proteomics in cardiovascular disease: clinical ...genomics mirnas maturity of research 3...

Proteomics in Cardiovascular Disease:Clinical Considerations

Christian Delles

1The screen versions of these slides have full details of copyright and acknowledgements

1


Dr. Christian DellesBHF Glasgow Cardiovascular Research Centre

Institute of Cardiovascular and Medical SciencesUniversity of Glasgow

2

Metabolomics

Proteomics

Transcriptomics

Systems Biology and "Omics"

DNA

mRNA

Protein

Metabolitessmall molecules

Genomics

miRNAs

Mat

urity

of r

esea

rch

3

Cardiovascular Continuum

Dzau V et al., Circulation2006

Risk factors

Oxidative and mechanical stressInflammation

Endothelial dysfunction

Atherothrombosis and progressive CV disease

Tissue injury(MI, stroke, renal

insufficiency,peripheral arterial

insufficiency)

Pathologicalremodeling

Target organ damage

End-organ failure(CHF, ESRD)

Death

Altered gene expressionAltered protein expression

Genome


Christian Delles


4

Proteomics

5

Proteomics (2)

Anderson NL & Anderson NG, Electrophoresis1998

The goal of proteomics is a comprehensive, quantitative description of protein expression and its changes

under the influence of biological perturbations such as disease or drug treatment

6

Publications on Proteomics

Van Eyk JE, Circ Res 2011


Christian Delles


7

Rationale for Proteomic Studies

• Research− Unravelling the pathophysiology

− Discovery of new pathways

• Clinical application: biomarker− Screening for disease

− Prediction of risk

− Diagnosis of (severity of) disease

− Monitoring of treatment

8

Samples

Tuñòn J et al., JACC 2010

9

Platforms

Tuñòn J et al., JACC 2010


Christian Delles


10Gerszten RE et al. Circ Res 2011

Platforms: LC-MS/MS

LC-MS/MSBiological samples (cases vs. control)

Data analysis

11Gerszten RE et al., Circ Res 2011

Technical ChallengesGenomics/transcriptomics Proteomics

• All possible features known

• Sample is static during analysis

• All features measured

• Robust means to amplify low numbers DNA or RNA (PCR)

• Signal not detected means feature not present

• All possible features not known

• Sample is dynamic during analysis

• 10-50% of features measured

• No protein PCR (analytics have to deal with enormous dynamic range)

• Signal not detected means either that feature not present or feature present but not detected

12

Post Translational Modifications

Van Eyk JE, Circ Res 2011


Christian Delles


13

Unravelling pathophysiology

Biomarker of disease

Pre-eclampsia

Organ damage

Biomarker of risk

14Iacobellis G et al., Nat Clin Pract Cardiovasc Med. 2005

Unravelling Pathophysiology: Epicardial Fat

15Salgado-Somoza A et al., Am J Physiol Heart Circ Physiol 2010

Unravelling Pathophysiology

Epicardial adipose tissue Subcutaneous adipose tissue

Catalase


Christian Delles


16Salgado-Somoza A et al., Am J Physiol Heart Circ Physiol 2010

Western Blot and IHCSAT

EAT

17

Function: NBT Assay of ROS

Salgado-Somoza A et al., Am J Physiol Heart Circ Physiol 2010

18



Pre-eclampsia

Organ damage

Biomarker of risk


Christian Delles


19

Urinary Proteomics: CE/MS PlatformCapillary electrophoresis coupled to mass spectrometry

UrineSample

Capillary Electrophoresis

Mass Spectrometry

Ionization

Report

Data Storageand Evaluation

Diagnosis Disease specificBiomarker pattern

Separation and analysis

of proteins and peptides

(>1,000)

Run time ~60 min

CE

• Fast

• Robust

• Inexpensive

• Reproducible

MS

• Resolution

• Scan speed

20

• Easily accessible

• Non invasive sampling

• Available in large quantities

• Urinary polypeptides are stable, yielding comparable datasets

• Urinary polypeptides display the “status” of the kidney, bladder, prostate and vascular architecture, are capable of depicting systemic diseases

Why Urine?

De Hortus SanitatisMainz, Germany, 1491

21

Urinary Proteomics: CE/MS Platform

Migration time (min)


Christian Delles


22

Methods

Cases ControlsCE/MSCE/MSDiagnostic pattern

Discriminatory biomarker

Discriminatory biomarker

Compiledpattern

Compiled pattern

Kolch et al., RCM 2004, 18: 2635-2636

Neuhoff et al., RCM 2004, 18: 149-156

Mischak et al., PROTEOMICS - Clinical Applications2007, 1: 148-156

23

CE/MS

Methods (2)

Cases ControlsDiagnostic pattern

Kolch et al., RCM 2004, 18: 2635-2636

Neuhoff et al., RCM 2004, 18: 149-156

Mischak et al., PROTEOMICS - Clinical Applications2007, 1: 148-156

24

Study cohort Samples CAD Control Primary Usage Secondary Usage

Biomarker discovery 586 204 382

CAD [9,10] (N=120†) 183 151 32 CAD markers SVM modelingUAP [10] (N=59) 59 35 24 SVM modeling n.a.CACTI [11] (N=33) 33 18 15 SVM modeling n.a.Additional controls [14] (N=153) 229 0 229 SVM modeling n.a.TRENDY, baseline [9,12] (N=17†) 14 0 14 Medication markers SVM modelingTRENDY, follow -up [9,12] 16 0 16 Medication markers SVM modelingFenofibrate, baseline [13] (N=26) 26 0 26 Medication markers SVM modelingFenofibrate, follow-up 26 0 26 Medication markers SVM modeling

Blinded cohort (N=138) 138 71 67 Validation n.a.

