Altered metabolism and resistance to obesity in long-lived mice producing reduced levels of IGF-I

Abstract

The extension of lifespan due to reduced insulin-like growth factor 1 (IGF-I) signaling in mice has been proposed to be mediated through alterations in metabolism. Previously, we showed that mice homozygous for an insertion in the Igf1 allele have reduced levels of IGF-I, are smaller, and have an extension of maximum lifespan. Here, we tested whether this specific reduction of IGF-I alters glucose metabolism both on normal rodent chow and in response to high-fat feeding. We found that female IGF-I-deficient mice were lean on a standard rodent diet but paradoxically displayed an insulin-resistant phenotype. However, these mice gained significantly less weight than normal controls when placed on a high-fat diet. In control animals, insulin response was significantly impaired by high-fat feeding, whereas IGF-I-deficient mice showed a much smaller shift in insulin response after high-fat feeding. Gluconeogenesis was also elevated in the IGF-I-deficient mice relative to controls on both normal and high-fat diet. An analysis of metabolism and respiratory quotient over 24 h indicated that the IGF-I-deficient mice preferentially utilized fatty acids as an energy source when placed on a high-fat diet. These results indicate that reduction in the circulating and tissue IGF-I levels can produce a metabolic phenotype in female mice that increases peripheral insulin resistance but renders animals resistant to the deleterious effects of high-fat feeding.

the reduction in type 1 insulin-like growth factor (IGF-I) signaling has been associated with increased lifespan in multiple laboratory models, including invertebrates and mice (7, 23, 45, 46). Many possible downstream effectors have been proposed as primary mechanisms that underlie this relationship; however, several commonalities in the response to caloric restriction and reduced IGF-I signaling have led to the suggestion that a shift in metabolism toward improving insulin sensitivity may be a key determinant of mammalian longevity (1). In support of this, long-lived mice deficient in growth hormone (GH), the primary effector of circulating IGF-I levels in mammals, or long-lived mice deficient in the receptor of GH display both a general enhancement of insulin sensitivity and reduced circulating levels of glucose and insulin (5, 15, 16, 28). However, this relationship has been less clearly defined in models in which IGF-I signaling has been targeted directly. In independent studies, reduction of the cellular receptor of IGF-I (IGF-IR) in mice has been reported to extend lifespan in female but not male mice (4, 21). Interestingly, both studies also reported a significant impairment in glucose metabolism in male mice that was not generally found in female mice. Furthermore, Garg et al. (19) reported that this same deficiency in IGF-IR renders both male and female mice susceptible to metabolic impairment caused by feeding a high-fat, high-sucrose diet. Similarly, mice lacking pregnancy-associated plasma protein A, a metalloprotease that degrades IGF-binding protein (IGFBP), have reduced local IGF-I signaling and extended longevity but show no significant alteration in insulin sensitivity on either a normal or high-fat diet (9–11). Collectively, these studies highlight a still unanswered question regarding the role of IGF-I in metabolic regulation and its relationship to longevity.

A better understanding of this question may also have profound consequences in the context of addressing the rising incidence of obesity and metabolic dysfunction that is occurring among all age groups of humanity worldwide. Serum IGF-I levels have been reported to be decreased in obese individuals compared with people of normal weight (38); however, obesity and metabolic dysfunction have been shown to elevate levels of free IGF-I, the metabolically active form of IGF-I (30, 42). These dynamic changes in IGF-I may be a consequence of hyperinsulinemia associated with metabolic disease (18). As a therapeutic approach, concomitant treatment with insulin and IGF-I has been tested in type 2 diabetes with clear beneficial effects on this metabolic dysfunction, although the relatively high effective doses used have limited clinical use (8, 35). In terms of direct manipulation of IGF-I, reduced adiposity has been reported in mice with targeted inactivation of hepatic IGF-I expression (53). Mice lacking hepatic IGF-I (and thus serum IGF-I) have also been shown to be susceptible to obesity and insulin resistance when fed a diet high in fat (52). Conversely, mice that express elevated levels of IGF-I preserve insulin sensitivity when fed a high-fat diet, although they are not protected from fat accumulation (40, 41). There is also evidence that the relationship between obesity and IGF-I levels appears to be stronger in young mice and reduced in mice with adult-onset obesity (12). IGF-I signaling has been shown to play a major role in the regulation of age-related diseases, including alterations in metabolic function that may promote obesity (49). Thus, it is important to clarify the metabolic impact of altered IGF-I to potentially identify strategies to prevent obesity and metabolic dysfunction.

