| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
University of Colorado Health Sciences Center, Center for Human Nutrition, Denver, Colorado 80262
Address all correspondence and requests for reprints to: Catherine Morin, University of Colorado Health Sciences Center, San Luis Valley Health Studies, 1016 West Avenue no. 4, Alamosa, Colorado 81101. E-mail: cathymorin{at}amigo.net
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
After 1 week of quarantine, rats were provided a semipurified diet and
fed ad libitum for 5 weeks (LF; Research Diets, New
Brunswick, NJ) (Table 1
) with 12% of
calories from fat (corn oil), 20% from protein (casein), and 68% from
carbohydrate (maltodextrins and cornstarch). The results presented in
this paper are part of a larger experiment involving several diets in
four age groups, which will be reported at a later date.
|
92.5% of
presurgery body weight on the day of study.
On the day of the experiment, extensions were added to catheters of
6–8 h fasted animals for ease of sampling and animals were allowed to
rest for 20 min. Following this, a baseline blood sample (for
preexperiment plasma glucose and insulin concentrations) was taken
along with blood pressure and hematocrit. Blood pressure was obtained
from the artery using a calibrated electronic blood pressure unit
(Stoelting, Wood Dale, IL). Then either basal or euglycemic,
hyperinsulinemic clamps were initiated. The basal study consisted of a
primed (12 µCi), continuous (0.1 µCi/min) infusion of HPLC purified
[3-3H]glucose in saline for a 90-min period. Euglycemic,
hyperinsulinemic clamps consisted of a primed, continuous infusion of
insulin (4 mU/kg·min) and [3-3H]glucose. A variable
glucose infusion (10 or 20% dextrose) was used to maintain plasma
glucose at baseline values. The glucose infusate was spiked with
[3-3H]glucose to a specific activity that was similar to
the plasma specific activity that would occur from the continuous
infusion alone. This was done to minimize changes in glucose specific
activity. The total experiment time was approximately 90 min during
which arterial blood was sampled at approximately 5-min intervals and
the glucose infusion rate was adjusted accordingly to maintain
euglycemia (
7 mM). As a relative index of
tissue-specific glucose uptake, a bolus injection of
2-deoxy-D-[1-14C]glucose (2DG, 40 µCi) was
administered via the carotid cannula at approximately 45 min
(steady-state glucose levels during both basal and clamp studies).
Blood samples were then taken at 2.5, 5, 10, 15, 20, 30, 35, 40, and 45
min. Circulating insulin concentration during the experiment was
determined from the final blood sample taken. No more than 12% of the
animals’ blood volume (assumed to be 8% of body weight) was taken.
After the last blood sample, the animal was anesthetized with sodium
pentobarbital (IV, 70 mg/kg) and the following tissues were removed and
immediately frozen with precooled clamps and then placed in liquid
nitrogen for subsequent tracer and metabolite analyzes: liver,
gastrocnemius, soleus, and biceps femoris. The epididymal (EPI),
retroperitoneal (RETRO), and mesenteric (MES) fat pads were removed,
weighed, and frozen. A portion of subcutaneous (SUB) fat, from the
region above the biceps femoris, was also removed and frozen. The sum
of the EPI, RETRO, and MES fat pads was used as a marker for body
composition as previous data (unpublished observations) have shown that
the sum of these fat pads is highly correlated with total carcass
lipid, r = 0.93. Because it has been hypothesized that visceral
fat may contribute more to the development of insulin resistance (8),
both a visceral fat pad, EPI, and a SUB fat pad (subscapular fat above
biceps femoris) were examined for differences in AT-TNF activity.
TNF bioassay
Under sterile conditions adipose tissue was minced and incubated
in DMEM containing 0.5% low endotoxin-fatty acid free albumin
(Sigma Chemical Co., St. Louis, MO) for 1 h. Medium
was collected and frozen at -70 C for later analysis.
