Quantitative Profiling of in Vivo Generated Cisplatin− DNA Adducts

Anal. Chem. 2009, 81, 9553–9560

Quantitative Profiling of in Vivo Generated Cisplatin-DNA Adducts Using Different Isotope Dilution Strategies D. García Sar,† M. Montes-Bayo´n,*,† E. Blanco Gonza´lez,† L. M. Sierra,‡ L. Aguado,‡ M. A. Comendador,‡ G. Koellensperger,§ S. Hann,§ and A. Sanz-Medel*,† Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, C/Julian Claverı´a 8, 33006 Oviedo, Spain, Department of Functional Biology (Genetic Area) and Oncology, University Institute (IUOPA), University of Oviedo, Oviedo, Spain, and Institute of Chemistry, BOKUsUniversity of Natural Resources and Applied Life Sciences, Muthgasse 18, 1190 Vienna, Austria Platinum compounds are the major group of metal-based chemotherapeutic drug used in current practice and still a topic of intense investigation. The relative contribution of structurally defined cisplatin adducts with DNA to induce apoptosis and the cellular processing of these lesions is still poorly understood mostly due to the lack of sensitive and accurate analytical tools for in vivo studies. In this regard, two novel sensitive and selective strategies are proposed here to quantify cisplatin-DNA adducts generated in Drosophila melanogaster larvae and in head and neck squamous cell carcinoma cultures. The methods involve the isolation and enzymatic digestion of the DNA in the samples exposed to cisplatin and further quantification by high-performance liquid chromatography with inductively coupled plasma mass spectrometric detection (HPLC-ICPMS). Two different strategies, based on isotope dilution analysis (IDA), have been attempted and evaluated for quantification: species-unspecific (the postcolumn addition of a 194Pt-enriched solution) and the species-specific (by means of a synthesized isotopically enriched cisplatin (194Pt) adduct). For the second approach, the synthesis and characterization of the cisplatin adduct in a custom oligonucleotide containing the sequence (5′-TCCGGTCC-3′) was necessary. The adducted oligo was then added to the DNA samples either before or after enzymatic hydrolysis. The results obtained using these two strategies (mixing before and after enzymatic treatment) permit to address, quantitatively, the column recoveries as well as the efficiency of the enzymatic hydrolysis. Speciesspecific spiking before enzymatic digestion provided accurate and precise analytical results to clearly differentiate between Drosophila samples and carcinoma cell cultures exposed to different cisplatin concentrations. After discovery of the cytotoxic effects of cisplatin (cisdiamminedichloroplatinum(II)) in the 1960s,1 the drug has de* To whom correspondence should be addressed. E-mail: montesmaria@ uniovi.es (M.M.-B.); [email protected] (A.S.-M.). † Department of Physical and Analytical Chemistry, University of Oviedo. 10.1021/ac901360f CCC: $40.75  2009 American Chemical Society Published on Web 11/03/2009

veloped successfully into one of the most commonly used anticancer agents. These days, cisplatin continues to be a firstline chemotherapy for multiple epithelial malignancies and metastatic cancers including lung or ovarian cancer.2 In spite of its success, two major drawbacks are associated to the clinical use of cisplatin in chemotherapy: the acquired resistance to the drug3 and the development of important unwanted side effects such as ototoxicity or nephrotoxicity.4 Trying to overcome such limitations, other Pt-based analogues have been developed and screened over the years. However, out of thousands of tested compounds, only a small fraction has entered clinical trials and merely three Pt drugs (cisplatin, carboplatin, and oxaliplatin) have been approved worldwide.5 The mechanism of action of cisplatin and its analogues is still not fully understood, although it is generally accepted that DNA platination is the ultimate event in the cytotoxic activity that finally yields to cell death.6 Therefore, the acquired resistance to the drug could derive from a DNA platination significantly lower in the resistant variants, compared to the respective sensitive cell lines, or perhaps from enhanced DNA repair in the resistant lines.7 An important feature of cancer cells with acquired drug-resistant phenotypes is the change in chromatin structure, a factor which could contribute to enhance DNA repair and increase DNA damage tolerance.8 Such events are still a topic of major interest and can only be resolved by means of sensitive and accurate analytical tools to approach in vivo studies. Thus, for a better understanding of the molecular resistance to Pt-based drugs or to compare the effect on the DNA of different Pt-analogues, it is ‡ Department of Functional Biology (Genetic Area) and Oncology, University of Oviedo. § BOKUsUniversity of Natural Resources and Applied Life Sciences. (1) Rosenberg, B.; Vancamp, L.; Krigas, T. Nature 1965, 205, 698–69. (2) Yang, P.; Ebbert, J. O.; Sun, Z.; Weinshilboum, R. M. J. Clin. Oncol. 2006, 24, 1761–1769. (3) Kelland, L. Nat. Rev. Cancer 2007, 7, 573–584. (4) Kim, Y. H.; Choi, B. K.; Kim, K. H.; Kang, S. W.; Kwon, B. S. Cancer Res. 2008, 68, 7264–7269. (5) Galanski, M.; Arion, V. B.; Jakupec, M. A.; Keppler, B. K. Curr. Pharm. Des. 2003, 9, 2078–2089. (6) Jung, Y. W.; Lippard, S. J. Chem. Rev. 2007, 107, 1387–1407. (7) Akiyama, S.; Chen, Z. S.; Sumizawa, T.; Furukawa, T. Anti-Cancer Drug Des. 1999, 14, 143–151. (8) Chekhun, V. F.; Lukyanova, N. Y.; Kovalchuk, O.; Tryndyak, V. P.; Pogribny, I. P. Mol. Cancer Ther. 2007, 6, 1089–1098.

