Magnetic Resonance in Medicine 48:322–330 (2002)
Mapping the Fiber Orientation in Articular Cartilage at Rest and Under Pressure Studied by 2H Double Quantum Filtered MRI Hadassah Shinar,1 Yoshiteru Seo,2 Kazuya Ikoma,2 Yoshiaki Kusaka,3 Uzi Eliav,1 and Gil Navon1* The one-dimensional 2H double quantum filtered (DQF) spectroscopic imaging technique was used to study the orientation of collagen fibers in articular cartilage. The method detects only water molecules in anisotropic environments, which in cartilage is caused by their interaction with the collagen fibers. A large quadrupolar splitting was observed in the calcified zone and a smaller splitting in the radial zone. In the transitional zone the splitting was not resolved and a small splitting was again detected in the superficial zone. From measurements performed at two orientations of the plug relative to the magnetic field it was deduced that in the calcified and radial zones the fibers are oriented perpendicular to the bone, bending at the transitional zone and flattening at the superficial zone. The effect of load applied to the cartilage– bone plug was monitored by the same technique. At low loads there is a small decrease in the quadrupolar splitting in the calcified zone, a marked decrease in the radial zone, and an increase of the splitting accompanied by a thickening of the superficial zone. Under high loads, while the thickening and the splitting of the superficial zone further increase, the splitting in the radial and calcified zones completely collapse. Pressure-induced changes in the thickness of the surface zone indicate flattening of the collagen fibers near the surface. The marked collapse of the splitting near the bone at high pressures may result from crimping of the collagen fibers. Magn Reson Med 48:322–330, 2002. © 2002 Wiley-Liss, Inc. Key words: DQF MRI; articular cartilage; compression; collagen orientation
Articular cartilage is a dense connective tissue that coats the ends of bones in their joints. It is mainly composed of water (⬃75%) and of a solid matrix of collagen fibrils (⬃15%) and proteoglycans (PG) (⬃10%). The fibrous, triple helix collagen molecules define the tissue’s shape and provide its tensile strength. The PG are composed of a central protein core with many glycosaminoglycan (GAG) sulfated sidechains. These are highly negatively charged and thus attract high concentrations of positive ions and water molecules. Scanning electron microscopy (SEM) has shown that the collagen fibers rise vertically from the bone through the radial zone, then bend and flatten, forming the
1
School of Chemistry, Tel Aviv University, Tel Aviv, Israel. Department of Physiology Kyoto Prefectural University of Medicine, Kyoto, Japan. 3 Department of Orthopedic Surgery, Murakami Memorial Hospital, Asahi University, Gifu, Japan. Grant sponsor: German Federal Ministry of Education and Research (BMBF) within the framework of German-Israeli Project Cooperation (DIP). *Correspondence to: Prof. Gil Navon, School of Chemistry, Tel Aviv University, Ramat Aviv, Tel Aviv, 69978, Israel. E-mail:
[email protected] Received 27 July 2001; revised 20 February 2002; accepted 9 March 2002. DOI 10.1002/mrm.10195 Published online in Wiley InterScience (www.interscience.wiley.com). 2
© 2002 Wiley-Liss, Inc.
superficial tangential zone (1–3). This structure, together with the large osmotic pressure in the tissue, is responsible for the remarkable compressive strength of the tissue. In conventional MR images, articular cartilage has a laminated appearance (4 –14). The number of laminae, their relative thickness and intensity, vary from study to study and from sample to sample and are strongly dependent on the orientation of the tissue in the magnetic field. In collagen-containing tissues, it has been shown that the transverse relaxation rate is dominated by the residual dipolar interaction (6,8,15), which is a result of the anisotropic motion of the water molecules. Thus, T2 depends on the orientation of the collagen fibers with respect to the magnetic field. The main function of cartilage is to withstand pressure. Direct visualization of the orientation of the collagen fibers in articular cartilage at rest and under applied load is obtained by SEM (2,3). MRI investigations of articular cartilage under various degrees of loadings were also reported by several authors (16 –20). The effects observed by MRI were related to the changes in the water T2 occurring under pressure. Since the deuteron is a spin I ⫽ 1 nucleus, its NMR spectra in anisotropic media are composed of two satellite transitions with a splitting equal to twice the quadrupolar interaction. The frequency difference between the two satellites in the 2H NMR spectrum of deuterated water interacting with oriented collagen fibers (21,22) depends on (1-3cos2⌰), where ⌰ is the angle between the director of the quadrupolar interaction and the magnetic field. In biological tissues, the amount of isotropic water is large relative to water molecules that experience anisotropic tumbling and thus the satellite transitions are masked by the isotropic signal. The double quantum filtering (DQF) technique is based on the existence of order in the tissue. The residual quadrupolar interaction allows the formation of even rank tensors and thus the application of DQF techniques. The signal of isotropic water is filtered out and only ordered structures are detected. In previous 2H DQF studies it was shown that in nasal cartilage the order of water molecules is local (23) and the residual dipolar interaction is insensitive to the orientation of the sample in the magnetic field. In articular cartilage the order is macroscopic (24,25). In these studies the orientation dependence of the lineshape was interpreted in terms of a Gaussian distribution of the angles between the local directors and the magnetic field and a Gaussian distribution of the quadrupolar interaction. It was shown that the aniostropic motion of the water molecules stems from their interaction with the oriented collagen fibers.
