Effect of metformin on life span and on the development of spontaneous
mammary tumors in HER-2/neu transgenic mice
Vladimir N. Anisimova,b,*, Lev M. Bersteina, Peter A. Egormina, Tatiana S. Piskunovaa,
Irina G. Popovicha, Mark A. Zabezhinskia, Irina G. Kovalenkoa, Tatiana E. Poroshinaa,
Anna V. Semenchenkoa,b, Mauro Provincialic, Francesca Rec, Claudio Franceschid
aDepartment of Carcinogenesis and Oncogerontology, N.N.Petrov Research Institute of Oncology, Pesochny-2, St Petersburg 197758, Russian Federation
bMax-Planck Institute for Demographic Research, Rostock 18059, Germany
cLaboratory of Tumor Immunology, I.N.R.C.A. Gerontology Research Department, Via Birarelli 8, Ancona, Italy
dDepartment of Experimental Pathology, University of Bologna, Bologna, Italy
Received 26 April 2005; received in revised form 21 June 2005; accepted 4 July 2005
Available online 24 July 2005
Studies in mammals have led to the suggestion that hyperglycemia and hyperinsulinemia are important factors both in aging and in the
development of cancer. Insulin/insulin-like growth factor 1 (IGF-1) signaling molecules that have been linked to longevity include DAF-2
and InR and their homologues in mammals, and inactivation of the corresponding genes is followed by increased life span in nematodes, fruit
flies and mice. It is possible that the life-prolonging effects of calorie restriction are due to decreasing IGF-1 levels. A search of
pharmacological modulators of insulin/IGF-1 signaling pathway (which mimetic effects of life span extending mutations or calorie
restriction) could be a perspective direction in regulation of longevity. The chronic treatment of female transgenic HER-2/neu mice with
metformin (100 mg/kg in drinking water) slightly decreased the food consumption but failed in reducing the body weight or temperature,
slowed down the age-related rise in blood glucose and triglycerides level, as well as the age-related switch-off of estrous function, prolonged
the mean life span by 8% (p!0.05), the mean life span of last 10% survivors by 13.1%, and the maximum life span by 1 month incomparison
with control mice. The demographic aging rate represented by the estimate of respective Gompertz’s parameter was decreased 2.26 times.
The metformin-treatment significantly decreased the incidence and size of mammary adenocarcinomas in mice and increased the mean
latency of the tumors.
q 2005 Elsevier Inc. All rights reserved.
Keywords: HER-2/neu; Transgenic mice; Mammary cancer; Biguanides
The potential link between aging and insulin/IGF-1
signaling has attracted substantial attention during last
years. The potential connection was evidenced by an
increase in incidence of insulin resistance and type 2
diabetes in accelerated aging syndromes, on the one side, as
well as by life span extension due to caloric restriction (CR)
in rodents, on the other. Concomitant reduction in plasma
insulin and plasma glucose levels, which implies increased
sensitivity to insulin, emerges as a hallmark of increased
longevity (Bartke et al., 2003). Hyperglycemia is an
important aging factor involved in generation of advanced
glycosylation end products (AGEs) (Facchini et al., 2000;
Elahi et al., 2002). There is evidence that hyperinsulinemia
favors accumulation of oxidized protein by reducing its
degradation as well as facilitates protein oxidation by
increasing steady-state level of oxidative stress (Facchini
et al., 2000). Untreated diabetics with elevated glucose
levels suffer many manifestations of accelerated aging, such
as impaired wound healing, obesity, cataracts, vascular and
microvascular damage (Dilman, 1994). It is important
to stress that hyperinsulinemia is an important factor
not only in aging but also in the development of cancer
Experimental Gerontology 40 (2005) 685–693
0531-5565/$ - see front matter q 2005 Elsevier Inc. All rights reserved.
*Corresponding author. Address: Department of Carcinogenesis and
Oncogerontology, N.N.Petrov Research Institute of Oncology, Pesochny-2,
St Petersburg197758,Russian Federation.Tel.: C7 812 596 8607; fax: C7
812 596 8947.
E-mail address: email@example.com (V.N. Anisimov).
(Dilman, 1994; Colangelo et al., 2002; Gupta et al., 2002;
Pollak et al., 2004).
The concept of CR mimetics is now being intensively
explored (Hadley et al., 2001; Mattson et al., 2001;
Weindruch et al., 2001). CR mimetics involve interventions
that produce physiological and anti-aging effects similar to
CR. It was suggested to use biguanide antidiabetics as a
potential anti-aging treatment (Dilman, 1994; Dilman and
Anisimov, 1980; Anisimov, 2003; Anisimov et al., 2003).
The antidiabetic drugs, phenformin and buformin, were
observed to reduce hyperglycemia and produce the
following effects: improved glucose utilization; reduced
free fatty acid utilization, gluconeogenesis, serum lipids,
insulin and IGF-1, and reduced body weight both in humans
and experimental animals (Dilman, 1994; Berstein, 2005).
