MultiModality NonInvasive Evaluation of PostMI LV Remodeling

1
Multi-Modality Non-Invasive Evaluation of
Post-MILV Remodeling

• Albert J. Sinusas, M.D.
• Professor of Medicine Diagnostic Radiology
• Yale University School of Medicine
• Heart Modeling Workshop
• IPAM - 2006

2
Multi-Modality Imaging

• Integration of molecular, physiological,
  mechanical information derived from multiple
  imaging sources
• Post-MI LV remodeling as example
• Interplay of regional mechanics and biological
  events in the setting of LV remodeling
• Role of targeted imaging of matrix
  metalloproteinases (MMPs)
• Role of targeted imaging of integrins (avß3)
• Translation from cellular events, to integrated
  processes in small animals, large animals and
  humans

3
Post MI Remodeling

• LV Remodeling - Changes in the structure,
  geometry and eventually function of the LV that
  occur following MI
• LV remodeling is an important independent risk
  factor following MI

4
Remodeling Following MI
Initial Infarct
Infarct Expansion (hours to days)
Global Remodeling (days to months)
Modified from Jessup and Brozena, NEJM 2003
3482007-18
5
Molecular Events Post-MI
Jugdutt BI. Circulation 1081395-1403 2003
6
Post MI RemodelingMajor Cellular Components

• Extracellular Matrix (ECM)
• Provides support structure for mechanical
  function
• Depot for signaling molecules (cytokines and
  growth factors)
• Integrins
• Modulate communication between ECM and
  intracellular cascade
• Matrix metalloproteinases (MMPs)
• Regulate matrix turnover

7
Matrix Metalloproteinases

• Family of Zn2 containing enzymes
• Secreted as proenzymes
• Activated by other MMPs, proteases and
  mechanotransduction
• Proteolytic activity regulated by Tissue
  Inhibitors of Matrix Metalloproteinases (TIMPs)
• Radiolabeling of an MMP inhibitor serves as a
  potential approach for evaluation of MMP
  activation

8
Representative MMP Classes Human Myocardium
Substrate/Function
Name
Number
Collagenases
Collagenase
3
Collagens I, II, III
Gelatinases
Gelatins,
collagens I, IV, V, VII and
basement membrane components
Stromelysins
Fibronectin,
laminin
collagens III, IV, IX,
,
and MMP activation
Membrane
-
type
MMPs
-1
MMP
-
14
Fibronectin,
laminin
Collagens I, II, III,
activates
proMMP
-
2 and
proMMP
-
13

9
(No Transcript)
10
Myocardial Fibrosis
Massons Trichrome Staining
INFARCT
NORMAL

• Matrix metalloproteinases (MMPs) degrade the
  extracellular matrix and modulate collagen
  synthesis
• Increased MMPs associated with increased
  fibrosis

11
Myocardial Collagen Matrix
LV Geometry and Coordinates Contraction
Rossi et al, Circulation, 1998
12
Courtesy of Francis G. Spinale, M.D., Ph.D.
13
MMP Levels Post-MI Remodeling Clinical
Evidence
14
LV Remodeling Independent Determinant of
Post-MI Survival
1.0
2
EDV lt 101 ml/m
N 97
0.9
Survivorship
0.8
2
EDV gt 101 ml/m
0.7
N 37
0.6
0
20
40
60
80
100
Months after infarction
Adapted from Gaurdon et al., JACC 38(1)33-40,
2001
15
Substudy - Randomized Aldactone Evaluation Study
(RALES)

• 261 pts CHF (NYHA class III-IV) and LVEF lt35
• Randomized to placebo or Spironolactone

Placebo Group
Zannad F et al. Circulation 1022700 2000
16
Substudy - Randomized Aldactone Evaluation Study
(RALES)
Zannad F et al. Circulation 1022700 2000
17
Collagen Metabolism Post-MI
Uusimaa P et al. Circulation 962565 1997
18
PIIINP Levels Post-MI and LVEF
Controls (n 16) and Pts s/p AMI (n 47)
Poulsen SH et al. Circulation 1011527 2000
19
MMP-9 LV Remodeling

• Patients s/p acute MI (n32)
• MI size CPK, or troponin I gt 2.5x normal
• Exclusion Criteria CABG lt 6 months, hx MI,
  cardiomyopathy, arrhythmia, valvular disease,
  systemic inflammatory diseases or neoplastic
  process
• RX PCI 66 Thrombolytics 25 Med 9
• Control Subjects (n53)
• Protocol

