Gene expression of paired abdominal adipose AQP7 and liver AQP9 in
patients with morbid obesity
Relationship with glucose abnormalities
Merce Mirandaa,b,⁎,1, Victòria Ceperuelo-Mallafréa,b,1, Albert Lecubeb,c, Cristina Hernandezb,c,
Matilde R. Chacona,b, Jose M. Fortd, Lluís Gallarta,b, Juan A. Baena-Fusteguerasd,
Rafael Simób,c, Joan Vendrella,b,e
aUnitat de Recerca, Hospital Universitari de Tarragona Joan XXIII, IISPV, 43007 Tarragona, Spain
bCIBER de Diabetes y Enfermedades Metabólicas Asociadas (CIBERDEM)
cDiabetes and Metabolism Research Group, Vall d'Hebron University Hospital, 08035 Barcelona, Spain
dEndocrinology Surgery Unit, Hospital Universitari Vall d'Hebrón, Institut de Recerca Vall d'Hebrón, Universitat Autònoma de Barcelona, 08035 Barcelona, Spain
eUniversitat Rovira i Virgili, 43201 Tarragona, Spain
Received 17 April 2009; accepted 16 June 2009
The trafficking of glycerol from adipose and hepatic tissue is mainly mediated by 2 aquaporin channel proteins: AQP7 and AQP9,
respectively. In rodents, both aquaporins were found to act in a coordinated manner. The aim was to study the relationship between adipose
AQP7 and hepatic AQP9 messenger RNA expression and the presence of glucose abnormalities simultaneously in morbid obesity. Adipose
tissue (subcutaneous [SAT] and visceral [VAT]) and liver biopsies from the same patient were obtained during bariatric surgery in 30 (21
male and 9 female) morbidly obese subjects. Real-time quantification of AQP7 in SAT and VAT and hepatic AQP9 gene expression were
performed. A 75-g oral glucose tolerance test was performed in all subjects. The homeostasis model assessment of insulin resistance and
lipidic profile were also determined. Visceral adipose tissue AQP7 expression levels were significantly higher than SAT AQP7 (P = .009).
Subcutaneous adipose tissue AQP7 positively correlated with both VAT AQP7 and hepatic AQP9 messenger RNA expression (r = 0.44, P =
.013 and r = 0.45, P = .012, respectively). The correlation between SAT AQP7 and liver AQP9 was stronger in intolerant and type 2 diabetes
mellitus subjects (r = 0.602, P = .011). We have found no differences in compartmental AQP7 adipose tissue distribution or AQP9 hepatic
gene expression according to glucose tolerance classification. The present study provides, for the first time, evidence of coordinated
regulation between adipose aquaglyceroporins, with a greater expression found in visceral fat, and between subcutaneous adipose AQP7 and
hepatic AQP9 gene expression within the context of human morbid obesity.
© 2009 Elsevier Inc. All rights reserved.
Adipose tissue has a major role as an energy storage organ
where many metabolic changes occur in response to the
whole-body energy balance. Under lipogenic conditions,
insulin increases glucose transport into the cell; and glucose
is converted to glycerol-3-phosphate. Likewise, lipoprotein
lipase activated by insulin recruits fatty acids from circula-
tion into the adipocytes; and both are esterified into
triglyceride (TG). Conversely, under lipolytic conditions,
catecholamines stimulate adrenergic receptors that translo-
cate hormone-sensitive lipase, which is a key enzyme in
hydrolyzing TG to free fatty acids (FFA) and glycerol. Free
fatty acids and glycerol are used for thermogenesis and
gluconeogenesis, respectively. In fasting states, gluconeo-
genesis from the liver is the main source of plasma glucose
; and about 22% of total glucose production comes from
glycerol in humans , the adipose tissue being the main
source of plasma glycerol. Trafficking of glycerol from
Available online at www.sciencedirect.com
Metabolism Clinical and Experimental 58 (2009) 1762–1768
E-mail address: email@example.com (M. Miranda).
