice.

The targeting vector was linearized and introduced into RW-4 embryonic stem (ES) cells by electroporation in the ESCore at Washington University School of Medicine. We screened for targeted ES cells by DNA blot of genomic DNA isolated from G418-resistant clones and digested with EcoR I. The blot was hybridized to a probe derived from a 0.5-kb Xba I-EcoR I fragment of the 3′-flanking region of the genomic clone and labeled by the random primer method. After hybridization, the membrane was exposed in a PhosphorImager cassette, and the bands were detected with a PhosphorImager (Molecular Dynamics). Four targeted ES clones were then transiently transfected with pTURBO-Cre in the ESCore at Washington University School of Medicine. The resultant clones were screened by DNA blot of genomic DNA after digestion with EcoR I. Two clones were selected for blastocyst injection by a core facility at Washington University School of Medicine. Chimeric mice resulted from each injection. The chimeras were bred to Black Swiss mice, and the offspring were screened for the targeted allele by PCRs. The genotype of the offspring was confirmed by DNA blot of genomic DNA (see above). Germline transmission occurred with chimeras of both ES cell lines. Initially, mice derived from both ES cell clones were screened and found to have indistinguishable characteristics. The data presented here were generated from mice of a mixed 129 × Black Swiss background. Some of the experiments were repeated with mice breed onto a C57BL/6 background over eight generations. No differences were found with these mice. Animal care and use were in accordance with Institutional Animal Studies Committee guidelines, and our animal protocol was approved by the Institutional Animal Studies Committee. The studies were begun at Washington University School of Medicine and finished at Children's Hospital of Pittsburgh. Animal Studies Committees at both institutions approved the protocol.

Animal diet.

The standard chow was PicoLab 5053 containing 11.9% of the energy as fat, 23.6% as protein, and 64.5% as carbohydrate or JL Rat and Mouse/Auto 4F 5K54 rodent diet (LabDiet, PMI Nutrition International) containing crude 12.2% of the energy as fat, 22.4% as protein, and 65.4% as carbohydrate. Ad libitum access to food and water was allowed. For the feeding experiments, the low-fat and high-fat diets were F5413 and F5144 (Bio-Serv), respectively. The low-fat diet contained 10.8% of the energy as fat, 22.8% as protein, and 66.4% as carbohydrate. The high-fat diet contained 58.0% of the energy as fat, 21.5% as protein, and 20.5% as carbohydrate. The diet containing enterostatin (166 μg/g food) was prepared as described elsewhere (38).

Food intake.

For measurement of food consumption, age- and sex-matched animals of the appropriate genotypes were housed individually in cages with wire-screen floors and no bedding. Preweighed food was placed in Pyrex food cups attached to the cage floor with spring clips. The food was replaced and weighed daily for the duration of the experiment, with careful monitoring of any spillage. For measurement of intake on the low-fat and the high-fat diet, 8-wk old mice of each sex were adapted to the specific diet for 3 wk, and intake was then measured daily for 5 days by weighing the food before it was placed in the cage and after 24 h. To determine whether there was a difference in intake of the low-fat diet after the animals were adapted to the high-fat diet, 8-wk-old mice of each sex were fed the high-fat diet for 3 wk and then switched to the low-fat diet. Food intake was measured daily for 5 days. To determine whether there was a diet preference of Ent^+/+ and Ent^−/− mice in a two-choice macronutrient diet, 8-wk old mice were adapted over 14 days to a cage containing two feeding cups, one containing the low-fat diet and the other the high-fat diet. On the night before the experiment, the mice were fasted but had ad libitum access to water. On the following morning, the feeding cups containing a weighed quantity of the low-fat or the high-fat diet were placed in the cage. The remaining food in each cup was weighed after 4 h. For each feeding paradigm, the kilocalories consumed per kilogram of body weight were calculated.

Injections with enterostatin.

Newborn pups were injected with enterostatin, Ala-Pro-Gly-Pro-Arg, or an unrelated peptide, Pro-Arg-Gly-Ala-Pro, every other day from birth to 10 days of age. After the skin was prepped with alcohol, 5 μg of peptide in 50 μl of saline were injected intraperitoneally with a 26-gauge needle. Cages were monitored at least twice a day for dead pups.

RNA and protein methods.

We isolated total RNA from the pancreas as described previously (19). The integrity of the RNA was assessed by denaturing agarose gel electrophoresis (6). RT-PCR was performed with oligo(dT) using the RETROscript kit (Ambion). The procolipase and colipase cDNAs were amplified using the Advantage cDNA PCR kit (Clontech). The primer sequences are given in Table 1. After the PCR, the samples were divided in half, and one-half was digested with Sma I. All samples were then separated by agarose gel electrophoresis, and the bands were detected by ethidium bromide staining.

