Quantitative determination of melphalan DNA adducts using HPLC—inductively coupled mass spectrometry

JOURNAL OF MASS SPECTROMETRYJ. Mass Spectrom. 2006; 41: 507–516Published online 15 March 2006 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jms.1009

Quantitative determination of melphalan DNA adductsusing HPLC – inductively coupled mass spectrometry

Michael Edler,1,2 Norbert Jakubowski2 and Michael Linscheid1∗

1 Humboldt-Universitat zu Berlin, Department of Chemistry, Brook-Taylor-Str. 2, 12489 Berlin, Germany2 ISAS – Institute for Analytical Sciences, P.O. Box 101352, 44013 Dortmund, Germany

Received 4 August 2005; Accepted 16 January 2006

For the quantification of Melphalan DNA adducts, an analytical approach based on the detection ofphosphorus using liquid chromatography combined with inductively-coupled-plasma mass spectrometry(ICP-MS) was developed. In reaction mixtures of native 2′-deoxynucleotides-5′-monophosphates andMelphalan, which were separated using reversed phase chromatography, phosphate adducts were foundas the most abundant modifications. Besides the phosphate adducts, several base alkylated adducts wereobserved. In calf thymus DNA incubated with Melphalan and enzymatically digested using Nuclease P1,the phosphate adducts as well as monoalkylated dinucleotides were found. The most abundant singleMelphalan adduct observed in DNA was a ring-opened adenosine monophosphate. Some dinucleotideadducts and the adenosine adduct were identified using electrospray ionization mass spectrometry(ESI-MS). Copyright 2006 John Wiley & Sons, Ltd.

KEYWORDS: DNA adducts; inductively-coupled-plasma mass spectrometry (ICP-MS); quantitative 31P monitoring;melphalan; LC/MS

INTRODUCTION

The formation of covalently bound adducts at the nucle-obases of the DNA can induce elimination of purine orpyrimidine bases. At these apurinic or apyrimidinic sites(ap-sites) strand breaks of the DNA can occur, which leadto numeric and structural chromosome aberrations likemicronuclei and sister chromatid exchange. The accumu-lation of mutations within a DNA strand increases theprobability of malfunctions in the metabolism of the cell,usually resulting in apoptosis.

Chemotherapy for cancer of the lymphatic and myeloicsystem (acute and chronic leukemia) can be based on thismechanism and drugs like Cyclophosphamide, Melphalanand Chlorambucil are frequently used. The high cytotoxicpotential of these substances is due to the two or morereactive centers that allow formation of interstrand cross-links.1 Furthermore, the nitrogen mustard derivatives havean immunosuppressive effect, which makes them suitablefor the palliative treatment of repulsive reaction after trans-plantations. Melphalan (MLP) (4-[bis(2-chloroethyl)amino]-L-phenylalanine), especially, is of great importance in thetreatment of chronic myeloic leukemia. In this drug, the alky-lating bis(2-chloroethyl)amino group is bound to the physi-ologically important amino acid L-phenylalanine. Because ofthis substitution, a fast admission of this cytostatic drug to the

ŁCorrespondence to: Michael Linscheid, Humboldt-Universitat zuBerlin, Department of Chemistry, Brook-Taylor-Str. 2, 12489Berlin, Germany. E-mail: [email protected]/grant sponsor: Fonds der Chemischen Industrie (FCI).

cells by active transport mechanisms should be ensured. Mel-phalan is assumed to alkylate the nucleophilic sites in DNAwithout metabolic activation via an intermediary formed‘aziridinium ion’.2,3

The use of alkylating agents holds considerable risks forthe patients, since they show low selectivity and themselveshave carcinogenic potential.4,5

To protect the patients, extensive diagnostics of theblood and liver parameters are necessary. Unfortunately,the diagnostically ascertainable changes of these diagnosticvalues appear with a time gap after the chemotherapy,hence any timely corrective measures are not possible.Determination of the plasma concentrations of the givendrugs is possible for many agents, but these analyses are notvery meaningful.6,7 An additional aggravating fact is that thedwell times of alkylating cytostatic drugs in blood plasmaare only a few minutes, therefore the concentrations of theactive agents can drop below the limits of detection after ashort duration.8 A determination of the metabolized drugsecreted in the urine would be practicable in some cases.However, the concentrations are without meaning becauseonly those shares of the given doses that have not developedany cytotoxic effect can be determined.