As a means to clarify the relationship between IGF-I, metabolism, and longevity, we have utilized a novel mouse mutant with specific reduction in levels of IGF-I. These IGF-I-deficient mice are distinct from other models of reduced IGF-I production in that they have reduced IGF-I in all tissues measured due to the direct manipulation of the Igf1 gene (22). This distinguishes these animals from mice in which the central, liver-specific production of IGF-I is ablated through gene deletion or loss of GH signaling or mice in which IGF-I signaling is decreased by reducing levels of the IGF-IR. We have reported previously that these IGF-I-deficient mice have reduced levels of circulating IGF-I, are smaller, and have a greater maximum, albeit not mean, lifespan than their littermate control mice (25). In this study, we tested the metabolic phenotype of these IGF-deficient mice and addressed whether global reduction in IGF-I directly alters metabolism. Surprisingly, we found that female IGF-I-deficient mice were relatively resistant to the obese phenotype (i.e., fat accumulation) that typically arises from a diet high in fat content. We also found that female IGF-I-deficient mice were relatively insulin resistant when maintained on normal chow; however, their response to high-fat diet was much less severe than that of control animals. The results suggest that a global reduction in IGF-I levels can in some ways protect against obesity caused by high-fat feeding and also highlight the complex interaction between IGF-I, metabolic processes, and longevity.

MATERIALS AND METHODS

Animals.

Mice used in this study were derived from mice homozygous for the Igf1 hypomorphic allele obtained from The Jackson Laboratory (Bar Harbor, ME) on a 129sv/C57/CD-1 heterogeneous genetic background (22). These mice were mated to C57BL/6J mice from The Jackson Laboratory to generate mice heterozygous for the mutant Igf1 allele. Heterozygous mice were mated to produce litters containing control mice, heterozygous mice, and mice homozygous for the Igf1 hypomorphic allele. Mice were housed as sibling groups of three to four mice in a high-efficiency particulate air-filtered positive-flow isolation rack. Food and water were provided ad libitum throughout the study. All animal care and procedures were in compliance with the National Institutes of Health and approved by the Institutional Animal Care and Use Committees at the University of Texas Health Science Center at San Antonio, The Pennsylvania State University, and Drexel University College of Medicine. Chow-fed animals were fed standard vegetable-derived chow (Harlan-Teklad 7912; Harlan Laboratories, Madison, WI); composition energy content (kcal/g) was as follows: protein 24%, carbohydrates 62%, fat 14%. The high-fat diet (TestDiet 58V8; TestDiet, Richmond, IN) used in this study was a defined diet containing 45% fat-derived calories. The composition of this defined diet in terms of source of calories was as follows: 18.1% protein, 35.8% carbohydrates, and 46.1% fat. All mice were monitored for food consumption and body weight weekly while on the high-fat diet. In this study, we have presented results only from female mice; however, we performed limited tests for most assays, which suggests that male mice responded in a manner similar to females (data not shown).

Tissue extracts for IGF-I measurements and Western blot analysis.

Tissue extracts for IGF-I measurements were prepared according to previously published methods (14). Equal amounts of proteins were analyzed for IGF-I concentration using a high-sensitivity ELISA (Immunodiagnostic Systems, Scottsdale, AZ), following the manufacturer's protocol utilizing an acid extraction method to release IGF-I from IGF-binding proteins.

Body composition.

Body composition of mice was analyzed using an EchoMRI 3-in-1 composition analyzer (Echo Medical Systems, Houston, TX). Body composition was determined by the formula [fat mass/(fat mass + lean mass)].

Glucose tolerance tests, insulin tolerance tests, and pyruvate tolerance tests.