Table 1.

Primers for mutation and for genotyping

	Sequence
Ent−/−
Sense	5′-GTGGCCTATGCAGGTCTTATTATCAACCTGGTAAGG-3′
Antisense	5′-GATAATAAGACCTGCATAGGCCACTGCCAAGGAGGGA-3′
Clps sense	5′-CTTTAAAGGCTCTCTCCCTCACTTGGC-3′
WT antisense	5′-GATAATAAGACCCCGGGGTCCGGGAGC-3′
Ent−/− antisense	5′-CCTTACCAGGTTGATAATAAGACCTGCA-3′
RT-PCR
Sense	5′-CTTGCCTTCTGCTGTCTGAACTTCCAG-3′
Antisense	5′-AACGGGCAATGCCCAGGATGGTGTCAT-3′

Ent^−/−, enterostatin-deficient; Clps, procolipase.

Protein extracts were made from the pancreas of adult mice as described elsewhere (5). The protein content of the extracts was determined by the bicinchoninic acid method (Pierce). After separation of 20 μg of protein by SDS-PAGE, immunoblot for colipase was done as described elsewhere (31). Fresh samples were assayed for total lipase activity, and heat-inactivated samples were assayed for the presence of colipase. Lipase and colipase were assayed by the pH-stat method, with tributyrin emulsified in 4 mM sodium taurodeoxycholate as substrate (29). Recombinant PTL was used in these assays (46).

Fecal fat analysis.

Eight-week-old female mice were adapted to the high-fat or the low-fat diet for 1 wk. The mice were then placed in a cage with a metabolic screen. They were given water but no food during a 4-h collection of feces. The collected stool was dried to a constant weight, and the fats were extracted from 100 mg of dry stool with chloroform-methanol (2:1) as described previously (28).

Analysis of lipid classes.

Extracted fecal fats were dissolved in 100 μl of chloroform, and 10 μl of each sample were spotted onto a silica G TLC plate. A standard mixture containing 10 μg each of oleate, monoolien, 1,2-diolein, 1,3-diolein, and triolein was also spotted. The plate was developed in two stages [chloroform-methanol-acetic acid (98:2:1) to 17 cm followed by hexane-ethyl ether-acetic acid (94:6:0.2) to the top of the plate] and air-dried, and the plate was dipped in a solution of 3% cupric acetate in 8% phosphoric acid for 3 s and charred at 130°C for 30 min to develop the image (2, 3). The intensity of the spots was determined by densitometry, and the ratio of cholesterol to fatty acid was determined for each sample.

Serum chemistries.

Before blood collection, mice were fasted for 5 h. Blood was obtained from the retroorbital venous plexus with a heparinized capillary tube. Glucose was measured with a Medisense Precision QID meter (Abbott Laboratories, Abbott Park, IL). Insulin was measured with the Rat/Mouse Insulin ELISA kit (Linco Research, St. Charles, MO). Cholesterol was measured with the Cholesterol E kit (Wako Chemicals, Richmond, VA). Triglycerides were measured with the Triglyceride Assay kit (Cayman Chemical, Ann Arbor, MI). In each case, the manufacturer's instructions were followed. HDL cholesterol was measured with lipid panel test strips using a CardioCheck analyzer following the manufacturer's instructions (Polymer Technology Systems, Indianapolis, IN). Non-HDL cholesterol was determined by subtraction of HDL cholesterol from total cholesterol.

Hormone analysis.

Insulin, amylin, leptin, ghrelin, gastric inhibitory polypeptide (GIP), glucagon-like peptide 1 (GLP-1), peptide YY (PYY), and pancreatic polypeptide (PP) were measured in mice that were weaned to the high-fat diet and acclimated for 4–5 wk. Blood was drawn in the morning after an overnight fast. After 2 days of recovery, the mice were fasted overnight, and the high-fat diet was provided on the following morning. Blood was drawn for analysis 1 and 4 h after food was introduced. The hormone levels were determined using a mouse gut hormone panel (LINCOplex) kit following the manufacturer's directions. The analytes were measured using a Luminex instrument (model 200).

Statistical analysis.

The data were analyzed by Student's t-test or the Mann-Whitney rank sum test. One-way ANOVA was used for pairwise multiple comparisons. Two-way ANOVA followed by the Holm-Sidak method for pairwise multiple comparisons, with the significance value for multiple comparisons set at 0.05, was used to compare the effect of two different factors on the means of multiple groups. The survival curves were analyzed by the Kaplan-Meier method. A log-rank test was used to compare the survival rates between the groups. The SigmaStat statistical package was used for all calculations.

RESULTS

Effect of endogenous enterostatin on mice.

