Disruption of Growth Hormone Receptor in Adipocytes Improves Insulin Sensitivity and Lifespan in Mice

Abstract

Growth hormone receptor knockout (GHRKO) mice have been used for 25 years to uncover some of the many actions of growth hormone (GH). Since they are extremely long-lived with enhanced insulin sensitivity and protected from multiple age-related diseases, they are often used to study healthy aging. To determine the effect that adipose tissue has on the GHRKO phenotype, our laboratory recently created and characterized adipocyte-specific GHRKO (AdGHRKO) mice, which have increased adiposity but appear healthy with enhanced insulin sensitivity. To test the hypothesis that removal of GH action in adipocytes might partially replicate the increased lifespan and healthspan observed in global GHRKO mice, we assessed adiposity, cytokines/adipokines, glucose homeostasis, frailty, and lifespan in aging AdGHRKO mice of both sexes. Our results show that disrupting the GH receptor gene in adipocytes improved insulin sensitivity at advanced age and increased lifespan in male AdGHRKO mice. AdGHRKO mice also exhibited increased fat mass, reduced circulating levels of insulin, c-peptide, adiponectin, resistin, and improved frailty scores with increased grip strength at advanced ages. Comparison of published mean lifespan data from GHRKO mice to that from AdGHRKO and muscle-specific GHRKO mice suggests that approximately 23% of lifespan extension in male GHRKO is due to GHR disruption in adipocytes vs approximately 19% in muscle. Females benefited less from GHR disruption in these 2 tissues with approximately 19% and approximately 0%, respectively. These data indicate that removal of GH’s action, even in a single tissue, is sufficient for observable health benefits that promote long-term health, reduce frailty, and increase longevity.

A strong body of evidence indicates that a reduction in growth hormone (GH) action is associated with slowed aging (as defined by reduced signaling in key aging-associated pathways, increased longevity, and a decrease in age-associated diseases) in mammals (1-3). Indeed, the world’s longest-lived laboratory mouse—which lived 1 week shy of 5 years—was created by global disruption of the GH receptor (GHR) gene (4, 5). This growth hormone receptor knockout (GHRKO) mouse line was produced in our laboratory in 1997 (6) to provide a mouse model completely and specifically deficient in GH signaling (as opposed to Ames and Snell dwarf mice with multiple pituitary deficits [7]). Since GHR is absent in all tissues in GHRKO mice, GH action is eliminated while other pituitary hormones remain intact (as reviewed in [5, 8]). Of note, GHRKO mice are GH insensitive with low insulin-like growth factor-1 (IGF-1) and high circulating GH due to a lack of GH’s feedback inhibition. They also have reduced body size with a substantial increase in the subcutaneous to visceral white adipose tissue ratio (ie, a higher proportion of “healthy fat”) (6, 9). Importantly, these mice exhibit extreme longevity that cannot be further extended with calorie restriction or rapamycin injection (10-12).

In addition to living longer, GHRKO mice are protected from numerous age-related diseases as they are resistant to cognitive decline (13, 14), age-associated loss of muscle strength and coordination (15), diabetic nephropathy (16), diet-induced hyperglycemia and hyperinsulinemia (17), and multiple forms of cancer (18-20). Mechanistically, GHRKO mice have been shown to have beneficial reductions to key aging-associated pathways including reduced mTOR complex-1 (mTORC1) activity (21), reduced cellular senescence (22), decreased insulin exposure (23), and resistance to multiple forms of stress (24). Furthermore, enhanced insulin sensitivity is a prominent phenotype of these mice, which may also partially explain their improved healthspan/lifespan (23, 25-28). Because of the aforementioned health benefits and extreme longevity, these mice have become a useful tool to better understand the role of GH in aging.