For glucose tolerance testing, mice were fasted for 6 h and given 1.5 g glucose/kg body wt by intraperitoneal (ip) injection (Sigma-Aldrich, St. Louis, MO). Blood glucose levels were measured 0, 15, 30, 60, and 120 min following injection. For insulin tolerance testing, mice were fasted for 6 h and given 0.75 U insulin/kg body wt by ip injection (Novolin; Novo Nordisk, Princeton, NJ). Blood glucose levels were measured 0, 15, 30, 60, and 90 min following injection. For pyruvate tolerance testing, mice were fasted overnight (18 h) and given 2 g sodium pyruvate/kg body wt by ip injection (Sigma-Aldrich). Blood glucose levels were measured 0, 15, 30, 60, and 120 min following injection. All blood glucose measurements were made using a OneTouch Ultra glucometer (LifeScan, Milpitas, CA).

Blood insulin measurements.

Insulin levels were measured in whole blood using the Ultra Sensitive Mouse Insulin ELISA kit (Crystal Chem, Downers Grove, IL), following the manufacturer's instructions.

Metabolic rate.

Oxygen (O₂) consumption and carbon dioxide (CO₂) production were measured concurrently with core body temperature and physical activity in six mice of each genotype and sex. Metabolic rate was determined via indirect calorimetry using a system designed in R. McCarter's laboratory, which has been described in detail elsewhere (29). Briefly, fresh air flow rate through the cage was held at 500 ml/min (Model R-2; AEI Technologies, Pittsburgh, PA), with 150 ml/min (TSI, Shoreview, MN) of air leaving the cage passing through a desiccant/membrane dryer (Model 4140; Perma Pure, Toms River, NJ) prior to entering gas analyzers. Gas composition was analyzed by a zirconia cell O₂ detector (Model S-3A/I; AEI Technologies) and an infrared CO₂ detector (Model CD-3A; AEI Technologies). Readings were taken every minute and recorded using Modular Animal Respirometry System software (AEI Technologies) on an independent PC.

RESULTS

When maintained on a standard chow diet, mice homozygous for the Igf1 allele mutation were smaller and leaner than their littermate controls. Differences in body weight were present from birth, as newborn female IGF-I-deficient mice were smaller than controls, and these differences persisted in adult females (Fig. 1A). Previously, it was shown that IGF-I levels in mutant mice were ∼46% of control levels, and the body weight of these mice was ∼60% of control levels (22). Here, too, we found that IGF-I levels were reduced in the IGF-I-deficient mice relative to controls (Fig. 1B), with skeletal muscle showing the greatest disparity between groups (IGF-I-deficient mice 58% of control), whereas liver (71% of control) and kidney (68% of control) also showed significant reductions in the IGF-I-deficient mice (Fig. 1B). Previously, we also reported significant reduction in levels of circulating IGF-I in the serum (25). These findings are consistent with studies on the relative contribution of IGF-I and GH to postnatal growth and body size in that IGF-I is thought to be responsible for 80–90% of body growth (26).

Fig. 1.

IGF-I-deficient mice are smaller and leaner than controls when maintained on standard rodent chow. Graphs represent average values obtained from female control and IGF-I-deficient mice. A: average body weights are presented for newborn and 6-mo-old mice maintained on standard rodent chow. B: average tissue levels of IGF-I. Tissue homogenates from 4- to 6-mo-old control mice (n = 5) and IGF-I-deficient mice (n = 7). C: average lean mass (open bar) and fat mass (black bar) for control and IGF-I-deficient mice. D: average body composition presented as %fat. For all graphs, error bars represent SE; *P < 0.05 for unpaired t-test.

We next addressed whether or not the small size of IGF-I-deficient mice was due to an equal reduction in both fat and lean mass. IGF-I-deficient mice showed a significant decrease in both lean mass and fat mass relative to their littermate controls (Fig. 1C). However, IGF-I-deficient mice were significantly leaner than control mice (Fig. 1D); that is, the reduction in fat mass of IGF-I-deficient mice compared with controls was greater than that of the reduction in lean mass. The relative lack of fat mass in these mice may highlight the importance of IGF-I in the differentiation of adipose tissue (20).