Short-term treatment effects [15] 193 n.a. n.a.

HIB 0 mg (N=55‡) 55 n.a. n.a. Drug interference n.a.HIB 300 mg 48 n.a. n.a. Drug interference n.a.HIB 600 mg 45 n.a. n.a. Drug interference n.a.HIB 900 mg 45 n.a. n.a. Drug interference n.a.

Long-term treatment effects [16] 44 n.a. n.a.

IRMA -2 Irbesartan baseline (N=11†) 11 n.a. n.a. Therapy monitoring n.a.IRMA -2 Irbesartan follow-up 11 n.a. n.a. Therapy monitoring n.a.IRMA -2 Placebo baseline(N=11†) 11 n.a. n.a. Therapy monitoring n.a.IRMA -2 Placebo follow -up 11 n.a. n.a. Therapy monitoring n.a.

Total (N=623) 961

Patients

Discovery

Adjustment for drug treatment

Effect of treatment

Blinded cohort

Delles C et al., J Hypertens 2010


Christian Delles


25

238 Biomarker Panel


26

Identification of Proteins

• Collagen type 1

• Collagen type 3

• Alpha-1-antitrypsin (AAT)

• Granin-like neuroendocrine peptide precursor (ProSAAS)

• Membrane associated progesterone receptor component 1

• Sodium/potassium-transporting ATPase gamma chain

• Fibrinogen-alpha-chain


27

Single Marker

Marker 1

CaseControl


Christian Delles


28

Two Markers

Marker 1

Marker 2

CaseControl

29

Three Markers

Marker 1

Marker 2

CaseControl

30 1 - Specificity1.00.80.60.40.20.0

Sen

sitiv

ity

1.0

0.8

0.6

0.4

0.2

0.0

CAD Controls

24 M

arke

rs

AUC 0.786

50 M

arke

rs

AUC 0.786AUC 0.882

238

Mar

kers

Better Discrimination with More Markers

CAD Controls


Christian Delles


31

Effect of Drug Therapy

10-week treatment with irbesartan

2-year treatment with irbesartan


32



Pre-eclampsia

Organ damage

Biomarker of risk

33

Pre-eclampsia

Hypertension

Proteinuria

> 20 weeks gestation

Without pre-existing hypertension

Maternal

• Oedema

• Headaches/blurred vision/seizures

• Renal failure

• Coagulation problems

• HELLP syndrome

Foetal

• Intra-uterine growth restriction

• Preterm delivery

• Death


Christian Delles


34

Normal Pregnancy

http://www.sgul.ac.uk/depts/immunology/~dash/troph/troph.htm

Unmodified artery Trophoblast-modified artery

Creation of a high-flow, low-resistance zone

Replacement of smooth muscle and endothelium

with trophoblasts

Invading trophoblasts

Loss of spirality

Endothelium

Smooth muscle (tunica media)

35

Maternal Risk Factors

Duckitt K & Harrington D, BMJ 2005

• Antiphospholipid syndrome, previous pre-eclampsia, pre-existing diabetes, twins vs. singleton, nulliparity, family history, raised BMI, age > 40 yrs, raised DBP

36

Delivery

n = 2385

Study Design

2407 pregnant women

Blood/urine/risk factors

Booking (Week 12-14)

Delivery

Outcome PE - yes/no

"Risk-factor" subgroup

n = 132

Weeks16 and 28


Christian Delles


37

Clinical Characteristics

Carty DM et al., Hypertension2011

Preeclampsian = 45

Controlsn = 86 P-valueMaternal characteristics

Age (years) 29±5 28±5 0.69Ethnicity

CaucasianOther

42 (93%)3 (7%)

81 (94%)5 (6%)

1.00

Body mass index (kg/m2) 29±5 29±5 0.76Nulliparous 39 (87%) 55 (64%) 0.01Smoker 4 (9%) 21 (24%) 0.04Previous pre-eclampsia 1 (2%) 6 (7%) 0.42Family history of pre-eclampsia (mother or sister) 8 (17%) 7 (8%) 0.14At bookingSystolic blood pressure (mmHg) 121±10 114±12 0.002Diastolic blood pressure (mmHg) 76±9 69±10 0.001Proteinuria (³ + on dipstick) 1 (2%) 2 (2%) 1.00Gestation at sampling (wks) 13.8±1.6 13.7±1.3 0.61Pregnancy OutcomeHighest systolic blood pressure (mmHg) 165±11 125±12 <0.001Highest diastolic blood pressure (mmHg) 101±8 71±9 <0.001Caesarean section 20 (44%) 10 (12%) <0.001Gestation at delivery (wks) 38.3±2 40±1.4 <0.001Birthweight (g) 3155±621 3558±416 <0.001Small for gestational age(<10th customised birthweight centile)

