seas SCIENCE
5 July 1985, Volume 229, pp. 60-62

Reexamination of Glucose-6-Phosphatase Activity in the
Brain in Vivo: No Evidence for a Futile Cycle

Thomas Nelson, Giovanni Lucignani, Steven Atlas,
Alison M. Crane, Gerald A. Dienel, and Louis Sokoloff

Copyright © 1985 by the American Association for the Advancement of Science

A
Reexamination of Glucose-6-Phosphatase Activity in the

Brain in Vivo: No Evidence for a Futile Cycle

Abstract. Glucose-6-phosphatase activity in the rat brain in vivo was estimated by
measuring the differential loss of tritium and carbon-14 from the glucose pool labeled
by a mixture of [27H] glucose and [U-C] glucose. The results provide no evidence of
significant dephosphorylation of glucose-6-phosphate and do not support the hy-
pothesis of a futile cycle involving glucose-6-phosphatase activity in the brain.

Tuomas NELSON

Giovanni LUCIGNANI

STEVEN ATLAS

ALISON M. CRANE

GERALD A. DIENEL

Louis SOKOLOFF

Laboratory of Cerebral Metabolism,
National Institute of Mental Health,
Bethesda, Maryland 20205

It is widely believed that the mammali-
an brain is not a gluconeogenic organ and
contains negligible glucose-6-phospha-
tase (G6Pase) activity (/). The presence
of G6Pase in the brain has been demon-
strated by histochemistry (2, 3), but
quantitative assays of the enzyme activi-
ty in cerebral tissues in vitro have usual-
ly shown it to be a small fraction of that
of known gluconeogenic organs, such as
liver and kidney (4, 5), and supported the
view that G6Pase activity has little, if
any, role in the carbohydrate metabolism
of the brain. This belief was recently
challenged by Huang and Veech (6), who
reported that there was sufficient G6Pase
activity in the rat brain to maintain a
steady-state hydrolysis of glucose-6-
phosphate (G6P) equal to at least 35
percent of its rate of formation by hexo-
kinase-catalyzed phosphorylation of glu-
cose. Since the brain has a very high rate
of glucose utilization (7), a rate of de-
phosphorylation of G6P equal to one-
third the rate of glucose phosphorylation
by hexokinase would represent a level of
G6Pase activity approaching that found
in the rat liver after feeding (8).

Huang and Veech injected a mixture
of [2-H] glucose and [U-'*C]glucose into
one carotid artery of rats, removed the
brains by freeze-blowing (9) at various
times up to 5 minutes later, separated
free glucose from the tissue, and deter-

60

mined its 7H/‘C ratio by liquid scintilla-
tion counting. The principle behind these
procedures is as follows. Any (2-?H]glu-
cose metabolized as far as fructose-6-
phosphate in the glycolytic pathway
loses most of the 3H label, but ['4C]fruc-
tose-6-phosphate retains the “C (8).
Some fructose-6-phosphate is converted
back to G6P by reversal of the rea tion
catalyzed by hexosephosphate isomer-
ase, and, if there is G6Pase activity, the
glucose moiety is returned to the free
glucose pool with its }4C but without 37H.
The ?H/4C ratio in the free glucose pool
should then decline progressively with
time if there is G6Pase activity. Huang
and Veech reported such a decline, sug-
gesting previously unsuspected levels of
G6Pase activity in the brain.

A key issue is the purity of the free
glucose pool in which they measured the
3H/4C ratio. Most products of glucose
metabolism beyond the G6P step lose the
3H but not the C label; if any of these
products were to contaminate the glu-
cose fraction, they would lead to low 7H/
4C ratios. In their experiments Huang
and Veech relied on Dowex 1-formate
and Dowex 1-borate column chromatog-
raphy of perchloric acid extracts of the
brain tissue for purification of the glu-
cose. In a subsequent study (/0) they
added derivatization of the glucose frac-
tion by hexokinase-catalyzed phospho-
rylation.

