Download Amino acid metabolism in human subcutaneous adipose tissue in vivo

Clinical Science (1991) 80, 471-474
471
Amino acid metabolism in human subcutaneous adipose
tissue in vivo
K. N. FRAYN, K. KHAN*, S. W. COPPACK
AND
M. EUA*
Sheikh Rashid Diabetes Unit, Radcliffe Infirmary, Oxford, UK and *MRC Dunn Clinical Nutrition Centre, Cambridge, UK
(Received 28 September/5 December 1990; accepted 20 December 1990)
SUMMARY
1. Arteriovenous differences for alanine, glutamate
and glutamine were measured across subcutaneous
adipose tissue and forearm muscle in normal subjects.
2. After an overnight fast, adipose tissue showed net
production of alanine and glutamine and uptake of
glutamate in each of 11 subjects.
3. In seven subjects, adipose tissue blood flow was
measured and the measurements were continued for 6 h
after eating a mixed meal. The pattern of amino acid
metabolism across the adipose tissue was remarkably
little disturbed after the meal, except for a short period of
apparent uptake of alanine as the concentration of that
amino acid rose.
4. The pattern of amino acid metabolism across
adipose tissue was qualitatively similar to that across the
forearm, although it differed quantitatively in that glutamate uptake was more prominent (compared with glutamine release) in the adipose tissue.
5. The rates of alanine and glutamine release observed
suggest that adipose tissue may play a substantial role in
the whole-body production of these amino acids.
Key words: adipose tissue, alanine, amino acids, forearm
metabolism, glutamate, glutamine.
INTRODUCTION
Although adipose tissue was at one time regarded as a
rather inert repository for excess energy, it is now recognized to have a dynamic pattern of metabolism, playing a
major role in regulating the flow of lipid energy in the
body on both a minute-to-minute and a longer term basis.
It has never been clear, however, whether adipose tissue
plays a quantitatively significant role in other aspects of
metabolism in the body. Glucose uptake by adipose tissue
Correspondence: Dr Keith N. Frayn, Sheikh Rashid Diabetes
Unit, Radcliffe Infirmary, Oxford OX2 6HE, UK
is insulin-sensitive in vitro [1], but adipose tissue probably
plays a very minor role in the whole-body disposal of a
glucose load [2, 3]. Similarly, although adipocytes consistently produce lactate in vitro [2], the contribution of
adipose tissue to the total amount of lactate released by
peripheral tissues in vivo is probably small [3].
Adipose tissue has long been recognized to have a
characteristic pattern of amino acid metabolism, but its
role in whole-body amino acid metabolism is again not
clear. Amino acids may act as substrates for lipogenesis
[4,5]. Branched-chain amino acids may be oxidized, and
their nitrogen transferred to form other amino acids for
release [6, 7]. Snell & Duff [6], for instance, showed
release of alanine from rat epididymal fat pads, and
suggested that this might make a significant contribution
to whole-body alanine production. Tischler & Goldberg
[7] have shown release of both alanine and glutamine by
rat adipose tissue. In contrast, isolated rat adipocytes have
been shown to utilize exogenous glutamine [8] at rates
which, if extrapolated to the whole body, appear physiologically significant.
In man, alanine and glutamine are quantitatively the
most important amino acids released by muscle both
before and after ingestion of a protein or mixed meal or
administration of glucose [9-11]. Glutamate, in contrast,
is consistently taken up by muscle, where it may enter the
citric acid cycle as 2-oxoglutarate or act as a precursor for
glutamine synthesis. Whether similar exchanges occur in
adipose tissue in vivo is unknown. We have therefore
investigated the quantitative exchange of these three
metabolically important amino acids across human
adipose tissue in vivo, before and after ingestion of a
mixed meal. The exchange was assessed by measurement
of arteriovenous differences across the subcutaneous
adipose depot together with measurement of blood flow.
For comparison, the simultaneous flux of these amino
acids was measured across the forearm.
Other measurements made in the same studies,
together with the measurements of blood flow, are
reported elsewhere [12].
