University of Groningen

An adipocentric view of the development of insulin resistanceSzalowska, Ewa

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite fromit. Please check the document version below.

Document VersionPublisher's PDF, also known as Version of record

Publication date:2011

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):Szalowska, E. (2011). An adipocentric view of the development of insulin resistance. Groningen: s.n.

CopyrightOther than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of theauthor(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons thenumber of authors shown on this cover page is limited to 10 maximum.

Download date: 07-06-2020

Introduction

1

An adipocentric view of the development of insulin resistance

Ewa Szalowska

Introduction

2

ISBN: 978-90-367-4849-0 Printed by Wӧhrmann Print Service Cover design: Ewa Szalowska Layout: Sacha van Hijum ©2011 Copyright Ewa Szalowska All rights reserved. No parts of this book may be reproduced, stored in retrieval systems, or transmitted in any form or by any means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the author. Printing of this thesis was supported by UMCG, Rijksuniversiteit Groningen

Introduction

3

RIJKSUNIVERSITEIT GRONINGEN

An adipocentric view of the development of insulin resistance

Proefschrift

ter verkrijging van het doctoraat in de

Medische Wetenschappen aan de Rijksuniversiteit Groningen

op gezag van de Rector Magnificus, dr. E. Sterken, in het openbaar te verdedigen op

woensdag 13 april 2011 om 13.15 uur

door

Ewa Szalowska

geboren op 14 mei 1975

te Wrocław, Polen

Introduction

4

Promotor: Prof. dr. R.J.Vonk Copromotores: Dr. H. Roelofsen

Dr. A. Hoek Beoorelingscommissie: Prof. dr. G.M.M. Groothuis

Prof. dr. B.H. Wolffenbuttel Prof. dr. M. Haluzik

Introduction

5

Introduction

6

Paranimfen: Marianne Schepers Kees Meijer Front cover: Histological staining of adipose tissue

Introduction

7

Contents Introduction 9 Scope of the thesis 20 Results and discussion 23 Concluding remarks and future perspectives 27 Summary 31 Samenvatting 34 References 38 Section I Optimization and development of proteomics technologies 42 Chapter 1 Fractional factorial design for optimisation of the SELDI protocol for human adipose tissue culture media 43 Chapter 2 Characterization of the human visceral adipose tissue secretome 63 Section II Role of adipose tissue in the development of insulin resistance 88 Chapter 3 Sub-chronic administration of stable GIP analogue in mice decreases serum LPL activity and body weight 89 Chapter 4 Adipokines and energy metabolism genes, but not proinflammatory genes are deregulated in patients with higher HOMA and lower HDL 107

Chapter 5 The “adipokine” resistin is more abundant in human liver than in adipose tissue and it is not upregulated by lipopolysaccharide 123 Chapter 6 Comparative analysis of the human hepatic and adipose tissue transcriptomes during LPS-induced inflammation leads to the identification of differential biological pathways and candidate biomarkers 145 Abbreviations 171 Acknowledgements 173 Curriculum Vitae 177

Introduction

8

Introduction

9

Introduction

Introduction

10

1. Insulin resistance

In an era of increasing obesity, which strongly predisposes to the development of insulin

resistance (IR) and related pathologies such as type 2 diabetes (T2D) and cardiovascular

diseases, adipose tissue became an object of multiple studies devoted to elucidate its role in

regulation of whole body energy metabolism[1].

IR is regarded as a state where peripheral organs involved in the regulation of energy

homeostasis (adipose tissue, liver, and muscles) are not able to maintain physiological blood

glucose concentration with normal insulin level. The increased demand for insulin can be

compensated with enhanced insulin production by pancreatic β -cells, however this

mechanism fails on a long –term basis due to β-cell dysfunction, and ultimately leads to T2D

[2] .

The gold standard for assessment and quantification of IR is the "hyperinsulinemic

euglycemic clamp," which measures the amount of glucose necessary to compensate for an

increased insulin level without causing hypoglycemia [3]. Due to the laborious character of

this test it is difficult to perform in clinical practice therefore are other alternative methods

used to asssess insulin sensitivity. One of them is the Homeostatic Model Assessment

(HOMA). This method measures fasting glucose and insulin levels to calculate IR. The

HOMA values correlate well with the results of the clamping studies [4] and can be used as

its surrogate.

As mentioned above, one of the factors predisposing to the development of IR is obesity.

However, it was reported that abdominal obesity independently of total obesity is strongly

associated with IR. Next to factors related to (visceral) obesity such as increased waist

circumference (WC), and high body mass index (BMI) , IR patients are characterized by

increased levels of free fatty acids (FFA) leading to ectopic fat depots, decreased high-density

lioprotein (HDL)-cholesterol, and increased low-density lipoprotein (LDL)-cholesterol.

Nowadays, it is also commonly accepted that low grade systemic inflammation with inreased

serum levels of CRP, IL-6, TNFα, and IL-1β closely predisposes to IR. However, it is not

fully elucidated if in humans inflammation is a cause or consequence of IR [5; 6].

Introduction

11

2. Adipose tissue as an endocrine organ and its link to insulin resistance

In a traditional view, adipose tissue was regarded as a tissue consisting of inert adipocytes

involved in storage of energy in form of triglycerides (TG). However, since the discovery of

leptin acting as a hormone via its receptors localized in peripheral tissues, adipose tissue

became recognized as an endocrine organ. Leptin acts via leptin receptors which have been

identified in the brain areas involved in feeding regulation, and several peripheral tissues

including adipose tissue, ovaries, testis, placenta, adrenal medulla, liver, pancreatic β- cells,

lung, jejunum, peripheral blood mononuclear cells, chondrocytes, heart, and skeletal muscle.

The known leptin functions are related to energy metabolism, reproduction, and

inflammation. For example in isolated skeletal muscles, liver and pancreas acute leptin

treatment increased fatty acid oxidation in an AMP-activated protein kinase (AMPK)

dependent manner and suppressed the hypothalamus signaling. Moreover, leptin modulates

hepatic gluconeogenesis and pancreatic β-cell function, and promotes energy expenditure.

Leptin is also known to modulate inflammatory functions by stimulating TNFα and IL-6

expression, but suppressing resistin and retinol binding protein 4 expression [7]. In addition

to leptin, the endocrine functions of adipose tissue are exerted by other adipose tissue derived

signaling molecules (adipokines) acting in different target organs via their receptors. For

example adiponectin is highly expressed in adipose tissue and it is an abundant plasma

protein [7]. Adiponectin binds to its own receptors: AdipoR1 and AdipoR2 which are

ubiquitously expressed [8]. The role of adiponectin in obesity, IR, and T2D is broadly

investigated and it was demonstrated that adiponectin displays insulin sensitizing properties

and can ameliorate systemic insulin resistance symptoms. In skeletal muscle and adipose

tissue adiponectin stimulates fatty acid oxidation and glucose uptake. In liver adiponectin

suppresses glucose production, and in the central nervous system it was shown to regulate

energy expenditure through activation of AMPK in the hypothalamus [7]. Additionally,

adiponectin exhibits strong anti-inflammatory, anti-diabetic, and anti-atherogenic actions [9].

Next to the established adipocyte-derived adipokines such as leptin and adionectin, there are

adipokines produced mostly by stromal vascular fraction of adipose tissue such as IL-6, IL-

1β, TNFα, resistin, visfatin, or retinol binding protein 4 [7]. The number of adipokines is still

growing with the recently discovered adipokines such as chemerin, omentin or cartonectin

[10; 11]. The known functions of selected adipokines are summarized in Table 1.

The adipose tissue constellation (complex tissue consisting of adipocytes, in a close

proximity of immune cells (macrophages, lymphocytes ) surrounded by a network of blood

Introduction

12

vessels, with endothelial cells [12]) implies on tight interactions between metabolic,

endocrine, and immune functions within the organ. Thereby it seems to be logical why

metabolic perturbations observed in adipose tissue of obese and IR patients are very often

associated with inflammation-like symptoms manifested by elevated production of pro-

inflammatory adipokines such as IL-6, TNFα or CRP. A comprehensive understanding of

endocrine actions of adipose tissue and exact feedback mechanisms regulating adipokines

expression by their target organs is just an emerging issue. At present, adipose tissue

endocrine functions are regarded as interplay between environment, genetic factors and

multi-organ interactions, schematically depicted in Figure 1. In future the Systems Biology

approach is expected to bring more insights into the dynamic and complex mode of action of

adipose tissue [13; 14].

Introduction

13

Table 1. Examples of adipokines that display dual inflammatory and immunological functions [10; 11].

Adipokines

with dual

function Metabolic function Immune function

Adiponectin insulin sensitizing, anti-diabetic anti-inflammatory

Leptin

regulation of satiety, energy

expenditure, pancreatic functions, and

intracellular lipid content immune-modulating

Resistin insulin-desensitizing

pro-inflammatory?, depending on the

cell types

TNFα insulin-desensitizing pro-inflammatory

IL-6, IL-8, IL-1β insulin-desensitizing pro-inflammatory

Visfatin

interfering with insulin-receptor

signaling pro-inflammatory

C3A TG synthesis, insulin sensitivity

adipose tissue macrophage infiltration

and cytokine production

Cartonectin regulation of adipokine secretion anti-inflammatory, acting on monocytes

Chemerin

Insulin sensitizing, enhances insulin

dependent glucose uptake andIRS-1

phosphorylation in humans

Anti-inflammatory effects on activated

macrophages expressing the chemerin

receptor

Omentin

Enhances insulin –stimulated glucose

transport and Akt phosphorylation-

insulin sensitizing in humans Not known

Retinol binding

protein 4

Insulin desensitizing in humans,

lifestyle modification and exercise

training reduces serum RBP4 levels Not known

Vaspin Insulin sensitizing Suppresses TNFα and resistin expression

Introduction

14

Figure 1. Adipose tissue endocrine functions as interplay between environment, life style, genes, and multi-organ interactions. Green arrows illustrate stimulation and red arrows suppression of gene expression, adopted from Rabe K. et al [10].

During the development of obesity and IR the homeostasis in adipose tissue can be

disturbed, what results in a changed adipokine profile with decreased adiponectin levels and

increased levels of leptin, several cytokines, visfatin and others [15-18]. Due to the altered

adipose tissue endocrine profile the released adipokines affect functions of the target organs,

what results in their altered metabolic functioning and these events synergistically can lead to

disturbance of metabolic and immune homeostasis resulting in systemic IR [19]. There is

evidence for implication of several mechanisms involved in the development of systemic IR

such as (a) excess of free fatty acids released from adipose tissue activating Toll like

receptors in other tissues and causing systemic inflammation [20], (b) GIP signaling

overstimulation in obesity [21] (c) oxidative stress and mitochondrial dysfunction [22; 23],

(d) endotoxemia and inflammation [24-27], (e) alternations in T-cells populations in adipose

tissue and infiltration of macrophages [28-30], however still the exact triggers leading to the

development of systemic IR remain unknown.

Introduction

15

3. Mechanisms leading to the development of insulin resistance

As indicate above several mechanisms are proposed to be involved in the development of IR;

in this thesis we will focus on the role of adipose tissue in the development of IR related to

glucose dependent insulinotropic polypeptide (GIP) signaling and inflammation.

A) GIP as a link between overnutrition, obesity and T2D

Glucose dependent insulinotropic polypeptide (GIP), also known as gastric inhibitory

polypeptide, is a gastrointestinal hormone secreted from gastrointestinal track upon fat

ingestion [31] and its serum level positively correlates with the intestinal glucose influx rate,

suggesting that GIP secretion is also influenced by glucose absorption [32].

GIP it is one of gastrointestinal hormones which acts on adipose tissue and pancreatic beta

cells thereby it is suggested as a link with obesity and T2D [31; 33]. Besides acting on the

beta cells and adipose tissue GIP exerts its functions on bone and central nervous system

[31]. Mainly due to the GIP’s beneficial actions on the beta cells this hormone gained a lot of

attention as an anti-diabetic factor. It was shown that GIP stimulates glucose induced insulin

secretion, enhances insulin gene transcription and biosynthesis, induces beta-cell neogenesis,

proliferation, and differentiation [31].

The GIP receptor (GIPR) knockout mice were resistant to obesity on a high fat diet and did

not accumulate fat in liver or muscles instead [21]. Besides the spectacular effects in the

GIPR knockout mice on preventing the development of obesity on a high fat diet there are

several other studies exploring this phenomenon in chemical/pharmacological GIPR

knockout mice [34-36]. For example chronic administration of the GIPR antagonist (Pro3)

GIP to adult diabetic ob/ob mice fed high-fat-diet resulted in substantial improvement in

metabolic status [34].

These observations might suggest that opposite actions, such as GIP oversignaling, could

contribute or accelerate the development of obesity and eventually IR and T2D. Indeed, GIP

was shown in vitro to stimulate lipoprotein lipase (LPL) activity, responsible for

accumulation of triglycerides (TG) in adipose tissue, to increase lipogenesis, fatty acids and

glucose uptake, to enhance insulin –induced fatty acid incorporation leading to hypertrophy

of adipocytes and possibly promoting obesity and eventually T2D in vivo [31]. However,

there is still not enough evidence that overstimulation of GIP signaling or elevated levels of

GIP in obesity function as a link between overnutrition and the development of obesity and

type T2D in vivo in humans [37; 38] . In contrast, there is an emerging view from clinical

Introduction

16

trials that application of GIP agonists (overstimulation of GIP signaling) led to weight loss,

blood pressure reduction, and, as expected, beta-cell function improvements [39].

B) Inflammation in adipose tissue-is it a cause (or consequence) of insulin resistance

Adipose tissue of obese animals and humans can display low grade inflammation with

upregulated proinflammatory proteins in both adipose tissue and blood such as TNFα, IL-1β,

IL-6, IL-8, PAI-1, SAA, CRP, MCP-1, [5; 5; 15; 25; 40; 41] and others. Addressing the

question if the whole body (systemic) IR is a cause or a consequence of inflammation

initiated in adipose tissue one may refer to studies described by Park et all [42] and Xu et all

[40]. In both studies IR was developed by diet induced obesity (DIO) in C57BL/6 mice.

Study of Park at all showed that IR occurred in heart after 1.5 week of high fat diet and it was

followed after 3 weeks by IR in adipose tissue, liver and skeletal muscles. Since week 1.5 a

significant increase in blood leptin and resistin levels were observed and hyperinsulinemia

occurred after 3 weeks. In studies conducted by Xu et all it was shown that in DIO a

significant increase in proinflammatory markers in adipose tissue but not other organs

occurred after 16 weeks of high fat diet, The comparison of these two studies suggests that

decreased glucose uptake indicative for IR occurs before inflammatory markers are up

regulated in adipose tissue thereby excluding inflammation as a pivotal event leading to IR.

However, in other studies it was shown that endotoxemia, associated with elevated LPS

levels can be a pivotal event leading to IR. It was shown that LPS induced inflammation

leads to IR in rodents and humans [24-27; 43-45]. In humans during short term endotoxemia

insulin sensitivity assessed by frequently sampled intravenous glucose tolerance testing

(FSIGT) was significantly decreased. Furthermore, DIO induced upregulation of multiple

proteins associated with IR , exampled by IL-6, TNFα, MCP-1 and CXCL10 and components

of insulin signaling pathway were affected: insulin receptor substrate-1 was decreased ,while

suppressor of cytokine signaling proteins (1 and 3) were markedly induced [24].

4. Omics technologies to be developed for the current research questions

Proteomics technologies involve experimental approaches that allow studying the total

proteome or specific group of proteins from the proteome such as secretome, focusing on

secreted proteins and peptides or phosphorylated proteins (phosphoproteomics). The

technologies used in proteomics are often based on mass spectrometry (MS). One of them is

Introduction

17

commonly used in biomarkers research Surface Enhanced Laser/Desorption Ionization

(SELDI) technology. In SELDI biological samples are applied to a protein chip with specific

biochemical binding properties on the surface such as hydrophobic, anion, cation, metal.

During the chip preparation, proteins with specific characteristics are bound to the chip

allowing then reduction of sample complexity. After washing the chip is treated with the

energy-adsorbing matrix. The matrix allows flying of the bound proteins upon laser-induced

ionization and time of flight (TOF) separation leads to determination of mass to charge ratio

(m/z) of the proteins. SELDI generates protein profiles which have to be analyzed with

statistical and bioinformatical tools in order to find differential proteins or protein profiles

which could be used as biomarkers. Further downstream analyses are required such as protein

digestion and peptide mass fingerprinting or MS/MS analysis to identify the differential

protein(s) [46].

The commonly used technique in the experimental cell lines research and proteomics is

Stable Isotope Labeling with Amino Acids in Cell Culture (SILAC). SILAC is a

straigforward approach where two cell populations (control and treatment) are grown in cell

culture media that are identical except that one of them contains a “heavy”and the other one a

“light”from of a particular amino acid (e.g. 13C and 12C labeled respectively). The labeled

amino acids are incorporated in 100% after a number of call devisions and after processing of

the cultures it is possible to apply MS based protein identification and to determine the

differences between two experimental conditions. Example of another MS-based technology

is CILAIR (Comparison of Isotope Labeled Amino Acid Incorporation Rates) [47] developed

for application in non-dividing cells and tissue cultures. In CILAIR experiments (adipose)

tissue samples are depleted from lysine and subsequently cultured in the presence of 13C-

labeled lysine. Newly synthesized proteins will incorporate the 13C-labeled lysine allowing to

discriminate between proteins changed in abundance due to the experimental conditions (13C-

labeled lysine) and proteins present a priori, derived form serum or tissue (12C- lysine). After

the culturing, the obtained biological samples (adipose tissue culture media) are concentrated

by ultra-filtration, fractionated by SDS-PAGE followed by in-gel digestion of excised bands.

The resulting peptide digests are analyzed by liquid chromatography coupled with mass

spectrometry (LC-MS/MS). The MS data are subsequently analyzed by software matching

the detected peptide masses into known proteins and performing protein identification. By

comparison of the C13\C12 ratios in identified proteins, information about quantitative

changes between experimental groups can be derived. The advantage of CILAIR in

comparison to a similar technique such as SILAC is that CILAIR does not require 100% label

Introduction

18

incorporation, which is only possible in dividing cells. In the tissue culture systems mostly

non-dividing cells are present and the only requirement for detection of altered proteins due

to the treatment is de novo protein synthesis.

Analysis of proteomics data requires application of complex bioinformatitcal tools. There are

several tools developed for analysis of proteomic data freely or commercially available. For

proteomic analysis the basic tools are for example: Expasy (Expert Protein Analysis System)

(http://www.expasy.org) the proteomics server of the Swiss Institute of Bioinformatics (SIB)

dedicated to the analysis of protein sequences and structures. UniProt

(http://www.uniprot.org) aims to provide the scientific community with a comprehensive,

high-quality and freely accessible resource of protein sequence and functional information

and databases for protein ontology. Other supporting tools such as SecretomeP 2.0 server

produces ab initio predictions of non-classical i.e. not signal peptide triggered protein

secretion. The method queries a large number of other feature prediction servers to obtain

information which is integrated into the final secretion prediction information and can be

used for prediction of secreted and intracellular proteins by means of algorithms

http://www.cbs.dtu.dk/services/SecretomeP/ [29; 48; 48]. STRING (Search Tool for the

Retrieval of Interacting Genes/Proteins) is being developed for reconstruction of functional

protein/gene networks based on known and predicted protein-protein interactions

(http://string.embl.de).

Another commonly used omics technology is the DNA microarray approach, which

facilitates the simultaneous quantification of thousands of mRNAs and provides

comprehensive information about their expression level under different experimental

conditions or in different patients groups.

DNA microarrays contain thousands of nucleotide probes fixed to a solid surface. Each of a

probe represents a short section of a gene or other DNA constituent, which can hybridize to

reverse transcribed mRNA sample or copy RNA (cRNA). These RNA probes are labeled

with a fluorescent dye which can be excited with a laser light. The resulting fluorescent

emission is representative for expression level of a certain mRNA and can be compared

within the analyzed samples. Next challenge in the DNA microarray technology is the

statistical analysis and the biological interpretation [49]. There are several software packages

which are developed for the statistical analysis such as GeneSpring GX developed by Agilent

Technologies, which can be used for many DNA microarray platforms (Affymetrix, Illumina,

Agilent) and can be used for identification of significantly affected genes. The biological

interpretation of the data can be performed by means of freely available bioinformatical

Introduction

19

resources such as DAVID [50; 51], STRING or commercial ones such as MetaCore

developed by GeneGo. These bioinformatical tools enable to perform GO analysis aiming to

extract the significantly affected (overrepresented) biological processes and to identify

affected biological pathways (DAVID, MetaCore), or to build gene functional networks

(MetaCore, STRING), which are built based on the information present in the literature from

all the known and predicted gene/protein interactions.

Omics technologies are broadly used to identify biomarkers for T2D [46; 49; 50; 52]. The

potential biomarkers should be specific to the disease and it should be possible to detect them

early in order to undertake actions preventing development of the disease. At present, there

are multiple candidate biomarkers for insulin resistance exampled by CRP, TNFα, IL-6

however their specificity for early detection of IR and clinical application still has to be

proven [19; 41].

Introduction

20

Scope of the thesis

The discoveries of leptin and other adipokines had greatly contributed to the recognition of

adipose tissue as an endocrine organ involved in regulation of the energy and inflammatory

homeostasis [53]. Nowadays it is generally acknowledged that impairment of adipose tissue

endocrine and metabolic characteristics correlated with obesity can lead to systemic IR and

eventually T2D.

The main aims of this thesis were:

(1) to explore role of adipose tissue in the development of insulin resistance related to:

(A) GIP action,

(B) clinical parameters,

(C) inflammation.

(2) to optimize and develop omics technologies for the specified above research questions

(1A) In order to study the GIP signaling in adipose tissue we applied stable GIP analogue to

mice and tested the hypothesis if excess of GIP can enhance or accelerate the development of

obesity and IR. (1B) In order to find the primary events in adipose tissue implicated in the

development of IR we studied the adipose tissue selected gene expression in association with

clinical parameters indicative for early IR such as high BMI and HOMA and low HDL in

non-diabetic women. (1C) Low grade inflammation associated with obesity and IR was

mimicked by application of LPS to human adipose tissue ex vivo in order to identify specific

candidate biomarkers indicative for inflammation/IR of adipose tissue.

(2) For this specific adipose tissue research several techniques had to be developed with a

special attention to novel proteomics technologies and integration of proteomics and

transcriptomics data.

Therefore the thesis is divided into two sections: (I) Development and optimization of

proteomics techniques; (II) Role of adipose tissue in the development of IR in relation to: (A)

GIP signaling (B) clinical parameters (C) inflammation leading to identification of

candidate biomarkers

Section I is devoted to the optimization of proteomic methodologies such as SELDI and LC-

MS/MS based proteomics technology for the purposes of investigation of the adipose tissue

secretome, and consists of 2 chapters: (1) “Fractional Factorial Design for Optimization of

the SELDI Protocol for Human Adipose Tissue Culture Media” and (2) “Characterization of

The Human Visceral Adipose Tissue Secretome”.

Introduction

21

In chapter 1 we performed optimization of SELDI sample preparation and application for

the SELDI protein chip protocol. One of the major challenges in SELDI technology is to find

out the most optimal sample preparation protocol where factors such as protein concentration,

salt concentration, pH , background reduction, type of applied matrix (energy absorbing

molecule (EAM)), and many others can be optimized. For the protocol optimization the

fractional factorial design was used.

In chapter 2 the characterization of the visceral adipose tissue secretome by means of LC-

MS/MS proteomic technology is described. The article describes optimization of several

adipose tissue culture set-ups aiming to remove highly abundant serum proteins such as

albumin or hemoglobin in order to apply adipose tissue culture media for LC-MS/MS

analysis. After selection of the best culture set up an experiment was performed where

adipose tissue samples were cultured in media containing 13C6,15N2 L-lysine to validate the

origin of the identified proteins (adipose tissue or serum derived). After SDS-PAGE

fractionation, and protein digestion with trypsin, proteins were identified by LC-MS/MS

technology. The characterisation of the adipose tissue secretome was performed in order to

deepen our understanding of the endocrine functions of the adipose tissue.

Section II is devoted to investigations of the role of adipose tissue in the development of IR

related to different mechanisms such as: (A) GIP signaling, (B) clinical parameters, (C)

inflammation. In Section IIA, Chapter 3 entitled:” Sub-Chronic Administration of Stable

GIP Analogue in Mice Decreases Serum LPL Activity and Body Weight” we described

experiments where we treated mice fed both chow or high fat diets with the stable GIP

analogue and followed the GIP effect on the development of obesity assessed by body mass,

serum LPL activity, serum biochemical parameters, and gene expression in adipose tissue.

The aim of these experiments was to evaluate the role of GIP on the development of obesity

and IR and in particular to investigate the role of GIP on the adipose tissue gene expression.

In Section IIB, Chapter 4 entitled “Expression of selected proinflammatory and metabolic

genes in human adipose tissue in relation to IR” we measured gene expression of subsets of

proinflammatory and metabolic genes in adipose tissue of non diabetic women and associated

the gene expression with anthropometric and biochemical characteristics such as body mass

index (BMI), waist circumference (WC), homeostasis model assessment index (HOMA),

glucose, insulin, and high density lipoprotein (HDL) serum levels. In this study we aimed to

find a link between adipose tissue gene expression related to altered anthropometric and

serum biochemical parameters indicative on early IR such as increased HOMA or decreased

Introduction

22

HDL levels in order to understand changes in the adipose tissue gene expression associated

with early events leading to the development of IR in humans.

In Section IIC, Chapter 5 “Resistin is More Abundant in Liver than Adipose Tissues and Is

Not Up-Regulated by Lipopolysachcaride” we evaluated carefully if indeed resistin can be

considered as an adipokine correlated with markers of inflammation in humans, as suggested

in the literature. There is evidence that serum resistin levels are elevated in obese patients,

however the role of resistin in IR and T2D remains controversial. We were interested whether

inflammation induces expression of resistin in organs involved in regulation of total body

energy metabolism, such as liver and adipose tissue. We therefore studied resistin gene

expression, protein abundance and localization in human healthy liver and human visceral

adipose tissue and analyzed the effect of LPS on resistin regulation.

In Section IIC, Chapter 6 “Comparative Analysis of The Hepatic and Adipose Tissue

Transcriptome During LPS-Induced Inflammation Leads to The Identification of Differential

Pathways and Candidate Biomarkers” we induced inflammation with LPS in adipose and

liver tissues to simulate endocrine activity of these organs as observed in vivo during

systemic low grade inflammation associated with IR. Our aim was to compare the inflamed

transcriptomes of both tissues in order to better understand their contribution to the

development of systemic inflammation and IR. Additionally the in silico predicted

inflammatory secretome (generated from the transcriptome data of adipose tissue) was

integrated with quantitative proteomics data (CILAIR).

Introduction

23

Results and Discussion

In Section I (Development and optimization of proteomics techniques) we described

optimization of protocols for adipose tissue culture media applicable for SELDI and LC-

MS/MS.

The SELDI protocol was optimized for the number of peaks and spectra quality by use of the

fractional factorial design. Using this statistical method we tested several factors derived

from the SELDI protocol in a relatively small number of experiments and identified critical

and not important factors. The protocol factors that significantly improved the SELDI spectra

were pretreatment of the sample, type of energy absorbing molecule (EAM), solvent of EAM,

saturation of EAM, and amount of EAM applied to the chip. Whereas, the molecular

concentration of the binding buffer did not influence the spectra quality and quantity (Chapter

1). Moreover, it is important to realize that different protein samples require individual

optimization experiments depending on the protein composition, background of the sample,

protein concentration etc.

The optimized protocol is intended for use in screening for differences in the adipose tissue

culture media derived from lean and obese patients in order to identify biomarkers associated

with early events of the development of IR.

The identification of the adipose tissue secretome by means of LC-MS/MS involved

optimization of experiments leading to removal of high abundant serum proteins (Chapter 2).

This approach resulted in identification of adipose tissue culture set up yielding a sample

(adipose tissue culture media) free of albumin and high abundant proteins. Addition of

labeled Lys (13C) into the culture media allowed distinguishing between proteins synthesized

by adipose tissue during the culture time from contaminating proteins (serum proteins or

proteins released by the tissue during preparation). Thereby, for the first time the human

visceral adipose tissue secretome was determined. In total 259 proteins were identified with

≥99% confidence; 108 proteins contained a secretion signal peptide of which 70 incorporated

the label and were considered secreted by adipose tissue. Within the secretome proteins such

as: adiponectin, adipsin, gelsolin, macrophage colony stimulating factor-1 (M-CSF), pigment

epithelium-derived factor (PEDF), plasma retinol binding protein (RBP), plasminogen

activator inhibitor-1 (PAI-1), and others were identified. Secreted proteins were classified

into functional groups such as: signaling/regulation, extra-cellular matrix, immune function,

degradation and other, which are in line with the present knowledge about the endocrine and

immuno-modulating actions of adipose tissue.

Introduction

24

The knowledge about the secretome of adipose tissue contributes to a better understanding of

the endocrine activities of adipose tissue and could facilitate identification of the specific

adipose tissue biomarkers related to IR.

In Section II (Role of adipose tissue in the development of obesity and IR) we investigated

different aspects related to adipose tissue actions and the development of IR such as (A)

effect of GIP oversignaling, (B) adipose tissue gene expression in respect to clinical

parameters and (C) inflammation. The application of the stable GIP analogue (Section IIA,

Chapter 3) did not accelerate the development of obesity thereby the hypothesis built on

observations in vitro that GIP enhances TG accumulation in adipose tissue and can promote

obesity was not confirmed. Our study shows that the effect observed in GIPR KO mouse can

not be simply reversed by mimicking the opposite conditions, as we tried to achieve here by

applying stable GIP agonist and thereby over- stimulating GIP signaling. However, our study

in mice is in line with results of the clinical trials showing that GIP agonists lead to weight

loss [39]. We observed a decrease in LPL activity in mice injected with the GIP agonist

which can be one of the mechanisms responsible for the body weight loss. Moreover, this

study led to identification of novel candidate genes regulated by GIP in adipose tissue.

Between these novel GIP targets are genes known to be involved in lipid metabolism which

is in line with the known role of GIP in lipid metabolism. No target genes related to

carbohydrate metabolism could be detected, which was hypothesized based on our previous

findings that GIP was highly correlated with the intestinal glucose influx rate. The exact

functions of the novel GIP target genes in adipose tissue have to be elucidated in the future.

In Section IIB in Chapter 4 we aimed to identify a link between changes the adipose tissue

gene expression associated with altered clinical parameters indicative on IR such as high

BMI and WC, increased HOMA or decreased HDL serum level. Moreover, we wanted to

determine differences in the selected gene expression between subcutaneous adipose tissue

(SAT) and omentum. As expected we found that leptin expression in omentum had positive

correlation with BMI and WC, adiponectin expression level in SAT had negative correlation

with HOMA and positive correlation with HDL levels. In respect to the other energy

metabolism genes, we observed that expression of LPL, GLUT4, PPARγ, INSR, SREBP1,

GLUCR, and GIPR was positively associated with HDL and negatively associated with

HOMA, indicating that there is a link between early events in IR and the energy metabolism

gene expression in fat tissue. The pro- inflammatory genes showed positive correlation only

with WC, which is a measure indicative for body fat distribution however the exact

relationship between WC and IR is not well understood [6]. In summary, our study showed

Introduction

25

that in early phases of the development of IR metabolic and not proinflammatory gene

expression is altered in adipose tissue. These findings suggest that energy metabolism

alternations in adipose tissue, but not inflammatory once can be the pivotal events inducing

IR. Moreover, the relationship between adiponectin and metabolic genes expression and HDL

levels suggest that HDL can be involved in regulation of expression of these genes. Indeed, it

was recently shown that adiponectin gene expression level and adipocyte metabolism are

controlled by HDL levels [54]. Furthermore the differential gene expression in SAT and

omentum indicates on different physiological roles of these both fat depots.

In Section IIC in chapters 5 and 6 we investigated inflammatory challenge towards

adipose tissue and aimed to identify candidate biomarkers for inflamed/IR adipose tissue.

Originally, based on the data obtained from studies in rodents, it was proposed that in humans

resistin is an adipokine and can be used as a biomarker indicative for inflammation and IR.

However, our studies did not support these data (Chapter 5). We showed that although

resistin is expressed in human adipose tissue, it is significantly higher abundant in human

liver on both gene and protein levels. In addition, during LPS induced inflammation resistin

gene and protein levels in adipose and liver tissues were not affected while known

proinflammatory cytokines such as IL1β, IL-6, and TNFα were significantly upregulated.

These results suggest that resistin is not directly linked to inflammation, in a similar manner

to IL-6 or TNFα, and it is highly disputable if resistin can be used as a biomarker indicative

for inflammation and IR of adipose tissue in humans.

These results prompted us to search for other biomarkers indicative for inflammation in

adipose tissue as described in (Chapter 6).

The inflammatory insult of both human adipose and human liver tissues ex vivo showed that

during inflammation adipose tissue displays a more proinflammatory profile compared to

liver, as assessed by number of GO processes and number of genes involved in inflammatory

processes. These finding is in line with the hypothesis that adipose tissue is the major

proinflammatory organ during the development of IR/T2D. Furthermore the comparative

analysis led to identification of common pathways involved in inflammation/IR for both

tissues and differential ones indicating on common and differential mechanisms involved in

induction of inflammation and presumably IR. For example in liver we observed upregulation

of Jak-STAT and NFκB signaling. In adipose tissue we found upregulation of SOCS

signaling and downregulation of PPARγ signaling. These signaling pathways are known to be

Introduction

26

associated with the development of IR in vivo and support our inflammatory ex vivo model to

study processes related to IR.

The study of the inflammatory predicted secretomes of adipose and liver tissues, revealed a

similar phenomenon: the adipose tissue predicted secretome contained more genes encoding

for secreted proteins related to inflammation compared to the liver (399 vs. 236). The

predicted secretomes of both tissues contained common and differential genes, indicating the

presence of common and tissue specific biomarkers related to inflammation/IR. Within the

common predicted secretome we identified IL6, IL-1β, visfatin, CCL3, and examples of the

differential predicted secretomes were SELE, TNFα, CSF2, CSF3 in adipose tissue and

CXCL9, CXCL3, CSF1 in liver tissue. Furthermore, the predicted adipose tissue secretome

was compared with data obtained from quantitative proteomics technology approach-

CILAIR. The comparison of transcriptomics and proteomics data showed a very good

correlation and a subset of genes predicted to be secreted by adipose tissue were identified in

the adipose tissue culture as differentially affected by LPS. These proteins were regarded as

top candidate biomarkers indicative for adipose tissue inflammation/IR and were exampled

by matrix metalopeptidase-1(MMP-1), pentraxin related gene product (PTX3), fractalkine

(CX3CL1), and PAI 1. The identified common and differential tissue specific biomarkers

and signaling pathways suggest common and differential mechanisms involved in

inflammation/IR in both tissues. Therefore, it has to be verified in the future, if the identified

candidate biomarkers can be applied for recognition of IR in human liver and/or adipose

tissue and if it could facilitate a more targeted, tissue specific treatment of IR.

Introduction

27

Concluding remarks and future perspectives

(Section I) The application of human adipose tissue culture in combination with omics

technology is a powerful model to study adipose tissue biology in men and gives advantage

above monoculture of adipocytes or cell lines models of adipocytes such as murine 3T3 cells.

By application of omics technologies is it possible to study changes in a large number of

proteins or genes in an unbiased way and to screen for potential biomarkers or differentially

expressed genes between experimental groups. The proteomics technology and other omics

techniques aim to get comprehensive view of changes in biological processes. Hypothesis

generated by means of omics data in ex vivo experiments have to be verified in biological

systems in vivo by means of functional genomics or molecular biology techniques. The

potential candidate biomarkers have to be validated in patients before their application in

clinical practice. In the future, omics technologies together with the simultaneous

development of bioinformatics and modeling tools will enable to generate more accurate

hypotheses by increasing the power of prediction of the interactions in biological systems and

thereby the validation studies will become more explicit. Moreover, the tight collaborations

of scientists from different disciplines and laboratories (physiologists, molecular biologists,

geneticists, biochemists, bioinformaticians) will be necessary to interpret the data, deliver a

comprehensive view and deepen our understanding on complex metabolic diseases by a

Systems Biology approach [55].

(Section II A) The finding that application of stable GIP analogue (D-Ala2-GIP) can lead to

body mass reduction by decreased LPL activity is a novel finding and supports application of

stable GIP agonists in the treatment of obesity and T2D. These findings are in contradiction

to the in vitro studies, proposing GIP as pro-obesity and pro-diabetic agent, but in agreement

with recent clinical studies. The exact mechanism behind the stable GIP agonist actions in

vivo should be explored in more details and the functions of the novel GIP targets in adipose

tissue will deepen our understanding of GIP actions on adipose tissue in vivo.

(Section IIB) The study aiming to identify early mechanisms related to IR in humans,

suggested that alternation in metabolic gene expression preceeds changes in proinflammatory

genes expression in adipose tissues of non-diabetic women, thereby suggesting that

inflammation is not a pivotal event in the development of IR in this group of patients.

Whether this finding can be translated into a general mechanism of the development of

IR/T2D in man remains to be elucidated. The outcome of similar studies performed in non

diabetic overweight and obese Pima Indians [17], [56] showed that in cultured adipocytes and

Introduction

28

peradipocytes/stromal vascular cells of obese, healthy subjects both proinflammatory and

genes involved in lipid and carbohydrate metabolism were altered. However, it is difficult to

compare our studies and those regarding the Pima Indians, because (i) differences in the

experimental set-up (ex vivo adipose tissue vs. cultured adipose tissue cells), and (ii) the

patient groups differed significantly in genetic background, and many other biochemical and

anthropometric characteristics, for example the average BMI for Pima Indians was ~50, while

in our studies ~33. Moreover we studied much smaller subset of genes compared to the above

mentioned study where DNA microarrays were used. Therefore, changes in both metabolic

and proinflammatory genes expression could be characteristic for certain patient groups, or

be associated with high BMIs (~50). At present, there are limited data about gene expression

in adipose tissue of non-diabetic, overweight and obese patients, so it is difficult to speculate

which changes in gene expression occur primarily during the development of IR/T2D in men.

However, in comparable studies in animal models, it was found that during development of

IR in adipose tissue there were not the metabolic genes and neither the proinflammatory ones

that were affected at first. In a very recent study of Kleemann et al in 2010 [57] it was

described that primarily genes involved in “energy derivation by oxidation” were

downregulated [57] (in week 6 and sustained until week 12 of high fat diet). The acute phase

response genes were upregulated in week 6 of the high fat diet and maintained until week 12

however the upregulation of inflammatory response genes (upregulated cytokines) occurred

in week 9, and this response was further intensified in week 12. The genes involved in

carbohydrate and lipid metabolism remained unchanged during the 12 weeks of high fat diet.

These findings indicate then that primary deviations in adipose tissue involve changes related

to mitochondrial activity and acute phase response and are followed by changes in genes

involved in chronic inflammation. Therefore these data suggest, that mitochondrial genes are

the primary targets during the development of insulin resistance and the resulted oxidative

stress and ROS production will be the pivotal event leading to the development of IR. A

similar study is needed to test this hypothesis in humans.

(Section IIC) The novel finding described in this thesis, that resistin is a non- adipose tissue

specific adipokine, unrelated to inflammation, brought a novel light in the resistin research

field and requires further studies aiming to elucidate the role of resistin in liver physiology. In

studies investigating human resistin serum levels in relation to liver pathologies, it was shown

that patients with liver cirrhosis, had increased resistin serum levels which correlated with the

severity of disease [58]. Furthermore the serum resistin level was inversely correlated with

insulin sensitivity and positively correlated with markers of inflammation and with portal

Introduction

29

hypertension [59]. At present there are only few studies dedicated directly to role of resistin

in liver (patho)physiology. Zhou L. et al [60] showed that in adult human hepatocytes (L-02

cells) resistin overexpression impaired glucose tolerance and Shenk C.G.et al [61] found that

resistin was present in human hepatocytes and its overexpression led to insulin resistance.

And in a recent study described by Bertolani et al[62] resistin overexpression was observed

during chronic injury and induced proinflammatory actions. The above mentioned studies

synergistically point out towards role of resistin in liver IR and inflammation. Further

development of this research field is necessary to deepen our understanding of the role of

resistin in humans and its link to liver inflammation/IR in both liver and adipose tissue.

The new findings that adipose tissue and liver display common and differential molecular

response towards inflammatory challenge, imply that there are common and tissue specific

pathways involved in the induction of systemic IR. This phenomenon is further reflected by

the presence of the common and tissue specific biomarkers indicative for inflammation/IR.

These findings led to the development of a new hypothesis that it should be possible, based

on the presence of differential biomarkers in serum, to distinguish between inflammation/IR

in adipose tissue and the liver, thereby facilitating tissue targeted treatment of IR. Such an

approach would require extensive knowledge about the physiology of liver and adipose tissue

and the tissues specific alternations during the development of IR. This subject has to be

explored further in both animal models of obesity and in humans.

To support our hypothesis, we refer to a recently described study preformed by Kleeman et al

58 [57]. During 12-weeks long high fat feeding, the gene expression in the liver, adipose

tissue, and muscles displayed distinct patterns of changes. For example, in the liver genes

involved in mitochondrion function, energy derivation by oxidation, lipid and carbohydrate

metabolic processes were upregulated from week 9 while in adipose tissue and muscles these

processes were downregulated at the same time. Interestingly, processes such as defense

response, inflammatory response, immune response, and acute inflammatory response were

downregulated in liver and muscles while in adipose tissue the same processes were

upregulated. These findings indicate that during the development of IR each tissue undergoes

specific changes which are related to its specific functions. Thereby synergistic however

distinct actions of different organs might result in the development of systemic IR.

Furthermore it was shown that by both pharmacological interventions (i.e. thiazolidinediones

(TZDs) and life style modification (i.e. exercise training) the whole body insulin sensitivity

can be improved [63]. Moreover, these different types of interventions and different drugs

can affect metabolism and gene expression in a tissue in a specific manner. For example it

Introduction

30

was shown that PPAR gamma agonists improve insulin sensitivity by preventing toxic

accumulation of lipids in skeletal muscle and by reversing hepatic steatosis [64; 65], while

exercise training induces several adaptations that may promote glucose uptake and fatty acid

oxidation, including mitochondrial biogenesis, improved insulin signal transduction, and

elevated GLUT4 protein content [66]. In another study [67] investigating the effects of

rosiglitazone, or exercise training , or both on lipid and glucose metabolism in high fat fed

rats showed that exercise training improved insulin stimulated glucose uptake and increased

rates of fatty acid oxidation in skeletal muscle. In contrast, rosiglitazone treatment increased

lipid accumulation and decreased insulin –stimulated glucose uptake in skeletal muscle.

However, in adipose tissue the same treatment increased GLUT4 and acetyl CoA expression.

The combination of both exercise training and rosiglitazone treatment decreased liver TG

content. The above mentioned data show that although both interventions can improve the

whole body insulin sensitivity, they produce divergent effects on protein expression and

triglyceride content in different tissues. In case patients might have a predisposition to the

development of insulin resistance in a certain tissue it would be very attractive to apply tissue

specific/targeted- treatment. Therefore, in future studies aiming to examine the effects of

antidiabetic drugs, or other types of therapies, with a potential to improve insulin sensitivity,

it would be beneficial to investigate their effects in respect to tissue-specific actions in order

to provide tissue specific treatment.

