Search

Iron oxide nanoparticles, with monocrystalline cores (generally magnetite or maghemite) and coated with organic polymer to increase chemical stability and solubility, have generated widespread interest as negative contrast (susceptibility) agents for magnetic resonance imaging (MRI) (1,2). For negative agents, the requisite T2 shortening depends on the geometry of the physiologic structures that take up the agent (2) but is relatively insensitive to the size and structure of the nanoparticles themselves. Nanoparticles also have a potential as positive agents for magnetic resonance angiography (MRA) (3–6), an application that depends on T1 shortening in the vasculature by individual nanoparticles, provided that T2 is not too short (7). To optimize their suitability for MRA, the agents must be synthesized and characterized reproducibly, the size of the magnetic core must be carefully controlled, and the interaction of the organic coating with nearby solvent molecules must be understood (8). Here we describe a set of physical techniques for measuring the parameters that determine MRA efficacy: the magnetic moment and diameter d of the iron oxide cores, and the interaction of organic coating with solvent. We study magnetization (8) as a function of magnetic field B0; l/Tl nuclear magnetic relaxation dispersion (NMRD) profiles (9) for B0 from 0.24 mT to 1.2 T (0.01–50 MHz) and 1/T2 at 0.47 T (20 MHz); and solute translational diffusivity by photon correlation spectroscopy (PCS) (10). As demonstrated earlier (8), magnetization data, being thermodynamic, yield and the saturation magnetization of the particles, Msat. Combined with total iron content and the known density of the core, one obtains d. When these results are combined with NMRD data—which depend on and d and yield certain correlation times—one can detect any water-opaque organic “rind” and measure the diffusivity D of outer sphere solvent, a measure of viscous interactions between solvent and polymer. These interactions can be quantitated by PCS, which yields the effective radius of gyration of the polymers (11), including contributions from both polymer and viscously entrained water molecules. This trio of measurements lets one characterize the physical properties of solute nanoparticles self-consistently, deduce their behavior in MRA applications, and guide the development of improved agents. We report on two preparations of coated iron oxide nanoparticles: (a) CLARISCAN, a Nycomed Amersham product in clinical trials (8); and (b) MION-46L, a laboratory preparation (11) of MION-46 obtained from Massachusetts General Hospital.

A preparation of monocrystalline iron oxide nanoparticles with an oxidized starch coating, currently in clinical trials (NC100150 Injection; CLARISCAN), was characterized by magnetization measurements, relaxometry, and photon correlation spectroscopy. By combining the results with a measure of iron content, one can obtain the size and magnetic attributes of the iron cores, including the relevant correlation times for outer sphere relaxation (tau(SO) and tau(D)), and information about the interaction of the organic coating with both core and solvent. The results are 6.43 nm for the iron oxide core diameter, a magnetic moment of 4.38x10(-17) erg/G, and a water-penetrable coating region of oxidized oligomeric starch fragments and entrained water molecules. The latter extends the hydrodynamic diameter to 11.9 nm and lowers the average diffusivity of solvent about 64% (which increases tau(D) accordingly). The nanoparticles show little size-polydispersity, evidenced by the lowest value of r(2)/r(1) at 20 MHz reported to date, an asset for magnetic resonance angiography.

RATIONALE AND OBJECTIVES: Large magnetic entities, with diameters in the range of 4 nm to 4 microns, are becoming of increasing interest for magnetic resonance imaging (MRI). The smaller are iron oxide nanoparticles, used for the RE system, and the larger are deoxygenated blood cells, for functional MRI. It can be useful to model such systems as magnetized solute spheres in water. Classical computations of 1/T2 have been reported for the larger particles, in the micron range, where the computational complexities are simplified by Monte Carlo methods. For smaller particles, the quantum mechanical (quantal) expressions for outer sphere relaxation, for both 1/T1 and 1/T2, have been available for some time, and are particularly simple to apply at MRI fields. The questions that arise, and which the author addresses, are how to interrelate the classical and quantal approaches and when to use which. METHODS: The author compares published results of Monte Carlo calculations of 1/T2 for diamagnetic polystyrene solute spheres of various sizes in water, made paramagnetic by addition of dysprosium-(DTPA)2-, with quantum mechanical outer sphere theory applied to the same system. The latter includes the usual assumption of motional narrowing and yields both 1/T1 and 1/T2. RESULTS: For particles with diameters less than about 1 micron, both approaches give identical results for 1/T2. For larger particles, the conditions for motional narrowing breakdown, and quantal theory overestimates 1/T2. In addition, in the particular system studied, relaxation becomes so effective near solute that there is insufficient time for all water molecules to experience their maximal effect. Classical theory handles this well whereas quantal theory does not. CONCLUSIONS: In comparing the classical and quantal approaches, one balances computational complexity but broader applicability with more limited but far simpler mathematics. In addition, because the quantal approach shows that 1/T1 and 1/T2 are intimately related, the author suggests, by analogy, how to extend classical methods to computation of 1/T1.

