Journal of Gerontology: BIOLOGICAL SCIENCES
1997, Vol. 52A. No. j . B10-BI9
Copyright 1997 by The Gerontohgical Society of America
A Study of Caloric Restriction and Cardiovascular
Aging in Cynomolgus Monkeys (Macacafascicularis):
A Potential Model for Aging Research
William T. Cefalu,1 Janice D. Wagner,2 Zhong Q. Wang,1 Audrey D. Bell-Farrow,'
Joel Collins,2 Darren Haskell,1 Robert Bechtold,3 and Timothy Morgan4
Departments of 'Internal Medicine, 'Comparative Medicine, 'Radiology, and 4Public Health Sciences,
The Bowman Gray School of Medicine of Wake Forest University.
Caloric restriction has been demonstrated to retard aging processes and extend maximal life span in rodents, and is
currently being evaluated in several nonhuman primate trials. We initiated a study in 32 adult cynomolgus monkeys to
evaluate the effect of caloric restriction on parameters contributing to atherosclerosis extent. Following pretrial
determinations, at which time a baseline measure of ad libitum (ad lib) dietary intake was assessed, animals were
randomized to an ad lib fed group (control) or a caloric restriction group (30% reduction from baseline intake). The
animals are being evaluated for glycated proteins, insulin, glucose, insulin sensitivity measures, and specific
measures of body fat composition by CT scans (e.g., intra-abdominalfat) over specified intervals. The results from the
first year of observation demonstrate a significant diet effect on body weight, and specifically intra-abdominal fat.
Further, insulin sensitivity has been significantly increased after 1 year of caloric restriction compared to the ad lib fed
group. These studies indicate that caloric restriction has a marked effect on a pathologic fat depot, and this change is
associated significantly with an improvement in peripheral tissue insulin sensitivity.
T has been shown that caloric restriction retards aging
ISnyder,
processes in rodents (Weindruch and Walford, 1988;
1989). This observation is supported by evidence
that food restriction increases life span, retards ageassociated physiological changes, and delays or prevents
most age-associated disease. The mechanisms by which
caloric restriction (CR) exerts its effects are unknown. However, if these findings are to be extrapolated to human
beings, it will first be necessary to test varying dietary
regimens in some higher species and evaluate the effect on
several aging processes and age-related diseases, particularly as they relate to human health. Age-related disease such
as atherosclerosis would be a valuable end point to study, but
this goal has been severely hampered by lack of a suitable
animal model. Cynomolgus monkeys have been shown in
multiple studies by our group to be an excellent model for
study of atherosclerosis and pathogenic mechanisms involved in the development of atherosclerosis (Rudel and
Pitts, 1978; Adams et al., 1983; Hamm et al., 1983).
Furthermore, it has been shown that caloric restriction can be
initiated safely and maintained without detrimental effects in
higher species such as nonhuman primates (Ingram et al.,
1990; Kemnitz et al., 1993). Therefore, with evidence
demonstrating safety of caloric restriction in nonhuman
primates and the availability of a nonhuman primate model
for an age-associated human disease (i.e., atherosclerosis), a
clinical trial to evaluate the effect of caloric restriction on the
pathogenesis and extent of atherosclerosis in a nonhuman
primate has been initiated.
Our specific hypothesis is that caloric restriction, over a
sustained period when compared with an ad libitum diet with
equivalent cholesterol intake, may reduce the progression of
BIO
atherosclerotic lesions by reducing age-associated increases
in vascular tissue and blood levels of early and advanced
glycated products (i.e., glycated protein and protein crosslinks), improving peripheral insulin sensitivity, and reducing intra-abdominal fat. To evaluate our hypothesis, we are
studying an aging adult nonhuman primate model of atherosclerosis using dietary intervention in the form of caloric
restriction compared to ad libitum food intake. The diets
have been designed so that the animals have identical cholesterol intake per body weight in order to evaluate the effect of
caloric restriction on atherosclerosis while maintaining similar plasma cholesterol levels. In this article, we describe the
study design for this ongoing nonhuman primate trial and
report the results from the first year of study.
METHODS
Animals
Thirty-two feral adult male monkeys (Macaca fascicularis) were acquired directly from Institut Pertanian (Bogar,
Indonesia). After arrival at our Center, age was assessed by
dentition, the animals were quarantined for 3 months,
treated for internal and external parasites, and tested for
tuberculosis. Routine tuberculin testing continues twice
yearly. All animals are housed socially in pairs except when
separated at meal time. Each monkey pair occupies a cage
constructed of stainless steel with dimensions of 61 cm wide,
71 cm deep, and 86.4 cm high. Each cage allows for the
placement of a sliding partition to separate the animals at
mealtime. They are fitted with containers for food and have
valves extending from the cage to allow water intake ad
libitum. All 32 pair-caged monkeys are housed in a single
Bll
CR AND CARDIOVASCULAR AGING IN MONKEYS
windowless room, 6 x 3.7 m in size. All rooms in the
animal building utilize 100% outside air when outside air
temperatures are above 40 °F. Below 40 °F, the system will
mix recirculated air and outside air to maintain the temperature at approx. 72 °F. During normal operations, all animal
housing areas maintain 10-15 air changes per hour and
utilize 40% niters for air nitration.
As animals were pair-caged, each pair was fed either a
control or CR diet.
Design of the Trial
The trial is a randomized one in which the independent
effect of caloric restriction and its interaction with insulin
resistance, intra-abdominal fat, blood and tissue glycation,
and the relationship to changes in atherosclerotic lesion
extent and composition are being evaluated. The design of
the trial is shown in Figure 1.
Beginning in the fourth month and throughout the remainder of the pretrial (baseline, months 4-6), all of the animals
were fed a moderately atherogenic diet (.25 mg cholesterol/
Cal) containing 30% of calories from fat. While on the ad lib
diet, caloric intake for each individual animal was assessed
by feeding a known allotment and weighing the uneaten
food. During the pretrial phase, measurements were taken
monthly for plasma total cholesterol and HDL-C concentrations. In addition, frequently sampled intravenous glucose
tolerance tests (FSIVGTT-Modified Minimal Model) were
done during the pretrial (baseline) to assess insulin sensitivity. Body composition was also measured at baseline employing CT scans and anthropometric measurements. Following the pretrial evaluations, all of the animals were
subjected to a baseline biopsy of the right femoral artery in
order to quantitate the levels of both early and advanced
glycated products. After the 6-month pretrial phase, animals
were randomized to continue the ad libitum diet or to
consume a caloric-restricted diet (30% reduction from baseline). The diets have been formulated so that cholesterol
intake per kg body weight is identical between groups, and
vitamin and mineral intake is approximately equal (see Diet
Composition).