Introduction

31

Summary

From the classical point of view adipose tissue is known for energy storage in the form of

triglycerides. However, during the last 15 years adipose tissue gained a lot of interest due to

its endocrine activity. Nowadays, adipose tissue is commonly accepted as an endocrine

organ secreting numerous hormonal factors called adipokines. These adipokines are involved

in divergent biological processes such as inflammation (IL-6, TNFα, IL-1β), energy

metabolism (adiponectin, leptin), reproduction (leptin) and many others. In obesity and

associated with it low grade systemic inflammation, the adipokine secretory profile changes,

which might result in deregulation of the metabolism of the adipose tissue itself and

eventually lead to systemic insulin resistance. However, the exact role of the adipose tissue in

the development of insulin resistance during these conditions is not completely known,

therefore in this dissertation our aims were: ( 1) to study the role of adipose tissue in the

development of systemic insulin resistance in relation to (A) nutritional overstimulation (GIP

signaling ), (B) clinical parameters involved in obesity and inflammation, (C) exogenous

inflammatory triggers and (2) to identify biomarkers specific for inflammation/insulin

resistance in adipose tissue by means of omics technologies such as various proteomics

techniques and DNA microarrays.

Glucose dependent insulinotropic polypeptide (GIP) was proposed as a link between

overnutrition and insulin resistance (IR), due to the fact that in several in vitro studies it was

shown to stimulate TG accumulation in the adipose tissue thereby promoting development of

obesity and insulin resistance. Additionally, it was shown that GIP receptor knockout mice

were protected from obesity on high fat diet. Therefore, in order to test the hypothesis that

access of GIP might accelerate development of obesity /IR, we performed experiments where

mice were injected with a GIP analogue during high fat or chow diets and monitored several

serum biochemical parameters and expression of subsets of proinflammatory and energy

metabolism genes in adipose tissue. Additionally, in order to identify GIP target genes in the

adipose tissue we performed DNA-microarray to screen for adipose tissue GIP target genes.

Results obtained from these studies did not confirm that excess of GIP leads directly to the

accelerated development of obesity or IR, thereby excluding GIP as a direct link between

overnutrition and IR in vivo. However, we identified several new GIP target genes in adipose

tissue: Apo –gene family members and other genes involved in lipid metabolism, and genes

with as yet unknown functions in the adipose tissue.

Introduction

32

In order to answer the question if inflammation in adipose tissue is a cause or consequence of

insulin resistance in humans, we studied gene expression of selected proinflammatory and

metabolic genes in adipose tissue in non diabetic women and their relation to several clinical

parameters indicative for early IR.

In our study group we found that the tested metabolic genes had altered expression associated

with parameters of obesity and insulin resistance. However, we did not find such correlations

for a subset of proinflammatory genes. These findings suggest that metabolic alternations in

adipose tissue precede the inflammation, thereby excluding inflammation as the pivotal event

leading to IR, at least in this particular patient group. Recent literature data [58] indicate that

dysfunction of mitochondria and the resulting overproduction of ROS could be the key events

initiating insulin resistance.

Despite the fact that there are several candidate biomarkers for systemic insulin resistance

their application in clinical practice as a supporting tool for early detection and its organ

specific origin is still futuristic. In this dissertation we aimed to validate if resistin is indeed a

good candidate biomarker (over)produced by the adipose tissue during inflammation/insulin

resistance. Our ex vivo studies did not confirm that resistin is induced by inflammation

(evoked by LPS), thereby excluding it as a biomarker indicative for inflammation / insulin

resistance. Moreover, we found that the human liver is an abundant source of resistin on both

gene and protein levels, thereby opening new avenues for the investigations devoted to the

role of resistin in the liver metabolism and its possible link to IR.

In order to find novel biomarkers and pathways indicative for inflammation/IR in adipose

tissue we studied in detail the changes in the adipose tissue secretome during inflammation

and compared the inflammatory adipose tissue secretome with the inflammatory liver

secretome. Our study led to the identification of differential pathways and biomarkers,

revealed by transcriptomic and proteomic approaches. The presence of these differential

biomarkers and pathways suggests tissue specific changes in response to

inflammation/insulin resistance, which could be applied for tissue specific detection and

treatment of IR. The adipose tissue specific biomarkers were represented by fractalkine,

tumor necrosis factor, pentraxin-related protein or interstitial collagenase (matrix

metallopeptidase 1) and the liver tissue specific biomarkers were for example chemokine (C-

X-C motif) ligand 9, chemokine (C-X-C motif) ligand 3, or follistatin-like 3 (secreted

glycoprotein).

In conclusion, in a search for the major players in insulin resistance we found that: (excess)

of GIP does not serve as a link between obesity and insulin resistance. Seeking for the

Introduction

33

primary changes in adipose tissue gene expression in the early stages of the development of

insulin resistance we found, that metabolic genes had altered expression in patients with

increased HOMA and decreased HDL serum levels while the proinflammatory genes were

unaffected in these patients. These findings suggest therefore, that metabolic alternations

might precede inflammatory ones in the early development of insulin resistance, and exclude

inflammation as a cause of IR in humans, but accommodate it as a consequence of IR.

Our finding that during inflammation adipose tissue displays a unique pattern of gene/protein

expression compared to the liver, suggests that the adipose tissue specific proteins could be

used as biomarkers to detect (adipose) tissue specific IR. Further investigations and

validation studies should explore the possibilities for the development of novel tissue specific

diagnosis of IR and thereby more targeted strategies for its treatment.

Introduction

34

Samenvatting

Van oudsher is vetweefsel bekend als een opslag plaats voor energie in de vorm van

triglyceriden. Gedurende de laatste 15 jaar is echter de interesse in vetweefsel verhoogd

doordat dit weefsel ook hormonen in de bloedbaan uitscheidt.Hierdoor wordt vetweefsel

nu gezien als een endocrien orgaan dat talloze hormonale factoren (adipokines) uitscheidt.

Adipokines zijn betrokken bij diverse biologische processen, zoals ontsteking (IL-6, TNFα,

IL-1β), energie huishouding (adiponectin, leptin), voortplanting (leptin). In vetzucht en

daaraan gekoppelde ontstekingen verandert het adipokine uitscheidings profiel waardoor het

metabolisme van het vetweefsel ontregeld kan worden. Dit zou tevens kunnen leiden tot de

ontwikkeling van systemische insuline resistentie. De exacte rol van vetweefsel in de

ontwikkeling van systemische insuline resistentie is echter niet duidelijk.

De onderzoeksdoelen gepresenteerd in deze dissertatie zijn:

(1) het onderzoeken van de rol van vetweefsel in de ontwikkeling van systemische insuline

resistentie in relatie tot (A) overvoeding (overstimulatie van GIP transmissie ketens), (B)

clinische parameters gerelateerd aan vetzucht en ontstekingsparameters, (C) LPS

geinduceeerde ontsteking en (2) het identificeren van biologische markers (indicatoren voor

een bepaalde conditie) die specifiek zijn voor ontstekingen / insuline resistentie in vetweefsel.

Deze indicatoren zijn bepaald door toepassing van zogenaamde omics technologieën

(moleculair biologische technieken die vele meetpunten opleveren) zoals diverse proteomics

technieken (bepalen eiwit profielen) en DNA microarrays (bepalen gen expressie).

Glucose afhankelijke insulinotropic poly peptide (GIP) is in wetenschappelijke studies

voorgesteld als een link tussen over-voeding en insuline resistentie. In diverse in vitro studies

is beschreven dat GIP de triglyceride ophoping in vetweefsel stimuleert, en daarmee vetzucht

en insuline resistentie (IR) bevordert. Tevens is aangetoond dat muizen waarbij de GIP

receptor (een specifieke bindings plaats voor GIP) was weggehaald, door een zogenaamd

knock-out experiment, beschermd waren tegen vetzucht indien ze een dieet hadden met een

hoog vet gehalte. Om de hypothese te testen dat toediening van GIP de ontwikkeling van

vetzucht / IR versneld, hebben we experimenten gedaan waarbij muizen zijn geïnjecteerd met

een GIP-achtige stof (GIP analoog) gedurende een controle dieet of een dieet met een hoog

vet gehalte. Vervolgens zijn diverse biochemische parameters in het serum bepaald alsmede

de expressie van een deel verzameling van genen betrokken bij energie metabolisme en

genen betrokken bij ontsteking in vet weefsel. Tevens is een DNA microarray analyse gedaan

om GIP doel-genen in vetweefsel te identificeren.

Introduction

35

De resultaten van bovengenoemde studies hebben niet bevestigd dat een overmaat aan GIP

leidt tot de versnelde ontwikkeling van vetzucht of IR. Hiermee is ook GIP als directe schakel

tussen overvoeding en IR in vivo niet aangetoond. We hebben echter wel een aantal nieuwe

GIP doel genen in vetweefsel geidentificeerd. Deze doel genen bestaan onder andere uit (i)

een aantal van de Apo-genen familie, (ii) genen betrokken bij metabolisme van vetten, en (iii)

genen met tot nu toe onbekende functie in vetweefsel.

Om de vraag te beantwoorden of in de mens een ontsteking in vetweefsel een oorzaak dan

wel een gevolg is van IR, hebben we de activiteit bepaald van een aantal genen betrokken bij

ontsteking en metabolisme in vetweefsel van niet- diabetische vrouwen en de relatie tussen

gen activiteit en clinische parameters van vroege IR geanalyseerd.

In de patiënten groep die we hebben bestudeerd, is een relatie gevonden tussen de veranderde

activiteit van metabole genen met parameters voor vetzucht en IR. We hebben echter geen

relatie kunnen vinden voor een selectie van genen betrokken bij ontsteking. Hieruit

concluderen we dat metabole veranderingen in vetweefsel voorafgaan aan een ontsteking, en

tevens dat ontsteking niet de initierende factor is in de IR ontwikkeling, in ieder geval in deze

patiënten groep. Recente literatuur [57] suggereert dat de disfunctie van mitochondrien en de

resulterende overproductie van ROS de belangrijkste factoren kunnen zijn bij de aanvang van

IR.

Ondanks het feit dat er diverse kandidaat biomarkers zijn voor systemische IR, is de klinische

toepassing hiervan als ondersteuning van vroege waarneming van (orgaan specifieke) IR iets

voor de toekomst. In deze dissertatie wilden we valideren of resistin een goede kandidaat

biomarker is die wordt (over) geproduceerd in vetweefsel gedurende ontsteking en IR.

In onze ex vivo studies hebben we niet aangetoond dat resistin wordt verhoogd door

ontsteking (door LPS geinduceerd), waardoor het waarschijnlijk niet een biomarker is voor

ontsteking / IR. Bovendien hebben we gevonden dat de menselijke lever een overvloedige

bron van resistin is zowel wat betreft gen activiteit als op eiwit niveaus. Dit opent nieuwe

mogelijkheden voor onderzoek naar de rol van resistin in lever metabolisme in relatie met

IR.

Vervolgens hebben we geprobeerd om nieuwe biomarkers en of metabole / regulatoire

cellulaire wegen te vinden die indicatief zijn voor weefsel-specifieke ontsteking / IR in

vetweefsel. Hiervoor hebben we de LPS geïnduceerde veranderingen vergeleken tussen de

door vetweefsel uitgescheiden eiwitten en lever uitgescheiden eiwitten (secretoom). Deze

studie heeft geleid tot de identificatie van diverse metabole / regulatoire cellulaire paden

waarvan de expressie verschilt tussen het vetweefsel en de lever. Tevens zijn vergeleken de

Introduction

36

biomarkers, afgeleid van het transcriptoom (bepalen gen activiteiten) en biomarkers afgeleid

van het proteoom (bepalen eiwit hoeveelheden) . De aanwezigheid van deze specifieke

biomarkers en cellulaire paden geeft indicatie dat weefsel -specifieke veranderingen

plaatsvinden bij ontsteking / IR. Deze zouden kunnen worden toegepast voor weefsel-

specifieke detectie en behandeling van IR. De vetweefsel-specifieke biomarkers bestonden

onder andere uit: fractalkine, tumor necrosis factor, pentraxin-gerelateerd eiwit en interstitial

collagenase (matrix metallopeptidase 1). Lever specifieke biomarkers waren onder andere:

chemokine (C-X-C motif) ligand 9, chemokine (C-X-C motif) ligand 3, en follistatin-like 3

(secreted glycoprotein).

Concluderend hebben we in een zoektocht naar de grootste spelers in IR gevonden dat een

(overmaat) aan GIP niet leidt tot een relatie tussen vetzucht en IR. Echter, GIP heeft een

effect op vetweefsel doordat het de activiteit veranderd van diverse genen betrokken bij vet

metabolisme, alsmede van genen die een nog onbekende functie hebben in het vetweefsel

metabolisme.

In de zoektocht naar de veranderingen in genen activiteit van vetweefsel in de vroege stadia

van ontwikkeling van IR hebben we gevonden dat metabole genen een veranderde activiteit

hadden in patiënten met verhoogd HOMA en verlaagde serum HDL spiegels, terwijl de

ontstekings gerelateerde genen een onveranderde expressie hadden in die situatie in deze

patiënten. Deze bevindingen suggereren dat metabole veranderingen de ontstekings

indicatoren voorafgaan, waardoor ontsteking als oorzaak voor IR in mensen minder voor de

hand liggend zou zijn, hoewel ontsteking wel het gevolg van IR zou kunnen zijn.

Onze bevinding dat vetweefsel een uniek patroon van genen activiteiten / eiwit hoeveelheden

laat zien ten opzichte van lever weefsel suggereert dat vetweefsel specifieke eiwitten zouden

kunnen dienen als biomarkers voor (vet) weefsel specifieke IR. Vervolg (validatie) studies

kunnen de mogelijkheden voor de ontwikkeling van nieuwe weefsel specifieke diagnose van

IR in kaart brengen en daarmee meer doelgerichte strategieen voor behandeling.

Introduction

37

References

[1] B.M.Spiegelman, J.S.Flier, Obesity and the regulation of energy balance Cell 104, (2001) 531-543. [2] V.G.Athyros, K.Tziomalos, A.Karagiannis, D.P.Mikhailidis, Preventing Type 2 Diabetes Mellitus:

Room for Residual Risk Reduction after Lifestyle Changes? Curr.Pharm.Des(2010). [3] R.Muniyappa, S.Lee, H.Chen, M.J.Quon, Current approaches for assessing insulin sensitivity and

resistance in vivo: advantages, limitations, and appropriate usage Am.J.Physiol Endocrinol.Metab 294, (2008) E15-E26.

[4] T.M.Wallace, J.C.Levy, D.R.Matthews, Use and abuse of HOMA modeling Diabetes Care 27, (2004) 1487-1495.

[5] T.Fulop, D.Tessier, A.Carpentier, The metabolic syndrome Pathol.Biol.(Paris) 54, (2006) 375-386. [6] S.Klein, D.B.Allison, S.B.Heymsfield, D.E.Kelley, R.L.Leibel, C.Nonas, R.Kahn, Waist

Circumference and Cardiometabolic Risk: a Consensus Statement from Shaping America's Health: Association for Weight Management and Obesity Prevention; NAASO, the Obesity Society; the American Society for Nutrition; and the American Diabetes Association Obesity.(Silver.Spring) 15, (2007) 1061-1067.

[7] S.Galic, J.S.Oakhill, G.R.Steinberg, Adipose tissue as an endocrine organ Mol.Cell Endocrinol. 316, (2010) 129-139.

[8] C.Buechler, J.Wanninger, M.Neumeier, Adiponectin receptor binding proteins--recent advances in elucidating adiponectin signalling pathways FEBS Lett. 584, (2010) 4280-4286.

[9] Y.Matsuzawa, Adiponectin: a key player in obesity related disorders Curr.Pharm.Des 16, (2010) 1896-1901.

[10] K.Rabe, M.Lehrke, K.G.Parhofer, U.C.Broedl, Adipokines and insulin resistance Mol.Med. 14, (2008) 741-751.

[11] A.Schaffler, J.Scholmerich, Innate immunity and adipose tissue biology Trends Immunol. 31, (2010) 228-235.

[12] G.S.Hotamisligil, Inflammation and metabolic disorders Nature 444, (2006) 860-867. [13] R.Breitling, Robust signaling networks of the adipose secretome Trends Endocrinol.Metab 20, (2009)

1-7. [14] V.Emilsson, G.Thorleifsson, B.Zhang, A.S.Leonardson, F.Zink, J.Zhu, S.Carlson, A.Helgason,

G.B.Walters, S.Gunnarsdottir, M.Mouy, V.Steinthorsdottir, G.H.Eiriksdottir, G.Bjornsdottir, I.Reynisdottir, D.Gudbjartsson, A.Helgadottir, A.Jonasdottir, A.Jonasdottir, U.Styrkarsdottir, S.Gretarsdottir, K.P.Magnusson, H.Stefansson, R.Fossdal, K.Kristjansson, H.G.Gislason, T.Stefansson, B.G.Leifsson, U.Thorsteinsdottir, J.R.Lamb, J.R.Gulcher, M.L.Reitman, A.Kong, E.E.Schadt, K.Stefansson, Genetics of gene expression and its effect on disease Nature 452, (2008) 423-428.

[15] K.Clement, D.Langin, Regulation of inflammation-related genes in human adipose tissue J.Intern.Med. 262, (2007) 422-430.

[16] M.Dolinkova, I.Dostalova, Z.Lacinova, D.Michalsky, D.Haluzikova, M.Mraz, M.Kasalicky, M.Haluzik, The endocrine profile of subcutaneous and visceral adipose tissue of obese patients Mol.Cell Endocrinol. 291, (2008) 63-70.

[17] Y.H.Lee, S.Nair, E.Rousseau, D.B.Allison, G.P.Page, P.A.Tataranni, C.Bogardus, P.A.Permana, Microarray profiling of isolated abdominal subcutaneous adipocytes from obese vs non-obese Pima Indians: increased expression of inflammation-related genes Diabetologia 48, (2005) 1776-1783.

[18] O.Poulain-Godefroy, C.Lecoeur, F.Pattou, G.Fruhbeck, P.Froguel, Inflammation is associated with a decrease of lipogenic factors in omental fat in women Am.J.Physiol Regul.Integr.Comp Physiol 295, (2008) R1-R7.

[19] M.E.Trujillo, P.E.Scherer, Adipose tissue-derived factors: impact on health and disease Endocr.Rev. 27, (2006) 762-778.

[20] H.Shi, M.V.Kokoeva, K.Inouye, I.Tzameli, H.Yin, J.S.Flier, TLR4 links innate immunity and fatty acid-induced insulin resistance J.Clin.Invest 116, (2006) 3015-3025.

[21] K.Miyawaki, Y.Yamada, N.Ban, Y.Ihara, K.Tsukiyama, H.Zhou, S.Fujimoto, A.Oku, K.Tsuda, S.Toyokuni, H.Hiai, W.Mizunoya, T.Fushiki, J.J.Holst, M.Makino, A.Tashita, Y.Kobara, Y.Tsubamoto, T.Jinnouchi, T.Jomori, Y.Seino, Inhibition of gastric inhibitory polypeptide signaling prevents obesity Nat.Med. 8, (2002) 738-742.

[22] A.Bloch-Damti, N.Bashan, Proposed mechanisms for the induction of insulin resistance by oxidative stress Antioxid.Redox.Signal. 7, (2005) 1553-1567.

Introduction

38

[23] L.E.Fridlyand, L.H.Philipson, Reactive species and early manifestation of insulin resistance in type 2 diabetes Diabetes Obes.Metab 8, (2006) 136-145.

[24] N.N.Mehta, F.C.McGillicuddy, P.D.Anderson, C.C.Hinkle, R.Shah, L.Pruscino, J.Tabita-Martinez, K.F.Sellers, M.R.Rickels, M.P.Reilly, Experimental Endotoxemia Induces Adipose Inflammation and Insulin Resistance in Humans Diabetes(2009).

[25] R.Shah, Y.Lu, C.C.Hinkle, F.C.McGillicuddy, R.Kim, S.Hannenhalli, T.P.Cappola, S.Heffron, X.Wang, N.N.Mehta, M.Putt, M.P.Reilly, Gene profiling of human adipose tissue during evoked inflammation in vivo Diabetes 58, (2009) 2211-2219.

[26] A.Virkamaki, I.Puhakainen, V.A.Koivisto, H.Vuorinen-Markkola, H.Yki-Jarvinen, Mechanisms of hepatic and peripheral insulin resistance during acute infections in humans J.Clin.Endocrinol.Metab 74, (1992) 673-679.

[27] A.Virkamaki, H.Yki-Jarvinen, Mechanisms of insulin resistance during acute endotoxemia Endocrinology 134, (1994) 2072-2078.

[28] S.P.Weisberg, D.McCann, M.Desai, M.Rosenbaum, R.L.Leibel, A.W.Ferrante, Jr., Obesity is associated with macrophage accumulation in adipose tissue J.Clin.Invest 112, (2003) 1796-1808.

[29] C.N.Lumeng, I.Maillard, A.R.Saltiel, T-ing up inflammation in fat Nat.Med. 15, (2009) 846-847. [30] S.Nishimura, I.Manabe, M.Nagasaki, K.Eto, H.Yamashita, M.Ohsugi, M.Otsu, K.Hara, K.Ueki,

S.Sugiura, K.Yoshimura, T.Kadowaki, R.Nagai, CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity Nat.Med. 15, (2009) 914-920.

[31] P.R.Flatt, Dorothy Hodgkin Lecture 2008. Gastric inhibitory polypeptide (GIP) revisited: a new therapeutic target for obesity-diabetes? Diabet.Med. 25, (2008) 759-764.

[32] R.E.Wachters-Hagedoorn, M.G.Priebe, J.A.Heimweg, A.M.Heiner, K.N.Englyst, J.J.Holst, F.Stellaard, R.J.Vonk, The rate of intestinal glucose absorption is correlated with plasma glucose-dependent insulinotropic polypeptide concentrations in healthy men J.Nutr. 136, (2006) 1511-1516.

[33] V.Marks, Human obesity: its hormonal basis and the role of gastric inhibitory polypeptide Med.Princ.Pract. 15, (2006) 325-337.

[34] V.A.Gault, N.Irwin, B.D.Green, J.T.McCluskey, B.Greer, C.J.Bailey, P.Harriott, F.P.O'Harte, P.R.Flatt, Chemical ablation of gastric inhibitory polypeptide receptor action by daily (Pro3)GIP administration improves glucose tolerance and ameliorates insulin resistance and abnormalities of islet structure in obesity-related diabetes Diabetes 54, (2005) 2436-2446.

[35] V.A.Gault, P.L.McClean, R.S.Cassidy, N.Irwin, P.R.Flatt, Chemical gastric inhibitory polypeptide receptor antagonism protects against obesity, insulin resistance, glucose intolerance and associated disturbances in mice fed high-fat and cafeteria diets Diabetologia 50, (2007) 1752-1762.

[36] P.L.McClean, N.Irwin, R.S.Cassidy, J.J.Holst, V.A.Gault, P.R.Flatt, GIP receptor antagonism reverses obesity, insulin resistance, and associated metabolic disturbances induced in mice by prolonged consumption of high-fat diet Am.J.Physiol Endocrinol.Metab 293, (2007) E1746-E1755.

[37] J.Kozawa, K.Okita, A.Imagawa, H.Iwahashi, J.J.Holst, K.Yamagata, I.Shimomura, Similar incretin secretion in obese and non-obese Japanese subjects with type 2 diabetes Biochem.Biophys.Res.Commun. 393, (2010) 410-413.

[38] I.R.Jones, D.R.Owens, S.D.Luzio, T.M.Hayes, Obesity is associated with increased post-prandial GIP levels which are not reduced by dietary restriction and weight loss Diabete Metab 15, (1989) 11-22.

[39] V.A.Fonseca, B.Zinman, M.A.Nauck, A.B.Goldfine, J.Plutzky, Confronting the type 2 diabetes epidemic: the emerging role of incretin-based therapies Am.J.Med. 123, (2010) S2-S10.

[40] H.Xu, G.T.Barnes, Q.Yang, G.Tan, D.Yang, C.J.Chou, J.Sole, A.Nichols, J.S.Ross, L.A.Tartaglia, H.Chen, Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance J.Clin.Invest 112, (2003) 1821-1830.

[41] R.B.Goldberg, Cytokine and cytokine-like inflammation markers, endothelial dysfunction, and imbalanced coagulation in development of diabetes and its complications J.Clin.Endocrinol.Metab 94, (2009) 3171-3182.

[42] S.Y.Park, Y.R.Cho, H.J.Kim, T.Higashimori, C.Danton, M.K.Lee, A.Dey, B.Rothermel, Y.B.Kim, A.Kalinowski, K.S.Russell, J.K.Kim, Unraveling the temporal pattern of diet-induced insulin resistance in individual organs and cardiac dysfunction in C57BL/6 mice Diabetes 54, (2005) 3530-3540.

Introduction

39

[43] A.O.Agwunobi, C.Reid, P.Maycock, R.A.Little, G.L.Carlson, Insulin resistance and substrate utilization in human endotoxemia J.Clin.Endocrinol.Metab 85, (2000) 3770-3778.

[44] P.D.Cani, J.Amar, M.A.Iglesias, M.Poggi, C.Knauf, D.Bastelica, A.M.Neyrinck, F.Fava, K.M.Tuohy, C.Chabo, A.Waget, E.Delmee, B.Cousin, T.Sulpice, B.Chamontin, J.Ferrieres, J.F.Tanti, G.R.Gibson, L.Casteilla, N.M.Delzenne, M.C.Alessi, R.Burcelin, Metabolic endotoxemia initiates obesity and insulin resistance Diabetes 56, (2007) 1761-1772.

[45] H.Sugita, M.Kaneki, E.Tokunaga, M.Sugita, C.Koike, S.Yasuhara, R.G.Tompkins, J.A.Martyn, Inducible nitric oxide synthase plays a role in LPS-induced hyperglycemia and insulin resistance Am.J.Physiol Endocrinol.Metab 282, (2002) E386-E394.

[46] E.M.Scott, A.M.Carter, J.B.Findlay, The application of proteomics to diabetes Diab.Vasc.Dis.Res. 2, (2005) 54-60.

[47] H.Roelofsen, M.Dijkstra, D.Weening, M.P.de Vries, A.Hoek, R.J.Vonk, Comparison of isotope-labeled amino acid incorporation rates (CILAIR) provides a quantitative method to study tissue secretomes Mol.Cell Proteomics. 8, (2009) 316-324.

[48] J.D.Bendtsen, L.J.Jensen, N.Blom, H.G.Von, S.Brunak, Feature-based prediction of non-classical and leaderless protein secretion Protein Eng Des Sel 17, (2004) 349-356.

[49] G.Sun, Application of DNA microarrays in the study of human obesity and type 2 diabetes OMICS. 11, (2007) 25-40.

[50] G.Dennis, Jr., B.T.Sherman, D.A.Hosack, J.Yang, W.Gao, H.C.Lane, R.A.Lempicki, DAVID: Database for Annotation, Visualization, and Integrated Discovery Genome Biol. 4, (2003) 3.

[51] d.W.Huang, B.T.Sherman, R.A.Lempicki, Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources Nat.Protoc. 4, (2009) 44-57.

[52] J.L.Sebedio, E.Pujos-Guillot, M.Ferrara, Metabolomics in evaluation of glucose disorders Curr.Opin.Clin.Nutr.Metab Care 12, (2009) 412-418.

[53] P.E.Scherer, Adipose tissue: from lipid storage compartment to endocrine organ Diabetes 55, (2006) 1537-1545.

[54] L.S.Van Linthout, A.Foryst-Ludwig, F.Spillmann, J.Peng, Y.Feng, M.Meloni, C.E.Van, U.Kintscher, H.P.Schultheiss, G.B.De, C.Tschope, Impact of HDL on adipose tissue metabolism and adiponectin expression Atherosclerosis 210, (2010) 438-444.

[55] E.E.Schadt, S.H.Friend, D.A.Shaywitz, A network view of disease and compound screening Nat.Rev.Drug Discov. 8, (2009) 286-295.

[56] S.Nair, Y.H.Lee, E.Rousseau, M.Cam, P.A.Tataranni, L.J.Baier, C.Bogardus, P.A.Permana, Increased expression of inflammation-related genes in cultured preadipocytes/stromal vascular cells from obese compared with non-obese Pima Indians Diabetologia 48, (2005) 1784-1788.

[57] R.Kleemann, E.M.van, L.Verschuren, A.M.van den Hoek, M.Koek, P.Y.Wielinga, A.Jie, L.Pellis, I.Bobeldijk-Pastorova, T.Kelder, K.Toet, S.Wopereis, N.Cnubben, C.Evelo, O.B.van, T.Kooistra, Time-resolved and tissue-specific systems analysis of the pathogenesis of insulin resistance PLoS.One. 5, (2010) e8817.

[58] S.Kakizaki, N.Sohara, Y.Yamazaki, N.Horiguchi, D.Kanda, K.Kabeya, K.Katakai, K.Sato, H.Takagi, M.Mori, Elevated plasma resistin concentrations in patients with liver cirrhosis J.Gastroenterol.Hepatol. 23, (2008) 73-77.

[59] E.Yagmur, C.Trautwein, A.M.Gressner, F.Tacke, Resistin serum levels are associated with insulin resistance, disease severity, clinical complications, and prognosis in patients with chronic liver diseases Am.J.Gastroenterol. 101, (2006) 1244-1252.

[60] L.Zhou, Y.Li, T.Xia, S.Feng, X.Chen, Z.Yang, Resistin overexpression impaired glucose tolerance in hepatocytes Eur.Cytokine Netw. 17, (2006) 189-195.

[61] C.H.Sheng, J.Di, Y.Jin, Y.C.Zhang, M.Wu, Y.Sun, G.Z.Zhang, Resistin is expressed in human hepatocytes and induces insulin resistance Endocrine. 33, (2008) 135-143.

[62] C.Bertolani, P.Sancho-Bru, P.Failli, R.Bataller, S.Aleffi, R.DeFranco, B.Mazzinghi, P.Romagnani, S.Milani, P.Gines, J.Colmenero, M.Parola, S.Gelmini, R.Tarquini, G.Laffi, M.Pinzani, F.Marra, Resistin as an intrahepatic cytokine: overexpression during chronic injury and induction of proinflammatory actions in hepatic stellate cells Am.J.Pathol. 169, (2006) 2042-2053.

[63] A.L.Hevener, D.Reichart, J.Olefsky, Exercise and thiazolidinedione therapy normalize insulin action in the obese Zucker fatty rat Diabetes 49, (2000) 2154-2159.

[64] B.M.Jucker, T.R.Schaeffer, R.E.Haimbach, M.E.Mayer, D.H.Ohlstein, S.A.Smith, A.R.Cobitz, S.K.Sarkar, Reduction of intramyocellular lipid following short-term rosiglitazone treatment in Zucker fatty rats: an in vivo nuclear magnetic resonance study Metabolism 52, (2003) 218-225.

Introduction

40

[65] P.D.Hockings, K.K.Changani, N.Saeed, D.G.Reid, J.Birmingham, P.O'Brien, J.Osborne, C.N.Toseland, R.E.Buckingham, Rapid reversal of hepatic steatosis, and reduction of muscle triglyceride, by rosiglitazone: MRI/S studies in Zucker fatty rats Diabetes Obes.Metab 5, (2003) 234-243.

[66] C.R.Bruce, J.A.Hawley, Improvements in insulin resistance with aerobic exercise training: a lipocentric approach Med.Sci.Sports Exerc. 36, (2004) 1196-1201.

[67] S.J.Lessard, D.A.Rivas, Z.P.Chen, A.Bonen, M.A.Febbraio, D.W.Reeder, B.E.Kemp, B.B.Yaspelkis, III, J.A.Hawley, Tissue-specific effects of rosiglitazone and exercise in the treatment of lipid-induced insulin resistance Diabetes 56, (2007) 1856-1864.

Chapter 1

41

Chapter 1

42

Section I

Optimization and development of proteomics

technologies

Chapter 1

43

Chapter 1

Fractional factorial design for optimisation of

the SELDI protocol for human adipose tissue

culture media

Ewa Szalowska

Sacha A.F.T. van Hijum

Han Roelofsen

Annemieke Hoek

Roel J.Vonk

Gerard J. te Meerman

Biotechnology Progress 2006

Chapter 1

44

Abstract

The early factors inducing insulin resistance are not known. Therefore, we are interested to

study the secretome of the human visceral adipose tissue as a potential source of unknown

peptides and proteins inducing insulin resistance.

Surface - enhanced laser desorption/ionization time-of-flight (SELDI-TOF) mass

spectrometry is a high throughput proteomics technology to generate peptide and protein

profiles (MS spectra). To obtain good quality and reproducible data from SELDI-TOF many

factors in the sample pretreatment and SELDI protocol should be optimized.

In order to identify the optimal combination of factors resulting in the best and the most

reproducible spectra we designed an experiment where factors were varied systematically

according to a fractional factorial design. In this study 7 protein chip preparation protocol

factors were tested in 32 experiments. The main effects of these factors and their interactions

contributing to the best quality spectra were identified by ANOVA. To assess the

reproducibility in a subsequent experiment the 8 protocols generating the highest quality

spectra were applied to samples in quadruplicates on different chips.

This approach resulted in the development of an improved chip protocol, yielding higher

quality peaks, and more reproducible spectra.

Introduction

Diabetes mellitus type 2 is becoming globally one of the major health problems in next

decades (1). Since obesity is recognized as one of the major risk factors in the development

of insulin resistance and eventually metabolic diseases such as diabetes (2,3), it became

important to investigate the endocrine functions of adipose tissue (2,4). Adipose tissue

consists of many cell types (adipocytes, endothelial cells, macrophages, connective tissue

cells) and secretes numerous peptides and proteins (adipokines). These adipokines affect

metabolic processes such as glucose uptake, lipolysis, lipogenesis of different tissues and

organs such as liver, muscle, and adipose tissue itself. Deregulation of adipokines secretion

may lead to metabolic alternations and eventually metabolic diseases (2,4). In future we aim

to analyse media derived from the human adipose tissue culture and compare protein profiles

Chapter 1

45

obtained from both healthy lean and obese people. Identified differences in the protein

profiles could lead to the discovery of relevant proteins related to obesity and possibly

involved in the development of insulin resistance [5,6].

In our studies we aim to find an optimal protocol for medium derived from human adipose

tissue culture and apply it in SELDI-TOF-MS (Surface- Enhanced Laser Desorption

Ionization Time-of-Flight Mass Spectrometry) technology developed by Ciphergen

Biosystems (Fremont, CA). This technology combines matrix- assisted laser

desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) and surface

chromatography. Ciphergen developed ProteinChip Arrays with different chromatographic

surfaces (e.g. weak cationic or strong anionic exchange, hydrophobic interphase) to bind

proteins from biological samples during the incubation process based on their biochemical

properties. After incubation, unbound proteins are removed by washing and the proteins

bound to the chip surface are analyzed by TOF mass spectrometry [7, 8, 9, 10].

SELDI TOF MS is a suitable technology for high throughput analysis of divergent biological

samples [11, 12, 13]. However the success of its application requires optimization of many

factors involved in sample pretreatment and chip protocol, e.g. : sample storage, protein

concentration applied on the chip, binding/washing buffer, type of applied matrix (energy

absorbing molecule-EAM), amount of applied EAM, age of EAM, number of EAM

application to the chip, and many others [14,15]. The scheme of the SELDI protocol is

depicted in Figure 1.

Chapter 1

46

Figure 1. Scheme of Chip Array protocol and tested factors. The major steps in the protocol of the Protein Chip Array. In the fractional design for optimization of the protocol we tested A, B, C, D, F, and G. In the second design for the reproducibility studies factors labeled with a star (*) were tested (C*, D*, F*, and G*). Factors in italics were not tested in factorial experiments, but the information about the applied levels were collected from prescreening experiments performed it the laboratory (results not shown).

Protein Chip array Protocol

Total protein amount loaded per spot

Molar concentration of the binding buffer B

Denaturation,no denat. A

Number of chip washes F*

Solvent of EAM C*

Type of EAM E

Saturation of EAM D*

Amount of EAM G*

Sample volume

Chip incubation with binding buffer

Sample binding

Water wash

Chip analysis in the ProteinChip Reader

Washing with binding buffer

pH of the binding buffer

EAM application

Chapter 1

47

To test all these factors one by one and by trial and error is expensive and laborious and does

not yield information on possible interactions.

Fractional factorial designs allow analyzing all major effects, determining main secondary

interactions, and confound some interactions to save on the number of experiments required

[16,17,18].

Here we describe the fractional factorial design for CM10 weak cation exchange chip

protocol for media derived from cultures of the human adipose tissue. We designed two

fractional factorial designs. The first design was to find the major factors involved in the

protocol preparation and to identify the protocols giving the best spectra quality. We analyzed

32 different combinations of conditions by analysis of variance (ANOVA) and found the

protocol generating the highest quality spectra. The second design investigated factors which

had major effects on the spectra reproducibility. In total 8 new protocols were applied to

samples in quadruplicates on 4 different chips. The protocol generating the most reproducible

spectra was selected.

The final protocol combines the findings both for the quality and reproducibility studies and

significantly contributes to improvement of generated spectra.

Materials and methods

Adipose tissue culture

About 5.5 g of human omental surgical biopsies were placed in transfer buffer (PBS

containing 5.5 mM glucose and 50 µg/ml Gentamycine), minced with scissors into pieces

(20-80 mg) and transferred to a sterile kitchen sieve with a cotton filter. On the filter the

tissue was washed thoroughly with 2 x 300 ml PBS at room temperature (RT), then briefly

with 200 ml PBS at 37ºC to remove residual blood components, leaked proteins, damaged

cells, and other low molecular debris. After washing, the fat was transferred to a 50 ml tube

with 40 ml PBS and centrifuged for 1 min at 1200 rpm to remove red blood cells and other

tissues containing insufficient amount of adipocytes to float. Afterwards the adipose tissue

was transferred to a filter and weighed again. The fat was cultured in 15 mm tissue culture

dishes. 1 g of fat was cultured in 10 ml M199 medium supplemented with 50 µg/ml

Gentamycine (Sigma) and 10 nM insulin (NovoNordisk) for 24 hrs at 37ºC in 5% CO2 The

culture medium was collected and stored at -20 ºC for further processing and analysis.

Chapter 1

48

Protein sample preparation

The culture media derived from human omental adipose tissue culture were concentrated in

Microcon Centrifugal Filter DevicesYM-3 (Millipore) according to the manufactures’

description at 15000 x g. After filtration the samples were resuspended in binding buffer (20

or 50nM NH4-Ac and 0.5% Triton, pH 4) to a final concentration of 200µg/ml. Protein

concentration was determined using a modified for 96 well plate Bradford Assay and

reagents from Bio-Rad Laboratories Ltd. Half of the concentrated media was treated with

20% ACN. The other half was untreated

Design of the Factorial Experiment

The SYSTAT 11 (Copyright © 2002 SYSTAT Software Inc.) statistical software package

was used to construct the fractional designs. The first design consisted of 7 dependent

variables (protocol factors) with two levels: 0 and 1(Table 1).

Table 1. 7 protocol factors were tested (A-G) in first fractional factorial design (quality studies). Factors labeled with * (C, D, F, G) were tested in the second fractional factorial design (reproducibility studies). Each factor was represented by two levels, defined here as level 0 and level 1. Abbreviations used in the table: ACN - acetonitrile, EAM – energy absorbing molecule, TFA – trifluoroacetic acid, SPA- sinapinic acid, CHCA - cyano -hydroxycinamic acid, DHBA - dihydroxybenzoic acid.

Symbol Factor Level 0 Level 1

A Denaturation ACN

pretreatment No ACN pretreatment

B Binding buffer 20 mM NH4Ac 50 mM NH4Ac

C* EAM solvent 50% ACN, 0.5%

TFA

30% ACN, 15% isopropanol, 0.5% TFA,

0.05% Triton

D* EAM saturation saturated 50% saturated

E Type EAM SPA CHCA/DHBA

F* Number of water

washes 1 2

G* Volume of EAM

depositions 1µl 0.5µl

Chapter 1

49

The factor A (denaturation/no denaturation), B (molar concentration of binding buffer), and F

(number of chip water washes) are related to the sample preparation and application/binding

to the chip. Factors C (solvent of EAM), D (saturation of EAM), E (type of EAM), and G

(amount of applied EAM) are related to the matrix preparation/ application. In total 32

different combinations of independent variables (protocol factors) and their levels were

applied to 4 chips (each chip had 8 spots). The independent variables and their levels were

randomly distributed on the spots using a random number table and each level of a factor is

represented 16 times, but each specific combination of 7 factors and their levels is

represented once. By performing the quality part of the experiment we aimed to estimate the

marginal contribution of each factor to the quality of spectra. Consequently we designed the

experiment in such a way that no information about all factors is available (e.g. lack of

replicates in the first design).

In the second design those factors which did not influence the quality were discarded. In this

way factors having significant effects on the spectra quality (see section Quality Criteria and

Ranking for details) were determined, in order to obtain more information on interactions and

confirm results from the first design with higher accuracy. These variables (C, D, F, and G;

the factors are depicted in the Table 1 and labeled with*) were analyzed in 8 spots (8

different combinations of tested factors at two levels) for the main effects and their

interactions. Each spot was replicated four times in order to assess the reproducibility of

obtained spectra.

Sample application on ProteinChip Array

Samples were applied to CM 10 weak cation exchange ProteinChip Arrays containing 8

spots present in a 96-well format bioprocessor (both from Ciphergen Biosystems). In total 8

chips were used in the two experiments. All chips were coming from the same lot number.

200 µl aliquots were applied in each well and covered with parafilm to avoid evaporation and

contamination during incubation (30 min. with shaking at 250 rpm on a circular shaker at

RT). After incubation the unbound material was removed by inverting the Bioprocessor and

striking it onto paper towels. Three 5- min washes with 200 µl binding buffer were performed

on a shaker at 250 rpm (RT). The chips were removed from the bioprocessor and briefly

washed in a 15 ml tube with Milli-Q water. Finally the chips were air-dried for 20 min (RT).

Chapter 1

50

Matrix preparation and application

Two types of energy absorbing molecules (EAM) were tested in these experiments: (i)

sinapinic acid (SPA), and (ii) a mix of (alpha)-cyano 4-hydroxycinamic acid (CHCA) and

dihydroxybenzoic acid (2, 3- DHBA); all EAM were purchased from Sigma-Aldrich).

Solutions of each of the EAMs were prepared approximately 10 min before application on the

dried spots containing bound proteins. Two solvents were tested (1) 50% ACN, 0.5% TFA

and (2) 30% ACN, 15% isopropanol, 0.5% TFA, 0.05% Triton. The solutions were prepared

in light-resistant 1.5 ml tubes. The tubes with the solvents were mixed vigorously using a

vortex. Afterwards the solutions were incubated for 5 min RT and vortexed again,

centrifuged 2 min at 14000 x g. For each spot EAM was applied twice (saturated) or (50%

saturated). Before first and second EAM application the spots were dried for 5 min. After

applying the matrix and an additional 10 min drying period, the chips were measured in a

SELDI-TOF mass spectrometer (PBS-II; Ciphergen Biosystems).

Chips measurement

Prior to the chip measurements, the SELDI-TOF mass spectrometer was calibrated using the

All-in-1 Peptide Standard for mass ranges up to 20 kDa and the protein molecular weight

calibrant kit for the mass range higher than 20 kDa (both from Ciphergen Biosystems). Each

chip was measured for the following mass ranges: (1) 3-20 kDa and (2) 20-150 kDa. Data

were captured according to an automated data collection protocol with proteinchip software

3.1 (Ciphergen). For the range 3-20 kDa the following settings were used: detector voltage

2050; detector sensitivity 9; 2 warming shots at laser intensity 235 (not collected); collection

shots at laser intensity 230, 5 shots were collected every five positions between 20 and 80;

high mass 20 kDa; mass optimization 3-16 kDa. For the range 20-150 kDa the settings were

as follows: detector voltage 2050; detector sensitivity 9; 2 warming shots at laser intensity

255 (not collected); collection shots at laser intensity 250, 5 shots were collected every five

positions between 20 and 80; high mass 150 kDa; mass optimization 20-70 kDa. After

baseline subtraction, peaks were assigned by automatic peak detection with the Ciphergen

ProteinChip Software (version 3.1). For the M/Z range from 3-20 kDa peaks with signal to

noise (S/N) ratio ≥3 were labeled, and for the M/Z range from 20-150kDa peaks with S/N-

ratio ≥5 were labeled. In addition to the automatic peak labeling we also labeled peaks

manually according the same criteria since the Ciphergen ProteinChip Software often omits

peaks in automatic labeling. Peak information was exported to Microsoft Excel. Analysis of

the spectra quality, reproducibility, and ranking was performed in Microsoft Excel.