322
2
H Double Quantum Filtered MRI
323
2 H DQF NMR was also utilized for the study of sciatic nerve (26) and blood vessels (27). In these tissues inner structures could be differentiated on the basis of their different quadrupolar splittings. By extending the DQF spectroscopic technique to DQF spectroscopic MRI (27), we obtained “histological” images of the blood vessel walls. It was also shown that the quadrupolar splitting depends on the strain of the blood vessels and “strain” images were calculated. In our previous study (25), average values and standard deviations of the quadrupolar splitting were obtained for a whole piece of cartilage. In the present work we observed the variation of the splitting within the cartilage as a function of the distance from the bone– cartilage interface by using 2H DQF spectroscopic imaging. The method was applied to cartilage plugs under incremental loading, giving information about the variation of the collagen fiber orientation in response to the applied pressure.
MATERIALS AND METHODS Cartilage and bone plugs 7– 8 mm in diameter were obtained from fresh bovine femoral lateral and medial condoyles and from the pattela. The cartilage and bone plugs were equilibrated for a minimum of 3 hr in deuterated saline and subsequently measured after a day or two. We have found that the 2H spectra were constant after this period of equilibration. The thickness of the various cartilage plugs varied from 1–3 mm. Cartilage was separated from the bone by cutting the plug above the calcified zone where the plug is soft and cutting is performed easily. For most measurements the plugs were placed in the NMR tube so that the surface of the cartilage was perpendicular to the magnetic field. For bovine Achilles tendon the same equilibrating procedure was employed. The Achilles tendon was measured with its long axis parallel to the external magnetic field. For NMR measurements samples were immersed in fluorinated oil (Fluorinert, FC-77, 3M, Minneapolis, MN). NMR measurements were recorded on an AMX-360 WB and AMX-300 WB Bruker NMR spectrometers operating at 2 H frequencies of 55.3 and 46.1 MHz, respectively. Both instruments were equipped with a 10-mm 2H-RF coil and a set of gradient coils with a maximum gradient strength of 200 G/cm. All experiments were conducted at room temperature (ca. 22°C). DQF spectra were recorded with the conventional pulse sequence: 90° – /2 – 180° – /2 – 90° – tDQ – 90° – t2 (Acq)
90° – /2 – 90° – t DQ – 90° – /2 – 90° – t ZQ – 90° – t 2 (Acq)
[2]
where tZQ is the zero quantum evolution time. While with the conventional DQF method the satellite signals are obtained in antiphase with each other, they are obtained within the same phase in the IP-DQF sequence. The pulse sequence consists of a DQ filter followed by a zero quantum filter. The one-dimensional (1D) DQF spectroscopic imaging pulse sequence and its coherence transfer diagram are given in Fig. 1. The imaging gradients are applied during tDQ, the DQ evolution time, which is relatively long for 2H water signals in connective tissues (26,27). For the measurements under static load, cartilage– bone plugs, 7 mm in diameter, were obtained from fresh bovine medial condoyles and equilibrated in deuterated saline. Plugs were placed in heavy wall NMR tubes (513-7TR9, Wilmad) with the cartilage surface on top of a glass filter. The pressure was applied from the bone side by calibrated nonmagnetic springs. Images were measured 30 min after the application of the load to allow for re-equilibration. The 2H DQF spectroscopic image was obtained with the phase encoding gradients applied perpendicular to the cartilage surface. RESULTS In order to demonstrate the applicability of the 1D DQF spectroscopic MRI method, the experiment was performed on a piece of bovine Achilles tendon. This tissue is mainly composed of collagen fibers uniformly aligned along its long axis. The 2H single pulse spectrum of bovine Achilles tendon, equilibrated in deuterated saline (Fig. 2a), is composed of a narrow singlet and two satellite transitions which are separated by approximately 2500 Hz. In the IP-DQF spectrum (Fig. 2b), only the two satellite transitions are observed, since ordered water molecules are detected by this technique and the intense signal of the isotropic water is eliminated. Thus, we attribute the broad satellite signals to water molecules interacting with the macroscopically oriented collagen fibers of the tendon.