Breast cancer is one of the most common cancers and is a
leading cause of mortality in women (Parkin et al., 2001).
The HER-2/neu oncogene encodes a 185 kDa (p 185)
receptor protein belonging to the epidermal growth factor
receptor family involved in organogenesis and epithelial
differentiation (Andrechek et al., 2000). Amplification and
mutation of HER-2/neu plays a pathogenetic role in several
malignancies, including carcinoma of the breast, ovary and
uterus (Weinstein et al., 2000). Overexpression of ErbB-2/
HER-2/neu occurs in 15–40% of human breast cancers
(Jones and Stern, 1999). Its appearance correlates with poor
prognosis and it is, therefore, an important target for
physiologic investigation and therapeutic intervention
(Weinstein et al., 2000).
In this paper, we for the first time present the results of
experiments with the antidiabetic biguanide metformin on
some ageing related biological parameters, on the survival
and on the spontaneous tumorigenesis in female transgenic
HER-2/neu mice. The similarly designed long-term
experiment with metformin performed on low-cancer
incidence mouse strain is in progress.
2. Material and methods
Homozygous HER-2/neu transgenic mice originally
obtained from Charles River (Hollister, CA) by the Italian
National Research Center for Aging (INRCA) were housed
and breed in the Department of Carcinogenesis and
Oncogerontology, N.N.Petrov Research Institute of Oncol-
ogy. The mice were kept 5–7 in polypropilene cages (30!
21!10 cm) under standard light/dark regimen (12 h light:
12 h darkness) at 22G2 8C and received standard laboratory
chow (Anisimov et al., 2002) and tap water ad libitum.
2.2. Experimental design
One-hundred and six female FVB/N HER-2/neu mice at
the age of 2 months were randomly divided into two groups.
Mice of the first group were given metformin (Siophor,
Berlin-Chemie, Menarini Group) with drinking water
(100 mg/kg) for five consecutive days every month, whereas
the mice of the second group were given tap water without
metformin and served as a control. This dose of metformin
is similar to used in our earlier experiments with phenformin
(Dilman and Anisimov, 1980) and equal to 300 mg/m2of
the surface area. Recalculation for human gives in average
510 mg/m2that much less than commonly used in clinical
practice (1.0–2.5 g per day). Once a week all mice were
palpated for detection of mammary tumors appearance. The
localization and the size of tumors were registered on the
special charts. Once a month all mice were weighted and,
simultaneously, the amount of consumed food and water
was measured, and the rate of the consumed water and food
mass (g) per mouse and per body weight unit were
calculated. Once in every 3 months, daily for 2 weeks
vaginal smears of the animals were cytologically examined
to estimate the estrus function. In the same period, rectal
body temperatures of the mice were measured with an
electronic thermometer, TPEM (KMIZ, Russia).
The time of appearance of mammary tumors was
evaluated by palpation and the neoplastic masses were
measured with calipers in the two perpendicular diameters.
Progressively growing masses of O3 mm in mean diameter
were regarded as tumors. Because some treated mice did not
display carcinomas in all mammary glands, the mean
was calculated as the cumulative number of incident
tumors/number of tumor-bearing mice. At the age of 5
and 9 months 10 mice from each group were sacrificed by
decapitation after overnight starvation. Samples of serum
were obtained and stored in the K20 8C for subsequent
analyzes. Other animals were observed until their natural
deaths. The date of each death was registered, and the mean
life span, the age at which 90% of the animals died, and the
maximum life span were estimated.
2.3. Metabolic and hormonal assays
The serum levels of total b- and pre b-lipoproteins were
estimated according to Ledvina (1960); glucose—by
enzymocolorimetric (glucose-oxidase) method with kits
from ‘Impact’ (Moscow, Russia); cholesterol and triglycer-
ides—by enzymocolorimetric method with kits of ‘Olvex’
(St Petersburg, Russia); insulin—by immune enzyme assay
(ELISA) with kits from Diagnostic Systems Laboratories,
Inc. (USA); total thyroxine and total triiodothyronine—by
immune enzyme assay with kits from HEMA-Medica,
2.4. RNA extraction and RT-PCR
The expression of mRNA for perforin and granzyme B
was evaluated in mammary tumor masses of mice exposed
or not-exposed to Metformin by RT-PCR. After
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693686
homogenization of tissue sample, RNA was extracted using
TRI-REAGENTeaccording to the manufacturer’s instruc-
tions (Sigma Chemical Co., USA). After treatment with
DNase (InvitroGen), RNA concentrations were determined
using the spectrophotometer absorbance at 260 nm (scien-
tific instruments UV1601 Shimadzu, Columbia, MD, USA).
