The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Department of Molecular Pharmacology, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. E-mail: firstname.lastname@example.org
Objective: To compare the effect of voluntary exercise on body weight, food consumption, and levels of serum proteins between wild-type and carboxypeptidase E-deficient (Cpefat/fat) mice.
Research Methods and Procedures: Study 1 consisted of three groups of female mice: Cpefat/fat mice with continuous access to exercise wheels for 3 weeks (n = 4); wild-type C57BKS mice with access to exercise wheels for 3 weeks (n = 4); and sedentary Cpefat/fat mice (n = 3). Activity, body weight, and food consumption were monitored for this period and a subsequent 9-week period without exercise wheels. Study 2 consisted of four groups of male mice (n = 6 to 7 each): Cpefat/fat mice with exercise wheels, wild-type mice with exercise wheels, and Cpefat/fat and wild-type mice without exercise wheels. Body weight and food consumption were measured over 4 weeks. Sera were collected, and the protein profile was determined by 2-dimensional gel electrophoresis and mass spectrometry.
Results: Cpefat/fat mice were moderately hyperphagic but lost weight during the initial exercise period because of greater energy expenditure. The effect of exercise was temporary, and the mice gained weight after the second week. Several serum proteins were found to be altered by exercise: haptoglobin was decreased by exercise in Cpefat/fat mice, and several kallikreins were increased by exercise in wild-type mice.
Discussion: The access to exercise wheels provided an initial weight loss in Cpefat/fat mice, but this effect was offset by elevated food consumption. The serum proteomics results indicated that Cpefat/fat and wild-type mice differed in their response to exercise.
Obesity is recognized as a chronic disorder that is a major risk factor for several human diseases, including non—insulin-dependent diabetes, cardiovascular disease, hypertension, and cancer (1,2). The fat mutation, discovered in 1972 in an inbred mouse, produces late-onset obesity and sterility (3). When the fat mutation was backcrossed onto the C57BKS strain, males (but not females) developed hyperglycemia at ∼14 weeks of age (3). The fat mutation was mapped to the carboxypeptidase E (CPE)1 gene, and a point mutation was found in the coding region; this mutation eliminates enzyme activity (4). CPE is the major carboxypeptidase that functions in the biosynthesis of neuroendocrine peptides, together with one or more endopeptidases and other enzymes (5,6,7).
The fat mutation causes a decrease in the levels of fully processed peptides and an increase in the levels of C-terminally extended peptide processing intermediates in brain, pituitary, and pancreatic islets (4,8,9,10). The precise cause of obesity in Cpefat/fat mice is not known, although it is likely caused by reduced levels of one or more peptides that function in the control of body weight. The regulation of body weight is complex, involving food intake, energy expenditure, and nutrient partitioning (1). Eight-week-old Cpefat/fat mice have been reported to eat the same amount of food as wild-type littermates, suggesting that the problem is with energy expenditure or storage (3). However, later ages were not previously examined for food intake. Nillni et al. (11) have found that Cpefat/fat mice are defective in their ability to thermoregulate.
Regular exercise has positive effects on a variety of biological systems (12,13). The focus of this study was to examine the effect of voluntary exercise on the maintenance of body weight by providing Cpefat/fat mice with activity wheels. If these mice were overweight because of reduced physical activity, it was expected that they would exercise less than wild-type littermates, as reported for genetically obese yellow agouti and ob/ob mice (14). Instead, the Cpefat/fat mice ran two to three times more than wild-type mice and temporarily reduced their body weight. Based on this result, additional experiments were performed to evaluate the effect of voluntary exercise on food consumption and serum levels of glucose. In addition, a proteomics approach was used to examine the effect of exercise on a variety of serum proteins in both Cpefat/fat and wild-type mice.
