cron-web.org

[A1CR]

The longevity of CRers is of potential interest. The below
is an overview
of how CR may be effective in prolonging life, and the
involvement of
hormesis receives some attention, apparently.

Masoro EJ.
Overview of caloric restriction and ageing.
Mech Ageing Dev. 2005 May 7; [Epub ahead of print]
PMID: 15885745

It has been known for some 70 years that restricting the
food intake of
laboratory rats extends their mean and maximum life span. In
addition, such
life extension has been observed over the years in many
other species,
including mice, hamsters, dogs, fish, invertebrate animals,
and yeast. Since
this life-extending action appears to be due to a restricted
intake of
energy, this dietary manipulation is referred to as caloric
restriction
(CR). CR extends life by slowing and/or delaying the ageing
processes. The
underlying biological mechanism responsible for the life
extension is still
not known, although many hypotheses have been proposed. The
Growth
Retardation Hypothesis, the first proposed, has been tested
and found
wanting. Although there is strong evidence against the
Reduction of Body Fat
Hypothesis, efforts have recently been made to resurrect it.
While the
Reduction of Metabolic Rate Hypothesis is not supported by
experimental
findings, it nevertheless still has advocates. Currently,
the most popular
concept is the Oxidative Damage Attenuation Hypothesis; the
results of
several studies provide support for this hypothesis, while
those of other
studies do not. The Altered Glucose-Insulin System
Hypothesis and the
Alteration of the Growth Hormone-IGF-1 Axis Hypothesis have
been gaining
favor, and data have emerged that link these two hypotheses
as one. Thus, it
may now be more appropriate to refer to them as the
Attenuation of
Insulin-Like Signaling Hypothesis. Finally, the Hormesis
Hypothesis may
provide an overarching concept that embraces several of the
other hypotheses
as merely specific examples of hormetic processes. For
example, the
Oxidative Damage Attenuation Hypothesis probably addresses
only one of
likely many damaging processes that underlie aging. It is
proposed that
low-intensity stressors, such as CR, activate ancient
hormetic defense
mechanisms in organisms ranging from yeast to mammals,
defending them
against a variety of adversities and, when long-term,
retarding senescent
processes.

1. Introduction
Early in the 20th century, Osborne et al. (1917) published a
paper in
Science, which showed that decreasing the food intake of
rats slowed growth
and increased length of life. This paper had little impact
because of the
poor quality of the survival component of the study, and
because Robertson
and Ray (1920) reported in the Journal of Biological
Chemistry that growth
rate and length of life are positively associated in mice.
However, in the
1930s, McCay et al. (1935) carried out well-executed studies
clearly showing
that markedly restricting food intake of rats, at or soon
after weaning,
resulted in life extension. Since then, restricting food
intake has been
observed to increase both the mean and maximum life span in
a spectrum of
rat and mouse strains and in many other species, including
yeast,
invertebrate animals, fish, hamsters, and dogs (Masoro, 2002).

McCay et al. (1939) also reported work suggesting that the
dietary factor
responsible for the increase in longevity is probably a
decreased intake of
energy. More recently, our studies, in which rats were fed
semi-synthetic
diets (Iwasaki et al., 1988 and Masoro et al., 1989),
strongly supported the
conclusion that the life extension is due to a reduced
caloric intake rather
than reduction in a specific nutrient.

Nevertheless, this view has been challenged because of
findings that show
that the restriction of a specific dietary component without
a decrease in
caloric intake can result in life extension. However,
research in our
laboratory on the male F344 rat indicates that such findings
do not
necessarily invalidate the conclusion that life extension in
response to a
reduction in food intake is primarily due to a decrease in
energy intake.
For instance, in an early study (Yu et al., 1985), we found
that when
dietary protein is reduced by 40% with no reduction in
energy intake, there
is a significant life extension in ad libitum-fed rats,
although not nearly
so marked as that resulting from a 40% reduction in food
intake. This led us
to believe that part of the life extension resulting from a
reduced food
intake was due to the decreased intake of protein. However,
in a subsequent
study (Masoro et al., 1989), we found that restricting the
caloric intake by
40% resulted in the same magnitude of life extension whether
or not protein
intake was restricted. The reason for the life extension by
the 40%
restriction of dietary protein in the earlier study probably
relates to the
fact that kidney failure in ad libitum-fed males of this rat
strain is the
major cause of death; in contrast, this disease process is
not a significant
cause of death in rats on energy-restricted diets. Since the
full
life-extension effect of food restriction occurs whether or
not protein is
restricted, it is clear that reduction in energy intake
rather than in
protein intake most likely plays the major role in the life
extension.

