The first example of MEEKC–ICP-MS coupling and its application for the analysis of anticancer platinum complexes

Abstract

MEEKC is a powerful electrodriven separation technique with many applications in different disciplines, including medicinal chemistry; however the coupling to highly sensitive and selective mass spectrometric detectors was limited due to the ion suppressive effect of the surfactant SDS. Herein, the first example of the coupling of MEEKC to ICP-MS is presented and an MEEKC method for the separation of Pt(II) and Pt(IV) anticancer drugs and drug candidates was developed. Different compositions of microemulsions were evaluated and the data were compared to those collected with standard ultraviolet/visible spectroscopy (UV/vis) detection. The MEEKC–ICP-MS system was found to be more sensitive than MEEKC–UV/vis and the analysis of UV/vis silent compounds is now achievable. Furthermore, the migration behavior of the Pt(II) and Pt(IV) compounds under investigation is correlated to their differences in structure.

1 Introduction

Traditionally, commercially available CE instruments have been equipped with UV/vis detectors, in the best case with a DAD device. These detectors provide sufficient sensitivity for many applications, however, with the disadvantage of low selectivity. In the last decade, CE hyphenated to mass spectrometers has gained considerable attention and the number of applications have increased steadily ^1-3. CE can be operated in different separation modes and more and more of these techniques are coupled to different mass spectrometric detectors, though some of these separation modes have been considered initially as limited in their compatibility to MS coupling. Today, CE–MS is applied in many different fields of research, including, e.g., the analysis of pharmaceuticals and their metabolites, often in difficult biological matrices, of environmental samples or of technical products ^{4, 5}. One of such applications is the analysis of platinum(II) compounds, often used in the chemotherapeutic schemes of cancer patients ⁶. Platinum complexes have been applied for more than 30 years in the clinic but their mode of action is still only partly known ⁷. The most important part in their activity appears to be related to the binding of such species to the DNA, whereas attachment to proteins, in particular to plasma proteins, being the first potential binding partners after intravenous administration, is thought to contribute to their side effects ⁸. These adverse effects comprise nephrotoxicity, ototoxicity, vomiting, etc., which are often observed during chemotherapy. However, the major limitation of healing tumor patients appears to be intrinsic and acquired resistance ⁷.

In recent years, capillary electrophoresis has become an established technique in metallodrug research ^9-11, especially for metallodrug–protein binding studies ^12-17. Initially, hydrolysis of the established anticancer agents cisplatin, carboplatin and oxaliplatin (neutral compounds which become positively charged after the exchange of the chlorido or biscarboxylato ligands by aqua groups) was studied in CZE mode ¹¹. Other major CE-based investigations include the binding of such compounds and related non-platinum anticancer agents to DNA model compounds ¹¹, i.e., nucleotides, under different conditions, often chosen with the aim to resemble physiological conditions ^{11, 18}. Furthermore, the binding to proteins was established in terms of rate and association constants ^{13, 19}. Whereas the use of unspecific UV/vis detection appears limited, coupling to MS, in particular to element-specific ICP-MS (which also becomes of increasing importance in proteomics ²⁰), provides an additional degree of selectivity and also minute amount sensitivity ^{12, 21}. Hyphenation to the soft and ICP-complementary ionization technique ESI-MS provided structural information, not available from ICP-MS ^{12, 21, 22}.

An important feature in the mode of actions of metallodrugs is the cellular uptake ²³. This process is often influenced by active transport into the cell but also passive accumulation by diffusion is known to play an important role, especially in the case of the platinum compounds ²⁴. The diffusion of drugs through the cell membrane is related to the lipophilicity of a compound, which can be related to log k values in MEEKC which relies on the different distribution of the analytes in microemulsions (oil droplets suspended in an aqueous buffer) [ref LCGC Europe Jan 2003, Mahuzier] . This allows efficient separation of neutral species as recently shown fordifferent isomers of new oxaliplatin derivatives In this case, the capacity factors were shown to be linearly related to the log P values determined by the rather laborious shake flask method ²⁵. To the best of our knowledge, no coupling of MEEKC to ICP-MS was reported so far. Herein, this new methodology was applied in anticancer metallodrug research in a proof of principle study and the potential was evaluated in the analysis of oxaliplatin and several new neutral Pt(IV) anticancer drug candidates.