We next examined the influence of reduced IGF-I production on glucose metabolism in IGF-I-deficient and control mice. Surprisingly, IGF-I-deficient mice exhibited higher fasting glucose levels (6-h fast) and a slight but significant glucose intolerance relative to control animals in response to glucose challenge (Fig. 2A). The average area under the curve for these glucose tolerance tests also differed significantly between groups (19,889 vs. 26,170 integrated AUC, P = 0.008). In addition, the fasting insulin level in blood was significantly higher in IGF-I-deficient mice (Fig. 2B). Together, these findings suggested that IGF-I-deficient mice may be mildly insulin resistant. To test this directly, IGF-I-deficient mice and control mice were subjected to insulin tolerance tests. IGF-I-deficient mice displayed features of insulin resistance when maintained on normal chow; that is, response to insulin tolerance tests was reduced in IGF-I-deficient mice (Fig. 2C). Again, the average area under the curve for insulin tolerance tests differed significantly between control and IGF-I-deficient mice (4,429 vs. 9,425 integrated AUC, P = < 0.001). Finally, IGF-I-deficient mice displayed higher rates of glucose production during pyruvate tolerance testing, suggesting that IGF-I-deficient mice had increased liver gluconeogenesis (Fig. 2D). Hepatic insulin resistance is associated with reduced inhibition of gluconeogenesis; these data together suggest insulin resistance in multiple tissues of IGF-I-deficient mice.

Fig. 2.

IGF-I-deficient mice exhibited altered glucose metabolism when maintained on standard rodent chow. Graphs represent average values obtained from female control (open bars or ○) and IGF-I deficient mice (black bars or ●) of ∼6 mo of age maintained on standard rodent chow. Error bars represent SE; *P < 0.05 for unpaired t-test. A: glucose tolerance test performed after a 6-h fast, as described in materials and methods; control: n = 10; IGF-I deficient: n = 9. B: fasting (6-h) insulin levels in whole blood; control: n = 5; IGF-I deficient: n = 5. C: insulin tolerance test performed after a 6-h fast, as described in materials and methods; control: n = 10; IGF-I deficient: n = 9. D: pyruvate tolerance test performed after an 18-h fast, as described in materials and methods; control: n = 7; IGF-I deficient: n = 9.

To examine the influence of reduced IGF-I production on the development of obesity, 6-mo-old female mutant and control mice were placed for 12 wk on a high-fat diet that provides 45% of calories from fat (see materials and methods). Surprisingly, IGF-I-deficient mice gained significantly less weight than controls throughout the course of this high-fat feeding regimen (Fig. 3A). Because the weights of IGF-I-deficient mice and controls at the beginning of the study were significantly different, we compared the weight gained by each mouse relative to its starting weight; on average, IGF-I-deficient mice were 60% the weight of controls at the onset of the high-fat diet. Even using this normalization, IGF-I-deficient mice gained significantly less weight than control mice (Fig. 3B). When adjusted for their smaller weight, food consumption was not markedly different between control and IGF-I-deficient mice (data not shown). Fat mass increased to a greater degree in control mice fed high-fat diets, whereas the lean mass of both IGF-I-deficient mice and control mice remained relatively unchanged (Fig. 3, C and D). Thus, the IGF-I-deficient mice remained relatively lean on the high-fat diet compared with controls.

Fig. 3.

IGF-I-deficient mice gained significantly less weight than control mice when fed a high-fat diet. A: average change in body weight from starting weight for control (n = 13) and IGF-I-deficient mice (n = 12) fed a high-fat diet. B: average change in body weight relative to starting weight. C: change in average lean mass from chow to high-fat fed. D: change in average fat mass from chow to high-fat fed. For C and D, graphs represent average values obtained from female control (n = 13) and IGF-I-deficient (n = 12) mice at ∼6 mo of age maintained for 12 wk on high-fat diet. Error bars represent SE; *P < 0.05 for unpaired t-test.

Glucose metabolism is strongly regulated by body composition, i.e., adiposity, and is a clear indicator of metabolic dysfunction. Compared with chow-fed mice, high-fat-fed control mice were significantly impaired in their responses to glucose tolerance and insulin tolerance tests (Fig. 4, A and B). High-fat-fed IGF-I-deficient mice showed no difference from control mice in glucose tolerance tests (38,003 vs. 37,640 integrated AUC, P = 0.92), suggesting that their relative impairment due to high-fat feeding was not to the same degree as that of control mice. IGF-I-deficient mice still differed from control mice in their response to insulin tolerance tests (12,152 vs. 15,429 integrated AUC, P = 0.002); however, it is interesting to note that high-fat-fed control mice were severely glucose intolerant and insulin resistant compared with their chow-fed controls, whereas chow-fed and high-fat-fed IGF-I-deficient mice differed minimally in these tests (Figs. 1 and 4). The IGF-I-deficient mice did maintain a slightly elevated fasting glucose level relative to controls in both tests. This elevated fasting glucose in IGF-I-deficient mice might be explained by our finding that gluconeogenesis, as measured by pyruvate challenge, continued to be greater in the IGF-I-deficient mice than in controls even on the high-fat diet (Fig. 4C).