[1]
where tDQ is the DQ evolution time. Both the lineshape and the intensity of the spectra depend on creation time . max, the creation time at which a maximum intensity is obtained, varies from one system to the other and can thus serve as a contrast mechanism by which water in different compartments can be independently detected. 2 H DQF in-phase (IP-DQF) spectra (28) were obtained with the following pulse sequence:
FIG. 1. The 1D 2H DQF spectroscopic MRI pulse sequence and coherence transfer diagram. Phase encoding gradients (G) were applied during the double quantum evolution period (tDQ).
324
Shinar et al.
FIG. 2. An SQ spectrum (a) and IP-DQF spectrum with ⫽ 200 s (b) of bovine Achilles tendon immersed in deuterated saline. The tendon was measure with its long axis parallel to the magnetic field. Pulse sequence shown in Eq. [2]. Spectra were measured on a Bruker AMX 360 WB spectrometer, operating at a 2H frequency of 55.3 MHz.
The narrow signal in the single pulse spectrum is assigned to the free, isotropically rotating water molecules. The frequency splitting between satellite transitions depends on the orientation of the tendon relative to the magnetic field. Maximum splitting was obtained when the long axis of the tendon, and thus the collagen fibers was parallel to the magnetic field, as was also previously reported (21,29). The 1D DQF spectroscopic image of bovine Achilles tendon is given in Fig. 3. The x-axis represents the spectral dimension (in Hz) and the y-axis the spatial dimension (in mm). The same frequency difference is obtained throughout the length the tendon, indicating that the order of the collagen fibers was constant within the image dimensions. The signal intensity of the image is almost constant in the range of 8 mm and tends to decrease in the outer regions representing the sensitivity profile of the RF coil. Thus, these results confirm that our pulse sequence and hardware, including the gradient system, are reliable for measuring 1D DQF spectroscopic imaging in the field of view of 8 mm. The 2H DQF spectra of a piece of articular cartilage– bone plug excised from bovine medial condoyle, equilibrated in deuterated saline, were obtained at various creation times, , in the range of 100 s to 100 ms (Fig. 4). The plug was placed in the NMR tube so that the surface of the cartilage was perpendicular to the external magnetic field. Two pairs of satellite transitions are evident from the spectra with frequency splittings of 1500 Hz and 500 Hz. These spectra represent a weighted-average of the spectra of the different layers of the cartilage– bone plug. The relative intensities of the two transitions depend on with a characteristic rise and decay time. As expected (26,27), the outer transition reaches a maximum at shorter creation time. Information about the variation of the splitting as a function of the distance from the bone surface is obtained from the 1D spectroscopic image (Fig. 5a) of one particular cartilage bone plug. Fifty bovine articular cartilage plugs,
FIG. 3. 2H DQF spectroscopic image of bovine Achilles tendon equilibrated in deuterated saline. The tendon was measure with its long axis parallel to the magnetic field. The image was recorded on a Bruker AMX 360 WB spectrometer. The DQF image was acquired with the following parameters: field of view 2 cm ⫻ 20 kHz, data matrix 64 ⫻ 2048, ⫽ 200 s. The x-axis represents the spectroscopic dimension and the y-axis is the spatial dimension.
taken from 10 different animals, were measured in the course of this study. Although the images obtained from the different plugs are similar, it should be emphasized that each plug has a characteristic appearance as well as a
FIG. 4. 2H DQF spectra of a cartilage-bone plug equilibrated in deuterated saline as a function of the creation time, . The cartilage surface was perpendicular to the magnetic field. Measurements were conducted on a Bruker AMX 300 WB spectrometer, operating at a 2H frequency of 46.08 MHz.
2
H Double Quantum Filtered MRI
FIG. 5. a: 2H DQF spectroscopic image of bovine articular cartilagebone plug, equilibrated with deuterated saline. This plug was the same as in Fig. 4. The x-axis represents the spectroscopic dimension and the y-axis is the spatial dimension. Field of view 0.7 cm ⫻ 20 kHz, data matrix 64 ⫻ 4096, ⫽ 100 s. The 2D FT was performed after zero-filling of data to 128 ⫻ 4096 without any filter function. b: 2H DQF spectra extracted from the image at selected distances from the articular surface.
