511
CIinicalScience (1983) 64,511-516
Insulin-induced glucose utilization influences triglyceride
metabolism
J. BAZELMANS, P. J. N E S T E L A N D C. N O L A N
Baker Medical Research Institute, Prahran, Melbourne,Australia
(Received 18 August 1982; accepted 23 November 1982)
Summary
1. We have investigated the possibility that the
effect of insulin on triglyceride metabolism is
related to the individual’s responsiveness to
insulin-mediated glucose utilization. Changes in
plasma triglyceride levels were determined during 2 h infusions of insulin with glucose that
maintained euglycaemia in 17 subjects, some of
whom were overweight and/or hypertriglyceridaemic.
2. Plasma triglyceride concentrations fell in
most subjects (mean f SD: 19.9 f 13.0%). The
percentage fall in plasma triglyceride was inversely related to body mass index (r = -0.64,
P < 0.01) and to basal triglyceride concentration
(r = -0.69, P < 0405), but directly to insulin
sensitivity (r = +0.48, P < 0.05), and was
unrelated to plasma free fatty acid concentration.
3. Since insulin sensitivity was also related to
body mass index and basal triglyceride level
stepwise regression analysis was carried out to
determine the influence of these three variables on
insulin-mediated lowering of plasma triglyceride.
The percentage fall in plasma triglyceride remained independently related to insulin sensitivity
(P< 0.05) and to body mass index (P< 0.05),
and these two variables accounted for 44% of the
fall in triglyceride.
4. Resistance to insulin (in terms of glucose
utilization) may therefore be one significant,
independent factor determining the plasma triglyceride concentration.
Correspondence: Dr P. J. Nestel, Baker Medical
Research Institute, Commercial Road, Prahran,3 181,
Melbourne, Australia.
Key words:
triglyceride.
insulin
sensitivity,
plasma
Abbreviation: FFA, free fatty acids.
Introduction
There is a well-recognized association between
hypertriglyceridaemia and hyperinsuliinaemia [ 131. This is partly due to interactions through
overweight, though insulin and triglyceride concentrations are nevertheless independently correlated to a degree that suggests a causal
relationship. Insulin appears to influence both the
production and removal of triglyceride; although
the former is more generally favoured to explain
the relationship between the two [31, there is more
direct experimental evidence for an effect on
removal. Insulin stimulates lipoprotein lipase
activity in adipose tissue in vitro 14, 51, which is
reflected in the hyperlipaemia and enzyme
deficiency of diabetic subjects [61.
In previous studies 171 we have observed an
inverse relationship between the diurnal fluctuations in the concentrations of plasma insulin
on the one hand and triglyceride and plasma free
fatty acids (FFA) on the other that suggested a
role for insulin in triglyceride removal. This was
seen most clearly during 2-3 day periods of
sucrose feeding: the postprandial rises in insulin
coincided with falls in triglyceride and FFA, a
relationship that reversed with fasting during the
night when insulin levels fell but triglyceride and
FFA concentrations rose. On the other hand, the
direct correlation between insulin and triglyceride levels that occurs in the postabsorptive
state might also reflect diminished tissue sensi-
0143-5221/83/050511-0652.00 @ 1983 The Biochemical Society and the Medical Research Society
J. Bazelmans, P. J. Nestel and C. Nolan
512
tivity to insulin, leading to impaired removal of
triglyceride. This is analogous to the state of
diminished glucose removal in subjects whose
tissues are under-responsive to insulin [81. Indeed, diminished sensitivity to insulin, hyperglycaemia and hypertriglyceridaernia are commonly associated, especially in the context of
overweight 191.
We have therefore examined this possibility
more closely by infusing insulin without significantly altering the blood glucose concentration.
This provided a measure of glucose utilization or
tissue sensitivity to insulin which could then be
correlated with the magnitude of any insulininduced fall in plasma triglyceride concentration.
