Effect of contrast agent charge on visualization of articular cartilage using computed tomography: Exploiting electrostatic interactions for improved sensitivity

Article abstract of DOI:10.1021/ja9053306

(Chemical Equation Presented) The synthesis and evaluation of a new class of cationic iodinated contrast agents for the imaging of cartilage using computed tomography (CT) are described. In direct comparisons with anionic contrast agents, the cationic contrast agents afforded higher equilibrium concentrations in the articular cartilage of ex vivo rabbit femurs and thus greater imaging sensitivity. Variations in CT intensity across the sample reflected the inhomogeneous distribution of glycosaminoglycans in the tissue as confirmed by histological analysis. We anticipate that this work represents the first step in the development of sensitive, nondestructive CT-based methods to characterize the biochemical properties of cartilage using cationic contrast agents.

Full text of DOI:10.1021/ja9053306

Published on Web 08/25/2009
Effect of Contrast Agent Charge on Visualization of Articular Cartilage Using
Computed Tomography: Exploiting Electrostatic Interactions for Improved
Sensitivity
Neel S. Joshi,† Prashant N. Bansal,‡,§ Rachel C. Stewart,‡ Brian D. Snyder,§ and
Mark W. Grinstaff*,†,‡
Departments of Chemistry, Biomedical Engineering, Boston UniVersity, Boston, Massachusetts 02215, and
Department of Orthopedic Surgery, Children’s Hospital, HarVard Medical School, Boston, Massachusetts 02215
Received June 29, 2009; E-mail: mgrin@bu.edu
Scheme 1. Synthesis of the +1 (CA1+), +2 (CA2+), and +4 (CA4+
)
Noncovalent interactions play significant roles in a myriad of
biological systems, including molecular recognition in nucleic acids,
Cationic Contrast Agentsa
stabilization of protein structures, and hemodynamical properties of
glycocalyx. These various noncovalent interactions (e.g., H-bonding,
π-stacking) have been used extensively to synthesize supramolecular
structures and as components of drug design using molecular model-
ing.1 Application of these principles to the design of tissue-specific
medical imaging agents is also of significant interest. For example,
contrast agents labeled with targeting antibodies and peptides have
been explored.2 Our focus is on exploiting Coulombic interactions
between ionic contrast agents and highly charged polysaccharides,3
such as those found in cartilage. Here, we report the synthesis of new
iodinated cationic computed tomography (CT) contrast agents, their
use for visualizing the spatial distributions of glycosaminoglycans
(GAGs) in articular cartilage (AC), and preliminary studies exploring
the effects of molecular charge on imaging efficacy.
AC is the smooth, hydrated tissue that lines the ends of long bones
in load bearing joints. The heavily sulfated and carboxylated polysac-
charides comprising the GAGs represent 5-10% by wt of AC and
a Reagents and conditions: (a) SOCl2, ∆; (b) AcCl, DMA; (c) DMA, DIPEA;
(d) CH2Cl2/TFA; (e) THF, ∆. Structures of anionic contrast agents 1 and 2 are
also shown.
are key components in conferring cartilage with its resistance to com-
pressive loads. The remaining mass of cartilage is composed mostly
of collagen (10-20%) and water (68-85%).4 It has been widely
recognized that the loss of GAGs from the AC is a hallmark of osteo-
arthritis, a degenerative joint disease in which wear or trauma results
in damage to the AC surface.5 Consequently, contrast agents capable
of assessing local variations in GAG content are of significant interest
for the study of cartilage biology and the diagnosis of cartilage diseases.
Due to the need for quantitative biochemical analysis of tissue
samples and cartilage health in patients, both magnetic resonance (MR)
and CT imaging modalities have been used to image cartilage ex ViVo
and in ViVo. To assess GAG content, both imaging modalities use
anionic contrast agents.6 Although these techniques represent the
current state of the art, they rely on the limited diffusion of the anionic
contrast agents into the target tissue. The contrast agents distribute
into the cartilage in inverse proportion to GAG in AC due to the
electrostatic repulsion between the contrast agent and the negative fixed
charge density of GAGs. We hypothesized that cationic contrast agents
would be electrostatically attracted to anionic GAGs and would
consequently result in a more sensitive technique for imaging cartilage.
Furthermore, our interest lies with CT imaging because it is more
widely accessible and less expensive, can image cartilage and bone
simultaneously, enables rapid 3D reconstruction of the tissue, and is
able to achieve higher spatial resolution over shorter acquisition times
compared to MRI. To investigate the effects of contrast agent charge
on CT imaging efficiency, we synthesized three new iodinated contrast
agent molecules: one having a single positive charge and three iodine
atoms (CA1+), a second having two positive charges and three iodine
atoms (CA2+), and a third having four positive charges and six iodine
atoms (CA4+). These molecules were compared to two commercial
formulations of anionic contrast agents, Cysto Conray II (1; iothala-
mate) and Hexabrix (2; ioxaglate). These two contrast agent molecules
contain 3 and 6 iodine atoms, respectively, but both bear a single
negative charge. Cysto Conray II was compared to CA1+ and CA2+,
and Hexabrix was compared to CA4+, to minimize differences in
molecular structure and enable a direct comparison between charge
and imaging sensitivity.
The three cationic iodinated X-ray contrast agents were each
synthesized in four steps from commercially available triodinated
precursors (Scheme 1). For each contrast agent, the appropriate acyl
chloride was generated using published protocols, then reacted with a
mono-Boc protected ethylenediamine, and subsequently deprotected
with TFA to afford the desired primary amino compounds. The final
contrast agents were soluble at concentrations up to 0.2 M in aqueous
solutions and bore positive charges, by virtue of their primary amine
substituents, at pH e 7. The solubility of CA1+, but not CA2+ and
CA4+, decreases at higher pH values (7.4); therefore the following
experiments were conducted at pH 7.0 so that all the contrast agents
could be compared without any solubility problems.
† Department of Chemistry, Boston University.
‡ Department of Biomedical Engineering, Boston University.
§ Harvard Medical School.
To evaluate the ability of the contrast agents to image AC, we
imaged an intact rabbit femur using microcomputed tomography
9
13234 J. AM. CHEM. SOC. 2009, 131, 13234–13235
10.1021/ja9053306 CCC: $40.75  2009 American Chemical Society