1These authors contributed equally to this work.
0026-0495/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
adipose and hepatic tissue is mainly mediated by 2 aquaporin
channel proteins: AQP7 and AQP9, respectively. AQP7 is
the sole described channel that permits the exit of glycerol
from adipocytes, and plasma glycerol is introduced into the
hepatocytes by the AQP9 [1,3]. Both aquaporins act in a
coordinated manner. In animal studies, feeding state reduces
the messenger RNA (mRNA) expression of the adipose
AQP7 and results in a reduction of glycerol release from
adipocytes . Feeding also reduces liver AQP9 mRNA
expression and glycerol-induced gluconeogenesis . How-
ever, obese and insulin-resistant db/db mice show increased
AQP7 and AQP9 mRNA levels (in mesenteric fat and liver,
respectively), despite hyperinsulinemia . Increased gly-
cerol release from adipocytes in parallel with increased
gluconeogenesis induced by the high glycerol levels in portal
vein results in hyperglycemia through the pathologic
induction of liver AQP9 .
Adult AQP7-knockout animals develop obesity by
increasing hypertrophic adipocytes in epididymal white
adipose tissue with higher intracellular glycerol content
compared with wild-type mice . Moreover, AQP7-
knockout mice develop severe insulin resistance associated
with obesity. These observations have led us to propose
AQP7 as a new factor influencing not only glycerol but
also glucose metabolism.
Studies on AQP7 in human obesity are scarce and have
shown differences according to the source of the adipose
tissue. In this regard, lower AQP7 mRNA levels have been
described in subcutaneous adipose tissue (SAT) samples
from subjects with severe obesity than those obtained in lean
subjects [6-8]. By contrast, higher AQP7 mRNA expression
has been found in visceral adipose tissue (VAT) from
massively obese subjects . AQP7 mRNA levels in type 2
diabetes mellitus (T2D) subjects were similar to those in
nondiabetic controls in both adipose depots [6,9]. With
regard to AQP9 gene expression in the liver, it appears to be
down-regulated by insulin, potentially via an “insulin-
responsive element” in the AQP9 promoter. In this sense,
there is only 1 study showing a down-regulation of this gene
in liver biopsies obtained from obese T2D patients .
The close regulatory mechanism depicted by these 2
aquaporins and their role in glycerol and glucose metabolism
would suggest a need to increase our knowledge of its
behavior in human obesity. To shed light on this issue, we
have studied paired AQP7 mRNA expression in adipose
(SAT and VAT) tissue and AQP9 mRNA hepatic expression
in biopsies from a cohort of morbidly obese patients. In
addition, the relationship between AQP7 and AQP9 mRNA
expression and the presence of glucose abnormalities were
2. Methods and materials
We recruited 30 consecutive morbidly obese subjects of
Caucasian origin who underwent gastric bypass surgery at
the University Hospital Vall d'Hebron (Barcelona, Spain).
All patients met the eligibility criteria established by the
guidelines of the National Institutes of Health Consensus
Conference . The preoperative evaluations included
assessment by an endocrinologist, a psychiatrist, and a
pneumonologist to identify and treat all comorbid medical
conditions before operation. None of the subjects presented
evidence of metabolic disease other than obesity, diabetes,
and dyslipidemia. No T2D patients were receiving glitazone
treatment. All hypolipidemic and oral hypoglycemic agents
were stopped at least 72 hours before the surgical procedure.
Before the surgical procedure, a 75-g oral glucose
tolerance test was performed on those patients in whom
diabetes was not previously diagnosed; and patients were
classified according to American Diabetes Association
The ethics committee approved the study, and informed
consent was obtained from all enrolled patients.
2.1. Anthropometric measurements
Height was measured to the nearest 0.5 cm and body
weight to the nearest 0.1 kg. Body mass index was calculated
as weight (kilograms) divided by height (meters) squared.