Studies of the cytotoxic effects (cell death and DNAsynthesis inhibition) of Melphalan to human cancer cellsshow a clear dose dependence.9 Furthermore, a correlationbetween the cytotoxicity and the form of applicationalso exists.10 Therefore, and to minimize acquired drugresistance,11 therapy accompanying the diagnostics shouldstart at the assumed reaction site of the alkylating cytostaticdrug, the DNA. The rapid determination of the adduct

Copyright 2006 John Wiley & Sons, Ltd.

508 M. Edler, N. Jakubowski and M. Linscheid

level could be an adequate measure for the course ofthe therapy to optimize the protocol for every patientindividually.11 Monitoring of DNA adducts in blood cellsrequires a sensitive analytical assay with a potential to detectone modification among about 108 native nucleotides.12,13

For the therapy accompanying diagnostics, detection of onemodified nucleotide among 104 to 107 native nucleotideswould be sufficient.14

The 32P postlabelling method shows this high detectioncapability (2 adducts among 1010 nucleotides)15 but is notsuitable for clinical diagnostics owing to several reasons: ithas high radioactivity, it is slow and no structure elucidationis possible.16 – 18

Recently it has been shown that for monitoring ofDNA adducts, liquid chromatography/electrospray ion-ization mass spectrometry (LC/ESI-MS) has the ability toprovide structure elucidation besides a very good detectioncapability,19 – 21 the structure of several complex adducts wasdetermined22 – 25 and the first quantitative results of adductdistribution in different tissues based on isotope dilutionwere published.26 For the therapy accompanying diagnostics,the quantitative determination is mandatory, which needswell characterized reference compounds to achieve the pre-cision required.27 Such standard compounds for calibrationare generally not available.

We have therefore introduced the LC/inductively-coupled-plasma mass spectrometry (ICP-MS) for thestructure independent quantitative determination ofnucleotides.28,29

MATERIALS AND METHODS

Sample preparationThe four 20-deoxynucleoside-50-phosphates of adenosine(dAMP), cytidine (dCMP), guanosine (dGMP) and thymi-dine (dTMP) were used as substrates for the reaction withMelphalan. The sodium salts of the nucleotides (Sigma,Taufkirchen, Germany) were dissolved in deionised waterto a final concentration of 10 mg ml�1. The solutions servedas the stock solutions for all reactions. The stock solutionof the Melphalan (Fluka, Taufkirchen, Germany) consistedof 1 mg of Melphalan dissolved in 1 ml dimethylsulfox-ide (DMSO).

For preparation of the Melphalan adducts, 20 µl ofeach stock solution were diluted with 140 µl of NaClsolution (15 mM NaCl, pH D 7.4) in a 0.5 ml reactionvessel and 40 µl of the Melphalan stock solution wereadded. The final concentrations of the nucleotides andMelphalan in the reaction mixture were 1 mg ml�1 and200 µg ml�1, respectively. For optimization of the LC andas a reference sample for comparison with DNA adducts,a reaction mixture consisting of the four nucleotides (5 µlof each nucleotide stock solution) was prepared in thesame way.

The reaction mixtures were incubated at 37 °C in athermocycler (Mastercycler Personal, Eppendorf AG, Ham-burg, Germany) for 24 h. After incubation the mixtureswere washed with 250 µl diethyl ether four times toremove DMSO. As internal standard, 200 µl of an aque-ous bis(4-nitrophenyl)phosphate solution (BNPP, Fluka,

Taufkirchen, Germany) with a concentration of 10 nmol ml�1

was added. A total volume of 1 ml was achieved byadding 600 µl of a 10 mM ammonium acetate solution(pH D 5).

With these samples, solid phase extractions (SPE) werecarried out. One-milliliter one-way-SPE cartridges filled with100 mg C18 absorbent (BondElut, Varian, Middelburg, NL)were used. The cartridges were rinsed successively with2 ml methanol and 1 ml 10 mM ammonia acetate solution(pH D 5) for cleaning and conditioning. After sampleadsorption, the cartridges were washed with 2 ml of theammonium acetate solution to remove the unmodifiednucleotides, the modified nucleotides were eluted from thecartridges with 2 ml methanol. The solvent was removedat 40 °C under nitrogen stream and the dried residuesredissolved in 200 µl of an ammonium acetate solution(10 mM, pH D 5) to yield the concentration of BNPP of10 nmol ml�1 again. The internal standard could be used ina one-point-calibration for the quantification of the modifiednucleotides.