Although the effect of enterostatin has been demonstrated in a variety of species, few studies have been done in mice. Before embarking on a study of Ent^−/− mice, we confirmed that endogenous enterostatin had a measurable effect in mice. Eight-week-old male mice were adapted to the high-fat diet for 7 days and then divided into two groups: the diet was continued without supplementation for the first group, and enterostatin was mixed in the food of the second group. When enterostatin was included in the diet, the mice showed a slower rate of weight gain (Fig. 1). By day 30, the mice fed enterostatin were significantly smaller (P = 0.002), and after 50 days weight loss was significant (P ≤ 0.001).

Fig. 1.

Growth of mice fed the high-fat diet supplemented with enterostatin. Eight-week-old C57BL/6 male mice were fed the high-fat diet with and without enterostatin (166 μg/g food) starting at time 0 and weighed at intervals. Values are means ± SD; n = 6 mice in each group.

Creation of an Ent^−/− mouse.

To create a targeting vector with a deletion of the codons encoding enterostatin, we started with a 5-kb BamH I-Not I fragment of the Clps gene subcloned into pGEM11. We previously showed that the fragment contains the entire gene, including ∼1 kb of promoter region, all three exons, the intervening introns, and ∼1 kb of 3′-flanking sequence (5). Overlapping oligonucleotides were designed to delete the 15-bp encoding enterostatin from the BamH I-Not I clone by PCR mutagenesis, as described in methods (Table 1, Fig. 2,A and B). Clones containing the deletion were identified by restriction mapping and confirmed by oligonucleotide sequence analysis.

Fig. 2.

Strategy for creating enterostatin-deficient mice. A: double-stranded DNA sequence of the region surrounding the nucleotides encoding enterostatin (identified by the smaller font). Uppercase letters, exon sequence; lowercase letters, intron sequence. Sequences of PCR primers used to create the deletion are presented above and below homologous genomic DNA sequence. Nucleotides encoding enterostatin (dashed lines) were missing from each primer. Hybridization of the 5′-end of each primer to regions flanking the deletion allows for overlap of primers. B: amino acid sequence for wild-type and enterostatin-deficient (mutant) procolipase in the 1-letter amino acid code starting with the signal peptide and continuing through the first few residues of the NH₂ terminus of colipase. Top: wild-type sequence, with enterostatin residues in gray letters; bottom: mutant sequence. C: schematic of homologous recombination event that occurred between the wild-type allele and the targeting vector to create the mutant allele. PGK-neoLoxP cassette cloned into the Xba I site in intron 1 contained flanking EcoR I sites. When the cassette was removed by Cre recombinase, an EcoR I site was left in intron 1.

The mutant Clps gene was cloned into the p1338 vector. We first removed the PGK-neoLoxP cassette from p1338 with EcoR I, purified the fragments from an agarose gel, and religated the EcoR I ends of the vector with T4 ligase. We then isolated a 2.4-kb Xba I fragment of the Clps gene that contains exons 2 and 3 from the BamH I-Not I Clps clone by restriction with Xho I and Apa I followed by agarose gel electrophoresis (Fig. 2C). This cDNA was ligated into the Xho I and Apa I sites of p1338 to create the 3′ arm of the targeting construct. We next ligated the 5′ arm into the vector by isolating the 1.4-kb BamH I-Xba I fragment containing the promoter region and exon 1 from the Clps gene. The isolated fragment was cloned into the BamH I and Spe I sites of p1338 containing the 3′ arm. Finally, the PGK-neoLoxP cassette was cloned back into the EcoR I site of p1338 containing both arms. The presence of each fragment and the orientation of the 5′ and 3′ arms were confirmed by restriction digest and by dideoxynucleotide sequence analysis. The final construct contained a PGK-neoLoxP cassette in a unique EcoR I site in the first intron (Fig. 2C). This allowed selection of transfected ES cells and the ability to remove the cassette from the transfected ES cells to prevent interference of the cassette with expression of the mutated gene. The construct was introduced into ES cells, and transgenic mice were created as described in methods.

Confirmation of the mutant allele.

Mice were initially screened for the mutant allele by PCR of genomic DNA. Because the products from the wild-type and mutant alleles were close in size, we used two separate PCRs. Each contained a primer derived from the sequence upstream from the mutation, Clps sense (Table 1, Fig. 3A). For the wild-type assay, an antisense primer with a 3′-end that included the sequence encoding enterostatin was used (Table 1, Fig. 3A). The mutant assay utilized an antisense primer that overlapped the sequence on either side of the region encoding enterostatin (Table 1, Fig. 3A). With these two reactions, we were able to identify the wild-type and mutant alleles and genotype the mice (Fig. 3B). We confirmed the PCR results by DNA blot of genomic DNA digested with EcoR I (Fig. 3C). The blot was hybridized to a probe derived from a 0.5-kb Xba I-EcoR I fragment from the 3′ end of the genomic clone (Fig. 2C). Because an EcoR I site was introduced into intron 1 of the mutant gene, the probe hybridized to a 3.0-kb fragment of the mutant allele. Since the EcoR I site is not present in intron 1, the probe hybridized to a fragment from the wild-type allele that migrated with the 23-kb marker (Fig. 3C).