Waist circumference was measured midway between the
lowest rib margin and the iliac crest.
2.2. Histologic studies of liver
Hematoxylin-eosin and trichrome stains of all liver
biopsies were reviewed by a pathologist without knowledge
of the clinical data and were then classified according to the
criteria of Brunt . The following parameters were graded
in the biopsies: (a) steatosis, 0 to 3; (b) hepatocyte
balloonization, 0 to 3; (c) lobular inflammation, 0 to 3; and
(d) portal inflammation, with or without different fibrosis
stages, 0 to 4.
2.3. Collection and processing of samples
All patients had fasted overnight, at least 12 hours before
undergoing the surgical procedure. Two experienced sur-
geons in abdominal surgery performed all the laparoscopic
Roux-en-Y gastric bypass procedures. Blood samples were
collected before the surgical procedure from the antecubital
vein, 20 mL of blood with EDTA (1 mg/mL) and 10 mL of
blood in silicone tubes. Fifteen milliliters of collected blood
was used for the separation of plasma. Plasma and serum
samples were stored at −80°C until analytical measurements
were performed. Five milliliters of blood with EDTA was
used for the determination of glycated hemoglobin (HbA1c).
During the surgical procedure, adipose tissue samples
from SAT and VAT were obtained, as well as a liver
biopsy from the same patient included in the study.
Adipose tissue samples were washed in phosphate-
buffered saline 1×, immediately frozen in liquid nitrogen,
and stored at −80° C. The liver biopsies were collected in
an RNA preservative solution (RNAlater; Sigma-Aldrich,
1763 M. Miranda et al. / Metabolism Clinical and Experimental 58 (2009) 1762–1768
St Louis, MO). After RNA later solution was removed, the
samples were immediately frozen in liquid nitrogen and
were stored at −80° C.
2.4. Analytical methods
Glucose, cholesterol, and TG plasma levels were
determined in a Hitachi 737 autoanalyzer (Boehringer
Mannheim, Marburg, Germany) using the standard enzyme
methods. High- and low-density lipoprotein (LDL) choles-
terol was quantified after precipitation with polyethylene
glycol at room temperature (PEG-6000). Plasma insulin was
determined by radioimmunoassay (Coat-A-Count Insulin;
Diagnostic Products, Los Angeles, CA) in all subjects of the
study, except in T2D patients treated with insulin. The
HbA1c was measured by chromatography microcolumn
(IsoLab, Akron, OH). Plasma high-sensitivity C-reactive
protein was determined by a highly sensitive immunone-
phelometry kit (Dade Behring, Marburg, Germany). Plasma
glycerol levels were analyzed using a free glycerol
determination kit, a quantitative enzymatic determination
assay (Sigma-Aldrich). Intra- and interassay coefficients of
variation were less than 6% and less than 9.1%, respectively.
Nonesterified free fat acid (NEFA) serum levels were
determined in an Advia 1200 autoanalyzer (Siemens,
Munich, Germany) using an enzymatic method developed
by Wako Chemicals (Neuss, Germany). Assay sensitivity
was 0.01 mEq/L, and the inter- and intraassay coefficients of
variation were lower than 8%.
The homeostasis model assessment of insulin resistance
was calculated as [glucose (milligrams per deciliter) ×
insulin (micro–international units per liter)]/405 .
2.5. Gene expression relative quantification
Four hundred to 500 mg of frozen adipose tissue and
200 mg of frozen liver tissue were homogenized with an
Ultra-Turrax 8 (Ika, Staufen, Germany). Total RNA was
extracted using an RNeasy Lipid Tissue Midi Kit (QIAGEN
Science, Hilden, Germany) for adipose tissues and an
RNeasy Midi Kit (QIAGEN) for hepatic biopsies, duly
following the manufacturer's instructions. Total RNA was
treated with 55 U RNase-free DNase (QIAGEN) before
column elution to avoid contamination with genomic DNA.
