Many heavy metals are considered to be harmful to humans. However, the toxic effects of some metals can be positively used to treat patients suffering from cancer. The first metal-containing anticancer agent was discovered in the 1960s by Rosenberg, Vancamp, and Krigas (1965). While investigating the possible effects of an electric field on growth processes in bacteria, these authors discovered that electrolysis products from platinum (Pt) electrodes produced an inhibition of the cell division process. After the identification of cisplatin (cis-diamminedichloridoplatinum(II)) (Fig. 1) as one of the species responsible for this anti-proliferative effect, the compound was successfully developed into one of the most widely used anticancer agents.
Unfortunately, the use of cisplatin is hampered by severe side effects, such as ototoxicity, nephrotoxicity and neurotoxicity and by the intrinsic and acquired resistance of several tumor types. These limitations have stimulated the search for other metal-containing cytotoxic compounds with better safety profiles and enhanced antitumor characteristics. Thousands of compounds have been synthesized and evaluated in the past 40 years and only few of these agents have entered clinical trials. Besides cisplatin, nowadays, carboplatin (cis-diammine(1,1-cyclobutanedicarboxylato)platinum(II)) and oxaliplatin ([(1R,-2R)-1,2-cyclohexanediamine-N,N′][oxalato(2-)-O,O′]platinum) (Fig. 1) have found important clinical applications in the treatment of cancer. Currently, also the orally administered satraplatin (platinum(IV) cis-dichloro-trans-bis(acetato-O)ammine (cyclohexanamine)) is under consideration for approval for the treatment of hormone refractory prostate cancer (Kelland, 2007). Furthermore, ruthenium (Ru) complexes (Fig. 2) are regarded as promising alternatives for Pt complexes. NAMI-A [Imidazolium-trans(imidazole)(dimethylsulfoxide) tetrachloro ruthenate(III)] (Rademaker-Lakhai et al., 2004), and KP1019 or FFC14A (Indazolium trans-[tetrachlorobis(1H-indazole)ruthenate(III)] (Scheulen et al., 2004; Hartinger et al., 2006) are the first Ru complexes that have finished phase I studies.
As the interest in metal-based anticancer agents grows, there is an increasing need for accurate and sensitive methods for the quantitative determination of these compounds. The determination of Pt or Ru in clinical samples of patients treated with these compounds is required to understand the pharmacokinetics and pharmacodynamics of these drugs. The clinical matrices of interest are usually either tissue or biological fluids such as blood, serum, plasma, plasma ultrafiltrate, and urine. Also adducts of the compounds and metabolites with DNA, proteins, and small molecules are of importance. In addition to the analysis of clinical samples, methods for determination of metal-based anticancer agents can be employed to assess occupational exposure of health care personnel working with these drugs. This can be done by monitoring, for example, blood or urine of personnel to measure the physical uptake of the drugs or by surface sampling to assess contamination of environments where the drugs are processed.
Because of the limited sample availability and the low drug concentrations present in these matrices, sensitive and specific methods are needed. Up to now, numerous techniques have been used for the study of Pt and Ru anticancer agents. The assays can be roughly divided into two groups. The first group comprises methods for the determination of total metal concentrations utilizing techniques such as atomic absorption spectrometry (AAS) (Vermorken, van der Vijgh, & Pinedo, 1980; Crul et al., 2001; Brouwers et al., 2006), voltammetry (Shearan & Smyth, 1988; Nygren et al., 1990; Gerl & Schierl, 2000; Gietema et al., 2000; Gelevert et al., 2001), differential pulse polarography (DPP) (Vrana, Brabec, & Kleinwachter, 1986), neutron activation analysis (NAA) (Xilei, Heydorn, & Rietz, 1992; Rietz, Heydorn, & Krarup-Hansen, 1994), X-ray spectrometry (Gorodetsky et al., 1993), X-ray fluorescence (XRF) (Stewart et al., 1982; Seifert et al., 1983; Jonson, Mattsson, & Unsgaard, 1988), inductively coupled plasma atomic emission spectrometry (ICP-AES) (Dominici et al., 1986), and inductively coupled plasma mass spectrometry (ICP-MS). The second group includes methods for the speciation of the various Pt or Ru species. Usually, a speciation technique such as high performance liquid chromatography (HPLC) is coupled to a diode array detector (Brandsteterova et al., 1991), electrochemical detector (Krull et al., 1983), UV detector (Bouma et al., 2003; Ehrsson & Wallin, 2003), or to an element specific detector such as ICP-MS. By combining speciation techniques with electrospray ionization mass spectrometry (ESI-MS) (Oe et al., 2002; Smith et al., 2003; Minakata et al., 2006) information on structural composition can be achieved.
In this review we will focus on ICP-MS. Since 1990, when the first ICP-MS assay for the analysis of Pt anticancer agents was published (Tothill, Matheson, & Smyth, 1990), ICP-MS has acquired increasing popularity in the field of analysis of metal-based anticancer drugs. It has been applied for the analysis of various Pt and Ru compounds (Figs. 1 and 2). The technique is highly sensitive and is applicable to a wide range of sample matrices including those of biological and environmental origin. As a result of the successful application of ICP-MS in the field of oncology, the number of publications on the quantitative analysis of Pt and Ru using ICP-MS and speciation techniques coupled to ICP-MS has increased tremendously over the last 20 years. The papers have appeared in a large range of scientific journals, covering the many disciplines this research comprises, for example, medicine, pharmacy, and chemistry.
The purpose of the current review is to provide a selected, systematic survey of publications describing the analysis of Pt and Ru using ICP-MS in the field of oncology. The focus is on the determination of the total metal concentration and on the speciation of Pt and Ru compounds in human biological fluids, DNA- and protein-adducts, and environmental samples. Problems encountered when developing an ICP-MS assay with or without combination with a speciation technique are discussed.
ANALYTICAL ICP-MS ASSAYS: GENERAL ASPECTS
As the name implies, ICP-MS is a combination of an inductively coupled plasma (ICP) with a mass spectrometer (MS) (Fig. 3). Typically, the sample is introduced into the ICP by a sample introduction system consisting of a peristaltic pump and a nebuliser, which generates a fine aerosol in a spray chamber. The spray chamber separates the small droplets from the large droplets. Large droplets fall out by gravity and exit through the drain tube at the end of the spray chamber, while the small droplets pass between the outer wall and the central tube and are eventually transported into the sample injector of the plasma torch using a flow of argon gas. The aerosol is then transported to the ICP, which is a plasma ion source. This plasma is formed by the application of a high voltage spark to a tangential flow of argon gas, which causes electrons to be stripped from their argon atoms. These electrons are caught up and accelerated into a magnetic field, formed by a radio frequency (RF) energy which is applied on a RF coil surrounding the plasma torch. This process causes a chain reaction of collision-induced ionization leading to an ICP discharge. The ICP reaches temperatures of 6,000–8,000 K. As the aerosol transits the plasma, the droplets undergo numerous processes which include desolvation, dissociation, atomization, and ionization (Jarvis, Gray, & Houk, 1992). Ions produced by the argon ICP are principally atomic and singly charged, making it an ideal source for atomic MS. Since the ICP works at atmospheric pressure and the MS requires a vacuum, an interface typically consisting of a coaxial assembly of two cones (sampler and skimmer cone) and a series of pressured differentials to allow efficient sampling of the atmospheric pressure plasma gases while minimally perturbing the composition of the sample gases.