Research Methods and Procedures
Wild-type (C57BKS) and mutant Cpefat/fat mice were initially obtained from the Jackson Laboratory (Bar Harbor, ME) and bred within the barrier facility at Albert Einstein College of Medicine. Food (rodent chow 5058; Labdiet, St. Louis, MO) and water were provided ad libitum; lights were on from 7:00 am to 7:00 pm. The experimental protocol was approved by the committee for animal experimentation of the Albert Einstein College of Medicine. At the start of each experiment, animals were 11 weeks old.
For evaluation of the effect of exercise on the body weight of female mice, three groups were tested. One group of Cpefat/fat mice (n = 3) was housed in standard mouse cages (28 × 18-cm floor, 12 cm height) without access to exercise wheels (“sedentary” group). The other two groups consisted of Cpefat/fat mice (n = 4) and wild-type mice (n = 4) individually housed in an activity cage (34 × 16-cm floor, 17 cm height) containing an exercise wheel (23.5 cm diameter) that recorded revolutions. After 3 weeks in the activity cage, the animals were transferred to a standard mouse cage for 9 weeks. Body weight and food consumption were measured every 2 days throughout the entire period.
The effect of voluntary exercise on male mice was carried out for 4 weeks. The male mice were divided into four groups. As with the female mice, the sedentary male wild-type (n = 7) and Cpefat/fat mice (n = 6) were individually housed in standard mouse cages. The groups of wild-type and Cpefat/fat mice (n = 7 each) with exercise wheels were housed individually in standard rat cages (48 × 27-cm floor, 20 cm height) equipped with small exercise wheels (14.5 cm diameter). Body weight and food consumption were measured every 2 days. After 4 weeks, the animals were killed by CO2 asphyxiation between 2:00 pm and 4:00 pm. The blood was collected, allowed to clot for 30 minutes at room temperature, and centrifuged at 1000g for 5 minutes, and the serum was stored at −80 °C until analysis. The fat pads (inguinal, mesenteric, perirenal, reproductive, and scapular) were dissected and weighed. Serum glucose was measured by the glucose oxidase method, using the glucose analyzer II (Beckman Instruments, Palo Alto, CA).
Two-Dimensional Gel Electrophoresis
Sera obtained from male mice were combined with the following reagents to give the indicated final concentrations: urea (7 M), thiourea (2 M), dithiothreitol (DTT; 65 mM), 3-[(cholamidopropyl) dimethylammonio]-1-propane sulfonate (CHAPS) (4% w/v), Tris base (40 mM), and immobilized pH gradient (IPG) buffer pH 4 to 7 (2% vol/vol), obtained from Amersham Biosciences (Uppsala, Sweden). After 1 hour of stirring at room temperature, 120 μg of serum protein was solubilized in 250 μL of a rehydration solution containing 8 M urea, 2% CHAPS, 20 mM DTT, a trace of bromophenol blue, and 0.5% vol/vol IPG buffer pH 4 to 7. Rehydration of IPG strips and isoelectric focusing (IEF) were carried out according to the programmed settings: 30 V for 6 hours; 60 V for 6 hours; 200 V for 1 hour; 500 V for 1 hour; 1000 V for 1 hour; and 8000 V for 3 hours (15). After IEF, the IPG strips were immediately equilibrated two times for 15 minutes in 10 mL of Tris-HCl buffer (50 mM, pH 8.8), 6 M urea, 30% w/v glycerol, 2% w/v sodium dodecyl sulfate, and a trace of bromophenol blue. DTT (1% w/v) was added to the first equilibration step, and iodoacetamide (4% w/v) was added to the second equilibration step. The two-dimensional sodium dodecyl sulfate polyacrylamide gel electrophoresis was stopped after 5 hours, and proteins were visualized by silver staining (16). Spot detection, spot editing, pattern matching, and spot quantification were performed using Genomic Solutions Investigator HT Analyzer v3.01 software (Genomic Solutions, Ann Arbor, MI). Statistical analysis was performed using Student's t test (Sigmaplot). The differentially expressed protein spots (p < 0.01, n ≥ 3) were excised from the two-dimensional gel electrophoresis gel, destained, reduced and alkylated, and digested with modified trypsin (16).