The recent paper by Zimmerman et al. (2003) reporting that
restriction of
dietary methionine markedly extends the life of rats has
also been heralded
as evidence that life extension in response to a reduced
food intake is due
to the restriction of specific nutrients rather than
calories. Those who
make this claim ignore the study of Masoro et al. (1989),
which clearly
shows that restriction of methionine intake is not involved
in the food
restriction-induced life extension.

Thus, although the restriction of various nutrients can
extend life, it is
most likely, based on current knowledge, that life extension
due to reduced
food intake results from the reduced intake of calories.
Indeed, the
life-prolonging manipulation of restricting food intake,
initially referred
to as food restriction (a good operational name) and then by
the vague name,
dietary restriction (DR), is now usually referred to by the
more specific
name, caloric restriction (CR). As just discussed, however,
there is not
total agreement that such a specific name is warranted.

2. CR and ageing
Does the extension of life mean that CR slows the ageing
process(es)? Before
addressing this question, we need to define what we mean by
ageing. In this
paper, ageing and senescence are used as a synonyms and are
defined as: the
deteriorative changes, during the adult period of life,
which underlie an
increasing vulnerability to challenges, thereby decreasing
the ability of
the organism to survive. Thus, the question becomes: does CR
slow senescence
processes? That CR increases the median length of life of a
population is
not an evidence that it slows the ageing processes, since
that would occur
if only premature deaths unrelated to ageing were prevented.
The fact that
CR increases the maximum life span of a population has long
been felt to
indicate that it retards senescence. However, Gavrilov and
Gavrilova (1990)
have questioned, on several grounds, the reliability of
using the maximum
length of life as an index of the rate of ageing. A more
reliable index is
the average age at death of the 10th percentile survivors,
which Lewis et
al. (1999) call the maximal length of life; they have shown
that the maximal
length of life is increased by CR in both genders of the
several mouse and
rat strains they studied.

Compared to ad libitum-fed rodents of the same age, the
physiologic
processes of old rodents maintained on a CR regimen are more
like those of
young rodents (Masoro, 2002). This has been viewed as
evidence that CR slows
senescent deterioration. However, there are two concerns
regarding this
interpretation. The first concern is that there are two
general ways in
which CR brings about a more youthful physiology at advanced
ages. One is by
slowing the age-associated change; e.g., CR does not
influence the serum
cholesterol level in young adult rats, but it markedly
attenuates the
age-associated increase in serum cholesterol level (Liepa et
al., 1980). The
other is by altering the physiologic activity in the young
adult, but not
altering the rate of age-associated change; e.g., CR
enhances hepatic
proteolytic capacity in young adult rats, but it does not
alter the
subsequent rate of the age-associated decrease in
proteolysis; thus, the
livers of old rats on a CR regimen have a proteolytic
capacity similar to
that of young ad libitum-fed rats (Ward, 1988). The first of
these two
general ways of modulating a physiologic process in animals
of advanced age
has intuitive appeal as an indicator of the slowing of the
rate of ageing.
It is not obvious that such can be said for the other
general way. The
second concern - the overriding one - is that studies aimed
to determine
biomarkers of ageing have yet to establish any age-change in
a physiologic
process as a valid biomarker of ageing. Indeed, the lack of
valid biomarkers
of ageing is a serious impediment to biogerontologic study;
however, the
findings of Miller (2001) indicate that it may be possible
for physiologic
processes to provide such biomarkers.

The influence of CR on psychologic parameters is unclear.
For example,
several studies indicate that CR retards or delays the
age-associated
decline in cognition while several other studies find CR to
be ineffective
in this regard (Masoro, 2002).

CR delays the onset and/or slows the progression of most
age-associated
diseases, including neoplastic diseases, degenerative
diseases, and immune
diseases (Maeda et al., 1985, Bronson and Lipman, 1991 and
Roe et al.,
1995). This action has also been viewed as evidence that CR
retards the
ageing processes. However, this conclusion is open to
question, too, because
of the ongoing debate on whether age-associated disease is
an integral part
of the ageing process (Hayflick, 2004 and Holliday, 2004).