After passing through the sampler and skimmer cones, several electrostatic lenses or ion optics focus the ions into the MS, where the ions are separated based on their mass-to-charge (m/z) ratios. Three main mass separation principles are used in ICP-MS systems: quadrupole, magnetic sector, and time of flight (TOF). The quadrupole is the most commonly used type in ICP-MS. It comprises of two pairs of parallel cylindrical rods. The voltages applied to these rods give a dynamic hyperbolic electric field, in which any ion above or below the set mass enters an unstable trajectory and is lost from the ion beam. By varying the voltages applied to these rods, a full mass spectrum can be obtained (Jarvis, Gray, & Houk, 1992). While the quadrupole MS is used in the majority of ICP-MS instruments, some systems utilize a magnetic sector or high resolution (HR) analyzer, typically employed when higher mass resolution is required (Giessmann & Greb, 1994). This analyzer uses a magnetic field, which is dispersive with respect to ion energy and mass and deflects different masses through different angles. The ions subsequently enter an electrostatic analyzer, which is dispersive with respect to ion energy and focuses the ions to the detector. In a TOF MS (Bings, 2005), a uniform electrostatic pulse is applied to all ions at the same time, causing them to be accelerated down a flight tube. Because lighter ions achieve higher velocities and arrive at the detector earlier than heavier elements, the arrival times of the ions are determined by their m/z ratios.
After passing the mass separator, the ions strike the active surface of the detector, typically an electron multiplier. Each ion which hits the channel electron multiplier generates a cascade of electrons leading to a discrete pulse. The pulses are counted and the output signal is given in counts per second.
One of the main limitations of ICP-MS is the appearance of interferences, which can be classified into two major groups. The first group comprises the spectral interferences, which arise from other elements (isobaric interferences), polyatomic ions (e.g., oxides) with the same m/z ratio as the analyte isotope, or doubly charged ions when half of their masses are similar to the mass of the analyte isotope. Elemental isobaric interferences can usually be avoided by choosing an interference free analyte isotope when the analyte of interest is not monoisotopic. Alternatively, because of the constant nature of isotope ratios for most of the naturally occurring elements, elemental isobaric interferences can be easiliy corrected mathematically by monitoring the intensity of an isotope of the interfering element which is free from spectral interferences (Jarvis, Gray, & Houk, 1992). For the three most abundant Pt isotopes (194Pt: abundance 33.0%, 195Pt: 33.8%, 196Pt: 25.2%), only 196Pt is subject to an isobaric interference (196Hg). However, this interference can be corrected online by monitoring 202Hg signals. For the most abundant isotopes of Ru (99Ru: 12.7%, 101Ru: 17.0%, 102Ru: 31.6%, 104Ru, 18.7%), 99Ru, 102Ru, and, 104Ru are subject to isobaric interferences of respectively 99Tc, 102Pd, and 104Pd, which can also be corrected online.
Polyatomic or molecular interferences can be produced by the combination of two or more atoms and/or ions leading to a molecule ion and are usually associated with either the argon plasma, atmospheric gases, or matrix components of the solvent or the sample. These interferences can be overcome by choosing an interference free isotope, removing the matrix (Evans & Giglio, 1993), using alternative sample introduction systems (Evans & Giglio, 1993), using mathematical corrections equations (Dams, Goossens, & Moens, 1995), employing cool plasma conditions (Vanhaecke et al., 2000), using a collision or reaction cell (Feldmann et al., 1999), or by using a high resolution mass analyzer (Nonose & Kubota, 2001). Elements with high masses, such as Pt, are less susceptible to molecular interferences than lower masses, such as Ru (Hsiung et al., 1997; Brouwers et al., 2007c). However, metal oxide interferences, which can occur as a result of incomplete dissociation of the sample matrix or from recombination within the plasma or the interface, can interfere with the analysis of Pt and Ru. Pt isotopes may be subject to interferences from hafnium oxides (Lustig et al., 1997; Rudolph et al., 2005) and tungsten oxides (Rudolph et al., 2005). Ru isotopes can be subject to oxide interferences from krypton, bromine, selenium, strontium, and rubidium. Oxide formation, though, can be minimized by optimizing the gas flow rate, pump rate, and ionization conditions of the plasma. Since metal oxide formation is typically controlled, via the plasma conditions, to be less than 2% and because hafnium background concentrations in biological samples are typically lower than Pt backgrounds (Rodushkin et al., 2004), hafniumoxides will not interfere significantly with Pt signals (Hann et al., 2003a). Prior to the development of an ICP-MS assay for Pt or Ru, however, background concentrations of the elements of potentially interfering metal oxides in the biological matrix should be investigated.
The last type of spectral interferences are the doubly charged ions, which are analyzed at half the mass of the element, since the mass spectrometer measures m/z ratios. Pt isotopes are not susceptible to interference of doubly charged ions as no element with a mass two times the mass of Pt exists. Ru isotopes, however, might be interfered by some doubly charged ions ([Pt]2+, [Hg]2+, and [Pb]2+). The formation of doubly charged ions can, however, be minimized by optimizing ICP-MS parameters such as lens voltages and plasma conditions. However, because background levels of heavy metals such as Pt might vary (Brouwers et al., 2007b) it is advisable to monitor the signals of these elements routinely during analysis.
The second group of interferences are the nonspectral interferences which can be broadly divided into two categories: first the physical signal suppression resulting from (un)dissolved solids or organics present in the matrix. Matrix components may have an impact on the droplet formation in the nebuliser or droplet size selection in the spray chamber, which can affect the transport efficiency and thus the signal intensity (Thomas, 2002). In the case of organic matrices, the viscosity of the sample that is aspirated is modified. In addition, the solids present in the matrix might lead to a deposition of solids on the cones and subsequently result in an altered ion transmission. Furthermore, undissolved solids can clog the nebuliser and torch. A decrease in these physical effects is possible by an adapted sample pretreatment (e.g., dilution), the use of proper calibration techniques (Jarvis, Gray, & Houk, 1992) preferably combined with the use of an internal standard (IS), or by adjustment of the sample introduction system.
The second category of nonspectral interferences are the matrix interferences (Thomas, 2002) which are caused by changes in the loading of the plasma or space-charge effects and result in signal alteration. An extensive loading of the plasma may effect the ionization efficiency of the analyte ions. High concentrations of easily ionisable matrix elements, such as sodium, might result in a decreased ionization efficiency of elements with higher ionization energies and thus a decreased signal of these elements. In general, the lower the degree of ionization of the analyte in the plasma, the greater the effect of a matrix component on the ion count rate of the element will be.
Space-charge effects are frequently seen in the analysis of light elements. The magnitude of signal suppression generally increases with decreasing atomic mass of the analyte ion. This is the result of a poor transmission of ions through the ion optics due to matrix induced space-charge effects. The high-mass matrix element will dominate the ion beam and pushes lighter elements out of the way resulting in a suppression of the signal.
It is difficult to measure and quantify nonspectral matrix interferences. Again, separation of the analytes from the matrix or dilution of samples may reduce this type of nonspectral interferences. Furthermore, internal standardization may be successful in reducing the interferences. The IS, however, must be closely matched in both mass and ionization energy because they are to behave equal to the analyte. Also, the use of matrix matched calibration standards or standard addition might correct the matrix interferences. Although the signal suppression of the analyte will be corrected by proper calibration methods, the actual space-charge effects will not be solved. The most common approach to reduce space-charge effects is to apply voltages to the individual ion lens components. This will steer the analyte ions through the mass analyzer while rejecting a maximum number of matrix ions (Thomas, 2002).
