0013-7227/79/1042-0350$02.
00
Endocrinology
Copyright © 1979 by The Endocrine Society
Daily Rhythms in Adrenal Responsiveness to
Adrenocorticotropin Are Determined Primarily by the
Time of Feeding in the Rat*
CHARLES W. WILKINSON,! JEANETTE SHINSAKO, AND MARY F. DALLMAN^
Department of Physiology and Metabolic Research Unit, University of California, San Francisco,
California 94143
ABSTRACT. These experiments were done to determine: 1)
whether feeding-related shifts in daily corticosterone rhythms
are dependent upon changes in ACTH rhythms, 2) whether
restricted feeding of rats results in abnormally high ACTH and
corticosterone levels (i.e. stress), and 3) whether changes in
either insulin or glucose levels might be the concomitants of
feeding that change adrenal responsiveness to ACTH. Young
male rats (80-90 g) on a 12-h light, 12-h dark cycle were allowed
access to one of three diets for 2 h/day beginning either at lights
off or lights on. The diets contained 3%, 4.5%, or 11% fat. A group
of rats had ad libitum access to the food containing 4.5% fat. On
day 20 of this regimen, rats were killed at 2- to 4-h intervals
during the next 24 h, and plasma ACTH, corticosterone, insulin,
and glucose were measured. Adrenal weight and corticosterone
content were also determined.
C IRCADIAN rhythms in plasma corticosteroid con-
centrations of many mammals are characterized by
the occurrence of maximal levels at approximately the
beginning of daily activity (1-5). In nocturnally active
murids, the circadian peak occurs at, or shortly before,
the onset of darkness (4, 5). It has been reported repeat-
edly that restriction of rats or mice to a brief period of
eating or drinking early in the light phase of the cycle
results in a shift of the daily plasma corticosterone peak
to the time immediately before eating (6-9). However, it
is debatable whether the shift in corticosterone repre-
sents a shift in the diurnal rhythm or the stress of fasting
which is relieved by food presentation. Circadian
rhythms of circulating ACTH in rats and man have been
reported to be in phase with those of corticosteroids (10,
11), and it has been assumed that the rhythm in corti-
Received July 3, 1978.
Address requests for reprints to: Dr. Charles W. Wilkinson, Depart-
ment of Physiology, University of California, School of Medicine, San
Francisco, California 94143.
* This work was supported in part by a NASA-University Consor-
tium Interchange Agreement NCA2-OR665-602 and by USPHS Grant
AM-06704.
f Recipient of Postdoctoral Fellowship AM-05681.
| Recipient of Research Career Development Award NS-00072.
350
Vol. 104, No. 2
Printed in U.S.A.
In none of these experiments was an ACTH rhythm demon-
strable by analysis of variance. Neither ACTH levels nor adrenal
and plasma corticosterone levels were higher in animals fed 2
h/day than in rats eating ad libitum. Peak corticosterone levels
occurred just before feeding, and the restricted feeding paradigm
appeared to sharpen the daily rhythms. However, there was also
an effect of the light-dark cycle on corticosterone measures.
Dietary fat content was directly related to increases in body
weight and mean insulin levels and inversely related to adrenal
responsiveness to ACTH. The data show that: 1) the time of
feeding determines the timing of the corticosterone rhythm in
the absence of a rhythm in ACTH, 2) restricted feeding is not a
stress, and 3) neither insulin nor glucose has a substantial
influence on adrenal responsiveness to ACTH. (Endocrinology
104: 350, 1979)
costeroids is driven by that in ACTH. However, in
ACTH-implanted, hypophysectomized rats, in which cir-
cadian ACTH cyclicity was presumed to have been abol-
ished, plasma corticosterone rhythms are maintained
(12). Furthermore, in intact rats under 12-h light, 12-h
dark (12:12 LD) conditions, the rhythms in plasma
ACTH are of much smaller amplitude than those of
plasma corticosterone, and often do not reach statistical
significance (13).
We have recently obtained evidence in rats that adre-
nal responsiveness to both endogenous and exogenous
ACTH is approximately 2.5 times greater at the peak
(lights off) of the daily corticosterone rhythm than at its
nadir (lights on) and that the ACTH and adrenal
rhythms are independent (13). Evidence for time of day
dependent changes in responsiveness of the human ad-
renal to ACTH has also been reported (14). In light of
these findings and in the absence of measurements of
plasma ACTH or adrenal corticosterone concentration in
previous investigations of restricted feeding, we have
attempted to determine the relationships among feeding-
induced reversals of plasma corticosterone rhythms and
rhythms in ACTH and adrenal responsiveness to ACTH.
Adrenal and plasma corticosterone concentrations and
 FOOD AND ADRENAL RESPONSIVENESS TO ACTH
plasma ACTH levels were measured at 2- to 4-h intervals
throughout 24 h in rats maintained for 20 days on feeding
periods of 2 h/day at either lights on or lights off.
Food ingestion is a potent stimulus to insulin secretion,
which in turn may be related to adrenal responsiveness
to ACTH. Human diabetics and alloxan-diabetic rats