DNA adducts of Melphalan were prepared by reaction ofMelphalan with double stranded DNA (Sigma, Taufkirchen,Germany). An aliquot of 40 µl of a DNA stock solution(5 mg calf thymus DNA in 1 ml deionised water) wasdiluted by adding 120 µl of the NaCl solution (15 mM,pH D 7.4). Adding 40 µl of the Melphalan stock solutioncompleted the reaction mixture. After sonification for 30 s,the reaction mixture was incubated at 37 °C for 24 h. Thenthe reaction mixture was cooled down to room temperatureand extracted with 250 µl diethyl ether four times. TheMelphalan–modified DNA was digested enzymatically withNuclease P1 (E.C.3.1.30.1).

To ensure that the enzymatic digestion of the DNA isnot hindered by Melphalan or its hydrolysis products, itis necessary to remove these substances from the reactionmixture; here we isolated the DNA instead. For precipitationof the DNA, 50 µl of 1 molar NaCl solution was addedto achieve a concentration of 0.2 mol l�1 in the reactionmixture. The DNA was precipitated at room temperatureby adding 500 µl of 2-propanol. After centrifugation at 4 °Cand 12.000 g the DNA pellet was washed twice with 100 µlaqueous ethanol (70%) and dried in vacuum.

The dried DNA was redissolved in 200 µl of a solutionconsisting of ammonium acetate (30 mM, pH 5.3), magnesiumchloride (15 mM), manganese chloride (100 µM) and zincchloride (100 µM). To achieve maximum enzyme activityat the hydrolysis, the DNA solution was heated to 97.5 °Cfor 30 s before 0.5 units of the Nuclease P1 were added. Thereaction solution for the enzymatic hydrolysis of the DNAwas incubated in the thermocycler at 37 °C for 16 h. Theenzyme was removed by centrifugal filtration (30 kDa cutoff, Millipore Corp, Bedford, USA) at 20 °C and 2500 g. Afteradding an aliquot of 50 µl of the internal standard solution tothe filtrate, the volume was filled up with 750 µl ammoniumacetate solution (10 mM, pH D 5) to 1 ml. Then a solid phaseextraction was carried out as described before. The isolatedadducts were redissolved in 50 µl of the ammonia acetatesolution. A preconcentration of adducts by a factor of 4 wasobtained.

Copyright 2006 John Wiley & Sons, Ltd. J. Mass Spectrom. 2006; 41: 507–516

Quantitation of Melphalan Adducts 509

HPLCFor chromatography, a metal-free four channel low-pressuregradient system (Knauer, Berlin, Germany) and a microborecolumn (250 ð 1 mm) filled with 5 µm particles MultospherRP 18 120A HP (Chromatographie Service, Langerwehe,Germany) were used. The flow rate was set to 31 µl min�1

(linear liquid velocity 1 mm s�1). An electrically driveninjection valve equipped with a 5 µl sample loop wasused for sample injection. The modified nucleotides wereseparated with a three-step gradient elution. Solvent A:Aqueous ammonium acetate solution (10 mM). Solvent B:water/methanol (10/90), 10 mM ammonium acetate. ThepH values of both eluents were adjusted to five byadding acetic acid. The gradient program is listed inTable 1.

DetectionThe UV detection was carried out at 258 nm using a pho-todiode array detector (K 2600, Knauer GmbH, Berlin,Germany) and detection of phosphorus at m/z D 31 witha Platform ICP-MS (Micromass Ltd, Manchester, UK). Amicro concentric nebuliser with membrane desolvation sys-tem (MCN 6000, CETAC Technology Inc, Omaha, USA)was used to interface the separation system to the ICP-MS. For the optimization of the phosphorus detection, theHPLC was replaced by a flow injection system consistingof a syringe pump (TSE GmbH, Bad Homburg, Germany)and an injection valve with a 5 µl sample loop. Deionisedwater delivered with a flow rate of 35 µl min�1 served aseluent. Aqueous standard solutions of ammonium dihydro-genphosphate and bis(4-nitrophenyl)phosphate (BNPP) inthe concentration range from 10 to 50 nmol ml�1 were used.The phosphorus detection was optimized for best signal tobackground ratio. The achievable limit of detection is about45 fmol phosphorus absolute. The operational parametersof ICP-MS detection are listed in Table 2. The determina-tion of adducts was based on the comparison of signalareas with the internal standard BNPP through out thisstudy.