K. N. Frayn et al.
472
METHODS
Seven normal subjects (four female, three male) were
studied (age range 28-42 years, body mass index 19-29
kg/m") after an overnight fast, having previously consumed their usual weight-maintaining diet. Their body fat
content, estimated from skinfold thicknesses as
described by Durnin & Womersley [13], was 10-19 kg;
their muscle mass, estimated as 40% body weight, was
19-29 kg. (These values were only used for approximate
whole-body extrapolations.) Cannulae were introduced,
as ~escribed previously [3,12], into a vein draining a hand
which was then warmed to provide arterialized blood,
retrogradely into an antecubital vein draining forearm
muscle, and into a vein draining the subcutaneous adipose
tissue of the anterior abdominal wall. A cuff was inflated
to 200 mmHg around the wrist to exclude hand blood
flow before taking samples from the antecubital vein.
Forearm blood flow was measured by strain-gauge
plethysmography [14] and adipose tissue blood flow by
clearance of 133Xe [15] immediately after taking each
sample. An additional four normal subjects (one male,
three female) aged 21-42 years, had arterialized blood
samples and adipose venous blood samples only taken
after overnight fast; blood flow was not measured.
After a rest period of at least 30 min, blood samples
were taken from all three sites as near simultaneously as
possible. Three sets of basal blood samples were taken;
mean results from these are presented below. Then,
between 0 and 20 min, the subjects ate a mixed meal as
described previously [12]. The meal had an energy value
Table 1. Amino acid concentrations and fluxes in seven
normal subjects
Values are means ± SEM. Post-prandial tissue fluxes are
based on the area under the flux-time curve from 0 to
360 min. All arteriovenous concentration differences
w~re sig~fi~antly different from zero (P<0.05 by
WIlco~on s Signed-rank test; P< 0.01 by paired r-test),
The differences between average post-prandial tissue flux
and basal flux were not significant.
Alanine Glutamine Glutamate
After an overnight fast
Bloodconcn. (,amol/I)
Arterialized
248± 16 509±21
Forearmvenous
290± 17 551 ± 18
Adiposevenous
271 ± 18 538±20
Tissueflux (nmol min- 1
100 ml: 1 tissue)
Forearm
97± 18
157±24
Adiposetissue
70± 15
90±22
Estimated whole-body
tissue flux (,amol/min)
Muscle
23±4
39±8
Adiposetissue
9±2
12±3
Post-prandial
Average tissue flux (nmol
min- [ 100 ml: 1 tissue)
Forearm
71±19139±35
Adiposetissue
51 ± 18
98 ± 11
167 ± 11
158 ± 11
152 ± 11
38±8
42±7
9±2
6±1
of 3.1 MJ, of which 41% came from fat and 47% from
carbohydrate. Further samples were taken (with respect to
the time of starting the meal) at 30, 60, 90 and 120 min,
then at hourly intervals until 6 h.
Blood samples (250 ,ul) were immediately deproteinized in 500 ,ul of sulphosalicylic acid (3.5%, w/v).
(Blood and acid were weighed.) After centrifugation, the
supernatant was frozen at - 70°C, before estimation of
alanine, glutamate and glutamine concentrations by
enzymic methods which were adapted for a Cobas Bio
centrifugal analyser (Roche Diagnostics, Welwyn Garden
City, Herts, U.K.) [11].
Tissue uptake or release of amino acids (expressed as
nmol min- I 100 ml [ tissue) was calculated as the
product of arteriovenous difference (,umolfl) and blood
flow (ml min -[ 100 ml- I tissue). Areas under curves were
calculated by using a trapezoidal method. Extrapolations
to whole-body rates of production and utilization (which
must be treated with caution as discussed further below)
were based on the estimates of fat and muscle mass for
each individual. (The amount of body fat as estimated
from skinfold thicknesses is not necessarily equal to the
amount of adipose tissue, but again these values were only
used for approximate extrapolations.) All results are
presented as means ± SEM for n = 7, unless otherwise
stated.