Chapter 1

51

Quality criteria (dependent variables) and ranking

The criteria to describe the quality of the generated spectra (dependent variables) were as

follows: (1) total number of peaks (NRP), (2) average signal to noise ratio (S/N) (3) shape of

the peaks (PC1); (4) resolution of the peaks (PC2). We aimed for the highest number of

detectable peaks, since we want to screen for as many peptides/proteins as can be detected. A

high signal to noise ratio is beneficial, since this reduces detection of high noise levels in

some cases recognized by ProteinChip Software as potential peaks and facilitates peak

detection. The shape of peaks was determined by the peak cleanness (PC) 1 formula 1:

∑=½

11

W

M

NPPC (1)

Where NP is the number of peaks, M is the mass of a peak, W ½ its width at half-height. The

summation is taken of all the detected peaks (Cordingley H.C. et al.2003). The PC1 value is

also supplied by the Ciphergen Protein Chip Software where it is named MZ resolution. The

applied criterion was to maximize the PC 1 value and thereby select a spectrum with the best

shaped peaks. Peak resolution was determined by the peak cleanness PC2 formula (2):

PC2=W2½½1

)12(2

+−

W

MM (2)

Where M is the mass of a peak, W½ its width at half-height (Cordingley H.C. et al.2003). A

higher value indicates better resolution between adjacent peaks.

After calculation of the above mentioned criteria, each spectrum was ranked for the total

number of peaks, the average signal to noise ratio, and the averages of PC1 and PC2 values.

The best spectrum was ranked as 1, second best as 2 and so on. Two spectra with identical

properties for a specific quality criterion obtained the same average rank. Ranks were

calculated for each quality criterion and for each spectrum.

In the reproducibility studies 8 different protocols were tested. Samples treated according to

each protocol were applied in quadruplicate on 4 different chips. The reproducibility of the

protocol was assessed based on coefficient of variation (CV) for average of mass deviation

(MS), signal to noise value (SN), peak cleanness (PC1), and peak resolution (PC2). The

spectra generating the lowest CV value for a particular variable were ranked as 1, second

lowest CV as 2 and so on. Ranks were calculated for each reproducibility criterion and for

each spectrum.

Chapter 1

52

Statistical analysis (ANOVA)

Two fractional factorial experiments were performed: in the first experiment we assessed the

major factors significantly influencing the spectrum quality and determine the major

interactions between these factors and their levels (7 protocol factors were tested in 32

protocols). We identified protocol factors and their levels yielding the highest quality of the

spectra.

For the reproducibility study a second fractional design was created. In this design we tested

4 selected factors (independent variables) in 8 new protocols. Each protocol was applied in

quadruplicate on four different chips Afterwards we calculated major effects and 2 way

interactions.

ANOVA was performed using SPSS 12.0.1 for Windows (SPSS Inc., Chicago, IL, USA)

using the general linear model; all other calculations were performed in Microsoft Excel.

In both experiments we analyzed the main effects and the two- and three way interactions

that could be estimated. Effects and interactions were considered significant at p ≤ 0.02. A

number of 2- and 3-way interactions could not be estimated due to the confounding that is

intrinsic to the fractional factorial design.

Results and discussion

Quality study

In the first fractional factorial experiment we tested 7 factors at two levels of the CM10 chip

protocol and obtained 32 spectra. All tested protocol factors and their levels are shown in

Table 1.

Our goal was to maximise the values obtained for the four quality criteria (dependent

variables): number of peaks (NR), signal to noise values (S/N), peak shape (PC1), and peak

resolution (PC2). The obtained spectra showed many differences regarding the quality

criteria. With ANOVA we determined the main protocol factors and some of their secondary

interactions contributing to the development of the best quality traces. Since we are interested

in finding biomarkers we analysed spectra in a wide range of masses from 3 to 150 kDa. All

major effects and interactions are shown in Table 2.

Chapter 1

53

Table 2. The main effects determined by ANOVA in the quality studies. All main effects and interactions with p ≤ 0.02 were considered as significant. A number of interactions between 2 factors and their levels were determined, others were confounded. The main effects and the major interactions were established for peptides and proteins in the mass range 3-150 kDa.

The analysis showed that factor A at level 1 (A1, no ACN pretreatment of the sample)

generates spectra with the highest number of peaks, what may be explained by protein

precipitation caused by ACN and depletion of some proteins from the sample, when it is

ACN treated. Better signal to noise values were obtained by applying E0 (SPA) but not E1

(DHBA/CHCA mixture). Application of 1 µl of matrix (G0), instead of 0.5 µl (G1) generates

better spectra regarding signal to noise value, what may be caused by better crystallization

process and therefore better ionization and flying of proteins. A clear 2 way interaction

affecting the S/N (signal to noise) was found for factors A1*E0 (no ACN pretreatment *

SPA, p=0.002), and D1*E0 (50% EAM * SPA, p=0.000). Almost all found interactions for

type of EAM clearly benefit SPA (E0) instead of CHCA/DHBA (E1) mixture. However for

PC1 value (peak shape) the best settings are achieved when a combination of factors D1*E1

(50% saturated CHCA) was applied (p=0.009). It is difficult to explain why a combination of

50% saturated solution of CHCA/DHBA yields overall better peak shapes. The analysis

showed that also for different analysed spectrum ranges the best settings for PC1 were

achieved when using D1*E1 (6-9, 10-20 (data not shown), and 3-20kDa). Moreover all 2-

way interactions found for 3-150 kDa range show better results for 50% saturated EAM

solution (D1) compared to 100% saturated EAM (D0). Mean rank values for all identified

combination of factors for the mass range 3-150 kDa are shown in Figure 2.

Spectrum

range kDa

Dependent variable

Main effect

Type III sum of squares

F value

P value 2-way

interactions

Type III sum of squares

F value

P value df

3-150

NRP A1 176.781 7.208 0.013 1

SN E0 G0

1378.125 300.125

48.891 10.647

3,13e-007 0.003

A1* E0 D1* E0

136.125 242.000

17.211 30.598

0.002 4.34e-008

1

PC1 D1 364.500 5.693 0.025 D1* E1 496.125 9.938 0.009 1 PC2 A1* E0 666.125 11.418 0.006 1

Chapter 1

54

Figure 2. Mean rank values for mass range 3-150 kDa for all significant 2 way interactions (p ≤ 0.02). A: Signal to noise A*E interaction,; B: Signal to noise D*E interaction; C: Peak shape D*E interaction; D: Peak resolution A*E interaction.

Since it is known that SELDI-TOF sensitivity varies for different spectrum ranges and even

specific proteins, it was also interesting to narrow down the analysed spectrum ranges and see

if indeed identified main effects and their interactions are unique for diverse spectrum areas.

We analysed the following mass ranges: 3-10, 6-9, 9-12.5, 10-20, 3-20, 20-150 kDa and

determined major effects and interactions influencing the quality (in Table 3 are shown data

for, 3-20, 20-150, 6-9 kDa spectrum ranges, data found for mass ranges 3-10, 9-12.5, and 10-

20 kDa are not shown ).

Chapter 1

55

Table 3. The main effects determined by ANOVA in the quality studies. All main effects and interactions with p ≤ 0.02 were considered as significant. A number of interactions between 2 factors and their levels were determined, others were confounded. The main effects and the major interactions were established for peptides and proteins, divided into three groups 3-20, 20-150 and 6-9 kDa.

For the 3-20 kDa range, the most sensitive of SELDI TOF measurement, NRP (number of

peaks) gives better results by applying factor A1 (no ACN pretreatment). C0 (EAM solvent

as 50% ACN and 0.5% TFA), D1 (50 % saturated EAM), and F0 (one water wash) yields

spectra with the highest number of peaks either. The higher NRP value (number of peaks)

achieved by application of 1 water wash (F0) instead of 2, could be explained by less

extensive removal of proteins bound to the chip surface and visualized in the spectra. Strong

2 way interactions were found for S/N (signal to noise) value; better results were achieved by

applying A0*D0 (no ACN pretreatment*100% EAM solution), p=0.004, and C0*D0 (50%

ACN, 0.5% TFA*100% EAM solution), p=0.011. PC1 (peak shape) value is improved by

applying D1 (50% saturated EAM, and D1*E1 (50% EAM*CHCA/DHBA, p=0.007) and

A0*C1 (ACN pretreatment*30% ACN, 15% isopropanol 0.5% TFA. 0.05% Triton, p=

0.020). PC2 values (peak resolution) are improved by applying A0 (ACN pretreatment,

p=0.001). This fact may be explained by a precipitation of some proteins while applying

ACN in the pretreatment of the sample, which could reduce the number of proteins bound to

the chip, thereby raising the resolution. For our purposes it is more relevant to increase the

Spectrum

range kDa

Dependent variable

Main effect

Type III sum of squares

F value P value 2-way

interactions

Type III sum of squares

F value P value df

3-20

NRP

A1 C0 D1 F0

242.000 36.125 78.125 45.125

39.177 5.848 12.648 7.305

1,80e-006 0.024 0.002 0.012

1

SN A0* D0 C0* D0

38.281 26.281

13.543 9.297

0.004 0.011

1

PC1 D1 810.031 21.322 0.000109 D1* E1 A0* C1

306.281 215.281

11.079 6.650

0.007 0.020

1

PC2 A0 D0

420.500 722.000

12.951 22.237

0.001 8,55e-005

1

20-150

NRP E0 300.125 73.877 8,65e-009 D1* E0 40.500 18.758 0.001 1

SN E0 G0

1391.281 294.031

51.410 10.865

2,07e-007 0.003

A1* E0 D1* E0 D0* G0

101.531 205.031 87.781

10.168 20.532 8.791

0.009 0.001 0.013

1

PC1 E1 166.531 7.201 0.013 1

PC2 D0 457.531 5.815 0.024 A1* E0 488.281 7.216 0.021 1

6-9

NRP E1* F0 16.531 7.707 0.018 1 SN E1 634.500 6.727 0.016 A1* E1 264.500 8.015 0.016 1

PC1

C1 D1 E0 G1

318.781 657.031 282.031 318.781

7.536 15.532 6.667 7.536

0.011 0.001 0.016 0.011

D1* E1 413.281 20.174 0.001 1

PC2 D0 648.000 13.344 0.001 D0* E1 450.000 12.807 0.004 1

Chapter 1

56

peak number than resolution. For this reason we chose no ACN pretreatment of the sample

(A1).

The mass range 20 -150 kDa clearly shows the advantageous effect of SPA (E0) on NRP

(number of peaks), S/N (signal to noise), but not PC1 (peak shape) values, here the

CHCA/DHBA mixture is better (E1). This finding for NRP and S/N values is expected since

it is known that SPA is a better EAM for high molecular weight proteins. It is intriguing why

the shape of peaks is improved with application of the CHCA/DHBA mixture. We cannot

offer an explanation for that. Only for this mass range the volume of EAM depositions has a

major effect on S/N (signal to noise); 1µl (G0) is better than 0.5 µl (G1). This fact is difficult

to understand theoretically, but it may be hypothesized that a larger volume of matrix

improves crystallization and ionization of larger proteins. For NRP (nr of peaks) and S/N

(signal to noise) we found a very strong interaction between 50% saturated EAM solution

and SPA (D1*E0) (p=0.0001 for both quality criteria). Signal to noise ratio is also improved

in protocols where no ACN pretreatment was combined with application of SPA (A1*E0,

p=0.009). For the PC2 (peak resolution) value the higher results are achieved by application

of saturated EAM (D0).

In these studies we generated in total 32 unique spectra with significant variation in quality.

ANOVA revealed that a number of main factors were associated with multiple quality criteria

(dependent variables) across different mass ranges (e.g. E0 (SPA) gives better results for

signal to noise (SN) values for mass ranges 3-150 and 20-150 kDa; number of peaks (NRP)

for the mass range 20-150 kDa; peak shape (PC1) for the mass range 6-9 kDa). Some factors

relate to very specific quality criteria of single spectra ranges (e.g. C0 (EAM solvent as 50%

ACN and 0.5% TFA) increases number of peaks (NRP) only for the mass range 3-20 kDa;

G1 (0.5µl of EAM) improves peak shape (PC1) only for the mass range 6-9 kDa). Moreover

the analysis showed that there is a clear difference between combinations of factors yielding

the best quality for small (3-20 kDa) and high mass ranges (20-150 kDa). In practice it means

that different protocols should be applied to obtain optimal results for low and high molecular

weight proteins. In Figure 3 the spectra from 3-20 and 20-150 kDa ranges obtained with the

best and the worst protocol settings are shown.

Chapter 1

57

Figure 3. SELDI MS spectra from the factorial design experiment for protocol optimization. The spectra are in the range 3-20 kDa (A and B) and 20-150 kDa (C and D). The pictures show the traces performed with the best (A and C) and the worst (B and D) protocol. Next to the spectrum symbol (A, B, C and D) factors levels yielding the particular trace are shown ( A0, B1 etc.).

0

10

20

30

0

10

20

30

5000 10000 15000

0

2.5 5

7.5 10

0

2.5 5

7.5 10

25000 50000 75000 100000 125000

5000 10000 15000 kDa

25000 100000 125000 kDa

30 20 10 0

30 20 10 0

10 5 0

10 5 0

Mass/ Charge (m/z)

Rel

ativ

e In

ten

sity

R

elat

ive

Inte

nsi

ty

A: A1, B0, C0, D0, E1, F1, G1 B: A0, B0, C1, D0, E0, F0, G1

C: A1, B0, C0, D1, E0, F1, G1 D: A0, B1, C1, D1, E1, F1, G1

Chapter 1

58

Reproducibility study

In the second fractional design used in a reproducibility study we tested following factors: C,

D, F, and G on both levels (0 and 1); factors A, B, and E were set up at level 1, 1, and 0

respectively. We generated 8 new protocols in SYSTAT 11, and each protocol was applied to

four distinct chips. We used ANOVA to identify main effects and interactions related to

spectrum reproducibility. We calculated coefficients of variation (CV) for mass deviation

(MS), signal to noise value (S/N), peak shape (PC1), and peak resolution (PC2). The obtained

CV values strongly indicate that reproducibility can vary considerably depending on the chip

protocol. For the range of 3-20 kDa we could improve the CV for MS, S/N, PC1, and PC2.

The improvement was 33% for the mass deviation 55.6% for signal to noise, 54% for peak

shape, and 25% for peak resolution.

For the range 20-150 kDa the improvement was 55% for mass deviation (MS), 47% for

signal to noise (S/N), 62% for peak shape (PC1), and 58% for peak resolution (PC2). All CV

values are in Table 4.

Table 4. CV values obtained in the reproducibility studies with the best and the worst protocol, for mass ranges 3-20 and 20-150 kDa.

Different main factors and different levels affecting different quality criteria were identified.

There was a clear distinction between pattern of factors affecting the reproducibility of 3-20

kDa range and 20 -150 kDa. For the range 3-20 kDa the highest reproducibility for CV of

mass deviation is achieved by applying EAM solvent at level 1 (C1); 50% saturated EAM

(D1); two water washes (F1). Two water washes (F1) and 0.5µl of EAM (G1) reduces the CV

for signal to noise. The CV for a peak shape (PC1) is improved by applying EAM solvent at

level 0 (C0), saturated EAM (D0), one water wash (F0), and 1µl of applied EAM (G0). The

Spectrum range kDa Dependent variable Best CV Worst CV

3-20

MS 0.03 0.09

SN 22.6 40.6

PC1 11.3 21

PC2 4.7 18.6

20-150

MS 0.11 0.20

SN 15 32

PC1 13.3 21.4

PC2 17 29.3

Chapter 1

59

CV for peak resolution (PC2) is improved similarly to the CV of PC1 by using C0 and F0,

however 0.5µl of EAM (G1) is better for the CV of PC2. For the range 20-150 kDa the

highest reproducibility regarding mass deviation is achieved by applying C0, D1, F1 (EAM

solvent at level 0; 50% saturated EAM, two water washes, respectively). D0, F1 and G0

(saturated EAM, two water washes, 1µl of applied EAM, respectively) improve the CV for

signal to noise values. The CV for peak shape in reduced in all protocols when two water

washes (F1) were applied. The CV for peak resolution were decreased in protocols where

50% saturated EAM (D1) and two water washes (F1) where applied. For all factors and

interactions in the reproducibility studies see Table 5.

Table 5. The main effects and 2 way interactions determined in the reproducibility studies by ANOVA. All main effects and interactions with p ≤ 0.02 were considered as significant. The main effects and the major interactions were determined for the mass ranges 3-20 and 20-150 kDa.

Spectrum

range kDa

Depend. variable

Main effect

Type III sum of squares

F value

P value

df

2-way int.

Type III sum of squares

F value

P value

df

3-20

MS C1 D1 F1

18.000 32.000 18.000

30.375 54.000 30.375

7.73e-006 6.63e-008 7.73e-006

1

C1D1 C1F1 C1G1 F1G1 D1F1 D1G0

58.000 38.000 38.000 38.000 50.000 42.000

19.333 7.389 7.389 7.389 12.963 8.909

5.43e-007 0.002 0.001 0.001

1.71e-005 0.000263

3

SN F1 G1

32.000 50.000

10.537 16.463

0.003 0.00038

1

C1F1 C1G1 F1G1 D0G1

84.000 84.000 84.000 84.000

9.333 9.333 9.333 9.333

0.000193 0.000193 0.000193 0.000193

3

PC1

C0 D0 F0 G1

98.000 32.000 8.000 18.000

220.500 72.000 18.000 40.500

1.64e-014 4.24e-009 0.00023

8.15e-007

1

C0D0 C0F0 C0G1 F0G1

132.000 124.000 124.000 124.000

34.222 26.303 26.303 26.303

1.66e-009 2.69e-008 2.69e-008 2.69e-008

3

PC2 C0 F0 G1

32.000 32.000 32.000

12.000 12.000 12.000

0.002 0.002 0.002

1

C0F0 C0G1 F0G1 D0F0 D1G1

96.000 96.000 96.000 64.000 64.000

12.444 12.444 12.444 5.744 5.744

2.36e-005 2.36e-005 2.36e-005

0.003 0.003

3

20-150

MS C0 D1 F1

32.000 8.000 32.000

54.000 13.500 54.000

6.63e-008 0.001

6.63e-008 1

C0D1 D1G1

8.000 16.000

6.500 26.000

0.005 6.27e-007

2

SN D0 F1 G0

32.000 32.000 32.000

13.500 13.500 13.500

0.001 0.001 0.001

1 C1D0 D0G0

64.000 32.000

26.000 6.500

6.27e-007 0.005

2

PC1 F1 72.000 27.000 1.79e-005 1 D1F1 40.000 13.000 0.00012 2

PC2 D1 F1

32.000 72.000

18.000 40.500

0.000232 8.15e-007

1 C0D1 D1F1 D1G1

40.000 40.000 64.000

13.000 13.000 52.000

0.00012 0.00012

8.19e-010 2

Chapter 1

60

In some cases is difficult to understand the reason why some factors at particular levels yield

more reproducible spectra. Moreover the optimal levels of protocol factors show in a number

of cases opposite parameters for the optimal outcome of a specific criterion for a particular

mass range due to interactions. In practice it means that when choosing the most reproducible

protocol compromises have to be made, and very often it may be necessary to prioritize some

criteria above others. For example in seeking for biomarkers the number of peaks is more

important than peak shape, but for a better visualization of a peak of interest, the resolution

and the peak shape would be more important then the total number of peaks in a spectrum.

Definitely there is a large difference between a characterization of 3-20 kDa range spectrum

and the 20-150 kDa range spectrum, which indicates that it is actually favorable to have

separate protocols for these two ranges of spectra in order to achieve an optimal outcome.

Conclusions

The final protocol we advocate for the screening within mass range 3-150 kDa is a

compromise between quality and reproducibility studies, but also between small and higher

mass ranges characteristics, prioritizing some criteria above others, such as number of peaks

above peak shape etc.. The optimal settings are: A1 (no ACN pretreatement), C1 (30% ACN,

15% isopropanol, 0.5%TFA, 0.05% Triton as EAM solvent), D1 (50% saturated EAM), E0

(SPA), F1 (two water washes), and G1 (0.5µl of EAM). The level of factor B does not have a

major effect on the analyzed criteria, and can be set at level 1 (50mM NH4Ac). In case of

interest in more narrowed-down spectrum range the protocol should be adjusted according to

the found characteristics for this specific mass range.

Many results found in these studies are as expected and can be scientifically explained but

some of them are difficult to understand. The systematic statistical approach results in a

protocol that generates high quality spectra and significantly improves the reproducibility

over previous protocols. It is possible that the outcome is specific for our experimental

condition. We recommend the application of fractional factorial design experiments for

different sample types and chip types because it is an efficient way to gain insight into the

major effects and especially their interactions involved in a particular preparation process.

Chapter 1

61

References

(1) Freeman, H.; Cox R.D. Type-2 diabetes: a cocktail of genetic discovery. Hum. Mol. Genet. 2006, 15, Suppl 1: R202-9.

(2) Esposito K.; Giugliano G.; Scuderi N.; Giugliano D. Role of adipokines in the obesity-inflammation relationship: the effect of fat removal. Plast. Reconstr. Surg. 2006, 118, 1048-1059.

(3) Bastard J.P.; Maachi M.; Lagathu C.; Kim M.; J.; Caronn M.; Vidal H.; Capeau J.; Feve B. Recent advances in the relationship between obesity, inflammation, and insulin resistance. 2006, 17, 4-12.

(4) Vendrell J.; Broch M., Vilarrasa N.; Molina A.; Gomez J.M.; Gutierrez C.; Simon I.; Soler J.; Richart C. Resistin, adiponectin, ghrelin, leptin and proinflammatory cytokines: relationships in obesity. 2004, 12, 962-971.

(5) Greenberg, A.S.; Obin, M.S. Obesity and the role of adipose tissue in inflammation and metabolism. Am J. Clin. Nutr. 2006, 83, 461S-465S.

(6) Scherer, P.E. From lipid storage compartment to endorcine organ. Diabetes 2006, 55, 1537-1545. (7) Merchant M.; Weinberger S.R. Recent advancements in surface-enhanced laser desorption/ionization-time

of flight-mass spectrometry. Electrophoresis 2000, 21, 1164-1177. (8) Bertucci F.; Birnbaum D.; Goncalves A. Proteomics of breast cancer: principles and potential clinical

applications. Mol. Cell Proteomics 2006, in press. (9) Clarke C.H.; Buchley J.A.; Fung E.T. SELDI-TOF-MS proteomics of breast cancer. Clin. Chem. Lab. Med.

2005, 43 , 1314-1320. (10) Issaq H.J.; Veenstra T.D.; Conrads T.P.; Felschow D. The SELDI-TOF MS approach to proteomics:

protein profiling and biomarker identification. Biochem. Biophys. Res. Commun. 2002, 292, 587-592. (11) Nomura F.; Tomonaga T.; Sogawa K.; Ohashi T.; Nezu M.; Sunaga M.; Kondo N.; Iyo M.; Shimada H.;

Ochiai T. Identification of novel and downregulated biomarkers for alcoholism by surface enhanced laser desorption/ionization-mass spectrometry. Proteomics 2004, 4, 1187-1194.

(12) Roelofsen H.; Balgobind R.; Vonk R. J. Proteomic analyzes of copper metabolism in an in vitro model of Wilson disease using surface enhanced laser desorption/ionization-time of flight-mass spectrometry. J. Cell Biochem. 2004, 93, 732-740.

(13) Cadieux P.A.; Beiko D.T.; Watterson J.D.; Burton J.P.; Howard J.C.; Knudsen B.E.; Gan B.S.; McCormick J.K.; Chambers A.F.; Denstedt J.D.; Reid G. Surface-enhanced laser desorption/ionization-time of flight-mass spectrometry (SELDI-TOF-MS): a new proteomic urinary test for patients with urolithiasis. J. Clin. Lab. Anal. 2004, 18, 170-175.

(14) Chertov O.; Biragyn A.; Kwak L.W.; Simpson J.T.; Boronina T.; Hoang V.M.; Prieto D.A.; Conrads T.P.; Veenstra T.D.; Fisher R.J. Organic solvent extraction of proteins and peptides from serum as an effective sample preparation for detection and identification of biomarkers by mass spectrometry. Proteomics 2004, 4 , 1195-1203.

(15) Laugesen S.; Roepstorff P. Combination of two matrices results in improved performance of MALDI MS for peptide mass mapping and protein analysis. J. Am. Soc. Mass Spectrom. 2003, 14, 992-1002.

(16) Cordingley H.C.; Roberts S.L.; Tooke P.; Armitage J.R.; Lane P.W.; Wu W.; Wildsmith S.E. Multifactorial screening design and analysis of SELDI-TOF ProteinChip array optimization experiments. Biotechniques 2003, 34, 364 -373.

(17) Park J.T.; Bradbury L.; Kragl F.J.; Lukens D.C.; Valdes J.J. Rapid optimization of antibotulinum toxin antibody fragment production by an integral approach utilizing RC-SELDI mass spectrometry and statistical design. Biotechnol. Prog. 2003, 22, 233-240.

(18) Baranda A.B.; Etxebarria N.; Jimenez R.M.; Alonso R.M. Improvement of the chromatographic separation of several 1,4-dihydropyridines calcium channel antagonist drugs by experimental design. J.Chromatogr. Sci. 2005, 43, 505-512.

Chapter 2

62

Chapter 2

63

Chapter 2

Characterization of the human visceral adipose

tissue secretome

Gloria Alvarez-Llamas

Ewa Szalowska

Marcel P. de Vries

Desiree Weening

Karloes Landman

Annemieke Hoek

Bruce H.R. Wolffenbuttel

Han Roelofsen

Roel J. Vonk

Molecular and Cellular Proteomics 2007

Chapter 2

64

Summary

Adipose tissue is an endocrine organ involved in storage and release of energy, but also in

regulation of energy metabolism in other organs via secretion of peptide and protein

hormones (adipokines). Especially visceral adipose tissue has been implicated in the

development of metabolic syndrome and type 2 diabetes. Factors secreted by the stromal-

vascular fraction contribute to the secretome and modulate adipokine secretion by adipocytes.

Therefore, we aimed at the characterization of the adipose tissue secretome rather than the

adipocyte cell secretome. The presence of serum proteins and intracellular proteins from

damaged cells, released during culture, may dramatically influence the dynamic range of the

sample and thereby identification of secreted proteins. Part of the study was therefore

dedicated to the influence of the culture set-up on the quality of the final sample. Visceral

adipose tissue was cultured in five experimental set-ups and the quality of resulting samples

was evaluated in terms of protein concentration and protein composition. The best set-up

involved one wash after the first hour in culture followed by two or three additional washes

within an eight hour period, starting after overnight culture. Thereafter, tissue was maintained

in culture for additional 48 to 114 hrs to obtain the final sample. For the secretome

experiment, explants were cultured in media containing 13C6,15N2 L-lysine to validate the

origin of the identified proteins (adipose tissue or serum derived). In total, 259 proteins were

identified with ≥99% confidence. 108 proteins contained a secretion signal peptide of which

70 incorporated the label and were considered secreted by adipose tissue. These proteins were

classified into five categories according to function. This is the first study on the (human)

adipose tissue secretome. The results of this study contribute to a better understanding of the

role of adipose tissue in whole body energy metabolism and related diseases.

Chapter 2

65

Introduction

Adipose tissue is a key organ for the regulation of energy metabolism. Besides its function as

an energy storage depot in the form of triglycerides, adipose tissue secretes a variety of

peptide and protein hormones (adipokines) involved in the regulation of energy metabolism

such as leptin, adiponectin, visfatin, retinol binding protein-4, adipsin, tumor necrosis factor α

(TNF-α) and interleukin 6 (IL-6) (1-3). Dysregulation of the production of adipokines and

free fatty acids contributes to the pathogenesis of diseases associated with energy metabolism

such as insulin resistance, metabolic syndrome and type 2 diabetes. Especially visceral

adipose tissue has been implicated in the development of these diseases (2-4). Therefore,

more insight into the visceral adipose tissue secretome will contribute to a better

understanding of its role in energy metabolism and related diseases and may lead to the

discovery of unknown peptides/proteins involved in regulation of energy metabolism and

new targets for therapy. Besides adipocytes, adipose tissue contains endothelial cells,

macrophages and fibroblasts (stromal fraction) which may modulate the overall peptide and

protein secretion pattern of the tissue via cross talk between the different cell types. For

example, factors secreted by macrophages have been shown to induce changes in the

secretion of adipokines, free fatty acids and glucose uptake by 3T3-L1 adipocytes (5). These

interactions between cells from the stromal fraction and adipocytes are necessary for

physiological functions of adipose tissue, and deregulation of this cross talk is regarded as an

important mechanism leading to insulin resistance and type 2 diabetes (6-9). Therefore, the

tissue secretome provides more relevant information for the in vivo situation than the

adipocyte cell secretome. To date, no studies have been published on the adipose tissue

secretome. Several studies investigated the human (10), mouse (11-14) and rat (15) adipose

tissue proteome, mostly using a two-dimensional gel electrophoresis (2DE) approach. Celis et

al. (16) analysed the human mammary adipose tissue proteome. The secretome from

adipocyte cells has been investigated by Kratchmarova et al. (17) and Wang et al. (18). They

studied changes in protein secretion during differentiation of the 3T3-L1 mouse pre-

adipocyte cell line to adipocytes. In another study in isolated rat adipocytes, Chen et al. (19)

identified 84 secreted proteins using 2DLC-MS/MS.

The major challenge for characterization of the adipose tissue secretome is the quality of the

secretome sample. The presence of serum proteins inside the tissue pieces that slowly diffuse

into the culture medium and the presence of intracellular proteins that are released from

damaged cells due to the cutting of the tissue, necessary for culture, can dramatically

Chapter 2

66

influence the dynamic range of the sample and thereby the detection of the secreted proteins.

Also the relevance of the identified proteins can be unclear if the source of the proteins

(secreted, serum or intracellular) is not clear due to a high level of contaminant proteins. This

has been a problem in previous studies with tissue explants where e.g. adiponectin, secreted

by the tissue, could not be reliably measured due to diffusion of serum adiponectin, as a

second source of adiponectin, from the tissue into the culture medium (20). In addition, the

duration of the culture influences the level of secreted proteins that accumulate in the

medium, but may also affect the function and breakdown of cells in the tissue. In view of

secretome complexity and considering that adipokines are expected in low concentrations

(ng/ml), the quality of the sample obtained from tissue culture is crucial to obtain high

quality, relevant secretome data. Specific removal of high-abundance serum proteins in

biological samples has been previously described (21), but does not resolve the problem of

contamination derived from intracellular proteins. Besides, low-abundance proteins that bind

to the high-abundance ones may also be removed at the same time. Therefore, we did not

consider this option but carefully evaluated the influence of the culture set-up on the quality

of the sample for secretome analysis. For this purpose, the influence of several culture set-ups

that varied in the number of washes and the distribution of washes in-time was investigated.

Protein concentration, dynamic range and composition of the resulting samples were then

evaluated. From these experiments the best culture set-up was chosen for further secretome

characterization. Proteins were then identified by LC-MS/MS after SDS-PAGE fractionation.

The resulting list of secreted proteins was validated by culturing adipose tissue in the

presence of stable isotope-labelled L-lysine. Proteins that contain a signal peptide sequence

and incorporate the label are derived from adipose tissue and not from an external source

such as serum.

Experimental Procedures

Adipose tissue culture

Human visceral adipose tissue explants were obtained from five women (age 25-64; BMI 19-

31) undergoing surgery for non-carcinogenic gynaecologic disorders. For each culture set-up

(see figure 1), adipose tissue from one subject was used. The study had the approval of the

local ethical committee.

Chapter 2

67

The adipose tissue culture protocol is based on Fried et al. (22). Briefly, adipose tissue

explants were transported from the operating room to the laboratory in transport buffer (PBS,

5.5 mM glucose, 50 µg/ml gentamicin) at room temperature. The following procedures were

carried out under a laminar flow hood, using sterile equipment. Immediately upon arrival, the

tissue was transferred to a Petri dish containing 20 ml of PBS and was finely minced in 20 to

80 mg pieces using scissors. The tissue pieces were extensively washed with 400 ml of PBS

over a filter containing sterile cotton bandage fabric. Thereafter, the tissue pieces were

transferred to a 75 cm2 culture flask containing 200 ml of PBS and were gently shaken for a

short period. Next the contents of the flask were poured over the filter and the tissue pieces

were washed with 300 ml of warm PBS (37°C). The tissue pieces were transferred to a tube

containing 50 ml of PBS and centrifuged for 1 minute at 277 g at room temperature to

remove red blood cells and debris. The tissue was then removed from the tube and the weight

was determined. 1 g of tissue was then placed in a Petri dish with 10 ml M199 (Gibco)

culture media supplemented with 50 µg/ml gentamicin. Adipose tissue pieces were cultured

at 37 ºC and 5% CO2 following the different culture set-ups (figure 1). The final secretome

sample and the media collected after every washing step in culture (figure 1) were stored at -

80 ºC.

Sample pre-treatment

Nine to ten ml adipose tissue culture medium was concentrated 25 to 30 fold by ultra-

filtration (Centriplus, 3 kDa cut-off, Millipore). The concentrated sample was used for total

protein concentration measurement (Bradford assay, Bio-Rad), SELDI-TOF-MS profiling

and protein identification by LC-MS/MS.

SELDI-TOF-MS protein profiling

Spots of a CM10 (weak cation exchanger) ProteinChip® array (Ciphergen Biosystems,

Fremont, CA), inserted into a bioprocessor, were pre-incubated twice with 200 µl binding

buffer (100 mM ammonium acetate, 0.05% Triton, pH 4.0) for five minutes at room

temperature with vigorous shaking. The buffer was then removed and 4-12 µl concentrated

sample (depending on protein concentration) was applied on every spot. Binding buffer was

added to a total volume of 100 µl per well (µg protein was adjusted so that the same protein

amount per spot was applied). After incubation for 30 minutes, the sample was removed and

the spots were washed three times with 200 µl binding buffer for five minutes followed by a

wash with 200 µl ultra pure water. The water was removed and the chip was allowed to air-

Chapter 2

68

dry before applying two times 0.5 µl of 5 mg α-cyano-4-hydroxy cinnamic acid (CHCA)

diethylamine salt dissolved in 1 ml of 50% ACN, 0.5% TFA. Mass analysis was performed

in a ProteinChip® reader (model PBS II, Ciphergen Biosystems) according to an automated

data collection protocol. Calibration was performed with the All-in-One peptide mix

(Ciphergen Biosystems) and spectra were obtained in the mass range 3.5-100 kDa at several

laser intensities (175-225).

Label experiment

An adipose tissue explant from a patient undergoing surgery for uterus myomatosus (47 years

old, BMI 22.9) was collected, cut, washed and divided into two Petri dishes as described

above. In this experiment, tissue was cultured from the start in lysine-free M199 media (ref.

22340 Lys-free, Gibco) in order to deplete lysine from other sources (serum). The media was

renewed after 1h, 21.5h, 25.5h and 29.5h. In the last wash (time point 29.5h), one dish

received normal M199 containing 70 mg/l of 12C6, 14N2 L-lysine and 60 nM insulin. The

other dish received M199 lysine-free media contained 70 mg/l labeled lysine (L-Lysine:

2HCl, U-13C6, 98%; U-15N2, 98%, Cambridge Isotope Laboratories, Inc., Andover MA, USA)

and 60 nM insulin. Tissue was maintained in culture for an additional 72 hrs. Thereafter,

media were collected and stored at -80 ºC until analysis. Normal and labeled media were then

mixed in 1:2 ratio and were concentrated by ultra-filtration before SDS-PAGE fractionation.

Protein identification by LC-MS/MS

Proteins present in the concentrated adipose tissue media sample were fractionated by SDS-

PAGE on a 4-12% Bis-Tris gel with a MOPS buffer system, according to manufacturer

protocol (NuPAGE®-Novex, Invitrogen, Carlsbad, CA, USA). Protein separation occurred

for 50 minutes at 200 V and visualization of bands was performed overnight by Coomassie

Brilliant Blue G-250 based staining (PageBlue Staining Solution, Fermentas). The whole lane

was excised into 28 bands which were processed for tryptic digestion. Each band was cut into

small pieces and stored at -20 ºC until analysis. Then, they were washed in ultra pure water

and dehydrated in ACN. In-gel reduction with dithiothreitol (for one hour at 60 ºC) and

carbamidomethylation with iodoacetamide (for 45 minutes at room temperature in the dark)

were performed. Gel pieces were subsequently washed with ultra pure water, 50% ACN and

pure ACN. Next, 0.1 µg trypsin in 50 mM ammonium bicarbonate was added and gel pieces

were allowed to rehydrate on ice for 20 minutes. Digestion was carried out overnight at

37 ºC.

Chapter 2

69

Separation of the resulting tryptic peptide mixtures was performed by nanoscale reversed-

phase liquid chromatography tandem mass spectrometry (LC-MS/MS). The Agilent 1100

nanoflow/capillary LC system (Agilent, Palo Alto, CA, USA) was equipped with a trapping

column (5 x 0.3 mm C18RP) (Dionex/LC Packings, Amsterdam, The Netherlands) and a

nanocolumn (150 x 0.075 mm, C18Pepmap) (Dionex/LC Packings). Peptides mixtures were

injected into the trapping column at a flow rate of 10 µl/min (3%ACN/0.1%FA). After 10

minutes the trapping column was switched into the nano flow system and the trapped

peptides were separated using the nano column at a flow rate of 0.3 µl/min in a linear

gradient elution from 95%A ( 3%ACN/0.1%FA) to 50%B (97%ACN/0.1%FA) in

50 minutes, followed by an increase up to 80% B in 3 minutes. The eluting peptides were on-

line electro-sprayed into the QStar XL Hybrid ESI Quadrupole time-of-flight tandem mass

spectrometer, ESI-qQTOF-MS/MS (Applied Biosystems, Framingham, MA;

MDSSciex,Concord, Ontario, Canada) provided with a nano spray source equipped with a

New Objective ESI needle (10 µm tip diameter). Typical values for needle voltage were 2 kV

in positive ion mode. Analyst QS 1.1 software (Applied Biosystems) was used for data

acquisition in the positive ion mode, typically with a selected mass range of 300-1200 m/z.

Peptides with +2 to +4 charge states were selected for tandem mass spectrometry, and the

time of summation of MS/MS events was set to be 2 seconds. The three most abundant

charged peptides above a 40 count threshold were selected for MS/MS and dynamically

excluded for 40 seconds with 100 ppm mass tolerance.

ProID 1.1 software (Applied Biosystems) (23) was used to identify proteins from the mass

spectrometric datasets according to Swiss-Prot database (May 2005, ~181000 entries). Mass

tolerance was set to 0.15 Da (MS) and 0.1 Da (MS/MS) and carboxamidomethylation and

methionine oxidation were chosen as modifications for database search.

Classification of identified proteins in terms of secretion pathways was performed according

to SecretomeP 2.0 Server (24). Those proteins with a signal peptide predicted by SignalP

were considered as secreted proteins via a classical pathway (Endoplasmic Reticulum/Golgi-

dependent pathway). If no signal peptide was predicted but the NN-score exceeded 0.6 value,

proteins were classified as secreted via non-classical pathway. Trans-membrane helices and

location were predicted according to TMHMM Server (25). MS spectra of identified peptides

which showed a lysine in the C-terminus were searched for shifts of 8, 4 or 2.666 m/z (singly,

doubly or triply charged ions, respectively). If a peptide incorporated the label, the derived-

protein was considered to be synthesized by adipose tissue.

Chapter 2

70

Results

Evaluation of culture set-up and sample quality

The standard adipose tissue culture protocol involves cutting of the tissue explants into small

pieces followed by several washing steps to remove serum and intracellular proteins before

culturing, as described in the Experimental Procedures section. Because of the cutting,

damaged cells will slowly lose their contents into the media. Furthermore, serum proteins still

present in the tissue pieces will diffuse out during culture. Therefore, additional washing

steps during culture were necessary to obtain a sample for secretome analyses containing

mainly adipose tissue-derived secreted proteins. In preliminary studies we evaluated the

protein composition of the adipose tissue media, cultured for 48 hrs with one washing step

after the second hour of culture (figure 1, set-up A). After in-solution digest of the final

sample and LC-MS/MS analysis, 42 proteins were identified with ≥95% confidence (results

not shown). According to the Human Protein Reference Database (HPRD) (26) none of them

could be directly related to a protein secreted by adipose tissue. Typical serum proteins

(albumin, hemoglobin and transferrin) and intracellular proteins (actin, histones and perilipin)

dominated the secretome composition.

To reduce the concentration of these high-abundance contaminating proteins, five different

culture set-ups were evaluated including the set-up mentioned above as control (see figure 1).

Chapter 2

71

Figure 1. Distribution of washing steps in time for the different culture set-ups assayed. An insert is included for visualizing the distribution of washes at the beginning of the culture period (first 6h).

The number of washes (replacement with fresh media) during tissue culture, as well as their

distribution in time, were varied in order to evaluate effects on the quality of the final sample

for secretome analyses. Processing of the visceral adipose tissue explants, obtained from the

five patients (A-E), was the same (described in the experimental section). During culture,

washing steps were performed by replacing media (10 ml) with 10 ml of fresh media,

according to the different schemes (A-E) depicted in figure 1. The media of the washing steps

and the final secretome samples were collected for further analyses. Protein concentration

was measured to evaluate the effectiveness of the different culture set-ups on removing (high-

abundance) serum and intracellular proteins. According to figure 2, protein concentrations in

the final media samples were the lowest in set-up D and E. This suggests set-ups D and E

were better in removing high-abundance proteins when compared to A and B while set-up C

gave an intermediate result.

0 20 40 60 80 100 120 140

Time (h)

A

B

C

D

E

0 1 2 3 4 5 6

Time (h)

Chapter 2

72

Figure 2. Total protein concentration measured in media collected from every washing step and in the final secretome sample. In x-axis, consecutive time points are displayed per experiment and every time point represents the culture time since the previous wash (periods of accumulation). Bars in black represent the initial protein concentrations and bars in grey refer to the final protein concentrations for every set-up. In figure 1, it is indicated the time points when the samples were collected.

The protein composition of the final samples obtained from each culture set-up was

investigated by SDS-PAGE (figure 3) and SELDI-TOF-MS (figure 4). When comparing the

lanes with samples of the indicated culture set-ups (figure 3), it was clear that samples from

set-ups A and B showed less uniformity in intensity of bands than those of set-ups C, D and

E.

0

5

10

15

20

25

30

2h 48h 1h

16.5

h

74h

30m

30m

30m

30m

30m

30m

48h 1h 21h 8h 48h 1h 15h 4h 4h

113.

5h

A B C D E

Pro

tein

co

ncen

tra

tion

(µg/

ml)

Chapter 2

73

Figure 3. SDS-PAGE (12% Bis-Tris gel) of final secretome samples obtained from the different culture set-ups as indicated in figure 1 (m: marker; A-E: final secretome samples).

We also used SELDI profiling to monitor changes in the dynamic range of the samples,

which are expected to occur if high-abundance contaminating proteins are removed more

efficiently. High-abundance proteins are likely to suppress ionization of low-abundance

proteins (27). As a consequence, with a certain laser energy, a lower total number of peaks

can be expected in a sample with large concentration differences compared to a sample with

small differences in concentration of individual proteins. We made use of this phenomenon to

evaluate the dynamic range of the final samples resulting from the different culture set-ups

(A-E). For this, spectra were obtained for every sample at four different laser intensities in

the mass range from 3.5 to 17.5 kDa. Figure 4.1 shows an example of spectra obtained for the

different set-ups with the lowest laser intensity (175). From these spectra it can already be

deduced that samples of set-up D and E showed considerable more peaks than samples from

the other three set-ups at this laser intensity. Total peak number (S/N>5) was also calculated

from all spectra and plotted against laser intensity (figure 4.2). At laser intensity 190 still no

Chapter 2

74

maximum is reached in the number of detected peaks for set-ups A, B and C, while the

number of peaks that could be detected with set-ups D and E plateaus at laser intensities 180

and 185, respectively. This indicates that the dynamic range of sample D and E is lower than

that of the other samples. We also monitored peak intensities of the α and β chain of

haemoglobin (predicted to be appearing at 15.1 and 15.9 kDa, respectively) and albumin (66

kDa) in the SELDI spectra which clearly showed a reduction in peak intensity (abundance) in

spectra from sample D and E compared to A, B and C (figure 4.3).

