Transcript General

Non-Invasive Rejection Diagnosis Using
Urine NMR Spectra
David Rush
Winnipeg Transplant Group
University of Manitoba
Immune Monitoring for Rejection of
Kidney Transplants
“…the clinical manifestations of acute rejection have changed with
present-day immmunosuppression. There are usually no local
symptoms, and the abnormalities are typically limited to
insidious, low-level dysfunction of the graft...”
“… systematic and repeated urinalyses performed in the absence of
substantial changes in graft function may provide a unique
opportunity to detect subclinical episodes of rejection that may
culminate in chronic rejection…”
Soulillou (NEJM (2001) 344:1006)
Editorial comment to Li et al (NEJM (2001) 344:945)
Surveillance for Acute Rejection
Standard of Practice: Serum Creatinine
Treatment
Diagnostic Threshold
Cr
Baseline Function
Strengths
Samples the Entire Graft
Rapid Turnaround Time
Non-invasive
Inexpensive
Widely Available
Inflammation
Weaknesses
Lacks Specificity
(Need a Biopsy to Diagnose Rejection)
Lacks Sensitivity
Immune Surveillance
Goal is to Develop a Biomarker in the Blood or Urine
Anti-HLA
Antibody
Capillary
Blood
GRAFT
CTL
Renal Tubule
Urine
Immune Surveillance
Probe for the Inflammatory Programs of Acute Rejection
Fas
CTL
Allorecognition
• Direct
• Indirect
Th
IL-15
APC
IL-2
IL-2
Th
Th
Th
Granzyme B
Perforin
Th
M
IFNg
Costimuli
TNFa
• B7:CD28
• CD40-CD40L
IL-4
IL-10
B
anti-HLA Ab
Immune Surveillance
Blood and Urine Biomarkers
Blood:
» PBMC RT-PCR CTL gene transcripts ( Fas, Granzyme, Perforin )
Vasconcellos et al (Transplantation 1998;66:562)
Urine:
» Flow cytometry to detect CD3 and HLA-DR on urine cells
Roberti et al (Transplantation 1995;59:495)
» RT-PCR CTL gene transcripts ( Granzyme, Perforin )
Li et al (NEJM 2001;344:945)
Immune Surveillance
Blood or Urine Biomarker Development

Limitations to the development of biomarker
» “Tarnished” Gold Standard (i.e. classification error of the biopsy)
» Lack of Specificity of any single biomarker


Biomarkers should distinguish Acute Rejection vs. Drug toxicity, Infection, ATN
Specificity could be improved by developing a:
» Donor antigen specific assay

Requires donor antigen source (e.g. donor spleen cells)
» Profile based on all components ( known / unknown ) in a blood or urine sample
 Requires strategies able to “profile” all components in a sample
Immune Surveillance
Donor Antigen Specific Biomarkers
Require Donor Cells for Analysis
Allorecognition
• Direct
• Indirect
» Flow Cross-match (anti-HLA Ab)
IL-2
O’Malley et al (ITS 1998 Abstr #1370)
Th
IL-15
APC
» ELISPOT Cytokine Assay
Heeger et al (J Immunol (1999) 163:2267)
Costimuli
• B7:CD28
• CD40-CD40L
» DTH Assay (“Tolerance Assay”)
VanBuskirk et al (J Clin Invest (2000) 106:145)
Immune Surveillance
Strategies to Profile all Components in the Blood or Urine
DNA
mRNA
Protein
Genome
Transcriptome
Proteome
> 100,000
mRNAs
> 1,000,000
Proteins
acgtacca
aggtaacg
cggtttttcgt
gtatctccctt
30,000 – 50,000
Genes
Immune Surveillance
Can Early Allograft Inflammation be
Detected by a Distinct Urine MR Spectral Profile?