Fig. 3.

Confirmation of mutant mRNA encoding colipase in the pancreas of enterostatin-deficient (Ent^−/−, mutant) mice. A: DNA sequence around the region encoding enterostatin (shown in smaller font) for wild-type and mutant alleles. Location of PCR primers is indicated by arrows. Wild-type antisense primer has its 3′ end in the region encoding enterostatin and cannot amplify the mutant allele. Ent^−/− antisense primer overlaps the deletion site and cannot efficiently prime from the wild-type allele. Sma I site in the region encoding enterostatin is identified by a horizontal line over the nucleotides. B: genomic DNA was isolated from the tail of mice. Primers in A and Table 1 were used to amplify wild-type (WT) and mutant [knockout (KO)] alleles. Because products were quite close in size, 2 separate reactions were done and separated by agarose gel electrophoresis followed by ethidium bromide staining. Corresponding reactions for Ent^+/+, Ent^+/−, and Ent^−/− mice are indicated below each lane. C: blot of genomic DNA. Tail DNA was digested with EcoR I and separated by agarose gel electrophoresis. After transfer of the DNA to a membrane, products were hybridized to a radiolabeled Xba I-EcoR I fragment from the 3′ end of the gene (Fig. 2B). Genotype is indicated below each lane. Positions of wild-type and mutant fragments are marked by arrows. D: total RNA was isolated from the pancreas and subjected to RT-PCR using primers that amplify the region encoding enterostatin (Table 1, RT-PCR sense and antisense). After RT-PCR, half of the reaction was subjected to restriction digestion with Sma I and separation by agarose gel electrophoresis. Nucleotide sequence around the Sma I site is shown above the gel, with bases encoding enterostatin underlined. Enterostatin-deficient allele does not contain the Sma I site, and digestion products will not be seen. Genotype and the presence (+) or absence (−) of Sma I digestion are indicated below each lane. Sizes of the bottom 3 markers are shown at left.

For the RT-PCR assay, we took advantage of an Sma I site that was removed from the DNA by deletion of the codons for enterostatin (Fig. 3, A and D). cDNA was made using total RNA from the pancreas of wild-type and Ent^−/− mice. The portion of the procolipase cDNA containing the codons for enterostatin or the deletion was amplified by primers flanking the region (Table 1). The resulting product was analyzed by agarose gel electrophoresis before and after digestion with Sma I. Only the wild-type fragment digested with Sma I. The fragment from the Ent^−/− mice did not contain an Sma I site as expected, confirming that the nucleotides encoding enterostatin were deleted.

We next determined whether the mutant allele produced functional colipase. We first extracted protein from the pancreas of Ent^+/+ and Ent^−/− mice and assayed for colipase by immunoblot and by an activity assay (Fig. 4). Protein immunoblot with an anti-human procolipase antibody detected a single major band in both samples (Fig. 4A). The band in the Ent^−/− sample migrated slightly faster than the band in the wild-type sample, suggesting that it has a smaller molecular size. A portion of the extract was tested for the ability to activate bile salt-inhibited PTL. PTL was in excess in both assays. Both extracts were able to reactivate PTL under these conditions (Fig. 4B). The amount of functional colipase in the pancreatic extract from the Ent^−/− mouse was not significantly lower than that in the Ent^+/+ extract (P = 0.114).

Fig. 4.

Functional colipase in the pancreas of Ent^−/− mice. A: protein blot of proteins extracted from the pancreas and separated by SDS-PAGE. Colipase was detected with a polyclonal antibody against human procolipase. B: protein was extracted from the pancreas of 8-wk-old mice, and colipase activity was measured. Values are means ± SD of 4 separate extractions. C: six 8-wk-old female mice were adapted to the low-fat or the high-fat diet for 1 wk before collection of feces. Fecal fat was determined for each diet for each of the enterostatin genotypes. Results from Clps^−/− mice are shown for comparison. There was no significant difference among the enterostatin genotypes on either diet (P > 0.05). D: fecal fats from samples of Ent^+/+ and Ent^−/− mice fed the high-fat diet were separated into lipid classes by TLC. First lane is a mixture of markers. Identity of the lipid standards is shown at left. MG, monoacylglycerol; FA, fatty acids; Ch, cholesterol; 1,2-DG, 1,2-diacylglycerol; 1,3-DG, 1,3-diacylglyderol; TG, triacylglycerol; CE, cholesterol esters. Intervening lanes of lower sample volumes were cropped from the image, and sample lanes were moved next to the marker lane.