Methods
The studies were carried out in 17 men, some
normolipidaemic and of normal weight, others
overweight with or without associated hypertriglyceridaemia (plasma triglyceride > 3 mmol/l)
(Table 1). In order to increase the variability in
insulin sensitivity further, half the subjects consumed a diet enriched in carbohydrate (approximately 60% of energy, derived two-thirds from
complex carbohydrates and one-third from sirnple sugars). The remainder consumed about 45%
of energy as carbohydrate of similar composition to that eaten by the others. All were
hospitalized in a metabolic ward for about 3
weeks. Constant body weight was maintained,
regular exercise was provided by daily 2 h walks
and smoking was prohibited before the insulin
infusion. To minimize the anxiety engendered by
the infusion, the subjects were familiarized with
the procedure during a simulated infusion carried
out several days beforehand. Written, informed
consent was obtained.
The insulin infusion was carried out in the
subjects after fasting overnight. Plasma insulin
levels were raised by a single increment of about
56 punits/ml above the fasting concentration by
constant infusion of crystalline insulin. The blood
glucose was maintained by simultaneously infusing 20% glucose at rates that were varied
according to blood glucose determinations made
every 5 rnin [lo]. Steady-state euglycaernia was
generally achieved within 90 min; insulin and
glucose infusions were then maintained for a
further 20 min.
Blood samples were obtained from a vein in the
opposite arm before beginning the infusion and
three times during the steady-state period. Plasma
total triglyceride concentration was measured by
an enzymic technique, plasma glucose by the
glucose oxidase method, plasma insulin by
radioimmunoassay [ 111 and plasma FFA by the
method of Novak [ 121.
Total body sensitivity to insulin was cal-
TABLE1. Details of 17 lest subjects: age. body weight. body mass index, plasma
triglyceride concentration and dietary carbohydrates
Subject
no.
I
2
3
4
5
6
7
a
9
10
I1
12
13
14
15
16
17
Mean
f
SD
Age
(years)
Weight
(kg)
55
a1
78
Body
mass
index'
Plasma
triglyceride
0.58
0.45
1.25
1.71
1.88
0.74
3.07
3.52
1.16
1.31
4.63
41
21
31
26
40
22
26
31
34
55
40
25
27
25
26
21
39
70
69
77
121
I22
58
I04
70
55
90
73
74
21.2
24.0
26.7
28.7
24.6
21.3
27.8
48.8
38.3
23.7
32.9
24.0
18.6
22.9
24.4
22.6
25. I
31
9
80.0
19.7
26.8
7.3
a1
ao
(mmol/l)
Weight/height' ( k g / m f ) .
1 .oa
0.88
0.47
1.74
1.42
1.29
1.60
0.88
Dietary
carbohydrate
(% energy)
59
63
60
59
61
62
64
45
45
44
45
59
61
43
45
44
45
Insulin sensitivity and plasma triglyceride
culated from the steady-state glucose utilization
related to the plasma insulin concentration, as
described by DeFronzo et al. [lo]. This was
compared with the percentage reductions in
plasma triglyceride levels from mean values of the
three measurements during the final 30 min of
insulin infusion when glucose utilization was in a
new steady state.
The rate of glucose infusion during insulin
infusion, corrected for changes in glucose space
[lo], equals glucose utilization (M,mg min-I
kg-I). Relating M to the steady-state insulin
concentration (punits/ml) provides a measure of
insulin sensitivity (M/z).
The changes in plasma triglyceride concentration have been expressed as percentage falls
from control (pre-infusion) levels; this took into
account the wide range of plasma triglyceride
values.
5 13
entirely to a reduction of triglyceride in very low
density lipoprotein (d < 1.006 g/ml).
A correlation matrix relating the measured
parameters in the basal state and after insulin
infusion is shown in Table 3. The percentage fall
in plasma triglyceride concentration was related
directly to the degree of insulin sensitivity (P <
0.05). It was also significantly and negatively
correlated with body mass index (P < 0.01),
basal triglyceride concentration ( P < 0.005)and
basal insulin concentration (P < 0.02). Despite a
substantial post-insulin fall in plasma FFA levels,
the reduction in triglyceride was neither significantly related to basal FFA levels nor, more
importantly, to the percentage fall in plasma FFA
during insulin infusion.
