Serum glucose, glucose tolerance, corticosterone and free fatty acids during aging in energy restricted mice

ELSEVIER SCIENCE IRELAND

Mechanisms of Ageing and Development 73 (1994) 209-221

Serum glucose, glucose tolerance, corticosterone and free fatty acids during aging in energy restricted mice Steven B. Harris a, Mark W. Gunion b, Mark J. Rosenthal c R o y L. W a l f o r d a aDepartment of Pathology, Center for the Health Sciences, University of California at Los Angeles, Los Angeles, CA 90024, USA bSepulveda Veterans Administration Hospital, B36, Building 5, 16111 Plummet, Sepulveda, CA 91343, USA "Research Service 151, Sepulveda Veterans Administration Hospital, 16111 Plummer, Sepulveda, CA 91343, USA

(Received 23 March; accepted 26 July 1993)

Abstract

Energy restriction, the only method known to increase maximum life span in laboratory animals, was used as a tool to test hypotheses regarding possible mechanisms of aging. Serum glucose and corticosterone (CS) concentrations in mice of a long-lived hybrid mouse strain, aged 7, 17, and 29 months, and on 50%, 80%, and 100% of ad libitum intake, were measured. Serum glucose and CS concentrations were also measured in response to intraperitoneal (i.p.) glucose challenge in mice at ages 7 and 29 months. Serum glucose and CS concentrations were also measured at several time points over 36 h, to assess their diurnal variation. There were no differences in single fasting glucose concentrations in 7- and 29omonth-old mice at the same degree of energy restriction, but energy restriction decreased glucose concentrations. Serum CS concentrations were generally increased restricted animals with respect to fully fed ones. Average serum glucose concentrations were found to be significantly decreased by dietary restriction. Glucose tolerance curves were unchanged by age in ad libitum fed or 50% restricted animals, but in 80% ad libitum groups, older animals showed evidence of decreased glucose tolerance with respect to young animals. For each age, peak serum glucose concentrations after i.p. glucose loading varied with degree of energy restriction, with more severely restricted animals showing less glucose tolerance. Average serum CS concentrations were elevated at 7 months by restriction, especially at night and long after feeding, but we found no differences with age or diet in average CS concentrations. Our serum glucose results sup* Corresponding author. 0047-6374/94/$07.00 © 1994 Elsevier Science Ireland Ltd. All rights reserved. SSDI 0047-6374(94)01391-K

210

S.B. Harris et aL / Mech. Ageing Dev. 73 (1994) 209-221

port the hypothesisthat nonenzymaticglycation is mechanisticallyinvolved in normal aging. Our serum CS results do not support the hypothesis that CS contributes significantlyto the pathophysiology of normal aging in mice. Key words:

Aging; Energy restriction; Dietary restriction; Glucose; Corticosterone

1. Introduction

Dietary restriction, and specifically energy restriction, has proven the only effective method of extending maximum life span in laboratory animals, and in delaying onset of many age-related pathologies and pathologic changes [1,2]. The mechanism(s) whereby energy restriction exerts these global effects is unknown but the overall findings suggest that dietary restriction may modify underlying process(es) in part responsible for aging [3]. Recently, two hypotheses concerning a proposed endocrinological basis of aging have been advanced which are amenable to investigation with chronic energy restriction models. Sapolski et al. [41 have noted that aged rodents hypersecrete corticosterone (CS) in response to stress, and also that glucocorticoids such as CS may damage brain centers involved in suppressing glucocorticoid release, thereby perhaps potentiating a cycle of damage. A number of common features-of glucocorticoid excess and aging have led these investigators to propose that a senescent cascade of glucocorticoid hypersecretion causes some of the changes of aging. Several hormonal control systems in the brain appear to be possibly susceptible to runaway feedback from hormone damage, and a neurohumoral hysteresis theory of aging has also been proposed, involving estrogen, glucose, and CS [5]. A hypothesis of Cerami [6] suggests that glucose is an important toxin in the body, and that long term damage to DNA and protein from nonenzymatic glycation may be responsible for some of the aspects of aging. In this theory, the rate of such damaging glycation would be directly dependent on concentrations of glucose in body fluids. In order to test some of the endocrine-based hypotheses about aging, we examined how certain metabolic and endocrine factors in rodents responded to a lifetime of chronic energy restriction. The present study measured basal serum concentrations of glucose and CS, as well as their response to glucose challenge, in severely restricted, mildly restricted, and ad libitum fed young and old mice. 2. Methods 2.1. M i c e

