54.full

7/28/2019 54.Full

1/8

http://dvr.sagepub.com/Diabetes and Vascular Disease Research

http://dvr.sagepub.com/content/2/2/54The online version of this article can be found at:

DOI: 10.3132/dvdr.2005.009

2005 2: 54Diabetes and Vascular Disease ResearchEleanor M Scott, Angela M Carter and John BC Findlay

The application of proteomics to diabetes

Published by:

http://www.sagepublications.com

can be found at:Diabetes and Vascular Disease ResearchAdditional services and information for

http://dvr.sagepub.com/cgi/alertsEmail Alerts:

http://dvr.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://dvr.sagepub.com/content/2/2/54.refs.htmlCitations:

What is This?

- May 1, 2005Version of Record>>

at MAHATMA GANDHI INST MED SCI on April 4, 2013dvr.sagepub.comDownloaded from
http://dvr.sagepub.com/http://dvr.sagepub.com/http://dvr.sagepub.com/content/2/2/54http://dvr.sagepub.com/content/2/2/54http://dvr.sagepub.com/content/2/2/54http://www.sagepublications.com/http://dvr.sagepub.com/cgi/alertshttp://dvr.sagepub.com/cgi/alertshttp://dvr.sagepub.com/subscriptionshttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsReprints.navhttp://www.sagepub.com/journalsPermissions.navhttp://dvr.sagepub.com/content/2/2/54.refs.htmlhttp://dvr.sagepub.com/content/2/2/54.refs.htmlhttp://dvr.sagepub.com/content/2/2/54.refs.htmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://dvr.sagepub.com/content/2/2/54.full.pdfhttp://dvr.sagepub.com/content/2/2/54.full.pdfhttp://dvr.sagepub.com/http://dvr.sagepub.com/http://dvr.sagepub.com/http://online.sagepub.com/site/sphelp/vorhelp.xhtmlhttp://dvr.sagepub.com/content/2/2/54.full.pdfhttp://dvr.sagepub.com/content/2/2/54.refs.htmlhttp://www.sagepub.com/journalsPermissions.navhttp://www.sagepub.com/journalsReprints.navhttp://dvr.sagepub.com/subscriptionshttp://dvr.sagepub.com/cgi/alertshttp://www.sagepublications.com/http://dvr.sagepub.com/content/2/2/54http://dvr.sagepub.com/

7/28/2019 54.Full

2/8

Abstract

Proteomics is the investigation of all the proteins andtheir various modifications making up a system, bethat a cell, tissue or organism. The techniques

involved in proteomics allow the global screening of com-plex samples of proteins and provide qualitative andquantitative evidence of altered protein expression. Thislends itself to the investigation of the molecular mecha-nisms underpinning disease processes and the effects oftreatment. This review describes the main techniques ofproteomics and how they have begun to be applied todiabetes research.

Diabetes Vasc Dis Res 2005;2:5460

Key words: proteomics, diabetes, two-dimensional gelelectrophoresis (2DE), mass spectrometry, MALDI, ESI,SELDI.

IntroductionDespite continued advances in our understanding of thecomplex molecular mechanisms underlying the develop-ment of diabetes mellitus (DM) and its complications, it isclear that our knowledge is still limited (for reviews, see1-3).Given the predicted explosion in the number of cases of

type 2 diabetes mellitus (T2DM) worldwide,4

continuedresearch is essential, particularly with a view to understand-ing the impact of environmental stimuli on the physiologicalprocesses contributing to the development and progressionof disease.

DM and its complications arise as a consequence ofdefects in a variety of different tissues.3 The liver, pancreasand skeletal muscle are intimately involved in glucosehomoeostasis and insulin resistance, and as such are impor-tant targets for research into the pathophysiology of diabetesand drug development. Adipose tissue is increasingly beingviewed as key in the development of insulin resistance, byvirtue of the activity of the hormones and cytokines it pro-duces, and as a result of the epidemiological link betweenobesity and the development of T2DM/insulin resistance.5,6

In the wider context of insulin resistance, dyslipidaemia is animportant component, involving adipose tissue and the liverin its pathogenesis and expression. Ultimately, the main clin-ical manifestation of diabetes is vascular disease:7 macrovas-cular ischaemic heart disease and stroke are the main caus-es of death, dramatically reducing life expectancy, andmicrovascular complications (retinopathy, nephropathy andneuropathy) are a major cause of morbidity.3

The field of diabetes research has expanded as the labo-ratory and bioinformatics tools for unravelling complex phe-notypes have evolved. Proteomics is the latest research tool tobe employed in this context, and it perhaps holds more

promise than the genetic analyses that have been prevalentfor the past decade or so. The aim of this review article is toprovide the reader with an overview of proteomic techniquesand their potential application to the study of diabetes.

What is proteomics?In a manner analogous to the terminology of genetic studies,the term proteome is used to describe the entire proteincomplement of a system and proteomics is the study of thiscomplement.8 Proteins are the principal mediators of everycellular process, and the proteome of an organism is farmore complex than its genome. Consequently, it is more

informative. The genome remains stable throughout life andthe genome isolated from one cell type is (for the most part)identical to that isolated from another. Thus, the genomecannot provide information regarding the impact of envi-ronmental and other stimuli on a particular organism or cell.

