Poster presented on World Molecular Imaging Congress 2011 in San Diego reporting my work on molecular MR imaging of liver fibrosis

1. Molecular MR Imaging of Moderate Liver Fibrosis
Miloslav Polasek1, Bryan C. Daniel T. Jamu K. Fuchs2,
Galen S. Schuhle1, Alford 1,
Loving 1, Ritika Uppal1, Alexander Guimaraes1, Kenneth K. Tanabe2, Peter Caravan1
1 A. A. Martinos Center for Biomedical Imaging, Massachusetts General Hospital and Harvard Medical School, 149 Thirteenth St., Suite 2301, Charlestown MA 02129
2 Massachusetts General Hospital, Surgical Oncology, WRN 401, 55 Fruit St., Boston, MA 02114
Non-fibrotic (Ishak 0) Figure 2. Rat DEN A) Non-fibrotic: Pre B) Non-fibrotic: post Figure 3. Liver
Introduction model.
DEN treated rats showed
S S
MRI in rat model.
Inversion recovery
Liver fibrosis is a common result of most types of liver injury such as hepatitis C, alcoholic liver disease and sequence before
increased collagen on L L
nonalcoholic steatohepatitis. Mild to moderate liver fibrosis is reversible by removing the cause of the insult. (left) and after (right)
However, if not detected and treated early, liver fibrosis can irreversibly progress into cirrhosis, liver cancer, trichrome staining. The
EP-3533 injection
and/or liver failure. Currently, fibrosis is assessed by liver biopsy, which is invasive, painful, and subject to model consistently
showed higher liver
sampling error. There is an ongoing need to develop non-invasive diagnostic methods to reliably assess liver Fibrotic (Ishak 3-4) resulted in Ishak 3-4 C) Fibrotic: Pre D) Fibrotic: post
signal enhancement
fibrosis at its early stages. Regardless of its cause, fibrosis is characterized by excess deposition of type I fibrosis. There was a 2.8-
S S in fibrotic rats. S =
collagen in the parenchyma. We hypothesized that an imaging agent providing a non-invasive measure of fold higher
Stomach, L = Liver
collagen would have broad applications in the assessment of fibrosis. EP-3533 is a gadolinium-based hydroxyproline L L
magnetic resonance (MR) contrast agent targeted to type I collagen [1,2] that was previously used to assess concentration in the
cardiac fibrosis [3]. Here we demonstrate that MR imaging with EP-3533 can detect moderate liver fibrosis in fibrotic rats (p<0.01).
two rodent models.
1.2 1.2
Methods 1
EP-3533
Control 1
Gd-DTPA
Signal enhancement
Signal enhancement
Animal models: Wistar rats were given 100 mg diethylnitrosamine (DEN) per kg weekly via i. p. injection for Fibrotic Fibrotic
0.8 0.8
normalized
normalized
four weeks. Strain A/J mice were administered carbon tetrachloride, CCl4 (0.1 mL of 40% solution in olive oil) Control
by oral gavage three-times a week for 20 weeks. Control groups received vehicle only. 0.6 0.6
Imaging probes: Compound EP-3533 was synthesized according to a published procedure [1]. Gd-DTPA
was used as a negative control. 0.4 0.4
MR imaging: MR Imaging was performed at 4.7 T. The imaging protocol consisted of 1) baseline T1-
0.2 0.2
weighted images (2D FLASH) and T1-maps; 2) injection of 100 µmol/kg Gd-DTPA and acquisition of
alternating T1-weighted images and T1-maps for 60 minutes; 3) a 24-48 hr washout period; 4) baseline 0 0
scanning followed by EP-3533 injection and acquisition of alternating T1-weighted images and T1-maps for 0 20 40 60 0 20 40 60
60 minutes. The animals were sacrificed 100 min after EP-3533 injection and tissue samples were collected. Time (min) Time (min)
Histology: Samples of liver were stained for collagen by Trichrome (rats) or Sirius red (mice). Slides were
Figure 4. Collagen probe EP-3533, but not Gd-DTPA, shows higher signal and slower
reviewed by a board certified pathologist blinded to the study and graded using the Ishak scoring system [4].
Image analysis: Images were analyzed using ImageJ and OsiriX (including Fit Toolbox plug-in). T1 values
washout in fibrotic liver compared to control.
were obtained by monoexponential fit of intensities using Microsoft Excel and Solver.
