Studies of longevity and aging in species as varied as the microscopic worm Caenoerhabditis elegans, the fly Drosophila melanogaster, and the mouse Mus musculus indicate that partial suppression of insulin/insulin-like growth factor-1 (IGF-1) signaling pathway positively correlates with extended longevity (Brown-Borg et al., 1996; Bartke et al., 2001; Coschigano et al., 2003; Tatar et al., 2003; Bartke, 2008). Loss-of-function mutations of the Prop1 or Pit1 genes which produce deficiency of GH, prolactin (PRL), and thyrotropin in Ames and Snell dwarf mice or elimination of the GH receptor gene in GH receptor/GH-binding protein knockout (GHRKO) mice causes precipitous decline in circulating IGF-1, significant life extension and increased insulin sensitivity (Brown-Borg et al., 1996; Coschigano et al., 2000, 2003; Bartke et al., 2001; Tatar et al., 2003; Bonkowski et al., 2006a).
Calorie restriction (CR) extends longevity in mammals (Weindruch & Sohal, 1997; Tatar et al., 2003), at least partially, by altering the insulin/IGF-1 signaling pathway (Weindruch & Sohal, 1997; Tatar et al., 2003; Al-Regaiey et al., 2005; Masternak et al., 2005a,b; Bartke et al., 2007; Masternak & Bartke, 2007). In Ames dwarf mice, CR further extends lifespan and enhances insulin sensitivity (Bartke et al., 2001), suggesting actions via similar although not identical pathways.
Surprisingly, the same dietary intervention altered neither insulin sensitivity nor longevity in the GHRKO mouse (Bonkowski et al., 2006a). One could therefore speculate that GHR knockout and CR might act via the same mechanisms. One of the consistently observed actions of CR is a decrease in the amount of adipose tissue, and there is considerable evidence that the amount of fat tissue is positively associated with insulin resistance. However, in sharp contrast to CR animals, GHRKO mice are obese in comparison to their normal siblings (Berryman et al., 2004; Bonkowski et al., 2006b), yet are insulin sensitive, healthy and long-lived (Coschigano et al., 2003; Bonkowski et al., 2006a). It has been demonstrated that the site of fat accumulation may be more critical for health than the overall amount of fat tissue. Visceral (intra-abdominal or ‘central’) obesity promotes insulin resistance and increases the risk of type 2 diabetes, dyslipidemia, and mortality (Carey et al., 1997; Wang et al., 2005; Nicklas et al., 2006; Ross et al., 2008). Peripheral obesity – i.e., the increased amount of subcutaneous fat – is associated with increased insulin sensitivity and lower risk of type 2 diabetes and dyslipidemia (Misra et al., 1997; Snijder et al., 2003; Tanko et al., 2003). In humans, omentectomy (removal of visceral fat) was shown to reduce insulin and glucose levels (Thorne et al., 2002), whereas liposuction targets only subcutaneous fat tissue and does not result in any improvement of metabolic parameters (Klein et al., 2004). Diet and exercise result in preferential loss of visceral rather than subcutaneous fat and promote improvement of metabolic parameters (Langendonk et al., 2006). In rodents, fat distribution is also important in regulating insulin signaling. Mice subjected to a high-fat diet are characterized by increased visceral fat accumulation with a parallel decrease of insulin sensitivity (Dubuc, 1976; Rebuffe-Scrive et al., 1993; Barzilai et al., 1998). Aging is typically accompanied by an increase of visceral fat and decrease in insulin sensitivity, while surgical visceral fat removal (VFR) in rats improved insulin sensitivity and, more importantly, increased longevity (Barzilai et al., 1998; Gabriely et al., 2002).
It is possible that obesity in the GHRKO mouse (Bonkowski et al., 2006b) represents a form of ‘healthy’ obesity, because it involves preferential accumulation of subcutaneous fat although visceral fat may also be increased. Different, and conceivably beneficial, functions of visceral fat in GHRKO mice could also explain many of their phenotypic characteristics as well as lack of responses of these animals to CR in terms of insulin signaling and longevity (Bonkowski et al., 2006a, 2009). The present study was undertaken to compare the responses to VFR in GHRKO and normal mice. We hypothesized that the benefits of this intervention previously shown in genetically normal animals will be absent or reversed in the GHRKO animals.
The key conclusion emerging from the present study is that in the absence of GH signals, function of visceral adipose tissue and its metabolic impact are profoundly altered.