Probably, the most accepted method of determining if a
manipulation has
influenced ageing is by Gompertzian and related analyses,
which provide an
index of the rate of ageing by assessing the rate of the
age-associated
increase in age-specific mortality. Using the Gompertzian
analysis, Holehan
and Merry (1986) assessed the average mortality rate
doubling time with data
from four published studies; they found it to be 102 days
for ad libitum-fed
rats and 203 days for CR-rats. Pletcher et al. (2000)
analyzed a study
carried out in our laboratory on male F344 rats (Yu et al.,
1982) and they
concluded that the increase in the longevity of the rats on
a CR regimen
results primarily from a decreased rate of increase in
age-specific
mortality. Thus, the findings on rats strongly support the
concept that CR
extends life by slowing the rate of ageing. However, at
least one study of
mice does not support this view. Weindruch et al. (1986)
found in their
study of female C3B10RF1 mice that CR delayed the start of the
age-associated increase in age-specific mortality rate, but
did not
influence the rate once the increase was underway. This
finding leads to the
interpretation that in this gender and strain of mice, CR
delays the start
of senescence, but does not slow it once underway.
Unfortunately, the number
of mouse strains analyzed in this way has been insufficient
to determine if
this finding is merely strain-specific. However, if
generally true of mice,
it would suggest that the underlying mechanism by which CR
influences aging
differs between mice and rats. In contrast, it is striking
that both among
rat strains and among mouse strains, CR causes similar
increases in the
median and maximal length of life (Lewis et al., 1999).
Further complicating
the picture is the study of Drosophila by Mair et al.
(2003), they found
that within two days of switching fully fed flies of any age
to a CR diet,
the age-specific mortality rate decreases to that of the
flies on a lifelong
CR regimen.

Indeed, there is other evidence that the nature of the
life-extending action
of CR may differ among species. For example, Lipman et al.
(1998) reported
that CR does not extend the life of F344 × BNFl rats when
initiated in late
middle age (18 months of age) or in old age (26 months of
age); they had
previously reported similar findings for the Long-Evans rat
strain (Lipman
et al., 1995). In contrast, Dhahbi et al. (2004) found that
CR increases the
life span of B6C3F1 mice when initiated at 19 months of age.

3. Biological mechanism: hypotheses
Since 1935, many mechanisms have been proposed as the
biological basis of
the life-prolonging and "anti-ageing" actions of CR.
Although none is
strongly supported by available evidence, most continue to
have advocates.
Indeed, it is entirely possible that the actions of CR
involve a combination
of suggested mechanisms. The following is a description and
evaluation of
what this author considers to be or to have been the major
hypotheses over
the years.

3.1. Retardation of Growth Hypothesis
In the 1935 paper cited above, McCay et al. proposed that CR
increases the
longevity of rats by retarding growth. This hypothesis held
sway until the
1980s with many gerontologists modifying it to include the
retardation of
development, because ageing was then viewed as a
continuation of
development. This hypothesis was challenged by studies
published in the
1980s by Roy Walford's group and our group. Weindruch and
Walford (1982)
reported that CR initiated in mice at 12 months of age
significantly extends
life, but not as markedly as CR initiated at or soon after
weaning. In our
study (Yu et al., 1985), the following four groups of male
F344 rats were
studied: rats fed ad libitum throughout life; rats in which
CR was initiated
at 6 weeks of age (2 weeks post-weaning); rats on CR from 6
weeks to 6
months of age (the rapid growth period) and fed ad libitum
thereafter; rats
fed ad libitum until 6 months of age (the age at which
skeletal development
is almost complete in this rat strain) and on CR for the
rest of life. The
longevity data from our study are summarized in Table 1.
There are two
important findings: first, when CR was limited to the rapid
growth period,
it did not markedly increase the age of the 10th percentile
survivors; and
second, when CR was initiated after the rapid growth period,
it was almost
as effective in increasing the age of the 10th percentile
survivors as CR
initiated at 6 weeks of age.

Table 1.

Age of initiation and time period of CR and longevity of
male F344 rats
=====================================
Dietary group Period on CR---Survival in days
---Median Tenth percentile
=====================================
1 None 701 822
2 From 6 weeks of age 1057 1226
3 6-26 weeks of age 808 918
4 From 26 weeks of age 941 1177
=====================================
Note: This table was generated from data in the report
of [YYu BP, Masoro
EJ, McMahan CA. Nutritional influences on aging of Fischer
344 rats: I.
Physical, metabolic, and longevity characteristics. J
Gerontol. 1985
Nov;40(6):657-70. PMID: 4056321 http://tinyurl.com/y9dm9z
unavailed paper].

For many years, these two studies were viewed as tests that
negated the
Retardation of Growth Hypothesis. Recently, however, this
hypothesis has
been resurrected because of studies such as that reported by
Miller et al.
(2002), which shows that in a genetically heterogeneous
mouse population,
body weight at 2 months of age is a significant predictor of
longevity;
indeed, body weight from 2 to 24 months of age was shown to
inversely
correlate with longevity, a correlation that becomes weaker
with increasing
age. As can be seen from Table 2, such a correlation was not
found in a
study with the genetically homogeneous inbred males of the
F344 rat strain
(Yu et al., 1982). Thus, it seems likely that the findings
of the inverse
correlation in the Miller et al. (2002) study relate to
genetics and are not
relevant in regard to CR. However, since the nature of the
life-extending
action of CR may differ between mice and rats, it is
possible that the
Retardation of Growth Hypothesis is applicable to mice, but
not to rats.

