Determination of Bismuth in Serum and Urine by Direct Injection Nebulization Inductively Coupled Plasma Mass Spectrometry HONGYAN LIa , BERNIE M. KEOHANEb , HONGZHE SUNa AND PETER J. SADLER*a aDepartment of Chemistry, University of Edinburgh, King’s Buildings, WestMains Road, Edinburgh, EH9 3JJ UK. E-mail: p.j.sadler@ed.ac.uk bDepartment of Chemistry, Birkbeck College, University of L ondon, Gordon House and Christopher Ingold L aboratories, 29 Gordon Square, L ondon, WC1H 0PP UK A sensitive method for the determination of Bi by direct Hg,15 Bi12,16,17 and Au18, owing to the large dead volume of the sample introduction system.injection nebulization inductively coupled plasma mass spectrometry (DIN-ICP-MS) in biological fluids is described. A microconcentric nebulizer, termed a direct injection nebulizer (DIN), which is placed inside the injector tube of the ICP The detection limit for Bi is 9.7 ng l-1 (ca. 46 pmol dm-3 ) with DIN compared with 17 ng l-1 (ca. 81 pmol dm-3) with torch, and eliminates the need for a spray chamber, was first described by LaFreniere and co-workers.19–21 It allows faster conventional pneumatic nebulization (PN). The absolute amount detectable by DIN-ICP-MS is about two orders of sample introduction and faster wash-out times, improving studies of memory-prone elements such as Hg.22 Reduced peak magnitude lower for DIN compared with PN (0.019 and 1.70 pg, respectively). Sample wash-out times were greatly broadening during flow injection has been observed due to the low dead volume of the DIN.Such a nebulizer has been reduced using DIN owing to minimization of the memory eVect for Bi. Using Tl as an internal standard, good coupled with ICP-AES,23 and with microwave-induced plasma mass spectrometry (MIP-MS),24 and Wiederin et al. have calibrations were obtained for Bi standards in 0.14 mol dm-3 nitric acid, serum and urine with comparable linearity between developed an improved DIN coupled with ICP-MS.25 DINICP- MS has previously been successfully used for the determi- the matrices, and these were used for the determination of Bi in serum and urine samples from animals dosed with the nation of Pt and Au in metalloproteins, injecting 20 ml of sample to minimize blockage of the nebulizer tip and sample antiulcer compound RBC (ranitidine bismuth citrate).This method is potentially useful in studies of the metabolism and cone.26 In the present paper, results are reported for the optimization biodistribution of Bi-containing drugs. of DIN-ICP-MS for the determination of Bi in serum and Keywords: Bismuth; direct injection nebulization; inductively urine matrices.The determination of Bi in rat serum and urine coupled plasma mass spectrometry; serum; urine samples from rats orally-dosed with RBC, a new antiulcer compound,3,4 was also investigated. Bismuth has been associated with medicine for more than 200 years. Clinical interest in the application of bismuth com- EXPERIMENTAL pounds has increased with the discovery of Helicobacter pylori Instrumentation and the usefulness of bismuth compounds in the treatment of infection caused by this microorganism. Bismuth subsalicylate A VG PlasmaQuad PQ2 ICP-MS instrument (VG Elemental, Winsford, Cheshire, UK) was used coupled with a CETAC (BSS; Pepto-Bismol) and colloidal bismuth subcitrate (CBS; De-Nol) have also been found to be eVective for the treatment Microneb 2000 direct injection nebulizer (CETAC Technologies, Omaha, NE, USA), which consisted of a micro- of peptic ulcers and to have a bactericidal eVect on Helicobacter pylori.1,2 Recently, a new antiulcer compound, ranitidine concentric nebulizer constructed inside the injector tube of the torch.A detailed description of the DIN is available else- bismuth citrate (RBC, developed by GlaxoWellcome),3,4 which combines the antisecretory action of ranitidine with the where.25 The carrier liquid was transported through a capillary made from fused silica (60 mm id×0.4 m long) directly into mucosal protectant and the bactericidal properties of bismuth, has been granted a product licence in the UK.Despite the the plasma, thereby avoiding the need for a spray chamber. Samples were injected