| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Experimental Studies |
1 From Institut für Kardiovaskuläre Physiologie, Klinikum der Johann Wolfgang Goethe-Universität, Theodor-Stern-Kai 7, D-60590 Frankfurt, Germany. Received March 1, 1999; revision requested April 5; final revision received October 25; accepted November 16. Supported by the German Science Foundation, SFB 553. Address correspondence to A.M. (e-mail: muelsch@em.uni-frankfurt.de).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: The influence of aqueous phantom solutions of nitrosyl-iron complexes on proton relaxation rates was analyzed for signal enhancement at conventional 1.5-T MR imaging. To induce formation of nitrosyl-iron complexes in a biologic tissue, isolated rat liver was perfused with a saline solution of the NO donor sodium nitroprusside (SNP), and the MR signal intensity was examined thereafter.
RESULTS: All investigated nitrosyl-iron complexes shortened the longitudinal, or T1, and transverse, or T2, relaxation times in a concentration-dependent fashion. Relaxivities were highest with a dinitrosyl-iron complex bound to albumin and with a water-soluble mononitrosyl-iron dithiocarbamate complex. The contrast properties of 240 µmol/L of a paramagnetic nitrosyl-iron complex were sufficient to substantially enhance the signal intensity of SNP-perfused rat livers at hydrogen 1 MR imaging.
CONCLUSION: Nitrosyl-iron complexes exhibit a contrast effect at MR imaging that can be exploited for NO imaging in living animals and patients with conventional 1H MR imaging techniques.
Index terms: Animals • Iron • Magnetic resonance (MR), contrast enhancement, 761.12143 • Magnetic resonance (MR), experimental studies, 761.12141 • Magnetic resonance (MR), relaxometry, 761.12146 • Nitrosyl-iron complex, 761.12143 • Phantoms
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In our current study, the suitability of paramagnetic nitrosyl-iron complexes for MR imaging of nitric oxide was evaluated. In many cell types, NO is produced from L-arginine by means of three NO synthase isoenzymes (7). NO serves diverse functions, including relaxation of vascular smooth muscle, inhibition of platelet aggregation, and neurotransmission. Thus, NO can be regarded as one of the most important biologic signal molecules (7,8). Because of the biologic and pharmacologic importance of NO, it is desirable to know the distribution of NO formation in target tissues.
Unfortunately, the extremely short biologic half-life and low steady-state levels of NO (<1 µmol/L) (9) have prevented direct noninvasive detection of its distribution in living organisms. Thus, assessment of NO formation in patients is restricted to ex vivo analyses of NO in gas samples obtained from gaseous body compartments, such as the lung (10), mouth (11), stomach (12), gut (13), and nasal cavities (14), and to ex vivo analyses of inorganic nitrate and nitrite (stable NO metabolites) in body fluids (15).
Recently, the first promising attempts have been made in the imaging of NO formation in living animals by using low-frequency (L-band) electron spin resonance (ESR) spectroscopy with gradient coils and back-projection techniques (16). To stabilize NO and render it visible at ESR imaging, iron-dithiocarbamate (DTC) complexes were applied to the animals (17) to form a paramagnetic nitrosyl-iron–DTC complex NOFe(DTC)2. This complex accumulated in tissues and/or body fluids, depending on the aqueous and/or lipid solubility of the complex. For instance, ESR imaging (18) and in vivo spectroscopy (19) in septic rats and mice revealed high levels of NOFe(DTC)2 formation in the liver. Therefore, in principle, endogenous paramagnetic nitrosyl-iron complexes should be visible in small animals at ESR imaging.