When results of these analyses are combined we conclude that one or two washing steps as

used in set-up A and B before isolation of the final media sample is clearly not sufficient. Set-

up D and E contained 3 or 4 washing steps, respectively, distributed over a period of 24 hrs

before the final sample was obtained. This clearly gave better results, also when compared to

set-up C. Here, six washing steps were applied but all in the beginning of the tissue culture

period. Based on these analyses we conclude that culture protocols D and E performed best.

Chapter 2

75

Figure 4. (4.1) SELDI-TOF-MS profiles (CM10 weak cation exchange chip) in the 3.5-17 kDa mass range at the lowest laser intensity (175); (4.2) graph of total peak number plotted against laser intensity by SELDI-TOF-MS, as indication of the dynamic range of final secretome samples; (4.3) SELDI-TOF-MS spectra zoomed-in at the mass range of hemoglobin and albumin (laser intensities 185 and 225, respectively) for the five set-ups.

A

B

C

D

E

5000 7500 10000 12500 15000

0

1

2

0

1

2

0

1

2

0

1

2

0

1

2

Mass (Da)

Pea

k in

tens

ity

0

1

2

0

1

2

0

1

2

0

1

2

0

1

2

4.1

0

5

10

15

20

25

170 175 180 185 190 195

laser intensity

tota

l pea

k nu

mb

er (

S/N

>5)

B

CD

E

A

4.2

A

B

C

D

E

14000 15000 16000

Pea

k in

tens

ity

Mass (Da)

Hemoglobin

0

1

2

00.5

11.5

00.250.500.75

00.10.20.3

00.250.500.75

15000

α β

50000 60000 70000 80000

0246

0246

0246

0246

0246

A

B

C

D

E

Mass (Da)

Pea

k in

tens

ity

4.3Albumin

Chapter 2

76

Secretome characterization

After the optimal culture set-up was established, this protocol was used for characterization

of the secretome. To determine whether identified proteins are secreted by adipose tissue or

are derived from serum that may still be present as a contaminate of the sample, a labelling

experiment was carried out. In this qualitative approach, an adipose tissue explant was

divided into two dishes. After the washing steps were performed in lysine-free media (see

Experimental Procedures section), the tissue was cultured for an additional 72 hrs. One dish

contained media supplemented with 13C6,15N2 L-lysine and the other dish contained normal

media containing 12C6,14N2 L-lysine. To facilitate detection of label incorporation by mass

spectrometry, media from both dishes were mixed in a 1:2 ratio in favour of the stable isotope

label. After concentration, the sample was fractionated by SDS-PAGE. The lane was excised

in 28 bands for tryptic digestion and digests were analysed by LC-MS/MS for protein

identification as described in the experimental section. The complete list of identified

proteins and peptide sequences obtained from the MS/MS data can be found in the

supplemental data file C. Peptide confidences > 5 are shown for each identified protein.

Confidence scores generated by the Pro ID software are explained in a paper by Tang et al.

(23). By this approach, a total of 297 proteins could be identified with ≥95% confidence.

Within this group, 259 proteins were identified with ≥99% confidence. Proteins, identified

with the highest confidence score (≥99%), were analysed for the presence of a signal peptide

using the SecretomeP 2.0 Server to determine whether they were secreted. This analyses

reveals that 108 out of 259 proteins were secreted following a classical pathway

(Endoplasmic Reticulum/Golgi-dependent pathway) (see tables 1 and 2). For all these

proteins a signal peptide was predicted by SignalP. Some of them (marked with an asterisk in

tables 1 and 2) contained trans-membrane helices located more than 40 amino acids away

from the N-terminus according to the TMHMM Server. This implies that they were anchored

to the membrane. However, because a signal peptide was predicted we considered them as

secreted. The possibility exists that they were released from the cell via cleavage of the extra-

cellular part of the protein. Proteins identified with ≥99% confidence that contained a signal

peptide were classified according to whether or not they incorporated the label (tables 1 and

2). Out 108 proteins containing a signal peptide, label incorporation could be confirmed for

70 of them (see table 1). These proteins were considered genuine adipose tissue secreted

proteins. For this, the MS spectra of the corresponding peptides were manually checked for

mass shifts due to label incorporation as described in the experimental section. Some proteins

Chapter 2

77

were only identified by peptides ending on an arginine at the C-terminus (no lysine present).

This leaves open the possibility that they are derived from adipose tissue although the label

incorporation could not be detected (see first group of non-labeled proteins in table 2).

Table 1. Secreted proteins (classical pathway; signal predicted by SignalP) that incorporated the label. They were identified after SDS-PAGE fractionation and tryptic digest analysis by LC-MS/MS (≥99% confidence).

ACCESSION NO. PROTEIN NAME

LABELED PROTEINS

Signalling / Regulatory function

gi|14916999 78 kDa glucose-regulated protein (GRP 78) (Immunoglobulin heavy chain binding protein)

gi|20137531 Adipocyte-derived leucine aminopeptidase (A-LAP) (ARTS-1) (Aminopeptidase PILS)

gi|2493789 Adiponectin §┼#

gi|117501 Calreticulin (CRP55) (Calregulin) §

gi|1171064 Cell surface glycoprotein MUC18 (Melanoma-associated antigen MUC18) *

gi|20177861 Complement C1q tumor necrosis factor-related protein 5

gi|23396772 Ectonucleotide pyrophosphatase/phosphodiesterase 2 (E-NPP 2)

gi|17865698 Endoplasmin (94 kDa glucose-regulated protein) (GRP94)

gi|23396609 Insulin-like growth factor binding protein 7 (IGFBP-7) (IBP-7) (IGF-binding protein 7) §

gi|117558 Macrophage colony stimulating factor-1 (CSF-1) (MCSF) (M-CSF) *

gi|9297107 Neuropilin-1 (Vascular endothelial cell growth factor 165 receptor) *

gi|118090 Peptidyl-prolyl cis-trans isomerase B (PPIase) (Rotamase) (Cyclophilin B)

gi|46576887 Periostin (PN) (Osteoblast-specific factor 2) (OSF-2)

gi|3024715 Peroxiredoxin 4 (Prx-IV) (Thioredoxin peroxidase AO372)

gi|20178323 Pigment epithelium-derived factor (PEDF) (EPC-1) §#

gi|124096 Plasma protease C1 inhibitor (C1 Inh) §

gi|62298174 Plasma retinol-binding protein (PRBP) (RBP) §

gi|129576 Plasminogen activator inhibitor-1 (PAI-1) (Endothelial plasminogen activator inhibitor) §

gi|401413 Von Willebrand factor precursor (vWF)

ECM

gi|24212664 Basement membrane-specific heparan sulfate proteoglycan core protein (HSPG) (Perlecan)

gi|115269 Collagen alpha 1(I) chain #

gi|115306 Collagen alpha 1(III) chain ┼#

gi|13878903 Collagen alpha 1(VI) chain ┼#

gi|62901508 Collagen alpha 1(XIV) chain (Undulin)

gi|728996 Collagen alpha 1(XV) chain

gi|45644997 Collagen alpha 1(XVIII) chain

gi|8039779 Collagen alpha 2(I) chain #

gi|115349 Collagen alpha 2(IV) chain §#

gi|27808647 Collagen alpha 2(VI) chain ┼#

gi|5921193 Collagen alpha 3(VI) chain ┼

gi|62510689 EGF-containing fibulin-like extracellular matrix protein 1 (Fibulin-3)

gi|2506872 Fibronectin (FN) (Cold-insoluble globulin)

gi|47115668 Galectin-3 binding protein (Lectin galactoside-binding soluble 3 binding protein)

gi|121116 Gelsolin (Actin-depolymerizing factor) (ADF) (Brevin) §┼# gi|20141592 Laminin alpha-4 chain

gi|126366 Laminin beta-1 chain (Laminin B1 chain) §

Chapter 2

78

gi|126369 Laminin gamma-1 chain (Laminin B2 chain) §

gi|62298084 Matrilin-2

gi|2506403 Microfibril-associated glycoprotein 4

gi|128199 Nidogen (Entactin) §┼

gi|52783472 Spondin-1 (F-spondin) (Vascular smooth muscle cell growth promoting factor)

gi|3915888 Tenascin (TN) (Hexabrachion) (Cytotactin) (Neuronectin) (GMEM)

gi|135717 Thrombospondin-1 §

gi|549136 Thrombospondin-2

gi|2498193 Transforming growth factor-beta induced protein IG-H3 (Beta IG-H3)

gi|2506816 Versican core protein (Large fibroblast proteoglycan) Immune function

gi|115205 Complement C1s subcomponent (C1 esterase) §

gi|38257345 Complement C2 (C3/C5 convertase) §

gi|116594 Complement C3 ┼#

gi|20141171 Complement C4 §

gi|61252057 Complement component C7

gi|584908 Complement factor B (C3/C5 convertase) (Properdin factor B) §

gi|3915626 Complement factor D (C3 convertase activator) (Properdin factor D) (Adipsin) §┼#

gi|48428995 Lysozyme C (1,4-beta-N-acetylmuramidase C)

Involved in degradation

gi|116856 72 kDa type IV collagenase (Matrix metalloproteinase-2) (MMP-2) §┼#

gi|112911 Alpha-2-macroglobulin (Alpha-2-M)

gi|115711 Cathepsin B (Cathepsin B1) (APP secretase) §

gi|115717 Cathepsin D §

gi|115741 Cathepsin L (Major excreted protein) (MEP)

gi|544413 Chitinase-3 like protein 1 (Cartilage glycoprotein-39)

gi|116852 Interstitial collagenase (Matrix metalloproteinase-1) (MMP-1) (Fibroblast collagenase)

gi|116863 Matrix metalloproteinase-9 (MMP-9) (92 kDa type IV collagenase) (92 kDa gelatinase) §

gi|135850 Metalloproteinase inhibitor 1 (TIMP-1) (Erythroid potentiating activity) (EPA) §

gi|6919941 Procollagen C-proteinase enhancer protein (PCPE) #

Other functions

gi|119576 Liver carboxylesterase 1 (Acyl coenzyme A:cholesterol acyltransferase)

gi|585223 Plasma glutathione peroxidase (GSHPx-P)

gi|61221730 Protein C19orf10 (Stromal cell-derived growth factor SF20) (Interleukin-25) #

gi|2507461 Protein disulfide-isomerase A3 (Disulfide isomerase ER-60) (ERp60)

gi|2507460 Protein disulfide-isomerase (PDI) (Prolyl 4-hydroxylase beta subunit)

gi|136191 Serotransferrin (Transferrin) (Siderophilin) (Beta-1-metal binding globulin) § * Signal peptide predicted by SignalP (24) but also transmembrane helices predicted by TMHMM Server (25) (more than one helices or one helix located more than 40 amino acids away from the N-terminus of the protein). § Secreted proteins also identified in (19). # Secreted proteins also identified in (18). ┼ Secreted proteins also identified in (17).

Chapter 2

79

Table 2. Secreted proteins (classical pathway; signal predicted by SignalP) that did not incorporate the label. They were identified after SDS-PAGE fractionation and tryptic digest analysis by LC-MS/MS (≥99% confidence).

* Signal peptide predicted by SignalP (24) but also transmembrane helices predicted by TMHMM Server (25) (more than one helices or one helix located more than 40 amino acids away from the N-terminus of the protein). § Secreted proteins also identified in (19). # Secreted proteins also identified in (18). ┼ Secreted proteins also identified in (17).

ACCESSION NO. PROTEIN NAME

NON-LABELED PROTEINS. IDENTIFICATION BASED ON ONLY –ARG ENDING PEPTIDES

gi|12643637 ADAMTS-4 (A disintegrin and metalloproteinase with thrombospondin motifs 4)

gi|416746 Azurocidin (Cationic antimicrobial protein CAP37) (Heparin-binding protein)

gi|23396490 Calsyntenin-1 *

gi|12643324 Cathepsin Z gi|116533 Clusterin (Complement-associated protein SP-40,40) (Complement cytolysis inhibitor) §

gi|3023630 Cystatin C §┼

gi|462007 Dermatopontin (Tyrosine-rich acidic matrix protein) (TRAMP)

gi|37537873 EMILIN 1 (Elastin microfibril interface-located protein 1)

gi|134635 Extracellular superoxide dismutase [Cu-Zn] (EC-SOD) # gi|30581038 Fibulin-1

gi|41017299 Latent transforming growth factor-beta-binding protein 2 (LTBP-2)

gi|126279 Leukemia inhibitory factor (LIF) (Differentiation-stimulating factor) (D factor)

gi|119292 Leukocyte elastase (Neutrophil elastase) (PMN elastase) (Bone marrow serine protease)

gi|20141203 Monocyte differentiation antigen CD14 (Myeloid cell-specific leucine-rich glycoprotein) §

gi|129825 Myeloperoxidase (MPO)

gi|8928569 Nidogen-2 (NID-2) (Osteonidogen) gi|50400889 Olfactomedin-like protein 1

gi|1346908 Pentraxin-related protein PTX3 (Pentaxin-related protein PTX3)

gi|62900717 Procollagen-lysine,2-oxoglutarate 5-dioxygenase 3 (Lysyl hydroxylase 3) (LH3)

gi|41017497 Prostaglandin-H2 D-isomerase (Lipocalin-type prostaglandin-D synthase)

gi|2501205 Protein disulfide-isomerase A6 (Protein disulfide isomerase P5)

gi|129283 SPARC (Secreted protein acidic and rich in cysteine) (Osteonectin) §#┼ gi|52783469 Spondin-2 (Mindin)

gi|3334154 Stanniocalcin 1 (STC-1)

gi|1351316 Tumor necrosis factor-inducible protein TSG-6 (TNF-stimulated gene 6 protein)

gi|13432109 Vascular endothelial-cadherin (VE-cadherin) (Cadherin-5) (7B4 antigen) (CD144 antigen) *

NON LABELED PROTEINS. IDENTIFICATION BASED ON –LYS ENDING PEPTIDES gi|57015285 Citrate synthase, mitochondrial precursor gi|3182940 Collagen alpha 1(XII) chain

gi|12643876 Fibulin-5 (FIBL-5) (Developmental arteries and neural crest EGF-like protein)

gi|6175096 Lactotransferrin (Lactoferrin)

gi|1708851 Laminin beta-2 chain (S-laminin) (Laminin B1s chain)

gi|20141464 Lumican (Keratan sulfate proteoglycan lumican) (KSPG lumican)

gi|2501336 Membrane copper amine oxidase (Vascular adhesion protein-1) (VAP-1) gi|129078 Mimecan (Osteoglycin) (Osteoinductive factor)

gi|1171700 Neutrophil gelatinase-associated lipocalin (NGAL) (P25) ┼

gi|51701718 Plexin B2 (MM1)

gi|113576 Serum albumin §

gi|9087217 Tenascin-X (TN-X)

Chapter 2

80

As an example, figure 5 shows two MS spectra obtained for two different proteins,

endoplasmin (figure 5A) and PAI-1 (figure 5B). The mass shift was clearly seen in both cases

(+4 m/z), although the different relative peak intensities for the non-labeled and labeled

peptides points to different incorporation rates of the label for these two proteins. Figure 5C

shows the MS/MS spectra from both the original and the labeled peptides from PAI-1, where

also a mass shift of +8 m/z was obvious.

Figure 5. Examples of different incorporation ratio in labeled proteins. (A) TOF-MS spectra of endoplasmin peptide; (B) TOF-MS spectra of PAI-1 peptide; (C) MS/MS spectra of both original and labeled peptide in PAI-1.

Secreted proteins were classified into groups with similar functions (table 1). For this,

information was obtained from Swiss-Prot and SOURCE (http://source.stanford.edu)

databases. The criteria for classification are not always uniform since many proteins have

more than one function which may place them in more than one functional category. In this

case, what was considered as the main function of the protein was used for classification in

ELISNASDALDK

638 640 642 644 646

m/z

0

20

40

60

80

100

120

140

160

Inte

nsity

, co

unts

638.35

638.85

639.36642.36

642.86639.86643.36

517.0 519.0 521.0 523.0

m/z

0

40

80

120

160

200

240

280

Inte

nsity

, co

unts

521.81

517.81

522.32518.31

522.82518.81

523.31519.31

FIINDWVK

A

B100 200 300 400 500 600 700 800

m/z

0

10

20

30

40

50

Inte

nsi

ty, c

ount

s

233.17

120.08

782.45

261.17 669.36343.15254.21 501.26

529.27

100 150 200 250 300 350 400 450 500 550 600 650 700 750 800m/z

2.0

6.0

10.0

14.0

18.0

22.0

Inte

nsi

ty, c

ount

s

233.17

120.08

774.45343.18

261.16 661.39501.29529.26

C

Chapter 2

81

one of the five categories: Signalling/regulatory, Extra-cellular matrix, Immune function,

Degradation and Other.

In supplementary data table A, non-secreted proteins (no signal peptide predicted by SignalP

according to SecretomeP) and non-classically secreted proteins (no signal peptide predicted

by SignalP but NN-Score>0.6, marked as °) are shown. In supplementary data table B,

classically secreted proteins identified with confidence between 95% and 99% are shown.

Discussion

Adipose tissue is recognized as an important organ for the regulation of the whole-body

energy metabolism through the secretion of adipokines. Although, major adipokines such as

leptin and adiponectin have been shown to be produced by adipocytes, also other cell types

that are part of the adipose tissue produce adipokines or influence production of adipokines

by adipocytes. Fain et al. (28) showed that the majority of adipokines, measured in their

study, was released by non-adipocyte cells in the tissue. This fact points out the relevance of

evaluating the adipose tissue secretome rather than the adipocyte cell secretome. In this paper

we describe the first proteomics study on the adipose tissue secretome. A major issue in

characterizing the adipose tissue secretome is that protein composition of the adipose tissue

culture media is highly dependent on the way the tissue culture is performed. This not only

has implications for proteomics studies but also for other studies where adipose tissue culture

is used to study individual adipokines by ELISA, since these peptide and protein hormones

are also present in serum and in the intracellular protein fraction which, as we demonstrate,

are the main sources of secretome contamination in adipose tissue culture. Therefore, we

established a tissue culture protocol that minimizes contamination of the secretome with

serum-derived and intracellular proteins. After analyses of five different culture set-ups we

conclude that a tissue culture protocol with one wash after the first hour in culture and two or

three additional washes after overnight culture within a period of eight hours followed by a

48 to 114 hrs incubation period (set-ups D and E, figure 1) provides the optimal culture

protocol to obtain a high quality sample for secretome analyses. With this set-up, total protein

concentration was reduced from approximately 17 µg/ml to around 4 µg/ml (set-ups D and E)

at the end of the culture period (figure 2). This reduction is considerably higher than that

observed for set-ups A and B (from about 25 µg/ml to about 18 µg/ml). Albumin and

hemoglobin (serum contaminating proteins) seemed to be highly reduced in set-ups D and E

Chapter 2

82

(figure 4.3) and the dynamic range of the sample has decreased considerably (figures 3 and

4.2) compared to set-ups A and B.

This new improved culture protocol was used to characterize the human adipose tissue

secretome. Proteins were identified and the presence of a signal peptide was investigated in

order to distinguish those that are secreted from intracellular proteins in the media, derived

from damaged cells. Furthermore, a metabolic labelling approach was designed to

differentiate between secreted proteins derived from adipose tissue and serum proteins that

still may be present in low levels as a contaminant of the sample. Proteins that contained a

signal peptide and incorporated the label were considered as genuinely secreted by adipose

tissue and not coming from an external source (serum, intracellular origin). With this

strategy, different incorporation rates of the label into the proteins were noticed, which can be

related to different turn-over rates. In summary, a total of 297 proteins were identified with

≥95% confidence. 259 proteins were identified with ≥99%, among which 108 secreted

proteins (tables 1 and 2). Out of 108 secreted proteins, 70 proteins incorporated the label

(table 1), i.e. adiponectin, adipsin, gelsolin, macrophage colony stimulating factor-1 (M-

CSF), pigment epithelium-derived factor (PEDF), plasma retinol binding protein (RBP),

plasminogen activator inhibitor-1 (PAI-1), among others. Interleukin-6 was also detected,

showing a clear incorporation of the label, although identified with ≥95% confidence,

(supplemental data, table B). Adipocyte fatty acid-binding protein (A-FABP) was identified

as non-classically secreted protein (supplemental data, table A) with label incorporation,

agreeing with a recent study by Xu et al. (29) where they claimed that, A-FABP is a

circulating biomarker released from adipocytes into the bloodstream, closely associated with

obesity and metabolic syndrome. For 38 proteins containing a signal peptide the label could

not be detected (table 2). However, 26 of these were identified with peptides ending on an

arginine in the C-terminus (no lysine present), which implies that it is still possible that these

proteins are secreted by adipose tissue although label incorporation could not be confirmed

(i.e. leukemia inhibitory factor (LIF), pentraxin-related protein PTX3, SPARC). Finally, the

label could not be detected in a group of 12 proteins, even though their peptides showed a

lysine in the C-terminus. This fact may be due to a low incorporation rate or because the

protein was derived from serum, e.g. albumin is part of this latter group (table 2).

In supplemental data, table A, non-secreted proteins (intracellular) or proteins predicted to be

secreted via a non-classical pathway are shown (151 proteins in total). Non-secreted proteins,

such as actin, histones, catalase and proteasome subunits, are intracellular proteins that may

Chapter 2

83

be derived from damaged cells in culture or from cleaved fragments of membrane proteins.

This, together with the fact that albumin was still detected, indicates that although with the

improved tissue culture set-up sample quality has improved considerably, it was not

completely free of (high-abundance) serum and intracellular proteins. Nevertheless, the

dynamic range of the sample improved considerably allowing a much more sensitive

secretome identification. The level of sensitivity that was reached with the described

procedure was in the low ng/ml range since adiponectin, which was identified with ≥99%

confidence, reached concentrations of around 10-20 ng/ml at the end of the culture period, as

determined by ELISA in similar experiments.

Three studies have been published thus far on the adipocyte cell secretome. We compared the

results of those studies with our results on the human visceral adipose tissue secretome. Chen

et al. (19) identified 84 proteins secreted by isolated rat adipocytes cells using 2DLC-

MS/MS. Of these proteins 29 were also identified in the present study (as indicated in tables

1 and 2). Wang et al. (18) identified 27 proteins that were secreted during mouse 3T3-L1

adipocyte cell differentiation by 2DE-MS(/MS). 16 proteins were also identified in our study

(as indicated in tables 1 and 2). Out of 20 secreted proteins identified by Kratchmarova et al.

during differentiation of 3T3-L1 preadipocytes by SDS-PAGE and LC-MS/MS (17), 13 were

also found in this study (as indicated in tables 1 and 2). Only five proteins are shared between

the three adipocyte cell studies and the present study on adipose tissue. These proteins are

adiponectin, SPARC, gelsolin, adipsin, and MMP-2. 68 proteins (classically secreted)

identified in the present study were not found in the other three studies. This may be

explained by the different starting materials (cells or tissue), the different proteomics

approaches that were followed and the different origins of the material (rat, mouse, human).

As mentioned before, although the adipocyte is the major cell type in adipose tissue, this

study shows that a substantial number of secreted proteins by the tissue are released by other

cell types such as macrophages and endothelial cells. In particular, some of the secreted

proteins identified in the present study were also identified in human macrophages by a 2D

gel approach (30). That is the case for endoplasmin, gelsolin, protein disulfide isomerase,

protein disulfide isomerase A3, calreticulin, cathepsin D and peroxiredoxin 4, among others.

In the same way, pentraxin-related protein 3 (PTX3) mRNA is expressed in the stromal-

vascular fraction of adipose tissue but not in fully differentiated adipocytes; it plays a role in

the regulation of resistance to pathogens and inflammatory reactions and the PTX3 gene can

be induced in adipocytes by TNF-α (31). Taken into account what has been reported

Chapter 2

84

previously (1, 7, 32-35) on proteins secreted by adipose tissue, we estimate that in this study

we identified 48 new proteins as being secreted by adipose tissue.

The functional classification of the identified secreted proteins, as shown in table 1, indicates

that 39% of the proteins are involved in the modulation of the extra-cellular matrix. An

example from this group is perlecan. It is a secreted proteoglycan, widely distributed as part

of the basement membrane that may bind lipoprotein lipase (LPL) and is localized close to

the cell surface, so that it can participate in triacylglycerol hydrolysis (36). Another example

from this group is versican core protein which plays a role in intercellular signaling and in

connecting cells with the extra-cellular matrix. It may also take part in the regulation of cell

motility, growth and differentiation (37). The second largest group (27%) consists of proteins

involved in signaling and regulation. Macrophage colony stimulating factor (MCSF) is part

of this group. It promotes human adipose tissue hyperplasia, it is up-regulated under

conditions favouring adipose tissue growth (obesity) and it is down-regulated by TNF-α (38).

Another example is neuropilin-1, which is expressed by adipocytes and is involved in

regulation of angiogenesis (39). Proteins classified as being involved in degradation (14%)

are e.g. cathepsin B, and D. These are lysosomal proteases but they can also be secreted. The

fact that a large part of the identified proteins is involved in the modulation of extra-cellular

matrix, protein degradation and regulation of cellular processes indicates that adipose tissue

is a very actively dividing tissue which is probably related to the demand to store energy in

the form of triglycerides. Therefore, the tissue has to be flexible to increase or decrease

storage capacity.

In conclusion, adipose tissue culture set-up has strong influence on the quality of the sample

of detection of secreted proteins. We show here that proteins secreted from adipose tissue can

be unequivocally identified by a qualitative labeling approach which allows distinguishing

the source of relevant proteins. For the future it will be interesting to compare adipose tissue

secretomes from lean and obese people to determine differences in protein expression which

may lead to the discovery of mechanisms involved in insulin resistance and type 2 diabetes.

For this, a quantitative labeling approach should be developed. We are currently working on

this topic.

Acknowledgements

This work was supported by the Netherlands Proteomic Centre (project 6.3.)

Chapter 2

85

References

1. Gimeno, R.E., and Klaman, L.D. (2005) Adipose tissue as an active endocrine organ: recent advances. Curr. Opin. Pharmacol. 5, 122-128.

2. Hutley, L., and Prins, J.B. (2005) Fat as an endocrine organ: relationship to the metabolic syndrome. Am. J. Med. Sci. 330, 280-289.

3. Rajala, M.W., and Scherer, P.E. (2003) Minireview: The adipocyte--at the crossroads of energy homeostasis, inflammation, and atherosclerosis. Endocrinology 144, 3765-3773.

4. Matsuzawa, Y. (2006) The metabolic syndrome and adipocytokines. FEBS Lett. 580, 2917-2921. 5. Permana, P.A., Menge, C., and Reaven, P.D. (2006) Macrophage-secreted factors induce adipocyte

inflammation and insulin resistance. Biochem. Biophys. Res. Commun. 341, 507-514. 6. Wellen, K.E., and Hotamisligil, G.S. (2003) Obesity-induced inflammatory changes in adipose tissue. J.

Clin. Invest. 112, 1785-1788. 7. Wellen, K.E., and Hotamisligil, G.S. (2005) Inflammation, stress, and diabetes. J. Clin. Invest. 115, 1111-

1119. 8. Xu, H., Barnes, G.T., Yang, Q., Tan,G., Yang, D., Chou, C.J., Sole, J., Nichols, A., Ross, J.S., Tartaglia,

L.A., and Chen, H. (2003) Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. J. Clin. Invest. 112, 1821-1830.

9. Scherer, P.E. (2006) Adipose tissue. From lipid storage compartment to endocrine organ. Diabetes 55, 1537-1545.

10. Corton, M., Villuendas, G., Botella, J.I., San Millan, J.L., Escobar-Morreale, H.F., and Peral, B. (2004) Improved resolution of the human adipose tissue proteome at alkaline and wide range pH by the addition of hydroxyethyl disulfide. Proteomics 4, 438-441.

11. Lanne, B., Potthast, F., Hoglund, A., Brockenhuus von Lowenhielm, H., Nystrom, A.C., Nilsson, F., and Dahllof, B. (2001) Thiourea enhances mapping of the proteome from murine white adipose tissue. Proteomics 1, 819-828.

12. Schmid, G.M., Converset, V., Walter, N., Sennitt, M.V., Leung, K.Y., Byers, H., Ward, M., Hochstrasser, D.F., Cawthorne, M.A., and Sanchez, J.C. (2004) Effect of high-fat diet on the expression of proteins in muscle, adipose tissues, and liver of C57BL/6 mice. Proteomics 4, 2270-2282.

13. Renes, J., Bouwman, F., Noben, J.-P., Evelo, C., Robben, J., and Mariman, E. (2005) Protein profiling of 3T3-L1 adipocyte differentiation and (tumor necrosis factor α-mediated) starvation. Cell. Mol. Life Sci. 62, 492-503.

14. Hansson, O., Strom, K., Guner, N., Wierup, N., Sundler, F., Hoglund, P., and Holm, C. (2006) Inflammatory response in white adipose tissue in the non-obese hormone-sensitive lipase null mouse model. J. Proteome Res. 5, 1701-1710.

15. Barcelo-Batllori, S., Corominola, H., Claret, M., Canals, I., Guinovart, J., and Gomis, R. (2005) Target identification of the novel antiobesity agent tungstate in adipose tissue from obese rats. Proteomics 5, 4927-4935.

16. Celis, J.E., Moreira, J.M., Cabezon, T., Gromov, P., Friis, E., Rank, F., and Gromova, I. (2005) Identification of extracellular and intracellular signaling components of the mammary adipose tissue and its interstitial fluid in high risk breast cancer patients: toward dissecting the molecular circuitry of epithelial-adipocyte stromal cell interactions. Mol. Cell. Proteomics 4, 492-522.

17. Kratchmarova, I., Kalume, D.E., Blagoev, B., Scherer, P.E., Podtelejnikov, A.V., Molina, H., Bickel, P.E., Andersen, J.S., Fernandez, M.M., Bunkenborg, J., Roepstorff, P., Kristiansen, K., Lodish, H.F., Mann, M., and Pandey, A. (2002) A proteomic approach for identification of secreted proteins during the differentiation of 3T3-L1 preadipocytes to adipocytes. Mol. Cell. Proteomics 1, 213-222.

18. Wang, P., Mariman, E., Keijer, J., Bouwman, F., Noben, J.P., Robben, J., and Renes, J. (2004) Profiling of the secreted proteins during 3T3-L1 adipocyte differentiation leads to the identification of novel adipokines. Cell. Mol. Life Sci. 61, 2405-2417.

19. Chen, X., Cushman, S.W., Pannell, L.K., and Hess, S. (2005) Quantitative proteomic analysis of the secretory proteins from rat adipose cells using a 2D liquid chromatography-MS/MS approach. J. Proteome Res. 4, 570-577.

Chapter 2

86

20. Halleux, C.M., Takahashi, M., Delporte, M.L., Detry, R., Funahashi, T., Matsuzawa, Y., and Brichard, S.M. (2001) Secretion of adiponectin and regulation of apM1 gene expression in human visceral adipose tissue. Biochem. Biophys. Res. Commun. 288, 1102-1107.

21. Chertov, O., Simpson, J.T., Biragyn, A., Conrads, T.P., Veenstra, T.D., and Fisher, R.J. (2005) Enrichment of low-molecular-weight proteins from biofluids for biomarker discovery. Expert Rev. Proteomics 2, 139-145.

22. Fried, S.K., and Moustaid-Moussa, N. (2001) Culture of adipose tissue and isolated adipocytes. Methods Mol. Biol. 155,197-212.

23. Tang, W.H., Halpern, B.R., Shilov, I.V., Seymour, S.L., Keating, S.P., Loboda, A., Patel, A.A., Schaeffer, D.A., and Nuwaysir, L.M. (2005) Discovering known and unanticipated protein modifications using MS/MS database searching. Analytical Chemistry 77, 3931-3946.

24. Bendtsen, J.D., Jensen, L.J., Blom, N., von Heijne, G., and Brunak, S. (2004) Feature based prediction of non-classical and leaderless protein secretion. Protein Eng. Des. Sel. 17, 349-356.

25. Krogh, A., Larsson, B., von Heijne, G., and Sonnhammer, E.L.L. (2001) Predicting transmembrane protein topology with a hidden Markov model: Application to complete genomes. Journal of Molecular Biology 305, 567-580.

26. Peri, S., Navarro, J.D., Amanchy, R., Kristiansen, T.Z., Jonnalagadda, C.K.,Surendranath, V., Niranjan, V., Muthusamy, B., Gandhi, T.K., Gronborg, M., Ibarrola, N., Deshpande, N., Shanker, K., Shivashankar, H.N., Rashmi, B.P., Ramya, M.A., Zhao, Z., Chandrika, K.N., Padma, N., Harsha, H.C., Yatish, A.J., Kavitha, M.P., Menezes, M., Choudhury, D.R., Suresh, S., Ghosh, N., Saravana, R., Chandran, S., Krishna, S., Joy, M., Anand, S.K., Madavan, V., Joseph, A., Wong, G.W., Schiemann, W.P., Constantinescu, S.N., Huang, L., Khosravi-Far, R., Steen, H., Tewari, M., Ghaffari, S., Blobe, G.C., Dang, C.V., Garcia, J.G., Pevsner, J., Jensen, O.N., Roepstorff, P., Deshpande, K.S., Chinnaiyan, A.M., Hamosh, A., Chakravarti, A., and Pandey, A. (2003) Development of human protein reference database as an initial platform for approaching systems biology in humans. Genome Research 13, 2363-2371.

27. Kratzer, R., Eckerskorn, C., Karas, M., and Lottspeich, F. (1998) Suppression effects in enzymatic peptide ladder sequencing using ultraviolet – matrix assisted laser desorption/ionization – mass spectrometry. Electrophoresis 19, 1910-1919.

28. Fain, J.N., Madan, A.K., Hiler, M.L., Cheema, P., and Bahouth, S.W. (2004) Comparison of the release of adipokines by adipose tissue, adipose tissue matrix, and adipocytes from visceral and subcutaneous abdominal adipose tissues of obese humans. Endocrinology 145, 2273-2282.

29. Xu, A., Wang, Y., Xu, J.Y., Stejskal, D., Tam, S., Zhang, J., Wat, N.M.S., Wong, W.K., and Lam, K.S.L. (2006) Adipocyte fatty acid-binding protein is a plasma biomarker closely associated with obesity and metabolic syndrome. Clinical Chemistry 52, 405-413.

30. Dupont, A., Tokarski, C., Dekeyzer, O., Guihot, A., Amouyel, P., Rolando, C., and Pinet, F. (2004) Two-dimensional maps and databases of the human macrophage proteome and secretome. Proteomics 4, 1761-1778.

31. Abderrahim-Ferkoune, A., Bezy, O., Chiellini, C., Maffei, M., Grimaldi, P., Bonino, F., Moustaid-Moussa, N., Pasqualini, F., Mantovani, A., Ailhaud, G., and Amri, E. (2003) Characterization of the long pentraxin PTX3 as a TNF-α-induced secreted protein of adipose cells. J. Lipid Res. 44, 994-1000.

32. Kershaw, E.E., and Flier, J.S. (2004) Adipose tissue as an endocrine organ. J. Clin. Endocrinol. Metab. 89, 2548-2556.

33. Krug, A.W., and Ehrhart-Bornstein, M. (2005) Newly discovered endocrine functions of white adipose tissue: possible relevance in obesity-related diseases. Cell. Mol. Life Sci. 62, 1359-1362.

34. Hauner, H. (2005) Secretory factors from human adipose tissue and their functional role. Proceedings of the Nutrition Society 64, 163-169.

35. Trayhurn, P. (2005) Endocrine and signalling role of adipose tissue: new perspectives on fat. Acta Physiol Scand 184, 285-293.

36. Wilsie, L.C., Chanchani, S., Navaratna, D., and Orlando, R.A. (2005) Cell surface heparan sulfate proteoglycans contribute to intracellular lipid accumulation in adipocytes. Lipids in Health and Disease 4, 1-15.

Chapter 2

87

37. Rahmani, M., Wong, B.W., Ang, L., Cheung, C.C., Carthy, J.M., Walinski, H., and McManus, B.M. (2006) Versican: signaling to transcriptional control pathways. Can. J. Physiol. Pharmacol. 84, 77-92.

38. Levine, J.A., Jensen, M.D., Eberhardt, N.L., and O’Brien, T. (1998) Adipocyte macrophage colony-stimulating factor is a mediator of adipose tissue growth. J. Clin. Invest. 101, 1557-1564.

39. Belaid, Z., Hubint, F., Humblet, C., Boniver, J., Nusgens, B., and Defresne, M. (2003) Differential expression of vascular endothelial growth factor and its receptors in hematopoietic and fatty bone marrow: evidence that neuropilin-1 is produced by fat cells. Haematologica 90, 400-401.

Chapter 3

88

Section II

Role of adipose tissue in the development of

insulin resistance

Chapter 3

89

Chapter 3

Sub-chronic administration of stable GIP

analogue in mice decreases serum LPL activity

and body weight

Ewa Szalowska

Kees Meijer

Niels Kloosterhuis

Farhad Razaee

Marion Priebe

Roel J. Vonk

Peptides 2011

Chapter 3

90

Abstract

GIP receptor knockout mice were shown to be protected from the development of obesity on

a high fat diet, suggesting a role of GIP in the development of obesity.

Our aims were to test the hypothesis if excess of GIP could accelerate development of obesity

and to identify GIP gene targets in adipose tissue.

Mice were kept on a chow or a high fat diet. During the last 2 weeks of the experiment mice

were injected with D-Ala2-GIP or PBS. Serum LPL activity and several biochemical

parameters (TG, FFA, cholesterol, glucose, insulin, resistin, IL-6, IL-1b, TNFa, GIP) were

measured. Fat tissue was isolated and QPCR was performed for a set of genes involved in

energy metabolism (LPL, GLUT4, FAS, SREBP1c) and inflammation (IL-6, IL-1b, TNFa).

A DNA Microarray was used to identify GIP gene targets in adipose tissue of the chow diet

group.

D-Ala2-GIP injection caused a significant decrease in both body weight and LPL activity

compared to PBS-injected animals. Serum biochemical parameters were not affected by D-

Ala2-GIP, with an exception for resistin and insulin. The set of inflammatory genes were

significantly decreased in adipose tissue in the D-Ala2-GIP injected animals on a chow diet.

A DNA microarray revealed that APO-genes and CYP- genes were affected by GIP treatment

in adipose tissue.

D-Ala2-GIP injections caused body weight loss related to decreased serum insulin level and

LPL activity. The identified GIP candidate gene targets in adipose tissue link GIP action to

lipid metabolism exerted by APO and CYP genes.

Chapter 3

91

Introduction

Glucose-dependent insulinotropic polypeptide (GIP) is one of gastrointestinal hormones

involved in the regulation of postprandial nutrient homeostasis 3. In response to glucose or

fat, GIP is secreted from K-cells in the proximal small intestine (duodenum and jejunum);

however only in the presence of glucose GIP stimulates insulin secretion and acts as

incretine. GIP acts via G-coupled GIP receptors (GIPR) which are expressed in pancreas,

adipose tissue, gastrointestinal tract, lung, brain, and bone. GIP is recognized as an anti-

diabetic hormone, since it acts on pancreatic beta cells and stimulates beta-cells proliferation,

growth, differentiation, and protects from apoptosis. Additionally GIP was shown to enhance

bone formation via stimulation of osteoblast proliferation and inhibition of apoptosis and it

may play a role in the central nervous system 2. Recent evidence indicates that GIP has an

emerging role in the development of obesity. The GIP receptor (GIPR) knockout mice on a

high fat diet were protected on a high fat diet from obesity, suggesting that GIP might

promote energy storage 20. The proposed mechanism behind the GIP-dependent fat deposition

in adipose tissue is an increase in lipoprotein lipase (LPL) activity leading to triglyceride

(TG) accumulation, shown in 3T3-L1 cells and human adipocytes 14. LPL is known to be

produced by many tissues such as adipose tissue, cardiac and skeletal muscles, macrophages,

and islets. Insulin is known to stimulate LPL activity, leading to accumulation of LPL-

catalyzed reaction products, fatty acids, and monoacylglycerol and partially taken up by the

tissues locally and stored as neutral lipids in adipose tissue, oxidized, or stored in skeletal and

cardiac muscles or as cholesteryl ester and TG in macrophages 29. Despite the evidence that

GIP activates LPL activity in vitro resulting in accumulation of TG in adipocytes, it was not

shown that infusion of the DPPIV-resistant GIP would lead to increased fat mass in vivo 13.

In addition, to the emerging role of GIP in the development in obesity, there is an interest to

use GIP for treatment of type 2 diabetes. However, one of factors limiting the use of native

GIP peptide is its short half-time in blood caused by rapid degradation by dipeptidyl

peptidase IV (DPPIV) and renal filtration. The use of GIP DPPIV stable analogues can

overcome these obstacles. One of these analogues is D-Ala2-GIP, which is equipotent to the

native GIP peptide and was shown to improve glucose tolerance in normal and obese diabetic

rats 11. In another study the D-Ala2-GIP was shown to increase serum resistin and leptin

levels in C57BL/6 mice fed a high fat diet, however no changes in body fat mass were

observed 18.

Chapter 3

92

In our studies we aimed to test the hypothesis, if excess GIP could accelerate development of

obesity on both chow and high-fat diets. Mice were intraperitoneally injected with D-Ala2-

GIP sub-chronically for 2 weeks and subsequently analyzed for body weight, various

biochemical serum parameters and subsets of metabolic and proinflammatory genes in

adipose tissue. Additionally we aimed to identify GIP gene targets in adipose tissue by means

of DNA microarray analysis in order to better understand how adipose tissue under GIP

action could influence the total body energy metabolism.

Materials and Methods

Animals and diets

Male C57BL/6 mice, age 8 weeks, were housed in a light- and temperature-controlled facility

(light on 7AM-7PM, 21ºC). The mice were fed standard laboratory chow ad libitum, n=20

(Harlan, The Netherlands or the high fat diet (HFD) ad libitum, n=20 (Harlan, The

Netherlands) and had free access to water for 11 weeks. In the last 2 weeks of the

experiments half of the animals from each diet group were intraperitoneally (ip) injected

with D-Ala2-GIP (0.12mg/kg) or PBS daily at 11AM. All experiments were approved by the

Ethics Committee for Animal Experiments of the University of Groningen.

GIP synthesis and sequencing

D-Ala2-GIP was synthetized by Solid phase Fmoc (N-(9-fluorenyl)methoxycarbonyl)

chemistry using a MilliGen 9050 peptide synthesizer (MilliGen/Biosearch, Bedford, MA), as

described previously 21. The peptide was purified by reversed-phase HPLC on a JASCO

HPLC System (Tokyo, Japan). The peptide was dissolved on 0.1% trifluoroacetic acid and

applied on a VYDAC C18-column (218TP, 1.0x25 cm, 10-µm particles, Vydac, Hesperia,

CA) equilibrated in 0.1% trifluoroacetic acid in 20 min at a flow rate of 4 ml/min. The

peptide sequence was confirmed by Orbitrap mass spectrometry analysis.