For identification of adducts, a time-of-flight massspectrometer from Micromass (Q-TOF Ultima, Micromass,Manchester, UK) was used. An electrospray with Z geometryserved as an ion source for this mass spectrometer. Optimiza-tion of the detection that was carried out with an aqueoussolution of adenosine (1 mg ml�1) showed that maximumsignal intensity can be obtained at �2.5 kV for negative ions.The operational parameters of the mass spectrometer andthe ion source are listed in Table 3.

Table 1. Gradient program

Time/min Eluent A (%) Eluent B (%)

0 89 1115 78 2242.5 45 5546.5 16 8450 16 8460 89 11

Table 2. Operational parameters for 31P detection(Platform ICP-MS)

Forward power 825–875 WArgon cooling gas flow 15 l min�1

Argon intermediate gas flow 1.8 l min�1

Argon nebuliser gas flow 0.75 l min�1

Argon sweep gas flow 3.0 l min�1

H2-addition in collision cell 2.0 ml min�1

Temperature of spraying chamber 70 °CTemperature of desolvation membrane 110 °C

Table 3. Operational parameters for ESI-MS (Q-TOF Ultima)

Ion-mode Electrospray Sprayvoltage

�2.5 kV

Polarity Negative Extractionlens voltage

40 V

Mass range 100–1000 Da – –Spray gasN2

350 l h�1 Sourcetemperature

80 °C

Desolvationgas N2

900 l h�1 Desolvationtemperature

250 °C

RESULTS AND DISCUSSION

The peaks in the chromatogram of the nucleotide mixtureafter reaction with Melphalan were assigned to adducts bycomparison with chromatograms of the single nucleotides(Fig. 1(a)–(d)). The complete assignment of the phosphorussignals is shown in Fig. 2.

The modified nucleotides elute in two distinct groups.The most abundant adducts appear in the first group bet-ween the 15th and 25th min. The second group (30–45 min)with minor adducts shows a much greater diversity indicat-ing that these nucleotides were modified at the nucleobases,whereas the first group corresponds to the alkylated phos-phate residues. There is more evidence for this on the basisof two additional experiments.

Firstly, Melphalan (40 µg) was hydrolyzed either in 1 mldeionised water and/or in phosphate buffer (NH4H2PO4

1/15 m, pH 7.4) at 37 °C for 1 h. Figure 3 shows the UVchromatograms of the two product mixtures recorded at258 nm.

Upon hydrolysis of Melphalan in phosphate buffer, largequantities of phosphate esters are formed. Those phosphateesters elute early because of their higher polarity. This showsthat Melphalan reacts easily with the phosphate residues.

In the second experiment, thymidine (200 µg), whichforms the largest quantity of adducts, was reacted withMelphalan (200 µg and 40 µg). Figure 4 shows the chro-matograms of this approximately equimolar reaction mixtureand in addition the reaction with a molar ratio of 5 to 1(thymidine to Melphalan).

The comparison of the chromatograms recorded bymeans of ICP-MS shows that the most obvious differenceis the concentration of the major adducts. A quantitativeassessment is given in Table 4.



Figure 1. LC-ICP-MS (31P) chromatograms of Melphalan adducts in the nucleotide mixture (top trace) and of single nucleotides; allwith internal standard (last signal). The phosphate adducts (first signal) and the base adducts (smaller following signals) are clearlydistinguishable. (a) dAMP, (b) dCMP, (c) dGMP, (d) dTMP.

Figure 2. LC-ICP-MS (31P) chromatogram of Melphalan adducts in the nucleotide mixture; the remaining nucleotides (first group),the phosphate adducts (second group, major signals) and the mixture of base adducts (blow up) are assigned.



Figure 3. Separation of hydrolysis products of Melphalan in water and in phosphate buffer; UV detection.

0 10 20 30 40 50 600

2×107

2×107

1×107

5×106

6,5*107 cps

dTMP : MLP, equimolardTMP : MLP, 5 : 1 molar

sign

al in

tens

ity /

cps

time / min

Figure 4. Melphalan (MLP)-thymidine adducts from equimolarand 5 : 1 (dTMP : Melphalan) reaction mixture; LC-ICP-MS (31P).