Randomization into caloric-restricted or ad libitum-fed
groups. — After the 6-month pretrial evaluations, the animals were assigned to diet groups using stratified randomization, based on TPC/HDL-C ratio, age, and body weight
measurements obtained during the pretrial. The randomization was designed to balance the pretrial values of total
plasma cholesterol to high-density lipoprotein cholesterol
concentration ratio (TPC/HDL-C), previously shown to be a
major prognostic factor with the degree of atherosclerosis.
32
Cynos
Diet Composition
The nutritional objective of the study is to provide sufficient food for control groups to meet their estimated energy
requirements (National Research Council, 1978) based on
their age and individual mean body weight. The caloricrestricted diet was introduced over a 3-month transition
period (90% of ad libitum intake during the first month, 80%
during the second month, 70% during the third month and
thereafter). The dietary restriction was based on each animal's calculated ad libitum intake during the pretrial, and
calorie-restricted animals were fed 30% less based on calorie
intake/kg weight.
Additional vitamin mixture, mineral mixture, betasitosterol, and crystalline cholesterol were added to the
caloric restriction diet so that the same amount of these
components was fed whether the animals were fed 100
calories/kg body weight (ad libitum) or 70 calories/kg body
weight (caloric restriction). Less dextrin and sucrose were
added to the caloric restriction diet so the amount of calories
provided from carbohydrate, and as a result the caloric
density of protein and fat, would be the same in the two
diets. Less calcium carbonate was added to the caloric
restriction diet so the ratio of calcium:phosphorus would be
the same in the two diets. Less alphacel (non-nutritive bulk)
was added to the caloric restriction diet to accommodate the
additional amounts of other ingredients.
The composition of each diet is outlined in Table 1.
The potential for error exists in determining the ad libitum
food intake of nonhuman primates due to their untidy feeding behaviors in captivity, as bits of food may be spread
throughout the cage. The following procedures are being
implemented to alleviate this problem.
Diet is kept frozen and thawed right before each animal is
fed. Food is offered in a stainless steel tray mounted on the
outside of each cage. Each animal is separated from its cage
mate by a sliding partition which prevents animals from
transferring food to each other or retrieving dropped food
from cage mate. At 1000 hr, each animal is fed its daily
calculated allotment weighed for each animal. At 1500 hr,
all unconsumed food is removed, the amount of uneaten
food is determined, the partitions are removed, and the
animals are allowed to interact. The amount consumed by
each monkey is then determined.
Measurement of insulin resistance, blood and tissue (i.e.,
skin collagen) glycation/advanced glycation, computed to-
6 months
4 years
6 months
PRETRIAL
TRIAL DETERMINATIONS
ATHEROSCLEROSIS
EVALUATIONS
Pretrial
Determinations - *
Randomize:
• TPC/HDL-C
Ratio
- • 3 0 % Cal Restrict
Necropsy
• Control Diet
Figure 1. Schematic representing study design for caloric restriction trial.
Endpoinls
B12
CEFALU ET AL.
Table 1. Composition of the Ad Libitum (Control) and Caloric Restriction Diets
Ad Libitum Diet
Ingredient
GmPer
100 gm
Casein, USP
Lactalbumin
Dextrin
Sucrose
Wheat Flour, self-rising
Wheat Bran
Alphacel
Lard
Beef Tallow
Butter, lightly salted
Safflower Oil (linoleic)
Corn Oil
Olive Oil
Dried Egg Yolk
Crystalline Cholesterol
Complete Vitamin Mix
Ausman-Hayes Min. Mix
Calcium Carbonate
Beta-sitosterol
7.00
7.00
6.70
6.00
40.00
7.00
4.86
1.80
1.00
2.00
2.00
2.00
1.80
3.00
0.00
2.50
5.00
0.34
0.00
Total
100.00
Cholesterol (mg/Calorie)
Composition
Protein (% of Calories)
Lipid (% of Calories)
Carbo(% of Calories)
Grams
Lipid
Caloric Restriction
Mg
Choi
0.40
0.32
1.80
1.00
1.62
2.00
2.00
1.80
1.84
12.78
1.71
1.09
4.38
87.84
0.00
95.02
0.25
GmPer
100 gm
7.00
7.00
6.00
5.90
40.00
7.00
2.60
1.80
1.00
2.00
2.00
2.00
1.80
3.00
0.041
3.57
7.14
0.13
0.014
100.00
Grams
Lipid
Mg
Choi
0.40
0.32
1.80
1.00
1.71
1.09
1.62
2.00
2.00
1.80
1.84
4.38
12.78
87.84
41.00
136.02
0.356
20.50
29.80
49.70
Calcium/Phosphorus
1.24
Fatty Acids:
Saturated (%)
Monounsaturated (%)
Polyunsaturated (%)
% of fat
30.00
38.50
31.50
mography (CT) assessment of abdominal fat and lipid levels
are being measured at multiple intervals over a 4-year
period. The data collected are summarized in Table 2. Those
data will be related to coronary artery atherosclerotic plaque
size and complications after the 4-year period.
Physiologic Evaluations
Insulin sensitivity. — Insulin sensitivity was determined
by the frequently sampled intravenous glucose tolerance test
(FSIVGTT; Cobelli and Thomaseth, 1987) using a thirdphase insulin infusion (Modified Minimal Model) (Finegood
et al., 1990; Cefalu et al., 1995a). Determinations were
made during the pretrial and at 6-month intervals thereafter.