Gadolinium analysis: Gadolinium contents were determined by ICP-MS measurement on tissue samples Mouse model Rat model
that were dissolved in concentrated nitric acid. 60 60
Figure 5. Half-life of the
p = 0.04
Hydroxyproline analysis: Hydroxyproline in tissue was quantified by HPLC analysis after a two-step p = 0.01 collagen probe in liver is
55
derivatization process of samples that were hydrolyzed in 6 M hydrochloric acid [5]. 55 p = 0.01 a biomarker of fibrosis.
50 Wash-out rates (half-lives) of
Half-life (min.)
Half-life (min.)
45 EP-3533 in liver are indicative
Results and Discussion 50
40 of the stage of liver fibrosis in
both animal models. The half-
Histological analysis revealed excessive deposition of collagen in the extracellular matrix in all fibrotic liver 45 35
specimens (Figures 1 and 2). Quantitative analysis of hydroxyproline (surrogate for collagen) confirmed that lives of Gd-DTPA did not show
30 statistically significant
the collagen content significantly increases with increasing stage of fibrosis (Figures 1 and 2). MR imaging 40
with collagen-binding agent EP-3533 could distinguish between fibrotic and healthy liver. EP-3533 was 25 differences between the
retained more in fibrotic livers compared to controls (Figure 4) leading to higher signal enhancement at later groups (data not shown).
35 20
time points (Figure 3 and 4). The wash-out rate (half-life) of EP-3533 from liver could distinguish moderate Ishak 0 Ishak 3-4 Ishak 5-6 Ishak 0
Control Ishak 3-4
Fibrotic
and advanced fibrosis from normal liver and was identified as a potential biomarker of liver fibrosis (Figure
5). MR imaging without contrast agent or with the commercial non-specific agent Gd-DTPA failed to
distinguish between the fibrotic and control groups (Figure 4). We found good correlation between the 3
concentration of gadolinium (EP-3533) and hydroxyproline (collagen) in liver (Figure 6). This indicates that
Conclusions 2.5
Gd (liver/blood ratio)
the prolonged wash-out kinetics of EP-3533 from fibrotic liver are due to binding to excessive collagen.
• The collagen-targeted contrast agent EP-3533 can distinguish 2
moderate and advanced fibrosis from normal liver in two distinct animal 1.5 R = 0.77
models.
Reversible • EP-3533 showed slower elimination from fibrotic liver compared to 1
Control
control.
0.5
• Positive correlation between gadolinium and hydroxyproline in liver
Fibrosis Progression Severe fibrosis/ supports proposed collagen-binding mechanism of action of EP-3533. 0
Fibrotic
Healthy liver cirrhosis • Our results demonstrate the possibility to use MR imaging for detection 0 200 400 600 800
of fibrosis in liver. Hydroxyproline (µg/g of tissue)
Figure 6. Correlation between
700 700 700 liver Gd and hydroxyproline
Hydroxyproline (µg/g)
Hydroxyproline (µg/g)
Hydroxyproline (µg/g)
600 600 600
500 500 500 supports collagen-binding
400
300
400
300
400
300
[1] Caravan P. et al.; Angew. Chem. Int. Ed., 2007, 46, 8171–8173. mechanism of action of
200
Healthy liver 200
Moderate fibrosis 200
Severe fibrosis [2] Caravan P. et al.; Chem. Commun., 2009, 430–432. EP-3533.
100
0 (Ishak 0)
100
0 (Ishak 3 – 4)
100
0 (Ishak 5 – 6) [3] Helm P. A. et al.; Radiology, 2008, 788–796.
[4] Ishak K, et al. J Hepatol 1995, 22, 696-699.
Figure 1. The amount of collagen in liver tissue increases with the stage of fibrosis. Staining [5] Hutson P. R. et al.; J. Chromatogr. B, 2003, 427–430.
of mouse liver with Sirius red shows an increasing amount of collagen in the liver parenchyma (collagen in red)
with increasing stage of fibrosis. This is demonstrated quantitatively by the concentration of hydroxyproline in
tissue which is significantly higher (p<0.01) in the moderate and severe fibrosis groups. Partial support from NIH (NIBIB) EB009062 is gratefully acknowledged.