Targeted disruption of growth hormone
receptor interferes with the beneficial
actions of calorie restriction
Michael S. Bonkowski*†, Juliana S. Rocha*, Michal M. Masternak*, Khalid A. Al Regaiey*‡, and Andrzej Bartke*‡§
Departments of *Internal Medicine–Geriatrics Research,†Pharmacology, and‡Physiology, Southern Illinois University School of Medicine,
Springfield, IL 62794
Edited by Cynthia J. Kenyon, University of California, San Francisco, CA, and approved April 6, 2006 (received for review January 6, 2006)
Reduced intake of nutrients [calorie restriction (CR)] extends lon-
gevity in organisms ranging from yeast to mammals. Mutations
affecting somatotropic, insulin, or homologous signaling path-
ways can increase life span in worms, flies, and mice, and there is
considerable evidence that reduced secretion of insulin-like
growth factor I and insulin are among the mechanisms that
mediate the effects of CR on aging and longevity in mammals. In
the present study, mice with targeted disruption of the growth
(GHRKO) mice] and their normal siblings were fed ad libitum (AL)
or subjected to 30% CR starting at 2 months of age. In normal
females and males, CR produced the expected increases in overall,
average, median, and maximal life span. Longevity of normal mice
subjected to CR resembles that of GHRKO animals fed AL. In sharp
contrast to its effects in normal mice, CR failed to increase overall,
life span only in females. In a separate group of animals, CR for 1
year improved insulin sensitivity in normal mice but failed to
further enhance the remarkable insulin sensitivity in GHRKO mu-
tants. These data imply that somatotropic signaling is critically
important not only in the control of aging and longevity under
conditions of unlimited food supply but also in mediating the
effects of CR on life span. The present findings also support the
notion that enhanced sensitivity to insulin plays a prominent role
in the actions of CR and GH resistance on longevity.
insulin-like growth factor I ? insulin ? longevity ? aging ? dietary restriction
Genes related to homologous signaling pathways in the yeast
Saccharomyces cerevisiae, the worm Caenorhabditis elegans,
and the fly Drosophila melanogaster play a key role in the
control of aging in these species (1–3). A moderate reduction
in the intake of nutrients [also known as calorie restriction
(CR)] is extremely effective in delaying aging and increasing
longevity in organisms ranging from yeast to mammals (3–5).
We have previously reported that CR produces an additional
increase in the life span of a long-lived hypopituitary mutant
mouse, the Ames dwarf, and alters the slope of its survival
curve similarly to the effects of CR in normal mice (6). This
result was counterintuitive because both Ames dwarfs and
normal animals subjected to CR have reduced insulin-like
growth factor I (IGF-I) and insulin levels and share other
phenotypic characteristics. Although C. elegans with a muta-
tion in the insulin?IGF-I homologous signaling pathway Daf
16?FOXO lived longer when subjected to CR (7, 8), CR failed
to further increase longevity in D. melanogaster with a chico
mutation that interferes with insulin?IGF-I signaling and
prolongs life (9). Interpretation of the findings obtained in
Ames dwarf mice is complicated by the fact that in addition to
growth hormone (GH) deficiency, these animals are deficient
also in prolactin and thyrotropin (10, 11), and the interaction
of CR with reduced lactogenic and?or thyroid hormone
utations affecting somatotropic and?or insulin signaling
can produce a marked increase of longevity in mice.
signaling could conceivably contribute to the results. It was
therefore of interest to examine the effects of CR in a different
long-lived mouse mutant in which the primary defect in
endocrine function is limited to the somatotropic axis.
GH receptor?GH-binding protein knockout (GHRKO) mice
were developed by Zhou et al. (12) by targeted disruption of the
Ghr?Ghrbp gene. These mutant mice do not express the GH
receptor, are GH resistant, and have profoundly suppressed
circulating levels of IGF-I and insulin, markedly increased life
span, and multiple indices of delayed aging, including increased
mortality rate doubling time (12–16).
Conflict of interest statement: No conflicts declared.
This paper was submitted directly (Track II) to the PNAS office.
Freely available online through the PNAS open access option.