have been reported to exhibit increased adrenocortical
responsiveness to stress or ACTH administration as well
as a flattening of circadian rhythms in plasma corticoste-
roids (15-17). Insulin treatment of alloxan-diabetic rats
returned adrenal responsiveness and circadian rhyth-
micity to normal (16). There is also evidence that circa-
dian rhythms in plasma insulin levels are 180° out of
phase with ACTH or corticosteroid rhythms in man (18)
and rats (7), although there are contradictory data for
the rat (19, 20). Therefore, we also measured plasma
glucose and insulin levels in these animals in a prelimi-
nary attempt to determine the specific concomitants of
eating and/or drinking that are capable of altering adre-
nocortical rhythms. In each of three separate experi-
ments, a different diet was provided to the animals in
order to investigate the effect of dietary fat content (or
caloric intake) on insulin concentrations and on the
rhythm of adrenal responsiveness to ACTH.
Materials and Methods
Male Sprague-Dawley rats, weighing 80-90 g (Simonsen,
Gilroy, CA), were received 2-3 days before the beginning of
each experiment and were housed two per cage in hanging wire
baskets in animal rooms maintained at 21-23 C. Rats were
exposed to a 12:12 LD schedule with lights on between
0900-2100 h (with the exception of the dark-fed animals in Exp
3, which were housed in a separate room with a reversed
lighting schedule). In each of three experiments, groups of 64-76
rats were given pellets of food inside their cages at lights on,
and equal numbers of animals were given food similarly at
lights off. All remaining food was removed from the cages 2 h
after its presentation. In one experiment (no. 2), a third group
of 76 rats fed ad libitum was included. All rats were given free
access to water throughout all experiments. Animals were
weighed at the beginning of the experiment and on alternate
days throughout. After 20 days on the 2-h feeding schedule,
groups of six to eight rats from each of the feeding conditions
were sacrificed at 8 or 12 times during a 24-h day. Groups were
killed 10 min before feeding, 10 min before food removal, and
at multiples of 2-h intervals from these times. Both animals in
a cage were killed by decapitation within 15 sec of opening the
cage, and trunk blood was collected in heparinized plastic tubes
and immediately chilled in ice. Left adrenals were removed and
kept moist in petri dishes on filter paper saturated with saline,
and terminal body weights were recorded. The blood was
centrifuged at 4 C shortly after collection and plasma was
separated and frozen in separate aliquots for 1) ACTH and
corticosterone, and 2) for insulin and glucose determinations.
ACTH was measured by RIA (21, 22). Corticosterone was
measured by a competitive protein-binding technique (23) using
351
human plasma as the source of transcortin. Insulin was mea-
sured by RIA using kits from Amersham-Searle and standards
of human insulin. Glucose was determined by a glucose oxidase
technique using a Beckman Autoanalyzer. For each hormone
assay, samples from individual groups within an experiment
were measured in each run of an assay so that interassay
variation was incorporated into individual group means and
variances. Wet weight of the adrenals was determined, and the
adrenals were homogenized (1:500, wt/vol) in a 20% ethanol-
isotonic saline solution for determination of adrenal corticos-
terone by competitive protein binding. In each of the three
experiments, a different diet (differing primarily in fat content,
as stated by the manufacturers), was fed to the rats. In Exp 1,
rats were fed Berkeley diet (Feedstuffs Processing, San Fran-
cisco, CA) containing 3% fat; in Exp 2, they were fed Purina
laboratory chow containing 4.5% fat; and in Exp 3, they were
given Purina mouse chow, containing 11% fat. In chronological
sequence Exp 3 followed Exp 1 and the change in diet was an
independent manipulation on the part of the Animal Care
Facility that was not discovered until the experiment was well
in progress. An intermediate diet was then selected by the
experimenters for use in Exp 2, the final experiment.
Data were processed for Bartlett's test for homogeneity of
variance, one- or two-way analysis of variance, Tukey's HSD
multiple range test, and linear correlation on an IBM 370
computer (24). One-way analyses were used to determine time
of day effects (i.e. presence or absence of daily rhythms) and
effects of dietary fat content. Paired comparisons between
groups receiving different diets were made using Tukey's HSD
test. Two-way ( 2 x 2 factorial) analyses were used to determine
effects and interactions of time of feeding and phase of the light
cycle. Only data obtained from animals (480 rats out of a total
of 500) for which all measurements were made successfully
were included in statistical analyses or in figures.
Results
Levels of hormones, adrenal responsiveness to ACTH,
and glucose during 24 h in rats fed ad libitum
The data obtained from rats fed ad libitum are shown
in Fig. 1. In addition to plasma ACTH, insulin, glucose,
and plasma and adrenal corticosterone concentrations