Monkeys were studied after an overnight fast. Two indwelling saphenous venous catheters were inserted after the
animals were anesthetized with ketamine HC1 (10 mg/kg
body weight). One catheter was used for blood sampling and
the other for glucose and insulin injections. Catheters were
kept patent by flushing with 1-2 ml of heparinized saline
between blood draws, and prior to the sampling, approx. 1-2
ml of blood was withdrawn to remove any residualized
heparinized saline. One milliliter blood samples were col-
20.50
29.80
49.70
1.23
% of cal
9.00
11.50
9.40
% of fat
30.10
38.50
31.40
% of cal
8.90
11.50
9.30
Table 2. Trial Evaluations
• Descriptive
Age
• Clinical
Body weight
Blood pressure
EKG
CT assessment of abdominal fat.
body measures
• Physiologic
Frequent sampling intravenous
glucose tolerance testing
(modified minimal model)
• Biochemical
Lipids — TPC, TG, LDL-C, HDL-C
Insulin
• Glycemic Indices
Tissue — Iliac artery glycation and
advanced glycation
Skin collagen — glycation and
advanced glycation
Blood — Glycated hemoglobin, fasting
glucose, serum fructosamine
lected at -10 and -1 minute, after which 0.5 g/kg glucose
was injected over 30 seconds, beginning at Time 0. After
glucose injection, blood samples (one ml) were taken at 2,4,
6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 28, 32, 40, 45, 50, 60,
70, 80, 90, 100, 110, 120, 140, 160, 180, 210 and 240
minutes post-glucose injection. After the 20-minute blood
draw, insulin (.015 units/kg BW) was injected. At each time
CR AND CARDIOVASCULAR AGING IN MONKEYS
point, glucose and insulin measurements were made. Glucose
was assayed by the glucose oxidase method on a glucose
Analyzer 2 (Beckman Instruments, Brea, CA). Insulin was
assayed by radioimmunoassay (Incstar Corp., Stillwater,
MN).
Clinical Evaluations
Adipose tissue distribution. — Computerized tomographic assessment of total abdominal, intra-abdominal, and
subcutaneous abdominal fat and body measures was done in
month -6 of the pretrial phase, month 0, and at 6-month
intervals throughout the study.
CT scan. — The monkeys were anesthetized with ketamine HCl (10 mg/kg administered intramuscularly [i.m.])
and acepromazine (0.1 mg/kg i.m.) prior to the CT scan.
The monkeys were positioned on their backs with legs
extended and arms placed adjacent to their trunks. A GE
CT9800 scanner was used for the procedure, and 120 kVp
and 100 MASA with a 2-second scan time were used as the
radiographic parameters. The CT density score used to
identify fat was previously determined and was found to be
-140 to -40 Hounsfield units, with the reference range from
-1000 (air) to + 1000 (dense bone) with zero being the
density of water (Laber-Laird etal., 1991). Intra-abdominal,
total, and subcutaneous fat was determined for a 1 cm scan
taken at the umbilicus. In a previous study, a single scan
taken at 1 cm levels 3 cm above or below the umbilicus was
representative of the entire abdominal fat distribution in this
species (Laber-Laird et al., 1991). Further, in human studies, intra-abdominal fat determined as a single scan at the
umbilicus was found to relate significantly to insulin resistance (Cefalu et al., 1995a). Therefore, a single scan at the
umbilicus was chosen because of the consistency and ease
with which it can be located. The quantity of fat per 1 cm
section is determined by counting the number of pixels in a
defined area within the specified density range. The area of
intra-abdominal fat is defined by outlining the transversalis
muscle with a cursor, and the area of total abdominal fat
defined by outlining the body margin. Subcutaneous abdominal fat is determined by subtracting the amount of intraabdominal fat from the amount of total abdominal fat.
Body measures. — A series of anthropometric measures
to characterize whole body and regional adiposity were
obtained. These include subscapular, triceps, and suprailiac
skinfold thickness using springloaded calipers (Shively et
al., 1987), waist and hip circumference, and body mass
index [BMI = body weight (kg)/trunk length (cm/100)2].
Blood pressure!heart rate. — Blood pressure and heart
rate were measured during the pretrial phase and at 6-month
intervals thereafter. Monkeys were sedated with 15 mg
ketamine HCl/kg body weight intramuscularly and placed on
their right side; the right arm was extended in a cephalad
direction until the upper arm was approximately at a right
angle to the vertical axis of the body. The upper arm
circumference was measured midway between the shoulder
and the elbow and the appropriate size of cuff used. Between
B13
8 and 18 minutes after sedation, at least three measurements
of systolic blood pressure, diastolic blood pressure, and
heart rate were taken. The average of these measurements
was used to determine the animal's blood pressure and heart
rate. Blood pressure measurements are made with a Dinamap Portable Adult/Pediatric and Neonatal Vital Signs Monitor (Model 8100), which uses the oscillometric technique to
measure systolic, diastolic, and mean arterial pressures and
heart rates noninvasively. This device computes and displays systolic, diastolic, and mean arterial pressures and
heart rate, eliminating subjective interpretation. The instrument is calibrated against a mercury manometer according to
the manufacturer's recommendation.
Electrocardiograms. — Electrocardiograms were obtained during month 6 of the pretrial phase and will be
obtained at study completion. A Hewlett-Packard electrocardiograph (Model 1500B) is used for cardiologic evaluation.
Electrocardiograms are recorded between 8 and 18 minutes
after ketamine HCl injection (15 mg ketamine HCl/kg body
weight, administered i.m.). The animals are placed in dorsal
recumbency and the standard six limb leads (I,II,III,AVR,
AVL, AVF) and three chest leads (V-l,V-4,V-6) recorded.
Biochemical Parameters
Total plasma cholesterol, triglycerides, and high density
lipoprotein cholesterol for all animals were measured during
months 4, 5, and 6 of the pretrial phase, and then every 3
months after starting the caloric-restriction or ad libitum
diets.
Cholesterol/triglyceride. — All specimens are analyzed
in the Lipid Analytic Laboratory. The laboratory is in compliance with the Cooperative Lipid Standardization Program. Cholesterol and triglyceride analyses are performed
using enzymatic methods on the Technicon RA-1000 analyzer. The methods for cholesterol (SM4-0139A85) and for
triglyceride (SM4- 0189K87, CPO Blank Method) are as
described by Technicon in their technical manual. The only
exception is the substitution in the cholesterol assay of the
BMD236691 reagent (Boehringer Mannheim) for the Technicon reagent. Our center is fully standardized with this
method and is in the continuing surveillance phase with
respect to the CDC Lipid Standardization Program.
