Characterization of degradation products of idarubicin through LC-UV, MSn and LC–MS-TOF studies

Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis journal homepage: www.elsevier.com/locate/jpba

Characterization of degradation products of idarubicin through LC-UV, MSn and LC–MS-TOF studies Dheeraj Kaushik, Gulshan Bansal ∗ Department of Pharmaceutical Sciences and Drug Research, Punjabi University, Patiala 147002, India

a r t i c l e

i n f o

Article history: Received 16 April 2013 Received in revised form 26 June 2013 Accepted 7 July 2013 Available online 24 July 2013 Keywords: Idarubicin Forced degradation LC-MS-TOF Degradation product Mass fragmentation

a b s t r a c t Idarubicin was subjected to forced degradation under the ICH recommended conditions of hydrolysis, oxidation, dry heat and photolysis to characterize its possible impurities and/or degradation products. The drug was found unstable to acid hydrolysis at 85 ◦ C and to alkaline hydrolysis, and oxidation at room temperature. The hydrolytic and oxidative degradation products were resolved with gradient and isocratic elution, respectively on an Inertsil RP18 (250 mm × 4.6 mm; 5 ␮) column with HCOONH4 (20 mM, pH 3.0) and acetonitrile. The drug degraded to two products (O-I and O-II) in oxidative condition, two products (K-I and K-II) in alkaline hydrolytic, and one product (A-I) in acidic hydrolytic conditions. The purity of each in the LC-UV chromatogram was ascertained through LC-PDA analysis. The products were characterized through +ESI-MSn studies on the drug and LC–MS-TOF studies on the degraded drug solutions. Based on the multistage mass fragmentation pattern of idarubicin and accurate mass analysis of the degradation products, the O-I, O-II and A-I were characterized as desacetylidarubicin hydroperoxide, desacetylidarubicin and deglucosaminylidarubicin, respectively. The products K-I and K-II were not characterized due to their low concentrations and/or extremely weak ionization. The mechanisms of degradation of idarubicin to these products were proposed and discussed. © 2013 Elsevier B.V. All rights reserved.

1. Introduction Idarubicin (IDA) (Fig. 1) is an anthracycline based antineoplastic drug used for the treatment of acute myelogenous leukemia and chronic lymphocyctic leukemia in adults [1]. It is a 4-demethoxy analog of daunorubicin and approved by US-FDA for better efficacy and less cardiotoxicity in comparison to other anthracycline antineoplastics. The higher activity of IDA vis-a-vis daunorubicin is attributed to the absence of 4-methoxy moiety in it [2,3]. It is official in United States Pharmacopeia [4] wherein no related substance or impurity is mentioned in the monograph. However, the ICH guidelines Q3A and Q3B require the characterization of all impurities (process and degradation related) that may be present in a drug substance or product [5,6]. This characterization is facilitated by forced degradation of the drug under varied conditions of hydrolysis, oxidation, dry heat, and photolysis [7]. Some analytical methods for determination of IDA along with its metabolites or in the presence of other members of this class are reported in literature [8,9]. Yang and Guo [10] have developed a HPLC method for determination of IDA and its related substances. But there is

∗ Corresponding author. Tel.: +91 175 3046255; fax: +91 175 2283073. E-mail addresses: [email protected], [email protected] (G. Bansal). 0731-7085/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jpba.2013.07.002

no report on degradation studies or degradation products of IDA. Hence, the present study has been designed to (i) conduct forced degradation study on IDA under the ICH prescribed conditions to identify all possible degradation products that may form under various conditions like hydrolysis, photolysis, dry heat, and oxidation; (ii) characterize the degradation products through spectral and/or LC–MS-TOF studies and (iii) establish its degradation pathways and intrinsic stability characteristics. The drug has been found stable to photolytic and thermal degradation but it undergoes extensive degradation in alkaline media. One acidic hydrolytic and two oxidative degradation products have been characterized and the most probable mechanisms of degradation are outlined and discussed.

