Regulation of tumour necrosis factor-alpha release from human adipose tissue in vitro

Abstract

Tumour necrosis factor-alpha (TNF-alpha), secreted by cells of the macrophage-monocyte lineage, has a well established role in inflammation and host-defence. The more recent discovery that adipocytes also secrete TNF-alpha has led to a substantial body of research implicating this molecule in the insulin resistance of obesity. However, little is known about the normal regulation of TNF-alpha release from human adipose tissue. In particular, it is not known whether adipocyte production of TNF-alpha is responsive to similar or different molecular regulators than those relevant to macrophages. TNF-alpha release from cultured human adipose tissue and isolated adipocytes was examined using an ELISA. Insulin, cortisol or the thiazolidinedione, BRL 49653, did not have a significant effect on TNF-alpha release from adipose tissue or isolated adipocytes. In contrast, lipopolysaccharide (LPS), a major stimulus of TNF-alpha protein production in monocytes and macrophages, resulted in a fivefold stimulation of TNF-alpha release from human adipose tissue. Significant stimulation of TNF-alpha release was also seen from isolated adipocytes, indicating that the increase in TNF-alpha release from adipose tissue in the presence of LPS is unlikely to be entirely attributable to contaminating monocytes or macrophages. Consistent with this observation was the finding that mRNA for CD14, a known cellular receptor for LPS, is expressed in human adipocytes. The increase in TNF-alpha protein release in response to LPS was blocked by an inhibitor of the matrix metalloproteinase responsible for the cleavage of the membrane-bound proform of TNF-alpha, indicating that this release represented regulated secretion and was not due to cell lysis. In conclusion, the regulation of TNF-alpha protein release from human adipose tissue and isolated adipocytes appears to be similar to its regulation in cell types more traditionally implicated in host defence. The production by the adipocyte of a range of molecules involved in host defence-TNF-alpha, factors D, B and C3, interleukin-6, and macrophage colony-stimulating factor--suggest that this cell type may make a significant contribution to innate immunity.

Full text

Regulation of tumour necrosis factor-alpha release from human
adipose tissue in vitro
C P Sewter, J E Digby, F Blows, J Prins and S O’Rahilly
Departments of Medicine and Clinical Biochemistry, University of Cambridge, Addenbrooke’s Hospital, Hills Road, Cambridge CB2 2QQ, UK
(Requests for offprints should be addressed to C P Sewter)
Abstract
Tumour necrosis factor-alpha (TNF-á), secreted by cells
of the macrophage-monocyte lineage, has a well estab-
lished role in inflammation and host-defence. The more
recent discovery that adipocytes also secrete TNF-á has
led to a substantial body of research implicating this
molecule in the insulin resistance of obesity. However,
little is known about the normal regulation of TNF-á
release from human adipose tissue. In particular, it is not
known whether adipocyte production of TNF-á is
responsive to similar or different molecular regulators than
those relevant to macrophages. TNF-á release from cul-
tured human adipose tissue and isolated adipocytes was
examined using an ELISA. Insulin, cortisol or the thiazo-
lidinedione, BRL 49653, did not have a significant effect
on TNF-á release from adipose tissue or isolated adi-
pocytes. In contrast, lipopolysaccaride (LPS), a major
stimulus of TNF-á protein production in monocytes and
macrophages, resulted in a fivefold stimulation of TNF-á
release from human adipose tissue. Significant stimulation
of TNF-á release was also seen from isolated adipocytes,
indicating that the increase in TNF-á release from adipose
tissue in the presence of LPS is unlikely to be entirely
attributable to contaminating monocytes or macrophages.
Consistent with this observation was the finding that
mRNA for CD14, a known cellular receptor for LPS, is
expressed in human adipocytes. The increase in TNF-á
protein release in response to LPS was blocked by an
inhibitor of the matrix metalloproteinase responsible for
the cleavage of the membrane-bound proform of TNF-á,
indicating that this release represented regulated secretion
and was not due to cell lysis.