HDL cholesterol. — For determination of HDL cholesterol concentration, we use the heparin-manganese precipitation procedure described in detail in the Manual of Laboratory Operations of the Lipid Research Clinics Program. The
only deviation from this procedure is that we utilize 2M
MnCl2 rather than 1M MnCl 2 , originally suggested for the
Lipid Research Clinics. The reason for this change is that we
have found that small amounts of LDL cholesterol fail to
precipitate in certain hyperlipoproteinemic monkeys and
human beings when 1M MnCl2 is used. This is completely
resolved with 2M MnCl2 without loss of HDL. For total and
HDL cholesterol determinations, the RA-1000 enzymatic
method is used with the Boehringer-Mannheim (BMD)
high-performance cholesterol reagent. The BMD's cholesterol reagent is used since it is made with Tris buffer which,
B14
CEFALUETAL.
unlike the phosphate-buffered reagents used by many other
manufacturers, does not cause positive interference in the
assay of HDL cholesterol when using heparin-manganese
precipitation.
Glycation Parameters
Soluble blood, skin, and arterial collagen. — Glycated
hemoglobin and fructosamine were determined at month 5
and 6 of the pretrial phase and every 3 months during the trial
phase. A femoral artery biopsy to determine levels of arterial
collagen glycation/advanced glycation was obtained in
month 6 of the pretrial period. A biopsy for skin collagen
glycation was obtained at month 6 of the pretrial period and
at 6-month intervals during the trial. The tissue samples
(skin and artery) are currently being stored at -70 °C and will
be batch analyzed at study completion.
Glycated hemoglobin. — Total glycated hemoglobin was
analyzed with automated affinity HPLC methodology (interassay c.v. = 1.2%, intraassay c.v. = 2.1%) performed
on a Primus CLC-330 HPLC (Primus Corporation, Kansas
City, MO) as previously reported (Cefalu et al., 1993,
1994). This method relies on boronate affinity columns to
non-enzymatically trap cis-diol groups of glycated proteins.
By virtue of measuring total glycation, the assay is free from
interference from abnormal hemoglobin as seen in hemoglobinopathies.
Total serum glycated proteins. — Total serum glycated
proteins were assessed with use of nitroblue colorimetric
methodology (2nd generation fructosamine assay) as determined on a Cobas Mira Chemistry Analyzer using Roche
reagents as described (Cefalu et al., 1991).
General Chemistries
Hematocrit, hemoglobin, white blood count, and indices
were determined on a model M430 Instrument (Coulter
Electrolosis, Hialeah, FL). Total protein/albumin were determined by the Cobas Mira Chemistry Analyzer using
Roche Reagents.
Statistical Methods
By adopting a randomized design, the issue of selection
bias in the allocation of subjects to intervention groups was
addressed. A double-blind design is preferred as it reduces
any bias in the reporting of subjective improvements and/or
occurrences of adverse effects as well as any investigator
bias in the collection of outcome measures. It was critical to
the trial conduct that persons conducting outcome evaluations were blinded to group assignments. All biochemical
assessments were made by technicians with no knowledge of
group assignments.
The effects of the CR diet on the trial evaluations measured at the specified intervals post-randomization were
determined by repeated measures analysis of covariance
(ANCOVA). Analysis of group differences was adjusted for
the prerandomization levels of the outcome measure being
tested in order to reduce the variance by taking into account
that part of the variance explained by prerandomization
predictors. All tests of hypotheses and reported p-values are
two-sided and unadjusted for multiple comparisons. Whenever a baseline value was used as a covariate in an ANCOVA
model, an interaction term between group and the covariate
was initially included to check the parallelism assumption.
If, and it was always the case, the interaction was not
significant at the .10 level of significance, the interaction
term was omitted and a standard parallel line ANCOVA was
used. Univariate histograms were plotted for each variable,
overall and by group using PROC UNIVARIATE with the
plot option in SAS. If moderate skewness was indicated, a
logarithmic transformation would have been used.
Estimates of intervention effects were obtained at each
follow-up observation. Tests of time of follow-up by intervention effects were conducted to test for consistency of
effects over the follow-up period. Average intervention
effects over the follow-up period were estimated and tested
for significance.
RESULTS
Table 3 demonstrates the clinical and biochemical measures determined for both the caloric-restricted and ad lib fed
groups during the pretrial assessments. As shown, there
were no significant differences noted in the baseline measures between groups. As randomization was based on
lipids, age, and body weight, the animals randomized to the
CR diet appeared to have less abdominal fat prior to initiating restriction. However, these differences in abdominal fat
were not statistically significant. All animals had normal
EKGs at baseline.
Caloric Intake
Ad lib caloric intake was assessed for 3 consecutive
months in the pretrial phase to get a baseline level for each
individual monkey. The amount of caloric restriction for
Table 3. Baseline Characteristics of Monkeys, Mean. ± SEM
Ad LibitumFed
(n = 16)
Age (years)
8 ± 1.1
Caloric
Restricted
(« = 16)
8.4± 1.2
Weight (kg)
5.6 ± .2
5.5 ± .2
Lipids (mg/dl)a
• Total cholesterol
• HDL cholesterol
• Triglycerides
283 ± 17
33 ± 3
21 ± 2
284 ± 14
35 ± 2
20 ± 1
Insulin Sensitivity ( x \iy
IxU'ml-')
2.9 ± .4
3.2± .4
Abdominal Fat (cm2)
• Total abdominal
• Intra-abdominal
• Subcutaneous
• Paraspinal
Blood Pressure (mmHg)
• Systolic
• Diastolic
• Mean arterial pressure
"Average of 3 values during pretrial.
3329
1775
1181
372
± 435
± 218
± 170
± 64
93 ± 3.6
53 ± 2.9
66 ± 3.1
2853
1502
1030
320
± 401
± 220
± 146
± 54
93 ± 14
52 ± 2.7
66 ± 2.9
CR AND CARDIOVASCULAR AGING IN MONKEYS
each animal randomized to the CR group was therefore
based on the individual animal's ad lib intake during the
pretrial. After randomization, the caloric-restricted diet was
introduced over a 3-month period, with 90% of ad lib diet
given in the first month, 80% in the second month, and 70%
in the third month and thereafter. As demonstrated in Figure
2A, compared to the ad lib-fed animals, there has been a
significant decrease in caloric intake for animals randomized
to the CR group during the first year of study (p < .001). At
the end of the first year, animals in the CR group have
averaged an ~ 34% decrease in dietary intake compared to
the ad lib-fed animals and an ~ 30% decrease when compared to their ad lib intake assessed during the pretrial.