Since it is possible that colipase is synthesized but not secreted from the pancreas in sufficient amounts to prevent fat malabsorption, we directly measured fat absorption in the Ent^−/− mice. On the low-fat or the high-fat diet, the amount of fecal fat did not significantly differ among the Ent^+/+, Ent^−/+, and Ent^−/− mice (Fig. 4C). The results from Clps^−/− mice, which malabsorb fat on the high-fat diet, are presented for comparison. We confirmed the quantitative results by TLC analysis of the lipid species in feces from the mice kept on the high-fat diet (Fig. 4D). The fecal fats from both genotypes contained detectable undigested cholesterol esters but minimal to no detectable levels of partially or undigested dietary triglycerides. Cholesterol was also detected in both samples, but the amounts did not vary appreciably between the two genotypes. The ratio of cholesterol to fatty acids was 2.1 ± 0.1 for Ent^+/+ and 1.9 ± 0.2 for Ent^−/− mice (P > 0.05).

The Ent^−/− mice of both sexes were fertile and showed no obvious deformities or malformations of major organs. Survival was normal in the Ent^−/− newborns compared with the wild-type pups (92% vs. 93%). At weaning, the body weight of the Ent^−/− pups was normal (Fig. 5). When weaned to the high-fat diet, female and male Ent^−/− mice gained weight at the same rate as their Ent^+/+ littermates (Fig. 5). Similarly, no difference in weight gain was observed on the low-fat diet (data not shown).

Fig. 5.

Weight gain of Ent^+/+ and Ent^−/− mice after they were weaned to the high-fat diet. A: mice were weaned to the high-fat diet and weighed weekly for 12 mo. Values are means ± SD; n = 16 male Ent^+/+, 15 male Ent^−/−, 16 female Ent^+/+, and 16 female Ent^−/− mice. There were no significant differences between Ent^+/+ and Ent^−/− mice of either sex. The same results were obtained with the null allele on the C57BL/6 background.

To determine whether enterostatin deficiency alters metabolism, we measured rectal temperatures and selected serum chemistries after the mice had been fed the high-fat diet for 6 mo (Table 2). The body temperature did not differ between Ent^+/+ and Ent^−/− mice. There was also no difference in serum glucose. On the other hand, the insulin levels were significantly higher in male than in female mice (P ≤ 0.001), but there was no effect of genotype (P = 0.079). Similarly, the triglyceride levels did not vary with genotype or sex. The total serum cholesterol levels did vary significantly by genotype, with higher levels in the Ent^−/− than Ent^+/+ mice (P = 0.004). Further analysis of the cholesterol levels revealed that the HDL cholesterol levels were significantly lower in the Ent^−/− mice independent of sex (P = 0.0102) and the non-HDL cholesterol levels were significantly higher in the Ent^−/− mice independent of sex (P ≤ 0.001).

Table 2.

Body temperature and serum chemistries of mice fed the high-fat diet

	Males		Females
	Ent+/+	Ent−/−	Ent+/+	Ent−/−
Temperature, °C	37.7±0.4	37.8±0.5	37.6±0.3	37.5±0.4
Glucose, mg/dl	159±29	162±42	163±30	165±18
Insulin, ng/ml	11.4±6.0*	12.8±5.9*	2.7±1.2	3.7±1.8
Triglyceride, mg/dl	180±37	207±87	186±31	188±46
Cholesterol, mg/dl
Total	208±43	257±55†	209±28	265±65†
HDL	168±37	112±36‡	139±36	129±27‡
Non-HDL	39±17	144±52§	70±33	136±52§

Values are means ± SD from 6-12 animals per measurement. Mice were weaned to the high-fat diet, and body temperature and serum chemistries were measured at 300 days of age. Non-HDL cholesterol was calculated by subtraction of HDL cholesterol from total cholesterol.

^*Significantly different from females (P < 0.001).

^†P = 0.004, Ent^−/− vs. Ent^+/+ independent of sex.

^‡P = 0.0102, Ent^−/− vs. Ent^+/+ independent of sex.

^§P < 0.001, Ent^−/− vs. Ent^+/+ independent of sex.

We next determined whether enterostatin deficiency had an effect on food consumption. Eight-week-old mice were adapted to the high-fat or the low-fat diet for 3 wk. Intake was then measured each day over a 5-day period. Neither sex nor enterostatin deficiency affected food intake on either diet (Fig. 6).

Fig. 6.