A total of 1.5 μg of RNA was reverse transcribed to
complementary DNA (cDNA) using a High-Capacity cDNA
Reverse Transcription Kit (Applied Biosystems, Foster City,
CA) in a final volume of 20 μL.
Real-time quantitative polymerase chain reaction was
performed with 1 μL of cDNA on a 7900HT Fast Real-
Time PCR System using Taqman Assays (Applied
Biosystems). Three replicate reactions per sample and
gene were performed. SDS software 2.3 and RQ Manager
1.2 were used to analyze the results with the comparative Ct
method (2−ΔΔCt). Data were expressed as an n-fold
difference relative to the calibrator (a mix of 3 tissues).
The Ctvalues for each sample were normalized with the
geometric mean of 2 endogenous controls: PPIA (cyclo-
philin A) and ACTB (β-actin).
2.5.1. Statistical analysis
Statistical analysis was performed using the SPSS/PC+
statistical package (version 15 for Windows; SPSS, Chicago,
IL). For clinical and anthropometric variables, normally
distributed data are expressed as mean value ± SD; and for
variables with no Gaussian distribution, values are expressed
as median (75th percentile). For statistical analysis of
expression variables that did not have a Gaussian distribu-
tion, values were logarithmically or inversely transformed.
Differences in clinical/laboratory parameters or expres-
sion variables between groups were compared by using
analysis of variance with a post hoc Scheffe correction.
Interactions among factors as well as the effects of covariates
and covariate interactions with factors were assessed by
general linear model univariate analysis. Associations
between quantitative variables were evaluated by Pearson
or Spearman (for non-Gaussian–distributed variables) cor-
relation analysis. Correction for confounding and interacting
variables was performed using a stepwise multiple linear
regression analysis. Results are expressed as multiple
correlation coefficient (R). Statistical significance occurred
if a computed 2-tailed probability value was b .050.
3.1. Aquaporin expression levels according to glucose
Clinical and laboratory variables of the study participants
are summarized in Table 1. The T2D obese patients were
significantly older than normoglycemic (NG) subjects
(50.14 ± 6.01 vs 41.08 ± 5.90 years, P = .018); and
therefore, this was taken into account in the statistical
analysis. Impaired glucose tolerance (IGT) and T2D subjects
showed significantly increased fasting glucose levels
compared with NG subjects. Low total and LDL cholesterol
levels in the T2D patients were mainly due to the high
percentage (43%) of preoperative statin treatments. Circulat-
ing serum glycerol and NEFA levels were not significantly
different between the studied groups (Table 1).
We did not find any statistical difference in mRNA
expression of AQP7 in adipose tissue (SAT and VAT) or
AQP9 in the liver according to glucose tolerance classifica-
tion, despite controlling for age (Table 1). Subjects with IGT
or T2D showed a strong positive correlation between
subcutaneous adipose AQP7 and AQP9 mRNA levels (r =
0.602, P = .011) (Fig. 1).
3.1.1. Aquaporin expression in the entire obese cohort
When the whole sample was considered (n = 30), AQP7
expression levels were compared in both adipose depots and
analyzed by sex using a univariate general linear model. The
results showed that AQP7 levels differed between adipose
depots (P = .009, with a partial η2of 11.5%, which reports
1764 M. Miranda et al. / Metabolism Clinical and Experimental 58 (2009) 1762–1768
the proportion of total variability attributable to depot origin)
(Fig. 2) but not with sex. No differences attributable to sex
were found for AQP9 expression levels (P = .158).
The SAT and VAT AQP7 adipose depots were positively
correlated (r = 0.449, P = .013) (Fig. 3). No other clinical or
analytical variables were found to be associated with AQP7
adipose expression. Subcutaneous adipose tissue AQP7 and
hepatic AQP9 showed a positive correlation (r = 0.459, P =
.012). Hepatic AQP9 expression also showed a negative
correlation with plasma TG levels (r = −0.399, P = .036).