We repeated the experiments of
Huang and Veech (6) but with special
efforts to ensure the purity of the glucose
fraction in which the *H/'*C ratio was
measured. Normal male Sprague-Daw-
ley rats (330 + 9 g) were anesthetized
with pentobarbital (30 mg/kg, intraperi-
toneally), and one femoral artery and
one external carotid artery were cathe-

terized in each animal. The carotid cath-
eter was inserted 5 mm past the carotid
bifurcation into the common carotid ar-
tery. Both catheters were secured by
ligatures. In one group of seven rats the
internal carotid artery contralateral to
the side of catheterization was ligated as
in the studies of Huang and Veech (6).
Because these animals occasionally ex-
hibited neurological and behavioral de-
fects at the time of the experiment, a
second group of eight animals was simi-
larly prepared except that the contralat-
eral internal carotid artery was not per-
manently ligated but was encircled with
a loose ligature that was temporarily
pulled taut during injection of the labeled
glucose through the carotid catheter to
ensure bilateral distribution of tracer.

Approximately 24 hours after surgery
a mixture of 50 pCi of {2-*H]glucose, 5
wCi of [U-4C]glucose (//), and 5.1 mM
glucose in 0.05 ml of 0.9 percent saline
was injected through the carotid cathe-
ter. At various times between 2 and 7
minutes after the injection the brain was
removed by freeze-blowing (9). Several
blood samples were drawn from the fem-
oral artery at various times during the
interval between the injection of labeled
glucose and freeze-blowing.

The frozen brains were powdered un-
der liquid nitrogen in a cryostat at —35°C
and perchloric acid (0.6M) extracts of
the brain tissue and plasma were pre-
pared. Acidic metabolites were removed
from the neutralized supernatant frac-
tions by passage through columns con-
taining 2 ml of Dowex AG 1-X8 formate
(200 to 400 mesh). After adjustment of
the pH of the effluent to 4.0, basic
metabolites were removed by cation-ex-
change chromatography on columns
with 2 ml of Dowex AG 50-X8 H* (200 to
400 mesh). The final effluent contained
approximately 96 percent of the glucose
originally present in the perchloric acid
extracts. Tritiated water released by the
metabolism of [2-*H]glucose was also
present.

The extracts, now depleted of all an-
ionic and cationic metabolic products of
glucose metabolism, were evaporated to
dryness to eliminate the PHJH,O and
chromatographed as 4-cm bands on
Whatman 3MMChr paper in the ascend-
ing direction with a solvent system con-
sisting of isobutyric acid, water, and
concentrated ammonium hydroxide
(66:33: 1 by volume). ['*C]Glucose stan-
dards were chromatographed in parallel
lanes on the same chromatographic pa-
pers. Bands migrating to the same posi-
tion relative to the solvent front (Rp) as
authentic ['4C]glucose were eluted with
water; this eluate contained 25 to 40

SCIENCE, VOL. 229
percent of the total 'C applied to the
chromatogram. Half of each of these
samples was assayed for *H and '4C
concentrations by liquid scintillation
counting calibrated with internal PH]to-
luene and ['4C]toluene standards; the
other halves were converted to G6P by
incubation with adenosine triphosphate,
yeast hexokinase, and MgCl, and chro-
matographed on paper as before. Bands
migrating with the same R; as authentic
[*C]G6P were eluted with water and
assayed to determine the 7H/"C ratios.

The °H/'4C ratios of the plasma and
brain extracts were normalized to the
3H/"4C ratio of the injectant used in each
experiment. The *H/'4C ratios of the
injectants were determined in samples
purified by the same procedures used for
the plasma and brain extracts to elimi-
nate radiochemical impurities that might
have contaminated the injected labeled
glucose. Best-fitting straight lines repre-
senting the normalized *H/"‘C ratios of
the labeled glucose and its G6P deriva-
tive with respect to time were calculated
by least-squares regression.

The results for the two groups did not
differ significantly and therefore were
pooled for statistical analysis. The glu-
cose fractions separated from plasma
showed a slight, progressive decline in
the 7H/"4C ratio below that of the injec-
tant over the 2- to 7-minute interval after
injection (Fig. 1A); this decline probably
reflects the consequences of glucose me-
tabolism throughout the body. The glu-
cose fractions from the brain showed a

3H7/74C in plasma
SH7'4c in injectant
S 2 :
@
—
«
>
@
e
eo
»

 

3H/'4¢ in brain
3H/'4c in injectant
° ° 7
o @ °
—_—_—_S
wo
°°
ae
eo
°y

i

y = 0.919 - 0.007x
r= -0.316 (P= 0.08)

y = 0.991 -0.009x
r=-0.536 (P= 0.04)

slightly greater progressive decline in
3H/"*C ratios, but the difference was not
statistically significant (Fig. 1B), indicat-
ing no differential loss of 7H and 'C in
the glucose pool in the brain.