Although each of our cells contains all the genes required tomake a complete human being, not all of these genes areexpressed in every cell or at all times. Therefore, whilst anygiven organism has one genome, by the nature of selectiveexpression of the genes in different cells and under differentstimuli, it can be considered to have many proteomes.

To complicate matters further, a protein coded for by asingle gene can occur in several different forms as a conse-quence of alternative splicing of mRNA transcripts and post-translational modifications (PTMs), see figure 1. At least 200different PTMs have been identified, including glycosylation,phosphorylation and non-enzymatic glycation (see DeltaMass, http://www.abrf.org/index.cfm/dm.home, a databaseof protein post-translational modifications).

A quick search using the word proteomics on Medlineis revealing, with obvious exponential growth in the publi-cation of articles utilising the technology of proteomics, par-ticularly in the last year. By characterising alterations in thepresence and types of proteins in target tissues and cells,proteomics holds the promise of increasing our understand-

ing of the mechanisms of disease and identification of targetsfor the development of novel drug therapies. Although thereare limited proteomics-based studies to date in the field of

REVIEW

The application of proteomics to diabetesELEANOR M SCOTT, ANGELA M CARTER, JOHN BC FINDLAY

Academic Unit of Molecular Vascular Medicine, The LIGHT Laboratories,Clarendon Way, University of Leeds, Leeds, LS2 9JT, UK.Eleanor Scott, Senior Lecturer in Medicine and Honorary ConsultantPhysician

Angela Carter, Principal Research Fellow in Molecular Vascular Medicine

Department of Biochemistry, Leeds University, Leeds, UK.John Findlay, Professor of Biochemistry

Correspondence to: Dr Eleanor ScottAcademic Unit of Molecular Vascular Medicine, The LIGHT Laboratories,Clarendon Way, University of Leeds, Leeds, LS2 9JT, UK.Tel: +44 (0)113 343 7719; Fax: +44 (0)113 343 7738E-mail: [email protected]

54 DIABETES AND VASCULAR DISEASE RESEARCH

http://dvr.sagepub.com/http://dvr.sagepub.com/http://dvr.sagepub.com/http://dvr.sagepub.com/

7/28/2019 54.Full

3/8

diabetes, those that have been performed offer an insightinto the power and impact that proteomics is likely to havein the future, as outlined in the following sections.

Proteomic technologiesProteomic analyses have two different aspects expressionproteomics and functional proteomics.9 Expression pro-teomics broadly involves the identification and quantifica-tion of proteins and characterisation of splice variants, PTMsand cellular localisation. Expression proteomics technologiesinvolve one or more separation steps followed by identifica-tion using mass spectrometry (figure 2). The classical pro-

teomic technique is two-dimensional gel electrophoresis(2DE), followed by protein identification by matrix-assistedlaser desorption ionisation (MALDI) mass spectrometry (MS).More recent developments include liquid chromatographycoupled to electrospray ionisation (ESI) MS/MS, and surface-enhanced laser desorption ionisation (SELDI) MS.

The above techniques are generally used for compre-hensive protein profiling. However, they can be precededby a variety of procedures for reduction of sample complex-ity, including multidimensional liquid chromatography, cellu-lar fraction, immunoprecipitation (IP) and affinity chro-matography, to name but a few.

Functional proteomics includes techniques for analysis ofprotein-protein interactions and protein networks such asthe two-hybrid systems, surface plasmon resonance (SpR)and nuclear magnetic resonance (NMR).10,11 It is beyond thescope of this review to cover all aspects of proteomics, andtherefore it will focus on providing a brief overview of massspectrometry-associated technologies and their applicationto the study of T2DM, to be outlined in the following sec-tions. For detailed descriptions of the various proteomictechnologies, see references.8-13

Mass spectrometry-associated proteomic analyses2DE and MALDI-MS

In this system, complex mixtures of proteins are separated bygel electrophoresis. In the first dimension, separation is basedon relative charge according to isoelectric point (isoelectric

focusing [IEF]); in the second dimension, separation is basedon molecular mass.14 Once separated, the proteins are visu-alised as an array of apparent spots by treating the gel withone of a variety of protein stains, including Coomasie Blue,silver and Sypro Ruby, or by pre-staining with Cy dyes (100sto 1,000s of spots can be resolved on a single gel). Thestained protein spots are imaged by specialised equipmentand differences in protein expression between samples canbe determined using specialised software.15 A variety ofanalysis programs are available which allow for comparisonof multiple samples by aligning gel images. Assessment canthen be made of changed protein patterns (often character-istic of PTMs) or levels (indicative of variant expression).

