Changes in Liver Lobule Glycogen Zonation During

Prolonged Fasting of Rats Previously Fed a 30%

Casein Diet and Adapted to a Controlled
Feeding Schedule1
WILLIAM

L. RICHARDS2

McArdle Laboratory for Cancer Research, 450 N. Randall

Avenue, University of Wisconsin, Madison, WI 53706

INDEXING
controlled

KEY WORDS

glycogen zonation

starvation

liver lobule

feeding and lighting

A recent study from this laboratory showed

that the daily cycle of liver glycogen concentration and liver lobule glycogen zonation
could be highly synchronized within the individual rats of a group if the rats were caged
individually,
fed a powdered
diet, and
adapted to a "2 + 22" controlled feeding and

followed by a resurgence to 20-40% of the

maximum levels observed in fed controls (25). Liver glycogen eventually reached a plateau of reaccumulation
and thereafter gradually decreased in conjunction with a depletion of body energy reserves (5). Partial
reaccumulation
of hepatic glycogen during

lighting schedule (1). In such rats, the daily

changes in glycogen zonation patterns were
shown to result from a proportionately higher
rate of glycogen turnover in periportal (P)
hepatocytes of the liver lobule than in midlobular (M) or centrilobular (C) hepatocytes
(P > M > C).
In earlier work with rats that were adapted
to diets containing 20-30% protein and 6070% carbohydrate,
prolonged
fasting re-

long-term fasting was also observed in mice

(6). None of these studies, however, determined whether liver glycogen changes during prolonged fasting were the result of uniform or asymmetrical glycogen concentration
changes within the liver lobule. The impor_
e 19g2American
Ins,.tute
ofNutrition
Received
forpublica,.on

^
tober
' Thisworkwas'"pp11
inP*
byN.I.H./N.C.I.
GrantsCA-HSM

I
j .
SUlted
m

(V.R.P.), CA-22484, and CA-07175.

Recipient of National Cancer Institute Fellowship 5-F32-CA05395-03

an

. .
initial

ill.depletion

r i
OI liver

i
glyCOgen

934

Downloaded from jn.nutrition.org by guest on March 15, 2014

ABSTRACT
Changes in the lobular distribution of liver glycogen were studied during
the prolonged fasting of young adult male Sprague-Dawley
rats that were previously
adapted to a 30% casein diet and to the 2 + 22 controlled feeding and lighting schedule.
The optical density of glycogen, revealed in fixed liver cryostat sections by the periodic
acid-Schiff procedure, was determined in the periportal (P), midlobular (M), and centrilobular (C) regions of the liver lobule at various times during a 196-hour fasting
period. The prolonged fast could be divided into 3 phases with respect to liver glycogen
variation: initial glycogen depletion, glycogen resurgence, and final glycogen depletion.
Glycogen was lost from all regions of the liver lobule during the initial glycogen depletion
phase. During the glycogen resurgence phase, a lobular glycogen concentration gradient
formed (P > M > C). The simultaneous occurrence of increasing liver glycogen, in
creasing liver tyrosine aminotransferase
activity, and decreasing body fat during gly
cogen resurgence suggests that the young adult rat does not spare body protein during
starvation. During the final glycogen depletion phase, liver glycogen was again present
in a lobular gradient (P > M > C) but the absolute amount of glycogen in each region
of the lobule was markedly reduced in comparison with rats killed during glycogen
resurgence.
J. Nutr. 112: 934-940, 1982.