Fig. 4.

IGF-I-deficient mice exhibited altered glucose metabolism when maintained on a high-fat diet. Graphs represent average values obtained from female control (○) and IGF-I-deficient (●) mice of ∼6 mo of age maintained for 12 wk on high-fat diet. Error bars represent SE; *P < 0.05 for unpaired t-test. A: glucose tolerance test performed after a 6-h fast, as described in materials and methods; control: n = 10; IGF-I deficient: n = 9. B: insulin tolerance test performed after a 6-h fast, as described in materials and methods; control: n = 10; IGF-1 deficient: n = 9. C: pyruvate tolerance test performed after an 18-h fast, as described in materials and methods; control: n = 7; IGF-I deficient: n = 9.

The reduced effect of high-fat feeding in IGF-I-deficient mice suggested that the global reduction in IGF-I may produce metabolic alterations toward preferred utilization of lipid resources. A cohort of mice was placed on a high-fat or control diet to specifically examine respiratory quotient (RQ). Differences in weight gain between control and IGF-I-deficient mice were very similar to the first cohort of animals (data not shown). Metabolic measurement performed over a 24-h period allowed calculation of RQ. These measurements demonstrated that the IGF-I-deficient mice exhibited a lower RQ, suggesting an increased utilization of fat as an energy source relative to control animals (Fig. 5). Continuous measurement over the 24-h period indicated that the IGF-I-deficient mice exhibited a lower RQ at all times; however, differences between the groups were greatest during the dark period (Fig. 5A). The percent of calories derived from fat was greatest in the IGF-I-deficient mice (Fig. 5B).

Fig. 5.

IGF-I-deficient mice had lower respiratory quotient (RQ) values on high-fat diet. A: RQ determined on the basis of oxygen consumption and carbon dioxide production continuously for a 24-h period on n = 6 of each genotype. Horizontal lines represent light cycle in animal rooms. B: %kcal derived from carbohydrate or fat consumption based on V̇o₂ consumption over 24 h. The overall difference in respiratory exchange ratio (RER) between the IGF-I-deficient mice and controls is significant at P < 0.005.

DISCUSSION

In this study, we describe the metabolic phenotype on both normal and high-fat diets of mice that harbor a hypomorphic allele of the Igf1 gene. Previously, we have shown that that maximum, although not mean, longevity is extended in these IGF-I-deficient mice. Our main finding is that although a reduction of IGF-I in these mice is associated with impaired glucose metabolism, it is also protective against fat accumulation caused by high-fat feeding. The IGF-I-deficient mice used in this study are distinct from other models of reduced IGF-I production in that they have reduced IGF-I in all tissues measured rather than loss of central, liver-specific production of IGF-I. The global reduction in IGF-I seems to contribute to alterations in glucose metabolism such that these mice are insulin resistant when fed a standard diet. However, the IGF-I-deficient mice show a significant advantage when they are fed a high-fat diet. This differential response is reflected in the lower fat mass of the IGF-I-deficient mice and in the lower overall weight gain in response to high-fat diet.

Although high-fat feeding promoted obesity and insulin resistance in control mice, these effects were relatively reduced in IGF-I-deficient mice. For example, the reduction in baseline glucose level 60 min following insulin challenge is reduced by 79% in control animals on the high-fat diet compared with those on a standard diet, whereas this difference is much smaller (30% reduction) in IGF-I-deficient mice. It must be noted that on both diets, IGF-I-deficient mice are less responsive to insulin than are control mice. One possible explanation may be that the altered metabolic state of the IGF-I-deficient mice provides these animals with the ability to preferentially utilize fatty acids as an energy source rather than having them be dependent on insulin-mediated glucose utilization. This interpretation is supported by the RQ measurements, which indicate that the IGF-I-deficient animals preferentially utilize fat as an energy source when provided a diet high in fat. Conversely, it is possible that this reflects an adaptation to the relative insulin resistance displayed by the IGF-I-deficient mice that is evident even on the control diet.