While knockout mice are extremely useful for elucidating global effects of a particular protein, elucidating direct actions on individual tissues has proven challenging. To this end, recent innovations in transgenic/knockout mouse technology using a Cre-Lox system has allowed researchers to specifically investigate tissue-specific activities by using cell type–specific promoter/enhancers. Accordingly, more than 20 new tissue-specific GHR “knockout” mouse lines have been generated in the past decade to allow researchers more precise tools for investigating the direct actions of GH in specific tissues/cell types (29-31). While these new GHR tissue-specific knockout lines have provided valuable data in more than 125 publications (30), only 2 of these publications have reported effects on lifespan (21, 32). This is likely because aging studies in these mice are extremely expensive as they require large numbers of mice and take up to 5 years to complete. However, considering the prominent role GH plays in the aging process, it would be of great interest to determine lifespan in these new tissue-specific mouse lines and help dissect GH’s tissue specific roles in aging. Therefore, we conducted the present study in recently generated adipocyte-specific GHR knockout (AdGHRKO) mice (33) to determine if disruption of GH action in a single insulin-sensitive cell type, that is, adipocytes, is sufficient to alter lifespan. In our original assessment of this mouse line, measurements were reported up to age 6 months and revealed that AdGHRKO mice have increased adiposity and enhanced insulin sensitivity compared to controls (33). In the present study we evaluate similar parameters at advanced ages along with measures of grip strength, frailty, and lifespan.

Materials and Methods

Animals/Housing

AdGHRKO mice used in this study have recently been characterized and described (33). Briefly, mice with lox P sites flanking exon 4 of the GHR (GHRflox/flox) were created by the Knockout Mouse Project as previously described (34, 35). Mice expressing adipocyte-specific Cre under the control of adiponectin promoter/enhancer, B6;FVB-Tg(Adipoq-cre)1Evdr/J (stock No. 010803), were purchased from Jackson Laboratory. AdGHRKO mice were generated by breeding GHRflox/flox mice with B6;FVB-Tg(Adipoq-cre)1Evdr/J mice. Male and female Cre-expressing mice were used as breeders with no differences observed in offspring.

All mice were bred at Ohio University (OU) and weaned into permanent housing no earlier than age 21 days and no later than age 28 days. At weaning, ear notching was performed for mouse identification and a portion (< 1 mm) of the distal tail was removed with scissors for genotyping. Mice were housed 3 or 4 per cage in ventilated microisolator cages using one-quarter-inch corn cob bedding (Bed O Cobs, produced by The Andersons) and given ad libitum access to water and rodent chow (ProLabRMH 3000; 14% of energy from fat, 60% from carbohydrates, and 26% from protein; PMI Nutrition International). The cages were maintained in a temperature- (22 ± 2 °C) and humidity-controlled room and exposed to a 14-hour light, 10-hour dark cycle. Fresh cages and bedding were exchanged once per week. The microbial status of the room was specific-pathogen free as determined by quarterly serology and semiannual parasitology reports from sentinel animals placed in used bedding. All procedures were approved by the OU Institutional Animal Care and Use Committee and fully complied with all federal, state, and local policies. Both male and female mice were used for all measures.

Estimation of Age of Death

Mice were examined daily for signs of poor health and euthanized if considered by an experienced technician unlikely to survive more than a week. A mouse was considered moribund if exhibiting more than one of the following signs: large or bleeding tumor, prolapsed anus or vagina, inability to eat or drink, severe lethargy, rapid weight loss over 1 or more weeks, severe balance or gait disturbance, or persistent ulceration larger than 1 cm² that does not resolve from trimming of nails. Eighteen total mice were euthanized because of deteriorating health conditions and for survival analysis were counted as deaths the day euthanasia was performed. The proportion of mice euthanized in this group are as follows: 7 female AdGHRKO mice, 3 female floxed littermate controls, 4 male AdGHRKO mice, and 4 male floxed littermate controls. All other mice were counted as dead the day the mice were found dead in the cage with cages being checked daily including weekends.