Blood-pool contrast agents in current use in clinical MRI are, in the main, small, highly stable, Gd3+-chelate complexes with a single inner-coordinated water molecule on the metal ion. Unfortunately, their relaxivities, and hence efficacy, diminish as the strength of the imaging field B 0 increases above -0.1 T, an effect associated with their rapid thermal rotation in solution (1). Restricting this rotation, for example, by rigid attachment of the small paramagnetic moiety to a macromolecule, can increase the relaxivity 10-fold in the 0.5-1.5-T range. The underlying physical mechanisms have been understood for several decades (2), and a quantitative comparison of the relaxivities expected for Gd(DTPA) 2and Gd(DOTA) 1bound to serum albumin, from theory, has been given in an earlier CMR symposium (3). There is, in addition, a range of relatively early experimental results that demonstrate the predicted enhancements, including measurements of solutions of binary complexes of Gd 3+ ions with macromolecules and ternary complexes that involve ligands such as DTPA. But what is of particular interest is the variety of approaches being pursued currently: covalent complexes of Gd(DTPA) 2with serum albumin (4); hydrophobic association of Gd(DTPA) 2derivatives with albumin (5-7); polymers of Gd(DTPA)2--containing monomers (8); attachment of Gd(DTPA) 2--, by a lipid tail, to phospholipid

The magnetic field dependence (NMRD profile) of 1/T1 of solvent protons in an aqueous solution of Gd(DTPA)2- was remeasured at 5, 15, 25, 30, and 35 degrees C. The data were reanalyzed with the usual low-field theory, using recently published values for tauM, the residence lifetime of the single inner-coordinated waters of solute Gd(DTPA)2-. (These tauM values are significantly longer than earlier estimates). Values were obtained for three dynamic parameters: tauR, the rotational relaxation time of solute ions, and tauSo and tauV, the low-field relaxation time of the Gd3+ magnetic moment and the related correlation time. These Gd(DTPA)2- values, together with recent results for tauM for Gd(DTPA-BMA)--a nonionic structural analog of Gd(DTPA)2- with an unusually long tauM--were used to calculate NMRD profiles at 5 and 35 degrees C. These profiles agree very well with new data given here for a solution of Gd(DTPA-BMA). This reaffirms the importance of knowing the temperature-dependent values of tauM a priori in order to obtain unambiguous quantitative theoretical analyses of NMRD profiles of chelates of known structure. Additionally, the theory of inner sphere relaxation is extended to high fields, at which the magnetic energy of a solute moment is greater than its thermal energy.

X-ray absorption and EPR spectroscopy have been used to probe the copper site structure in galactose oxidase at pH 4.5 and 7.0. the results suggest that there are no major differences in the structure of the tetragonal Cu(II) site at these pH values. Analysis of the extended X-ray absorption fine structure (EXAFS) indicates that four N,O scatterers are present at approximately 2 {Angstrom}; these are presumably the equatorial ligands. In addition, the EXAFS data establish that oxidative activation to produce the active-site tyrosine radical does not cause major changes in the copper coordination environment. Therefore results obtained on the one-electron reduced enzyme, containing Cu(II) but not the tyrosine radical, probably also apply to the catalytically active Cu(II)/tyrosine radical state. Solvent water exchange, inhibitor binding, and substrate binding have been probed via nuclear magnetic relaxation dispersion (NMRD) measurements. The NMRD profile of galactose oxidase is quantitatively consistent with the rapid exchange of a single, equatorial water ligand with a Cu(II)-O separation of about 2.4 {Angstrom}. Azide and cyanide displace this coordinated water. The binding of azide and the substrate dihydroxyacetone produce very similar effects on the NMRD profile of galactose oxidase, indicating that substrates also bind to the active site Cu(II)more » in an equatorial position.« less

RATIONALE AND OBJECTIVES. Water content and waterproton relaxation rates are reported for fresh, histologically characterized, surgical specimens of calcified human intracranial meningiomas and compared with results for noncalcified meningiomas from an earlier study and with calcium hydroxyapatite (CaHA) suspensions to elucidate the influence of calcification on magnetic resonance imaging (MRI) signal intensity of calcified meningiomas. METHODS. The magnetic field dependence of 1/T1 of water protons (nuclear magnetic relaxation dispersion profile) and dry weights are reported for 38 calcified nonhemorrhagic and 3 hemorrhagic specimens of known histologic subtype, a subset of the 67 specimens measured earlier. Calcification was considered mild or heavy when the dry weight was within or above the range for noncalcified meningiomas. Preliminary 1/T1 profiles for pure Call A and a single high-field 1 /T2 value also are reported. RESULTS. The ranges of dry weights and of low-field 1/T1 values were twice as large for calcified as for noncalcified meningiomas. No correlation was found between low-field 1/T1 and either histologic subtype or dry weight. Mild calcification produced the highest low-field 1/T1 values; the most heavily calcified tumor had slightly increased low-field 1/T1. Calcium hydroxyapatite increases low-field 1/T1 significantly but not high-field 1/T1; high-field 1/T2 is large. For calcified hemorrhagic meningiomas, increases in both low-field and high-field 1/T1 were seen. CONCLUSION. For mild calcification, MRI signal voids result from an increased high-field 1/T2; for heavier calcification, reduced proton density (from excluded water) becomes of increasing importance. Cellular CaHA appears to brighten the signal in T1-weighted MRI in the presence of hemorrhage.