The results of this study are consistent with previous reports that modulating IGF-I can dramatically affect glucose metabolism in mice under both normal and high-fat diet conditions. For example, overexpression of IGFBP-1 effectively reduces IGF-I concentrations and also causes elevation of plasma insulin levels and reduction in insulin-stimulated glucose transport and glycogen synthesis in skeletal muscle, which is indicative of mild insulin resistance (36). This genetic mutation also causes glucose intolerance later in life (13). In addition, brain-specific overexpression of IGFBP-6 causes glucose intolerance and insulin resistance both on a control diet and on a high-fat diet and promotes increased weight gain (3). As mentioned above, reduction of IGF-I signaling by reducing levels of IGF-I receptors has also been associated with impairments in glucose metabolism under both normal and high-fat-fed conditions (4, 19, 21). Conversely, mice lacking IGFBP-3, -4, and -5, and thus displaying increased levels of IGF-I, show significant enhancement in several measurements of glucose metabolism, including glucose tolerance (31). Increased production of IGF-I alone can improve glucose metabolism and alleviate insulin resistance caused by high-fat diets, although weight gain is increased (41). Overall, these data are consistent with our finding that reduction of IGF-I in IGF-I-deficient mice may lead to development of insulin resistance under both normal and high-fat conditions.

Furthermore, our finding that IGF-I-deficient mice show elevated gluconeogenesis under both standard and high-fat feeding conditions may account for the elevated fasting glucose levels in the IGF-I-deficient mice. Treatment of mice with IGF-I has been shown to reduce glucose production from the liver (34), and increased levels of IGFBP-1 can increase gluconeogenesis (37). However, overexpression of IGFBP-2 in mice seems to reduce hepatic glucose production, reducing the expression of proteins important for gluconeogenesis such as phosphoenolpyruvate carboxykinase and glucose-6-phosphate dehydrogenase (37).

Downstream targets of the IGF-I receptor are key players in the response to high-fat diet and obesity. In many ways, the phenotype of IGF-I-deficient mice is similar to that of S6 kinase-null (S6K^−/−) mice described by Um and colleagues (47, 48). S6K^−/− mice do not show weight gain when placed on a high-fat diet, and given that the S6 kinase is a common substrate for both the IGF-I receptor and the insulin receptor, it may be that a reduction in IGF-I levels produces changes in insulin responsiveness. S6K^−/− mice also show reduced insulin signaling in young animals and reduced body size and adiposity similar to the IGF-I-deficient mice used in this study (48). Additionally, recent studies indicate that hypothalamic S6 kinase signaling is an important determinant of responses to high-fat feeding (33) and mediates metabolic shifts in gluconeogenesis (27). In IGF-I-deficient mice, S6 kinase signaling would be expected to be decreased, and thus, there may be good rationale to further study the relationship between downstream IGF-I responses and metabolic phenotypes seen in these animals.

A recent study of mice lacking both the insulin and the IGF-I receptor specifically in adipose tissue also revealed a phenotype remarkably similar to the IGF-I-deficient mice used in our work (6). The insulin and IGF receptor fat knockout mice displayed a reduced adiposity, were protected from weight gain on a high-fat diet, were resistant to glucose intolerance caused by a high-fat diet, and had increased metabolic rate. IGF-I-deficient mice have a similar phenotype in terms of response to high-fat diet and the ability to maintain glucose homeostasis as well as an increased ability to utilize fat as an energy source more readily than the control animals. It seems likely that the change in overall metabolic function may regulate both the resistance to weight gain and altered glucose metabolism in both models. For example, the shift in the ratio of lean mass to fat mass in the IGF-I-deficient mice may allow preferential delivery of free fatty acids to muscle, which can utilize the free fatty acids as an energy source, gaining energy from β-oxidation. Interestingly, this previous study utilized male mice, whereas our findings were generated in female IGF-I-deficient mice. In limited studies of male mice, our results indicate similar findings to those obtained in female mice. However, it will be of interest to confirm whether sex difference can significantly alter these outcomes in future studies.