For protein identification, protein spots of interest areexcised from the gel, fragmented by proteases (most oftentrypsin) and the resulting mixture of peptides is then spottedonto a MALDI-MS plate. The samples are then dried andcoated with MALDI matrix to promote peptide ionisation forMS analysis. This involves separation of peptides accordingto their time of flight (TOF), which is dependent upon massto charge (m/z) ratio. The resultant peptide m/z fingerprintis then compared against databases (such as Mascot,www.matrixscience.com) of theoretical peptide fingerprintsof all proteins to identify the protein(s) of interest.16

Liquid chromatography coupled to MS (LC-MS)In its simplest form, LC-MS involves the proteolytic digestion(most usually with trypsin) of complex mixtures of proteins,

REVIEW

Figure 1. One gene can code for many proteins as a result ofsplicing and post-translational modifications. This has effectson the proteins function and alters the phenotype.Gene-environment interactions also alter protein behaviour.Thus a proteome cannot be predicted from the genome

PhenotypeGenotype

DNA mRNA

Splice variants

Protein Post-translationalmodifications

Protein exists in several forms,each with different functions

Figure 2. Principles of proteomics

Complex proteinmixture

Protein orpeptideseparation

LC

Proteinseparation2DE

ProteinseparationSELDI chip

Stained gel spotsimaged and analysed

Spots excised and digested

Mass spectrometry

ESI-MS/MSMALDI-MS/MS

Mass spectrometry

MALDI-MS

Mass spectrometry

SELDI-MS

Database search

Protein identification

Database search


Protein profile analysis

Biomarker discovery

Database search


VOLUME 2 ISSUE 2 . MAY 2005 55


7/28/2019 54.Full

4/8

which are then simplified by one or more liquid chromato-graphy steps based on different peptide properties such ascharge, hydrophobicity or pH. A combination of strongcation exchange chromatography followed by reversedphase chromatography is commonly used; see Wang andHanash for a review of the variety of multidimensional chro-matography approaches used in proteomics.9

Peptide analysis is (most usually) by ESI tandem MS(MS/MS). In MS/MS analysis the peptides are furtherresolved in MS mode: selected peptides are isolated andsubjected to further fragmentation in a gas-containing colli-sion cell before being introduced into a second phase of MSanalysis to provide sequence information.10 In tandem MSanalysis, different combinations of mass analyser can be cou-pled, including TOF, ion trap, quadrupole and Fourier trans-form ion cyclotron.10 ESI is most commonly coupled to iontrap and/or quadrupole mass analysers.

NanoLC systems are a recent development which enablespotting onto MALDI plates for analysis, and tandem MALDI

systems (either TOF/TOF or quadrupole/TOF) are increas-ingly being utilised for peptide identification.10 With MS/MSanalysis, protein identification is more certain as linearsequences of six or more residues can be enough to identi-fy the protein unambiguously. Often sequences can beobtained from more than one peptide.17,18

SELDI-MSIn SELDI analysis the sample to be analysed is applied to aprotein chip of defined binding property (e.g. hydrophobic,cation, anion, metal). This on-chip fractionation allows forretention of proteins of potential interest and reduction in

sample complexity. After washing the chip to removeunwanted proteins, the energy-adsorbing matrix is addedprior to laser-induced ionisation of bound proteins and TOFseparation to determine m/z. For a review of SELDI tech-nologies, see Tanget al.13

SELDI has most commonly been applied to clinical pro-teomics as a means of identifying protein profiles which dif-ferentiate different clinical groups, and in its simplest formprovides no protein identification. SELDI analysis relies oncomplex bioinformatics and statistical tools to mine the pro-tein m/z data to define the protein profiles that are uniqueto a particular patient group, particularly if unique markersare present in low quantities. Further downstream analysesare required, including protein digestion and peptide massfingerprinting or MS/MS analysis, to obtain protein identifi-cation to inform further functional analyses.13

Advantages and limitations of these techniques2DETwo-dimensional gel electrophoresis and protein analysis byspot-picking and trypsin digestion followed by MALDI-MSpeptide analysis provides a number of advantages over othermethods. 2DE is the best validated method for analysis ofwhole proteins; there is a variety of stains available, whichcan be utilised to obtain quantitative data using a variety of

imaging and software packages for gel analysis. 2DE has theadvantage of providing visualisation of the proteins and con-sequently, at present, it is the best method for identification

of proteins which have a variety of splice variants and post-translational modifications.

However, 2DE also has a number of limitations.Membrane proteins and extremely acidic, basic orhydrophobic proteins do not seem to be readily resolved by2DE (in these cases 1D PAGE is often utilised, although thiswill not provide any information regarding post-translationalmodifications). In addition, 2DE does not have sufficient res-olution for examining the whole dynamic range of proteomecoverage simultaneously, as the currently available stainsenable detection of proteins in the low ng range andabove.10,11,15 This is particularly disadvantageous where theamount of sample available for analysis is limited.

For abundant samples (such as plasma), a variety of frac-tionation methods can be used prior to 2DE to increase pro-teome coverage. For instance, IEF fractionation with pre-selection of narrow IEF ranges, or 2D liquid chromatographyin conjunction with the use of large 2D gel formats (toincrease sample loading), enable increased proteome cover-

age by improving the detection of low abundance pro-teins.10,11

LC-MS/MSLC-MS/MS allows for high-throughput applications, withidentification of many hundreds of proteins in 24 hours, andit can cover a wider dynamic range of protein concentrationthan 2DE. Limitations of this method include the fact thatthe MS analysis does not give complete peptide coverage ofa protein. Consequently, it does not readily lend itself toidentification of splice variants and PTMs in complex mix-tures.