plotted against time of day, adrenal responsiveness (ad-
renal corticosterone/log ACTH) is also shown in the
figure. The use of this measure is justified by reports that
adrenal synthesis and secretion of corticosterone are
directly proportional to the logarithm of the ACTH
concentration under conditions of bioassay for ACTH
(25,26). Because the responsiveness measure is a function
of adrenal rather than plasma corticosterone concentra-
tion, it is not directly affected by changes in the rate of
clearance of corticosterone from the circulation. Al-
though ACTH levels were highest just before lights out
and were significantly lower in the sample obtained just
before lights on, one-way analysis of variance did not
reveal a significant rhythm in ACTH. By contrast, sig-
nificant time of day effects were found for adrenal re-
352 WILKINSON, SHINSAKO, AND DALLMAN
40
Plasma
ACTH 20
(pg ml)
0
80
Adrenal
B 40
0900 2100 0900
Time of day (hours)
FIG. 1. Plasma ACTH, corticosterone (B), insulin, and glucose concen-
trations of rats with free access to food containing 4.5% fat. Also plotted
are adrenal corticosterone concentrations and adrenal responsiveness
to ACTH (micrograms of B per g divided by log picograms of ACTH
per ml). Each point represents the mean (±SEM) value for 5-10 rats. In
several cases, the SEM is too small to be visible beyond the dimensions
of the symbol. The two points connected by dotted lines at the right of
each line are repetitions of the first two points at the left. Asterisks at
the far right indicate significant (P < 0.0001) time of day effects by
analysis of variance. For ACTH, P = 0.60; for glucose, P = 0.43.
sponsiveness to ACTH and for the corticosterone meas-
ures. Adrenal responsiveness and corticosterone levels
increased at 1700 h and remained elevated for the first 6
h after lights out. Plasma corticosterone levels peaked at
a level of 14 /xg/lOO ml in the sample obtained just before
lights out. Minima in adrenal responsiveness and the
corticosterone measures occurred in the samples col-
lected just before lights on. Plasma insulin concentrations
rose sharply from a nadir at 1900 h to a relatively stable
plateau achieved between 2100-0900 h, during the time
when rats normally eat. There was no rhythm in plasma
Endo • 1979
Vol 104 • No 2
glucose levels; however, the lowest point coincides with
the minimum in insulin concentration at 1900 h.
Levels of hormones, adrenal responsiveness to ACTH,
and glucose levels during 24 h in rats fed for 2 h at
lights on or lights off
Mean values of data for each of the three experiments
are plotted as a function of the time of day in Fig. 2. One-
way analyses of variance revealed significant time of day
effects for all variables except ACTH in each feeding
condition in every experiment. A significant rhythm in
ACTH was never obtained in these studies. However,
mean ACTH levels generally appeared to be highest
during the 6 h before feeding and lowest 12 h later.
Rhythms in adrenal responsiveness and corticosterone
reached their maxima during the hours just before feed-
ing. In five of the six restricted feeding groups, the daily
minimum in adrenal responsiveness to ACTH occurred
at the first time point after lights on, as did the minima
for adrenal and plasma corticosterone concentrations in
the three dark-fed groups. Minima for light-fed rats were
less predictable, but for all variables except ACTH they
occurred during the first 8 h after lights on.
In Exp 2 (4.5% fat diet) a direct comparison can be
made between hormone levels from rats eating this diet
ad libitum and rats allowed to feed only 2 h/day. Mean
ACTH levels ranged between 14-38 pg/ml in all three
conditions (Fig. 1 vs. Fig. 2, ). Peak values of adrenal
and plasma corticosterone levels and adrenal responsive-
ness to ACTH did not differ between rats feeding ad
libitum and those fed for 2 h/day. Peak adrenal corticos-
terone levels were 5.1 ± 0.8, 6.4 ± 0.9, and 5.4 ± 0.8
/xg/100 mg, and peak plasma corticosterone levels were
14.1 ± 3.0,19.4 ± 1.6, and 16.2 ± 2.8 /xg/100 ml in the rats
fed ad libitum, for 2 h at lights off, and for 2 h at lights
on, respectively. No significant differences in ACTH,
adrenal corticosterone, or adrenal responsiveness
rhythms could be demonstrated among the ad libitum
and the two restricted feeding groups using two-way
analyses of variance.
Figure 3 shows pooled data from the three experiments
plotted as a function of time of feeding rather than time
of day. On the left side of the figure are shown data for
the eight times common to all three experiments. Adrenal
corticosterone content is included as well as corticoster-
one concentration. One-way analyses of variance of the
combined data from all three experiments revealed sig-
nificant rhythms for all measures for both light-fed and
dark-fed groups, with the exception of plasma ACTH in
the dark-fed group (Table 1). Peak values of ACTH,
plasma and adrenal corticosterone concentrations, adre-
nal responsiveness, and adrenal corticosterone content
all occurred at the onset of feeding regardless of its
relationship to the light cycle. Again, in the pooled data,
 FOOD AND ADRENAL RESPONSIVENESS TO ACTH 353