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9553

necessary to accurately address the formation of Pt-DNA adducts with the specific drug. In this regard, a variety of analytical methods have been described for qualitative assessment of the presence of Pt-DNA adducts in a given cell including immunochemical methods9,10 or coupled techniques such as highperformance liquid chromatography11 (HPLC) or capillary electrophoresis (CE)12 with mass spectrometric detectors (MS).13 For quantitative purposes, 32P-postlabeling assay14 and LC-MS with stable isotope internal standard (using 13C- or 15N-labeled species)15 are the preferred strategies to detect and quantify bulky DNA adducts in subjects exposed to genotoxic agents. The postlabeling analysis protocol involves DNA hydrolysis, adduct enrichment, 32P postlabeling, and chromatographic separation.14 For detection, the online radioactivity measurement of 32P offers excellent sensitivity features once the DNA adducts have been separated via LC. One important limitation of this method is the variable yield of the different enzymatic reactions (necessary for DNA hydrolysis, to remove the phosphodiester bond and to form the 32P-labeled species) that makes absolute quantitative determination of DNA adducts rather complex. The methods based on liquid chromatography-electrospray ionization mass spectrometry (LC-ESI-MS) strategies allow more precise quantitative results of the specific adduct and a time reduction in the analysis.16 However, the lack of commercially available adducted species can be a major limitation for this quantification strategy. Therefore, since the quantitative analysis of DNA-cisplatin adducts in vivo is still a remaining challenge in bioanalytical chemistry, in this work we present an alternative possibility for precise and accurate Pt-DNA adducts determination based on inductively coupled plasma mass spectrometry (ICPMS) methodologies. The final aim is to overcome the actual limitations of the existing methods (semiquantitative and complex) and to test the suitability for in vivo studies. For this purpose, we have proposed the use of ICPMS which is probably the most sensitive and robust elemental detector for Pt detection associated to biomolecules.17,18 Among others, an additional advantage of using ICPMS detection is the possibility to conduct isotope dilution analysis (IDA) analysis and, thus, to obtain accurate quantitative information about the trace elements associated to biomolecules. This can be done by using two different strategies: species-specific19-21 or species-unspecific (or (9) Veal, G. J.; Dias, C.; Price, L.; Parry, A.; Errington, J.; Hale, J.; Pearson, A. D. J.; Boddy, A. V.; Newell, D. R.; Tilby, M. J. Clin. Cancer Res. 2001, 7, 2205–2212. (10) Liedert, B.; Pluim, D.; Schellens, J.; Thomale, J. Nucleic Acids Res. 2006, 34, e47. (11) Iijima, H.; Patrzyc, H. B.; Dawidzik, J. B.; Budzinski, E. E.; Cheng, H. C.; Freund, H. G.; Box, H. C. Anal. Biochem. 2004, 333, 65–71. (12) Ku ¨ ng, A.; Strickman, D. B.; Galanski, M.; Keppler, B. J. Inorg. Biochem. 2001, 86, 691–698. (13) Iannitti-Tito, P.; Weimann, A.; Wickham, G.; Sheil, M. M. Analyst 2000, 125, 627–634. (14) Nagy, E.; Cornelius, M. G.; Moller, L. Mutagen 2009, 24, 183–189. (15) Chen, H. J. C.; Chang, C. M. Chem. Res. Toxicol. 2004, 17, 963–971. (16) Monien, B. H.; Muller, C.; Engst, W.; Frank, H.; Seidel, A.; Glatt, H. Chem. Res. Toxicol. 2008, 21, 2017–2025. (17) Esteban-Fernandez, D. E.; Montes-Bayon, M.; Blanco Gonzalez, E.; Gomez Gomez, M. M.; Palacios, M. A.; Sanz-Medel, A. J. Anal. At. Spectrom. 2008, 23, 378–384. (18) Garcı´a Sar, D.; Montes-Bayo´n, M.; Agu ado Ortiz, L.; Blanco, E.; Sierra, M.; Sanz-Medel, A. Anal. Bioanal. Chem. 2008, 390, 37–44. (19) Rottmann, L.; Heumann, K. G. Anal. Chem. 1994, 66, 3709–3715.

