Journal of Analytical Toxicology, Vol. 16, July/August 1992
Mass Spectral Characterizationof Three Anthracycline Antibiotics:A Comparisonof ThermosprayMass Spectrometryto Particle Beam Mass Spectrometry J o s e p h Bloom 1,2., Paul L e h m a n 1, Mervyn Israel 3, O s v a l d o Rosario 2, and Walter A. K o r f m a c h e r l t
1U.S. Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas 72079 ; 2Department of Chemist~ University of Puerto Rico, Rio Piedras, Puerto Rico 00931; and 3Department of Pharmacology and Medicinal Chemist~ Health Science Center, University of Tennessee at Memphis, Memphis, Tennessee 38163
Abstract I Mass spectral results for three anthrecycllnes, doxorublcln, daunorublcln, and carmlnomycln are compared by using thermosprsy (TS) or particle beam (PB) [electron Ionization (El) and chemical Ionization (Cl)] Instruments. Typically, positive Ion TS mass spectrometry (MS) provided Intense [MH ]+ Ions and some fragment Ions, while PBMS in the El mode provided only fragment ions. Significant [MH ]+ Ions were observed for carmlnomycln and daunorublcln when analyzed using PBMS in the positive Ion Cl mode. Under negative Ion detection, TSMS yielded Intense [M-HI- Ions for these compounds while PBClMS resulted In significant M- Ions. Fragment ions observed In all three anthrecycllnee under positive and negative ion detection by TSMS and PBClMS are due mainly to the cleavage of glycosidic bond, loss of H=O, and the loss of the side chain (COCH2R=) from the aglycone.
Introduction Anthracyclines are antibiotic substances whose structure consists of a napthacenedione chromophore in conjunction with one or more glycoside units. Several of these products exhibit potent antitumor activity. Doxorubicin (AdriamycinTM) is one of the most widely used drugs in medical oncology because of its broad spectrum of antitumor activity (1). Daunorubicin, closely related in structure, sees more limited use, principally against hemotologic malignancies. Carminomycin was previously used clinically but was found inferior to adriamycin and daunorubicin. In general, the use of anthracyclines in cancer chemotherapy is limited by toxicity, including acute bone marrow suppression and an insidious dose-dependent delayed cardiomyopathy (2). This has led to the search for improved antitumor antibiotics and semisynthetic derivatives and a continuing need for improved analytical techniques to study the pharmacology of these substances. Typically, the methods of analysis for anthracyclines have utilized HPLC with UV (3), fluorescence (4-6), or electrochemical detection (7). While these techniques are valuable, none of them provide any structural information about the com. Current address: Medical University of South Carolina, 171 AshleyAve., Charleston, SC 29425 t Author to whomcorrespondence should be addressed. Currentaddress: Department of Drug Metabolism, Schering-PIough Research, 60 Orange St., Bloomfield, NJ 07003
pounds. In addition, these compounds are not amenable to analysis by gas chromatography/mass spectrometry (GC/MS) or by conventional electron impact ionization mass spectrometry (ELMS). Previous reports have documented the analysis of one or more of these compounds using fast atom bombardment (FAB), MS (8), or desorption chemical ionization (DCI) MS (9,10). For comparison, we decided to investigate the utility of particle beam mass spectrometry (PBMS) and thermospray mass spectrometry (TSMS) for providing structural information and characteristic fragment ions that could be used not only in the detection but also in the identification of these compounds. Both TS and PB were developed as techniques to interface high-performance liquid chromatography (HPLC) and mass spectrometry. TSMS has the advantage of being a new ionization technique, while PBMS has the advantage of providing data either in the ElMS mode or the chemical ionization (CI) MS mode. In TS there is a controlled, partial vaporization of the effluent as it flows through an electrically heated capillary tube. As a result of heating, the liquid is nebulized and partially vaporized, and any unvaporized solvent and sample are carried into the ion source as microdroplets or particles forming a supersonic jet (11). When the ions are in solution, they are carried into the ion source in these charged microdroplets, where they are ejected to the gas phase (12). For neutral compounds, the solvent rapidly evaporates, leaving a desolvated analyte, which is then ionized in the ion source and extracted to the mass spectrometer for detection (11). In particle beam MS, a pneumatic nebulizer generates an aerosol from the HPLC effluent. As this aerosol passes through a heated desolvation chamber, the mobile phase evaporates from the droplets, forming a mixture of mobile phase solvent vapor and particles. This mixture of vapor and particles enters a twostage momentum separator in which the low momentum vapor molecules are pumped away and the high momentum particles are carried into the mass spectrometer source where they will be ionized and analyzed (13,14). Because of the differences between these two techniques, HPLC/PBMS and HPLC/TSMS are complementary techniques for the characterization of a wide variety of compounds of environmental and biological interest (11,15-28). This report describes the results obtained for the characterization of doxorubicin, daunorubicin, and carminomycin by PBMS and TSMS in both the positive and negative ionization modes.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
223
Journal of Analytical Toxicology,Vol. 16, July/August 1992
The results demonstrate the applicability of these techniques for the direct (flow injection) analysis of these compounds and suggest that either of these MS interfaces could be chosen for the combined HPLC-MS analysis of these compounds.