There are also issues regarding quantitation with thisapproach, although new methods for relative quantitationhave been established, most notably with the introduction ofstaple isotope tagging of proteins, including isotope codedaffinity tags (ICAT, Applied Biosystems, www.applied-biosystems.com). ICATs comprise a biotin tag with either alight or heavy (deuterium or 13C) linker chain and a reactivegroup, which covalently binds to cysteine residues. Labellingtwo samples to be compared with the two forms of ICAT andmixing the samples prior to LC-MS/MS enables relativequantification of peptides by comparing the relative abun-dance of light and heavy ICAT-tagged peptides. The tagenables separation of biotin-labelled peptides prior to LC-MS/MS analysis in order to reduce sample complexity. Thedisadvantage of this system is that it only allows for pairwisecomparisons and is selective for cysteine-containing pep-tides. Also, the presence of the ICAT tag can impair ionisa-tion, resulting in reduced sensitivity.

A recent development which can compare up to four dif-ferent samples directly is the introduction of iTRAQ(Applied Biosystems, www.appliedbiosystems. com), stableisobaric reagents which are amine-specific and thereforehave the advantage that they label all peptides in a sample.The iTRAQ reagents comprise a peptide-reactive groupwhich covalently binds to lysine and N terminal moieties of

peptides, a variably sized reporter group and a balancegroup (which maintains equal peptide mass in MS mode).

The advantage of this system over the ICAT system is that

REVIEW



7/28/2019 54.Full

5/8

in MS mode peptides labelled with the different iTRAQreagents will be observed as a single peak whereas in MS/MSmode they are readily ionised, the balance group is lost andthe reporter groups are detected in the low m/z regionbetween 113 and 119; comparison of the reporter groupintensities provides relative quantification of the differentlabelled samples.

Despite these advances, the technology is not sufficient-ly developed at present to enable high-throughput analysisof large clinical case control studies using this tagged proteinapproach.

SELDIThere has been much discussion of late of the relativeadvantages and disadvantages of SELDI-based analyses. Itis clear that whilst the SELDI approach holds muchpromise, major issues, including reproducibility of resultsboth within and between laboratories, need to be resolvedif it is to become widely accepted as a tool in clinical pro-

teomics. In addition, there is still the need to develop strin-gent and validated bioinformatics and statistical tools formining SELDI data for predictive protein m/z profiles.19 Theuse of SELDI analysis solely to identify protein profiles asbiomarkers for disease without subsequent identification ofthe protein biomarkers should also be questioned. For clin-ical research it will be important to determine protein bio-marker identities to enable functional proteomic analysesto be carried out, which enhance both our knowledge ofdisease processes and the potential for therapeutic inter-vention.

Other issuesRegardless of the approach chosen for proteomic analysis,stringent procedures for the taking of samples and their pro-cessing (including storage where relevant) prior to proteom-ic analysis must be followed in order to ensure high-quality,reliable results. The quality of the results will be a reflectionof the quality of the samples analysed. The HumanProteome Organisation (www.HUPO.org) has been estab-lished to address many of the issues related to standardisa-tion of proteomic analysis.20 However, it is beyond the scopeof this review to cover these issues.

The application of proteomics to diabetes researchAs outlined in the introduction, there are key organs and tis-sues known to contribute to the development of DM and itscomplications. To define the underlying pathophysiologicalprocesses which contribute to disease development and toidentify key pathways to target in the development of noveltherapeutics will require a variety of approaches. These willinclude, for example, analysing tissue- and cell-specific pro-tein profiles under conditions which mimic DM and differ-ent environmental stimuli thought to contribute to the devel-opment of DM; and analysing samples from patients withDM, with insulin resistance or impaired glucose toleranceand comparing them to healthy subjects. Outlined below

are some of the key tissues and organs involved in DM andhow the few proteomic analyses to date have been appliedin investigating these areas.

PancreasInsulin is synthesised in and secreted from the cells of theislets of Langerhans in the pancreas.3 The main stimulator ofinsulin release is glucose, and once insulin is released itexerts its effects by binding to the insulin receptor that isexpressed on target cell membranes. Insulin receptor-ligandinteractions stimulate tyrosine kinase activity, which is essen-tial for insulins action. The post-receptor signalling mecha-nisms for insulin ultimately result in glucose being carriedinto cells across the cell membrane by glucose transporterproteins (GLUTs).2 Type 1 diabetes mellitus (T1DM) is char-acterised by a pancreatic insulin secretion defect21,22 where-as T2DM is characterised by insulin resistance in target tis-sues, often combined with a defect in pancreatic insulinsecretion.23

Despite the fact that the pancreas is a key target organ,few proteomics studies have been performed and, to date,none have been done in human cells in relation to dia-betes. The human pancreas has been explored taking a pro-

teomics approach and a reference map of 302 proteins hasbeen identified.24 The same has been done in mouse islets,25

and both of these will provide a useful reference for futureproteomics studies of the pancreas in diabetes. Cytokinesreleased from macrophages are known to be toxic to isletcells, and this is thought to be one of the mechanismsinvolved in the development of type 1 DM in geneticallypredisposed individuals.21 On this basis proteomic studieshave been performed on rat islet cells,26 exposing them tothe cytokine interleukin (IL)-1. More than 2,000 proteinswere identified, of which 105 had altered expression fol-lowing exposure to IL-1. The functional relevance of these

changes to human disease is currently unclear.Proteomic analysis of murine pancreatic islet cells hasidentified proteins that are potentially involved in insulinresistance in mice since Lep/Lep mice were found to havealtered actin-binding proteins, which may contribute to isletcell dysfunction and reduced insulin secretion.27 Using a pro-teomics approach, the effect of treatment has been studied.