Abbreviations: AL, ad libitum; CR, calorie restriction; GH, growth hormone; GHRKO, GH
receptor?GH-binding protein knockout; IGF-I, insulin-like growth factor I; ITT, insulin
§To whom correspondence should be addressed at: Geriatrics Research, Southern Illinois
University, 801 N. Rutledge, P.O. Box 19628, Springfield, IL 62794-9628. E-mail:
© 2006 by The National Academy of Sciences of the USA
and GHRKO mice that were fed AL or subjected to 30% CR. Animals were
weighed weekly starting at 2 months of age.
Time course of changes in body weight for male and female normal
May 16, 2006 ?
vol. 103 ?
no. 20 ?
CR was started at 2 months of age and continued throughout the
remainder of the study, producing the expected reduction in
body weight in both normal and GHRKO mice (Fig. 1). In males,
the relative decrease in body weight in response to CR was
comparable in normal and GHRKO mice, whereas in females,
the response to CR was more pronounced in normal animals
than in GHRKO animals.
As expected, under conditions of ad libitum (AL) feeding,
GHRKO mice lived much longer than the normal (control)
animals. Analysis of longevity data collapsed across sexes re-
vealed that, in normal mice, CR produced the expected signif-
icant increase in overall survival (log-rank test) and average,
median, and maximal longevity (Fig. 2 and Table 1). In contrast,
CR did not affect the overall survival or average or median
longevity of GHRKO mice and produced a smaller (7.8% in
GHRKO mice vs. 17.2% in normal mice) although statistically
significant increase in the maximal longevity of GHRKO ani-
mals. It was particularly striking that median longevity was
mice, and it was nearly identical (1,188 vs. 1,181 days) in
GHRKO-AL and GHRKO-CR mice.
When male and female longevity data were analyzed sepa-
rately, it became evident that the effects of CR on maximal
longevity were sexually dimorphic in GHRKO mice but not in
normal mice. In normal males (Fig. 3), CR increased overall and
maximal survival, and median longevity was increased by 18.8%.
In GHRKO males, CR did not affect overall, average, or
maximal longevity, whereas median longevity was increased by
1.2%. In normal females (Fig. 4), CR increased each of the
examined parameters, including a 28% increase in median
longevity. In GHRKO females, CR did not affect overall,
average, or median longevity (median longevity was 1% shorter
in GHRKO-CR mice than in GHRKO-AL mice) but produced
a significant increase in maximal longevity. The increase in
maximal longevity was proportionally smaller than that mea-
remainder of life span. Data from males and females are combined.
Table 1. Longevity characteristics in normal and GHRKO mice that were fed AL or subjected
to 30% CR
Measures of longevity
Males and females
Log-rank test of
Log-rank test of
Log-rank test of
887 ? 29*
1,163 ? 38*
1,028 ? 45†
1,363 ? 20†
1,139 ? 30‡
1,362 ? 11†
1,167 ? 31‡
1,468 ? 15‡
862 ? 39*
1,113 ? 16*
963 ? 64*
1,271 ? 11†
1,145 ? 38†
1,334 ? 20†
1,171 ? 27†
1,320 ? 17†
916 ? 43*
1,158 ? 59*
1,090 ? 63†
1,382 ? 10†
1,133 ? 46†
1,359 ? 14†
1,154 ? 63†
1,481 ? 10‡
significantly different (P ? 0.05).
were fed AL or subjected to 30% CR starting at 2 months of age. CR was
maintained throughout the life span.
Kaplan–Meier survival plot of male normal and GHRKO mice that
www.pnas.org?cgi?doi?10.1073?pnas.0600161103 Bonkowski et al.
sured in normal females (9% in GHRKO females vs. 19% in
In a separate group of normal and GHRKO mice, 30% CR for
1 year produced the expected decrease in body weight that was
statistically significant in each sex?genotype group (data not
before and 15, 30, and 60 min after insulin injection. In normal
male and female mice, CR significantly increased sensitivity to
insulin challenge. Insulin sensitivity in GHRKO animals was
greater than in normal mice and was not further increased by CR
in either sex (Fig. 5).
The main finding of the present study is the striking difference
in the response of GHRKO and normal mice to an identical
regimen of CR. In normal mice, 30% CR produced the expected
robust increase in longevity. GHRKO mice that were fed AL
lived much longer than normal mice and thus resembled normal
CR animals. Subjecting GHRKO mice to CR failed to produce
any further extension of overall, median, or average longevity.