LIVER GLYCOGEN ZONATION DURING STARVATION

tance of studies on changes in the lobular

distribution of liver enzymes and other con
stituents has been emphasized (1, 7). Thus,
in the present investigation, rats were adapted
to the 2 + 22 controlled feeding and lighting
schedule and were then subjected to pro
longed fasting in order to determine its effect
on liver lobule glycogen distribution during
the initial glycogen depletion, glycogen re
surgence, and final glycogen depletion phases
of starvation.
MATERIALS AND METHODS

filter paper, frozen on dry ice, and stored at

70
until cryostat sections were cut. The
remaining liver was quickly frozen in liquid
nitrogen and stored at 70
until chemical
glycogen assays were performed. Last of all,
the perirenal, retroperitoneal, and epididymal fat pads were resected and weighed and
their combined weights were calculated. The
prefeeding weight of rats that had been on
the 2 + 22 schedule for 27 days was 189
4 g. By the end of the experiment, body
weights of fed control rats and fasted rats
were 206 11 g and 127 3 g, respectively.
Glycogen determination. Methods for the
assay of glycogen in liver homogenates and
the histochemical demonstration of liver gly
cogen by the periodic acid-Schiff (PAS) pro
cedure were described previously (1).
Microdensitometry. The lobular distribu
tion of liver glycogen, revealed in cryostat
sections by the PAS reaction, was determined
by measuring optical densities at 560 nm
with the instrument described previously (1).
With a field aperture of 0.147 mm, four op
tical density readings were made adjacent to
the central vein and portal vein of a given
lobule and a third set of three readings was
made in a region midway (midlobular) be
tween each central vein/portal vein pair.
Three lobules were examined in the median
lobe and in the right lateral lobe of each rat.
For each rat, the optical density readings
from each lobular region (24 readings from
C and P, 18 from M) were averaged to obtain
the mean optical density of periportal, mid
lobular, and centrilobular glycogen. For the
3-6 rats of each group, the mean optical den
sities of glycogen in the P, M, and C regions
of each animal were used to calculate the
mean SE (8) for the group.
Tyrosine aminotransferase (TAT) assay.
Homogenates (5% wt/vol) of liver were
made in 50 mM KPO4, pH 7.5:0.05 mM pyridoxal phosphate: 1.0 mM disodium ethylenediaminetetraacetic acid: 100 mM KCL:1.0
mM dithiothreitol in a Potter-Elvehjem homogenizer. The homogenates were centrifuged for 1 hour at 105,000 X g, and the re
sultant supernatant solutions were assayed for
'The

30% casein diet, purchased from Teklad Test Diets (Madison. WI),

had the following composition in percent by weight: casein. 30; cornstarch.

15; sucrose. 40; hydrogenated vegetable oil. 10; salt miUSP XIV, 4; vitamin
fortification mi(Tetlad No. 40060), 1

Downloaded from jn.nutrition.org by guest on March 15, 2014

Animals. Sixty-six male rats (21 days postpartum) of the Sprague-Dawley strain
(Sprague-Dawley Co., Madison, WI) were
caged individually in a windowless air-con
ditioned room with an inverted and dis
placed lighting schedule in which lights were
on from 2030 to 0830 daily. The cages had
wire bottoms to minimize coprophagy, and
food was supplied in small metal cups set on
the cage floor. Water was available ad libi
tum throughout the experiment. The rats
were fed the powdered 30% casein diet3 de
scribed previously (1). Food was available ad
libitum for 1 day, during the first 8 hours of
the dark period (8 + 16 schedule) for 7 days,
and during the first 2 hours of the dark period
(2 + 22 schedule) for 26 days. On the 27th
day of the 2 + 22 schedule, 3 rats were killed
45 minutes before the beginning of the meal
and 57 rats were fed a final 2-hour meal. Six
rats, termed "fed controls", continued to be
fed a daily 2-hour meal throughout the
course of the experiment and were killed
along with rats that had been without food
for 196 hours. The 57 rats that received no
further meals were killed or died as follows:
three rats were found to be moribund late
in the fasting period and were killed at the
times indicated in figures 1 and 2; nine rats
died during the prolonged fast, between 100
hours and 196 hours, and were excluded from
the data set; six rats were killed at the last
time point of the prolonged fast (196 hours);
three rats, of the 39 remaining, were killed
at each of the 13 other time points indicated
in figures 1 and 2. Just after the rats had been
weighed and then killed by decapitation,
portions of the median and right lateral liver
lobes were resected, mounted on pieces of

