Download Insulin-Mediated Increases in the HDL Cholesterol/Cholesterol

Insulin-Mediated Increases in the HDL
Cholesterol/Cholesterol Ratio in Humans
Craig N. Sadur and Robert H. Eckel
Downloaded from http://atvb.ahajournals.org/ by guest on May 10, 2017
The role of insulin in the regulation of lipoprotein cholesterol distribution was
studied in 18 human volunteers who were of normal weight, normolipemic, and glucose-tolerant. We used the euglycemic clamp technique and made comparisons with
six saline-infused controls. While total cholesterol and low density lipoprotein cholesterol levels fell similarly by 6 hours in both groups (p = NS), there was a greater
decrease in triglyceride levels (p < 0.02) but less decrease in high density lipoprotein
cholesterol levels (p < 0.02) in the insulin-infused group. These changes resulted in
both a higher high density lipoprotein cholesterol/total cholesterol ratio (p = 0.01) and
a higher high density lipoprotein cholesterol/low density cholesterol ratio (p = 0.02) in
the insulin-infused group. The timing of this effect of insulin implies a mechanism
unrelated to high density lipoprotein turnover. Thus, insulin infusion during euglycemia appears to alter cholesterol distribution favorably in normal subjects.
(Arteriosclerosis 3:339-343, July/August 1983)
to examine the insulin effect in a more direct manner
would be beneficial. In the present study, the euglycemic clamp technique has provided a tool for assessment of the effect of insulin on lipoprotein cholesterol distribution in human subjects.
he role of insulin in the regulation of lipoprotein
T
cholesterol metabolism is unclear. In vivo, diseases at the extremes of serum insulin concentration
have been associated with alterations in lipoprotein
cholesterol distribution. In obesity, a state where serum insulin is high, there are increases in total, very
low density lipoprotein (VLDL), and low density lipoprotein (LDL) cholesterol and a decrease in high
density lipoprotein (HDL) cholesterol.12 Type I diabetes mellitus, where serum insulin is low, is associated with altered lipoprotein metabolism which normalizes with improved glycemic control.3"5 In fact,
some studies have demonstrated increased levels of
HDL cholesterol,67 HDL cholesterol/total cholesterol, or HDL cholesterol/LDL cholesterol ratios8 in insulin-treated diabetics. Although the lipoprotein alterations seen in the insulinopenic patient could have
resulted from insulin deficiency alone, a direct effect
of insulin on lipoprotein cholesterol metabolism has
not been proven because other metabolic alterations
are also corrected (e.g., ketonemia). Thus, a method
Methods
The study group included 18 subjects who received intravenous insulin and glucose. Each person
was average in height and within 20% of ideal body
weight according to the Metropolitan Life Insurance
Company Standards.9 The body mass index (wt/ht2)
was 2.06 x 10" 3 ± 0.05 kg/cm2 (x ± SEM). The
subjects ranged in age from 20 to 45 years, x = 29.2
± 1.4 years. There were 15 women and three men.
The control group, composed of four women and two
men, received intravenous infusions of 0.9% (normal) saline at approximately 100 to 150 ml/hr. They
were also normal in height and weight with a body
mass index of 2.17 x 10~3 ± 0.05 kg/cm2 and with
an age range of 23 to 40 years, x = 29.0 ± 2.5
years. Individuals were excluded if they were taking
drugs with known effects on glucose or lipid metabolism, such as oral contraceptives, other estrogenic
compounds, diuretics, or other antihypertensive
medications. All subjects were free of acute or chronic illnesses. Two individuals were euthyroid while
taking thyroxine for full thyroid hormone replacement. Physical examinations confirmed the normal
From the Division of Endocrinology, Department of Medicine,
University of Colorado Health Sciences Center, Denver, Colorado.
Support for this work was provided by National Institutes of
Health Grants AM 26356 and RR-00051 to the General Clinical
Research Center at the University of Colorado Health Sciences
Center.
Address for reprints: Craig N. Sadur, M.D., B151, 4200 East
Ninth Avenue, Denver, Colorado 80262.