2 Materials and Methods

Instrumentation

CE separations were carried out on an HP^3D CE system (Agilent, Waldbronn, Germany) equipped with an on-column diode-array detector. Detection was either carried out by UV/vis at 200 nm or with an Agilent 7500ce ICP-MS interfaced utilizing a CETAC CEI-100 microconcentric nebulizer ²⁶. For all measurements with UV-detection, capillaries of 48.5 cm total length (40 cm effective length; 50 μm ID) were used (Polymicro Technologies, Phoenix, AZ, USA), in the case of ICP-MS detection the capillary length was extended to 60 cm. Capillary and sample tray were thermostatted at 37 °C. The sample injection was carried out by hydrodynamic injection at 10 mbar for 15 s or at 25 mbar for 4 s and the separation voltage was +20 kV or +25 kV for UV/vis and ICP-MS detection, respectively. New capillaries were conditioned with 0.1 M HCl, water, 0.1 M NaOH, and again with water (10 min each). As a daily routine, the capillary was flushed with 0.1 M NaOH, water, and BGE for 5 min each before the first run. Before each injection, the capillary was purged with 0.1 M NaOH, water and the BGE for 2 min each. The operational values for the CE–ICP-MS interface are shown in the Table 1. The nebulizer was employed in self-aspiration mode with the sheath liquid closing the electrical circuit and spraying a fine aerosol. Analyses were only started if a sufficiently stable signal (RSD ¹¹⁵In < 5%) was attained. Instrument control as well as data analysis was carried out using ChemStation software.

Table 1.

ICP-MS operational parameters.

Sampler	Ni (0.1 mm orifice)
Skimmer	Ni (0.4 mm orifice)
Plasma RF power, W	1500
Isotopes recorded	34S, 79Br, , 115In, 195Pt
Nebulizer gas flow rate, L/min	1.14
Sheath liquid	10 mM phosphate buffer with 20 ppb In
BGE	microemulsion: 1.43% SDS, 91.15% phosphate buffer (10 mM, pH 7.4), 0.85% heptane, 6.57% butanol

Reagents and Solutions

Sodium dodecyl sulfate (SDS), dodecanophenone, 1-bromododecane, 1-butanol and sodium monohydrogenphosphate were purchased from Sigma-Aldrich (Vienna, Austria). Sodium hydroxide solution (0.1 M), hydrochloric acid, heptane sodium dihydrogenphosphate and ammonium bicarbonate were obtained from Fluka (Buchs, Switzerland), DMSO from Fisher Scientific. The ICP-MS tuning solution containing lithium, yttrium, cerium, thallium and cobalt in 2% HNO₃ (each 10 mg/L) was obtained from Agilent Technologies (Vienna, Austria) and the ¹¹⁵In standard was form CPI international (Santa Rosa, USA). High purity water used throughout this work was obtained from a Millipore Synergy 185 UV Ultrapure Water system (Molsheim, France). Oxaliplatin 1 and the Pt(IV) compounds 2–12 (Figure 1) were synthesized as described elsewhere ^27-30.

Figure 1.

Structures of the platinum anticancer drug oxaliplatin 1 and Pt(IV) drug candidates 2–12.

Microemulsion preparation

For the preparation of the MEEKC solutions, different ratios of SDS, 1-butanol, heptane and a small portion of buffer were mixed by sonication for 5 min. The residual buffer was then slowly added to the mixture. Phosphate buffer (10 and 20 mM for MEEKC–UV/vis and –ICP-MS, respectively; pH 7.4) was used for the preparation of the microemulsions. The solution was kept for 24 h at room temperature before use.