Food Food
Lights on M Lights off Lights on Lights off
50
Plasma
ACTH 25
(pg/ml)
0

8
Adrenal
corticosterone 4
(>jg/IOO mg)

24

Plasma 16
corticosterone
(>ig/IOOml) 8
0L

Adrenal
responsiveness
/ \ 3 \
/ jjg B /lOOmg \
\logpgACTH/ml/ 0
0900 2100 0 9 0 0 0900 2100 0900
Time of day (hours)
FIG. 2. ACTH and adrenal measures for rats fed for 2 h at lights off (left) or for 2 h at lights on (right). Each line represents the mean values from
a different experiment plotted as a function of time of day. , Data from rats fed the 3% fat diet; , data from rats fed the 4.5% fat diet;
• - •, data from rats fed the 11% fat diet (five to eight rats per point).

ACTH and corticosterone levels resemble those observed
in ad libitum fed rats (Fig. 1). The nadir for each of the
above measures, except for plasma ACTH and corticos-
terone in the light-fed group, occurred at the first time
point after lights on regardless of the feeding schedule.
The right side of Fig. 3 illustrates pooled data from Exp
2 and 3 in which samples were taken at 12 2-h intervals.
These data follow patterns similar to those shown at the
left of the figure, with minima and maxima occurring in
the same relationship to the light cycle and feeding. One-
way analyses of variance of these data did not reveal
rhythms in ACTH in either light- or dark-fed rats,
whereas rhythms in all other variables are highly signif-
icant.
For all of the adrenal measures, levels obtained during
the dark phase of the light cycle are consistently higher
than those obtained during the light phase of the cycle,
although the daily maxima of the rhythms are synchro-
nized with feeding time. The differences between dark-
ness and light are shown more clearly when all data
obtained from a given feeding group at three time points
during the dark phase are pooled and the mean values
are compared with similarly calculated mean values ob-
tained during the light phase of the cycle (Fig. 4). The
mean values (obtained from the data shown in Fig. 3,
left) do not include data from either the 0900- or the
2100-h time points because these points represent dark-
light transition times (Fig. 3, C>). Two-way analyses of
variance (light phase by feeding time; Table 2) show
significant light-dark differences for all variables except
ACTH, whereas there are no significant effects of feeding
time for any variable. There are significant feeding time-
light phase interactions for both plasma corticosterone
concentration and plasma to adrenal corticosterone ratio,
suggesting a feeding-related component in dark-light dif-
ferences in metabolism, distribution, or binding of corti-
costerone.
Plasma insulin and glucose levels (Fig. 5) were lowest
in the samples obtained immediately before feeding. In-
sulin levels show a steep rise with a peak at the next time
point just before the removal of food; the levels then
gradually fall throughout the remainder of the 24 h.
Analyses of variance revealed significant time of day
effects for both insulin and glucose. Table 3 shows that
there is a low but significant positive correlation between
the logarithm of plasma ACTH concentration and adre-
nal corticosterone concentration both in the rats fed ad
libitum and those fed for only 2 h/day. There is also a
354 WILKINSON, SHINSAKO, AND DALLMAN Endo • 1979
Vol 104 • No 2

Food Food
50
Plasma
ACTH 25
(pg/ml)
0

8
Adrenal
corticosterone 4
(>jg/IOOmg)
0

24

Plasma 16
corticosterone
(^g/100 ml) 8
\x ^4, A/
'•9'
0

6
Adrenal
responsiveness
•
/ jjg B/IOOmg \
\logpgACTH/ml/

720
Adrenal
corticosterone content 3 6 0
(ng/adrenal)
0

Plasma-adrenal
1.
corticosterone ratio

jig/IOOml \

yug/IOOmg/ 12 18
Time after food presentation (hours)
FIG. 3. ACTH and adrenal measures pooled across the three experiments are plotted on the left as a function of time after food presentation for
the eight sampling times common to all of the experiments (means of 17-24/group ± SEM). Right, Pooled data for each of the 12 time points
common to Exp 2 and 3 (means of 9-16/group ± SEM). , Results from rats fed during the first 2 h of light; , results from rats fed during the
first 2 h of darkness. € , Samples collected at the time just before a light-dark transition; • , samples collected during the dark period; O, those
coUected during the light. Results of analyses of variance of these data are shown in Tables 1 and 2.

low negative correlation between plasma insulin levels
and both adrenal corticosterone concentration and ad-
renal responsiveness to ACTH which only achieves sig-
nificance in the rats fed 2 h/day. Despite the statistical
significance of the insulin correlations with adrenal cor-
ticosterone and responsiveness, neither r2 value exceeds
0.035. In other words, less than 4% of the variability in
adrenal corticosterone or adrenal responsiveness can be
attributed to changes in circulating insulin levels. Glucose
values show a nonsignificant positive correlation with
adrenal corticosterone levels in the ad libitum groups (r
= +0.119, P = 0.156) and a significant negative correla-
tion with adrenal corticosterone in the restricted feeding
groups (r = -0.265, P < 0.001). Similar inconsistent
correlations were found between glucose and adrenal
responsiveness measures. The inconsistencies in the signs
of the glucose correlations with adrenal measures indi-
cate that changes in glucose levels are not responsible for
diurnal variations in adrenal responsiveness to ACTH.
Adrenal and plasma corticosterone levels are highly cor-
related with one another, and r2 values are 0.640 and
0.558 for the ad libitum and 2-h feeding groups, respec-
tively. Thus it appears that reversal of the plasma corti-
costerone rhythm by reversal of the feeding schedule is
determined primarily by changes in adrenal corticoster-
one secretion, but diurnal fluctuations in metabolism,
 FOOD AND ADRENAL RESPONSIVENESS TO ACTH 355

TABLE 1. Results of analyses of variance of pooled data from Exp 1-3 (data shown in Fig. 3, left)

Dark fed Light fed

Effect; of time after feeding on:
df df
Plasma ACTH 1.08 (7, 153) N.S. 2.97 (7, 144) <0.01
Adrenal corticosterone concentration 11.65 (7, 153) <0.0001 8.97 (7, 144) <0.0001
Adrenal responsiveness to ACTH 8.25 (7, 153) <0.0001 5.60 (7, 144) <0.0001
Adrenal corticosterone content 12.02 (7, 153) <0.0001 8.94 (7, 144) <0.0001
Plasma corticosterone 17.04 (7, 153) <0.0001 15.72 (7, 144) <0.0001
Plasma corticosterone/adrenal corticosterone 2.57 (7, 153) <0.02 6.89 (7, 144) <0.0001

Light-fed Dark-fed Light-fed

Adrenal
responsiveness
Plasma ACTH
/ >ig B/IOOmg \
(pg/ml)
llogpgACTH/miy

Adrenal
Adrenal corticosterone corticosterone content
(.pg/IOOml) (ng/adrenal/IOOgBW)