TABLE2. Metabolic measurements in the basal state
and during the steady-state insulin infusion
Results
In most subjects the plasma triglyceride concentration had fallen at the first sampling during
the new steady state of glucose utilization, that is
after a mean period of 89 min of insulin infusion.
Triglycerides fell on average from 1.60 f 0.27
(SE) mmol/l to 1.35 f 0.27 (P < 0.001) and then
further to 1.33 ? 0.27 and 1.30 ? 0.27 mmol/l
after 99 and 109 min of insulin infusion respectively. The overall percentage fall was 19.9 ?
13.0 (SD) for the 17 studies (Table 2). Plasma
FFA fell by 49.5 f 15.1% (Table 2). The
reduction in plasma triglyceride was due almost
Mean f SBM
Basal plasma triglyceride (mmol/l)
Basal plasma free fatty acids (pmolll)
Basal plasma insulin (punitslml)
Basal blood glucose (mmolh)
Triglyceride reduction. (%)
FFA reduction* (%)
Insulin sensitivity ( M / I ) t
Average of three values taken after 89, 99 and 109 min
of insulin infusion.
t M / I = mass of glucose utilized mg min-I kg-' in
relation to plasma insulin concentration.
TABLE3. Simple correlation coeflcients (r)relating the insulin-induced percentage
fall in plasma triglyceride, insulin sensitivity, body weight. basal triglyceride level,
basal insulin level and basal and percentagefall in FFA
N.S.,not significant.
Variable X
Variable Y
r
P
Fall in triglyceride concn. (%)
Insulin sensitivity ( M / I )
Body mass index
Basal triglyceride
Basal insulin
Basal FFA
Fall in FFA (%)
Body mass index
Basal insulin
Basal triglyceride
Basal FFA
Fall in FFA (%)
Basal insulin
Basal FFA
Body mass index
Basal insulin
Basal FFA
+0.482
-0.643
-0.698
-0.587
-0.306
-0.184
-0.560
-0.512
-0.535
-0,629
-0.153
+0.594
+0.472
+0.629
+0.827
+0.689
<0.05
<0.01
Insulin sensitivity ( M / I )
Basal triglyceride
Body mass index
1.60 f 0.24
437 f 38.0
9 f 2.0
4.9 f 0.2
19.9 ? 3.2
49.5 f 3.7
13.0 ? 2.0
<0405
(0.02
N.S.
N.S.
<0.02
<0.05
(0.05
<0.001
N.S.
<0.01
N.S.
<0*01
<0401
<0.005
514
J. Bazelmans. P. J. Nestel and C. Nolan
TAELE4 . Stepwise regression analysis relating the insulin-induced percentage fall in
plasma triglyceride to insulin sensitivity. body mass index, basal triglyceride and basal
insulin concentrations
N.S., Not significant.
Source of variation
F
P
Contribution
by variable($
(%)
I . Insulin sensitivity (Mlr)
2. Body mass index
3. Basal plasma triglyceride
4. Basal plasma insulin
5.1&2
6. 1 , 2 & 3
I . 1,2,3&4
4.543
5.020
3.613
0.1419
Insulin sensitivity ( M / I ) was negatively correlated with body mass index (P < 0.02), basal
insulin (P < 0.05),basal triglyceride (P < 0.05)
and basal FFA (P < 0.001) concentrations.
Body mass index was also significantly correlated positively with basal insulin (P < 0.001)
and basal FFA (P < 0.005) concentrations.
These simple correlations emphasized the close
interrelationships between body mass index, the
concentrations in plasma of triglyceride, FFA
and insulin and the overall sensitivity to insulin as
measured by the utilization of glucose (M/I). Of
immediate relevance was the finding that insulin
‘sensitivity’ measured in terms of plasma triglyceride reduction was apparently influenced by
insulin sensitivity (as M/I), body mass index and
plasma triglyceride and plasma insulin concentrations. Increasing adiposity and hypertriglyceridaemia therefore impeded insulin mediated
reduction of triglyceride levels.