To ensure that the °H/"4C ratios repre-
sented uncontaminated glucose, portions
of the fractions derived from plasma,
brain, and injectants were derivatized to
G6P by hexokinase-catalyzed phospho-
rylation. The labeled G6P was separated
by paper chromatography and assayed to
determine the *H/'4C ratios. The frac-
tions from plasma so treated showed a
statistically insignificant, time-depen-
dent trend toward reduced 7H/"4C ratios
(Fig. 2A), and the fractions from brain
tissue exhibited no greater change in 7H/
'4C ratio with time (Fig. 2B). In contrast
to the results of Huang and Veech (6),
there was no evidence of differential loss
of 7H and '4C from glucose in the brain
after intracarotid injection of [2-2H]glu-
cose and [U-“C]glucose.

It appears that Huang and Veech inad-
equately purified the fraction extracted
from brain tissue which they designated
as glucose and in which they measured
the 3H/'4C ratio. The labeled glucose in
brain tissue gives rise to many metabolic
prc lucts, most of which lose the 7H but
not '4C. Experiments in our laboratory
have confirmed their results but have
demonstrated that the sequential column
chromatography of perchloric acid ex-
tracts of brain tissue on Dowex AG 1-X8
formate and Dowex AG 1-X8 borate
does not isolate glucose uncontaminated

eel eo &*
>

»
t

0.84

0.6]

a
———_—— oF
6 8 0

 

Time (minutes)

by other labeled products. The eluate
from the Dowex AG 1-X8 borate column
does have a low *H/'*C ratio but con-
tains several labeled components other
than glucose when chromatographed on
paper. At 7 to 8 minutes after the injec-
tion, 40 to 50 percent of its '4C content
are in contaminants that contain little
tritium, but the 7H/'4C ratio in its glucose
component is similar to that of the injec-
tant. When the eluate from the Dowex
AG 1-X8 borate column is chromato-
graphed on a cation-exchange column,
the *H/*C ratio in the effluent is in-
creased. Treatment of the effluent with
glucose oxidase, which converts glucose
to gluconic acid, and separation of this
derivative by anion-exchange chroma-
tography raises the 7H/'4C ratio to that of
the plasma and close to that of the injec-
tant. Hexokinase-catalyzed phosphoryl-
ation of the eluate from the Dowex AG 1-
X8 borate column is inadequate to en-
sure purity of the glucose; hexokinase is
less specific than glucose oxidase and
would phosphorylate any fructose or glu-
cosamine that might be present in that
fraction.

Our results do not support the finding
of significant G6Pase activity in the brain
in vivo. The activity of the small amount
of enzyme that is present is limited by
intracellular compartmentation. In the
cell, G6Pase is present on the inner sur-
faces of the cisterns of the endoplasmic
reticulum (ER) (3, 12); the G6P is formed
in the cytosol. In tissues with high glu-
coneogenic activity there is a specific

y =1.060- 0.011x
r=-0.32 (P=0.09)

Ad,

y= 1.022-0.011x
r= -0.359 (P= 0.16)

Fig. 1 (left). Time course of 3H/'4C ratios in glucose purified from plasma (A) and brain (B) after intracarotid injection of a mixture of [2-
3H]glucose and {U-'C]glucose. Symbols: (A) results for animals with carctid ligation contralateral to side of carotid catheterization and (@)
results for animals with contralateral carotid obstruction only during the injection. The r represents the product-moment correlation coefficient of

the normalized 7H/'4C ratio on the ordinate with respect to time; P represents the probability value of the correlation coefficient.

Fig. 2

(right). Time course of H/C ratios in G6P derivatized from the labeled glucose isolated from plasma (A) and brain (B) in Fig. 1. The symbols are

explained in the legend to Fig. 1.
5 JULY 1985

61
62

carrier in the ER that transports the
substrate across the membrane to the
phosphatase (12, 13). Karnovsky et al.
(/4) found that this carrier is absent in
the brain and that G6P gains access to
the phosphatase only by slow diffusion
across the ER membrane. This would
further slow the phosphohydrolytic ac-
tivity of whatever G6Pase is present in
the brain.