Matrix-Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry and Database Search
The tryptic peptides were desalted with ZipTip-C18 reverse-phased pipette tips, and 0.5 μL was mixed with 1.5 μL of a saturated solution of α-cyano-4-hydroxycinnamic acid in 50% acetonitrile/0.1% trifluoroacetic acid. Analysis was performed in the delayed-extraction linear positive mode of the Voyager-DE STR matrix-assisted laser desorption ionization time-of-flight mass spectrometer (MALDI-TOF MS; PerSeptive Biosystems, Framingham, MA). For each sample, the spectra produced from 200 laser shots were accumulated. Internal calibration of the MALDI-TOF mass data used the masses of two trypsin autolysis products: [M+H]+ 842.51 and [M+H]+ 2211.10. Identification of proteins was performed with the programs ProFound (http:220.127.116.11prowl-cgiProFound.exe) and MS-FIT (http:prospector.ucsf.edu), using the criteria of >25% coverage and a minimum of four matching tryptic fragments with a mass tolerance of 100 ppm.
Female Cpefat/fat mice given access to an exercise wheel showed a reduction in body weight of slightly >10% within 1 week (Figure 1). After the first week, there was a slight increase in body weight during the remainder of the time with the wheel, although this change was not statistically significant. In contrast, Cpefat/fat mice without access to an exercise wheel gained ∼2 g/wk during this 3-week period from age 11 to 14 weeks. After the 3-week period with the exercise wheel, the Cpefat/fat mice gained weight, slowly at first and then more rapidly, and were within a few grams of the body weight of the sedentary Cpefat/fat mice within 4 to 5 weeks after removal of the exercise wheel (Figure 1). Wild-type female mice lost ∼5% body weight during the first 2 days with the exercise wheel but regained it by the end of the 3-week exercise period (Figure 1).
The wild-type mice ran ∼7 to 8 km/wk for the first 2 weeks with the exercise wheel and slightly less for the third week (Figure 2). In contrast, the Cpefat/fat mice ran ∼17 to 22 km/wk during this period (Figure 2). There was a large variability in the running performance of each mouse; two of the Cpefat/Cpefat mice were comparable with the wild-type mice, whereas the other two Cpefat/fat mice used in the study ran extreme amounts. (One Cpefat/fat mouse ran over 6 km in a single night!)
The food consumption of the wild-type mice paralleled their physical activity; the mice ate 20% to 40% more food during the 3-week period with the exercise wheel than during the subsequent 3-week periods (Figure 3). In contrast, the Cpefat/fat mice ate 20% to 40% more food after removal of the exercise wheel relative to the amount they ate while the exercise wheel was present (Figure 3). The amount of food consumed by Cpefat/fat mice that were never provided access to an exercise wheel generally matched that of the exercising Cpefat/fat and wild-type mice (Figure 3).
Because there is a tendency for male Cpefat/fat mice to develop type 2 diabetes (3), the effect of 4 weeks of voluntary exercise was examined in male Cpefat/fat and wild-type mice. For this study, mice were individually housed in rat cages that were outfitted with small exercise wheels or in standard mouse cages without exercise wheels. Although this equipment did not provide quantitation of the amount of exercise, it was advantageous in that many cages could be analyzed simultaneously, and a larger number of animals were tested in a single experiment. The male Cpefat/fat mice with exercise wheels lost weight during the initial 1 to 1.5 weeks with the exercise wheels but gained ∼5%/wk for the remainder of the period (Figure 4). Sedentary Cpefat/fat mice gained 5%/wk during the 4-week period from 11 to 15 weeks of age (Figure 4). Wild-type males also lost ∼5% of their body weight during the initial 1 to 1.5 weeks with the exercise wheels and gained it back during the remainder of the study (Figure 4).