Stepwise regression analysis was carried out of
the factors that appeared to affect the triglyceride fall. This is summarized in Table 4,
which shows that insulin sensitivity (MI)
and
body mass index were the two most significant
factors and apparently accounted for about 44%
of the variation.
Discussion
This study suggests a close link in the body’s
capacity to clear glucose and triglyceride from
plasma. The maintenance of euglycaemia during
the infusion of insulin permits controlled estimates of insulin-stimulated glucose utilization
[lo]. It also appears to provide a measure of the
influence of insulin on the plasma triglyceride
concentration. The finding that insulin-regulated
glucose clearance and the fall in plasma triglyceride are so significantly correlated has not
t0.05
<0.05
<o. 1
N.S.
Multiple
correlation
coefficients
23
21
12
0.3
44
56
57
0.6596
0.7469
0.7504
been demonstrated previously and raises the
question of mechanisms. Stepwise regression
analysis suggested that about 23% of the fall in
triglyceride was independently attributable to
insulin sensitivity.
The observation that impaired glucose
tolerance and hypertriglyceridaemia are often
associated 11-31, most commonly but not exclusively in fat people, has been variously
interpreted. Reaven et al. [2] believe the sequence
to begin with tissue resistance to insulin
(measured as glucose removal) leading to oversecretion of insulin which in turn stimulates
triglyceride production in the liver. In their
studies, the flux of triglyceride through plasma
correlated with the degree of insulin resistance,
increasing with obesity and diminishing with
weight reduction [ 131; since circulating insulin
levels rose and fell correspondingly they have
concluded that insulin regulates plasma triglyceride through its rate of production. More recent
studies with isolated liver cells, however, suggest
the reverse, namely that insulin may suppress the
release of hepatic triglyceride [ 141. Insulin also
stimulates triglyceride removal through its control of tissue lipoprotein lipase formation [SI.
Nilsson-Ehle et al. [ 151 have shown a significant
rise in human adipose tissue lipoprotein lipase
within half an hour of a glucose meal, which
would favour removal of triglyceride. Furthermore, Goldberg et al. [161 have shown that the
plasma triglyceride response to high carbohydrate diets is related to reciprocal changes in
adipose lipoprotein lipase activity during meals. It
is therefore very likely that both formation and
removal of triglyceride are influenced by insulin,
which is reflected in the diurnal fluctuations in
plasma triglyceride and insulin concentrations.
We have shown previously that, on a fat-free diet,
plasma triglyceride and insulin levels vary
Insulin sensitivity and plasma triglyceride
reciprocally [71; more recently Hayford et al.
[171 have observed this even with mixed food
intake. Thus triglyceride and insulin concentrations tend to be inversly related throughout the
24 h cycle of eating and fasting, despite the
consistent finding of a direct correlation between
the concentrations of the two in fasting, morning
samples.
In the present study we have also observed a
highly significant direct correlation between the
plasma insulin and triglyceride levels taken in the
fasting state (Table 3). Yet by raising the insulin
modestly (to levels reached during meals), plasma
triglycerides fell rapidly. Furthermore those who
were least sensitive to insulin in terms of glucose
utilization showed the smallest percentage falls in
triglyceride. Put another way, those who were
most resistant to the actions of insulin maintained
their triglyceride at higher levels.
Circulating free fatty acid concentrations were
also suppressed rapidly after the infusion of
insulin and since the fractional incorporation of
plasma FFA into triglyceride is rapidly reduced
as FFA mobilization from adipose tissue is
suppressed by insulin [181, an early effect on
triglyceride production might have been expected. However, statistical analysis failed to
show a correlation between the reductions in
plasma FFA and triglycerides. That is not to say
that the halving of a major substrate did not
influence triglyceride production, and we have
observed a rapid increase in the triglyceride
concentration in lean subjects when FFA
mobilization was increased acutely
191.