Body Composition

Body composition was determined using a nuclear magnetic resonance imaging Bruker Minispec (Bruker Corp) (36, 37). To minimize stress, measurements were limited to once every 3 months starting at age 3 months and ending at age 24 months, with 34 to 42 mice per group.

Insulin Tolerance Tests

The 0 time point of the assay was collected at 1500 hours following a 6-hour fast. Mice were given 0.050 U/mL of recombinant human insulin (Humulin-R; Eli Lilly & Co) at 0.01 mL/g body weight. Blood was collected at 0, 15, 30, 45, 60, and 90 minutes by cutting 1 mm from the tip of the tail. Blood glucose measurements were collected using OneTouch Ultra glucometers and test strips (LifeScan). Insulin tolerance tests (ITTs) were performed on mice that were approximately aged 18 months, with 20 per group.

Blood Measurements

Plasma measurements were performed from blood samples collected at age 18 months following an overnight 12-hour fast. Insulin, C-peptide, interleukin-6 (IL-6), resistin, leptin, monocyte chemoattractant protein-1 (MCP-1), and tumor necrosis factor α (TNF-α) were measured using a mouse metabolic magnetic bead panel (RRID: AB_2801416; catalog No. MMHMAG-44K; MilliporeSigma) and a Luminex 200 System (catalog No. 40-012; MilliporeSigma) as previously described (38). Adiponectin was measured using a mouse high-molecular-weight and total adiponectin enzyme-linked immunosorbent assay kit (RRID: AB_2801466; catalog No. 47-ADPMS-E01; ALPCO) according to the manufacturer’s instructions, with 7 to 11 mice per group.

Measures of Frailty

Frailty and functional measures were evaluated at 20 months and again between 28 and 30 months of age by measuring grip strength, motor coordination, and by grading clinical signs of deterioration by use of a mouse frailty index (39). Forelimb grip strength was assessed using 2 distinct methods: 1) an automated grip strength force meter (catalog No. 1027SM; Columbus Instruments) (40); and 2) a grip strength procedure (Deacon’s weightlifting) that uses progressively increasing weights (20.1 g, 32.2 g, 44.7 g, 57.2 g, 69.5 g) and a scoring system (41). For both methods, measures were taken at 20 months and 28 months of age, with mice trained 4 days in a row before testing and the average of 5 measurements recorded. Motor coordination was assessed by using an automated 4-lane rotarod (Columbus Instruments). Briefly, at age 20 and 29 months, mice were pretrained 4 days in a row before testing by placing them on a moving rotarod at 10 rpm until they performed at this speed for 120 seconds. The test phase consisted of 5 trials separated by 15 minutes intertrial intervals. The testing apparatus was set to accelerate from 4 to 40 rpm in 300 seconds. One mouse was then placed on each lane and the rotarod was started with an initial rotation of 4 rpm. The rotational velocity was set to increase every 10 seconds and the latency to fall was recorded in seconds. A mouse clinical frailty index system developed by the Howlett laboratory was used to evaluate frailty at age 20 and 30 months as described previously (39). A member of the Howlett lab (E.S.B.) traveled to OU and trained our laboratory members how to grade the mice for clinical signs of deterioration using the frailty index.

Statistical Analysis

Using a 2-sided 5% significance level, we performed the log-rank test for each sex individually and with both sexes combined to determine whether survival for AdGHRKO mice differs from that of control mice. For evaluation of maximal survival, a Fisher exact test at 90% survival was conducted. For grip strength, motor coordination, frailty index scores, and measures of circulating peptides (insulin, c-peptide, leptin, adiponectin, resistin, IL-6, MCP-1, TNFα), 2-way analysis of variance (genotype and sex) was conducted. For body composition data collected over 2 years of life including body weight, fat mass, lean mass, and fluid mass (all of which had missing values due to deaths at later time points), a mixed-effects model with Geisser-Greenhouse corrections was used. For ITTs, 2 statistical methods were used: 1) unpaired t test of the area under the curve to compare overall insulin tolerance between genotypes, and 2) 2-way analysis of variance (genotype and time) was used combined with the Bonferroni multiple comparison to compare genotype means at each time point.