Study Design:
» Gold Standard: Protocol Biopsy (months 1, 2, 3 and 6)
» Urine: Collected at time of Protocol Biopsy and stored at -80°C

Study Population:
» “Normal” Urine Spectra:
– Transplant Patients with Normal Histology by Protocol Biopsy
» “Rejection” Urine Spectra:
– Transplant Patients with Acute Rejection by Protocol Biopsy
Developing an MR Biomarker Makes No Assumption as
to What Target is Important
“Normal”
Spectra
“Rejection”
Spectra
Classifier
“Rejection”
INFORMATICS
Rate-Limiting Step is Analysis of the Spectral Profile
 1H
MR spectra
» 0.5-4.5 and 6.5-9.5 ppm
» 1690 data points / spectra

Multivariate classification strategy:
» Optimal region selector (data reduction)
» Bootstrap cross-validation
» Linear Discriminant Analysis (LDA) classifier
1H
MR Biomarkers Developed from the Urine Spectra
Correctly Identify Allograft Histology
Normal vs Rejection Histology
1st Generation
2nd Generation
3rd Generation
(33 vs 35)
(70 vs 41)
(81 vs 46)
6
6+5
6+6
Sensitivity
88%
98%
91%
Specificity
93%
96%
95%
PPV
93%
98%
95%
NPV
96%
96%
91%
Crispness
75%
96%
94%
Spectral Regions
A Biomarker for Rejection Must Be Specific
1200
Creatinine (mmol/L)
Weeks Post-Transplant
1000
800
600
400
Simulect ™
Neoral ™
MMF ™
Prednisone
200
0
0
1
2
Biomarker
N
N
Biopsy
i0t0
(ATN)
3
4
5
6
N
N
N
i1t0
7
8
9
N
i0t0
10
11
12
N
N
i1t0
The Biomarker for Rejection May Precede
the Histologic Diagnosis of Rejection
1200
Creatinine (mmol/L)
Weeks Post-Transplant
1000
800
Simulect ™
Neoral ™
MMF ™
Prednisone
600
400
Steroids
200
0
0
Biomarker
Biopsy
1
2
3
4
5
Rj Rj Rj N
i2t3
(SC)
6
7
8
N
N
i0t0
9
10
11
12
N
i0t0
The Biomarker for Rejection can Persist After
Allograft Function Returns to Baseline
500
Creatinine (mmol/L)
Weeks Post-Transplant
400
300
200
Neoral ™
MMF ™
Prednisone
100
Steroids
0
0
Biomarker
Biopsy
1
12
18 19 20
Rj Rj Rj Rj Rj Rj Rj Rj Rj Rj Rj
N N N
2
3
4
i2t2
(SC)
5
6
7
8
9
10
11
i3t2
(CL)
Urine 1H MR Biomarker
Precedes and Persists after the Diagnosis of Rejection

46 patients had 154 protocol biopsies
» 31/154 biopsies had diagnosis of Acute Rejection
» 24/31 had a urine sample prior to the biopsy
– 18/24 the urine MR classifier for rejection was present 1-2 weeks prior
to the biopsy.
» 15/24 had urine samples collected after the biopsy
– 9/15 the urine MR classifier for rejection disappeared within 4 weeks
and was confirmed by repeat protocol biopsy.
– 4/15 the urine MR classifier for rejection persisted at for 4 weeks and a a
repeat protocol biopsy confirmed the persistence of rejection.
– 2/15 have rejection classifier at last follow up (not biopsied)
Case Presentation from Yesterday
500
Weeks Post-Transplant
Creatinine (mmol/L)
400
300
200
Simulect
Neoral ™
MMF ™
Prednisone
100
Steroids
Steroids
0
0
1
2
3
4
5
6
7
8
9
10
11
12
Biomarker
N
Rj
Rj
Rj
Biopsy
i1t0
i1t1
i2t2
(SC)
i2t2
(SC)
14
Conclusions

Subclinical renal allograft rejection appears to have a distinct urine
1H MR spectrum

Resolution of subclinical rejection may correlate with the
disappearance of the spectrum and vice versa

Repeated, frequent urine spectral analysis may establish whether
there is a link between subclinical acute rejection and the
development of chronic rejection

Monitoring of urine 1H MR spectra may assist in drug withdrawal
and tolerance protocols
Collaborators
UNIVERSITY OF MANITOBA
Peter Nickerson
John Jeffery
Sylvia Dancea
NRC INSTITUTE FOR BIODIAGNOSTICS
Roxanne Deslauriers
Raymond Somorjai
Miriam Glogowski
Tony Shaw