different depth. In almost all plugs, we could distinguish between a larger splitting near the bone and a smaller splitting in the radial zone as well as a small splitting in the superficial zone. Average values for these splittings (value ⫾ SD) are: calcified zone: 1320 ⫾ 230 Hz (n ⫽ 34), radial zone 450 ⫾ 80 Hz (n ⫽ 30), and superficial zone 160 ⫾ 40 Hz (n ⫽ 24). As in the previous example (Fig. 3), the x-axis represents the spectral dimension and the y-axis the spatial dimension. The frequency splitting of the two satellite transitions is evident from the image as well as
325
from the individual spectra (Fig. 5b) obtained from different locations between the cartilage surface and the bone. Near the bone a large splitting of approximately 1500 Hz is detected. Above this zone, and moving towards the surface, a second, much smaller splitting on the order of 600 Hz is detected. This splitting decreases continuously and is completely lost about 200 m beneath the surface. In the surface layer, a much smaller splitting (approximately 200 Hz) is observed. The origin of the large splitting is evident from the spectroscopic image in Fig. 6. Here, the cartilage was separated from the bone above the calcified zone. Both pieces—the cartilage and the bone with the calcified zone—were placed in the NMR tube with a Teflon spacer between them. As can be seen from the figure, the region with the large splitting can be identified with the calcified zone. No DQF signal could be observed from the bone. The effect of the orientation on the 2H DQF spectroscopic image was studied on a sample of patellar cartilage detached from the bone (Fig. 7) and a number of samples of articular bone– cartilage plugs (Fig. 8). For the detached patellar cartilage, only one pair of satellites is observed at each location along the cartilage. When the magnetic field was parallel to the cartilage surface, the splittings near the bone and in the radial zone were approximately half the values obtained when the magnetic field was perpendicular to the cartilage surface (Fig. 7). In the intermediate zone no splitting was observed at both orientations of the plug relative to the magnetic field. In the superficial zone, the splitting was slightly bigger when the surface of cartilage was parallel to the magnetic field. The situation is more complicated for the articular bone– cartilage plug. Two pairs of satellite transitions and a small narrow signal in the center are evident at most of the locations on the cartilage. The narrow central signal can be due to a third site, with a very small splitting, or to some leakage of the signal of free water molecules through the DQ filter. Thus, in this case we used a line fitting procedure to extract the splittngs. The procedure is based on the following assumptions: 1) at each spatial location there are three different types of quadrupolar interactions, each representing a different compartment. These compartments are in slow exchange between them. 2) Each compartment is characterized by the number of spins and by the mean value of the Gaussian distribution of the quadrupolar interaction. As a result of the fitting, the quadrupolar splittings, their distributions, and the relative amount of spins contributing to each compartment is obtained. The values of the two splittings, as a function of the distance from the articular surface, at two orientations of the plug relative to the magnetic are given in Fig. 8. As in the case of the patellar cartilage, a factor of approximately 2 is observed for each of the splittings in the calcified and radial zones between the two plug orientations. Pressure-induced changes in DQF spectroscopic images were observed under stepwise application of static loads up to 1.0 MPa (0, 0.15, 0.3, 0.5, and 1.0 MPa). After each application of pressure a 30-min interval was allowed in order to reach a new equilibrium and subsequently 1D 2H DQF spectroscopic images were measured. A typical set of images of cartilage at various static loads is shown in Fig. 9a. Before the application of any load the large splitting in
326
Shinar et al.
FIG. 6. 2H DQF spectroscopic image of a combination of the two separated parts of a bovine articular cartilage and bone plug. The plug was equilibrated with deuterated saline and the cartilage was detached from the bone, above the calcified zone. The two parts were placed in the NMR tube, so that both the surface of the bone and the surface of the cartilage are perpendicular to the magnetic field and the Teflon spacer is between them. Field of view 1.0 cm ⫻ 20 kHz, data matrix 128 ⫻ 2048, ⫽ 400 s. Spectra extracted from the image indicated by arrows.
the calcified zone and the smaller splitting in the radial zone are clearly observable. When the load is increased the thickness of the cartilage, as seen from the vertical scale, decreases gradually. Upon application of low loads, 0 – 0.3 MPa, there is a slight decrease of the splitting at the calcified zone, a marked decrease of the splitting at the radial zone, and a significant increase of the splitting and a thickening of the superficial zone. At higher pressures (0.5–1.0 MPa), while the increase of the splitting at the surface and the thickening of this zone continues, the splitting at the calcified region completely collapses. Individual spectra, extracted from the image at a load of 1.0 MPa, are given in Fig. 9b. After releasing the pressure and re-equilibrating the bone– cartilage plug in deuterated saline, the image obtained is identical to the original image before the application of pressure, except for a reduction in the intensity. This is caused by the fact that the first image was originally measured on a larger plug, which was then cut down to fit into the pressure tube. Strain is defined as the percent difference in the thickness between the unloaded and the loaded cartilage rela-
tive to the unloaded one. Plugs were obtained from different positions on the condoyle and the pressure changes were monitored for each plug. The stress–strain relationships of the weight-bearing region and nonweight bearing region were almost the same. The mean and standard error of strain were 28.7 ⫾ 2.1, 43.4 ⫾ 1.7, 52.7 ⫾ 1.7, 62.6 ⫾ 1.0% (n ⫽ 12) for the static loads of 0.15, 0.3, 0.5, and 1.0 MPa, respectively. DISCUSSION In this study we have introduced a method that enables determination of the orientation and the degree of order of the collagen fibers at each spatial location on a cartilage plug. The 2H DQF NMR technique depicts exclusively water having residual quadrupolar splitting caused by the anisotropic reorientation of the water molecules interacting with the oriented collagen fibers. The magnitude of the splitting and its dependence on the orientation of the sample relative to the magnetic field are a measure of the direction of the collagen fibers and their degree of hydra-
2
H Double Quantum Filtered MRI
FIG. 7. The 2H quadrupolar splitting as a function of the distance from the cartilage surface at two orientations of the cartilage plug relative to the magnetic field. Patellar cartilage– bone plug was equilibrated in deuterated saline and the cartilage was detached from the bone above the calcified zone. 2H DQF spectroscopic images were obtained with the cartilage surface perpendicular (■) and parallel (F) to the magnetic field. The quadrupolar splitting at each location was measured from the extracted spectra. Measurements were performed on a Bruker AMX 360 WB spectrometer, operating at a 2H frequency of 55.3 MHz. Field of view 1.0 cm ⫻ 20 kHz, data matrix 128 ⫻ 2048, ⫽ 400 s.