into the carrier stream via an injection medicinal interest, the metabolism of bismuth drugs is still not well understood. This is partially due to the lack of suitable valve constructed from polyether ether ketone (PEEK) material to provide a metal-free system.Transport of liquid required techniques to detect bismuth at trace or ultra-trace levels (sub-mg l-1). a backpressure of 17×105 Pa provided by a gas displacement pump (GDP). Although ETAAS has been used to determine bismuth in serum, blood5–7 and tissues,8 this method is hampered by the The DIN-ICP-MS system was optimised for Bi sensitivity on a daily basis using a 500 ml sample loop which gave a occurrence of matrix eVects and the need for pretreatment of samples prior to analysis.9 Since 1983, ICP-MS has been steady-state signal for approximately 5 min.The instrument operating conditions are listed in Table 1. The nickel cones commercially available for multi-element determinations at low and sub-mg l-1 levels in biological materials.10–12 Total showed no signs of erosion when water was nebulized through the DIN (as other workers have found25) so they were used dissolved solids must be less than 0.2% to avoid blockage of the nebulizer tip and sampling cone.This is usually achieved for the present work. A 20 ml loop was used for all experiments to minimize the amount of protein injected. by sample dilution with 0.14 mol dm-3 nitric acid12,13 or alternative diluents.14 Long rinse times have been observed, The timings of the load and injection sequences for DIN were adjusted to minimize peak tailing. A 10 s injection time particularly for analytes with known memory eVects such as Journal of Analytical Atomic Spectrometry, October 1997, Vol. 12 (1111–1114) 1111Table 1 DIN-ICP-MS operating conditions Control rat serum was labelled as S(C) and serum samples from two rats dosed with RBC (200 mg kg-1, 1 h) as S(Bi1) VG PlasmaQuad PQ2 ICP-MS and S(Bi2). Similar procedures were used for rat urine samples, ICP torch CETAC B-23 and were labelled as U(C), control and U(Bi) and UF(Bi) for Argon flow rates— fed and fasted rats, respectively, dosed with RBC.These *Outer gas 14 l min-1 samples were diluted with 0.14 mol dm-3 nitric acid before *Intermediate gas 0.7 l min-1 analysis. *Aerosol carrier gas 0.15 l min-1 GDP-DIN parameters— *DIN nebulizer gas (argon) 5.4×105 Pa Reagents pressure High-purity water was obtained with an Elga UHQ water *GDP pressure 17×105 Pa *Liquid flow rate 100 ml min-1 purification system (resistivity of 18 MV cm). The Bi and Tl standard solutions (10 000 mg ml-1 in 5% nitric acid) were Plasma forward power 1350 W supplied by Aldrich (Gillingham, Dorset, UK). Concentrated Reflected power <10 W *Sampling position 10 mm from front of load nitric acid (67%, Suprapur grade) was from Merck coil (Lutterworth, Leicestershire, UK).The Bi and Tl standard ICP-MS interface— solutions were freshly prepared prior to analysis with 0.14 mol Sampler 1.00 mm Nicone dm-3 nitric acid as diluent. Skimmer 0.75 mm Nicone Analyser pressure 2×10-4 Pa RESULTS * These values were adjusted daily to maximize ion signal intensity.Detection Limits A comparison was made between the detection limits for Bi in 0.14 mol dm-3 nitric acid using DIN and conventional pneu- was adequate for all samples as the steady signal was reached matic nebulization (PN). The detection limit, calculated as after 3 s and remained for 10 s before tailing oV. The load three times the standard deviation of the signal produced from time, and also the rinse time, was set at 60 s for Bi aqueous ten replicates of a 0.14 mol dm-3 nitric acid blank solution, standards to allow the signal to drop to background levels was determined to be 9.7 ng l-1 (0.019 pg) for DIN and 17 ng prior to the next sample injection.Biological samples required l-1 (1.70 pg) using conventional PN. The LODs for concen- rinse times of 120 s to reach background levels, probably tration are comparable with PN but the absolute amounts are because of memory eVects from RBC or adsorption of proteins nearly two orders of magnitude lower owing to the small and salts in the sample introduction system. A summary