It is possible that the metabolic and antiobesogenic effects of IGF-I reduction in these mice are due at least in part to alterations of other endocrine factors, primarily GH. As noted previously, levels of GH in these IGF-I-deficient mice are significantly elevated, which is likely due to compensatory feedback for the low levels of IGF-I (22). GH can cause insulin resistance directly by altering the activity of insulin-stimulated phosphatidylinositol 3-kinase, thereby reducing insulin signaling (44). The elevated levels of GH present in IGF-I-deficient mice may then contribute to the insulin-resistant phenotype as a result of the interaction between GH receptor signaling and insulin signaling (16). GH can also elevate blood glucose levels by stimulating gluconeogenesis, further contributing to hyperglycemia (43). The growth and expansion of adipose tissue can also be dramatically inhibited by GH; this hormone can block adipocyte differentiation and accumulation of triglycerides as well as stimulate lipolysis (39). Moreover, GH transgenic mice are resistant to the accumulation of body fat associated with high-fat feeding (32). To better understand the role of IGF-I in metabolic regulation, it will be crucially important to delineate those effects we report in these IGF-I-deficient mice that are due to increased GH and those that are due specifically to IGF-I deficiency.

GH and IGF-I have an intricate regulatory relationship, and this may also partially explain the relative dichotomy between longevity and metabolism displayed by long-lived mice deficient in both GH and IGF-I signaling and those deficient in IGF-I signaling alone (4, 5, 15, 16, 19, 21, 28). Perhaps the impairment of or lack of effect on glucose metabolism induced by deficiency in IGF-I alone contributes to the relatively smaller effect on lifespan found in GH-deficient mice that typically have an improvement in metabolic function. On the other hand, the phenotypes of the IGF-I-deficient mice used here have some similarities and some differences compared with mice with altered GH signaling both on control and on high-fat diet. Long-lived mice lacking GH receptor (GHRKO) are insulin sensitive yet moderately obese compared with control mice on normal diets (28). A study comparing GHRKO mice and mice that produce excess amounts of GH demonstrated that the presence of excess GH produced a shift in energy utilization that drives a disproportionate gain in lean body mass over fat mass (2). These results are consistent with protection against obesity induced by a high-fat diet but not against metabolic dysregulation in mice producing excess GH (32). However, GHRKO mice showed gains in relative body weight and reduction in insulin sensitivity on high-fat diet similar to those of control mice, indicating that a reduction in GH signaling alone does not provide protection against high-fat diet (2). It has become apparent that tissue-specific changes in GH signaling can influence the response to high-fat diets. For example, mice harboring a muscle-specific deletion of the GH receptor gene are resistant to weight gain induced by high-fat feeding and display a phenotype similar to the IGF-I-deficient mice described in our study (50, 51). Interestingly, the muscle-specific loss of GH receptor induces a shift in the relative utilization of fatty acids, as reflected in a shift in the RQ. In contrast, liver-specific deletion of the GH receptor reduces circulating IGF-I levels and promotes animal leanness but does not alter the metabolic utilization of substrates (24). However, deletion of the GH receptor in the liver causes insulin resistance, glucose intolerance, and elevated free fatty acids in the circulation, which can be reversed by the restoration of GH but not IGF-I signaling (17). Clearly, tissue-specific effects of GH and IGF-I signal responses exist, and it will be important in future studies to target their clarification in terms of their role in metabolism and longevity.

GRANTS

This work was supported by National Institutes of Health Training Grant T32-AG021890-05, Grant No. AG-223343 from the National Institute on Aging, funds from the Drexel University Aging Initiative, and Grant No. AG-02243 to C. Sell. A. B. Salmon was supported by the the Geriatric Research, Education, and Clinical Center of the South Texas Health Care System and a grant from the American Federation of Aging Research.

DISCLOSURES

The authors have no financial interests related to the work presented in this article to declare.

AUTHOR CONTRIBUTIONS

A.B.S. and C.S. conception and design of research; A.B.S., C.L., Y.I., S.M.P., R.M., and C.S. performed experiments; A.B.S., C.L., Y.I., S.M.P., R.M., and C.S. analyzed data; A.B.S., R.M., and C.S. interpreted results of experiments; A.B.S. and C.S. prepared figures; A.B.S. and C.S. drafted manuscript; A.B.S., C.L., Y.I., S.M.P., R.M., and C.S. edited and revised manuscript; A.B.S., C.L., Y.I., S.M.P., R.M., and C.S. approved final version of manuscript.