Plasma-adrenal
corticosterone ratio
Plasma corticosterone
(>ig/IOOmg)
Q JZL
/ jjg/IOOml
\ p g/IOOmg n
I I Light ME Dark
FIG. 4. Mean values from the three sample collections during the darkness and from the three sample collections during the light period in light-
fed and dark-fed groups shown in Fig. 3. Data from samples collected just before a light-dark transition are not included. Each bar represents data
from 55-59 rats. Analysis of these data (Table 3) shows an effect of the phase of the light cycle independent of the time of feeding for all measures
except for plasma ACTH.

distribution, and/or binding also contribute to the circa-
dian variation in plasma corticosterone levels.
Effect of diet on mean levels of hormones, adrenal
responsiveness to ACTH, glucose, and body weight in
rats fed 2 h/day
The effects of differences in dietary composition are
illustrated in Fig. 6. The data shown are overall mean
values for the eight time points common to each of the
three experiments. Weight gained during the 20 days of
the experiment, plasma insulin levels, and plasma glucose
concentrations all increase with increasing dietary fat
content. However, mean adrenal corticosterone concen-
tration was highest in the group fed the low fat diet and
lowest in the rats fed the high fat diet, despite the fact
that mean plasma ACTH concentration was highest in
the latter group. Adrenal responsiveness to ACTH, rela-
tive adrenal weight, and adrenal corticosterone content
were all inversely related to the dietary fat content,
increase in body weight, and plasma insulin levels. Dif-
ferences in adrenal responsiveness to ACTH and adrenal
corticosterone content between Exp 2 and 3 (4.5% and
11% dietary fat content) were not significant when data
from only eight time points were compared; however, the
mean values were significantly different on both meas-
ures when all data from these experiments were included
in the analyses. .
Discussion
The results of these experiments confirm earlier obser-
vations that daily rhythms in plasma corticosterone con-
centration can be synchronized by restricted feeding or
watering schedules. Feeding-dependent rhythms char-
acterized by maxima at the time points immediately
preceding food presentation are also demonstrated for
adrenal corticosterone concentration and content and
adrenal responsiveness to ACTH. Significant rhythms in
plasma ACTH concentration were found only when data
from all three experiments were pooled and then only for
the light-fed groups. In this case also, values were highest
in samples taken just before the onset of feeding. The
absence of an ACTH rhythm in these experiments may
have resulted from the fact that animals on both feeding
schedules as well as the animals fed ad libitum were all
housed in the same room in Exp 1 and 2. The disturb-
ances associated with food presentation and removal may
356 WILKINSON, SHINSAKO, AND DALLMAN Endo • 1979
Vol 104 • No 2

Food Food
en
o

.001
SIS
a.

.00
90
o
cj
" © © Plasma
V V
"co I 1 insulin 45
s §
u <-
CO' CO t o
CN CN CM
(nU/ml)
•a CN CN CN 0
CO " "

CO
w- ^ w- 180
Hcon Plasma
a © to >o glucose 90
i in ©
© CN rH (mg/100 ml)
cu
0 6 12 18 12 18
CU

<0.00(
bo

<0.05
Time after food presentation (hours)

NS
ron

a,
•c aa FIG. 5. Mean levels of plasma insulin and glucose in rats fed 2 h daily
u
t o tO CD either at lights on ( ) or at lights off ( ). Left, Pooled results
CN CN CN
o CN CN CN form the 8 time points common to Exp 1-3; right, mean data from the
u
12 sampling times common to Exp 2 and 3. All values are plotted as a
•51 _
***
IO •<}•
n oi H
Hiriin
CT>
function of time after food presentation.

cu
TABLE 3. Correlations between some of the variables presented in
eron

a. t/3 g CO

CO
Figs. 1 and 2
O «->
.« C
CO CO CO Pooled 2-h
o c CN CN CN Ad libitum
CJ O CN CN CN
Feeding condition feeding (n =
„ o (n = 75)
c 405)
co co i o
2?
•a TJ; O> - H
H CN
Tf O Log ACTH:adrenal corticosterone 0.299" 0.213"
CO Insulin:adrenal corticosterone -0.183'' -0.187*
veness

0.0001

a Insulin:(adrenal corticosterone/log -0.161'' -0.177"

NS
NS

a, ACTH)
Io sh V Adrenal corticosterone:plasma corti- 0.800'' 0.747''
a t j co to to
( N CN CN
costerone
cu < ^S CN CN CN
u
o rt rt -H" " P < 0.005.
c '—' '—' " ^ h
cu
ha
m I-H --i P< 0.001.
fc<co © O
(NO© ' P = 0.058.
CN

cu
c ©
£ a. CO § CO lier findings of nycthemeral rhythms in adrenocortical
ost<
tior

V responsiveness to ACTH which are independent of

tj -*2
O ^"
CO CO tO
CN CN CN
ACTH rhythms (13).
CN CN CN
8g •a
Comparison of the data obtained from rats fed ad
con
renal