Results

Evaluation of Lifespan

Lifespan of male and female AdGHRKO mice (153 mice in total with the following numbers per group: 37 female AdGHRKO mice, 34 female floxed littermate controls, 40 male AdGHRKO mice, and 42 male floxed littermate controls) was evaluated. All mice in this cohort were specifically bred for this longevity study, and no mice were removed from the study for use in other studies. There were no accidental deaths due to handling/experimental procedures. Male AdGHRKO mice had a statistically significant increase in lifespan (P = .039) vs control mice (Fig. 1A). In contrast, female AdGHRKO mice had a similar lifespan (P = .071) to that of controls (Fig. 1B). When sexes were combined, AdGHRKO mice significantly outlived (P = .015) their control littermates (Fig. 1C). Mean survival for AdGHRKO vs controls was increased 6% (944 days for AdGHRKO vs 891 days for controls) for males and 3% (828 days for AdGHRKO vs 806 days for controls) for females. Median survival for AdGHRKO vs controls was 963 vs 934 days in males, 876 vs 826 days in females, and 912 vs 869 when the sexes were combined. Maximal lifespan as determined by Fisher exact test was not significantly altered in males (6 of 40 surviving past 90% for AdGHRKO compared to 2 of 42 for controls; P = .150) or females (5 of 37 surviving past 90% for AdGHRKO compared to 2 of 34 for controls; P = .432).

Figure 1.

Kaplan-Meier survival plots for AdGHRKO mice. AdGHRKO mice (red) and littermate controls (black) are shown for males (A, top left), females (B, top right), and combined sexes (C, bottom). P values are provided for log-rank tests and shown in bold if statistical significance is reached (P < .05). AdGHRKO, adipocyte-specific growth hormone receptor knockout.

Grip Strength, Rotarod, and Frailty Index

Grip strength was evaluated in older adult mice (aged 20 months) and again at an advanced age (aged 28 months) using 2 separate methods, force meter, and Deacon’s weightlifting. Using the force meter method (Fig. 2A), no genotype effect was observed (P = .895) between AdGHRKO and control groups at age 20 months. However, at age 28 months, a statistically significant genotype effect was observed (P = .008) with AdGHRKO mice having greater grip strength than controls (Fig. 2B). Using Deacon’s weightlifting method of evaluating grip strength (Fig. 2C), similar results were observed in grip strength as no genotype effect was detected (P = .561) at age 20 months but at age 28 months, a statistically significant genotype effect became apparent (P = .011) with AdGHRKO mice having a greater grip strength (Fig. 2D). Rotarod testing at age 20 months (Fig. 2E) and again at age 29 months (Fig. 2F) revealed no difference in motor coordination between genotypes at either age (P = .987 and P = .245, respectively). Using the mouse clinical frailty index, we evaluated frailty at ages 20 and 30 months (Fig. 2G and 2H). Similar to grip strength results, no genotype effect was observed at age 20 months (P = .715) but at advanced age (age 30 months), AdGHRKO mice exhibited less frailty than controls (P = .015).

Figure 2.

Various measures of frailty in AdGHRKO vs control. Force meter grip strength at A, age 20 and B, 28 months. Decan’s weightlifting grip strength at C, age 20 and D, 28 months. Rotorod measured at E, age 20 and F, 29 months. Frailty index at G, age 20 and H, 30 months. Values are gives as mean ± SEM in addition to individual values. P values from 2-way analysis of variance appear in the boxes to the right. AdGHRKO, adipocyte-specific growth hormone receptor knockout.

