Finding Biomarkers for Transplantation

1
Finding Biomarkers for Transplantation

• Raymond Ng
• (Computer Science iCapture, UBC
• rng_at_cs.ubc.ca)

2
Overview of Application

• Focus on the Genome Canada project entitled
  Better Biomarkers Of Acute and Chronic Allograft
  Rejection (www.allomark.ubc.ca)
• Led by Drs. Paul Keown, Bruce McManus and Rob
  McMaster
• 3-year project starting January 2005
• 9.1 million over 3 years including contributions
  from Novartis and IBM

3
Definitions

• Acute Rejection
• Injury that typically occurs within weeks or a
  few months after a solid organ is transplanted
• Chronic Rejection
• Injury that occurs over time to a transplanted
  organ
• This injury occurs mostly in the blood vessels of
  the organ
• Accommodation
• Absence of either form of rejection
• (current means of detecting rejection can be very
  invasive, e.g., frequent biopsies)

4
The Overall Goal

• To establish effective, minimally-invasive and
  affordable markers that reliably predict
  rejection of heart, liver, and kidney allografts

5
How do we accomplish this goal?

• Determine patterns of gene expression in white
  blood cells that react specifically to the
  transplanted organ
• Identify protein biomarkers in the plasma
• Put the identified gene and protein markers
  together, and use new mathematical tools to
  determine the best predictors of and diagnostics
  for rejection

6
Milestones

• Year 1
• To find possible biomarkers in blood that predict
  rejection
• Year 2
• To evaluate how well the biomarkers found in Year
  1 predict rejection in a separate set of patients
  
• Years 3-5
• To use the biomarkers in clinical trials to
  further test their ability to predict rejection
• To also use these biomarkers to personalize
  existing immunosuppressive treatment
• To identify novel targets for new drug development

7
Discovery Strategy
DE NOVO Patients lt1 year post-transplant
CURRENT Patients 1-5 years post-transplant
Acute Rejection
Chronic Rejection
Accommodation
Blood, Urine, Tissue Immunology Laboratory
BioLibrary Dr. Paul Keown
Anonymized Data Biomarker Database Dr. Raymond Ng
RNA Extraction from Blood, Alloreactive T Cells
and Biopsy Tissue Jack Bell Research Centre Dr.
Alice Mui
Plasma Depletion Jack Bell Research Centre Dr.
Robert McMaster
Pax-gene Blood
ITRAQ Analysis UVic Genome BC Proteomics
Platform Victoria, BC Dr. Christoph Borcher
RNA Amplification and Affymetrix GeneChip
Analysis Microarray Core Laboratory, Childrens
Hospital, LA Dr. Tim Triche
8
Proteomics Team
Rob McMaster Lead
Ross MacGillivray Co-Lead
Janet Wilson-McManus Martha Casey-Knight
Jack Bell Axel Bergman
UVic Christoph Borcher Derek Smith
Novartis Andreas Scherer, Georges Imbert, Nelson
Guerreiro, Stephan Gatzek
9
Plasma Biomarker Discovery
Plasma protein level variation Plasma protein
levels may vary from patient to patient when
compared to pooled normal plasma Highlights of
experimental design Analysis of the same
patients plasma over multiple time points
will reduce variation Use of a large number of
patients gt100
10
(No Transcript)
11
An Example Protein Biomarker
12
The Major Plasma Proteins
Leigh Anderson
Leigh Anderson
Dynamic range of plasma proteins 1 pg/ml to 50
mg/ml (1010 range)
13
Biomarkers in TransplantationDiscovery Strategy
Proteomics Analysis
Plasma Jack Bell Research Centre, Dr. Robert
McMaster
BioLibrary
ITRAQ Analysis UVic Genome BC Proteomics
Platform Victoria, BC
14
iTRAQ Labeling (Applied Biosystems)
12
12
14
16
14
114
13
13
14
16
14
115
13
13
15
17
14
116
13
13
15
18
15
117
15
Plasma Biomarker Discovery iTRAQ Experimental
Design
Each depleted plasma sample is digested with
trypsin
All 4 samples are pooled
Quantitative 2DLC / MS/MS analysis
Protein identification and differential
expression analysis
16
Plasma Biomarker Identification by iTRAQ
Technology
Heart acute rejection plasma normal plasma
number of proteins
increase
decrease
iTRAQ ratio (117114)
17
Informatics Team
Raymond Ng Robert Balshaw Co-Leads
Janet Wilson-McManus Martha Casey-Knight
Data Management
Data Analysis
iCAPTURE Mark Wilkinson Nina Opushneva, Wendy
Alexander, Joe Comeau, Andrew Ferris
iCAPTURE Bruce McManus, Mark Wilkinson,
Zsuzsanna Hollander, Andrew Ferris
Trainees Benjamin Good, Gabriella Cohen
Freue, Jon Carthy
Novartis Andreas Scherer, Mischa Reinhardt
Novartis Andreas Scherer, Peter Grass
Epicenter Tim Triche, Jonathan Buckley
UBC Wyeth Wasserman
IBM Paul Moody, Tony Li Mahendran Maliapen, Agata
Szewczyk
IBM Jeff Betts, O.K. Baek, Kareem Saad, Usha
Reddy, Prasanna Athma
18
50,000 Genes, ESTs, Proteins
Preliminary gene / protein selection
10,000 Genes, Proteins
Secondary gene / protein selection
Biological Knowledge
Analytical Strategy
Clinical Data
60-120 Genes / Proteins
Statistical model building
Biological Knowledge
Clinical Data
lt10 Predictive Markers
19
Filtering Methods
20
Questions Methods
21
Data Validation Checks

• Out of range values
• Data inconsistencies
• performed within and across visits for an
  individual patient, as well as across all
  patients in the study
• Warnings (flags) generated when manual
  verification is required
• More to discuss later in the Frontiers session
  about QC issues

22
Concluding Remarks

• Consistency of our sample handling is vital for
  this project
• Depletion of plasma is key to our workflow. The
  more depleted the plasma is, the more sensitive
  our method becomes in order to identify a
  potential Biomarker in pg/mL range
• A time-course analysis adequately supported by
  the four channels of iTRAQ

23
Plasma Biomarker Discovery Using iTRAQ Technology
MS/MS spectra
114
115
Peptide quantitation reporter tags
117
Protein identification Plasminogen
peptide FVTWIEGVMR
116

24
Plasma Protein Identification by iTRAQ Dynamic
Range
pg per mL
mg per mL
Thats 100pg/mL Range with MALDI
Apo E
Specific to ESI
Specific to MALDI
Common to both methods
Leak from column