2. Experimental 2.1. Drug and chemicals Idarubicin hydrochloride (IDA) was procured as gift samples from Strides Arcolabs Pvt. Ltd. (Bangaluru, India) and used without further purification. Sodium hydroxide (NaOH), hydrochloric acid (HCl), hydrogen peroxide (H2 O2 , 30%) and ammonium formate were purchased from Loba Chemical Pvt. Ltd. (Mumbai, India). Methanol, formic acid and acetonitrile (HPLC grade) were purchased from Merck Specialist Pvt. Ltd. (Mumbai, India). HPLC grade water was obtained from the Direct Ultra water purification

124

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

of each stressor separately (i.e., 0.1 M HCl, 0.1 M NaOH and H2 O) in transparent glass vials. For solid state photolytic study, the drug was spread as a thin layer in petri-plates. These vials as well as the petri-plates were placed at a distance of 9 from the light source in the photolytic chamber for 14 days during which the total UV and white light exposure equaled about 200 Wh m−2 and 1.2 million lux h, respectively. A parallel set of same vials and petri-plates was kept in dark under similar conditions of temperature and RH for the same period of time to serve as dark control. Each degraded sample was refrigerated till analysis. 2.4. LC-UV method and sample preparation

Fig. 1. Structure of idarubicin (IDA).

system in the laboratory (Bio-Age Equipment and Services, SAS Nagar, India). 2.2. Equipments Hydrolytic and thermal forced degradations were carried out using high precision water bath and hot air oven equipped with digital temperature control capable of controlling temperature within range of ±1 ◦ C and ±2 ◦ C, respectively (Narang Scientific Works, New Delhi, India). Photodegradation was carried out in a photostability chamber (KBF 240, WTB Binder, Tuttlingen, Germany) capable of controlling temperature and relative humidity (RH) within a range of ±2 ◦ C and ±5% RH, respectively. The chamber was equipped with light sources as described in Option 2 in the ICH guideline Q1B [11]. The chamber was set at a temperature of 25 ◦ C and RH of 55%. The forced degradation samples were analyzed on a Waters HPLC system (Milford, MA, USA) consisting of binary pumps (515), dual wavelength detector (2487) and Rheodyne manual injector. The data were acquired and processed in Empower 2 software. The chromatographic separations were achieved on Inertsil RP18 (250 mm × 4.6 mm; 5 ␮) column. The mobile phase was degassed using ultrasonic bath (570/H ELMA, Germany). LC-PDA analysis was performed on Waters binary HPLC system (Milford, MA, USA) consisting of pumps (515), auto injector (2707), and PDA detector (2998). MSn studies on IDA were carried out using positive mode of electrospray ionization (+ESI) on LTQ-XL ion trap quadrupole mass spectrometer (Thermo Scientific, Germany). LC–MS-TOF studies were carried out in +ESI mode on micrOTOF-Q11 mass spectrometer (Bruker Daltonics GmbH, Germany), which was controlled by microTOF control software ver.2.0. LC part of the LC–MS comprised of Agilent 1100 series LC system (Agilent Technologies Inc, CA, USA) controlled by Hystar (Ver. 3.1) software. A splitter was placed before the mass detector, allowing entry of only 35% of the eluent. Column used for LC–MS study was same as that for LC-UV study. 2.3. Forced degradation study The hydrolytic degradation studies on IDA were carried out in water, 0.1 M NaOH and 0.1 M, 1 M and 2 M HCl at 80 ◦ C for 8 h. The alkaline hydrolytic degradation study was also carried out in 0.1 M and 0.01 M NaOH at 80 ◦ C for 4 h and in 0.01 M NaOH at 40 ◦ C for 2 h. For oxidative degradation, about 0.1 g of IDA dispersed in 100 ml of 30% H2 O2 was placed in dark at room temperature (30 ± 5 ◦ C) for 24 h. Thermal degradation was carried out on solid drug sealed in amber color vials by exposure to 50 ◦ C for 30 days. The photolytic degradation study was carried out on the drug in solid as well as solution state. For solution state photolytic studies, 2 ml of a 0.1% (w/v) solution of IDA in acetonitrile was mixed with 3 ml

