Methodology for applied 4 MHz RF hyperthermia concomitant with 31P NMR spectroscopic monitoring of murine tumours.

It has been generally found that solid tumours in vivo are more susceptible to destruction by heat than normal tissues. Hyperthermia has, thus, been employed in the treatment of cancer either applied alone or in combination with other modalities such as chemotherapy and radiotherapy. However, the critical mechanism(s) by which heat sensitizes and kills cells in the solid tumour remains poorly defined. Magnetic resonance spectroscopic monitoring of tumour metabolism during application of hyperthermia may provide important insight into the response to hyperthermic challenge. The implementation of dual antenna-coil methodology that provides for NMR spectroscopic monitoring (31P at 121 MHz) concomitant with applied 4 MHz RF hyperthermia in murine tumours is described herein, in some detail. This technology, which does not require advanced (and expensive) magnetic resonance imaging systems, should be readily adaptable by other laboratories with an interest in murine tumour models.

Therapeutic efficacy as predicted by quantitative assessment of murine RIF-1 tumour pH and phosphorous metabolite response during hyperthermia: an in vivo 31P NMR study.

Described herein are the initial findings from an 'in-magnet' 31P NMR compatible hyperthermia system capable of concurrently heating and monitoring the metabolic response of murine tumours; the murine radiation induced fibrosarcoma (RIF-1) was employed for these studies. At thermal doses sufficient to raise tumour temperature to 41.5 and 43 degrees C for a period of 30 min, a marked and rapid decrease in nucleoside triphosphate concentration and in pH was observed during the heating period, while inorganic phosphate concentration increased significantly but more gradually. These 31P NMR determined metabolic indices remained depressed/elevated throughout a 1.5 h post-hyperthermia monitoring period. Importantly, these metabolic indices correlated significantly with specific growth delay. This suggests a possible role for NMR spectroscopy in early assessment, and perhaps control, of therapeutic response to hyperthermia.

Impaired prostate tumorigenesis in Egr1-deficient mice.

The transcription factor early growth response protein 1 (EGR1) is overexpressed in a majority of human prostate cancers and is implicated in the regulation of several genes important for prostate tumor progression. Here we have assessed the effect of Egr1 deficiency on tumor development in two transgenic mouse models of prostate cancer (CR2-T-Ag and TRAMP). Using a combination of high-resolution magnetic resonance imaging and histopathological and survival analyses, we show that tumor progression was significantly impaired in Egr1-/- mice. Tumor initiation and tumor growth rate were not affected by the lack of Egr1; however, Egr1 deficiency significantly delayed the progression from prostatic intra-epithelial neoplasia to invasive carcinoma. These results indicate a unique role for Egr1 in regulating the transition from localized, carcinoma in situ to invasive carcinoma.

Sepsis increases intracellular free calcium in brain.

Using magnetic resonance methods and a clinically relevant rodent model of sepsis, we have made in vivo measurements of increased intracellular calcium in a pathologic state in the CNS. The intracellular calcium concentration was increased nearly twofold in septic rat brain compared with controls (p < 0.0001). This result, in a fully intact functioning mammalian system, ties together a previous spectrum of indirect evidence from numerous laboratories suggesting an important role for elevated intracellular calcium in sepsis. In addition, levels of the proinflammatory cytokine tumor necrosis factor-a were elevated threefold in septic rat brain (p < 0.02), and electron microscopic examination revealed scattered injury in approximately 0.25% of glial cells. These findings are discussed in light of the current understanding of the pathophysiology of sepsis.

Quantification of ion transport in perfused rat heart: 133Cs+ as an NMR active K+ analog.