9554

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

postcolumn)22,23 IDA. Eventually, this elemental information obtained can be converted into biomolecule-specific information, if the stoichiometry of the target element-biomolecule association is known. In the present work both possibilities, species-unspecific and species-specific IDA strategies, are investigated with the aim to obtain precise and accurate quantitative results on possible DNA-cisplatin adducts induced in vivo. The different IDA strategies have been first tested in Drosophila melanogaster after their exposure to cisplatin at biologically relevant concentrations. This is the model organism for most genomic studies due to the high homology of the DNA sequence to the human DNA. Additionally, since the final aim of this work is to establish a good method for addressing possible acquired resistance to the drug in human cells, the developed strategy has been also tested in cultures of brain and neck squamous carcinoma cell lines. EXPERIMENTAL SECTION Instrumentation. In the postcolumn studies, the elemental detection of Pt in the column effluent was performed using an ICPMS 7500ce equipped with a collision cell system from Agilent Technologies (Agilent, Tokyo, Japan). The second part of the study dealing with the synthesis and characterization of the GG-194cisplatin spike was done using an Elan DRC II (PESCIEX, Ontario, Canada) and a sector field ICPMS Thermo Element 2 (Thermo Fischer, Bremen, Germany). The Agilent system was fitted with a MicroMist concentric nebulizer (0.4 mL min-1) with EzyFit, EzyLok arm, and EzyLok connector (Glass Expansion, Melbourne, Australia), whereas the Elan and Element introduction system consisted of a PFA nebulizer (PFA-ST, Elemental Scientific Inc., Omaha, NE) and a cyclonic spray chamber (PE-SCIEX, Japan). The HPLC separations were carried out using also two different HPLC pumps: a dual-piston liquid chromatographic pump (Shimadzu LC-10AD, Shimadzu Corporation, Kyoto, Japan) equipped with a sample injection valve Rheodyne, model 7125 (Cotati, CA), fitted with a 20 µL injection loop and an inert HPLC gradient system (Rheos 2000, Flux Instruments AG, Basel, Switzerland) in combination with a metal-free autosampler (BioLC AS50, Dionex, Sunnyvale, CA) with a 20 µL sample loop. For all the quantitative analysis of the DNA samples an Agilent Zorbax XDB-C18 narrow-bore (2.1 mm × 150 mm, 5 µm) was used. The characterization of the synthesized GG-194cisplatin was done on a similar C18 column (Waters Xterra C18 narrow-bore, 2.1 mm × 150 mm, 5 µm). The ESI quadrupole time-of-flight (ESI-Q-TOF) instrument used for this study was a QStar XL model (Applied Biosystems) equipped with the ion-spray source and using N2 as nebulization gas. The sample was introduced at 5 µL min-1, and the TOF scanned from m/z 200-1200. The instrument was calibrated daily, and the measurements were taken in positive mode using 4.5 kV. (20) Heumann, K. G.; Rottmann, L.; Vogl, J. J. Anal. At. Spectrom. 1994, 9, 1351–1355. (21) Del Castillo Busto, M. E.; Montes-Bayo´n, M.; Sanz-Medel, A. Anal. Chem. 2006, 78, 8218–8226. (22) Bru ¨ chert, W.; Kru ¨ ger, R.; Tholey, A.; Montes-Bayon, M.; Bettmer, J. Electrophoresis 2008, 29, 1451–1459. (23) Koellensperger, G.; Meelich, K.; Galanski, M.; Stefanka, Zs.; Stingeder, G.; Hann, S. J. Anal. At. Spectrom. 2008, 23, 29–36.

