Interesting Test of Allopurinol on Gout Patients

Determination of Volatile Constituents of Human Blood and Tissue Specimens by Quantitative High Resolution Mass Spectrometry

             A technique utilizing multiple peak integrated ion current mass spectrometry at high resolution has been designed to enable the simultaneous determination of up to five compounds present in human blood plasma and muscle tissue. Apart from simple desiccation, no chemical extraction of the specimen or isolation of the compounds to be determined is required. The compounds are determined at the 10-100 ppm level to and accuracy of ±20% with a consumption of only 5 mg of desiccated material for a duplicate analysis. The method is illustrated by comparative estimations of purine metabolites in normal and gouty individuals.

             Experimental details for estimating up to five compounds simultaneously are presented together with a discussion of the accuracy attainable with particular reference to mixtures of the purines hypoxantine I, xanthine II, uric acid III, and their isomer allopurinol IV and oxypurinol V .

  Figure 1. Partial high resolution mass spectrum of substances evaporated from human muscle, showing peaks attributable to uric acid   a. Relative intensities of uric acid peaks from muscle                                                       b. Relative intensities of same peaks from pure uric acid                          c. Nominal mass of multiplet  

Experimental

    Materials. Specimens of normal skeletal muscle were obtained from patients undergoing minor surgical operations. None of these subjects ad suffered previously from any disease of purine metabolism. Specimens from untreated and allopurinol-treated gout pout patients were the same as those obtained earlier for microscopic investigation. Ten-milliliter samples of blood were withdrawn from healthy volunteers. Each specimen was immediately transferred to a heparinized container, centrifuged at 4ƒC and the plasma layer separated for investigation. The specimens were dehydrated and 2-3 mg portions admitted to the mass spectrometer.

    Qualitative Observations. Before attempting quantitative measurements, it was necessary to ensure that the characteristic ions, on which subsequent quantitative analyses if I- V were based, were completely separated from background ions of the same nominal mass arising from the fragmentation of other constituents of the specimen. In Figure 1, the relative intensities if the ions having the correct atomic composition to be derived from uric acid are compared, at a resolving power of 20,000 (10% valley definition), with a standard mass spectrum. With the exception of m/e 69 which was not used for quantitative measurements, the agreement was satisfactory. This indicated further that interference from unresolved ions of almost the same precise mass but derived from sources other than uric acid was negligible. Similar conclusions could be drawn about the characteristic ions of the other four compounds in this study.

    Quantitative Measurements. To obtain abundance data of high sensitivity, independent of vaporization temperature and inlet fractionation, for a series of ions from the same sample, a technique involving multiple peak selection and ion current integration was devised.

    Setting Up procedure. After the mass spectrometer resolving power was adjusted to some suitable value between 15,000 and 20,000, a mixture of the substances to be determined was admitted. Each channel of the peak selector was then adjusted to display on the mass spectrometer oscilloscope one of the ion beams to be measured. The ion accelerating voltage sweep was then decreased to exclude all multiplet peaks except that of interest. At a sweep speed of 2cm/sec, this display could be recorded on a fast response pen recorder.

    Measurement Procedure. After admission of the sample, the vaporization temperature was raised until the most volatile component (hypoxanthine at 23ƒC) began to appear. To avoid excessive tailing of the less volatile uric acid, the vaporization temperature was increased during the run as shown in Figure 2 to a maximum of 300ƒC. Recording was continued until all the material of interest had evaporated. Typical evaporation profiles for the molecule ions of hypoxanthine, xanthine and uric acid evaporated from normal muscle are shown in Figure 2. The areas of each profile could then be related to the absolute concentration of its precursor molecule. Allopurinol and oxipurinol were distinguished from their equally volatile structural isomers hypoxanthine and xanthine, respectively, on the basis of the differing relative intensities of m/e 52.0187 and 54.0218 for the two pairs of compounds.