RATIONALE AND OBJECTIVES Resected meningiomas were examined by relaxometry and light microscopy to evaluate the potential of magnetic resonance imaging (MRI) for identifying histologic subtypes and for discriminating among benign, radiation therapy-induced, and malignant meningiomas. METHODS The magnetic field dependence of 1/T1 of water protons (nuclear magnetic relaxation dispersion [NMRD] profile) and the water content (dry weight) were measured for 67 specimens, and the data were compared with histology. Only noncalcified, nonhemorrhagic meningiomas are reported. RESULTS No correlations were found between NMRD profiles, dry weight, and any histologic subtype, in contrast to an analogous study of astrocytomas. Rather, meningiomas have a broader variability of dry weight and 1/T1 than related parenchyma but a much narrower range than all grades of astrocytomas. The mean value of 1/T1, at all fields, is slightly higher in meningiomas--and the mean water content about the same--as adult cortical gray matter. CONCLUSION Meningiomas are frequently isointense with cortex, and histologic subtypes cannot be differentiated at any magnetic field strength by MRI using only T1- or proton density-weighted MRI.

This work presents the anisotropic pseudorotation (APR) model, which is a new dynamic model for the interpretation of experimental nuclear spin‐lattice relaxation times in paramagnetic (S=1) complexes of low symmetry. It comprises two dynamic processes active in modulating the zero‐field splitting interaction (ZFS). Reorientation of the complex modulates a static zero‐field splitting, defined as a measure of the asymmetry in the equilibrium geometry at the paramagnetic site. Local motions of the ligands surrounding the paramagnetic site further contribute a rapidly fluctuating (transient) zero‐field splitting interaction. This dynamic model is evaluated within a general theoretical framework capable of dealing with the electron‐spin system in the low‐ and high‐magnetic field limits for both Redfield and slow‐motion cases, i.e., where the motions inducing electron‐spin relaxation and the electron‐spin relaxation itself are characterized by the same time scale. The dynamic model is characterized and discuss...

Homogeneous soft tissue, as regards its magnetic relaxation properties, is well-modeled by solutions of cross-linked protein (see Koenig and Brown, Prog. NMR Spectr. 22, 487 (1991)). Interactions at the solute-solvent interface alter the hydrodynamics of solvent water, and also couple the solute and solvent proton Zeeman energy reservoirs, giving hydrodynamic and cross-relaxation contributions to water proton relaxation that respond differently to deuteration of solvent. We report measurements of the magnetic field dependence of 1/T1 of water protons in cross-linked bovine serum albumin (BSA), for partially deuterated solvent and, in order to separate these two contributions, of 1/T1 of deuterons. The major experimental finding is that, in addition to recently identified water-binding sites on protein (covering approximately 0.2% of the surface) with water lifetimes of about 1 microsecond, there is another group of sites with lifetimes of about 23 ns, covering approximately 2% of the surface, which are evident in both proton and deuteron data. In addition, we have formulated a theory of interfacial proton-proton magnetic interactions which--with these four parameters, plus two that quantify the protein-water coupling at each site--can account for all the proton and deuteron data, in both native and cross-linked BSA.

The magnetic field dependence (dispersion profile) of 1/ T 1 of solvent water protons of solutions of the mammalian eye lens protein α-crystallin has a reversible nonlinear dependence on concentration which, as protein concentration increases above ∼20 wt.%, changes rapidly from a profile characteristic of mobile protein solute to a profile characteristic of rotationally immobilized protein (e.g., chemically cross-linked). From quantitative comparisons of new measurements of K at 200.1 MHz, the rate of water-to-protein interfacial magnetization transfer, with earlier data for α-crystallin and bovine serum albumin (BSA), we conclude, for α-crystallin, that: (i) Brownian rotation is slowed by intermolecular interactions at unexpectedly low concentrations; (ii) K is mediated by the newly reported, long-lived, protein hydration sites with the same surface density as BSA; and (iii) 14 N peaks seen in solvent 1/ T 1 arise from interfacial magnetization transfer plus diffusion to protein NH magnetization sinks.

We report results for proton 1/T1, 1/T2, and K, the rate of magnetization transfer from solvent to solute, for 5 and 10 wt. % solutions of bovine serum albumin, both native and chemically cross-linked, in undeuterated and approximately 50% deuterated water, at 4.7 T (200.1 MHz) and 19 degrees C. At this field, although K > 1/T1 for the cross-linked samples, magnetization transfer contributes little to 1/T1 directly. Therefore K was measured using off-resonance irradiation of the protein protons. The data for all the samples can be fit using a theoretical model for magnetization transfer, with three parameters: the intrinsic longitudinal relaxation rates of solute and solvent protons, and K. The magnitude of K is so large that the newly-identified, long-lived (approximately 1 microseconds) hydration sites (S.H. Koenig, R.D. Brown III, and R. Ugolini, Magn. Reson. Med., 29, 77 (1993)) must be invoked to account for K, as is necessary to explain the differential effects of cross linking on the magnetic field dependence of 1/T1 of protons and deuterons and the large 1/T1 and 1/T2 values below approximately 20 MHz in immobilized systems. Although these sites are few in number, their long resident lifetime becomes the correlation time for magnetization transfer when protein is immobilized, accounting for the large value of K. Recent data from several laboratories have shown that cross-linked protein, as used here, is a good model for 1/T1 and 1/T2 of tissue, as a function of temperature and magnetic field.