Proper ion balance between intra- and extracellular compartments is necessary for normal physiological function. Conversely, alterations in membrane ion transport occur in numerous pathological states. As a noninvasive, nondestructive spectroscopic technique, nuclear magnetic resonance (NMR) offers a powerful approach to the study of ion balance in intact biological systems. Unfortunately, rare NMR active nuclides that are isotopes of the 100% naturally abundant 23Na+ and 39K+ are not available for tracer kinetic studies of Na1 and K+ transport. However, Cs is a biologically active analog of K+, and the 100% naturally abundant NMR active 133Cs+ nuclide can be employed to examine K+ transport (Davis, D. G., E. Murphy, and R. E. London. Biochemistry 27: 3547-3551, 1988). The distinguishing feature of 133Cs+ is that it naturally gives two separate well-resolved NMR resonances for intra- and extra-cellular 133Cs+, permitting study of the time course changes of either of these compartments independent of the other. In this report, the experimental procedures and compartmental modeling formalism are developed that allow quantitative analysis of Cs+ membrane transport in the perfused rat heart. Intracellular 133Cs+ is shown to be 100% visible by solution-state NMR methods and its influx transport to be markedly inhibited by ouabain, a confirmation of findings previously reported by others. Intracellular 133Cs+ spin-lattice and spin-spin relaxation times at 7 T were determined to be 2.1 +/- 0.3 (SD)s (n = 8) and 0.065 +/- 0.007 (SD) s (n = 8), respectively, for T1 and T2. The rate constant for Na(+)-K(+)-ATPase pump dominated intracellular influx was measured to be 0.25 +/- 0.07 (SD) min-1 (n = 27) and that for efflux 0.005 +/- 0.001 (SD) min-1 (n = 14). The rate constant for 133Cs+ equilibration in the extracellular space at supraphysiological perfusate flow rate (20 ml/min) was found to be 4.6 +/- 0.9 (SD) min-1 (n = 20). Thus extracellular diffusion limitations do not dominate the 133Cs+ transport measurements.

Inhibition of ion transport in septic rat heart: 133Cs+ as an NMR active K+ analog.

Sepsis, the systemic response to severe infection, and the resulting multiorgan failure it induces are major contributors to intensive care unit morbidity and mortality. A number of abnormalities in ion transport processes and intracellular free Na+ ([Na+]i) and K+ ([K+]i) concentrations have been reported to occur during sepsis/endotoxemia. An effect of sepsis on the NA(+)-K(+)-ATPase may be an important contribution to changes in intracellular ion balance and the resultant pathophysiology of the disorder. The purpose of this study was to examine the effect of sepsis on the Na(+)-K(+)-ATPase in the isolated perfused rat heart using 133Cs+ nuclear magnetic resonance (NMR). Cs+ is a K+ analog, and 133Cs-NMR offers the opportunity to examine Na(+)-K(+)-ATPase activity in the intact organ via tracer kinetics. Sepsis was induced in halothane-anesthetized male Sprague-Dawley rats using the cecal ligation and perforation (CLP) model. Twenty-four to thirty-six hours after surgery, hearts from CLP or sham-operated rats were perfused with Krebs-Henseleit buffer containing 1.25 mM Cs+. The influx rate constant for Cs+ was decreased by 24% in septic rat hearts, i.e., 0.25 +/- 0.08 (SD) min 1 for controls and 0.19 +/- 0.04 (SD) min-1 for septic animals (P = 0.003). There was no difference for Cs+ efflux [0.005 +/- 0.001 (SD) min-1 for controls and 0.005 +/- 0.002 (SD) min-1 for septic animals; P = 0.8]. These results are consistent with an inhibition of the Na(+)-K(+)-ATPase pump during sepsis/endotoxemia. A decrease in the activity of the Na(+)-K(+)-ATPase pump may be responsible for or contribute to the changes in [Na+]i and [K+]i during the disorder.

Tumor 31P NMR pH measurements in vivo: a comparison of inorganic phosphate and intracellular 2-deoxyglucose-6-phosphate as pHnmr indicators in murine radiation-induced fibrosarcoma-1.