tion. The 2H DQF spectrum of a cartilage– bone plug is a sum of spectra in the different zones of the plug. In the DQF spectroscopic image the spectrum from each zone is obtained giving information on the spatial variation of the order of the collagen fibers. We have found that when the surface of the cartilage– bone plug is perpendicular to the external magnetic field, a large splitting is obtained near the bone. This splitting decreases slowly towards the radial zone. At the bottom part of the radial zone, in most cases, two different splittings are evident, both decreasing towards the intermediate zone. At a certain location only one splitting is evident. The fact that we are observing at least two pairs of satellite signals at each location indicates that there are at least two different water pools in slow exchange on the NMR timescale. This may be due to water interacting with collagen fibers of different diameters, density, and degree of hydration at the same location on the plug. The variation of each of the splittings, as a function of the location, could have been interpreted as an increase in the spread of orientation of the collagen fibers. Yet the fact that the splitting changes approximately by a factor of 2 upon 90° rotation of the cartilage sample is an indication that the collagen fibers are well oriented throughout the calcified and radial zones. Thus, the variation of the splitting is probably a result of changes in the fraction of water molecules interacting with the collagen fibers and the pool of water molecules exchanging with this fraction. In the transitional zone the splitting vanishes and the signal intensity is relatively low. When the orientation of the plug in the magnetic field is changed by 90°, so that the surface is parallel to the field, the splitting in the transitional zone is again not resolved. These results indicate that the fibers in this zone are not uniformly oriented and for part of their length they are aligned at or near the magic
327
angle. Water molecules interacting with fibers oriented at the magic angle relative to the magnetic field are not expected to contribute to the DQF signal. Indeed, a very low-intensity DQF signal is detected from the transitional zone and the splitting could not be resolved. In most cases, a small splitting appears again at the superficial zone. Unlike the splitting of the radial zone, the value of the splitting at the superficial zone increases slightly upon 90° rotation, so that the surface is parallel to the field. This behavior is in line with tangentially aligned fibers at the surface (1–3). When the magnetic field is parallel to the surface the angle between the fibers and the magnetic field can be of any value between 0 –90°, so that a wide distribution of splittings is expected, ranging from twice that of the radial zone or that of the surface zone in the perpendicular orientation between the surface and the magnetic field, to zero for the fibers with magic angle orientation. The intensity of the spectrum obtained from the surface is low, since in the present resolution of approximately 0.11 mm the volume represented by the spectrum is only partially occupied by the cartilage tissue. The 2H DQF results agree well with the results obtained by SEM (1–3), where the cartilage is divided into four zones: calcified, radial, transitional, and superficial. The electronic micrographs show clearly that in the calcified and radial zones the fibers are perpendicular to the surface. Moreover, it was shown that in the calcified zone the radially oriented collagen fibers are arranged in closely packed bundles with small calcium deposits (1). Collagen fibers of different diameters are evident in SEM (30). In the radial zone, the orientation of the collagen fibers persists, but there are many linking fibers between adjacent fibrils. Indeed, as was shown by SEM (30), the diameter of the
FIG. 8. The 2H quadrupolar splitting as a function of the distance from the cartilage surface at two orientations of the cartilage plug relative to the magnetic field. Articular cartilage– bone plug was equilibrated in deuterated saline. 2H DQF spectroscopic images were obtained with the cartilage surface perpendicular (■,䊐) and parallel (F,E) to the magnetic field. The quadrupolar splitting at each location were obtained from a line-fitting procedure, described in the text. Measurements were performed on Bruker AMX 360 WB spectrometer, operating at a 2H frequency of 55.3 MHz. Field of view 1.0 cm ⫻ 20 kHz, data matrix 64 ⫻ 2048, ⫽ 100 s.