Maximal longevity of GHRKO mice, estimated by the mean age
of death of the oldest 10% in each group, was extended by CR
very modestly and only in females. Although the magnitude
of the effect of CR on life span is strain-dependent (17), failure
to observe life extension in CR mice is extremely rare if not
unprecedented (18). In particular, the present findings in
GHRKO mice contrast sharply with significant extensions of
longevity in normal (control) animals from the same line (Table
1 and Fig. 2). These findings also differ from the results obtained
in long-lived Ames dwarfs, which share many phenotypic char-
acteristics with GHRKO mice and were exposed to an identical
CR protocol in an earlier study in our laboratory (6). Both
GHRKO and Ames dwarf mice are characterized by extremely
low levels of IGF-I in peripheral circulation, reduced postnatal
growth, delayed puberty, diminished body size (19), reduced
plasma insulin and glucose, and enhanced sensitivity to injected
insulin (20, 21). Because normal siblings of GHRKO mice
exhibited a robust longevity response to CR in the present study,
it appears exceedingly unlikely that differences in the responses
of Ames dwarf and GHRKO mice to CR are due to distinct
genetic backgrounds or differences in some hard-to-define en-
vironmental factors between studies conducted several years
Phenotypic differences between GHRKO and Ames dwarf
mice include opposite alterations in prolactin levels; varying
degrees of suppression of thyroid hormone levels and body core
temperature; major differences in reproductive status, particu-
larly in females; and differences in body composition, antioxi-
dant enzymes, and end-of-life pathology (19, 22). From the
present data, we cannot conclude which of these differences may
have contributed to their divergent responses to 30% CR.
Absence of additional extension of life by CR in GHRKO
animals in the present study could have been due to the
particular regimen of CR chosen or to the ‘‘ceiling effect’’ (i.e.,
maximal life extension) in GHRKO animals that were fed AL.
Clancy et al. (9) elegantly demonstrated that, in Drosophila,
different severities of CR are optimally effective for life exten-
sion in wild-type and in long-lived mutant flies. Although we do
not know whether a different (milder) regimen of CR may have
allowed GHRKO mice to respond by life extension, the 30% CR
restriction used in the present study is not considered severe.
were fed AL or subjected to 30% CR starting at 2 months of age. CR was
maintained throughout the life span.
Kaplan–Meier survival plot of female normal and GHRKO mice that
time points.*, P ? 0.05 compared with AL controls within phenotype.
Results of ITT in normal and GHRKO mice that were fed AL or subjected to 30% CR between 2 and 12 months of age. All groups were randomly fed
Bonkowski et al.
May 16, 2006 ?
vol. 103 ?
no. 20 ?
Moreover, an identical regimen of CR was previously shown to
extend life span in another long-lived mutant that resembled
GHRKO mice in body size, insulin sensitivity, and other char-
acteristics (ref. 6; detailed above).
It is exceedingly unlikely that GHRKO mutation produces the
maximal possible life extension and thus that CR in animals of
this genotype causes near starvation rather than a beneficial
reduction in nutrient intake. Autopsy findings and studies of
body composition indicate that GHRKO-CR mice are not
depleted of adipose tissue. Extrinsic factors including starvation
unmistakably alter the shape of a survival curve (17, 23), and
there was no indication of such alterations in GHRKO-AL and
GHRKO-CR mice. In fact, 90% of the GHRKO-CR mice
remained alive at 925 days of age (Fig. 2).
An important implication of the present findings is that
targeted disruption of the GH receptor and the resulting GH
resistance mimic some of the effects of CR in mice, although
numerous differences also exist (24, 25). Moreover, these
findings imply that the GH receptor or GH receptor-
dependent GH signaling is extremely important and may
possibly be required for CR to extend mammalian longevity.
Thus, CR-induced suppression of the somatotropic axis
emerges as one key (or perhaps the key) candidate for
mechanisms linking CR with increased longevity.
It is also interesting to view the present results in a broader
context of interactions between ‘‘longevity genes’’ and CR.