cDNA was synthesized from 1 mg RNA incubating RNA
with dNTP (0.5 mM), Oligo dT (12.5 ng/ml), First Strand
Buffer (1X), M-MLV Reverse Transcriptase (10 U/ml),
Rnase Inhibitor (1 U/ml) and DTT (0.01 M) all from
Gibco BRL, in a final volume of 20 ml. The samples were
incubated at 37 8C for 1 h and 95 8C for 10 min.;
subsequently cDNA was frozen at K20 8C until use. PCR
was performed incubating 5 ml (0.25 mg) cDNA with a
reaction mixture containing: PCR Buffer (1X), MgCl2
(1.5 mM), dNTP (200 mM), specific forward and reverse
primers (0.4 mM of each), Taq DNA Polymerase (1 U/ml) in
a total volume of 50 ml (all from Roche Diagnostics GmbH,
Germany). The samples were incubated in a GeneAmp PCR
System 9700 (Perkin Elmer) for a total of 35 cycles for
perforin and granzyme B, and 30 cycles for b-actin. Each
cycle consisted of: 1 min at 94 8C, 1 min. at 60 8C, 1 min at
72 8C for perforin and granzyme B; 1 min at 94 8C, 2 min at
63 8C, 2 min at 72 8C for b-actin.
The primers for perforin, granzyme B and b-actin were
purchased from Roche Diagnostics (GmbH, Germany)
using DNA published cDNA sequences. The perforin
fragment of 486 bp was defined by the forward primer: 50-
GGTGGAGTGGAGGTTTTTGTACC and the reverse
primer: 50-CAGAATGCAAGCAGAAGCACAAG. The
granzyme B fragment of 531 bp was defined by the forward
primer: 50-CCTGAAGGAGGCTGTGAAAGAATC and
the reverse primer: 50-CCCTGCACAAATCATGTTT
AGTCC. Mouse b-actin fragment of 349 bp was evaluated
by the forward primer: 50-TGGAATCCTGTGGCATCC
ATGAAAC and the reverse primer: 50-TAAAACGCAG
CTCAGTAACAGTCCG. The PCR products and a mol-
ecular weight standard (DNA molecular weight marker
VIII, Roche Diagnostics) were visualized after electrophor-
esis in a 1.5% agarose gel containing 1 mg/ml ethidium
bromide (EtBr). Densitometric analysis was performed
using the Gel Doc 2000 (BioRad Laboratories, Italy).
2.5. Pathomorphological examination
All animals were autopsied. Site, number and size of
mammary tumors and their metastases in lungs were
checked. All tumors, as well as the tissues and organs
with suspected tumor development were excised and fixed
in 10% neutral formalin. After the routine histological
processing the tissues were embedded into paraffin. Thin
histological sections of 5–7 mm were stained with haema-
toxylin and eosine and were microscopically examined.
Tumors were classified according to International Agency
for Research on Cancer recommendations (Turusov
and Mohr, 1994).
2.6. Statistical methods
Experimental results were statistically processed by
the methods of variation statistics with the use of
STATGRAPH statistic program kit. The significance of
the discrepancies was defined according to the Student t-
criterion, Fischer’s exact method, c2, non-parametric
Wilcoxon–Mann–Whitney and Friedman RM Anova on
Ranks. Student–Newman–Keuls Method was used for all
pairwise multiple comparisons. Coefficient of correlation
was estimated by Spearman’ method (Goubler, 1978).
Differences in tumor incidence were evaluated by the
Mantel–Haenszel log-rank test.
Parameters of Gompertz model were estimated using
maximum likelihood method, non-linear optimization pro-
cedure (Fletcher, 1987) and self-written code in ‘Matlab’;
confidence intervals for the parameters were obtained using
the bootstrap method (Davison and Hinkley, 1997).
For experimental group the Cox’s regression model
(Cox, 1972) was used to estimate relative risk of death and
tumor development under the treatment compared to the
control group: h(t,z)Zh0(t) exp(zb), where h(t,z) and h0(t)
denote the conditional hazard and baseline hazard rates,
respectively, b is the unknown parameter for treatment
group, and z takes values 0 and 1, beingan indicator variable
for two samples—the control and treatment group.
Semiparametric model of heterogeneous mortality
(Semenchenko et al., 2004) was used to estimate the
influence of the treatment on frailty distribution and
3.1. Age-related body weight dynamics
The body weight of mice in both control and metformin-
treated groups increased with age, exceeding by 9 months
the body weight of 3-month-old animals by 66.2% in the
control group, and by 64.1% in the group treated with
metformin. There was no difference in the mean body
weight of mice exposed and non-exposed to the drug during
the period of observation (data not shown).
Effect of metformin on the age-related food consumption dynamics in
female HER-2/neu mice
GroupDaily food consumption (g/mouse)
3 months4 months6 months 9 months
aSignificant difference with control (p!0.05).
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693687
3.2. Age-related dynamics of food
and water consumption
The amount of food and water daily consumed by mice
during the period of observation was similar in the control
group and in metformin-treated group (Table 1). However,
at the age of 6 months the food consumption was decreased
by 21.4% in metformin-treated group.