Body Measures
Body weight. — As demonstrated in Figure 2B, body
weight was noted to decrease in the CR animals at the onset
of caloric restriction. There has been a significant diet effect
on body weight in the CR group when compared to the ad
lib-fed group (p < .05). CR animals have lost an average 0.3
kg of body weight since randomization, and at the end of the
first year there is an 8% difference in body weight between
animals.
Body composition. — Figures 3A and 3B demonstrate the
change in total abdominal fat (TAF) and intra-abdominal fat
(IAF) assessed for both diet groups with CT scanning. As
demonstrated, monkeys in both groups had an increase in
TAF and IAF over the 6-month pretrial preceding randomization. Since randomization, there has been a significant
diet effect on both TAF and IAF as demonstrated. No
significant diet effect was seen for paraspinal fat, and the
change in subcutaneous fat approached significance at the
end of the first year (p = .07). A significant diet effect (p <
.05) was observed for both waist circumference and BMI at
the end of the first year (data not shown).
Lipid Measurements
Figures 4A and 4B demonstrate the total plasma cholesterol and HDL-C levels observed during the first year of
study. No diet effect was noted for these lipid parameters.
Further, there was no diet effect noted for TPC: HDL-C ratio
or triglyceride levels.
Carbohydrate Metabolism
Table 4 demonstrates parameters assessing carbohydrate
metabolism during the study. There has been no significant
change in levels for glycated hemoglobin or fructosamine in
A- TOTAL ABDOMINAL FAT
A -DAILY DIETARY INTAKE
_ 220
(0
E 200
I Fat (cm'
4800
3
^
180
™ 160
C
"" 140
ital i
™ 120
0>
* p < .05
" p < .001
CR Initiated
100
8
-2
10
1
1800"
CR Initiated
ADLIB
CAL RES
2600
U) 6.26.0-
2200
5.8-
5.6
5.4O
5.2-
0
2
,
4
6
p < .05
8 10 12 14
B-INTRA-ABDOMINAL FAT
6.4-
m
I
I
"
I
f»on-8-6-4-2
12
B- BODY WEIGHTS
•D
1
T
o
80
-4
3800-
o 2800
i.
Q
B15
n
1800
U.
•o
XI
<
1400
ADLIB
CAL RES
g
5.0-
C
1000
Diet effect, *p < .05
4.8
CR
Initiated
* p < .001 vs ad lib
4.6"
-4
-
2
0
2
4
6
8
10
12
Months of Trial
Figure 2. Demonstrates daily dietary intake (A) and body weight (B) in
CR and control fed groups. As demonstrated, a significant reduction in
dietary intake has been observed for CR animals. Data are mean ± SE.
600
-8-6-4-2
0
2
4
6
8
10 12 14
Months of Trial
Figure 3. Total abdominal (A) and intra-abdominal fat mass (B) assessed
with a single CT scan at the umbilicus. A significant diet effect has been
noted for both parameters. Data are mean ± SE.
CEFALUETAL.
B16
either dietary group. Table 4 also demonstrates the measurements of fasting glucose and insulin in both dietary groups.
Although an increase in insulin for the ad lib-fed animals and
a decrease in fasting glucose in the CR animals was noted at
the end of 12 months, these changes were not statistically
significant. Figure 5 demonstrates the insulin sensitivity
parameter obtained with the modified minimal model technique. There has been a significant diet effect to improve
insulin sensitivity in the CR animals. Compared to baseline,
an increase in insulin sensitivity of 50% has occurred. No
significant diet effect was noted for glucose effectiveness
(data not shown).
550"
A-TOTAL PLASMA CHOLESTEROL
ig/dl)
450
h
350"
TPC
500-
250
400"
300"
200150
mo8
10
12
14
General Health/Cardiovascular Measures
Despite a significant reduction in calories, the CR animals
are maintaining hematologic values (hemoglobin, hematocrit, white blood count), total protein, albumin levels, BUN,
and creatinine levels that are no different from the ad lib
group (data not shown). No difference in blood pressure or
mean arterial pressure has been noted.
DISCUSSION
The long-term goal of this study is to assess the independent effect of chronic caloric restriction on a specific measure of atherosclerosis, i.e., coronary artery plaque extent,
in a nonhuman primate model. Caloric restriction is postulated to alter pathogenic mechanisms contributing to atherosclerosis. Therefore, in this ongoing trial, we are evaluating
the role of chronic caloric restriction in modifying cardiovascular risk factors (e.g., insulin resistance, adipose tissue
distribution) and glycation/advanced glycation (for both
blood and tissue proteins), all of which have been strongly
implicated in contributing to cardiovascular disease. In addition, we propose to evaluate the relationship of these
changes to changes observed in arterial plaque/size complications determined at study completion.
The focus of this article is to present the study design,
methodology, and results from the first full year of caloric
restriction for this ongoing trial. The results from the first
year suggest that after one year of a significant 30% reduction in total caloric intake for the calorie-restricted animals
ADLIB
CAL RES
B- HDL-CHOLESTEROL
45
INSULIN SENSITIVITY
10
CO
40-
9
8
O)
E,
35
ADLIB
CAL RES
7
6"
30
O
c
25
Q
Z
-o—
-•—
CO
5"
43-
20
2
15
* p < .05
" p = .005
1
10
-2
0
2
4
6
8
10
12
-2
14
2
4
6
8
10
12
14
Months of Trial
Months of Trial
Figure 4. Total plasma cholesterol (TPC) (A) and HDL-Cholesterol (B)
for both dietary groups. Value at month 0 is average of 3 values obtained
during pretrial. Calorie restriction was initiated at month 0. No significant
diet effect has been observed. Data are mean ± SE.
Figure 5. Demonstrates insulin sensitivity for both dietary groups. As
shown, a significant increase in insulin sensitivity has been observed for CR
animals. S, units = ( x lO^min'fjLLHml). Data are mean ± SE.
Table 4. Carbohydrate Measures, Mean ± SD
Baseline"
AdLib
Glycated hemoglobin (%)
Fructosamine (|xMol/L)
Glucose (mg/dl)
Insulin (u.U/ml)
•Average of 3 monthly values during pretrial.