The long-lived C3BIORF l hybrid strain of mouse has been used in previous dietary restriction studies in this laboratory [8,9]. These mice are bred from C57BL10.R111 and C3H.Sw/Sn. lines obtained originally from Jackson Laboratories, Bar Harbor, Maine. Female hybrid progeny were weaned at 21-28 days of age, individually caged in plastic cages on wood chip bedding, and assigned to one of three dietary regimens. Three age cohorts were studied, mice 7-, 17- and 29-month-

S.B. Harris et al./Mech. Ageing Dev. 73 (1994) 209-221

211

old at the beginning of the study. Mice in each of three age cohorts, born within a 6-week period, were maintained under conventional (non-barrier) conditions, with temperature (20-24°C), humidity (50-60%) and 12 h. (06:00-18:00 h) lighting constant throughout the study. Survival curves for these mice under these conditions have been shown to be well rectangularized [8,9]. To monitor for infections, sentinel mice were kept in the same room as experimental mice, and sentinel serum samples were screened every 6 months for antibody titers against 11 common pathogens. Positive titers were not found during this study. 2.2. Diets C3H.SW/Sn dams were fed Purina Laboratory Chow TM (Ralston Purina, St. Louis, MO) during gestation and nursing of the hybrid study animals. Based on previous life-extending dietary restriction studies in our laboratory [8,9], mice at weaning were assigned to one of three diet groups. (1) The fully fed (ad libitum fed) group ate diet C (specified in Ref. 9), a 20% casein-purified diet. These mice were fed 10.0, 10.0 and 15.0 g of diet C given Monday, Wednesday, and Friday, respectively. This-represented a total feeding of 126 kcal/week (526 kJ/week), but food was not completely consumed, and on inspection some food was always present in the cages of these animals. These mice were obese by inspection after about 6 months of age. (2) The mildly restricted group was restricted to approximately 80% of the ad libitum intake, which prevented obesity. They received 86 kcal/week (360 kJ/week), given as 3.5 g daily on Monday through Thursday, and a 10.0-g feeding on Friday. (3) The severely restricted group ate restricted diet R, a 35% casein, vitamin and mineral enriched purified diet [9]. They received 49 kcal/week, given as 3.5 g Monday and Wednesday, and 7.0 g on Friday). The compositions of diets C and R were chosen so that mildly and severely energy restricted groups ate very nearly the same weekly amounts of protein, fat, vitamins, and minerals, with restricted animals differing only in carbohydrate (energy), and fiber intake [9]. Fully fed animals ate greater quantities of all food components. All feedings were done in the morning at 08:00 h, and no food was routinely visible in the cages of either mildly or severely restricted animals within a few hours after feeding. 2.3. Study 1: Baseline serum glucose and corticosterone, 24 h after feeding Nine groups of eight animals were studied, consisting of cohorts of three different ages (7, 17, and 29 months) from each of the three lifetime diet groups (Fully Fed, Mildly Restricted, and Severely Restricted). Blood was collected from 10:00 to 12:00 h, 26-28 h from last feeding in all animals. To minimize animal stress and its possible attendant CS and glucose increases, blood was collected in the room housing the animals, but in an adjacent room area. Each animal was not disturbed until the time of collection, then in turn each cage was removed from its place and the animal rapidly transferred directly from cage to an airtight jar containing anaesthetic agent (enflurane, USP, Anaquest/BOC Health Care, Madison, WI). The animal was removed from anaesthesia immediately after losing the righting reflex (20-30 s), and bled from the retro-orbital venous plexus by means of a 339-/zl capacity heparinized capillary tube (350 mm x 1.08 mm, Monoject/Sherwood Medical, St. Louis, MO). Total time from removal of an animal cage from the rack to completion of collection

212

S.B. Harris et al./ Mech. Ageing Dev. 73 (1994) 209-221

of blood was typically less than 60 s. To avoid disturbing still-to-be-bled animals, animals which had been bled were not returned to the racks until completion of the experiment. Previous pilot studies demonstrated no significant differences in serum CS and glucose concentrations between suddenly decapitated animals and animals anesthetized in this manner (data not shown), nor were significant differences seen in serum concentrations of glucose or CS in animals bled at the beginning of the series for each day, and those bled at the end. Animals were weighed after the drawing of blood. Collected blood was pipetted immediately into plastic microcentrifuge tubes on ice, allowed to clot for several hours, then microcentrifuged for 5 min and the serum removed for storage at -79°C until assay. For each type of serum assay done in this study and those below (glucose, insulin, CS, free fatty acid), all samples from each study were assayed together in a single assay run, in duplicate, and the assay repeated for any sample in which duplicate values failed to agree within 10%.