935

936

RICHARDS
I DARK

3LIGHT

GLYCOGEN

DISTRIBUTION

PORTAL VEIN
<MID-LOBULAR
CENTRAL

TYROSINE

Downloaded from jn.nutrition.org by guest on March 15, 2014

LBULAR

AREA

VEIN

AMINOTRANSFERASE

II

20

60

80
HOURS

i
l
i
100
WITHOUT

l
i
120
FOOD

aoo

Fig. 1 Effect of prolonged fasting on liver glycogen concentration,

liver lobule glycogen distribution,
liver
tyrosine aminotransferase
activity, and fat pad weights. In the top panel, the dashed line labeled "if fed" represents
a repetition of the minimum and maximum values of glycogen concentration
observed during the first dark and
light periods in rats fed the final 2-hour meal. All values are the mean SE. A, individual rats killed when moribund
Six rats were killed at the final time point (196 hours) and three rats were killed at each other time point.

937

LIVER GLYCOGEN ZONATION DURING STARVATION

| DARK

40

| LIGHT

60
80
100 120 140
HOURS WITHOUT FOOD

160

180

200

>

Fig. 2 Effect of prolonged fasting on liver weight and liver weight as percent of body weight. All values are
the mean SE.A and A, individual rats killed when moribund. Six rats were killed at the final time point (196
hours) of the fed control group and of the fasting group. Three rats were killed at each other time point.

tyrosine aminotransferase
scribed previously (9).

activity

as de

RESULTS

In figure 1, the variations in liver glycogen

concentration and in lobular glycogen zonation patterns can be divided into two com
ponents: 1) the daily glycogen cycle in mealfed rats (a 2-hour meal followed by 22 hours
without food), 2) the variation of glycogen
during prolonged fasting (22-196 hours with
out food).
The period of prolonged fasting can be
subdivided into three phases: 1) initial gly
cogen depletion phase (22-52 hours without
food), 2) glycogen resurgence phase (52-106
or more hours without food), 3) final glyco
gen depletion phase (>106 hours without
food).
The daily glycogen cycle in meal-fed rats.
The daily cycle of glycogen followed the
pattern expected in rats adapted to the 2
+ 22 meal-feeding schedule and to the 30%
casein diet (fig. 1). The minimum daily gly
cogen level was attained just before the meal.
Relative to the end of the meal, the liver
glycogen level rose to a high value by 4 hours,
dropped to one-half of the highest value by
10 hours, and returned to its minimum level

by 21.25 hours. TAT activity was markedly

increased at 4 hours, returned to the prefeeding level by 10 hours, and remained at
this level through 21.25 hours. Liver weight
as percent of body weight (grams liver/100
g body weight) changed (fig. 2) in parallel
with changes in the liver glycogen concen
tration (fig. 1). Fat pad weights did not
change significantly during this period. At
each of the time points of the normal daily
glycogen cycle, the periportal and midlobular regions of the liver lobule contained sim
ilar amounts of glycogen while the centrilobular region contained slightly less glycogen
than the other two regions.
Period of prolonged fasting. Initial gly
cogen depletion phase (22-52 hours without
food). During the 22- to 52-hour period, gly
cogen was lost from all regions of the liver
lobule and became nearly undetectable (fig.
1). A peak of TAT activity occurred 28 hours
after the end of the final meal. This peak
coincided with the middle of the dark period,
as in fed rats, but the peak was of lower
magnitude. Thereafter, TAT activity de
creased markedly and no further cyclic
change was evident. Liver weight decreased
rapidly during this period (fig. 2) but fat pad
weights did not change (fig. 1).

Downloaded from jn.nutrition.org by guest on March 15, 2014

20

938

RICHARDS

centrilobular

region. X167.