To test the strength of these associations, we constructed a
linear regression model for each aquaporin to analyze the
Fig. 1. Correlation between SAT AQP7 and AQP9 mRNA levels in IGTand
T2D subjects (arbitrary units) (r = 0.602, P = .011).
Clinical, anthropometric, and analytical characteristics (units) and relative mRNA levels (arbitrary units) according to glucose tolerance classification
NG (n = 12) IGT (n = 11)T2D (n = 7)
Anthropometric and analytical characteristics
Waist circumference (cm)
Antihypertensive treatment (%)
Hypolipidemic agents (%)
Oral hypoglycemic agents (%)
Fasting glucose (mg/dL)
HDL cholesterol (mg/dL)
LDL cholesterol (mg/dL)
Relative mRNA levels (arbitrary units)
41.1 ± 5.9
43.3 ± 5.1
15,8 ± 7,4
3,48 ± 1,55
5.25 ± 0.34
43.0 ± 12.8
117.0 ± 25.8
131.0 ± 51.0
0.64 ± 0.18
48.2 ± 8.5
44.5 ± 5.6
20,6 ± 11,5
5,41 ± 3,10
5.76 ± 0.44
49.5 ± 7.3
134.3 ± 43.7
109.8 ± 25.2
0.63 ± 0.14
50.1 ± 6.0⁎
43.2 ± 4.1
11,0 ± 7,5
4,25 ± 2,53
7.72 ± 2.28†‡
39.1 ± 7.1
69.5 ± 14.1⁎§
120.5 ± 24.5
0.79 ± 0.27
1.02 ± 0.45
1.33 ± 0.75
1.13 ± 0.27
1.36 ± 0.61
0.83 ± 0.35
1.30 ± 0.51
BMI indicates body mass index; HDL, high-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; HOMA-IR, homeostasis model assessment of
Differences vs NG:⁎P b.05;†P b.01.
Differences vs IGT:‡P b.05;§P b.01.
Fig. 2. Levels of expression of VAT AQP7 relative to SAT AQP7 (n = 30,
mean ± SD;⁎P = .009).
1765 M. Miranda et al. / Metabolism Clinical and Experimental 58 (2009) 1762–1768
main effects, as well as an interaction term for sex, age, and
hypolipidemic or hypoglycemic treatment effects, for each
aquaporin. Subcutaneous adipose tissue AQP7 and VAT
AQP7 were positively related with SAT AQP7 as dependent
variable (R = 0.508, P = .019) and with VAT AQP7 as
dependent variable (R = 0.627, P = .004). Hepatic AQP9
expression remained negatively associated with plasma TG
concentration and with SAT AQP7 expression (R = 0.623,
P = .008 and P = .010, respectively).
We did not find any relationship between the hepatic
mRNA levels of AQP9 and the degree of hepatic steatosis
The present study provides, for the first time, evidence of
a coordinated regulation between subcutaneous adipose
AQP7 and hepatic AQP9 gene expression within the context
of human morbid obesity. Moreover, SAT AQP7 and hepatic
AQP9 mRNA levels are positively correlated. In addition, a
negative association between liver AQP9 mRNA expression
and circulating TG levels is described independently of
glucose metabolic status.
It has been suggested that adipose tissue has 2
functionally different preadipocytes. These 2 “pools”
(VAT and SAT) of adipose tissue depots receive differential
sympathetic innervations and have different enlargement
rates with a characteristic release of hormones and
metabolites specific for each tissue compartment .
Adipose tissue distribution depends on many factors, and
sex is one of the main contributors for explaining body fat
distribution. Thus, it is well known that estrogens increase
the size and number of subcutaneous adipocytes and
attenuate lipolysis . In a previous report in which
AQP7 expression was analyzed in both SAT and VAT in a
small cohort of morbidly obese subjects, sex dimorphism
was suggested, showing increased expression levels in
women . However, in our study, we failed to find
differences attributable to sex, in accordance with our
previous observation in nonsevere obese subjects .