Wild-type mice ate 40% more food when given access to the exercise wheels, and there was no difference between the first 2 weeks and the last 2 weeks of the study (Figure 5). In contrast, Cpefat/fat mice given access to wheels ate equal amounts of food as the sedentary Cpefat/fat mice for the first 2 weeks of the study, but the exercising mice ate 30% more for the last 2 weeks (Figure 5). Interestingly, sedentary Cpefat/fat mice ate 30% to 40% more food than sedentary wild-type mice at all age groups examined in this study (11 to 14 weeks of age).
Analysis of the body fat of the 15-week-old male mice revealed that the Cpefat/fat mice with exercise wheels averaged 5.7 grams of fat, whereas those mice lacking exercise wheels averaged 8.8 grams of fat in the major fat pads (inguinal, scapular, perirenal, reproductive, and mesenteric). When expressed as a percentage of total mass, the sedentary mice had 20.8 ± 2.9% body fat, whereas the mice with wheels had only 15.5 ± 1.5% body fat (SE; n = 7). In contrast, wild-type sedentary mice had only 7.9 ± 0.9% body fat, and wild-type mice with exercise wheels had 6.6 ± 1.0% body fat. All of the various fat pads analyzed were 2- to 3-fold higher in the Cpefat/fat mice compared with wild-type mice.
Because the male, but not female, Cpefat/fat mice develop hyperglycemia at ∼14 to 16 weeks of age (3), serum levels of glucose were measured in the 15-week-old mice. Wild-type mice had serum glucose levels of 170 ± 9 mg/dL for the sedentary mice and 186 ± 10 mg/dL for the mice with exercise wheels. Sedentary Cpefat/fat mice had serum glucose levels of 298 ± 78 mg/dL, whereas those with exercise wheels had levels of 250 ± 42 mg/dL. This difference was not statistically significant at the p < 0.05 level.
A proteomics approach was used to explore whether voluntary exercise affected the levels of serum proteins in wild-type and/or mutant Cpefat/fat mice. Altogether, 20 proteins were detected on two-dimensional gel electrophoresis that were substantially altered by exercise (Figure 6). Several proteins (spots 15 to 20) were present at extremely low levels in the wild-type mice (Figure 6) but were much more abundant in the sedentary Cpefat/fat mice (Figure 7). The levels of all six of these protein spots were substantially decreased by exercise (Figure 7A and 7B). Quantitation of the relative levels of these proteins in three exercising and three sedentary Cpefat/fat mice showed a statistically significant decrease in all six proteins (Table 1). Interestingly, the levels of these proteins in wild-type mice were not significantly altered by exercise (Table 1). Tryptic fingerprinting using MALDI-TOF MS identified all six of these proteins as the β-chain of haptoglobin. Additional proteins were detected that increased with exercise in the wild-type mice (Figure 8). Eight of these spots were identified as kallikreins (kallikrein-5, −11, −16, −22, and −24) or kallikrein-like proteins, such as the α subunit of nerve growth factor (αNGF). Six of the proteins that increased with exercise could not be identified from the tryptic fingerprinting MALDI-TOF MS analysis.
Table 1. Proteins in mice serum that are altered by exercise
Accession no. (NCBI)
WT (con.) (mean ± SD)
WT (ex.) (mean ± SD)
Fat mice (con.) (mean ± SD)
Fat mice (ex.) (mean ± SD)
WT mice (Ex./Con.)
Fat mice (Ex./Con.)
No exercise (Fat/WT)
p < 0.05,
p < 0.01 difference between the two groups used to calculate the indicated ratio: exercise (ex.) vs. sedentary control (con.), or Cpefat/fat (Fat) vs. wild-type (WT).
A major finding of this study was that Cpefat/fat mice voluntarily used an exercise wheel to temporarily reduce their body weight. Despite being overweight at the start of the study (11 weeks of age), the Cpefat/fat mice ran two to three times more than the wild-type mice. The increased exercise was presumably responsible for the initial decrease in body weight observed in both male and female Cpefat/fat mice during the first 1 to 2 weeks with the exercise wheel. Wild-type male and female mice that were offered exercise wheels also decreased their body weight by a small percentage during the first week, but the difference between sedentary and exercising Cpefat/fat mice was greater than the difference between sedentary and exercising wild-type mice.