Energy intake of Ent^+/+ and Ent^−/− mice fed the high-fat and the low-fat diet. Eight-week-old mice of each sex were adapted to the diet for 3 wk, and intake was measured daily for 5 days by weighing the food before it was placed in the cage and after 24 h. Values are means ± SE; n = 30 mice in each group. There were no significant differences between diets, between sexes, and between genotypes. The same results were obtained with the null allele on the C57BL/6 background.

Next, we tested the effect of adaptation to the high-fat diet on intake when the mice were switched to the low-fat diet. This paradigm investigates the possibility that enterostatin affects satiety when the animal is offered a low-fat diet after adaptation to a high-fat diet. Mice were adapted to the high-fat diet for 3 wk, and then intake of the diet was measured over 5 days. The mice were switched to the low-fat diet, and intake was measured over the next 5 days. There was no significant effect of genotype or diet on intake (Fig. 7). In the female mice, food intake was higher after the switch to the low-fat diet (P ≤ 0.001). Enterostatin deficiency did not affect the increased intake of the female mice.

Fig. 7.

Intake of the low fat-diet after adaptation to the high-fat diet by Ent^+/+ and Ent^−/− mice. Eight-week-old mice of each sex were adapted to the high-fat diet for 3 wk and then switched to the low-fat diet. Food intake was measured daily for 5 days. Values are means ± SE of cumulative intake; n = 25 Ent^+/+ males, 30 Ent^−/− males, 45 Ent^+/+ females, and 45 Ent^−/− females fed the high-fat diet and 25 Ent^+/+ males, 30 Ent^−/− males, 45 Ent^+/+ females, and 45 Ent^−/− females fed the high-fat diet. There was a significant effect of diet for the female mice, with greater intake of the low-fat than the high-fat diet: *P = 0.00001. There was no effect of genotype (P = 0.68) or interaction between genotype and diet (P = 0.25).

Since enterostatin inhibits fat intake when rats are given a choice of macronutrient diet, we measured the effect of enterostatin deficiency on choice between the low-fat and the high-fat diet (36). Mice were adapted to a two-choice diet for 14 days. On the night before the experiment, they were fasted. In the morning, they were provided with two feeding cups, one containing the high-fat diet and the other the low-fat diet. Their intake of the two-choice diet was measured over a 4-h period (Fig. 8). Although it appears that mice of both sexes and genotype preferred the high-fat diet, there was no significant effect of genotype, diet, or genotype × diet by ANOVA. Importantly, enterostatin deficiency did not affect diet choice after the effects of differences in diet were taken into account (P = 0.798 for males and P = 0.816 for females).

Fig. 8.

Diet preference of Ent^+/+ and Ent^−/− mice in a 2-choice macronutrient diet. Eight-week-old mice were adapted over 14 days to a cage with 2 feeding cups, one containing the low-fat diet and the other the high-fat diet. On the night before the experiment, mice were fasted but had ad libitum access to water. On the following morning, feeding cups containing a weighed quantity of the low-fat or the high-fat diet were placed in the cage. Remaining food in each cup was weighed after 4 h, and kilocalories consumed per kilogram of body weight were calculated. Values are means ± SE of intake over 4 h; n = 14 Ent^+/+ males, 25 Ent^−/− males, 13 Ent^+/+ females, and 19 Ent^−/− females fed the high-fat diet and 14 Ent^+/+ males, 25 Ent^−/− males, 13 Ent^+/+ females, and 19 Ent^−/− females fed the low-fat diet. There was no significant effect of genotype or diet or interaction between genotype and diet for either sex.

To determine whether other hormones compensated for enterostatin deficiency, we measured the serum levels of insulin, amylin, leptin, ghrelin, GIP, GLP-1, PYY, and PP in mice after fasting and after consumption of the high-fat diet. The mice were weaned to the high-fat diet. At 8 wk of age, they were fasted overnight and blood was drawn. After 2 days of recovery, they were again fasted overnight and given the high-fat diet in the morning. Blood was drawn 1 and 4 h after the diet was placed in the cage. Leptin, insulin, GIP, PYY, and PP significantly increased after the animals were fed, but there was no difference between genotypes (Table 3). Amylin, ghrelin, and GLP-1 did not change with feeding.

Table 3.