Chemicals. Cisplatin, calf thymus DNA, and ammonium acetate were purchased from Sigma-Aldrich (St. Louis, MO). All the safety concerns were considered when dealing with this chemical. The customized oligonucleotide (231.98 µg) was synthesized by Invitrogen (Invitrogen, Barcelona, Spain) with the sequence 5′-TCCGGTCC-3′ (MW ) 2362.6 g/mol), and the Trizol reagent was also obtained by Invitrogen. Isotopically enriched Pt solution (10 ng g-1 in 1% sub-boiling nitric acid) with relative abundances 194Pt 81.5%, 195Pt 14.8%, 196Pt 3.33% was obtained from Teknolab (Drøbak, Norway) and used for postcolumn isotope dilution studies. Isotopically enriched cisplatin was synthesized from 194PtI2(NH3)2 (see Koellensperger et al.23) according to a procedure given by Dullin et al.24 For this purpose, isotopically enriched metallic Pt (194Pt 87.34%, 195Pt 10.6%, 196Pt 1.5%) was purchased from Chemotrade (Duesseldorf, Germany). The mobile phases were prepared by using 18 MΩ · cm distilled deionized water obtained from a Milli-Q system (Millipore, Bedford, MA). For HPLC separation, methanol (HPLC grade, Merck, Darmstadt, German)/water linear gradients (5 min in 0% MeOH and then from 0% to 15% in 30 min) were used in the presence of 20 mM ammonium acetate (pH ) 6.5). Injection volumes of 20 µL and mobile phase flows of 0.24 mL min-1 were used in both C18 columns. Enzymatic hydrolysis of DNA samples was conducted by incubation with Nuclease S1 (GE Healthcare, Barcelona, Spain). The activation buffer to dissolve the enzyme was provided by the manufacturer and contained 10 mM sodium acetate (pH ) 4.6), 150 mM sodium chloride, 0.05 mM zinc sulfate, and 50% glycerol. To remove the excess of enzyme, membrane ultracentrifugation devices were used (Centricon YM-10 centrifugal filter devices, Millipore). Synthesis of the Platinated Oligonucleotide (5′-TCCGGTCC-3′). The purchased oligonucleotide (5′-TCCGGTCC-3′) was obtained as a lyophilized powder (no purification of the oligo was performed after synthesis), and it was dissolved in 2 mL of Milli-Q water. An aliquot of 200 µL of the dissolved oligonucleotide was mixed with 20 µL of cisplatin solution (75 µg mL-1) with either natural or isotopically altered abundances, and the solutions were vigorously mixed and left to react for 24 h at 37 °C. After 24 h, 30 µL of the Nuclease S1, previously dissolved on 500 µL the activation buffer, is added to the reaction mixture and left to react at 45 °C for 14 h. Sample Treatment, DNA Isolation, and Enzymatic Hydrolysis. The in vivo assays were carried out with the yellow and white Oregon K strains (OK-y) of D. melanogaster and cell cultures from brain tumor. Strains were maintained at 24 °C in standard baker-yeast medium with cycles of 12 h of light. Groups of 100 virgin females were mass-mated to 60-80 males, and after 2 days, they were allowed to lay eggs for 24 h in bottles containing 30 mL of Carolina instant Drosophila medium (Carolina Biological Supply Company, U.S.A.) hydrated with the same volume of water. Larvae were fed in this medium until reaching the required age for experiments. Unless stated differently, all the experiments were carried out at 24 °C.

Third instard OK-y larvae obtained as described elsewhere18 were treated adding 1.5 mL of the cisplatin solutions (0, 0.25, 0.50, and 0.75 mM) or of the solvent to each bottle (surface treatment). After 12 and 24 h of treatment larvae were collected using a 20% sucrose solution and a nylon gauze and were frozen at -80 °C until further analysis. DNA from frozen larvae was isolated with the phenol– chloroform method as previously described.18 Briefly, frozen larvae in Eppendorf tubes were smashed with sterile plastic pestles in homogenization buffer (0.1 M NaCl; 30 mM Tris-HCl, pH 8.0; 10 mM EDTA; 10 mM 2-mercaptoethanol; 0.5% Triton X-100; 0.15 mM spermine/spermidine). After centrifugation, the pellet was resuspended in extraction buffer with 200 µg/mL-1 proteinase K and 2% SDS (sodium dodecyl sulfate) and incubated at 50 °C for 1 h. Then 0.2 vol of 8 M potassium acetate was added to the solution at room temperature, and the mix was incubated in ice for 30 min. After one extraction with chloroform, the nucleic acids were precipitated with absolute ethanol during 30 min, dissolved in buffer (10 mM Tris-HCl, pH 7.5 and 1 mM EDTA), and incubated 30 min first with 20 µg/mL-1 RNase and later with 50 µg/mL proteinase K and 0.1% SDS at 37 °C. The solution was extracted twice with phenol-CIA (chloroform-isoamil alcohol) and once with CIA, and the DNA was precipitated with 0.1 vol of 3 M sodium acetate and 2 vol of absolute ethanol. DNA was resuspended in water. The established human head and neck squamous cell carcinoma (HNSCC) cell line used in this study was kindly provided by Dr. M. D. Chiara (Servicio de Otorrinolaringologı´a, Instituto Universitario de Oncologı´a del Principado de AsturiassIUOPA). Cells were grown in DMEM medium supplemented with 10% fetal bovine serum, 2 mmol/L-1 L-glutamine, 20 mmol/L HEPES (pH 7.3), and 100 µmol/L-1 nonessential amino acids according to Canel et al.25 The cell cultures were exposed to 0, 5, 10, and 20 µM cisplatin for 2 h at 37 °C. The DNA was isolated by using Trizol reagent suitable for isolation of RNA, DNA, and proteins in a simple procedure. In this case, a preconcentration step by speed-vac treatment was necessary before enzymatic hydrolysis. In all the DNA samples, the double-stranded DNA was converted in singlestranded by heating at 90 °C for 30 min before hydrolysis. Postcolumn Isotope Dilution Analysis. For postcolumn IDA, a solution containing 10 ng g-1 194Pt was prepared in 1% nitric acid and continuously introduced through a T piece (at 0.1 mL min-1, using a peristaltic pump) at the end of the column, and the mixture was nebulized into the plasma. The intensity chromatograms (counts/s) were converted, after adequate mathematical treatments,26 into mass flow chromatograms (ng min-1) using the online isotope dilution equation. The amount of Pt was obtained in each chromatographic peak by integration using Origin 7.5. Mass bias was measured daily before and after the analysis using the exponential model. Species-Specific IDA. For the species-specific spiking mode, the isotopically labeled platinated oligonucleotide (194Pt) prepared as described before was added to the DNA samples. Two strategies were evaluated: (i) addition before enzymatic hydrolysis and (ii) addition of the hydrolyzed platinated oligo into the hydrolyzed DNA samples. When the platinated oligo was used without enzymatic hydrolysis, it was first purified and