Our previous findings (Al-Regaiey et al., 2005; Masternak et al., 2005a,b; Bonkowski et al., 2006a,b; Bartke et al., 2007; Masternak & Bartke, 2007) imply that there may be different roles of the adipose tissue as an endocrine organ in GHRKO mice in comparison to their normal controls. The differences could involve an altered pattern of adipokine production and release, and indeed, plasma adiponectin levels are elevated in GHRKO mice. In support of this interpretation, the levels of adiponectin in epididymal fat from GHRKO mice were higher than the levels measured in subcutaneous fat from the same animals or in the subcutaneous or epididymal fat from normal controls. Subcutaneous fat is believed to be an important source of adiponectin present in the circulation. However, high levels of adiponectin in epididymal fat of GHRKO mice would support the hypothesis that epididymal fat from GHRKO mice has different biological function than the same adipose tissue in normal controls. Functional differences between visceral fat of normal and GHRKO mice also include reduced levels of resistin and IL-6 in perinephric fat of GHRKOs and downregulation of IL-6 in the epididymal fat of these animals. Because adiponectin is anti-inflammatory and promotes insulin sensitivity while IL-6 is proinflammatory, the observed alterations in adipokines levels in the visceral fat of GHRKO vs. normal mice support the hypothesis that these differences represent an important mechanism that regulates whole-body insulin sensitivity and health of these long-living animals.
Additionally, the analysis of genes involved in lipid metabolism indicated that relative expression of IR, PPARγ, PPARα, PGC1α, SERBP, and HSL was increased in epididymal fat from GHRKO as compared to normal mice without significant genotype effects in other fat depots. This indicates increased lipogenesis in GHRKO mice and could explain increased fat accumulation in these animals. However, GHRKO mice have also increased fat accumulation at subcutaneous and perinephric sites where we did not find any alterations in the expression levels of the same genes. Presumably, the differences in the metabolism and secretory function of epididymal fat are among important mechanisms of differential regulation of insulin sensitivity in GHRKO vs. normal mice with perinephric fat having no or a lesser role. Based on the present data and previous observation that GHRKO mice do not benefit from CR (Bonkowski et al., 2006a), we hypothesize that elevation of anti-inflammatory adiponectin and expression of genes promoting insulin sensitivity together with suppression of IL-6 levels represent mechanisms linking alterations in intra-abdominal adipose tissue biology with whole-body insulin sensitivity in these long-living animals. Following these findings, we decided to compare the effects of VFR in GHRKO and normal animals.
The present results demonstrate that surgical removal of visceral fat produces disparate, in many cases opposite effects in the genetically GH-resistant GHRKO as compared to normal mice. The net result is improvement of insulin sensitivity and induction of metabolic traits associated with increased longevity in normal animals and induction of insulin resistance and detrimental metabolic changes in GHRKO animals. These largely unexpected findings indicate that secretory activity of visceral fat of GHRKO mice contributes to the metabolic profile that favors extended longevity of these mutants. They also provide a likely explanation for the coexistence of obesity with enhanced insulin sensitivity and increased lifespan in these animals. Moreover, the present findings identify a likely reason why GHRKO mice do not benefit from CR by improved insulin signaling or a further extension of longevity.
In addition to their remarkable longevity (Tatar et al., 2003; Bonkowski et al., 2006a), GHRKO mice have extended ‘healthspan’ as indicated by protection from age-related decline of cognitive function and by reduced incidence and delayed occurrence of cancer (Ikeno et al., 2009). Similar findings have been recorded in hypopituitary Ames and Snell dwarf mice that are GH deficient (Ikeno et al., 2003). The physiological alterations observed in GH deficiency and GH resistance resemble in many ways the effects of CR. However, the lifespan of Ames dwarf mice can be further extended by CR (Bartke et al., 2001), suggesting that CR and GH deficiency do not affect longevity through identical mechanisms. Surprisingly, the GH-resistant GHRKO animals gained no further life extension benefit from CR (Bonkowski et al., 2006a). This unexpected finding raised a possibility of substantial overlap between mechanism responsible for extension of longevity by CR and by GH resistance. However, in contrast to animals subjected to CR, GHRKO mice are obese in comparison to normal littermates (Berryman et al., 2004; Bonkowski et al., 2006b) and yet are insulin sensitive and long-lived. Berryman et al. (2004) determined that the obesity of GHRKO mice was primarily because of increased subcutaneous fat depots with a trend toward increased visceral obesity in GHRKO mice at the age of 2 years. In our colony with a heterogeneous genetic background (Panici et al., 2009), GHRKO mice have similarly increased subcutaneous fat but also exhibit increased visceral obesity that is usually associated with increased insulin resistance and risk of diabetes. This sharply contrasts with insulin sensitivity and extension of healthspan and lifespan in GHRKO animals (Ikeno et al., 2003; Bonkowski et al., 2006a; Muzumdar et al., 2008; Panici et al., 2009). These counterintuitive findings and differences in biology of adipose tissue in GHRKO and normal mice lead us to speculate that the role of VF in the regulation of insulin/glucose metabolism in the GHRKO mice differs from its role in normal animals.