Studies in C. elegans suggest that some of the mutations that
extend life act through mechanisms similar and perhaps identical
to those that are responsible for the effects of CR (7) but that
genes related to the insulin-like pathway do not (7, 8). Results
obtained from Drosophila suggest considerable similarity be-
tween the actions of CR and genes related to the insulin-like
signaling pathway, although outcomes of specific studies may
depend on the severity of CR imposed (9). In the mouse,
previous studies in Ames dwarf mice suggested that this muta-
tion (Prop1df) and CR affect longevity by distinct but likely
overlapping mechanisms (6). In support of this conclusion, the
effects of Prop1dfand CR on life span are additive, whereas their
effects on the slope of the survival plot are different (6). In
contrast, close similarity of survival of normal-CR, GHRKO-
AL, and GHRKO-CR mice in the present study would suggest
that targeted disruption of the Ghr?Ghrbp gene and CR affect
longevity by the same mechanism. This conclusion was not
expected from our earlier studies that identified major differ-
ences in the effects of CR and GHRKO mutation on a wide
profile of hepatic gene expression (24) and on the expression of
selected insulin- and IGF-I-related genes in the liver (25),
skeletal muscle (26), and heart (27). However, in support of the
present findings, analysis of liver gene expression revealed a
significant interaction between CR diet and GHRKO mutation,
whereas significantly fewer genes were altered by the CR regimen
in GHRKO mice compared with normal control littermates (24).
On the basis of the present and previous results, we suggest
that the failure of CR to increase overall, median, or average
longevity in GHRKO mice is related to its failure to improve
insulin sensitivity in these mutant animals. There are numerous
indications that insulin and insulin release and actions play a
major role in the control of mammalian aging. We and others
have previously suggested that increased sensitivity to insulin
(measured directly or implied by concomitant reductions in
plasma insulin and glucose levels) accounts for, or importantly
contributes to, extended longevity of Ames dwarf (21), Snell
dwarf (28), and GHRKO (20) animals as well as mice exposed
to CR (29). In an attempt to identify reasons for the failure of
CR to extend longevity of GHRKO mice, we examined the
responses to injected insulin [the so-called insulin tolerance test
(ITT)] in a separate group of GHRKO and normal mice
subjected to an identical CR regimen for 12 months (Fig. 5). CR
produced the expected improvement in insulin sensitivity of
normal mice. Insulin sensitivity of GHRKO mutants was greatly
increased in comparison with normal mice, as expected from
previous studies (21), and was not further enhanced by CR. We
have previously reported that in GHRKO mice, but not in
normal mice, CR enhances the expression of hepatic genes
related to gluconeogenesis and fails to further reduce hepatic
levels of phosphorylated Akt, an important mediator of insulin
action (25). More recently, we reported that liver expression of
insulin receptor and insulin receptor substrate 2 were elevated in
normal-CR, GHRKO-AL, and GHRKO-CR mice when com-
pared with normal-AL control animals (26).
A recent report of increased life span in transgenic Klotho
mice that are insulin-resistant (30) raises an interesting possi-
signal, regardless of its underlying causes (31). Thus, effects of
mild insulin resistance in Klotho transgenics and adipose-
knockout) mice (32) may overlap the effects of reduced insulin
and IGF-I release in GHRKO, dwarf, and CR animals.
We conclude that, in the absence of a functional GH receptor,
CR does not affect most of the examined measures of longevity
in mice. The failure of CR to increase average life span of
GHRKO mice or maximal life span of GHRKO males was
associated with its failure to further increase insulin sensitivity in
these animals. We suspect that the divergent effects of CR on
gene expression and levels of signaling molecules in GH and
insulin target organs may underlie differences between normal
and GHRKO mice in deriving a longevity benefit from CR.
A broader implication of the present findings is that insulin
signaling emerges as an important determinant of mammalian
aging and longevity. This implication raises the issue of the
potential (and, we believe, likely) impact of insulin-resistant
states, including the current ‘‘epidemic’’ of metabolic syndrome
on life expectancy in this society and others. We suspect that
research efforts to develop ‘‘CR mimetics’’ for pharmacological
intervention in the aging process may be more productive if they
focus on targets in the GH?IGF-I?insulin signaling axis.