RESULTS
Fasting state
In all 11 subjects studied, alanine and glutamine concentrations in the blood draining adipose tissue were
higher than in arterialized blood, whereas the reverse was
t~e for glutamate (P< 0.001 for each amino acid by
SIgn test). Further analysis refers only to the seven
subjects in whom blood flow was measured, and from
whom antecubital samples were taken.
Concentrations of both alanine and glutamine were
higher in forearm venous and adipose venous blood than
in arterialized blood in all subjects (Table 1). For glutamate, the concentration was higher in arterialized blood
than in venous blood from either site; concentrations in
forearm venous and adipose venous blood were similar.
T.he ratio of glutamine release to glutamate uptake was
different for the two tissues. Mean values for the ratio
were: adipose tissue, 2.7 ± 1.0; forearm, 6.1 ± 2.4
(P< 0.05 by Wilcoxon's signed-rank test).
. Extrapolations to the whole-body (assuming adipose
tIssu.e and muscle masses all to behave like the depots
studied here) showed that in the overnight-fasted state
the potential contribution of adipose tissue was not
insignificant in comparison with that of muscle particularly in the case of glutamate uptake (Table 1). '
Responses to ingestion of a meal
50 ± 11
78+21
After the meal, alanine concentrations in arterialized
blood (Fig. 1) rose from a mean baseline value of
248 ± 16 ,umol/l to a peak of 414 ± 25 ,umol/l at 90 min
Amino acid metabolism in adipose tissue
200
(a)
::::.
'0
8
.3 175
=
u
l::
0
u
v
OJ
150
1
"0
0
0
125
~
100
600
(b)
l __l __.L--l
::::.
'0
8
.3 550
...
=
'"
473
under the curve for alanine release [nmol min - I 100 ml" I
tissue x time (min)] was calculated from 0 min to 360 min,
as this measure is less susceptible to non-steady state
effects (Table 1). For both tissues, the areas were significantly greater than zero (P<0.05 by Wilcoxon's signedrank test) showing consistent release averaged over the
whole of the post-prandial period. Averaged rates of
production were very similar to those in the baseline
period (Table 1).
Blood concentrations of both glutamine and glutamate
(Fig. 1) were relatively steady after the meal. Again, in
each case the areas under the curves (for release or uptake
respectively, from 0 to 360 min) were consistently different from zero (P< 0.05 by Wilcoxon's signed-rank test).
Rates of glutamine release were steady throughout for
both tissues. There was a tendency (not significant) for
glutamate uptake by both tissues to increase after the
meal (e.g. the mean adipose tissue uptake rate at 60 min
was 133 ± 60 nmol min- 1 100 ml- 1 tissue).
u
l::
0
u
v
l::
.~
~
"0
0
0
500
DISCUSSION
450
~
400
450
(e)
<;
'0
§.
= 350
u
l::
0
o
V
.~
-a
250
"0
0
0
~
150
L--'---I~_L.-_..L..._...L..._....L.._.....L.._.....J
-40 0
60
120
180
240
300
360
Time (min)
Fig. 1. Blood concentrations of glutamate (a), glutamine
(b ) and alanine (c) in seven normal subjects, after an overnight fast ( - 40 to 0 min) and after ingestion of a mixed
meal at time zero. Results are shown as means ± SEM.
Some error bars are omitted for clarity. A--A,
Arterialized blood; .- - -., forearm venous blood;
• - - . adipose venous blood. Alanine concentrations in
forearm venous blood were very similar to those in
adipose venous blood and are not shown for clarity.
(P<0.02 by Wilcoxon's signed-rank test). There was a
transient switch to apparent alanine uptake in both forearm and adipose tissue, although this may represent an
artefact of the non-steady state. For the last 4 h of the
study, both tissues were again releasing alanine. The area
Although there has been controversy, as outlined above,
about the handling of glutamine by adipose tissue in vitro,
these results show that adipose tissue in vivo is a consistent net exporter of both alanine and glutamine. It
should be remembered that results obtained from
measurement of arteriovenous differences reflect net
exchange across a tissue, and the possibility of simultaneous uptake and release of amino acids cannot be
excluded. Although the tissue studied by this technique
includes a proportion of skin, the metabolic contribution
of skin appears to be small [3]. The possible role of cells
other than adipocytes in the exchanges we observed is
considered below.