Sample preparation

The analytes were dissolved in the microemulsions to give concentrations between 0.4 and 0.6 mg/ml (UV/vis detection) and 0.1 mg/ml (ICP-MS detection). For the MEEKC–UV/vis studies, DMSO and dodecanophenone were used as markers for the EOF and the microemulsion droplets, respectively: to 1 ml of the sample solution 1 μl DMSO and 10 μl dodecanophenone (18 mg/ml in methanol) were added. In the case of ICP-MS detection DMSO (³⁴S) and 1-bromododecane (⁷⁹Br) were used as markers [1 ml of the sample contained 3 μl DMSO and 1 μl 1-bromododecane (diluted 1 : 1 with 1-butanol)].

Calculations

The retention factor, k, is defined as the ratio n_me/n_aq (n_me = total number of moles of solute in the microemulsion phase, n_aq = total number of moles of solute in the aqueous phase) and this mass distribution coefficient can be calculated according to Eq. (1) [t_o = migration time of an unretained substance (EOF marker), t_me = migration time of the microemulsion, t_R = solute migration time].

$$k=\frac{t_R-t_O}{t_O(1-t_R∕t_{me})}$$ (1)

3 Results and discussion

Oxaliplatin 1 {[(R,R)-cyclohexane-1,2-diamine-κ²N,N](oxalato-κ²O,O)platinum(II), Figure 1} has been used in clinics for several years and more recently Pt(IV) compounds moved into the focus of interest, due to potential oral administration which would improve the living quality of cancer patients significantly. Several examples of Pt(IV) complexes underwent clinical trials, most recently satraplatin, but so far none of them made it to clinical application. One of the chemical features of the compound class exploitable for drug development is that they are relatively inert and are not significantly metabolized after administration but can be reduced in the hypoxic environment of the tumor. The axial ligands were found to influence the redox potential of the Pt^IV/II couples and eventually after reduction structurally related compounds to those used in clinics are formed. Recently, the synthesis of a series of such Pt(IV) compounds was reported and promising in vitro anticancer activity was demonstrated (Figure 1) ^28-30.

Oral administration of drug compounds does not only require a certain degree of chemical inertness but also sufficient bioavailability, a factor which is also dependent on the lipophilicity of a drug. In a recent report, the lipophilicity as expressed by the k factor was related to the structures of oxaliplatin derivatives by using a MEEKC–UV/vis method ²⁵.

MEEKC–MS has only been introduced recently, since the coupling with ESI-MS is limited due to the suppressive effect of the surfactant SDS on the electrospray ionization ³¹. However, the development of atmospheric pressure photo ionization (APPI) with ionization by high-energy photons ³² enabled the hyphenation of the two methods ³³, and different pharmaceuticals were analyzed with this technique ^{34, 35}.

The interface between the CE system and the ICP-MS does not rely on electrospray but the samples are transferred in a constant gas flow to the ion source. It appears as the presence of the surfactant does not influence the ionization efficiency. Note that metal complexes are known to have potential to coordinate to the anionic surfactant ³⁶ which makes them difficult to analyze by ESI-MS in such a matrix. In contrast, ICP-MS is a metal selective and highly sensitive method and the analysis is not altered by such processes, making this technique perfectly suited to study compounds present at low concentration ^{20, 21}.

3.1 Analysis of oxaliplatin by MEEKC–ICP-MS, method development and initial performance tests

Many parameters influence the separation performance of an MEEKC system and the complexity of the microemulsion allows optimization in several directions. Since the developed method is aimed to be eventually used in the analysis of samples in biological environment, the pH of the buffer composition was set to 7.4 and phosphate buffer was used for the preparation of the microemulsion with the surfactant SDS. In order to generate the oil droplets, heptane was added and 1-butanol was the co-surfactant. Before injection, the sample was diluted with the microemulsion in order to not disrupt the separation system ³⁷.