Period of prolonged fasting. Glycogen re

surgence phase (52-106 hours without food).
During the 52- to 106-hour period, the liver
glycogen concentration (fig. 1) increased
from a nearly undetectable level to a level
that was about 20% of the maximum level
attained in fed rats and 1.4 times greater than
the minimum level found just before a meal
in rats on a regular meal-feeding schedule.
In the starving rat, glycogen resurgence oc
curred asymmetrically (figs. 1,3) in the three
regions of the liver lobule (P > M > C). The
maximum differential occurred after 82
hours without food at which time the con
centrations of glycogen in the midlobular and
centrilobular regions of the liver lobule were,
respectively, 66% and 47% of the concentra
tion in the periportal region. The resurgence
in liver glycogen was accompanied by an in
crease in liver TAT activity and by a decrease
in fat pad weights (fig. 1). Liver weight (fig.
2) remained at the decreased value attained
during the initial glycogen depletion phase
of prolonged fasting.
Period of prolonged fasting. Final gly
cogen depletion phase (>106 hours without
food). During this period, liver glycogen was
depleted faster in some animals than in oth

ers. Three rats, killed when moribund (143,

168, and 191 hours without food), had lower
liver glycogen levels (fig. 1), fat pad weights
(fig. 1), and liver weights (fig. 2) than nonmoribund rats that were killed at 196 hours.
Also between 100 and 196 hours nine rats
died, presumably from depletion of energy
sources. Because we could not determine the
length of time these animals had been dead
before discovery, they were excluded from
the data set.
During the 106- to 196-hour period, liver
TAT activity (fig. 1) increased to a level near
that observed 4 hours after the end of a meal
in rats on a regular meal-feeding schedule.
Data from the moribund rats killed at 143,
168, and 191 hours suggest that the increase
in TAT activity between 106 and 196 hours
was gradual. In rats killed at 196 hours, liver
glycogen (fig. 1) was again present in a lob
ular gradient (P > M > C) but the absolute
amount of glycogen in each region of the
lobule was markedly reduced in comparison
with rats killed during glycogen resurgence.
Fat pad weights decreased during the 106to 196-hour period (fig. 1) but liver weight
(fig. 2) remained at the decreased value that
was reached during the initial glycogen de-

Downloaded from jn.nutrition.org by guest on March 15, 2014

Fig. 3 Asymmetrical resurgence of liver lobule glycogen during prolonged fasting. In a rat that was killed 106
hours after the end of its final meal, a lobular glycogen gradient was revealed by the periodic acid-Schiff reaction
na 6 imliver cryostat section postfixed for 5 minutes at 2in Rossman's fluid (1) P, periportal region; C,

LIVER GLYCOGEN ZONATION DURING STARVATION

pletion phase and maintained during the glycogen resurgence phase of prolonged fasting.
DISCUSSION

dual rat variation with respect to the daily

glycogen variation cycle (1), they were sub
jected to a period of prolonged fasting. The
present data confirm the existence of initial
glycogen depletion, glycogen resurgence,
and final glycogen depletion phases during
prolonged fasting (2-5). Goldstein et al. pro
vided evidence that the partial hepatic gly
cogen reaccumulation during starvation could
be of physiological significance for glucose
homeostasis and showed that glycogen reaccumulation was associated with changes in
hepatic synthetase I and phosphorylase ac
tivities that would be expected to favor gly
cogen synthesis (2, 11).
Novel data in the present report demon
strate that glycogen is lost from all regions
of the liver lobule during the initial phase of
prolonged fasting but that the periportal cells
are the most active in accumulating glycogen
during the glycogen resurgence phase of pro
longed fasting. The resultant lobular gradient
of glycogen concentration (P > M > C) exists
throughout the phase of glycogen resurgence.
The asymmetry of lobular glycogen accu
mulation (P > M > C) during gluconeogenesis in rats receiving a high protein diet (1)
and in fasting rats during glycogen resurg
ence correlates well with the primary local
ization of gluconeogenic enzymes in peri
portal hepatocytes (7, 12, 13).
The three main precursors for gluconeogenesis are lactate, glycerol, and amino acids
(14). In the glycogen resurgence phase of
prolonged fasting, increased TAT activity
can be taken as an index of the utilization
of amino acids for gluconeogenesis (15), and
decreasing fat pad weight serves as evidence
of triglycridelipolysis to free fatty acids and
the gluconeogenic precursor glycerol. In
creases in TAT activity result from increases
in plasma amino acids and corticosterone
(15); Inoue et al. (16) have demonstrated that
plasma corticosterone gradually loses its circadian rhythm during fasting to assume a flat
pattern of higher levels between the 3rd and
7th day of total fasting. In the present in
vestigation, the simultaneous occurrence of
decreasing fat pad weight, increasing TAT
activity, and increasing liver glycogen during
the glycogen resurgence phase of prolonged
fasting suggests that amino acids (from proteolysis ) and glycerol (from lipolysis) become

