Simultaneous determination of Pt and I by ICP-MS for studies of the mechanism of reaction of diiodoplatinum anticancer complexes Marina Patriarca,† Nicole A. Kratochwil and Peter J. Sadler* Department of Chemistry, University of Edinburgh, Joseph Black Building, King’s Buildings, West Mains Rd., Edinburgh, UK EH9 3JJ. E-mail: P.J.Sadler@ed.ac.uk Received 14th October 1998, Accepted 19th January 1999 The determination of Pt by ICP-MS in environmental and biological samples is well documented and generally performed after dissolution in dilute HNO3.On the other hand, I is poorly ionised in the plasma and, at low pH, memory eVects and instability arise from the formation of potentially volatile species, such as I2 and HI, depending on the oxidation state of I. In order to investigate the role of iodo ligands in the design of Pt anticancer complexes, we have optimised conditions for the simultaneous determination of Pt and I. Standards and samples were diluted in 10 mM KOH and improved ion extraction into the quadrupole was achieved by means of an additional pump (S-Option@), leading to about a 2-fold increased sensitivity.The limits of detection in water were 10 and 0.6 ng l-1 for I and Pt, respectively, but increased to 23 ng l-1 for I and 2.2 ng l-1for Pt in KOH (10 mM). The analysis of certified reference materials yielded the following results: 0.84±0.05 mg g-1 I (certified value: 0.81±0.05 mg g-1) in BCR CRM 063R ‘Skimmed Milk Powder’, and 0.121±0.006 mg ml-1 Pt (indicative value: 0.12 mg ml-1) in NIST SRM 2670 ‘Toxic Metals in Urine’. Pt5I ratios ranging from 0.240 to 1.035 were measured with an accuracy of 101.3±2.4%.The determination of the Pt5I ratio in the low Mr fraction of reaction mixtures of diiodo Pt complexes and human albumin provided evidence for the release of iodide and for diVerent kinetics for the reactions of diiodo Pt(IV) and Pt(II) complexes. Introduction The capability of inductively coupled plasma mass spectrometry (ICP-MS) for multi-elemental detection with almost unrivalled sensitivity is of particular interest in biomedical research for the investigation of the functions of biological molecules with aYnity for more than one element (e.g.the metal transport proteins metallothionein and transferrin) and studies of the mechanism of action of newly designed drugs involving metal centres. The full exploitation of this technique, however, requires careful optimisation of analytical param- Fig. 1 Structures of the diiodo Pt complexes used in this work: (1) eters, especially when elements with diVerent chemistry are to trans,cis-[Pt(en)(OH)2I2] and (2) [Pt(en)I2]. be determined. As part of current work on the role of iodo ligands in the design of Pt anticancer complexes,1–3 we have investigated The determination of Pt by ICP-MS is relatively straightoptimum conditions for the simultaneous determination of Pt forward,5 the only reported spectral interferences being due to and I by ICP-MS. Release of iodide may be an important step oxides of less common elements, such as hafnium, ytterbium in the mechanism of reaction of diiodo Pt complexes (Fig. 1) and tungsten.6 In contrast, I is poorly ionised in the plasma with human albumin, the major protein in blood plasma.3 The and the analytical performance is aVected by the oxidation determination of the Pt5I ratio in the protein and low Mr state of I.Poor stability and memory eVects occur at low pH fractions of the reaction mixture can provide evidence for the in the presence of iodide and have been attributed to the release of I and clarify the mechanism of reaction.formation of volatile species such as HI and I2.7,8 Therefore, Previous work has been reported on the determination of the dilution of standards and samples in alkaline media is either I or Pt in biological matrices by diVerent techniques, recommended.7,9–11 Both inorganic and organic