Decreased body temperature after VFR in normal animals resembles the effect observed in CR mice, which would agree with the hypothesis that VFR mimics CR (Barzilai et al., 1998; Gabriely et al., 2002; Muzumdar et al., 2008). Supporting this hypothesis, the lack of the body temperature alteration in GHRKO mice after VFR coincides with the previous findings that responses to CR are greatly attenuated in GHRKO mice (Bonkowski et al., 2006a, 2009). More importantly, the disparate effects of VFR on ITT, GTT, insulin and glucose levels (Gesing et al., 2011) in GHRKO as compared to normal mice indicate that eliminating or decreasing the amount of VF tissue by surgical or dietary interventions is not beneficial to GHRKO mice.
Consistent with the lipolytic actions of GH, lipolysis measured in vitro in adipose tissue from GHRKO mice was reduced in comparison to the values measured in fat derived from normal animals. Importantly, there was no difference in lipolysis levels between subcutaneous and epididymal fat from GHRKO mice, while epididymal fat from normal mice had much greater lipolysis level than subcutaneous fat from the same animals. It could be concluded that in terms of lipolysis, epididymal fat from GHRKO mice shares some characteristics with subcutaneous fat. Additional studies are needed to identify mechanisms responsible for these somewhat counterintuitive findings.
There is a well-established association between metabolic disorders and fat accumulation at ectopic sites. Surprisingly, no differences were detected in intra-muscular or intra-hepatic fat accumulation between normal and highly insulin-sensitive GHRKO mice. However, surgical VFR promoted reduction of intramuscular fat in normal mice, which likely contributed to the observed improvement in whole-body insulin sensitivity and glucose tolerance. Unexpectedly, the same procedure in GHRKO animals promoted an increased accumulation of lipids in skeletal muscle, which may be one of the key causes of decreased whole-body insulin sensitivity in GHRKO-VFR mice. Apparently VF influences fat distribution between the adipose tissue and tissues that normally store small amounts of lipid.
The analysis of the IR activation in skeletal muscle indicated that in normal animals, VFR enhances responses to insulin in this tissue. Decrease of the inhibitory IRS-1 phosphorylation at IRS1307ser, after VFR in normal mice, also implies improved insulin signaling and resembles findings in CR mice (Bonkowski et al., 2006a). Importantly, muscle levels of p-IRS1307ser in GHRKO mice were not affected by VFR (the present study) or by CR (Bonkowski et al., 2006a, 2009). The lack of alterations in the insulin signaling pathway in skeletal muscle of GHRKO mice after VFR corresponds to the lack of improvement in whole-body insulin sensitivity.
Following the demonstration of different responses of glucose metabolism responses to VFR in normal and GHRKO mice, we investigated the impact of this intervention on metabolic rate and transitions of fuel selection between fat and carbohydrate oxidations by measuring VO2 and RQ. RQ is a dimensionless ratio comparing the volume of carbon dioxide (VCO2), an organism produces over a given time, to its oxygen consumption (VO2). This ratio (RQ = VCO2/VO2) gives an estimate of the primary metabolic fuel source, which is typically either fat (RQ = 0.7) or carbohydrate (RQ = 1) oxidation. Thus, RQ varies inversely with lipid oxidation; a higher fasting RQ, which indicates lowered fat oxidation, is linked to body weight gain, metabolic inflexibility, and insulin resistance (Snitker et al., 1998). In the normal mouse, VFR decreased RQ, indicating increased fat oxidation. This novel effect of VFR suggests that the visceral fat that was removed was somehow impeding fat metabolism in the normal mice. Interestingly, the opposite was seen in the GHRKO mouse after VFR; the RQ of GHRKO-VFR mice was increased, indicating increased carbohydrate utilization compared to GHRKO sham animals. This suggests that visceral fat in GHRKO mice promotes the metabolism of fats, perhaps through adiponectin production.