TNF activity was measured in a bioassay that utilizes WEHI cells, a murine fibroblastic cell line that is very sensitive to TNF (9). In this paper, AT-TNF activity was defined as the TNF bioactivity measured with this method. Samples containing AT-TNF were aliquoted into a 96-well plate in triplicate, in serial dilution, with the WEHI cells. After 48 h the number of live cells was quantified using Alamar blue (Alamar Biosciences, Inc. Sacramento, CA), a dye that is reduced by live cells only. Cytotoxicity curves for each sample were then generated based on the percent survival of cells and the log of sample dilution. TNF activity is expressed as pg/ml relative to that of the recombinant TNF standard curve (R&D Systems, Minneapolis, MN). AT-TNF containing samples incubated with a TNF antibody (R&D Systems) showed no reduction in live cells confirming that AT-TNF was the only cytotoxic protein secreted from these adipose tissue samples. The assay typically had a 2% interassay coefficient of variation.
TNF protein
TNF protein was measured with a rat ELISA (Biosource International, Camarillo, CA) which was previously validated in
our laboratory (6).
RNA-PCR
Homogenized tissue was incubated in a guanidine-containing
solution, Trizol (Gibco BRL, Baltimore, MD). The RNA was
then purified using centrifugation and ethanol precipitation. After
DNase 1 treatment, RNA was quantified and frozen in 10 ug aliquots.
RNA-PCR was performed using Ambion, Inc.’s QuantumRNA kit
(Austin, TX). In this system quantitation is based on 18S mRNA levels
which are titrated to match levels of the mRNA of interest. The TNF
primers used were: 5'-TACTGAACTTCGGGGTGATTGGTCC-3',
5'-CAGCCTTGTCCCTTGAAGAGAACC-3'. RNA and random hexamers were mixed and
heated to 65 C for 10 min. RT was performed in 50 mM KCl,
10 mM Tris-HCL, pH 8.3, 1 mM deoxynucleotide
triphosphates, 5 mM MgCl2, 1 U/µl RNase
inhibitor, and 300 U RT in a 20 µl reaction. For every RNA sample, a
no-RT control was run to assure no DNA contamination. All no-RT
reactions were negative indicating that all PCR products were of RNA
origin. From each RT reaction, 4 µl was added to each PCR reaction
(25 pmol of each primer, 1.5 mM MgCl2, 2
mM deoxynucleotide triphosphates, 18S primers and
Competimers, and 4 U Taq polymerase in 50 µl total). The
PCR reaction was amplified for 40 cycles (94 C, 30 sec; 60 C, 30 sec;
72 C, 1 min; Perkin Elmer Corp. 9600), as this was the
number of cycles previously determined to be in the exponential range
for TNF and 18S with Competimers at 7:3 ratio amplification
(unpublished observations). PCR products were run on a 4% gel
(NuSieve, FMC, Rockland, ME), and the bands were then quantified by
densitometer.
Cell sizing and number
Fat cell size was determined by measuring the diameter of 50
collagenase-treated cells (10) under a microscope. Fat cell number
(FCN) was determined using the following formula: number of
cells/g = avg cell size (pl) x 0.95 (ng lipid/1 pl) x
g/109 ng. This product was then multiplied by
the number of grams of tissue to get FCN. We have previously found this
method correlates well with FCN determinations (r > 0.9) based on
lipid content (unpublished data). Nevertheless, lipid content was
measured in all the fat samples (11) and used to calculate FCN and TNF
activity/106 cells. This method of calculation produced
different absolute numbers but did not change the results or
interpretation. Due to the variation, this measurement introduced, we
have chosen to present the data based on the method using the equations
listed above.