A satisfying result for the analysis of oxaliplatin was obtained with an MEEKC system consisting of 1.43% SDS, 0.85% heptane, 6.57% 1-butanol and 91.15% phosphate buffer (10 mM, pH 7.4; Figure 2). Microemulsions of such composition were demonstrated to possess high stability ³⁸. Due to the low flow through the capillary, the interface between the CE system and the ICP-MS requires addition of a sheath liquid (phosphate buffer containing 20 ppb In), which is also used to introduce an external standard in order to monitor the stability of the spraying conditions. The oxaliplatin sample was spiked with DMSO and 1-bromododecane to mark the EOF and the microemulsion droplets, respectively, and the ICP-MS was operated to record in parallel ¹⁹⁵Pt, ³⁴S and ⁷⁹Br, in order to detect all three sample components. Furthermore, the capacity factors for oxaliplatin 1 and the Pt(IV) complexes 2, 5, 9 and 10 were determined for MEEKC–UV/vis and –ICP-MS. The values were found to be very similar (Table 2), proofing also that 1-bromododecane is a suitable but UV/vis-silent marker for the microemulsion droplets.

Figure 2.

MEEKC–ICP-MS of oxaliplatin 1. Shown are the traces of ¹⁹⁵Pt, the EOF marker (³⁴S) and the marker of the microemulsion droplets (⁷⁹Br).

Table 2.

Comparison of the log k values obtained for 1, 2, 5, 9 and 10 by MEEKC–UV/vis and –ICP-MS with 1.44% SDS, 6.48% 1-butanol, 0.82% heptane, 91.26% phosphate buffer as microemulsion.

Compound	Detection method
	UV/vis	ICP-MS
oxaliplatin 1	−0.863	−0.890
2	−1.130	−1.100
5	0.553	0.616
9	0.098	0.170
10	0.105	0.061

In order to determine the limits of detection (LOD) and of quantification (LOQ) for the MEEKC–UV/vis and the –ICP-MS systems, calibration curves with oxaliplatin samples were prepared. A linear calibration curve with high correlation coefficients was obtained for both detection methods: in the case of UV for the peak area of oxaliplatin as a function of the analyte concentration, covering the range of 0.35–5 mM (Figure 3), in the case of ICP-MS detection for the peak area of oxaliplatin devided by the peak area of 1-bromododecane as a function of the analyte concentration, covering the range of 0.125–1 mM. The LODs and LOQs were significantly lower when ICP-MS detection was used (Table 3), however, addition of an internal standard was required in that case, whereas the MEEKC–UV/vis system was easier to handle in terms of operation.

Figure 3.

Calibration graph for the quantification of oxaliplatin 1 with the MEEKC–UV/vis and –ICP-MS systems.

Table 3.

Limits of detection (LOD) and quantification (LOQ) as determined by MEEKC–UV/vis and –ICP-MS for the anticancer drug oxaliplatin 1.

Method	LOD / mM	LOQ / mM
CE–UV/vis	0.66	2.18
CE–ICP-MS	0.24	0.92

3.2. Separation of oxaliplatin and platinum(IV) complexes with UV detection

In order to maintain the stability of the separation system, oil-in-water microemulsions require specific ratios of the respective components. These ratios can only be varied within a very narrow range and a standard microemulsion consists of approximately 90% aqueous buffer, 3% surfactant, 6% co-surfacant, and ~1% oil ³⁷. The separation conditions of oxaliplatin, its Pt(IV) precursor 12 and a series of Pt(IV) anticancer drug candidates, were optimized by changing the composition of the microemulsion (Figure 4). The microemulsion consisting of 1.44% SDS, 6.48% 1-butanol, 0.82% heptane and 91.26% phosphate buffer was identified as the best compromise with regard to separation performance and analysis time for the separation of 1–8, 10 and 12 (9 and 11 were not added to the sample mixture due to expectable overlaps based on the log k determination; Table 4). Higher SDS concentrations resulted in significantly longer analysis times and peak broadening, whereas the lowering of the SDS concentration to 0.85% caused the disruption of the separation power. These studies have to be considered as preliminary studies with regard to biological applications (therefore all experiments were carried out at 37 °C, and the phosphate buffer was adjusted to pH 7.4), e.g., for the detection of oxaliplatin and its potentially orally administerable Pt(IV) precursor in biological tissue.