An explanation for the strong appearance of the phos-phate adducts in the studies conducted by the Esmans’group23 and by us here may be that the adducts were earlieroverlooked because generally the nucleobases or nucleosideswere investigated. It required LC/MS to detect and identifythem because they were not expected to play any major

Table 4. Differentiated quantification of thymidine adducts

Molar ratiodTMP : MLP

Major adducts Con-centration/nmol ml�1

Minoradducts

Equimolar 316 63.05 : 1 22.5 38.9

role. A quantitative analysis of the adduct formation in thereaction mixture of the single nucleotides is given in Fig. 5.This quantification seems to prove that thymidine formsmore Melphalan adducts than do adenosine, cytidine andguanosine.

Hoes et al. by means of ESI-MSn already showed thealkylation of phosphate moieties by Melphalan in syntheticreaction solutions. For thymidine the phosphate adduct waseven verified as the major reaction product. However, theyalso found that most adducts are formed at the purinebases.30 Here in a reaction mixture of nucleotides, wefound a remarkable high yield of approximately one-thirdof adducts formed at the phosphates. This contradictioncan be explained by the common depurination induced byalkylation at N7 of guanosine and N7 or N3 of adenosine.The depurinated adducts are invisible for ICP-MS, whichleads to an intrinsically distorted image of the syntheticreaction mixtures and is the only reason thymidine appearsas the major target despite its low nucleophilic activity. The



0

5

10

15

20

25

dAMP dCMP dGMP dTMP dNMP

addu

cts

per

1000

nuc

leot

ides

adducts total phosphate adducts

Figure 5. Quantitative distribution of Melphalan adducts in the reaction mixture of nucleotides; LC-ICP-MS (31P); for discussion seetext.

Figure 6. Comparison of ICP-MS (31P) with UV detection for the separation of the guanosine Melphalan adducts; the phosphatecontaining adducts are much less than those visible with UV detection, mainly caused by depurination.

comparison of the ICP-MS trace with the UV trace of theguanosine Melphalan adducts, as shown in Fig. 6, supportsthis explanation.

The chromatogram recorded at 258 nm shows intensesignals at the 23rd min and in the retention time windowbetween the 35th and 45th min, which obviously contain nophosphorus. Those signals cannot be assigned to Melphalan(tmsr D 52.5 min) or its hydrolysis products (tmsr D 14 min,tmsr D 33.5 min) and presumably they are guanosineadducts without ribosyl phosphate. The alkylation at thenucleophilic centers of thymidine, however, does not lead toweakening of the glycosidic bond and cleavage of the ribosylphosphate. Thus, thymidine phosphate is apparently mostfrequently modified.

The comparison of the chromatograms derived fromDNA sample and the nucleotide mixture, (Fig. 7(a)) showsthat a complete quantification of the Melphalan DNAadducts is not possible because a large portion of the adductscannot be directly correlated. Only similarities exist in theretention time window between the 15th and the 17th minas well as at the 21st min. Those signals correspond to thephosphate adducts of cytidine, guanosine and thymidine.

The phosphate adduct of adenosine cannot be assignedwithout ambiguity because of a coeluting compound. Inthe retention time window around the 30th min, wherethe alkylated nucleobases are expected, no signals weredetected in the DNA sample. In Fig. 7(b), a blank sampleshows the location of the unmodified nucleotides andthe low level of background at the retention times ofinterest.

The two intense signals eluting at the 8th min and the 12thmin correspond to thymidine and adenosine, respectively,as additionally confirmed with UV spectra. Two abundantsignals without correlation appear in the DNA sample atthe 18th and the 27th min. The broad signal of the latterone indicates that this is because of the coelution of severalnucleobase-adducts.