Body Composition Over Time

Body composition was measured every 3 months from ages 3 to 24 months. Body weights were statistically significantly increased in male (P = .003) and female (P = .002) AdGHRKO mice compared to controls (Fig. 3A and 3B). Likewise, AdGHRKO fat mass was significantly increased (P < .001) in both sexes (Fig. 3C and 3D). Lean mass as well as fluid mass were unchanged between genotypes regardless of sex (Fig. 3E-3H).

Figure 3.

Body composition over 2 years in AdGHRKO mice. A and B, Body weight, C and D, fat mass, E and F, lean mass, and G and H, fluid mass was evaluated every 3 months from age 3 to 24 months in males (left side) and females (right side). Values are mean ± SEM. # indicates P less than .05 for genotype in repeated-measures analysis of variance. AdGHRKO, adipocyte-specific growth hormone receptor knockout.

Insulin Tolerance, Fasting Insulin, and C-Peptide

Several measures of glucose homeostasis were measured at age 18 months. Male AdGHRKO mice had significantly enhanced insulin tolerance compared to controls (Fig. 4A) at 45, 60, and 90 minutes post insulin injection as well as by comparison using area under the curve. In contrast, female AdGHRKO mice showed a mild improvement that became significant only at the 90-minute time point (Fig. 4B) and was not significant using area under the curve. Fasting levels of circulating insulin (Fig. 4C) and c-peptide (Fig. 4D) were significantly reduced in the AdGHRKO genotype vs controls (P < .001 and P = .006, respectively) with a statistically significant effect of sex on both measures (P < .003 and P = .001, respectively); however, no interaction between genotype and sex was observed.

Figure 4.

Insulin sensitivity, plasma insulin, and c-peptide levels at age 18 months. A and B, Insulin tolerance test (ITT) results and corresponding area-under-the-curve values are shown for males and females, respectively. C, Fasting plasma insulin and D, c-peptide levels are shown for males and females, respectively. Values are gives as mean ± SEM in addition to individual values. P values from 2-way analysis of variance appear in boxes to the right. AdGHRKO, adipocyte-specific growth hormone receptor knockout.

Circulating Cytokines/Adipokines

Fasting serum cytokine levels were measured at age 18 months. No effect of genotype was seen for leptin, IL-6, MCP-1, or TNFα (Fig. 5A and 5D-5F). In contrast, the AdGHRKO genotype had statistically significantly reduced levels of adiponectin and resistin (Fig. 5B and 5C). A significant effect of sex was seen for leptin, adiponectin, and resistin (P = .001, P = .001, and P = .013, respectively) but no interaction between genotype and sex was observed.

Figure 5.

Plasma cytokine levels at age 18 months. Circulating levels of A, leptin; B, adiponectin; C, resistin; D, IL-6; E, MCP-1; and F, TNFα. Values are gives as mean ± SEM in addition to individual values. P values from 2-way analysis of variance appear in the boxes to the right and post hoc P values are shown when statistical significance is reached (P < .05). AdGHRKO, adipocyte-specific growth hormone receptor knockout; IL-6, interleukin-6; MCP-1, monocyte chemoattractant protein-1; TNFα, tumor necrosis factor α.

Discussion

Global GHRKO mice are remarkably long-lived with enhanced insulin sensitivity. Enhanced insulin sensitivity in GHRKO mice is not surprising since GH is well known to have diabetogenic (or anti-insulin) activity, and removal of GH action would remove this diabetogenic activity, promoting enhanced insulin sensitivity. To test the hypothesis that removal of GH action in a single insulin-sensitive cell type, adipocytes, might partially replicate the enhanced insulin sensitivity and increased lifespan observed in global GHRKO mice, we performed aging studies on AdGHRKO mice. Our results show that disrupting the GHR gene in adipocytes improves insulin sensitivity and increases lifespan in AdGHRKO mice. AdGHRKO mice also exhibit increased fat mass, reduced circulating levels of insulin, c-peptide, adiponectin, resistin, and reduced frailty scores with improved grip strength at advanced ages. Female AdGHRKO mice were less protected by the adipose growth hormone receptor knockout as insulin tolerance was altered at only a single time point of the ITT, fat mass gain was less pronounced than males, and the increase in lifespan did not reach statistical significance (P = .07).