Table 2. Body weight and the length of life of ad
libitum-fed male F344
rats (n = 115)
=====================================
Age (month)---Body weight-longevity correlation
---r P
=====================================
2 ?0.14 n.s.
4 ?0.15 n.s.
=====================================
Note: This table was generated from data in the report
of [Yu BP, Masoro
EJ, Murata I, Bertrand HA, Lynd FT. Life span study of SPF
Fischer 344
male rats fed ad libitum or restricted diets: longevity,
growth, lean body
mass and disease. J Gerontol. 1982 Mar;37(2):130-41. PMID:
7056998 paper
http://tinyurl.com/y9yc2f which is not pdf-availed].

3.2. Reduction of Body Fat Hypothesis
Berg and Simms (1960) hypothesized that CR's life extension
is due to a
decrease in body fat content. They based this view on the
fact that body fat
is associated with premature death in humans and on the
assumption that CR
decreases the body fat content of rodents, which they did
not measure. It
was subsequently shown that CR does decrease body fat
content of rats and
mice (Bertrand et al., 1980, Harrison et al., 1984 and
Garthwaite et al.,
1986) and that it is particularly effective in decreasing
visceral fat
(Barzilai and Gupta, 1999). CR has similar effects on body
fat of rhesus and
cynomolgus monkeys (Hansen and Bodkin, 1993, Lane et al.,
1997, Cefalu et
al., 1997 and Colman et al., 1999).

Gerontologists did not embrace this hypothesis because in
the 1960s and
1970s, most accepted the Retardation of Growth Hypothesis as
fact. However,
many nutritionists viewed it favorably until two studies,
published in the
1980s, provided a strong case against the validity of the
Reduction of Body
Fat Hypothesis. Bertrand et al. (1980) reported no
correlation between the
body fat mass and the length of life of ad libitum-fed male
F344 rats and a
positive correlation in male rats of this strain maintained
on a CR regimen.
Harrison et al. (1984) compared obese (ob/ob) mice with lean
mice that were
congenic except for the ob/ob locus. The length of life of
the ad
libitum-fed ob/ob mice was less than that of the ad
libitum-fed lean mice,
but ob/ob mice on a CR regimen lived longer than the ad
libitum-fed lean
mice, even though they had a 48% fat content compared to 22%
for the lean
mice. Indeed, the ob/ob mice on a CR regimen lived at least
as long as CR
lean mice (13% body fat).

These findings led to a dismissal of the Reduction of Body
Fat Hypothesis
until 1999 when Barzilai and Gupta (1999) revisited it in a
theoretical
paper, which contained no supporting empirical data.
Subsequently, Blüher et
al. (2003) reported that the FIRKO mouse (in which the
insulin receptor is
"knocked out" only in the adipose tissue) exhibits life
extension and a
decreased body fat mass. Although they had not studied CR,
the authors came
to the surprising conclusion that a reduction in fat mass,
rather than a
restriction of calories, underlies the food
restriction-induced life
extension. Recently, the Guarente group has drawn a similar
conclusion based
on research carried out during the past 5 years in their and
David
Sinclair's laboratories, which indicates that sirtuin
proteins play an
important roll in the life-extending action of CR. In 2000,
the Guarente
group (Lin et al., 2004) reported that a sirtuin protein,
SIR2, is required
for CR to extend the replicative life span of a yeast
species (Saccharomyces
cerevisiae) and that this action involves the deacetylase
activity of this
protein. Subsequently, the Guarente group and the Sinclair
group reported
findings indicating that it is likely that sirtuin protein
deacetylase
activity is involved in CR-induced life extension in
invertebrate animal
species as well as in mammalian species (Picard et al., 2004
and Wood et
al., 2004). In addition, Picard et al. (2004) found that
sirtuin deacetylase
activity decreases fat deposition and increases fat
mobilization, thereby
decreasing fat mass in mammalian white adipose tissue. Based
on this finding
and without studying CR, the Guarente group concluded that a
major factor in
CR's life-extending action is the reduction of body fat
mass. Although these
findings are interesting, some empirical support for their
relevance in
CR-induced life extension is sorely needed before the
Reduction of Body Fat
Hypothesis can be reinstated. Unfortunately, the mouse study
by Miller et
al. (2002), which showed an inverse correlation of body mass
and longevity,
did not include the measurement of body fat content.