2.4. Study 2: Glucose tolerance testing and CS stress response For this study 7- to 8- and 29- to 30-month-old animals were used from-the Fully Fed, Mildly Restricted, and Severely Restricted groups. Each of the resulting six groups consisted of a pool of 12-20 animals, from which animals were selected sequentially to be bled on a rotating basis, as often as once a week. Animals-were bled (as described above) from 10:00 to 12:00 h, all groups having been fed the previous morning. Animals were briefly removed from their cages, weighed, and either bled immediately (for the zero time measurement) or else given a glucose load of 1.5 mg glucose per gram body weight by. i.p. injection of a 150 g/1 glucose solution. After injection, animals were returned to their cages for a specified time (30, 60, or 120 rain) before being bled. Thus, the glucose tolerance curve measurement at each time point represents the mean of serum concentration values for six to eight animals which had each been bled only once on a given day. To minimize daily and procedural sources of variance, one animal to represent each of the four time points (0, 30, 60, or 120 min) was selected from each of the six age/diet groups on each testing day, for a total of 24 animals and samples taken each day. In addition, an injection and bleeding schedule was chosen to thoroughly mix animals from the six groups and four test time points on each testing day. For Study 2, collected blood was pipetted immediately into small plastic microcentrifuge tubes holding 5 ~tl aproteinin 1 mg/ml (Sigma, St. Louis, MO) vortexed, and allowed to clot on ice. After all blood had been collected, serum was collected as previously described and stored until assay. 2.5. Study 3." Diurnal variation of glucose, corticosterone and jhee fatty acids This study followed Study 2 by approximately a month, and these measurements were conducted on the same pools of young (8-month) and old (30-month) fully fed, mildly restricted, and severely restricted animals used in study 2. Animals were fed normally on their diets, and blood and serum was collected as in Study 2, from six animals in each of the six groups, for concentrations of glucose, CS, and free fatty acids, at 6-h intervals for 36 h, beginning after a feeding. As in Study 2, all animals in each pool were bled several times during this study, but animals were selected in sequence and each animal was bled not more frequently than once per week.

S.B. Harris et al. / Mech. Ageing Dev. 73 (1994) 209-221

213

2. 6. Assays Serum glucose was measured with an enzymatic (hexokinase) assay and free fatty acids using a commercial reagent kit (Wako, Osaka, Japan). Serum corticosterone (ICN Biomedicals, Costa Mesa, CA) and insulin (Novo, Denmark) were measured using commercial R I A kits. 2. 7. Data analysis Data are presented as means 4- S.E.M. The data were analyzed using multiple regression. Comparisons between groups of animals were made using a t-test or (for post hoc comparisons) Duncan's test. 3. R e s u l t s

3.1. Animal weights At all ages, body weights differed significantly between group's eating differing amounts of food (Table 1, Fig. 1). Body weight was stable throughout life in the Severely Restricted mice, but rose slowly with aging in the Mildly Restricted group. The Fully Fed group showed significant weight gain between 7 and 17 months, but none thereafter. 3.2. Study 1: Mid-morning serum glucose and CS, 26 h after feeding Within each intake group, age did not have an effect on significant effect on fasting mid-morning serum glucose concentrations, save that severely restricted animals showed a significant rise in serum glucose at 17 months, while mildly restricted animals showed a significant fall in fasting glucose at the same age. Mice in the fully fed group showed no significant fasting glucose change with aging (Table 2, Fig. 1). Dietary restriction depressed average fasting glucose; concentrations in all three age groups. Average glucose concentrations for fasting Fully Fed, Mildly Restricted, and Severely Restricted groups were 113, 92, and 68 mg/dl respectively. Diet (energy intake) also had a significant effect on glucose concentrations at each age, except between the two restricted groups at 17 months, and between the mildly restricted and fully fed groups at 29 months.

Table 1 Body weights of mice of different ages under varying degrees of dietary restriction since weaning Feeding

50% Ad libitum Mildly restricted Ad libitum fed

Age 7 Months

17 Months

29 Months

21 ± 0.5a 28 -4- 0.9b 36 q- 1.6¢

21 ± 0.6a 32 ± 0.8 b 51 .4- 1.6c

23 .4- 1.1 a 39 .4- 1.5b 48 .4- 2.6¢

Estimated as 100-110 kcal/week. Each group represents eight mice. Values are mean weights in ± S.E.M. Means in each column not sharing a common superscript are significantly different (P < 0.05). grams

214

S.B. Harris et al./Mech. Ageing Dev. 73 (1994) 209-221

120

g

,-, 300

loo o

200

~ o

loo

ao °°

a0

40

i

i

i

I

|

7

i

!

I

i

i

!

17

~d

I

~__..----.~-.-.~

o

O

.

29 o

.

7

Age (months)

.