Uncertainty regarding the intracellular/extracellular distribution of inorganic phosphate (P(i)) in tumors has raised concerns that pH calculated from the tumor P(i) chemical shift may not accurately represent the intracellular pH (pHin). This issue was addressed in subcutaneously transplanted murine radiation induced fibrosarcoma-1 by directly comparing pH measured via P(i) with pH measured via the in situ generated intracellular xenometabolite 2-deoxyglucose-6-phosphate (2DG6P). In 131 comparative measurements employing eight tumor-bearing mice under both control and hyperglycemic conditions (the latter to extend the range of tumor pH examined), the pH as derived from either 2DG6P or P(i) showed only a small, but statistically significant, difference (0.07 +/- 0.11 SD; P = 0.0001). Scatter in the comparative analysis over the pH range examined (ca. 5.5-7.5) was not uniform. Above pH 6.6, 2DG6P indicated a pH lower than that of P(i) by 0.088 +/- 0.105 SD (n = 107, P = 0.0001); below pH 6.6, 2DG6P indicated a pH essentially identical to and not statistically different from that of P(i) (mean difference 0.003 +/- 0.128 SD (n = 24, P = 0.92)). Evidence is presented in support of this differential arising from a systematic measurement error due to peak overlap between 2DG6P and endogenous phosphomonoester species. These results support the use of P(i) as a tumor 31P NMR pHin indicator, at least in RIF-1 tumors under control and hyperglycemic conditions.

Increased intracellular Ca2+: a critical link in the pathophysiology of sepsis?

Severe bloodstream-borne infection--i.e., sepsis--and the resulting multiorgan failure are now the most common cause of death in many intensive care units. One of the most fundamentally important and controversial issues concerning the pathophysiology of sepsis is the role of intracellular free calcium concentration ([Ca2+]i) in this disorder. Because of the critical role of calcium as an intracellular second messenger and as a potential cellular toxin, resolution of this issue is crucial. Using 19F NMR spectroscopy and the calcium indicator 5,5'-difluoro-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetate we demonstrate in the intact perfused organ, the rat thoracic aorta, that [Ca2+]i in aortic smooth muscle is increased > 2-fold during sepsis. Furthermore, we determined that sodium dantrolene, a drug that decreases release of calcium from the sarcoplasmic reticulum and that is lifesaving in malignant hyperthermia (a disorder due to increased [Ca2+]i), is able to reduce the elevated [Ca2+]i in sepsis to control values when added in vitro or when given in vivo to the animal. These results suggest that an increase in [Ca2+]i is an early event in sepsis and that increased [Ca2+]i may be responsible for, or contribute to, cellular injury. Dantrolene may offer a therapeutic strategy in the treatment of sepsis.

Effect of methylation inhibitors on gene expression in HL-60 cells.

The methylation inhibitors Neplanocin A (Nep A), 3'-deazaadenosine (dzAdo), and 3-deaza(+/-)aristeromycin (Dari) were tested for their effect on the expression of histone H2B, actin, and the protooncogenes c-myc, and v-fos. Nep A and Dari bind to the S-adenosylhomocysteine hydrolase resulting in the accumulation of S-adenosylhomocysteine, while dzAdo served as a substrate for the enzyme. With dzAdo, inordinant amounts of 3-deazaadenosylhomocysteine (dzAdoHcy) accumulated in the cell, provided L-homocysteine (Hcy) was present. When added at sublethal concentrations, the methylation inhibitors had little or no effect on c-myc, v-fos, histone H2B, or actin expression, nor did any significant number of the drug-treated cells demonstrate myeloid characteristics. However, growth and gene expression were markedly inhibited upon the addition of Hcy and dzAdo. One of the earliest effects of dzAdoHcy on HL-60 cells was the disappearance of c-myc mRNA. Within 1 h of the addition of dzAdo and Hcy, only trace amounts of c-myc mRNA were detectable. After 4-5 h v-fos, histone H2B, and actin mRNAs also decreased to about 40% of control levels. Differences in the stability of preexisting mRNAs would appear to account for these results. Within 1 h following the addition of dzAdo and Hcy, the synthesis of rRNA and mRNA were completely blocked as measured by the incorporation of [3H]uridine.

