• Open Access

The growth hormone receptor gene-disrupted mouse fails to respond to an intermittent fasting diet


  • Oge Arum,

    1. Department of Internal Medicine, Southern Illinois University-School of Medicine, Springfield, IL 62794, USA
    Search for more papers by this author
  • Michael S. Bonkowski,

    1. Department of Internal Medicine, Southern Illinois University-School of Medicine, Springfield, IL 62794, USA
    Search for more papers by this author
    • *

      Present address: Michael S. Bonkowski, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.

  • Juliana S. Rocha,

    1. Laboratory of Cellular Biology, Department of Morphology, Institute of Biological Sciences, Federal University of Minas Gerais, Belo Horizonte, Brazil
    Search for more papers by this author
  • Andrzej Bartke

    1. Department of Internal Medicine, Southern Illinois University-School of Medicine, Springfield, IL 62794, USA
    Search for more papers by this author

Oge Arum, Southern Illinois University-School of Medicine, 801 N. Rutledge St., Rm. 4324, PO Box 19628, Springfield, IL 62794-9628, USA. Tel.: (217) 545 2193; fax: (217) 545 8006; e-mail: oarum@siumed.edu


The interaction of longevity-conferring genes with longevity-conferring diets is poorly understood. The growth hormone receptor gene-disrupted (GHR-KO) mouse is long lived; and this longevity is not responsive to 30% caloric restriction, in contrast to wild-type animals from the same strain. To determine whether this may have been limited to a particular level of dietary restriction, we subjected GHR-KO mice to a different dietary restriction regimen, an intermittent fasting diet. The intermittent fasting diet increased the survivorship and improved insulin sensitivity of normal males, but failed to affect either parameter in GHR-KO mice. From the results of two paradigms of dietary restriction, we postulate that GHR-KO mice would be resistant to any manner of dietary restriction; potentially due to their inability to further enhance insulin sensitivity. Insulin sensitivity may be a mechanism and/or a marker of the lifespan extending potential of an intervention.


In different animal species, aging can be postponed and longevity increased by dietary restriction (DR) or mutations of genes involved in somatotrophic/insulin signaling. The mechanisms by which these mutations and DR affect aging are apparently overlapping, although not identical, and dietary intake can interact with longevity genes in various ways (Piper & Bartke, 2008). In mice, we have previously reported that an identical regimen of DR produces further longevity extension in long-lived Ames dwarf mice (Bartke et al., 2001), but is ineffective in another long-lived mutant, the growth hormone (GH) resistant Ghr−/− (GHR-KO) mouse (Zhou et al., 1997; Coschigano et al., 2000, 2003; Bonkowski et al., 2006).

Because, in Drosophila melanogaster, longevity genes can alter energy intake requirements for optimal survival (Clancy et al., 2002; Wang et al., 2009) it was of interest to determine if GHR-KO mice will respond to a different DR regimen, intermittent fasting (IF) (Anson et al., 2005).

Results and discussion

We determined the gender-specific caloric intake for ad libitum (AL) vs. IF mice at approximately 7–8, 14–15 and 32–33 months of age. All animals were fed AL for the first 8–10 weeks of age. Subsequently, the mice were either fed AL every day (AL group) or every other day (IF group). Average food intake was numerically reduced in almost all IF groups, but the difference between AL and IF animals was statistically significant only in normal males at 14–15 months of age (23% reduction, < 0.02); thus confirming the reported compensatory feeding of IF animals during their AL phase (Anson et al., 2003; Mattson et al., 2003).

Intermittent fasting had no effect on the growth trajectory assessed by body weight of normal females or GHR-KO males, but reduced growth of normal males and female GHR-KO mice (Fig. 1A,B). This contrasts with no significant attenuation of growth trajectory in C57Bl/6 mice on IF (Goodrick et al., 1990; (Anson et al., 2003).

Figure 1.

