Obesity-resistant (A/J) and obesity-prone (C57BL/6J) mice were weaned onto low-fat (LF) or high-fat (HF) diets and studied after 2, 10, and 16 wk. Despite consuming the same amount of food, A/J mice on the HF diet deposited less carcass lipid and gained less weight than C57BL/6J mice over the course of the study. Leptin mRNA was increased in white adipose tissue (WAT) in both strains on the HF diet but to significantly higher levels in A/J compared with C57BL/6J mice. Uncoupling protein 1 (UCP1) and UCP2 mRNA were induced by the HF diet in brown adipose tissue (BAT) and WAT of A/J mice, respectively, but not in C57BL/6J mice. UCP1 mRNA was also significantly higher in retroperitoneal WAT of A/J compared with C57BL/6J mice. The ability of A/J mice to resist diet-induced obesity is associated with a strain-specific increase in leptin, UCP1, and UCP2 expression in adipose tissue. The findings indicate that the HF diet does not compromise leptin-dependent regulation of adipocyte gene expression in A/J mice and suggest that maintenance of leptin responsiveness confers resistance to diet-induced obesity.
- uncoupling proteins
- adipose tissue
accumulation of excess adipose tissue is a known risk factor for hypertension, heart disease, and non-insulin-dependent diabetes mellitus. The epidemic of obesity in Western society is associated with consumption of high-fat (HF) diets. A subset of the obese population appears sensitive to the diabetogenic effects of HF diets in that their obesity progresses to insulin resistance and diabetes. A similar variation in sensitivity to dietary fat has been documented in mice (19, 43, 44, 54), where particular strains readily develop an obese/diabetic syndrome after chronic consumption of HF diets. These susceptible strains provide excellent models to study the developmental pathophysiology of an obesity syndrome that is highly analogous to human obesity (5, 44, 54). Contrasting fat-sensitive with fat-resistant mouse strains serves the dual purpose of illustrating the molecular adaptations used to avoid obesity and of identifying genes that are sensitive to dysregulation by HF diets. The common factor in mice that are resistant to dietary fat is an ability to avoid obesity despite increased caloric density, but the cellular adaptations that confer this resistance are poorly understood.
The ability of uncoupling proteins (UCP) to modulate energetic efficiency comes from their ability to short-circuit the mitochondrial proton gradient that drives ATP synthesis (30,35, 39). A cardinal property of adipose tissue is that its thermogenic capacity is directly related to the amount of UCP expressed in it (31, 42). It is well established that leptin regulates expression of the UCP1 gene (4, 11, 31), whereas dietary factors may play an important role in controlling expression of UCP2 (1, 15, 47) and UCP3 (25, 51, 53) and regulating leptin responsiveness. Thus the ability of an animal to respond to increased caloric density may be dictated by its ability to increase thermogenic capacity. Using mouse strains that differ in their sensitivity to HF diets, we show that fat-resistant A/J mice, but not fat-sensitive C57BL/6J mice, increase thermogenic capacity in brown (BAT) and white adipose tissue (WAT) in response to HF diet. These differences may be the product of strain-specific changes in expression of and subsequent responses to circulating leptin.
MATERIALS AND METHODS
EDTA, sodium cholate, Triton X-100, BSA, guanidinium thiocyanate, TES, sucrose, and other common chemicals were from Sigma Chemical (St. Louis, MO). T1 RNase and TRIzol LS reagent were from Life Technologies (Gaithersburg, MD); T7 RNA polymerase, SP6 RNA polymerase,Taq DNA polymerase, Moloney murine leukemia virus reverse transcriptase, and the pGEM-3Z cloning vector were from Promega (Madison, WI); and the T7 Megashortscript kit was purchased from Ambion (Austin, TX). 2-Mercaptoethanol was acquired from J. T. Baker (Phillipsburg, NJ); oligonucleotide primers were prepared by the DNA Core Facility at the Medical University of South Carolina; Na[125I] and α-[32P]cytidine triphosphate were purchased from Du Pont NEN Radiochemicals (Boston, MA); and Immobilon-P polyvinylidine fluoride membranes were from Millipore (Bedford, MA). A semipurified HF diet was prepared by Research Diets (New Brunswick, NJ) to contain 36% fat by weight (44), and the low-fat (LF) control diet contained 5% fat, as described in detail previously (44, 46). The fat source for the diets was coconut and soybean oil (44). Male A/J and C57BL/6J mice were obtained from Jackson Laboratories (Bar Harbor, ME) at weaning. Serum insulin and leptin levels were estimated using kits for rodent insulin and mouse leptin obtained from Linco Research Labs (St. Charles, MO).