In conclusion, the regulation of TNF-á protein release
from human adipose tissue and isolated adipocytes appears
to be similar to its regulation in cell types more tradition-
ally implicated in host defence. The production by the
adipocyte of a range of molecules involved in host defence
– TNF-á, factors D, B and C3, interleukin-6, and
macrophage colony-stimulating factor – suggest that this
cell type may make a significant contribution to innate
immunity.
Journal of Endocrinology (1999) 163, 33–38
Introduction
Tumour necrosis factor-alpha (TNF-á) is a proinflam-
matory cytokine secreted by several cell types such as
monocytes (Waage & Bakke 1988), macrophages
(Mijatovic et al. 1997) and adipocytes (Hotamisligil et al.
1995). It is produced as a 26 kDa active membrane-bound
precursor that is proteolytically cleaved by a matrix
metalloproteinase (TNF-á converting enzyme (TACE))
to release a 17 kDa soluble form (Gearing et al. 1994). Its
expression and production in macrophages have been
shown to be increased on exposure to endotoxins such as
lipopolysaccharide (LPS) (Mijatovic et al. 1997), a compo-
nent of the cell wall of Gram-negative bacteria. TNF-á
has an important role as a mediator of the acute phase
response and, in addition to an immunological function, it
has been shown to have a variety of effects on lipid
metabolism and adipocyte function. These include the
stimulation of lipolysis through increases in hormone-
sensitive lipase expression (Sumida et al. 1997) and inhi-
bition of lipoprotein lipase (Fried & Zechner 1989),
induction of preadipocyte dedifferentiation (decreasing
CCAAT/enhancer binding protein (Stephens & Pekala
1991) and peroxisome proliferator-activated receptor
gamma (PPARã) (Zhang et al. 1996)), promotion of
adipocyte apoptosis (Prins et al. 1997) and induction of
insulin resistance (decreasing GLUT4 (Hauner et al. 1995)
and insulin-stimulated autophosphorylation of insulin
receptor and phosphorylation of insulin receptor
substrate-1 (Feinstein et al. 1993)). Expression of TNF-á
in adipose tissue has been shown to be increased in rodent
and human obesity and to correlate with insulin resistance
(Hotamisligil et al. 1995, Kern et al. 1995). As a result of
these observations, the paracrine effects of TNF-á released
by adipocytes have been proposed to underlie the link
between obesity and insulin resistance (Hotamisiligil et al.
1993). Despite the potential pathogenic importance of
adipocyte TNF-á in obesity remarkably little is known
33
Journal of Endocrinology (1999) 163, 33–38
0022–0795/99/0163–0033  1999 Society for Endocrinology Printed in Great Britain
Online version via http://www.endocrinology.org

about the normal role of TNF-á in human adipose tissue.
We reasoned that a better definition of the molecular
regulators of TNF-á in human adipose tissue might
provide clues to the function of this cytokine in normal
human fat cell biology.
Materials and Methods
Acquisition of human tissue
Omental and subcutaneous adipose tissue biopsies
(approximately 3 g) were obtained from patients under-
going elective open surgery. None of the patients had
diabetes or severe systemic illness. Cambridge Local
Research Ethics Committee approval was obtained, and all
patients gave their informed consent. The tissue donor
group consisted of 11 women (ages 49·215·2 years;
body mass index (BMI) 25·94·3 kg/m
2
) and five men
(ages 71·28·8 years; BMI 264·5 kg/m
2
).