Glucose kinetics
Basal rates of glucose appearance (Ra) and disappearance (Rd)
were estimated by isotope dilution (12). Rates of endogenous glucose Ra
(Endo Ra) and Rd during the euglycemic, hyperinsulinemic clamp were
determined as previously described (13). Samples were collected under
steady-state conditions to avoid underestimation of Endo Ra. Values for
glucose and insulin concentration, glucose specific activity, Endo Ra,
and Rd are the average of three samples taken over the final 10 min
steady-state period. Glucose kinetic data presented are the average of
three time points taken under steady-state conditions. Steady state
conditions were defined as <1.0%/min change in glucose specific
activity. Experiments were excluded if these steady-state conditions
did not exist, or if the coefficient of variation (CV) of the plasma
glucose level during the last 45 min of the clamp was >10%. The CV of
the plasma glucose level over this time period for included experiments
was 5.1±0.6%, and was not significantly different among groups. Data
are reported as means ± SEM.
Tissue specific glucose uptake was estimated in individual tissues (Rg') using the accumulation of phosphorylated 2DG, based on the fact that 2DG is trapped in most tissues, except for liver, and undergoes negligible further reaction. The decay curve of plasma 2DG following a bolus injection was determined over a 45-min period and the specific activity was integrated. The integrated specific activity was divided into the tissue phosphorylated 2DG level to yield Rg'. The use of Rg' as a relative index of glucose uptake in individual tissues is based on the assumption that any difference between 2DG and glucose is unaffected by the experimental conditions (7, 14).
Analytical methods
Plasma tracer samples were deproteinized with
Ba(OH)2 and ZnSO4 and stored at 0EC overnight.
An aliquot was dried to eliminate 3H2O,
reconstituted with distilled water, and 3H and
14C disintegrations per min were determined by liquid
scintillation spectrometry (Beckman Coulter, Inc.
Instruments, Fulterton, CA). Skeletal muscle and adipose tissue
14C-phosphorylated 2DG was determined on homogenates using
ion-exchange chromatography and liquid scintillation counting (14).
Plasma glucose levels were determined by the glucose oxidase method (15) using a Beckman Coulter, Inc. glucose analyzer (Fullerton, CA). Plasma insulin was measured by RIA (Linco Research, Inc., St. Louis, MO).
Data analysis
Data were analyzed by two-way ANOVA. When significant
differences (P < 0.05) were found among groups
pairwise multiple comparisons were made following the
Student’s-Newman-Keuls method (SigmaStat, Jandel, San Rafael, CA).
Forward stepwise regression was used to determine which independent
variables best explained the observed differences in AT-TNF activity.
Data are expressed as means ± SEM
(SEM).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). Fasting
plasma glucose did not change with age (Table 1). Fasting plasma
insulin increased significantly with age (Table 1). Each of the four
fat pads excised increased significantly with age (data for EPI and SUB
presented in Table 2). The sum of EPI,
MES, and RETRO fat pads increased significantly with age (W =
11.7 ± 1.1, Y = 13.1 ± 2.1, M = 25.8 ± 2.9,
O = 47.2 ± 4.5; P < 0.001). Both fat cell
size and number increased significantly with age in both the EPI and
SUB fat pads (Table 2).
|
).
Under hyperinsulinemic, euglycemic conditions, the glucose infusion
rate (GIR), was significantly decreased with age (W = 22.1 ±
5.2, Y = 23.4 ± 0.3, M = 17.1 ± 0.9, O =
15.6 ± 4.2 mg/kg·min; P = 0.008). Both the rate
of endogenous glucose appearance and the rate of glucose disappearance
significantly decreased with age (P = 0.03).
|
). Basal glucose uptake per
106 cells increased with age in SUB only (P
= 0.03). When glucose uptake is expressed per fat pad, then glucose
uptake increased significantly with age in EPI and SUB, largely because
the fat pads significantly expanded with age. Under hyperinsulinemic
conditions, glucose uptake per gram was significantly decreased in the
Older group. However, this decrease was not apparent when glucose
uptake was expressed per 106 cells. As in the basal state,
glucose uptake per fat pad increased during hyperinsulinemia in EPI and
SUB. Under both basal and insulin-stimulated conditions, glucose uptake
was directly related to fat cell volume in both EPI and SUB (greater
than r = 0.77, P < 0.001 in all cases).
|
).