Day 1 measurements performed within 72 hrs of MI
Bonnema D, et al. Circulation 2005 (abstract)
20
LV Remodeling Post-MI
Bonnema D, et al. Circulation 2005 (abstract)
21
Temporal Changes MMP-9Post-MI
Bonnema D, et al. Circulation 2005 (abstract)
22
LV Remodeling Changes in MMP-9
35
35


plt0.05
30
30
25
25

20
20
? EDV at day 90
15
15
10
10
5
5
0
0

• lt 35
• MMP-9
• 6hrs to 5 days

• gt 35
• MMP-9
• 6hrs to 5 days

Bonnema D, et al. Circulation 2005 (abstract)
23
MMP Levels Post-MI Remodeling Experimental
Evidence
24
Wall Stress and Matrix RemodelingRat Infarct
Model



80
Infarct
70
60
50
Wall Stress (Kdyn/cm2)
40
Border
30
Remote
20
10
Sham
0
Day 1
Day 7
Day 14
Day 21
Rohde LE et al., JACC 33835 1999
25
Wall Stress and Matrix RemodelingRat Infarct
Model
70
60


50
40
Wall Stress (Kdyn/cm2)
30
20
10
(n33)
(n15)
(n29)
(n19)
0
MMP-9
Macrophages
lt10 cells/hpf
gt10 cells/hpf
p lt0.01 vs lt10 cells/hpf
Rohde LE et al., JACC 33835 1999
26
Occlusion (1hr) / ReperfusionDual
Immunohistochemistry
Canine Model 5 hrs post-reperfusion
MMP-9
Neutrophils
Lindsey M et al. Circulation 1032181 2001
27
Pacing-Induced Stretch Regional MMP9 Activity

• Canine model of LV epicardial pacing
• Pre-systolic shortening and late systolic stretch

Garcia RA et al. AJP 288H1080, 2005
28
MMP9 ActivationGelatin Zymography
REMOTE REGION
PACED REGION
200

180
160

p lt 0.001 vs LVR
140
120

p lt 0.01 vs LVR 92kDa
Band Intensity (arbitrary units)
100
80

60

40
20
0
LVR 92kDa
LVR 86kDa
LVP 92kDa
LVP 86kDa
Garcia RA et al. AJP 288H1080, 2005
29
Pacing-Induced Stretch
Collagen Degradation
Myeloperoxidase Activity
Garcia RA et al. AJP 288H1080, 2005
30
Region- and Type-Specific Induction of MMPs
PostMI
Porcine Model 8 wks post-MI
Wilson EM et al., Circulation 1072857 2003
31
Post-MI Relationship MT1-MMP and TIMP-4 levels
? ES chord lengths
Porcine Model 8 wks post-MI
TIMP-4
MT1-MMP
Wilson EM et al., Circulation 1072857 2003
32
MMP Inhibition in Post-MI Remodeling
Porcine Model MMP-1 MMP-7 sparing
inhibitor PGE-530742 (10 mg/kg PO TID)
Yarbrough WM et al., Circulation 1081753 2003
33
MMP Inhibition Post-MI Collagen Staining and
Content
Porcine Model MMP-1 MMP-7 sparing
inhibitor PGE-530742 (10 mg/kg PO TID)
Picosirius Red
Collagen Content
Yarbrough WM et al., Circulation 1081753 2003
34
MMP Activation Post MI

• Region-specific change in MMPs and TIMPs occurred
  post-MI
• MMP activation contribute to progressive adverse
  LV remodeling post-MI
• Mechanistic relationship between LV remodeling
  and MMP activation in the myocardium post-MI
• Selective inhibition of MMPs can favorably
  influence LV remodeling post-MI

35
Hypotheses

• Enhanced MMP activation occurs post MI and is
  associated with LV remodeling
• Radiolabeled MMP inhibitor targeted at MMPs could
  provide a novel approach for evaluation of the
  changes in MMP activation and localization
  post-MI, which could guide therapy

36
Radiotracer MMP Imaging

• Macrocyclic peptidomimetic
• Based on the structure of pharmacological MMP
  inhibitors
• Broad-spectrum
• In vitro assays confirm specificity to several
  MMPs involved with post-MI remodeling
• Binds to exposed catalytic domain of active MMP
• Provides index of MMP activation in vivo
• No MMP inhibitory activity

37
111In-Labeled MMP Tracers
RP-782
RP-788 (control)
Bristol-Myers Squibb Medical Imaging
38
Murine Post-Infarct Model
CONTROL
1 Wk S/P MI
MI 6
C 7
MI 6
5 mm
5 mm
5 mm
39
Methods RP782

