Experimental31p nmr study of the influence of ionic strength on the apparent dissociation constant of mgatp

The classical method for³¹P NMR determination of intracellular free magnesium concentration ([Mg _free ²⁺ ]) requires an accurate knowledge of the apparent dissociation constant (K _D ) of MgATP. There is a large difference between the previously determined values ofK _D . Although the value of 50 µM, determined by a³¹P NMR method, is now largely accepted, a value of 86 µM has more recently been measured with a fitting method derived from the original one, and with a different ionic strength. The purpose of our study was to assess if the cause of the difference between these two previously reportedK _D values was due to the measuring method or to the ionic strength value used.Working at pH=7.2,T=37°C, and [KCl]=0.25 M, we performedK _D measurements with the original³¹P NMR method and with the fitting method. The results (67±13 µM and 61±20 µM, respectively) were not significantly different. Then, with the first method, we measured K_D at [KCl]=0.12 M and found a value of 19±5 µM. We conclude that the main cause of difference between theK _D values measured by³¹P NMR reside in the disparity of ionic strength values used for their measurement. OurK _D measurements at [KCl]=0.25 and 0.12 M demonstrate the importance of the ionic strength value used for imitating the intracellular medium on the absolute value of ([Mg _free ²⁺ ]) measured by³¹P NMR spectroscopy.Address for correspondence: Université Catholique de Louvain, Unité CPMC, Bâtiment Lavoisier, Place Louis Pasteur n°1, B-1348 Louvain-la-Neuve, Belgium. Additional reprints of this chapter may be obtained from the Reprints Department, Chapman & Hall, One Penn Plaza, New York, NY 10119.     

MRS studies on calcium and zinc in metabolizing cerebral slices exposed to glutamate and NMDA

The¹⁹F-NMR calcium indicator 5FBAPTA has been used to measure changes in free intracellular calcium, [Ca²⁺],, in superfused guinea pig cerebral cortical slices. Increases in [Ca²⁺], have been observed with depolarization and the combined insult of hypoxia and hypoglycemia. These increases in [Ca²⁺]_i, can be explained by the excitotoxic hypothesis which proposes a central role for the neurotransmitter glutamate and the NMDA receptor in cell death. We, therefore, investigated the effects of glutamate and NMDA directly on [Ca²⁺]_i, in cerebral cortical slices with³¹P-NMR to monitor the energy state. In the presence of glutamate (±Mg) a new resonance was observed in the¹⁹F spectra and was attributed to the Zn-5FBAPTA complex. The zinc peak appears with or just following an increase in [Ca²⁺]_i, and with a drop in PCr. Similar results were obtained on exposing the tissues to NMDA. Zinc has been reported to be enriched in various parts of the brain and have a multiplicity of possible roles. This is the first time zinc has been shown in actively metabolizing tissue. The observation that the zinc resonance does not appear with depolarization or hypoxia and hypoglycemia suggests that the mechanisms of damage in these latter insults are not solely attributable to the release of excitotoxins.     

Potential pitfall in the determination of free [Mg2+] by31P NMR when using the β/α-ATP peak height ratio method

Recently, Clarke et al., (Clarke K, Kashiwaya Y, King MT, Gates D, Keon CA, Cross HR, Radda GK, Veech RL. The β/α peak height ratio of ATP. A measure of free [Mg _free ²⁺ ] using³¹ P NMR, J. Biol. Chem. 1996;271:21142–21150.) reported a new method to noninvasively determine the concentration of intracellular free magnesium ([Mg _free ²⁺ ]) based on the measurement of the peak height ratioh _β/α of the β- and α-ATP signals in³¹P NMR spectra.h _β/α varies with Mg_free ²⁺], however, the study presented here shows thath _β/α also strongly depends on the homogeneity of the static magnetic field. For this reason, we performed at a magnetic field strength of 1.5 T³¹P NMR measurements of solutions that mimic intracellular medium. The magnetic field homogeneity was varied by changing the currents in the shim coils, and the effect onh _β/α is demonstrated with and without proton decoupling. In both cases,h _β/α strongly depends on the magnetic field homogeneity and can therefore lead to a pitfall in the determination of [Mg_free ²⁺].     