Experimental animal protocol.
Fifty-six A/J and C57BL/6J mice were obtained at weaning, and eight animals of each strain were killed at time 0. One-half of the remaining animals of each strain were randomly assigned to receive either the LF or the HF diet (24 per strain per diet). The four groups of mice received their respective diets for 16 wk, and eight representative mice from each group were killed after 2, 10, or 16 wk on diet. At 2-wk intervals, food consumption was monitored in randomly selected groups of eight mice for 24-h periods. Animals were housed in a controlled environment at 22°C on a 12:12-h light-dark cycle with free access to food and water. Body weights were obtained twice weekly. Blood samples were obtained from each mouse at the time it was killed. Thereafter, interscapular BAT (IBAT), epididymal WAT (EWAT), and retroperitoneal WAT (RWAT) were carefully dissected from each animal for preparation of total RNA or isolation of adipocytes.
In an additional experiment, mice from each strain/diet combination were injected with vehicle or CL316,243 (1 μg· day−1·g body wt−1) for 2 days before they were killed. The tissues were harvested and processed as described above, and the mice were studied after 2 and 10 wk on the respective diets.
Preparation of RNA.
After dissection, the interscapular, epididymal, and retroperitoneal fat pads were homogenized with TRIzol LS reagent using an Ultraturax (Tekmar, Cincinnati, OH) according to manufacturer's specifications. Total RNA was isolated and purified as previously described (21).
Ribonuclease protection assay.
RNA probes complementary to mRNA were produced by RT-PCR using total RNA from mouse IBAT for UCP1 (5′-3′:F, caatctgggcttaacgggt; R, tgaaactccggctgagaag), UCP3 (5′-3′:F, ccaccatggctgtgaagttcctg; R, gggtgtacacctgcttgacggagtc), and β3-adrenergic receptor (5′-3′:F, accccagtgcagccaacacca; R, cgcaaccagtttcgcccaagg). Reverse transcribed RNA from mouse EWAT was used for UCP2 (5′-3′:F, cagttctacaccaagggtc; R, aggtcataggtcaccagctca), and the UCP2 probe was shortened to 143 bp by use of a Sma I digest, which cut the fragment at a site corresponding to nucleotide 741. The respective fragments were purified and cloned into the pGEM-3Z riboprobe vector containing transcriptional start sites 5′ and 3′ to the multiple cloning site (Promega, Madison, WI). The identities of the cloned fragments were confirmed by sequencing, and the probes corresponded to nucleotides 7–300 for UCP1, 741–884 for UCP2, 219–467 for UCP3, and 630–870 for the β3-adrenergic receptor. The probe for mouse leptin was obtained from M. Daniel Lane. The respective probes were labeled and used in our modification (11, 21) of the ribonuclease protection assay described by Granneman and Lahners (26). The protected fragments were quantitated by comparison to known amounts of sense strand RNA produced by SP6 transcription of the linearized plasmids and incubated simultaneously with labeled probe. After digestion with 300 U T1 RNase (Life Technologies), sense strand standards and protected fragments were separated on 6% polyacrylamide/8 M urea gels and visualized by autoradiography. Detected bands were quantitated by scanning laser densitometry (Molecular Dynamics, Sunnyvale, CA). Bands were standardized to the amount of 18S rRNA by cohybridizing with a riboprobe complementary to the 18S rRNA (nucleotides 715–794). Estimated concentrations of each mRNA were then determined by reverse calibration from standard curves generated from the known amounts of sense strand standards that were included in each assay.