(24) Dullin, A.; Dufrasne, F.; Gelbcke, M.; Gust, R. Arch. Pharm. (Weinheim, Ger.) 2004, 337, 654–667.

(25) Canel, M.; Secades, P.; Rodrigo, J. P.; Cabanillas, R.; Herrero, A.; Suarez, C.; Chiara, M. D. Clin. Cancer Res. 2006, 12, 3272–3279.

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9555

Figure 1. HPLC-ICPMS (Agilent 7500 CE) chromatogram for the synthesized adduct GG-cisplatin obtained by incubation of cisplatin with the oligonucleotide 5′-TCCGGTCC-3′ followed by enzymatic hydrolysis. Postcolumn addition of the 194Pt spike can be observed in the gray trace.

characterized by matrix-assisted laser desorption ionization (MALDI) and gel electrophoresis (GE)-ICPMS as described by Bru ¨ chert et al.22 RESULTS AND DISCUSSION Postcolumn Isotope Dilution Analysis. In the search for accurate quantitative analysis of DNA adducts with cytotoxic compounds such as cisplatin we have focused on the most commonly found cisplatin adduct in vivo involving the intrastrand cross-link between adjacent guanine units of the same DNA strand. For quantification of this species, elemental IDA offers extraordinary favorable features since isotope ratios (the only parameter to be monitored) can be measured by modern ICPMS with high precision and accuracy. Among the described IDA modes, the more straightforward postcolumn addition of the spike (or speciesunspecific) was first evaluated.23 In this case, for obtaining accurate results about the specific Pt-DNA adducts it is necessary to have the knowledge of the metal/biocompound stoichiometry and of the column recovery. Such recovery studies were conducted by determination of the total Pt in a solution of synthesized GG-cisplatin adduct standard (obtained by platination of a custom oligonucleotide). Pt was measured before (by direct calibration) and after its elution from the column (by the postcolumn method using a 194Pt spike as shown in Figure 1). For the postcolumn measurements, a standard solution of enriched 194Pt (10 ng mL-1) (the so-called “spike”) was continuously added through a T piece located after the analytical column and the mixture introduced into the ICPMS. By using the online IDA calculations,20 it was possible to address that 68% ± 2% of the injected Ptwaselutingfromthecolumn.Thesoughtspecies(GG-cisplatin) eluted at 40 min (see Figure 1) and accounts for 84% of the total eluting Pt in the chromatogram. As previously stated, the stoichiometry between the heteroelement and its biomolecule has to be known in order to obtain species quantitative information by unspecific IDA. Previous studies from our group18 confirmed the findings of other authors 9556