3.3. Age-related dynamics of estrous
function in mice
The length of estrous cycles in the control female
HER-2/neu mice was not significantly changed with age,
whereas it was increased between 2 and 5 months of age in
mice exposed to metformin (Table 2). The relative number
of long estrous cycles was increased in metformin-treated
mice with age and was higher than that in the controls. A
similar pattern of age-related changes in the rate of mice
with regular and irregular estrous cycles in both groups was
3.4. Age-related dynamics of body
temperature in mice
There was no difference in average body temperature
between control and metformin-treated groups during the
entire period of observation (data not shown).
3.5. Survival and longevity of female HER-2/neu
According to the log-rank test (Cox and Oakes, 1996)
distributions of life span in the control and experimental
groups were identical with probability pZ0.000573
(c2Z11.9 at 1 degree of freedom). Survival dynamics in
control and metformin-treated mice are shown in Fig. 1.
Only one control mouse (3%) survived the age of 10 months
whereas 9 mice (28%) survived this age in metformin-
treated group (p!0.001) (Table 3).
Metformin treatment significantly increased mean
(C8.0%) and maximal life span (C9%) of female
HER-2/neu transgenic mice (Table 4). Mean life span of
long-living individuals (last 10% of survivors) was signifi-
cantly greater in the group exposed to metformin (C13.1%)
compared to the control one. Parameter a of the Gompertz
model, which is interpreted as the rate of aging, was lower
Effect of metformin on age-related dynamics of estrus functional parameters in HER-2/neu mice
Age (months)No. of miceLength of estrous
Rate of estrous cycles of various length (%) Fraction of mice with
regular cycles (%)
!5 days 5–8 days
The difference with the controls of corresponding age in the control group.
ap!0.05 (Student’s t-test).
bp!0.05 (Fischer’s exact test).
Effect of metformin on survival distribution in female HER-2.neu mice
Group No. of survivors at the age of
6 mo 7 mo8 mo 9 mo10 mo 11 mo12 mo
aThe difference with the corresponding age in the control group is
significant: p!0.01 (Fischer’s exact test).
150 200 250 300350
No. of mice, %
Fig. 1. Effect of metformin on survival curves in female transgenic HER-
2/neu mice. Ordinate, number of mice, %; abscissa, age, days; 1, control; 2,
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693688
than in controls, whereas the mortality rate doubling time
(MRDT) was increased in metformin-treated group of mice.
The difference between the control and the experimental
group was statistically significant (Table 4).
3.6. Spontaneous tumor development in female
HER-2/neu transgenic mice
According to the long-rank test (Caraci et al., 2003) age-
related distributions of the first mammary adenocarcinoma
(MAC) occurrence are identical in control and metformin-
treated groups (pZ0.000385; c2Z12.6 at 1 degree of
freedom). As shown in Fig. 2, a slower kinetic of tumor
incidence was found in metformin-treated mice in compari-
son with control animals (P!0.001), with a 50% of tumor-
bearing mice revealed at 173 or 194 days of age in control or
metformin-treated mice, respectively. All mice were
bearing a tumor at day 207 or day 226 in control or
metformin groups, respectively. Metformin did not
influence the distribution of the mean number of tumors
per mice, however, it decreased significantly the mean size
and accumulation of MAC (p!0.05) (Table 5; Fig. 3).
Half of control mice were bearing 9 to 10 tumors per
mouse (47%), whereas in the group treated with metformin
only 23.5% of mice were bearing this number of MAC per
animal (p!0.05) (Fig. 4).
3.7. Effect of metformin on metabolic and hormonal
parameters in transgenic HER-2/neu mice
Between the age of 5 and 9 months the average value of
serum level of glucose was increased by 2.1 times in the
control group and by 1.25 times in the group treated with
metformin (Table 6). The level of total cholesterol and total
lipoproteins was 45% lower in 5-month-old animals treated
Effect of metformin on parameters of life span in female HER-2/neu mice
Number of mice
Mean life span, days
Mean life span of last
10% of survivors
Maximum life span,
Aging rate a (daysK1)b
297G7.3 336G2.7 (C13.1%)a
0.0762 (0.072; 0.078)0.03373 (0.031;
MRDTc, days 9.10 (8.89; 9.62)
aThe difference with controls is significant: p!0.05.
bConstant a in the Gompertz equation: RZR0(exp) at, where R0Z
mortality at tZ0.
cMRDT, mortality rate doubling time, days. 95% confidence limits are
given in parentheses.
Effect of metformin on mammary adenocarcinomas development in female
Number of mice
Number of tumor-bearing mice (%)
Age of the first mammary tumor
Total number of mammary
Number of mammary tumors per
Mean tumor size (cm)
Number of mice with metastases of
mammary adenocarcinoma into
Maximum size of the metastases
Mean latency of the 1st mammary
The difference with controls is significant: * p!0.05.(Student’s t-test).
120140 160 180200220 240
No. of tumorbearing mice, %
Fig. 2. Effect of metformin on tumor yield curves in female transgenic
HER-2/neu mice Ordinate, number of tumor-bearing mice, %; abscissa,
age, days; 1, control; 2, metformin.