Unfortunately, an inherent shortcoming of ESR imaging is the limited penetration of high radio frequencies (X band, L band) into biologic tissue due to absorption (20). Consequently, in the clinical arena, alternative approaches must be investigated. Theoretically, since paramagnetic nitrosyl-iron complexes can be expected to affect the proton MR signal, we assessed the influence of several of these artificially and naturally formed complexes on the proton nuclear spin relaxation rate to establish an MR imaging technique for NO.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
1 kPa) and nitrogen-gassed solutions of FeSO4·7H2O (1 mL, 72 mmol/L) and neutralized GSH (19 mL, 76 mmol/L) under an atmosphere of pure NO gas (PNO 60 kPa) in a Thunberg-type glass tube (21). The molar ratio of Fe2+ ions and GSH was 1:20, and the concentration of the complex was 3.6 mmol/L. Aqueous solutions with different concentrations (50–1,200 µmol/L) of DNIC-GSH 1:20 were obtained by means of dilution with neutralized GSH (72 mmol/L) in Hepes buffer (15 mmol/L, pH 7.5). The high excess of thiol was required to keep the complex in the monomeric paramagnetic state. DNIC-GSH 1:2 was similarly prepared, but with only a twofold molar excess of GSH to Fe. This complex was a diamagnetic dimer.
DNIC-di-L-cysteine complex Fe(NO)2 (cysteine)2 1:20 was prepared as described previously, but neutralized L-cysteine (72 mmol/L) was used instead of GSH. Solutions with different concentrations (90–1,100 µmol/L) were obtained by means of dilution with neutralized cysteine (72 mmol/L) in Hepes buffer (15 mmol/L, pH 7.5).
Bovine serum albumin (BSA)–bound DNIC Fe(NO)2(BSA) was prepared as described previously (22). Fatty acid–free BSA (1 mmol/L) was dissolved in Hepes buffer (15 mmol/L, pH 7.5) and incubated with DNIC-GSH (1:20, 0.4 mmol/L) for 5 minutes. The suspension was diluted with BSA (1 mmol/L) in Hepes buffer solution to obtain solutions with different concentrations (50–400 µmol/L) of DNIC-BSA.
Mononitrosyl-iron complex (MNIC) with N-methyl-D-glucamine DTC (MGD) Fe(NO)(MGD)2 was obtained by mixing MGD (30 mmol/L) with DNIC-GSH (1:2, 600 µmol/L). Solutions of different concentrations (50–500 µmol/L) were prepared by means of dilution with neutralized GSH (72 mmol/L) dissolved in Hepes buffer (15 mmol/L, pH 7.5).
BSA (fatty acid–free, >96% albumin), FeSO4·7H2O, L-cysteine, and reduced GSH were purchased from Sigma (Deisenhofen, Germany). Sodium nitroprusside (SNP) was obtained from Merck (Darmstadt, Germany). NO gas was synthesized by means of reaction of FeSO4·7H2O with NaNO2 in 0.1 mol/L HCl and was purified by means of low-temperature high-vacuum (P 1.3 Pa) distillation. Andrei L. Kleschyov, PhD, (Laboratoire de Pharmacologie Cellulaire et Moléculaire, Université Louis Pasteur, Strasbourg, France) kindly provided MGD.
Preparation of Samples for MR Imaging and ESR Spectroscopic Measurements
For the MR imaging experiments, plastic syringes (10-mm diameter) were filled with 2 mL of each solution, immediately frozen in liquid nitrogen, and stored at -70°C. For cryogenic ESR spectroscopic measurements, smaller (0.5-mL) samples were prepared as described previously (23).
Treatment of Rats
Rats were maintained in the department animal house in accordance with the guidelines of German law for animal welfare and as approved by the local authorities. Six adult male rats (strain Wistar-Kyoto, aged 3 months) were killed with an overdose of nembutal (180 mg/kg). The abdomen was quickly opened, and the hepatic portal vein was revealed. A cannula was tied into the hepatic portal vein, and 50 mL of 30 mmol/L SNP dissolved in saline was infused over 2 minutes. Well-perfused liver regions turned pale and were excised, placed into plastic syringes, and stored on ice until measurements were performed at MR imaging. Small samples from a well-perfused lobe of the liver were shock frozen for an assessment of DNIC formation at cryogenic ESR spectroscopy. Rat liver perfused with a saline solution without SNP served as a negative control.