To confirm the correlations, a DNA sample was examinedusing ESI-MS. An aliquot was separated under similarchromatographic conditions and an ESI-QTOF was used fordetection. This data confirm the aforementioned assignment.The intense signal at the 18th min belongs to a componentwith molecular mass of 599 Da ([M � H]� D 598, Figs 8 and 9)and can be assigned to a ring-opened adenosine-Melphalan



0 10 20 30 40 50 60

8×106

7×106

6×106

5×106

4×106

3×106

2×106

1×106

0

nucleotides Melphalan adducts DNA Melphalan adducts

time / min

sign

al in

tens

ity /

cps

(a)

8×106

6×106

4×106

2×106

1×107

00 10 20 30 40 50 60

time / min

internal standardBNPP

native nucleotides

sign

al in

tens

ity /

cps

(b)

Figure 7. LC-ICP-MS (31P): Chromatogram of the Melphalan adducts from DNA and a nucleotide mixture. For comparison, anICP-MS chromatogram of an unmodified DNA sample after enzymatic digestion with nucleases P1 and solid phase extraction (blanksample) is given.

Figure 8. ESI-MS of the compound eluting at 18 min in the DNA sample shown in Fig. 7.

adduct, already proposed by Hoes et al. on the basis ofcollision induced dissociation data.31

An analogous guanosine Melphalan adduct has not beenfound. This indicates that adenosine and guanosine differ intheir follow up reaction on the alkylation at the N7 positionof the purine structure; in guanosine, the ribosyl phosphateis readily lost instead.

In the retention time window between the 26th and30th min, Melphalan adducts consisting of two nucleotidesand one Melphalan moiety appear. The averaged ESI massspectrum between 800 and 950 Da of this retention timewindow is given in Fig. 10. The mono-isotopic molecule ion

Table 5. Melphalan adducts mono-isotopic masses [M–H]/Da

2nd NucleotidedAMP-

MLPdCMP-

MLPdGMP-

MLPdTMP-MLP

dAMP 893.249 – – –dCMP 869.237 845.225 – –dGMP 909.244 885.232 925.239 –dTMP 884.237 860.225 900.232 875.225

masses arising from the possible combinations are listed inTable 5. Recently, the possible structures were proposed.5,25



Figure 9. Formation of the ring-opened adenosine adduct after reaction with Melphalan.

Figure 10. Averaged mass spectrum (800 and 950 Da) of the compounds eluting between the 26th and 30th min.

The observed molecular ions suggest that these adductsare not cross-linked nucleobase pairs of the DNA. Anintrastrand cross-link of neighboring nucleobases can alsobe excluded, since those adducts would be 18 amu lessowing to the second alkylation reaction. Therefore, themost probable structures seem to be dinucleotides thatare monoalkylated at a nucleobase or a phosphate group.Particularly, the alkylation of phosphate groups (tri-ester

formation) is expected to cause protection against thedigestion by the phosphodiesterase Nuclease P1. Such modi-fications might not be recognized as substrates and thereforeskipped, resulting in DNA fragments of structures such as50-pdNpMLP�OH�dN-OH-30. Even though dimer adducts orcross-links were found by other researchers as well,5,25 thequestion concerning their existence in vivo and their originremains unanswered.



CONCLUSION

In this study, we could show that the ICP-MS is an appro-priate quantitative detector for the 31P element specificdetection of the alkylated nucleotides and dinucleotides.They can be quantified regardless of the chemical natureof the modification because the ICP-MS provides a struc-ture independent sensitivity. The analytical system can becalibrated with readily available phosphates as internal stan-dard when using liquid chromatography or flow injection.The absolute detection limit of 45 fmol corresponds to thedetection of three modified nucleotides among 107 nativenucleotides (on the basis of the use of 50 µg DNA). TheMelphalan adducts at the phosphate groups of the fournucleotides can be separated by reversed phase chromatog-raphy using a three-step gradient elution with methanolas organic modifier. The separation and detection limit wasimproved using solid phase extraction to remove the unmod-ified nucleotides almost completely. The Melphalan adductswere assigned by comparison of retention data and a surpris-ingly high amount of phosphate adducts in nucleotides andin DNA was found. With ESI-MS, the previous finding wasconfirmed that a ring-opened adenosine adduct is a majormodification in DNA. Furthermore, we could demonstratethat the enzymatic digestion of alkylated DNA is incom-plete. It remains for future work to show that the methodpresented here is capable of handling real life samples aswell. However, we are confident that the clean-up protocol isspecific enough to remove all unwanted contaminants. Onecan envisage that the LC-ICP-MS could be the method ofchoice to accompany cancer therapy in terms of monitoringthe direct impact of a drug to the target DNA. Detailed struc-ture information as provided by LC/ESI-MS sheds lighton mechanisms and reactions, but is not mandatory in thecontext of chemotherapy.