Adenylylcyclase assay on adipocyte plasma membranes.
Adipocytes were isolated from the epididymal fat pads and used to prepare plasma membranes, as described in detail previously (21, 22). The plasma membranes were used in adenylylcyclase assays to assess the functional coupling of the β3-adrenergic receptor to its effector system among the treatment groups, as described previously (8,20, 21).
Methods of analysis.
The estimated concentrations of UCP1, UCP2, UCP3, leptin, and β3-adrenergic receptor mRNA were obtained by reverse calibration from standard curves, as described in Ribonuclease protection assay. Group means for mRNA estimates, growth data, and plasma hormone concentrations were analyzed by means of a three-way factorial design with strain, diet, and time as main effects. The strain × diet × time interaction was tested by residual variance (animal within strain × diet × month) as the error term, and in the absence of such interactions, interest shifted to strain × diet differences within each time point. Post hoc testing of group means within each time point was made with the Bonferroni correction, which used the pooled error term to calculate standard errors. Protection against Type I errors was set at 5% (α = 0.05).
Growth and food consumption.
Body weights of mice within each strain were similar at the beginning of the study (Fig. 1), as were the weights of epididymal fat pads from representative mice of each strain (Table 1). Fat pad weights were also similar between the strains after mice consumed the LF diet for 2, 10, and 16 wk (Table 1). In mice mice consumed the LF diet, body weights began to diverge after 4–5 wk, such that the C57BL/6J mice were heavier (P < 0.05) at all subsequent time points (Fig.1). The response between mouse strains on the HF diet was more pronounced, and Fig. 1 illustrates that the C57BL/6J mice grew faster and to higher weights than the A/Js (P < 0.05). The strain difference was evident as early as 2 wk, and the fat pad weights in the C57BL/6J mice were higher (P < 0.05) than those of the A/Js at the latter two time points (Table 1).
Strain differences in growth and fat deposition were not complemented by differences in food consumption, because A/J mice ate more (or similar amounts) of the LF diet than C57BL/6J mice throughout the study (Fig. 2 A). For instance, at 2, 4, 6, and 10 wk, A/J mice consumed significantly (P < 0.05) more LF diet than C57BL/6J mice (Fig. 2). Similar consumption rates were observed between the strains at the other time points (Fig.2 A). The increased caloric density of the HF diet led to lower food consumption on a weight basis, but the amount of diet consumed did not differ between C57BL/6J and A/J mice at any time point (Fig. 2 B). Thus the faster growth and greater fat deposition that occurred in C57BL/6J mice cannot be attributed to greater food consumption. Subjective evaluations based on observations of relative physical activity were made throughout the study, and the C57BL/6J mice were consistently more active than the A/J mice. Therefore, the apparently higher efficiency of C57BL/6J compared with A/J mice cannot be explained by differences in relative physical activity.
Strain differences in growth were also reflected by differences in serum insulin. Irrespective of diet, serum insulin concentrations were higher in C57BL/6J than in A/J mice at all except the 0 time point (Table 1). The difference between strains was relatively small in mice on the LF diet, but the HF diet produced a much more significant increase in serum insulin in C57BL/6J compared with A/J mice (Table 1). C57BL/6J mice on the HF diet also developed hyperglycemia at 16 wk, whereas the A/J mice maintained normoglycemia throughout the study (Table 1).
Serum leptin levels were similar between mouse strains at the start of the experiment and increased modestly and to the same extent over time in both groups on the LF diet (Fig. 2). Regardless of strain, plasma leptin levels were higher at each time point in mice on the HF compared with the LF diet. The HF diet-induced increase in plasma leptin was comparable at all except the 2-wk time point, when levels in the A/J mice were significantly higher (P < 0.05) than in C57BL/6J (Table 1). Similar comparisons at the later two time points also suggest higher plasma leptin levels in A/J compared with C57BL/6J mice, but the variability of estimates precluded detection of these differences. If plasma leptin is expressed as a function of EWAT size (29), the results suggest that A/J mice produced more leptin per unit of adipose tissue than C57BL/6J mice at each time point (Table 1).