Materials and Methods
Animals. GHRKO and normal mice were produced in our
breeding colony derived from GHRKO animals kindly provided
by J. J. Kopchick (Ohio University, Athens). Phenotypically
normal siblings of GHRKO mice served as controls for this
study. Animals were housed under temperature- and light-
controlled conditions (20–23°C and 12-h light?12-h dark cycle)
and were fed Lab Diet Formula 5001 (Ralston Purina). Sentinel
animals were sent for bacterial and viral testing every 3 months,
and the results were uniformly negative. All animal protocols for
this study were approved by the Animal Care and Use Com-
mittee of Southern Illinois University.
Longevity Study. All animals (n ? 37–43 per genotype per diet
group) were studied beginning at 8 weeks of age. In both
longevity and long-term (1 year) studies, animals were gradually
placed on 30% CR by receiving 90% of the amount of food
consumed by AL controls in the initial week, 80% the following
week, and 70% throughout the rest of the studies. Animals in the
longevity study were checked daily for health and survival and
were handled for cage changes and weekly body weight mea-
surements only. Animals that appeared to be near death (listless,
unable to walk, and cold to the touch) or had large bleeding
tumors or neoplastic growth approaching 10% of body weight
were euthanized, and the date of euthanasia was considered the
date of death. As of this writing, 147 animals had died, 13 had
been euthanized, and 1 was still alive.
www.pnas.org?cgi?doi?10.1073?pnas.0600161103Bonkowski et al.
ITT. In a separate group of mice, CR was imposed as described Download full-text
above. After 12 months of CR, ITTs were conducted. All groups
(n ? 8–10 per genotype per diet per sex) were randomly fed
(100% AL) the night before the test. The following morning,
food was removed, and basal glucose was determined by using a
glucometer (Lifescan; Johnson & Johnson, New Brunswick, NJ)
in blood obtained by removing the tip of the tail. Porcine insulin
(Sigma) was injected i.p. at 0.75 units?kg of body weight. Blood
was subsequently sampled at 15, 30, and 60 min thereafter for
Statistical Analysis. Kaplan–Meier survival curves were used for
survival analysis, with a log-rank test to evaluate significance of
differences between groups. Median life span represents the age
at which 50% of the population within groups remained alive.
Maximal life span was calculated as the mean age of the oldest
10% of the population within that group, and significance was
tested by using two-way ANOVA and independent t tests.
Results of ITTs were plotted as mean percentage change from
baseline within experimental groups. Repeated measures
ANOVA was used to determine the interaction of the main
effect variables, phenotype and diet. Independent t tests were
used to detect significant differences at specified time points. All
graphs were developed by using PRISM 4 (GraphPad, San Diego).
All statistical analyses were conducted by using SPSS 12 (SPSS,
Chicago). ? was set at 0.05 for determination of significance.
With the exception of log-rank longevity data, all values are
reported as mean ? SEM throughout the figures and text.
We thank Jacob Panici and Marty Wilson for laboratory assistance and
Steve Sandstrom for editorial assistance. This work was supported by
National Institute on Aging Grants AG19899 and U19 AG023122, the
Ellison Medical Foundation, and the Southern Illinois University Ge-
riatrics Medicine and Research Initiative.
1. Longo, V. D. & Finch, C. E. (2003) Science 299, 1342–1346.
2. Tatar, M., Bartke, A. & Antebi, A. (2003) Science 299, 1346–1351.
3. Kenyon, C. (2005) Cell 120, 449–460.
4. Weindruch, R. & Sohal, R. S. (1997) N. Engl. J. Med. 337, 986–994.
5. Guarente, L. & Picard, F. (2005) Cell 120, 473–482.
6. Bartke, A., Wright, J. C., Mattison, J. A., Ingram, D. K., Miller, R. A. & Roth,
G. S. (2001) Nature 414, 412.
7. Lakowski, B. & Hekimi, S. (1998) Proc. Natl. Acad. Sci. USA 95, 13091–13096.
8. Houthoofd, K., Braeckman, B. P., Johnson, T. E. & Vanfleteren, J. R. (2003)
Exp. Gerontol. 38, 947–954.