AcknowledgementsWe thank I. Feldmann (ISAS) for expert technical support. Financialsupport by the Fonds der Chemischen Industrie (FCI) is gratefullyacknowledged.

REFERENCES1. Hemminki K. DNA adducts of nitrogen mustards and ethylene

imines. IARC Sci Publ. 1994; 125: 313.2. Thomas DEV, Connors TA. Molecular aspects of anti-cancer drug

action, Verlag Chemie GmbH: Weinheim, 1983.3. Osborne MR, Lawley PD, Crofton-Sleigh C, Warren W.

Products from alkylation of DNA in cells by melphalan: humansoft tissue sarcoma cell line RD and Escherichia coli WP2. ChemBiol Interact. 1995; 97: 287.

4. Kinlen LJ. Incidence of cancer in rheumatoid arthritis and otherdisorders after immunosuppressive treatment. Am J Med. 1985;78: 44.

5. Balcome S, Park S, Quirk Dorr DR, Hafner L, Phillips L,Tretyakova N. Adenine-containing DNA-DNA cross-links ofantitumor nitrogen mustards. Chem Res Toxicol. 2004; 17: 950.

6. Muckenschnabel I, Bernhardt G, Spruss T, Buschauer A. Aversatile high-performance liquid chromatography method forthe measurement of melphalan tailored to the optimization ofhyperthermic isolated limb perfusion. Eur. Jl Pharm. Sci. 1997; 5:129.

7. Kalhorn TF, Ren S, Howald WN, Lawrence RF, Slattery JT.Analysis of cyclophosphamide and five metabolites from

human plasma using liquid chromatography-mass spectrometryand gas chromatography-nitrogen-phosphorus detection. JChromatogr B Biomed Sci Appl. 1999; 732: 287.

8. Chen R, Chen X, Mitchell DW, Hofstadler SA, Wu Q,Rockwood AL, Sherman MG, Smith RD. Trapping, Detectionand Mass Determination of Coliphage T4 DNA Ions of 108 Daby Electrospray Ionisation FTICR-MS. Anal. Chem. 1995; 67: 1159.

9. Hjertner O, Borset M, Waage A. Comparison of the effects of 2-chlorodeoxyadenosine and melphalan myeloma cell lines. LeukRes. 1996; 20: 155.

10. Pinguet F, Bressolle F, Culine S, Fabbro M, Astre C, Chevil-lard C. Influence of the schedule of exposure on the cytotoxiceffect of melphalan on human 8226 and A2780 cells. Eur J Cancer.1999; 35: 1402.

11. Zhang J, Ye Z, Lou Y. Metabolism of melphalan by rat livermicrosomal glutathione S-transferase. Chem Biol Interact. 2005;152: 101.

12. Savela K, Hemminki K. Analysis of cigarette-smoke-inducedDNA adducts by butanol extraction and nuclease P1-enhanced32P-postlabeling in human lymphocytes and granulocytes.Environ Health Perspect. 1993; 101: 145.

13. Moller L, Grzybowska E, Zeisig M, Cimander B, Hemminki K,Chorazy M. Seasonal variation of DNA adduct pattern in humanlymphocytes analyzed by 32P-HPLC. Carcinogenesis 1996; 17: 61.

14. Trams EG, Nadkarni MV, Smith PK. On the mechanism of actionof the alkylating agents. I. Interaction of alkylating agents withnucleic acids. Cancer Res. 1961; 21: 560.

15. Reddy MV, Randerath K. Nuclease P1-mediated enhancementof sensitivity of 32P-postlabeling test for structurally diverseDNA adducts. Carcinogenesis 1986; 7: 1543.

16. Hemminki K, Forsti A, Lofgren M, Segerback D, Vaca C,Vodicka P. Testing of quantitative parameters in the 32P-postlabelling method. IARC Sci Publ. 1993; 51.

17. Steinberg JJ, Cajigas A, Brownlee M. Enzymatic shot-gun 50-phosphorylation and 30-sister phosphate exchange: a two-dimensional thin-layer chromatographic technique to measureDNA deoxynucleotide modification. J Chromatogr. 1992; 574: 41.