• 9 control mice
• 14 mice surgical infarction
• Injected at 1 wk S/P MI
• MMP-targeted radiotracer (111In-RP782)
• Negative control (111In-RP788)
• MIBI was injected to assess myocardial perfusion
• Microautoradiography
• Tissue well counting
• Immunofluorescence and in situ zymography

40
In Situ Zymography
MMP-2 / MMP-9 Immunofluorescence
NC
Green MMP-2
Control
Red MMP-9
100 µm
100 µm
NC
7 day Post-MI
Su H et al. Circulation 2005
100 µm
100 µm
41
Murine Model Post-MI Remodeling
Normal
1 wk S/P MI
RP782
RP788
RP782
RP788
a
c
e
g
Morphology
MI43
MI06
C 05
C 09
d
f
h
b
RV
LV
Autoradiography
Su H et al. Circulation 1123157-67 2005
42
111In-labeled MMP Radiotracer Myocardial Uptake
plt 0.05
vs.
RP788, plt0.05
vs.
Control
5


4
3
Myocardial Activity Ratio
Infarct/ Remote
2
1
(n 8)
(n 5)
(n 4)
(n 4)
0
RP782
RP788
MI
Control
Su H et al. Circulation 1123157-67 2005
43
Method RP805

• 4 control mice (C57BL/6)
• 15 mice with surgical infarction
• 99mTc labeled compound targeted at MMPs
• 99mTc-RP805 (1.3?0.3 mCi i.v.)
• 1 wk (n5), 2 wks (n5), and 3 wks (n5) S/P MI
• Dynamic In vivo planar pinhole imaging during 15
  min post-injection, static image at 75 min
• 201Tl (0.24 ? 0.05 mCi) injected i.v. followed by
  pinhole imaging to evaluate perfusion
• Gamma well counting of heart and selected organs

44
Hybrid MicroSPECT/CT
Su H et al. Circulation 1123157-67 2005
45
MicroSPECT / CTMouse 1 wk S/P MI
Tl-201 green RP805 (MMP) red
46
MicroSPECT / CTMouse 1 wk S/P MI
Tl-201 green 0.25mCi RP805 (MMP) red 2 mCi
0.5 mm 30 sec/step 32 steps
47
99mTc-RP805 and 201Tl Myocardial Uptake Control
Post-MI Mice
Su H et al. Circulation 1123157-67 2005
48
99mTc-RP805 and 201Tl Myocardial Uptake (ID/g)
Control Post-MI Mice
Su H et al. Circulation 1123157-67 2005
49
Canine Porcine StudiesMMP ImagingRP-805

• Canine
• 24 hrs S/P MI
• 1 wk S/P MI
• Porcine
• Control, 1, 2, 4 wks S/P MI
• In vivo imaging
• Ex vivo imaging

50
In Vivo Dual Isotope SPECTCanine Infarct 24 Hrs
after 2 Hr LAD Occlusion
Short Axis
Tl-201
RP-805
Vertical Long Axis
Horizontal Long Axis
Tl-201
RP-805
CMMPB
51
In Vivo Dual Isotope SPECTCanine Infarct 1 Week
after 2 Hr LCX Occlusion
Short Axis
Tl-201
RP-805
Vertical Long Axis
Horizontal Long Axis
Tl-201
RP-805
CMMPA
52
Change in MMP Distribution
1 Day - AMI
1 Week - IMI
53
MMP Activity Post-MI Relationship to Strain and
LV Remodeling

• In vivo imaging of myocardial MMP activation in
  relationship to abnormal strain patterns could
  provide prognostic information in patients
  following MI
• Serial SPECT/CT imaging of an MMP-targeted
  radiotracer will allow noninvasive spatiotemporal
  tracking of MMP activity in post-MI myocardium
• MMP activation will correlate with regional
  myocardial strain

54
99mTc RP805 / 201Tl SPECT/CT
Surgical MI Yorkshire pigs LCX occlusion 4
groups 1 wk post-MI (n6), 2 wk post-MI
(n5), 4 wks post-MI (n5), Control (n5)
RP805
Tl-201
MRI
Dual Isotope
SPECT/CT
SPECT/CT
//
//
90 min
120 min
55
SPECT/CT Protocol

• In vivo SPECT/CT performed with GE Hawkeye
• RP805 (23.6 3.2mCi)
• 201Tl (2.6 0.7mCi)
• SPECT/CT Reconstruction
• batch reconstructed using an iterative algorithm
• CT attenuation correction
• reoriented into short and long axes