The drug and its UV active degradation products were resolved on Inertsil RP18 (250 mm x 4.6 mm; 5 ␮) column at ambient temperature (30 ◦ C) using 254 nm as detection wavelength. The hydrolytic degradation products were resolved with mobile phase A (HCOONH4 , 20 mM, pH 3.0) and mobile phase B (acetonitrile) in gradient mode (0–20 min; A 90%, B 10% → 21–21 min; A 60%, B 40% → 21–80 min; A 60%, B 40% → 80–81 min; A 90%, B 10%). The oxidative degradation products were resolved through isocratic elution with mobile phase A and mobile phase B (70:30, v/v). In both the methods, the mobile phase flow rate was 0.7 ml min−1 and the injection volume was fixed at 20 ␮l. The solid drug samples from thermal and photolytic conditions were rendered into solutions (1 mg ml−1 ) in methanol and were analyzed using the gradient elution method. Each degraded drug solution as well as solution of solid drug samples was diluted up to 10 times with mobile phase A. The acid and alkali hydrolyzed solutions were neutralized before dilution. Each dilute sample was filtered through nylon membrane (0.45 ␮) before analysis. The LC-UV analysis of each sample was preceded by the corresponding blank. The LC-PDA studies were carried out to check the purity of IDA and each degradation product peak resolved in the LC-UV chromatograms. 2.5. +ESI-MSn and LC–MS-TOF studies Five stage mass (MS5 ) spectra of IDA were recorded in +ESI mode using appropriately chosen precursor ions and ionization potentials (18.0–28.0 V). The operating conditions for recording MS scan of IDA were optimized as follows; end plate offset voltage, −500 V; capillary voltage 4500 V; collision cell RF, 400.0 vpp nebulizer, 1.2 bar; dry gas, 6.0 l min−1 and dry temperature, 200 ◦ C. The same operating conditions were employed for recording mass spectra of IDA and degradation products at the ionization potentials of 10 and 15 V during LC–MS-TOF analysis. All MS5 and LC–MS-TOF spectra were recorded in the range of 50–1000 m/z. 3. Results and discussion 3.1. LC-UV method A HPLC method for determination of IDA as reported by Wall et al. [12] was taken as a lead to develop a method for separation of IDA and its UV active degradation products in a single run. After numerous experimental trials, IDA and the degradation products were optimally resolved on an Inertsil RP18 (250 mm × 4.6 mm; 5 ␮) column using gradient elution with mobile phase A (water:acetonitrile, 90:10, pH 2.5) and mobile phase B (methanol: water, 55:45, pH 2.5) at a flow rate of 1 ml min−1 . While the alkaline degradation products were resolved with mobile phase A, the acidic and oxidative degradation products were resolved with mobile phase B. However, the LC–MS studies of the degradation sample using these chromatographic conditions did not show any peak in the total ion chromatogram (TIC). Even an increase in ionization energy did not produce any peak in TIC. It was attributed

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

125

Fig. 2. Chromatogram of standard solution of IDA (A), of IDA solutions subjected to 30% H2 O2 (B) and to 0.1 M NaOH and 2 M HCl (C).