In summary, surgical removal of visceral fat – generally considered to be ‘bad fat’– in normal animals broadly mimics the effects of CR in agreement with previous studies (Muzumdar et al., 2008). CR acts on adipose tissue to decrease fat content, but its actions are much more complex at the whole-animal level. The present study indicates that surgical VFR also produces complex alterations in whole-body insulin sensitivity/glucose regulation, oxygen consumption, RQ, ectopic fat distribution, and insulin signaling in skeletal muscle. The absence of beneficial effects of VFR in GHRKO mice indicates that the same endocrine organ plays a different role in GHRKO and N mice. Indeed, the present findings strongly suggest that VF has an unexpectedly important and positive role in regulating insulin action and perhaps also longevity in GHRKO mice. Finally, disparate effects of VFR in normal and GHRKO mice could explain why CR, known to target mainly adipose tissue, affects insulin sensitivity and longevity in normal, but not in GHRKO mice.
Collectively, the present findings suggest that GH signals in adipose tissue could be among key regulators of insulin action and longevity, and targeting suppression of this specific pathway in only white adipose tissue might produce beneficial CR-like effects in mammals. These findings, together with previous findings on alterations in expression of various genes in different organs of long-living mutants and CR animals (Al-Regaiey et al., 2005, 2007; Masternak et al., 2005a,b,c,d, 2006, 2009; Wang et al., 2006, 2007; Masternak & Bartke, 2007), indicate that a search for pharmaceutical interventions specifically targeting VF rather than working globally could be rewarding in terms of metabolic, anti-diabetic and longevity benefits. It might also reduce the wide list of potentially undesirable side effects.
Materials and methods
Experimental animals and animal maintenance
Normal and GHRKO male mice used in this study were produced in our breeding colony and developed by crossing 129Ola/BALB/c normal (GHR+/−) animals generously provided by Dr. J.J. Kopchick with mice derived from crosses of C57BL/6J and C3H/J strains; these mice were then maintained as a closed colony with inbreeding minimized by avoiding brother × sister mating (Panici et al., 2009). All animal protocols for this study were approved by the Southern Illinois University Laboratory Animal Care and Use Committee. The animals were housed under temperature- and light-controlled conditions (20–23 °C, 12-h light/12-h dark cycle) and were provided ad libitum with nutritionally balanced diet (Rodent Laboratory Chow 5001; 23.4% protein, 4.5% fat, 5.8% crude fiber; LabDiet PMI Feeds, Inc., St. Louis, MO, USA). To produce GHRKO (−/−) mice, knockout (−/−) males were mated with heterozygous (+/−) females. Heterozygous normal (+/−) males were used as controls based on previous data indicating that there are no significant phenotypic differences between normal homozygous and heterozygous mice (Coschigano et al., 2003).
Surgical visceral fat removal
At the age of about 5 months, GHRKO and normal littermates (N) were subjected to VFR or sham (S) surgery. All animals in these experiments were males. The animals were anesthetized with ketamine/xylazine, shaved, and prepared in the usual sterile fashion. Additionally, animals were supplied with ibuprofen in drinking water starting 2 days before and up to 3 days after the surgery. In the VFR group, for technical reasons, only the major fat pads epididymal and perinephric were removed, with no attempt to remove mesenteric or omental fat. The epididymal fat pads were removed using blunt dissection through a vertical midline incision, and then perinephric fat pads were removed via flank incisions. We removed as much epididymal or perinephric fat as was possible without compromising blood supply to the testes and to the adrenals. For sham operations, the abdominal cavity and both sides of the back were incised, and the VF was mobilized, but not removed. We performed three independent experiments at different time points with VFR to be able to measure different parameters.
Ten days postsurgery, animals were subjected to ITT as described below (n = 5–6). Fifteen days after surgery, GTT were performed as described below (n = 5–6).
At 55 days after VFR, after fasting over-night (c. 14 h), animals were anesthetized with ketamine/xylazine. Approximately 200–300 μL of blood was collected by cardiac puncture. After bleeding, half of the animals from each experimental group were injected with a high dose of insulin (10 IU kg−1 of body weight) or saline through the liver portal vein to stimulate the insulin signaling pathway following previously described protocol (n = 8–10) (Bonkowski et al., 2009). Exactly 2 min after insulin injection and sacrifice muscle, tissues were collected and kept on dry ice until moved to −80 °C freezer. At the time of tissue collection 55 days after surgery of VFR, there was no visible fat regrowth.
Assessment of blood chemistry
Plasma was obtained from blood collected by cardiac puncture and was used for assessment of adiponectin using Mouse Adiponectin ELISA and leptin using Mouse Leptin ELISA following manufacturer’s protocols (Linco Research Inc., St. Charles, MO, USA and IDS, Inc., Fountain Hills, AZ, USA).
