Proteomica Clinica

I.R.C.C.S. Istituto Superiore di Sanità (HUB Italiano)

Coordina le attività di 23 biobanche presenti sul territorio nazionale tra cui:

¾ Definition of the gold standard protocol for the chosen biological samples:

STANDARDIZATION OF URINE ANALYSIS

Cancer Biomarker Discovery Sample Sources – Dilution of Markers with Distance from Tumor

Neoplastic Body

Sample Type

Cytology

Tissue Fluids

Biomarker Concentration

High

Medium

Low

Seminal Plasma

Serum

Biopsy

Sample(s) suitable for Nipple Aspirates

Plasma

LCM Fine Needle proteomic analysis

Urine Aspirates

Biologial samples and Kidney biomarkers

Sample type

Biomarkers amount

Biopsy specimens

High

intermediate

Urine

Low

Serum/Plasma

Urine advantages

Less complex then serum and plasma

Un-invasive recruitment in large amount

Higher presence of kidney derived molecules

STANDARDIZATION OF URINE ANALYSIS

Fresh sample or frozen one?

Time controlled collection or not?

Urine Storage at RmT: how long can it stay?

Between‐subject differences?

Can protease

Inhibitors efficiently prevent protein degradation?

CONCLUDING REMARKS

The timing of the collection can affect the proteomic profile

Urine are stable up to 2 hours at RmT after collection

Repeated freeze and thaw cycles do not alter the protein profiles

Large inter‐individual variability

J. Chromatogr. B. Analyt. Technol. Biomed. Life Sci., 856 (2007) 205‐213

Papale M, Di Paolo S, Gesualdo L et Al¾ Definition of the clinical question and identification of the patients to be enrolled:

BIOMARKER DISCOVERY STUDIES FOR:

• DIAGNOSIS
• PROGNOSIS
• THERAPY MONITORING

Study Design (pre - clinical considerations)

Pre-Analytical Factors Analytical Factors

Pre-Analytical Factors Analytical Factors

• Lab technique.
• Clinical inclusion/exclusion criteria.
• Automation vs. manual protocols.
• Subject demographics (race, age).
• Fractionation protocols.
• Sample size.
• Randomization during processing.
• Sample type (EDTA plasma, Citrate plasma, serum, lysate).

• Array binding and washing
• Sample collection Protocol (time protocols for clotting, time 'on the bench', transport, aliquoting)
• Array reading protocols (Baseline subtraction, noise definition, normalisation)
• Data pre-processing
• Sample Storage (age, temperature, freeze-thaw cycles)
• Classification approaches
• Study Design (Clinical aspects)
• Two types of studies:
  • Prospective study: enroll patients and examine what happens to them.
    • Case-controlled (clinical trials)
    • Cohort - patients self select their group
    • Longitudinal
  • Retrospective study: choose patients from a databank and examine what happened to them.
    • Case-controlled
    • Cross-sectional
• Identification and validation of the candidate biomarkers

THE AIMS OF A CLINICAL PROTEOMIC STUDY

Biomarker Discovery and Validation

Work plan

• Discovery phase
  • Develop assay conditions
  • Find candidate markers
  • Develop

classification rules.

• Validation phase
  • Large sample set size.
  • Subset of assay conditions.
  • Refine classification rules.
• Identification phase
  • Identify a candidate biomarker and validate its association with the condition you are studying by alternative approaches

PROTEOMICA CLINICA: LA NOSTRA ESPERIENZA

• INDIVIDUAZIONE DI NUOVI BIOMARCATORI DI URINARI DI CARCINOMA RENALE (RC)
• DIAGNOSI DIFFERENZIALE DELLA NEFROPATIA DIABETICA
• FARMACO-PROTEOMICA DELLA NEFROPATIA DI BERGER (IgAN)

Cross-comparison between tissue and urine proteomics allowed to identify RCC-specific urinary biomarkers

Massimo Papale, Margherita Gigante, Clelia Prattichizzo, Maria Teresa Rocchetti, Grazia Vocino, Michele Battaglia, Loreto Gesualdo and Elena Ranieri

Renal Cell Carcinoma (RCC)

Most common solid tumor of the kidney that accounts about 3% of all adult malignancies

Histological Classification (Mainz):

RCC: anatomy and histological characteristics

RCCDiagnosis

Therapy

Complete Blood Count (CBC)

Ultrasound examination

CT with contrast

NMR Radicalnephrectomy

Skeletal scintigraphy

Urography

Arteriography

Urinalysis

Identification of RCC Biomarkers: a current challenge

AIMS OF THE STUDY

• To identify the best sample for biomarkers validation
• To validate the specificity of the identified biomarkers for RCC