Adipose tissue culture
Whole adipose tissue was separated from surrounding
blood vessels, diced finely and washed twice in Hanks’
balanced salt solution (HBSS) containing 0·2% BSA. Small
pieces of tissue were then placed in sterile tissue culture
plates containing pre-equilibrated medium 199 containing
10% fetal bovine serum, 2 mM -glutamine, 100 U/ml
penicillin, 0·1 mg/ml streptomycin, 33 µM biotin,
17 µM pantothenic acid. Different test compounds
were also added to the medium: human insulin (7 µM),
cortisol (1 mg/ml), LPS (50 µg/ml), a thiazolidinedione
BRL 49653 (10
-7
M) (from SmithKline Beecham
Pharmaceuticals, Harlow, Essex, UK), a metalloproteinase
inhibitor BB3103 (10 µM) (from British Biotechnology,
Oxford, UK) and dimethyl sulphoxide. The plates were
then incubated under standard conditions (37 C/5%
CO
2
). After 20 h, medium was collected from each well
and frozen at 80 C before measurement of TNF-á
protein levels by ELISA. The tissue samples were collected
from each well, blotted onto filter paper to remove excess
medium, and weighed. TNF-á levels were expressed as
pg TNFá per mg adipose tissue per hour and fold increase
or decrease relative to the control (medium only). Unless
specifically indicated, all reagents were obtained from
Sigma Chemical Company, Poole, Dorset, UK.
Adipocyte isolation and culture
Adipose tissue biopsies were taken under sterile conditions
at the time of surgery and placed in normal saline. The
transport time to the laboratory was less than 1 h. The
tissue was diced finely into 1–2 mm pieces. All instru-
ments were sterile disposable or autoclaved. The diced
tissue was digested in a collagenase solution containing
HBSS, 3 mg/ml collagenase (type II) and 1·5% BSA. At
least three volumes more digest solution than the volume
of tissue was used. The digestion was carried out in a
shaking waterbath at 37 C for approximately 1 h. After
filtration through a 260 µm steel mesh, the adipocytes
were allowed to float to the surface and collected using a
pasteur pipette. After two washes in HBSS, aliquots of
adipocytes were dispensed into tissue culture plates con-
taining pre-equilibrated 199 medium and test compounds
(see adipose tissue culture). After 20 h, infranatants were
harvested from beneath the adipocyte layer and frozen at
80 C before measurement of TNF-á by ELISA. Re-
sults are expressed as pg TNFá per µl adipocyte suspension
per hour or percentage increase or decrease relative to the
control (medium alone).
TNF-á protein measurement
TNF-á levels were determined in culture medium using
a commercially available ELISA kit (Genzyme Diag-
nostics, Cambridge, MA, USA). The manufacturer’s
procedure for enhanced sensitivity was followed. The
assay range was 4–512 pg/ml, with a limit of sensitivity
of 500 fg/ml. The intra-assay coefficient of variation was
9·67%, and the inter-assay coefficent of variation was
12·47%.
Detection of CD14 mRNA by reverse
transcription-polymerase chain reaction
RT-PCR was carried out on RNA extracted from human
adipocytes, preadipocytes and monocyte cells. Adipocyte
and preadipocyte cells were isolated by collagense
digestion of whole adipose tissue as described above.
Monocyte cells were isolated from whole blood using
HISTOPAQUE-1077 from Sigma. The manufacturer’s
procedure was followed. RNA was extracted using a
guanidium thiocyanate–phenol technique (Tri Reagent,
Sigma). RNA samples were treated with RNase inhibitor
and DNase and quantified by spectrophotometry. The
integrity of RNA samples was assessed by agarose gel
electrophoresis and ethidium bromide staining. The
reverse transcription reaction was performed in a 20 µl
volume containing 5 mM MgCl
2
,1 reverse tran-
scriptase buffer (Promega, Cambridge, UK), 1 mM
dNTPs (Pharmacia Biotech, St Albans, Herts, UK), 10 ng
random hexamers (Promega). The reactions were incu-
bated at 65 C for 5 min to remove any residual DNAase
A activity and allowed to cool to room temperature. One
microlitre Moloney murine leukemia virus reverse tran-
scriptase (Promega) was added to the reactions (except the
control) and incubated at 37 C for 1 h. The control
reaction, containing no reverse transcriptase, allowed for
the detection of any possible genomic DNA contami-
nation in the RNA samples after PCR. A further control
containing all reagents except RNA was also included to
C P SEWTER and others · Regulation of human adipose TNF release34
Journal of Endocrinology (1999) 163, 33–38

allow the detection of any possible DNA/RNA con-
tamination of reagents. PCR was then performed on
200 ng first-strand cDNA in a 50 µl volume containing
1 mM MgCl
2
, 0·2 µM each of CD14 sense (5-
GGAAAGAAGCTAAAGCACTT-3) and CD14 anti-
sense (5-TTTAGAAACGGCTCTAGGTT-3) primers,
0·15 mM dNTPs, 1  NH
4
reaction buffer, 2·5 units
Advantage 2 Taq polymerase (Clontech, Palo Alto, CA,
USA). Cycling parameters were: 94 C for 1 min, 50 C
for 1 min, 68 C for 1 min (30 cycles). PCR products
were confirmed as CD14 by nucleotide sequencing.