ESR Spectroscopic Measurements
The concentration of paramagnetic nitrosyl-iron complexes was determined at ESR spectroscopy. Spectra were recorded at 77 K and at room temperature with a CW-ESR (EPR 300E; Bruker, Rheinstetten, Germany) spectrometer that operated at the X band (9.6 GHz). For measurements at 77 K, samples were placed into a quartz Dewar flask (5-mm inner diameter) filled with liquid nitrogen. For measurements at room temperature, 50 µL of the sample was placed into a quartz capillary tube (1-mm inner diameter). Instrument settings were the following: modulation frequency, 100 kHz; modulation amplitude, 0.5 mT (0.05 mT for DNIC-GSH at 298 K); time constant, 655 msec (2,621 msec for DNIC-GSH at 298 K); microwave power, 20 mW; and sweep time, 167 seconds (670 seconds for DNIC-GSH at 298 K).
The concentrations of individual solutions were determined by comparing the ESR spectra with those of concentration standards recorded with identical instrument settings.
MR Imaging
For the MR imaging experiments, the solutions containing the nitrosyl-iron complexes were thawed and thermally equilibrated for 20 minutes at room temperature. Thereafter, the tubes containing the nitrosyl-iron complexes were fixed on a flat cylindrical (diameter, 25 cm; height, 10 cm) Plexiglas MR imaging head phantom filled with a paramagnetic solution (1.25 g of NiSO4 and 5 g of NaCl in 1,000 g of H2O) and were positioned just in the middle of the coil. MR images were obtained at room temperature by using a 1.5-T Magnetom unit (Siemens Medical Systems, Erlangen, Germany) and a standard head coil.
Spin-lattice relaxation, or T1, times were determined by using a spin-echo sequence with an echo delay time and six different repetition times (TRs) msec/echo time (TE) msec (150–3,600/15). Spin-spin relaxation, or T2, times were determined by using a sequence with 16 echoes (2,000/50–800). Transverse single-section images (section thickness, 10 mm) through multiple tubes that contained solutions of different concentrations were acquired with four (T1) and seven (T2) phase-cycled acquisitions. The matrix was 256 x 256, with a 150 x 150 field of view.
Different T1-weighted (300–500/15) and T2-weighted (2,000–3,000/70–80) images of the SNP- and saline-perfused liver samples were obtained within 1 hour after perfusion for maximal contrast enhancement.
After the MR imaging experiments were finished, the solutions containing nitrosyl-iron complexes were frozen in liquid nitrogen, and their concentrations were again verified at cryogenic ESR spectroscopy. Changes in concentrations of nitrosyl-iron complexes were taken into account by using mean intermediate concentrations in the calculation of relaxivities.
Data Analysis
Mean intensity values were obtained in circular regions of interest that were identical to the cross-sectional area of the sample syringe in each image obtained with the different TRs and TEs. Both authors independently placed the regions of interest, which yielded similar results.
A simplified expression for the functional dependence of the MR imaging signal intensity S on the intrinsic properties (proton spin density N(H), T1, T2) has been derived. On the basis of the assumption of a monoexponential dependence on T1 and T2, the following equation is generally used with a spin-echo sequence (24): S(TE,TR) = N(H)exp(-TE/T2) x {1 - exp[-(TR - TE)/T1]}.
The term on the right reflects the T1 dependence, and the exponential expression on the left reflects the T2 effects. If a TE much smaller than the T2 is chosen, the T2-dependent term may be neglected, and the image is T1-weighted. T1 is obtained by means of least-squares fitting (Marquardt-Levenberg algorithm) of the mean intensity values versus TR. T2 is obtained by choosing a TR much larger than the T2. As a result, the term on the right may be neglected; the image is now T2-weighted, and the mean intensity values are fitted versus the TE (Marquardt Levenberg algorithm).
When a paramagnetic agent is added to an aqueous solution, its contribution to proton relaxation may be represented by the following equation (25): 1/T(1,2)o = 1/T(1,2)d + 1/T(1,2)p, where T(1,2)o is the observed relaxation time, T(1,2)d is the relaxation times of the diluent, and T(1,2)p is the paramagnetic contribution to T1 and T2 relaxation. Values for 1/T(1,2)p were calculated for each of the solutions by using the measured solvent relaxation times. Since the paramagnetic relaxation rates 1/T(1,2)p are expected to be directly proportional to the concentration of the paramagnetic agent, the following equation holds: 1/T(1,2)p = M(1,2) x [C], where [C] is the concentration, and r1 and r2 are the T1 and T2 relaxivities of the paramagnetic species, respectively. Relaxivity is a measure of the ability of a paramagnetic species to influence relaxation rates and is expressed in units of liters per millimole per second.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
= 2.043 and g|| = 2.01 in the frozen state. At room temperature, this ESR signal transformed into a narrow isotropic signal at g = 2.03 with 13-line hyperfine splitting (Fig 1, A).