Although it is highly likely that the fall in FFA
contributed to the fall in triglyceride, the data
nevertheless show that this factor could not
account for the significant correlation between
insulin sensitivity and the plasma triglyceride
reduction.
The effect of body weight, which appeared
independently to account for a further 21% of the
insulin-induced fall in triglyceride concentration,
might reflect lesser inhibition of FFA flux in the
obese [ 191 and impaired re-esterification of
released fatty acids. This technique for estimating
insulin sensitivity does not identify the organs and
tissues which reflect the overall result. The liver in
obesity shows reduced sensitivity to the actions of
insulin with respect to glucose metabolism [201,
insulin-mediated uptake of glucose by the human
forearm is diminished in obese subjects 1211 and
enlarged human adipocytes have reduced insulinbinding capacity [221. The widespread effects of
insulin on lipid and glucose metabolism need to
be considered in the interpretation of our results.
Within adipose tissue and probably muscle the
515
inhibition of FFA efflux reflects the sum of
increased re-esterification of FFA through enhanced glucose uptake and diminished triglyceride lipolysis, which may be independent of
glucose utilization [231. Further, although insulin-stimulated glucose removal occurs to an
important extent in the liver direct hepatic
removal of non-dietary triglyceride is thought to
be minor, at least in normolipidaemic subjects.
Thus the effects of insulin on glucose utilization,
FFA and triglyceride are mediated through both
unrelated and related mechanisms and the overall
short-term effects of insulin are represented by
the present data. The strength of the statistical
analysis must be viewed in this context, and
although it indicates a close association between
insulin-induced glucose and triglyceride removal
it cannot by itself pinpoint the mechanism
through which this occurs. Nevertheless fatty
acids that are released during lipoprotein lipase
mediated breakdown of triglyceride require the
presence of insulin for re-esterification with
glycerol derived from glucose. Since this reesterification prevents intracellular FFA accumulation, which inhibits lipoprotein lipase activity
[241, greater insulin sensitivity and greater
glucose uptake are likely to enhance also the
removal of triglyceride.
Aknowledgments
This work was supported in part by a grant from
the National Heart Foundation of Australia.
References
I1I EATON,R.P.& NYE,W.H.R. (1973) The relationship between
insulin secretion and triglyceride concentration in endogenous
lipemia. Journal of Laboratory and Clinical Medicine, 81,
682-695.
REAVEN, G.M., LERNER, R.L., STERN, M.P. & FARQUHAR,
J.W. (1967) Role of insulin in endogenous hypertriglyceridemia. Journal of Clinical Invesligafion,46,1756-1767.
[Sl TOBEY,T.A., GREENFIELD,
M.,KRAEMER, F. & REAVEN,
G.M.(198 1) Relationship between insulin resistance, insulin
secretion, very low density lipoprotein kinetics, and plasma
triglyceride levels in normotriglyceridemic man. Melabolism,
I21
30,165-171.
I41 AUSTIN, W. & NESTEL, PJ. (1968) The effect of glucose and
insulin in vilro on the uptake of triglyceride and on lipoprotein
lipase activity in fat pads from normal, fed rats. Biochimlca et
Biophysica Acla, 164,59-63.
IS1 PYKXLISTO, OJ., SMITH, P.H. & BRUNZELL,J.D. (1975)
Determinants of human adipose tissue lipoprotein lipase. Effect
of diabetes and obesity on basal and'diet-induced activity.
Journal of Clinical Investigalion, 56,1108-1 111.
I61 TASKINEN,
M A . & NIKKILA,E.A. (1979) Lipoprotein lipasc
activity of adipose tissue and skeletal muscle in insulindeficient
human diabetes. Diaktologia, 17,35 1-356.