Combination of ICP-MS Detection With Speciation Techniques
ICP-MS can be used as a Pt or Ru specific detector for several speciation technologies. ICP-MS has several advantages over other methods of detection including a wide linear dynamic range, low detection limits, potential for isotope determinations, and multi-element capability. Moreover, the signal intensities are independent of the chemical structure of the analyte incorporating Pt or Ru and hence the method does not require standards of each analyte/metabolite/adduct. ICP-MS can provide quantitative information for structurally noncorrelated metal compounds.
Following development of an ICP-MS assay and before implementation into routine use, the assay needs to be validated to demonstrate that it is suitable for its intended use. Validation is required to ensure the performance of the method. As chromatography is widely used in bioanalysis, validation guidelines have already been extensively described for speciation methods (U.S. Food and Drug Administration et al., 2001). In contrast, no such guidelines are available for ICP-MS. This has led to some discrepancies concerning the definition of validation parameters in literature describing ICP-MS based bioanalytical assays. No stringent procedure is followed for the assessment of limit of detections (LODs), lower limit of quantifications (LLOQs), precision, accuracy, and linearity in the field of ICP-MS. The LOD and LLOQ for instance can be obtained by several approaches such as; signal-to-noise ratios, the standard deviation of the noise, or the standard deviation of the noise and slope of the calibration curve (Rosing et al., 2000). For reported ICP-MS assays it is not always defined which approach has been used. Furthermore, the LOD, LLOQ, and calibration range are reported either in the processed sample matrix (the final matrix entering the ICP-MS) or in the unprocessed sample matrix. The difficulty is, that the matrix in question is not always clearly defined. Another intricacy is that concentrations of compounds are commonly reported in weight per volume (w/v) instead of molar concentrations (moles/v). In case of an elemental detection technique like ICP-MS, it therefore is pivotal to report whether the metal or the metal-containing compound is used for calculation of the concentrations. Unfortunately, this is not always clear from the reported data. Because of these issues it is difficult to compare assays based on their detection limits and other validation parameters.
In our opinion, procedures followed in, for example, the FDA guidelines could, as far as applicable for ICP-MS, serve as an example for the development of a guideline for the validation of ICP-MS assays in biological matrices (U.S.Food and Drug Administration et al., 2001). Validation parameters could include assessment of the LLOQ, carry-over, linearity, specificity, accuracy, precision, crossanalyte/IS interference, and stability.
ANALYTICAL ICP-MS ASSAYS: TOTAL METAL DETERMINATION
Metal-Based Anticancer Agents in Biological Fluids/Cells
After an intravenous infusion, metal-based anticancer compounds form a variety of hydrolyzed intermediates in the blood (Desoize & Madoulet, 2002; Alessio et al., 2004). These reactive species become rapidly partitioned into plasma protein-bound metal, free plasma metal, tissue metal, white blood cell metal, and erythrocyte-sequestered metal. The free metal fraction is generally considered as the pharmacologically active metal fraction (Calvert, Judson, & van der Vijgh, 1993; Graham et al., 2000). Because of the rapid biotransformation and reactivity of the biotransformation products, investigation of the pharmacokinetics of the intact parent compounds or the metabolites is technically difficult and not feasible in routine analysis. Consequently, the assessment of total Pt or Ru, rather than the analysis of the parent compound (e.g., cisplatin) and its metabolites (e.g., aquated cisplatin), is a generally accepted approach for the analysis of the pharmacokinetics of metal-based anticancer agents (De Waal, Maessen, & Kraak, 1990; Yang, Hou, & Jones, 2002) in different biological matrices. The analysis of the total metal content by an elemental technique such as ICP-MS gives insight into the distribution of the drug irrespective of the molecular composition of the drug and its metabolites. Since the first application of ICP-MS for an oncology research question in 1990 (Tothill, Matheson, & Smyth, 1990), ICP-MS has become an accepted and commonly used technique for the analysis of Pt anticancer agents. Table 1 summarizes the literature in which ICP-MS is used as the analytical technique to analyze total Pt and Ru in biological fluids and tissue. Biological fluids predominantly studied are plasma or serum, which contain the protein-bound and free metal fraction, ultrafiltered plasma (pUF), ultracentrifuged plasma (pUC), or protein precipitated plasma (pP), which contain the free metal fraction, and urine which contains metal eliminated by the kidney. The tissues that are primarily studied in addition to tumor cells include renal and nerve tissue, which are of interest due to the renal and neurotoxicity of Pt agents. The capability of ICP-MS to measure ultra-trace Pt levels, allows the evaluation of long-term Pt retention after treatment with Pt agents (Robbins et al., 1992; Tothill et al., 1992b; Ding et al., 1999; Screnci et al., 2000) as well as the determination of Pt levels in small amounts of tissue samples.
Table 1. Total metal determination in biological fluids and cells (NS, not specified; NA, not applicable)
Metal-Based Anticancer Agents Bound to DNA
The mechanism of action of Pt compounds is still not completely understood. It is, however, generally accepted that DNA platination is the ultimate event in the cytotoxic activity of Pt anticancer agents. The hydrolyzed products of the Pt compounds are believed to primary attack the nucleophilic N7 positions from guanine (G) and adenine (A) leading to the formation of monofunctional adducts and bifunctional intra- and interstrand crosslinks (Bose, 2002; Boulikas & Vougiouka, 2003). The four major cisplatin-DNA adducts are: Pt-G (monofunctionally bound cisplatin), Pt-GG (intrastrand crosslink on pGpG sequences), Pt-AG (intrastrand crosslink on pApG sequences), and G-Pt-G (intrastrand crosslinks on pG(pN)pG and interstrand crosslinks) (Fichtinger-Schepman et al., 1985; Eastman, 1987). Pt-GG and Pt-AG represent respectively 60–65% and 20–25% of the total amount of adducts formed. platinum-DNA (Pt-DNA) adducts affect the DNA replication and transcription and, thereby, inhibit tumor growth. As a consequence, in addition to the analysis of Pt in biological fluids and cells, the quantification of Pt-DNA adducts is of major interest. For cisplatin, only 1% of the Pt molecules that enter the cells actually bind to nuclear DNA (Reedijk, 1999; Centerwall et al., 2006). This issue illustrates the need for sensitive techniques to quantify the level of Pt bound to DNA. The high sensitivity of ICP-MS allows the determination of Pt-DNA adducts in a small number of cells. Table 2 summarizes literature in which ICP-MS was used for quantification of the total amount of Pt-DNA adducts in peripheral blood mononuclear cells (PBMCs) or tissues from patients (Bonetti et al., 1996; Liu et al., 2002a; Cooper et al., 2004) or rodents (McDonald et al., 2005; Rice et al., 2006) after treatment with Pt agents. This table also summarizes the quantification of Pt-DNA adducts in various cell types (Walker et al., 1999; Bible et al., 2000; Azim-Araghi et al., 2001; John et al., 2003; McDonald et al., 2005; Yamada et al., 2005; Centerwall et al., 2006; Ta et al., 2006; Bjorn et al., 2007; Zhang & Zhao, 2007) after in vitro incubation with Pt.
Table 2. Determination of total amount of metal bound to DNA (NS, not specified)
Metal-Based Anticancer Agents in Environmental Samples
Another application of ICP-MS for the analysis of total Pt is the monitoring of personnel working with Pt anticancer drugs and the monitoring of the contamination of environments where these drugs are prepared and administered. Because Pt agents play a major role in the treatment of cancer, large amounts of these agents are processed, for example, in hospital pharmacies. Considering the numerous publications regarding the monitoring of the potential exposure of personnel, apparently, the potential health risks for persons manipulating cytotoxic drugs are a concern. Another source of contamination of the environment, which might effect the health of individuals is the release of metal-based anticancer agents by hospitals into waste water. Considerable portions of Pt drugs are eliminated via the patients urine (Lenz et al., 2005) into the waste water. The low concentrations present in biological samples from personnel and in environmental samples such as surface wipes, air filters, and waste water, make ICP-MS to a commonly used method for the quantification of Pt in these samples. Table 3 summarizes the literature published in this field.