Downloaded from jn.nutrition.org by guest on March 15, 2014

In a study of rat liver glycogen and glycogen zonation patterns during prolonged
fasting, it is desirable to synchronize, as much
as possible, the daily glycogen cycles of in
dividual rats prior to beginning the pro
longed fast. The present data, in conjunction
with previous data from this laboratory (1,
10), demonstrate the reproducibility of the
daily cycle of glycogen and glycogen zona
tion patterns in individually caged male
Sprague-Dawley rats adapted to a powdered
30% casein :55% carbohydrate diet and to the
2 + 22 controlled feeding and lighting sched
ule. With this regimen, daily changes in liver
glycogen can be characterized by a mini
mum concentration occurring between 1
hour before to 1 hour after the beginning of
the daily meal (1) and by a maximum con
centration occurring 4-6 hours after the end
of the 2-hour meal (10). Thus, in the present
experiment, the minimum and maximum
aspects of the daily glycogen cycle were con
cisely described by determining glycogen at
45 minutes before the beginning and 4 hours
after the end of the 2-hour meal, respec
tively. The minimum (1.4%) and the maxi
mum (7.8%) glycogen concentrations found
in the present experiment agree with the
minimum (1.1-1.7%) and the maximum
(8.4%) concentrations determined previously
under this regimen (1, 10). The small differ
ences between the glycogen concentrations
of the periportal, midlobular, and centrilobular regions of the liver lobule during this
daily cycle also confirm previous find
ings (1).
Liver weight expressed as percent of body
weight changed in parallel with daily gly
cogen concentration changes. The increased
level of TAT activity that was observed 4
hours after the end of a 2-hour meal may not
represent the true activity peak since an ear
lier report (10) showed that the peak of TAT
activity occurs about 2 hours before the peak
of glycogen concentration, i.e., at 2 hours
after the end of the meal.
After the rats were adapted to this re
producible regimen that minimizes indivi

939

940

RICHARDS

ACKNOWLEDGMENTS

The author wishes to thank Dr. Van Potter

for helpful discussions and critical reading
of the manuscript; Mrs. Nancy P. Dast for
excellent technical and editorial assistance;
Dr. James A. Gurr for performing tyrosine
aminotransferase assays; Mrs. Lona Barsness
(McArdle Histology Laboratory) for cryostat
section preparation, fixation, and staining;
Mr. Mark Kleinschmidt and Mr. Richard
Tefo for maintaining the feeding regimen
and assisting with animal kills; and Kristi
Traeder for accurate and rapid typing.
LITERATURE

CITED

1. Richards, W. L. & Potter, V.

microdensitometry
of glycogen
of rats adapted to a controlled
to 30, 60, or 90% casein diets.
85.