56
MRI Strain Analysis

• Cine gradient-echo breathhold MR images acquired
  within hrs of SPECT/CT
• 16 5mm SA slices, apex to mem. septum
• Semiautomated segmentation of SA slices
• Computation of regional displacements using
  shape-based analysis
• Mechanical model and FEA
• Calculation of cardiac-specific strains (radial,
  longitudinal, circumferential)

57
Postmortem Analysis

• Hearts cut into 5mm thick short-axis slices
• Ex vivo imaging of slices
• Each slice cut into radial pies
• Gamma well-counting of each pie
• Myocardial activity expressed as
• non-ischemic
• injected dose

58
In Vivo SPECT
Control
1 Wk s/p MI
201Tl
99mTc MMP
2 Wks s/p MI
4 Wks s/p MI
201Tl
99mTc MMP
McAteer J, et al. Circulation 2005 (abstract)
59
Ex vivo Images - 1 wk post-MI
201Tl
RP805
60
Postmortem Analyses
2
3
4
1
B
5
16
15
6
14
7
I
8
13
9
12
11
10
I
I
B
250
300
250
200
200
150
RP805 (kcnts)
201Tl (kcnts)
150
100
100
50
50
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Segment
61
99mTc-RP805 and 201Tl Myocardial Uptake (ID/g)
Control Post-MI Pigs
McAteer J, et al. Circulation 2005 (abstract)
62
Cine-MRPig 4 wks S/P MI
PMMPDA
63
Surface Segmentation
Endocardial Surface
Epicardial Surface
64
3D DISPLACEMENTS From Shape Based Matching
W
t2 t1

• Confidence measure is computed based on variation
  of shape similarity metric over window W.

65
LV MATERIAL MODEL TRANSVERSELY ISOTROPIC LINEAR
ELASTIC
Material Matrix
Stiffness (Youngs modulus) fiber stiffness (Ef
) cross-fiber stiffness (Ecf ) (Ef
Ecf 3.51) Incompressibility (Poissons ratio)
? (0.4, nearly
incompressible)
66
NORMAL STRAIN PATTERN - DOG
CARDIAC-SPECIFIC STRAINS
RADIAL
LONGITUDINAL
CIRCUMFERENTIAL
FIBER-SPECIFIC STRAINS
CROSS-FIBER
FIBER
Shortening
Lengthening
Sinusas AJ, et al. Am J Physiol 281H698-714 2001
67
Myocardial ES Strain ShearPig 4 wks S/P MI
Radial
Circumferential
Longitudinal
RC Shear
CL Shear
RL Shear
68
1 wk post-MI MR Strain
LL
RR
CC
69
MMP Activation / Strain
PMMPDA
70
Radial Strain MMP Activity1 wk post-MI
Radial Strain
RP-805 Uptake
Song et al. ASNC 2005
71
MR Strain vs MMP Activity 1 wk post-MI
72
MR Strain vs MMP Activity 1 wk post-MI
Sahul Z et al. ACC 2006 (in press)
73
Summary - Pig Studies

• Increased MMP activity can be detected
  non-invasively with a MMP targeted radiotracer
• Maximum MMP activity was observed at 1 wk
  post-MI, although remained elevated at 2 and 4
  wks
• MMP activity was increased 4-5 fold in the
  central infarct region early post-MI, and
  correlated with non-invasive indices of regional
  myocardial strain

74
Conclusion Pig Studies

• Demonstrated non-invasive in vivo Hybrid SPECT/CT
  imaging of an MMP targeted radionuclide, RP-805,
  permits serial monitoring of spatiotemporal MMP
  activation in a clinically relevant model of
  post-MI remodeling
• MMP targeted imaging provides an approach for
  tracking novel therapeutic interventions directed
  at MMP inhibition and reduction of post-MI
  remodeling

75
MMP Imaging - Summary

• MMP-targeted radiotracer (RP782 RP805)
  specifically localized in regions of MMP
  activation post-MI, and may have potential
  clinical value for non-invasive imaging of MMP
  activation post-MI and for prediction of LV
  remodeling

76
Magnetic ResonanceDiffusion Tensor Imaging

• Diffusion Tensor Imaging (DTI) is an MRI modality
  sensitive to random movement of water molecules
• When this movement is hindered by membranes and
  macromolecules, diffusion becomes anisotropic
• In highly structured tissues such as muscle
  fibers, this anisotropy can characterize local
  tissue structure
• Diffusion Tensor eigenvectors have been shown to
  correlate with tissue architecture