Table 3. Determination of total amount of metal in environmental samples (NS, not specified)
For ICP-MS analysis, biological samples cannot be analyzed directly, but require a pretreatment to reduce the matrix effects of endogenous compounds, such as cell constituents, proteins, salts, and lipids. The development of ICP-MS methods for metals in biological matrices is generally focused on the selection of an appropriate sample pretreatment and the selection of calibration procedures to avoid and compensate matrix effects. Additionally, instrumental modifications can be used to further optimize the assay. For the analysis of low concentrations of metal it is important to consider that, whatever sample pretreatment procedure is used, special care has to be taken to avoid contamination of samples. A careful selection of pretreatment devices and reagents should be performed. Glassware should be avoided as it may contain considerable amounts of Pt. Moreover, sample pretreatment needs to be performed in a dedicated area to prevent environmental Pt or Ru originating from, for example, pollution by car exhaust catalysts (Barefoot, 1997), from interfering with the analysis. Besides the prevention of contamination it is relevant that blanks do not contain detectable levels of analyte. Screening of background levels is, therefore, necessary.
Metal-based anticancer agents in biological fluids/cells
As was mentioned before, the direct determination of metals in biological matrices is problematic. High protein contents can easily block the nebuliser and torch and deposit on the cones and thereby affect the performance of the method. The most commonly used pretreatment method for liquid biological samples is dilution. This method is employed in order to lower the concentration of dissolved solids. For the analysis of Pt and Ru in liquid biological samples, dilution with water (Morazzoni et al., 1995; Sessa et al., 1998, 2000; Carr, Tingle, & McKeage, 2002; Liu et al., 2002a,b), diluted HCl (Tothill, Matheson, & Smyth, 1990), diluted HNO3 (Casetta et al., 1991; Allain et al., 1992; Gamelin et al., 1995, 1997; Morrison et al., 2000; Siim et al., 2003; Turci et al., 2003; Bettinelli et al., 2004; Mason et al., 2005; Brouwers et al., 2006, 2007c; Ziegler, Mason, & Baxter, 2002), and reagents such as a mixture of edta diammonium salt and Triton X-100 in water (Johnsson et al., 1998; Screnci et al., 2000; Brouwers et al., 2007c) have been frequently used to reduce the solid content below the for ICP-MS required less than 0.2% (g/v) (McCurdy & Potter, 2001). Dilution factors are a compromise between a minimum sample dilution to assure low quantification limits and a maximum reduction of total dissolved solid content. A simple dilution, however, could be problematic in samples with high protein contents such as whole blood, plasma, or serum because the acids present in most ICP-MS diluents might precipitate proteins which may clog up the nebuliser. Another possibility to reduce the matrix effects is by protein precipitation. Using this method, however, protein-bound elements such as Pt, will not be analyzed which can be an undesired effect.
Although various studies used dilution as sample pretreatment procedure for protein containing fluids without reporting precipitation problems, Tothill, Matheson, and Smyth (1990) showed that addition of a small amount of acid to plasma which was diluted with water resulted in protein precipitation. Hence, instead of dilution, several applications based their pretreatment on acid digestion where after, prior to analysis, samples were diluted with a proper diluent. In bioanalysis of metal-based anticancer drugs, there are several methods commonly used for digestion. In general, the sample is heated with several combinations of concentrated acids. Tothill, Matheson, and Smyth (1990) digested plasma by using 70% HNO3 (v/v) and heated the mixture at 100°C until dryness. Another study used a similar procedure for the digestion of whole blood and plasma (Morrison et al., 2000). Nygren et al. (1990) digested samples by dry ashing and wet ashing. For the first procedure, a combination of concentrated HNO3 and HCl (aqua regia), and temperatures up to 800°C were used to obtain a dried and ashed sample. For wet ashing, the samples were heated with concentrated acids (HNO3/HClO4) resulting in a solution of ashed sample and acid. The methods were in agreement, indicating that no sample was lost due to volatility after dry ashing. An open vessel wet ashing procedure (HNO3/H2O2/HCl) was used in another study to digest urine (Hann et al., 2003a). HCl was used to provide chloro-complexes of Pt and thereby reduce the memory effects in the sample introduction system that were observed when diluting in 1% HNO3 (v/v). A more aggressive microwave digestion procedure was used for serum using concentrated HNO3 in combination with elevated pressure and temperature levels (Hann et al., 2003b). The authors used digestion with subsequent dilution instead of a simple dilution because the total dilution factor was reduced by using the first method.
The pretreatment of tissue samples always includes digestion followed by dilution. The digestion procedure can be performed in several ways, varying from heating at high temperatures in concentrated HNO3 (Tothill, Matheson, & Smyth, 1990; Tothill et al., 1992a,b; Minami et al., 1994; Minami, Ichii, & Okazaki, 1995; Screnci et al., 2000; Liang & Huang, 2002; Siim et al., 2003; Ghezzi et al., 2004; McDonald et al., 2005; Centerwall et al., 2006; Rice et al., 2006; Samimi et al., 2007) to more complex procedures such as digestion with concentrated HNO3 and 30% H2O2 (Ding et al., 1999; Bible et al., 2000) at elevated temperatures or microwave digestion with HNO3, H2O2, and HCl (Hann et al., 2003a). Perry and Balazs (1994) compared the digestion of cells with sodium hydroxide with the sonication of cells followed by digestion with concentrated HNO3. Digestion with sodium hydroxide, however, was not adequate because it resulted in significant signal suppression. The addition of 0.9% (g/v) sodium chloride to the cell lines, with subsequent incubation of the suspension in a vacuum oven at 120°C also appeared to be an appropriate digestion technique for cells (Hanada et al., 1998a,b), indicating that the addition of acid is not a requisite. The matrix effect of the high level of dissolved sodium chloride was reduced by dilution with water. An alternative for digestion of solid biological samples is the use of laser ablation ICP-MS (LA-ICP-MS), which can directly analyze solid samples, requiring no further sample pretreatment. Kidney sections from mice were analyzed for Pt distribution using this technique (Zoriy et al., 2007). The advantage of this technique is that information can be obtained on the distribution of metal among the tissue. The use of this technique, however, requires the availability of an advanced laser ablation system, which is not present in most clinical laboratories.
Metal-based anticancer agents bound to DNA
The analysis of Pt in DNA samples is preceded by the isolation of DNA from PBMCs or tissue samples. For this purpose, commercially DNA kits are available. However, instead of using commercially available DNA kits, a standard isolation procedure can also be used. In general, cells are washed with phosphate buffered saline, suspended in a buffer containing Tris-HCl, NaCl, and edta disodium salt, and lysed with sodium dodecylsulfate and protease. After removing the proteins, the DNA is ethanol or isopropanol precipitated, where after it is dissolved in a water containing diluent. Subsequently, in some applications, samples were diluted with triton in water (Bonetti et al., 1996) or with a HCl solution (McDonald et al., 2005) and analyzed directly. Others diluted samples in a HCl (Walker et al., 1999) or HNO3 (Azim-Araghi et al., 2001; John et al., 2003; Cooper et al., 2004) solution and heated it prior to analysis. A more extensive wet digestion procedure with concentrated HNO3 and overnight heating was performed by Rice et al. (2006). Yamada et al. (2005) digested DNA with concentrated HNO3 and 30% H2O2. Zhang and Zhao (2007) have described an assay for which the DNA was sonicated prior to analysis. To decide which technique is suitable, experiments which assess the matrix effect and recovery of the metals from the resulting solution should be performed. The lack of certified reference compounds makes a proper validation difficult. Even though, Yamada et al. demonstrated that matrix effects of various amounts of DNA did not effect Pt quantification. None of the other studies described validation procedures.