We argue that, for solutions of immobilized protein and for tissue, the dependence of 1 / T1 of solvent protons on B0 at low fields and 1 / T1p on B1 for all B0 are both manifestations of the same underlying phenomena: magnetization transfer between mobile water protons and solid-state broadened protein proton levels. Broadening causes rapid mixing of spin orientation within the transverse plane, at all B0, unless B1 is greater than the protein internal field; this affects 1 / T1p of solvent protons by magnetization transfer. Similarly, decreasing B0 below the internal field mixes all orientations of magnetization, which affects the solvent proton low-field 1 / T1 and high-field 1 / T2.© 1992 Academic Press,Inc.

Covalent attachment of methoxypoly(ethylene glycol) (MPEG) 5000 to the surface of unilamellar liposomes composed of egg phosphatidylcholine and dioleoylphosphatidylethanolamine (DOPE) (8:2) containing paramagnetic chelates, either entrapped within the interior volume of the liposomes, or associated with the membrane surface, had no effect upon the measured spin-lattice relaxation rates (1/T1) for water in these systems. 31P-NMR studies indicate no destabilization of dioleoylphosphatidylcholine (DOPC)/(DOPE) (1:1) vesicles following attachment of MPEG. However, in DOPC/DOPE (1:3) mixtures, covalent modification with MPEG results in a destabilization of multilamellar vesicles into smaller vesicular structures. These results indicate that covalent attachment of poly(ethylene glycol) to liposomal magnetic resonance agents may prove a useful method for increasing their utility as vascular MR agents by extending their lifetime in the circulation, without decreasing the relaxivity of paramagnetic species associated with the liposome, but that the presence of PEG covalently attached to the membrane surface may modify the polymorphic phase behavior of the lipid system to which it is covalently linked.

The 1/T1 NMRD profiles of lipid vesicles with the paramagnetic ion Gd attached via a chelate covalently linked to the membrane surface show a peak at approximately 20 MHz indicating that fluctuations of approximately 10(-8) s contribute to the form of the dispersion profile. If the correlation time for fluctuations of the paramagnetic chelate on the membrane surface is much less than the correlation time for rotation of the lipid vesicle, it would be expected that the measured 1/T1 relaxation rate for solvent protons should be invariant with vesicle size above a certain minimum vesicle diameter. We show that this is indeed the case for vesicles in the size range 50 to 400 nm average diameter and discuss general design considerations for the preparation of vesicle-associated MR contrast agents based upon paramagnetic chelates either trapped within the vesicle interior or attached to the membrane surface.

The diffusive permeability to water molecules, Pd, of lipid vesicles with entrapped paramagnetic solute ions can be determined rapidly from analysis of the magnetic field dependence (nuclear magnetic relaxation dispersion, or NMRD profile) of T1 of exterior solvent water protons. Such data yield tau, the mean lifetime of solvent molecules inside the vesicles, from tau = (fT1Para) - T1Ves, where f is the volume fraction of entrapped water, T1Para is the observed T1 corrected for buffer background, and T1Ves is the relaxation time of water protons in the entrapped solution. For small spherical unilamellar vesicles of inner radius R, Pd = R/3 tau, f can be obtained accurately from knowledge of both the concentration of Gd(DTPA)2- in the solution in which the vesicles were formed and the average concentration of ions in the final sample. At low temperatures, in the limit of slow exchange, T1Para becomes independent of field and tau = fT1Para; the observation of a field-independent profile is a control that confirms that no paramagnetic material is external to the vesicles. We have measured T1Para, using a field-cycling relaxometer, for suspensions of POPC (1-palmitoyl-2-oleoyl-lecithin) vesicles with 100-500 mM entrapped Gd(DTPA)2- and membrane concentrations of cholesterol ranging from 0 to 40 mol %. These profiles, which span the field range 0.01-50 MHz proton Larmor frequency, were taken at 5, 15, 25, and 35 degrees C. Concentrations of Gd(DTPA)2- were determined independently by both ICP analyses and NMRD methods. Values for Pd for vesicles with 100 mM Gd(DTPA)2- and outer diameters 100 nm +/- 20%, as determined by quasielastic light scattering, are 63, 47, 24, 16, and 8.7 x 10(-4) cm s-1, at 25 degrees C, for cholesterol concentrations of 0, 10, 20, 30, and 40%, respectively. The corresponding activation enthalpies are 14, 14, 14, 17, and 17 kcal/M. Comparison with 2H NMR studies of deuterated POPC vesicles with no cholesterol at 20 degrees C, and with 10% at 40 degrees C, which yielded the same order parameter for the palmitoyl acyl chains, gives no indication of a correlation between order parameter and permeability.

A non-cyclic tetraazatricarboxylate ligand, tris(4-carboxy-3-methyl-3-azabutyl)amine (H3L), derived from tris(2-aminoethyl)amine, was synthesised and purified. Its macroscopic and microscopic protonation behaviour was studied using potentiometry and proton NMR titrations, respectively. Stability constants for its complexes with Ni2+, Cu2+, Zn2+, Cd2+ and Gd3+ were obtained potentiometrically and compared with those of related ligands. The structure, protonation and dynamics of the complexes of H3L with Zn2+ and La3+ were investigated by proton NMR spectroscopy as a function of pH and the results were in accord with the potentiometric data for the complexes of Zn2+ and Gd3+. A nuclear magnetic relaxation dispersion study compared the water-proton relaxation parameters of [GdL] and [Gd(dtpa)]2–(H5dtpa =N,N,N′,N″,N″-diethylenetriaminepentaacetic acid), concluding that their water co-ordination numbers are the same but the former complex has a shorter electron-spin relaxation time.