3.3. Reduction in Metabolic Rate Hypothesis
Sacher (1977) proposed that CR extends the life span by
decreasing the
metabolic rate. This hypothesis was based on the many
studies showing that
reduction of food intake in humans decreases the metabolic
rate (for a
review of these studies see Garrow, 1978) and the work of
Rubner (1908)
showing a mammalian interspecies inverse relationship
between species
longevity and its metabolic rate per kilogram body mass.
Pearl (1928)
proposed the "rate of living theory of ageing," which
extends Rubner's
concept to include "longevity differences within the same
species".

The findings of ongoing research in our laboratory, using
male rats of the
F334 strain as the animal model, were not in accord with
Sacher's
hypothesis. We found that following the initiation of CR,
there was only a
brief period of reduced food intake per gram body mass; this
was followed by
a lifetime of intake greater per gram body weight in the CR
rats than in the
ad libitum-fed rats (Masoro et al., 1982). Indeed, assuming
that the
lifetime intake of calories is a valid index of lifetime
caloric
expenditure, CR rats had a markedly greater lifetime caloric
expenditure per
gram body weight than did ad libitum-fed rats (Table 3), a
finding clearly
not in accord with the concepts of Rubner or Pearl.

Table 3. Lifetime caloric intake Rat group Mean length of
life days
======================
Mean lifetime caloric intake (kcal/g body mass)
======================
Ad libitum-fed 701 91.5
CR 986 133.5
======================
Note: This table was generated from data in the report
of Yu et al.
(1982) [ http://tinyurl.com/y9yc2f ].

Expanding our study, McCarter and Palmer (1992) determined
the lifetime 24-h
metabolic rate of male F344 rats under usual living
conditions. A brief
period of reduced metabolic rate followed the initiation of
CR at 6 weeks of
age. However, the metabolic rate of rats on a CR regimen
quickly increased
to a rate higher per gram body weight than that of those fed
ad libitum,
while on a per gram lean body mass, the rate was the same as
those fed ad
libitum. The transient decrease in metabolic rate relates to
the period of
time in which the lean body mass falls following initiation
of CR. After a
relatively brief period, the decrease in lean body mass
matches the decrease
in caloric intake and, at this point, the lean body mass is
stable. Duffy et
al. (1991) reported similar findings for mice. In two
long-term studies of
CR in rhesus monkeys, the metabolic rate per gram of lean
body mass
decreased following the initiation of CR, but rose to that
of non-restricted
monkeys as the CR regimen continued (Lane et al., 1996 and
Ramsey et al.,
1996). Other studies, such as the rhesus monkey study of
DeLaney et al.
(1999), have shown that CR results in a sustained decrease
in the metabolic
rate. Nevertheless, strongly weighing against the Reduction
of Metabolic
Rate Hypothesis is the fact that there are rodent studies
showing that CR
can extend life without a long-term reduction in metabolic rate.

In spite of these findings, this hypothesis still has many
advocates. And
one does, indeed, encounter the following question when
dismissing it: when
comparing rats on CR to those fed ad libitum, is it
appropriate to normalize
the metabolic rate for body size by expressing the findings
per unit of
either body mass or lean body mass? The confounder is that
the relative
sizes of tissues and organs within the rodent are not the
same for animals
on CR as those fed ad libitum (Yu et al., 1982). Greenberg
and Boozer (2000)
tried to address this confounder by expressing the metabolic
rate per unit
of the combined mass of heart, kidneys, brain, and liver;
using this
normalization method, they found that 22-month-old male F334
rats on a CR
regimen and those fed ad libitum have the same metabolic
rate. However,
Gallagher et al. (2003) point out that the methods so far
used in CR studies
to normalize for body size are based on dubious assumptions.
They assert
that what is required are in vivo measurements of the
specific metabolic
rate of individual organs and tissues. Although the question
of the effect
of CR on metabolic rate remains yet to be answered, the
recent report of
Speakman et al. (2004) indicates that Sacher's Reduction of
Metabolic Rate
Hypothesis is not likely to be correct. They reported a
positive correlation
between metabolic rate and longevity in mice, a finding that
challenges both
Sacher's hypothesis and Pearl's "rate of living theory of
aging".