12 (27%) 3 (3%) <0.001

38

Pregnancy Specific Proteomic Biomarkers


39

Week 28: Pre-eclampsia Specific Biomarkers



Christian Delles


40

Week 28: Pre-eclampsia Specific Biomarkers (2)

Controlsn = 17

Casesn = 18

P<0.001

-2.5

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0C

lass

ifica

tion

fact

or


41

-2.5

-2.0

-1.5

-1.0

-0.50.0

0.5

1.0

1.5

2.0

Cla

ssifi

catio

n fa

ctor

Week 12-16 Week 28

P = 0.031

Controls

-2.5

-2.0

-1.5

-1.0-0.5

0.0

0.5

1.0

1.5

2.0

Cla

ssifi

catio

n fa

ctor

P < 0.001

Week 12-16 Week 28

Cases

Change in Classification Factor


42

• Collagen alpha-1 (I) chain

• Collagen alpha-1 (III) chain



• Collagen alpha-3 (IX) chain

• Fibrinogen alpha chain

• Ig kappa chain V-IV region B17

• Retinol-binding protein 4

• Uromodulin

Sequencing



Christian Delles


43Blumenstein M et al., Proteomics2009

Plasma Proteome: DIGE and LC-MS/MS

Samples from20 weeks gestation

44

Pre-eclampsia and Cardiovascular Risk

Carty DM et al., J Hypertens 2010

45

Serum Proteomics: 2D-LC-MS/MS


Christian Delles


46 Rasanen J et al., J Proteome Res 2010

Serum Proteomics: 2D-LC-MS/MS (2)

47 Buhimschi IA et al., AJOG 2008

Urinary Proteomics: SELDI

48Buhimschi IA et al., AJOG 2008

Urinary Proteomics: SELDI (2)

Samples at time of severe diseaseSome longitudinal samples


Christian Delles


49



Pre-eclampsia

Organ damage

Biomarker of risk

50

Stroke

Dawson J et al., PloS One 2012

51

Stroke (2)

Dawson J et al., PloS One 2012

Diagnostic accuracy Stroke severity


Christian Delles


52

Heart Failure

Kuznetsova T et al. Eur Heart J 2012

53



Pre-eclampsia

Organ damage

Biomarker of risk

54

Chronic Kidney Disease PatternControlsCKD

CE migration time [min] CE migration time [min]

Training set

Cases Controls

n = 23030 ANCA,30 MGN,22 MCD,44 IgAN,25 FSGS,

58 DN,

21 SLE

n = 379379 C

CKD pattern (n = 273 biomarkers):

Fragments of• Various collagens

• Plasma proteins (serum albumin, transthyretin, alpha-1-antitrypsin, alpha-1B-glycoprotein, alpha-2-HS-glycoprotein, antithrombin-III, apolipoprotein A-I, beta-2-microglobulin, fibrinogen alpha)

• Clusterin

• Uromodulin

• Na/K-transporting ATPase gamma chain

• Psoriasis susceptibility 1 candidate gene 2

• Prostaglandin-H2 D-isomerase

• Proprotein convertase subtilisin/kexin type 1 inhibitor

• Polymeric-immunoglobulin receptor

• Osteopontin

• Neurosecretory protein VGF

• Membrane associated progesteronereceptor component 1

• CD99 antigen

• Ig lambda chain C regions

Good DM et al., Mol Cell Protomics 2010, Jantos-Siwy J et al., J Proteome Res 2009


Christian Delles


55

Diabetic Nephropathy

Normal Diabetic

Increase in ECM and collagen in diabetes and diabetic nephropathy

Normal Thick

56

Diabetic Nephropathy (2)

early detection?

57

Study Flow ChartProteomic prediction and renin angiotensin aldosterone system inhibition prevention

of early diabetic nephropathy in type 2 diabetic patients with normoalbuminuria

TEST

Spironolactone Placebo

Type 2 diabetesNormoalbuminuria

N = 3280

Positive (at risk)n = 656 (approx)

Negative (low risk)n = 2624 (approx)

Standard

Screened populationN = 7000

End of trial

156 weeksRandomization


Christian Delles


58

Integrating "omics" data

59

Systems Biology and "Omics"

Metabolomics

Proteomics

Transcriptomics

Genomics

DNA

mRNA

Protein

Metabolitessmall molecules

miRNAs

Mat

urity

of r

esea

rch

60

Summary and Conclusions

• Proteomics has a potential to elucidate pathophysiological mechanisms and as a clinical biomarker of disease and risk

• A range of different proteomic techniques is available; Results from experiments using different platforms do not necessarily overlap

• Rigorous standards for phenotypic characterisation and replication of results have to be applied to clinical proteomic studies

• Integration of data across "omics" modalitiesremains challenging


Christian Delles


61

BHF Glasgow Cardiovascular Research Centre

62

proteomics in cardiovascular disease: clinical ...genomics mirnas maturity of research 3...

Documents