100 150200 250300
Cumulative No. of tumors
Fig. 3. Effect of metformin on cumulative number of tumors in female
transgenic HER-2/neu mice Ordinate, cumulative number of tumors-
bearing mice, abscissa, age, days; 1, control; 2, metformin.
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693689
with metformin than in 9-month-old mice; the level of
triglycerides decreased with age in both groups. The level of
insulin in the serum of control mice was increased in
average by 3.6 times between the age of 5 and 9 months and
by 2.4 times in the metformin-treated group. The serum
levels of thyroxine and triiodothyronine were similar in both
groups at the age of 9 months (Table 6).
We calculated coefficients of correlation between some
individual metabolic and hormonal parameters in mice at
the age of 9 months. It was shown that body weight
significantly correlated with the level of insulin in the serum
of mice (rZ0.842, p!0.01) and was rather high between
the body weight and serum glucose level (rZ0.523) and
between body weight and number of MAC per mouse
between the serum insulin level and the latency of MAC
A negativecorrelationwas observed
3.8. Effect of metformin on expression of granzyme B
and perforin mRNA in mammary adenocarcinomas
of transgenic HER-2/neu mice
To evaluate the presence of cytotoxic lymphocytes in
mammary tumors, tumor masses from control and
metformin-treated mice were analyzed for the content of
mRNA encoding granzyme B and perforins. As shown in
Fig. 5, mRNA for both cytolytic molecules was not
detectable in control mice whereas it was significantly
increased in four out of the five mice treated with
metformin. The reversed transcribed and amplified RNA
was normalized for b-actin expression in individual
samples. The relative expression of the perforin or
granzyme B gene in the tumor masses from metformin-
treated mice, as determinated by densitometric analysis, was
0.2G0.1 or 0.4G0.3 (MeanGSD), respectively.
3.9. Mathematical modeling of the results
As it is shown in Table 7, metformin treatment
significantly decreases relative risk of death and MAC
development in HER-2/neu transgenic mice.
Semiparametric model of heterogeneous mortality
(Semenchenko et al., 2004) was used to estimate the
influence of the treatment on frailty distribution and
SeðxÞ Zð1CrgðScðxÞK s2K1Þ
Parameter s2indicates the presence of heterogeneity in
the control population. Effects of changes in the baseline
hazard are controlled by parameters a and b. Changes in
Fig. 4. Effect of metformin on distribution of mice with multiple mammary
tumors Ordinate, total number of tumor-bearing mice, %; 1, 4–6 tumors per
mouse; 2, 7–8 tumors per mouse; 3, 9–10 tumors per mouse. The difference
between parameters 1 and 2, 1 and 3 in controls is significant, p!0.05; the
difference between 1 and 2 in metformin group is significant, p!0.05; the
difference between parameters 2 in control group and metformin-treated
group is significant, p!0.05; the difference between parameters 3 in
control group and metformin-treated group is significant, p!0.05.
Metabolic parameters in female transgenic HER-2/neu mice of different age treated or not treated with metformin
5 months 9 months5 months9 months
Total cholesterol (mM)
Total lipoproteins (ext.
Thyroxine (T4) (ng/ml)
Triiodothyronine (T3) (ng/
nZ5 animals per group. NE, not estimated. The difference with the parameter in 5-months-old mice of the same group is significant (Student t-test) is
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693 690
parameter a reflect permanent (constant) decrease or
increase of the baseline hazard, depending on whether a is
greater or less than zero. Parameter b describes the
amplification or disappearance of the a-effect, according
to whether b is greater or less than zero. Effects of changes
in the frailty distribution are controlled by parameters r and
g. Parameter r!1 shows an increase in the average
robustness, while rO1 indicates an accumulation of frail
individuals in the population. Parameter gs1 shows an
increase (gO1) or decrease (g!1) in the population
heterogeneity. It was shown that the intact control group
is slightly heterogeneous. The group subjected to metformin
treatment has lower baseline hazard, and it is also more frail
on average and more heterogeneous compared to the intact
control group (Table 8).
Our experiments have shown that the long-term
treatment with the antidiabetic biguanide metformin slowed
aging rate and increased life span of all as well as 10% most
long-living female transgenic HER-2/neu mice. The
treatment with metformin inhibited mammary tumor
development and increased their latency that seems the
Fig. 5. Effect of metformin on expression mRNA of granzyme and perforin in mammary adenocarcinomas of transgenic HER-2/neu mice.