3.5
153
64
36
± .6
± 14
± 5.2
± 18
12 months
Restricted
3.5
161
66
34
± .7
± 17
± 8
± 13
AdLib
3.4
145
64
54
± .6
± 20
± 12
± 28
Restricted
3.3
151
59
35
± .6
± 19
± 10
± 16
CR AND CARDIOVASCULAR AGING IN MONKEYS
compared to the ad lib fed group: (a) a significant decrease in
weight in the caloric restriction group has occurred accompanied by a significant decrease in central adiposity as
assessed by specific determination of abdominal fat depots
with CT scans and by anthropometric measurements; (b)
successful maintenance of total plasma cholesterol/HDL
cholesterol concentrations has occurred despite a reduction
in calories of 30% in the CR group; (c) no detrimental effects
of CR on general chemistry profiles or cardiovascular measurements have been observed; and (d) a significant improvement in insulin sensitivity has been observed, while no
significant change has been observed for glycated proteins.
We are aware of two other ongoing trials of caloric
restriction in aging nonhuman primates (Ingram et al., 1990;
Kemnitz et al., 1993), plus additional studies that have
evaluated weight maintenance in delaying the onset to diabetes in rhesus monkeys (Hansen and Bodkin, 1993). These
trials have provided significant knowledge regarding the
effects and validity of caloric restriction in primates, and as
they are ongoing, will continue to provide important information regarding long-term maintenance of this nutritional
regimen. Our study shares similarities with the other trials,
as reported observations from the other nonhuman primate
trials have greatly influenced our study design. Since 1987,
investigators at the National Institute on Aging (NIA) have
evaluated the effect of caloric restriction in nonhuman primates at several ages across the life span. This study contains four major cohorts of animals, two male rhesus groups,
one cohort of male squirrel monkeys, and a group of female
rhesus monkeys. Findings from this cohort have been published in several reports (Ingram et al., 1990, 1993; Lane et
al., 1995a, 1995b). A second study at the University of
Wisconsin (UW) was initiated in 1989 and is designed to
evaluate the effect of adult-onset CR in male rhesus monkeys (Kemnitz et al., 1993). In a third study, Hansen and
Bodkin (1993) at the University of Maryland (UM) have
compared calorically restricted weight-stabilized adult rhesus monkeys to ad-lib fed monkeys and evaluated factors
related to the development of diabetes.
Our ongoing trial differs from the above studies primarily
as we are using cynomolgus monkeys (Macaca fascicularis), whereas the other three sites are evaluating either
rhesus (Macaca mulatta) (NIA, UW, UM sites) or squirrel
monkeys (Saimiri sciureus) (NIA site). Our choice for cynomolgus monkeys derives from the extensive literature demonstrating their validity as an excellent model for human
atherosclerosis (Rudel and Pitts, 1978; Adams et al., 1983;
Hamm et al., 1983). Further, there appear to be differences
in susceptibility to atherosclerosis related to physiological
aspects of aging [i.e., development (puberty) and surgical
menopause] in these primates (Adams et al., 1983;
Weingand et al., 1986). As such, this study is unique in that
it is the first study evaluating chronic caloric restriction in a
higher species on atherosclerosis. In order to independently
evaluate the effect of caloric restriction on atherosclerosis, it
is imperative that the study be designed to significantly
reduce calories, yet not to alter the cholesterol content of the
diet, as that will alter a major prognostic factor for atherosclerosis, i.e., total plasma cholesterol:HDL-C ratio. This
was achieved by supplementing the caloric-restricted diet
B17
with crystalline cholesterol (see diet composition). As such,
cholesterol intake was similar between groups and has resulted in similar lipid profiles (Figure 4A and 4B). Therefore, this particular study design will allow the independent
effect of caloric restriction on atherosclerosis to be evaluated
as we have achieved similar plasma cholesterol levels between groups.
Another major difference from the other CR trials is the age
of dietary initiation. In the NIA study, caloric restriction was
initiated in juvenile rhesus (age .6-1 yr) and squirrel monkeys (age 1-4 yr), young adult rhesus (age 3-5 yr) and adult
squirrel monkeys (age 5-10 yr), and old rhesus (> 18-25 yr)
and old squirrel monkeys (>10 yr). In the UW study, fully
adult rhesus monkeys were studied as the age range at
initiation was 8-14 years. The University of Maryland study
reported results in older rhesus monkeys with an average age
of 17.9 for the weight-stabilized group and 18.1 for the
control group after 5-9 years of diet titration. However, our
study endpoint of atherosclerosis has influenced the age at
which we initiated caloric restriction. The age range of 6-9
years was chosen based on evidence from our institution that
demonstrated that juvenile (2.5-3.5 years) and adult (6-12
years) monkeys that had consumed comparable atherogenic
diets resulting in similar plasma lipid concentrations showed
marked differences in coronary artery lesion characteristics
and extent (Weingand et al., 1986). Adult animals (n = 16)
developed more extensive lesions characterized predominantly as proliferative atherosclerotic plaques, while juveniles (n = 10) had minimal lesions that consisted primarily of
fatty streaks. These differences in lesion characteristics and
extent were not explained by differences in blood pressure. It
was uncertain if these quantitative and qualitative differences
in susceptibility to diet-induced atherosclerosis were due to
intrinsic age-related changes in the arterial wall and/or related
to the endocrine and metabolic changes associated with puberty. This observation has been referred to as "juvenile
protection" and has influenced our decision about the age
range at study initiation.
The diet and feeding strategy has also varied considerably
between studies and resulted in different observations for
body weight for the first year of study. For example, in the
NIA study, the amount of food given was based on body
weight and National Research Council (NRC)-estimated
energy requirements. It was reported that food intake assessments revealed an actual reduction of 22-24% (Ingram et
al., 1990) rather than the 30% sought on the basis of
allotment. Nevertheless, the DR animals had a substantial
reduction in body weight gain. In contrast, investigators at
the University of Wisconsin (Kemnitz et al., 1993) determined ad libitum intakes for each animal during the baseline
period, and caloric-restricted intakes were individually determined. They reported that actual intakes in the control
animals over the first year decreased, and they assessed that
the original baseline intakes were overestimates because of a
transient increase in intake due to the greater palatability of
the defined diet. As such, the average intake of DR animals
represented a reduction of 38% from their average baseline
intake, but a reduction of only 18% from the average intake
of controls. Despite this, and in contrast to the NIA study,
body weights in the UW study actually declined over the first
B18
CEFALUETAL.