The report of Huang and Veech (6) has
led to debate on the role of G6Pase in the
brain and to speculations about futile
cycles in cerebral tissue (/5). It has also
been used to argue against the validity of
the deoxyglucose method for measuring
local utilization of glucose in the brain in
animals and humans (6, /4), a method
that in its earliest form assumed negligi-
ble loss of deoxyglucose-6-phosphate in
the brain during the experimental period
(16). It now seems that such speculations
and extrapolations are without founda-
tion.

References and Notes

1, H. G. Hers and C. De Duve, Bull. Soc. Chim.
Biol. 32, 20 (1950); G. Weber and A. Cantero,
Cancer Res. 15, 105 (1955); H. G. Hers, Le
Métabolisme du Fructose (Editions Arscia,
Brussels, 1957), p. 102; H. A.:Krebs and M.
Woodford, Biochem. J. 94, 436 (1965); K. G.
Prasannan and K. Subrahmanyam, Endocrinol-
ogy 82, 1 (1968); M. C. Scrutton and M. F.
Utter, Annu. Rev. Biochem. 37, 249 (1968); W.
Colilla, *R. A. Jorgenson, R. C. Nordlie, Bio-
chim, Biophys. Acta 377, 117 (1975); L. Stryer,
Biochemistry (Freeman, San Francisco, 1975),
p. 385; A. L. Lehninger, Biochemistry (Worth,
New York, ed. 2, 1975); p. 628; A. L. Lajtha, H.
S. Maker, D. O. Clarke, in Basic Neurochemis-
try, G. Siegal et al., Eds. (Little, Brown, Bos-
ton, ed. 3, 1981); p. 335; E. Smith et al., Eds.,
Principles of Biochemistry: General Aspects
(McGraw-Hill, New York, ed. 7, 1983), p. 430.

2. H. B. Tewari and G. H. Bourne, J. Histochem.
Cytochem. 11, 121 (1963); A. Petrescu and M.
Alexianu, Rev. Roum. Neurol. 2, 321 (1965); N.
N. Sharma, Acta Histochem. 27, 165 (1967); S.
I. Rosen, ibid. 36, 44 (1970); H. R. Stephens and
E. B. Sandborn, Brain Res. 113, 127 (1976); S.
Y. A. Al-Ali and N. Robinson, Histochem. J.
14, 311 (1982); R. D. Broadwell, A. M. Cataldo,
ec a J. Histochem. Cytochem. 31, 818

. S. Y. A. Al-Ali and N. Robinson, Histochemis-
try 72, 107 (1981).

. J. M. Anchors and M. L. Karnovsky, J. Biol.
Chem. 250, 6408 (1975).

. R. C. Nordlie, Life Sci. 24, 2397 (1979).

. M.-Ta. Huang and R. L. Veech, J. Biol. Chem.
257, 11358 (1982).

. B. K. Siesjé, Brain Energy Metabolism (Wiley,
New York, 1978).

. J. Katz, P. A. Walls, R. Rognstad, J. Biol.
Chem. 253, 4530 (1978).

. R. L. Veech et al., J. Neurochem. 20, 183
(1973); R. L. Veech and R. A. Hawkins, in
Research Methods in Neurochemistry, N.
Marks and R. Rodnight, Eds. (Plenum, New
York, 1974), vol. 2, pp. 171-182.

10. M.-Ta Huang and R. L. Veech, Trans. Am. Soc.

Neurachem. 15 (No. 1), 188 (1984).

11. The radionuclides p-[2-H(N)]glucose (specific
activity, 24.0 Ci/mmol), p-[U-“C]glucose (14.4
mCi/mmol), and v-[1-'C]G6P (51.4 mCi/mmol)
were purchased from New England Nuclear.

12. L. M. Ballas and W. J. Arion, J. Biol. Chem.
252, 8512 (1977).

13. W. J. Arion et al., ibid. 255, 10396 (1980).

14. M. L. Karnovsky, personal communication.

15. J. L. Fox, Science 224, 143 (1984).

16. L. Sokoloff et al., J. Neurochem. 28, 897 (1977);

, in Advances in Neurochemistry, B. W.

Agranoff and M. H. Aprison, Eds. (Plenum,

New York, 1982), vol. 4, pp. 1-82.

7 February 1985; accepted 19 April 1985

wo ont AN SF WwW

 

SCIENCE, VOL. 229