WAT gene expression.
To examine the basis for the apparent differences in plasma leptin between A/J and C57BL/6J mice, we examined leptin mRNA in WAT from both strains. Leptin mRNA levels in EWAT were comparable between strains at the beginning of the study and changed little over time in either group consuming the LF diet (Table 2). In contrast, differences were clearly evident in mice consuming the HF diet, because leptin mRNA levels were significantly higher (P < 0.05) in A/J compared with C57BL/6J mice at the 2-wk time point (Table 2). This difference was also evident at 10 and 16 wk, because leptin mRNA levels were 50–100% higher in A/J than in C57BL/6J mice on the HF diet (Table 2). Leptin mRNA levels were also higher (P < 0.05) in RWAT of A/J compared with C57BL/6J mice (0.043 ± 0.006 vs. 0.020 ± 0.006 fmol mRNA/μg RNA) consuming the HF diet at the 10-wk time point. Using two different depot sites, we confirmed the results that leptin expression is unexpectedly higher in A/J compared with C57BL/6J mice consuming the HF diet.
Additional experiments were conducted to establish the basis for this difference in leptin expression between the mouse strains. Mice on each of the diets were treated with the selective β3-adrenergic receptor agonist CL316,243 to test for differences in inhibitory regulation of leptin expression. The results from C57BL/6J mice on the LF diet were consistent with previous work (9, 18, 23, 49) and showed that CL316,243 produced a significant (P < 0.05) downregulation of leptin mRNA in mice at 2 and 10 wk (Fig.3, A and B). The opposite response was seen in A/J mice on the LF diet, where CL316,243 failed to decrease leptin mRNA in WAT at either time point (Fig. 3,A and B). Consumption of the HF diet transformed the response to CL316,243 in both groups, but the change in response was dependent on time. This produced a strain × diet × time interaction (P < 0.05) and indicates that the effect of each diet over time differed between the strains. For instance, in C57BL/6J mice, the HF diet blunted the inhibition of leptin mRNA expression by CL316,243 at both time points, although a modest but significant (P < 0.05) effect could be detected at 10 wk (Fig. 3 B). In contrast, the HF diet enabled the response in A/J mice, such that CL316,243 produced a significant reduction (P < 0.05) in leptin mRNA at both 2 and 10 wk (Fig.3 B).
Measurements of plasma leptin in the aforementioned mice at the 2-wk time point were consistent with leptin mRNA levels in tissues from the corresponding animals. For example, plasma leptin at 2 wk was higher (P < 0.05) in A/J mice (LF 10.6 ± 2.5, HF 27.6 ± 2.8 pg/ml) than in C57BL/6J mice (LF 4.44 ± 0.28, HF 21.6 ± 2.3 pg/ml), regardless of diet. Moreover, in mice consuming the HF diet, CL316,243 decreased plasma leptin in A/Js from 27.6 ± 2.8 to 7.99 ± 1.3 pg/ml (P < 0.05) but failed to lower leptin levels in C57BL/6J mice (vehicle 21.6 ± 2.3, CL316,243 18.4 ± 4.3 pg/ml) on the same diet. Together, these results illustrate a fundamental difference in the way leptin is regulated in adipose tissue between A/J and C57BL/6J mice.