Experimental
The standards used were doxorubicin and daunorubicin (Sigma). Carminomycin was a gift from Bristol Laboratories. All the standards were dissolved in methanol at a concentration of 0.3q3.5 mg/mL. The TSMS analysis was performed with a Finnigan-MAT TSQ 70 triple quadrupole mass spectrometer by using positive and negative ion detection, in the filament-off and discharge-off mode. Typically, the vaporizer temperature used was 80~ with a block (aero) temperature of 250~ Introduction of the sample was via direct injection with an ISCO LC-5000 syringe pump and a Rheodyne 7125 injector equipped with a 20-I.tL loop at a flow rate of 1.25 mL/min. The mobile phase was 50:50 acetonitrile-(water, 0.2M ammonium acetate (AA), adjusted to pH 4.5 with acetic acid). The PBMS analysis was performed on a Hewlett-Packard 5988A quadrupole mass spectrometer provided with a HewlettPackard 59980A particle beam interface. The El spectra were acquired at 70 eV, with a source temperature of 250~ The CI spectra were obtained in the positive and negative ion modes with methane as the reagent gas adjusted to 0.50 torr and an analyzer pressure of 2.1 x 104 torr. The desolvation chamber temperature was held at 65~ Thc helium nebulizer pressure was set to 35 psi. The mobile phase used was 50:50 acetonitrile-(water, 0.2M AA, adjusted to pH 4.5 with acetic acid) delivered at a flow rate of 0.5 mL/min from a Hewlett-Packard 1090M HPLC system employing a 10-~L injector loop. The introduction of the sample to the interface was via direct injection. The desorption chetnical ionization (DCIMS) analysis was performed on a t:innigan 4023 GC/MS quadrupole mass spectrometer under both positive and negative ion detection. The reagent gas used was methane at a nominal ionizer pressure of 0.25 ton (33 Pa) and a source pressure of 3.4 x 10-5 torr. Samples were analyzed via a direct-exposure probe incorporating a plat-
0
OH
0
inum filament in which the ramp current was increased linearly from 0 to 3 amperes in 60 s. The source temperature was set to 240~ The acetonitrile and methanol used were HPLC grade from Fisher Scientific. The water used was obtained from a Millipore system. The ammonium acetate was purchased from Fluka and the acetic acid was purchased from Fischer Scientific. All mobile phases were filtered through a 0.45-~ma nylon filter (Lida Co.) before use.
Results
and
Discussion
Positive ion m o d e
Figure l shows the structures of the three compounds discussed in this study. The TS mass spectrum for doxorubicin is shown on Figure 2(a); in this mass spectrum the protonated molecule ([MH] § can be seen at m/z 544 with a relative intensity of 20%. Figure 3 shows the proposed scheme for the major fragmentation pathways observed for the three compounds discussed in this report. This scheme is based, in part, on fragmentation schemes proposed by Dass et al. (8) and Monneret and Sellier (9). The fragment ions seen for the TS mass spectral analysis 337
100-
a
[MH]*
ot
544
397 379|
1
363
100
b [MR}* 528
rr
383 337
1,,
0
514
I00-
C R1
O
[MH]*
O] n CH~ H3
NH 2
Doxorubicin (Adriamycin): Daunorubicin (daunomycin): Carminomycin:
R1 = OCH3, R 2 = OH R1 = OCH3, R2=H R 1 = OH, R2=H
Figure 1. Structures of the anthracyclines discussed in this report.
224
349 255 ] 36r 21, L ~ ao7 | 1384.o2 Li,.,, L . . . . ~ -~L,L 9 I L 200 a6o ,6o
*gs t s~o
660
m/z Figure 2. Positive ion TS mass spectrum of (a) doxorubicin, (b) daunoru-
bicin, and (c) carminomycin. These mass spectra were obtained from the direct injection of about 8 pg of each compound.
Journal of Analytical Toxicology, Vol. 16, July/August 1992
of doxorubicin at m/z 397 and 379 can be attributed to be the ions A2 and B. The base peak in the TS mass spectrum of doxorubicin is the C ion (Figure 3). The TS mass spectrum for daunorubicin shown in Figure 2b includes the protonated molecule at m/z 528 as a strong peak. Fragment ions at m/z 383, 381, and 363 can be assigned to the At, A2, and B ions, respectively (see Figure 3). The TS mass spectrum for carminomycin is shown in Figure 2c; the protonated molecule, m/z 514, is seen as the base peak. The small peak at m/z 496 can be assigned to loss of H20, i.e. the [MH-H20]* ion. In addition, peaks at m/z 367, 349, and 307 can be assigned to be A2, B, and D ions, respectively. Thus, under TSMS conditions, all three of these anthracyclines displayed a protonated molecule and followed pathways a2 and b to give ions A2 and B, respectively. On the other hand, the relative ratios of these ions differed significantly for these three anthracyclines under TSMS conditions, and additional minor peaks were also observed. When analyzed by PBMS under methane CI conditions, only a weak protonated molecule was observed for daunorubicin and carminomycin, and for doxorubicin it was not observed at all (see Figure 4). Each of these anthracyclines displayed a prominent Al ion under these conditions. In addition, daunorubicin exhibited a D ion and carminomycin exhibited an A2 ion while both com-
pounds showed a B ion as a major peak. Therefore, the results for these anthracyclines, when analyzed by PBMS under methane CI conditions, were significantly different from those obtained by TSMS. For comparative purposes, these compounds were also analyzed by methane desorption chemical ionization (DCI) MS. The results for the DCIMS analysis are shown in Table I along with those obtained for TSMS and PBCIMS. As expected, the results obtained for these anthracyclines under methane DCIMS were similar to those obtained by methane PBCIMS. In each case, the protonated molecule was either very weak or not observed, and major peaks were observed corresponding to the Ai and B ions. These methane DCIMS data are similar to those reported previously by Smith (10) for these compounds. When analyzed by PB under EIMS conditions, the mass spectra for these anthracyclines included a molecular ion that was either of very low intensity or not observed at all (see Figure 5). For all three anthracyclines, the base peak corresponded to the ion formed after cleavage of the glycosidic bond and cleavage of 100t a 22,
t,ll2i '
3,,
o- & ~ , , , ~ . u - , , , . . , , a . L
so=
...... J,l..~,
.>. 36:3 383
100
"~
b
T II
....
C"
100|'1 C
J I,=
5~8
aT9
2513
349 /
1
o" ,.._,J,l,., ...~..LI.,.,.,._ .1 200
-
L..1 ,
....
360
,
.....
,,,
.
4()0
s14
5()0
m/z 9
c
Figure 4. Positive ion PBCI mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. These mass spectra were obtained from the direct injection of about 4 lag of each compound.
Figure 3. Proposed fragmentation pathways for the protonated molecules of the three anthracyclines.