to the absence of ionization facilitators (such as inorganic salts) in the mobile phase. Hence, water in both the mobile phases was replaced with ammonium acetate or ammonium formate buffer. Ammonium acetate buffer of different strengths (10, 20 and 50 mM) with varied pH (2.5–5) did not resolve the degradation product whereas ammonium formate buffer (20 mM, pH 2.5) afforded moderate resolution. Further replacement of methanol with acetonitrile in mobile phase B improved the elution and resolution of the oxidative products. The best resolution was obtained with mobile phase B composed of ammonium formate buffer (20 mM, pH 3) and acetonitrile (60:40, v/v) but the run time required was more than 80 min. In order to shorten the run time, acetonitrile was increased to 60% which separated the acid hydrolyzed product from the drug within 60 min but the oxidation products were merged with the drug peak. Hence, these gradient chromatographic conditions were selected for separation of alkali and acid degradation products of IDA whereas a separate isocratic HPLC method was developed to resolve the oxidative degradation products as disclosed in Section 2.4. The LC-UV chromatograms of standard, hydrolytically degraded and oxidized IDA are given in Fig. 2. 3.2. Forced degradation behavior IDA was detected as a sharp peak at 24–26 min in both the isocratic as well as gradient methods (Fig. 2A). LC-UV chromatograms of the IDA standard solution as well as of each degraded solution was compared with those of the corresponding and similarly treated blank to locate the peaks due to degradation products. The drug degraded significantly to two products (O-I and O-II) after exposure to oxidative stress at room temperature in dark for 24 h (Fig. 2B). It was extensively degraded in 0.1 M NaOH at 80 ◦ C after 8 h. Exposure of the drug to mild alkaline hydrolytic conditions (i.e., 0.01 M NaOH; degradation temperature, 40 ◦ C or ambient; and degradation time, 4 h or 30 min) produced the same pattern

of degradation impurities which was hard to resolve. Despite the extent and pattern of drug degradation being independent of severity of the alkaline hydrolytic conditions, the UV absorption spectra of variedly alkali degraded drug solutions were found to be similar to that of the standard drug solution. It suggested that though the drug is highly susceptible to hydrolysis in alkaline medium, the degradation was not accompanied with any change in chromophore of the drug. The drug was stable when exposed to 0.1 M HCl at 80 ◦ C for 8 h. Increase in acid strength to 1 M resulted in insignificant degradation with formation of a single product (A-1) as a trace peak. However, further increase in acid strength to 2 M significantly degraded the drug to A-1 after 8 h. It suggested that though the acid strength did not affect the pattern of drug degradation but it severely affected the extent of degradation. The two alkali hydrolyzed products (K-I and K-II) and the single acid hydrolyzed product (A-1) were optimally resolved in a single run (Fig. 2C). No degradation was seen in water at 80 ◦ C for 8 h, thermal stress for 30 days, and the photolytic stress. Based on these results, the drug has been found to be extremely unstable in alkaline medium, unstable in oxidative and strong acidic media but stable in neutral medium as well as in dry heat and photolytic conditions. The LC-PDA analysis of the degradation samples revealed that purity angles of the peaks due to IDA, A-I, O-I, O-II, K-I and K-II were less than their

Table 1 Peak purity data. Analyte peak

Purity angle

Purity threshold

IDA A-I O-I O-II K-I K-II

0.784 0.629 1.063 0.847 0.731 0.852

2.538 1.693 2.206 2.729 2.460 2.758

126

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

Fig. 3. Five stage mass fragmentation spectra of IDA.

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

127

Fig. 4. Proposed mass fragmentation pattern of the IDA and its degradation products.

purity thresholds (Table 1). It suggested that all these peaks were pure and no other product co-eluted with these peaks.

3.3. Mass fragmentation pattern of IDA Some reports on mass spectrometric analysis of IDA are available in literature [13–15] but there is no study on multistage mass spectrometric analysis and hence no mass fragmentation pattern of the drug is known so far. In the present study, five stage mass (MS5 ) spectra of IDA were recorded (Fig. 3) to outline its mass fragmentation pattern. The drug was detected at m/z 498 as [M+H+ ] ion (parent ion) corresponding to its molecular mass of 497 Da (Fig. 3A). The various product ions formed from different precursors ions in all five stages are summarized in Table 2. The product ions m/z 351, 333, 291 and 148 have also been detected in CI-MS, CID-MS and +ESI-MS spectra of idarubicin as reported by Beijnen et al. [13],