bases, such as including spectrophotometry, voltammetry, gas chromatogra- NH3,7–9 KOH or NaOH, at concentrations between 50 and phy, neutron activation analysis, atomic absorption spec- 70 mM,11,12 tetramethylammonium hydroxide11 and a mixture trometry, inductively coupled plasma optical emission of water-soluble tertiary amines10 have been used to eliminate spectrometry and ICP-MS.4,5 However, few of these allow the memory eVects.simultaneous determination of both elements with comparable In this paper, we focused on the validation of a method for sensitivity. The use of diVerent techniques for the determi- the simultaneous determination of I and Pt, which could nation of the two elements would be prone to the introduction provide a comparable degree of reliability for both elements of a wider range of uncertainty and would require larger and allow the accurate determination of their ratio.Previous volumes of samples.reports have indicated influences of the matrix on the analytical behaviour of individual elements, according to specific characteristics. For example, the addition of organics, such as †On leave from Istituto Superiore di Sanita`, Rome, Italy. J. Anal. At. Spectrom., 1999, 14, 633–637 633glycerol,13 Triton X-100,14 methanol15,16 and organic amines,10 exchange–reverse osmosis system (Elga, Bucks., UK) and used in all experiments. enhances selectively the sensitivity of elements with ionisation potentials>9 eV, including I, by increasing the torch tempera- Diiodo Pt complexes, [Pt(en)I2] and trans,cis- [Pt(en)(OH)2I2], were synthesised according to previously ture.Therefore, since Pt and I diVer in ionisation potential (8.70 eV; 10.08 eV) and degree of ionisation (62%; 29%),17 described procedures.1 Recombinant human albumin (rHA, batch R970103) was supplied by Delta Biotechnology particular attention was paid to the influence of increasing concentrations of KOH on signal intensity and to the eYcacy (Nottingham, UK).of internal standardisation. Procedures Experimental ICP-MS analysis. Working standard solutions containing both I and Pt in a molar ratio of 251 were prepared freshly Instrumentation each day as follows. A stock standard solution at a concentration of 100 mg l-1 was made up from KI in ultrapure water, A PlasmaQuad 3 inductively coupled plasma mass stored frozen at -20 °C and replaced every 30 d.A solution spectrometer from VG Elemental (Winsford, UK) was used containing I (1.3 mg l-1) and Pt (1.00 mg l-1) was prepared for all measurements. The instrumental apparatus included a by dilution of the stock solutions (I 100 mg l-1, Pt high eYciency interface (S-Option@), which improves sensi- 1000 mg l-1) in water, from which working standard solutions tivity by increasing the eYciency of the extraction of ions in at concentrations of 0, 1.3, 3.2, 6.4 and 12.7 mg l-1 I and 0, the plasma through to the mass spectrometer itself.This 2.5, 5.0 and 10.0 mg l-1 Pt, were prepared by dilution with feature is achieved by means of improved pumping of the 10 mM KOH containing 20 mg l-1 Te as internal standard expansion chamber via a high capacity rotary pump, which (diluent). The certified reference materials were prepared for reduces the pressure within the expansion region to less analysis as follows. Portions of about 200 mg of CRM 063R than 1 mbar.