339
340
ARTERIOSCLEROSIS VOL 3, No 4, JULY/AUGUST 1983
Downloaded from http://atvb.ahajournals.org/ by guest on May 10, 2017
health of all subjects. Each ingested 40 g/m2 of glucose and had a normal 3-hour oral glucose tolerance
curve, according to the guidelines of the National
Diabetes Data Group.10 In addition, all subjects were
normal in tests for: tri-iodothyronine resin uptake,
total thyroxine, thyroid-stimulating hormone, electrolyte panel, calcium, liver panel, phosphorus, magnesium, complete blood count, and urinalysis. Fasting
serum cholesterol and triglyceride (TG) were normal
for all subjects.
All studies were approved by the Human Subjects
Committee and were performed in the Clinical Research Center at the University of Colorado Health
Sciences Center. Informed consent was obtained
from all participants studied. For 2 days before the
study, each subject was fed an isocaloric formula
feeding containing 45% carbohydrate, 40% fat, and
15% protein. After an overnight fast, on the morning
of the euglycemic clamp study, two intravenous lines
were placed in opposite arms of the study subjects,
one for infusion of fluids and one for sampling of
blood. A heating pad was placed over the arm where
the samples were obtained from venous blood. Plasma glucose was determined after plasma was immediately separated with a Beckman Microfuge
(Beckman Instruments, Incorporated, Palo Alto,
California) by the glucose oxidase technique using
the Beckman glucose analyzer (Beckman Instruments, Incorporated, Clinical Instruments Division,
Fullerton, California). In the insulin group, basal plasma glucose values were determined on the morning
of the euglycemic clamp study and served as the
standard, determining the varying amounts of glucose infusion needed to maintain euglycemia during
the rest of the study. Although glucose was similarly
determined throughout the saline infusion, glucose
was not infused.
Basal and serial adipose tissue biopsies (at 20,80,
180, and 360 minutes after the beginning of the infusions) were performed for the measurement of lipoprotein lipase activity as previously described.11
Most of these data have been published and will not
be included in the present paper.
Insulin was infused into the study subjects in an
exponentially decreasing manner over the first 10
minutes and then at either 40 or 120 mU/m2 per
minute to obtain a steady-state serum insulin level of
75 ± 5 or 271 ± 27 ^U/ml, respectively. Glucose
infusion was begun at 4 minutes after the start of the
insulin infusion. At that point and throughout the
study, a blood sample was drawn at least every 5
minutes to test for plasma glucose concentration in
order to determine the changes in glucose infusion
rate necessary to maintain euglycemia. In addition,
these subjects received saline infusions similar to
those received by the control subjects.
A potassium supplement of 30 mEq of carbohydrate-free KCI solution was given to all subjects the
evening preceding the study and at approximately
1.5 and 3.5 hours after the start of the insulin infusion
to maintain normokalemia.
TG was determined by an enzymatic method.12
The coefficient of variation of TG was 8%.
Total cholesterol was also measured enzymatically13 with a coefficient of variation of 4%.
HDL cholesterol was measured in plasma from
which VLDL and LDL were precipitated by 4% sodium phosphotungstate.14 The HDL cholesterol was
measured in the supernatant after centrifugation in a
Beckman TJ-6 centrifuge, 1500 g for 30 minutes at
4°C. The coefficient of variation for HDL cholesterol,
including analytical and precipitation variation, was
8%.
LDL cholesterol was calculated from the formula of
Friedewald, et al.15
Serum insulin levels were measured by a double
antibody radioimmunoassay with the technique of
Desbuquois and Aurbach.16
The Mann-Whitney U test for unpaired observations was used for statistical analyses.
Results
Baseline Data
The baseline data of both the 18 study subjects and the six control subjects are displayed in
Table 1. Basal plasma glucose values in all subjects
ranged from 71 to 99 mg/dl. Basal lipid determinations were similar in both high-insulin (n = 7) and
low-insulin (n = 11) infusion groups. Whereas TG,
HDL cholesterol, and HDL cholesterol/total cholesterol were similar in both the insulin and saline control groups, total cholesterol and LDL cholesterol
were significantly higher in the control group (p <
0.05) while HDL cholesterol/LDL cholesterol was
significantly lower in the control group (p < 0.05).