of volume injected (20 ml compared with 1 ml for PN).the instrumental and operating parameters is given in Table 1. Memory EVects Data Acquisition Three diVerent Bi standard solutions (10, 50 and 100 mg l-1 in Time resolved analysis (TRA) software (VG Elemental) was 0.14 mol dm-3 nitric acid) were introduced into the plasma used for continuous element monitoring during analysis. Shortby either PN or DIN with flow rates of 1 ml min-1 and 100 ml term stability tests with DIN-ICP-MS were conducted in the min-1 for PN and DIN, respectively, and 20 ml of the standard peak-jumping mode using a 160 ms dwell time, three measurewere injected into the plasma, to compare rinse times for Bi. ments per peak and 30 s acquisition time, which gave RSDs The profiles produced from both modes of sample introduction of less than 1% for ten replicates of a 10 mg l-1 multiinto the ICP-MS instrument are shown in Fig. 1. In Fig. 1(a), element standard. the standard was introduced after 50 s and the steady-state Data were acquired in the peak-jumping mode and for each signal was reached after a further 60 s. Long wash-out times scan, data were stored as a unique time slice. Both Bi (m/z were necessary, particularly for the 100 mg l-1 standard which 209) and Tl as the internal standard (m/z 205) were measured using a 0.5 s time slice, 10 ms dwell time and 3 measurements per peak.A 20 ml loop was used for all sample injections in order to minimize amounts of proteins and salts injected and avoid blockage of the sampling cones and nebulizer tip. The timings of the load and injection sequences for DIN were adjusted to minimize peak tailing. The injection time was set to 10 s, and the load time was 60 s for aqueous Bi standards and 120 s for biological samples, with three injections per sample.Sample Collection and Preparation Rat blood samples (supplied by GlaxoWellcome) were taken with a poly(propylene) intravenous catheter mounted on a metallic needle and collected in higher-purity quartz tubes to avoid contamination. Samples of 20 ml of blood were collected and no anticoagulant was added. Serum was separated by centrifugation for 30 min at 3500 rev min-1 and 4 °C and decanted into polyethylene screw-cap containers. Afterwards, the serum samples were stored at #-20 °C in a freezer until required for analysis. Before analysis, the samples were Fig. 1 (a) Bi wash-out times for PN at three diVerent concendefrosted, and then a known volume was transferred into a tration levels, and (b) Bi wash-out times for DIN at three diVerent concentration levels. polyethylene calibrated flask using a polyethylene pipette. 1112 Journal of Analytical Atomic Spectrometry, October 1997, Vol. 12took 300 s to reach background levels.Introduction and wash- Bismuth in Urine and Serum Samples out times were greatly reduced with DIN as shown in Fig. 1(b). Samples of serum and urine were obtained from rats which A steady-state signal was reached after 3 s and the wash-out had received oral doses of RBC of 200 mg kg-1 body mass, time for all Bi standards was less than 30 s. which is an elevated, non-therapeutic dose. The samples were diluted with 0.14 mol dm-3 nitric acid to achieve a Bi concentration within that of the calibration curve.The Bi concen- Standard Calibration Curves trations obtained for diVerent rat serum and urine samples DIN-ICP-MS responses for diVerent concentrations of Bi were before and after dosage of RBC are given in Table 3. measured and the calibrations for Bi in 0.14 mol dm-3 nitric Background levels of Bi from the control samples are 0.54 mg acid, serum and urine (dilution factor five) were prepared. The l-1 (0.0026 mM) for serum and about 0.56 mg l-1 (0.0027 mM) correlation coeYcient for the calibration in 0.14 mol dm-3 for urine.nitric acid was 0.9998, and the RSDs ranged from 1.6% at 20 mg l-1 to 0.2% at 100 mg l-1. The slopes for the three calibrations varied, with a marked decrease in sensitivity for DISCUSSION the urine samples which suggested that the Bi signal was Recently, ICP-MS has become an important technique for the suppressed because of matrix eVects. Serum and urine samples determination of total elemental concentrations in body fluids were therefore spiked with Tl (10 mg l-1) and varying concento monitor metabolism in vivo and study the bioavailability of trations of Bi (20, 40 and 80 mg l-1) and the Bi concentrations elements.27 were determined by DIN-ICP-MS. With an internal standard The determination of Bi in biological samples using present, the measured concentrations are close to the expected ICP-MS and conventional nebulization has previously been values for both serum and urine samples (Table 2).The calireported. 