Concurrent quantification of tissue metabolism and blood flow via 2H/31P NMR in vivo. I. Assessment of absolute metabolite quantification.

In a series of three papers, we demonstrate and validate an approach for concurrent absolute quantification in situ of blood flow and energy metabolism with a modification of the NMR method for absolute concentration determination put forth by Thulborn and Ackerman [J. Magn. Reson. 55, 357 (1983)] and later expanded upon by Tofts and Wray. In this first paper of the series, we briefly review the theoretical basis for the concentration measurement and present, for the first time, a successful paired validation of metabolite quantification via 31P surface-coil NMR through corroborative in vitro enzymatic assays. The paired radiolabeled microsphere validation of blood flow measurement via 2H surface-coil NMR employing D2O as a freely diffusible tracer and the concurrent determination of blood flow and energy metabolism in a septic rat model are presented in the accompanying second and third paper to complete the series. In this article a classical RF tank circuit is employed to describe the effect of conductive sample loading on the NMR receiver by considering its apparent series resistance. It is shown in an easily visualized generalizable manner that the effect of sample loading on the observed NMR signal intensity can be accounted for quantitatively by monitoring changes in 90 degrees pulse width at constant power at a fixed reference point, i.e., Ssample = Sphantom (PW90phantom/PW90sample). In a series of paired experiments the absolute concentrations of high energy phosphates obtained from resting rat leg muscle (n = 4) in situ (NMR) and in vitro (enzymatic) were determined as follows: [PCr]NMR = 17.2 +/- 0.8 SD, [PCr]enzymatic = 17.3 +/- 2 SD, [ATP]NMR = 5.1 +/- 0.8 SD, [ATP]enzymatic = 5.0 +/- 0.2 SD mmol/kg tissue wet wt. Results of these two independent methods of concentration determination were not statistically different (P = 0.94 and P = 0.74 respectively) and serve to rigorously validate the Thulborn approach for absolute quantification of phosphorous metabolites in situ via NMR. Furthermore, these results strongly suggest that ATP and PCr in resting rat leg muscle under normal physiologic conditions are 100% NMR visible. The free cytosolic [ADP]NMR was estimated from the creatine kinase reaction equilibrium expression to be 0.022 +/- 0.003 SD mmol/kg tissue wet wt.

Concurrent quantification of tissue metabolism and blood flow via 2H/31P NMR in vivo. III. Alterations of muscle blood flow and metabolism during sepsis.

In the conclusion of this series of reports, the application of 31P/2H NMR to investigate the pathophysiology of sepsis in rat hindlimb muscle is demonstrated. Sepsis decreased muscle [PCr] by 18%, 18 +/- 4 SD vs 22 +/- 4 SD mmol/kg tissue wet wt (P = 0.01) in control rats but [ATP] was unchanged, 6 mmol/kg tissue wet wt (P = 0.2). The derived free cytosolic [ADP] in the two groups was similar, [ADP]septic = 0.023 +/- 0.004 SD and [ADP]control = 0.021 +/- 0.003 SD mmol/kg tissue wet wt, and not statistically different (P = 0.14). Likewise [Pi] in the septic and control groups was not statistically different, [Pi]septic = 1.1 +/- 0.5 SD and [Pi]control = 1.2 +/- 0.4 SD mmol/kg tissue wet wt (P = 0.2). Septic rats presented the symptom of respiratory alkalosis evidenced by elevated blood pH. Sepsis decreased muscle blood flow by 33%, P = 0.003, but examination of individual subjects did not demonstrate a correlation with the reduction in [PCr]. Thus, a metabolic energy deficit caused by cellular ischemia/hypoxia is not a likely cause of cellular abnormality in rat hindlimb muscle during sepsis.