77
Fiber Orientation
Jackowski M et al. SCMR 2006
78
Postmortem MR DTI Fiber Connectivity
Fiber connectivity established using
streamlining method
P
P
P
S
S
S
L
L
L
A
A
A
apex
apex
apex
Primary Eigenvector - Myofiber Architecture
Jackowski M et al. SCMR 2006 (in press)
79
Sheet Architecture
Orientation of the diffusion tensor eigenvectors
at different transmural locations.
Primary eigenvector
Secondary eigenvector
Tertiary eigenvector
longitudinal
Endocardial surface
circumferential
radial
Epicardial surface
Jackowski M et al. SCMR 2006
80
Sheet Orientation
L
S
longitudinal
circumferential
Projected tertiary eigenvector
Sheet angle
radial
Jackowski M et al. SCMR 2006
81
Postmortem MR DTI Sheet Connectivity
A
A
P
P
apex
apex
Tertiary Eigenvector - Sheet Architecture
Jackowski M et al. SCMR 2006
82
MicroSPECT-CT Imaging 7 days post MI
Apical
Basal
Apical
Basal
CT
201TlPerfusion
99mTcavß3 integrin
Fusion
Wild Type
MMP-9 Null
Dobrucki WL, et al. J Nucl Cardiol 12S15 2005
83
NC100692 UptakeMurine MI 1 Wk MMP-9 K.O. vs
Control
700

Control (n9)
600
500
MMP9 K.O. (n6)
400
NC100692 Uptake ( non-ischemic)
300

200
100
0
0 40
41 60
61 80
gt 80
Tl-201 perfusion ( non-ischemic)
Dobrucki WL, et al. J Nucl Cardiol 12S15 2005
84
Imaging AngiogenesisChronic Canine Model MI
SA
HLA
VLA
BASELINE
201Tl
99mTc NC100692
1 WK S/P MI
201Tl
99mTc NC100692
2 WK S/P MI
201Tl
99mTc NC100692
CHFCNC09
85
Integrated Molecular Imaging Summary

• Combined imaging of MMP activation and markers of
  apoptosis or angiogenesis with changes in
  regional and global mechanics should lead to a
  better understanding of the pathophysiological
  processes associated with post-MI remodeling
• Targeted imaging approaches may allow for risk
  stratification and evaluation of novel
  therapeutic interventions like new MMP inhibitors
  directed at preventing post-MI remodeling

86
FUTURECONSIDERATIONS

• Evaluate transgenic mice with selective MMP or
  TIMP KOs or overexpression
• Develop and evaluate more selective MMP targeted
  radiotracers
• Development Hot spot quantitative approaches
• Relate MMP activation to changes in perfusion and
  regional mechanics (strain)
• Integrate fiber and sheet architecture in the
  analysis of strain
• Use radiotracer images to impose a tissue
  constraint on mechanical model

87
Acknowledgement

• Yale University Experimental Nuclear Cardiology
• YiHwa Liu, Ph.D.
• Staff Donald Dione, Jennifer Hu, Patti
  Cavaliere, Christi Hawley, Chris Weyman, Donna
  Dione
• Post Doctoral Fellows Haili Su, Wawrzyniec
  Dobrucki, James Song, Zakir Sahul, Leszek
  Kalinowski, Matt Brennen, Suman Tandon, Choukri
  Mekkaoui
• Students Jodi-Ann McKain, Conroy Chow, Jarod
  McAteer
• VAMC Molecular Labororatory
• Mehran Sadeghi, M.D.
• Yale Pathology
• Joseph Madri, M.D., Ph.D.
• GE Healthcare
• Marivi Mendizabal, Ph.D.

• Medical University of South Carolina
• Francis Spinale, M.D., Ph.D.
• Merry Lindsey, Ph.D.
• Yale University - Radiology
• James Duncan, Ph.D.
• Lawrence Staib, Ph.D.
• Todd Constable, Ph.D.
• Xenophon Papademetris, Ph.D.
• Marcel Jackowski, Ph.D.
• Maolin Qui, Ph.D.
• Alex Pinus, Ph.D.
• Bristol-Myers Squibb
• D. Scott Edwards, Ph.D.
• Michael Azure, Ph.D.
• Padmaja Yalamanchili, Ph.D.
• Richard Liu, Ph.D.
• Edward H. Cheesman, Ph.D.
• Simon Robinson, Ph.D.