Table I. Summary of Positive Ion TS, PBCl, and DCl Mass Spectral Results for the Three Anthracycllnes Relative Intensities of Major ions Compound
Method
[MH] §
A1
Az
B
C
D
Doxorubicin Daunorubicin Carminomycin
TS TS TS
544(22) 528(65) 514(100)
399(10) 383(20) -
397(20) 361 (15) 367(10)
379(10) 363(100) 349(15)
337(100) 337(5)
321 (5) 307(5)
Doxorubicin Daunorubicin Carminornycin
PBCI PBCI PBCI
526(2) 514(2)
399(36) 383(100) 369(100)
. . 367(20)
. . 363(95) 349(50)
Doxorubicin Daunorubicin Carminomycin
DCI DCI DCI
514(5)
383(45) 369(85)
-
379(30) 363(100) 349(100)
Other
496(5), 402(5), 384(5) 255(10), 214(5)
-
321(25) -
224(100) + Others** 323(30) 309(45), 253(70)
337(65) -
321 (20) 321 (20) 307(20)
205(100) 206(20) 205(10)
-
" Dash (-) indicates that the ion was not observed or had e relative intensity of less than 5 %
"~ To~ many to l|st, see the mass spectrum.
225
J o u r n a l o f A n a l y t i c a l T o x i c o l o g y , Vol. 16, J u l y / A u g u s t
100
the side chain including the R2 group. In addition to the base peak, ions corresponding to additional fragmentation of these compounds were also observed.
338
la
28,
I
0
Negative ion mode
Ill i
E
,-- 1~176 t b
3~ i
"~ -~
C....
rr
3 1
382
L,. . . . . .
,
100
325
200
1992
360
56o
460
m/z Figure 5. PBEI mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. These mass spectra were obtained from the direct injection of about 4 pg of each compound.
335
a {M-H]" 542
377
413
c.,~
k~
,
.4~2,
As shown in Figure 6, when analyzed by TSMS in the negative ion mode, all three anthracyclines produced intense [M-H]ions. For daunorubicin and carminomycin this ion was the base peak, while for doxorubicin the relative intensity of this ion was 40%. The base peak for doxorubicin is observed at rrdz 335, which is assumed to be the aglycone anion with a loss of the side chain (COCH2R2). In the TS mass spectrum of doxorubicin, the fragment ion at rrdz 413 is assumed to be caused by cleavage at the glycosidic bond with a retention of the oxygen atom by the aglycone; the same loss is observed for daunorubicin and carminomycin at m /z 397 and 383, respectively. Also, all three compounds showed fragment ions at m/z 377, 361, and 347 for doxorubicin, daunorubicin, and carminomycin, respectively, from a cleavage of the glycosidic bond and the loss of H20. Under FABMS conditions in the negative ion mode, Dass et al. (8) provided a scheme that summarizes the fragmentation pattern for anthracyclines which can be used for comparison to the TSMS results. When analyzed by PBCIMS with negative ion detection, all three anthracyclines produced significant M- ions (see Figure 7). For doxorubicin this ion was the base peak, and for carminomycin this ion had a very high relative intensity (95%). The relative intensity of the M- ion for daunorubicin was 30%. For daunorubicin and carminomycin, the base peak corresponded to the aglycone anion. Characteristic fragment ions, which are also observed by other MS techniques, are seen in these mass spectra. As expected, these PBCIMS results are similar to those obtained by us and the ones reported by Smith (10) for the negative ion methane DCIMS conditions, although Smith (10) reported the M- ion as the base peak for daunorubicin and carminomycin.
526
~00
M-
[M-HI"
543
a 0j '
361
|
396
..... I~,
,Jt~ll iI J.., |,,,, , . . . . . t . . . . .
J
t
397
L~L
0
/
)1~176 t
512
100-
[M-H}"
~
383
382 3
M 527 !
o"
nM" 368
513
100t C 347
02~ 0
360
, Ll
460
'7's6o
s6o
m/z Figure 6. Negativeion TS mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. These massspectrawere obtainedfrom the direct injection of about 8 ~ of each compound.
226
%60
'
........ IJLJ ..... 360 460
5'60
m/z Figure 7. Negative ion PBCI mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. Thesemass spectra were obtained from the direct injection of about 4 lag of each compound.
Journal of Analytical Toxicology, Vol. 16, July/August 1992
Conclusion The mass spectra of the three anthracyclines and the positive ion TS and PBCI mass spectra, in particular, have similar fragmentation patterns with significant characteristic fragment ions observed with both techniques. Also, these ions are similar to those described in the literature for these compounds when analyzed by other MS ionization techniques (8-10). These data show the capability of either TSMS or PBMS for the characterization of these compounds. When comparing TS mass spectra and PBCI mass spectra, the [MH] § ion has a higher relative intensity with TSMS and less fragmentation is observed than with PBCIMS. In the comparison of PBEIMS and PBCIMS, the use of the latter for the analysis of these compounds is preferred, because the spectral data obtained via PBCIMS was of greater utility than the PBEIMS data. In addition, the TIC signal for the PBCIMS analysis was also more intense when compared to the PBEIMS analysis. Under negative ion conditions, the [M-H]- ion was the major ion for TSMS and the M- ion was a significant peak under PBCIMS conditions. In general, negative ion analysis provided an ion in the molecular ion region of greater relative intensity than was obtained in the positive ion mode for all three anthracyclines. Also, characteristic fragment ions were observed for both TSMS and PBCIMS in the negative ion mode for all the compounds. Therefore, both TSMS and PBMS (EI or CI) techniques can be applied for the analysis of these anthracyclines. The choice between TSMS and PBMS may be determined by the HPLC conditions needed for the analysis, the equipment that is available, or the structural information that is desired.
Acknowledgments The authors thank Joaquin Abian for helpful discussions, and Pat Fleischer and Cindy Hartwick for typing the manuscript. This work was presented, in part, at the 39th ASMS Conference on Mass Spectrometry and Allied Topics, May, 1991, Nashville, Tennessee.