.

.

.

.

.

.

.

~ .

.

17

29

Age (months)

50 40 o

30 20

o

M S

/ S-

S

10

m

0

|

7

i

i

i

~

i

!

i

i

i

17

i

i

29

Age (months) Fig. 1. Body weights, midmorning serum corticosterone and serum glucose levels of mice of different ages under varying degrees of dietary restriction since weaning.

T h e r e w a s a s i g n i f i c a n t d r o p in m i d - m o r n i n g f a s t i n g C S c o n c e n t r a t i o n s b e t w e e n 7- a n d 2 9 - m o n t h m i c e in t h e r e s t r i c t e d g r o u p s , b u t n o s u c h a g e - r e l a t e d c h a n g e f o r t h e fully f e d a n i m a l s , w h i c h h a d m u c h l o w e r C S c o n c e n t r a t i o n s . All g r o u p s s h o w e d a d e c l i n e in m i d - m o r n i n g C S b e t w e e n 17 a n d 29 m o n t h s ( T a b l e 3, Fig. 1).

Table 2 Serum glucose levels of mice of different ages under varying degrees of dietary restriction since weaning Feeding

50% Ad libitum Mildly restricted Ad libitum fed

Age 7 Months

17 Months

29 Months

58 ± 6 a 94 4- 4 b 123 ± II c

84 ± 3 a 74 + 4 b I l l 4- 8 c

63 ± 3 a 107 ± 5 b 105 ± 8 b

Estimated as 100-110 kcal/week. Each group represents eight mice. Values are mean glucose concentrations in milligrams per deciliter ± S.E.M. Means in each column not sharing a common superscript are significantly different (P < 0.05).

215

S.B. Harris et al./Mech. Ageing Dev. 73 (1994) 209-221

Table 3 Midmorning serum corticosterone levels of mice of different ages under varying degrees of dietary restriction since weaning

Feeding

50% Ad libitum Mildly restricted Ad libitum fed

Age 7 Months

17 Months

29 Months

398 ± 35a 130 ± 23b 72 ± 15c

365 ± 45a 365 ± 48a 108 ± 28b

210 ± 35a 57 ± 11b 42 ± 6b

Estimated as 100-! I0 kcal/week. Each group representseight mice. Valuesare mean serumcorticosterone levels in nanograms/ml ± S.E.M. Means in each column not sharing a common superscript are significantly different (P < 0.05).

On average, dietary restriction increased mid-morning fasting CS concentrations substantially. Mean CS serum concentrations (across all ages) for the fully fed, mildly restricted, and severely restricted groups were 74, 184 and 324 ng/ml respectively. Severely restricted animals of each age showed a significantly greater CS than fully fed animals, and also a significantly greater CS than mildly restricted animals at 7 and 29 months. 3.3. Study 2: Glucose tolerance test and stress CS secretion test

The results of the glucose tolerance tests and CS measurements during the test (a measure of stress CS secretion) are shown in Fig. 2. Dietary effects proved to be more important than aging effects. Fully fed and young mildly restricted animals did not show a significant rise in glucose from baseline during the test, although a trend is evident. Old mildly restricted animals and both groups of severely restricted animals showed a significant (P < 0.05) glucose rise at 30 and 60 min, but their glucose concentration values had returned to baseline by 120 min. Neither severely restricted nor fully fed animals showed a difference between age groups in glucose tolerance, but in the mildly restricted groups, older animals showed a significantly larger glucose rise (less glucose tolerance) than younger animals. No effect of glucose injection was seen with our insulin assay (Fig. 2), although in general fully fed animals had significantly higher insulin concentrations throughout than the two restricted groups. No clear-cut effect of glucose injection on free fatty acid (FFA) concentrations was seen in our assay for most groups of animals (Fig. 2). Young severely restricted animals initially had significantly higher F F A concentrations than old animals, and these were suppressed significantly by glucose. Corticosterone (CS) secretion in response to the glucose injection is shown in Fig. 2. Mildly restricted animals did not respond significantly, while only old animals responded among the severely restricted group (P < 0.05 at 60 min). Both young and old animals responded significantly (P < 0.05) in the fully fed groups, with the older animals having a significantly higher CS concentration than younger ones at 60 min.

S.B. Harris et al./ Mech. Ageing Dev. 73 (1994) 209-221

216

260

Gluco,e

to~

o .o o~.

~

Pi I00 o~

50

o ~

o

Y

-

o -

O~d

Young

ZSO

c,lucose

250

200

'

ZOO

150

(e)

0.5 o.o

o.o

-

0

Young,,(?) oxd:('t)

]

i

l

1.0 , ~, / - : - -

o"

° ° ° °

~

T

'°°1/-% v

~

o

'.

o

8oo

El ,oo

*p<,o5

800

I00 0

120

°Y:

I

!