The Cpefat/fat mice resemble Otsuka Long Evans Tokushima Fatty (OLETF) rats in several ways. Both develop late-onset obesity and hyperglycemia (3,17). OLETF rats given access to exercise wheels voluntarily exercise to reduce their body weight (18), as was also seen in the Cpefat/fat mice. However, OLETF rats exercise to the point of attaining “normal” body weights (18), whereas the Cpefat/fat mice only temporarily reduced their body weight by 5% to 10% when given access to exercise. OLETF rats have a mutation within the cholecystokinin A receptor gene (19), although it is not clear whether this alone explains their phenotype, because mice with a disruption in the cholecystokinin A receptor neither gain weight nor develop hyperglycemia (20). Cholecystokinin is involved in the feedback loop from stomach to brain, which signals when the stomach is full. The biosynthesis of this peptide requires a carboxypeptidase, and in Cpefat/fat mice, the level of mature cholecystokinin is reduced in both brain and intestine (10). Thus, Cpefat/fat mice would be expected to partly resemble the OLETF rats. However, the Cpefat/fat mice have alterations in many other peptides (21), and some of these are presumably involved in the regulation of body weight by mechanisms that differ from those of cholecystokinin.
Body weight is controlled by several mechanisms, many of which involve peptide neurotransmitters (1). Ultimately, body weight is the balance between caloric intake and energy expenditure, either as physical activity or general metabolism. It has been previously reported that Cpefat/fat mice are not hyperphagic at 8 weeks of age, suggesting that the obesity in these mice is caused by energy expenditure (3). However, in this study, sedentary male Cpefat/fat mice were found to be moderately hyperphagic when measured at 11 to 14 weeks of age, and female Cpefat/fat mice exposed to an exercise wheel for 3 weeks ate significantly more than wild-type mice for many weeks after removal of the wheel. While the increased food consumption in the Cpefat/fat mice contributes to their body weight gain, it is unlikely that this alone is responsible for the increase. For example, after removal from the exercise wheel, the female Cpefat/fat mice gained 2 to 3 g/wk but ate only ∼10 grams more food per week than the wild-type mice. Thus, it is likely that a defect in nutrient partitioning contributes to obesity in the Cpefat/fat mice, together with moderate hyperphagia. The recent finding that Cpefat/fat mice are deficient in their ability to maintain body temperature when placed in a cold environment (11) suggests that obesity in these animals results, in part, from a decreased metabolic rate.
Exercise improves glucose homeostasis by a variety of mechanisms, including enhanced insulin action on the glucose transport system, reduced gluconeogenesis, improved blood flow to skeletal muscle, and normalization of the blood lipid profile (22). Male Cpefat/fat mice typically develop type 2 diabetes at ∼14 to 16 weeks of age (3). Although we found a tendency for a reduction in serum glucose levels with exercise, the difference in glucose levels between exercising and sedentary mice observed in this study was not statistically significant. However, a number of serum proteins were significantly affected by exercise in wild-type and Cpefat/fat mice. One protein that migrated on two-dimensional gel electrophoresis as a series of related spots, presumably because of posttranslational modifications, and was consistently higher in the serum of the sedentary Cpefat/fat mice relative to the other three groups was identified as the β chain of haptoglobin. Levels of β-haptoglobin were 55% to 70% lower in the Cpefat/fat mice that exercised compared with sedentary mice. Haptoglobin has been found to be expressed in high levels in white adipose tissue from three different genetically obese mouse lines: db/db, ob/ob, and yellow agouti (23). Increased levels of serum haptoglobin in humans has been linked to obesity (24) and to future weight gain (25). In humans, serum haptoglobin is known to be decreased by long-term running (26). Haptoglobin binds to free hemoglobin formed from hemolysis, and the resulting haptoglobin-hemoglobin complex is subsequently destroyed in the liver (27). Exercise can induce hemolysis, especially in long-distance runners, presumably because of mechanical damage to red blood cells as they pass through capillaries of the foot during the footstrike (28). The decreased levels of haptoglobin observed after exercise in this study may, therefore, reflect the binding of haptoglobin to hemoglobin and clearance by the liver. Alternatively, the lower levels of serum haptoglobin may be caused by the reduction in body fat observed in Cpefat/fat mice with access to exercise wheels. It is possible that both of these factors contribute to the decrease in serum haptoglobin observed in the exercising Cpefat/fat mice in this study.