Metal-based anticancer agents in environmental samples
Processing of environmental samples such as surface wipes is focused on the ability to remove all the Pt present on the surface and on the efficiency to extract all the Pt from the swab. In most assays, swabs were wetted with a diluent such as 30 mM NaOH (Ziegler, Mason, & Baxter, 2002; Mason et al., 2003, 2005) or water (Mason et al., 2003; Brouwers et al., 2007a) which was then used to wipe the surface. Subsequently, Pt was extracted from the swab with water (Ziegler, Mason, & Baxter, 2002; Mason et al., 2003, 2005) or diluted HCl (Brouwers et al., 2007a) under, respectively, constant mixing or sonication. Extraction efficiencies of Pt from the swabs were assessed in all studies and were higher than 87%. Only one study reported the actual Pt recovery from the spiked surface, which exceeded 50% for all Pt compounds tested (Brouwers et al., 2007a). The extraction of Pt from gloves was done using water (Ziegler, Mason, & Baxter, 2002; Mason et al., 2003, 2005), leading to Pt recoveries of 67%. Air filters, which were used in a demistifier (Wittgen et al., 2006) and in a clean room (Mason et al., 2005), were extracted with respectively 0.9% NaCl in a n-propanol water mixture (75:25 v/v) and water.
In addition to an adequate pretreatment procedure, the calibration method should be optimized during the development of a reliable ICP-MS assay. In general, calibration procedures are focused on the compensation of matrix effects. As mentioned before, a frequently applied method to circumvent matrix interference problems is internal standardization. An IS can be used to normalize the analyte signal and thereby correct for matrix effects and instrumental signal drifts. It was shown that to correct appropriately for nonspectral matrix effects, the mass number and ionization potential of the IS should be close to that of the analyte (Thompson & Houk, 1987; Vanhaecke et al., 2000; De Boer et al., 2004). For the analysis of Pt, internal standardization is the most commonly used method to overcome matrix interferences. Several IS are described for Pt in literature. In order of decreasing popularity the following IS are used: iridium (191Ir or 193Ir, IP1; 9.1), indium (113In or 115In, IP1; 5.8), europium (153Eu, IP1; 5.7), bismuth (203Bi, IP1; 7.3), gold (197Au, IP1; 9.2), thallium (203Tl, IP1; 6.1), rhodium (103Rh, IP1; 7.5), and rhenium (187Re, IP1; 7.9). Considering the mass and first ionization potential of Pt (195Pt, IP1; 9.0), iridium and gold seem to be the most adequate ISs. Gold, however, tends to suffer from severe memory effects (Casetta et al., 1991). Furthermore, Tothill, Matheson, and Smyth (1990) encountered differences in chemistry between gold and Pt. These issues point out that gold is a less suitable IS than iridium. Even though the resemblance between Pt and the IS is mentioned all over in literature, the frequent use of indium suggests that stable results are achieved using this IS and that a deviating mass and ionization potential is not always critical. Ding et al. (1999) compared several ISs for the quantification of Pt in cells. They, however, concluded that indium did not properly correct for matrix effects leading to a Pt recovery of 87–92% in the reliable quantification range. Iridium and bismuth, in contrast, did show acceptable results (101–108%). Hann et al. (2003a) showed that rhenium and indium did not correct for matrix suppression of Pt in digested urine. The inconsistent results point out that, when developing a new assay, it is advisable to test several ISs for each matrix. For the analysis of Ru (IP: 7.4), yttrium (98Y, IP; 6.4) and germanium (72Ge, IP; 7.9) have been used as IS (Polec-Pawlak et al., 2006; Brouwers et al., 2007c). They both corrected well for matrix effects.
Internal standardization is often used in combination with external calibration, preferably with matrix matched calibration standards. In addition to spiking the analyte in the biological matrix (Morrison et al., 2000; Bettinelli et al., 2004; Brouwers et al., 2006, 2007c; Zoriy et al., 2007) some authors use alternative matrices containing the major components of the biological matrix, such as saline solutions (Nygren et al., 1990; Gamelin et al., 1997, 1998) and artificial plasma (Allain et al., 1992). Others use calibrants in water (Hann et al., 2003a) or in diluted acid (Casetta et al., 1991; Hanada et al., 1998a,b; Ding et al., 1999; Screnci et al., 2000; Yamada et al., 2005). When significant matrix effects are observed in the samples, it is, however, advisable to use matrix matched standards instead of solely water or diluted acid. An alternative method for external standardization is standard addition, whereby the standard is added to the sample at multiple levels. The advantage is that the spiked sample undergoes the same matrix effect as the analyte present in the unknown sample. The disadvantage is that a large amount of sample is needed and that the method is time consuming. Perry and Balazs (1994) used this method for the determination of Pt in cell lines. Unfortunately, no validation results were shown. Yamada et al. (2005) obtained equivalent results for external standardization in diluted acid and standard addition for the determination of DNA-bound Pt. A last solution to overcome the matrix effects is the isotope dilution mass spectrometry (IDMS) technique. With this technique, the sample is spiked with an enriched isotope of the element of interest. The analysis of isotope ratios in the unspiked and spiked sample, as well as in the spike itself, lead to the quantification of the analyte. Hann et al. (2003a) compared IDMS with enriched 196Pt to external quantification with aqueous standards and standard addition for the analysis of three Pt levels in urine. Quantification with external calibrants revealed incorrect results for all spiked levels, regardless whether internal standardization with rhenium or indium was used or not. This could be due to the aqueous matrix used for calibration that might have resulted in different matrix effects of calibrants and samples. Standard addition was only found to be suitable for samples spiked at higher levels, whereas IDMS resulted in a correct quantification of all levels.
The majority of instruments used in the determination of the total amount of metals originating from anticancer agents in biological matrices are quadrupole based instruments. Few applications used double focusing ICP-MS for Pt analysis (John et al., 2003; Hann et al., 2003a; Cooper et al., 2004; Zoriy et al., 2007). Hann et al. (2003a) examined that for digested serum and microdialysates, the low and high resolution mode resulted in similar concentrations. This indicates that in these two matrices, no spectral interferes were present and that the use of the high resolution mode in routine analysis was not required. John et al. (2003) compared a double focusing ICP-MS with a quadrupole ICP-MS for the determination of Pt bound to DNA. The problem of oxide interference was minor for the quadrupole ICP-MS because oxide formation was low. For the double focusing ICP-MS, however, oxide formation was higher (>5%) leading to a higher contribution of hafnium interferences to the Pt signal. The double focusing ICP-MS was more sensitive than the quadrupole instrument. Analysis times for the first instrument, on the other hand, were significantly longer.