Insulin tolerance test (ITT) and glucose tolerance test (GTT)
For ITT, nonfasted mice were injected i.p. with 0.75 IU insulin per kg of body weight. Blood glucose levels were measured at 0, 15, 30, and 60 min using a OneTouch Ultra glucometer. The data for ITT are presented as a percentage baseline of glucose. Fourteen hours fasted mice underwent GTT by intraperitoneal (i.p.) injection with 2 g of glucose per kilogram of body weight. Blood glucose levels were measured at 0, 15, 30, 45, 60, and 120 min.
Body temperature measurements
Body temperature was measured using Implantable Programmable Temperature Transponders (IPTT) provided by Bio Medic Data Systems (BMDS) Inc., Seaford, DE, USA. IPTT 300 transponders were implanted subcutaneously in the inter-scapular region using transponder trocor kit. Temperatures from the transponders were read using the DAS-6007 Smart Probe.
BMDS radio-telemetric probe reader
After transponder implantation, mice were allowed to heal for 2 weeks before measurements were taken. Mice were acclimated to the probes by taking mock measurements not used for data. Measurements were taken at two different time points: 6 am and 6 pm. Data were compiled and expressed as averages of each time point.
mRNA expression was analyzed by real-time PCR (RT–PCR) using the Smart Cycler instrument (Cepheid, Sunnyvale, CA, USA) with iQ™ SYBR Green Supermix (Bio-Rad, Hercules, CA, USA). Details of the procedure and the list of primers were reported previously (Masternak et al., 2005a,b,c, 2006, 2009).
Subcutaneous, epididymal, and perirenal adipose tissues were surgically removed from 9- to 9.5-month-old male GHRKO and normal mice. The tissues were rinsed three times with PBS buffer. Tissue pieces (about 100 mg) were incubated in Dulbecco’s modified Eagle medium (GIBCO, Invitrogen Corp., Carlsbad, CA, USA) with or without 10% FBS (GIBCO, Invitrogen Corp.) at 37 °C. Aliquots of the medium were collected after being cultured for 24 h and investigated for free glycerol content using commercial kit (Sigma, St. Louis, MO, USA).
Enzyme-linked immunosorbant assay for IR and insulin receptor substrate 1 (IRS-1) in skeletal muscle
The level of total IR and IRS1 and the phosphorylated forms (pY1158 IR and IRS1307ser) were analyzed by commercially available ELISA kit (Invitrogen, Carlsbad, CA, USA) according to provided protocol.
Magnetic resonance imaging (MRI)
Frozen tissue samples were thawed to 4 °C and analyzed using an EchoMRI 3-in-1 composition analyzer (Echo Medical Systems, Houston, TX, USA). Tissue composition was determined by the formula [fat mass/(fat mass + lean mass)].
Adult (6–7 month old) male GHRKO and normal littermates (N) that had been subjected to either VFR or sham (S) surgery (n = 8 per phenotype) were measured by indirect calorimetry using the PhysioScan Metabolic System (AccuScan Instruments, Inc., Columbus, OH, USA). This system utilizes zirconia and infrared sensors to monitor oxygen (O2) and carbon dioxide (CO2), respectively, inside of respiratory chambers in which individual mice were tested. All comparisons are based on animals studied simultaneously in eight different chambers connected to the same O2 and CO2 sensors in an effort to minimize the effect of environmental variations and calibration on data. After a 24-h acclimation period, mice were monitored in the metabolic chambers for 24 h with ad libitum access to standard chow (Laboratory Diet 5001) and water; and then for a second 24-h period without food. Gas samples were collected and analyzed every five minutes per animal, and the data were averaged for each hour. Output parameters include oxygen consumption (VO2, mL kg−1 min−1) and respiratory quotient (RQ, VCO2/VO2).
A simple one-way anova was used to analyze visceral fat deposition. Body temperature, blood chemistry, intra-hepatic and intra-muscular fat deposition, and insulin receptor mRNA and protein data were analyzed using two-way anovas. A three-way anova was used to analyze phosphorylation levels of IR and IRS1. A repeated measures anova was used to determine the interaction and main effects for ITT and GTT data followed by Fisher’s LSD tests for pairwise comparisons. Student’s t-test was used to analyze RQ, and two-way repeated measures anova was used to analyze VO2. α was set at 0.05 for determination of significance, and all values are reported as mean ± SEM throughout the figures and text.