References 1. R.H. BI~m and SK, Carter. A new anticazcer dru~ with significar~t clinicat activity. Ann. Intern, Med. 80:249-59 (1974). 2. M, Israel, R. Seshadri, Y. Koseki, T.W. Swealman, and J.M. Idriss. Amelioration of adriamycin toxicity through modification of drug DNA binding properties. Cancer Treat. Rev. 14:163-68 (1987). 3. M. Israel, W.J. Pegg, P.M. Wilkinson, and MB. Garnick. Liquid chromatographic analysis of adriamycin and metabolites in biological fluids. J. Liquid Chromatogr. 1:795-809 (1978). 4. A. Machine, J. Martinez Lanao, F. Gonzalez Lopez, and A. Dominquez-Gil Hurle. Biomed. Chrom. 4:154-56 (1990). 5. J, DeJong, P.A. Massen, A. Akkerdaas, S.I. Cheung, HM. Pinedo, and W.J.F. Van Der Vijgh. Sensitive method for the determination of daunorubicin and all its known metabolites in plasma and heart by high-performance liquid chromatography with fluorescence detection. J. Chromatogr. 529: 359-68 (1990), 6. M.J.M. Ooesterbaan, R.JM. Dirks, T.B. Vree, and E. Van Der Kleijn. Rapid quantitative determination of seven anthracyclines and their hydroxy metabolites in body fluids. J. Chromatogr. 306:323-32 (1984).
7. CM. Riley, A.K. Runyan, and J. Graham-Pole. Determination of doxorubicin in plasma and urine by high performance liquid chromatography with electrochemical detection (HPLC-EC). Application to the clinical pharmacokinetics of doxorubicin in patients with osteogenic sarcoma. Anal. Lett. 20:97-116 (1987). 8. C. Dass, R. Seshadri, M Israel, and D.M. Desiderio. Fast atom bombardment mass spectrometric analysis of anthracyclines and ar~thracyclinones. Biomed. Environ. Mass Spectrom. 17:37-45 (1988). 9, C. Monneret and N. Sellier. Desorption chemical ionization mass spectrometry of anthracyclines and of trisaccharides related to aclacinomycin A and marcellomycin. Biomed. Environ. Mass Spectrom. 13:319-28 (1986). 10. R.G. Smith. Characterization of anthracycline antibiotics by desorption chemical ionization mass spectrometry. Anal. Chem. 54:2006-2008 (1982). 11. A.L. Yergey, C.G. Edmonds, I.A.S. Lewis, and M.L Vestal, In LiquidChro-
matography/Mass Spectrometry Techniques and Applications, Plenum, New York and London, 1990, p. 31. 12. M.L. Vestal. Ionization techniques for non-volatile molecules. Mass Spectrometry Reviews 2:447-80 (1983). 13. R.C. Willoughby and R.E Browner. Monodisperse aerosol generation interface for combining liquid chromatography with mass spectrometry, Anal. Chem. 56:2626-31 (1984). 14. R.F. Browner, P.C. Winkler, D.D. Perkins, and L.E. Abbey. Aerosols as microsampie introduction media for mass spectrometry. Microchem. J. 34: 15-24 (1986). 15. P. Newton. Liquid chromatography-mass spectrometry: Essential tool for drug research. LC GC8:706-14 (1990). 16. R.D. Voyksner, C.S. Smith, and P.C. Knox, Optimization and application of particle beam high-performance liquid chromatography/mass spectrometry to compounds of pharmaceutical interest. Biomed. Environ. Mass Spectrom. 19:523-34 (1990). 17. T.A. Bellar, T.D. Behymer, and W.L. Budde, Investigation of enhanced ion abundances from a carrier process in high-performance liquid chromatography particIe-beam mass spectrometry. J_ Am. Soc. Mass Specfrom. 1:92-98 (1990). 16. G.B. Kenion, G.T. Carter, J. Ashraf, M M Siegel, and DB. Borders, Qualitative analysis of pharmaceuticals by thermospray liquid chromatography/mass spectrometry. ACS Symp. Series 420:140-65 (1990), 19. J. lida and T. Murata. Formic acid-ammonium formate buffer system for thermospray liquid chromatography/mass spectrometry~ Anal. Sci. 6" 269-72 (1990). 20. T,D. Behymer, T.A. Belier, and W.L. Budde. Liquid chromatography/particle beam/mass spectrometry of polar compounds of environmental interest. Anal. Chem. 62:1686-90 (1990). 21. C.E. Parker, S. Verma, K.B. Tomer, R.L. Reed, and D.R. Buhler. Determination of Senecio alkaloids by thermospray liquid chromatography/mass spectrometry. Biomed. Environ. Mass Spectrom. 19:1-12 (1990). 22. R.D. Voyksner, E,D. Bush, and D. Brent. Derivatization to improve thermospray HPLC/MS sensitivity for the determination of prostaglandins and thromboxane B2. Biomed. Environ. Mass Spectrom. 14:523-31 (1987). 23. W.A. Korfmacher, J.P. Freeman, T.A. Getek, J. Bloom, and C.L. Holder. Analysis of rat urine for metabolites of pyrilamine via high-performance liquid chromatography/thermospray mass spectrometry and tandem mass spectrometry. Biomed. Environ. Mass Spectrom, 19:191-201 (1990). 24. W.A. Korfmacher, T.A. Getek, E,B, Hanson, Jr., and C.E. Cerniglia.Direct analysis of microbial extracts containing metabolites of ethylenediaminetype antihistamines via high performance liquid chromatography-thermospray mass spectrometry. Anal. Biochem. 185:136-42 (1990). 25. L.D. Betowski, W,A. Kcffmacher, J.O. Lay, Jr., D.W_ Potter, and J.A. Hinson. Direct analysis of rat bile for acetaminophen and two of its conjugated metabolites via thermospray liquid chromatography/mass spectrometry. Biomed. Environ. Mass Spectrom. 14:705-709 (1987). 26. W.A. Korfmacher, T.A. Getek, E.B. Nansen, Jr,, and J. Bloom. Characterization of diphenhydramine, doxylamine, and carbinoxamine using high performance liquid chromatography-thermospray mass spectrometry. LC GC 8: 538-41 (1990). 27. W.A. Korfmacher, C.L. Holder, L.D. Betowski, and RK. Mitchum. Characterization of doxylamine and pyrilamine metabolites via thermospray mass spectrometry and tandem mass spectrometry. Biomed. Environ. Mass Spectrom. 15:501-508 (1988). 28. R. Voyksner, T. Pack, C. Smith, H. Swaisgood, and D. Chen. Techniques for enhancing structural information from high-performance liquid chromatography/mass spectrometry. ACS Symp. Series 420:14-39 (1990). Manuscript received July 15, 1991 ; revision received December 9, 1991.