Peroxisome proliferator-activated receptors (PPARs) aretranscription factors which through gene regulation have keyroles in glucose and lipid homoeostasis. PPAR agonists (thi-azolidinediones) are increasingly used in clinical diabeticpractice to improve insulin sensitivity. Treatment with thia-

zolidinediones in Lep/Lep mice leads to increased expres-sion of carboxypeptidase B, a protein involved in insulin pro-cessing, which suggests that PPAR agonists may normaliseglucose homoeostasis in part through improved insulin pro-cessing.27

LiverThe liver is pivotal in the regulation of fatty acid, lipoproteinand carbohydrate metabolism, playing a key role in glucosehomoeostasis.3,28 In the fasted state the liver maintains a sup-ply of glucose by gluconeogenesis and glycogenolysis. In thefed state insulin switches this off by inhibiting hepatic glu-cose production and stimulating glycogen synthesis. In

insulin resistance, there is a failure to suppress hepatic glu-cose production and glycogenolysis. This is accompanied byfat accumulation within the hepatocytes, with increased free

REVIEW



7/28/2019 54.Full

6/8

fatty acid uptake and synthesis of triglycerides, leading toworsening of insulin resistance.

As discussed above, PPARs are nuclear receptorsinvolved in regulating lipid metabolism and glucosehomoeostasis. PPAR is preferentially expressed in the liverand its activation leads to increased oxidation of fattyacids.29 It is therefore a target for the treatment of dyslipi-daemia, and PPAR agonists normalise elevated plasmatriglycerides and hyperglycaemia in insulin-resistant mice.29

Utilising a proteomics approach, the effect of a PPAR ago-nist on the liver proteins in insulin-resistant mice has beeninvestigated. It was shown to upregulate at least 16 proteinspots: when these were identified by MS, 14 were found tobe components of peroxisomal fatty acid metabolism, con-firming that PPAR induces fatty acid -oxidation andupregulates enzymes involved in lipogenesis.29,30

MuscleSkeletal muscle is the main site of glucose uptake and stor-

age by insulin. Reduced oxidative enzyme capacity and dys-function in skeletal muscle mitochondria, with impairedphosphorylation in insulin-resistant skeletal muscle, suggeststhat defects in AMP-activated protein kinase pathways areinvolved in the development of insulin resistance.31 Thereare many potential enzymes and pathways involved andconsequently this field is difficult to study.2 Proteomics offersthe techniques to evaluate many of these proteins simulta-neously.

By using the 2DE approach, skeletal muscle biopsiesfrom 10 patients with T2DM and 10 controls were com-pared in one analysis.32 Eight proteins were overexpressed in

T2DM skeletal muscle compared to controls. They were allfound to be linked to increased cellular stress and alterationsin skeletal muscle mitochondrial metabolism. In addition,two proteins that have a crucial role in ATP synthesis (CKBand ATP synthase) were underexpressed in the T2DM sam-ples. At this stage it is not clear whether these are primarydefects or changes resulting from T2DM. Future studiesusing healthy first-degree relatives of patients with T2DMmay help to differentiate between primary defects and sec-ondary effects.

Adipose tissueWeight maintenance depends on balancing energy inputwith energy expenditure, which is coordinated by signalsbetween the brain (hypothalamus/arcuate nucleus) and adi-pose tissue. Obesity arises when there is an imbalancebetween energy input and energy expenditure, resulting inadipocyte hypertrophy (increased cell size) and hyperplasia(increased cell number). The consequences of obesityinclude increased secretion of adipocyte proteins, includinginflammatory mediators, angiotensinogen and plasminogenactivator inhibitor-1 (PAI-1), and these proteins have beenassociated with the development of T2DM.33

Obesity is rapidly becoming one of the major health careissues of the 21st century. This is mirrored by increases in the

pathological consequences of obesity, with classical T2DMbeing diagnosed in children and adolescents for the firsttime. There is growing evidence that the adipocytes of obese

individuals produce and secrete proteins which promote thedevelopment of T2DM.34,35 It is clear that abnormaladipocyte function is associated with T2DM, although itremains unclear what the underlying mechanisms promotingthe onset of T2DM are and why some obese individuals donot go on to develop complications such as T2DM.