These differences became insignificant when the
data were matched for gender.
Table 1. Baseline Data Before Insulin or Saline
Infusion
Age (yrs)
Weight (kg)
Wt/Ht2 x 10" 3 (kg/cm2)
Glucose (mg/dl)
Total cholesterol (mg/dl)
LDL cholesterol (mg/dl)
TG (mg/dl)
HDL cholesterol (mg/dl)
HDL cholesterol/
total cholesterol (%)
HDL cholesterol/
LDL cholesterol (%)
Insulin
Saline
29.2+1. 4
57.7 + 2. 2
2.06 ± 0 . 05
84.2±1. 7
159 + 5
88±5
74±7
56±3
29.0 ±2.5
63.8±3.3
2.17±0.05
86.8 + 2.6
179±8
108 + 6
90±11
53±6
35 + 2
30 + 3
67±6
50 + 4
Values are means ± SEM. TG = triglyceride; HDL =
high density lipoprotein; LDL = low density lipoprotein.
INSULIN INCREASES HDL CHOLESTEROL/CHOLESTEROL
Sadur and Eckel
341
Glucose
Saline infusion resulted in a gradual fall in plasma
glucose over the 6-hour infusion period to 92% ±
3% of basal. While insulin was infused into study
subjects, plasma glucose was maintained by a variable glucose infusion and ranged throughout the
study from 94% to 103% of basal value.
Total Cholesterol
Downloaded from http://atvb.ahajournals.org/ by guest on May 10, 2017
For total cholesterol and all remaining lipid measurements, no significant differences between insulin
at low (40 mU/m2 per minute) and high (120 mU/m2
per minute) infusion rates were found. Thus, all insulin data are combined measurements for both insulin
infusion rates. Although women appeared to have a
greater decrease in total cholesterol by 6 hours during both the insulin infusions (men = - 1 0 ± 2 mg/dl
vs women = - 2 1 ± 3 mg/dl) and the saline infusions (men = - 9 ± 8 mg/dl vs women = - 2 1 ± 5
mg/dl), these differences were not statistically significant. The combination of men and women, as shown
in Figure 1 A, revealed similar 6-hour declines in both
infusion groups (saline = - 1 7 ± 5 mg/dl vs insulin
= - 1 9 ± 3 mg/dl, p = NS).
-15
-25
5
-5
LDL Cholesterol
As shown in Figure 1 B, there were no significant
differences between the two groups in the decreases
in LDL cholesterol levels.
Trlglycerldes
TG levels decreased more in the study subjects
than in the controls (Figure 1 C). The p values for
the differences in TG for the two groups at the 80minute, 180-minute, and 380-minute time-points
were < 0.01, < 0.05, and < 0.02, respectively.
There was no gender difference in basal or subsequent TG changes for either saline or insulin infusions.
-15
B
o
-to
HDL Cholesterol
The change in HDL cholesterol is seen in Figure
1 D. In both study subjects and controls, there was a
fall in HDL cholesterol with the infusions, but the
insulin-infused group had no further decrease after
the first 80 minutes. Whereas the differences between the two groups were insignificant at 80-minute
and 180-minute time-points, the difference at 380
minutes was statistically significant (p < 0.02).
Again, there was no gender difference between responses during insulin and saline infusions.
-40
5
-5
Figure 1. Changes in total cholesterol (A), LDI cholesterol (B), triglycerides (C), and HDL cholesterol (D) are plotted by minutes during saline (•—•) and insulin (o—-o)
infusions. *p s£ 0.01, t p < 0.05, tP *= 0.02. Data are mg/dl
(mean ± SEM).