Brain T2 relaxation times correlate with regional cerebral blood volume

We previously reported cerebellar and putaminal transverse relaxation time (T2) differences in children with ADHD and in adults with childhood trauma. As brain T2 can be altered by deoxyhemoglobin concentration ([dHb]) and because [dHb] is proportional to regional cerebral blood volume (rCBV), at steady state we attributed those differences to rCBV changes. Studies in other species have established a correlation between T2 and rCBV; however this has yet to be demonstrated in human brain. Echo planar imaging (EPI) T2 relaxometry and dynamic susceptibility-contrast (DSC) MRI were used to measure T2 and rCBV in 11 healthy adults. Significant T2-rCBV correlations were observed in both cerebellar vermis and putamen (r=0.759,p=0.007;r=0.782,p=0.004, respectively). These correlations predict 9±3% and 10±3% rCBV changes, respectively, for each 1-msec change in T2. Consequently, brain T2 measurements may be useful for estimating steady-state rCBV.     

Is absolute noninvasive temperature measurement by the Pr[MOE-DO3A] complex feasible

Recently, the feasibility of the praseodymium complex of 10-(2-methoxyethyl)-1,4,7,10-tetraaza-cyclododecane-1,4,7-triacetate (Pr[MOE-DO3A]) for non-invasive temperature measurement via¹H spectroscopy has been demonstrated. Particularly the suitability of the complex for non-invasive temperature measurements including in vivo spectroscopy without spatial resolution as well as first spectroscopic imaging measurements at low temporal resolution (≥4 min) and high temporal resolution (breath hold, ∼20 s) has been shown. As of today, calibration curves according to the particular experimental conditions are necessary. This work aims to clarify whether the Pr[MOE-DO3A] probe in conjunction with¹H-NMR spectroscopy allows non-invasive absolute temperature measurements with high accuracy. The measurement results from two different representative media, distilled water and human plasma, show a slight but significant dependence of the calibration curves on the surrounding medium. Calibration curves in water and plasma were derived for the temperature dependence of the chemical shift difference (F) between Pr[MOE-DO3A]'s OCH₃ and water withF=−(27.53±0.04)+(0.125±0.001)^* T andF=−(27.61±0.02)+(0.129±0.001)^* T. respectivel, withF in ppm andT in °C. However, the differences are minuscule even for the highest spectral resolution of 0.001 ppm/pt, so that they are indistinguishable under practical conditions. The estimated temperature errors are ±0.18°C for water and ±0.14°C for plasma and with that only slightly worse than the measurement accuracy of the fiber-optical temperature probe (±0.1°C). It can be concluded that the results obtained indicate the feasibility of the¹H spectroscopy method in conjunction with the Pr[MOE-DO3A] probe for absolute temperature measurements, with a maximum accuracy of ±0.2°C.     

Quality assurance of PASADENA hyperpolarization for 13C biomolecules

Object  Define MR quality assurance procedures for maximal PASADENA hyperpolarization of a biological ¹³C molecular imaging reagent. Materials and methods  An automated PASADENA polarizer and a parahydrogen generator were installed. ¹³C enriched hydroxyethyl acrylate, 1-¹³C, 2,3,3-d₃ (HEA), was converted to hyperpolarized hydroxyethyl propionate, 1-¹³C, 2,3,3-d₃ (HEP) and fumaric acid, 1-¹³C, 2,3-d₂ (FUM) to hyperpolarized succinic acid, 1-¹³C, 2,3-d₂ (SUC), by reaction with parahydrogen and norbornadiene rhodium catalyst. Incremental optimization of successive steps in PASADENA was implemented. MR spectra and in vivo images of hyperpolarized ¹³C imaging agents were acquired at 1.5 and 4.7 T. Results  Application of quality assurance (QA) criteria resulted in incremental optimization of the individual steps in PASADENA implementation. Optimal hyperpolarization of HEP of P = 20% was achieved by calibration of the NMR unit of the polarizer (B ₀ field strength ± 0.002 mT). Mean hyperpolarization of SUC, P = [15.3 ± 1.9]% (N = 16) in D ₂O, and P = [12.8 ± 3.1]% (N = 12) in H ₂O, was achieved every 5–8 min (range 13–20%). An in vivo ¹³C succinate image of a rat was produced. Conclusion  PASADENA spin hyperpolarization of SUC to 15.3% in average was demonstrated (37,400 fold signal enhancement at 4.7 T). The biological fate of ¹³C succinate, a normally occurring cellular intermediate, might be monitored with enhanced sensitivity.     