Novel adipocyte proteins (adiponectin, for example)continue to be identified and characterised, and conse-quently an understanding of both the cellular and secretedadipocyte protein complement and the factors associatedwith pre-adipocyte differentiation will provide insights intothe mechanisms involved in the regulation of adiposity andthe pathogenesis of obesity and T2DM. It has been suggest-ed that adipocytes have a finite capacity for lipid storage(and consequently a finite size) and that fully lipid-ladenadipocytes secrete factors to promote pre-adipocyte prolif-eration and differentiation into mature adipocytes in orderto facilitate further lipid storage. Some individuals appear tohave a defect in this differentiation process, leading to

increased hypertrophic adipocytes which have been report-ed to secrete excess proteins involved in inflammatoryprocesses (e.g. tumour necrosis factor [TNF]), consequent-ly contributing to the pathological consequences of obesity.

Limited proteomic analyses have been carried out in themurine 3T3 L1 cell line to determine variations in cellularand secreted proteins during differentiation.36-39An increasein a variety of mitochondrial proteins, which was associatedwith observed differences in mitochondrial morphology, wasobserved during differentiation and treatment with thePPAR agonist rosiglitazone.37 Analyses of membrane andsecreted proteins during differentiation identified a number

of proteins not previously known to be secreted byadipocytes38,39 and membrane proteins selectively expressedby adipocytes.36

Analysis of the cellular and secreted proteins of pre-adipocytes and adipocytes from normoglycaemic individualsand from individuals with insulin resistance, impaired glu-cose tolerance and T2DM to determine differences in quan-tity and structural modifications (splice variants, glycationetc.) will provide invaluable insights into the mechanismsunderlying the pathogenesis of T2DM.

HeartDiabetic cardiomyopathy is a recognised complication ofDM, leading to congestive cardiac failure. Proteomics hasbeen applied to the study of cardiac proteins in a type 1 DMmouse model.40 Twenty proteins were altered in mice withcardiomyopathy compared to controls, 12 of them mito-chondrial proteins. This may reflect oxidative stress damag-ing the mitochondria in DM and contributing to the devel-opment of cardiomyopathy.

Kidney and urineDiabetic nephropathy is the most common cause of renalfailure worldwide.41 Using a mouse model of type 1 DMnephropathy, the renal proteome was compared to that of

non-diabetic mice.42 Whole kidneys were lysed and proteinsseparated by 2DPAGE, then identified by MALDI-TOF. Ofthose proteins that were altered, it was noted that elastase

REVIEW



7/28/2019 54.Full

7/8

IIIB was reduced and accompanied by an increase in mono-cyte neutrophil elastase inhibitor, with accumulation ofelastin throughout the kidney. Subsequent examination ofhuman renal biopsies by immunohistochemistry of elastinconfirmed increased elastin expression, suggesting it has arole in the pathophysiology of diabetic nephropathy.

It is widely accepted that early identification of patientswho are at risk of developing diabetic nephropathy is impor-tant, and urine is currently screened for the presence ofmicroalbumin.41 Proteomic techniques have thus beenapplied to examine the urine of healthy subjects and tocompare the proteome seen with that of the proteome ofthe urine of diabetic subjects with and without albumin-uria.43 Diabetic subjects without albuminuria had noticeablydifferent polypeptide patterns to controls, and in those sub-

jects with albuminuria a different polypeptide profile againwas seen.

Using MS/MS, three of the discriminating polypeptideshave been identified, and they are small fragments of pro-

teins known to be affected by the presence of renal disease.Because of their small size it would seem possible that theycould serve as very early indicators of renal damage, beforethe glomerular filter has become damaged enough to letthrough the albumin that we currently use to identify renaldisease.

Plasma and serumExtracellular fluids contain many secreted proteins and thesechange in response to the state of the organism. With diseaseit is likely that there may be changes in the protein compo-sition of extracellular fluids. In comparison to the tissues out-

lined above, plasma (and serum) also has the advantage ofbeing easy to obtain in relatively large quantities, which is anobvious advantage for clinical research. Patients are alsomore likely to provide consent for a blood sample in prefer-ence to tissue biopsies. Plasma is a valuable resource for pro-teomic analysis, since it comprises components of virtuallyevery physiological and pathological process occurring in thehuman body.

Since T2DM is associated with defects in multiple organs,proteomic analysis of plasma can provide much informationrelevant to the study of diabetes and its complications andhas the potential to provide diagnostic markers. As Andersonand Anderson describe: The human plasma proteomeholds the promise of a revolution in disease diagnosis andtherapeutic monitoring provided that major challenges inproteomics and related disciplines can be addressed. Plasmais not only the primary clinical specimen but also representsthe largest and deepest version of the human proteome pre-sent in any sample.44 It is apparent that at present we haveidentified only a minor fraction of the total protein variantspresent in plasma, estimated to be in the order of 500,000(excluding the immunoglobulin fraction) if the differentsplice variants and PTMs are included.44

The main challenge in the proteomic analysis of plasmais the wide dynamic range of protein concentration (in the

region of 10 orders of magnitude), which has implicationsfor the analysis of lower concentration components.44

Initiatives such as the Human Proteome Organisation

(HUPO) Plasma Proteome Project (PPP) are enabling thestandardisation of plasma proteomic analyses in the clinicalcontext.45 Furthermore, improved systems for easy samplefractionation, image analysis and data analysis, and methodsfor depletion of high abundance proteins, have made possi-ble the study of plasma samples from clinical cohorts. Formost applications the first step in the proteomic analysis ofplasma (or serum) involves depletion of the most abundantproteins (in particular albumin and immunoglobulin) by avariety of methods to enable analysis of the less abundantproteins. Additional fractionation steps will also increaseplasma proteome coverage.