Serum biochemical measurements

Plasma insulin, IL-6, TNFa, and resistin concentrations were determined using a multiplex

assay (MILLIPLEX MAP Mouse Serum Adipokine Panel assay, Millipore, Amsterdam, The

Netherlands). Plasma GIP level was determined using ELISA (Rat/Mouse GIP (total) ELISA

Kit, Millipore, Amsterdam, The Netherlands). For the GIP measurements, plasma was

supplemented directly after collection with DPP IV protease inhibitor according to the

manufacturer’s description (Millipore, Amsterdam, The Netherlands). Plasma NEFA, TG,

Chapter 3

93

and cholesterol concentrations were determined using commercially available kits (Roche

Diagnostics, Mannheim, Germany and Wako Chemicals, Neuss, Germany).

LPL serum measurement

LPL in serum was measured using a LPL activity assay kit (Roar Biomedical Inc.NY, USA)

according to the manufacturer’s protocols.

RNA isolation and cDNA synthesis

RNA was extracted from adipose tissue using RNeasy Lipid Tissue Mini Kit (Qiagen, Venlo,

The Netherlands) according to the manufacturer’s instructions. RNA extraction from livers

was performed using Tri reagent (Sigma-Aldrich, St. Louis, MO). The RNA concentration

was determined by Nano Drop ND-1000 Spectrophotometer (Isogen IJsselstein, The

Netherlands). The quality of total RNA from adipose tissue was evaluated by capillary

electrophoresis using an Agilent 2100 Bioanalyzer (Agilent Technologies, Palo Alto, Calif.).

cDNA synthesis was performed from total RNA with QuantiTect Reverse Transcription Kit

(Qiagen, Venlo, The Netherlands) according to the manufacturer’s instructions.

Quantitative Real-Time PCR (QPCR)

Expression of the genes of interest was quantified by QPCR on an ABI prism 7900 HT

(Applied Biosystems) with the following cycling conditions: 15 min 95°C followed by 40

cycles of 15s 95 ºC and 1 min 60 ºC. Reactions were performed in 10 µl and contained 20 ng

cDNA, 1x TaqMan PCR Master Mix (Applied Biosystems, Foster City, CA), 250 nM of each

probe and 900 nM of each primer. For each gene, a standard curve was generated and

efficiency of the primers was determined as described previously 22. For each primer pair the

efficiency was about 95 %. Specific primer sets for actin β (ACTB), IL-1β, IL-6, TNFα,

lipoprotein lipase (LPL), solute carrier family 2 (facilitated glucose transporter) member 4

(GLUT4), sterol regulatory element binding protein 1c (SREBP1C), and fatty acid synthase

(FAS) were developed with Primer Express 1.5 (Applied Biosystems). The sequences for

forward (F), reverse (R) primers and the probe (P) were (5’-3’): ACTB (F) AGC CAT GTA

CGT AGC CAT CCA; (R) TCT CCG GAG TCC ATC ACA ATG; (P) TGT CCC TGT ATG

CCT CTG GTC GTA CCA C; IL-1β (F) ACC CTG CAG CTG GAG AGT GT ; (R) TTG

ACT TCT ATC TTG TTG AAG ACA AAC C; (P) CCC AAG CAA TAC CCA AAG AAG

AAG ATG GAA ; IL-6 (F) CCG GAG AGG AGA CTT CAC AGA ; (R) AGA ATT GCC

ATT GCA CAA CTC TT; (P) ACC ACT TCA CAA GTC GGA GGC TTA ATT ACA;

TNFα (F); (R); (P); LPL (F) AAG GTC AGA GCC AAG AGA AGC A; (R) CCA GAA

AAG TGA ATC TTG ACT TGG T; (P) CCT GAA GAC TCG CTC TCA GAT GCC CTA

Chapter 3

94

CA; GLUT4 (F) CTC ATG GGC CTA GCC AAT G; (R) GGG CGA TTT CTC CCA CAT

AC; (P) CAT TGG CGC CTA CTC AGG GCT AAC ATC;

SREBP1c (F) GGA GCC ATG GAT TGC ACA TT; (R) CCT GTC TCA CCC CCA GCA

TA; (P) CAG CTC ATC AAC AAC CAA GAC AGT GAC TTC C;

FAS (F) GGC ATC ATT GGG CAC TCC TT; (R) GCT GCA AGC ACA GCC TCT CT; (P)

CCA TCT GCA TAG CCA CAG GCA ACC TC.

Data were analyzed with SDS 2.0 software (Applied Biosystems). For each sample, the

QPCR reaction was performed twice in triplicate and the averages of the obtained threshold

cycle values (CT) were processed for further calculations. For normalization ACTB was

used. Relative expression was calculated with ∆(∆(CT))-method 19.

Statistical analysis

QPCR experiments were performed with n=10 adipose tissue. Insulin, IL-6, TNFα, resistin,

and GIP multiplex or ELISA measurements in plasma were in technical duplicates for n=10

animals of each experimental groups. Biochemical measurements were performed in

technical duplicates for n=10 mice in each of the experimental groups. Kruskal Wallis (KW)

test was applied to compare the different experimental groups and p-value <0.05 was

considered significant.

Illumina Human WG6-v2 Microarray Analysis

The Illumina platform was used for the gene expression analysis in adipose tissue. Adipose

tissue obtained form 8 mice (4 mice fed chow diet, injected with PBS and 4 mice fed HFD

injected with D-Ala2-GIP) were used in the DNA microarray experiment. Biotin- labeled

cRNA was generated from high-quality total RNA with the Illumina TotalPrep RNA

amplification kit (Ambion). Briefly, 50 ng of total RNA was reversely transcribed with an

oligo(dT) primer containing a T7 promoter. The first- strand cDNA was used to make the

second strand. The purified second-strand cDNA, along with biotin UTPs, was subsequently

used to generate biotinylated, antisense RNA of each mRNA in an in vitro transcription

reaction. The size distribution profile for the labeled cRNA samples was evaluated by

Bioanalyzer. After RNA labeling, 1.5ug of purified, labeled cRNA from each sample was

hybridized at 55ºC overnight with a Human-8 v2 expression Illumina Beadchip targeting

22000 transcripts. The beadchip was washed the following day. Streptavidin-Cy3 was used to

develop a signal, and each chip was scanned with an Illumina Bead Array Reader.

The preprocessing of Illumina data was performed using the BeadStudio package with default

settings. The background was subtracted and quantile normalization performed. Probes with

“absent” signals in all samples (lower than or near to background levels) were removed from

Chapter 3

95

further analysis. To select the candidate targets of GIP in adipose tissue we concentrated on

genes which change at least more than 5-fold in abundance in the D-Ala2-GIP injected

animals compared to the placebo injected mice.

Results

Effect of D-Ala2-GIP on body weight during chow and high fat diet

In order to study the effect of GIP on the body mass, mice were fed with chow diet and HFD

for 11 weeks. During the first 9 weeks animals fed with the HF diet increased body weight

faster compared to the chow diet group, as expected (Figure 1).

24

26

28

30

32

34

36

38

0 2 4 6 8 10 12

bo

dy

ma

ss

(g

)

weeks

HFD GIP

HFD PBS

CHD GIP

CHD PBS

Figure 1. Mice were fed with chow and high fat diet for 11 weeks. During the last 2 weeks of the dietary interventions D-Ala2-GIP or PBS were daily, intraperitoneally injected. X axis shows time frame of the experiment in weeks. Y axis shows mice body weight in grams. Indicated p values are calculated for body weight loss between PBS and D-Ala2-GIP injected animals.

In the last two weeks (week 10 and 11) of dietary interventions D-Ala2-GIP or PBS were

injected, what resulted in body weight reduction in animals fed with both chow and HFD.

After 2 weeks of injections there was a significant difference in body weight loss between D-

Ala2-GIP and PBS injected animals both in the chow and HFD groups (p=0.008 and p=0.01

respectively) (Figure 2).

Chapter 3

96

0.0

1.0

2.0

3.0

4.0

5.0

6.0

CHD PBS CHD GIP HFD PBS HFD GIP

we

igh

t re

du

cti

on

(g

)

Figure 2. Columns in the figure represent body weight loss after 2 weeks of D-Ala2-GIP or PBS i.p. injections. On the X axis the analyzed groups of animals are depicted. The Y axis represents body weight loss in grams in the D-Ala2-GIP injected animals compared to PBS injected mice. The difference between CHD PBS and CHD GIP and HFD PBS and HFD GIP were significant (p=0.008 and p=0.01 respectively). LPL activity in serum

With the aim to identify factors which could be implicated in the body weight reduction we

analyzed serum LPL activity. The serum LPL activity was measured in both chow and HF

diet groups after 2 weeks of administration of D-Ala2-GIP or PBS. Animals injected with D-

Ala2-GIP had significantly lower LPL activity compared to the PBS injected animals in both

chow and HF diet groups (Figure 3).

0

20

40

60

80

100

120

140

CHD PBS CHD GIP HFD PBS HFD GIP

rela

tiv

e L

PL

se

rum

ac

tiv

ity

(%

)

Figure 3. The chart represents relative LPL activity (Y axis) in mice on a chow diet injected with PBS or D-Ala2-GIP, depicted on the chart as CH PBS and CH GIP respectively; and mice fed with a high fat diet injected with PBS or D-Ala2-GIP depicted on the chart as F PBS and F GIP respectively. The difference between CHD PBS and CHD GIP and HFD PBS and HFD GIP were significant (p=0.001 and p=0.009 respectively).

Chapter 3

97

Biochemical serum parameters. In order to find systemic effects of the D-Ala2-GIP we measured several biochemical serum parameters such as triglyceride (TG), cholesterol, free fatty acids (FFA), resistin, insulin, IL-6, TNFα , and GIP serum levels. We did not detect any significant effect of the D-Ala2-GIP administration on most of the indicated parameters, except for a significantly lower resistin level in the D-Ala2-GIP injected animals on the HF diet, (p=0.04),( Figure 4A); significantly lower insulin level of the D-Ala2-GIP injected animals in the chow diet group ( p=0.016 ),( Figure 4B) ; and higher GIP serum level in the D-Ala2-GIP-injected animals on the HF diet,( p=0.002),( Figure 4C). Data for parameters which did not change significantly are not shown.

0

500

1000

1500

2000

2500

3000

3500

CHD PBS CHD GIP HFD PBS HFD GIP

resi

stin

(p

g/

ml)

0

100

200

300

400

500

600

700

800

900

CHD PBS CHD GIP HFD PBS HFD GIP

insu

lin

(p

g/

ml)

0

20

40

60

80

100

120

140

160

CHD PBS CHD GIP HFD PBS HFD GIP

GIP

(p

g/

ml)

Figure 4. (A) Resistin serum levels in the animal experimental groups. The difference CHD PBS vs. HFD PBS and HFD PBS vs. HFD GIP were significant (p=0.001 and p=0.04 respectively). (B) Insulin serum levels in the animal experimental groups. The difference CHD PBS vs. CHD GIP and CHD GIP vs. HFD GIP were significant (p=0.02 and p=0.005 respectively). (C) GIP serum levels in the animal experimental groups. The difference between HFD PBS vs. HFD GIP and CHD PBS vs. HFD PBS were significant (p=0.002 and p=0.005 respectively).

Metabolic and pro-inflammatory genes expression in adipose tissue

Adipose tissue was collected after D-Ala2-GIP or PBS injections and gene expression was

analyzed by QPCR. As expected there was a significant difference between chow and HF

diets in respect to the metabolic and the pro-inflammatory gene expression; the

proinflammatory genes (IL-1β, IL-6, and TNFα) were significantly upregulated in obese

animals (Figures 5 A-C).

4A 4B

4C

Chapter 3

98

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f IL

1b

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f TN

Fa

Figure 5. Relative gene expression of IL-1b (A), IL-6 (B), and TNFa (C) in adipose tissue (Y axis) in the experimental animals groups (x axis). The differences in RGE for IL-1b, IL-6, and TNFa between CHD and HFD groups were significant p<0.05.

The energy metabolism genes (GLUT4, LPL, SREBP1c) were significantly downregulated in

the animals fed with the HF diet (Figures 6 A-C), except for expression of FAS, which was

significantly higher in animals fed with HF diet compared to animals fed with chow diet

(Figure 6D).

0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0.045

0.05

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f G

LU

T4

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f L

PL

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f S

RE

BP

1c

0

0.5

1

1.5

2

2.5

3

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f F

AS

Figure 6. Relative gene expression of selected energy metabolism genes (GLUT4 (A), LPL (B), SREBP1c (C),

FAS (D) in adipose tissue (Y axis) in the experimental animals groups (x axis). The differences between the

CHD and HFD were significant (p<0.05).

The only significant effect of GIP was found for the proinflammatory genes expression (IL-

1β, IL-6, and TNFα), which had significantly decreased expression in adipose tissue in the

chow diet group, (Figures 7 A-C).

0

1

2

3

4

5

6

7

CHD PBS CHD GIP HFD PBS HFD GIP

RG

E o

f IL

6

5A 5B 5C

6A 6B

6D 6C

Chapter 3

99

0

0.00005

0.0001

0.00015

0.0002

0.00025

0.0003

0.00035

CHD PBS CHD GIP

RG

E I

L1

b

0

0.00001

0.00002

0.00003

0.00004

0.00005

0.00006

CHD PBS CHD GIP

RG

E I

L6

0

0.0001

0.0002

0.0003

0.0004

0.0005

0.0006

0.0007

0.0008

0.0009

0.001

CHD PBS CHD GIP

RG

E T

NFa

Figure 7. Relative gene expression (RGE) of IL-1b (A), IL-6 (B), and TNFa (C) in adipose tissue (Y axis) in the chow diet animals (x axis), the difference is significant p=0.05, p=0.02, p=0.02 for of IL-1b, IL-6, and TNFa respectively.

Identification of GIP gene targets in adipose tissue

In order to identify candidate GIP gene targets in adipose tissue, a DNA microarray was

performed in the chow diet group. The DNA microarry analysis revealed that the most

changed (mostly upregulated) genes in adipose tissue belonged to Apo family such as

APOA1, APOA2, APON (…) and cytochrome P450 supergene family of enzymes such as

CYP1A2, CYP2A5, CYP2C37, CYP2D10 (…). Additionally, we observed upregulation of

ABC transporters, SERPINA gene family and other genes encoding for transporters:

SLC22A1, SLC27A5. The results are summarized in Table1 and Table 2.

7C

7B 7A

Chapter 3

100

Table 1. The most differentially affected genes affected by D-Ala2-GIP injections in adipose tissue of the chow diet group. The DNA microarray was performed for 4 mice injected with PBS (PBS 1-4) and compared with 4 mice injected with D-Ala2-GIP (GIP 1-4). In the table Avg. PBS and Avg. GIP stand for average signal from the DNA microarray in both groups, FC stands for fold change.

Target gene ID

PBS 1 PBS 2 PBS 3 PBS 4 Avg. PBS

GIP 1 GIP 2 GIP 3 GIP 4 Avg. GIP

FC GIP vs.

PBS

ABCG5 3.3 1.9 2.6 17 6.2 39.2 172.8 205.1 13.7 107.7 17.4

ABCG8 -2.5 -2.6 -0.2 2.8 -0.625 7.2 100.3 116.3 2 56.5 90.3

APOA1 390.8 349.2 4.5 30.1 193.65 964.5 4533.8 3378 929.9 2451.6 12.7

APOA2 343.2 261 3.6 29.9 159.43 1512.1 10053.1 7442.8 1145.4 5038.4 31.6

APOC3 85 90 10.7 20.3 51.5 388.3 3711.8 2355.5 359.6 1703.8 33.1

APOC4 93.3 79.8 92.7 23.5 72.325 227.2 1447.8 1246.3 206.8 782 10.8

APOF 16.5 7.1 3.1 0.3 6.75 60.5 452.9 207.9 28.6 187.5 27.8

APOH -6.9 9.4 7.2 -16.9 -1.8 26.3 197.6 95.2 40.5 89.9 49.9

APON -2.6 3.3 1.4 -1.7 0.1 14.5 40.4 22.7 -1.9 18.9 189.3

C9 -5.5 6.7 3.4 -3.7 0.225 29.7 104.2 71.2 8.2 53.3 237

CPS1 5.3 -1.8 -10.1 1.1 -1.375 31.6 289.3 73.9 2.8 99.4 72.3

CRP 71.1 23.5 8.4 19.9 30.725 134.3 625.1 444.1 74.3 319.5 10.4

CYP1A2 -2.7 17.2 -8 -4.9 0.4 37.7 132.3 101.3 46.3 79.4 198.5

CYP24A1 45.4 19.9 6.2 11.8 20.825 0.6 -2.5 6.5 7.2 3 -10

CYP2A5 -0.3 1.7 2.3 5.5 2.3 8.8 99.6 48.1 8.3 41.2 17.9

CYP2A5 -7.2 -1.9 -0.8 10 0.025 72.2 712.4 454 16.2 313.7 12548

CYP2C37 -0.2 1.3 -7.2 -8.2 -3.575 67.3 593.1 416.2 31.4 277 77.5

CYP2C50 -4.3 -7.9 -0.1 -13.7 -6.5 53.4 541.9 341.2 22 239.6 36.9

CYP2C70 -5.5 2.8 -7.4 -0.9 -2.75 74.4 467.2 387.6 24 238.3 86.7

CYP2D10 8.2 7.7 -8.1 -1.2 1.65 90.9 582.8 459.2 54.7 296.9 179.9

CYP2D9 2.6 11.5 6.5 8 7.15 167.7 992.5 993.8 121.5 568.9 79.6

CYP4F14 0.1 -0.4 -7.4 -0.8 -2.125 32.9 165.7 107.1 6.8 78.1 36.8

SERPINA10 -7 -6.2 -0.7 5.5 -2.1 26 136.9 83.7 6.4 63.3 30.1

SERPINA12 4.9 7.6 -4.8 -1.7 1.5 25.1 102 103.6 7 59.4 39.6

SERPINA1A 2.7 36.4 1.4 8.9 12.35 79.2 536.5 229.5 77 230.6 18.7

SERPINA1B 280.9 169.4 -1.4 50.1 124.75 1167.4 8021.5 4171 907.1 3566.8 28.6

SERPINA1C 82.8 69.8 0.2 0.1 38.225 391.5 3197.6 1573.4 391.3 1388.5 36.3

SERPINA1D 211 167.9 -3.4 50.9 106.6 1015.6 6938 3646.7 916.5 3129.2 29.4

SERPINA3M 1.9 17.6 7.2 15.6 10.575 41 216.5 109.3 19.3 96.5 9.1

SERPINA6 7.5 3.2 -8.6 0.5 0.65 13.5 109.6 54.8 4.5 45.6 70.2

SLC22A1 10 2.1 27.7 4.8 11.15 84.7 343.2 600.6 22.2 262.7 23.6

SLC27A5 1.2 0.5 -13.6 -6.6 -4.625 34.7 338.6 203.4 10.3 146.8 31.7

SLC38A3 4.7 -3.2 6.7 -7.4 0.2 19.9 144.6 77.5 5.6 61.9 309.5

Chapter 3

101

Table 2. GIP candidate target genes and their functions

Target ID Gene functional description

Full name

ABCG5

The protein encoded by this gene is a member of the superfamily of ATP-binding cassette (ABC) transporters. ABC proteins transport various molecules across extra- and intra-cellular membranes. ABC genes are divided into seven distinct subfamilies (ABC1, MDR/TAP, MRP, ALD, OABP, GCN20, and White). This protein is a member of the White subfamily. In humans, this protein functions as a half-transporter to limit intestinal absorption and promote biliary excretion of sterols; however, the function of the mouse gene has not been determined. Mutations in the human gene have been associated with sterol accumulation, atherosclerosis, and sitosterolemia.

ATP-binding cassette, sub-family G (WHITE), member 5

ABCG8 Mutations in ABCG8 cause sitosterolemia, an inborn error in metabolism characterized by high plasma sterol concentrations (16

ATP-binding cassette, sub-family G (WHITE), member 8

APOA1 APOA1 is a major component of the high-density lipoprotein complex and has anti-inflammatory effects 31 apolipoprotein A-I

APOA2 In mice, amyloidogenic type C apolipoprotein A-II (apoA-II) forms amyloid fibrils n age associated-amyloidosis 15,24 apolipoprotein A-II

APOC3

ApoC3 inhibits LPL activity. High levels of plasma ApoC3 cause hypertriglyceridemia, while the absence of ApoC3 leads to the reduced plasma TG levels, and resistance to diet induced obesity. The gene encoding for ApoC3 in liver is regulated by insulin, bile acids, retinoids, statins, and fibrates. In adipose tissue little is known about functions of ApoC3, but recently it was shown that apoC3 gene expression increases during adipogenesis and is augmented by retinoid X receptor (RXR) agonists 27 apolipoprotein A-III

APOC4

ApoC-IV overexpression may perturb lipid metabolism leading to lipid accumulation. HCV core protein may modulate ApoC-IV expression through Ku antigen and PPARgamma/RXRalpha complex in human. In mouse the role is unknown 12 apolipoprotein A-IV

APOF Overexpression of murine ApoF significantly reduced total cholesterol levels by 28%, high density lipoproteins by 27% and phospholipid levels by 19%.17 apolipoprotein F

APOH These results are compatible with a role for apolipoproteins in lipid metabolism and transport in the developing lung in association with the sex difference in surfactant lipid synthesis23 apolipoprotein H

APON Unknown apolipoprotein N

C8G Complement and coagulation cascades complement component 8, gamma polypeptide

C9 Complement and coagulation cascades complement component 9

CPS1 a ligase enzyme located in the mitochondria involved in the production of urea carbamoyl-phosphate synthetase 1

CRP CRP plays a crucial role in the induction, amplification, and prolongation of inflammatory processes, including atherosclerotic lesions, c-reactive protein C-reactive protein

CYP1A2 The major P450 enzyme involved in metabolism of drugs and exogenous toxins, drug and steroid (especially estrogen) metabolism (CYP1A2),

cytochrome P450, family 1, subfamily a, polypeptide 2

CYP24A1 Xenobiotics, vit D, drugs and steroid metabolism

cytochrome P450, family 24, subfamily a, polypeptide 1

CYP2A5 Oxidative stress and xenobiotics metabolism cytochrome P450, family 2, subfamily a, polypeptide 5

CYP2C37, CYP2C50 xenobiotics metabolism

cytochrome P450, family 2. subfamily c, polypeptide 37 and polypeptide 50 resp.

CYP4F14 arachidonic acid and fatty acid metabolism cytochrome P450, family 4, subfamily f, polypeptide 14

SERPINA10, 12, 1A, 1B, 1C, 1D, 3M, 6

Serpins belong to a group of proteins with similar structures that were primarily identified as a set of proteins inhibiting proteases. Most of serpins control proteolytic cascades in processes such as coagulation and inflammation, and other serpins were shown to have diverse functions such as storage (ovalbumin), hormone carriage proteins and tumor suppressor genes. Recently it was shown that vaspin (SERPINA 12) has insulin sensitizing properties and it is expressed in human adipose tissue 28

serine (or cysteine) peptidase inhibitor, clade A (alpha-1 antiproteinase, antitrypsin), member 10, 12, 1A, 1B, 1C, 1D, 3M, 6 resp

SLC22A1\ OCT1

SLC22A1 plays a role in the hepatic uptake of metformin which is one of the most widely prescribed drugs for the treatment of type 2 diabetes 25

solute carrier family 22 (organic cation transporter), member 1\organic cation transporter 1

SLC27A5\ FATP5

Protein encoded by this gene has been shown to be a multifunctional protein that in vitro increases both uptake of fluorescently labeled long-chain fatty acid (LCFA) analogues and bile acid/coenzyme A ligase activity on overexpression. In FATP5 knockout mice it was shown that in livers there were alternations in lipid homeostasis coupled with a decreased in uptake of dietary LCFAs. The role of FATP5 in adipose tissue is not known 6

solute carrier family 27 (fatty acid transporter), member 5/fatty acid transport protein 5

SLC38A3 Protein encoded by SLC38A3 it is an amino acid transporter. Insulin decreases expression of SLC38A3 , however the dietary restricted mice have increased expression of SLC38A3 9

solute carrier family 38, member 3

Chapter 3

102

Discussion

In the present study we aimed to investigate the role of GIP in the development of obesity in

mice. We tested if excess of GIP applied subchronically by intraperitoneal injections might

contribute to the accelerated development of obesity. We injected mice on chow and HFD

daily with a stable, DPPIV-resistant analogue of GIP (D-Ala2-GIP) and measured selected

serum biochemical parameters and expression of sets of metabolic and pro-inflammatory

genes in adipose tissue. Additionally, we aimed to identify GIP gene targets in adipose tissue

in order to deepen our understanding of the role of GIP in adipose tissue metabolism.

As expected, animals which were fed for 9 weeks with a chow diet were lean and mice fed a

HFD became obese. In weeks 10 and 11 animals were fed the same diets as during the 9

weeks and were daily injected with D-Ala2-GIP or PBS. During this period (week 10 and 11)

mice lost body weight, however the body weight loss was significantly higher in mice

injected with the stable GIP analogue compared to placebo injected animals, in both chow

and HFD groups. These findings are in contradiction to other experiments performed in

C57BL/6 mice fed a HFD and injected subcutaneously (twice-daily) for 8 weeks with D-Ala-

GIP where no significant change in body mass was observed 18. It is difficult to explain this

difference, especially because the applied concentrations of the drug were identical (24

nM/kg) and other experimental factors such as gender, age, housing conditions were very

similar. In the experiment of Lemont et al. the D-Ala2-GIP was applied twice daily and for 8

weeks, but it is unlikely that these factors explain the difference. Here we can only speculate,

that the difference in genetic variation of the animals used in our and Lemont’s study plays a

significant role. In our experiments, the decrease in body mass in both D-Ala2-GIP and PBS

injected animals could be partly explained by stress caused by daily injections; however, the

significant higher body weight loss in GIP injected animals indicates that D-Ala2-GIP has an

effect on body weight. The observed significant higher body weight reduction in GIP injected

animals was coupled with significantly lower LPL activity in serum. This suggests that the

stable GIP analogue effect might be mediated by a decrease of LPL serum activity, resulting

in lower TG accumulation in adipose and/or other tissues and eventually decreased body

mass. Moreover to support this hypothesis, we observed that insulin level was significantly

reduced in the D-Ala2-GIP injected animals fed chow diet (p=0.016) and a similar trend,

however not significant, was observed in animals fed HF diet. Insulin is known to increase

LPL activity; thereby the combined decreased insulin serum level and the decreased LPL

Chapter 3

103

activity observed in our study can be involved in the mechanism responsible for the

decreased body weight reduction upon D-Ala2-GIP action. These findings might be in

contradiction to earlier in vitro studies demonstrating that native GIP increases LPL activity

and triglyceride (TG) accumulation in differentiated 3T3-L1 cells and human subcutaneous

adipocytes 14. However, these and our results can not be compared directly since native and

modified GIP analogue were used respectively, and in vitro studies can not be directly

extrapolated to the in vivo situation. Moreover, recently, it was reported that in clinical trials,

application of GIP agonists led to weight loss, blood pressure reduction, and, as expected,

beta-cell function improvements 8 thereby suggesting an anti-diabetic role for the

overstimulation of GIP signaling in vivo.

Besides LPL activity we measured several serum biochemical parameters in order to analyze

systemic effects of D-Ala2-GIP. We did not find any effect on TG, cholesterol, and FFA

levels in D-Ala2-GIP injected animals. We also could not confirm previous observations that

GIP injections lead to increased resistin serum levels 10. However, in agreement with a

previous report 26, we observed that resistin serum levels were significantly elevated in obese

mice. This effect, though, was not seen in the GIP injected animals on a HFD, suggesting

that GIP might reverse the effect of a HFD on resistin levels. In contradiction to findings of

Miyawaki et al. 20, the obese animals did not have elevated GIP serum levels. Thereby, this

observation does not support the hypothesis that GIP acts as a direct link between

overnutrition and obesity.

As anticipated, the selected metabolic and proinflammatory gene expression analysis showed

that obese animals had significantly upregulated proinflammatory- and significantly

downregulated energy metabolism genes expression in adipose tissue compared to mice fed

chow diet. These findings are indicative for local inflammation and insulin resistance in

adipose tissue of obese animals. However the D-Ala2- GIP significantly decreased

proinflammatory gene expression in adipose tissue of lean animals. These data suggest that

GIP might have indirect anti-inflammatory actions in lean animals which are absent in obese

mice. The absence of this effect in obese animals could be due a very strong pro-

inflammatory action of the HFD which can not be compensated by GIP. We also anticipated

that in fat tissue of obese animals angiogenic processes, requiring upregulation of

“proinflammatory” cytokines, were activated, in order to support expansion of fat tissue

necessary for storage of TGs, and associated with these processes angiogenesis 5.

Subsequently, we wondered if the local inflammation in adipose tissue of animals fed a HFD

would result in increased serum concentration of proinflammatory cytokines. We did not

Chapter 3

104

detect any differences in serum levels for theses cytokines between lean and obese animals.

These findings are in line with the hypothesis that the local inflammation in adipose tissue

might precede systemic inflammation and insulin resistance in obesity 1,4,7 or it could mean

that the observed upregulation of proinflammatory cytokines is characteristic for

angiogenesis necessary for the expansion of fat tissue related to overnutrition as stated above 5.

The DNA-microarray data analysis revealed novel GIP candidate targets. Within these genes

we identified several APO family genes known from their involvement in lipid metabolism in

the liver, however, not reported yet in the same context in adipose tissue metabolism. For

example APOA1 was reported to be the major component of the high-density lipoprotein

complex and has anti-inflammatory effects 31. ApoC-IV overexpression may perturb lipid

metabolism leading to lipid accumulation resulting in steatosis in humans; in mice the role of

ApoCIV is largely unknown 12. Murine ApoF significantly reduced total cholesterol levels,

high density lipoproteins and phospholipid levels 17. Also other candidate GIP target genes

are known to be involved in lipid metabolism such as SLC27A5 6, ABCG5, ABCG8, 16

ApoH 23. The role of other genes such as SERPINA-group family is not established yet in

mice in relation to insulin resistance. However it was recently shown that human vaspin

(SERPINA 12) exhibits insulin sensitizing properties and its mRNA decreased with

worsening of diabetes and body weight loss. Vaspin mRNA increased upon treatment with

pioglitazone and improved glucose tolerance in obese ICG mice 28. Further research is needed

to elucidate the link between GIP, insulin sensitization and SRPINA group family of proteins.

The cytochrome P450 supergene family of enzymes is involved in metabolism of steroids,

fatty acids, vitamin D, but also the degradation of drugs and exogenous toxins 30. In our

studies we could not exclude that the upregulation of the CYP450 genes was due to

exogenous effect of D-Ala2-GIP, which had to be metabolized and excreted due to its

toxicity; however we did not observed any changes in animals indicative for obvious toxic

actions of D-Ala2-GIP.

In summary, our studies showed for the first time that in vivo application of D-Ala2-GIP

induces significant body weight reduction, which can be explained by decreased LPL serum

activity and decreased insulin serum levels. Moreover, the identified novel candidate GIP

target genes in adipose tissue linking GIP actions to lipid metabolism exerted by APO,

SERPINA or ABCG genes bring new insights into adipose tissue metabolism in relation to

GIP. The exact mechanism connecting GIP and its candidate target genes in lipid metabolism

has to be established.

Chapter 3

105

Acknowledgements

We would like to thank Henk van der Molen for excellent technical assistance in animal

handling.

References

1. Andersson, C. X.; Gustafson, B.; Hammarstedt, A.; Hedjazifar, S.; Smith, U. Inflamed adipose tissue,

insulin resistance and vascular injury. Diabetes Metab Res Rev 2008 Nov;24(8):595-603. 2. Asmar, M.; Holst, J. J. Glucagon-like peptide 1 and glucose-dependent insulinotropic polypeptide: new

advances. Curr Opin Endocrinol Diabetes Obes 2010 Feb;17(1):57-62. 3. Badman, M. K.; Flier, J. S. The gut and energy balance: visceral allies in the obesity wars. Science

2005 Mar;307(5717):1909-1914. 4. Bastard, J. P.; Maachi, M.; Lagathu, C.; Kim, M. J.; Caron, M.; Vidal, H.; Capeau, J.; Feve, B. Recent

advances in the relationship between obesity, inflammation, and insulin resistance. Eur Cytokine Netw 2006 Mar;17(1):4-12.

5. Cao, Y. Adipose tissue angiogenesis as a therapeutic target for obesity and metabolic diseases. Nat Rev Drug Discov 2010 Feb;9(2):107-115.

6. Doege, H.; Stahl, A. Protein-mediated fatty acid uptake: novel insights from in vivo models. Physiology (Bethesda ) 2006 Aug;21:259-268.

7. Flowers, J. B.; Oler, A. T.; Nadler, S. T.; Choi, Y.; Schueler, K. L.; Yandell, B. S.; Kendziorski, C. M.; Attie, A. D. Abdominal obesity in BTBR male mice is associated with peripheral but not hepatic insulin resistance. Am J Physiol Endocrinol Metab 2007 Mar;292(3):E936-E945.

8. Fonseca, V. A.; Zinman, B.; Nauck, M. A.; Goldfine, A. B.; Plutzky, J. Confronting the type 2 diabetes epidemic: the emerging role of incretin-based therapies. Am J Med 2010 Jul;123(7):S2-S10.

9. Gu, S.; Villegas, C. J.; Jiang, J. X. Differential regulation of amino acid transporter SNAT3 by insulin in hepatocytes. J Biol Chem 2005 Jul;280(28):26055-26062.

10. Hansotia, T.; Maida, A.; Flock, G.; Yamada, Y.; Tsukiyama, K.; Seino, Y.; Drucker, D. J. Extrapancreatic incretin receptors modulate glucose homeostasis, body weight, and energy expenditure. J Clin Invest 2007 Jan;117(1):143-152.

11. Hinke, S. A.; Gelling, R. W.; Pederson, R. A.; Manhart, S.; Nian, C.; Demuth, H. U.; McIntosh, C. H. Dipeptidyl peptidase IV-resistant [D-Ala(2)]glucose-dependent insulinotropic polypeptide (GIP) improves glucose tolerance in normal and obese diabetic rats. Diabetes 2002 Mar;51(3):652-661.

12. Kim, E.; Li, K.; Lieu, C.; Tong, S.; Kawai, S.; Fukutomi, T.; Zhou, Y.; Wands, J.; Li, J. Expression of apolipoprotein C-IV is regulated by Ku antigen/peroxisome proliferator-activated receptor gamma complex and correlates with liver steatosis. J Hepatol 2008 Nov;49(5):787-798.

13. Kim, S. J.; Nian, C.; McIntosh, C. H. Resistin is a key mediator of glucose-dependent insulinotropic polypeptide (GIP) stimulation of lipoprotein lipase (LPL) activity in adipocytes. J Biol Chem 2007 Nov;282(47):34139-34147.

14. Kim, S. J.; Nian, C.; McIntosh, C. H. Activation of lipoprotein lipase by glucose-dependent insulinotropic polypeptide in adipocytes. A role for a protein kinase B, LKB1, and AMP-activated protein kinase cascade. J Biol Chem 2007 Mar;282(12):8557-8567.

15. Korenaga, T.; Fu, X.; Xing, Y.; Matsusita, T.; Kuramoto, K.; Syumiya, S.; Hasegawa, K.; Naiki, H.; Ueno, M.; Ishihara, T.; Hosokawa, M.; Mori, M.; Higuchi, K. Tissue distribution, biochemical properties, and transmission of mouse type A AApoAII amyloid fibrils. Am J Pathol 2004 May;164(5):1597-1606.

Chapter 3

106

16. Kruit, J. K.; Drayer, A. L.; Bloks, V. W.; Blom, N.; Olthof, S. G.; Sauer, P. J.; de, H. G.; Kema, I. P.; Vellenga, E.; Kuipers, F. Plant sterols cause macrothrombocytopenia in a mouse model of sitosterolemia. J Biol Chem 2008 Mar;283(10):6281-6287.

17. Lagor, W. R.; Brown, R. J.; Toh, S. A.; Millar, J. S.; Fuki, I. V.; de, l. L.-M.; Yuen, T.; Rothblat, G.; Billheimer, J. T.; Rader, D. J. Overexpression of apolipoprotein F reduces HDL cholesterol levels in vivo. Arterioscler Thromb Vasc Biol 2009 Jan;29(1):40-46.

18. Lamont, B. J.; Drucker, D. J. Differential antidiabetic efficacy of incretin agonists versus DPP-4 inhibition in high fat fed mice. Diabetes 2008 Jan;57(1):190-198.

19. Livak, K. J.; Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 2001 Dec;25(4):402-408.

20. Miyawaki, K.; Yamada, Y.; Ban, N.; Ihara, Y.; Tsukiyama, K.; Zhou, H.; Fujimoto, S.; Oku, A.; Tsuda, K.; Toyokuni, S.; Hiai, H.; Mizunoya, W.; Fushiki, T.; Holst, J. J.; Makino, M.; Tashita, A.; Kobara, Y.; Tsubamoto, Y.; Jinnouchi, T.; Jomori, T.; Seino, Y. Inhibition of gastric inhibitory polypeptide signaling prevents obesity. Nat Med 2002 Jul;8(7):738-742.

21. Oudhoff, M. J.; Bolscher, J. G.; Nazmi, K.; Kalay, H.; van 't, H. W.; Amerongen, A. V.; Veerman, E. C. Histatins are the major wound-closure stimulating factors in human saliva as identified in a cell culture assay. FASEB J 2008 Nov;22(11):3805-3812.

22. Pfaffl, M. W. A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 2001 May;29(9):e45.

23. Provost, P. R.; Boucher, E.; Tremblay, Y. Apolipoprotein A-I, A-II, C-II, and H expression in the developing lung and sex difference in surfactant lipids. J Endocrinol 2009 Mar;200(3):321-330.

24. Sawashita, J.; Kametani, F.; Hasegawa, K.; Tsutsumi-Yasuhara, S.; Zhang, B.; Yan, J.; Mori, M.; Naiki, H.; Higuchi, K. Amyloid fibrils formed by selective N-, C-terminal sequences of mouse apolipoprotein A-II. Biochim Biophys Acta 2009 Oct;1794(10):1517-1529.

25. Shu, Y.; Sheardown, S. A.; Brown, C.; Owen, R. P.; Zhang, S.; Castro, R. A.; Ianculescu, A. G.; Yue, L.; Lo, J. C.; Burchard, E. G.; Brett, C. M.; Giacomini, K. M. Effect of genetic variation in the organic cation transporter 1 (OCT1) on metformin action. J Clin Invest 2007 May;117(5):1422-1431.

26. Steppan, C. M.; Bailey, S. T.; Bhat, S.; Brown, E. J.; Banerjee, R. R.; Wright, C. M.; Patel, H. R.; Ahima, R. S.; Lazar, M. A. The hormone resistin links obesity to diabetes. Nature 2001 Jan;409(6818):307-312.

27. Takahashi, Y.; Inoue, J.; Kagechika, H.; Sato, R. ApoC-III gene expression is sharply increased during adipogenesis and is augmented by retinoid X receptor (RXR) agonists. FEBS Lett 2009 Jan;583(2):493-497.

28. Wada, J. Vaspin: a novel serpin with insulin-sensitizing effects. Expert Opin Investig Drugs 2008 Mar;17(3):327-333.

29. Wang, H.; Eckel, R. H. Lipoprotein lipase: from gene to obesity. Am J Physiol Endocrinol Metab 2009 Aug;297(2):E271-E288.

30. Yang, X.; Zhang, B.; Molony, C.; Chudin, E.; Hao, K.; Zhu, J.; Gaedigk, A.; Suver, C.; Zhong, H.; Leeder, J. S.; Guengerich, F. P.; Strom, S. C.; Schuetz, E.; Rushmore, T. H.; Ulrich, R. G.; Slatter, J. G.; Schadt, E. E.; Kasarskis, A.; Lum, P. Y. Systematic genetic and genomic analysis of cytochrome P450 enzyme activities in human liver. Genome Res 2010 Aug;20(8):1020-1036.

31. Yoo, K. H.; Kim, Y. N.; Lee, M. J.; Seong, J. K.; Park, J. H. Identification of apolipoproteinA1 reduction in the polycystic kidney by proteomics analysis of the Mxi1-deficient mouse. Proteomics 2009 Aug;9(15):3824-3832.

Chapter 4

107

Chapter 4

Adipokines and energy metabolism genes, but

not proinflammatory genes are deregulated in

patients with higher HOMA and lower HDL

Ewa Szalowska

Gerard J. te Meerman Annemieke Hoek

Roel J. Vonk

In preparation

Chapter 4

108

Abstract

Context: The adipose tissue (AT) gene expression changes related to obesity may lead to

alternations in the metabolism of the tissue itself and contribute to the development of

systemic insulin resistance (IR). In has been suggested that one of the pivotal events leading

to systemic IR is inflammation of AT and it was shown that in AT of obese subjects

expression of proinflammatory genes was upregulated, while the energy metabolism genes

were downregulated. These changes were predominant in visceral fat compared to

subcutaneous fat tissue (SAT).

Objective: We studied expression of subsets of proinflammatory and energy metabolism

genes in omentum and SAT in non diabetic women in order to test if: (1) AT displays

proinflammatory or/and metabolic alternations in early stages in the development of IR and

(2) if both kind of AT display differential expression for the selected genes indicating on their

divergent functions.

Methods: Relative gene expression in AT was determined by QPCR. The basic biochemical

serum parameters were measured by means of standard laboratory techniques.

Results: Adiponectin and metabolic genes expression was decreased in women characterized

by higher HOMA and low HDL level, compared to the counterpart groups; no difference in

expression of pro-inflammatory genes was observed in these groups. Proinflammatory genes

expression was increased in women with waist circumference above 88 cm. The SAT and

omentum displayed differential gene expression for some of the analyzed genes.

Conclusions: Based on the altered AT gene expression and assuming that high HOMA and

low HDL level are related to the developing IR, we postulate that metabolic perturbations

precede AT inflammation in early stages of the development of IR. Furthermore, the

differential gene expression between SAT and omentum indicates on different functions for

these fat depots.

KEYWORDS: insulin resistance, inflammation, energy metabolism and proinflammatory and

energy metabolism gene expression, subcutaneous adipose tissue, omentum

Chapter 4

109

Introduction

Obesity is associated with an increased risk factor for the development of insulin resistance

(IR) and type 2 diabetes (T2D). Adipose tissue (AT) acts as an endocrine organ secreting a

variety of factors that exert paracrine and endocrine effects and plays an important role in the

maintenance of the whole-body energy homeostasis [1; 2]. In obesity AT releases increased

amounts of proinflammatory cytokines, and has perturbed secretion pattern of adipokines

what could contribute to the systemic IR [3-7]. It was shown in humans and rodents, that

overnutrition cause oxidative stress which contributes to the infiltration of macrophages into

AT, causing a shift in fat tissue physiology from an organ predominantly involved in the

energy metabolism towards a significant source of proinflammatory cytokines released to the

blood such as TNF-α, IL-6, and IL-1β [8; 9; 9-12]. Therefore, inflammation of AT is

regarded as one of the key events leading to the development of systemic IR and in its early

stages is manifested by upregulated expression of several proinflammatory genes in AT.

Moreover, IR patients are characterized by low serum level of high density lipoprotein (HDL)

and elevated levels of serum glucose, insulin, triglycerides (TG), cholesterol, low density

lipoprotein (LDL) and high values for homeostasis model assessment (HOMA), waist

circumference (WC) and body mass index (BMI) [13].

Several studies have shown that the distribution of body fat is a critical factor for the

determination of insulin sensitivity. Lean individuals with more subcutaneous adipose tissue

(SAT) are more insulin sensitive than lean individuals with more fat distributed centrally

(omentum and visceral fat). Furthermore, it was shown in most but not all studies, that intra-

abdominal adipose tissue expresses more genes encoding for secretory proteins and has a

more proinflammatory character than SAT [14; 15].

In our studies, we aimed to find out if during early stages of insulin resistance AT

inflammation precedes its metabolic perturbations. Therefore, we studied proinflammatory

and energy metabolism genes expression in non-diabetic women. In order to characterize

patients we determined serum biochemical parameters (glucose, low density lipoprotein

(LDL), high density lipoprotein (HDL), cholesterol, triglycerides (TG), and insulin) and

anthropometric factors such as body mass index (BMI) and waist circumference (WC).