Mechanisms of31P relaxation in phosphorus metabolites

Measurements have been made of the longitudinal relaxation timeT ₁ of³¹P for the individual resonances of the metabolites AMP, ADP, ATP, P_ir and PCr (phosphocreatine) in H₂O and D₂O solutions from 5 to 60°C at various concentrations and at frequencies of 40 MHz (2.3 T) and 120 MHz (7 T). The contributions of dipolar, chemical shift anisotropy, and spin-rotation mechanisms have been separated, and activation parameters of the underlying molecular reorientations have been determined.     

Creatine loading and resting skeletal muscle phosphocreatine flux: a saturation-transfer NMR study

³¹P saturation-transfer nuclear magnetic resonance spectroscopy was used to study skeletal muscle phosphocreatine (PCr) flux in healthy male volunteers. Data analysis included consideration of effects from incomplete saturation and radiofrequency spillover. Spectra were recorded from the resting gastrocnemius muscle before and after 6 days of creatine monohydrate (Cr·H₂O) intake (20 g/day). Parallel to an improved muscle performance during maximal intermittent exercise following Cr·H₂O supplementation, the concentration of PCr increased(P = 0.01) by 23% (34.9 ± 2.8 mmol/ 1 vs. 28.6 + 2.7 mmol/1), whereas other metabolites were unaffected (inorganic phosphate: 4.3± 1.4 mmol/1, free intracellular Mg²⁺: 1.1 ±0.7 mmol 1, cytosolic pH: 7.04 ± 0.02). Forward and reverse fluxes through the creatine kinase (CK) reaction did not change significantly from their baseline levels (v_for: 11.8 ± 5.4 mmol/1 per second vs. 15.3 ± 6.8 mmol 1 per second. (v_rev: 9.5 ± 3.4 mmol/1 per second vs. 10.9 ± 3.7 mmol/ 1 per second). The rate of PCr resynthesis in resting muscle is not limited by the CK reaction, which is near equilibrium. Consequently, the post-load increase in total creatine has no effect on the unidirectional CK reaction rates.     

Severity of delayed (“secondary”) cerebral energy failure after acute hypoxia-ischemia is related to the time integral of acute ATP depletion

The aim of this study was to reproduce the delayed (secondary) cerebral energy failure previously described in birth-asphyxiated newborn infants and to investigate relationships between primary insult severity and the extent of the delayed energy failure. Phosphorus (³¹P) magnetic resonance spectroscopy (MRS) at 7 T was used to study the brains of 12 newborn piglets during an acute, reversible, cerebral hypoxic-ischemic episode which continued until nucleotide triphosphates (NTP) were depleted. After reperfusion and reoxygenation, spectroscopy was continued for 48 h. High-energy metabolite concentrations returned to near normal levels after the insult, but later they fell as delayed energy failure developed. The time integral of NTP depletion in the primary insult correlated strongly with the minimum [phosphocreatine (PCr)]/[inorganic orthophosphate (P_i)] observed 24–48 h after the insult. (Linear regression analysis gave slope –8.04 h^–1; ordinate intercept=1.23;r=0.92;P<0.0001.) This model is currently being used to investigate the therapeutic potential of various cerebroprotective strategies including hypothermia.     

Physiological constraints on changes in pH and phosphorus metabolite concentrations in ischemically exercising muscle: implications for metabolic control and for the interpretation of31P-magnetic resonance spectroscopic studies

Relationships between pH and the concentrations of phosphocreatine (PCr), inorganic phosphate (Pi), and lactate during ischemic exercise depend on passive buffering, proton consumption as a consequence of net PCr breakdown, the control of glycogenolysis, (particularly in relation to the concentration of Pi, a substrate of glycogen phosphorylase that is produced by net PCr breakdown), and the creatine kinase equilibrium. The author analyzes the implications of these relationships for the interpretation of³¹P-magnetic resonance spectroscopic data and for the control of glycogenolysis. For realistic adenosine diphosphate (ADP) concentrations, given the constraints of the creatine kinase equilibrium, the pH must be near-linear with lactate, with an apparent buffer capacity (i.e., the ratio of lactate accumulation to pH change) that is nearly twice the true buffer capacity (i.e., the ratio of net proton loading to pH change). The implications for glycogenolytic control depend on adenosine triphosphate (ATP) turnover, but an upper limit of activation of glycogen phosphorylase (i.e., the amount of thea form) that would permit no increase in ADP concentration can be calculated. Phosphorylase activation during ischemic exercise seems approximately proportional to the power output, consistent with calcium stimulation of phosphorylaseb kinase. In simulations, ADP concentration is highly sensitive to this proportionality, as (unlike in purely oxidative exercise) ADP concentration is not known to participate in any closed feedback loops in ischemic exercise.     