Insulin-stimulated glucose uptake/gram was significantly decreased in
vastus, gastrocnemis, and soleus muscles (Table 5). When
insulin-stimulated glucose uptake was adjusted for total muscle mass,
glucose uptake increased with age in the soleus muscle while the
gastrocnemius muscle showed a trend (P = 0.053) toward
increased glucose uptake in Y and M but not in O.
|
). TNF activity/106 cells
paralleled TNF activity/pad, and correlated with EPI cell size (r
= 0.76, P < 0.001) (Fig. 1B). In SUB, TNF activity/pad
significantly increased with age (Fig. 2A). SUB-TNF activity/106
cells also correlated with SUB cell size (r = 0.58,
P = 0.007) (Fig. 2B).
|
|
). These relationships were
similar whether TNF activity was expressed on a per pad or a per cell
basis (only pad correlations shown). However, as noted before TNF
activity correlated positively with fasting plasma insulin levels
(r 0.6, P < 0.001).
|
AT-TNF mRNA did not differ between basal and clamp conditions (data not shown), thus data were combined for further analysis. AT-TNF mRNA did not significantly change with age (W:0.29 ± 0.07, Y:0.33 ± 0.07, M:0.41 ± 0.07, O:0.55 ± 0.08 arbitrary units; P = 0.06). AT-TNF protein also showed no changes with age (W:25 ± 4, Y:20 ± 13, M:18 ± 5, O:12 ± 2 pg/10 cells; P = 0.21). This was true whether the data were expressed per cell or per pad (W:340 ± 62, Y:263 ± 60, M:316 ± 95, O:349 ± 138 pg/EPI pad; P = 0.90). No differences were noted between basal and clamp levels. Secreted levels of AT-TNF protein correlated with AT-TNF activity (r = 0.64, P < 0.0001) but not with TNF mRNA levels (r = -0.03, P = 0.9).
In regression analysis, using fasting plasma insulin, fat cell size, and Rg/106 cells, only cell size predicted TNF activity in both EPI and SUB (EPI: R2 = 0.58, P < 0.001; SUB: R2 = 0.34, P = 0.007). Neither fasting insulin levels or the respective fat tissue glucose uptake contributed to this prediction of TNF activity.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In this study, neither EPI nor SUB TNF activity correlated significantly with any of the parameters of insulin resistance measured during the clamp. It is possible that an increased sample size would result in a significant negative relationship between SUB TNF activity and whole body glucose disappearance as measured during the clamp. However, SUB TNF activity would still only explain approximately 14% of the variance in Rd. Of note, EPI TNF activity showed no relationship to Rd. Thus, adipose tissue TNF activity may not be functionally related to the insulin resistance of aging.
Other studies that have tried to define the role of AT-TNF in insulin resistance have had disparate results. Although some studies have observed elevated TNF mRNA and protein with increased adiposity and insulin levels (1, 16) our previous data showed that a high fat diet resulted in increased AT-TNF activity, but that this increase did not correlate with either fat mass gained or insulin levels (6). This divergence among TNF, obesity, and insulin resistance was also noted in TNF knockout transgenic mice that do not produce TNF from any source (4, 17). Insulin action, as measured by either glucose or insulin challenge, appeared to be similar between the knockouts and wild-type mice. When the mice were challenged with either a high fat diet or gold thio-glucose injections, which typically produce obesity and insulin resistance, the transgenic mice became insulin resistant, albeit to a lesser degree (17). Furthermore, the TNF knockout mice failed to become as insulin resistant as their wild-type littermates as they aged (4). These data suggest that TNF plays a limited role in determining insulin action. Several in vivo studies, using TNF neutralization methods, have also failed to find an association between TNF and insulin resistance (2, 3). There could be several explanations for conflicting data. One likely explanation is that the action(s) of TNF, as with most cytokines, may be reproduced or compensated for, by other cytokines under certain conditions. It should be noted that AT-TNF activity correlates well with adipose cell size. Increased cell size has frequently been associated with insulin resistance (18). Thus, elevated AT-TNF may be coincidentally, but not functionally, associated with insulin resistance.