3.4. Oxidative Damage Attenuation Hypothesis
Harman (1956) proposed that ageing is due to damage caused
by free radicals.
With the recognition that the metabolic use of oxygen is the
major
biological source of free radicals (e.g., hydroxyl and
superoxide radicals)
as well as other damaging reactive oxygen molecules (e.g.,
hydrogen
peroxide), Harman's theory evolved into the oxidative stress
theory of
ageing. It also gave rise to the mitochondrial theory of
ageing, since
mitochondria are the major source of reactive oxygen
molecules (Beckman and
Ames, 1998 and Barja, 2000). These reactive oxygen molecules
can damage
important biological molecules, including DNA, proteins and
lipids, thereby
altering cellular functions. Currently, many believe that
the accumulation
of oxidative damage is the primary basis of ageing. Although
several
investigators have theorized that CR retards aging and
extends life by
slowing the age-associated increase in oxidative damage,
Sohal and Weindruch
(1996) have provided an especially clear and succinct
presentation of this
concept, which I call the Oxidative Damage Attenuation
Hypothesis.

Indeed, it is well established that CR retards the
age-associated
accumulation of oxidatively damaged molecules in rodents;
the reader can
find a review of many of these studies in Yu (1996). There
is also some
evidence that CR acts the same way in monkeys (Zainal et
al., 2000). This
attenuation of the accumulation of oxidative damage must be
due to either a
decreased rate of generation of reactive oxygen molecules,
or to increased
efficiency of protective processes, or to an increase in
repair activity, or
to a combination of these processes. While several studies
have showed that
CR decreases the formation of reactive oxygen molecules by
isolated
mitochondria and microsomes from CR rodents, Feuers et al.
(1993) point out
that little is known about this action in a functioning
intact rat or mouse.
Of course, it is risky to draw physiologically relevant
conclusions solely
from in vitro studies. Although many studies have shown that
CR increases
the activity or retards the age-associated decrease in
activity of enzymes
like catalase, superoxide dismutase, and glutathione
peroxidase, which
protect rodents from oxidative damage, other studies have
shown just the
opposite effect (for examples see Richardson, 1991 and
Luhtalta et al.,
1994). Obviously, the action of CR on the antioxidant
enzymes is far from
simple.

The ability of CR to bolster the non-enzymatic antioxidant
defenses, such as
increasing the levels of reduced glutathione, has been
clearly shown (Armeni
et al., 1998). CR is also known to enhance the ability to
repair oxidatively
damaged DNA (Guo et al., 1998) and to replace damaged
proteins (Lewis et
al., 1985).

Thus, it seems beyond question that CR protects rodents from
damage caused
by oxidative stress. Is this action the major mechanism
underlying the
life-prolonging and senescence-retarding actions of CR? The
answer to this
question depends on whether oxidative stress plays a major
role in ageing.
There is a body of evidence in support of a major role for
oxidative damage
in ageing: the classic study of Arking et al. (1991) showed
that flies
selected for postponed senescence have an increased
resistance to oxidative
stress. Also, the study of Migliaccio et al. (1999) showed
that mice with a
homozygous mutation in the p66shc gene exhibit a markedly
increased life
span and an enhanced resistance to oxidative stress. Female
mice
heterozygous for the disruption of the IGF-1 receptor gene
also have an
extended life span and increased resistance to oxidative stress
(Holzenberger et al., 2003).

However, there also have been studies indicating, at least
in some
instances, that oxidative stress is not an important factor
in the
occurrence of senescence. Hauck et al. (2002), reporting
work on growth
hormone receptor/binding protein gene knockout mice, found
that although
there is an increase in life span compared to the wild type,
the knockout
mice are more susceptible to damage from paraquat. Van
Remmen et al. (2003)
found that in mice deficient in Mn-superoxide dismutase, the
increased
oxidative stress/damage does not affect life span and other
age-sensitive
parameters. Orr et al. (2003) reported that the life span of
long-lived
Drosophila is not increased by the over-expression of the
following
antioxidant enzymes: CuZn-superoxide dismutase,
Mn-superoxide dismutase,
thioredoxin reductase, and catalase. Thus, it remains an
open question
whether CR's ability to attenuate oxidative damage plays a
major role in its
life-extending action.

3.5. Altered Glucose-Insulin System Hypothesis
Although by 1990, it had been known for some time that
fasting levels of
plasma glucose and insulin are lower in rodents on a CR
regimen, the
question had not been addressed as to whether such is the
case under usual
daily living over a lifetime. Therefore, a lifetime
longitudinal study on
male F344 rats was carried out in our laboratory (Masoro et
al., 1992). We
found that throughout the lifetime, CR decreased the mean
24-h plasma
glucose concentration by about 15 mg/dl and the insulin
concentration by
about 50%. Moreover, the rats on the CR regime used glucose
as fuel at the
same rate per unit of metabolic mass as did the rats fed ad
libitum, despite
the lower plasma glucose and markedly lower plasma insulin
levels. We
concluded that CR either increased glucose effectiveness or
insulin
responsiveness or both, and proposed that the lifetime
maintenance of low
levels of glucose and markedly low levels of insulin played
a major role in
the life-extending and related actions of CR (Altered
Glucose-Insulin System
Hypothesis).