-10
0
80
180
Mln
380
342
ARTERIOSCLEROSIS VOL 3, No 4, JULY/AUGUST 1983
Miscellaneous
o
aj
o
o
Insulin-stimulated adipose tissue lipoprotein lipase failed to correlate with changes in cholesterol
distribution. There was also no relationship between
the fall in serum triglyceride and the changes in HDL
cholesterol. In addition, the amount of glucose infused did not predict the changes observed.
O
"5
o
O
Q
I
o
©
Downloaded from http://atvb.ahajournals.org/ by guest on May 10, 2017
"5
o
Discussion
-2
15
10
5
2
.2
0
r 1
S -.0
80
B
ISO
Mln
380
Figure 2. Changes in the HDL cholesterol/total cholesterol ratio (A) and HDL cholesterol/LDL cholesterol ratio
(B) are plotted by minutes during saline (•—•) and insulin
(o—-o) infusions. *p s= 0.01, %p ^ 0.02. Data are shown
as percentages (mean ± SEM).
HDL Cholesterol/Total Cholesterol
When expressed as the ratio of HDL cholesterol/
total cholesterol, insulin infusion resulted in a higher
absolute increase in the ratio at 380 minutes when
compared to saline infusion, 2.8 ± 1.2% vs - 1 . 2 ±
0.8%, respectively (Figure 2 A, p = 0.01). Once
again, gender did not affect the changes that occurred during the infusions. In men, the 6-hour HDL
cholesterol/total cholesterol ratio changed - 2 . 0 ±
0.0% with saline infusion vs 2.0 ± 1.2% with insulin;
in women, the change was - 0 . 8 ± 1.3% vs 3.0 ±
1.4%.
HDL Cholesterol/LDL Cholesterol
As with HDL cholesterol/total cholesterol, the
study group had a statistically significant increase in
the HDL cholesterol/LDL cholesterol ratio by 380
minutes after the start of the infusion (Figure 2 B, p =
0.02).
In this study the euglycemic clamp technique has
provided a tool for examination of the effect of insulin
on lipoprotein cholesterol distribution in 18 normal
subjects. Because the response was similarly affected at the two insulin infusion rates, 40 and 120 mU/
m2 per minute, the results from the two groups were
combined.
In both saline-infused and insulin-infused subjects, a progressive decrease in total cholesterol occurred over 380 minutes. LDL cholesterol also fell
similarly in the two groups during the same time
course because of a greater decrease in TG levels
but a smaller decrease in HDL cholesterol levels in
insulin-infused subjects. Some caution in the interpretation of the calculated LDL cholesterol data is
needed because the maintenance of VLDL composition was assumed but not proven. Because substantial intravenous volumes were infused in both
groups, a dilutional effect is likely. The insulin effect
on trigycerides, however, was more than dilutional
and has been reported previously.11
Most important, insulin administration reduced the
decrease in HDL cholesterol level, particularly during
the latter part of the infusion. This effect was not
associated with the insulin-stimulatory effect on adipose tissue lipoprotein lipase activity, the decrease
in triglyceride level, or the amount of glucose infused
(data not shown). Moreover, when the data were
expressed as the ratios of both HDL cholesterol/total
cholesterol and HDL cholesterol/LDL cholesterol, insulin administration resulted in higher ratios than
those found with normal saline infusions. Some investigators have felt that these ratios provide the
most meaningful lipid assessment of antiatherogenic
risk.17'18 Because the half-life of apo A-l and apo A-ll
is measured in days,19 the effects of the hormone on
HDL production or clearance would be unlikely explanations for the differences noted between saline
and insulin administration.