In addition to the mass spectrometer, the type of sample introduction system can also greatly affect the sensitivity of the instrument. Furthermore, it can determine the amount of sample needed. Conventional pneumatic nebulisers (concentric or cross flow), which are often used for Pt or Ru analysis (Nygren et al., 1990; Allain et al., 1992; Perry & Balazs, 1994; Gamelin et al., 1995; Morazzoni et al., 1995; Azim-Araghi et al., 2001; Liu et al., 2002a; John et al., 2003) operate at a flow rate of 0.5–1 mL/min (Todoli & Mermet, 2006). The high flow rate in combination with the time required to carry out a complete signal reading (1–5 min), requires a high sample volume (>1 mL). Obviously, this volume is much larger than the available sample volume in many clinical applications. Therefore, undigested samples need to be diluted with a large volume of diluent, which reduced the detection limit of the method. In addition to this limitation, the transport efficiency of common sample introduction systems is low. With common pneumatic nebulisers, only about 1% of the total sample volume is actually transported into the plasma (Jarvis, Gray, & Houk, 1992). To surmount these issues, specially designed systems (micro nebulisers) have been developed (Bjorn et al., 2007), which operate at low sample uptake rates (0.1–0.2 mL/min) and thereby only use a limited sample volume. In addition, ultrasonic nebulisers have been developed, which have a high efficiency, independent of the gas flow. Thus more analyte can be transported to the ICP resulting in lower detection limits. A 25-fold improve of the detection limit for ultrasonic nebulisation (USN) compared to standard concentric nebulisation was observed by Morrison et al. (2000) for the determination of Pt in pUF. USN was also used by Turci et al. (2003). Hann et al. (2003a) tested two introduction systems, USN and micro concentric nebulisation (MCN), for the determination of Pt in urine. USN and MCN revealed signal intensities in the same range, probably because of reduced nebulisation efficiency of the USN due to matrix effects of the digested urine sample. LODs for USN were slightly better than for MCN. MCN based nebulisers were also applied by Brouwers et al. (2006, 2007c) and John et al. (2003). An alternative method for continuous sample introduction is flow injection analysis (FIA) (Ding et al., 1999), in which a discrete sample volume is injected into a continuously flowing carrier stream. A 10-fold improvement of the detection limit was observed compared to continuous nebulisation because no sample dilution was needed using FIA. Another method to decrease the required sample volume and increase the transport efficiency is electrothermal vaporization (ETV) (Ziegler, Mason, & Baxter, 2002; Mason et al., 2005). The sample is deposited into an electrically conductive vaporization cell where the sample is dried and vaporized. An argon gas flow subsequently carries the sample vapor to the plasma.
ANALYTICAL ICP-MS ASSAYS: SPECIATION OF METAL-BASED ANTICANCER AGENTS
Because of the complex nature of Pt and Ru solution chemistry and the lack of suitable certified reference materials to identify the species, often, no priority was given to the speciation of the compounds in clinical samples. However, in addition to the determination of total metal concentrations, the determination of the parent metal compounds and their metabolites is valuable. Besides, the speciation of metal compounds can be applied to investigate the metabolism of the compounds in the body and to characterize adducts of the metals with endogenous species, such as proteins and DNA, to improve knowledge on the mechanism of action of metal-based anticancer agents. Furthermore, it may be used to study the stability of metal-based anticancer agents. In environmental samples, speciation can be practical to study the composition of the molecules in, for example, surface samples and waste water. While most of the work has involved the speciation of cisplatin and cisplatin-adducts, oxaliplatin, carboplatin, satraplatin, ormaplatin, ZD0473, BBR3464, and the Ru compounds NAMI-A and KP1019 have received attention too.
Speciation of Metal-Based Compounds and Metabolites
After administration of Pt drugs, Pt compounds rapidly form a variety of reactive intermediates in the blood stream, including hydrolyzed products. Speciation analysis of these intermediates is gaining interest. Methods applying ICP-MS have been developed for the speciation and quantification of the parent drugs and metabolites of cisplatin (Zhao et al., 1993; Bell et al., 2006), BBR3463 (Vacchina et al., 2003), satraplatin (Cairns, Ebdon, & Hill, 1994, 1996; Galettis et al., 1999; Carr, Tingle, & McKeage, 2002), oxaliplatin (Allain et al., 2000), and ZD0473 (Smith et al., 2003) in matrices such as whole blood, red blood cells, plasma, pUF, pP, and urine. A summary of the assays is presented in Table 4. In addition to the investigation and quantification of the individual Pt agents and their metabolites, speciation analysis with ICP-MS detection has been used to assess the hydrofobicity of several Pt compounds (Screnci et al., 2000).
Table 4. Speciation of metal-based compounds and metabolites (NS, not specified; NA, not applicable)
Speciation of Reaction Products of Metal-Based Anticancer Compounds With DNA and Proteins
As mentioned before, it is generally accepted that DNA platination is the ultimate event in the cytotoxic mechanism of action of Pt anticancer agents. Speciation of the various Pt-DNA adducts formed can be used to gain more insight into the exact cytotoxic mechanism of Pt compounds. Pt-DNA adducts, however, are not the only reaction products of Pt compounds that are interesting to study. Before Pt compounds can bind to DNA, they must pass from the blood through the cytosol of the cell. Reactive Pt complexes can bind to various constituents in the blood or in cells. Among the potential Pt-binding sites are proteins and other compounds containing thiol donor ligands such as cysteine, glutathione, and methionine (Wang, Lu, & Li, 1996; Reedijk, 1999). The exact role that Pt binding to proteins such as albumin and transferrin plays in the mechanism of drug action, remains unclear. Interactions with proteins, however, might play an important role in drug efficacy and side effects (Timerbaev et al., 2004). Methionine is important because of its large concentrations and reactivity. There is evidence that the nephrotoxicity of cisplatin is increased in the presence of its reaction products with methionine (Heudi, Cailleux, & Allain, 1998). Glutathione is believed to be involved in the cellular detoxification of Pt compounds. There is an indication that the polymorphisms of genes encoding glutathione transferase are relevant for clinical response and development of toxicity (Stoehlmacher et al., 2002; Lecomte et al., 2006; Oldenburg et al., 2007). Thus, the detection and identification of reaction products of Pt-containing drugs with thiol compounds is important in studies of toxicities.
The cytotoxic mechanism of Ru compounds is still largely unknown. In contrast to the view that DNA is the main target of these agents, DNA-independent mechanism are also suggested (Hartinger et al., 2006; Khalaila et al., 2006). The binding of NAMI-A to DNA is far weaker than that of Pt complexes (Pluim et al., 2004). Conversely, NAMI-A was shown to bind tightly to serum proteins (Kratz et al., 1994; Messori et al., 2000), suggesting that the binding of Ru compounds to plasma proteins is of utmost importance for its cytotoxic effect. There is experimental evidence that the Ru moiety is transferred into the tumor predominantly via the transferrin pathway (Pongratz et al., 2004). Hence, the investigation of the interaction of Ru compounds with serum proteins is relevant.
Research has utilized methods for speciation and ICP-MS quantification of biotransformation and reaction products of Pt and Ru agents with DNA, proteins and thiol compounds. These methods are summarized in Table 5.
Table 5. Speciation of reaction products of metal-based anticancer compounds with DNA and proteins (NS, not specified; NA, not applicable)
Speciation of Metal-Based Anticancer Compounds in Environmental Samples
Environmental samples, such as hospital waste water may contain Pt agents and their metabolites. For the determination of the toxicity of Pt in the environment, the determination of total Pt is not sufficient. It is of interest to investigate the molecular form in which Pt is present and, furthermore, to assess the ability of cleaning systems to remove these molecules from the water circulation. Falter and Wilken (1999) investigated the chromatographic behavior of cisplatin and carboplatin in water in the presence of several anions. The results of this study are used for the development of an extraction procedure for cisplatin and carboplatin from environmental matrices. Hann et al. studied low concentrations of cisplatin (Hann et al., 2003b, 2005), oxaliplatin (Hann et al., 2005), and carboplatin (Hann et al., 2005) and their degradation products in water containing varying concentrations of chloride and in human urine in order to support the development of elimination procedures as well as toxicological studies (Hann et al., 2003a, 2005). Lenz et al. (2007) investigated the metabolism and adsorption of Pt compounds in biological waste water. These methods are summarized in Table 6.