We resolve the observed magnetic field-dependent relaxivity of Gd(DTPA)2- into inner and outer sphere contributions by comparison with theory. After noting that the outer sphere part, which contributes comparable to both 1/T1 and 1/T2, is identical to the "susceptibility" effect, we demonstrate the transition from the outer sphere relaxivity of small complexes, e.g., Gd(DTPA)2-, to that of large magnetic particulates, e.g., magnetite, used as T2 agents. This transition is expressed as a shift from a quantum to a classical mechanical description of relaxivity. We next show how to extrapolate the inner sphere contribution of Gd(DTPA)2- to similar small agents with greater structural symmetry, e.g., Gd(NOTA)-, by lengthening the electronic relaxation time and to ternary macromolecular complexes of small agents with protein by lengthening the rotational relaxation time. Finally, we indicate how relatively large lipid vesicles containing either small paramagnetic complexes or magnetized particulates may be considered analogs of Gd(DTPA)2-, with analogous inner and outer sphere contributions.

The sections in this article are 1 Introduction: Rabi and NMR in Atomic and Molecular Beams 2 Bloch, Purcell, and NMR in Condensed Matter 3 ESR in Beams, NMR in Water, and the Anomalous Electron g-Value 4 Relaxometry and Relaxometers 5 Biogrpahical Sketch

The sections in this article are 1 Introduction 2 Proton 1/T1 NMRD Profiles of Tissue and Model Systems 3 Proton and Deuteron NMRD Profiles in Model Systems with Deuterated Solvent 4 Theory 5 Data and Theory 6 Cytoplasmic Ordering and Cytoplasmic Chaos 7 Perspectives 8 Biographical Sketches Related Articles

The sections in this article are 1 Introduction to Relaxometry 2 Protein–Water Interfaces 3 Lipid–Water Interactions 4 Quantification of the Dynamics of Water on the Molecular Scale: Results from Magnetic Solutes 5 Water as a Viscous Vacuum 6 Concluding Remarks 7 Biographical Sketch Related Articles

Organically coated magnetic monocrystalline iron oxide nanoparticles are being considered as contrast agents for T1-weighted magnetic resonance imaging (MRI) and magnetic resonance angiography (MRA) (1–4). For such Tl-weighted applications, imaging efficacy is strongly dependent on the physical characteristics of the individual nanoparticles (5). In particular, three stringent physical requirements must be satisfied. First, the magnetic cores must be of an optimal size with essentially a monodisperse size distribution. If the cores are too small, rl (the Tl relaxivity) will be too small for practical applications as T1 agents; if the cores are too large, r2 (the T2 relaxivity) may be so large relative to rl that the T1 efficacy of the particles will be diminished. Second, the high-field magnetization of the cores, typically close to its saturation value, must be sufficient to relax water protons effectively, also implying that the organic coating must not compromise access of solvent to the core. This high-field limit is related to both the total iron content of the cores and their geometry. Third, the nanoparticles cannot become agglomerated in vivo. Given that r2 is very sensitive to agglomeration and rl is not, agglomeration preferentially increases r2, thereby diminishing the efficacy of the nanoparticles as Tl agents. In the current work, the relative physical characteristics of two distinct nanoparticle preparations are compared with respect to their efficacy for Tl-weighted MRI and MRA: MION-46L and NC100150 Injection. 1/Tl nuclear magnetic relaxation dispersion (NMRD) profiles suffice for such a comparison, and magnetization data (as discussed by Koenig et al [6] in this supplement) are not needed.

Nuclear Magnetic Relaxation field-cycling relaxometry is a technique, able to report on water mobility in tissues. By means of this technique, post-mortem specimens from both controls and idiopathic Parkinson's disease patients have been investigated. Results show different relaxometric behavior between the groups, which is consistent with protein aggregation in Parkinson's disease specimens.

A laboratory-scale synthesis of NC100150 (iron oxide particles with an oxidized starch coating) was characterized by magnetization measurements (vibrating sample magnetometry, VSM), relaxometry (1/T 1 NMRD profiles and 1/T 2 at 10 and 20 MHz), and dynamic light scattering (photon correlation spectroscopy, PCS). The results were related to give a self-consistent physical description of the particles: a water-impenetrable part making up 12% of the total particle volume, 82% of this volume consisting of an iron oxide core and the remaining 18% consisting of an oxidized starch rind; and, a water-penetrable part making up 88% of the total particle volume, consisting of oxidized starch polymers and entrained water molecules. Relating the magnetization to the relaxometry results required that the oxidized starch coating slows the diffusivity of solvent water molecules in the vicinity of the iron oxide cores. The effect of the organic coating on water diffusivity, not previously considered in the application of relaxation theory to iron oxide nanoparticles, is supported by the much greater (factor of about 2) diameter obtained from the dynamic light scattering measurements in comparison to that obtained from the magnetization measurements. The present work shows that three physical techniques—VSM, relaxometry, and PCS—are needed for properly assessing iron oxide nanoparticles for use as contrast agents for magnetic resonance angiography (MRA). It is also shown that NC100150 has a narrow range of diameters and the smallest value ofr 2/r 1 reported to date, an asset for MRA.