328
Shinar et al.
FIG. 9. a: 2H DQF spectroscopic images of bovine articular cartilage– bone plug, equilibrated in deuterated saline at various applied loads: 0. 0.15, 0.3, 0.5, and 1.0 MPa as indicated on the figure. The cartilage– bone plug was placed in the NMR tube with the cartilage on top of a glass filter and the pressure was applied by calibrated springs on the bone side. Images were measured 30 min after the application of pressure. At the end of the experiment the plug was taken out of the tube, re-equilibrated in deuterated saline for 12 hr, and measured again. Imaging parameters are the same as in Fig. 5. b: Individual spectra extracted from the image at 1.0 MPa load as a function of the distance from the surface.
fibrils decreases with the distance from the bone and this may be the reason for a higher degree of hydration and reduced splitting. We have adopted the zonal visualization of the cartilage plug and the results are discussed on this basis. We find distinct differences in the splitting and in the orientation of the collagen fibers in the different zones. The relative depth of the different zones, as obtained from our 2H NMR data, is in good agreement with the results of SEM. In a number of MRI studies (4,7,10,13,14,31–36), the variation of the T2 of the water protons was measured as a
function of the distance from the bone. These variations are responsible for the lamellar appearance of the cartilage in MRI. As mentioned above, proton T2 in connective tissues is dominated by the dipolar interaction, which, like the quadrupolar interaction of water deuterons, is related to the fiber orientation and the degree of hydration. The measured value of the proton T2 includes molecules that do not have any interaction with the collagen fibers and their motion is strictly isotropic. This pool of free water does not contribute to the 2H DQF signal and therefore the DQF technique represents the fiber orientation more faith-
2
H Double Quantum Filtered MRI
fully. However, single-pulse experiments on cartilage plug immersed in Fluorinert indicated that the contribution of the isotropic water is relatively small. Indeed, the spectral distribution of the 2H quadrupolar splitting measured in the present work show similar trends to that of proton T2. As in the case of 2H quadrupolar splitting (Fig. 5b), the proton transverse relaxation rate, 1/T2, is largest near the bone, decreases gradually in the radial zone, has its minimum in the transitional zone, and increases again in the superficial zone (7,10,34). In accordance with our results, the 1H T2 results do not show any orientation dependence in the transitional zone. When the cartilage is depleted of its proteoglycans, the spatial dependence (37) as well as the orientation dependence of the splitting (38) follow the same trend as in the intact tissue, with very small changes in the value of the splitting. Although it was assumed by some authors (10,12,39) that the water 1H transverse relaxation time is determined by the proteoglycans, no experimental result was given to support this notion. In fact, Regatte et al. (18,19) found only an insignificant change in the 1H T2 after depletion of 30% of the PG. Moreover, a correlation found between biochemically determined collagen content and MR imaging data (36) suggest that collagen has a more determinant effect on T2. The small effect of PG depletion on the 2H quadrupolar splitting and proton T2 indicates that the PG have only a secondary effect in determining the anisotropy of the water molecules. It is interesting to note that, in contrast to the 2H and 1H results, 23 Na quadrupolar splitting is very sensitive to the proteoglycans concentration, increasing about twofold upon complete PG depletion (40). However, it was suggested that that this was not a result of increased order but rather a decrease in the amount of bulk Na⫹ and thereby an increase in the fraction of Na⫹ bound to the collagen fibers. One of the merits of the technique introduced here lies in its ability to measure the orientation of the collagen fibers when the cartilage plug is in its native state. No fixation, mechanical slicing, or enzymatic treatment which may result in structural changes are employed. The 1D 2H DQF spectroscopic imaging technique is particularly useful for the measurement of the orientation of the collagen fibers under load. The collagen fibers’ orientation can be monitored on the same piece of cartilage following the application of pressure and its release. The 2H DQF spectroscopic image presented here revealed dramatic changes in the quadrupolar splitting in the various parts of the plug when pressure is applied to the plug. We have found that the splitting on the surface of the cartilage increases and the splitting in the deep part and near