The metabolic pattern of the adipose tissue, in terms of
these three amino acids, was qualitatively similar to that of
the forearm, although the relative rates of glutamate
uptake and of alanine or glutamine release were different.
It is unlikely, for anatomical reasons as well as from the
composition of the blood, that there is significant contamination of the adipose tissue venous blood by muscle
drainage [3]. We therefore conclude that human subcutaneous adipose tissue in vivo is a net exporter of both
alanine and glutamine, and a consumer of glutamate.
In whole-body terms, the contribution of adipose tissue
appears to be significant. Extrapolations from individual
adipose tissue or muscle beds to the whole body can only
be taken as approximate because of regional differences
in metabolism and blood flow (see references cited in [12]).
These extrapolations, however, suggest that after an overnight fast adipose tissue contributes about one-third as
much as muscle to whole-body alanine and glutamine
production, and more than one-half as much as muscle to
glutamate uptake. Recent measurements of muscle blood
flow by the xenon technique suggest that plethysmography, as used in these studies, may overestimate forearm muscle blood flow (A. Kurphad & M. Elia,
unpublished work). In this case, the role of adipose tissue
474
K. N. Frayn et al.
relative to that of muscle would be even greater than estimated here.
After the meal, although blood alanine concentrations
rose, there was no indication that this reflected increased
release from either muscle or adipose tissue. In fact there
was a period of apparent alanine uptake by both tissues,
as noted in other work [11]. The calculations based on
areas under the flux-time curves for the post-prandial
period showed average rates of release very similar to
those during the baseline period. Although average rates
of amino acid release and uptake across the tissues
changed little after the meal, the role of adipose tissue
relative to muscle in glutamate uptake became somewhat
greater.
Since the nitrogen leaving adipose tissue in the form of
alanine (1 mol per mol) plus glutamine (2 mol per mol) is
greater than the nitrogen entering in the form of glutamate
(l mol per mol), it appears that there is another source of
nitrogen; for example, the catabolism of branched-chain
amino acids as shown in vitro [6, 7] or the catabolism of
proteins. Similar arguments apply to the carbon skeletons
of alanine and glutamine, which may be derived from
other amino acids or from glucose.
Adipose tissue in vivo is not a pure preparation of
adipocytes. The contribution of skin is inseparable, albeit
small. Several other cell types in adipose tissue may be
involved in amino acid exchange, including fibroblasts,
vascular (endothelial) cells, macrophages, erythrocytes
and leucocytes. However, since several of these cell types,
including skin, possess a high glutaminase activity, and/or
have been shown to use glutamine avidly in vitro, it seems
unlikely that they could be responsible at least for the
glutamine production [16-18]. No production of alanine
is seen on storage of blood for several minutes [19], so
again it seems unlikely that the changes could be caused
by the metabolic activity of blood cells during passage
through the tissue.
In conclusion, we find human adipose tissue in vivo to
have a pattern of amino acid metabolism which is qualitatively similar to that of skeletal muscle. Rates of alanine
and glutamine release and of glutamate uptake by adipose
tissue are not insignificant compared with those of skeletal muscle.
ACKNOWLEDGMENTS
We thank S. M. Humphreys and M. J. Kirk for technical
assistance. The Sheikh Rashid Diabetes Unit is supported
by the Oxford Diabetes Trust. S.W.c. held an MRC Training Fellowship.
REFERENCES
1. Taylor, R, Husband, DJ., Marshall, S.M., Tunbridge,
WM.G. & Alberti, KG.M.M. Adipocyte insulin binding and
insulin sensitivity in 'brittle' diabetes. Diabetologia 1984;
27,441-6.