    Calibrations. The mass spectrometer was calibrated for each experiment by recording the evaporation profiles for two sets of mixtures of known compositions, one set consisting of compounds I-III only, the other comprising IV and V. In order to assess the effect of the carrier media on the precision of the analysis, separate mixtures of I-III were diluted with deionised water, anhydrous sodium sulfate, and desiccated muscle from which all endogenous purine had been evaporated. The concentrations of these mixtures ranged from 10-1000 ng/mg. Because of the very low solubility in water of compounds IV and V, sodium sulfate only was used as their diluent. Calibration mixtures containing similar concentrations of caffeine (1:3:5 tri-N-methyl xanthine VI) in both methanol and sodium sulfate were used to compare the precision of determination and detection limits of a relatively nonpolar purine with those obtainable for the more polar compounds I-III.

Figure 2.  Evaporation profiles of hypoxanthine, xanthine, and uric acid.  Ordinates denote the instantaneous molecular ñ ion intensities measured sequentially on three channels at the probe temperatures indicated

Results

    Instrument Performance.  To achieve successful measurements of integrated abundance for a number of ion beams simultaneously and at high resolution, the outputs of the analyser power supplies and ion current amplifier had to remain constant to a high degree over prolonged period of time. With the present arrangement, the stability of the analyzer, including the peak preselector, was sufficient to maintain a peak in the centre of the display oscilloscope to within a distance equivalent to a ±5 ppm mass difference for a period about 8 hours without intermediate adjustment. Over the same period of time, changes in mass spectrometer sensitivity remained less than 5% as measured by the constancy of the ion currents derived from the internal standards. A prerequisite for this performance was the control of the ambient temperature to within ±2ƒC. The routine examination of biological samples caused only a slight (about 20%) deterioration in sensitivity, at constant 20,000 resolving power, over a period of six months. Throughout this time, the ion source was baked for about six hours each nigh. Since the spectrometer was calibrated on a day-to-day basis, the analytical accuracy was not affected.

    Calibrations. Linear relationships were found between the areas of the molecular-ion profiles of compounds I-V and the corresponding weights of material evaporated over the concentration range investigated. In Figure 3 the mean percentage deviations of the experimental points from the appropriate regression lines are plotted against concentration expressed as ng/mg of sample (ppm) of caffeine, hypoxanthine, xanthine, and uric acid both in solution and in solid mixtures with sodium sulfate. The use of purine-free muscle as a diluent gave results indistinguishable from those obtained with sodium sulfate. Part of the total observed error is attributable to inaccuracies in the preparation and admission to the mass spectrometer of the standard solutions and mixtures. The maximum estimated errors from these sources are also shown in Figure 3 (dash curve).

  Figure 3.  Dependence of analytical precision on concentration and diluent for caffeine, hypoxanthine, xanthine, and uric acid in solution and solid mixtures  

    Analyses. The analyses of some typical mixtures of the 5 compounds in known concentrations are presented in Table II. With very few exceptions, the overall error for concentrations above 100 ppm was within ±10% and ±20% for concentrations less than 100 ppm. The concentration of hypoxanthine, xanthine and uric acid in the plasma of a series of normal subjects are known in Table III. Also the mean measured oxypurine (hypoxanthine plus xanthine) and uric acid levels are compared with those obtained by enzymatic spectrophotometric methods. In Figure 4 the concentrations of uric acid in the muscle of gout patients before and during allopurinol treatment are compared with those in normal muscle

Discussion 

    Precision of Measurement. An examination of Figure 3 shows that for a relatively nonpolar material such as caffeine, there is no essential difference in the precision of measurement of a solution or a solid mixture of the same concentration provided the latter is thoroughly mixed. The minimum detectable concentration of caffeine in each case was about 0.5 ppm under the present conditions of measurements. It was not possible, of course, to increase the sensitivity considerably by simply increasing the mass spectrometer gain. It was not possible, however, to introduce subnanogram quantities of material into the mass spectrometer in a sufficiently reproducible manner for proper calibration. Consequently the precision of measurements under such conditions was very poor and any figure quoted for a minimum detectable concentration became questionable. A more useful concept, therefore, for quantitative purposes is the smallest measurable concentration (2 ppm in the case o caffeine) which could be measured with a precision of ±10% or better. At least half the observed error could be accounted for by handling inaccuracies. The remainder was most probably a combination of variation in mass spectrometer sensitivity (contributing not more than ±5% for concentrations of caffeine less tan 20 ppm) and errors in the integration procedure. Several methods of integration of the evaporation profile were attempted (e.g., weighting a cut-out of the profile shape and numerical integration by Simpsonís rule), but the most convenient was to determine simply the sum of the recorded ordinates. Provided at least ten ordinates were obtained per profile, this method was as satisfactory as the others.