31P NMR studies of human soleus and gastrocnemius show differences in theJ γβ coupling constant of ATP and in intracellular free magnesium

Localized proton decoupled³¹Pin vivo NMR spectroscopy of the human calf muscle was performed using a 1.5-T whole-body imager and the slice selective two-dimensional chemical-shift-imaging (2D-CSI) technique. The³¹P-³¹P coupling constants and the chemical shifts of ATP were compared in gastrocnemius and soleus. Significant differences were found in the coupling constantJ : (18.1±0.7) Hz versus (17.1±0.6) Hz (means ± SD,P<10^–5). Differences were also observed in the chemical shift separation between the - and -ATP signal: (8.498±0.023) ppm versus (8.522 ± 0.022) ppm (p<0.001) in gastrocnemius and soleus, respectively. Ahigher [MgATP]/[ATP_free] ratio and a significantly higher level of intracellular free magnesium of (0.52±0.06) mM in gastrocnemius versus (0.46 ± 0.05) mM in soleus (p<0.001) can be derived based on and K _D ^MgATP . Heterogeneity needs to be taken into account in clinical studies on magnesium by NMR methods in calf muscle. The coupling constantJ provides additional information, possibly on enzymatic processes, and correlates with [Mg _free ^2– ]. The detailed analysis of muscles with different fiber type characteristics lends support to the significance of this parameter in evaluating metabolism. The data reported can be used as prior knowledge for fits in which the coupling constants are set to a fixed value.     

T2 Relaxation time study of the pig renal allografts

A previous study showed the possibility of the dependence of theT ₂ maps measured by Siemens Magnetom MR imager 1.5 T on the viability of the cadaveric kidney. For theT ₂ relaxometry study the pig model was used. Ten pig renal allografts were examined during 68 h after removal. Significant differences were found in the change of the initialT ₂ ⁱⁿⁱ values in the histogram within the first 20 h [₁=T ₂ ⁱⁿⁱ (2)–T ₂ ⁱⁿⁱ (20)] or 8 h [₂=T ₂ ⁱⁿⁱ (2)–T ₂ ⁿⁱ (8)] after removal between the groups of the damaged kidneys (₁=20.2±6.8 ms, ₂=19.4±3.0) and the kidneys without apparent damage (₁=9.4±8.5 ms, ₂=5.4±5.8).     

Large increases in1H metaboliteT 2's after cerebral hypoxia-ischemia correlate with ATP depletion

In vivo proton (¹H) magnetic resonance spectroscopy (MRS) can measure cerebral metabolite concentrations and nuclear relaxation times. Function of the sodium (Na⁺)/potassium (K⁺) pump in cell membranes depends on adequate adenosine triphosphate (ATP) levels: intracellular Na⁺ is normally extruded in exchange for extracellular K⁺. Low ATP will cause pump dysfunction and loss of K⁺ accompanied by influx of Na⁺and water. Raised intracellular water may increase molecular mobility and this might be detectable as increased apparent transverse relaxation times (T ₂'s).¹H-MRS of the brains of newborn piglets during acute hypoxia-ischemia revealed enigmatic increases in the peak area of creatine + phosphocreatine (Cr) relative to those of choline-containing compounds (Cho) andN-acetylaspartate (NAA). Interleaved¹H and phosphorus (³¹P) MRS showed that theT ₂'s of both Cr and lactate (Lac) increased during acute hypoxia-ischemia and these changes correlated with reductions in nucleotide triphosphate (NTP; largely ATP). Within 50 h of metabolic recovery from the primary insult, as delayed energy failure developed, theT ₂'s of Cho, Cr, NAA, and Lac increased greatly. TheseT ₂ changes also correlated with NTP depletion. These observations demonstrate important relationships betweenT ₂'s and function of the ATP-dependent Na⁺/K⁺ pump.     