516
J. Barelmans, P. J. Nestel and C. Nolan
171 BARTER,P.J., CARROLL,
K.F. & NESTEL,P J . (1971) Diurnal
1161 GOLDBERG,A.P., CHAIT, A. & BRUNZELL,J.D. (1980)
fluctuations in triglyceride, free fatty acids and insulin during
sucrose consumption and insulin infusion in man. Journal of
Clinical Investigation, 50,583-59 1.
181 KOLTERMAN,
O.G., INSEL, J., SAEKOW,
M. & OLEFSKY,J.M.
(1980) Mechanisms of insulin resistance in human obesity.
Journal of Clinical Investigation, 65, 1272-1284.
191 NESTEL,P.J. & GOLDRICK,B. (1976) Obesity: changes in lipid
metabolism and the role of insulin. Clinical Endocrinology and
Metabolism, $ 3 13-335.
IlOl DEFRONZO,R.A., SOMAN,V., SHERWIN,
R.S., HENDLER,R.
& FELIG,P. (1978) Insulin binding to monocytes and insulin
action in human obesity, starvation, and refeeding. Journal of
Clinical Investigation, 62,204-2 13.
V., LAU,K.S., GOTTLIEB,C.W. & BLEICHER,S.J.
1111 HERBERT,
(1965) Coated charcoal immunoassay of insulin. Journal of
Clinical Endocrinology and Metabolism, 25,1375-1384.
1121 NOVAK,M. (1965) Colorimetric ultramicro method for the
determination of free fatty acids. Journal of Lipid Research. 6,
Postprandial adipose tissue lipoprotein lipase activity in
primary hypertriglyceridemia. Metobolism, 29,223-229.
1171 HAYFORD,
J.T., DANNEY,M.M. & THOMPSON,R.G. (1979)
Triglyceride-integrated concentration: relationship to insulinintegrated concentration. Metobolism, 28, 1078-1085.
1181 NESTEL, P.J. (1967) Relationship between FFA flux and
TGFA influx in plasma before and during the infusion of
insulin. Metabolism, 16, 1123-1 132.
1191 NESTEL, P.J., ISHIKAWA,
T. & GOLDRICK,R.B. (1978)
Diminished plasma free fatty acid clearance in obese subjects.
Metabolism. 27,589-597.
1201 FELIG,P., WAHREN,
J., HENDLER,R. & BRUNDIN,
T. (1974)
Splanchnic glucose and amino acid metabolism in obesity.
Journal of Clinical Investigation, 53,582-590.
1211 RABINOWITZ,D. & ZIERLER,K.L. (1962) Forearm metabolism in obesity and its response to intra-arterial insulin.
Characterization of insulin resistance and evidence for adaptive hyperinsulinism. Journol of Clinical Investigation, 41,
431-433.
1131 OLEFSKY,J., REAVEN,G.M. & FARQUHAR,
J.W. (1974)
Effects of weight reduction on obesity. Journal of Clinical
Investigation. 53,64-76.
G. (1981) The role of insulin on
1141 PATSCH,W. & SCHONFELD,
2 173-2 18 1.
1221 HARRISON.L.C., MARTIN,F.I.R. & MELICK,R.A. (1971)
lipoprotein assembly and secretion in primary rat liver cell
cultures. Clinical Research, 2 9 , 8 3 9 ~ .
P., CARLSTROM,
S. & BELFRAGE,P. (1978)
1151 NILSSON-EHLE,
Effects of ethanol intake on lipoprotein lipase activity in
adipose tissue of fasting subjects. Lipids, 13.433-437.
Correlation between insulin receptor binding in isolated fat
cells and insulin sensitivity in obese human subjects. Journol of
Clinical Investigation, 58, 1435-1441.
1231 STEINBERG,
D. & KHOO,J.C. (1977) Hormone-sensitive lipase
of adipose tissue. Federation Proceedings, 36, 1986-1990.
E.A. & PYKALIST~),
0. (1968) Regulation of adipose
1241 NIKKILA,
tissue lipoprotein lipase synthesis by intracellular free fatty
acid. L f e Sciences, 7,1303-1309.