Table 6. Speciation of metal-based anticancer compounds in environmental samples (NS, not specified)
The development of speciation methods using ICP-MS for Pt or Ru detection is generally focused on the selection of a speciation system that is capable of separating the compounds of interest and that is compatible with ICP-MS. This compatibility can be improved by instrumental modifications. In addition, it is relevant that the mobile phase is not reactive with the Pt compounds and their metabolites.
High performance liquid chromatography (HPLC) combined with ICP-MS detection is a versatile speciation technique. The chromatographic columns used in the investigations published in literature on the speciation of metal-based compounds in combination with ICP-MS included reversed phase, size exclusion, ion-exchange, and ion-pair chromatography (Table 5). The ion-pairing reagents were anionic and cationic surfactants.
ICP-MS can be used as an off-line or on-line detector. When online detection is used, the coupling of the chromatographic method to the ICP-MS is achieved by connecting the outlet of the column to the liquid sample inlet of the nebuliser. The feasibility of coupling HPLC-ICP-MS is mainly affected by the composition and flow rate of the mobile phases used to perform chromatographic separation. The high amounts of organic solvents frequently used in reversed or normal phase HPLC results in physical changes of the plasma. This might cause plasma instability or even extinguishing of the plasma (Zoorob, McKiernan, & Caruso, 1998). Further problems are encountered with organic solvents when performing gradient elution. As the eluent composition changes, the nature of the plasma changes, which could lead to a variable sensitivity during the gradient elution and thus calibration problems (Cairns, Ebdon, & Hill, 1996). Furthermore, the presence of high levels of organic solvent or dissolved solids (e.g., salts) can result in constriction of the cone orifices owing to build-up of solids. High solid contents can also lead to clogging of the nebuliser. These issues affect the robustness, sensitivity, and precision of the technique. Therefore, the organic solvent load and salt content of the mobile phase should be kept to a minimum.
Reversed Phase Chromatography (RPC)
RPC is the most commonly used speciation mechanism in liquid chromatography and consists of a hydrophobic stationary phase bonded to a solid support. The mobile phases are less hydrophobic, usually water containing different amounts of organic modifiers such as methanol or acetonitril. The sample compounds are partitioned between the mobile phase and stationary phase. Analyte retention is determined by the affinity for each of the two phases and can be altered by changes in, for example, pH. The main limitation of RPC-HPLC-ICP-MS is that most organic modifiers are not ICP-MS compatible. Despite this limitation, however, RPC is the most frequently used technique in the speciation of metabolites of metal-based anticancer agents. Several approaches have been used to surmount the compatibility problems. Screnci et al. (2000) avoided the use of large amounts of organic solvent and used a mobile phase consisting of 100% water to assess differences in hydrophobicity of several Pt compounds in order to evaluate relationships between hydrophobicity, Pt accumulation in dorsal root ganglia, and neurotoxicity. For satraplatin, however, methanol was added. Allain et al. (2000) also used isocratic elution with a mobile phase containing a large percentage of water (90%) for the speciation of oxaliplatin biotransformation products. Hann et al. (2003b) hyphenated RPC-HPLC to ICP-MS for the analysis of cisplatin metabolites in waste water. They used gradient elution from 1 mM sodium hydroxide to water, entering the ICP-MS through a Tee. The other entrance of the Tee was used to introduce the enriched 196Pt isotope for online IDMS. Because of low salt concentrations and the absence of organic solvent, no compatibility issues were met. A gradient elution with up to 12% methanol, used for the speciation of several Pt anticancer agents in hospital waste water, was described by the same authors (Hann et al., 2005). Smith et al. (2003) reported the speciation and quantification of metabolites of ZD0473 using isocratic elution with a mobile phase containing 20% of methanol. The deposition of carbon on the cones was prevented by the addition of oxygen to the nebuliser gas. Cairns, Ebdon, and Hill (1994, 1996) used an eluent containing water and acetonitrile for the speciation of metabolites of satraplatin. To allow a gradient elution with acetonitrile concentrations up to 95%, an interface was developed to desolvate the HPLC eluent prior to introduction to the plasma (Cairns, Ebdon, & Hill, 1996). The interface consisted of a common concentric nebuliser, a heated spray chamber, a membrane desolvator, and a condenser. Galettis et al. (1999) also developed a method for the quantification of satraplatin metabolites. They, however, managed to develop and validate a method with gradient elution using methanol concentrations from 25% to 45%, without the necessity to desolvate the mobile phase. Methanol was chosen because of its lower vapor pressure and carbon loading compared to acetonitrile. The gradient from 25% to 45% of methanol suppressed Pt signals by 70%. Because not all metabolites showed a similar signal suppression, to achieve reliable results, this approach does require the availability of reference compounds of all the metabolites tested. Metabolites should be quantified using calibration curves for each separate metabolite. Another application using similar chromatographic conditions was described with offline ICP-MS detection (Carr, Tingle, & McKeage, 2002).
In addition to the speciation of metabolites of metal-based anticancer agents, RPC proved to be suitable for studying the reaction products of cisplatin with methionine (Heudi, Cailleux, & Allain, 1998), using an isocratic elution with an aqueous mobile phase. The interaction of cisplatin with DNA nucleotides was also studied (Garcia Sar et al., 2006) using isocratic elution with low amounts of methanol. The investigators tested two columns with different characteristics and optimal flow rates. The eluent from the C8 column (0.2 mL/min) was introduced into the ICP-MS via a micronebuliser to facilitate the introduction of low flow rates, whereas for the C16 column (1 mL/min) a common concentric nebuliser was used. Both methods were compatible with ICP-MS because low amounts of methanol and isocratic elution were used. With the C8 column, however, a better speciation of nucleobases was achieved.
Reversed Phase Ion-Pairing Chromatography (RPIPC)
RPIPC is used for the speciation of ionic or ionisable compounds for which an ion-pair (IP) is formed between the solute ion and an appropriate counter ion. The resulting IP is partitioned between the mobile and stationary phase. The mobile and stationary phases used in RPIPC are similar to those employed in RPC-HPLC, although an IP reagent is added to the mobile phase. The IP reagent contains a polar and non-polar moiety.
Because of their positive charges, Pt metabolites can be successfully separated using RPIPC. Zhao et al. (1993) applied RPIPC to retain ionic and neutral cisplatin derivatives. Hydrolysis products of cisplatin were separated using sodium dodecylsulfate (SDS) or heptanosulfonate as IP reagent. It was shown that SDS had low activity with cisplatin and its metabolites (El Khateeb et al., 1999). Heptanosulfonate, which resulted in a better speciation, however, did shift the equilibrium of hydrolysis when the analytes were diluted in the mobile phase. Because the shift was slow, this effect could be avoided by diluting the sample immediately before injection. Heptanosulfonate was also used for the speciation of the reaction products of cisplatin with thiol compounds. The organic content of the mobile phases was purposely kept low (3% n-propanol for the hydrolysis products and 10% methanol for the thiol compounds) to ensure the stability of the plasma. SDS in a 3% methanol containing mobile phase was also used in another application for the speciation of cisplatin and monohydrated cisplatin (Bell et al., 2006). Although the mobile phase changed the overall appearance of the argon plasma, the Pt counts were only slightly suppressed and thus adequate results were achieved.