Gut hormone levels in mice fed the high-fat diet

	Ent+/+			Ent−/−
	Fast	1 h	4 h	Fast	1 h	4 h
Leptin, ng/ml	8.4±2.8a	9.9±3.0	22.8±4.7a	8.1±1.2b	9.2±3.0	24.7±3.1b
Amylin, ng/ml	0.10±0.05	0.12±0.03	0.11±0.06	0.10±0.08	0.12±0.09	0.11±0.08
Insulin, ng/ml	0.25±0.03c	0.68±0.13c	0.73±0.11c	0.24±0.06d	0.85±0.20d	0.60±0.09d
GLP-1, pg/ml	90.1±2.0	89.2±4.6	94.3±8.7	94.0±9.2	89.2±4.8	85.6±10.6
GIP, ng/ml	0.19±0.06e	1.44±0.13e	0.82±0.13e	0.18±0.05f	1.40±0.20f	0.64±0.14f
Ghrelin, pg/ml	17.4±3.3	16.6±1.8	18.1±3.2	16.3±0.5	16.5±0.8	17.1±2.4
PYY, pg/ml	77.5±3.5g	183±24.9g	138±8.2g	79±6.5h	220±24.8h	120±12.7h
PP, pg/ml	NDi	89.6±19.3i	ND	NDj	92.2±15.3j	ND

Values are means ± SE; n = 11. Eight-week-old Ent^+/+ and Ent^−/− mice were fasted overnight, and blood was drawn. Mice were allowed to recover for 2 days and again fasted overnight and fed the high-fat diet in the morning. Blood was drawn 1 and 4 h after the start of feeding. Amount of hormone was determined by Luminex.

GLP-1, glucagon-like peptide 1; GIP, gastric inhibitory polypeptide; PYY, peptide YY; PP, pancreatic polypeptide; ND, none detected.

GIP levels were detected in 3 of 11 Ent^+/+ and in 7 of 11 Ent^−/− mice after fasting. Leptin levels differ significantly between fasting and 4 h of feeding:

^aP < 0.001 for Ent^+/+ mice and

^bP < 0.001 for Ent^−/− mice. Insulin increased with feeding:

^cP < 0.001 at 1 h and P = 0.004 at 4 h for Ent^+/+ mice and

^dP < 0.001 at 1 h and P = 0.001 at 4 h for Ent^−/− mice. GIP increased with feeding:

^eP < 0.001 at 1 h and P = 0.048 at 4 h for Ent^+/+ mice and

^fP < 0.001 at 1 h and P = 0.004 at 4 h for Ent^−/− mice. PYY increased with feeding:

^gP < 0.001 at 1 h and P < 0.001 at 4 h for Ent^+/+ mice and

^hP < 0.001 at 1 h and P = 0.01 at 4 h for Ent^−/− mice. PP increased with feeding:

ⁱP < 0.001 at 1 h for Ent^+/+ mice and

^jP < 0.001 at 1 h for Ent^−/− mice. No significant effects of genotype were observed for any of the hormones.

Since survival was decreased in the Clps^−/− pups, we sought to determine whether enterostatin contributed to this phenotype. We mated Clps^−/− mice and intravenously injected the resultant Clps^−/− pups with enterostatin or an unrelated peptide every 2 days starting from birth and continuing until 10 days of life (Fig. 9). The enterostatin injections improved the survival of the Clps^−/− pups compared with the pups injected with the unrelated peptide (P ≤ 0.001).

Fig. 9.

Survival of Clps^−/− pups after treatment with enterostatin. Pups were injected intraperitoneally with 5 μg of enterostatin or 5 μg of an unrelated, control peptide in 50 μl of saline every other day starting on postnatal day 1 and ending on postnatal day 10. Pups were monitored twice daily for survival. Differences between the 2 groups were significant (P ≤ 0.001).

To determine whether enterostatin injections altered the weight gain of the survivors, we weighed the pups at 4 and 10 days of age. At 4 days of age, weight gain was greater in the enterostatin-treated Clps^−/− pups than in the control peptide-treated Clps^−/− pups: 3.3 ± 0.6 vs. 2.4 ± 0.6 g (P = 0.001). By 10 days of age, the weight of the enterostatin-treated Clps^−/− pups was not different from that of the control peptide-treated Clps^−/− pups: 5.9 ± 0.8 and 5.1 ± 1.4 g (P = 0.17).

DISCUSSION

A number of studies from various investigators have implicated enterostatin as a mediator of dietary fat intake (12, 13, 23). Exogenous administration of enterostatin selectively suppresses fat intake in rats after central or peripheral dosing (20–23, 25, 27, 33, 41). On the basis of these studies, we predicted that deletion of enterostatin would have effects on fat intake and weight gain. Our present studies did not substantiate these predictions. In fact, mice lacking enterostatin have no obvious alterations in appetite or weight regulation, allowing us to conclude that endogenous enterostatin is not absolutely required for satiety or the regulation of fat intake.

