Measurement of tissue potassium in vivo using 39K nuclear magnetic resonance

39K nuclear magnetic resonance (NMR) spectra were readily obtained, in vivo, from rat muscle, kidney, and brain in 5-10 min with signal-to-noise ratios of approximately 20:1. Quantitation of the K+ signal was achieved by reference to an external standard of KCl/dysprosium nitrate as well as by reference to the proton signal from tissue water. In vitro NMR studies of isolated tissue showed a K+ visibility (NMR K+/total tissue K+) of 96%, 62 +/- 8%, 47 +/- 1.9%, 45 +/- 3.5%, and 43 +/- 2.5% for blood, brain, muscle, kidney, and liver, respectively. Absolute tissue K+ was determined by flame photometry of acid-digested tissue. Changes in tissue K+ status by chronic K+ depletion or acute K+ loading produced changes of 39K NMR signal intensity that were equal to changes of absolute tissue K+. Acidosis, alkalosis, mannitol, or RbCl infusion did not significantly change the NMR K+ signal. These results indicate that the changes in K+ detected by NMR were specifically and accurately detected. To investigate the factors that affect the 39K NMR signal, the effects of liver homogenate on 39K NMR signal intensity were studied. Addition of homogenate produced a 60% loss of signal intensity, suggesting that a large portion of cell K+ may be only 40% visible. Addition of RbCl to undiluted homogenate increased the NMR K+ signal by 11 +/- 2 mumol/g. Addition of H2O or NaCl had no effect, suggesting that Rb+ was replacing K+ in sites of low (less than 40%) NMR visibility. These results demonstrate that 39K NMR experiments can be performed using intact organs. To explain the lack of detectable K+ and changes in K+ NMR visibility, a three compartment model is proposed.