All sequencing was carried out using Dideoxy termi-
nator chemistry (Perkin-Elmer, Foster City, CA, USA)
and electrophoresed on an ABI 377 automated DNA
sequencer.
Statistical analysis
A paired Wilcoxon non-parametric test was used for
statistical analysis. Results are expressed as means...
Results
Mean release of TNF-á from whole adipose tissue
and isolated adipocytes in culture medium was
0·0680·02 pg/mg adipose tissue per hour and
0·0170·01 pg/µl adipocyte suspension per hour respect-
ively. Stability of TNF-á protein in the medium was
confirmed by spiking medium containing whole adipose
tissue with known amounts of TNF-á and quantitating
TNF-á protein concentrations after a 20-h incubation
relative to an unspiked control (data not shown). There
was no correlation between TNF-á protein release and
BMI or age and no consistent gender differences
were seen.
Exposure to insulin (7 µM), cortisol (1 mg/ml) or BRL
49653 (10
-7
M) had no significant effect on TNF-á release
from either whole adipose tissue (Fig. 1) or isolated
adipocytes (Fig. 2). In contrast, exposure to LPS resulted
in a mean fivefold (1·85 ...) increase in TNF-á
release from adipose tissue after 20 h (relative to controls)
(P=0·0134). Figure 3 shows that BB3103 (10 µM), a
matrix metalloproteinase inhibitor, was able to inhibit
LPS-stimulated TNF-á release from adipose tissue over
the 20-h period by 903·8% (P=0·002). BB3103
(10 µM) inhibited basal (unstimulated) TNF-á release
from adipose tissue by 3017·4% (P=NS). The inhibi-
tory effect of BB3103 suggests that the increase in release
of TNF-á from adipose tissue exposed to LPS is due to a
bona fide effect of LPS and not due to cell lysis. LPS also
had a stimulatory effect on TNF-á release from isolated
adipocytes (Fig. 2), although to a lesser extent compared
with whole adipose tissue (1·50·15-fold relative to
control; P=0·0469). In the presence of 10 µM BB3103,
this increase was inhibited by 7013·75% (P=0·03)
(Fig. 4). Under identical conditions, exposure of human
monocyte-macrophages to LPS resulted in a 50-fold
increase in TNF-á release relative to unstimulated controls
(P=<0·0001 data not shown).
Figure 1 TNF-á release into the incubation medium from human
adipose tissue after a 20-h incubation period in medium alone
(Control) or medium containing 50 µg/ml LPS (n=14), 7 µM insulin
(n=11), 1 mg/ml cortisol (n=11),or10
7
M BRL 49653 (n=11). In
each experiment, samples were incubated in duplicate. Data are
expressed as means
S.E.M. *P=0·0314 compared with Control.
Figure 2 TNF-á release into the incubation medium from isolated
human adipocytes after a 20-h incubation period in medium alone
(Control) or medium containing 50 µg/ml LPS (n=7), 7 µM insulin
(n=3), 1 mg/ml cortisol (n=3),or10
7
M BRL 49653 (n=4). In
each experiment, samples were incubated in duplicate. Data are
expressed as means
S.E.M. *P=0·0469 compared with Control.