In addition to haptoglobin, four other proteins were detected that were expressed at higher levels in the serum of sedentary Cpefat/fat mice compared with wild-type mice. Of these, three proteins (spots 2, 3, and 14) could not be identified using the tryptic fingerprinting technique; the fourth (spot 9) was identified as a fragment of kallikrein-16. Although spots 2 and 3 could not be identified in this study, their isoelectric point of 5.5 and molecular weight of 42 to 44 kDa are very close to those for apolipoprotein A-IV, which can be detected on two-dimensional gel electrophoresis of serum (29,30). Hepatic apolipoprotein A-IV gene expression is greatly up-regulated in ob/ob mice compared with wild-type mice (31). Apolipoprotein A-IV is thought to play a role in the regulation of food intake (32). Thus, it is possible that spots 2 and/or 3 represent apolipoprotein A-IV, although without confirmatory sequencing or tryptic fingerprinting data, this association remains speculative.
In this study, 14 protein spots on the two-dimensional gel electrophoresis were found to be significantly elevated in the serum of wild-type mice provided exercise wheels compared with sedentary wild-type mice. Interestingly, a number of these proteins were identified as kallikreins or prokallikreins (these could not be readily distinguished because of the small difference in size), kallikrein fragments, or αNGF, which has substantial amino acid sequence homology to kallikrein-like serine proteases. It is not clear why serum levels of these kallikrein proteins are elevated during exercise in wild-type mice only and not in Cpefat/fat mice. Kallikreins function in the processing of kininogen and are involved in maintaining blood pressure. It is possible that the increase in blood pressure during aerobic exercise causes an increase in circulating levels of kallikreins. Alternatively, chronic exercise typically lowers the resting blood pressure, and it is possible that activation of the kallikrein system is involved in this effect (33,34). Several studies have found that exercise raises the level of kallikrein activity in human or rat serum, although the precise form(s) of kallikrein responsible for this increase were not determined (35,36). The tryptic fingerprinting proteomics approach easily distinguishes the different kallikrein gene products. However, the proteomics approach does not provide a reflection of the enzymatic activity of the protein; kallikreins are initially produced as inactive precursors that are activated by selective proteolysis. Thus, the combination of the proteomics approach used in this study with the enzymatic assays used in previous studies provides a more complete picture of the regulation of these enzymes on exercise.
In summary, we found that genetically obese Cpefat/fat mice temporarily reduced their body weight by 5% to 10% when given an opportunity to exercise, but then proceeded to gain weight during the remainder of the period with exercise wheels. The Cpefat/fat mice were moderately hyperphagic and lost weight during the initial exercise period because of substantially greater energy expenditure. However, the mice compensated by eating more during the third and fourth weeks with exercise wheels, and this contributed to weight gain. Although serum glucose levels were not significantly altered in the exercising vs. the sedentary Cpefat/fat mice, a number of other proteins were found to be substantially altered by exercise in either the Cpefat/fat or wild-type mice. Some of these proteins may be useful as markers for type 2 diabetes, cardiovascular disease, or other diseases for which obesity elevates the risk and exercise reduces it.
This work was supported by NIH Grants DK-51, 271 and DA-04, 494 (L.D.F.). The Laboratory for Macromolecular Analysis of the Albert Einstein College of Medicine is supported, in part, by Cancer Center Core Grant CA13330. Special thanks to Chunhui Fang, Reeta Biswas, and Arun Fricker for assisting with the measurements of body weight, food consumption, and activity, and Fa-Yun Che for helpful discussions.