Likewise, visceral fat showed greater AQP7 levels than
SAT, confirming a previous observation in isolated visceral
adipocytes in which AQP7 was overexpressed in compar-
ison with subcutaneous adipocytes . Increased lipolytic
activity of visceral adipocytes may be partly responsible for
these differences. In comparison with subcutaneous fat,
VAT is more sensitive to catecholamine-induced lipolysis
and less sensitive to the antilipolytic effects of insulin. In
fact, increased expression of lipoprotein lipase and
hormone-sensitive lipase in visceral vs subcutaneous fat
has been reported [16,17].
We have found no differences in compartmental AQP7
adipose tissue gene expression according to glucose
metabolic status. It may be worth considering that, in
morbid obesity, the presence of insulin resistance does not
influence the greatest expression of AQP7 observed in
VAT. In fact, this observation has been made for other
lipolytic genes, like adipose triglyceride lipase, which
catalyzes the initial step in TG hydrolysis and has a close
regulation with insulin. This enzyme is not differentially
expressed between visceral and subcutaneous fat, despite
being related to insulin sensitivity independently of body
fat mass and fat distribution .
Similar assumptions may be deduced from the analysis
of hepatic AQP9 expression in our morbidly obese cohort.
Glucose tolerance distribution did not influence AQP9
liver expression. However, the design of our study does
not permit to discard an effect of insulin on human AQP9
mRNA liver expression. In fact, most of our patients
showed a notable degree of insulin resistance before
surgery, which could contribute to explaining the absence
of differences in AQP9 gene expression observed after
glucose tolerance classification. Furthermore, a close
correlation between subcutaneous AQP7 and hepatic
AQP9 expression was observed mainly in intolerant and
T2D patients. AQP7 and AQP9 mRNA expression has
been shown to be coordinately regulated by plasma
concentrations of insulin in rodents, in accordance with
nutritional condition, such as fasting and refeeding. These
facts indicate the possible involvement of aquaglyceropo-
rins in pathophysiologic glucose metabolism . Our work
supports these observations, showing a robust correlation
between SAT AQP7 and AQP9 liver expression mainly in
patients with an altered glucose metabolism. It remains to
be determined whether AQP7 and AQP9 protein levels
have similar profiles to mRNA because the present study
was conducted only by examining mRNA levels.
Visceral adipose tissue has drainage over the portal
venous system with a direct flow to the liver, and it is
reasonable to assume that the main dependence for AQP9
Fig. 3. Correlation between SAT and VAT AQP7 mRNA levels (arbitrary
units) (r = 0.449, P = .013).
1766 M. Miranda et al. / Metabolism Clinical and Experimental 58 (2009) 1762–1768
expression comes from visceral fat. We have no satisfactory
explanation for this preferential association between SAT
and hepatic aquaglyceroporins; however, the strong rela-
tionship between SAT and VAT aquaporins (26% of SAT
AQP7 variability was explained by VAT AQP7 levels)
may be partly responsible for this association. We know
that the transversal design of our study does not permit to
infer mechanistic conclusions, and we are conscious that
many other variables could be related with fat and hepatic
aquaporin coordination. Therefore, one is tempted to
speculate as to an alternative glycerol secretion channel
in adipocytes, as has been suggested by many authors
[18-20]. To date, 4 aquaglyceroporins have been described
(AQP3, 7, 9, and 10), with glycerol transport ability, close
to diffusion rates and with a low conductance to water.
Although the expression of AQP3 in mice liver  and
AQP9 in pig adipose tissue  has been described, to
our knowledge, in humans no other aquaglyceroporin in
adipose tissue and liver than AQP7 and AQP9 has been
described. Moreover, we have not found AQP10 expression
in adipose tissue (data not shown).