12 months. On an average, the ad lib-fed monkeys gained a
kg of weight (an increase of approx. 9% in BW) from
baseline, while the DR animals lost approx. 0.5kg (approx.
4.5% loss in body weight). In the UM study (Hansen and
Bodkin, 1993), animals were weighed weekly and caloric
intake was adjusted based on any prior change in body
weight to maintain a stable weight in each animal. Our study
is similar to the UW study in that the amount of caloric intake
calculated was based on ad libitum intake for each animal
determined in the pretrial phase. The 30% reduction in
calories for animals randomized to the CR diet was therefore
based on the ad lib diet determined for each individual
animal. The CR animals had daily caloric intake at consistently 30% less than the calculated ad lib intake for each
animal, and averaged 34% less than the control group.
Therefore, the level of dietary restriction achieved for the
first full year of our study is greater than that reported for the
other primate trials.
The NIA study uses both anthropometric measures and
DEXA scans to monitor body fat content as does the Wisconsin group. Our trial differs in that we assessed abdominal
obesity with CT scans allowing separation into intraabdominal, subcutaneous, and paraspinous fat. Our data
suggest not only a decrease in total abdominal fat mass, but a
reduction in the fat depot that has been most significantly
linked to cardiovascular disease and cardiovascular risk
factors in both human and nonhuman primates, i.e., intraabdominal fat mass (Kissebah et al., 1988; Shively and
Clarkson, 1988; Despres et al., 1989). It is of interest in our
study that the intra-abdominal fat mass showed the most
significant change secondary to caloric restriction, p <
.001), whereas changes in subcutaneous abdominal fat mass
approached significance (p = .07) in the CR animals and the
paraspinous fat mass did not appear to be affected by diet.
Further, this finding is of great interest from a gerontologic
perspective as we have recently reported that intraabdominal fat mass accumulates with age and it is the
abdominal fat depot that can most readily explain the variance in insulin resistance with age (Cefalu et al., 1995a).
The safety of caloric restriction was demonstrated in our
study as has been demonstrated in the previous trials. Although we were successful in significantly reducing calories
by over 30% in the CR group, there has been no detrimental
effect on hematocrit, hemoglobin, total protein, albumin, or
renal function. The results from the first full year also failed
to demonstrate any changes in blood pressure between
groups.
Our results failed to show a significant improvement in
glycated proteins between groups. These data agree with the
results obtained from both the NIA study (Cutler et al.,
1992) and University of Wisconsin study (Kemnitz et al.,
1994), although both trials have demonstrated improvements in carbohydrate metabolism. If lessons can be learned
from the rodent model, we would expect decreases in early
glycated products as measured on short-lived blood proteins
be observed first (i.e., glycated hemoglobin and serum
fructosamine), and would ultimately lead to reduced substrate available for advanced glycation as measured in longlived tissue proteins. Such was the pattern we recently
reported in a CR rodent (Cefalu et al., 1995b). However, a
more recent article by Lane et al. (1995a) reported no change
in glycated hemoglobin after 7 years of CR. The reasons as
to why no change in glycated proteins has been seen in the
reported long-term trials while an improvement in insulin
sensitivity has occurred are not currently known.
Finally, our most significant finding is that insulin sensitivity was increased in our animals after 1 year of caloric
restriction. Although there was no significant difference at
baseline between dietary groups, there was a significant diet
effect on insulin sensitivity. Our finding of improved insulin
sensitivity in chronic caloric restriction states confirms the
observations first reported by the University of Wisconsin's
investigators (Kemnitz et al., 1994), and indeed our methods
for assessing insulin sensitivity were similar. As such, although the UW investigators did not show a significant
improvement in insulin sensitivity after the first year, it may
have been that the percent of restriction was milder then
intended. After the decision was made to adjust the allotment
provided to the restricted animals at 18 months, the investigators demonstrated a significant increase in insulin sensitivity at 6 and 12 months later (i.e., 24 and 30 months of trial).
Therefore, as demonstrated first by the University of Wisconsin group and confirmed by investigators at the NIA
(Lane et al., 1995a), insulin sensitivity can be significantly
improved in nonhuman primates secondary to chronic CR.
The mechanism by which this occurs is currently not known
but will be the focus of future investigations.
In conclusion, the results from the first year of study have
demonstrated that a 30% reduction in calories can be safely
achieved while maintaining plasma lipid levels. Further, a
significant decrease in body weight was observed in caloricrestricted animals associated with a significant decrease in
intra-abdominal fat. We observed a significant improvement
in insulin sensitivity at one year, but no significant change in
glycated proteins. As observed with other nonhuman primate trials, a more prolonged period of observation will be
necessary to determine if significant improvement in glycated products and maintenance of improved insulin sensitivity
can be demonstrated, as well as if these changes are associated with diminished progression of atherosclerosis.
ACKNOWLEDGMENTS
This work was supported by grants R29 AG10816 and K01 AGOO578
from the National Institutes of Health to Dr. Cefalu.
Address correspondence to Dr. William T. Cefalu, Department of Internal Medicine, Bowman Gray School of Medicine, Wake Forest University,
Winston-Salem, NC 29157-1047.
REFERENCES
Adams, M.R.; Clarkson, T.B.; Kaplan, J.R.; Koritnik, D.R. Ovariectomy,
social stress and coronary artery atherosclerosis in cynomolgus monkeys. Circulation 68(suppl III):III-188; 1983.
Cefalu, W.T.; Bell-Farrow, A.D.; Petty, M.; Izlar, C ; Smith, J.A. Clinical
validation of a new, 2nd generation fructosamine assay. Clin. Chem.
37:1252-1256; 1991.
Cefalu, W.T.; Bell-Farrow, A.D.; Wagner, J. Role of glycated proteins in
the detection and monitoring of diabetes in cynomolgus monkeys. J.
Lab. Anim. Sci. 43:73-77; 1993.
Cefalu, W.T.; Wang, Z.Q.; Bell-Farrow, A.D.; Kiger, F.D.; Izlar, C.
Glycohemoglobin measured by automated affinity HPLC correlates
with both short-term and long-term antecedent glycemia. Clin. Chem.
40:1317-1321; 1994.