Thereafter, we applied a specific cut off value for the analyzed parameters, in order to divide

patients into low or high group for a specific parameter, for example low or high HDL level

group. Then, we aimed to identify which biochemical or anthropometric parameters are

Chapter 4

110

related to altered AT gene expression. Moreover, we wanted to asses if the AT gene

expression changes were fat depot specific, therefore the analyzed genes were studied both in

SAT and omentum.

Materials and Methods

Study subjects

In the study 38 Caucasian women undergoing surgery because of benign gynecological

problems were included. The women were in general good health, as determined by medical

history and physical examination. All the individuals had no history or symptoms of T2D or

other inflammatory diseases. The research protocols were approved by the medical ethical

committee of the University of Groningen, University Medical Center Groningen.

Participants gave their written informed consent.

Anthropometric measurements, blood collection and adipose tissue sampling

Anthropometric examination of the subjects was performed 1 day before the surgical

intervention. The patients were weighted, and the body mass index was calculated. Blood

samples for biochemical and hormonal measurements were taken after overnight fasting, at

the day of operation, before start of anesthesia. Anthropometric and biochemical

characteristics of the subjects are summarized in Table 3.

The subcutaneous biopsies were taken at the place of the incision, midline lower abdomen

above the symfysis and under the umbilicus. The biopsies of the omentum were taken at the

lower edge of the omentum, both types of biopsies were taken by means of surgical scissor.

Human omental and subcutaneous surgical biopsies were placed in sterile transfer buffer

(TB) (PBS containing 5.5 mM glucose and 50 µg/ml Gentamycine) in the operating room ,

and transferred within 10 minutes to the laboratory where the biopsies were snap frozen in

liquid nitrogen and stored in -80ºC till further processing.

Serum adipokines, insulin and biochemical analysis

Total serum insulin, cholesterol, triglycerides, HDL-cholesterol, LDL cholesterol, and

glucose were determined in the Laboratory Centre of the University Medical Centre

Groningen (UMCG) with standard laboratory methods.

Chapter 4

111

Homeostasis model assessment (HOMA-IR) index was calculated as previously described

(Matthews et al 1985) using the following formula: fasting serum insulin (mIU/l) x fasting

serum glucose (mmol/l)/22.5.

RNA isolation and cDNA synthesis

RNA was extracted from adipose tissue using RNeasy Lipid Tissue Mini Kit (Qiagen, Venlo,

The Netherlands) according to the manufacturers’ instructions. The RNA concentration was

determined by Nano Drop ND-1000 Spectrophotometer (Isogen Ijsselstein, The Netherlands).

First-strand cDNA synthesis was performed from total RNA with QuantiTect Reverse

Transcription Kit (Qiagen, Venlo, the Netherlands) according to the manufacturers’

instructions.

Real-time quantitative PCR

Quantification of the expression of the genes of interest was done by QRPCR on ABI prism

7900 HT (Applied Biosystems) with the following cycling conditions: 15 min 95°C followed

by 40 cycles of 15s 95 ºC and 1 min 60 ºC. Reactions were performed in 10 µl volume and

contained 20 ng cDNA, 1x TaqMan PCR Master Mix (Applied Biosystems, Foster City, CA),

250 nM of each probe and 900 nM of each primer. For each gene, a standard curve was

generated and efficiency of the primers was determined as described previously [16]. For

each primer pair the efficiency was about 95 %. Specific primer sets were developed with

Primer Express 1.5 (Applied Biosystems) and the sequences are given in Table 1. Primers for

CRP were purchased from Applied Biosystems (the Netherlands). Data were analyzed with

SDS 2.0 software (Applied Biosystems). For each sample, the RT-PCR reaction was

performed twice in triplicate and the averages of the obtained threshold cycle values (CT)

were processed for further calculations. For normalization B2M was used. Relative

expression was calculated with ∆ (∆ (CT))-method [17]. The functional and nomenclature

summary of the studied genes is presented in Table 2.

Chapter 4

112

Table 1. Sequences of primers and probes used in Taq-Man Q-PCR.

Gene symbol

Forward primer (5’-3’) Reverse primer (5’-3’) Probe (5’-3’)

ADIPOQ AGG CCG TGA TGG CAG AGA T GTC TCC CTT AGG ACC AAT AAG ACC T ATC TCC TTT CTC ACC CTT CTC ACC AGG G

LEP TCA CCA GGA TCA ATG ACA TTT CAC AGC CCA GGA ATG AAG TCC AAA C CGC AGT CAG TCT CCT CCA AAC AGA AAG TCA

RETN GCC GGA TTT GGT TAG CTG AG GAG GAG GAG ACA GAG AGC TTT CAT CCA CCG AGA GGC GCC TGC AG

PPARγ GAT GTC TCA TAA TGC CAT CAG GTT GGA TTC AGC TGG TCG ATA TCA CT CCA ACA GCT TCT CCT TCT CGG CCT G

GLUT4 GCTGTGGCTGGTTTCTCCAA CCCATAGCCTCCGCAACATA CGAGCAACTTCATCATTGGCATGGGTT

LPL TGGAGATGTGGACCAGCTAGTG CAGAGAGTCGAT GAAGAGATGAATG CTCCCACGAGCGCT

SREBP1C GGA TTG CAC TTT CGA AGA CAT G AGC ATA GGG TGG GTC AAA TAG G CAG CTT ATC AAC AAC CAA GAC AGT GAC TTC CC

FASN GCAAATTCGACCTTTCTCAGAAC GGACCCCGTGGAATGTCA CCCGCTCGGCATGGCTATC

IR CAACGGGCAGTTTGTCGAA GCAGCCGTGTGACTTACAGATG ACTCATAGTCACTGCCAGAAAGTTTGCCCG

GLUR TGCACTGCACCCGCAAT GCACGGAGCTGGCTTTCA CGCGAATCTGTTTGCGTCCT

GIPR GGCCTTTCTGGACCAAAGG TGTGGCGAGAGACAGGGAGTA TTGGAGCGGTTGCAGGTCATGTACAC

GHRR CCTTCCACGTAGGGCGATATT TGATCTGAGCAATCTCCAAGGA TTTTCCAAATCCTTTGAGCCTG

GLP-1R CCCATTCTCTTTGCCATTGG GCACATGAGATTGGCCTTCA TGTTCGGGTCATCTGCATCGTGGT

CD 68 GCTTCTCTCATTCCCCTATGGA ATGTAGCTCAGGTAGACAACCTTCTG CAGCTTTGGATTCATGCAGGACCTCC

CD 163 TGCAGAAAACCCCACAAAAAG CAAGGATCCCGACTGCAATAA AACAGGTCGCTCATCCCGTCAGTCA

IL-1β CTGATGGCCCTAAACAGATGAAG GGTCGGAGATTCGTAGCAGCTGGAT TTCCAGGACCTGGACCTCTGCCCTC

IL-6 CCAGGAGCCCAGCTATGAAC CCCAGGGAGAAGGCAACTG CCTTCTCCACAAGCGCCTTCGGT

TNFα CTCGAACCCCGAGTGACA AGCTGCCCCTCAGCTTGA CCTGTAGCCCATGTTGTAGCAAACC

B2M TGACTTTGTCACAGCCCAAGATA AATCCAAATGCGGCATCTTC TGATGCTGCTTACATGTCTCGATCCCA

Chapter 4

113

Table 2. Functional and nomenclature summary of the studied genes.

Gene symbol

Gene name Function description

AD

IPO

KIN

ES

ADIPOQ adiponectin decreases glucose, FFA, TG, increases glucose uptake and fatty acid oxidation in muscle; increased adiponectin levels have many beneficial effects on insulin sensitivity and inflammation, plasma levels are decreased in diabetic compared to non diabetic people

LEP leptin antidiabetic, increases the whole body glucose metabolism, decreases glycogen content in liver and stimulates fatty acid oxidation in skeletal muscle; decreases during fasting and it is restored after feeding; plasma levels correlate with the level of adiposity;

RETN resistin correlates with markers of inflammation and CVD

GE

NE

S IN

VO

LVE

D IN

EN

ER

GY

ME

TA

BO

LIS

M

PPARγ Peroxisome proliferator- activated receptor gamma

regulator of adipocyte differentiation

GLUT4 Solute carrier family 2, member 4

insulin-regulated facilitative glucose transporter

LPL Lipoprotein lipase expressed in heart, muscle, adipose tissue, has the dual functions of triglyceride hydrolase and ligand/bridging factor for receptor-mediated lipoprotein uptake

SREBP1C Sterol regulatory element binding transcription factor 1c

transcription factor that binds to the sterol regulatory element-1 (SRE1), which is flanking the low density lipoprotein receptor gene and some genes involved in sterol biosynthesis

FASN Fatty acid synthase its main function is to catalyze the synthesis of palmitate from acetyl-CoA and malonyl-CoA, in the presence of NADPH, into long-chain saturated fatty acids

INSR Insulin receptor mediates insulin actions; there are two transcript variants encoding different isoforms of INSR

GLUR Glucagon receptor mediates glucagon actions, mostly expressed in liver and kidneys, but also adipose tissue, heart , spleen, and gastrointestinal tract

GIPR Glucose dependend insulinotropic peptide receptor

mediates effects of GIP, involved in the regulation of the lipid metabolism by for example activating LPL activity in adipocytes

GHRR Ghrelin receptor mediates ghrelin effects such as regulation of appetite

GLP-1R Glucagon like peptide-1 receptor

mediates effects of GLP-1 such as decreasing food intake by increasing satiety

PR

OIN

FLA

MM

AT

OR

Y

GE

NE

S

CD 68 CD 68 molecule macrophage marker

CD 163 CD 163 molecule marker for activated macrophages, increases during endotoxemia

CRP C-reactive protein expressed by human monocytes and tissue macrophages, involved in clearance of cellular debris, promotes phagocytosis, and mediates the recruitment and activation of macrophages.

IL-1β Interleukin 1 beta produced by activated macrophages, mediator of the inflammatory response, involved in a variety of cellular activities, including cell proliferation, differentiation, and apoptosis.

IL-6 Interleukin 6 cytokine ,marker of many inflammatory diseases such as T2D, CVD, arthritis

TNFα Tumor necrosis factor alpha proinflammatory cytokine, secreted by macrophages, involved in cell proliferation, differentiation, apoptosis, lipid metabolism, and coagulation; implicated in diseases, including autoimmune diseases, IR, and cancer.

Chapter 4

114

Statistical analysis

The statistical analysis was performed using SPSS 12.0.1 for Windows (SPSS Inc., Chicago,

IL, USA). The obtained values for the relative gene expression were ranked, the lowest value

obtained the lowest rank (1), the next low value rank 2, etc, and the highest values obtained

the highest rank. To determine differences in changes of relative gene expression (dependent

variables) against anthropometric and biochemical parameters (independent variables)

Kruskal Wallis test was used. The cut off values for dependent variables were set as depicted

in Table 2. Statistical significance was assigned to p ≤0.05.

Results

1. Anthropometric and biochemical characteristics of the study group and applied cut-

off values for the studied parameters

The patients were classified in three groups for BMI (lean, overweight, obese), two groups

for waist circumference (WC), and lower and higher levels group for the following

biochemical and hormonal parameters: HOMA, glucose, insulin, triglycerides (TG), total

cholesterol, LDL, and HDL. The cut off values for the groups are given in Table 3. The

selected cut off values for the HOMA, insulin, cholesterol, LDL, group were set based on the

median value for the specific parameter within the group. The cut off values for HDL,

glucose, TG were set according to the WHO Clinical Criteria for Metabolic Syndrome [13].

Chapter 4

115

Table 3. Anthropometric and biochemical characteristics of the study group and applied cut-off values for the studied parameters.

2. Identification of differentially expressed genes in relation to anthropometric and

biochemical parameters

By means of Q-PCR gene expression in SAT and omentum was determined in relation to the

biochemical and anthropometric parameters. The statistical analysis revealed significantly

changed genes for the investigated parameters. Within the BMI-group the expression of

leptin in omentum and PPARγ, LPL, and IL1β in SAT was significantly different. For the

WC-group the omentum had different expression of leptin, LPL and IL-6 and the SAT,

within the same group, had different expression for resistin, IL-1β, and IL-6. The HOMA-

group had significantly changed expression in SAT for adipokines and several energy

metabolism genes while in omentum expression of resistin, LPL, GLUR and CD163 were

changed. The TG-group showed changes associated with adipokines genes in SAT and in

omentum only resistin expression was changed. The HDL group had changed expression in

SAT of adipokines and several metabolic genes, whereas in omentum leptin and LPL

expression were altered. For the LDL-group only the resistin expression in omentum was

significantly changed. Within the cholesterol group leptin in omentum and resistin in SAT

expression was changed. The results are summarized in Table 4.

INDEPENDENT VARIABLES CUT OFF FOR INDEPENDENT VARIA BLES AND NUMBER (N) OF SUBJECTS WITHIN THE GROUP

Total number of subjects N=38

BMI N=15 BMI<25 N=16 25<BMI<30

N=7 BMI>30

Waist circumference (WC) cm N=16 WC<88cm N=22 WC>88cm

Age (years) N=8 30-39 N=17 40-49 N=8 50-59 N=5 60-69

Fat tissue type N=38 SAT N=38 omentum

HOMA N=26 HOMA<2.6 N=12 HOMA>2.6

Glucose (mM) (G) N=19 G <5.5 N=19 G>5.5

Insulin (pg/ml) (I) N=17 I<250 N=21 I>250

TG(mM) N=16 TG <1 N=22 TG >1

Cholesterol (mM) (CH) N=23 CH<3 N=15 CH>3

LDL (mM) N=16 LDL<4 N=22 LDL>4

HDL (mM) N=27 HDL<1.3 N=11 HDL>1.3

Chapter 4

116

Table 4. The significantly changed gene expression in SAT and omentum (om.) for the analyzed biochemical (glucose, insulin, HOMA, HDL, LDL, total cholesterol, TG) and anthropometric parameters (BMI, WC). Kruskal-Wallis test was used to calculate p-value signifying the difference in a gene expression for a particular parameter, p value ≤0.05 was considered significant. None of the analyzed gene had changed expression for both “glucose” and “insulin” groups.

BMI WC HOMA TG HDL LDL Cholesterol

ADIPOQ SAT NS NS p=0.009 p=0.05 p=0.02 NS NS

LEPTIN SAT. NS NS p= 0.03 p=0.003 p=0.006 NS NS

LEPTIN om. p=0.001 p=0.002 NS NS p=0.02 NS p=0.03

RETN SAT. NS p=0.05 NS NS p=0.006 NS p=0.04

RETN om. NS NS p=0.05 p=0.01 NS p=0.04 NS

PPARg SAT p=0.05 NS p=0.03 NS NS NS NS

SREBP1c SAT NS NS p=0.05 NS p=0.01 NS NS

LPL SAT. p=0.03 NS p=0.04 NS p=0.03 NS NS

LPL om. NS p=0.04 p=0.05 NS p=0.05 NS NS

GLUT4 SAT. NS NS p=0.04 NS NS NS NS

INSR SAT NS NS p=0.05 NS NS NS NS

GLUR SAT NS NS p=0.02 NS p=0.05 NS NS

GLUR om. NS NS p=0.01 NS NS NS NS

GIPR SAT NS NS p=0.05 NS p=0.03 NS NS

GLP-1R SAT NS NS NS NS p=0.03 NS NS

GHRR SAT NS NS NS NS p=0.005 NS NS

CD68 SAT NS NS NS NS p=0.05 NS NS

CD163 om. NS NS p=0.05 NS NS NS NS

Il-1β SAT p=0.04 p=0.02 NS NS NS NS NS

IL-6 SAT NS p=0.03 NS NS NS NS NS

IL-6 om. NS p=0.005 NS NS NS NS NS

TNFα SAT NS NS NS NS p=0.03 NS NS

3. Direction of changes in gene expression in relation to the analyzed parameters

The relative gene expression values were ranked and the significant up- and downwards

trends of the gene expression for the analyzed parameters associated with IR such as

increased BMI, WC, HOMA and decreased HDL levels were determined. Expression of

leptin, LPL and proinflammatory genes (RETN, IL-1β, and IL-6) was increased in women

with WC above 88cm. The increased HOMA index was associated with decrease in

expression of adipokines (ADIPQ, leptin, and RETN) and several metabolic genes (PPARγ,

SREBP1c, LPL, GLUT4, INSR, GLUR and GIPR). The decreased HDL levels, similarly as

the HOMA, were associated with decreased expression of adipokines and energy metabolism

genes (SREBP1c, LPL, GLUR, GIPR, GHRR). The results are summarized in Table 5.

Table 5. The relative gene expression (RGE) mean rank value for the studied anthropometric and biochemical parameters. p-value ≤ 0.05 was considered significant and was calculated with Kruskal-Wallis test.

Chapter 4

117

Significantly, fat depot specific changed genes

Mean relative gene expression rank values for the studied biochemical and anthropometric parameters

BMI BMI<25 25<BMI<30 BMI>30 LEPTIN omentum 12.16 25.36 26.67

PPARg SAT 22.56 15.14 12 LPL SAT 25.83

19.25

12.57

14 IL1b SAT 7.3 14.5 Waist circumference (WC) cm WC<88cm WC>88cm

LEPTIN omentum 12.75 23.54 RETN SAT 15.5 22.25

LPL omentum 15.1 22.36 Il-1β SAT 6.78 12.95 IL-6 SAT 7.11 12.6

IL-6 omentum 9 13.97 HOMA HOMA<2.7 HOMA>2.7

ADIPOQ SAT 21.93 10.38 LEPTIN SAT s 24.24 14.6

RETN omentum 20.38 10.2 PPARg SAT s 21.79 8.88 SREBP1c SAT 14.06 7.83

LPL SAT 23.41 15.4 LPL omentum 23.24 14.3 GLUT4 SAT 12.4 7.5 INSR SAT 16.64 9.86 GLUR SAT 17.57 8.71

GLUR omentum 17.24 6.8 GIPR SAT 18.16

10.57

CD163 omentum 14.24 6.5 HDL (mM/L) HDL<1.3 HDL>1.3 ADIPOQ SAT 16.93 25.82 LEPTIN SAT 19.12 31.5

LEPTIN omentum 16.52 25.23 RETN SAT 17.63 28.47

SREBP1c SAT 9.5 16.7 LPL SAT 18.41 27.07

LPL omentum 18.54 25.44 GLUR SAT 13.4 19.7

GLUR omentum 12.2

21.8 GIPR SAT 13.65 21.25

GHRR SAT 14.1 24 CD68 SAT 9.36 14.29 TNFα SAT 16.72 24.85

4. Differential gene expression in SAT and omentum

Moreover, in order to explore the hypothesis about fat depot specific differences we analyzed

which genes displayed significant differential expression in SAT and omentum. The results

are summarized in Table 6.

Chapter 4

118

Table 6. Differential gene expression between SAT and omentum. The gene expression difference between SAT and omentum was significant p≤ 0.05.

Significantly, fat depot specific changed genes

Mean relative gene expression rank values for the studied biochemical and anthropometric parameters

TISSUE SAT Omentum leptin 50.77 29

PPARg 42.39 31.14 GIPR 35.64 28 GHRR 39.75 28.59 CD163 19.09 29.08

Discussion

We analyzed expression of several genes in both types of adipose tissue (SAT and omentum)

in relation to different anthropometric (BMI, WC) and biochemical (insulin, HOMA, glucose,

insulin, TG, HDL, LDL, total cholesterol) parameters in 38 women. The women used in our

study were in general good health and non-diabetic. The patients were divided into low and

high level groups in respect to each of the analyzed parameters, for example low and high

HDL level. We applied the cut off values for HDL, glucose, BMI, and WC according to the

WHO Clinical Criteria for Metabolic Syndrome [13]. The cut off values for HOMA, LDL,

total cholesterol, and insulin were based on the median value for the specific parameter

within the group because these values are not defined within criteria for metabolic syndrome

[13] and are not known for IR patients. We aimed to investigate if in our group of patients

inflammation or energy metabolism perturbations in AT were associated with early

symptoms of IR and if these changes were fat-depot specific. Therefore, we studied

expression of a subset of proinflammatory and energy metabolism genes in both SAT and

visceral adipose tissue in relation to factors indicative for IR, such as high HOMA index,

high LDL serum levels, low HDL levels, high WC, and BMI. Additionally, in order to

quantify accumulation of macrophages in AT-macrophages’ specific markers gene

expression was measured (CD68 and CD163).

The gene expression analysis revealed that patients with increased HOMA index and

decreased HDL serum levels had decreased expression of genes involved in energy

metabolism and adipokines, without changes in proinflammatory genes expression.

Assuming, that higher HOMA and lower HDL are indicative for IR, this observation

indicates that in very early stages of the development of IR a metabolic disarrangement, but

not a proinflammatory one occurs in AT. These data were in consistence with previously

Chapter 4

119

described findings; it was reported that several genes involved in insulin signaling pathway,

exampled by INSR, and other energy metabolism genes such as: SREBP1c, GLUT4, FASN

had decreased expression in obese patients compared with lean ones [15; 18; 19].

Furthermore, it was shown by Rudovich et al [20] that GIPR gene expression was decreased

in AT of IR women assessed by HOMA index. It has to be elucidated in future, if higher

HOMA and lower HDL are indeed representative for metabolic perturbations in AT. Based in

our results we speculated that HDL serum level can affect AT gene expression. To support

this hypothesis, it was recently shown that adiponectin gene expression level and adipocyte

metabolism was controlled by HDL levels [21].

Previously, it was reported that obesity and IR are associated with macrophage accumulation

responsible for upregulation of pro inflammatory cytokines [9; 22] [23]. However, in our

study we did not observe changes in AT proinflamamtory gene expression related to factors

indicating on IR such as higher HOMA, lower HDL levels or obesity. Moreover, we did not

observe increased infiltration of macrophages assessed by levels of CD68 and CD163 in

neither SAT and omentum. Thereby, this observation contradicted the hypothesis that

infiltration of macrophage into AT triggers inflammation and is the pivotal event leading to

IR in humans. However, the cytokines expression was increased in women with WC above

88 cm. The WC above 88cm is considered as one of indirect measures of IR and is one of the

criteria for the definition of metabolic syndrome, but at present a direct link between WC and

IR is unknown and WC remains a speculative measure of IR [24].

There is an ongoing scientific discussion if different fat depots have divergent physiological

functions reflected in for example differential gene expression profile. Some studies have

demonstrated that omentum exhibits more harmful profile by expressing more

proinflammatory genes than SAT [25]. However, other studies failed to confirm this

hypothesis [15], [26]. In our study, we could identify few genes such as leptin, PPARγ,

GIPR, and GHRR which were significantly higher expressed in SAT compared to omentum.

This observation can be indicative for functional differences between SAT and omentum.

Moreover, we observed that direction of the genes expression changes in SAT and omentum,

related to the investigated anthropometric and biochemical parameters, were similar, but

mostly genes in SAT and not in the omentum reached significant difference. This could mean

that SAT might be more prone to alternations of gene expression during early development of

IR, compared to omentum. This phenomenon implies that omentum and SAT have different

metabolic functions and play divergent roles in the development of IR. We also can not rule

Chapter 4

120

out the possibility, that the investigated group of patients had its unique characteristics which

can not be extrapolated to the general mechanism of the development of IR in humans.

In summary, this study demonstrates that high HOMA and low HDL are associated with

metabolic, but not proinflammatory disarrangements in the adipose tissue gene expression.

Assuming that these parameters can be used as indicators for IR, we postulate the early

events in IR include metabolic perturbations preceding inflammatory changes, thereby

excluding inflammation as key event leading to IR. How these finding can be extrapolated to

the mechanisms responsible for the development of IR in humans remains to be elucidated in

the future.

There are differences in the abundance for some genes expressed between SAT and

omentum, supporting the hypothesis that these two adipose tissue depots have different

functions. Moreover, based on the significant changes in the energy metabolism genes

expression in SAT and not in omentum, the SAT seems to be more prone to metabolic

alternations and thereby is an important player in early events associated with IR.

References

[1] R.H.Eckel, S.M.Grundy, P.Z.Zimmet, The metabolic syndrome Lancet 365, (2005) 1415-1428. [2] M.Fasshauer, S.Kralisch, M.Klier, U.Lossner, M.Bluher, A.M.Chambaut-Guerin, J.Klein, R.Paschke,

Interleukin-6 is a positive regulator of tumor necrosis factor alpha-induced adipose-related protein in 3T3-L1 adipocytes FEBS Lett. 560, (2004) 153-157.

[3] S.Schinner, W.A.Scherbaum, S.R.Bornstein, A.Barthel, Molecular mechanisms of insulin resistance Diabet.Med. 22, (2005) 674-682.

[4] M.Sowers, H.Zheng, K.Tomey, C.Karvonen-Gutierrez, M.Jannausch, X.Li, M.Yosef, J.Symons, Changes in body composition in women over six years at midlife: ovarian and chronological aging J.Clin.Endocrinol.Metab 92, (2007) 895-901.

[5] B.M.Spiegelman, J.S.Flier, Adipogenesis and obesity: rounding out the big picture Cell 87, (1996) 377-389.

[6] K.A.Virtanen, P.Iozzo, K.Hallsten, R.Huupponen, R.Parkkola, T.Janatuinen, F.Lonnqvist, T.Viljanen, T.Ronnemaa, P.Lonnroth, J.Knuuti, E.Ferrannini, P.Nuutila, Increased fat mass compensates for insulin resistance in abdominal obesity and type 2 diabetes: a positron-emitting tomography study Diabetes 54, (2005) 2720-2726.

[7] P.E.Scherer, Adipose tissue: from lipid storage compartment to endocrine organ Diabetes 55, (2006) 1537-1545.

[8] B.V.Howard, Insulin, insulin resistance, and dyslipidemia Ann.N.Y.Acad.Sci. 683, (1993) 1-8. [9] S.P.Weisberg, D.McCann, M.Desai, M.Rosenbaum, R.L.Leibel, A.W.Ferrante, Jr., Obesity is

associated with macrophage accumulation in adipose tissue J.Clin.Invest 112, (2003) 1796-1808.

[10] H.Xu, G.T.Barnes, Q.Yang, G.Tan, D.Yang, C.J.Chou, J.Sole, A.Nichols, J.S.Ross, L.A.Tartaglia, H.Chen, Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance J.Clin.Invest 112, (2003) 1821-1830.

Chapter 4

121

[11] S.Cinti, G.Mitchell, G.Barbatelli, I.Murano, E.Ceresi, E.Faloia, S.Wang, M.Fortier, A.S.Greenberg, M.S.Obin, Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans J.Lipid Res. 46, (2005) 2347-2355.

[12] K.E.Wellen, G.S.Hotamisligil, Inflammation, stress, and diabetes J.Clin.Invest 115, (2005) 1111-1119. [13] S.M.Grundy, H.B.Brewer, Jr., J.I.Cleeman, S.C.Smith, Jr., C.Lenfant, Definition of metabolic

syndrome: Report of the National Heart, Lung, and Blood Institute/American Heart Association conference on scientific issues related to definition Circulation 109, (2004) 433-438.

[14] K.Maeda, K.Okubo, I.Shimomura, K.Mizuno, Y.Matsuzawa, K.Matsubara, Analysis of an expression profile of genes in the human adipose tissue Gene 190, (1997) 227-235.

[15] M.Dolinkova, I.Dostalova, Z.Lacinova, D.Michalsky, D.Haluzikova, M.Mraz, M.Kasalicky, M.Haluzik, The endocrine profile of subcutaneous and visceral adipose tissue of obese patients Mol.Cell Endocrinol. 291, (2008) 63-70.

[16] M.W.Pfaffl, A new mathematical model for relative quantification in real-time RT-PCR Nucleic Acids Res. 29, (2001) e45.

[17] K.J.Livak, T.D.Schmittgen, Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method Methods 25, (2001) 402-408.

[18] T.E.Graham, B.B.Kahn, Tissue-specific alterations of glucose transport and molecular mechanisms of intertissue communication in obesity and type 2 diabetes Horm.Metab Res. 39, (2007) 717-721.

[19] O.Poulain-Godefroy, C.Lecoeur, F.Pattou, G.Fruhbeck, P.Froguel, Inflammation is associated with a decrease of lipogenic factors in omental fat in women Am.J.Physiol Regul.Integr.Comp Physiol 295, (2008) R1-R7.

[20] N.Rudovich, S.Kaiser, S.Engeli, M.Osterhoff, O.Gogebakan, M.Bluher, A.F.Pfeiffer, GIP receptor mRNA expression in different fat tissue depots in postmenopausal non-diabetic women Regul.Pept. 142, (2007) 138-145.

[21] L.S.Van, A.Foryst-Ludwig, F.Spillmann, J.Peng, Y.Feng, M.Meloni, C.E.Van, U.Kintscher, H.P.Schultheiss, G.B.De, C.Tschope, Impact of HDL on adipose tissue metabolism and adiponectin expression Atherosclerosis 210, (2010) 438-444.

[22] C.A.Curat, V.Wegner, C.Sengenes, A.Miranville, C.Tonus, R.Busse, A.Bouloumie, Macrophages in human visceral adipose tissue: increased accumulation in obesity and a source of resistin and visfatin Diabetologia 49, (2006) 744-747.

[23] K.Clement, D.Langin, Regulation of inflammation-related genes in human adipose tissue J.Intern.Med. 262, (2007) 422-430.

[24] S.Klein, D.B.Allison, S.B.Heymsfield, D.E.Kelley, R.L.Leibel, C.Nonas, R.Kahn, Waist circumference and cardiometabolic risk: a consensus statement from shaping America's health: Association for Weight Management and Obesity Prevention; NAASO, the Obesity Society; the American Society for Nutrition; and the American Diabetes Association Diabetes Care 30, (2007) 1647-1652.

[25] J.N.Fain, Release of interleukins and other inflammatory cytokines by human adipose tissue is enhanced in obesity and primarily due to the nonfat cells Vitam.Horm. 74, (2006) 443-477.

[26] B.L.Wajchenberg, D.Giannella-Neto, M.E.da Silva, R.F.Santos, Depot-specific hormonal characteristics of subcutaneous and visceral adipose tissue and their relation to the metabolic syndrome Horm.Metab Res. 34, (2002) 616-621.

Chapter 5

122

Chapter 5

123

Chapter 5

The “adipokine” resistin is more abundant in

human liver than in adipose tissue and it is not

upregulated by lipopolysaccharide

Ewa Szalowska

Marieke G.L. Elferink

Annemiek Hoek

Geny M.M. Groothuis

Roel J. Vonk

Journal of Clinical Endocrinology and Metabolism 2009

Chapter 5

124

Abstract

Context: Resistin is an adipokine correlated with inflammatory markers and is predictive for

cardiovascular diseases (CVD). There is evidence that serum resistin levels are elevated in

obese patients, however the role of resistin in insulin resistance (IR) and type 2 diabetes

(T2D) remains controversial.

Objective: We addressed the question whether inflammation may induce expression of

resistin in organs involved in regulation of total body energy metabolism, such as liver and

adipose tissue (AT).

Methods: Human liver tissue, subcutaneous adipose tissue (SAT) and omentum were

cultured in the absence/presence of lipopolysaccharide (LPS). The resistin and cytokine

mRNA and protein expression levels were determined by real-time PCR (RTPCR), ELISA

and Multiplex Technology respectively. The localization of resistin in human liver was

analyzed by immunohistochemistry.

Results: Resistin gene and protein expression was significantly higher in liver than in AT.

Exposure of human AT and liver tissue in culture to LPS did not alter resistin concentration;

however, concentrations of IL-1β, IL-6 and TNFα were significantly increased in these

tissues. In liver resistin colocalizes with markers for Kupffer cells, for a subset of endothelial

and fibroblasts like cells.

Conclusions: High level of resistin gene and protein expression in liver compared to AT

implicates that resistin should not be considered only as an adipokine in humans.

LPS induced-inflammation does not affect resistin protein synthesis in human liver and AT.

This suggests that elevated serum resistin levels are not indicative for inflammation of AT or

liver in a manner similar to known inflammatory markers such as IL-1β, IL-6 or TNFα.

Keywords: resistin, inflammation, insulin resistance, T2D, CVD, human liver and adipose

tissue

Chapter 5

125

Introduction

Human resistin (12.5-kD) is known as an adipokine and belongs to a family of small

cysteine-rich secretory proteins, named FIZZ (Found In Inflammatory Zone) or resistin-like

molecules (1). The human resistin gene (Retn) is a homolog of the mouse resistin gene. In

mice resistin is almost exclusively expressed in white adipose tissue and has been shown to

be involved in glucose intolerance and hepatic IR (1-3). These findings suggested that resistin

could be a link between obesity and T2D. However most of the human studies failed to show

this link (4-6) and only a few studies showed that resistin serum levels are elevated in obese

patients and could be related to IR and T2D (7,8).

IR and T2D are associated with obesity and low grade inflammation with upregulated

cytokines and chemokines (IL-1β, IL-6, IL-8 and TNFα). LPS is a compound of the cell wall

of Gram-negative bacteria that has been demonstrated to induce inflammatory reactions and

upregulate many cyto- and chemokines. Besides its role in inflammation, LPS is known to

trigger hyperglycemia and IR in rats and humans (9-11). Recently it was shown that mice

infused with LPS and on a chow diet gain body weight, develop liver IR and have a

dysregulated inflammatory tone which leads to T2D (12,13). Conflicting data are published

concerning the regulation of resistin mRNA by proinflammatory agents such as LPS, TNF-α,

IL-6 both for rodents and humans (2,14-20). In humans serum resistin level is correlated with

markers of inflammation and is predictive for CVD (2,21).

Recently, resistin was found to be expressed at a low level in human liver and to be

upregulated during severe fibrosis (19) and in hepatocytes where it induces IR (22). In

addition it was reported that adipose tissue of patients with nonalcoholic fatty liver disease

had significantly higher resistin expression than healthy individuals and obese people, and

resistin serum levels were positively correlated with severity of the liver disease (20). These

results suggest that resistin expression may be induced indirectly by LPS possibly as a result

of increased cytokine exposure.

In the present study we compared resistin mRNA and protein expression in organs involved

in regulation of total body energy metabolism: liver and adipose tissue (subcutaneous adipose

tissue= SAT and omentum=om.) and identified by immunohistochemistry in which liver cells

resistin was expressed. Furthermore we tested if resistin is upregulated during LPS-induced-

inflammation in human liver, SAT and omentum. We studied resistin and cytokine gene

Chapter 5

126

expression, protein synthesis and secretion upon stimulation with LPS in cultured human

liver slices, SAT and omentum.

Materials and Methods

Human liver tissue

Human liver tissue (n=20) was obtained and prepared as described previously (23,24). The

donors of livers for the quantification experiments and immunohistochemistry stainings were

females, aged 40-49 years, with body mass index (BMI) 18.3-27.7. The donors of livers used

in LPS experiments were males, aged 16-48, with BMI 23.1-27.7. The information about the

medical history was not available. The liver tissue used for quantification studies by RTPCR

or ELISA was snap frozen in liquid nitrogen and stored at -80°C. For immunohistochemistry,

tissue was frozen in isopentane at -80°C and stored at -80°C. The research protocols were

approved by the Medical Ethical Committee of the UMCG and the donors gave their

informed consent.

Preparation and incubation of liver slices

Human liver slices were prepared and incubated as described previously (24,25). Liver slices

were incubated at 37°C in Williams Medium E in the presence or absence of 100 µg/ml LPS.

At 5, 24, and 48 h after incubation, slices were frozen in liquid nitrogen and, as the media

samples, and stored at –80°C.

Human adipose tissue

In the study SAT and omentum biopsies were obtained from in total 21 Caucasian women

undergoing surgery because of benign gynecological problems. The women were in general

good health, had no history or symptoms of T2D or inflammatory diseases. The subjects were

aged between 30 and 45 years, with BMI ranging from 23 to 29. The SAT biopsies were

taken at the place of the incision, midline lower abdomen above the symfysis and under the

umbilicus. The omentum biopsies were taken at the lower edge of the omentum. Biopsies

were taken by means of scissors. The AT biopsies used for quantification studies by RTPCR

or ELISA were snap frozen in liquid nitrogen and stored in -80°C. The research protocols

Chapter 5

127

were approved by the Medical Ethical Committee of the UMCG and patients gave written

informed consent.

Adipose tissue culture

The human AT surgical biopsies were processed as described previously (26). The resulting

pieces of fat were cultured in six-well plates. Per well 0.5 g of fat tissue was cultured in 5ml

M199 medium (Gibco) supplemented with 50 µg/ml Gentamycine (Sigma) for 24 hrs at 37ºC

in 5% CO2 in the absence and presence of 100 µg/ml LPS. After the incubation the tissues

were frozen in liquid nitrogen and, as the media, stored in -80ºC.

Adipose tissue and liver tissue extracts

50 mg of liver tissue and 300 mg of fat tissue were homogenized in 500 µl lysis buffer

(Sigma) supplemented with inhibitors of proteases (Complete, Roche) in a homogenizer

(Precellys 24, Bertin Technology) with settings: 6500g, 2x 15 sec, 8 ºC. The tissue extracts

were centrifuged 5 min 12000g at 8°C and the resulting supernatant was aliquoted and stored

in -80ºC.

Protein assay

Protein concentration was determined with the Bradford assay according to the manufactures

instructions (Biorad).

Cytokine and resistin analysis

TNFα, IL-1β and IL-6 were analyzed by Bio-Plex Human Cytokine Assay manufactured by

BioRad (The Netherlands). The detection limits for IL-1β, IL-6, and TNFα were 0.8, 1.1, and

3 respectively. The intra- and inter-assay variation for IL-1β, IL-6, and TNFα were 2, 6, 5%

and 8.6, 7.2, and 5.3% respectively. Resistin was measured with Human Resistin Elisa Kit,

manufactured by LINCO (The Netherlands). The detection limit was 0.312 ng/ml and the

intra- and inter-assay variation was 4 and 18% respectively. The measurements were

performed according to the manufactures’ protocols.

RNA isolation and cDNA synthesis

RNA was extracted from AT using RNeasy Lipid Tissue Mini Kit (Qiagen, Venlo, the

Netherlands) according to the manufacturers’ instructions. RNA extraction from human liver

slices was performed as described previously (24). The RNA concentration was determined

Chapter 5

128

by Nano Drop ND-1000 Spectrophotometer (Isogen Ijsselstein, The Netherlands). cDNA

synthesis was performed from total RNA with QuantiTect Reverse Transcription Kit (Qiagen,

Venlo, the Netherlands) according to the manufacturers’ instructions.

Real-time quantitative PCR

Quantification of the expression of the genes of interest was done by QPCR on a ABI prism

7900 HT (Applied Biosystems) with the following cycling conditions: 15 min 95°C followed

by 40 cycles of 15s 95 ºC and 1 min 60 ºC. Reactions were performed in 10 µl volume and

contained 20 ng cDNA, 1x TaqMan PCR Master Mix (Applied Biosystems, Foster City, CA),

250 nM of each probe and 900 nM of each primer. For each gene, a standard curve was

generated and efficiency of the primers was determined as described previously (27). For

each primer pair the efficiency was about 95 %. Specific primer sets for β-2-microglobulin

(B2M), actinβ (ACTB), and ribosomal protein S18 (RPS18), IL-1β, IL-6, TNFα and resistin

were developed with Primer Express 1.5 (Applied Biosystems). Primers for CRP were

purchased from Applied Biosystems. The sequences for forward (F), reverse (R) primers and

the probe (P) were (5’-3’): RPS18 (F)CGGCTACCACATCCAAGGA;

(R)CCAATTACAGGGCCTCGAAA; (P)CGCGCAAATTACCCACTCCCGA; B2M

(F)TGACTTTGTCACAGCCCAAGATA; (R)AATCCAAATGCGGCATCTTC;

(P)TGATGCTGCTTACATGTCTCGATCCCA; ACTB (F)AGCGCGGCTACAGCTTCA;

(R)CGTAGCACAGCTTCTCCTTAATGTC; (P)ATTTCCCGCTCGGCCGTGGT; IL-1 β

(F)CTGATGGCCCTAAACAGATGAAG; (R)GGTCGGAGATTCGTAGC;

(P)TTCCAGGACCTGGACCTCTGCCCTC ; IL-6 (F)CCAGGAGCCCAGCTATGAAC ;

(R)CCCAGGGAGAAGGCAACTG; (P)CCTTCTCCACAAGCGCCTTCGGT; TNFα

(F)CTCGAACCCCGAGTGACAA; (R)AGCTGCCCCTCAGCTTGA;

(P)CCTGTAGCCCATGTTGTAGCAAACC; Resistin (F)GCCGGATTTGGTTAGCTGAG;

(R)GAGGAGGAGACAGAGAGCTTTCAT; (P) CCACCGAGAGGCGCCTGCAG;

Data were analyzed with SDS 2.0 software (Applied Biosystems). For each sample, the RT-

PCR reaction was performed twice in triplicate and the averages of the obtained threshold

cycle values (CT) were processed for further calculations. For normalization B2M, ACTB,

and RPS18 were used. Relative expression was calculated with ∆ (∆ (CT))-method (28).

Chapter 5

129

Cryosectioning and immunohistochemistry staining

Cryostat sections were prepared from human liver tissue as described previously (24). For

colocalization studies the rabbit anti-human-resistin 1:20 (Santa Cruz) was incubated with a

mouse anti-human- CD68 1:150 (DAKO ) or mouse anti-human CD31 1:50 DAKO) or

mouse anti-human actin 1:400 (Sigma). The antibodies were diluted in phosphate buffered

saline (PBS) supplemented with 5% normal human serum and applied 1 h in a humidity

chamber at RT. A single primary antibody for control staining was applied with the same

dilutions. After being rinsed with PBS endogenous peroxidase blocking was conducted with

0.037% H2O2 in PBS for 1 h. Afterwards sections were washed with PBS and sections

incubated with anti-resistin were incubated with secondary antibodies goat anti-rabbit IgG

conjugated with peroxidase (GARPO; DAKO) diluted 1:100 in PBS, supplemented with 5%

normal human serum for 30 minutes in a humidity chamber. After PBS washing the same

sections were incubated with tertiary antibodies rabbit anti-goat conjugated with peroxidase

(RAGPO, DAKO) diluted 1:100 in PBS with 5% normal human serum for 30 minutes in a

humidity chamber at RT. The peroxidase activity was visualized using 3, 3-

diamonobenzidine tetrahydrochloride (DAKO) for 10 min. Afterwards sections were washed

in PBS and incubated briefly with 0.1M Tris/HCl, 2mM MgCl2 pH 8.2. Sections stained with

anti-CD31, anti-CD68 and anti-actin were subsequently incubated with goat anti-mouse

conjugated with alkaline phosphatase (GAMAF, DAKO) 1:100 in PBS supplemented with

5% normal human serum for 30 minutes in a humidity chamber at RT. After washing in

PBS, the alkaline phosphatase reaction was conducted in a buffer containing 100ml 0.1M

Tris HCl pH 8.2 with 2 mM MgCl2, 20 mg Napthol AS-MX phosphate, 100 mg Fast Blue BB

and 48 mg levamisole for 30 minutes in a 37°C water bath. After the final PBS washing step,

the sections were covered with gelatin and a cover glass.

Light microscopy

Light microscopy images were taken with Olympus BX41 using Olympus Soft Imaging

System Cell^D at the magnification of X40.

Statistical analysis

QRTPCR experiments in Results 3 were performed with n=5 human livers and n=7 AT.

QRTPCR experiments in Results 3 were performed with n=10 livers and n=10 AT.

Immunostainings were performed in n=5 different livers. Cytokine analysis and resistin

Chapter 5

130

measurements in cultured tissues were done in duplicate for n=4 livers and n=4 AT. Results

were compared using Kruskal Wallis (KW) test or paired t-test; p value <0.05 was considered

significant.

Results

1. Relative gene expression and protein abundance of resistin in human liver and

AT

To determine resistin expression 5 human liver tissues and 7 AT biopsies (SAT and

omentum) were used. The housekeeping genes used in the normalization process were

selected according to the algorithm described in GeNorm (29). The best housekeeping genes

were B2M and ACTB. In addition we used the commonly applied RPS18 for normalization.