|
= 2.035 and g|| = 2.02 in the frozen state and an isotropic triplet with a hyperfine splitting constant of 1.4 mT in liquid phase at g = 2.04 (Fig 1, C).
Relaxation Times and Relaxivities of MNIC and DNIC
Cross-sectional MR images of aqueous solutions with different concentrations of DNIC-GSH and DNIC-BSA are shown in Figure 2. All nitrosyl-iron complex solutions exhibited concentration-dependent differences in signal intensity at MR imaging that were influenced by the recording conditions. On T1-weighted images, the nitrosyl-iron complexes increased the signal intensity, whereas on T2-weighted images, they decreased the signal intensity. Representative graphs of signal intensity versus TR or TE and of the relaxation rates versus the concentration of DNIC-BSA are shown in Figure 3.
|
|
|
|
SNP-perfused Rat Liver
To induce formation of DNIC in a biologic tissue, isolated rat liver was briefly perfused at room temperature with a saline solution of the NO donor and antihypertensive agent SNP (1.5 mmol). Perfusion with SNP induced the formation of a typical DNIC ESR spectrum that was similar to that of DNIC-BSA (Fig 1, B) in the liver tissue. Since the anisotropic nature of the cryogenic EPR signal did not change at room temperature, this DNIC was bound to macromolecules (ie, liver proteins). The mean concentration of DNIC achieved with this procedure in liver tissue amounted to 310 µmol/L ± 110 (n = 3), whereas in the control livers (n = 3), DNIC was undetectable.
On the T1-weighted images (Fig 4), the signal intensity of the SNP-perfused tissue markedly increased by about 30% (from 562 msec ± 21 to 394 msec ± 84; n = 3) because of a reduction in the T1 compared with that of the control tissue. As shown in Figure 4, the signal intensity differences were artificially enhanced by means of linear extension of the gray scale. On T2-weighted images, the signal intensity of the liver that contained DNIC decreased because of a reduction in the T2 to about 80% of that of the control tissue (from 36 msec ± 4 to 28 msec ± 2; n = 3).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We found that phantom solutions of DNIC-cysteine, DNIC-GSH, DNIC-BSA, and MNIC-MGD affected the signal intensity on MR images in a concentration-dependent fashion; these effects produced a decrease in the T1s and T2s of MR-excited proton spins (Fig 3, A and B). Therefore, paramagnetic nitrosyl-iron complexes exhibit the classic properties of MR imaging contrast agents.
However, the potency for contrast enhancement differed considerably. MNIC-MGD and DNIC-BSA exhibited relaxivities that were severalfold higher than that those of DNICs with low-molecular-weight thiols (L-cysteine and GSH) (Table). This finding indicates that the electromagnetic coupling between the paramagnetic center and the solvent protons is much stronger in protein-bound DNIC and MNIC than in low-mass DNIC with cysteine and GSH ligands.
One explanation may be that the tetragonal geometry of the last two DNICs (Fig 5) imposes a high-energy barrier to the coordination of water to the central iron (26). In contrast, the slightly different coordination geometry of the other two complexes allows water molecules to penetrate into the coordination sphere. The molecular structure of MNIC is a square-based pyramid with NO at the tip, whereas DNIC-BSA is a flattened tetragon because of the bulky proteinacious ligand that distorts the ideal tetragonal coordination of the central iron. Since the signal intensity at MR imaging is related to the concentration of the contrast agent and to the product of the magnetic moments of the interacting species but is inversely proportional to r6 (where r is the distance between the paramagnetic center and the water protons), the ability of water to access the paramagnetic center critically determines the relaxivity. However, even if the water were totally prevented from having direct access to the center, relaxation can nevertheless occur by means of outer-sphere effects.