Several possibilities could explain the lack of detectable effects of enterostatin deficiency on food intake or weight gain. Endogenous enterostatin may not have a role in energy regulation, and the effects of enterostatin administration to animals represent a pharmacological perturbation of food intake. The observation that endogenous enterostatin is effective when delivered by multiple routes argues against the delivery method perturbing the system. Still, the delivery of exogenous enterostatin may not reflect the physiological functions of enterostatin. Since there are fewer studies of the effects of enterostatin in mice (27), it is possible that rats, the most common animal model for enterostatin research, have a more robust response to enterostatin. Alternatively, other hormones may uniquely adapt to congenital enterostatin deficiency through upregulation during development and compensate for the absence of enterostatin. In a single-feeding paradigm, we did not find any significant difference in multiple gut hormones between wild-type and enterostatin-deficient mice. Still, it is highly likely that compensatory mechanisms exist. In either case, the effect of enterostatin deficiency would be obscured. In this regard, our observations are similar to the lack of obvious phenotype in mice deficient in the agouti-related protein or neuropeptide Y (7, 37). As with enterostatin, abundant evidence suggests that injection of these peptides has effects on feeding behavior. If there is redundancy, phenotypes may be revealed only in mice with null alleles for two or more factors (4).

The Ent^−/− mice did differ from Ent^+/+ mice in their serum cholesterol profile. The Ent^−/− mice had higher total serum cholesterol, lower HDL cholesterol, and higher non-HDL cholesterol, a profile that puts the Ent^−/− mice at increased risk for cardiovascular disease. This observation supports an earlier report demonstrating that orally administered enterostatin decreased serum cholesterol at doses that did not affect food intake or involve corticosteroid pathways (43). Further investigations are required to determine the mechanism by which enterostatin influences cholesterol metabolism. The possibilities include altered liver synthesis of cholesterol, changes in the regulation of reverse cholesterol transport, and changes in the protein composition of the lipoprotein particles. The finding of no obvious impairment in food intake in the Ent^−/− mice supports the notion that the effect of enterostatin on cholesterol levels is not secondary to enterostatin's anorexogenic effect. A recent study demonstrated that oral administration of enterostatin did not reduce serum cholesterol levels if it was administered simultaneously with a cholecystokinin receptor antagonist, lorglumide, suggesting that enterostatin may lower serum cholesterol levels through a cholecystokinin-dependent mechanism (44).

In the present study, we observed several differences between the male and female mice independent of the genotype. First, insulin levels were higher in the male mice fed the high-fat diet (Table 2). This suggests that the male mice developed insulin resistance at a faster rate than the females. Other studies reported similar findings in rats and mice (14, 15, 17, 18). In these studies, females did not develop insulin resistance or the development lagged behind the appearance of insulin resistance in males. We also noted a sex difference in our study to determine whether enterostatin influences hyperphagia when mice are adapted to the high-fat diet and switched to the low-fat diet (Fig. 7). Intake on the low-fat diet was increased only in female mice. Although most feeding studies utilize male animals, a few have compared males and females and have demonstrated differences in feeding behavior between the sexes. The interaction of estrogen or estradiol and testosterone may be critical for these sex differences (16, 45). Enterostatin status did not influence either observation.

Because the Clps^−/− mice had a striking phenotype of decreased pup survival, we attempted to determine whether enterostatin deficiency contributed to the decreased survival and growth failure of the Clps^−/− mice. Thus we injected Clps^−/− pups with enterostatin to determine whether growth and survival were affected. We found that endogenous enterostatin significantly improved survival of the Clps^−/− pups. Furthermore, enterostatin injections improved weight gain at 4 days but did not alter weight at 10 days compared with Clps^−/− pups injected with a control peptide. Because most of the Clps^−/− pups died between 4 and 10 days of age, the pattern of weight gain likely reflects our previous observations that nonsurviving Clps^−/− pups were significantly smaller than the survivors at 4 days of age and that the proportion of nonsurvivors was greater in the control group at 4 days of age (5).

These findings suggest that enterostatin may critically influence energy homeostasis early in life. The effect could occur through alterations in feeding behavior or in respiratory quotient or metabolic rate, all of which can be affected by enterostatin (1, 24, 38). Alternatively, enterostatin may alter response to a stress, consistent with the report that enterostatin increases sympathetic activity (34). The amygdala is a site in the brain that is important for pup survival by directing the learning of feeding by the pup (42). Enterostatin is known to influence feeding behavior by targeting the amygdala and by activating amygdala neurons that have functional and anatomic projections to feeding centers in the brain, such as the arcuate nucleus (26, 27). Such an activation of feeding pathways may be important for pup survival. Clps^−/− pups have an additional stress, i.e., energy loss through steatorrhea, that is not present in Ent^−/− pups. The effect of enterostatin in Clps^−/− pups suggests that, even in the face of the apparent normalcy of Ent^−/− mice, subtle changes in the physiology of adult enterostatin-deficient mice may become apparent with additional testing. Even so, we can conclude that enterostatin is not essential for the regulation of fat intake or of growth under the physiological conditions tested in this study.

GRANTS

This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-53100.