Clearly the transgenic studies indicate that TNF plays a yet undefined role in insulin resistance, albeit this role is limited and only during specific environmental conditions. Our data suggests that TNF derived from adipose tissue is not functionally related to the development of insulin resistance. Thus, the source of TNF that influences whole body insulin resistance may be from some other organ, such as muscle or pancreas. TNF has been found to be made in muscle, liver, and recently pancreas, all of which are regulated by insulin (19, 20).
It would be easier to understand the role of AT-TNF in insulin resistance if the actions of AT-TNF were known. The two current postulates are that TNF induces insulin resistance through its actions on either FFA and/or insulin receptor phosphorylation (4). In vitro data suggests that TNF could regulate these pathways as well as glucose uptake (21, 22). However, in vivo studies are less clear. The transgenic TNF knockout mice had similar levels of insulin receptor phosphorylation despite the absence of TNF. When mice were fed a high fat diet, receptor phosphorylation was significantly reduced, but this also occurred in mice without TNF. Thus, it is unclear if TNF functions exclusively through receptor phosphorylation to induce insulin resistance. The differences between in vivo and in vitro data may be explained by the dose dependency of TNF’s effects. Low doses of TNF appear to have different effects due to the actions of counter-regulatory hormones (23). Rats infused with low doses of TNF displayed no effect on insulin action despite the fact that the circulating levels of TNF were higher than those found in genetically obese, insulin resistant rats (24). In the current study, AT-TNF activity correlated with cell size. Larger fat cells are also associated with disparate rates of lipolysis and glucose uptake (18, 25).
Alternatively, TNF may regulate some other aspect of adipose tissue metabolism. TNF has been shown to act as either a proliferative agent or to increase angiogenesis (26). In fat tissue, TNF could regulate preadipocyte number, which is known to increase with age and with a high fat diet (27, 28). A decrease in preadipocytes or a lack of increase in preadipocyte number could explain the differences in fat pad weight in the transgenic models.
As anticipated, insulin resistance increased with age. In the basal state, fasting plasma insulin increased with age. Under hyperinsulinemic conditions, the glucose infusion rate (GIR), used as a measure of whole body insulin resistance, indicated that insulin action decreased between 2 and 4 months, with no further decreases thereafter. This pattern of change has been observed previously (29). In the present study, the decrease in Rd was not accompanied by a significant decrease in in vivo glucose uptake in the fat or muscle tissues measured. Other studies have had similar findings suggesting that other tissues or organs contribute to the drop in Rd (30). The increase in insulin resistance was initially due to a decrease in insulin-stimulated glucose disappearance, but Older animals also demonstrated a decreased ability of insulin to suppress endogenous glucose appearance. This suggests that insulin resistance in the liver and/or kidney followed the age-associated decrease in peripheral insulin action. A reduction in insulin suppression of hepatic glucose metabolism has been noted previously (31).
Historically, fat tissue has been thought to become insulin resistant
with age. Typically, these data have been expressed on a per gram
basis. This is most likely due to a decrease in fat cell number per
gram, as fat cells hypertrophy. When glucose uptake was calculated on a
per cell basis, there was no decrease with age. Because fat pads
continue to increase in size with age, glucose uptake by fat actually
increased with age. In fact, as the animal aged and gained fat mass,
fat tissue contributed more, as a percentage, to the total peripheral
glucose uptake (W 2%, Y 1.2%, M 2.8%, O 3.2). Notably,
the absolute amount of glucose taken up by fat is still quantitatively
small and these percentage differences may not be significantly
different.