Parameters estimates of the Cox’s regression model for the relative risk of
death and tumor development under the treatment with metformin
compared to the control group
Fit of semiparametric model on parameters of survivals of female HER-
2/neu mice treated with metformin as compared to the controls
ParametersEstimated values Confidential intervals
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693691
These observations are in agreement with the data
previously obtained with antidiabetic biguanides phenfor-
min and buformin. It was shown in earlier studies that
phenformin and buformin increase the life span and inhibit
spontaneous carcinogenesis in female C3H/Sn mice
(Dilman and Anisimov, 1980) and female outbred rats,
inhibit mammary carcinogenesis induced by 7,12-dimethyl-
benzo(a)anthracene or N-nitrosomethylurea in female rats,
colon carcinogenesis induced by DMH in rats as well as
carcinogenesis induced by some other chemical carcinogens
and ionizing radiation (Anisimov, 2003; Anisimov et al.,
2003). Treatment with metformin prolongs life span of rats
(G. Roth, personal communication) and mice (McCarty,
2004) and inhibits pancreatic carcinogenesis in hamsters
(Schneider et al., 2001). We believe that the results of our
new experiment with metformin in low cancer strain of mice
(which is in progress) will be useful for better understanding
the mechanisms of the geroprotective effect of the drug.
Like other biguanides, metformin slows down the age-
related disturbances in the estrus function of rodents. It is
noteworthy that metformin improves menstrual regularity,
leading to spontaneous ovulation, and enhances the
induction of ovulation with clomiphene citrate in women
with polycystic ovary syndrome (Awartani and Cheung,
2002; Nestler et al., 2002).
The use of phenformin in humans has been limited the
last two decades because of a potential association with
lactic acidosis. Metformin does not increase risk for lactic
acidosis or for increased lactate levels in type 2 diabetes
(Kruse, 2004), but has some adverse effects, including renal
insufficiency (Nisbet et al., 2004) and gastrointestinal side
effects (Krentz et al., 1994).
Recently, we have shown that antidiabetic drug
2-a) benzimidazol dihydrochloride) also inhibit spon-
taneous carcinogenesis in NMRI and HER-2/neu mice and
DMH-induced colon carcinogenesis in rats (Popovich et al.,
Thus, at least some antidiabetic drugs could be potent
geroprotectors and anticarcinogens.
Several years ago, it was originally suggested to use
biguanide antidiabetics as mimetics of caloric restriction
(CR) and a potential anti-aging treatment (Dilman, 1994).
Although it is known that free radicals are produced during
metabolic reactions, it is largely unknown which factor(s),
of physiological or pathophysiological significance, modu-
late their production in vivo. It has been suggested that
hyperinsulinemia may increase free radicals and therefore
promote aging, independent of glycemia (Dilman, 1994;
Facchini et al., 2000, 2001). Plasma levels of lipid
hydroperoxides are higher, and antioxidant vitamins are
lower in individuals who are resistant to insulin-stimulated
glucose disposal but otherwise glucose tolerant, nonobese,
and normotensive (Facchini et al., 2001). This finding
indicates that enhanced oxidative stress is present before
causeofthe increaseinthe life span.
diabetes ensues and therefore cannot simply be explained by
overt hyperglycemia. There is substantial evidence support-
ing the hypothesis that selective resistance to insulin-
stimulated (muscle) glucose disposal and the consequential
compensatory hyperinsulinemia trigger a variety of meta-
bolic effects, likely resulting in accelerated oxidative stress
and aging (Dilman, 1994; Facchini et al., 2000).
The anti-diabetics biguanides inhibit fatty acid
oxidation, suppress gluconeogenesis in the liver, increase
the availability of insulin receptors, inhibit monoamine
oxidase (Muntoni, 1999), increase sensitivity of hypotha-
lamo-pituitary complex to negative feedback inhibition,
reduce excretion of glucocorticoid metabolites and
dehydroepiandrosterone-sulfate (Dilman, 1994). Recently,
it was shown that metformin decreases platelet super-
oxide anion production in diabetic patients (Gargiulo et
It is noteworthy that experiments in yeast and C. elegans
show that the life extension by CR is not a mechanical
output of low calories and consequence of a reduction in
ROS or AGE formation, but a process that is highly
regulated, triggering metabolic shift toward respiration that
activates the regulator SIR2 (Koubova and Guarente, 2003).
In yeast and worms, life span is extended by extracopies of
SIR2/Sir-2.1 gene (Tissenbaum and Guarente, 2001), by
SIR2 orthologue, Sirt1 (sirtuin 1) (Picard et al., 2004), or by
small molecule sirtuin-1 agonists, e.g. resveratrol (Howitz et
al., 2003). In mammals, it is suggested that SIRT1 is a key
regulator of cell defences and survival in response to stress
(Brunet et al., 2004; Motta et al., 2004). Recently it was
shown that the expression of mammalian Sir2 (SIRT1) is
induced in CR rats as well as in human cells that were
treated with serum from these animals (Cohen et al., 2004).
Long-lived mutant mice and CR rodents are protected from
cancer despite attenuating apoptosis possibly because their
cells possess increased defences and repair mechanism and
they retain the ability to undergo apoptosis if the damage is
beyond repair (Cohen et al., 2004). It was observed that
phenformin inhibits proliferation and induced enhanced and
transient expression of the cell cycle inhibitor p21 and
apoptosis in human tumor cells lines (Caraci et al., 2003).