12,16 Owing to an extensive memory eVect for Bi with bration curves for Bi in 0.14 mol dm-3 nitric acid, serum and PN, a long rinse time is needed and it normally takes at least urine are shown in Fig. 2. 10 min before an acceptable background level is obtained after the determination of a Bi standard solution with a concentration 100 mg l-1.16 In the present paper, for the first time, Table 2 Accuracy of Bi determinations for serum and urine samples the determination of Bi in biological samples by ICP-MS using with addition of 10 mg l-1 Tl as an internal standard for three DIN is reported.The important improvement with DIN is replicates (mean±s) that the memory eVect of Bi is dramatically reduced since the Bi concentration/mg l-1 DIN fits directly into the torch, replacing the injection tube Recovery and avoiding the use of a spray chamber and aerosol transfer Sample Expected Measured (%) tubing.With DIN, the rinse time was 60 s for Bi standards Urine 20 18.25±0.99 91.3 and 120 s for biological samples. This allowed the determi- 40 35.54±0.76 88.9 nation of Bi in these samples to be made more rapidly and 80 71.50±1.46 89.4 accurately. Serum 20 17.28±0.98 86.4 For PN, sample volumes of 2–3 ml are required, but with 40 36.09±1.41 90.2 DIN this was reduced to 20 ml without loss of sensitivity.15 80 72.89±1.32 91.1 The reduction of sample size decreased matrix eVects and also reduced the possibility of blockages of the sampling cone and nebulizer tip.The sampling eYciency of DIN is much higher than for conventional sample introduction, since the nebulizer tip is placed only a few millimeters from the base of the plasma. The aerosol has a small cross-sectional area when it enters the plasma, which ensures that nearly all of the aerosol is injected into the axial channel.22,25 After elevated, non-therapeutic, oral dosage with the Bi antiulcer compound RBC, high levels of Bi were observed in both urine and serum, with significantly higher levels in urine. It has been reported that an increase of Bi in human whole blood is observed after intake of Bi antiulcer drugs (CBS).28,29 Thiol-containing ligands such as glutathione bind strongly to Bi and are probably involved in its transport, for example through the membranes of red blood cells.30 Recently it has been found that Bi can bind strongly to human serum transferrin, an iron transport protein (80 KDa),31 although the data given in the present paper suggest that the Bi levels are not high enough to saturate transferrin (#35 mM in serum)32 even Fig. 2 Plots of concentration versus intensity ratio for Bi: # in after elevated doses of Bi. However, metal metabolism can 0.14 mol dm-3 nitric acid; D in rat urine; and % in rat serum with Tl added as an internal standard. vary between diVerent animal species. Table 3 Concentrations of Bi in urine and serum samples before and after oral dosage of the rat with RBC Serum* Urine* Bi concentration S(C) S(Bi1) S(Bi2) U(C1) U(C2) U(C3) U(Bi1) U(Bi2) UF(Bi) mg l-1 0.54 3532 2340 0.42 0.63 0.56 11411 17660 20064 mM 0.0026 16.9 11.2 0.0021 0.003 0.0027 54.6 84.5 96.0 *Sample labels: C control, Bi after dosing, S serum and U urine; numbers indicate samples from diVerent animals (see under Experimental). Journal of Analytical Atomic Spectrometry, October 1997, Vol. 12 111310 McLaren, J. W., Beauchemin, D., and Berman, S.S., Anal. Chem., ICP-MS has previously been used for on-line monitoring 1987, 59, 610. studies of trace element species in protein matrices separated 11 Gelinas, Y., Youla, M., Beliveau, R., Schmit, J. P., and Ferraris, by HPLC.33,34 It is possible to couple DIN-ICP-MS with J., Anal. Chim. Acta, 1992, 269, 115. microcolumn chromatography which operates at the same 12 Vanhoe, H., Dams, R., and Versieck, J., J. Anal. At. Spectrom., flow rate as DIN (100 ml min-1).This approach is likely 1994, 9, 23. 