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

about the bidentate GG-cisplatin complex as the major adduct with clinical relevance resulting from the interaction of the drug with the DNA in somatic cells.18 However, other Pt-containing species are formed in the real sample as can be seen in Figure 2 where the Pt and the P traces are superimposed. Figure 2A corresponds to the DNA extract from D. melanogaster exposed to 0.5 mM cisplatin for 24 h; it shows that together with the sought species at 40 min (GG-cisplatin) other significant unknown Ptcontaining species are exhibited in the chromatogram. The intense Pt compounds eluting at around 10 and 40 min were analyzed by ESI-Q-TOF obtaining the molecular mass of 575.068 and 904.155 Da, respectively, and a good Pt isotope pattern (see Figure 2B). Such molecular masses correspond to the monoadducted species G-cisplatin and the bisadduct GG-cisplatin whose structure is shown in Figure 2C (MS/MS data available). Therefore, the monoand bisadducted cisplatin-guanine seem to be the most abundant Pt-containing species in the chromatogram. The species eluting at about 35 min was also studied by ESI-Q-TOF, but it was not possible to obtain conclusive evidence about its structure. Other Pt-containing compounds at the void volume could be ascribed to partially digested DNA or other cisplatin associations. Once the two major species were characterized, the postcolumn IDA quantitative strategy was conducted in the different real DNA samples. Table 1 shows the obtained results for the concentration the monoadduct, bisadduct, and total eluted Pt in the different samples (corrected for column recovery, 68%) of D. melanogaster exposed for 12 and 24 h to cisplatin. Very good relationships between the monodentate and bidentate adduct levels and the total concentration of cisplatin used for larvae exposure were encountered at 24 h. However, the monoadduct formed at 12 h seems to be independent from the concentration used for exposure (similar for 0.25 and 0.5 mM) and similar to the maximum concentration at 24 h. Thus, concentrations about 40 ng of monoadduct per mg of DNA might be the level that the larvae can deal before the equilibrium between the mono- and the bisadducted forms shifts toward the bisadduct formation. In terms of analytical figures of merit, the whole method provided detection limits (LDs) of about 0.5 pmol for the bisadduct (GG-cisplatin). These LDs are comparable to those achieved by LC-MS but slightly higher than those provided by 32P postlabeling. However, the method is sensitive enough to detect the adducted species in vivo. The calculated precision for a triplicate of the same DNA sample of the whole method (digestion plus analysis by HPLC-ICPMS) ranged from 6% to 12% for all the concentrations assayed, demonstrating the extraordinary potential of the proposed strategy for absolute determination of DNA cisplatin adducts in real life samples. With the same methodology, samples from head and neck squamous carcinoma cell cultures were analyzed. In this case, the level of exposure was considerably lower in order to maintain drug concentration below toxicity levels. The obtained quantitative results are also given in the lower part of Table 1. As before, the concentration of the GG-cisplatin adduct increased with the exposure concentrations. Figure 3 shows the chromatogram obtained for the cell cultures exposed to 10 µM cisplatin; here only a few Pt-containing species can be detected, in contrast to the larvae sample, but also with the GG-cisplatin adduct at about 40 min being the main species found. The species eluting at 37

Figure 2. Analysis of the real samples: (A) chromatographic profile by HPLC-ICPMS (Agilent 7500 CE) showing the P and the Pt traces of the extracted DNA from a D. melanogaster sample exposed to 0.5 mM cisplatin for 24 h; (B) ESI-TOF of the Pt adducts detected by HPLC-ICPMS at 10 and 40 min, respectively, showing a corresponding m/z 575.068 and 904.155; (C) structure of the major species found corresponding to the G-cisplatin monoadduct and GG-cisplatin bisadduct.

min (also observed in Figure 2A but showing there lower relative intensity) was detected in all the cell culture samples and could be ascribed to the same GG-cisplatin adduct where the phosphodiester bond between guanines has been cleaved by the enzymatic hydrolysis. Due to the low signal intensity, this could not be confirmed by ESI-Q-TOF. Species-Specific Isotope Dilution Analysis of GG-Cisplatin Adducts. In spite of the fact that postcolumn IDA permits cisplatin adducts quantification in real samples, the use of species-specific isotope dilution should offer important advantages in terms of accuracy of the analytical results. In order to evaluate so, two different possibilities of species-specific IDA, summarized in

Figure 4, have been studied: (i) synthesis of the GG-cisplatin adduct by platination with isotopically enriched cisplatin (194Pt) of the custom oligonucleotide, followed by enzymatic hydrolysis; this synthesized GG-cisplatin (enriched in 194Pt) is mixed with the hydrolyzed DNA samples and analyzed by HPLC-ICPMS (see route B in Figure 4); (ii) identical to the first one but adding the platinated (194Pt) oligo into the extracted DNA samples prior to digestion. Then, the enzymatic hydrolysis is conducted in the mixture and the cisplatin adducts are determined by (26) Sariego Mun ˜iz, C.; Marchante-Gayo´n, M.; Garcia Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2001, 16, 587–592.