    In the case of the more polar purines I-III, the precision of measurement was less and showed marked dependence on the method of  introduction. When aqueous solutions were used, the minimum measurable concentration was as high as 30 ppm while the error increased much more rapidly for smaller amounts than was observed for solutions of caffeine. In addition the solid residue from solutions of caffeine I-III usually required excessively high temperatures (>350ƒC) for vaporization and gave profiles with prolonged tailing. This appeared to be due to adsorption of the polar purines on the surface of the gold sample container of the inlet. These gold containers were normally regarded as having negligible surface activity but especially after prolonged cleaning the surface activity was very pronounced.  Surface deactivation took place after a container has been used for 3-4 samples, provided it was not heated to more than 300ƒC in between.  Consistent evaporation profiles usually could then be obtained but the mechanism of the deactivation process remains unclear.  Containers made from borosilicate glass proved equally prone to surface activity and were not improved by standard silanization procedures.

     By contrast, the profiles of I-III from muscle, plasma, or sodium sulfate exhibited much less tailing, were of a more consistent area, and required lower vaporization temperatures than did those from aqueous solutions.  This was taken as evidence that essentially complete evaporation of the purines took place and only negligible amounts were retained by the tissue and plasma proteins.  This was further indicated in Figure 3, where the minimum measurable concentration of I-III  in a solid mixture was now 10 ppm.  Their error curve was much more closely parallel to that of caffeine indicating a much reduced adsorption of the polar purines on the surface of the sampke container.  Some adsorption on the sodium sulfate undoubtedly took place, however, accounting for the dereased measurement precision compared with caffeine if the solid mixture was allowed to stand for  more than 24 hours. Best results were obtained from mixtures used immediately after preparation.

     The error curves of IV and V in sodium sulfate mixtures were closely similar to those of I-III and the same considerations apply.

  Figure 4.  Comparison of muscle uric acid levels (vertical lines) of 6 normal subjects, 5 gout patients before having allopurinol treatment, and 6 gout patients during allopurinol therapy

    Overall Accuracy.  Contributions to the total overall error (Table II) in the simultaneous determination of compounds I-V probably arose from two main sources besides the calibration errors.  Nevertheless, the results shown in Table II indicated that with very few exceptions, the total error in the analysis of the test mixtures at concentrations down to 50 ppm was approximately twice the calibration error for a given component.  Consequently the error in measuring such five-component mixtures at the 10-ppm level was probably no more than ± 20%.  When compared with the natural variation of purine levels within a group of normal indivuals (Table III and Figure 4), this error is not considered to be excessive.

    Measurements on Natural Specimens.  In Table III, the mean measured plasma oxypurine (hypoxanthine plus xanthine) level for five normal individuals is 0.17 ± 0.08 mg/100 ml measured by enzymatic procedures.  The corresponding figure for uric acid of 2.5 ± 0.7 mg/100 ml cannot be compared with the normal range of 4-6 mg/100 ml.  The enzymatic method measures the total uric acid and sodium urate in plasma but at blood pH 7.4, almost all the uric acid is present as the sodium salt and, hence, undetectable by mass spectrometry.  However there is good evidence that certain of the plasma proteins can form loosely bound complexes with uric acid.  For normal individuals, the concentration of uric acid bound in this way has recently been reported to be 2.9 ± 0.4 mg/100 ml which compares well with the figure obtained by mass spectrometry⁸.