227
Mass Spectral Characterizationof Three Anthracycline Antibiotics:A Comparisonof ThermosprayMass Spectrometryto Particle Beam Mass Spectrometry J o s e p h Bloom 1,2., Paul L e h m a n 1, Mervyn Israel 3, O s v a l d o Rosario 2, and Walter A. K o r f m a c h e r l t
1U.S. Food and Drug Administration, National Center for Toxicological Research, Jefferson, Arkansas 72079 ; 2Department of Chemist~ University of Puerto Rico, Rio Piedras, Puerto Rico 00931; and 3Department of Pharmacology and Medicinal Chemist~ Health Science Center, University of Tennessee at Memphis, Memphis, Tennessee 38163
Abstract I Mass spectral results for three anthrecycllnes, doxorublcln, daunorublcln, and carmlnomycln are compared by using thermosprsy (TS) or particle beam (PB) [electron Ionization (El) and chemical Ionization (Cl)] Instruments. Typically, positive Ion TS mass spectrometry (MS) provided Intense [MH ]+ Ions and some fragment Ions, while PBMS in the El mode provided only fragment ions. Significant [MH ]+ Ions were observed for carmlnomycln and daunorublcln when analyzed using PBMS in the positive Ion Cl mode. Under negative Ion detection, TSMS yielded Intense [M-HI- Ions for these compounds while PBClMS resulted In significant M- Ions. Fragment ions observed In all three anthrecycllnee under positive and negative ion detection by TSMS and PBClMS are due mainly to the cleavage of glycosidic bond, loss of H=O, and the loss of the side chain (COCH2R=) from the aglycone.
Introduction Anthracyclines are antibiotic substances whose structure consists of a napthacenedione chromophore in conjunction with one or more glycoside units. Several of these products exhibit potent antitumor activity. Doxorubicin (AdriamycinTM) is one of the most widely used drugs in medical oncology because of its broad spectrum of antitumor activity (1). Daunorubicin, closely related in structure, sees more limited use, principally against hemotologic malignancies. Carminomycin was previously used clinically but was found inferior to adriamycin and daunorubicin. In general, the use of anthracyclines in cancer chemotherapy is limited by toxicity, including acute bone marrow suppression and an insidious dose-dependent delayed cardiomyopathy (2). This has led to the search for improved antitumor antibiotics and semisynthetic derivatives and a continuing need for improved analytical techniques to study the pharmacology of these substances. Typically, the methods of analysis for anthracyclines have utilized HPLC with UV (3), fluorescence (4-6), or electrochemical detection (7). While these techniques are valuable, none of them provide any structural information about the com. Current address: Medical University of South Carolina, 171 AshleyAve., Charleston, SC 29425 t Author to whomcorrespondence should be addressed. Currentaddress: Department of Drug Metabolism, Schering-PIough Research, 60 Orange St., Bloomfield, NJ 07003
pounds. In addition, these compounds are not amenable to analysis by gas chromatography/mass spectrometry (GC/MS) or by conventional electron impact ionization mass spectrometry (ELMS). Previous reports have documented the analysis of one or more of these compounds using fast atom bombardment (FAB), MS (8), or desorption chemical ionization (DCI) MS (9,10). For comparison, we decided to investigate the utility of particle beam mass spectrometry (PBMS) and thermospray mass spectrometry (TSMS) for providing structural information and characteristic fragment ions that could be used not only in the detection but also in the identification of these compounds. Both TS and PB were developed as techniques to interface high-performance liquid chromatography (HPLC) and mass spectrometry. TSMS has the advantage of being a new ionization technique, while PBMS has the advantage of providing data either in the ElMS mode or the chemical ionization (CI) MS mode. In TS there is a controlled, partial vaporization of the effluent as it flows through an electrically heated capillary tube. As a result of heating, the liquid is nebulized and partially vaporized, and any unvaporized solvent and sample are carried into the ion source as microdroplets or particles forming a supersonic jet (11). When the ions are in solution, they are carried into the ion source in these charged microdroplets, where they are ejected to the gas phase (12). For neutral compounds, the solvent rapidly evaporates, leaving a desolvated analyte, which is then ionized in the ion source and extracted to the mass spectrometer for detection (11). In particle beam MS, a pneumatic nebulizer generates an aerosol from the HPLC effluent. As this aerosol passes through a heated desolvation chamber, the mobile phase evaporates from the droplets, forming a mixture of mobile phase solvent vapor and particles. This mixture of vapor and particles enters a twostage momentum separator in which the low momentum vapor molecules are pumped away and the high momentum particles are carried into the mass spectrometer source where they will be ionized and analyzed (13,14). Because of the differences between these two techniques, HPLC/PBMS and HPLC/TSMS are complementary techniques for the characterization of a wide variety of compounds of environmental and biological interest (11,15-28). This report describes the results obtained for the characterization of doxorubicin, daunorubicin, and carminomycin by PBMS and TSMS in both the positive and negative ionization modes.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher's permission.
223
Journal of Analytical Toxicology,Vol. 16, July/August 1992
The results demonstrate the applicability of these techniques for the direct (flow injection) analysis of these compounds and suggest that either of these MS interfaces could be chosen for the combined HPLC-MS analysis of these compounds.