R. (1980) Scanning
zonation in the livers
feeding schedule and
Am. J. Anat. J57, 71-

2. Goldstein, D. E. & Curnow, R. T. (1978) Effect

of starvation on hepatic glycogen metabolism and
glucose homeostasis. Metab. Clin. Exp. 27, 315-323.
3. Harrison, M. F. (1953) Effect of starvation on the
composition of the liver cell. Biochem. J. 55, 204211.
4. Mirski, A., Rosenbaum, I., Stein, L. & Wertheimer,
E. (1938) On the behaviour of glycogen after diets
rich in protein and in carbohydrate. J. Physiol. (Lon
don) 92, 48-61.
5. Parrilla, R. (1978) Flux of metabolic fuels during
starvation in the rat. Pfluegers Arch. 374, 3-7.
6. Boutwell, R. K., Brush, M. K. & Rusch, H. P.
(1948) Some physiological effects associated with
chronic caloric restriction. Am. J. Physiol. 154, 517524.
7. Jungermann, K. & Sasse, D. (1978) Heterogeneity
of liver parenchyma! cells. Trends Biochem. Sci. (rf.
ed.) 3, 198-202.
8. Schefler, W. C. (1979) Statistics for the Biological
Sciences, 2nd ed., Addison-Wesley, Menlo Park, CA.
9. Diamondstone, T. I. (1966) Assay of tyrosine transaminase activity by conversion of p-hydroxyphenylpyruvate to p-hydroxybenzaldehyde.
Anal. Bio
chem. 16, 395-401.
10. Hopkins, H. A., Bonney, R. J., Walker, P. R., Yager,
J. D., Jr. & Potter, V. R. (1973) Food and light
as separate entrainment signals for rat liver enzymes.
Adv. Enzyme Regul. 11, 169-191.
11. Goldstein, D. E., Sutherland,
C. A. & Curnow,
R. T. (1978) Altered mechanism of glucagon-mediated hepatic glycogenolysis during long-term star
vation in the rat. Metab. Clin. Exp. 27. 1491-1498.
12. Guder, W. G. & Schmidt, U. (1976) The distri
bution of pyruvate kinase and phosphoenolpyruvate
carboxykinase (GTP) in the liver lobule of fed and
starved rats. Hoppe-Seylers Z. Physiol. Chem. 357,
1793-1800.
13. Schwarz, V. G. (1978) Quantitative investigations
of the zonal distribution of SDH, G6Pase, and malic
enzyme activity in liver parenchyma.
Acta Histochem. 62, 133-141.
14. Newsholme, E. A. & Start, C. (1973) Regulation
of carbohydrate metabolism in liver. In: Regulation
in Metabolism, chapt. 6, John Wiley and Sons, Lon
don.
15. Wurtman, R. J. (1970) Diurnal rhythms in mam
malian protein metabolism. In: Mammalian Protein
Metabolism, (Munro, H. N., ed.), vol. IV, pp. 445479, Academic Press, New York.
16. Inoue, K., Takahashi, K. & Takahashi, Y. (1976)
Influence of change in feeding regime and food de
privation on circadian rhythm of adrenal cortical
activity in rats. Folia Endocrino!. Jpn. 52, 898-907.

Downloaded from jn.nutrition.org by guest on March 15, 2014

simultaneously available as precursors for

gluconeogenesis in these starving young adult
rats. This supports the data of Parrilla (5),
which indicate that the young adult rat does
not spare body protein during starvation.
However, the absence of protein sparing in
these animals may be age-specific, i.e., it is
clear that death due to starvation is a function
of the size of the fat depots, which are neg
ligible in weanlings, moderate in young
adults, and increasing with age. [In rats over
2 years old and weighing up to 750 g, rats
could survive as long as 42 days during star
vation (Van R. Potter, personal communi
cation).]
The present data, demonstrating the asym
metry of liver lobule glycogen resurgence in
the fasting rat (P > M > C), are consistent
with the hypothesis (1) that periportal and
midlobular hepatocytes are more metabolically responsive and active than centrilobular
hepatocytes.