500

',

4oo ]

:

~

I. 1

8oo 5o0

.

%,----i'

~ ~ ~

2OO

30 80 90 TIX(E ( m i n )

;

•

500

0

y

0

,

1000 q

•a,

o

1

0.0

u,

,ooJ

8 ,oo

1

0

..p<.01

,ooI ~

(B)

1.5

0.6

:-~ °00

~-

Young

2.0 m~

o~

:~. ooo

-

o - old (e)

2.5

"p<05 ,,p<.Ol

;

Y

3.0

ooo :

150

50 ¥ o

2.5 ] 2.0

e luco,~. 0

100

5o

~l.o ]

2.0

oo.~

~/J ~ Y0 ~ O:

~oo

o

(8)

Z~

z~ ~

SEVERELY RESTRICTED

MILDLY RESTRICTED

FULLY FED

'°° 1

,oo

~,~

zoo

1oo o

o

30

80

°o

TFdE ( m i n )

120

o

30

80

T[UE

go

1~o

(rain)

Fig. 2. Serum levels of glucose, insulin, free fatty acids, and corticosterone during intraperitoneal glucose challenge of mice of young (Y. 7-8 months old) and old (O, 29-30 months old) mice on three different diets: severely restricted 50% of ad libitum, mildly restricted to 80% of ad libitum, and fully fed (ad libitum). *P < 0.05 for differences between young and old at the same time point. **P < 0.01. Times are in minutes post injection (dotted line).

3.4. Study 3: Diurnal variation o f glucose, corticosterone, and free f a t t y acids Mean serum glucose values over a 36-h period are shown for y o u n g and old groups in Fig. 3. A d i u r n a l p a t t e r n in which CS c o n c e n t r a t i o n s are elevated at night is seen in all groups, and is most p r o n o u n c e d in the fully fed group. A l t h o u g h there was a trend t o w a r d lower glucose c o n c e n t r a t i o n s in older animals in each diet group, no significant effect o f aging was seen. As expected, diet and feed timing (every d a y vs. every other day feeding) exerted p r o f o u n d effects on the b l o o d glucose c o n c e n t r a t i o n s o f animals at each age. The mean serum glucose c o n c e n t r a t i o n over 36 h for y o u n g fully fed, mildly restricted, and severely restricted animals were 103, 94, and 59 mg/dl, respectively. M e a n glucose c o n c e n t r a t i o n s over 36 h for the c o r r e s p o n d i n g old animal groups were 93, 85 and 58 mg/dl. M e a n 36 h glucose c o n c e n t r a t i o n values did not differ with age within any diet group. M e a n 36 h, glucose c o n c e n t r a t i o n s were significantly less ( P < 0.0t) for severely restricted animals than for other animals at both ages, but at each age

217

S.B. Harris et al./ Mech. Ageing Dev. 73 (1994) 209-221 FULLY FED __~

rood

~_. "0

~,o ~

UlLDLY RESTRICTED ToEd

-:': :~:ii!!iiii l

i

°0

SEVERELY RESTRICTED

-,- ~ood

THd

"i@

......

go

~0

o

0

Y - Young (6) o - Old ( e )

80 o

7 - Younz (El 0

Old

~r~.

Igo

Y - Y o u n g (61 0 - Old (0)

'(e)

i".:. .~

mu

Q

0

,

,p~.O~ ,,p~.Ol

1200

o

$

"p<.06

"p¢.06 ~

t

•

*'p<.Ot

11100

g ~ eoo 1

l 1.0 t ~

,,0 ]

,,o ] 600

o

400

~ ~oo o

.

°oo

o~

loo o

i

i

i!iiiiiiiiiiiiii!;ili~i~!~i;i~ii

i Lldht

~

Lldht

lOO

400

..~i

i Lllht

Lllht

o~

~

400

,o. 1

1o0

i

i 8

°

14

20

o

O °0

82

TIME SINCE FEEDING (hr)

0

U 2

8

14

II0

°

°

TIME ~INCE FEEDING ( h r )

0

iN:

TBEIN 8

14

80

B4

88

SINCE FEEDING (hr)

Fig. 3. Serum levels of glucose, insulin, free fatty acids and corticosterone measured over 36 h for mice of two different ages: young (Y, 7-8 months-old) and old (O, 29-30 months old) three different diets: severely restricted, 50% of ad libitum, mildly restricted to 80% of ad libitum, and fully fed (ad libitum). *P < 0.05 for-differences between young and old at the same time point. **P < 0.01. Times are in hours post feeding. Dark legend, night cycle; dotted line, feeding.