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9557

Table 1. Concentration of the Main Cisplatin Adducts (Shown in Figure 2) Found in D. melanogaster Samples and Tumor Cell Cultures Exposed to Different Drug Concentrationsa

D. melanogaster 24 h exposure D. melanogaster 12 h exposure

cisplatin concentration (mM)

monoadduct (ng add/mg DNA)

bisadduct (ng add/mg DNA)

total Pt (ng Pt/mg DNA)

0 0.25 0.5 1 0 0.25 0.5

0 12.5 ± 0.9 24.6 ± 3.6 39.0 ± 5.8 0 38.5 37.5

0 10.7 ± 1.6 19.0 ± 2.8 26.9 ± 2.7 0.0 37.5 86.9

0 28.7 ± 3.0 56.2 ± 5.4 84.3 ± 8.2 0 49.6 64.6

bisadduct (ng add/mg DNA)

total Pt (ng Pt/mg DNA)

0 0.99 2.30 4.73

0 1.4 3.2 6.7

cisplatin concentration (µM) cell cultures 2 h exposure

a

monoadduct (ng add/mg DNA)

0 5 10 20

Standard deviations calculated for three different digestions of the same DNA sample.

Figure 3. HPLC-ICPMS (Agilent 7500 CE) chromatogram of the DNA extracted from head and neck squamous cell carcinoma (HNSCC) cell line exposed to 10 µM cisplatin for 2 h.

HPLC-ICPMS as before (see route C in Figure 4). The first strategy would compensate for column recoveries (losses in the column would affect sample and spike in the same way). The second strategy should allow us to take into account losses during hydrolysis and eventually to calculate the enzymatic hydrolysis yield. The first step in both cases was the platination of the oligonucleotide with isotopically enriched cisplatin (194Pt). Following route B in Figure 4, the platinated oligo is enzymatically hydrolyzed and the obtained product was purified by HPLC. The corresponding fraction (eluting at 40 min, as in Figure 1) was then collected and preconcentrated. This fraction was characterized in terms of relative isotope abundances and exact concentration by sector field ICPMS. Additionally, the stability of this synthesized GG-194cisplatin adduct in terms of isotope exchange was also evaluated by mixing it with natural cisplatin at molar ratios (194adduct/natural cisplatin) of 1:1 and 1:3. The mixture was then analyzed by HPLC-ICPMS in order to separate the excess of cisplatin from the sought adduct, containing the altered isotope abundances. Table 2 shows the measured isotope ratios of the GG-194cisplatin peak before and after mixing with natural cisplatin. As can be seen in the table, 9558

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

the isotope ratios are statistically identical in the three cases. In other words, measurable isotope exchange between species should not happen when the spike (GG-194cisplatin) is mixed with the DNA samples even if some unbound cisplatin is still present. Once the GG-194cisplatin-enriched standard was appropriately characterized, it was spiked into the previously digested DNA samples and the mixture was well-homogenized before final determination by HPLC-ICPMS. The third strategy (route C in Figure 4) consisted in the addition of the platinated oligonucleotide (with 194cisplatin) to the DNA sample before enzymatic hydrolysis. For this purpose, the platinated oligonucleotide was separated from the excess of unreacted oligo as previously described,22 and the structure of the final product was characterized by MALDI-TOF (data not shown). The concentration and isotope abundances of the resulting spike were also determined by ICPMS.22 Then, the platinated oligonucleotide was mixed with the DNA sample and the mixture was enzymatically digested and analyzed by HPLC-ICPMS. In this case, any losses of analyte or incomplete enzymatic hydrolysis after spike addition will affect sample and spike in the same way and will be, therefore, compensated. Figure 5 shows the obtained chromatogram of D. melanogaster exposed to 0.5 mM cisplatin where it is possible to observe that the signal intensities corresponding to 194Pt and 195Pt are very similar (theoretical isotope ratio is 1.03) on each chromatographic peak beside on this at about 40 min corresponding to the GG-cisplatin species (due to the presence of the enriched cisplatin only in that species). By applying the isotope dilution equation to peak areas on each isotope at 40 min it was possible to obtain quantitative levels of the GG-cisplatin adduct formed in vivo in the samples. This strategy was applied again to analyze the samples from D. melanogaster exposed to increasing concentrations of cisplatin. Figure 6 summarizes, in a comparative way, the quantitative results obtained with the different strategies, namely, postcolumn (without recovery correction, blue dots), species-specific (adding the spike after hydrolysis, pink dots), and species-specific (adding the spike before hydrolysis, green dots). As can be observed, the three methods present a similar trend: the level of the GG-cisplatin

Figure 4. Schematic diagram of the different applied workflows for the quantitative profiling of the cisplatin-DNA adduct in the samples of D. melanogaster: (A) postcolumn route, (B) species-specific route but adding the spike just before column injection, and (C) species-specific route adding the spike before enzymatic hydrolysis.