Regulation of human adipose TNF release · C P SEWTER and others 35
Journal of Endocrinology (1999) 163, 33–38

Previous reports on the study of LPS induction of
TNF-á protein production in monocytes and macrophages
have shown that this is mediated by CD14, a cell-surface
receptor for LPS. CD14 mRNA was detected by RT-
PCR in isolated adipocytes, preadipocytes and monocytes
(Fig. 5).
Discussion
Although a large body of research on TNF-á action in
metabolically important tissue such as muscle and fat has
recently emerged, little information has been forthcoming
on the regulation of TNF-á release from adipose tissue,
and even less on TNF-á release from human adipose
tissue. In this study, we have demonstrated that TNF-á
release from human adipose tissue is responsive to LPS but
not to insulin, cortisol or a thiazolidinedione. LPS also
stimulated TNF-á release from isolated adipocytes, albeit
less robustly than in whole adipose tissue. The LPS-
stimulated release of TNF-á was prevented by an inhibitor
of the metalloproteinase responsible for the cleavage of the
membrane-bound form of TNF. Finally, adipocytes were
found to express mRNA encoding CD14, a known
receptor for LPS.
As in unstimulated monocytes and macrophages
(Crawford et al. 1997), TNF-á release from human
adipose tissue and isolated adipocytes under control con-
ditions occurred at low levels. This is consistent with our
previously published data concerning TNF-á mRNA
expression which indicated that this gene is expressed at
very low abundance in human adipocytes (Montague et al.
1998). Interestingly, we were unable to demonstrate any
correlation between adipose tissue TNF-á release and the
BMI of the subjects studied. This is in contrast to previous
studies which demonstrated a 2·5-fold increase in fat tissue
TNF-á expression in obese (BMI=39·91·4 kg/m
2
)
compared with lean (BMI=21·40·3 kg/m
2
) individuals
(Hotamisiligil et al. 1995). The narrow range of BMIs in
the present study (20–30 kg/m
2
) would not have allowed
us to detect an effect of marked obesity. This is consistent
with our previously published data concerning TNF-á
mRNA expression in adipocytes from a study population
with a similar range of BMI values (Montague et al. 1998).
Circulating TNF-á protein concentrations have been
reported to correlate with hyperinsulinaemia in humans
Figure 3 TNF-á release from human adipose tissue after a 20-h
incubation in medium alone (Control), medium plus BB3103
(10 µM), medium plus LPS (50 µg/ml) plus BB3103 vehicle, or
medium containing LPS (50 µg/ml) plus 10 µM BB3103 (n=10). In
each experiment, samples were incubated in duplicate. Data are
expressed as means
S.E.M. *P=0·0078, LPS+BB3103 vehicle
compared with Control; **P=0·002, LPS+BB3103 vehicle
compared with LPS+BB3103.
Figure 4 TNF-á release from isolated human adipocytes after a
20-h incubation in medium alone (Control), medium plus LPS
(50 µg/ml) plus BB3103 vehicle, or medium containing LPS
(50 µg/ml) plus 10 µM BB3103 (n=2). In each experiment,
samples were incubated in duplicate. Data are expressed as
means
S.E.M. *P=0·0497, LPS+BB3103 vehicle compared with
Control; **P=0·03, LPS+BB3103 vehicle compared with
LPS+BB3103.
C P SEWTER and others · Regulation of human adipose TNF release36
Journal of Endocrinology (1999) 163, 33–38

(Hotamisligil et al. 1995). In addition, insulin has been
reported to have a stimulatory effect on TNF-á release
from monocyte and macrophages (Gepner-Atlan et al.