mean glucose concentrations did not significantly differ (P > 0.05) between fully fed and mildly restricted groups (comparison done with Duncan's test for age*diet interaction collapsed across time). Mean serum CS values over 36 h are shown for young and old groups Fig. 3. A pronounced diurnal variation is apparent in mildly restricted and fully fed animals, with CS peaks occurring near the beginning o f the dark phase o f the 24 h cycle. For these two groups, old animals exhibited a significantly earlier peak CS concentration than younger animals. In young severely restricted animals, CS rose significantly 36 h after feeding (P < 0.01), but this response was blunted in older animals and did not reach significance with respect to CS around the time o f feeding. Mean 36 h CS concentrations did not significantly differ (P < 0.05) between diet or age groups. Mean serum F F A concentrations measured over 36 h are shown in Fig. 3. For all diet groups there was a trend for young animals to exhibit higher F F A concentrations than older animals, these differences between age groups becoming significant (P < 0.05) for the fully fed and severely restricted groups (which are fed every 48 h) at 26 h post feeding.

218

S.B. Harris et al./Mech. Ageing Dev. 73 (1994) 209-221

4. Discussion

We found that for each level of energy restriction that fasting glucose concentrations do not change greatly with senescence in our strain, in general accordance with studies in rats [10,11], although we did find evidence of a mid-life glucose peak in our mildly restricted animals, perhaps corresponding to something also reported in rats [12]. We found that mice restricted to 50% ad libitum intake had significantly lower serum glucose concentrations than mice fed ad libitum or at 80% of ad libitum, with differences between groups ranging from 26 to 43 mg/dl. Absolute fasting serum glucose concentrations were as low as 32 mg/dl in restricted mice. Reported work in restricted F344 rats has found plasma glucose concentrations in animals energy restricted to 60% of ad libitum to be an average of 15 mg/dl lower than ad libitum fed controls. Previous studies in this laboratory [14] using mice eating a purified diet have shown similar glucose reductions. This characteristic serum baseline glucose reduction is independent of composition of diet, and is seen not only in rodents but also in energy restricted non-human primates (R. Weindruch, personal communication) and energy restricted humans [15]. No saline controls were used for this series of i.p. glucose tolerance tests, therefore glucose serum increase results reported here formally include effects from the stress of injection. However, subsequent work in our laboratory (not yet published) has found that i.p. saline injection itself does not significantly raise blood glucose concentrations in this strain under these conditions, so our glucose tolerance results are probably indicative of glucose effect only. The present study found evidence for a relative age-related impairment of glucose tolerance in mice, evident only in the mildly restricted animals. For groups not severely restricted, older mice in our study had time-averaged glucose concentrations less than younger mice (this did not reach significance), a finding seen less often in published-studies with rats. Both rats and mice undergo islet cell hypertrophy with advancing age, but may display contrasting patterns of glucose tolerance. In a study of glucose homeostasis in aging C57BL/6J mice, Leiter et al. [16] found that old mice had improved glucose clearances (i.p. loading) and lower non-fasting glucose concentrations than mature younger mice, and this was associated with increased pancreatic islet size and increased stimulated insulin secretion capacity in the older mice. By contrast, rats have been reported to exhibit worsening glucose tolerance between 1 and 2, and 9 and 12 months of age due to insulin insensitivity [17,18], and perhaps also a delay in first phase insulin release [11]. In examining the effect of lifetime energy restriction on glucose handling in mice we found that severe restriction decreases glucose tolerance, and prevents the agerelated decrement in glucose handling that occurs with some levels of energy restriction in our mouse strain. In Sprague-Dawley rats, 12 months of energy restriction improved glucose tolerance [11 ], and prevented much of the 'normal' age-associated exocrine cell mass enlargement in the pancreas, as well as age-associated insulin resistance [17]. However, maximal stimulated insulin output per islet was still decreased with age under chronic energy restriction [17]. Similarly, Koizumi et al. [14] have reported that lifetime energy restriction decreases pancreatic islet cell mass in the