One explanation for the failure of mice to downregulate leptin mRNA in response to CL316,243 could be compromised expression or function of the β3-adrenergic receptor. β3-Adrenergic receptor mRNA levels were similar in EWAT from A/J and C57BL/6J mice (0.23 ± 0.03 and 0.28 ± 0.02 fmol mRNA/μg RNA) at the beginning of the study and were comparable at 10 wk in mice on both diets (A/J LF 0.36 ± 0.09, A/J HF 0.48 ± 0.09, C57BL/6J LF 0.38 ± 0.09, C57BL/6J HF 0.31 ± 0.09 fmol/μg RNA). In contrast, after 16 wk on the HF diet, β3-adrenergic receptor mRNA levels were significantly higher (P < 0.05) in A/J compared with C57BL/6J mice (0.73 ± 0.05 vs. 0.50 ± 0.05 fmol/μg RNA). Measurements of mRNA were complemented by adenylylcyclase assays in adipocyte membranes from each group to compare functional activity of the β3-adrenergic receptor. When the β3-selective agonist CL316,243 was used, maximal adenylylcyclase activation was significantly lower (P < 0.05) in adipocyte membranes from C57BL/6J compared with A/J mice (190 ± 9.7 vs. 268 ± 12.9 pmol cAMP·min−1·mg protein−1) after 16 wk on the HF diet. In contrast, maximal adenylylcyclase activation was comparable between HF-fed mice at 2 wk (A/J 151 ± 12, C57BL/6J 139 ± 9 pmol cAMP·min−1·mg protein−1) and at 10 wk (A/J 130 ± 8, C57BL/6J 141 ± 7 pmol cAMP·min−1·mg protein−1). The observed changes in functional activity of the β3-adrenergic receptor in adipocyte membranes from each group parallel the observed changes in β3-adrenergic receptor mRNA among the groups. These findings indicate that compromised β3-adrenergic receptor-mediated signaling develops in C57BL/6J mice only after long-term consumption of the HF diet and do not predate the development of altered leptin regulation or obesity in these mice.
Because WAT is one of the primary sites of UCP2 expression, UCP2 mRNA levels were measured in EWAT to test the hypothesis that UCP2 is differentially upregulated between the mouse strains in response to a HF diet. UCP2 mRNA levels were similar between A/J and C57BL/6J mice (0.27 ± 0.06 and 0.30 ± 0.02 fmol mRNA/μg RNA) at the beginning of the study and were unchanged from these initial levels in C57BL/6J mice, regardless of diet (Fig.4 A; Table 2). In contrast, UCP2 mRNA levels in A/J mice were significantly increased (P < 0.05) from these initial levels at 2, 10, and 16 wk on the LF diet (Fig. 4 A; Table 2). Moreover, weaning A/J mice onto the HF diet produced a further significant increase (P < 0.05) in UCP2 mRNA levels at each time point (Fig. 4 B; Table 2). Compared with C57BL/6J mice on the HF diet, UCP2 mRNA levels in A/J mice were two- to fourfold higher at each time point (Fig. 4 B; Table 2). These results indicate a clear strain difference in the adaptive response of WAT to HF diets.
UCP1 mRNA expression was compared in RWAT from each strain, because previous studies had shown significant potential for induction in this site after cold exposure (27). UCP1 mRNA was readily detected in RWAT from A/J mice at the beginning of the study but was below the detection limit in C57BL/6J mice. This difference was also evident at the 2- and 10-wk time points, when significant levels of UCP1 mRNA were observed in RWAT from A/J mice (Fig.5; Table 2). As before, UCP1 mRNA levels were near the detection limit in tissue from C57BL/6J mice at all time points surveyed. In contrast to UCP2, the HF diet had no apparent effect on UCP1 mRNA expression in this tissue (Fig. 5; Table 2). The HF diet was also without effect on UCP1 mRNA expression in RWAT from A/J mice, levels being comparable between mice on each diet at 2 and 10 wk (Fig. 5; Table 2).
BAT gene expression.
BAT UCP1 mRNA levels were similar between C57BL/6J and A/J mice (4.22 ± 0.22 and 3.74 ± 0.12 fmol UCP1 mRNA/μg RNA) at the start of the experiment and between A/J and C57BL/6J mice on the LF diet at all but the initial time point over the course of the study (Fig. 6; Table 2). UCP1 mRNA levels were also similar between the mouse strains after 2 wk on the HF diet, but at both 10 and 16 wk, UCP1 mRNA was significantly higher (P < 0.05) in BAT from A/J compared with C57BL/6J mice (Fig. 6; Table 2). Thus A/J mice responded to the HF diet by increasing UCP1 mRNA levels, whereas the C57BL/6J mice failed to respond in a comparable manner.