Experimental
The standards used were doxorubicin and daunorubicin (Sigma). Carminomycin was a gift from Bristol Laboratories. All the standards were dissolved in methanol at a concentration of 0.3q3.5 mg/mL. The TSMS analysis was performed with a Finnigan-MAT TSQ 70 triple quadrupole mass spectrometer by using positive and negative ion detection, in the filament-off and discharge-off mode. Typically, the vaporizer temperature used was 80~ with a block (aero) temperature of 250~ Introduction of the sample was via direct injection with an ISCO LC-5000 syringe pump and a Rheodyne 7125 injector equipped with a 20-I.tL loop at a flow rate of 1.25 mL/min. The mobile phase was 50:50 acetonitrile-(water, 0.2M ammonium acetate (AA), adjusted to pH 4.5 with acetic acid). The PBMS analysis was performed on a Hewlett-Packard 5988A quadrupole mass spectrometer provided with a HewlettPackard 59980A particle beam interface. The El spectra were acquired at 70 eV, with a source temperature of 250~ The CI spectra were obtained in the positive and negative ion modes with methane as the reagent gas adjusted to 0.50 torr and an analyzer pressure of 2.1 x 104 torr. The desolvation chamber temperature was held at 65~ Thc helium nebulizer pressure was set to 35 psi. The mobile phase used was 50:50 acetonitrile-(water, 0.2M AA, adjusted to pH 4.5 with acetic acid) delivered at a flow rate of 0.5 mL/min from a Hewlett-Packard 1090M HPLC system employing a 10-~L injector loop. The introduction of the sample to the interface was via direct injection. The desorption chetnical ionization (DCIMS) analysis was performed on a t:innigan 4023 GC/MS quadrupole mass spectrometer under both positive and negative ion detection. The reagent gas used was methane at a nominal ionizer pressure of 0.25 ton (33 Pa) and a source pressure of 3.4 x 10-5 torr. Samples were analyzed via a direct-exposure probe incorporating a plat-
0
OH
0
inum filament in which the ramp current was increased linearly from 0 to 3 amperes in 60 s. The source temperature was set to 240~ The acetonitrile and methanol used were HPLC grade from Fisher Scientific. The water used was obtained from a Millipore system. The ammonium acetate was purchased from Fluka and the acetic acid was purchased from Fischer Scientific. All mobile phases were filtered through a 0.45-~ma nylon filter (Lida Co.) before use.
Results
and
Discussion
Positive ion m o d e
Figure l shows the structures of the three compounds discussed in this study. The TS mass spectrum for doxorubicin is shown on Figure 2(a); in this mass spectrum the protonated molecule ([MH] § can be seen at m/z 544 with a relative intensity of 20%. Figure 3 shows the proposed scheme for the major fragmentation pathways observed for the three compounds discussed in this report. This scheme is based, in part, on fragmentation schemes proposed by Dass et al. (8) and Monneret and Sellier (9). The fragment ions seen for the TS mass spectral analysis 337
100-
a
[MH]*
ot
544
397 379|
1
363
100
b [MR}* 528
rr
383 337
1,,
0
514
I00-
C R1
O
[MH]*
O] n CH~ H3
NH 2
Doxorubicin (Adriamycin): Daunorubicin (daunomycin): Carminomycin:
R1 = OCH3, R 2 = OH R1 = OCH3, R2=H R 1 = OH, R2=H
Figure 1. Structures of the anthracyclines discussed in this report.
224
349 255 ] 36r 21, L ~ ao7 | 1384.o2 Li,.,, L . . . . ~ -~L,L 9 I L 200 a6o ,6o
*gs t s~o
660
m/z Figure 2. Positive ion TS mass spectrum of (a) doxorubicin, (b) daunoru-
bicin, and (c) carminomycin. These mass spectra were obtained from the direct injection of about 8 pg of each compound.
Journal of Analytical Toxicology, Vol. 16, July/August 1992
of doxorubicin at m/z 397 and 379 can be attributed to be the ions A2 and B. The base peak in the TS mass spectrum of doxorubicin is the C ion (Figure 3). The TS mass spectrum for daunorubicin shown in Figure 2b includes the protonated molecule at m/z 528 as a strong peak. Fragment ions at m/z 383, 381, and 363 can be assigned to the At, A2, and B ions, respectively (see Figure 3). The TS mass spectrum for carminomycin is shown in Figure 2c; the protonated molecule, m/z 514, is seen as the base peak. The small peak at m/z 496 can be assigned to loss of H20, i.e. the [MH-H20]* ion. In addition, peaks at m/z 367, 349, and 307 can be assigned to be A2, B, and D ions, respectively. Thus, under TSMS conditions, all three of these anthracyclines displayed a protonated molecule and followed pathways a2 and b to give ions A2 and B, respectively. On the other hand, the relative ratios of these ions differed significantly for these three anthracyclines under TSMS conditions, and additional minor peaks were also observed. When analyzed by PBMS under methane CI conditions, only a weak protonated molecule was observed for daunorubicin and carminomycin, and for doxorubicin it was not observed at all (see Figure 4). Each of these anthracyclines displayed a prominent Al ion under these conditions. In addition, daunorubicin exhibited a D ion and carminomycin exhibited an A2 ion while both com-
pounds showed a B ion as a major peak. Therefore, the results for these anthracyclines, when analyzed by PBMS under methane CI conditions, were significantly different from those obtained by TSMS. For comparative purposes, these compounds were also analyzed by methane desorption chemical ionization (DCI) MS. The results for the DCIMS analysis are shown in Table I along with those obtained for TSMS and PBCIMS. As expected, the results obtained for these anthracyclines under methane DCIMS were similar to those obtained by methane PBCIMS. In each case, the protonated molecule was either very weak or not observed, and major peaks were observed corresponding to the Ai and B ions. These methane DCIMS data are similar to those reported previously by Smith (10) for these compounds. When analyzed by PB under EIMS conditions, the mass spectra for these anthracyclines included a molecular ion that was either of very low intensity or not observed at all (see Figure 5). For all three anthracyclines, the base peak corresponded to the ion formed after cleavage of the glycosidic bond and cleavage of 100t a 22,
t,ll2i '
3,,
o- & ~ , , , ~ . u - , , , . . , , a . L
so=
...... J,l..~,
.>. 36:3 383
100
"~
b
T II
....
C"
100|'1 C
J I,=
5~8
aT9
2513
349 /
1
o" ,.._,J,l,., ...~..LI.,.,.,._ .1 200
-
L..1 ,
....
360
,
.....
,,,
.
4()0
s14
5()0
m/z 9
c
Figure 4. Positive ion PBCI mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. These mass spectra were obtained from the direct injection of about 4 lag of each compound.
Figure 3. Proposed fragmentation pathways for the protonated molecules of the three anthracyclines.