In human liver there was on average 8-fold (p= 0.011 KW), 23-fold (p=0.03 KW) and 3.5-

fold (p= 0.036 KW) higher expression of resistin then in SAT and omentum when normalized

with B2M (Fig. 1A), ACTB (Fig 1B), and RPS18 (Fig. 1C) respectively. There was no

significant difference between SAT and omentum in resistin mRNA expression. Resistin

protein levels were analyzed in both types of AT and liver extracts. The resistin

concentrations were on average 18-fold higher in liver than in SAT and omentum (p=0.0009

and p=0.008 respectively KW) while normalized per mg of tissue (Fig 1D). The resistin

concentration normalized on 1 µg of total protein was on average 2.6-times higher in liver

then in SAT and omentum (p=0.038, and p=0.05 resp. KW, Fig. 1E). There was no

significant difference in resistin protein abundance between SAT and omentum while

normalized on mg of tissue or µg of total protein (Fig 1D and 1E respectively).

Chapter 5

131

*#

0.00

1.00

2.00

3.00

4.00

SAT om. liver

resi

stin

ng

pe

r m

g o

f

tiss

ue

snap frozen biopsies

*#

0.0

10.0

20.0

30.0

SAT om. liver

resi

stin

ng

pe

r μ

g o

f

pro

tein

snap frozen biopsies

Figure 1.The relative gene expression of resistin in human adipose tissue (snap frozen biopsies) and liver (snap

frozen biopsies) normalized with B2M (1A), ACTB (1B) and RPS18 (1C). The expression in liver is expressed

relative to that in adipose tissue, which was set to 1. Resistin concentration in tissue extracts of adipose tissue

and liver tissue was normalized per mg of tissue weight (1D) and per µg of protein (1E). Kruskal-Wallis test was

used to calculate p-value signifying the difference in resistin gene expression between adipose tissue (n=7) and

liver (n=5) and resistin concentration between fat tissue (n=7) and liver extracts (n=5); p-value below 0.05 was

considered significant and indicated with * (liver vs. SAT) and with # (liver tissue vs. omentum).

2. Localisation of resistin in human liver

Immunohistochemical staining shows that resistin is present in Kupffer cells (colocalization

with CD68 a marker for the liver residual macrophages) (Fig. 2A), in a subset of endothelial

cells stained with CD31 (Fig. 2B) and occasionally in actin positive fibroblast like cells

colocalizing with actin beta (Fig 2 C).

snap frozen biopsies

*#

0

2

4

6

8

10

SAT om. liver

Re

lati

ve

ge

ne

ex

pre

ssio

n

vs

B2

M

snap frozen biopsies

*#

0

10

20

30

40

SAT om. liverR

ela

tiv

e e

xp

ress

ion

vs

AC

TB

snap frozen biopsies

*#

0

1

2

3

4

5

SAT om. liver

Re

lati

ve

ex

pre

ssio

n v

s

18

S

Fig. 1E

Fig. 1A

Fig. 1D

Fig. 1C Fig. 1B

Chapter 5

132

Figure 2. Light microscopy pictures of double immunostaining for resistin (red) with (A) Kupffer cells marker CD68 (blue); (B) endothelial cells marker CD31 (blue); (C) actin positive fibroblasts like cells (blue) in untreated liver tissue (isopentane frozen biopsies). The arrows indicate selected examples of double staining of resistin with one of the liver tissue cells markers.

3. Influence of LPS on cytokine and C-reactive protein (CRP) gene expression in

human liver and adipose tissue culture

After 24 h of incubation with LPS IL-1β and IL-6 were significantly upregulated in liver (Fig

3A and 3B resp.). TNFα was significantly upregulated after 5 h in liver (Fig 3C) and after 24

h the TNFα expression returned to the normal level (Fig 3D). SAT and omentum responded

to LPS in significant upregulation of IL-1β, IL-6, and TNFα, Figures 3 E-J. CRP was not

significantly changed in liver, SAT and omentum at any of the tested time points (data not

shown).

Chapter 5

133

Figure 3. The effect of LPS on the mRNA level of IL-1β (3A), IL-6 (3B), TNFα 5h (3C), and TNFα 24h (3D) in human liver slices. In SAT the effect of LPS on the mRNA level of IL-1β, IL-6, and TNFα is depicted in figures 3E-3G respectively. In omentum the effect of LPS on the mRNA level of IL-1β, IL-6, and TNFα is depicted in figures 3H-3J respectively. A paired t-test was used to calculate p-value signifying the difference between untreated and treated tissues; p-value below 0.05 was considered significant and indicated with *.

liver *

0

10

20

30

40

50

60

control LPS

Fo

ld i

nd

uc

tio

n I

L-1

β 2

4h

liver*

-1

0

1

2

3

4

5

6

7

control LPS

Fo

ld i

nd

uc

tio

n I

L-6

24

h

liver *

0

1

2

3

4

5

6

7

control LPS

Fo

ld i

nd

uc

tio

n T

NF

α 5

h

liver

0

0.5

1

1.5

2

2.5

control LPS

Fo

ld i

nd

uc

tio

n T

NF

α 2

4h

SAT *

0

50

100

150

200

cont. LPS

Fo

ld i

nd

uc

tio

n I

L-1

β

SAT*

0

2

4

6

8

10

12

14

16

cont. LPS

Fo

ld i

nd

uc

tio

n I

L-6

SAT*

0

1

2

3

4

5

6

cont. LPS

Fo

ld i

nd

uc

tio

n T

NF

α

omentum*

0

5

10

15

20

25

30

35

40

45

cont. LPS

Fo

ld i

nd

uc

tio

n I

L-1

β

omentum*

0

2

4

6

8

10

12

14

16

cont. LPS

Fo

ld i

nd

uc

tio

n I

L-6

omentum*

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

cont. LPS

Fo

ld i

nd

uc

tio

n T

NF

α

Fig. 3J Fig. 3I Fig. 3H

Fig. 3G Fig. 3F Fig. 3E

Fig. 3D Fig. 3C Fig. 3B Fig. 3A

Chapter 5

134

4. Influence of LPS on cytokine protein synthesis and secretion in human liver and

adipose tissue culture

After 24 h of LPS treatment IL-1β (4A), IL-6 (4B), and TNFα (4C) were significantly

upregulated in liver slices. The secretion of IL-1β, IL-6 and TNFα in liver culture media was

also significantly increased upon LPS treatment in comparison to untreated culture media

(Fig 4D, 4E, and 4F respectively). The LPS treatment in SAT and omentum cultures

significantly increased concentration of IL-1β, IL-6, and TNFα in tissues extracts (5A, 5C, 5E

resp.), and media (Fig 5B, 5D, and 5F resp.).

Figure 4. The effect of 24h LPS treatment on the IL-1β (A, D), IL-6 (B, E), and TNFα (C, F) protein

concentration in liver slices (A, B, C) and liver culture media (D, E, F). The liver slices were cultured for 24 h

in the absence or presence of LPS (100µg/ml). The paired t-test was used to calculate p-value signifying the

difference between untreated and treated media or slices; p-value equal or below 0.05 was considered significant

and indicated with *.

*

0

200

400

600

800

control LPS

IL-1

βp

g/m

g

liver slices

Fig. 4D

Fig. 4CFig. 4BFig. 4A

*

0

500

1000

1500

2000

control LPS

IL-6

pg

/mg

liver slices

*

0

100

200

300

400

control LPS

TN

Fα

pg

/mg

liver slices

*

0

100

200

300

400

control LPS

IL-6

pg

/mg

liver media*

0

1

2

3

4

5

6

control LPS

IL-1

βp

g/m

l

liver media

Fig. 4FFig. 4E

*

0

10

20

30

40

50

control LPS

TN

Fα

pg

/mg

liver media

Chapter 5

135

*

#

0

50

100

150

200

250

300

cont. LPS

IL-1

β p

g/m

lAT extracts

SAT

omentum*

#

0

5

10

15

20

cont. LPS

IL-1

β p

g/m

l

AT media

SAT

omentum

*

#

0

100

200

300

400

cont. LPS

IL-6

pg

/ml

AT extracts

SAT

omentum *

#

0

50

100

150

200

cont. LPS

IL-6

pg

/ml

AT media

SAT

omentum

*

#

0

10

20

30

cont. LPS

TN

Fα

pg

/ml

AT extracts

SAT

omentum

*

#

-5

0

5

10

15

cont. LPS

TN

Fα

pg

/ml

AT media

SAT

omentum

Figure 5. The effect of LPS on the IL-1β (5A, 5B), IL-6 (5C, 5D), and TNFα (5E, 5F) protein concentration in AT extracts and AT culture media. Human SAT and omentum explants (5A, 5C, 5E), and SAT and omentum culture media (5B, 5D, 5F) were cultured for 24 h in the absence or presence of LPS (100µg/ml). The paired t-test was used to calculate p-value signifying the difference between untreated and treated tissues or media; p-value equal or below 0.05 was considered significant and indicated with * (control SAT vs. LPS treated SAT) and with # (control omentum vs. LPS treated omentum).

Fig. 5F Fig. 5E

Fig. 5D Fig. 5C

Fig. 5B Fig. 5A

Chapter 5

136

5. Influence of LPS on resistin gene expression and protein level in adipose tissue

and liver

Resistin gene expression was measured in SAT and omentum after 24 h of LPS treatment and

in liver slices after 5, 24 and 48 h of LPS treatment. In SAT and omentum resistin mRNA

expression was significantly upregulated 16-fold, p=0.009 and 7-fold, p=0.005 respectively

(Fig 6 B). In liver slices there was no significant effect of LPS on resistin gene expression

after 5, 24 and 48 h (Fig 6A) although in 2 out of 10 livers an upregulation of 2.5- till 7- fold

at 24 and 48 hrs was observed. Resistin protein level in liver and AT extracts (Fig 6C, 6D

resp.) and concentration in liver and AT culture media was not significantly altered by the 24

h LPS treatment (Fig 6E and F respectively).

Chapter 5

137

0

1

2

3

4

cont. LPS

Fold

in

du

ctio

n r

esi

stin

liver slices

5h

24h

48h

0

0.2

0.4

0.6

0.8

1

1.2

1.4

cont. LPS

res

isti

n n

g/m

l

liver media

Figure 6. The effect of LPS on the resistin mRNA expression in liver (6A), SAT and omentum (6B). The liver slices were treated without/with LPS for 5, 24, and 48 hrs. The SAT and omentum were treated without/with LPS for 24hrs. The resistin protein level in liver slice extracts (6C), SAT and omentum (6D) extracts upon 24 hrs of incubation in absence or presence of LPS. The resistin protein level in liver culture media (6E), SAT and omentum culture media (6F) upon 24 hrs of incubation in absence or presence of LPS. The paired t-test was used to calculate p-value signifying the difference between untreated and treated tissues; p-value equal or below 0.05 was considered significant and indicated with * (SAT control vs. SAT LPS treated) and # (omentum control vs. omentum LPS treatment).

*

#

0

5

10

15

20

25

con. 24h LPS 24hFo

ld i

nd

uc

tio

n r

esi

stin

AT

SAT

omentum

0

1

2

3

4

5

cont. 24h LPS 24h

resi

stin

ng

/mg

liver slices

Fig. 6D Fig. 6C

Fig. 6B Fig. 6A

0

0.02

0.04

0.06

0.08

0.1

cont. 24h LPS 24h

resi

stin

ng

/ml

AT extracts

SAT

omentum

0

0.001

0.002

0.003

0.004

cont. LPS

res

isti

n n

g/m

l

AT media

SAT

Fig. 6F Fig. 6E

Chapter 5

138

Discussion

In the present study we aimed to (1) quantify resistin mRNA and protein abundance in human

liver, SAT and omentum; (2) identify in which cell types in the liver resistin protein is

localized, and (3) study resistin gene and protein regulation during LPS-induced

inflammation, in liver, SAT and omentum.

Resistin mRNA showed significantly higher relative gene expression in liver compared to

SAT and omentum. The resistin gene mRNA expression in SAT and omentum was not

significantly different. We realize that relative quantification in two different organs (liver

and AT) is problematic because (1) it is very difficult to find a housekeeping gene that is

equally expressed in the two different tissues, and (2) the selection of the commonly used

housekeeping genes is based on assumptions and observations made in different biological

systems (30). In order to diminish this problem we decided to use 3 different housekeeping

genes. Moreover the unnormalized Ct values obtained for the liver (avg. Ct = 29, n=5) were

lower than in AT (avg. Ct=34, n=7), when the same amount of cDNA is used, which supports

our normalized gene expression data and indicatives higher resistin expression in liver than in

AT. We also compared the CD68 mRNA level in AT (both SAT and omentum) and liver and

there was no significant difference between any of these tissues (data not shown), indicating

that the difference in resistin expression between AT and liver is not caused by a difference in

the number of macrophages/Kupffer cells coexpressing resistin.

The protein data obtained for liver and AT extracts are in consistence with the gene

expression data and demonstrate that resistin is more abundant in the liver. The resistin level

in liver tissue extracts normalized per mg of tissue is 18- fold higher than in SAT and

omentum and the resistin level in liver tissue extracts normalized per total protein is 2.5-fold

higher in liver than in both types of AT. As the amount of protein per mg tissue differs

widely between fat and liver tissue it seems more relevant to normalize on weight of tissue

instead of normalizing on total protein level for a proper estimation of the physiological

levels of resistin in these tissues.

The localization studies showed presence of resistin in Kupffer cells (90% of the cells were

positive). This finding was not unexpected because resistin is also present in human

circulating macrophages and macrophages of AT (6,31). Previously it was suggested that

hepatocytes could produce resistin (19,22), but we could not confirm this observation.

Furthermore, we detected resistin in a subset of endothelial cells (5%). This finding is

Chapter 5

139

consistent with data linking resistin with inflammation and a profound role of endothelial

cells in inflammatory processes. In addition it was previously shown that resistin induces

human endothelial cell proliferation and migration, promotes capillary-like tube formation

thereby implying a role in angiogenesis-associated vascular disorders (21,32). The presence

of resistin in actin positive fibroblasts cells, which can become activated stellate cells during

inflammation, is in agreement with former findings showing that resistin expression is linked

to areas of inflammatory cell accumulation (19).

Our next aim was to challenge SAT, omentum and liver with LPS in order to induce an

inflammatory reaction and analyze if resistin expression both on mRNA and protein level is

altered. We previously validated the human liver slice model for study of inflammatory

reactions (24,33,34). The liver slices cultured with LPS for 24 h significantly upregulated IL-

1β and IL-6. TNFα was significantly upregulated after 5 h but after 24 h the level returned to

basal, confirming earlier data (24,34). The cytokines mRNA expression data were consistent

with the protein data. Upon LPS treatment we detected significant upregulation of IL-1β, IL-

6 and TNFα in both liver slices and culture media. In our study CRP, the classical acute phase

protein and the most extensively studied systemic marker of inflammation (35,36) was not

upregulated by LPS treatment at any of the tested time points on mRNA level either. In

contrast, it was shown previously that CRP was upregulated by LPS in peripheral blood

mononuclear cells (PBMC) both on mRNA and protein level in vivo (37). Previously we

found that CRP is upregulated in human liver slices by incubation with high concentration of

TNFα (15 ng/ml) (33) . Therefore we speculate that in the present study, the level of TNFα in

the liver tissue does not reach the level necessary for CRP induction.

The LPS-treatment of SAT and omentum for 24 hrs resulted in upregulation of mRNA for

IL-1β, IL-6 and TNFα. Also in AT the expression of CRP was not affected by LPS. The

mRNA data for IL-1β, IL-6 and TNFα were consistent with the protein data obtained for

cultured SAT and omentum and culture media. Based on these results we concluded that

treatment of AT in vitro with LPS is sufficient to induce an inflammatory reaction and

mimics inflammatory processes in vivo with upregulated cytokines. Furthermore LPS-treated

omentum synthesized and secreted significantly higher levels of cytokines than SAT what is

in line with the literature data pointing out towards more proinflammatory character of

omentum than SAT (38-40).

SAT and omentum treated for 24h with LPS responded with a significant upregulation of

resistin mRNA. However the resistin protein level in SAT and omentum tissue extracts and

Chapter 5

140

media was not affected by LPS. In the liver we did not detect a significant effect of LPS on

resistin gene expression and protein level in tissue or media at any of the tested time points.

Based on these results we could not find evidence that resistin is significantly upregulated in

human liver during 24hrs LPS-induced inflammation despite upregulation of IL-1β, IL-6 or

TNFα.

On the other hand we observed 2 livers responded to LPS by increased resistin protein

secretion after 48 hrs (5- and 3-fold), but not after 24h. The resistin mRNA expression was

increased in the same livers after 24 and 48 hrs (on average 2.5- and 7-fold respectively).

Although these results were obtained in only 2 out of 10 livers and were not significant for

n=10, this observation suggests that resistin could indeed be an inflammatory marker but a

conditional one, because the increase in resistin mRNA expression/synthesis occurred in

some livers only. The finding that resistin mRNA was significantly upregulated by LPS in

SAT and omentum without any protein changes might be explained by assuming that also in

AT resistin protein production occurs at 48hrs. We also could not exclude a proteolytic

resistin degradation, but this would imply increased proteolytic activity after the LPS

treatment compared to untreated tissues/media.

In summary, the remarkable high resistin mRNA expression and protein content in fresh liver

tissue compared with SAT and omentum suggests that resistin should not be considered only

as an adipokine and that the resistin-related research should be more committed towards its

role in the liver physiology. Furthermore higher resistin secretion in liver slices during

incubation compared to AT indicates that liver may contribute more to the resistin blood level

than AT. In the liver resistin is present in Kupffer cells (~90%), a subset of endothelial cells

(~10%) and actin positive fibroblasts like cells (~5%). During LPS induced inflammation

omentum and in a minor extend SAT, in addition to the liver, contribute to the total blood

concentrations of IL-1β, IL-6 and TNFα which are indicative for inflammation in vitro and in

vivo. Resistin mRNA expression was significantly upregulated during inflammatory reactions

in SAT and omentum, but not liver tissue. Further research is needed to reveal the exact

mechanism of the resistin regulation. Therefore the application of resistin as an inflammatory

marker in T2D and CVD should be considered critically.

Chapter 5

141

Acknowledgements

We would like to thank to Anne-miek van Loenen-Weemaes and Alie de Jager-Krikken for

excellent technical assistance and Prof Klaas Poelstra discussing the light microscopy images.

References

1. Steppan CM, Bailey ST, Bhat S, Brown EJ, Banerjee RR, Wright CM, Patel HR, Ahima RS, Lazar MA

2001 The hormone resistin links obesity to diabetes. Nature 409:307-312. 2. Pang SS, Le YY 2006 Role of resistin in inflammation and inflammation-related diseases. Cell Mol

Immunol 3:29-34. 3. Muse ED, Lam TK, Scherer PE, Rossetti L 2007 Hypothalamic resistin induces hepatic insulin

resistance. J Clin Invest 117:1670-1678. 4. Savage DB, Sewter CP, Klenk ES, Segal DG, Vidal-Puig A, Considine RV, O'Rahilly S 2001 Resistin

/ Fizz3 expression in relation to obesity and peroxisome proliferator-activated receptor-gamma action in humans. Diabetes 50:2199-2202.

5. Nagaev I, Smith U 2001 Insulin resistance and type 2 diabetes are not related to resistin expression in human fat cells or skeletal muscle. Biochem Biophys Res Commun 285:561-564.

6. Patel L, Buckels AC, Kinghorn IJ, Murdock PR, Holbrook JD, Plumpton C, Macphee CH, Smith SA 2003 Resistin is expressed in human macrophages and directly regulated by PPAR gamma activators. Biochem Biophys Res Commun 300:472-476.

7. McTernan PG, McTernan CL, Chetty R, Jenner K, Fisher FM, Lauer MN, Crocker J, Barnett AH, Kumar S 2002 Increased resistin gene and protein expression in human abdominal adipose tissue. J Clin Endocrinol Metab 87:2407.

8. McTernan CL, McTernan PG, Harte AL, Levick PL, Barnett AH, Kumar S 2002 Resistin, central obesity, and type 2 diabetes. Lancet 359:46-47.

9. Virkamaki A, Yki-Jarvinen H 1994 Mechanisms of insulin resistance during acute endotoxemia. Endocrinology 134:2072-2078.

10. Sugita H, Kaneki M, Tokunaga E, Sugita M, Koike C, Yasuhara S, Tompkins RG, Martyn JA 2002 Inducible nitric oxide synthase plays a role in LPS-induced hyperglycemia and insulin resistance. Am J Physiol Endocrinol Metab 282:E386-E394.

11. Agwunobi AO, Reid C, Maycock P, Little RA, Carlson GL 2000 Insulin resistance and substrate utilization in human endotoxemia. J Clin Endocrinol Metab 85:3770-3778.

12. Cani PD, Amar J, Iglesias MA, Poggi M, Knauf C, Bastelica D, Neyrinck AM, Fava F, Tuohy KM, Chabo C, Waget A, Delmee E, Cousin B, Sulpice T, Chamontin B, Ferrieres J, Tanti JF, Gibson GR, Casteilla L, Delzenne NM, Alessi MC, Burcelin R 2007 Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes 56:1761-1772.

13. Cani PD, Bibiloni R, Knauf C, Waget A, Neyrinck AM, Delzenne NM, Burcelin R 2008 Changes in gut microbiota control metabolic endotoxemia-induced inflammation in high-fat diet-induced obesity and diabetes in mice. Diabetes 57:1470-1481.

14. Kaser S, Kaser A, Sandhofer A, Ebenbichler CF, Tilg H, Patsch JR 2003 Resistin messenger-RNA expression is increased by proinflammatory cytokines in vitro. Biochem Biophys Res Commun 309:286-290.

Chapter 5

142

15. Lu SC, Shieh WY, Chen CY, Hsu SC, Chen HL 2002 Lipopolysaccharide increases resistin gene expression in vivo and in vitro. FEBS Lett 530:158-162.

16. Lehrke M, Reilly MP, Millington SC, Iqbal N, Rader DJ, Lazar MA 2004 An inflammatory cascade leading to hyperresistinemia in humans. PLoS Med 1:e45.

17. Rajala MW, Lin Y, Ranalletta M, Yang XM, Qian H, Gingerich R, Barzilai N, Scherer PE 2002 Cell type-specific expression and coregulation of murine resistin and resistin-like molecule-alpha in adipose tissue. Mol Endocrinol 16:1920-1930.

18. Fasshauer M, Klein J, Neumann S, Eszlinger M, Paschke R 2001 Tumor necrosis factor alpha is a negative regulator of resistin gene expression and secretion in 3T3-L1 adipocytes. Biochem Biophys Res Commun 288:1027-1031.

19. Bertolani C, Sancho-Bru P, Failli P, Bataller R, Aleffi S, DeFranco R, Mazzinghi B, Romagnani P, Milani S, Gines P, Colmenero J, Parola M, Gelmini S, Tarquini R, Laffi G, Pinzani M, Marra F 2006 Resistin as an intrahepatic cytokine: overexpression during chronic injury and induction of proinflammatory actions in hepatic stellate cells. Am J Pathol 169:2042-2053.

20. Pagano C, Soardo G, Pilon C, Milocco C, Basan L, Milan G, Donnini D, Faggian D, Mussap M, Plebani M, Avellini C, Federspil G, Sechi LA, Vettor R 2006 Increased serum resistin in nonalcoholic fatty liver disease is related to liver disease severity and not to insulin resistance. J Clin Endocrinol Metab 91:1081-1086.

21. Reilly MP, Lehrke M, Wolfe ML, Rohatgi A, Lazar MA, Rader DJ 2005 Resistin is an inflammatory marker of atherosclerosis in humans. Circulation 111:932-939.

22. Sheng CH, Di J, Jin Y, Zhang YC, Wu M, Sun Y, Zhang GZ 2008 Resistin is expressed in human hepatocytes and induces insulin resistance. Endocrine 33:135-143.

23. Brouwers MA, Peeters PM, de Jong KP, Haagsma EB, Klompmaker IJ, Bijleveld CM, Zwaveling JH, Slooff MJ 1997 Surgical treatment of giant haemangioma of the liver. Br J Surg 84:314-316.

24. Elferink MG, Olinga P, Draaisma AL, Merema MT, Faber KN, Slooff MJ, Meijer DK, Groothuis GM 2004 LPS-induced downregulation of MRP2 and BSEP in human liver is due to a posttranscriptional process. Am J Physiol Gastrointest Liver Physiol 287:G1008-G1016.

25. Olinga P, Merema M, Hof IH, de Jong KP, Slooff MJ, Meijer DK, Groothuis GM 1998 Effect of human liver source on the functionality of isolated hepatocytes and liver slices. Drug Metab Dispos 26:5-11.

26. Alvarez-Llamas G, Szalowska E, de Vries MP, Weening D, Landman K, Hoek A, Wolffenbuttel BH, Roelofsen H, Vonk RJ 2007 Characterization of the human visceral adipose tissue secretome. Mol Cell Proteomics 6:589-600.

27. Pfaffl MW 2001 A new mathematical model for relative quantification in real-time RT-PCR. Nucleic Acids Res 29:e45.

28. Livak KJ, Schmittgen TD 2001 Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402-408.

29. Vandesompele J, De PK, Pattyn F, Poppe B, Van RN, De PA, Speleman F 2002 Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 3:RESEARCH0034.

30. de Jonge HJ, Fehrmann RS, de Bont ES, Hofstra RM, Gerbens F, Kamps WA, de Vries EG, van der Zee AG, te Meerman GJ, ter EA 2007 Evidence based selection of housekeeping genes. PLoS ONE 2:e898.

31. Curat CA, Wegner V, Sengenes C, Miranville A, Tonus C, Busse R, Bouloumie A 2006 Macrophages in human visceral adipose tissue: increased accumulation in obesity and a source of resistin and visfatin. Diabetologia 49:744-747.

32. Mu H, Ohashi R, Yan S, Chai H, Yang H, Lin P, Yao Q, Chen C 2006 Adipokine resistin promotes in vitro angiogenesis of human endothelial cells. Cardiovasc Res 70:146-157.

33. Nijsten MW, Olinga P, The TH, de Vries EG, Koops HS, Groothuis GM, Limburg PC, ten Duis HJ, Moshage H, Hoekstra HJ, Bijzet J, Zwaveling JH 2000 Procalcitonin behaves as a fast responding acute phase protein in vivo and in vitro. Crit Care Med 28:458-461.

Chapter 5

143

34. Elferink MG, Olinga P, Draaisma AL, Merema MT, Bauerschmidt S, Polman J, Schoonen WG, Groothuis GM 2008 Microarray analysis in rat liver slices correctly predicts in vivo hepatotoxicity. Toxicol Appl Pharmacol 229:300-309.

35. Casas JP, Shah T, Hingorani AD, Danesh J, Pepys MB 2008 C-reactive protein and coronary heart disease: a critical review. J Intern Med 264:295-314.

36. Pepys MB, Hirschfield GM, Tennent GA, Gallimore JR, Kahan MC, Bellotti V, Hawkins PN, Myers RM, Smith MD, Polara A, Cobb AJ, Ley SV, Aquilina JA, Robinson CV, Sharif I, Gray GA, Sabin CA, Jenvey MC, Kolstoe SE, Thompson D, Wood SP 2006 Targeting C-reactive protein for the treatment of cardiovascular disease. Nature 440:1217-1221.

37. Haider DG, Leuchten N, Schaller G, Gouya G, Kolodjaschna J, Schmetterer L, Kapiotis S, Wolzt M 2006 C-reactive protein is expressed and secreted by peripheral blood mononuclear cells. Clin Exp Immunol 146:533-539.

38. Wajchenberg BL 2000 Subcutaneous and visceral adipose tissue: their relation to the metabolic syndrome. Endocr Rev 21:697-738.

39. Wajchenberg BL, Giannella-Neto D, da Silva ME, Santos RF 2002 Depot-specific hormonal characteristics of subcutaneous and visceral adipose tissue and their relation to the metabolic syndrome. Horm Metab Res 34:616-621.

40. Poulain-Godefroy O, Lecoeur C, Pattou F, Fruhbeck G, Froguel P 2008 Inflammation is associated with a decrease of lipogenic factors in omental fat in women. Am J Physiol Regul Integr Comp Physiol 295:R1-R7.

Chapter 6

144

Chapter 6

145

Chapter 6

Comparative analysis of the human hepatic and

adipose tissue transcriptomes during LPS-

induced inflammation leads to the

identification of differential biological

pathways and candidate biomarkers

Ewa Szalowska

Martijn Dijkstra

Marieke G.L. Elferink

Desiree Weening

Marcel de Vries

Marcel Bruinenberg

Annemieke Hoek

Han Roelofsen

Geny M.M. Groothuis

Roel J. Vonk

Submitted

Chapter 6

146

Abstract

Background

Insulin resistance (IR) is accompanied by chronic low grade systemic inflammation, obesity,

and deregulation of total body energy homeostasis. We induced inflammation in adipose and

liver tissues in vitro in order to mimic inflammation in vivo with the aim to identify tissue-

specific processes implicated in IR and to find biomarkers indicative for tissue-specific IR.

Methods

Human adipose and liver tissues were cultured in the absence or presence of LPS and DNA

Microarray Technology was applied for their transcriptome analysis. Gene Ontology (GO),

gene functional analysis, and prediction of genes encoding for secretome were performed

using publicly available bioinformatics tools (DAVID, STRING, SecretomeP). The

transcriptome data were validated by proteomics analysis of the inflamed adipose tissue

secretome.

Results

LPS treatment significantly affected 667 and 484 genes in adipose and liver tissues

respectively. The GO analysis revealed that during inflammation adipose tissue, compared to

liver tissue, had more significantly upregulated genes, GO terms, and functional clusters

related to inflammation and angiogenesis. The secretome prediction led to identification of

399 and 236 genes in adipose and liver tissue respectively. The secretomes of both tissues

shared 66 genes and the remaining genes were the differential candidate biomarkers

indicative for inflamed adipose or liver tissue. The transcriptome data of the inflamed adipose

tissue secretome showed excellent correlation with the proteomics data.

Conclusions

The higher number of altered proinflammatory genes, GO processes, and genes encoding for

secretome during inflammation in adipose tissue compared to liver tissue, suggests that

adipose tissue is the major organ contributing to the development of systemic inflammation

observed in IR. The identified tissue specific functional clusters and biomarkers might be

used in a strategy for the development of tissue-targeted treatment of insulin resistant

patients.

Chapter 6

147

Background

Adipose tissue is an important metabolic and endocrine organ that secretes numerous

biologically active proteins (adipokines) such as leptin, adiponectin, many cytokines, and

chemokines [1]. During the development of obesity, adipose tissue undergoes a switch from

being mainly a metabolic organ towards an organ that shows substantial pro-inflammatory

activity, associated with decreased insulin sensitivity, declined expression of adiponectin and

enhanced production of pro-inflammatory cytokines and chemokines. These processes are

believed to lead to low-grade inflammation and eventually systemic insulin resistance (IR)

and type 2 diabetes (T2D) [2]. However, it is not yet understood how the change in the

inflamed adipose tissue transcriptome and secretome leads to the development of IR. In

addition to adipose tissue, the liver as an important metabolic and endocrine organ secreting

many hormones, chemokines and cytokines, is also affected in obesity [3,4]. In a fatty liver,

inflammation with activated NF-κB signaling and upregulated cytokines (IL-6, TNFα, and

IL-1β) seems to be a pivotal event leading to the development of liver insulin resistance and

non-alcoholic fatty liver disease (NAFLD) which both strongly predispose to the

development of systemic IR and T2D. Except for the few proteins known to be produced and

secreted by the liver during inflammation little is known about other protein factors which

alone or by interacting with the secretome of inflamed adipose tissue could contribute to the

development of systemic inflammation and insulin resistance in humans [5-8].

Lipopolysachcaride (LPS) is a compound of the cell wall of Gram-negative bacteria which

induces inflammatory reactions and upregulates many cyto- and chemokines via TLRs.

Besides its role in inflammation LPS triggers hyperglycemia and IR in rats and humans [9-

12] and induces weight gain and liver IR in mice [13,14].

In our studies, we aimed to identify molecular processes affected during inflammation in

human AT and LT in order to better understand their roles in the inflammation- related

development of IR/T2D in vivo. Therefore we challenged human adipose tissue (omentum)

and liver tissue slices with LPS and analyzed gene expression changes by DNA microarray

technology and performed Gene Ontology (GO), gene functional classification/clustering

analysis by means of publicly available bioinformatics tools Database for Annotation,

Visualization, and Integrated Discovery (DAVID) and Search Tool for the Retrieval of

Interacting Genes/Proteins (STRING).

Chapter 6

148

Additionally, we aimed to compare the secretomes of adipose and liver tissues during

inflammation in order to better understand how these two organs can contribute to the

development of systemic inflammation and IR. The transcriptome data were used to predict

genes encoding for secreted proteins, by means of SecretomeP. The comparative analysis of

the predicted secretomes led to the identification of differential candidate biomarkers for the

inflamed adipose tissue and the inflamed liver tissue. Significantly changed genes detected in

the adipose tissue secretome, but not in the inflamed liver tissue secretome were considered

as the top candidate biomarkers related to inflammation of adipose tissue and these

transcriptome data were confirmed by proteomics analysis of the inflamed adipose tissue

culture medium.

The identified biological processes and biomarkers indicative for the inflamed adipose tissue

or the inflamed liver tissue might be used for tissue-specific diagnosis of insulin resistance

related to inflammation and thereby facilitate more targeted treatment of insulin resistant

patients.

Methods

Human liver tissue

Human liver tissue (n=5) was obtained and prepared as described previously [15]. The donors

of livers were healthy males aged 16–34 years, with BMI 23.1–27.7. The information about

the medical history was not available. The research protocols conformed the Helsinki

Declaration, were approved by the local Medical Ethical Committee of the UMCG, and

patients gave written informed consent to participate in the study.

Preparation and incubation of liver slices

Human liver slices were prepared and incubated as described previously [15]. Liver slices

were incubated at 37°C in Williams Medium E in the presence or absence of 100 µg/ml LPS.

24 h after incubation, slices were frozen in liquid nitrogen and stored at –80°C.

Chapter 6

149

Human adipose tissue

Omentum AT biopsies were obtained from 7 Caucasian women undergoing surgery because

of benign gynecological problems. The women were in general good health, had no history or

symptoms of T2D or inflammatory diseases. The subjects were aged between 30 and 45

years, with BMI ranging from 23 to 29.The omentum biopsies were taken at the lower edge

of the omentum using scissors. The research protocols conformed the Helsinki Declaration,

were approved by the local Medical Ethical Committee of the UMCG, and patients gave

written informed consent to participate in the study.

Preparation and incubation of adipose tissue biopsies

The human AT surgical biopsies were processed as described previously [16,17]. In our

studies AT was cultured in the absence/presence of LPS (100µg/ml) for 24 hours. After the

culture time the fat tissue was snap-frozen in liquid nitrogen and stored in -80ºC until further

processing.

RNA isolation

RNA was extracted from adipose tissue using RNeasy Lipid Tissue Mini Kit (Qiagen, Venlo,

The Netherlands) according to the manufacturer’s instructions. RNA extraction from human

liver slices was performed as described previously [15]. The RNA concentration was

determined by Nano Drop ND-1000 Spectrophotometer (Isogen Ijsselstein, The Netherlands).

The quality of total RNA was evaluated by capillary electrophoresis using an Agilent 2100

Bioanalyzer (Agilent Technologies, Palo Alto, Calif.).

Illumina Human WG8-v2 Microarray Analysis

The Illumina platform was used for the gene expression analysis in adipose tissue. Biotin-

labeled cRNA was generated from high-quality total RNA with the Illumina TotalPrep RNA

amplification kit (Ambion). Briefly, 50 ng of total RNA was reversely transcribed with an

oligo(dT) primer containing a T7 promoter. The first- strand cDNA was used to make the

second strand. The purified second-strand cDNA, along with biotin UTPs, was subsequently

used to generate biotinylated, antisense RNA of each mRNA in an in vitro transcription

reaction. The size distribution profile for the labeled cRNA samples was evaluated by

Bioanalyzer. After RNA labeling, 1.5ug of purified, labeled cRNA from each sample was

hybridized at 55ºC overnight with a Human-8 v2 expression Illumina Beadchip targeting

Chapter 6

150

22000 transcripts. The beadchip was washed the following day. A signal was developed

during incubation with Streptavidin-Cy3, and each chip was scanned with an Illumina Bead

Array Reader.

The preprocessing of Illumina data was performed using the BeadStudio package with default

settings. The background was subtracted and quantile normalization performed. Probes with

“absent” signals in all samples (lower than or near to background levels) were removed from

further analysis. To identify the differentially expressed genes in LPS treated samples versus

controls an eBayes test was performed and Benjamini Hochberg corrected false discovery

rate (FDR) ≤0.05. Probes with fold change ≥2 were used for further analysis. The

calculations were performed in R, a language for statistical computing and graphics (www.R-

project.org).

Affymetrix Human Genome U133 Plus 2.0 Array Analysis

The Affymetrix platform (55000 transcripts) was used for the liver tissue gene expression

analysis. Double-stranded cDNA was synthesized from 1.5 µg total RNA using the One-

Cycle Target Labeling Kit (Affymetrix Santa Clara, CA), and used as a template for the

preparation of biotin-labeled cRNA using the GeneChip IVT Labeling Kit (Affymetrix Santa

Clara, CA). Biotin-labeled cRNA was fragmented at 1 µg/µl following the manufacturer's

protocol. After fragmentation, cRNA (10µg) was hybridized at 45°C for 16 hours to the

Human Genome U133 Plus 2.0 array (Affymetrix, Santa Clara, CA). Following

hybridization, the arrays were washed, stained with phycoerythrin-streptavidin conjugate

(Molecular Probes, Eugene, OR), and the signals were amplified by staining the array with

biotin-labeled anti-streptavidin antibody (Vector Laboratories, Burlingame, CA) followed by

phycoerythrin-streptavidin. The arrays were laser scanned with a GeneChip Scanner 3000 7G

(Affymetrix, Santa Clara, CA) according to the manufacturer's instructions. Data was saved

as raw image file and quantified using GCOS (Affymetrix).

Probe set summarization was performed using the RMA algorithm. Subsequently, baseline

subtraction was performed setting the baseline to the median of all samples. To identify the

differentially expressed genes in LPS treated samples versus controls an eBayes test was

performed and Benjamini Hochberg corrected false discovery rate (FDR) ≤0.05. Probes with

fold change ≥2 were used for further analysis. The calculations were performed in R, a

language for statistical computing and graphics (www.R-project.org).

Chapter 6

151

Gene Functional Classification Analysis

The significant transcriptomes of AT and LT were uploaded to Database for Annotation,

Visualization, and Integrated Discovery (DAVID) Bioinformatics Resource where the Gene

Functional Classification tool was applied to generate clusters of functionally related genes.

Additionally the Functional Annotation Clustering tool was used to generate clusters of

overrepresented Gene Ontology (GO) terms [18,19]. The HG-U133 Plus 2 and

HUMANREF-8 V2 0 R3 11223162A were used as a background for the GO analysis of liver

tissue and adipose tissue respectively. The GO terms after correction for FDR at p≤0.05

(Benjamini Hochberg) were selected for further analysis and interpretation.

Gene networks and pathways identification

The significant transcriptomes of adipose and liver tissues were uploaded to Search Tool for

the Retrieval of Interacting Genes/Proteins 8.2 (STRING) where networks based on known

and predicted protein-protein interactions were built and clustered into functional categories

[20].

Secretome prediction

From the significant transcriptome data obtained for adipose and liver tissues, the secretome

prediction was performed with in-house developed software, which retrieved the

information about the predicted secretomes from SecretomeP [21]. Genes were considered to

belong to the secretome when they encoded for proteins with a predicted signal peptide

(present in proteins that are secreted via the classical endoplasmic reticulum/Golgi-dependent

pathway) or when their Neuronal Network (NN) score exceeded the value of 0.5, which

classifies them as secreted via the non-classical pathway. Genes encoding for proteins which

did not have a signal peptide nor had the NN-score below 0.5 were considered as genes

encoding for intracellular proteins and were discarded from the final secretome analysis.

Adipose tissue culture for the quantitative proteomics analysis

Quantitative secretome analysis was performed by Isotope-labeled Amino Acid Incorporation

Rates (CILAIR) as described previously [22]. Briefly, 6g of fat tissue was used from one

patient and divided into six Petri dishes containing 10 ml of lysine-free M199 medium

(reference number 22340 Lys-free, Invitrogen) to deplete lysine from other sources (blood in

the tissue) and supplemented with 50 µg/ml gentamicin. The tissue was incubated for 24 h.

Chapter 6

152

After this period, fresh M199 containing 70 mg/liter 13C-labeled lysine (L-

[13C6,14N2]lysine (Invitrogen) was added to all dishes for the next 24 hours to allow

incorporation of the label into newly synthesized proteins, in the absence (3 dishes) or

presence (3 dishes) of LPS (100 µg/ml). CILAIR is based on the incorporation rate of 13C-

labeled lysine in newly synthesized secreted proteins. If this rate is different between two

conditions for a specific protein the change in expression of this protein can be calculated by

comparing the heavy/light ratios for the two conditions. After the 24 h incubation, media

were collected and stored at –80ºC until further processing. The sample preparation and

protein identification by liquid chromatography coupled to mass spectrometry was performed

as described previously [22]. ProteinPilot 2.0 software (Applied Biosystems) was used to

analyze the mass spectra using the UniprotKB/Swiss-Prot database (release 54, January 2008,

276,256 entries). The settings used in the analysis were the same as described previously

[22].

CILAIR data analysis

The statistical analysis to detect differences in the secretome of LPS-treated vs. control

adipose tissue cultures was performed with in-house generated software that was developed

using the open source MOLGENIS toolbox [23]. A two-sided unpaired Student’s t-test was

applied, and multiple testing correction was performed to control the false discovery rate

(FDR) at FDR < 0.05.

The applied criteria for the proteins predicted to be secreted were the same as described

above for the transcriptome data.

Chapter 6

153

Results

Functional gene annotation analysis

The transcriptome data analysis revealed that in adipose tissue 667 genes were significantly

affected (322 were upregulated and 345 were downregulated) after exposure to LPS. In liver

tissue we detected 483 significantly changed genes (283 were upregulated and 200 were

downregulated). The overlapping significant transcriptome shared by both tissues consisted

of 82 transcripts. The significantly changed genes found in adipose tissue and liver tissue

which were not present on both platforms were discarded from further analysis (47 and 42

respectively). Functional gene annotation analysis of significantly upregulated genes in

adipose tissue (including the overlapping genes with the liver tissue significant

transcriptome) led to the identification of functional groups such as: chemokines; growth and

differentiation of hematopoietic precursors; (anti)apoptosis; modulation of immune response;

T-, B-, leukocytes, and NK-cells activation, suppression of cytokine signaling (SOCS),

extracellular matrix remodeling, and upregulation of numerous transporters, (supplementary

Table 1A). Within the downregulated gene functional groups we identified:

lysosomal/endosomal system activity, basement membrane components, extracellular matrix

components, cell adhesion and migration, deoxy-ribonucleases activity, and detoxification,

(supplementary Table 1B). A similar analysis was performed for liver tissue and within the

upregulated gene functional groups we identified : chemokines; matrix remodeling;

(anti)apoptosis; cell adhesion and migration; T- and NK- cell activity; and breakdown of

extracellular matrix/tissue remodeling, (supplementary Table 1C). The functional

classification of the downregulated genes led to identification of groups such as: amino acid

metabolism, membrane activity, redox/detoxification reactions, cell adhesion and

mitochondrial functions, (supplementary Table 1D). Additionally, in order to better visualize

the similarities and differences between the adipose tissue and liver tissue transcriptomes

during inflammation we performed gene functional network reconstruction in STRING,

Figures 1-5.