O> © CN
t ~ CO —>
libitum (Fig. 1) and those allowed to feed for only 2 h at
fc
HCOO
CN lights off (Figs. 2 and 3) suggests strongly that restricted
CO CO CO
feeding does not represent a stressful stimulus to the rat.
•2 I K a. Z 2 2 ACTH levels are similar and low in the two conditions,
h
o CO CO t O
CN CN CN
and plasma and adrenal corticosterone levels were not
CN CN CN
-a
CO
i-T —T <-T different between the rats fed at lights off and those with
s
© --1 CN continuous access to food in the sample obtained just
O t - H
© © © before lights out in Exp 2 (diet containing 4.5% fat).
?1 Adrenal responsiveness to ACTH and adrenal and
plasma corticosterone levels have decreased by 4 h after
lights out in freely eating animals. It is likely that these
c
O .bp •3
SJJ
rats ate most at the onset of dark as well, since rats fed
3 S ad libitum eat more food during the 2-h interval imme-
« O diately after lights out than during any other interval
.§ SP
CN CO during the 24-h day (27). The most prominent effect on
"s "o ti hormone levels of restricting feeding to a 2-h period each
CJ O L.
$2 cu cu a)
its it: £ day is a sharpening and narrowing of the period of peak
adrenal activity. This is particularly evident in the
have constituted an environmental stimulus of sufficient marked decrease in corticosterone that occurs within 2 h
magnitude to disrupt a rhythm in ACTH. The more of food presentation in the rats fed at lights on (Figs. 2
robust rhythms in adrenal responsiveness to ACTH and and 3). Because there are no significant differences in the
in corticosterone were not disturbed, supporting our ear- maximum levels of adrenal responsiveness to ACTH or
 FOOD AND ADRENAL RESPONSIVENESS TO ACTH 357

Increase
•»
40
Plasma ACTH
in body weight
(pg/ml)
(g/2Odays)
Q
Adrenal
corticosterone Adrenal FIG. 6. The effect of dietary fat content
(.ug/IOOmg) o (caloric density) on body weight, ACTH,
responsiveness
adrenal measures, insulin, and glucose
jjgB/IOOmg \ expressed as overall 24-h means of data
Plasma logpgACTH/ml I from the eight common time points from
corticosterone Exp 1-3. For each measure, group means
Oig/I00ml)
are significantly (P < 0.05) different by
Tukey's HSD test from those with a
different number of asterisks. (*), Group
mean that is not different from either of
Plasma insulin
(.uU/ml)
Relative adrenal size

(mg/100 gBW) [D_D the other two group means. Exp 1, n =

120; Exp 2, n = 102; Exp 3, n = 91.

Adrenal
Plasma glucose 300
corticosterone content
(mg/IOOml)
(ng/adrenal/IOOg) 0
3.0 4.5 11.0 3.0 4.5 11.0

Dietary fat content (%) Dietary fat content (%)