C O M M U N I C A T I O N S
(microCT40, Scanco Biomedical, Switzerland). In these ex ViVo
experiments, [I], pH, and osmolality were kept constant across all of
the contrast agent solutions to determine the effects of molecular
structure and charge on imaging. In each experiment, the distal end of
a single rabbit femur was immersed in one contrast agent, imaged
using µCT, immersed in saline solution for 24 h to remove the contrast
agent, imaged to confirm contrast removal, then exposed to a second
contrast agent, imaged, and subjected to histological analysis. This
procedure is different from that envisioned in ViVo where the contrast
agent would be delivered via an intra-articular injection.
Images obtained from the rabbit femur studies show that the cationic
contrast agents afforded higher X-ray attenuation values and more
specific imaging for the cartilage tissue as compared to the anionic
contrast agents (Figure 1c). At the low contrast agent concentrations
used in this study (15 mg of I/mL of solution, versus 160-300 mg of
I/mL for a typical CT arthrography procedure), the anionic contrast
agents were largely excluded from the cartilage ECM, resulting in
lower attenuation for the tissue. By comparison, the cationic contrast
agents achieved higher equilibrium concentrations in the tissue,
allowing for facile differentiation between bone, cartilage, and air. The
cationic contrast agents had 1.6 (CA1+), 2.4 (CA2+), and 2.9 (CA4+)
times higher mean attenuation values for cartilage than their anionic
counterparts, indicating that increasing cationic charge yielded higher
affinities for the anionic fixed charge density of GAGs.
AC is distinctly stratified in its organization, with the superficial
layer being comprised mostly of highly oriented collagen fibrils, while
the middle and deep zones have higher GAG content. In addition to
increasing the overall attenuation for cartilage, the distribution of the
cationic contrast agents also reflects the inhomogeneous distribution
of GAGs in each sample. As we hypothesized, the equilibrium
distribution of the cationic contrast agents appears to be dominated
by electrostatic attraction, such that contrast agent concentration varies
proportionately with GAG content.7 Images obtained after immersion
in cationic contrast agents (Figure 1c) had lower attenuation values
closer to the cartilage/air interface (superficial zone) and higher
attenuation values closer to the cartilage/bone interface (middle and
deep zones). This trend is the opposite of what is seen in published
data obtained with anionic contrast agents, whose distributions show
an inverse relationship with GAG content.8 The images obtained with
1 and 2 in this study failed to show this trend, primarily due to the
lower concentration of contrast agent used. A sample histological
section obtained from one of the femurs used in this study shows the
natural distribution of GAG for the femoral groove (Figure 1b). A
reconstruction of a rabbit femur imaged with CA4+ highlights the
Figure 2. 3D reconstruction of femur after exposure to CA4+. (a) 60 slices
of distal femur, reconstructed. Zoomed-in images of the femoral groove (b)
and the medial condyle (c) show that the cartilage can easily be segmented
from the bone (dashed line) and that the distribution of the cationic contrast
agent reflects the normal distribution of GAG in AC (low GAG in the superficial
zone and high GAG in the middle and deep zones). Scale bar ) 1 mm.
ability of cationic contrast agents and CT to facilitate monitoring of
changes in cartilage attenuation, thickness, and morphology in three
dimensions as well as the trabecular architecture of the underlying
bone (Figure 2), a task that is difficult to accomplish with MRI.
Taken together, the data presented here represent a compelling case
for the continued development of cationic CT contrast agents. Our ongoing
experiments are directed toward demonstrating that these cationic contrast