Kinetics of PCr to ATP and β-ATP to β-ADP phosphoryl conversion are modified in working rat skeletal muscle after training

Kinetics of phosphoryl transfers from PCr to γ-ATP and from β-ATP to β-ADP were measured by magnetization transfer in an in vivo³¹P NMR experiment in working rat skeletal hind leg muscles. Two groups were examined. One group was submitted to a 6-week training program of treadmill running. The other group was composed of sedentary animals. Metabolic oxidative capacity and mechanical performance were improved greatly by training as shown previously. Phosphoryl transfer of PCr→γ-ATP or β-ATP→β-ADP total fluxes were identical in resting trained and untrained muscles. Under stimulation, the flux of creatine kinase transfer was significantly inhibited by 23% compared with resting level in untrained muscles; by contrast, it was not inhibited and maintained at the high resting level in trained muscles. Thus physiological changes probably linked to a decrease of the production of anions, which could inhibit creatine kinase, were able to maintain creatine kinase flux. The flux of β-ATP to β-ADP transfer were enhanced largely in working muscles from 1.4±0.8 and 2±0.8 at rest to 4±1.6 and 6.6±2.7 mM s⁻¹ for untrained and trained muscles respectively; the effect was more pronounced in trained than in untrained muscles. These results showed an acceleration of phosphoryl turnover in working muscles after training, which could contribute to improve oxidative and mechanical performances. Such kinetic measurements of phosphoryl conversion may provide information on ATP turnover in pathophysiologic situations where ADP accumulates because of impaired ATP synthesis (mitochondrial myopathies, lower perfusion level).     

Quantification of metabolites from single-voxelin vivo 1H NMR data of normal human brain by means of time-domain data analysis

We present here a combination of time-domain signal analysis procedures for quantification of human brainin vivo ¹H NMR spectroscopy (MRS) data. The method is based on a separate removal of a residual water resonance followed by a frequency-selective time-domain line-shape fitting analysis of metabolite signals. Calculation of absolute metabolite concentrations was based on the internal water concentration as a reference. The estimated average metabolite concentrations acquired from six regions of normal human brain with a single-voxel spin-echo technique for theN-acetylaspartate, creatine, and choline-containing compounds were 11.4±1.0,6.5±0.5, and 1.7±0.2 mmol kg^–1 wet weight, respectively. The time-domain analyses ofin vivo ¹H MRS data from different brain regions with their specific characteristics demonstrate a case in which the use of frequency-domain methods pose serious difficulties.     

Ions,transport, and energetics in normal and diseased skeletal muscle

Magnetic resonance spectroscopy (MRS) has highlighted the relationship between intracellular ionic homeostasis and the control of muscle energetics. In skeletal muscle, the oxidative rate of ATP synthesis is largely controlled by ADP, the concentration of which is determined by the creatine kinase equilibrium that includes the concentration of H⁺. At the onset of aerobic dynamic exercise, ATP is maintained largely by glycolysis, producing lactic acid and PCr breakdown. Vasodilation follows and ATP synthesis becomes predominantly oxidative. During recovery, PCr resynthesis gives a measure of mitochondrial function, and pH recovery reflects the Na⁺:H⁺ antiport activity. Dynamic³¹P MRS measurements can be used to derive quantitative information about the processes (fluxes and concentrations) described above. In diseased muscle (e.g., mitochondrial myopathy, dystrophy, hypertension) specific changes are observed in some of the control functions, ionic fluxes, or mitochondrial oxidative rates.     

Use of spin-traps during warm ischemia-reperfusion in rat liver: comparative effect on energetic metabolism studied using31P nuclear magnetic resonance