To date, only one proteomic analysis of serum has beenreported in relation to DM.46 However, as no clinical char-acterisation of the study or control group was made it is dif-ficult to determine and interpret the relevance of their find-ings. Proteomics of serum, plasma or even peripheral bloodcells has potential utility for closer understanding of diseasemechanisms, detecting disease markers, new targets for drug

treatment, and the mechanisms involved in drug action. Itcould help to develop new markers for the presence of dis-ease and may eventually allow more effective screening ofdiabetic subjects who are at risk of diabetic complications.

ConclusionProteomic technologies will enable the analysis and com-parison of protein expression profiles in tissues and plasmaor serum from people with diabetes and normal controls, inanimal models of disease and in response to drug treatment.The concept of proteomics is straightforward but putting itinto practice remains challenging due in part to the complex

dynamic nature of the proteome and the seemingly limitlessdata that could be generated. It is also expensive and time-consuming, and should not be undertaken lightly. If appliedappropriately, however, proteomics offers the ability to iden-tify proteins involved in the development of insulin resis-tance, dysglycaemia and DM, and an understanding of howenvironmental factors can influence the development andprogression of disease. It has the potential to identify proteinmarkers of susceptibility to micro- and macrovascular dis-eases and to help our understanding of the mechanisms ofdrug action, whilst identifying new targets for therapeuticdevelopment.

Conflict of interestNone declared.

References1. Kelly MA, Rayner ML, Mijovic CH, Barnett AH. Molecular aspects of

type 1 diabetes. Mol Pathol 2003;56:1-10.2. White MF. IRS proteins and the common path to diabetes.Am J Physiol

Endocrinol Metab 2002;283:E413-E422.3. Pickup J, Williams G. Textbook of Diabetes. Oxford: Blackwell Science,

1997.4. Zimmet P. The burden of type 2 diabetes: are we doing enough?

Diabetes Metab 2003;29:6S9-18.5. Hsueh WA, Law R. The central role of fat and effect of peroxisome pro-

liferator-activated receptor-gamma on progression of insulin resistanceand cardiovascular disease.Am J Cardiol 2003;92:3J-9J.

6. Goldfine AB, Kahn CR. Adiponectin: linking the fat cell to insulin sen-sitivity. Lancet 2003;362:1431-2.

7. Stern MP. Do non-insulin-dependent diabetes mellitus and cardiovas-

REVIEW



7/28/2019 54.Full

8/8

cular disease share common antecedents?Ann Intern Med 1996;124:110-16.

8. Banks RE, Dunn MJ, Hochstrasser DF et al. Proteomics: new perspec-tives, new biomedical opportunities. Lancet 2000;356:1749-56.

9. Wang H, Hanash S. Multi-dimensional liquid phase based separationsin proteomics.J Chromatogr B Analyt Technol Biomed Life Sci 2003;787:11-18.

10. Aebersold R, Mann M. Mass spectrometry-based proteomics. Nature2003;422:198-207.

11. Zhu H, Bilgin M, Snyder M. Proteomics.Annu Rev Biochem 2003;72:783-812.12. Tyers M, Mann M. From genomics to proteomics. Nature 2003;422:

193-7.13. Tang N, Tornatore P, Weinberger SR. Current developments in SELDI

affinity technology. Mass Spectrom Rev 2004;23:34-44.14. O'Farrell PH. High resolution two-dimensional electrophoresis of pro-

teins.J Biol Chem 1975;250:4007-21.15. Sivakumar A. 2D gels and bioinformatics an eye to the future. In Silico

Biol 2002;2:507-10.16. Henzel WJ, Watanabe C, Stults JT. Protein identification: the origins of

peptide mass fingerprinting. J Am Soc Mass Spectrom 2003;14:931-42.

17. Mann M, Wilm M. Error-tolerant identification of peptides in sequencedatabases by peptide sequence tags.Anal Chem 1994;66:4390-9.

18. Perkins DN, Pappin DJ, Creasy DM, Cottrell JS. Probability-based pro-tein identification by searching sequence databases using mass spec-trometry data. Electrophoresis 1999;20:3551-67.

19. Baggerly KA, Morris JS, Coombes KR. Reproducibility of SELDI-TOFprotein patterns in serum: comparing datasets from different experi-ments. Bioinformatics 2004;20:777-85.

20. Hanash S. Building a foundation for the human proteome: the role ofthe Human Proteome Organization.J Proteome Res 2004;3:197-9.

21. Nerup J, Mandrup-Poulsen T, Helqvist S et al. On the pathogenesis ofIDDM. Diabetologia 1994;37(suppl 2):S82-S89.

22. Nerup J, Mandrup-Poulsen T, Molvig J, Helqvist S, Wogensen L, EgebergJ. Mechanisms of pancreatic beta-cell destruction in type I diabetes.Diabetes Care 1988;11(suppl 1):16-23.