Because the β3-adrenergic receptor mediates many of the effects of sympathetic stimulation of BAT, it plays a central role in mediating the effects of sympathetic regulation of UCP1 expression. Therefore, we compared β3-adrenergic receptor mRNA levels among the groups to determine whether compromised receptor expression could account for strain differences in UCP1 induction. β3-Adrenergic receptor mRNA levels were 1.7-fold higher in A/J compared with C57BL/6J mice (0.73 ± 0.04 vs. 0.44 ± 0.04 fmol mRNA/μg RNA) at the beginning of the study and were essentially unchanged in A/J and C57BL/6J mice (0.67 ± 0.05 vs. 0.46 ± 0.03 fmol mRNA/μg RNA) after 16 wk on the HF diet. This finding suggests that diet-induced differences in UCP1 mRNA levels between the strains are not due to diet-induced changes in β3-adrenergic receptor expression in BAT.
BAT UCP3 mRNA levels were similar between A/J and C57BL/6J mice (0.017 ± 0.002 and 0.018 ± 0.003 fmol UCP3 mRNA/μg RNA) at the start of the study and changed little within the groups during the progression of the study (Table 2). In contrast to UCP1, UCP3 mRNA was unaffected by diet in both mouse strains and was expressed at fairly uniform levels across mouse strain and diet (Table 2). Considered together, the results are inconsistent with the hypothesis that BAT UCP1 and UCP3 mRNA are regulated by a common signal unless the threshold for induction is substantially different between the two proteins.
The translated product of the ob gene regulates food consumption by communicating the status of peripheral adipose stores to the brain (17, 55). Recent evidence indicates that leptin also regulates energy expenditure through the sympathetic nervous system by enhancing UCP1 expression in both BAT and WAT (2, 10, 11, 32,33). The consensus is that leptin is the afferent signal in a feedback loop between adipose tissue and the brain that acts to regulate both energy intake and expenditure. This paradigm predicts that animals will respond to diets of high caloric density by reducing food intake and increasing energy expenditure. In the present work, C57BL6/J mice failed to fulfill this prediction and are among several strains of mice classified as “sensitive” to HF diets (5, 43, 44). In contrast, A/J mice are classified as “fat resistant,” because they do not develop the associated diabetic pathologies on HF diets and become only moderately obese (43, 44). The present findings are consistent with these classifications and show that C57BL/6J mice grow faster and deposit significantly more fat than A/J mice at similar or lower rates of food consumption. Taken together, these findings demonstrate a fundamental difference in energy requirements for body weight maintenance between these mouse strains and argue that the ability of leptin to match rates of food intake and energy utilization was differentially compromised by the HF diet. Therefore, the major goal of the present study was to examine the regulation of leptin expression and function as a basis for the differing propensities of C57BL/6J and A/J mice to become obese.
Plasma leptin levels were initially similar between the mouse strains, but after 2 wk on the HF diet, leptin levels were higher in A/J than in C57BL/6J mice. The data suggest that plasma leptin was also higher in A/J mice at 10 and 16 wk, but the variability of the observations precluded detection of this difference. Surwit et al. (45) reported a similar disproportionate increase in plasma leptin in A/J compared with C57BL/6J mice after 2–4 wk on HF diet. This is surprising, given the greater fat deposition in C57BL/6J compared with A/J mice, and it suggests that leptin expression per unit of adipose tissue was actually higher in A/J compared with C57BL/6J mice. To test this hypothesis, we compared leptin mRNA levels in two WAT depot sites from mice of each strain. The findings from these studies are consistent with our hypothesis at all time points. These differences are particularly important because they occurred during the initial 10 wk of the study, when fat deposition was clearly occurring at a faster rate in C57BL/6J compared with A/J mice. Given its demonstrated role in regulating energy utilization and fat oxidation (10,11, 32, 33, 52), the higher leptin expression in A/J compared with C57BL/6J mice may have contributed to the observed differences in fat accumulation by increasing peripheral energy utilization in A/J mice. Support for this suggestion comes from Surwit et al., who reported that core temperatures were higher in A/J vs. C57BL/6J mice consuming HF diets. Lower body temperature is a distinguishing characteristic ofob/ob mice, and replacement of the missing leptin restores body temperature to normal (40), ostensibly by increasing UCP1 expression (11) and thermogenic activity in BAT. Similar observations were evident in the present study, where the higher leptin expression in A/J mice on HF diets was mirrored by higher UCP1 mRNA levels in BAT from these mice. These findings support the concept that disproportionately higher leptin expression in A/J compared with C57BL/6J mice lessens their energetic efficiency by increasing nonshivering thermogenesis.