Finally, in our study we demonstrate that metformin-treated
mice have in their tumoral mammary glands a significant
increase of cytotoxic lymphocytes producing granzyme B
and perforins. Perforins (perforating protein) and associated
granule proteases (granzymes) are cytolytic molecules
located in the cytoplasmic granules responsible of the
cytolytic machinery of lymphocytes (Henkart, 1994). This
lymphocyte infiltrate may be related to the antitumoral
effect exerted by metformin in this transgenic tumor model.
The exact mechanisms involved in this effect need to be
further investigated, even if an increased lymphocyte
metabolism determined by the metformin-induced changes
in glucose metabolism, may be suggested (Dilman, 1994;
Frauwirth and Thompson, 2004).
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693 692
Acknowledgements Download full-text
This article was supported in part by grant # 05-04-48110
from the Russian Foundation for Basic Research, and by
grant # NSh-2293.2003.4 from the President of Russian
Andrechek,E.R.,Hardy,W.R.,Siegel, P.M.,Rudnicki,M.A., Cardiff,R.D.,
2000. Amplification of the neu/erbB-2 oncogene in a mouse model
mammary tumorigenesis. Proc. Natl Acad. Sci. USA 97, 3444–3449.
Anisimov, V.N., Khavinson, V.Kh., Popovich, I.G., et al., 2002. Inhibitory
effect of the peptide epitalon on the development of spontaneous
Anisimov, V.N., 2003. Insulin/IGF-1 signaling pathway driving aging and
cancer as a target for pharmacological intervention. Exp. Gerontol. 38,
Anisimov, V.N., Semenchenko, A.V., Yashin, A.I., 2003. Insulin and
longevity: antidiabetic biguanides as geroprotectors. Biogerontology 4,
Awartani, K.A., Cheung, A.P., 2002. Metformin and polycystic ovary
syndrome: a literature review. J. Obstet. Gynecol. Can. 24, 393–401.
Bartke, A., Chandrashekar, V., Dominici, F., et al., 2003. Insulin-like
growth factor 1 (IGF-1) and aging: controverses and new insights.
Biogerontology 4, 1–8.
in the prevention and treatment of cancer. Cancer Lett. 224, 203–212.
Brunet, A., Sweeney, L.B., Sturgill, F., et al., 2004. Stress-dependent
regulation of FOXO transcription factors by the SIRT1 deacetylase.
Science 303, 2011–2015.
Caraci, F., Chisari, M., Frasca, G., et al., 2003. Effects of phenformin on the
proliferation of human tumor cell lines. Life Sci. 74, 643–650.
Cohen, H.Y., Miller, C., Bitterman, K.J., et al., 2004. Calorie restriction
promotes mammalian cell survival by inducing the SIRT1 deacetylase.
Colangelo, L.A., Gapstur, S.M., Gann, P.H., Dyer, A.R., Liu, K., 2002.
Colorectal cancer mortality and factors related to the insulin resistance
syndrome. Cancer Epidemiol. Biomarkers Prev. 11, 385–391.
Cox, D., 1972. Regression models and life-tables (with discussion). J. R.
Stat. Soc. Ser. B (Methodological) 34, 187–220.
Cox, D.R., Oakes, D., 1996. Analysis of Survival Data. Chapman & Hall,
Davison, A.C., Hinkley, D.V., 1997. Bootstrap Methods and Their
Application. Cambridge University Press, Cambridge, England.
Dilman, V.M., 1994. Development, Aging and Disease. A New Rationale
for an Intervention. Harwood Academic Publishers, Chur.
Dilman, V.M., Anisimov, V.N., 1980. Effect of treatment with phenfromin,
dyphenylhydantoin or L-DOPA on life span and tumor incidence in
C3H/Sn mice. Gerontology 26, 241–245.
Elahi, D., Muller, D.C., Egan, J.M., Andres, R., Veldhuist, J., Meneilly,
G.S., 2002. Glucose tolerance, glucose utilization and insulin secretion
in aging. Novartis Found. Symp. 242, 222–242.
Facchini, F.S., Hua, N.W., Reaven, G.M., Stoohs, R.A., 2000. Hyper-
insulinemia: the missing link among oxidative stress and age-related
diseases? Free Radic. Biol. Med. 29, 1302–1306.
Facchini, F.S., Hua, N., Abbasi, F., Reaven, G.M., 2001. Insulin resistance as a
Fletcher, R., 1987. Practical Methods of Optimization, second ed. Wiley,
Frauwirth, K.A., Thompson, C.B., 2004. Regulation of T lymphocyte
metabolism. J. Immunol. 172, 4661–4665.
Gargiulo, P., Caccese, D., Pignatelli, P., et al., 2002. Metformin decreases
platelet superoxide anion production in diabetic patients. Diabetes
Metab. Res. Rev. 18, 156–159.
Goubler, E.V., 1978. Computing Methods of Pathology Analysis and
Recognition. Meditsina, Leningrad.