S.B. Harris et al. / Mech. Ageing Dev. 73 (1994) 209-221

219

mouse strain, diet, and lab conditions used in the present study, and others have reported that chronic energy restriction abolishes the normal exocrine hypertrophy noted with aging in another mouse strain [16]. These findings, together with present results, suggest that severe restriction blunts maximum insulin secretion in young and old mice. The commonly used terminology of 'better' or 'poorer' glucose tolerance, when used to describe energy restricted mice, does not necessarily correlate, respectively with better or poorer physiologic status. Although we found poorer glucose tolerance in restricted mice, we also found average lower glucose concentrations in these animals, indicating that in vivo the improved glucose handling ability of lessrestricted animals is over-ridden by their increased food consumption. Glucose is a potential mediator for multiple age-related free radical and protein pathologies [19]. Although our findings support the hypothesis that damaging glycation of proteins may contribute to the pathology of aging, they do not rule out more subtle effects of lowered serum glucose concentrations, such as possible glucose-mediated effects by trans-acting factors [7]. In the present study we found a decline in morning CS concentrations with aging in mice, in accord with previous studies in mice, but did not find an age-related decline in 36-h averaged CS concentrations. In both fully fed and severely restricted animals we also noted a tendency for CS concentrations raised by stress (ip glucose injection) to delay return to baseline in older animals, a phenomenon noted in many rat studies which has been cited as evidence in CS-related theories of aging [4,5]. In the present study we noted significant changes in diurnal secretion of CS in response to diet. Gallo et al. [20] found that 11 weeks of mild restriction (80% of ad libitum) produced a post-feeding suppression of diurnal CS concentrations in rats. Dietary restriction of rats on a once-a-day feeding schedule typically shifts the CS peak to the pre-feeding time, with variable effects on total CS concentrations [21,22]. In the present study we noted the same phenomenon of rising CS concentrations with fasting, associated with the rapidly rising free fatty acid concentrations of starvation. However, because our restriction regimen involves every-other-day feeding, our severely restricted animals did not exhibit normal CS circadian rhythms. In addition, because the relatively longer interval of food deprivation resulted in an elevated CS toward the end of our 48-h feeding-cycle severe restriction regimen, we did not find lower average CS concentrations in our severely restricted mice, and indeed in Study 1 found, across all ages, higher mid-morning CS levels 26 h post feeding. Masoro et al. [23] have failed to find the average lower CS concentrations in their restricted rats required by CS theories of aging, and indeed found the opposite. Since both the every-day feeding regime of Masoro et al. and our own feeding regime have been proven effective at significantly extending maximum lifespan [8,24] these results argue against the hypothesis [4,5] that average lower CS concentrations mediates any of the beneficial effect of dietary restriction in rodents [23]. The present study underscores the difficulties for diet/restriction aging studies which lie in separating metabolic and endocrine effects of altered feeding times and hour by hour energy intake, from effects attributable purely to more important changes in mean energy consumption over many days. For example, although in this

220

S.B. Harris et a l . / Mech. Ageing Dev. 73 (1994) 209-221

study the fully fed animals were obese and food was always present in this group's cages to gross inspection, F F A results suggest that these 'ad libitum' animals nevertheless ate more on days of feeding, and less on intervening days. Moreover, while complete food deprivation for 24 h in rodents results in drastic changes in F F A , CS, and presumably many other parts o f metabolism, it is probably that these changes do not need to be present for the anti-aging effects o f calorie restriction to manifest: in contrast to our every-other-day regime, many restriction studies have shown a good anti-aging effect for every-day feeding, and the study of Nelson and Halberg found the full anti-aging effect o f restriction in mice even when six small meals per day were fed [25]. Future-studies attempting to isolate the metabolic effects which mediate the anti-aging effect o f dietary restriction must take care to use several feeding regimes to separate out the large metabolic and h o r m o n a l effects which are due to the altered feeding patterns or periods o f starvation imposed by some restricted diet regimes, since these effects may have nothing to do with the energy restriction anti-aging mechanism. F o r example, in the case o f serum corticosterone concentrations in rodents, the remarkable robustness of the anti-aging effect in the face of many different dietary restriction regimes which would be expected to induce many different kinds of corticosterone secretion patterns [23,25], argues that this system is probably not extensively casually involved in the anti-aging effect. In the case o f glucose, it will be useful to examine average serum glucose concentrations in energy restriction models where energy restriction is carried out by means o f very frequent small feedings [25].

5. Acknowledgements Our thanks to Paula Wilhelmi for skilled technical assistance with this study, and Augusto Tayag for expert animal care. This study was supported by N I H grant AG-00424.

