Magnetic resonance imaging of gadolinium-labeled monoclonal antibody polymers directed at human T lymphocytes implanted in canine brain.

Two different murine monoclonal anti-human T cell antibodies, that were coupled to gadolinium (Gd), bind specifically to human T lymphocyte cells implanted in canine brain. This binding was at a concentration of Gd sufficient to detect the implanted cells and to distinguish them from the surrounding brain tissue with magnetic resonance imaging (MRI) at a field strength of 1.5 Tesla. These Gd-labeled immunoglobulin preparations did not bind bovine T cells at a concentration sufficient to be detected on MRI. A protein solution containing the immunoglobulins (100 micrograms), gelatin (2 mg), and bovine serum albumin (2.5 mg) was reacted with the dianhydride of diethylenetriaminepentaacetic acid (DTPA); the DTPA serves as a metal chelator and as a protein crosslinking agent. The DTPA-protein complex was reacted with Gd chloride. There were approximately 10 DTPA residues per protein molecule in the modified protein mixture. Isolated human or bovine monocytes (approximately 12 million cells) were implanted in the brains of anesthetized dogs in a volume of 40 microliters. The blood-brain barrier was then disrupted by the intra-arterial injection of hyperosmotic mannitol, and the Gd-labeled antibodies were injected through a catheter placed at the branch of the internal and external carotid arteries. The brains were imaged 48 to 72 hours later. The MRI scans revealed a markedly decreased T1 relaxation time with a high signal intensity (TE = 25 msec, TR = 200 msec) related to the human T cell implants. There was no evidence of decreased T1 at the site of the bovine T cells. Neither control murine gamma globulin coupled to Gd-DTPA nor anti-human T cell antibodies uncoupled to Gd modified the MRI contrast of the human T cells in the brain.