Gupta, K., Krishnaswamy, G., Karnad, A., Peiris, A.N., 2002. Insulin: a
novel factor in carcinogenesis. Am. J. Med. Sci. 323, 140–145.
Sherman, S.S., Starke-Reed, P.E., 2001. Human implications of caloric
restriction’s effect on laboratory animals: an overview of opportunities
for research. J. Gerontol. Ser. A 56A (Special issue I), 5–6.
Henkart, P.A.,1994.Lymphocyte-mediatedcytotoxicity: two pathwaysand
multiple effector molecules. Immunity 1, 343–349.
Howitz, K.T., Bitterman, K.J., Cohen, H.Y., et al., 2003. Small molecule
activators of sitruins extend Saccharomyces cerevisiae lifespan. Nature
Jones, F.E., Stern, D.F., 1999. Expression of dominant-negative ErbB2 in
the mammary gland of transgenic mice reveals a role in lobuloalveolar
development and lactation. Oncogene 18, 3481–3490.
Koubova, J., Guarente, L., 2003. How does calorie restriction work? Genes
Dev. 17, 313–321.
Krentz, A.J., Ferner, R.E., Balley, C.J., 1994. Comparative tolerability
profiles of oral antidiabeticagents. Drug Saf. 11, 223–241.
Kruse, J.A., 2004. Review: metformin does not increase risk for lactic
acidosis orincrease lactatelevelsin type2 diabetes.ACP J. Club 141,7.
Ledvina, M., 1960. Estimation of b-lopoprotins in the blood serum with
turbidimetric method. Lab. Delo 3, 13–17.
Mattson, M.P., Duan, W., Lee, J., 2001. Progress in the development of
caloric restriction mimetic dietary supplements.. J. Anti-Aging Med. 4,
McCarty, M.F., 2004. Chronic activation of AMP-activated kinase as a
strategy for slowing aging. Med. Hypotheses 63, 334–339.
Motta, M.C., Divecha, N., Lemieux, M., et al., 2004. Mammalian SIRT1
represses forkhead transcription factors. Cell 116, 551–563.
Muntoni, S., 1999. Metformin and fatty acids. Diabetes Care 22, 179–180.
Nestler, J.E., Stovall, D., Akhther, N., Iorno, M.J., Jakubowicz, D.J., 2002.
Strategies for the use of insulin-sensitizing drugs to treat infertility in
women with polycystic ovary syndrome. Fertil. Steril. 77, 209–215.
Nisbet, J.C., Strurtevant, J.M., Prins, J.B., 2004. Metformin and serious
adverse effects. Med. J. Aust. 180, 53–54.
Parkin, D.M., Bray, F.I., Devesa, S.S., 2001. Cancer burden in the year
2000. The global picture. Eur. J. Cancer 37, S4–S66.
Picard, F., Kurtev, M., Chung, N., et al., 2004. Sirt1 promotes fat
mobilization in white adipocytes by repressing PPAR-g. Nature 429,
Pollak, M.N., Schernhammer, E.S., Hankinson, S.E., 2004. Insulin-like
growth factors and neoplasia. Nat. Rev. Cancer 4, 505–518.
Popovich, I.G., Zabezhinski, M.A., Egormin, P.A., Tyndyk, M.L., Anikin,
I.V., Spasov, A.A., Semenchenko, A.V., Anisimov, V.N., 2005. Insulin
in aging and cancer: new antidiabetic drug Diabenol as geroprotector
and anticarcinogen. Int. J. Biochem. Cell Biol. 37, 1117–1129.
Schneider, M.B., Matsuzaki, H., Harorah, J., Ulrich, A., Standlop, J., Ding,
X.Z., Adrian, T.E., Pour, P.M., 2001. Prevention of pancreatic cancer
induction in hamsters by metformin. Gastroenterology 120, 1263–1270.
Semenchenko, G.V., Anisimov, V.N., Yashin, A.I., 2004. Stressors and
antistressors: how do they influence life span in HER-2/neu transgenic
mice? Exp. Gerontol. 39, 1499–1511.
Tissenbaum, H.A., Guarente, L., 2001. Increased dosage of a sirK2 gene
extends life span in Caenorhabditis elgans. Nature 410, 227–230.
Pathology of tumours in laboratory animals. In: Turusov, V.S., Mohr, U.
(Eds.), 1994. Tumours of the Mouse, second ed IARC Scientific
Publication, 111, vol. I. IARC, Lyon.
Weindruch, R., Keenan, K.P., Carney, J.M., et al., 2001. Caloric restriction
mimetics: metabolic intervention. J. Gerontol. Biol. Sci. 56A (Special
Issue 1), 20–33.
Weinstein, E.J., Kitsberg, D.I., Leder, P., 2000. A mouse model for breast
cancer induced by amplification and overexpression of the neu
promoter and transgene. Mol. Med. 6, 4–16.
V.N. Anisimov et al. / Experimental Gerontology 40 (2005) 685–693693