Figure 5. HPLC-ICPMS (Agilent 7500) chromatogram of the extracted DNA D. melanogaster exposed to 0.5 mM cisplatin for 12 h when the platinated oligonucleotide (194cisplatin) was spiked before enzymatic hydrolysis. The magnification shows the real scale of the species eluting between 0 and 35 min. 194Pt is off-axis for clarity. Table 2. Obtained Isotope Ratios in the Peak of the GG-194Cisplatin Adduct When Naturally Abundant Cisplatin Was Added in Equimolar Concentration (1:1) and in Excess (1:3)a

R (195Pt/194Pt) a

GG-194cisplatin

GG-194cisplatin + natural cisplatin (1:1)

GG-194cisplatin + natural cisplatin (1:3)

GG-194cisplatin after speed-vac

0.157 ± 0.002

0.154 ± 0.003

0.159 ± 0.004

0.152 ± 0.003

The results are shown also after the spike is preconcentrated before mixing with the sample by using a speed-vac.

adduct per milligram of DNA increased linearly with the cisplatin doses (up to 0.5 mM cisplatin). From an analytical point of view, the different results provided by the three workflows are wellcorrelated with the suspected sources of error: the potential

GG-cisplatin losses in the different steps of the procedure, particularly due to incomplete enzymatic hydrolysis and/or due to undesired retention in the chromatographic column. Additionally, and most importantly, these sources of error can be Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

9559

Figure 6. Summary of the obtained quantitative results for GG-cisplatin with the different workflows summarized in Figure 4. Blue dots correspond to route A, pink spots to route B, and green spots to route C.

quantitatively addressed. Figure 6 results demonstrate that incomplete hydrolysis accounts to 23% ± 3% losses, whereas retention in the column can go as high as 33% ± 4%. This incomplete elution of the species has been often documented when running phosphate-containing molecules27 that exhibit high polarity and might interact with the silanol groups of the packing material. CONCLUSIONS Three sensitive and selective IDA-based approaches have been developed aiming at a more precise and accurate quantification of cisplatin-DNA nucleobases adducts. The obtained analytical figures of merit demonstrated that this step-by-step quantitative methodology provides an advantageous alternative to existing chemical techniques to study such interactions from a quantitative perspective. In comparison to the 32P-postlabeling method (involving, at least, two enzymatic hydrolyses), the developed strategies provide lower sensitivity but higher accuracy. With regard to LC-MS methods, the sensitivity is comparable with the developed strategies, but the lack of commercially available standard complicates the quantitative dimension in LC-MS. Here, a classical speciation postcolumn method has shown to provide estimative concentration results for the main cisplatin adducts in the samples, but its accuracy is clearly affected by the lack of quantitative yields in critical steps such as enzymatic digestion and HPLC column recovery. The use of an appropriate species-specific spike strategy, on the other hand, is more accurate but implies the synthesis and characterization of an (27) Navaza, A. P.; Encinar, J. R.; Carrascal, M.; Abia´n, J.; Sanz Medel, A. Anal. Chem. 2008, 80, 1777–1787.

9560

Analytical Chemistry, Vol. 81, No. 23, December 1, 2009

isotopically labeled GG-194cisplatin spike. When this purified GG-194cisplatin adduct is added to DNA samples before column separation, it permits us to determine the Pt losses during the chromatographic separation, providing similar results to those obtained by conventional postcolumn IDA once they are corrected for previously determined column recovery. Additionally, if the species-specific spiking is done before enzymatic digestion, it is possible to correct for the enzymatic hydrolysis yield as well. In brief, although the three strategies would permit us to determine DNA-cisplatin adducts at biologically relevant levels, the species-specific spiking mode conducted before the hydrolysis should be the method of choice for direct unprecedented accuracy in in vivo experiments. Moreover, judicious choice and combinations of speciesspecific and species-unspecific IDA strategies, as illustrated here, could open the door to the development of reliable analytical methodologies for more accurate quantification of polyisotopic (semi)metals associated to DNA. ACKNOWLEDGMENT The authors acknowledge J. Bettmer and Marta Sierra Zapico for very valuable discussions and Kristof Meelich and Markus Galanski for the synthesis of enriched cisplatin. Additionally, financial support from the Ministry of Science and Innovation is acknowledged through the projects CTQ2007-60206/BQU and CTQ2006-02309. Received for review June 23, 2009. Accepted October 12, 2009. AC901360F