RATIONALE AND OBJECTIVES: Many magnetic resonance imaging (MRI) agents are Gd(III)-based; its half-filled f-shell has an S-ground state and hence a long electronic relaxation time, leading to comparably large effects on 1/T1 and 1/T2 of water protons with no shift in the water-proton resonance frequency. 1/T1 and 1/T2 nuclear magnetic relaxation dispersion (NMRD) profiles of the Dy(III) aquo ion and its chelates have been reported recently. Dy(III) ions differ magnetically from Gd(III); the large spin-orbit interaction of its non-S-ground state reduces the electronic relaxation time 100-fold, and can have a large effect on proton 1/T2 and resonance frequency. Relaxation theory is well-developed and applicable to both ions but, for Dy(III), the phenomena are more wide-ranging. Recent interpretations have suggested that the data are anomolous, requiring a new mechanism for their explanation. The authors explain published Dy(III) data in terms of known theory, guided by experience with Gd(III) agents. METHODS: For fields below 1 T, the authors incorporate the shortened electronic relaxation time into the usual low-field theory for magnetic dipolar interactions between water protons and Dy(III) magnetic moments. Both inner- and outer-sphere relaxations are included. At higher fields (and unusual for small single-ion agents) one must include dipolar interactions of protons with the magnetization of the Dy(III) moments. This "Curie magnetization" causes a quadratic dependence of 1/T1 on field, and--through dipolar-induced shifts--an even greater quadratic dependence of 1/T2. RESULTS: All published data can be explained by magnetic dipolar interactions. For Dy(III), the Curie term has a longer correlation time than the low-field term, namely, the rotation of solute for 1/T1 and the even longer water exchange lifetime tau M for 1/T2. This exchange modulates the shift, producing phenomena not seen with Gd(III). CONCLUSIONS: Relaxation by Dy(III) chelates can be explained by the same well-established theory of dipolar interactions used for their Gd(III) analogs. Interestingly, for MRI applications, tau M should be long for Dy(III)-based agents and short for Gd(III)-based agents.

A series of linear copolymers of DTPA-class Gd3+ conjugates, linked by alpha, omega-alkyldiamides with a varying number (n) of methylenes separating the amide function, were synthesized. Surprisingly, their relaxivities at all fields increased with increasing n. At MRI fields and 35 degrees C, the relaxivities of the n = 10 and n = 12 polymers were unexpectedly high, similar to those of rigid dendrimer-based Gd3+ chelates. The magnetic field dependence of solvent proton 1/T1 was measured for aqueous urea-free and urea-containing polymer solutions. The results for urea-free solutions imply an increase of rigidity (required for high relaxivities) with increasing n, arising from hydrophobic interactions of the methylene groups with solvent. This hypothesis is supported by a large decrease in the relaxivities upon addition of urea, which is known to weaken hydrophobic interactions. The relaxivities are also independent of polymer concentration, indicating that the hydrophobic interactions are intramolecular.

The magnetic relaxation rate 1/T1 of tissue water protons was measured over a wide range of magnetic field strengths (NMRD profile) for 92 fresh surgical specimens of astrocytomas to search for correlations of 1/T1 with tumor histology, as determined by light microscopy, and to assess the diagnostic potential of NMRD profiles for grading astrocytomas. A third goal was to elucidate the molecular determinants of 1/T1. Each specimen was histologically graded and inspected for evidence of mineral deposits (Ca, Fe); its dry weight was determined and expressed in % of original wet weight. To minimize variability not directly related to tumor grade, this initial report is limited to NMRD profiles of 47 non-calcified, non-hemorrhagic, untreated astrocytomas. For these, the mean value of 1/T1 at very low magnetic field strengths was found to increase with increasing grade of malignancy; no clear correlation could be demonstrated at high fields where most imaging is done. The spread of 1/T1 for different grades of malignancy is large, however, and the overlap significant, even at the lowest field, so that astrocytomas can not be graded by NMRD profiles alone. Average 1/T1 and average dry weight increase with grade of malignancy; but the variability of 1/T1 among specimens of the same dry weight is large, indicating that at least one other cellular parameter, not variable in normal tissue, influences 1/T1 strongly. We hypothesize that this parameter reflects changes at the molecular level in size distribution, mobility, or intermolecular interaction of cytoplasmic proteins. Which specific changes are induced by malignant transformation in astrocytomas remains to be investigated.

Publisher Summary This chapter discusses relaxometry and the source of contrast in magnetic resonance imaging (MRI). MRI has had great clinical influence in imaging the head, with adult white matter appearing much brighter than the gray matter under the usual imaging protocols used for the brain—the gray–white contrast is as great as that between any two soft tissues of the body. Approximately one-half of clinical MRI involves the administration of contrast agents. The majority of agents in use are blood-pool agents. There are additional possibilities, including lipid-specific agents that attach to myelin and other membranes, agents for in vivo spectroscopy, agents that alter their nuclear magnetic relaxation dispersion (NMRD) profiles on binding or when metabolized, agents suited to the gut, and tumor-specific agents. In all cases, it is anticipated that in vitro NMRD studies would be the first step toward understanding and predicting the behavior of these agents in the MRI context.