‘Skimmed Milk Powder’ were weighed and dissolved overnight Instrumental settings, operating conditions and acquisition in 40 g of the diluent to obtain clear solutions; the moisture parameters are given in Table 1. Lens settings were optimised content of CRM 063R was determined according to the every day at mass 127 with a solution containing 12.7 mg l-1 manufacturer’s instructions and used to correct results to dry I and 9.75 mg l-1 Pt in KOH (10 mM). mass. After reconstitution, the SRM 2670 ‘Toxic Metals in Urine’ (Elevated Level ) was diluted 1+24 with 10 mM KOH Reagents containing 20 mg l-1 Te.Potassium iodide (99.5%) and single element standards for ICP at 1000 mg l-1 of Pt (Pt in 20% HCl) and Te (Te in Studies of the reactions of diiodo Pt complexes with human HCl), all ‘Aristar’ grade, from Merck (Leics., UK), were used albumin. The diiodo Pt complexes were separately incubated for calibration. KOH was of analytical-reagent grade from with rHA at a molar ratio of 151 (100 mM) in 100 mM Fisher Scientific (Loughborough, UK). The reference mate- NaCl–10 mM NaH2PO4 at 37 °C, pH 7.4.Aliquots (900 ml ) rials NIST SRM 2670 ‘Toxic Metals in Urine’ (National were taken from the reaction mixture (10 ml ) at 0, 0.5, 1, 1.5, Institute of Standards and Technology, Gaithersburg, MD, 2.5, 6.0, 20.5 and 24 h. To evaluate the distribution of Pt and USA), and BCR CRM 063R ‘Skimmed Milk Powder’ (EU I among the high and low Mr fractions at diVerent stages of Institute for Reference Materials and Measurements, Geel, the reaction with rHA, the two fractions were separated by Belgium) were used to assess the accuracy of Pt and I ultrafiltration using a Centricon-30 concentrator (10 min at determination, respectively.Ultrapure water (resistivity 2516 g, 4 °C). Since significant binding of iodide to human 18 MV, TOC <5 ppb) was obtained by a combined ion serum albumin has been shown to occur (Ka1=6.15×103),18 reversibly bound I was removed from the protein fraction by repeated washing with the reaction buVer, i.e.after the first Table 1 VG PlasmaQuad 3 instrumental settings, operating conditions centrifugation, 900 ml of buVer were added to the protein and acquisition parameters fraction and the centrifugation was repeated. This step was repeated three times for each sample, the last centrifugation Operating conditions— Plasma: Rf power Forward: 1350 W carried out for 15 min and all ultrafiltrates were combined.Reflected: <2 W The final volumes of both fractions were recorded. Ar gas flow rates Plasma: 13.1 l min-1 Determinations of Pt and I were carried out by ICP-MS on Auxiliary: 0.80 l min-1 both fractions of all samples, after dilution with 10 mM KOH Nebuliser: 0.80 l min-1 containing 20 mg l-1 Te. Sample Peristaltic pump Minipuls 3, (Gilson) introduction: Nebuliser Meinhard concentric glass nebuliser Results Spray chamber type Scott double-pass type, Choice of experimental conditions water cooled at 5 °C Sample uptake rate 0.6 ml min-1 Initial studies were carried out to assess the feasibility of the Vacuum: Expansion stage 0.8 mbar (S-Option@ on) simultaneous determination of both I and Pt in KOH and to Intermediate stage 10-5 mbar determine the minimum concentration of KOH needed.The Analyser stage 3.4×10-6 mbar signal intensity profile obtained for a 10 mg l-1 solution of KI Sampling depth 5 mm from load coil in 50 mM KOH (Fig. 2) showed a flat top peak with little Acquisition Isotopes 127I, 195Pt, 126Te fronting and a shorter tail, as compared with 10 mM KOH, parameters: Time per sweep 0.2 s thus indicating a more eVective stabilisation of I species, but Acquisition mode Peak jumping poorer sensitivity. The estimated time for complete wash-out Dwell time 10.24 ms Points per peak 3 was 3 min in 50 mM KOH and 8 min in 10 mM KOH. The Uptake 120 s analysis of a set of standard solutions, prepared in either 10, Acquisition time 10 s 20 or 50 mM KOH, indicated a drop in signal intensity with Rinsing time 420 s increasing concentration of KOH, aVecting Pt and I to diVerent Replicates 5 extents (Table 2).Total dissolved solids increased from 0.06% 634 J. Anal. At. Spectrom., 1999, 14, 633–637Fig. 2 Signal intensity profiles for 10 mg l-1 I in 10 and 50 mM KOH. Table 2 Changes of signal intensity (%) for I and Pt in increasing concentrations of KOH relative to 10 mM KOH (mean±s, n=5) KOH concentration Fig. 4 Long-term stability profile for 127I, 195Pt and 126Te. 10 mM 20 mM 50 mM Te (internal standard) in 10 mM KOH, studied over a period I 100 89.1±1.1 68.9±1.7 of 10 h, by repeated analysis of a solution containing all three Pt 100 84.8±3.2 61.5±0.8 elements, was satisfactory for a period of about 5 h, after which diVerences between the analytical behaviour of Pt and Te became evident (Fig. 4). in 10 mM KOH to 0.28% in 50 mM KOH.In addition, poorer long-term stability, higher imprecision, and higher blank levels Analytical performance were observed when 20 or 50 mM KOH was used for the dilution of standards and samples. Therefore, 10 mM KOH In standard sensitivity mode, the limits of detection (LODs, was used for all subsequent work. To prevent memory eVects, 3s of ten measurements of the blank) were 40 ng l-1 for I and the rinsing time between samples was increased to 7 min, using 0.5 ng l-1 for Pt in water, but increased to 58 ng l-1 for I and water rather than KOH to limit the build-up of salts and 4.9 ng l-1 for Pt in KOH (10 mM).When the S-Option@ was damage to the torch. This approach proved to be eVective in used, the LODs for I improved to 10 ng l-1 in water and the range 0–12.7 mg l-1 I. 23 ng l-1 in KOH (10 mM). For Pt, the LODs were 0.6 ng l-1 The variations in signal intensity with the plasma in water and 2.2 ng l-1 in KOH (10 mM). The average concentemperature are shown in Fig. 3 for both Pt and I, in terms trations of I and Pt in the blank (10 mM KOH, 20 mg l-1 Te) of absolute intensity, normalised for concentration and isotope were 0.35 mg l-1 and 10.3 ng l-1, respectively.abundance, for rf power varying from 1050 to 1500 W. Both The analysis of certified reference materials yielded the elements show a maximum at 1450 W, but with little variation following results. For the CRM 063R ‘SkimmedMilk Powder’, within the range 1350–1450 W (+12 and +5%, respectively).we found an I value of 0.84±0.05 mg g-1 (6.0% RSD, n=5) In addition, the signal-to-noise ratio reached maximum values as compared with the certified value of 0.81±0.05 mg g-1. We for both I and Pt between 1350 and 1400 W. With improved found no detectable Pt in this CRM for which no Pt value is ion extraction (S-Option@), we observed an increase in signal reported. The Pt value measured in the SRM 2670 ‘Toxic intensity of 2.6-fold for I and 1.2-fold for Pt.Optimisation of Metals in Urine’ was 0.121±0.006 mg ml-1 (4.7% RSD, n= lens settings at mass 127 proved to be critical to achieve 9) as compared with an indicative value of 0.12 mg ml-1. No maximum sensitivity and good short-term stability for I, I value is reported for this SRM, but we measured an I content whereas Pt was less sensitive to optimisation of lens settings of 0.244±0.014 mg ml-1 (5.7% RSD, n=9). The precision of at either mass 127 or 195. The long-term stability of I, Pt and the Pt5I ratio measured in the SRM 2670 was 0.496±0.013 (RSD 2.7%, n=9).The accuracy of the determination of the Pt5I ratio was tested by analysing a solution of the diiodo Pt(IV) complex, trans,cis-[Pt(en)(OH)2I2], which has a stoichiometric Pt5I ratio of 0.5, and in solutions of the same complex spiked with diVerent amounts of either I or Pt standards, to obtain Pt5I ratios ranging from 0.240 to 1.035 (Table 3). The regression analysis between measured ( y) and expected (x) values gave y=-0.018+1.050 x, r2=0.998.The within-day precisions for I, Pt and Pt5I ratio Table 3 Accuracy of the determination of the Pt5I ratio Measured Pt5I ratio Expected Pt5I ratio Mean s RSD (%) n 0.240 0.245 0.001 0.4 3 0.324 0.329 0.005 1.6 3 0.500 0.492 0.006 1.2 13 0.767 0.762 0.008 1.1 3 Fig. 3 Molar response for 127I, 194Pt and 195Pt as a function of 1.035 1.089 0.009 0.9 3 rf power. J. Anal. At. Spectrom., 1999, 14, 633–637 635tivity for both elements, lower blank values for Pt, and reduced content of total dissolved solids, which in turn ensures better stability.Tellurium has been used as an internal