CR has been found to reduce plasma glucose and insulin
concentrations in
fasting rhesus monkeys (Kemnitz et al., 1994 and Lane et
al., 1995). In
addition, CR increases insulin sensitivity in rhesus and
cynomolgus monkeys
(Kemnitz et al., 1994, Bodkin et al., 1995, Lane et al.,
1995 and Cefalu et
al., 1997).

In recent years, the focus has been on the insulin component
of the
glucose-insulin system. The major reasons for this emphasis
are findings
that loss-of-function mutations of the insulin signaling
system result in
life extension in three species: C. elegans (Kenyon et al.,
1993 and Wolkow
et al., 2000), D. melanogaster (Clancy et al., 2001 and
Tatar et al., 2001),
and mice (Blüher et al., 2003). Clearly, by maintaining
plasma insulin at a
markedly low level throughout life, CR is in effect
decreasing insulin
signaling.

3.6. Alteration of the Growth Hormone-IGF-1 Axis Hypothesis
In the early 1990s, Bill Sonntag's group reported that CR
results in
markedly lower levels of plasma insulin-like growth factor-1
(IGF-1) in rats
and mice (Breese et al., 1991, Sonntag et al., 1992 and
D'Costa et al.,
1993). The possible significance of this finding was not
recognized until
some years later when the Ames and Snell Dwarf mice were
found to exhibit
life extension (Brown-Borg et al., 1996, Bartke et al., 2001
and Flurkey et
al., 2001). Among other endocrine characteristics, these
dwarf mice exhibit
an inability to secrete growth hormone and consequently low
plasma levels of
IGF-1. Furthermore, Coschigano et al. (2000) have studied
mice in which the
growth hormone receptor/binding protein gene has been
disrupted, and found
that they exhibit high levels of plasma growth hormone, very
low levels of
plasma IGF-1, and life extension. These and related studies
led to the view
that the reduction of plasma IGF-1 in rodents on CR regimens
may play an
important role in its life-extending action, which I term
the Alteration of
the Growth Hormone-IGF-1 Axis Hypothesis.

Further support for the foregoing hypothesis comes from the
fact that
nematodes and fruit flies do not have separate receptors for
insulin and
IGF-1. Thus, the genetic studies that implicate reduction in
insulin
signaling in the life extension of these species also apply
to IGF-1
signaling.

Holzenberger et al. (2003) have reported findings that at
first sight appear
to provide direct support for the role of the attenuating
IGF-1 signaling in
life extension of mice. They studied male and female mice
heterozygous for
the disruption of IGF-1 receptor gene and found that the
females with this
disruption have a statistically significant increase in mean
length of life
compared to wild type females, while in the case of the male
mice, the
increase was not statistically significant. However, there
are several
concerns about this study and its use to draw conclusions
about the role of
decreased IGF-1 signaling in the actions of CR. First, since
only a small
number of mice was studied, to conclude that reduction in
IGF-1 signaling
increases longevity in the female but not in the male is a
precarious claim.
Second, if such a gender difference were truly the case, it
would indicate
that the reduction in IGF-1 signaling does not play an
important role in
life extension by CR, because it has been found that CR
increases the length
of life in a variety of rat and mouse strains with no
evidence that the
extent of the effect relates to gender (Lewis et al., 1999).
In addition,
there are other issues that are not addressed in the
Holzenberger et al.
paper. The wild type 129/Sv strain of mice used in this
study had a much
shorter life (mean length of life less than 19 months in the
females) than
most mice strains, and it is not clear from the paper
whether this is
typical of this mouse strain. In fact, while the mice were
maintained in a
conventional facility (rather than in specific-pathogen-free
facility), the
problem of infectious disease was neither discussed, nor
were any pathology
data presented. Indeed, it seems essential to know whether
the life
extension relates to an effect on a specific disease process
in the female
mice.

3.7. Hormesis Hypothesis
Hormesis refers to the phenomenon whereby a usually detrimental
environmental agent (radiation, chemical substance, etc.)
changes its role
to provide beneficial effects when administered at low
intensities or
concentrations (Furst, 1987). In regard to biological
gerontology, I have
modified the definition somewhat as follows: hormesis is the
beneficial
action resulting from the response of an organism to a
low-intensity
stressor. These beneficial actions include: increased
longevity, retardation
of senescent deterioration, retardation of age-associated
diseases, and
enhanced coping with intense stressors. I proposed that CR's
ability to do
all of these is due to hormesis and termed it the Hormesis
Hypothesis
(Masoro, 1998).