2. Marin, P., Rebuffe-Scrive, M., Smith, U. & Bjorntorp, P.
Glucose uptake in human adipose tissue. Metab. Clin. Exp.
1987;36,1154-60.
3. Frayn, KN., Coppak, SW., Humphreys, S.M. & Whyte, P.L.
Metabolic characteristics of human adipose tissue in vivo.
Clin. Sci. 1989; 76,509-16.
4. Herrera, M.G. & Renold, AE. Amino acid and protein
metabolism. In: Renold, AE. & Cahill, G.F., eds. Washington, D.C.: American Physiological Society, 1965: 375-83.
(Handbook of physiology, section 5. Adipose tissue.)
5. Feller, D.D. Conversion of amino acids to fatty acids. In:
Renold, AE. & Cahill, G.F., eds. Washington D.C.:
American Physiological Society, 1965: 363-73. (Handbook
of physiology, section 5. Adipose tissue.)
6. Snell, K & Duff, D.A Alanine release by rat adipose tissue
in vitro. Biochern. Biophys. Res. Commun. 1977; 77,
925-32.
7. Tischler, M.E. & Goldberg, AL. Leucine degradation and
release of glutamine and alanine by adipose tissue. J. BioI.
Chern. 1980;255,8074-81.
8. Kowalchuk, J.M., Curi, R & Newsholme, EA. Glutamine
metabolism in isolated incubated adipocytes of the rat.
Biochem. J. 1988; 249, 705-8.
9. Elia, M. & Livesey, G. Effects of ingested steak and infused
leucine on forelimb metabolism in man and the fate of the
carbon skeletons and amino groups of branched-chain
amino acids. Clin. Sci. 1983; 64, 517-26.
10. Elia, M., Neale, G. & Livesey, G. Alanine and glutamine
release from the human forearm: effects of glucose administration. Clin. Sci. 1985; 69,123-33.
11. Elia, M., Schlatmann, A, Goren, A & Austin, S. Amino
acid metabolism in muscle and in the whole body of man
before and after ingestion of a single mixed meal. Am. J.
Clin. Nutr. 1989; 49, 1203-10.
12. Coppack, S.W, Fisher, RM., Gibbons, G.F. et al. Postprandial substrate deposition in human forearm and
adipose tissues in vivo. Clin. Sci. 1990; 79, 339-48.
13. Durnin, J.Y.G.A & Womersley, J. Body fat assessed from
total body density and its estimation from skinfold thickness: measurements of 481 men and women aged from 16
to 72 years. Br. J. Nutr. 1974; 32, 77-97.
14. Greenfield, AD.M., Whitney, R.I. & Mowbray, J.F.
Methods for the investigation of peripheral blood flow. Br.
Med. Bull. 1963; 19, 101-9.
15. Larsen, O.A, Lassen, N.A & Quaade, F. Blood flow
through human adipose tissue determined with radioactive
xenon. Acta Physiol. Scand. 1966; 66, 337-45.
16. Zielke, H.R, Ozand, P.T., Tildon, J.T., Sevdalian, DA. &
Cornblath, M. Growth of human diploid fibroblasts in the
absence of glucose utilization. Proc. Natl. Acad. Sci. U.S.A
1976; 73, 4110-4.
17. Newsholme, E.A, Crabtree, B. & Ardawi, M.S.M. Glutamine metabolism in lymphocytes: its biochemical, physiological and clinical importance. Q. J. Exp. Physiol. 1985; 70,
473-89.
18. Keast, D., Nguyen, T. & Newsholme, E.A Maximal activities of glutaminase, citrate synthase, hexokinase, phosphofructokinase and lactate dehydrogenase in skin of rats and
mice at different ages. FEBS Lett. 1989; 247, 132-4.
19. Foster, KJ., Alberti, KG.M.M., Hinks, L. et al. Blood intermediary metabolite and insulin concentrations after an
overnight fast: reference ranges for adults, and interrelations. Clin. Chern. 1978; 24, 1568-72.