Measurement of circulating glycerol levels underesti-
mated the glycerol released into the portal vein by adipose
tissue before its clearance by the liver . This could help
interpret the absence of correlation observed between
aquaglyceroporin expression in both adipose and hepatic
tissues, and circulating glycerol levels, although the
negative association between plasma TG levels and AQP9
expression reveals an indirect role of AQP9 in glycerol
metabolism, in agreement with the metabolic findings
described in AQP9-knockout mice . Besides, FFA
release into circulation by adipocyte lipolysis is also
undervalued because of the reesterification of the FFA in
adipose tissue. Thus, in humans, the recycling in this tissue
has been estimated to be as high as 40% . Finally, it is
important to note that we did not find any relationship
between AQP9 expression in the liver and the degree of
hepatic steatosis or fibrosis. Therefore, it seems that AQP9
expression is not essential in regulating fatty liver deposits.
In agreement with these findings, Rojek et al  found no
apparent histologic abnormalities in the liver between
control mice and AQP9-knockout mice.
One limitation to our study is the fact that protein levels
were not measured, which would show whether they are
correlated with gene expression.
In summary, we have found that glucose tolerance status
does not seem to influence adipose AQP7 and hepatic
AQP9 expression in morbidly obese patients. However, a
close relationship between SAT AQP7 and hepatic AQP9
was found mainly in subjects with glucose metabolic
abnormalities. Adipose aquaglyceroporins appeared to be
regulated coordinately with a greater expression found in
visceral fat in our morbidly obese cohort. More mechanistic
studies are needed to further explore the implication and
response of AQP7 and AQP9 in physiologic and pathologic
This work was supported by FIS 07/1024, FIS08/1195
awarded by the Spanish Instituto de Salud Carlos III,
Ministerio de Sanidad y Consumo. CIBER de Diabetes y
Enfermedades Metabólicas Asociadas is an ISCIII project.
The authors gratefully acknowledge the invaluable colla-
boration of the Servei de Recursos Científics i Tecnics of
Universidad Rovira i Virgili and Elsa Maymo for technical
assistance. Dr Matilde R Chacon is supported by a
fellowship awarded by the Fondo de Investigacion Sanitaria
 Kuriyama H, Shimomura I, Kishida K, Kondo H, Furuyama N,
Nishizawa H, et al. Coordinated regulation of fat-specific and liver-
specific glycerol channels, aquaporin adipose and aquaporin 9.
 Baba H, Zhang XJ, Wolfe RR. Glycerol gluconeogenesis in fasting
humans. Nutrition 1995;11:149-53.
 Wintour EM, Henry BA. Glycerol transport: an additional target for
obesity therapy? Trends Endocrinol Metab 2006;17:77-8.
 Kishida K, Kuriyama H, Funahashi T, Shimomura I, Kihara S, Ouchi
N, et al. Aquaporin adipose, a putative glycerol channel in adipocytes.
J Biol Chem 2000;275:20896-902.
 Hibuse T, Maeda N, Funahashi T, Yamamoto K, Nagasawa A,
Mizunoya W, et al. Aquaporin 7 deficiency is associated with
development of obesity through activation of adipose glycerol kinase.
Proc Natl Acad Sci U S A 2005;102:10993-8.
 Ceperuelo-Mallafré V, Miranda M, Chacón MR, Vilarrasa N, Megia A,
Gutiérrez C, et al. Adipose tissue expression of the glycerol channel
aquaporin-7 gene is altered in severe obesity but not in type 2 diabetes.
J Clin Endocrinol Metab 2007;92:3640-5.
 Marrades MP, Milagro FI, Martínez JA, Moreno-Aliaga MJ.
Differential expression of aquaporin 7 in adipose tissue of lean and
obese high fat consumers. Biochem Biophys Res Commun 2006;339:
 Prudente S, Flex E, Morini E, Turchi F, Capponi D, De Cosmo S, et al.