The hypothesis that high caloric density mobilizes a more vigorous thermogenic response in A/J compared with C57BL/6J mice is also supported by measurements of UCP2 mRNA, where levels were significantly higher in WAT from A/J mice. These findings are similar to results reported by Fleury et al. (15) and Surwit et al. (47), with two subtle but important differences. First, our results show that UCP2 mRNA expression in adipose tissue was comparable between the strains at weaning, with strain and diet-associated differences developing thereafter. Second, the present work shows that UCP2 mRNA levels were consistently higher in A/J than in C57BL/6J mice after 2, 10, and 16 wk on the HF diet. The previous study (15) had found a similar difference at early time points but actually found higher levels of UCP2 mRNA in C57BL/6J compared with A/J mice later in that study. Although UCP2 has been shown to lower yeast membrane potential and uncouple respiration (15, 34), the functional significance of increased UCP2 mRNA in the present system is unknown. If increased mRNA levels are matched by increased protein expression, then it seems likely that decreased metabolic efficiency would result. In contrast to UCP1 and UCP2, no evidence was found to support a role for UCP3 in diet-induced obesity.
The mechanism for increased UCP2 expression in A/J compared with C57BL/6J mice is unknown. Although initial reports suggested that UCP2 was regulated by leptin (56) or by norepinephrine (3), a number of other studies have questioned these findings (11, 15, 24,25). For instance, we showed that UCP2 was unchanged in retroperitoneal WAT from mice in which leptin treatment induced UCP1 expression in the same tissue (11). Therefore, it seems unlikely that the higher leptin expression in A/J mice is the basis for the observed strain differences in UCP2 mRNA. What seems more likely is that UCP2 expression was differentially affected between strains by dietary fat. Support for this suggestion comes from reports showing that UCP2 expression is responsive to changes in circulating free fatty acids (25, 41), and from the work of Aubert et al. (1), who showed that peroxisome proliferator-activated receptor-γ2 (PPARγ2) agonists increased UCP2 mRNA in adipocytes. Given that fatty acids can transcriptionally activate PPARγ-sensitive genes (16) and the recent report showing that PPARγ expression level influenced adipocyte gene expression (36), the possibility of strain differences in signaling through this pathway warrants further study.
Observed differences in circulating leptin and tissue mRNA suggest that regulation of this gene differs between the two mouse strains. Based on recent reports showing that leptin expression is inhibited by adrenergic stimulation of adipose tissue (9,10, 48, 49), we explored the possibility that inhibitory regulation of leptin expression was differentially compromised in mice of each strain by the HF diet. The results from studies using a selective β3-adrenergic receptor agonist were surprising and showed that the HF diet completely transformed the inhibitory response in each mouse strain. For instance, CL316, 243 failed to reduce leptin mRNA in A/J mice consuming the LF diet, whereas A/J mice on the HF diet showed a robust agonist-induced reduction in leptin mRNA. In contrast, the CL316,243-mediated reduction in leptin mRNA noted in C57BL/6J mice on the LF diet was essentially absent in C57BL/6J mice on the HF diet. A potential explanation for diet-induced transformation of the response could be changes in β-adrenergic receptor expression. Collins et al. (7) examined this question in C57BL/6J and A/J mice fed HF diets for 16 wk and found greater reductions in β3-adrenergic receptor mRNA and function in WAT from C57BL/6J mice. We found comparable reductions in β3-adrenergic receptor mRNA and function in C57BL/6J mice at 16 wk, but the observed differences in inhibitory regulation of leptin expression occurred at earlier time points (2 and 10 wk), when we found no evidence of compromised expression or function of the receptor. Although leptin expression involves a number of hormonal inputs (12-14, 38), a recent study showed that reduced expression of PPARγ in adipose tissue had a profound effect on leptin expession levels (36). Leptin expression was disproportionately high in relation to adipose tissue mass, producing leaner mice with elevated metabolic rates (36). In that study, the authors showed that changes in regulatory factors and systems that impact adipocyte gene expression are capable of transforming obesity-prone mice into mice that are resistant to diet-induced obesity (36).