of adrenal or plasma corticosterone between light- and
dark-fed groups, it seems most likely that the time of
food presentation, rather than the possible stress of food
deprivation, determines the adrenal rhythms. In support
of this conclusion is the report that starvation of rats for
as long as 3 days does not result in increased plasma
corticosterone levels (27).
These experiments do not address the question of
whether some aspect of feeding 1) acts as a Zeitgeber to
entrain an endogenous oscillation in the adrenal, a pos-
sibility suggested by reports of adrenal rhythms obtained
in culture (28), or 2) acts as an oscillating driving force to
which the adrenal responds passively. Because rats on
restricted feeding schedules drink as much as 90% of
their daily water intake during the feeding period (29),
these experiments do not differentiate between concom-
itants of eating, as opposed to drinking, as the primary
determinants of adrenal rhythms. Restriction of access
to water has also been shown to resynchronize adrenal
rhythms (6).
The strength of the coupling between time of feeding
and/or drinking and corticosterone rhythms is shown by
the fact that reversal of the light cycle results in concom-
itant reversals in ad libitum food intake and plasma
corticosterone (30), whereas the resynchronization of the
rhythm in ACTH follows a longer time course (31). Also,
in blinded rats, the free-running phase shifts in food
intake and corticosterone rhythms follow the same time
course (32), and corticosterone rhythms disappear in rats
that have been food deprived for 5 or more days (27).
Further, pharmacological and surgical manipulations
that abolish adrenal rhythms also abolish feeding and
drinking rhythms in animals fed ad libitum. Treatment
with parachlorophenylalanine abolishes rhythms in food
and water intake (33) and in plasma and adrenal corti-
costerone but does not affect rhythms in ACTH (13, 34).
Lesions of the suprachiasmatic nuclei abolish circadian
rhythms in eating, drinking, locomotor activity, and
plasma corticosterone (35-37), but restricted feeding
schedules synchronize adrenal rhythms in these animals
(38).
Our experiments demonstrate also that the phase of
the light cycle affects adrenocortical rhythms. Adrenal
corticosterone concentration and content and adrenal
responsiveness to ACTH, but not plasma ACTH concen-
tration, are higher during the dark than during the light
phase of the light cycle, regardless of the time of feeding
in food-restricted rats (Fig. 4). Thus there are two inter-
acting influences on the daily variation in adrenal re-
sponsiveness to ACTH: the time of feeding and the light
cycle. Interactions of feeding effects with the light cycle
have also been reported by others (39, 40). By changing
adrenal responsiveness to ACTH, feeding and light cycles
amplify the adrenal expression of the circadian rhythm
in circulating ACTH by a mechanism that may be me-
diated by direct neural influences on the adrenal cortex
or by alterations in the circulating levels of other hor-
mones or metabolites.
358 WILKINSON, SHINSAKO, AND DALLMAN
In addition to the effects of feeding and light cycles on
adrenal responsiveness to ACTH, a third factor that
results in even greater excursions in the rhythm in plasma
corticosterone is the time of day difference in metabo-
lism, distribution, or binding of corticosterone (see Fig.
4). Plasma to adrenal corticosterone ratios during the
darkness are significantly higher than during the light
phase of the cycle, indicating more rapid disappearance
of corticosterone from the blood during the light than
during the darkness. This consistent finding (13) supports
previous reports of nycthemeral rhythms in the disap-
pearance rate of corticosterone from plasma (41).
Finally, the diet consumed had a significant effect on
mean daily levels of adrenal corticosterone, relative ad-
renal weight, and adrenal responsiveness to ACTH. Rats
were placed on the restricted feeding schedule at a time
of extremely rapid growth. In none of the experiments
was growth rate normal. From observation of the stom-
achs of rats at the end of the feeding period, it seems
likely that stomach capacity limited the amount of food
consumed during the 2-h period of food availability. It is
likely, therefore, that the three dietary groups differed
primarily in the calories consumed, and differences in
corticosterone levels and adrenal responsiveness may
reflect different degrees of chronic malnutrition (42). The
inverse relationship obtained between the mean insulin
levels for each dietary group and the corresponding mean
values of adrenal responsiveness to ACTH may be a
coincidental finding reflecting independent relationships
of insulin and adrenal responsiveness to caloric intake.
The low negative correlation obtained between insulin
levels and adrenal responsiveness to ACTH suggests that
insulin may exert a very minor inhibitory influence on
the adrenal cortex but that it is not the primary deter-
minant of adrenal responsiveness to ACTH. Results of
preliminary experiments directly testing the acute effect
of insulin on adrenal responsiveness to ACTH in vivo
and in vitro support this conclusion (Wilkinson, C. W.,
and M. F. Dallman, unpublished). Neither circulating
insulin levels nor circulating ACTH levels are good pre-
dictors of circadian variations in resting adrenal or
plasma corticosterone concentrations.
Acknowledgments
We thank Mr. Brian Matsumura for corticosterone assays and Ms.
Anne Aldrich for glucose assays.
References
1. Mills, J. N., Human circadian rhythms, Physiol Rev 46: 128,1966.
2. Migeon, C. J., A. B. French, L. T. Samuels, and J. Z. Bowers,
Plasma 17-hydroxycorticosteroid levels and leucocyte values in the
rhesus monkey, including normal variation and the effect of ACTH,
Am J Physiol 182: 462, 1955.
3. McNatty, K. P., M. Cashmore, and A. Young, Diurnal variation in
plasma cortisol levels in sheep, J Endocrinol 54: 361,1972.
Endo • 1979
Vol 104 • No 2
4. Guillemin, R., W. E. Dear, and R. A. Liebelt, Nycthemeral varia-
tions in plasma free corticosteroid levels of the rat, Proc Soc Exp
BiolMed 101: 394, 1959.
5. Halberg, F., J. H. Galicich, F. Ungar, and L. A. French, Circadian
rhythmic pituitary adrenocorticotropic activity, rectal temperature,
and pinnal mitosis of starving, dehydrated C mice, Proc Soc Exp
Biol Med 118: 414, 1965.
6. Johnson, J. T., and S. Levine, Influence of water deprivation on
adrenocortical rhythms, Neuroendocrinology 11: 268, 1973.