Chapter 6

154

Figure 1. The gene functional clusters identified for the significant, overlapping adipose and liver tissue transcriptomes. The overlapping (shared) upregulated adipose tissue and liver tissue significantly changed transcriptome. Within the overlapping network we identified functional clusters related to : T and B cell activation and functioning (pink); matrix remodeling (green); interleukin 7 receptor activity (yellow); mobilization of T-lymphocytes and monocytes (red); (anti)apoptosis/inflammation(blue).

Chapter 6

155

Figure 2. The gene functional clusters identified for the significant, upregulated adipose tissue transcriptome. The upregulated adipose tissue network contained 8 functional clusters: regulation of cytokine signaling (grey); glucocorticoid receptor signaling (red); acute phase response (pink); growth and differentiation of hematopoietic cells (dark blue); plasminogen activation system (bright blue); IL-10 signaling (light blue); apoptosis (green); cell adhesion (yellow).

Chapter 6

156

Figure 3. The gene functional clusters identified for the significant, downregulated adipose tissue transcriptome. The downregulated adipose tissue network had 6 functional clusters: PPARγ signaling (red); cellular defense against toxic compounds (green); redox reactions (dark blue); innate immune system (bright blue); G-receptor signaling (black); Wnt-signaling (pink).

Chapter 6

157

Figure 4. The gene functional clusters identified for the significant, upregulated liver tissue transcriptome. The liver tissue upregulated network consisted of 6 clusters: JAK-STAT signaling (light blue); NFκB signaling (blue); extracellular matrix remodeling (bright blue); chemo-attraction of T- and NK-cells (black); innate immune system (red); ROS production (pink).

Chapter 6

158

Figure 5. The gene functional clusters identified for the significant, downregulated liver tissue transcriptome. The downregulated liver tissue network contained 7 functional clusters: redox reactions (blue); cytochrome P450 (red); cell-cell adhesion (black); amino acids metabolism (light blue); metabolism (pink and blue); leukocytes functioning (green); immune reactions (bright blue).

Gene Ontology analysis

Additionally we performed GO ontology analysis. In adipose tissue we identified more

upregulated GO terms compared to liver tissue (106 vs. 36) and for the down-regulated GO

terms we detected 2 and 19 in adipose tissue and liver tissue respectively. The significantly

upregulated GO terms were divided into broad categories such as “inflammation”,

“development”, “signaling”, “metal ion homeostasis”, ‘secretion’ and “angiogenesis” and

within the downregulated GO categories we distinguished: “extracellular region”, “amino

acid metabolism”, and “polysaccharide binding”. The GO terms identities within the GO

categories are presented in the supplementary Tables 2A-D. Adipose tissue had more

upregulated GO terms belonging to “inflammation”, “development” and “angiogenesis”

compared to liver tissue and had additional terms such as: “signaling”, “metal ion

homeostasis” and “secretion”, (Figure 6).

Within the downregulated GO categories in adipose tissue we detected “extracellular region

while in liver tissue- “amino acid metabolism” and “polysaccharide binding”, (Figure 6).

Chapter 6

159

0 10 20 30 40 50

inflammation

development

signaling

metal ion homeostasis

secretion

angiogenesis

Number of significantly affected GO terms

GO analysis-upregulation (GO terms)

LT

AT

0 2 4 6 8 10

extracellular region

aa metabolism

inflammation/binding

Number of significantly affected GO terms

GO analysis-downregulation (GO terms)

LT

AT

Figure 6. GO analysis of the significant adipose and liver tissues transcriptomes. The number of significantly enriched upregulated and downregulated GO terms in adipose tissue (AT) and liver tissue (LT) upon LPS treatment. The GO terms were categorized into broader GO categories such as: angiogenesis, secretion, metal ion homeostasis, signaling, development, inflammation, amino acid (aa) metabolism, and extracellular region.

When analyzing individual genes within the GO categories, a similar picture emerged -in

general the larger number of genes belonging to the identified GO categories was altered in

adipose tissue compared to liver tissue Figure 7. The names and Entrez IDs of genes up- and

down- regulated in both tissues for each GO category are given in supplementary Tables 3A–

B.

Chapter 6

160

0 50 100 150 200

inflammation

development

signaling

metal ion homeostasis

secretion

angiogenesis

Number of significanltly affected genes

GO analysis-upregulation (genes)

LT

AT

0 50 100 150 200

extracellular region

aa metabolism

inflammation/binding

Number of significantly affcted genes

GO analysis-downregulation (genes)

LT

AT

Figure 7. Gene count analysis for the identified GO categories. Number of genes significantly upregulated and

downregulated in adipose tissue (AT) and liver tissue (LT) within GO categories (angiogenesis, secretion, metal

ion homeostasis, signaling, development, inflammation, amino acid (aa) metabolism, and extracellular region).

Chapter 6

161

The differentially expressed genes and secretome prediction

Subsequent analysis of the significant transcriptome data was performed in order to select

genes predicted to encode for secreted proteins (the predicted secretome). The analysis

revealed that adipose tissue and liver tissue share 66 genes predicted to encode for secreted

proteins (46 were upregulated and 20 were downregulated). In the adipose tissue predicted

secretome we identified additional 333 significantly changed genes encoding for secreted

proteins (138 transcripts were upregulated and 195 -were downregulated) and within the liver

tissue predicted secretome we identified 170 different genes encoding for secreted proteins

(80 were upregulated and 90 were downregulated).

In our studies we were mostly interested in the upregulated genes as they could be the best

candidate biomarkers measurable in human serum. The information about gene expression of

the highest upregulated genes in adipose and liver tissues is summarized in Table 1. The

presented genes were subdivided in three categories: the first category contained genes which

were significantly upregulated in both tissues (p≤0.05, FC≥2) as the best candidate

biomarkers for the inflamed adipose and liver tissues. The second category contained genes

significantly upregulated in adipose tissue (p≤0.05, FC≥2), but not changed in liver tissue, as

the best candidate biomarkers for the inflamed adipose tissue. The third category contained

genes significantly upregulated in liver tissue (p≤0.05, FC≥2) and unchanged in adipose

tissue (p>0.05) as the best source of candidate biomarkers for the inflamed liver tissue. The

entire list of genes encoding for the predicted inflammatory secretomes of adipose and liver

tissues is given in supplementary material in Tables 4A-C.

Chapter 6

162

Table 1. The most differential predicted secretome of adipose and liver tissues. The highest upregulated genes (p≤0.05) upon LPS treatment in adipose tissue (AT) and liver tissue (LT) are indicated in bold and are underlined and unchanged genes (p>0.05) are in italics. In the last two columns of the table fold change (FC) of gene expression in AT and LT is given.

ACCESSION NAME4 GENE SYMBOL AT FC LT FC

P01584 INTERLEUKIN 1, BETA IL1B 20 100

P10147 CHEMOKINE (C-C MOTIF) LIGAND 3 CCL3 19.8 10.7

Q96DR8 SMALL BREAST EPITHELIAL MUCIN MUCL1 17.4 5.8

P18510 INTERLEUKIN 1 RECEPTOR ANTAGONIST IL1RN 12.5 6.5

P78556 CHEMOKINE (C-C MOTIF) LIGAND 20 CCL20 11.4 7.5

P16619 CHEMOKINE (C-C MOTIF) LIGAND 3-LIKE 1 CCL3L1 10.7 34.54

P42830 CHEMOKINE (C-X-C MOTIF) LIGAND 5 CXCL5 7.8 30.3

P35354 PROSTAGLANDIN-ENDOPEROXIDE SYNTHASE 2 PTGS2 6.8 7.9

P13501 CHEMOKINE (C-C MOTIF) LIGAND 5/RANTES CCL5 6.3 68

P05120 SERPIN PEPTIDASE INHIBITOR, CLADE B (OVALBUMIN), MEMBER 2 SERPINB2 6 8

P01583 INTERLEUKIN 1, ALPHA IL1A 5.8 4.3

O14625 CHEMOKINE (C-X-C MOTIF) LIGAND 11 CXCL11 5.4 20.4

P05231 INTERLEUKIN 6 (INTERFERON, BETA 2) IL6 4.5 17.8

P08254 MATRIX METALLOPEPTIDASE 3 (STROMELYSIN 1, PROGELATINASE) MMP3 4 20.6

P09038 FIBROBLAST GROWTH FACTOR 2 (BASIC) FGF2 3.7 3.5

P39900 MATRIX METALLOPEPTIDASE 12 (MACROPHAGE ELASTASE) MMP12 3.7 5.8

P09341 CHEMOKINE (C-X-C MOTIF) LIGAND 1 CXCL1 3 23.5

P02778 CHEMOKINE (C-X-C MOTIF) LIGAND 10 CXCL10 2.9 17

P80162 CHEMOKINE (C-X-C MOTIF) LIGAND 6 (GRANULOCYTE CHEMOTACTIC PROTEIN 2) CXCL6 2.8 22

P10144 GRANZYME B (GRANZYME 2, CYTOTOXIC T-LYMPHOCYTE-ASSOCIATED SERINE ESTERASE 1) GZMB 2.8 6.1

O60462 NEUROPILIN 2 NRP2 2.6 2.6

P10145 INTERLEUKIN 8 IL8 2.5 6

P28845 HYDROXYSTEROID (11-BETA) DEHYDROGENASE 1 HSD11B1 2.4 2.5

P13500 CHEMOKINE (C-C MOTIF) LIGAND 2 CCL2 2.3 3.7

P16581 SELECTIN E (ENDOTHELIAL ADHESION MOLECULE 1) SELE 105.1 -2.5

P04141 COLONY STIMULATING FACTOR 2 (GRANULOCYTE-MACROPHAGE) CSF2 82.5 1

Q92629 SARCOGLYCAN, DELTA (35KDA DYSTROPHIN-ASSOCIATED GLYCOPROTEIN) SGCD 58.8 -2.5

Q9BYE3 LATE CORNIFIED ENVELOPE 3D LCE3D 46 -1.4

P02763 OROSOMUCOID 1 ORM1 26.6 1

O14944 EPIREGULIN EREG 25 -1.2

P22894 MATRIX METALLOPEPTIDASE 8 (NEUTROPHIL COLLAGENASE) MMP8 22.2 1.3

Q00604 NORRIE DISEASE (PSEUDOGLIOMA) NDP 19.5 -1.1

P07357 COMPLEMENT COMPONENT 8, ALPHA POLYPEPTIDE C8A 18.6 1

P09919 COLONY STIMULATING FACTOR 3 (GRANULOCYTE) CSF3 16.3 -1.1

P78423 FRACTALCINE CX3CL1 6.4 1.6

P01375 TUMOR NECROSIS FACTOR (TNF SUPERFAMILY, MEMBER 2) TNF 6 1.7

Q9UHD0 INTERLEUKIN 19 IL19 5.6 1.2

O14896 INTERFERON REGULATORY FACTOR 6 IRF6 5.3 -1.4

P26022 PENTRAXIN-RELATED GENE, RAPIDLY INDUCED BY IL-1 BETA PTX3 4 1.5

P03956 MATRIX METALLOPEPTIDASE 1 (INTERSTITIAL COLLAGENASE) MMP1 3.4 -1.1

P03950 ANGIOGENIN, RIBONUCLEASE, RNASE A FAMILY, 5 ANG 3.2 -2

Q9BY76 ANGIOPOIETIN-LIKE 4 ANGPTL4 3.2 1

P19875 CHEMOKINE (C-X-C MOTIF) LIGAND 2 CXCL2 3.1 1.6

P05121 SERPIN PEPTIDASE INHIBITOR/ PLASMINOGEN ACTIVATOR INHIBITOR TYPE 1) MEMBER 1 SERPINE1/pai1 3 1.6

P10124 PROTEOGLYCAN 1, SECRETORY GRANULE SRGN 2.7 1

Q96RQ9 INTERLEUKIN 4 INDUCED 1 IL4I1 2.4 1

Q9Y5U4 INSULIN INDUCED GENE 2 INSIG2 2.3 -1.2

P12643 BONE MORPHOGENETIC PROTEIN 2 BMP2 2.2 1.5

P02735 SERUM AMYLOID A1 SAA1 2.2 1.1

Q07325 CHEMOKINE (C-X-C MOTIF) LIGAND 9 CXCL9 1.8 69.3

P19876 CHEMOKINE (C-X-C MOTIF) LIGAND 3 CXCL3 3.6 20.6

O95633 FOLLISTATIN-LIKE 3 (SECRETED GLYCOPROTEIN) FSTL3 1 15.1

Q13113 PDZK1 INTERACTING PROTEIN 1 PDZK1IP1 1.9 12.9

Q9NRD8 DUAL OXIDASE 2 DUOX2 -2 5.5

Q8WWX9 SELENOPROTEIN M SELM 1.6 3.9

O94808 GLUTAMINE-FRUCTOSE-6-PHOSPHATE TRANSAMINASE 2 GFPT2 1.7 3.6

P13164 INTERFERON INDUCED TRANSMEMBRANE PROTEIN 1 (9-27) IFITM1 -1.1 3.5

P09603 COLONY STIMULATING FACTOR 1 (MACROPHAGE) CSF1 -1.6 3.3

P12544 GRANZYME A (GRANZYME 1, CYTOTOXIC T-LYMPHOCYTE-ASSOCIATED SERINE ESTERASE 3) GZMA -1.4 3.3

P25774 CATHEPSIN S CTSS 1.4 3.2

P24001 INTERLEUKIN 32 IL32 1.6 3

P31431 SYNDECAN 4 (AMPHIGLYCAN, RYUDOCAN) SDC4 1.9 2.7

P03973 SECRETORY LEUKOCYTE PEPTIDASE INHIBITOR SLPI -1.4 2.7

P09237 MATRIX METALLOPEPTIDASE 7 (MATRILYSIN, UTERINE) MMP7 -1.6 2.6

Q5VY09 IMMEDIATE EARLY RESPONSE 5 IER5 1.6 2.5

O75976 CARBOXYPEPTIDASE D CPD 1.3 2.2

Chapter 6

163

Transcriptomics and proteomics data comparison and candidate biomarkers

identification

In order to validate biomarkers related to inflamed adipose tissue, we performed a similar

experiment using a quantitative proteomics approach (CILAIR), and analyzed the secreted

proteins in the adipose tissue culture media (secretome). In the CILAIR experiment we

identified 192 proteins with incorporated label in medium of LPS treated tissue and 209 in

medium of untreated adipose tissue. 178 proteins had incorporated label in both conditions

and could thus be compared quantitatively. The statistical analysis revealed that 23 proteins

were significantly changed in abundance in the secretome by LPS treatment. Comparison

with the gene expression data for adipose tissue showed excellent correlation between

proteomics and transcriptomics data (Pearson’s correlation r2 = 0.78; Table 2). Within the 23

significantly affected proteins we selected those which were significantly affected by LPS in

adipose tissue, on both gene and protein level, but not changed in the liver tissue

transcriptome, and those proteins were considered as the best candidate biomarkers for

inflamed adipose tissue. We propose: PTX3, MMP1, SERPINE1, and CX3CL1 as the top

candidate biomarkers related to the inflamed adipose tissue. The results are summarized in

Table 2.

Chapter 6

164

Table 2. Significantly changed secreted proteins in adipose tissue culture media and the corresponding identified genes in adipose tissue and liver tissue upon LPS-treatment. The significantly changed proteins (p≤0.05, FC>1.2) and the significantly changed genes (p≤0.05, FC>2) are represented in bold. The insignificantly affected genes are depicted in italics. The top candidate biomarkers related to inflamed adipose tissue are depicted as underlined. FC stands for fold change.

NAME SYMBOL ADIPOSE TISSUE

FC-protein

ADIPOSE TISSUE FC-

transcriptome

LIVER TISSUE-FC

transcriptome

Granulocyte colony-stimulating factor CSF3 10.3 16 14

Leukemia inhibitory factor LIF 2.3 7.1 2.4

Fractalkine CX3CL1 4.3 6.4 1.6

Tumor necrosis factor TNF 3.8 6 1.7

Plasminogen activator inhibitor 2 SERPINB2 3.1 6 8

Interleukin-6 IL6 1.6 4.5 17.8

Pentraxin-related protein PTX3 1.9 4 1.6

Interstitial collagenase MMP1 1.7 3.4 -1.1

Tumor necrosis factor-inducible gene 6 protein TNFAIP6 5.4 3.1 17.9

Plasminogen activator inhibitor 1 SERPINE1 1.7 3 1.6

C-C motif chemokine 2 CCL2 6.9 2.3 3.7

CD44 antigen CD44 2.4 1.7 5.8

Insulin-like growth factor-binding protein 4 IGFBP4 -2.5 -1.1 1.2

Adipocyte enhancer-binding protein 1 AEBP1 -1.4 -1.2 1.2

Cystatin-C CST3 -3.1 -1.2 -1.1

Versican core protein VCAN -2 -1.6 -1.4

Collagen alpha-1(VI) chain COL6A1 -3.3 -1.6 -1.4

Transforming growth factor-beta-induced protein ig-h3 TGFBI -2.5 -1.6 1

Legumain LGMN -3.1 -2 1.2

Gelsolin GSN -2.5 -2 1.1

Cathepsin B CTSB -1.2 -2 -1.2

Lysozyme C LYZ -3.3 -2.5 -3.3

Alpha-2-macroglobulin A2M -3.3 -3.3 -1.6

Chapter 6

165

Discussion

In the present study we evoked LPS induced inflammation in adipose and liver tissues in

vitro in order to mimic IR caused by inflammation in vivo. We aimed to compare the changes

in the inflamed transcriptomes and secretomes of both tissues in order to (1) better understand

contribution of the inflamed adipose and liver tissues to the development of insulin resistance

and (2) to identify candidate biomarkers indicative for tissue specific inflammation/IR.

The gene functional classification analysis revealed that both adipose and liver tissue share

common response mechanisms that are activated during inflammation (chemokine signaling,

(anti)apoptosis, extracellular matrix remodeling, adhesion and migration of different immune

cells involved in inflammatory reactions). Although functional clustering led to identification

of the same functional groups, both tissues had a different set of genes within one functional

group, suggesting tissue-specific inflammatory signaling. The significantly upregulated

adipose tissue transcriptome contained additional gene functional categories belonging to

SOCS and several transporters (supplementary Table 1A). The SOCS signaling was shown

previously to be involved in induction of insulin resistance during acute inflammation in

human adipose tissue [24] and our ex vivo data are in line with these in vivo findings. The

analysis of the down regulated functional groups pointed out towards redox/detoxification

processes affected in both tissues and mitochondrial functions observed in liver tissue. These

processes could contribute to the enhanced reactive oxygen species (ROS) production

recognized as one of the mechanisms implicated in the development of IR/T2D [13].

Furthermore, adipose tissue had downregulated genes involved in the extracellular matrix

activity which is involved in multiple processes including modulation of immune response. In

liver tissue downregulation of genes involved in amino acid metabolism and polysaccharide

binding were observed. There are reports about changed amino acids concentrations in

animal models of obesity and obese humans [25,26], however interpretation of this ex vivo

finding in relation to these reports is not unequivocal.

The additional network identification for the common (overlapping) and differential

adipose and liver tissue transcriptomes was in line with the data obtained from the gene

functional analysis and distinguished the common and differential networks. Several of these

networks were described previously in the literature for their role in induction of IR thereby

supporting our model system to study the inflammation related insulin resistance in vivo. For

Chapter 6

166

example chemokine signaling and matrix remodeling were found for both tissues [27,28];

SOCS and PPARγ were changed in adipose tissue [29,30]; upregulated Jak-STAT signaling

and NFκB were identified previously in IR liver [31,32].

The GO analysis and gene count revealed that adipose tissue had more LPS-induced

upregulated GO terms and genes related primarily to “inflammation”, “angiogenesis”, and

“development”. Moreover, the predicted secretome studies showed that the adipose tissue

predicted inflammatory secretome is more abundant compared to the liver tissue secretome.

This observation indicates that adipose tissue is more active during inflammation, compared

to liver tissue, and supports the hypothesis that adipose tissue plays the major role in the

development of inflammation-related IR [2].

The predicted secretome analysis. The microarray data analysis of both tissues revealed

that adipose and liver tissues have numerous overlapping LPS-responsive genes which

protein products are predicted to be secreted. Among these genes we identified several known

markers associated with insulin resistance such as IL-6, IL-1β, IL-8, and PAI 1. Other

proteins known to be upregulated during insulin resistance by adipose tissue [33] such as

RANTES , MCP1, PLAUR, CXCL5, were found in our studies to be upregulated in both

adipose- and liver tissues. Additionally in both tissues we found genes, previously shown to

be regulated in adipose tissue in relation to insulin resistance: CXCL1, CXCL10, CXCL11,

ICAM1, TNFAIP6 [34], FGF2, IL6 [35], and ICAM1, IL-1 [36]. Although TNFα is known to

be involved in the development of insulin resistance in both adipose tissue and the liver, it

was only significantly upregulated in adipose tissue. However, we observed that 3 out of 5

livers had upregulated expression of TNFα and previously we showed that in liver tissue in

vitro, TNFα mRNA level was significantly upregulated after 5 hrs while after 24hrs the

TNFα mRNA level returned to basal values. [15,37].

Furthermore, the comparative analysis of adipose and liver tissues secretomes in vitro

provides a source of candidate biomarkers related to tissue specific inflammation/insulin

resistance. Similarly to Shah et al. [34], we identified in the inflamed adipose tissue

secretome genes such as: SELE, CD274, ORM1, PLA1A, SLAMF1, CX3CL1, OSM, TNF,

C19ORF59, PTX3, IER3, CCL8, CXCL2, SERPINE1, BMP2, FAM107A, GPX3. Moreover

we identified genes of yet unknown functions such as: C14ORF162, C20ORF59 or genes

implicated in other than insulin resistance inflammatory diseases: epiregulin, IL-19 or

sarcoglycan [38-40].

Chapter 6

167

The analysis of the predicted secretome of inflamed liver tissue revealed several significantly

changed genes with a known- and an unknown- relationship to insulin resistance.

Identification of biomarkers indicative for inflamed liver tissue could be a useful tool in a

diagnosis of NAFLD patients, where the only “golden standard” is an invasive liver biopsy

[41]. Biomarkers previously associated with liver diseases and identified in our samples were

among others: ANGPTL3, IGFBP2, SDC4, IL1RN [7,42]. Examples of other pro-

inflammatory proteins affiliated with inflammation but not liver insulin resistance were

cathepsin S [33] or granzyme A [43]. In future it has to be validated if the other most

differentially regulated genes between both tissues such as: SGCD, LCE3D, EREG, NDP and

CXCL9, FSTL3, PDZK1IP1 could be used as biomarkers related to insulin resistance of

adipose or liver tissues respectively.

Comparison of transcriptomics and proteomics data. Finally the transcriptome data

encoding for the adipose tissue inflammatory secretome was validated and compared with the

protein data of the inflamed adipose tissue culture medium. The analysis showed that the

transcriptome data were in line with the proteomics data, in respect to observed upwards and

downwards fold changes (FC) for genes and their corresponding protein products. However

the FC derived from the proteomics experiment cannot be directly compared with the FC of

the transcriptome experiment due to substantial technical differences between both

technologies. By combination of the comparative transcriptome analysis and proteomics

technology we identified matrix metalopeptidase-1(MMP-1), pentraxin related gene product

(PTX3), fractalkine (CX3CL1), and PAI 1 as the potential set of biomarkers for the inflamed

adipose tissue. We believe that such an approach could result in more specific diagnosis for a

tissue specific insulin resistance related to inflammation, than the use of single biomarkers.

One of the shortcomings of our study was the use of two different DNA microarray

platforms, since the data used here were generated in two different laboratories. However,

previous studies comparing human Affymetrix and Illumina platforms show that the obtained

results, using the same human material, are highly comparable, especially for genes which are

predicted to be differentially expressed [44]. Furthermore in our studies we compared only

genes which were significantly affected and present on both platforms; therefore genes which

were not present on both platforms were excluded from the analysis and we did not compare

intensities of corresponding genes since they would be different due to the platform specific

design.

Chapter 6

168

Conclusions

In summary, our in vitro approach showed that LPS-induced inflammation in adipose and

liver tissues, results in upregulation of inflammatory processes and downregulation of

metabolic pathways and redox/detoxification reactions, which could synergistically

contribute to the deregulation of energy homeostasis leading to insulin resistance.

Furthermore, our study implies that adipose tissue is more active during inflammation

compared to the liver, based on identification of higher number of GO terms and genes

involved in inflammation and angiogenesis, and a number of genes predicted to encode for

secreted proteins. It has to be validated in the future if the identified tissue specific molecular

pathways and the identified tissue specific candidate biomarkers can be used for tissue

specific diagnosis of insulin resistance in patients. We believe that such an approach might

facilitate more targeted treatment of insulin resistant patients.

Acknowledgements

We would like to thank Susanne Bauerschmidt and Jan Polman (Merck, Oss, The

Netherlands) for the performance and elaboration of the Affymetrix data, Prof. Rainer

Breitling for critical discussions and Heleen de Weerd for bioinformatical assistance.

References

1. Rasouli N, Kern PA: Adipocytokines and the metabolic complications of obesity. J Clin Endocrinol Metab 2008, 93:S64-S73.

2. Hotamisligil GS: Inflammation and metabolic disorders. Nature 2006, 444:860-867. 3. Shoelson SE, Lee J, Goldfine AB: Inflammation and insulin resistance. J Clin Invest 2006,

116:1793-1801. 4. Wellen KE, Hotamisligil GS: Inflammation, stress, and diabetes. J Clin Invest 2005, 115:1111-1119. 5. Breitling R: Robust signaling networks of the adipose secretome. Trends Endocrinol Metab 2009,

20:1-7. 6. Karalis KP, Giannogonas P, Kodela E, Koutmani Y, Zoumakis M, Teli T: Mechanisms of obesity and

related pathology: linking immune responses to metabolic stress. FEBS J 2009, 276:5747-5754. 7. Younossi ZM, Gorreta F, Ong JP, Schlauch K, Del GL, Elariny H, Van MA, Younoszai A, Goodman

Z, Baranova A et al.: Hepatic gene expression in patients with obesity-related non-alcoholic steatohepatitis. Liver Int 2005, 25:760-771.

8. Edens MA, Kuipers F, Stolk RP: Non-alcoholic fatty liver disease is associated with cardiovascular disease risk markers. Obes Rev 2009, 10:412-419.

9. Virkamaki A, Yki-Jarvinen H: Mechanisms of insulin resistance during acute endotoxemia. Endocrinology 1994, 134:2072-2078.

Chapter 6

169

10. Virkamaki A, Puhakainen I, Koivisto VA, Vuorinen-Markkola H, Yki-Jarvinen H: Mechanisms of hepatic and peripheral insulin resistance during acute infections in humans. J Clin Endocrinol Metab 1992, 74:673-679.

11. Sugita H, Kaneki M, Tokunaga E, Sugita M, Koike C, Yasuhara S, Tompkins RG, Martyn JA: Inducible nitric oxide synthase plays a role in LPS-induced hyperglycemia and insulin resistance. Am J Physiol Endocrinol Metab 2002, 282:E386-E394.

12. Agwunobi AO, Reid C, Maycock P, Little RA, Carlson GL: Insulin resistance and substrate utilization in human endotoxemia. J Clin Endocrinol Metab 2000, 85:3770-3778.

13. Bashan N, Kovsan J, Kachko I, Ovadia H, Rudich A: Positive and negative regulation of insulin signaling by reactive oxygen and nitrogen species. Physiol Rev 2009, 89:27-71.

14. Cani PD, Amar J, Iglesias MA, Poggi M, Knauf C, Bastelica D, Neyrinck AM, Fava F, Tuohy KM, Chabo C et al.: Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes 2007, 56:1761-1772.

15. Elferink MG, Olinga P, Draaisma AL, Merema MT, Faber KN, Slooff MJ, Meijer DK, Groothuis GM: LPS-induced downregulation of MRP2 and BSEP in human liver is due to a posttranscriptional process. Am J Physiol Gastrointest Liver Physiol 2004, 287:G1008-G1016.

16. Cani PD, Bibiloni R, Knauf C, Waget A, Neyrinck AM, Delzenne NM, Burcelin R: Changes in gut microbiota control metabolic endotoxemia-induced inflammation in high-fat diet-induced obesity and diabetes in mice. Diabetes 2008, 57:1470-1481.

17. Alvarez-Llamas G, Szalowska E, de Vries MP, Weening D, Landman K, Hoek A, Wolffenbuttel BH, Roelofsen H, Vonk RJ: Characterization of the human visceral adipose tissue secretome. Mol Cell Proteomics 2007, 6:589-600.

18. Dennis G, Jr., Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA: DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol 2003, 4:3.

19. Huang dW, Sherman BT, Lempicki RA: Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc 2009, 4:44-57.

20. Jensen LJ, Kuhn M, Stark M, Chaffron S, Creevey C, Muller J, Doerks T, Julien P, Roth A, Simonovic M et al.: STRING 8--a global view on proteins and their functional interactions in 630 organisms. Nucleic Acids Res 2009, 37:D412-D416.

21. Bendtsen JD, Jensen LJ, Blom N, Von HG, Brunak S: Feature-based prediction of non-classical and leaderless protein secretion. Protein Eng Des Sel 2004, 17:349-356.

22. Roelofsen H, Dijkstra M, Weening D, de Vries MP, Hoek A, Vonk RJ: Comparison of isotope-labeled amino acid incorporation rates (CILAIR) provides a quantitative method to study tissue secretomes. Mol Cell Proteomics 2009, 8:316-324.

23. Swertz MA, Jansen RC: Beyond standardization: dynamic software infrastructures for systems biology. Nat Rev Genet 2007, 8:235-243.

24. Mehta NN, McGillicuddy FC, Anderson PD, Hinkle CC, Shah R, Pruscino L, Tabita-Martinez J, Sellers KF, Rickels MR, Reilly MP: Experimental Endotoxemia Induces Adipose Inflammation and Insulin Resistance in Humans. Diabetes 2009.

25. Wijekoon EP, Skinner C, Brosnan ME, Brosnan JT: Amino acid metabolism in the Zucker diabetic fatty rat: effects of insulin resistance and of type 2 diabetes. Can J Physiol Pharmacol 2004, 82:506-514.

26. She P, Van HC, Reid T, Hutson SM, Cooney RN, Lynch CJ: Obesity-related elevations in plasma leucine are associated with alterations in enzymes involved in branched-chain amino acid metabolism. Am J Physiol Endocrinol Metab 2007, 293:E1552-E1563.

27. Feuerer M, Herrero L, Cipolletta D, Naaz A, Wong J, Nayer A, Lee J, Goldfine AB, Benoist C, Shoelson S et al.: Lean, but not obese, fat is enriched for a unique population of regulatory T cells that affect metabolic parameters. Nat Med 2009, 15:930-939.

28. Nishimura S, Manabe I, Nagasaki M, Eto K, Yamashita H, Ohsugi M, Otsu M, Hara K, Ueki K, Sugiura S et al.: CD8+ effector T cells contribute to macrophage recruitment and adipose tissue inflammation in obesity. Nat Med 2009, 15:914-920.

Chapter 6

170

29. Baranova A, Collantes R, Gowder SJ, Elariny H, Schlauch K, Younoszai A, King S, Randhawa M, Pusulury S, Alsheddi T et al.: Obesity-related differential gene expression in the visceral adipose tissue. Obes Surg 2005, 15:758-765.

30. Gomez-Ambrosi J, Catalan V, ez-Caballero A, Martinez-Cruz LA, Gil MJ, Garcia-Foncillas J, Cienfuegos JA, Salvador J, Mato JM, Fruhbeck G: Gene expression profile of omental adipose tissue in human obesity. FASEB J 2004, 18:215-217.

31. Larsen L, Ropke C: Suppressors of cytokine signalling: SOCS. APMIS 2002, 110:833-844. 32. Lebrun P, Van OE: SOCS proteins causing trouble in insulin action. Acta Physiol (Oxf) 2008,

192:29-36. 33. Clement K, Langin D: Regulation of inflammation-related genes in human adipose tissue. J Intern

Med 2007, 262:422-430. 34. Shah R, Lu Y, Hinkle CC, McGillicuddy FC, Kim R, Hannenhalli S, Cappola TP, Heffron S, Wang X,

Mehta NN et al.: Gene profiling of human adipose tissue during evoked inflammation in vivo. Diabetes 2009, 58:2211-2219.

35. Gomez-Ambrosi J, Catalan V, ez-Caballero A, Martinez-Cruz LA, Gil MJ, Garcia-Foncillas J, Cienfuegos JA, Salvador J, Mato JM, Fruhbeck G: Gene expression profile of omental adipose tissue in human obesity. FASEB J 2004, 18:215-217.

36. Nair S, Lee YH, Rousseau E, Cam M, Tataranni PA, Baier LJ, Bogardus C, Permana PA: Increased expression of inflammation-related genes in cultured preadipocytes/stromal vascular cells from obese compared with non-obese Pima Indians. Diabetologia 2005, 48:1784-1788.

37. Szalowska E, Elferink MG, Hoek A, Groothuis GM, Vonk RJ: Resistin is more abundant in liver than adipose tissue and is not upregulated by lipopolysaccharide. J Clin Endocrinol Metab 2009.

38. McIntyre E, Blackburn E, Brown PJ, Johnson CG, Gullick WJ: The complete family of epidermal growth factor receptors and their ligands are co-ordinately expressed in breast cancer. Breast Cancer Res Treat 2009.

39. Commins S, Steinke JW, Borish L: The extended IL-10 superfamily: IL-10, IL-19, IL-20, IL-22, IL-24, IL-26, IL-28, and IL-29. J Allergy Clin Immunol 2008, 121:1108-1111.

40. Sandona D, Betto R: Sarcoglycanopathies: molecular pathogenesis and therapeutic prospects. Expert Rev Mol Med 2009, 11:e28.

41. Byrne CD, Olufadi R, Bruce KD, Cagampang FR, Ahmed MH: Metabolic disturbances in non-alcoholic fatty liver disease. Clin Sci (Lond) 2009, 116:539-564.

42. Yilmaz Y, Ulukaya E, Atug O, Dolar E: Serum concentrations of human angiopoietin-like protein 3 in patients with nonalcoholic fatty liver disease: association with insulin resistance. Eur J Gastroenterol Hepatol 2009.

43. Pardo J, Aguilo JI, Anel A, Martin P, Joeckel L, Borner C, Wallich R, Mullbacher A, Froelich CJ, Simon MM: The biology of cytotoxic cell granule exocytosis pathway: granzymes have evolved to induce cell death and inflammation. Microbes Infect 2009, 11:452-459.

44. Barnes M, Freudenberg J, Thompson S, Aronow B, Pavlidis P: Experimental comparison and cross-validation of the Affymetrix and Illumina gene expression analysis platforms. Nucleic Acids Res 2005, 33:5914-5923.

171

Abbreviations

AdipoR1 adiponectin receptor 1 AMPK AMP-activated protein kinase ANOVA analysis of variance Apo apolipoprotein BMI body mass index CILAIR Comparison of Isotope Labeled Amino Acid Incorporation Rates CRP C-reactive protein CSF1 colony stimulating factor 1 CVD cardiovascular disease CX3CL1 chemokine (C-X3-C motif) ligand 1 CXCL10 C-X-C motif chemokine 10 CoA coenzyme A DAVID Database for Annotation, Visualization, and Integrated Discovery DIO diet induced obesity EAM energy absorbing molecule FC fold change FDR false discovery rate FFA free fatty acids FSIGT frequently sampled intravenous glucose tolerance testing GIP glucose dependent insulinotropic polypeptide GIPR glucose dependent insulinotropic polypeptide receptor GIPR KO glucose dependent insulinotropic polypeptide receptor knockout GLUT4 solute carrier family 2 (facilitated glucose transporter) member 4 GLUCR glucagon receptor GO gene ontology HDL high-density lipoprotein HOMA homeostasis model assessment IL-1b interleukin 1 b IL-6 interleukin -6 INSR insulin receptor IR insulin resistance Jak-STAT Janus kinase/signal transducers and activators of transcription LC liquid chromatography M-CS macrophage-colony stimulating factor MCP-1 monocyte chemotactic protein-1 MMP-1 matrix metalloproteinase-1 MS mass spectrometry m/z mass to charge ratio NFκB nuclear factor kappa-light-chain-enhancer of activated B cells LDL low-density lpoprotein LPL lipoprotein lipase LPS lipopolysachcaride PAI-I plasminogen activator inhibitor I PEDF pigment epithelium-derived factor PPARγ peroxisome proliferator-activated receptor gamma PTX3 pentraxin-related protein RBP4 retinol binding protein 4

172

ROS reactive oxygen species SAA serum amyloid A SAT subcutaneous adipose tissue SELDI surface-enhanced laser desorption/ionization SILAC stable isotope labeling by/with amino acids in cell culture SOCS suppressors of cytokine signaling SREBP1 sterol regulatory element-binding proteins STRING Search Tool for the Retrieval of Interacting Genes/Proteins T2D type 2 diabetes TG triglycerides TNFα tumor necrosis factor a TOF time of flight TZDs thiazolidinediones WC waist circumference VAT visceral adipose tissue

173

Acknowledgements

Dear Roel, starting from the beginning …as you may not remember anymore... I met you by coincident while having a meeting with someone else; at that time you had a PhD position which I decided to take. It was a pioneering project involving human adipose tissue and proteomics. With hardly any equipment except for a brand new SELDI machine and the only bench at the Pediatrics’ lab we were scrutinizing SELDI limits and searching for biomarkers in human adipose tissue culture media. And looking from this early moment to the present situation, it is actually hard to believe, how things changed and developed further.

I think working in your lab worked out for me very well, because of its pioneering character, such a situation stimulates to explore a lot of possibilities which you would not think of while working in a well established lab with well defined research questions. Dear Roel, I really enjoyed especially the last years of my PhD studies, when we gained a lot of experience, and most of techniques were established. I appreciate that you supported me in my choices and gave me a lot of freedom in doing what I thought was interesting to do. And your guidance helped me to keep focus. Thanks for the Keystone adventure; it was a pleasure to be there and I will never forget these beautiful mountains (and lectures of course) and the great time we spent there!

Dear Han and Annemieke, you both were involved in the adipose tissue project from the beginning. Han, I appreciate your efforts into the gradual development of proteomics in our lab, ending up with the CILAIR. Thanks a lot for your contribution and discussions within last years. And of course, I can not forget, you are a great skier and thanks for your tips which upgraded my skills and my first time on the black piste.

Dear Annemieke, I admire your enthusiasm and the passion you have for your work as a scientist and as medical doctor. Without your involvement we could not investigate the human adipose tissue and could not make this book.

Dear Marianne, you were also present from the beginning of “our” lab and for some while you were the only colleague struggling with the SELDI. Thank you for all the little things you helped me with, for the small talks and for being my paranymph.

Later we were joined shortly or permanently by other colleagues who worked also on the adipose tissue projects: Desiree (you developed into a multitasking technician impossible to substitute), Aldona, Suzette, Mariska, Jenneke, Karl-Loes, and Gloria-thank you all for your contributions and efforts.

Of course I can not omit the Pediatrics colleagues: Prof. Folkert Kuipers, Aldo, Marijke, Torsten, Fjodor, Dirk-Jan, Janine, Thierry, Meike, Jelske, Frans, Dirk-Jan, Renzee, Juul, Klary, Henk, Hilde, and co-members. I really enjoyed being part of your lab for a while and I learnt a lot from you guys!

174

After we moved to the new location several members enriched our team whose joint work contributed to this thesis. Marcel B. we all appreciate your knowledge and professionalism in the way how you deal with the mass spectrophotometers. Dear Fahrad, since you joined our lab you brought a lot of energy and enthusiasm. Within a short period of time it looked like you knew half of the UMCG employees. I wish you can realize your scientific dream once! Dear Marion, you are a very kind person. I value your scientific opinion and passion you have for your work. I wish you success in your further career. Martijn, my great appreciation for your both bioinformatical and non-bioinformatical, very often ad hoc, input. Not to be forgotten are my (inter)national room mates. Dear Saed, we had a kind of a tough start, but I am really happy that after all things worked out for us both quite well-a lot of success in your “second” life…, well things can only get better ... Kees it was a pleasure that you could join the GIP project and we could work for a while together. I appreciate your practical and intellectual contributions & thank you for being my paranymph. Dear Nicolai, we did not really were roommates but it felt like this. Your arrival to the Medical Biomics lab filled the gap for a social evening’s organizer. Ich wunsche dir, dein Buch rechtzeitig zu liefern, obwohl der Weg darnach kein Honiglecken ist. Moreover thanks for nice time to roommates and other colleagues -Jouke, Andrea, Heleen, Ma-ye, Marcel, Coby, Tao, Hong -waai, Sulima, Diederick.

My very special thanks go to Dr. Marieke Elferink and Prof. Geny Groothuis from the department of Pharmacokinetics, Toxicology, and Targeting. Dear Marieke and dear Geny, from my perspective I think it is very special how our collaboration developed-the primary aim we met for did not really work out, but the others resulted in publication(s). I really enjoyed working with you both, and I am grateful for your involvement, warmness, and stimulating discussions. I hope our collaboration will be successfully continued.

My very first paper would not see light if I would not have met Gerard te Meerman during one of the courses for PhD students. After that we had contact more frequently and you were always willing to explain to me diverse statistical issues. Thank you Gerard for your time, enthusiasm, and patience.

From the Department of Genetics I would like to thank Prof. Cisca Wijmenga for setting up the Illumina facility which we could make use of. Of course, the help of Marcel Bruinenberg with practical Illumina issues is highly appreciated. I see myself lucky I could meet Martin Wapenaar, a great teacher, you Martin explained me a lot of topics related to DNA microarrays and help me to plan my very first arrays experiments. Thanks a lot for your time, openness and stimulating discussions!

175

Our floor colleagues are also on the thank you list. Prof. Marten Hofker I would like to express my appreciation for your willingness to cooperate. Marcel thanks for your cooperating attitude as well, it is a pity there was not enough time to explore all the scientific possibilities. Dear Jana, yes still we have to finish the TUB project and it was nice to collaborate with you and to get to know you. Niels I am very happy that after all difficulties we were able to finish the animal experiment and publish the results. It was fun to do this animal experiment with you. Thanks for your enthusiasm and help, you are a great improviser and I really learnt from you a lot. I also would like to thank Henk van der Molen who was guiding us during beginning of animal experiments.

My words of appreciation go also to Prof. Ingrid Molema. We met shortly, but I learnt from you a bit about angiogenesis and of course in your lab I could practice stainings of adipose tissue. Henk thanks a lot for making the adipose tissue sections for me and your help in other experiments.

Dear Jolien, you made few times my life easier. While living in Arnhem and still working in Groningen you spent (too much) time getting all these signatures on the copyright forms… Thanks also for your help and time we spent together in our department!

Dearest Sacha, Filip, and Nadia ………………………buziaki, you guys make my days!

So this is briefly how we managed to get to the

HHHHAAAAPPPPPllll-EEEEND!

Ewa

176

177

Curriculum Vitae

Ewa Szalowska was born on 14th of May 1975, in Wroclaw, Poland. She studied

Biotechnology/Biochemistry at the University of Wroclaw, Poland. As a student she visited

Laboratory of Microbial Physiology led by Prof. Lubbert Dijkhuizen where she performed

one of her Master Projects. After graduating in Poland she came back to the Netherlands and

was appointed as a PhD student in the Department of Developmental Genetics in Haren.

After 2 years she resigned this job and started a new PhD position under the supervision of

Prof. Roel Vonk in the Department of Pediatrics. The appointment was continued under the

same supervision in the Department of Medical Biomics in the UMCG/University of

Groningen.

Since 2010 she is appointed as a postdoc at RIKILT/University of Wageningen within the

Netherlands Toxicogenomics Center project. She is working on pharmacologically induced

liver pathologies (cholestasis, steatosis, and necrosis) in order to study mechanisms behind

these processes and to identify biomarkers of early liver injury. Moreover, she will continue

to study adipose tissue physiology and its interactions with the liver in relation to

environmental pollutants and type 2 diabetes.