Protein in water solution increases magnetic relaxation rates of solvent nuclei to an extent that depends on magnetic field strength and molecular weight. Koenig and Schillinger (J. Biol. Chem. 244, 3283 (1969)) showed that a small fraction of the water molecules in the first hydration shell, bound irrotationally with a residence lifetime in the range 0.1 to 10 microseconds, would account for the phenomena. No experiments, as yet, have proven the existence of such long-lived waters, nor yielded a value for their lifetime. Analogous measurements on solutions of both denatured and cross-linked protein give data different from that of native protein, but much like results for tissue. By comparing proton and deuteron relaxation rates in solutions of native and cross-linked protein, it is possible to demonstrate the existence of these relatively long-lived waters; the data indicate that 1% of a monolayer of the waters of hydration of protein have lifetimes that cluster near 1 microsecond and, it is argued, are held in place by multiple hydrogen bonds. Assigning shorter lifetimes for waters held by fewer bonds, it is possible to develop a unified view of relaxation of water nuclei in protein solutions and in tissue, and to relate it to recent crystallographic data on hydrated protein.

From analyses of the magnetic field dependence of 1/T1 (NMRD profiles) of water protons in solutions of calf lens alpha-crystallin at several concentrations, we find two regimes of solute behavior in both cortical and nuclear preparations. Below approximately 15% vol/vol protein concentration, the solute molecules appear as compact globular proteins of approximately 1,350 (cortical) and approximately 1,700 (nuclear) kD. At higher concentrations, the effective solute particle size increases, reversibly, as evidenced by the appearance of spectra-like 14N peaks in the NMRD profiles and a change in the field and temperature dependence of 1/T1. At these higher concentrations, the profiles are very similar to those of calf gamma II-crystallin, a crystallin that undergoes an analogous transition near approximately 15% protein (Koenig, S. H., C.F. Beaulieu, R. D. Brown III, and M. Spiller, 1990. Biophys. J. 57:461–469). By comparison with recent analyses of NMRD results for solutions of immobilized proteins as models for the transition from protein solutions to tissue (Koenig, S. H., and R. D. Brown III. 1991. Prog. NMR Spectr. 22:487–567), we argue that alpha-crystallin solute behaves as aggregates approximately greater than 50,000 kD as protein concentration is progressively increased above 15%. Finally, the concentration dependence of the NMRD profiles of alpha- and gamma II-crystallin can readily explain recent osmotic pressure data, in particular the intersection of the respective pressure curves at approximately 23% vol/vol (Vérétout, F., and A. Tardieu. 1989. Eur. Biophys. J. 17:61–68).

The relative brightness of adult white matter in T1-weighted MRI arises from myelin, but the mechanisms responsible remain to be clarified. Koenig et al. [Magn. Reson. Med.14, 482 (1990)] conjectured that the cholesterol of myelin (∼ 30% of its lipid) was responsible. We present 1/T1 and magnetization transfer contrast imaging data [Wolff and Balaban, Magn. Reson. Med.10, 135 (1989)] on a model system— 50% lipid—50% water by weight, with the lipid one-half phosphatidyl cholinc (PC) and one-half cholesterol—and a control in which the lipid is all PC. The differences between the model and control samples mimic the myelin contribution to white matter in both experiments. © 1991 Academic Press, Inc.

The remarkable technological advances in magnetic resonance imaging (MRI), combined with the increasing utility of MRI as a clinical diagnostic modality, have generated a renewed interest in paramagnetic ions and their macromolecular complexes in vivo. The connection should be rather clear: contrast in MRI depends in the main on the relaxation rates of the protons of mobile water molecules of tissue, and these rates can be altered significantly by the introduction of tracer amounts of paramagnetic centers. Indeed, Bloch and collaborators, in their discovery of proton magnetic resonance in liquids over four decades ago (Bloch, 1946; Bloch et al., 1946), added Fe3+ ions to water to increase the proton relaxation rate to a convenient value.

The authors have measured the magnetic field dependence of 1/T1 (nuclear magnetic relaxation dispersion, or NMRD profiles) of water protons of histologically characterized samples of astrocytomas, meningiomas, and lymphomas. The goal was to elucidate the determinants of 1/T1 of brain tumors at the cellular level and, in particular, to search for a possible correlation of the profiles with neoplastic properties, including degree of malignancy. Because of the recently demonstrated contribution of myelin to 1/T1 of white matter, careful histologic analyses were performed to correct for its presence. The range of magnitude of the profiles of differing types and grades of tumors correlates with the range of water content of these tumors; the correlation of water content with cellularity (density of cell nuclei in a histologic preparation), in turn, produces correlations of 1/T1 with tumor type. For all the tumors studied, 1/T1 is proportional to solids content; the constant of proportionality is relatively insensitive to tumor type and, for astrocytomas, grade of malignancy; and is about the same as that of normal gray matter. For low- and intermediate-grade astrocytomas that contain myelin, the myelin-specific contribution to 1/T1 has to be considered to make manifest the underlying correlations, which are best demonstrated at low fields, where the background contribution of water and dissolved oxygen is minimal. At high fields, where most imaging is done, a change in oxygen partial pressure, as for example from ischemia in very malignant tumors, is sufficient to alter 1/T1 significantly, reducing the intrinsic correlation between histology and 1/T1.