6. References 1 R. Weindruch and R.L. Walford, The Retardation of Aging and Disease by Dietary Restriction, Charles C. Thomas Springfield, Ill, 1988, 2 E.J. Masoro, Minireview: food restriction in rodents: an evaluation of its role in the study of aging, J. Gerontol. Biol. Sci., 43 (1988) B59-B64. 3 R.L. Walford, S.B. Harris and R. Weindruch, Dietary restriction and aging: historical phases, mechanisms, And current directions. J. Nutr.. 117 (1987) 1650-1654. 4 R.M. Sapolsky, L.C. Krey and B.S. McEwen, The neuroendocrinology of stress and aging: the glucocorticoid cascade hypothesis. Endocrinol. Rev., 7 (1986) 284-301. 5 C.V. Mobbs, Neurotoxic effects of estrogen, glucose, and glucocorticoids: neurohumoral hysteresis is and its pathological consequences during aging. In: Review of Biological Research in Aging. WileyLiss New York, 1990, pp. 201-228. 6 A. Cerami, Hypothesis: Glucose as mediator of aging. J. Am. Geriatr. Sot., 33 (1985) 626-634. 7 R.L. Walford and M. Crew, How dietary restriction retards aging: an integrative hypothesis, Growth. Dev, and Aging. 53 (1990) 139-141. 8 R. Weindruch, R.L. Walford, S. Fligiel and D. Guthrie, The retardation of aging mice by dietary restriction: longevity, cancer, immunity and lifetime energy intake. J, Nutr., 116 (1986) 641-654. 9 S.B. Harris, R. Weindruch, G.S. Smith, M.R. Mickey and R.L. Walford, Dietary restriction alone and in combination with oral ethoxyguin/2-mercaptoethylamine in mice. J. Gerontol.. 45 (1990) B141-BI47.

S.B. Harris et al. / Mech. Ageing Dev. 73 (1994) 209-221

221

10 J.E. Klimas, Intestinal glucose absorbtion during the lifespan in a colony of rats. J. Gerontol., 23 (1968) 529-532. 11 N. Kalant, J. Stewart and R. Kaplan, Effect of diet restriction on glucose metabolism and insulin responsiveness in aging rats. Mech. Aging, Dev., 46 (1988) 89-104. 12 E.J. Masoro, C. Compton, B.P. Yu and H. Bertrand, Temporal and compositional dietary restrictions modulate age-related changes in serum lipids. J. Nutr., 113 (1983) 880-892. 13 E.J. Masoro, M.S. Katz and C.A. McMahan, Evidence for the glycation hypothesis of aging from the food-restricted rodent model. J. Gerontol., 44 (1989) B20-B22. 14 A. Koizumi, Y. Wada, M. Tsukada, J. Hasegawa and R.L. Walford, Low blood glucose levels and small islets of Langerhans in the pancreas of calorie restricted mice. Age, 12 (1989) 93-96. 15 R.L. Walford, S.B. Harris and M.W. Gunion, The calorically restricted low-fat, nutrient-dense diet in Biosphere 2 significantly lowers blood glucose, total leukocyte count, cholesterol, and blood pressure in humans. Proe. Soc. Natl. Acad. Sci., USA, 89 (1992) 11533-11537. 16 E.N. Leiter, F, Premdas, D.E. Harrison and S.C. Lipson, Aging and glucose homeostasis in C57BL/6J male mice. FASEB J., 2 (1988) 2807-2811. 17 E. Reaven, D. Wright, C.E. Mondon, R. Solomon, H. Ho and G.M. Reaven, Effect of age and diet on insulin secretion- and insulin action in the rat. Diabetes, 32 (1983) 175-180. 18 E. Bracho-Romero, Effect of aging and weight on plasma glucose and insulin responses in the rat. J. Am. Geriatr. Soc., 25 (1977) 299-302. 19 B.S. Kristal, An emerging hypothesis: synergistic induction of aging by free radicals and maillard reactions. 3". Gerontol., 47 (1992) B107-B114. 20 P.V. Gallo and J. Weinberg, Corticosterone rhythmicity in the rat: interactive effects of dietary restriction and schedule of feeding. J. Nutr., 111 (1981) 208-218. 21 A. Armario, J.L. Montero and T. Jolin, Chronic food restriction and the circadian rhythms of pituitary-adrenal hormones, growth hormone, and thyroid hormone. Ann. Nutr. Metab., 31 (1987) 81-87. 22 J. Stewart, M.J. Meaney, D. Aitken, L. Jensen and N. Kalant, The effects of acute and life-long food restriction on basal and stress-induced serum corticosterone levels in young and aged rats. Endocrinology, 123 (1988) 1934-41. 23 F. Sabatino, E.J. Masoro, C.A. McMahan and R.W. Kuhn, assessment of the role of the glucocorticoid system in aging processes and in the action of food restriction. J. Gerontol., 46 (1991) BI71-BI79. 24 E.J. Masoro, B.P. Yun and H.A. Bertrand, Action of food restriction in delaying the aging process. Proc. Natl. Acad. Sci. USA, 79 (1982)4239-4241: 25 W. Nelson and F. Halberg, Meal timing,-circadian rhythms, and life span of mice. J. Nutr., 116 (1986) 2244-2253.