A functional variant of the adipocyte glycerol channel aquaporin 7
gene is associated with obesity and related metabolic abnormalities.
 Catalán V, Gómez-Ambrosi J, Pastor C, Rotellar F, Silva C, Rodríguez
A, et al. Influence of morbid obesity and insulin resistance on gene
expression levels of AQP7 in visceral adipose tissue and AQP9 in
liver. Obes Surg 2008;18:695-701.
 Gastrointestinal surgery for severe obesity: National Institutes of
Health Consensus Development Conference Statement. Am J Clin
 American Diabetes Association. Standards of medical care in diabetes.
Diabetes Care 2007;30:S4-S41.
 Brunt EM. Nonalcoholic steatohepatitis definition and pathology.
Semin Liver Dis 2001;21:3-16.
 Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF,
Turner RC. Homeostasis model assessment; insulin resistance and
β-cell function from fasting plasma glucose and insulin concentrations
in man. Diabetologia 1985;28:412-9.
 Mattsson C, Olsson T. Estrogens and glucocorticoid hormones in
adipose tissue metabolism. Curr Med Chem 2007;14:2918-24.
 Sjöholm K, Palming J, Olofsson LE, Gummesson A, Svensson PA,
Lystig TC, et al. A microarray search for genes predominantly
expressed in human omental adipocytes: adipose tissue as a major
production site of serum amyloid A. J Clin Endocrinol Metab 2005;90:
1767M. Miranda et al. / Metabolism Clinical and Experimental 58 (2009) 1762–1768
 Berndt J, Kralisch S, Klöting N, Ruschke K, Kern M, Fasshauer M, Download full-text
et al. Adipose triglyceride lipase gene expression in human visceral
obesity. Exp Clin Endocrinol Diabetes 2008;116:203-10.
 Wajchenberg BL. Subcutaneous and visceral adipose tissue: their
relationto the metabolicsyndrome.
 Kondo H, Shimomura I, Kishida K, Kuriyama H, Makino Y,
Nishizawa H,et al. Humanaquaporinadipose(AQPap) gene.Genomic
structure, promoter analysis and functional mutation. Eur J Biochem
 Large V, Peroni O, Letexier D, Ray H, Beylot M. Metabolism of
lipids in human white adipocyte. Diabetes Metab 2004;30:
 Maeda N, Funahashi T, Hibuse T, Nagasawa A, Kishida K, Kuriyama
H, et al. Adaptation to fasting by glycerol transport through
aquaporin 7 in adipose tissue. Proc Natl Acad Sci U S A 2004;
 Patsouris D, Mandard S, Voshol PJ, Escher P, Tan NS, Havekes LM,
et al. PPARalpha governs glycerol metabolism. J Clin Invest 2004;
 Li X, Lei T, Xia T, Chen X, Feng S, Chen H, et al. Molecular
characterization, chromosomal and expression patterns of three
aquaglyceroporins (AQP3, 7, 9) from pig. Comp Biochem Physiol B
Biochem Mol Biol 2008;149:468-76.
 Coppack SW, Chinkes DL, Miles JM, Patterson BW, Klein S. A
multicompartmental model of in vivo adipose tissue glycerol kinetics
and capillary permeability in lean and obese humans. Diabetes 2005;
 Rojek AM, Skowronski MT, Füchtbauer EM, Füchtbauer AC, Fenton
RA, Agre P, et al. Defective glycerol metabolism in aquaporin 9
(AQP9) knockout mice. Proc Natl Acad Sci U S A 2007;104:3609-14.
 Reshef L, Olswang Y, Cassuto H, Blum B, Croniger CM, Kalhan SC,
et al. Glyceroneogenesis and the triglyceride/fatty acid cycle. J Biol
1768 M. Miranda et al. / Metabolism Clinical and Experimental 58 (2009) 1762–1768