Table I. Summary of Positive Ion TS, PBCl, and DCl Mass Spectral Results for the Three Anthracycllnes Relative Intensities of Major ions Compound
Method
[MH] §
A1
Az
B
C
D
Doxorubicin Daunorubicin Carminomycin
TS TS TS
544(22) 528(65) 514(100)
399(10) 383(20) -
397(20) 361 (15) 367(10)
379(10) 363(100) 349(15)
337(100) 337(5)
321 (5) 307(5)
Doxorubicin Daunorubicin Carminornycin
PBCI PBCI PBCI
526(2) 514(2)
399(36) 383(100) 369(100)
. . 367(20)
. . 363(95) 349(50)
Doxorubicin Daunorubicin Carminomycin
DCI DCI DCI
514(5)
383(45) 369(85)
-
379(30) 363(100) 349(100)
Other
496(5), 402(5), 384(5) 255(10), 214(5)
-
321(25) -
224(100) + Others** 323(30) 309(45), 253(70)
337(65) -
321 (20) 321 (20) 307(20)
205(100) 206(20) 205(10)
-
" Dash (-) indicates that the ion was not observed or had e relative intensity of less than 5 %
"~ To~ many to l|st, see the mass spectrum.
225
J o u r n a l o f A n a l y t i c a l T o x i c o l o g y , Vol. 16, J u l y / A u g u s t
100
the side chain including the R2 group. In addition to the base peak, ions corresponding to additional fragmentation of these compounds were also observed.
338
la
28,
I
0
Negative ion mode
Ill i
E
,-- 1~176 t b
3~ i
"~ -~
C....
rr
3 1
382
L,. . . . . .
,
100
325
200
1992
360
56o
460
m/z Figure 5. PBEI mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. These mass spectra were obtained from the direct injection of about 4 pg of each compound.
335
a {M-H]" 542
377
413
c.,~
k~
,
.4~2,
As shown in Figure 6, when analyzed by TSMS in the negative ion mode, all three anthracyclines produced intense [M-H]ions. For daunorubicin and carminomycin this ion was the base peak, while for doxorubicin the relative intensity of this ion was 40%. The base peak for doxorubicin is observed at rrdz 335, which is assumed to be the aglycone anion with a loss of the side chain (COCH2R2). In the TS mass spectrum of doxorubicin, the fragment ion at rrdz 413 is assumed to be caused by cleavage at the glycosidic bond with a retention of the oxygen atom by the aglycone; the same loss is observed for daunorubicin and carminomycin at m /z 397 and 383, respectively. Also, all three compounds showed fragment ions at m/z 377, 361, and 347 for doxorubicin, daunorubicin, and carminomycin, respectively, from a cleavage of the glycosidic bond and the loss of H20. Under FABMS conditions in the negative ion mode, Dass et al. (8) provided a scheme that summarizes the fragmentation pattern for anthracyclines which can be used for comparison to the TSMS results. When analyzed by PBCIMS with negative ion detection, all three anthracyclines produced significant M- ions (see Figure 7). For doxorubicin this ion was the base peak, and for carminomycin this ion had a very high relative intensity (95%). The relative intensity of the M- ion for daunorubicin was 30%. For daunorubicin and carminomycin, the base peak corresponded to the aglycone anion. Characteristic fragment ions, which are also observed by other MS techniques, are seen in these mass spectra. As expected, these PBCIMS results are similar to those obtained by us and the ones reported by Smith (10) for the negative ion methane DCIMS conditions, although Smith (10) reported the M- ion as the base peak for daunorubicin and carminomycin.
526
~00
M-
[M-HI"
543
a 0j '
361
|
396
..... I~,
,Jt~ll iI J.., |,,,, , . . . . . t . . . . .
J
t
397
L~L
0
/
)1~176 t
512
100-
[M-H}"
~
383
382 3
M 527 !
o"
nM" 368
513
100t C 347
02~ 0
360
, Ll
460
'7's6o
s6o
m/z Figure 6. Negativeion TS mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. These massspectrawere obtainedfrom the direct injection of about 8 ~ of each compound.
226
%60
'
........ IJLJ ..... 360 460
5'60
m/z Figure 7. Negative ion PBCI mass spectrum of (a) doxorubicin, (b) daunorubicin, and (c) carminomycin. Thesemass spectra were obtained from the direct injection of about 4 lag of each compound.
Journal of Analytical Toxicology, Vol. 16, July/August 1992
Conclusion The mass spectra of the three anthracyclines and the positive ion TS and PBCI mass spectra, in particular, have similar fragmentation patterns with significant characteristic fragment ions observed with both techniques. Also, these ions are similar to those described in the literature for these compounds when analyzed by other MS ionization techniques (8-10). These data show the capability of either TSMS or PBMS for the characterization of these compounds. When comparing TS mass spectra and PBCI mass spectra, the [MH] § ion has a higher relative intensity with TSMS and less fragmentation is observed than with PBCIMS. In the comparison of PBEIMS and PBCIMS, the use of the latter for the analysis of these compounds is preferred, because the spectral data obtained via PBCIMS was of greater utility than the PBEIMS data. In addition, the TIC signal for the PBCIMS analysis was also more intense when compared to the PBEIMS analysis. Under negative ion conditions, the [M-H]- ion was the major ion for TSMS and the M- ion was a significant peak under PBCIMS conditions. In general, negative ion analysis provided an ion in the molecular ion region of greater relative intensity than was obtained in the positive ion mode for all three anthracyclines. Also, characteristic fragment ions were observed for both TSMS and PBCIMS in the negative ion mode for all the compounds. Therefore, both TSMS and PBMS (EI or CI) techniques can be applied for the analysis of these anthracyclines. The choice between TSMS and PBMS may be determined by the HPLC conditions needed for the analysis, the equipment that is available, or the structural information that is desired.
Acknowledgments The authors thank Joaquin Abian for helpful discussions, and Pat Fleischer and Cindy Hartwick for typing the manuscript. This work was presented, in part, at the 39th ASMS Conference on Mass Spectrometry and Allied Topics, May, 1991, Nashville, Tennessee.