standard for I determination7,9 and is among the group of elements with higher ionisation potential; a disadvantage is that it is a multiisotopic element and the most abundant isotopes (130Te, 128Te) are both prone to isobaric interferences with Xe. The suitability of Te as an internal standard was verified by studying the long-term stability of the signal for a period of 10 h.This showed that 126Te closely followed the analytical behaviour of I, but diVerences from Pt became increasingly evident after 5 h. In addition, we used frequent recalibration (every 20 samples) and the analysis of control samples to monitor drift and variation of performance. For analysis over a long period of time, an additional internal standard should be considered. In this case, 197Au, 193Ir or isotope dilution with a minor Pt isotope could be suitable choices.The quest for higher sensitivity and lower detection limits has stimulated the development of alternative methods to improve these analytical performances, in particular novel methods for sample introduction, including direct injection, Fig. 5 Pt5I ratio in the protein (a) and low Mr (b) fractions of the ultrasonic and microconcentric nebulisers. In this study, we reaction mixture containing trans,cis-[Pt(en)(OH)2I2] (6) or applied a relatively new instrumental option, which provided [Pt(en)I2] (&) and rHA as a function of time.a substantial improvement in sensitivity, especially for the most diYcult ion (I ). Reported LODs for I range from 100 to 4000 ng l-1,7–9,11,16,19 but values of 15 and 18 ng l-1, in determinations, measured as RSD (n=13), were 0.9, 0.9 and NaOH and a mixture of tertiary amines, respectively, have 1.2%, respectively. Between-day precisions were determined by recently been reported.10 The LOD of our method is 23 ng l-1, repeated analysis of trans,cis-[Pt(en)(OH)2I2] on diVerent days which compares well with this last report.Literature data for (n=10) and were 2.3, 2.2 and 3.7%, respectively, for I, Pt and LODs of Pt range from 14 to 50 ng l-1,20–23 using conventional the Pt5I ratio. Attempts to reduce the KOH concentration to instrumentation, although LODs of 0.02 ng l-1 can be 1 or 0.1 mM yielded lower values for the Pt5I ratio obtained by double-focusing magnetic sector field ICP-MS.24 (0.484±0.004 and 0.442±0.004, respectively) and were Our LOD for Pt (2.2 ng l-1) is, therefore, much lower than therefore abandoned.previously reported values for quadrupole ICP-MS, and even lower LODs can be measured in water (0.5 ng l-1). Kinetic studies The method allows the reliable determination of the Pt5I ratio with an overall precision of 3.7% and an average accuracy Initial experiments were carried out to assess the extent of non-specific binding of iodide to albumin.After 24 h incu- of 101.3±2.4%. We studied the kinetics of the interactions between both diiodo Pt(II) and Pt(IV) complexes and rHA, by bation of trans,cis-[Pt(en)(OH)2I2] with rHA, the Pt5I ratio in the highMr fraction was measured before and after repeated determining the ratio of the two elements in the protein and low Mr fractions of samples taken from the reaction mixture (×3) washing with the reaction buVer.The Pt5I ratio was 0.645 initially, but increased to 4.08 when reversibly bound over a period of 24 h. The determination of the ratio between the two elements rather than their absolute amount or concen- iodide was removed by repeated washing. Fig. 5 shows the variations of the Pt5I ratio in the protein tration overcomes problems associated with uncertainties such as recovery, dilution and volume measurements. and low Mr fractions of the reaction mixture between the diiodo Pt(IV) and Pt(II) complexes with rHA over 24 h.In the The rapid decrease in the Pt5I ratio in the low Mr fraction of the reaction mixture indicates the presence of increasing two fractions, the Pt5I ratio varies rapidly