CR AND CARDIOVASCULAR AGING IN MONKEYS
Cefalu, W.T.; Wang, Z.Q.; Werbel, S.S.; Bell-Farrow, A.; Crouse, J.R.;
Hinson, W.H.; Terry, J.G.; Anderson, R. Contribution of visceral fat
mass to the insulin resistance of aging. Metabolism 44:954—959; 1995a.
Cefalu, W.T.; Bell-Farrow, A.D.; Wang, A.Q.; Sonntag, W.; Fu, M-X.;
Baynes, J.W.; Thorpe, S.R. Caloric restriction (CR) decreases agedependent accumulation of the glycoxidation products N' (Carboxymethyl)lysine (CML) and pentosidine (P) in rat skin. J. Gerontol. Biol.
Sci. 50A:B337-B341; 1995b.
Cobelli, C ; Thomaseth, K. The minimal model of glucose disappearance:optimal input studies. Mathematical Biosci. 83:127-155; 1987.
Cutler, R.G.; Davis, B.J.; Ingram, D.K.; Roth, G.S. Plasma concentrations of glucose, insulin, and percent glycosylated hemoglobin are
unaltered by food restriction in rhesus and squirrel monkeys. J. Gerontol. Biol. Sci. 47:B9-B12; 1992.
Despres, J-P.; Moorjani, S.; Ferland, M.; Tremblay, A.; Lupien, F.J.;
Nadeau, A.; Pinault, S.; Theriault, G.; Bouchard, C. Adipose tissue
distribution and plasma lipoprotein levels in obese women. Importance
of intra-abdominal fat. Arteriosclerosis 9:203-210; 1989.
Finegood, D.T.; Hramiak, I.M.; Dupre, J. A modified protocol for estimation of insulin sensitivity with the minimal model of glucose kinetics in
patients with insulin-dependent diabetes. J. Clin. Endocrinol. Metab.
70:1538-1549; 1990.
Hamm, T.E., Jr.; Kaplan, J.; Clarkson, T.B.; Bullock, B.C. Effects of
gender and social behavior on the development of coronary artery
atherosclerosis in cynomolgus monkeys. Atherosclerosis 48:221-233;
1983.
Hansen, B.C.; Bodkin, N.L. Primary prevention of diabetes mellitus by
prevention of obesity in monkeys. Diabetes 42:1809-1814; 1993.
Ingram, D.K.; Cutler, R.G.; Weindruch, R.; Renquist, D.M.; Knapka,
J.J.; April, M.; Belcher, C.T.; Clark, M.A.; Hatcherson, C D . ; Marriott, B.M.; Roth, G.S. Dietary restriction and aging: the initiation of a
primate study. J. Gerontol. Biol. Sci. 45:B148-B163; 1990.
Ingram, D.K.; Lane, M.A.; Cutler, R.G.; Roth, G.S. Longitudinal study of
aging in monkeys: effects of diet restriction. Neurobiol. Aging 14:687688; 1993.
Kemnitz, J.E.; Weindruch, R.; Roecker, E.B.; Crawford, K.; Kaufman,
P.L.; Ershler, W.B. Dietary restriction of adult male rhesus monkeys:
design, methodology, and preliminary findings from the first year of
study. J. Gerontol. Biol. Sci. 48:B17-B26; 1993.
Kemnitz, J.W.; Roecker, E.B.; Weindruch, R.; Elson, D.F.; Baum, S.T.;
Bergman, R.N. Dietary restriction increases insulin sensitivity and
B19
lowers blood glucose in rhesus monkeys. Am. J. Physiol. 266:E540E547; 1994.
Kissebah, A H . ; Peiris, A.N.; Evans, D.J. Mechanisms associating body
fat distribution to glucose intolerance and diabetes mellitus: window
with a view. Acta Med. Scand. 723(Suppl):79-89; 1988.
Laber-Laird, K.; Shively, C.A.; Karstaedt, N.; Bullock, R.C. Assessment
of abdominal fat deposition in female cynomolgus monkeys. Int. J.
Obesity 15:213-220; 1991.
Lane, M.A.; Ingram, D.K.; Cutler, R.G.; Knapka, J.J.; Barnard, D.E.;
Roth, G.S. Dietary restriction in nonhuman primates: progress report on
the NIA study. Ann. NY Acad. Sci. 673:36-45; 1992.
Lane, M.A.; Ball, S.S.; Ingram, D.K.; Cutler, R.G.; Engel, J.; Read, V.;
Roth, G.S. Diet restriction in rhesus monkeys lowers fasting and
glucose-stimulated glucoregulatory endpoints. Am. J. Phys. 268:941948; 1995a.
Lane, M.A.; Baer, D.J.; Tilmont, E.M.; Rumpler, W.V.; Ingram, D.K.;
Roth, G.S.; Cutler, R.G. Energy balance in rhesus monkeys (Macaca
mulatto) subjected to long-term dietary restriction. J. Gerontol. Biol.
Sci. 50A:B295-B302; 1995b.
National Research Council. Nutrient requirements of nonhuman primates.
Washington, DC: National Academy of Sciences, 1978.
Rudel, L.L.; Pitts, L.L. II. Males-female variability in the dietary
cholesterol-induced hyperlipoproteinemia of cynomolgus monkeys
(Macaca fascicularis) J. Lipid Res. 19:992-1003; 1978.
Shively, C.A.; Clarkson, T.B. Regional obesity and coronary artery atherosclerosis in females: a nonhuman primate model. Acta Med. Scand.
723(Suppl):71-78; 1988.
Shively, C.A.; Clarkson, T.B.C.; Miller, L.C.; Weingand, K.W. Body fat
distribution as a risk factor for coronary artery atherosclerosis in female
cynomolgus monkeys. Arteriosclerosis 7:226-231; 1987.
Snyder, D.L., ed. Dietary restriction and aging. New York: Alan R. Liss,
1989.
Weindruch, R.; Walford, R.L. The retardation of aging and disease by
dietary restriction. Springfield, 1L: Charles C Thomas, 1988.
Weingand, K.W.; Clarkson, T.B.; Adams, M.R.; Bostrom, A.D. Effects
of age and/or puberty on coronary artery atherosclerosis in cynomolgus
monkeys. Atherosclerosis 62:137-144; 1986.
Received February 26, 1996
Accepted April 22, 1996