 Gender-specific effects of intermittent fasting (IF) on body weight (BW) and insulin sensitivity. GHR-KO mice and their littermate controls [derived from 129/Ola founders, provided by J. J. Kopchick (Ohio University, Athens, OH), and outbred to Balb/c, C57Bl/6J and C3HJ strains] were bred in a closed colony, housed under standard conditions (12-h light/12-h dark cycling and 20–23 °C), and fed Lab Diet Formula 5001 (23% protein, 4.5% fat, 6% fiber) (Nestle Purina, St Louis, MO, USA). All animals were fed AL for the first 8–10 weeks of age. Thereafter, the mice were either fed AL every day (AL group) or every other day (IF group). Mice were weighed in the morning after a feeding day, approximately 16–20 h after the IF group had been fed. Animal protocols were approved by the Animal Care and Use Committee of Southern Illinois University. (A) Body weight trajectory for male mice shows IF reduces BW for only littermate control mice. (B) Body weight trajectory for female mice shows IF reduces BW solely for GHR-KO mice. Unpaired Student’s t-test (Microsoft Excel, Redmond, WA, USA) was employed to compare food consumption data. Body weight-gain data was contrasted with repeated measures anova (SPSS, Chicago, IL, USA). Cohorts of mice not allocated for the lifespan assay, but similarly subjected to IF for 10 months, were tested for insulin tolerance (ITT). After AL access to food overnight, food was removed, 0.75 international units (IU)/kg body weight of porcine insulin (Sigma-Aldrich, St Louis, MO, USA) was injected intra-peritoneally (IP) and tail-blood glucose was determined with a glucometer (Lifescan; Johnson & Johnson, New Brunswick, NJ, USA) at 0, 15, 30 and 60 min. (C) Insulin tolerance test for male mice depicts amelioration of insulin resistance of the littermate control mouse by IF. (D) ITT for females shows IF has no effect on insulin sensitivity for either phenotype. Repeated measures anova was used for comparisons of data from the ITT (SPSS). Graphs were generated with Excel (Microsoft, Redmond, WA, USA). All measures of central tendency are arithmetic means, and all depictions of variation (error bars) represent standard deviations.

Insulin sensitivity, and the glucose homeostasis that it connotes, is a common correlate of longevity (Russell & Kahn, 2007; Bartke, 2008; Masternak et al., 2009). Results of insulin tolerance tests (ITT) revealed that IF ameliorated the innate insulin resistance of male littermate control mice (Fig. 1C), but had no effect on insulin sensitivity of normal females (Fig. 1D) or GHR-KO mice of either gender [repeated measures analysis of variance (anova) P-value for male normal mice on AL (N-AL) vs. male GHR-KO on AL (KO-AL) < 0.001 (Bonferroni’s post-hoc test corrected); P-value for male N-AL vs. male normal mice on IF (N-IF) < 0.05 (Student–Newman–Keuls’post-hoc test corrected); P-value for female N-AL vs. female KO-AL < 0.05 (Student–Newman–Keuls’post-hoc test corrected; data log10 transformed)]. Female mice for which blood glucose levels increased after insulin injection, presumably indicative of an endocrine response to the stress of handling and injection, were excluded from the analysis; with their inclusion, the differences in insulin sensitivity between the female KO-AL and the female N-AL was not statistically significant.

Intermittent fasting increased the lifespan of normal male mice but did not affect the longevity of GHR-KO males [median survival: male N-AL = 851 days, male N-IF = 1010 days, male KO-AL = 1178 days and male GHR-KO on IF (KO-IF) = 1157 days; log-rank (Mantel-Cox) test P-value for male N-AL vs. male N-IF = 0.0002] (Fig. 2A). The results from the median lifespan were previously published by Bonkowski et al. (2009), and here we include the completed lifespan curves.

Figure 2.

 Sexual dimorphism of effect of intermittent fasting (IF) on survivorship. For the survivorship assay, all animals were fed AL for the first 8–10 weeks of age. Henceforth, the mice were either fed AL every day (AL group) or every other day (IF group). Mice found either moribund or with a neoplasm approximately 10% of their total body volume were euthanized, and the date of euthanasia was recorded as the date of death. (A) IF confers longevity to male littermates. (B) For females, IF has no effect on either phenotype. Log-rank (Mantel-Cox) test analysis was utilized to compare the overall survivorship data with graphpad prism 5.01 (GraphPad Software Inc., La Jolla, CA, USA); and maximal lifespan, of the 90th percentile of the appropriate control population, was analyzed with quantile regression exploiting an exact unconditional (Score Statistic) test (http://www.stat.ncsu.edu/exact/); for the latter, a binomial model was used, the hypothesis was two-sided, 99.9000% confidence interval assigned, and the test statistic employed was Fisher’s exact-Boschloo test statistic. Graphs were generated with Excel (Microsoft). All measures of central tendency are arithmetic means. Mice physically lost from the study, euthanized due to senescence-independent matters, or included in a supplemental cohort in the survivorship assay account for the discrepancies in the sample sizes (within subgroup) between the body weight data and the survivorship data.

When maximal lifespans (survivorship when only the longest-lived decile of the appropriate control population remained) were contrasted, they concurred with the mean lifespan results, with only male littermate control mice on IF exhibiting increased longevity relative to their AL counterparts (exact unconditional homogeneity/independence test P-value for male N-AL vs. male N-IF = 0.0151; P-value for male N-AL vs. male KO-AL = 0.0028; female N-AL vs. female KO-AL = 0.0030) (Berger, 1994, 1996; Wang et al., 2004; Arum & Johnson, 2007; Miller et al., 2007).