Figure 4.

Effect of the SDS surfactant concentration on the separation of the platinum complexes 1–8, 10 and 12 in MEEKC mode. (A) 1.89% SDS, 5.94% 1-butanol, 0.59% heptane, 91.59% phosphate buffer; (B) 1.44% SDS, 6.48% 1-butanol, 0.82% heptane, 91.26% phosphate buffer; (C) 0.85% SDS, 6.68% 1-butanol, 1.06% heptane, 91.41% phosphate buffer.

Table 4.

log k values for 1–12 as obtained by MEEKC analysis with different microemulsion compositions [(A) 1.89% SDS, 5.94% 1-butanol, 0.59% heptane, 91.59% phosphate buffer; (B) 1.44% SDS, 6.48% 1-butanol, 0.82% heptane, 91.26% phosphate buffer; (C) 0.85% SDS, 6.68% 1-butanol, 1.06% heptane, 91.41% phosphate buffer] and the solubility of the platinum complexes in water.

Compound	log k			Solubility in water in mg/mL
	A	B	C
Oxaliplatin 1	−0.700	−0.863	−1.113	4.0
2	−1.082	−1.130	−1.301	0.2
3	−0.572	−0.563	−0.816	0.5
4	0.034	−0.059	−0.242	-
5	0.749	0.553	0.371	-
6	−0.444	−0.516	−0.673	0.3
7	−0.275	−0.340	−0.481	0.1
8	−0.349	−0.415	−0.566	0.5
9	0.110	0.098	−0.141	0.2
10	0.199	0.105	−0.136	0.1
11	−0.518	−0.551	−0.761	0.6
12	−0.010	−0.166	−0.370	2.0

In an attempt to correlate the solubility of the analytes in aqueous solution with the log k values, as determined from analyses with three different MEEKC systems, no clear-cut relationship was observed (Table 4). All MEEKC systems show the same order of migration (Figure 5), and separation of the compounds in structurally related groups allows drawing the following conclusions: in the case of compounds 2–5 the increase of lipophilicity by changing from the methyl ester to ethyl, propyl and butyl rests is clearly reflected by increasing log k values, as was also observed for the methyl and ethyl esters 8 and 9. In the case of the propyl amides of the butane- 6 and pentanedicarboxylate 7 derivatives the latter possesses, as expected, a more lipophilic character. Compounds 10 and 11 are difficult to compare to any of the other compounds due to their structural difference. However, the Pt(IV) precursor 12 for oxaliplatin 1 is significantly more lipophilic than the clinically established compound, as in this case also reflected by its lower water solubility.

Figure 5.

Taking the log k values as a measure for trends of migration behavior of the Pt complexes in different microemulsion compositions.

4 Conclusions

MEEKC was found to be an efficient separation system for the analysis of Pt(II) and Pt(IV) anticancer agents. The application of MEEKC was for a long time limited in the choice of detectors and only recently it was hyphenated to mass spectrometric detectors with atmospheric pressure photo ionization. In this paper the coupling of MEEKC with the element-specific ICP-MS is presented for the first time. The high selectivity and sensitivity of the ICP-MS was used to detect in parallel the platinum compounds and the EOF microemulsion droplet markers DMSO and 1-bromododecane, respectively. Initial performance tests revealed higher sensitivity of the ICP-MS system as compared to UV/vis. However, MEEKC–UV/vis was the more stable system in terms of operation and high linearity of the calibration curve was obtained without addition of internal standards. An advantage of the ICP-MS technology is that UV/vis silent compounds (such as 1-bromododecane) can be detected without prior derivatization.

When comparing the lipophilicity of the Pt compounds, a clear-cut trend is only observable as long as minor structural changes are made, e.g., from methyl to ethyl, propyl and butyl esters. To establish a more general rule, a broader series of compounds is required.