Three-dimensional NMR microscopy of rat spleen and liver.

Three-dimensional microscopic NMR images of spleen and liver specimens from rats injected with dextran magnetite particles and from controls were obtained at 4.7 T, using a specially designed probe in conjunction with a 3D filtered back projection reconstruction algorithm. All of the images were reconstructed as 64(3) arrays with (25 microns) 3 isotropic voxels. With the aid of the MR contrast agent, the red pulp and marginal zone of the spleen and the portal triad of the liver could be distinguished from the surrounding tissue in T2-weighted images. For mature rat spleen, natural contrast in T2-weighted images was found to distinguish the same features. Histological examinations of the tissues with and without contrast agent were also performed using an optical microscope. Microscopic NMR images, despite their lower resolution, clearly revealed many features seen in the optical images.

Dextran magnetite as a liver contrast agent.

The superparamagnetic particle dextran magnetite was studied as a liver tumor contrast agent for magnetic resonance imaging (MRI). The effects of dextran magnetite on the longitudinal (T1) and transverse (T2) relaxation times in liver, spleen, and an implanted rat liver tumor were measured at 0.47 T (IBM/Bruker PC-20 relaxometer) over the dose range of 23 to 69 mumol Fe/kg. Dextran magnetite substantially reduced the T2 of the liver and spleen, but not of the tumor, thereby providing a basis for improved tumor imaging. The T1 of the tumor was not affected following injection of dextran magnetite in the dose range studied, while the spleen T1 was reduced substantially more than the T1 of the liver. Histological studies using the iron reaction for Prussian blue clearly showed dextran magnetite in the liver and spleen, but not in the tumor. While dextran magnetite was sequestered in macrophages in both liver and spleen, the distribution in the liver was more diffuse (70 microns average particle separation) than that in the spleen (25 microns separation). The lack of a T1 effect in the liver is consistent with the fact that a majority of the water in the tissue cannot diffuse to the relaxational centers on the time scale of the liver's intrinsic T1 (280 ms). In the spleen, however, the dextran magnetite is more densely packed in the red pulp allowing a significant fraction of the water to be relaxed by a T1 mechanism. Spin-echo images of the implanted tumor (mammary adenocarcinoma. R3230AC) in the livers of Fischer 344 rats were obtained at 0.50 T (Siemens Magnetom). The tumor-to-liver contrast was improved for both T1 and T2-weighted spin-echo images after intravenous injection of the dextran magnetite contrast agent. The contrast determined from these images agreed with that predicted by the measured T1 and the T2 (Hahn spin-echo) values. In addition, gradient-echo T2-weighted images with good contrast were obtained in a much shorter imaging time than was needed for T2-weighted spin-echo images. These results demonstrate that the MRI contrast enhancement observed with dextran magnetite is based on its selective uptake and distribution in the macrophages in the liver and spleen and that this agent has substantial potential as a superparamagnetic MR contrast agent.

NMR study of water exchange across the hepatocyte membrane.

An understanding of the cellular permeability for water is needed to evaluate MR images of complex tissues, such as liver, and to interpret the effects of contrast agents. To obtain data essential for such an understanding we measured water exchange across the isolated rodent hepatocyte membrane by proton NMR relaxation with dextranmagnetite as a relaxation agent. The results are treated as water exchange in a two-compartment system, and possible reasons for deviations from that behavior are analyzed. The mean residence time of intracellular water was approximately 40 ms at 37 degrees C. We found the lower limit for the diffusional permeability of the hepatocyte membrane to be 8 x 10(-3) cm s-1. These results, combined with consideration of hepatic anatomy indicate that the failure to observe effects on the T1 of liver from particulate contrast agents such as magnetite, Gd-starch, and liposome encapsulated Mn2+ is due to the localization of these agents in the Kupffer cells. Also, the nonexponential T1 decay observed in normal liver is unlikely to be due to slow exchange of water between compartments.