in opposite directions within the first 2.5 h, indicating the release of iodide at amounts of free iodide and provides evidence for the release of I at an early stage of the reaction of both complexes with an early stage of the reaction. The pattern of variation of the Pt5I ratio in the protein fraction (from 1 to 4) also suggests albumin.The Pt5I ratios for the protein fraction suggest that the release of the two I ligands occurs at diVerent stages. The that the two iodide ligands are released from Pt in separate stages. study of Pt5I ratios has highlighted diVerences in the kinetics of the reactions of the diiodo Pt(IV) and Pt(II) complexes with albumin. A more detailed discussion of the reaction Discussion mechanisms will be reported elsewhere.3 Previous work has shown that I is best determined at alkaline pH, to prevent severe memory eVects and signal instability.7,9,10 Acknowledgements The drawbacks of this approach are mainly the deterioration of analytical performances, because of the wearing of glassware We thank the EU (Marie Curie Research Training Grant, and build-up of deposits on cones, and higher blank values, ERB4001GT963865 to NAK, and COST Action D8), BBSRC due to impurities in the reagents.In this work, we used KOH, and EPSRC, for their support to this work, and Delta which could be obtained with a higher purity than NaOH. Biotechnology for the gift of recombinant human albumin. Ammonia was avoided, because it forms stable complexes with Pt. The comparison of diVerent concentration of KOH has References shown that at least 10 mM KOH is necessary for the stabilisation of the I signal. Under these conditions, the complete 1 N. A. Kratochwil, M. Zabel, J.J. Range and P. J. Bednarski, elimination of residual memory eVects requires a rather long J. Med. Chem., 1996, 39, 2499. 2 N. A. Kratochwil, Z. Guo, P. S. Murdoch, J. A. Parkinson, washing time (7 min) but has the advantages of higher sensi- 636 J. Anal. At. Spectrom., 1999, 14, 633–637P. J. Bednarski and P. J. Sadler, J. Am. Chem. Soc., 1998, 120, 14 M. J. Campbell, C. Demesmay and M. Olle, J. Anal. At. Spectrom., 1994, 9, 1379. 8253. 15 E. H. Larsen and S. Stu�rup, J.Anal. At. Spectrom., 1994, 9, 1099. 3 N. A. Kratochwil, A. I. Ivanov, M. Patriarca, J. A. Parkinson and 16 E. H. Larsen and M. B. Ludwigsen, J. Anal. At. Spectrom., 1997, P. J. Sadler, unpublished work. 12, 435. 4 Encyclopedia of Analytical Science, ed. A. Townshend, 17 R. S. Houk, Anal. Chem., 1986, 58, 97A. P. J.Worsfold and S. J. Haswell, Academic Press, London, 1995. 18 D. C. Carter and J. X. Ho, Adv. Protein Chem., 1994, 45, 153. 5 M. Balcerzak, Analyst, 1997, 122, 67R. 19 G. Ra�dlinger and K. G. Heumann, Anal. Chem., 1998, 70, 2221. 6 H. Mukai, Y. Ambe and M. Morita, J. Anal. At. Spectrom., 1990, 20 T. Minami, K. Hashii, I. Tateyama, E. Kadota, Y. Tohno, 5, 75. S. Tohno, M. Utsumi, M. Yamada, M. Ichii, K. Namikawa and 7 H. Vanhoe, F. Van Allemeersch, J. Versieck and R. Dams, Y. Okazaki, Biol. Trace Elem. Res., 1994, 42, 253. Analyst, 1993, 118, 1015. 21 T. Minami, M. Ichii and Y. Okazaki, Biol. Trace Elem. Res., 1995, 8 M. Haldimann, B. Zimmerli, C. Als and H. Gerber, Clin. Chem., 48, 37. 1998, 44, 817. 22 J. Christodoulou, M. Kashani, B. M. Keohane and P. J. Sadler, 9 H. Baumann, Fresenius’ J. Anal. Chem., 1990, 338, 809. J. Anal. At. Spectrom., 1996, 11, 1031. 10 Y. Gelinas, A. Krushevska and R. M. Barnes, Anal. Chem., 1998, 23 W. R. L. Cairns, C. W. McLeod and B. Hanck, Spectroscopy, 70, 1021. 1997, 12, 16. 11 S. Stu�rup and A. Buchert, Fresenius’ J. Anal. Chem., 1996, 354, 24 J. Bergerow, M. Turfeld and L. Dunemann, J. Anal. At. 323. Spectrom., 1997, 12, 1095. 12 P. Schramel and S. Hasse, Mikrochim. Acta, 1994, 116, 205. 13 P. Allain, L. Jaunault, Y. Mauras, J. M. Mermet and T. Delaporte, Anal. Chem., 1991, 63, 1497. Paper 8/07981A J. Anal. At. Spectrom., 1999, 14, 633–6