|
Using SNP-perfused rat liver as a model tissue, we were able to demonstrate that the concentration (300 µmol/L on average) of protein-bound DNIC generated in the tissue with this procedure was sufficient to induce a large change (increase on T1-weighted images, decrease on T2-weighted images) in the signal intensity on MR images compared with that of saline-perfused control liver (Fig 4). It can be estimated that, with a threefold lower concentration of DNIC (100 µmol/L), the difference in signal intensity will equal the error of signal intensity determination on T1-weighted images. This limit may be improved by selecting smaller tissue volumes of interest and by using appropriate pulse sequences. Thus, the detection limit for DNIC at MR imaging may approach the tissue concentrations of endogenously formed DNIC detected in infected rabbit liver (100 µmol/L) (22). But this limit is still an order of magnitude higher than that observed in tumor patients (4 µmol/L in affected liver tissue) (22).
In contrast to organic free-radical nitroxides, the DTCs used in the formation of MNIC were less toxic and, therefore, better suited for application in patients. For example, DTC has long been used as a deterrent to alcohol consumption (29) and is an investigational drug in the inhibition of human immunodeficiency viral progression (30). By choosing a DTC with an appropriate solubility, we can select the body compartments in which NO is trapped and detected. The hydrophilic iron-MGD complex and its nitrosyl complex mainly distribute into the blood and are excreted in the urine (31), while lipophilic iron-diethyldithiocarbamate and its NO-iron complex accumulate in organ tissues, apparently in cell membranes and other lipid structures (32). Another advantage of MNIC and DNIC is that endogenous iron is chelated and included in the contrast agent; thus, this process prevents the potentially harmful action of the free iron released from various intracellular sources due to excessive NO formation (33).
It is obvious that DTCs have to be applied to raise concentrations of nitrosyl-iron complexes in NO-exposed tissues of patients above the detection limit at MR imaging; only then will sufficient amounts of iron-DTCs be provided for efficient trapping of NO. Thus, the lower relaxivity of MNIC compared with that of DNIC-BSA may well be compensated by the high concentrations of MNIC achieved in the NO-exposed tissue. Data in patients are not available, but in septic mice treated with diethyldithiocarbamate, up to 460 µmol/L of MNIC were observed in the liver (34). In a recent study (35), MR imaging "spin-trapping" of NO with MGD was applied in rats treated with lipopolysaccharide to induce a sepsis-like syndrome with high levels of endogenous NO formation. MNIC-MGD accumulated in the liver, which displayed substantially enhanced signal intensity on T1-weighted MR images. The signal enhancement was eliminated by means of pharmacologic inhibition of NO synthesis and was reproduced with intraperitoneal injection of synthetic MNIC-MGD. Although this study demonstrated the feasibility of MR imaging of nitrosyl-iron complexes in animal models, further investigation of clinical applicability is clearly warranted.
Practical application: Isolated tissues exposed to high levels of NO can be imaged with 1H MR imaging techniques by using the contrast properties of paramagnetic NO-iron complexes after NO is trapped as a stable nitrosyl-iron complex. The contrast effect of nitrosyl-iron complexes can be exploited for in vivo NO imaging with conventional 1H MR imaging techniques to substantially broaden the application of MR imaging.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, A.M.; study concepts, A.M.; study design, A.M., B.F.; definition of intellectual content, A.M., B.F.; literature research, B.F.; experimental studies, B.F.; data acquisition and analysis, B.F.; statistical analysis, B.F.; manuscript preparation and editing, A.M., B.F.; manuscript review, A.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. M. Pastor, C. Planchamp, S. Pochon, V. Lorusso, X. Montet, J. Mayer, F. Terrier, and J.-P. Vallee Kinetics of Gadobenate Dimeglumine in Isolated Perfused Rat Liver: MR Imaging Evaluation Radiology, October 1, 2003; 229(1): 119 - 125. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
), 190 (
), 90 (
), and 0 (
) µmol/L. C, Graph of reciprocal relaxation times (T) versus concentration DNIC-BSA show the values of r1 (M1, 
) and r2 (M2, 