Although the effect of the applied IP reagents on cisplatin metabolism in the previous publications seemed to be insignificant, the possible effect of IP reagents should be taken into account when developing RPIPC methods for the speciation of Pt compounds.
Size Exclusion Chromatography (SEC)
SEC is used to separate molecules according to their effective size in solution using a stationary phase gel with pores of a particular dimension. Molecules that are too large to enter pores elute first, while small molecules interact with the stationary phase and elute depended on their size. This technique is useful for the speciation of proteins. The mobile phases used for SEC are usually buffered solutions such as Tris–HCl or phosphate buffers (Xie et al., 2007). Some applications, however, use high saline solutions as a mobile phase, which limits the hyphenation of SEC to ICP-MS. Several methods for the speciation and quantification of Pt bound to proteins have been reported. Allain et al. (2000) demonstrated the binding of oxaliplatin to γ-globulins, albumin, and hemoglobin using an aqueous mobile phase with 150 mM sodium hydroxide. No compatibility problems were described. The interaction of cisplatin with hemoglobin was also extensively studied with SEC using a mobile phase of 10 mM ammonium bicarbonate (Mandal, Teixeira, & Li, 2003; Mandal, Kalke, & Li, 2004). A similar procedure was utilized to separate oxaliplatin and oxaliplatin bound to holo-transferrin (Zhao, Mandal, & Li, 2005). The interaction of carboplatin with plasma proteins analyzed by SEC was also reported (Xie et al., 2007). For this application a mobile phase of 50 mM Tris–HCl was used, because a mobile phase containing phosphate buffer resulted in an unstable plasma. Szpunar et al. (1999) investigated various SEC systems for the speciation of the protein-bound and unbound fractions of Pt and Ru drugs prior to online ICP-MS detection. The mobile phase used (30 mM Tris–HCl buffer) was well tolerated by the ICP-MS. The speciation of Pt and Ru compounds using SEC was complicated because parent compounds and metabolites were partly retained on the stationary phase. Furthermore, the authors demonstrated that when using SEC it was not possible to separate albumin- and transferring-bound drug because of a proximity in molar masses. Another speciation mechanism such as anion-exchange could solve this limitation. Other authors also mentioned that SEC alone could not adequately distinguish between metallo-proteins that show only small differences in amino acid sequence. They, therefore, combined SEC with anion-exchange chromatography coupled to ICP-MS to separate and quantify Ru-albumine and Ru-transferrin adducts (Hartinger et al., 2005; Sulyok et al., 2005).
Ion-Exchange Chromatography (IEC)
IEC is based on the interactions of charged functional groups of the stationary phase with charged analytes. Two types of IEC exist; cation-exchange, where positively charged analytes react with anionic sites on the column and anion-exchange, where cationic sites on the column are used to react with negatively charged analytes. Mobile phases generally consist of an aqueous salt buffer, which might cause difficulties when hyphenating IEC to ICP-MS. Additionally, an organic modifier is often added to the mobile phase. The pH of the mobile phase is of great interest because it affects the dissociation of the compounds analyzed.
A cation-exchange method with ICP-MS detection was reported for the speciation of BBR3464 and drug related products (Vacchina et al., 2003). Because of the strong cationic character of the compounds, RPC-HPLC appeared to be unsuitable. Speciation of the compounds could be achieved by the addition of an IP reagent and 35% acetonitrile to the mobile phase. This, however, resulted in a progressive loss of sensitivity. The use of cation-exchange chromatography, with a gradient from 20 to 200 mM pyridine, resulted in a proper separation. Another cation-exchange method was applied to separate cisplatin, its metabolites, methionine, and their adducts (Stefanka et al., 2004). The method employed gradient elution from 0.01 to 0.05 mM HCl and the mobile phase did not affect cisplatin kinetics. No problems for the hyphenation of cation-exchange chromatography to ICP-MS for any of the described applications were reported.
Falter and Wilken (1999) applied a solvent-generated anion exchanger for the separation of cisplatin and carboplatin. The column was pretreated with hexadecyltrimethylammonium bromide (HTAB) and speciation was carried out by the use of a gradient containing 10–50% methanol in water. The high amounts of organic solvent could be used because of the application of an ultrasonic nebuliser followed by desolvation, minimizing the solvent load entering the ICP. Anion-exchange was also applied to separate the cisplatin reaction products with DNA after enzymatic digestion. The negatively charged nucleotides could be separated using a stationary phase with quaternary ammonium groups (Morrison et al., 1993; Azim-Araghi et al., 2001). The platinated DNA was digested and eluted off the column by a linear salt gradient of NaCl. Fractions were collected and, after dilution, analyzed using ICP-MS. The interaction of cisplatin with guanosine monophosphate has been investigated by Hann et al. (2001) using anion-exchange chromatography coupled to sector field ICP-MS. The column was also functionalized with quaternary ammonium groups. The mobile phase consisted of a carbonate buffer system with 5% acetonitril to reduce retention times. To avoid matrix interferences, only 22% of the mobile phase was directed to the ICP-MS. Before entering the introduction system the mobile phase was diluted by a make-up flow containing 1% HNO3 and IS.
Speciation Techniques Other Than Liquid Chromatography
In addition to methods using liquid chromatography, few other speciation methods hyphenated to ICP-MS detection for the analysis of metal-based anticancer drugs have been described. Gel electrophoresis (SDS–PAGE) was used for the speciation of serum proteins after reaction with Pt (Lustig et al., 1999; Carr, Tingle, & McKeage, 2002). The bands from the gel were cut out and dissolved in 70% HNO3 and then heated at 90°C before being diluted with water (Carr, Tingle, & McKeage, 2002) or they were extracted with aqua regia (Lustig et al., 1999). The Pt content of the bands was measured by ICP-MS. Timerbaev et al. (2004) coupled capillary electrophoresis to ICP-MS using an interface consisting of a microconcentric nebuliser to study the interaction of cisplatin with albumine. A similar method was applied to study the interaction of KP1019 with albumine and transferrin (Polec-Pawlak et al., 2006).
CONCLUSIONS AND PERSPECTIVES
The successful application of ICP-MS in oncology has had an enormous impact on the field of quantitative analysis of metal-based anticancer agents from biological and environmental samples. ICP-MS provides enormous sensitivity, which makes the technique applicable to study metal pharmacokinetics, metal accumulation in cells, and DNA- and protein-binding. Furthermore, environmental monitoring to assess the exposure of hospital personnel to metal-based anticancer agents can be studied using this technique. Pretreatment of the samples depends on the composition of the sample and on the available instrument and is generally focused on the reduction of matrix effects. In the absence of a suitable matrix removal pretreatment procedure, matrix effects can be circumvented by using appropriate calibration methods.
In addition to the analysis of total metal content of samples, the hyphenation of speciation techniques to ICP-MS has extended the application of ICP-MS to the analysis of parent compounds, metabolites, and adducts of metal-based anticancer agents. The development of speciation methods using ICP-MS is commonly focused on the selection of a speciation system that is capable of separating the compounds of interest and that is compatible with ICP-MS.
Publications discussed in this paper appeared in the last 17 years. During this period, the application of ICP-MS to study metal-based anticancer agents increased rapidly. Currently available ICP-MS instruments are capable of quantifying picograms of metallic elements, and thus can detect ambient Pt or Ru reliably in humans. With this capability, current ICP-MS technologies offer great potential for continued investigations of metal-based oncology treatments and the elucidation of their anticancer mechanisms.