13 Mulligan, K. J., Davidson, T. M., and Caruso, J. A., J. Anal. At. to be useful in studies of the distribution of Bi and other Spectrom., 1990, 5, 301. metallo-drugs in biological samples. 14 Delves, H. T., and Campbell, M. J., J. Anal. At. Spectrom., 1988, 3, 343. 15 Powell, M. J., Quan, E. S. K., Boomer, D. W., and Wiederin, CONCLUSION D. R., Anal. Chem., 1992, 64, 2253. 16 Vanhoe, H., Versieck, J., Vanballenberghe, L., and Dams, R., Clin.It has been shown that DIN-ICP-MS is a sensitive method Chim. Acta, 1993, 219, 79. for the determination of Bi in biological fluids with a detection 17 Williams, C. A., Abou-Shakra, F. R., and Ward, N. I., Analyst, limit for absolute amounts about two orders of magnitude 1995, 120, 341. lower than conventional PN. The use of DIN greatly reduced 18 Mauras, Y., Premel-Cabic, S., and Allian, P., Clin. Chim. Acta, memory eVects. Bismuth antiulcer drugs are widely used and 1993, 218, 201. 19 LaFreniere, K. E., Fassel, V. A., and Eckels, D. E., Anal. Chem., it has been demonstrated that Bi in biofluids can readily be 1987, 59, 879. determined using DIN-ICP-MS. Since the required flow rates 20 LaFreniere, K. E., Rice, G. W., and Fassel, V. A., Anal. Chem., are compatible with microcolumn chromatography, it should 1984, 56, 289. be possible to extend this work to speciation studies. It is 21 LaFreniere, K. E., Rice, G. W., and Fassel, V. A., Spectrochim.likely that such work will considerably improve knowledge of Acta, Part B, 1985, 40, 1495. 22 Powell, M. J., Quan, E. S. K., Boomer, D. W., and Wiederin, the biochemical pharmacology of bismuth. D. R., Anal. Chem., 1992, 64, 2253. 23 Gjerde, D. T., Wiederin, D. R., Smith, F. G., and Mattson, B. M., The authors thank GlaxoWellcome and the EPSRC for their J. Chromatogr., 1993, 640, 73. support of this work, the Committee of Vice-Chancellors and 24 Giglio, J. J., Wang, J., and Caruso, J.A., Appl. Spectrosc., 1995, Principals for an Overseas Research Student Award (to H.L.), 49, 314. 25 Wiederin, D. R., Smith, F. G., and Houk, R. S., Anal. Chem., Drs. W. Jenner and M. Dunne (GlaxoWellcome) for supplying 1991, 63, 219. urine and serum samples, and Drs. G. Klinkert and C. Donaghy 26 Christodoulou, J., Kashani, M., Keohane, B. M., and Sadler, P. J., for helpful comments. J. Anal. At. Spectrom., 1996, 11, 1031. 27 Taylor, A., Branch, S., Crews, H.M., Halls, D. J., and White, M., J. Anal. At. Spectrom., 1996, 11, 103R. REFERENCES 28 Hespe, W., Stall, H. J. M., and Hall, D. W. R., L ancet, 1988, 2, 1258. 29 Froomes, P. R. A., Wan, A. T., Keech, A. C., McNeil, J. J., and 1 Marshall, B. J., Am. J. Gastroenterol., 1991, 86, 16. McLean, A. J., Eur. J. Clin. Pharmacol., 1989, 37, 533. 2 WagstaV, A. J., Benfield, P., and Monk, J. P., Drugs, 1988, 36, 132. 30 Sadler, P. J., Sun, H., and Li, H., Chem. Eur. J., 1996, 2, 701. 3 McColm, A. A., McLaren, A., Klinkert, G., Francis, M. R., 31 Li, H., Sadler, P. J., and Sun, H., J. Biol. Chem., 1996, 271, 9483. Connolly, P. C., Grinham, C. J., Campbell, C. J., Selway, S., and 32 Bell, J. D., Brown, J. C. C., and Sadler, P. J., Chem. Brit., 1988, Williamson, R., Aliment. Pharmacol. T herap., 1996, 10, 241. 24, 1021. 4 Sadler, P. J., and Sun, H., J. Chem. Soc., Dalton T rans., 1995, 1395. 33 Mason, A. Z., Storms, S. D., and Jenkins, K. D., Anal. Biochem., 5 Slikkerveer, A., Helmich, R. B., Edelbroek, P. M., van der Voet, 1990, 186, 187. G. B., and de WolV, F. A., Clin. Chim. Acta., 1991, 201, 17. 34 Owen, L. M. W., Crews, H. M., Hutton, R. C., and Walsh, A., 6 Dean, S., Tscherwonyi, P. J., and Riley, W. J., Clin. Chem., 1992, Analyst, 1992, 117, 649. 38, 119. 7 Rooney, R. C., Analyst, 1976, 101, 749. Paper 7/00269F 8 Slikkerveer, A., Helmich, R. B., and de WolV, F. A., Clin. Chem., Received January 10, 1997 1993, 39, 800. 9 Wan, A. T., and Froomes, P., At. Spectrosc., 1991, 12, 77. Accepted April 9, 1997 1114 Journal of Analytical Atomic Spectrometry, October 1997, Vol. 12