agents are able to convey quantitative information about the biochemical
characteristics of AC. The suitability of these contrast agents for in ViVo
applications remains to be determined, and issues such as toxicity,
administration method, and radiation dose will be the focus of future
studies. However, their ability to characterize ex ViVo cartilage samples is
clearly evident. Currently, obtaining data about the spatial distribution of
biochemical components in tissue samples is largely accomplished using
histology, which is destructive and time-consuming, and thus the use of
these contrast agents in conjunction with CT imaging will result in a readily
available, nondestructive alternative to histology. We anticipate that the
ability to acquire quantitative information about cartilage thickness,
morphology, and localized GAG content will aid in the diagnosis and
treatment of osteoarthritis as well as the evaluation of new disease
modifying osteoarthritis drugs and tissue engineered therapies.
Acknowledgment. We thank Nipun Patel for help with image
processing and the Coulter Foundation for generous support.
Supporting Information Available: Full ref 5, detailed synthetic
procedures, preparation of contrast agents, imaging protocol, and statistical
analysis. This material is available free of charge via the Internet at http://
pubs.acs.org.
References
(1) (a) Atwood, J. L.; Davies, J. E. D.; MacNicol, D. D.; Vo¨gtle, F.; Lehn,
J. M. ComprehensiVe Supramolecular Chemistry; Pergamon: Oxford, 1996.
(b) Brunsveld, L.; Folmer, B. J. B.; Meijer, E. W. MRS Bull. 2000, 25, 49.
(c) Longmire, M.; Choyke, P. L.; Kobayashi, H. Curr. Top. Med. Chem.
2008, 8, 1180. (d) Gissot, A.; Camplo, M.; Grinstaff, M. W.; Barthe´le´my,
P. Org. Biomol. Chem. 2008, 6, 1324. (e) Gund, P.; Andose, J. D.; Rhodes,
J. B.; Smith, G. M. Science 1980, 208, 1425.
(2) (a) Boswell, C. A.; Brechbiel, M. W. Nucl. Med. Biol. 2007, 34, 757. (b)
Kersemans, V.; Kersemans, K.; Cornelissen, B. Curr. Pharm. Des. 2008,
14, 2415.
(3) (a) Yin, M.; Shen, J.; Pflugfelder, G. O.; Mullen, K. J. Am. Chem. Soc.
2008, 130, 7806. (b) Winalski, C. S.; Shortkroff, S.; Schneider, E.; Yoshioka,
H.; Mulkern, R. V.; Rosen, G. M. Osteoarthritis Cartilage 2008, 16, 815.
(c) Van Ootegham, S. P.; Smith, R. G.; Daves, G. D., Jr.; El-Mazati, A.;
Chan, E. M.; Olsen, G. D.; Riker, W. K. J. Med. Chem. 1976, 19, 1349.
(4) Mow, V. C.; Huiskes, R. Basic orthopaedic biomechanics & mechano-
biology, 3rd ed.; Lippincott Williams & Wilkins: Philadelphia, 2005.
(5) Felson, D. T.; et al. Ann. Intern. Med. 2000, 133, 635.
(6) (a) Palmer, A. W.; Guldberg, R. E.; Levenston, M. E. Proc. Natl. Acad. Sci.
U.S.A. 2006, 103, 19255. (b) Silvast, T. S.; Jurvelin, J. S.; Lammi, M. J.;
Toyras, J. Osteoarthritis Cartilage 2009, 17, 26. (c) Bashir, A.; Gray, M. L.;
Hartke, J.; Burstein, D. Magn. Reson. Med. 1999, 41, 857. (d) Kneeland,
J. B.; Reddy, R. J. Magn. Reson. Imaging 2007, 25, 339.
Figure 1. Comparison of anionic and cationic contrast agents. (a) Transverse
image of ex ViVo rabbit femur. Zoomed-in images in (c) highlight the
bone-cartilage-air interfaces at the femoral groove (white box). (b) Histologi-
cal section of femur in 2 vs CA4+ comparison stained with GAG-specific
Safranin-O dye (red color). Scale bar ) 1 mm. (c) Representative images from
a pairwise comparison of contrast agents, each in a single femur sample (mean
cartilage attenuation in Hounsfield Units ( SD). All pairwise comparisons (1
vs CA1+; 1 vs CA2+; 2 vs CA4+) were statistically significant (p < 0.0001)
using a student’s t test comparison. Scale bar ) 1 mm.
(7) This correlation between GAG and contrast agent concentration has been
confirmed in our lab (see Supporting Information).
(8) Kallioniemi, A. S.; Jurvelin, J. S.; Nieminen, M. T.; Lammi, M. J.; Toyras,
J. Phys. Med. Biol. 2007, 52, 1209.
JA9053306
9
J. AM. CHEM. SOC. VOL. 131, NO. 37, 2009 13235