The present study demonstrates that targeting GHR in a single insulin-sensitive cell type, adipocytes, is sufficient to improve insulin sensitivity, reduce late-life frailty, and increase lifespan. This effect was relatively mild compared to the health benefits reported in global GHRKO mice, which exhibit larger increases in lifespan (5, 10, 11, 42). This may partially be due to the background strain used in this study since C57BL/6J mice are already relatively long-lived with lower GH/IGF-1 compared to other mouse strains; thus, further gains in lifespan are more difficult to achieve when C57BL/6J mice are used. More specifically, in a large-scale aging study conducted by the Jackson Laboratories, the C57BL/6J strain had the fourth lowest levels of IGF-1 and the second longest lifespan out of 32 mouse strains (43). In agreement, while GHRKO mice exhibit a 26% to 55% increase in average lifespan in males and a 16% to 38% increase in females, the smallest lifespan increases reported were from GHRKO mice with a C57BL/6J background (26% for males and 16% for females) (10). Regardless, it is remarkable that knockdown of GHR in a single cell type, adipocytes, had a measurable effect on whole-body insulin sensitivity and lifespan, considering GHR are found on most cells in the body. In attempts to estimate the contribution that adipocyte GHR gene disruption has on the extended lifespan of GHRKO mice, we compared the mean lifespan increases for both lines. For AdGHRKO mice, mean lifespan was increased 6% (944 days for AdGHRKO vs 891 days for controls) for males and 3% (828 days for AdGHRKO vs 806 days for controls) for females. Since the present study used a C57BL/6J background strain, we can compare our results with that of Coschigano and colleagues in 2003 (10) (same C57BL/6J background as well as the same animal facility). The removal of GHR in adipocytes alone contributes approximately 19% to 23% of the lifespan extension seen in C57BL/6J mice with global knockout of GHR (3 of 16 = 19% in females, 6 of 26 = 23% in males). Stated another way, 77% to 81% of lifespan extension in GHRKO mice is due to the removal of GHR in tissues other than adipocytes. To put GHRKO lifespan extension into human perspective (based on 2019 US Centers for Disease Control and Prevention estimates of 76.3 for US male life expectancy and 81.4 for US females), a 26% to 55% extension in lifespan equals 20 to 42 years of extra life for men and a 16% to 38% increase equals 13 to 31 years for women. Even the mild 6% and 3% increases seen in AdGHRKO mice seems less trivial if extended to humans, with 4.6 and 2.4 years of extra life for men and women, respectively. However, this experimental paradigm using tissue-specific knockouts is not possible to recapitulate in humans. Rather, these tissue-specific GHRKO lines are meant to allow for the dissection of GH action in living systems and approximation of tissue contribution. Furthermore, despite the growing evidence in mice and numerous other organisms that suggest that blocking GH action promotes healthy aging, the question of whether GH deficiency or maintaining GH levels are beneficial for lifespan or healthspan in humans remains controversial.