the bone is eventually lost. This is an indication that, concomitantly with the loss of water and the shrinkage of the tissue in response to the load, there is a reorganization of the collagen fiber architecture. The results obtained indicate a flattening and an increase in the width of the superficial zone as a result of the application of pressure. This is in agreement with SEM results (3). The marked collapse of the splitting near the bone may be a result of the crimping of the collagen fibers in this zone. The effect of pressure was found to be fully reversible (41). Our results are in agreement with results obtained by 1H T2 measurements. Rubenstein et al. (16) and Gru¨ nder et al. (20) found a thickening of the
329
superficial, low T2 lamina with a gradual decrease of the T2 in the deep parts of the cartilage. The recovery to the original image (4,16) found after the release of the pressure is again in agreement with the present work. Regatte et al. (19) also reported shortening of the T2 upon compression. However, the spatial distribution of the T2 was not given. The agreement between the results obtained by proton MRI and those obtained with 2H DQF is expected since both the 2H quadrupolar splitting as well as the 1H transverse relaxation rate are affected by the fiber orientation and the degree of hydration. However, the 2H quadrupolar splitting is more specific, since factors such as viscosity are expected to affect the 1H T2 and not the quadrupolar interaction. REFERENCES 1. Jeffery AK, Blunn GW, Archer CW, Bentley G. Three-dimensional collagen architecture in bovine articular cartilage. J Bone Joint Surg 1991; 73–B:795– 801. 2. Notzli H, Clark JM. Deformation of loaded articular cartilage prepared for scanning electron microscopy with rapid freezing and freeze-substitution fixation. J Orthop Res 1997;15:79 – 86. 3. Kaab MJ, Ito K, Clark JM, Notzli HP. Deformation of articular cartilage collagen structure under static and cyclic loading. J Orthop Res 1998; 16:743–751. 4. Lehner KB, Rechl HP, Gmeinwieser JK, Heuck AF, Lukas HP, Kohl HP. Structure, function, and degeneration of bovine hyaline cartilage: assesment with MR imaging in vitro. Radiology 1989;170:495– 499. 5. Modl JM, Sether LA, Huaghton VM, Kneeland JB. Articular cartilage: correlation of histologic zones with signal intensity at MR imaging. Radiology 1991;181:853– 855. 6. Rubenstein JD, Kim JK, Morova-Protzner I, Stanchez PL, Henkelman RM. Effects of collagen orientation on MR imaging characteristics of bovine articular cartilage. Radiology 1993;188:219 –226. 7. Freeman DM, Bergman G, Glover G. Short TE MR microscopy: accurate measurement and zonal differentiation on normal hyelin cartilage. Magn Reson Med 1997;38:72– 81. 8. Xia Y, Farquhar T, Burton-Wurster N, Lust G. Origin of cartilage laminae in MRI. J Magn Reson Imag 1997;7:887– 894. 9. Cova M, Toffanin R, Frezza F, Pozzi-Mucelli M, Mlynarik V, PozziMucelli RS, Vittur F, Dalla-Palma L. Magnetic resonance imaging of articular cartilage: ex-vivo study on normal cartilage correlated with magnetic resonance microscopy. Eur Radiol 1998;8:1130 –1136. 10. Xia Y. Relaxation anisotropy in cartilage by NMR microscopy (MRI) at 14-m resolution. Magn Reson Med 1998;39:941–949. 11. Uhl M, Ihling C, Allmann KH, Lauberberger J, Tauer U, Adler CP, Langer M. Human articular cartilage: in vitro correlation of MRI and histologic findings. Eur Radiol 1998;8:1123–1129. 12. Gru¨ nder W, Wagner M, Werner A. MR-microscopic visualization of anisotropic internal cartilage structures using magic angle techniques. Magn Reson Med 1998;39:376 –382. 13. Kim DJ, Suh JS, Jeong EK, Shin KH, Yang WI. Correlation of laminated MR appearance of articular cartilage with histology, ascertained by artificial landmarks on the cartilage. J Magn Reson Imag 1999;10:57– 64. 14. Xia Y. Heterogenity of cartilage laminae in MR imaging. J Magn Reson Imag 2000;11:686 – 693. 15. Henkelman MR, Stanisz, GJ, Kim JK, Bronskill MJ. Anisotropy of NMR properties of tissues. Magn Reson Med 1994;32:592– 601. 16. Rubenstein JD, Kim JK, Henkelman RM. Effects of compression and recovery on bovine articular cartilage: appearance on MR images. Radiology 1996;210:843– 850. 17. Herberhold C, Stammberger T, Faber S, Putz R, Englmeier KH, Reiser M, Eckstein F. An MR-based technique for quantifying the deformation of articular cartilage during mechanical loading in an intact cadaver joint. Magn Reson Med 1998;39:843– 850. 18. Kaufman JH, Regatte RR, Bolinger L, Kneeland JB, Reddy R, Leigh JS. A novel approach to observing articular cartilage deformation in vitro via magnetic resonance imaging. J Magn Reson Imag 1999;9:653– 662. 19. Regatte RR, Kaufamn JH, Noyszewski EA, Reddy R. Sodium and proton MR properties of cartilage during compression. J Magn Reson Imag 1999;10:961–967.