It is reasonable to assume that differences in inhibitory regulation of leptin expression are not the sole basis for the differences noted in the present work. Our results, however, make a strong case that leptin expression and its regulation are fundamentally different in WAT from A/J and C57BL/6J mice. Because cAMP-dependent mechanisms are also important in regulating leptin release from adipocytes (18), it will be important to establish whether the observed strain differences in gene transcription extend to leptin release from the adipocyte.
Strain and diet-dependent differences in adipocyte gene expression could also be related to differences in the proliferative state within adipocyte depots. For instance, diet-induced differences in hyperplastic growth within white fat depot sites could impact circulating leptin levels through an increase in leptin expressing cells. Although cell numbers were not determined in the present study, Surwit et al. (43) found a 40% increase in adipocytes in white fat depots from C57BL/6J mice that had consumed the high-fat diet for 16 wk. In contrast, no evidence of hyperplasia was found in A/J mice on the same diet (43). Therefore, these findings are inconsistent with the suggestion that increased leptin expression in A/J mice is caused by adipocyte hyperplasia; rather, these data support the conclusion that leptin expression per adipocyte is higher in WAT from A/J compared with C57BL/6J mice.
Recent evidence suggests that diet-induced obesity is associated with the development of leptin resistance (28, 37,50). Our studies show that, despite robust increases in circulating leptin, strain differences in body weight and fat accretion occurred at similar rates of food consumption. These findings suggest that strain differences in energy utilization reflect either a relative leptin insufficiency or a compromised ability of leptin to induce thermogenic capacity and activity in C57BL6/J compared with A/J mice. The present studies do not distinguish between these possibilities, but it should be noted that several mouse models of diet-induced obesity display peripheral but not central leptin resistance (6,50). Leptin resistance may also occur through diet-induced effects on signaling components in target tissues. An example would be the diet-induced reduction of β3-adrenergic receptor expression in C57BL/6J mice, which could limit the ability of adipose tissue to respond to sympathetic stimulation. This seems unlikely in the present study, because differential fat accumulation occurred well before any change in β-adrenergic signaling efficacy. Our findings are more consistent with changes in downstream signaling components that modify the set points for leptin release between the two mouse strains. The present studies do not exclude central leptin resistance as a contributing mechanism of diet-induced obesity; rather they demonstrate fundamental differences in the regulation of adipocyte gene expression between A/J and C57BL/6J mice. It will be important in future studies to assess the relative importance of altered leptin expression vs. central and peripheral mechanisms of leptin resistance in various models of diet-induced obesity.
The authors acknowledge the excellent technical assistance of Isabel Frampton and Jami Kelley.
This work was supported by a research grant from the American Diabetes Association (T. W. Gettys), National Institute of Diabetes and Digestive Kidney Diseases Grant DK-53981 (T. W. Gettys), and Research Grant No. 9800699 from the US Department of Agriculture NRICGP/USDA (T. W. Gettys).
Address for reprint requests and other correspondence: T. W. Gettys, 916-G Clinical Science Bldg., Medical University of South Carolina, 96 Jonathan Lucas St. Charleston, SC 29425 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
- Copyright © 2000 the American Physiological Society