Detection of free radicals by electron spin resonance (ESR) proves the involvement of reactive oxygen species (ROS) in reperfused organ injuries. Spin-traps are known to ameliorate hemodynamic parameters in an isolated postischemic heart. The effects of 5 mmol/L DMPO (5,5-dimethyl-1-pyrroline-N-oxide) or DEPMPO (5-(diethylphosphoryl)-5-methyl-1-pyrrolineN-oxide) on intracellular pH (pH_in) and ATP level were evaluated by³¹P nuclear magnetic resonance on isolated rat liver submitted to 1 hour of warm ischemia and reperfusion. At the end of the reperfusion period, during which pH_in recovered to its initial value (7.16±0.03) in all groups, the ATP recovery level (expressed in percentage of initial value) was similar in controls and DEPMPO (60%±5%,n=6 and 54%±4%,n=6, respectively), but only 37%±1% in DMPO-treated livers (n=6) (p<0.05 versus controls andp<0.05 versus DEPMPO). Oxidative phosphorylation was not affected by an addition of nitrones on isolated mitochondria extracted from livers not submitted to ischemia-reperfusion. In contrast, mitochondria extracted at the end of the ischemia-reperfusion showed an impairment in the phosphorylation parameters, particularly in the presence of DMPO. Mass spectrum of ischemic liver perchloric acid extracts evidenced probable catabolites in treated groups. The differences in the effect of the two nitrones on energetic metabolism may be explained by the production of deleterious catabolites by DMPO as compared to DEPMPO. Even though a specific radical scavenging effect could be operative in the liver, our results indicate that catabolic effects were predominant. The absence of deleterious effects of DEPMPO in contrast to DMPO on the liver energetic metabolism was evidenced, allowing the use of DEPMPO for ESR detection.     

Tissue oxygen tension measurements in the Shionogi model of prostate cancer using 19 F MRS and MRI

Objectives: To investigate changes in tumour tissue oxygenation throughout the tumour growth–regression–relapse cycle in an androgen-dependent animal tumour model. Materials and methods: ¹⁹F T₁ relaxometry of Perfluoro-15-Crown-5-Ether was used to measure in vivo partial oxygen pressure (pO₂) of Shionogi tumours on a 2.35-T MR scanner. Perfluoro-15-Crown-5-Ether was administered as an emulsion injected intravenously or as a neat compound injected directly into the tumour. Non-localized, tumour ¹⁹F T₁ measurements, made at multiple time points throughout the tumour cycle, were translated into pO₂ levels. Results: No correlation between tumour size and pO₂ values was found. Values of pO₂ for growing tumours (50 ± 30 torr) were significantly lower than for regressing and relapsing tumours after 9 days post-castration (70 ± 10 torr, p<0.05). Maximum pO₂ values (90 ± 30 torr) were reached between fifth and eighth day post-castration, when tumour pO₂ was significantly higher than both pre-castration (p<0.001) and after 9 days post-castration (p<0.05). Conclusion: We demonstrate that longitudinal pO₂ measurements in vivo are feasible. Values of pO₂ for growing androgen-dependent tumours were significantly lower than for regressing and relapsing androgen-independent tumours. These results have potential clinical importance in optimizing the timing of chemotherapy and/or radiotherapy of hormone dependent tumours.     

3T Renal 23Na-MRI: effects of desmopressin in patients with central diabetes insipidus

Purpose

The purpose of this prospective study was to assess physiologic changes in the renal corticomedullary ²³Na-concentration ([²³Na]) gradient with ²³Na-MRI at 3.0T in patients with central diabetes insipidus (CDI) before and after intranasal administration of 20 μg desmopressin (DDAVP).

Methods and materials

Four patients with CDI (all male, mean age 60.2 years) were included in this IRB-approved study. For ²³Na-imaging, a 3D density adapted, radial GRE-sequence (TE = 0.55 ms; TR = 120 ms; projections = 8,000; spatial resolution = 5 × 5 × 5 mm³) was used in combination with a dedicated ²³Na-coil and reference phantoms. The corticomedullary [²³Na] gradient (in mmol/L/mm) was calculated pixel-by-pixel along a linear region-of-interest (ROI) spanning from the renal cortex in the direction of the medulla. Mean ± SDs of [²³Na] were calculated for each patient as well as for the entire group.

Results

Mean [²³Na] increased along the corticomedullary gradient from the cortex (pre-DDAVP 38.0 ± 6.3 mmol/L vs. post-DDAVP 30.7 ± 3.5 mmol/L) to the medulla (pre-DDAVP 71.6 ± 14.8 mmol/L vs. post-DDAVP 59.7 ± 10.8 mmol/L). The overall mean decrease of [²³Na] after DDAVP administration was 17.1 ± 1.1 %.

Conclusion

²³Na-MRI with state-of-the-art techniques at 3T depicts the physiologic renal response to the administration of desmopressin in patients with central diabetes insipidus.