Clinical features

EXPERIMENTAL DESIGN

SERUM PROFILING

URINE PROFILING

CONFIRMATION

DISCOVERY

TISSUE IMAGING

TISSUE PROFILING

LOCALIZATION

SELDI PROFILING ON URINE, TISSUES AND SERA

5000

7500

10000

12500

20 Node

Class = DN15 C08586_0 <= 19.726

Class Cases %DN 31 43.110 Other CKD 41 56.9

W = 72.000

N = 725 C08586_0 <= 19.726 C08586_0 > 19.726

Terminal Node 2

Node 1 Class = DN

Class = Other CKD C013593_ <= 3.010

Class Cases % Class Cases %30 DN 0 0.0 DN 31 60.8

Other CKD 21 100.0 Other CKD 20 39.2

W = 21.000 W = 51.000

N = 21 N = 51

20 C013593_ <= 3.010 C013593_ > 3.010

10 Node 3 Terminal

Class = Other CKD Node 4

C03086_1 <= 1.605 Class = DN

Class Cases % Class Cases %0 DN 1 6.7 DN 30 83.3

Other CKD 14 93.3 Other CKD 6 16.7

W = 15.000 W = 36.000

N = 15 N = 36

10 C03086_1 <= 1.605 C03086_1 > 1.605

Terminal Terminal

Node 2 Node 3

Class = DN Class = Other CKD

5 Class Cases % Class Cases %

DN 1 100.0 DN 0 0.0

Other CKD 0 0.0 Other CKD 14 100.0

W = 1.000 W = 14.000

N = 1 N = 14

0 5000 7500 10000 12500

SAMPLE PREPARATION AND MULTIVARIATE ANALYSIS

DATA ACQUISITION

ANALYSIS ®(BPS )

RCC Vs. Healthy Subjects (URINE)

SELDI MASS SPECTRA

5000 7500 10000 12500

20uA 15 RCC

INTENSITY 503020 HS100 MASS m/z

(Da)Fifty-seven mass peaks resulted differentially excreted (p-value < 0.05) between RCC and healthy subjects (mass range: 3000 e 30000 Da)

Classification and Regression Tree (CART) Analysis

Node 1

Class = RCC M0_MxC04128_0 <= 14.796

Class Cases %

METHODCTRL 12 46.2

RCC M0_Mx 14 53.8

W = 26.000

N = 26

SELDI dataset analysis by multivariate classification software may provide rapid, simplified pattern analysis for discovery of multiple biomarkers

C04128_0 <= 14.796 C04128_0 > 14.796

Terminal Node 2

Node 1 Class = RCC M0_Mx

Class = CTRL C03736_9 <= 4.991

Class Cases % Class Cases %

CTRL 11 91.7 CTRL 1 7.1

Basic principle of CART (Classification and Regression Trees) analysis

RCC M0_Mx 1 8.3 RCC M0_Mx 13 92.9

W = 12.000 W = 14.000

N = 12 N = 14

peak intensity values are used to define a single splitting rule that best•

SELDI C03736_9 <= 4.991 C03736_9 > 4.991

segregates the training set by phenotype. Terminal Terminal

Node 2 Node 3

tree describing the best

• The Class = CTRL
• Class Cases %
• CTRL 1 100.0
• CTRL 0 0.0
• Class = RCC M0_Mx
• Class Cases %
• RCC M0_Mx 0 0.0
• RCC M0_Mx 13 100.0
• W = 1.000
• W = 13.000
• N = 1
• N = 13

ARE THERE COMPARABLE REDUCED OR INCREASED RCC MARKERS IN KIDNEY AND URINE?

RCC Vs "normal" tissue RCC Vs HS urine

Differentially expressed mass peaks

Common markers (confident RCC biomarkers)

Differently expressed urine biomarkers in tissue profiling

3600

3700

3800

3900

8500

8750

9000

9250

8500

8750

9000

9250

3600

3700

3800

3900

500

200

400

150

300

100

200

50

100

0

500

200

400

150

300

100

200

50

100

0

500

200

300

150

200

100

500

0

500

200

300

150

200

100

100

500

0

400

200

300

150

200

100

100

500

0

400

200

300

150

200

100

100

500

0

3600

3700

3800

3900

8500

8750

9000

9250

8500

8750

9000

9250

3800 3900
3600 3800 8500 9000 8500 9000
3600 3800 12580 10060 uA 75uA 40 50
2520 0080 12510060 uA 75uA 40 50 HS20 25
Urine 0080 12510060 uA 75uA 40 50
20 2500 8012510060 uA 75uA 40 50
20 2500 8012560 100uA 75uA 40 50
20 2500 8012560 100uA uA 75uA 40 50
RCC250 080 12560 100uA uA 75uA 40 50
20 2500 8012560 100uA uA 75uA 40 50
3600 3800
3600 3800 8500 9000 8500 9000
3736 m/z 8755 m/z
ANDAMENTO DEL PICCO DI 3736 m/z nelle urine e nei tessuti RCC
URINE TESSUTO
CTRL RCC Tex CTRL Tex RCC
CART analysis
Classification Tree for RCC and HS (URINE)
Node 1
Class = RCC M0_MxC08755_0 <= 9.139
Class Cases %CTRL 12 46.2
RCC M0_Mx 14 53.8
W = 26.000
N = 26 1.0
C08755_0 <= 9.139 C08755_0 > 9.139 0.8
Node 2 Node 3
Class = CTRL Class = RCC M0_MxC03736_9 <= 25.320 C03736_9 <= 7.575 Sensitivity
Class Cases % Class Cases % 0.6
CTRL 10 76.9 CTRL 2 15.4
RCC M0_Mx 3 23.1 RCC M0_Mx 11 84.6
W = 13.000 W = 13.000 0.4
N = 13 N = 13 ROC Integral
0.764
C03736_9 <= 25.320 C03736_9 > 25.320 C03736_9 <= 7.575 C03736_9 >

7.575 0.2

Terminal Terminal Terminal Terminal

Node 1 Node 2 Node 3 Node 4

0.0

Class = CTRL Cl