7. de Gasquet, P., and E. Pequignot, Changes in adipose tissue and
heart lipoprotein lipase activities and in serum glucose, insulin and
corticosterone concentrations in rats adapted to a daily meal, Horm
Metab Res 5: 440, 1973.
8. Krieger, D. T., Food and water restriction shifts corticosterone,
temperature, activity and brain amine periodicity, Endocrinology
95: 1195, 1974.
9. Nelson, W., L. Scheving, and F. Halberg, Circadian rhythms in
mice fed a single daily meal at different stages of lighting regimen,
JNutr 105: 171, 1975.
10. Cheifetz, P., N. Gaffud, and J. F. Dingman, Effects of bilateral
adrenalectomy and continuous light on the circadian rhythm of
corticotropin in female rats, Endocrinology 82: 1117, 1968.
11. Krieger, D. T., W. Allen, F. Rizzo, and H. P. Krieger, Characteri-
zation of the normal temporal pattern of plasma corticosteroid
levels, J Clin Endocrinol Metab 32: 266, 1971.
12. Meier, A. H., Daily variation in concentration of plasma corticoste-
roid in hypophysectomized rats, Endocrinology 98: 1475, 1976.
13. Dallman, M. F., W. C. Engeland, J. C. Rose, C. W. Wilkinson, J.
Shinsako, and F. Siedenburg, Nycthemeral rhythm in adrenal
responsiveness to ACTH, Am. J Physiol, in press.
14. Forsham, P. H., V. DiRaimondo, D. Island, A. P. Rinfret, and R. H.
Orr, Dynamics of adrenal function in man, Ciba Found Colloq
Endocrinol 8: 279, 1955.
15. Lentle, B. C, and J. P. Thomas, Adrenal function and the compli-
cations of diabetes mellitus, Lancet 2: 544, 1964.
16. L'age, M., W. Fechner, J. Langholz, and H. Salzmann, Relationship
between plasma corticosterone and the development of ketoacidosis
in the alloxan diabetic rat, Diabetologia 10: 131, 1974.
17. L'age, M., J. Langholz, W. Fechner, and H. Salzmann, Disturbances
of the hypothalamo-hypophysial-adrenocortical system in the al-
loxan-diabetic rat, Endocrinology 95: 760, 1974.
18. Lakatua, D. J., E. Haus, and F. Halberg, Habitual circadian timing
of growth hormone (STH), adrenocorticotrophic hormone (ACTH),
insulin, cortisol and glucose in human serum, In Aschoff, J., F.
Ceresa, and F. Halberg (eds.), Chronobiological Aspects of Endo-
crinology, Verlag, Stuttgart, 1974, p. 185.
19. Jolin, T., and A. Montes, Daily rhythm of plasma glucose and
insulin levels in rats, Horm Res 4: 153, 1973.
20. Bellinger, L. L., V. E. Mendel, and G. P. Moberg, Circadian insulin,
GH, prolactin, corticosterone, and glucose rhythms in fed and
fasted rats, Horm Metab Res 7: 132,1975.
21. Dallman, M. F., D. DeManincor, and J. Shinsako, Diminishing
corticotrope capacity to release ACTH during sustained stimula-
tion: the twenty-four hours after bilateral adrenalectomy in the rat,
Endocrinology 95: 65,1974.
22. Rees, L. H., D. M. Cook, J. W. Kendall, C. F. Allen, R. M. Kramer,
J. G. Ratcliffe, and R. A. Knight, A radioimmunoassay for rat
plasma ACTH, Endocrinology 89: 254, 1971.
23. Murphy, B. E. P., Some studies of the protein-binding of steroids
and their application to the routine micro and ultramicro measure-
ment of various steroids in body fluids by competitive protein-
binding radioassay, J Clin Endocrinol Metab 27: 973,1967.
24. Nie, N. H., C. H. Hull, J. G. Jenkins, K. Steinbrenner, and D. H.
Bent, Statistical Package for the Social Sciences, ed. 2, McGraw-
Hill, New York, 1975.
25. Vernikos-Danellis, J., E. Anderson, and L. Trigg, Changes in adre-
nal corticosterone concentration in rats: method of bio-assay for
ACTH, Endocrinology 79: 624, 1966.
26. Saffran, M., and A. V. Schally, In vitro bioassay of corticotropin:
modification and statistical treatment, Endocrinology 56: 523,
1955.
27. Inoue, K., K. Takahashi, and Y. Takahashi, Influence of change in
 FOOD AND ADRENAL RESPONSIVENESS TO ACTH
feeding regime and food deprivation on circadian rhythm of adrenal
cortical activity in rats, Folia Endocrinol Jap 52: 898, 1976.
28. Shiotsuka, R., J. Jovonovich, and J. A. Jovonovich, In vitro data 36.
on drug sensitivity: circadian and ultradian corticosterone rhythms
in adrenal organ cultures, In Aschoff, J., F. Ceresa, and F. Halberg
(eds.), Chronobiological Aspects of Endocrinology, Verlag, Stutt- 37.
gart, 1974, p. 255.
29. Zucker, I., Light-dark rhythms in rat eating and drinking behavior,
Physiol Behav 6: 115, 1971.
30. ter Haar, M. B., Circadian and estrual rhythms in food intake in 38.
the rat, Horm Behav 3: 213, 1972.
31. Siedenburg, F., J. Shinsako, and M. F. Dallman, Separation of
circadian rhythms in circulating ACTH and adrenal corticosterone 39.
(B) by a 180° phase shift in light cycle, Program of the 59th Annual
Meeting of The Endocrine Society, Chicago, 1977, p. 272 (Abstract).
32. Takahashi, K., K. Inoue, and Y. Takahashi, No effect of pinealec- 40.
tomy on the parallel shift in circadian rhythms of adrenocortical
activity and food intake in blinded rats, Endocrinol Jap 23: 417,
1976. 41.
33. Borbely, A. A., J. P. Huston, and P. G. Waser, Physiological and
behavioral effects of parachlorophenylalanine in the rat, Psycho-
pharmacologia 31: 131, 1973. 42.
34. Vernikos-Danellis, J., P. Berger, and J. D. Barchas, Brain serotonin
and pituitary-adrenal function, Prog Brain Res 39: 301, 1973.
35. Stephan, F. K., and I. Zucker, Circadian rhythms in drinking
359
behavior and locomotor activity of rats are eliminated by hypotha-
lamic lesions, Proc NatlAcad Sci 69: 1583, 1972.
Moore, R., and V. Eichler, Loss of circadian adrenal corticosterone
rhythm following suprachiasmatic lesions in the rat, Brain Res 42:
201, 1972.
Raisman, G., and K. Brown-Grant, The 'suprachiasmatic syn-
drome': endocrine and behavioural abnormalities following lesions
of the suprachiasmatic nuclei in the female rat, Proc R Soc Lond
[Biol] 198: 297, 1977.
Krieger, D. T., H. Hauser, and L. C. Krey, Suprachiasmatic nuclear
lesions do not abolish food-shifted circadian adrenal and tempera-
ture rhythmicity, Science 197: 398, 1977.
Philippens, K. M. H., H. von Mayersbach, and L. E. Scheving,
Effects of the scheduling of meal-feeding at different phases of the
circadian system in rats, J Nutr 107: 176, 1977.
Sulzman, F. M., C. A. Fuller, and M. C. Moore-Ede, Comparison of
synchronization of primate circadian rhythms by light and food,
Am J Physiol 234: R130, 1978.
Saba, G. C, P. Saba, A. Carnicelli, and V. Marescotti, Diurnal
rhythm in the adrenal cortical secretion and in the rate of metab-
olism of corticosterone in the rat, Acta Endocrinol 44: 409, 1963.
Rao, K. S. J., Endocrine responses to malnutrition in man. In
Assenmacher, I., and D. S. Farner (eds.), Environmental Endocri-
nology, Springer-Verlag, Berlin, 1978, p. 321.