330 20. Gru¨ nder W, Kanowski M, Wagner M, Werner A. Visualization of pressure distribution within loaded joint cartilage by application of anglesensitive NMR microscopy. Magn Reson Med 2000;43:884 – 891. 21. Dehl RE, Hoeve CAJ. Broad-line NMR study of H2O and D2O in collagen fibers. J Chem Phys 1969;50:3245–3251. 22. Migchelsen C, Berendsen HJC. Proton exchange and molecular orientation of water in hydrated collagen fibers. An NMR study of H2O and D2O. J Chem Phys 1971;59:296 –305. 23. Sharf Y, Eliav U, Shinar H, Navon G. Detection of anisotropy in cartilage using 2H double quantum filtered NMR spectroscopy. J Magn Reson 1995;B107:60 – 67. 24. Shinar H, Eliav U, Knubovetz T, Sharf Y, Navon G. Measurement of order in connective tissues by multiple quantum filtered NMR spectroscopy of quadrupolar nuclei. Q Magn Reson Biol Med 1995;2:73– 82. 25. Shinar H, Sharf Y, Maroudas A, Navon G. Orientation dependent 2H residual quadrupolar interaction in ex-vivo human articular cartilage. Workshop on magnetic resonance of connective tissues and biomaterials, Philadelphia, 1998. 26. Shinar H, Seo Y, Navon G. Discrimination between the different compartments in sciatic nerve by 2H double-quantum filtered NMR. J Magn Reson 1997;129:98 –104. 27. Sharf Y, Seo Y, Eliav U, Akselrod S, Navon G. Mapping of the strain exerted on blood vessel walls using deuterium double quantum filtered MRI. Proc Natl Acad Sci USA 1998;95:4108 – 4112. 28. Eliav U, Navon G. A study of dipolar interactions and dynamic processes of water molecules in tendon by 1H and 2H homonuclear and heteronuclear multiple quantum-filtered NMR spectroscopy. J Magn Reson 1999;137:295–310 29. Tsoref L, Shinar H, Navon G. Observation of 1H double quantum filtered signal of water in biological tissues. Magn Reson Med 1998;39: 11–17. 30. Clark JM. The organization of collagen in cryofractured rabbit articular cartilage: a scanning electron microscopic study. J Orthop Res 1985;3: 17–29.
Shinar et al. 31. Dardzinski BJ, Mosher TJ, Shizhe L, Van Slyke MA, Smithe MB. Spatial variation of T2 in human articular cartilage. Radiology 1997;205:546 – 550. 32. Fragonas E, Mlynarik V, Jellus V, Micali F, Piras A, Toffanin R, Rizzo R, Vittur F. Correlation between biochemical composition and magnetic resonance appearance of articular cartilage. Osteoarthritis Cartilage 1998;6:24 –32. 33. Frank LR, Wong EC, Lyh WM, Ahn JM, Resnick D. Articular cartilage in the knee: mapping of physiologic parameters at MR imaging with local gradient coil—preliminary results. Radiology 1999;210:241–246. 34. Mosher TJ, Dardzinski BJ, Smith MB. Human articular cartilage: influence of aging and early symptomatic degeneration on the spatial variation of T2 — preliminary findings at 3T. Radiology 2000;214:259 –266. 35. Nieminen MT, Toyras J, Rieppo J, Hakumaki JM, Silvennoinen J, Helminen HJ, Jurvelin JS. Quantitative MR microscopy of enzymatically degraded articular cartilage. Magn Reson Med 2000;43:676 – 681. 36. Watrin A, Ruaud JPB, Olivier PTA, Guingamp NC, Gonord PD, Netter PA, Blum AG, Guillot GM, Gillet PM, Loeuille HJ. T2 mapping of rat patellar cartilage. Radiology 2001;219:395– 402. 37. Keinan-Adamsky K, Shinar H, Eliav U, Seo Y, Navon G. 23Na spectroscopic MRI studies of intact and proteoglycan depleted cartilage. In: Proc ISMAR 14th Meeting, Rhodes, Greece, 2001. 38. Shinar H, Eliav U, Scheneiderman R, Maroudas A, Navon G. 2H and 23 Na multiple quantum filtered NMR spectroscopy and diffusion measurements of human articular cartilage. In: Proc 3rd Annual Meeting SMR, Nice, France, 1995. p 432. 39. Goodwin DW, Zhu H, Dunn JF. In vitro MR imaging of hyaline cartilage. Am J Roentgenol 2000;174:405– 409. 40. Danziger O, Shinar H, Eliav U, Navon G. Differentiation between the action of different enzymes on the structure of articular cartilage using multiple quantum filtered 23Na NMR. In: Proc 7th Annual Meeting ISMRM, Philadelphia, 1999. p 1522. 41. Broom ND, Myers DB. A study of the structural response of wet hyaline cartilage to various loading situations. Connect Tissue Res 1980;7:227– 237.