Nonglycolytic acidification of murine radiation-induced fibrosarcoma 1 tumor via 3-O-methyl-D-glucose monitored by 1H, 2H, 13C, and 31P nuclear magnetic resonance spectroscopy.

The effects of 3-O-methyl-D-glucose (3-OMG) on subcutaneously implanted murine radiation-induced fibrosarcoma 1 tumor were examined with 2H, 13C, and 31P nuclear magnetic resonance (NMR) in situ. Using 31P NMR, changes in tumor high-energy phosphate metabolism were monitored for 2.5 h after i.p. administration of 3-OMG (8.1 g/kg body weight); tumor pH decreased by a mean maximum of 0.52 +/- 0.05 (SE) (n = 10), [PCr] decreased by 54%, [NTP] decreased by 35%, and [Pi] increased by 36%. Tumor blood flow, as measured by 2H NMR monitoring of D2O washout kinetics, decreased by 40% at 1 h and by 47% at 2 h after 3-OMG injection (n = 4). This substantial tumor acidification (pH decrease much greater than 0.1), expected to require a glycolytic substrate (Hwang et al., Cancer Res., 51: 3108-3118, 1991), is surprising in light of the previously documented metabolically inert nature of 3-OMG. In situ 13C NMR spectroscopy, following [6-13C]3-OMG i.p. injection, examined the possibility of the glycolytic metabolism of 3-OMG. However, only the C-6 resonance of 3-OMG was detected (n = 6); no resonances from [6-13C]3-OMG-6-phosphate or [3-13C]lactate were observed. These results confirmed that 3-OMG was not metabolized in radiation-induced fibrosarcoma 1 tumor. At the completion of the in situ 13C NMR experiments, tumors were freeze clamped, and perchloric acid extraction was performed. High-resolution 1H NMR measurement of lactate concentrations showed no statistically significant difference in control tumor extracts (from mice not receiving i.p. injection; n = 5) and in tumor extracts from mice administered i.p. [6-13C]3-OMG (n = 5), indicating that there was no significant increase in lactate level in the tumor extracts from mice administered i.p. 3-OMG due to increased plasma glucose concentration. The results of these 1H and 13C NMR studies indicated that the radiation-induced fibrosarcoma 1 tumor acidification caused by i.p. administration of 3-OMG was not due to a direct (3-OMG----lactate) or an indirect (systemic glucose----lactate) increase in tumor lactic acid levels.

Modulation of murine radiation-induced fibrosarcoma-1 tumor metabolism and blood flow in situ via glucose and mannitol administration monitored by 31P and 2H nuclear magnetic resonance spectroscopy.