1996, Hahn & Filkins 1993). In these studies, however,
insulin did not have any significant effect on release of
TNF-á protein from either adipose tissue or isolated
adipocytes. Cortisol has previously been shown to
markedly suppress the endotoxin-induced increase in
TNF-á protein production in human macrophages
(Beutler et al. 1986). No significant suppression of adipo-
cyte TNF-á by cortisol was seen in this study, although
there was a trend towards a reduction. The low basal levels
of TNF-á release may have made an inhibitory effect
difficult to detect, and future studies should consider the
effects of cortisol on LPS-induced secretion. Thiazolidin-
edione compounds are agonists at the PPARã receptor
(Berger et al. 1996), a nuclear hormone receptor highly
expressed in fat (Chawla et al. 1994). They act as insulin
sensitisers (Saltiel & Olefsky 1996) and, in obese rodents
treated in vivo have been demonstrated to reduce adi-
pocyte TNF-á expression (Peraldi et al. 1997). In contrast,
no effects of the thiazolidinedione, BRL 49653, on
TNF-á release from fat tissue were seen in the present
study. This apparent discrepancy may relate to species
differences, to in vivo and in vitro differences, or to the fact
that only fat from non-obese humans was studied.
LPS is a cell-wall component of Gram-negative bacteria
that powerfully stimulates TNF-á protein production in
monocytes and macrophages (Beutler et al. 1986). In this
study, we have demonstrated that LPS also stimulates
TNF-á release from human adipose tissue and from
isolated human adipocytes. The greater stimulation of
TNF-á from whole adipose tissue than from isolated
adipocytes may suggest that cells other than adipocytes are
responsible for most of the LPS response. However,
precise quantitative comparisons are difficult, as the iso-
lated adipocytes have inevitably undergone an isolation
procedure that may impair their subsequent responsiveness
in vitro. In this regard, Hotamisligil et al. (1993) have
reported that the majority of TNF-á expression in fat
tissue is in adipocytes, with a minor contribution from the
stromovascular fraction.
We have shown that the stimulatory effect of LPS on
TNF-á release can be blocked by BB3103, a matrix
metalloproteinase inhibitor, and that this effect is seen in
both whole adipose tissue and isolated adipocytes. These
findings suggest that the increase in TNF-á levels seen
on exposure to LPS occurs via normal physiological
mechanisms and is unlikely to represent a non-specific
lytic effect of LPS. Our results also demonstrate, for the
first time, that the release of TNF-á from adipocytes
occurs through the same metalloproteinase-mediated
mechanism as that seen in more classical TNF-á secreting
cells.
The finding that LPS stimulates TNF-á production
from human adipocytes suggests a role for adipocyte TNF
in host defence. This adds to a growing body of evidence
pointing towards a contribution of adipose tissue to innate
immunity. Thus adipocytes are a major source of several
components of the complement system including factors D
(adipsin), B and C3 (Peake et al. 1997). Levine et al. (1998)
have recently shown that adipocytes produce macrophage
colony-stimulating factor, a proliferative stimulus for mac-
rophages, and that this is upregulated in growing adipose
tissue, as seen in ‘creeping fat’ that surrounds inflammatory
intestinal lesions such as in Crohn’s disease. Adipose tissue
is also known to be a source of interleukin-6 (Fried et al.
1998), a pleiotropic cytokine that has a systemic role in
infections and inflammatory disorders as part of the acute
phase response. Finally, leptin, a major secretory product of
adipocytes, has important effects on lymphocyte function
(Lord et al. 1998). In addition, Gainsford and co-workers
(1996) found that addition of leptin to culture medium
enhanced cytokine production and phagocytosis of
Leishmania parasites by murine peritoneal macrophages.
Thus our finding that LPS is a significant secretagogue
for TNF-á release from human adipose tissue and adi-
pocytes adds to this growing body of evidence for a role of
the adipocyte in innate immunity. These findings do not,
however, exclude additional roles for adipocyte TNF-á in
the control of other aspects of fat cell biology, such as
differentiation, apoptosis and lipid metabolism.