References 1. R.H. BI~m and SK, Carter. A new anticazcer dru~ with significar~t clinicat activity. Ann. Intern, Med. 80:249-59 (1974). 2. M, Israel, R. Seshadri, Y. Koseki, T.W. Swealman, and J.M. Idriss. Amelioration of adriamycin toxicity through modification of drug DNA binding properties. Cancer Treat. Rev. 14:163-68 (1987). 3. M. Israel, W.J. Pegg, P.M. Wilkinson, and MB. Garnick. Liquid chromatographic analysis of adriamycin and metabolites in biological fluids. J. Liquid Chromatogr. 1:795-809 (1978). 4. A. Machine, J. Martinez Lanao, F. Gonzalez Lopez, and A. Dominquez-Gil Hurle. Biomed. Chrom. 4:154-56 (1990). 5. J, DeJong, P.A. Massen, A. Akkerdaas, S.I. Cheung, HM. Pinedo, and W.J.F. Van Der Vijgh. Sensitive method for the determination of daunorubicin and all its known metabolites in plasma and heart by high-performance liquid chromatography with fluorescence detection. J. Chromatogr. 529: 359-68 (1990), 6. M.J.M. Ooesterbaan, R.JM. Dirks, T.B. Vree, and E. Van Der Kleijn. Rapid quantitative determination of seven anthracyclines and their hydroxy metabolites in body fluids. J. Chromatogr. 306:323-32 (1984).
7. CM. Riley, A.K. Runyan, and J. Graham-Pole. Determination of doxorubicin in plasma and urine by high performance liquid chromatography with electrochemical detection (HPLC-EC). Application to the clinical pharmacokinetics of doxorubicin in patients with osteogenic sarcoma. Anal. Lett. 20:97-116 (1987). 8. C. Dass, R. Seshadri, M Israel, and D.M. Desiderio. Fast atom bombardment mass spectrometric analysis of anthracyclines and ar~thracyclinones. Biomed. Environ. Mass Spectrom. 17:37-45 (1988). 9, C. Monneret and N. Sellier. Desorption chemical ionization mass spectrometry of anthracyclines and of trisaccharides related to aclacinomycin A and marcellomycin. Biomed. Environ. Mass Spectrom. 13:319-28 (1986). 10. R.G. Smith. Characterization of anthracycline antibiotics by desorption chemical ionization mass spectrometry. Anal. Chem. 54:2006-2008 (1982). 11. A.L. Yergey, C.G. Edmonds, I.A.S. Lewis, and M.L Vestal, In LiquidChro-
matography/Mass Spectrometry Techniques and Applications, Plenum, New York and London, 1990, p. 31. 12. M.L. Vestal. Ionization techniques for non-volatile molecules. Mass Spectrometry Reviews 2:447-80 (1983). 13. R.C. Willoughby and R.E Browner. Monodisperse aerosol generation interface for combining liquid chromatography with mass spectrometry, Anal. Chem. 56:2626-31 (1984). 14. R.F. Browner, P.C. Winkler, D.D. Perkins, and L.E. Abbey. Aerosols as microsampie introduction media for mass spectrometry. Microchem. J. 34: 15-24 (1986). 15. P. Newton. Liquid chromatography-mass spectrometry: Essential tool for drug research. LC GC8:706-14 (1990). 16. R.D. Voyksner, C.S. Smith, and P.C. Knox, Optimization and application of particle beam high-performance liquid chromatography/mass spectrometry to compounds of pharmaceutical interest. Biomed. Environ. Mass Spectrom. 19:523-34 (1990). 17. T.A. Bellar, T.D. Behymer, and W.L. Budde, Investigation of enhanced ion abundances from a carrier process in high-performance liquid chromatography particIe-beam mass spectrometry. J_ Am. Soc. Mass Specfrom. 1:92-98 (1990). 16. G.B. Kenion, G.T. Carter, J. Ashraf, M M Siegel, and DB. Borders, Qualitative analysis of pharmaceuticals by thermospray liquid chromatography/mass spectrometry. ACS Symp. Series 420:140-65 (1990), 19. J. lida and T. Murata. Formic acid-ammonium formate buffer system for thermospray liquid chromatography/mass spectrometry~ Anal. Sci. 6" 269-72 (1990). 20. T,D. Behymer, T.A. Belier, and W.L. Budde. Liquid chromatography/particle beam/mass spectrometry of polar compounds of environmental interest. Anal. Chem. 62:1686-90 (1990). 21. C.E. Parker, S. Verma, K.B. Tomer, R.L. Reed, and D.R. Buhler. Determination of Senecio alkaloids by thermospray liquid chromatography/mass spectrometry. Biomed. Environ. Mass Spectrom. 19:1-12 (1990). 22. R.D. Voyksner, E,D. Bush, and D. Brent. Derivatization to improve thermospray HPLC/MS sensitivity for the determination of prostaglandins and thromboxane B2. Biomed. Environ. Mass Spectrom. 14:523-31 (1987). 23. W.A. Korfmacher, J.P. Freeman, T.A. Getek, J. Bloom, and C.L. Holder. Analysis of rat urine for metabolites of pyrilamine via high-performance liquid chromatography/thermospray mass spectrometry and tandem mass spectrometry. Biomed. Environ. Mass Spectrom, 19:191-201 (1990). 24. W.A. Korfmacher, T.A. Getek, E,B, Hanson, Jr., and C.E. Cerniglia.Direct analysis of microbial extracts containing metabolites of ethylenediaminetype antihistamines via high performance liquid chromatography-thermospray mass spectrometry. Anal. Biochem. 185:136-42 (1990). 25. L.D. Betowski, W,A. Kcffmacher, J.O. Lay, Jr., D.W_ Potter, and J.A. Hinson. Direct analysis of rat bile for acetaminophen and two of its conjugated metabolites via thermospray liquid chromatography/mass spectrometry. Biomed. Environ. Mass Spectrom. 14:705-709 (1987). 26. W.A. Korfmacher, T.A. Getek, E.B. Nansen, Jr,, and J. Bloom. Characterization of diphenhydramine, doxylamine, and carbinoxamine using high performance liquid chromatography-thermospray mass spectrometry. LC GC 8: 538-41 (1990). 27. W.A. Korfmacher, C.L. Holder, L.D. Betowski, and RK. Mitchum. Characterization of doxylamine and pyrilamine metabolites via thermospray mass spectrometry and tandem mass spectrometry. Biomed. Environ. Mass Spectrom. 15:501-508 (1988). 28. R. Voyksner, T. Pack, C. Smith, H. Swaisgood, and D. Chen. Techniques for enhancing structural information from high-performance liquid chromatography/mass spectrometry. ACS Symp. Series 420:14-39 (1990). Manuscript received July 15, 1991 ; revision received December 9, 1991.
227