The present study demonstrates that age-associated insulin resistance is characterized by reductions in both hepatic and nonhepatic insulin action. In the rat, tissues other than fat and muscle appear to contribute to the reductions in insulin’s action on glucose removal. Aging is also associated with higher AT-TNF activity. However, this increase is not significantly related to insulin action but instead to increased fat cell size. Thus, as rats age they become more insulin resistant and have higher AT-TNF activity levels but the two are not related to each other.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Current address: Arizona State University, P.O. Box 870404, Tempe,
Arizona 85287.
Received February 19, 1998.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
: direct role in
obesity-linked insulin resistance. Sci 259:87–91 antibody (CDP571) on
insulin sensitivity and glycemic control in patients with NIDDM.
Diabetes 45:881–885[Abstract]
gene. Diabetes 46:1526–1531[Abstract]
-induced inhibition
of insulin signaling. J Clin Invest 100:1863–1869[Medline]
activity. Endocrinology 138:4665–4671 function. Nature 389:610–614[CrossRef][Medline]
gene and protein expression
in adult feline myocardium after endotoxin administration. J Clin
Invest 96:1042–1052
. J Clin Invest 100:1853–1862[Medline]
-induced insulin resistance in vivo and its
prevention by Troglitazone. Diabetes 46:1678–1683[Abstract]
. Am J Path 151:557–563[Abstract]
This article has been cited by other articles:
 |
|
 |
|
  M. R. Jackman, A. Steig, J. A. Higgins, G. C. Johnson, B. K. Fleming-Elder, D. H. Bessesen, and P. S. MacLean Weight regain after sustained weight reduction is accompanied by suppressed oxidation of dietary fat and adipocyte hyperplasia Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2008; 294(4): R1117 - R1129. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Tchkonia, T. Pirtskhalava, T. Thomou, M. J. Cartwright, B. Wise, I. Karagiannides, A. Shpilman, T. L. Lash, J. D. Becherer, and J. L. Kirkland Increased TNF{alpha} and CCAAT/enhancer-binding protein homologous protein with aging predispose preadipocytes to resist adipogenesis Am J Physiol Endocrinol Metab, December 1, 2007; 293(6): E1810 - E1819. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. S. MacLean, J. A. Higgins, M. R. Jackman, G. C. Johnson, B. K. Fleming-Elder, H. R. Wyatt, E. L. Melanson, and J. O. Hill Peripheral metabolic responses to prolonged weight reduction that promote rapid, efficient regain in obesity-prone rats Am J Physiol Regulatory Integrative Comp Physiol, June 1, 2006; 290(6): R1577 - R1588. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. H. Shen, J.-H. Zhou, S. R. Broussard, R. W. Johnson, R. Dantzer, and K. W. Kelley Tumor Necrosis Factor {alpha} Inhibits Insulin-Like Growth Factor I-Induced Hematopoietic Cell Survival and Proliferation Endocrinology, July 1, 2004; 145(7): 3101 - 3105. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Tollet-Egnell, P. Parini, N. Stahlberg, I. Lonnstedt, N. H. Lee, M. Rudling, A. Flores-Morales, and G. Norstedt Growth hormone-mediated alteration of fuel metabolism in the aged rat as determined from transcript profiles Physiol Genomics, January 15, 2004; 16(2): 261 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Zhang, K.-Y. Guo, P. A. Diaz, M. Heo, and R. L. Leibel Determinants of leptin gene expression in fat depots of lean mice Am J Physiol Regulatory Integrative Comp Physiol, January 1, 2002; 282(1): R226 - R234. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. J. Pagliassotti, E. C. Gayles, D. A. Podolin, Y. Wei, and C. L. Morin Developmental stage modifies diet-induced peripheral insulin resistance in rats Am J Physiol Regulatory Integrative Comp Physiol, January 1, 2000; 278(1): R66 - R73. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Desvergne and W. Wahli Peroxisome Proliferator-Activated Receptors: Nuclear Control of Metabolism Endocr. Rev., October 1, 1999; 20(5): 649 - 688. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