The first issue that needs to be addressed is whether CR is
a low-intensity
stressor. Strong support comes from findings in both rats
and mice that CR
results in the daily elevation of circadian peak plasma free
corticosterone
levels throughout the life span (Sabatino et al., 1991 and
Han et al.,
1995).

The next issue is whether rodents on a CR regimen exhibit an
enhanced
ability to cope with intense stressors. Indeed, CR has been
found to have
beneficial actions in this regard. CR attenuates the loss in
body weight of
rats following surgery for the implantation of jugular
cannulae (Masoro,
1998), and it increases the ability of rats to survive
intense heat stress
(Heydari et al., 1993). The inflammatory reaction following
the injection of
carageenan into the footpad is attenuated in mice on a CR
regimen (Klebanov
et al., 1995). CR also protects rodents from the damaging
action of toxic
drugs (Berg et al., 1994, Duffy et al., 1995 and Keenan et
al., 1997).

Does CR's ability to increase the resistance of rodents to
acute, intense
stressors have any relevance to its ability to retard
senescence and extend
life? The Disposable Soma Hypothesis of Ageing poses that
less energy is
used for somatic maintenance than needed for indefinite
survival (Kirkwood,
1977). Thus, with increasing age, there is an accumulation
of damage caused
by both endogenous stressors, such as the reactive oxygen
molecules produced
during fuel utilization, and a spectrum of environmental
factors, such as
chemical toxins, infectious agents, etc. By promoting the
hormetic
processes, it is proposed that CR attenuates the rate of
accumulation of
damage from these various agents, thereby retarding
senescent deterioration
and extending life.

What are the organismic, cellular, and molecular mechanisms
involved in the
hormetic action of CR? At an organismic level, the daily
elevation of blood
glucocorticoids could well be one. It is well known that
glucocorticoids
play a key role in enabling mammalian species to cope with
stressors (Munck
et al., 1984). Indeed, Chung et al. (2001) proposed the
Inflammation
Hypothesis of Aging, which postulates that inflammatory
processes play a key
role in aging. The daily moderate elevation of plasma free
corticosterone
induced by CR in rats and mice would be expected to have a
significant
anti-inflammatory action.

At the cellular and molecular level, another possibility is
an increase in
the activity of genes that protect cells from the damaging
action of harmful
agents (Papaconstantinou et al., 1996). Indeed, CR has been
shown to
increase the induction of hepatic HSP 70, one of these
protective proteins,
in response to heat stress (Heydari et al., 1993).

Finally, studies of S. cervisiae by Anderson et al. (2003)
detailed the
afferent hormetic pathway by which CR extends the
replicative life of this
yeast species. They found that not only is a functional Sir2
gene needed for
CR-induced replicative life extension but, in addition, a
functional PNC1
gene is required. The PNC1 gene encodes a protein with
nicotinamidase
activity and CR acts to increase the amount of this enzyme.
The deacetylase
activity of the SIR2 protein involves the generation of
nicotinamide, which
is an inhibitor of the deacetylase activity of SIR2 protein.
By reducing the
level of nicotinamide, the PNC1 protein increases the
deacetylase activity
of the SIR2 protein, and it is this deacetylase activity
that plays a key
role in the CR-induced increase in the replicative life span
of this yeast
species. They also found that this same pathway is involved
in the
heat-stress- and osmotic-stress-induced extension of
replicative life span
of S. cervisiae. Thus, a spectrum of low-intensity stressors
shares a common
hormetic pathway. Moreover, as discussed above, there is
growing evidence
that sirtuin proteins are also involved in the
life-extending action of CR
in other species, including mammals.

These recent findings provide strong support for the
Hormesis Hypothesis, a
concept that embraces many of the other proposed hypotheses.
Indeed, many of
those hypotheses, such as the Oxidative Damage Attenuation
Hypothesis,
appear to be merely specific components of the hormetic
mechanism underlying
life extension by CR. However, there is concern that
although many
low-intensity stressors increase both mean and maximum
length of life in
many species, some apparent stressors do not. For example,
Holloszy et al.
(1985) found that voluntary exercise increased the mean but
not the maximum
length of life of rats. In still another study, Holloszy and
Smith (1986)
reported that cold exposure increased neither mean nor
maximum length of
life of rats. The explanation may simply be that hormesis is
a response to
low-intensity stressors and that high-intensity stressors
are usually
detrimental rather than beneficial. Indeed, it is not
evident that voluntary
exercise is a stressor for rats, while Holloszy and Smith's
cold exposure
protocol may have been too intense to induce hormesis.