Acknowledgements
This study was supported by the British Diabetic Associ-
ation and the Wellcome Trust. J P is a Wellcome Trust
International Travelling Fellow. The authors would like to
thank Professor Ken Siddle for helpful discussions. The
authors would also like to acknowledge the support of
Dominic Corkill from British Biotechnology Limited. The
assistance of members of the departments of Surgery and
Figure 5 Detection of CD14 mRNA by RT-PCR in isolated adipocytes (A), monocytes (M) and
preadipocytes (PA). C, RT-PCR reactions in which reverse transcriptase enzyme was omitted; L,
1 kb ladder; RB, RT-PCR amplification mixture without template as a negative control.
Regulation of human adipose TNF release · C P SEWTER and others 37
Journal of Endocrinology (1999) 163, 33–38

Obstetrics and Gynaecology of Addenbrooke’s Hospital
and the University of Cambridge, in particular Peter
Friend, Neville Jamieson, John Williamson and Andrew
Prentice, is greatly appreciated.
References
Berger J, Bailey P, Biswas C, Cullinan CA, Doebber TW, Hayes NS,
Saperstein R, Smith RG & Leibowitz MD 1996 Thiazolidinediones
produce a conformational change in peroxisomal proliferator-
activated receptor-ã: binding and activation correlate with
antidiabetic actions in db/db mice. Endocrinology 137 4189–4195.
Beutler B, Krochin N, Milsark I, LuedkeC&Cerami A 1986
Control of cachectin (tumor necrosis factor) synthesis: mechanisms
of endotoxin resistance. Science 232 977–979.
Chawla A, Schwarz E, DimaculanganD&Lazar M 1994 Peroxisome
proliferator-activated receptor (PPAR) gamma: adipose-
predominant expression and induction early in adipocyte
differentiation. Endocrinology 135 798–800.
Crawford E, Ensor J, KalvakolanuI&Hadsay J 1997 The role of 3
poly (A) tail metabolism in tumor necrosis factor-alpha regulation.
Journal of Biological Chemistry 272 21120–21127.
Feinstein R, Kanety H, Papa M, LunenfeldB&Karasik A 1993
Tumor necrosis factor-alpha suppresses insulin-induced tyrosine
phosphorylation of insulin-receptor and its substrates. Journal of
Biological Chemistry 268 26055–26058.
FriedS&ZechnerR1989 Cachectin tumor necrosis factor decreases
human adipose-tissue lipoprotein-lipase messenger-RNA levels,
synthesis, and activity. Journal of Lipid Research 30 1917–1923.
Fried S, BunkinD&Greenberg A 1998 Omental and subcutaneous
adipose tissues of obese subjects release interleukin-6: depot
difference and regulation by glucocorticoid. Journal of Clinical
Endocrinology and Metabolism 83 847–850.
Gainsford T, Willson T, Metcali D, Handman E, McFarlane C, Ng
A, Nicola NA, Alexander WS & Hilton DJ 1996 Leptin can induce
proliferation, differentiation, and functional activation of hemo-
poietic cells. Proceedings of the National Academy of Sciences of the
USA 93 14564–14568.
Gearing A, Beckett P, Christodoulou M, Churchill M, Clements J,
Davidson AH, Drummond AH, Galloway WA, Gilbert R, Gordon
JL, Leber TM, Mangan M, Miller K, Nayee P, Owen K, Patel S,
Thomas W, Wells G, Wood LM & Woolley K 1994 Processing of
tumor necrosis factor-á precursor by metalloproteinases. Nature 370
555–557.
Gepner-Atlan C, Bongrand P, Farnarier C, Xerri L, Choux R,
Gauthier J-F, Brue T, Vague P, Grob J-J & Vialettes B 1996
Insulin-induced lipoatrophy in type I diabetes. Diabetes Care 19
1283–1285.
HahnE&Filkins J 1993 Insulin restores plasma tumor-necrosis-factor
(TNF) in the dexamethasone protected endotoxic rat. FASEB
Journal 7 A420.