Possibly, an effect of insulin on cholesterol transfer between two lipoprotein particles played a role. In
vitro studies have revealed that esterified cholesterol
can be transferred from HDL to VLDL, an effect mediated by a specific transfer protein.20 This phenomenon also occurs with triglyceride transferred from
VLDL to HDL. Both these transfers are reversible21
and are probably independent of each other.20 Although HDL cholesterol, the HDL cholesterol/total
cholesterol ratio, and the HDL cholesterol/LDL cho-
INSULIN INCREASES HDL CHOLESTEROL/CHOLESTEROL
Downloaded from http://atvb.ahajournals.org/ by guest on May 10, 2017
lesterol ratio, did not change significantly at the 80and 180-minute time-points in the study group when
compared to the control group, the 380-minute timepoints revealed significant changes with the insulin
infusion, consistent with the temporal course of lipid
transfers decribed in vitro. Thus, perhaps insulinmediated increases in the HDL cholesterol/total cholesterol and HDL cholesterol/LDL cholesterol ratios
could be due to alterations in cholesterol transfer to
or from HDL, preventing HDL cholesterol concentrations from falling during the infusion period.
The evidence that insulin affects cholesterol metabolism in humans is indirect. A favorable effect of
insulin on cholesterol metabolism is best exemplified
by Type I diabetics who demonstrate a reduction in
total cholesterol after improvement of the hypoinsulinemic state.3"5 After insulinization, lipids return to
normal. In fact, higher levels of circulating free insulin
concentrations found in treated Type I diabetics22'M
may relate to higher levels of HDL cholesterol67 or
HDL cholesterol/total cholesterol8 found in these
patients.
An unfavorable effect of insulin on cholesterol metabolism would have also been possible. For instance, high carbohydrate diets, which have been
associated with reductions in total and LDL cholesterol2425 have also resulted in increases in VLDL 2425
and decreases in HDL cholesterol.24 In addition,
obese subjects who are hyperinsulinemic have increases in total cholesterol, VLDL cholesterol, and
LDL cholesterol with lower HDL cholesterol concentrations than normal-weight controls. 12 Thus, although the present study was a short-term one and
involved a nonphysiologic administration of insulin,
we observed an antiatherogenic influence of insulin
on lipoprotein cholesterol distribution in normal human subjects. This effect was predominantly on HDL
cholesterol and suggests that mechanisms independent from lipoprotein turnover may be important
in mediating this phenomenon.
Acknowledgments
We sincerely thank Orville G. Kolterman for insulin radioimmunoassays; Debbie Gwinner, Deborah J. Krinitzsky, and Judy Prasad for technical assistance; and Joni Foti for secretarial assistance.
References
1. Garrison RJ, Wilson PW, Castelll WP, Felnlelb M, Kannel
WB, McNamara PM. Obesity and lipoprotein cholesterol in
the Framingham Offspring Study. Metabolism 1980;29:
1053-1060
2. Sadur CN, Eckel RH. Insulin-mediated lipoprotein metabolism in obesity. Diabetes 1982;31:156A
3. Pletrl A, Dunn FL, Raskin P. The effect of improved diabetic
control on plasma lipid and lipoprotein levels. Diabetes
1980;29:1001-1005
Index Terms:
insulin
• euglycemic clamp
Sadur and Eckel
343
4. Sosenko JM, Breslow JL, Mlettlnen OL, Gabbay KH. Myperglycemia and plasma lipid levels. N Engl J Med 1980;
302:650-654
5. Eckel RH, McLean E, Alters JJ, Cheung MC, Blerman EL.
Plasma lipids and microangiopatny in insulin-dependent diabetes mellitus. Diabetes Care 1981;4:447-453
6. Nikklla EA, Hormlla P. Serum lipids and lipoproteins in insulin-treated diabetes. Demonstration of increased high density
lipoprotein concentrations. Diabetes 1978;27:1078-1086
7. Klujber L, Molnar D, Kardos M, Jaszai V, Soltesz GY,
Mestyan J. Metabolic control, glycosylated haemoglobin and
high density lipoprotein cholesterol in diabetic children. Eur J
Pediatr 1979; 132:289-297
8. Eckel RH, Albers JJ, Cheung MC, Wahl PW, Llndgren FT,
Blerman EL. High density lipoprotein composition In insulindependent diabetes mellitus. Diabetes 1981 ;30:132-138
9. Metropolitan Lite Insurance Company. New weight standards for men and women. Statistical Bulletin 1959,40:1-4
10. National Diabetes Data Group. Classification and diagnosis
of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:1039-1057
11. Sadur CN, Eckel RH. Insulin stimulation of adipose tissue
lipoprotein lipase: Use of the euglycemlc clamp technique. J
Clin Invest 1982;69:1119-1125
12. Stavropoulos WS, Crouch RD. A new colorimetric procedure for the determination of serum triglycerides (abstr). Clin
Chem 1974;20:857
13. Richmond W. Preparation and properties of a cholesterol
oxidase from norcadia Sp. and its application to the enzymatic assay of total cholesterol in serum. Clin Chem
1973;19:1350-1356
14 Lop«s-Vlrella MF, Stone P, Ellis S, Colwell J A. Cholesterol
determinations in high-density lipoproteins separated by
three different methods. Clin Chem 1977;23:882-884
15. Frtedewald WT, Levy Rl, Fredrlckson DS. Estimation of the
concentration of tow-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem
1972:18:499-502
16. Desbuquols B, Aurbach GD. Use of the polyethylene glycol
to separate free and antibody-bound peptide hormones in
radioimmunoassays. J Clin Endocrinol Metab 1971,33:732738
17. Burslem J, Schonfeld G, Howald MA, Weldman SW, Miller
JP. Plasma apoprotein and lipoprotein lipid levels in vegetarians. Metabolism 1978;27:711-719
18. Wltztum J, Schonfeld G. High density lipoproteins. Diabetes
1979;28:326-336
19. Blum CB, Levy Rl, Elsenberg S, Hall M III, Goebel RH,
Bercnan M. High density lipoprotein metabolism in man. J
Clin Invest 1977;60:795-807
20. Hopkins GJ, Barter PJ. Transfers of esterified cholesterol
and triglycende between high-density and very low-density
lipoproteins: In vitro studies of rabbits and humans. Metabolism 1980;29:546-550
21. Barter PJ, Lally Jl. In vitro exchanges of esterified cholesterol between serum lipoproteins fractions: Studies of humans
and rabbits. Metabolism 1979;28:230-236
22. Hedlng LG. Determination of total serum insulin (IRI) in insulin-treated diabetic patients. Diabetologia 1972;8:260-266
23. Rasmussen SM, Hedlng LG, Parbst E. Serum IRI in insulintreated diabetics during a 24-hour period. Diabetologia 1975;
11:151-158
24. Falko JM, Schonfeld G, Wltztum JL, Kolar JB, Salmon P.
Effects of short-term high carbohydrate, fat-free diet on plasma levels of apo c-ll and apo c-lll and on apo c subspecies in
human plasma lipoproteins. Metabolism 1980:29:654-661
25. Ginsberg HN, Le N-A, Mellsh J, Steinberg D, Brown WV.
Effect of a high carbohydrate diet on apoprotein-B catabolism
in man. Metabolism 1981:30:347-353
• HDL cholesterol
• cholesterol
• triglycende
Insulin-mediated increases in the HDL cholesterol/cholesterol ratio in humans.
C N Sadur and R H Eckel
Downloaded from http://atvb.ahajournals.org/ by guest on May 10, 2017
Arterioscler Thromb Vasc Biol. 1983;3:339-343
doi: 10.1161/01.ATV.3.4.339
Arteriosclerosis, Thrombosis, and Vascular Biology is published by the American Heart Association, 7272 Greenville
Avenue, Dallas, TX 75231
Copyright © 1983 American Heart Association, Inc. All rights reserved.
Print ISSN: 1079-5642. Online ISSN: 1524-4636
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://atvb.ahajournals.org/content/3/4/339
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Arteriosclerosis, Thrombosis, and Vascular Biology can be obtained via RightsLink, a service of the Copyright
Clearance Center, not the Editorial Office. Once the online version of the published article for which permission
is being requested is located, click Request Permissions in the middle column of the Web page under Services.
Further information about this process is available in the Permissions and Rights Question and Answerdocument.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Arteriosclerosis, Thrombosis, and Vascular Biology is online
at:
http://atvb.ahajournals.org//subscriptions/