9. Clancy, D. J., Gems, D., Hafen, E., Leevers, S. J. & Partridge, L. (2002) Science
10. Bartke, A. (1979) in Genetic Variation in Hormone Systems, ed. Shire, J. G. M.
(CRC, Boca Raton, FL), pp. 113–126.
11. Sornson, M. W., Wu, W., Dasen, J. S., Flynn, S. E., Norman, D. J., O’Connell,
S. M., Gukovsky, I., Carriere, C., Ryan, A. K., Miller, A. P., et al. (1996) Nature
12. Zhou, Y., Xu, B. C., Maheshwari, H. G., He, L., Reed, M., Lozykowski, M.,
Okada, S., Cataldo, L., Coschigamo, K., Wagner, T. E., et al. (1997) Proc. Natl.
Acad. Sci. USA 94, 13215–13220.
13. Coschigano, K. T., Clemmons, D., Bellush, L. L. & Kopchick, J. J. (2000)
Endocrinology 141, 2608–2613.
14. Bartke, A., Chandrashekar, V., Bailey, B., Zaczek, D. & Turyn, D. (2002)
Neuropeptides 36, 201–208.
15. Coschigano, K. T., Holland, A. N., Riders, M. E., List, E. O., Flyvbjerg, A. &
Kopchick, J. J. (2003) Endocrinology 144, 3799–3810.
16. de Magalhaes, J. P., Cabral, J. A. S. & Magalhaes, D. (2005) Genetics 169,
17. Weindruch, R. & Walford, R. L. (1988) in The Retardation of Aging and Disease
by Dietary Restriction (Thomas, Springfield, IL), pp. 31–71.
18. Forster, M. J., Morris, P. & Sohal, R. S. (2003) FASEB J. 17, 690–692.
19. Bartke, A., Coshigano, K., Kopchick, J., Chandrashekar, V., Mattison, J.,
20. Liu, J. L., Coschigano, K. T., Robertson, K., Lipsett, M., Guo, Y., Kopchick,
J. J., Kumar, U. & Liu, Y. L. (2004) Am. J. Physiol. Endocrinol. Metab. 287,
21. Dominici, F. P., Hauck, S., Argentino, D. P., Bartke, A. & Turyn, D. (2002) J.
Endocrinol. 173, 81–94.
22. Bartke, A. (2005) Endocrinology 146, 3718–3723.
23. Austad, S. N. (2001) in Handbook of the Biology of Aging, eds. Masoro, E. J. &
Austad, S. N. (Academic, San Diego), pp. 3–22.
24. Miller, R. A., Chang, Y., Galecki, A. T., Al Regaiey, K., Kopchick, J. J. &
Bartke, A. (2002) Mol. Endocrinol. 16, 2657–2666.
25. Al Regaiey, K. A., Masternak, M. M., Bonkowski, M., Sun, L. & Bartke, A.
(2005) Endocrinology 146, 851–860.
26. Masternak, M. M., Al Regaiey, K. A., Rosario Lim, M. M., Jimenez-Ortega,
V., Panici, J. A., Bonkowski, M. S. & Bartke, A. (2005) Exp. Gerontol. 40,
27. Masternak, M. M., Al Regaiey, K. A., Del Rosario Lim, M. M., Jimenez-
Ortega, V., Panici, J. A., Bonkowski, M. S., Kopchick, J. J., Wang, Z. & Bartke,
A. (March 6, 2006) Exp. Gerontol., 10.1016?j.exger. 2006.01.009.
28. Hsieh, C. C., DeFord, J. H., Flurkey, K., Harrison, D. E. & Papaconstantinou,
J. (2002) Mech. Ageing Dev. 123, 1245–1255.
29. Masoro, E. J. (2005) Mech. Ageing Dev. 126, 913–922.
30. Kurosu, H., Yamamoto, M., Clark, J. D., Pastor, J. V., Nandi, A., Gurnani, P.,
McGuinness, O. P., Chikuda, H., Yamaguchi, M., Kawaguchi, H., et al. (2005)
Science 309, 1829–1833.
31. Bartke, A. (2006) Trends Endocrinol. Metab. 17, 33–35.
32. Bluher, M., Kahn, B. B. & Kahn, C. R. (2003) Science 299, 572–574.
Bonkowski et al.
May 16, 2006 ?
vol. 103 ?
no. 20 ?