Hauner H, Petruschke T, Russ M, RohrigK&Eckel J 1995 Effects
of tumor necrosis factor-alpha (TNF-alpha) on glucose transport and
lipid metabolism in newly-differentiated fat cells in culture.
Diabetologia 38 764–771.
Hotamisligil G, Shargill N & Speigelman B 1993 Adipose expression
of tumor necrosis factor-á: direct role in obesity-linked insulin
resistance. Science 259 87–91.
Hotamisligil G, Arner P, Caro J, AtkinsonR&Spiegelman B 1995
Increased adipose tissue expression of tumor necrosis factor-á in
human obesity and insulin resistance. Journal of Clinical Investigation
95 2409–2415.
Kern P, Saghizadeh M, Ong J, Bosch R, DeemR&SimsoloR1995
The expression of tumor necrosis factor in human adipose tissue.
Journal of Clinical Investigation 95 2111–2119.
Levine J, Jensen M, EberhardtN&O’Brien T 1998 Adipocyte
macrophage colony-stimulating factor is a mediator of adipose tissue
growth. Journal of Clinical Investigation 101 1557–1564.
Lord G, Matarese G, Howard J, Baker R, BloomR&LechlerR
1998 Leptin modulates the T-cell immune response and reverses
starvation-induced immunosuppression. Nature 394 897–901.
Mijatovic T, Kruys V, Caput D, DefranceP&HuezG1997
Interleukin-4 and -13 inhibit tumor necrosis factor alpha mRNA
translational activation in lipopolysaccaride-induced mouse
macrophages. Journal of Biological Chemistry 22 14394–14398.
Montague C, Prins J, Sanders L, Zhang JL, Sewter CP, Digby J,
Byrne CD & O’Rahilly S 1998 Depot-related gene expression in
human subcutaneous and omental adipocytes. Diabetes 47
1384–1391.
Peake P, O’Grady S, PussellB&Charlesworth J 1997 Detection and
quantification of the control proteins of the alternative pathway of
the complement in 3T3-L1 adipocytes. European Journal of Clinical
Investigation 27 922–927.
Peraldi P, Xu M & Spiegelman B 1997 Thiazolidinediones block
tumor necrosis factor-á-induced inhibition of insulin signaling.
Journal of Clinical Investigation 100 1863–1869.
Prins J, Niesler C, Winterford C, Bright NA, Siddle K, O’Rahilly S,
Walker NI & Cameron DP 1997 Tumor necrosis factor-á induces
apoptosis of human adipose cells. Diabetes 46 1939–1944.
SaltielA&Olefsky J 1996 Thiazolidinediones in the treatment of
insulin resistance and type II diabetes. Diabetes 45 1661–1669.
StephensJ&Pekala P 1991 Transcriptional repression of the GLUT4
and C/EBP genes in 3T3-L1 adipocytes by tumor necrosis
factor-alpha. Journal of Biological Chemistry 266 21839–21845.
Sumida M, Shiosaka T, Nagar A, Isshikimasuda M, Okuda H &
Hamada M 1997 Suppresive effect of tumor necrosis factor-alpha
on adipogenic cell differentiation and on gene expression of
hormone-sensitive lipase. Journal of Clinical Biochemistry and Nutrition
22 1–11.
WaageA&Bakke O 1988 Glucocorticoids suppress the production of
tumor necrosis factor by lipopolysaccaride-stimulated human
monocytes. Immunology 63 299–302.
Zhang B, Berger J, Hu E, Szalkowski D, White-Carrington S,
Spiegelman BM & Moller DE 1996 Negative regulation of
peroxisome proliferator-activated receptor-ã gene expression
contributes to the anti-adipogenic effects of tumor necrosis factor-á.
Molecular Endrocrinology 10 1457–1466.
Received 29 December 1998
Revised manuscript received 28 April 1999
Accepted 5 May 1999
C P SEWTER and others · Regulation of human adipose TNF release38
Journal of Endocrinology (1999) 163, 33–38