The hyperglycemia-induced in situ metabolism and blood flow changes produced in s.c. implanted murine radiation-induced fibrosarcoma-1 tumors, grown on the flanks of female C3H/HeJ mice, were examined with 31P and 2H nuclear magnetic resonance. Initial experiments verified a hyperglycemic tumor acidification similar to that reported earlier with a different substrain of mice, C3H/AnF (J.L. Evelhoch et al., Proc. Natl. Acad. Sci. USA, 81: 6496-6500, 1984). Changes in the tumor pH, phosphorus metabolites, and blood flow were then compared after administration of saline, glucose, or mannitol (a nonmetabolizable glucose analogue) using a mole-equivalent dose of the sugars (i.e., 0.8 mmol/20g mouse). Neither saline (n = 8) nor mannitol (n = 6) administration had any marked effect upon tumor pH, whereas glucose administration produced a mean maximum tumor pH reduction of 0.74 +/- 0.09 (SE; n = 9) during the 2.5 h post-glucose injection. No significant changes in high energy phosphate concentrations were observed during the same period after saline injection. After glucose injection, the [phosphocreatine] gradually decreased by 64% (P = 0.0001). After the initial 1 h post-glucose injection, the [inorganic phosphate] increased by 58% (P = 0.0001), and the [nucleoside triphosphates] decreased by 29% (P = 0.0001) during the following 1.5 h. After mannitol injection, while there was no change in [inorganic phosphate] over time (P = 0.37), the [phosphocreatine] decreased by 33% (P = 0.0001) and the [nucleoside triphosphates] decreased by 21% (P = 0.0015) within 20 min, then both the [phosphocreatine] and [nucleoside triphosphates] remained at constant levels during the following 2 h. In parallel experiments, the volumetric rate of tumor blood flow and perfusion was measured by 2H nuclear magnetic resonance monitoring of 2H2O washout kinetics (S-G. Kim and J. J. H. Ackerman, Cancer Res., 48: 3449-3453, 1988); tumor blood flow decreased by 80% (P = 0.0001, n = 11), 60% (P = 0.0031, n = 4), and 20% (P = 0.058, n = 10) at 2 h after glucose, mannitol, or saline injections, respectively. These results suggest that anaerobic glycolysis is a requirement for hyperglycemic tumor acidification. However, the decrease in tumor blood flow accompanying hyperglycemic acidification suggests that flow reduction also may be a contributing or a required cofactor for acidification via inhibition of lactic acid egress.(ABSTRACT TRUNCATED AT 400 WORDS)

Sepsis does not impair tricarboxylic acid cycle in the heart.

Sepsis has been reported to cause mitochondrial dysfunction and inhibition of key enzymes that regulate the tricarboxylic acid (TCA) cycle. We investigated the effect of sepsis on high-energy phosphates, glycolytic and TCA cycle intermediates, and specific amino acids that are involved in regulating the size of the TCA cycle pool during changes in metabolic state of the heart. Sepsis was induced in 12 female rats by the cecal ligation and perforation technique under halothane anesthesia; seven control rats underwent cecal manipulation without ligation. At 36-42 h postsurgery, the rats were reanesthetized, the chest was opened, and the hearts were freeze-clamped. Perchloric acid extracts of the hearts were analyzed with fluorometric enzymatic methods and 31P nuclear magnetic resonance spectroscopy. There were no significant differences in the levels of the TCA cycle intermediates or high-energy phosphates between the septic and control rats. The major metabolic changes were the 28% decrease in alanine and the 31% decrease in glutamate in the septic hearts compared with control (P less than 0.05 and P less than 0.005, respectively). Phosphocholine, a component of membrane phospholipids, was increased by 91% in the septic hearts (P less than 0.01). We conclude that sepsis does not impair the TCA cycle or induce significant cellular ischemia in the heart. The increase in phosphocholine may represent significant cellular membrane disruption during sepsis.

Quantification of regional blood flow by monitoring of exogenous tracer via nuclear magnetic resonance spectroscopy.

Deuterium and fluorine nuclear magnetic resonance spectroscopy have been employed for quantification of regional blood flow in concert with nonradiative, exogenous, freely diffusible tracers such as D2O and freon gas. Typically, the tracer residue washout was monitored by NMR over time following tracer administration by bolus injection or inhalation. The theory, including compartmental analysis, required to quantitatively derive volumetric tissue blood flow and perfusion is reviewed herein. Applications of NMR tissue blood flow measurement techniques to tumor, muscle, liver, and brain are presented with discussion of the advantages and disadvantages of NMR methods.

Long-term culturing of TPA-induced differentiated HL-60 cells results in increased levels of lytic enzymes.