23. Chiasson JL, Rabasa-Lhoret R. Prevention of type 2 diabetes: insulinresistance and beta-cell function. Diabetes 2004;53(suppl 3):S34-S38.

24. Hu L, Evers S, Lu ZH, Shen Y, Chen J. Two-dimensional protein data-base of human pancreas. Electrophoresis 2004;25:512-18.

25. Sanchez JC, Chiappe D, Converset V et al. The mouse SWISS-2D PAGE

database: a tool for proteomics study of diabetes and obesity.Proteomics 2001;1:136-63.26. Sparre T, Christensen UB, Gotfredsen CF et al. Changes in expression

of IL-1 beta influenced proteins in transplanted islets during develop-ment of diabetes in diabetes-prone BB rats. Diabetologia 2004;47:892-908.

27. Sanchez JC, Converset V, Nolan A et al. Effect of rosiglitazone on thedifferential expression of diabetes-associated proteins in pancreaticislets of C57Bl/6 lep/lep mice. Mol Cell Proteomics 2002;1:509-16.

28. Collier JJ, Scott DK. Sweet changes: glucose homeostasis can be alteredby manipulating genes controlling hepatic glucose metabolism. MolEndocrinol 2004;18:1051-63.

29. Edvardsson U, Alexandersson M, Brockenhuus VL et al. A proteomeanalysis of livers from obese (ob/ob) mice treated with the peroxisome

proliferator WY14,643. Electrophoresis 1999;20:935-42.30. Edvardsson U, von Lowenhielm HB, Panfilov O, Nystrom AC, Nilsson F,

Dahllof B. Hepatic protein expression of lean mice and obese diabeticmice treated with peroxisome proliferator-activated receptor activators.Proteomics 2003;3:468-78.

31. Winder WW, Hardie DG. AMP-activated protein kinase, a metabolicmaster switch: possible roles in type 2 diabetes. Am J Physiol 1999;277:E1-E10.

32. Hojlund K, Wrzesinski K, Larsen PM et al. Proteome analysis reveals

phosphorylation of ATP synthase beta-subunit in human skeletal mus-cle and proteins with potential roles in type 2 diabetes. J Biol Chem2003;278:10436-42.

33. Fruhbeck G, Gomez-Ambrosi J, Muruzabal FJ, Burrell MA. Theadipocyte: a model for integration of endocrine and metabolic signal-ing in energy metabolism regulation. Am J Physiol Endocrinol Metab2001;280:E827-E847.

34. Considine RV, Nyce MR, Morales LM et al. Paracrine stimulation ofpreadipocyte-enriched cell cultures by mature adipocytes.Am J Physiol1996;270:E895-E899.

35. Gabriely I, Ma XH, Yang XM et al. Removal of visceral fat preventsinsulin resistance and glucose intolerance of aging: an adipokine-medi-ated process? Diabetes 2002;51:2951-8.

36. Xu A, Choi KL, Wang Yet al. Identification of novel putative membraneproteins selectively expressed during adipose conversion of 3T3-L1cells. Biochem Biophys Res Commun 2002;293:1161-7.

37. Wilson-Fritch L, Burkart A, Bell G et al. Mitochondrial biogenesis andremodeling during adipogenesis and in response to the insulin sensitiz-er rosiglitazone. Mol Cell Biol 2003;23:1085-94.

38. Choi KL, Wang Y, Tse CA, Lam KS, Cooper GJ, Xu A. Proteomic analy-sis of adipocyte differentiation: Evidence that alpha2 macroglobulin isinvolved in the adipose conversion of 3T3 L1 preadipocytes.Proteomics 2004;4:1840-8.

39. Kratchmarova I, Kalume DE, Blagoev B et al. A proteomic approach foridentification of secreted proteins during the differentiation of 3T3-L1preadipocytes to adipocytes. Mol Cell Proteomics 2002;1:213-22.

40. Shen X, Zheng S, Thongboonkerd V et al. Cardiac mitochondrial dam-age and biogenesis in a chronic model of type 1 diabetes. Am J PhysiolEndocrinol Metab 2004;287:E896-E905.

41. Marshall SM. Recent advances in diabetic nephropathy. Clin Med2004;4:277-82.

42. Thongboonkerd V, Barati MT, McLeish KR et al. Alterations in the renalelastin-elastase system in type 1 diabetic nephropathy identified by

proteomic analysis.J Am Soc Nephrol 2004;15:650-62.43. Mischak H, Kaiser T, Walden M et al. Proteomic analysis for the assess-ment of diabetic renal damage in humans. Clin Sci (Lond) 2004;107:485-95.

44. Anderson NL, Anderson NG. The human plasma proteome: history,character, and diagnostic prospects. Mol Cell Proteomics 2002;1:845-67.

45. Omenn GS. The Human Proteome Organization Plasma ProteomeProject pilot phase: reference specimens, technology platform compar-isons, and standardized data submissions and analyses. Proteomics2004;4:1235-40.

46. Zhang R, Barker L, Pinchev D et al. Mining biomarkers in human serausing proteomic tools. Proteomics 2004;4:244-56.

REVIEW


54.full

Documents