The remarkable success of magnetic resonance imaging of adult brain relates to the unusually large ratio of the longitudinal relaxation rates 1/T1 of white and gray matter, approximately 2:1 at physiological temperature and traditional imaging fields. Several investigators have conjectured that myelin is the source of the greater 1/T1 of white matter without, however, suggesting details of the molecular mechanisms responsible. From measurements of the magnetic field dependence of 1/T1 (NMRD profiles) of adult and neonatal gray and white matter at 5 and 35 degrees C, we find a thermally activated contribution to the NMRD profile of adult white matter that is not present in the profiles of either adult gray or neonatal gray and white matter. We attribute this contribution to myelin and develop a quantitative model that accounts for the unique relaxation behavior of myelinated white matter. We find that myelin water, 15% of the total, has a relatively short T1 that arises from an unexpectedly large interaction with myelin lipid; when cast in terms of an interaction over the entire myelin bilipid-water interface, it is sevenfold greater than the analogous protein-water interfacial interaction. Its magnitude remains to be accounted for, but cholesterol, known to alter the relaxation rates of lipid protons, may play an important role. The contribution of myelin to 1/T1 at physiological temperatures is attributed to thermally activated transmembrane diffusion of water and, hence, more rapid mixing of axonal and the rapidly relaxing myelin water molecules.

Nitroxide stable free radicals (nitroxides) have potential utility as MRI contrast-enhancing agents with the additional capability of reflecting redox metabolism. In order to gain a better understanding of their potential interactions in vivo, we have studied the longitudinal NMRD profiles (1 / T1 as a function of field strength) and ESR spectra for lipophilic and aqueous-soluble nitroxides in blood, plasma, and plasma components. Typical water-soluble nitroxides do not interact appreciably with blood, plasma, or plasma proteins. Fatty acid nitroxides do interact physically with blood, predominantly by intercalation within red blood cell membranes and binding to albumin. The latter interaction results in significantly enhanced relaxivity for the nitroxide/HSA complex. Relaxation of water protons in this case is dominated by inner sphere processes, ostensibly due to water molecules hydrogen bonded to nitroxide moieties. The rotational reorientation time for the complex, the electronic relaxation time, and the exchange time for the water molecule reversably bound to the nitroxide, all appear significantly to influence the correlation time (∼ 16 ns) for this inner sphere contribution. Inc. © 1990 Academic Press, Inc.

Despite the demonstrated utility of measurements of the magnetic field dependence of the magnetic relaxation rates of solvent protons in solutions of metalloproteins as an indicator of biochemical changes, it is becoming increasingly evident that quantitative comparisons of such data with the theory of relaxation, limited by the approximations and assumptions usually made, yield results for the strength of the solvent-paramagnetic ion interaction that generally do not make chemical sense. These results, when expressed as the number of solvent-donated ligands of the ions, usually give too large a value, typically by about twofold. It has been suggested by several investigators that a comparison of proton and deuteron relaxation rates could resolve the problem. Data are presented for the longitudinal relaxation rates of solvent protons and deuterons over more than four decades of magnetic field (from 0.01 to 270 MHz proton Larmor frequency) for solutions of Mn2+-concanavalin A, a protein for which the physical biochemistry is thoroughly documented, one that should be particularly tractable for such comparisons. The main conclusion is that, in the general case, there is no decade of magnetic field over which the mathematical criterion of best agreement of data with theory can be relied upon to yield quantitatively correct biochemical results; rather, biochemistry must still be a guide for elucidating relaxation pathways.

Earlier studies of the magnetic field dependence of the nuclear spin magnetic relaxation rate of solvent protons in solutions of diamagnetic proteins have indicated that this dependence (called relaxation dispersion) is related to the rotational Brownian motion of solute proteins. In essence, the dispersion is such that 1/T1 (the proton spin-lattice relaxation rate) decreases monotonically as the magnetic field is increased from a very low value (approximately 10 Oe); the dispersion has a point of inflection at a value of magnetic field which depends on protein size, shape, concentration, temperature, and solvent composition. The value of the proton Larmor precession frequency nu(c) at the inflection field appears to relate to tau (R), the rotational relaxation time of the protein molecules. We have measured proton relaxation dispersions for solutions of various proteins that span a three-decade range of molecular weights, and for one sample of transfer ribonucleic acid. We have also measured deuteron relaxation dispersions for solutions of three proteins: lysozyme, carbonmonoxyhemoglobin, and Helix pomatia hemocyanin with molecular weight 900 000. A quantitative relationship between both proton and deuteron dispersion data and protein rotational relaxation is confirmed, and the point is made that magnetic dispersion measurements are of very general applicability for measuring the rotational relaxation rate of macromolecules in solution. It has been previously shown that the influence of proton motion on the relaxation behavior of the solvent is not due to exchange of solvent molecules between the bulk solvent and a hydration region of the protein. In the present paper, we suggest that the interaction results from a long range hydrodynamic effect fundamental to the situation of large Brownian particles in an essentially continuum fluid. The general features of the proposed mechanism are indicated, but no theoretical computations are presented.