Analysis of the AdGHRKO line compared with lifespan data from other GHR tissue-specific knockout lines should give greater insight into which tissues provide the greatest benefit when targeting GH action. To date, only 2 other tissue-specific GHR knockout mouse lines—a liver-specific GHRKO line (LiGHRKO) and a muscle-specific GHRKO line (MuGHRKO) (21, 32, 35)—have had longevity data reported. Lifespan in LiGHRKO mice is unchanged. Interestingly, these mice have a 90% reduction in circulating IGF-1 and reduced circulating IGF-1 is thought to be a primary reason why GHRKO mice are long-lived. However, there are many complex factors to consider when interpreting these results. LiGHRKO mice have a unique phenotype since liver is a key organ in the GH/IGF-1 axis (~ 90% of circulating IGF-1 comes from GH-stimulated IGF-1 secretion in liver) (35, 44). Furthermore, LiGHRKO mice lack GHRs in hepatic tissue; however, all other tissues still express GHR. Thus, the high circulating GH levels in these mice result in aberrant feedback inhibition and extrahepatic acromegaly (45). High circulating GH, while unable to act on the liver (perpetuating the low circulating IGF-1), is fully capable of increasing local IGF-1 in all other nonhepatic tissues. Hence, a unique divergence in the GH/IGF-1 axis exists in LiGHRKO mice with elevated circulating “endocrine” GH, low endocrine IGF-1, and elevated autocrine/paracrine IGF-1 (35). In fact, LiGHRKO mice have an increase in local IGF-1 within muscle and adipose tissue (35)and impaired glucose homeostasis (35, 44, 46). Thus, we can gather from LiGHRKO mice that decreased circulating IGF-1 is not the only hormone responsible for extended lifespan in global GHRKO mice.

When GHR is specifically disrupted in muscle tissue, a significant increase in lifespan is observed (32). More specifically, mean lifespan was increased 5% (839 days for MuGHRKO vs 798 days for controls) for males and 0% (796 days for MuGHRKO vs 798 days for controls) for females. Since MuGHRKO mice are also from a C57BL/6J background, we can compare mean lifespan to that of GHRKO data from Coschigano et al (10). Based on this estimate, the removal of GHR in skeletal muscle contributes approximately 19% of the lifespan extension seen in male GHRKO mice and does not appear to contribute to female GHRKO lifespan extension. This sex-specific extension in male lifespan is not surprising since MuGHRKO males exhibited a greater enhancement to insulin sensitivity. These results are similar to results seen in the present study as AdGHRKO males have a more pronounced improvement to insulin sensitivity and increased lifespan. The sexual-dimorphic improvements to insulin sensitivity in AdGHRKO and MuGHRKO male mice might be explained, in part, by sex differences in baseline glucose homeostasis. That is, male C57BL/6J controls have elevated fasting insulin levels and impaired insulin sensitivity relative to female controls, as seen in the present study, and as reported by others (47, 48). Furthermore, female C57BL/6J control mice are relatively protected from high-fat diet–induced alterations in glucose homeostasis compared to males (49). Thus, the reason why females do not achieve significant improvements to lifespan when GHR is disrupted specifically in muscle or adipose tissue may be that further improvements in glucose homeostasis are more difficult to achieve in females. These sex-specific alterations in the glucose homeostasis of control mice may be due to estrogen, as estrogen can increase phosphorylation of insulin signaling intermediates IRS and Akt (50) and activate GLUT-4 and glucose uptake into muscle (51). Furthermore, estrogen can influence adipose tissue development and function, which may also contribute to sex-specific differences in glucose homeostasis (52).

Our findings that tissue-specific disruption of GHR in adipocytes extends male lifespan are in agreement with previous findings in muscle (as discussed earlier); however, they are in contrast with other previous findings when GH action is knocked down globally after birth. More specifically, when adult-onset isolated GH deficiency (AOiGHD) is induced in mice using diphtheria toxin–mediated destruction of GH producing somatotrophs at age 10 to 12 weeks, female but not male AOiGHD mice exhibit increased lifespan (53). Similarly, global knockdown of GHR using a tamoxifen-inducible system at age 1.5 months (1.5mGHRKO mice) (54) or at age 6 months (6mGHRKO mice) (55) extends female but not male lifespan. The reason males did not live longer in these inducible lines remains unclear; however, it could be due to the loss of sex-dependent GH circadian pattern. In males (mice and humans), the secretory pattern of GH is more pulsatile compared to a more tonic pattern in females. Thus, it is possible that males could be more adversely affected by methods that inhibit GH action (due to the reduction of secretion levels or the loss of GHR). Accordingly, circulating levels GH and IGF-1 are normal in AdGHRKO and MuGHRKO lines, but significantly altered in AOiGHD, 1.5mGHRKO, and 6mGHRKO lines.