Intermittent fasting did not increase mean or maximal lifespan of female normal or GHR-KO mice [median survival: female N-AL = 990 days, female N-IF = 1049 days, female KO-AL = 1316 days, and female KO-IF = 1325 days; log-rank (Mantel-Cox) test P-value for female N-AL vs. female KO-AL = 0.0056] (Fig. 2B).

The key novel finding of the present study is that IF fails to affect insulin sensitivity and average as well as maximal longevity of male GHR-KO mice, even though it significantly increases all of these parameters in normal males from the same strain. These observations indicate that the previously reported inability of 30% reduction of caloric intake to cause further increases in longevity or insulin sensitivity in these long-lived mutants (Bonkowski et al., 2006) was not limited to that particular regimen of DR, and that this presumably applies to DR in general. As it would address whether it was simply the lengthened fasting schedule or that and a minor reduction in caloric intake that was responsible for the IF effects observed in this report, a further IF study involving mice fed daily with a diet isocaloric to that of IF mice is warranted.

We have also found striking sexual dimorphism in the response of normal animals from this strain to IF, with significant improvements in insulin sensitivity and longevity in males only. Gender-based differential responses to caloric restriction have been reported (Willott et al., 1995; Turturro et al., 1999; Forster et al., 2003; Porter et al., 2004; Bonkowski et al., 2006; Valle et al., 2007; Selman et al., 2008). Although C57Bl/6J male mice on IF exhibit hyperphagia, and thus gain weight at a rate very similar to that of AL counterparts (Anson et al., 2003), IF has been reported to decrease body weight gain substantially in female C57Bl/6J mice (Barrows & Kokkonen, 1978); moreover, the incidence of mammary tumors in female rats is also decreased by IF (Carlson & Hoelzel, 1946). Yet, to the best of our knowledge, this is the first documentation of a sexually dimorphic effect of IF on insulin sensitivity or survivorship.

Association of the effects of IF on insulin sensitivity and longevity across the genotype × gender groups supports the suggested causative link between the two: male littermates have improvements in both upon IF treatment, but the other three subgroups do not have any effect on either. Of further note in this regard are the concordant results on female littermates in both ITT and survivorship assays in the two related studies: in Bonkowski et al. (2006), 30% CR-induced insulin sensitivity in normal females concurred with enhanced survivorship; in this report, IF failed to affect both insulin sensitivity and survivorship in females. Together with the results from this report of IF allaying insulin resistance and increasing survival in male littermates, these results further buttress the hypothesis that insulin sensitivity may be a mechanism and/or a marker of the lifespan extending potential of an intervention.

The manner in which increased insulin sensitivity might increase lifespan may involve multiple mechanisms including reduction of postprandial blood glucose content, decreased potential for hyperglycemic cytotoxicity, whether via Maillard reactions (Thorpe & Baynes, 1996; Cerami & Ulrich, 2001) or other means (Cerami, 1985; Ceriello, 2001), increased rate of ATP production that would provide the cellular currency to resist the postprandial oxidative surge (Charpentier et al., 2006; Jenkins et al., 2006) and to increase the levels of damage-preventing and/or repairing constituents, and/or altered stress-responsive kinase Erk signaling. Another potential mechanism is enhanced protein turnover, particularly through autophagy, induced by conditions of hypoinsulinemia, such as CR or decreased GH signaling (Cuervo, 2008). Reduced basal insulin concentrations themselves, independent of insulin sensitivity, blunt the insulin-mediated shift in metabolism (towards anabolism), increase autophagy, and lead to (i) the preservation of cellular energetics (for efficient production of ATP to be used for maintenance and repair processes), (ii) the degradation of perniciously effete proteins and subcellular organelles (most notably mitochondria), (iii) the enhancement of innate and acquired immunity and (iv) protection from malignant neoplasia (Mizushima et al., 2008); these actions may engender longevity by counterbalancing the age-associated decline in autophagy (Cuervo et al., 2005).


We thank Dr Rafael de Cabo, Dr Michal M. Masternak, Jacob A. Panici, Adam Spong, and Reyhan Westbrook for scientific assistance; and Steve Sandstrom and Pam Barnett for copyediting. This work was supported by National Institute on Aging Grants AG19899, U19 AG023122, and 3R01AG019899-07S1, as well as a Senior Scholar Award in Aging from The Ellison Medical Foundation and The Glenn Foundation for Medical Research.