In Vivo Qualitative Assessments of Articular Cartilage in the Rabbit Knee With High-Resolution MRI at 3 T Didier Laurent, 1 * James Wasvary, 1 Elizabeth O’Byrne, 1 and Markus Rudin 2 Proteoglycan (PG) loss and disruption of the collagen frame- work in cartilage are early events associated with osteoarthritis (OA). The feasibility of in vivo high-resolution MRI assessments probing both macromolecules was explored in articular carti- lage of the rabbit knee. One-millimeter thick coronal images were obtained at 3 T with a 97  97 m 2 pixel size. A 22% decrease in the magnetization transfer (MT) exchange rate along with an 2-fold greater Gd(DTPA) 2- -induced decrease in T 1 relaxation time were measured in response to papain injec- tion 1 day prior to the MRI session, indicative of an alteration of collagen integrity and PG depletion, respectively. A two-point method was tested as an alternative to the more time-consum- ing multipoint method typically used to measure T 1 changes. Kinetics of Gd(DTPA) 2- uptake were observed with a 10-min time resolution. The diffusive transport of Gd(DTPA) 2- was char- acterized by a T 1 decrease 2-fold faster in papain-treated knees. These data suggest that kinetics of tracer diffusion may be used as an informative marker of PG loss, in addition to the amplitude of T 1 variations. When applied to a relevant OA model, the combination of MT and Gd(DTPA) 2- -MRI may help in identifying new active compounds during efficacy studies on cartilage protection. Magn Reson Med 50:541–549, 2003. © 2003 Wiley-Liss, Inc. Key words: cartilage; collagen; degeneration; gadolinium; mag- netization transfer; osteoarthritis; papain; proteoglycan; rabbit Osteoarthritis (OA) is a degenerative disease characterized by failure of the cartilage matrix, ultimately leading to joint instability. Early changes in the biochemical compo- sition of articular cartilage, such as proteoglycan (PG) de- pletion and loosening of the collagen network (1), play an important role in the etiology and progression of the dis- ease. Large aggregating PGs account for about 40% of the tissue dry weight (2). Their major function is hydration of the cellular matrix, which in turn provides much of the cartilage resiliency through electrostatic repulsion. Due to abundant negative charges on glycosaminoglycan sidechains, a fixed charge density (FCD) on the order of –200 mM has been reported in adult articular cartilage (3). The resulting osmotic pressure induces resistance to loss of fluid under compression. This implies that a decrease in FCD that is likely to expose cartilage to dehydration pre- disposes the joint to OA. Collagen, which accounts for up to 50 – 60% of tissue dry weight (4), functions as a load- bearing material, providing not only a smooth surface for joint articulation but also extra stiffness to restrain stress increase in the PGs (5). PG loss may result in a structural change of the collagen matrix. For instance, water can accumulate in the intrafibrillar space of collagen as the lateral space between collagen fibrils increases due to the decrease in osmotic pressure (6). However, this is most likely not sufficient to compensate for the loss of water associated with PG depletion itself. As collagen and PGs obviously work in concert, a combination of methods that can noninvasively follow both macromolecules in parallel would be useful to quantify the severity of cartilage abnor- mality as the disease progresses. Recently, the delayed gadolinium-enhanced MRI tech- nique emerged as an in vivo method capable of detecting PG loss in OA tissue (7). In principle, this contrast imaging method allows measurement of cartilage FCD. The as- sumption is that the negatively charged gadolinium com- plex (Gd(DTPA) 2- ) penetrates the interstitial fluid of carti- lage to reach an equilibrium concentration that is governed by 1) the Gd(DTPA) 2- -concentration gradient and 2) elec- trostatic interactions (inversely proportional to the FCD). Due to its paramagnetic properties, the distribution of Gd(DTPA) 2- in degraded cartilage is reflected by a decrease in the T 1 relaxation time of water, especially where tissue PGs are depleted. The local Gd(DTPA) 2- concentration can be derived from local T 1 maps, which are usually com- puted from an inversion recovery (IR) series. This method, although accurate, is rather time-consuming, in particular when high spatial resolution is required for studying fine structures. In addition, the poor time resolution of the IR method may not allow kinetic assessment of Gd(DTPA) 2- diffusion into the cartilage. To account for these shortcom- ings a two-point method to estimate the T 1 relaxation time has been applied (8 –10). The exchange of magnetization between the bulk water pool and the water pool bound to macromolecules has been demonstrated to be a good measure of the integrity of the macromolecular network in cartilage (11). In a magne- tization transfer (MT) experiment the broad resonance from bound water is selectively saturated, which results in a decrease in the amplitude of the bulk water signal. The degree of the signal reduction depends on the exchange rate of water between the two states and on the T 1 relax- ation time of bulk water. Several in vitro studies have demonstrated that the MT effect in cartilage is dominated by the contribution of collagen, while the influence of PGs is significantly smaller (11–14). In a goat model of acute cartilage degeneration, we have previously reported a close relationship between the ex- tent of papain-induced PG loss and the Gd(DTPA) 2- -in- duced T 1 decrease (15). In the same animals, MT measure- ments evaluated the integrity of the collagen structure (13). In the present study these methods of assessing PG loss and collagen integrity are applied to the rabbit knee. Rab- 1 Novartis Institute for Biomedical Research, East Hanover, New Jersey. 2 Novartis Pharma AG, Basel, Switzerland. *Correspondence to: Didier Laurent, Ph.D., Novartis Institute for Biomedical Research, In Vivo MRI Laboratory, One Health Plaza, Bldg. 437/1319, East Hanover, NJ 07936-1080. E-mail: didier.laurent@pharma.novartis.com Received 4 November 2002; revised 8 April 2003; accepted 11 May 2003. DOI 10.1002/mrm.10566 Published online in Wiley InterScience (www.interscience.wiley.com). Magnetic Resonance in Medicine 50:541–549 (2003) © 2003 Wiley-Liss, Inc. 541