After exposure to 12-O-tetradecanoylphorbol-13-acetate (TPA), cells of the promyelocytic leukemia cell line, HL-60, differentiate into macrophage-like cells. Within 24 h the cells adhere to the surface of the culture flask and increase production of nonspecific esterases. The intracellular concentration of the serine proteases increases two- to threefold within 4 days and continues to increase as the cells develop into mature macrophages. The acid hydrolases, lysozyme and beta-glucuronidase, were secreted by the differentiated cells. Both the intracellular and extracellular concentrations of these enzymes continued to increase as the cells matured. The fully differentiated cells readily phagocytized opsonized yeast cells. Phagocytosis had little effect on the secretion of acid hydrolases, while intracellular proteases increased significantly. The fully differentiated HL-60 cells resembled normal macrophages regarding all parameters studied. Viability of the differentiated cells exceeded 50% when cultured for 30 days. Therefore, these cells should prove to be a useful tool for the study of macrophage function with respect to microorganisms that are resistant to destruction by phagocytic cells.

Multicompartment analysis of blood flow and tissue perfusion employing D2O as a freely diffusible tracer: a novel deuterium NMR technique demonstrated via application with murine RIF-1 tumors.

Deuterium NMR is employed in concert with multicompartment kinetic analysis for measurement of tissue blood flow and perfusion through a bolus administration of D2O as a freely diffusible tracer. The traditional single-compartment and two-compartment in-parallel flow models with no tracer recirculation are briefly discussed. The two-compartment in-series flow model with recirculation is developed to account for reflow of the stable (slowly excreted) deuterium tracer. With this model a monoexponential tracer washout curve is predicted. The rate of blood flow and tissue perfusion is readily extracted by three-parameter monoexponential analysis of the residue decay curve. A three-compartment model with recirculation, incorporating one compartment in-series with two compartments in-parallel, is developed for analysis of biexponential tracer washout curves. With this model the flow rates through the two in-parallel compartments (i.e., fast and slow) and the volume fractions of these two compartments are obtained by five-parameter biexponential analysis of the residue decay curve. Application of these multicompartment tracer-recirculation flow models is demonstrated with in situ determinations of murine RIF-1 tumor blood flow and tissue perfusion. The blood flow rates determined by deuterium NMR and analyzed by the multicompartment flow models agree well with those determined by others using radiolabels. A companion article (S.-G. Kim and J.J.H. Ackerman, Cancer Res. 48, 3449-3453, 1988) discusses in more depth the practical aspects of applying these multicompartment models to tumor blood flow measurement.

Quantitative determination of tumor blood flow and perfusion via deuterium nuclear magnetic resonance spectroscopy in mice.

Murine RIF-1 tumor blood flow and perfusion were quantified by deuterium NMR using D2O as a freely diffusible tracer. After direct intratumor injection of D2O saline solution, the tracer (HOD) residue from the tumor was detected by deuterium NMR and the deuterium residue washout time course was then analyzed employing multicompartment flow models (S-G. Kim and J.J.H. Ackerman, manuscript submitted for publication). The mean tumor blood flow and perfusion rate was 18.5 +/- 8.5 SD ml/(100 g.min) (n = 46) when analyzed by a two-compartment in-series flow model. A number of tumors (n = 15 out of 61 total) showed a biexponential deuterium tracer washout curve. Application of a three-compartment flow model (S-G. Kim and J.J.H. Ackerman, manuscript submitted for publication) fitted the biexponential residue decay data well and yielded a mean tumor blood flow of 15.7 +/- 9.7 SD, fast- and slow-flow components of 36.8 +/- 19.8 SD and 9.7 +/- 5.8 SD ml/(100 g.min), and a fast-flow component fraction of 21 +/- 10 SD%. Small tumors of less than 0.5 cm3 had faster blood flow, 21.1 +/- 8.4 SD ml/(100 g.min) (n = 27), than large tumors of greater than 1.0 cm3, 9.4 +/- 2.9 SD ml/(100 g.min) (n = 13). The NMR measurement of tumor blood flow and perfusion was not dependent on the number of direct intratumor injection sites and was found reproducible upon repeated measurements of individual tumors. Good agreement with previous in situ photon activation H215O flow determinations was observed.