For the tamoxifen-induced lines, there are several additional points. First, it is important to note that 6mGHRKO males had a trending increase in lifespan (P = .088), in addition to several significant health improvements, including enhanced insulin sensitivity and reduced cancer (55). Thus, it is possible that larger numbers of mice may have yielded a considerable increase in male lifespan. Second, tamoxifen-inducible disruption of GHR produces a complicated phenotype with certain tissues, such as liver, having nearly complete GHR disruption while other tissues (eg, kidney, muscle, and heart) have partial GHR disruption (54, 55). As a result, reduction in local IGF-1 expression varies and is not completely replicated between sexes (ie, IGF-1 expression in 6mGHRKO liver and perigonadal white adipose tissue are higher in females vs males while IGF-1 expression is lower in kidney of females vs males). Third, the tamoxifen used to induce these temporal knockdowns has the potential to cause negative health outcomes in mice, especially in males (56-58). However, despite only partial knockdown of GHR and possible sex-specific effects of tamoxifen treatment, disruption of GHR after birth produces observable health benefits in both sexes and is consistent with the growing evidence that targeting GH action extends healthy aging. To this end, we believe that future studies using improved inducible systems that more efficiently and ubiquitously knock down GHR without negative health effects or alternative treatments to block GH action safely and efficiently (such as the use of GHR antagonists) will produce even greater improvements to health and lifespan in both sexes.

AdGHRKO mice are resistant to late-life frailty as demonstrated in 3 of 4 different tests used. More specifically, grip strength is significantly enhanced/preserved in AdGHRKO mice at advanced ages (30 months) compared to controls. Similarly, AdGHRKO mice had significantly better mouse clinical frailty index scores at age 30 months vs controls. This finding partially replicates what is observed in global GHRKO mice as they also are resistant to decline in grip strength, balance, and motor coordination at advanced ages; however, the clinical frailty index has not been used on GHRKO mice (15). Furthermore, this effect is not limited to GHRKO mice as GH-deficient Ames dwarf mice exhibit reduced frailty with advanced ages (59). Thus, reduced GH action in mice by multiple methods (GH deficiency or GHR ablation) appears to attenuate late-life frailty.

In conclusion, disruption of GHR in adipocytes increased fat mass, altered adipokine profiles, reduced frailty, enhanced glucose homeostasis, and increased lifespan in male mice. Female AdGHRKO mice had a milder phenotype that was similar to our previous reports for muscle-specific MuGHRKO mice, which also exhibited enhanced insulin sensitivity and increased lifespan in males. While it is not possible to recapitulate this experimental paradigm in humans, these tissue-specific GHRKO lines allow for the dissection of GH action in living systems and approximation of tissue contribution. To this end, an estimated 19% to 23% of lifespan extension in GHRKO is predicted to result from GHR disruption in adipocytes vs 0% to 19% in muscle (female to male, respectively).

Glossary

Abbreviations

AdGHRKO

adipocyte-specific growth hormone receptor knockout

AOiGHD

adult-onset isolated growth hormone deficiency

GH

growth hormone

GHRKO

growth hormone receptor knockout

IGF-1

insulin-like growth factor-1

IL-6

interleukin-6

ITT

insulin tolerance test

MCP-1

monocyte chemoattractant protein-1

MuGHRKO

muscle-specific GHRKO line

OU

Ohio University

TNFα

tumor necrosis factor α

Financial Support

This work was supported by the National Institutes of Health (grant No. R01AG059779); the State of Ohio’s Eminent Scholar Program, which includes a gift from Milton and Lawrence Goll; the AMVETS; and by the Diabetes Institute at Ohio University.

Disclosures

The authors have nothing to disclose.

Data Availability

Original data generated and analyzed during this study are included in this published article or in the data repositories listed in “References.”