18. Vaca CE, Lofgren M, Hemminki K. Some quantitativeconsiderations about DNA adduct enrichment procedures for32P-postlabelling. Carcinogenesis 1992; 13: 2463.

19. Wolf SM, Vouros P. Application of capillary liquidchromatography coupled with tandem mass spectrometricmethods to the rapid screening of adducts formed by the reactionof N-acetoxy-N-acetyl-2-aminofluorene with calf thymus DNA.Chem Res Toxicol. 1994; 7: 82.

20. Rindgen D, Turesky RJ, Vouros P. Determination of in vitroformed DNA adducts of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine using capillary liquid chromatography/electrospray ionization/tandem mass spectrometry. Chem ResToxicol. 1995; 8: 1005.

21. Vanhoutte K, Van Dongen W, Hoes I, Lemiere F, Esmans EL,Van Onckelen H, Van den Eeckhout E, van Soest REJ, Hud-son AJ. Development of a Nanoscale Liquid Chromatogra-phy/Electrospray Mass Spectrometry Methodology for theDetection and Identification of DNA Adducts. Anal.Chem. 1997;69: 3181.

22. Van den Driessche B, Lemiere F, Van Dongen W, Esmans EL.Alkylation of DNA by melphalan: investigation of capillaryliquid chromatography-electrospray ionization tandem massspectrometry in the study of the adducts at the nucleosidelevel. J Chromatogr B Analyt Technol Biomed Life Sci. 2003; 785: 21.

23. Van den Driessche B, Van Dongen W, Lemiere F, Esmans EL.Implementation of data-dependent acquisitions in thestudy of melphalan DNA adducts by miniaturized liquidchromatography coupled to electrospray tandem massspectrometry. Rapid Commun Mass Spectrom. 2004; 18: 2001.

24. Van den Driessche B, Lemiere F, van Dongen W, Esmans EL.Structural characterization of melphalan modified 20-oligodeoxynucleotides by miniaturized LC-ES MS/MS. J AmSoc Mass Spectrom. 2004; 15: 568.



25. Van den Driessche B, Lemiere F, Witters E, Van Dongen W,Esmans EL. Implications of enzymatic, acidic and thermalhydrolysis of DNA on the occurrence of cross-linked melphalanDNA adducts. Rapid Commun Mass Spectrom. 2005; 19: 449.

26. Van den Driessche B, Esmans EL, Van der Linden A, VanDongen W, Schaerlaken E, Lemiere F, Witters E, Berneman Z.First results of a quantitative study of DNA adducts of melphalanin the rat by isotope dilution mass spectrometry using capillaryliquid chromatography coupled to electrospray tandem massspectrometry. Rapid Commun Mass Spectrom. 2005; 19: 1999.

27. Yang Y, Nikolic D, Swanson SM, van Breemen RB. Quantitativedetermination of N7-methyldeoxyguanosine and O6-methyldeoxyguanosine in DNA by LC-UV-MS-MS. Anal Chem.2002; 74: 5376.

28. Siethoff C, Feldmann I, Jakubowski N, Linscheid M. Quantita-tive determination of DNA adducts using liquid chromatog-raphy electrospray ionization mass spectrometry and liquid

chromatography high-resolution inductively coupled plasmamass spectrometry. Journal of Mass Spectrometry. 1999; 34: 421.

29. Edler M, Jakubowski N, Linscheid M. Styrene oxide DNAadducts: quantitative determination using P-31 monitoring.Analytical and Bioanalytical Chemistry. 2005; 381: 205.

30. Hoes I, Lemiere F, Van Dongen W, Vanhoutte K, Esmans EL,Van Bockstaele D, Berneman Z, Deforce D, Van denEeckhout EG. Analysis of melphalan adducts of 20-deoxynucleotides in calf thymus DNA hydrolysates by capillaryhigh-performance liquid chromatography-electrospray tandemmass spectrometry. J Chromatogr B Biomed Sci Appl. 1999; 736: 43.

31. Hoes I, Van Dongen W, Lemiere F, Esmans EL, VanBockstaele D, Berneman ZN. Comparison between capillary andnano liquid chromatography-electrospray mass spectrometryfor the analysis of minor DNA-melphalan adducts. J ChromatogrB Biomed Sci Appl. 2000; 748: 197.


Quantitative determination of melphalan DNA adducts using HPLC—inductively coupled mass spectrometry

Documents