Sleno et al. [14] and Lachatre et al. [15], respectively. A critical study of the MS5 spectra revealed that the parent ion directly fragmented to m/z 480, 454, 369, 351, 148 in MS2 spectrum (Fig. 3B). The heaviest fragment m/z 480 further fragmented to m/z 462, 436, 368 and 333 in MS3 Spectrum (Fig. 3C). The m/z 462 fragmented to m/z 444 and 388 in MS4 spectrum (Fig. 3D) and the m/z 436 fragmented to single fragment of m/z 291 in MS5 spectrum (Fig. 3E). Based on this

Table 2 Precursor (parent) ions and product (daughter) ions in MS5 studies. MSn stage 1

MS MS2 MS3 MS4 MS5

Precursor ion (m/z)

Product ions (m/z)

– 498 480 462 436

498 (Parent ion) (100%) 480, 454, 369, 351, 333 (100%), 291, 148 462, 444, 436, 368, 333(100%), 291 444, 388 291 (100%)

128

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

Fig. 5. LC–MS-TOF spectra of IDA and its degradation products. The ‘*’, ‘**’ and ‘***’ represent the parent, M+Na+ and M+K+ ion peaks, respectively.

data, mass fragmentation pattern of IDA was outlined (Fig. 4). Fragmentation of the parent ion to m/z 480 and 454 by the loss of 18 Da and 44 Da was attributed to the loss of H2 O and CH3 CHO, respectively. The fragments m/z 369 and 351 were proposed to form by fragmentation of the glucosamine moiety, similarly as fragments of m/z 415 and 397 have been reported to form from parent ion of doxorubicin [14]. Taking a clue from the same report, the m/z 148 was assigned to the glucosamine ion. In MS3 spectrum, the m/z 462, 436, 368 and 333 were proposed to form by loss of a water molecule, CH3 CHO molecule and by fragmentation of the glucosamine moiety similarly as in MS2 spectrum. In MS4 , the m/z 462 fragmented to m/z 444 by the loss of a H2 O molecule and to m/z 388 due to fragmentation of the glucosamine chain. In MS5 , the m/z 436 fragmented to the single product ion of m/z 291 which was possible due to the loss of glucosamine moiety. In addition to the MS5 study, the drug was also analyzed through LC–MS-TOF to generate an accurate mass

spectral data. The TOF spectrum of the drug (Fig. 5A) showed parent ion at m/z 498.1748 in addition to Na+ and K+ adduct ions at m/z 520.1566 (M+Na+ ), m/z 536.1300 (M+K+ ), respectively. The major fragments were noted at m/z 351.0850, 333.0735, 291.0632. These masses were found to match very closely with accurate masses of the fragments of m/z 351, 333 and 291 (Fig. 4) which supported the proposed structural assignments to these ions. 3.4. Characterization of degradation products The hydrolyzed and oxidized drug solutions were analyzed on LC–MS-TOF instrument to characterize the degradation products through the mass spectral data. The two oxidative degradation products (O-I and O-II) and the single acid hydrolyzed product (A-I) were detected in TIC similarly as in LC-UV chromatogram. However, the alkali degraded products (K-I and K-II) were not detected

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

129

Table 3 LC–MS-TOF spectral data of the idarubicin and its degradation products.

Analyte peak IDA

Observed Mass (Da) 498.1748 [M+H+] 351.0850

Mass difference 147.0898 18.0115

333.0735 42.0103 291.0632 IDA

498.1748

O-II

454.1502 [M+H+] 325.0716

Most probable composition (Theoretical mass; Tolerance)a For the mass difference For the Observed mass C26H28NO9+ (498.1758; -1.1) C6H13NO3 (147.0895; +0.2) C20H15O6+ (351.0863; -1.3) H2O (18.0105; +0.9) C20H13O5+ (333.0757; -2.2) C2H2O (42.0105; -0.2) C18H11O4.+ (291.0651; -1.9)

44.0246

CH3CHO (44.0262; -1.6)

129.0786

C6H11NO2 (129.0789; -0.3)

147.0786

C6H13NO3 (147.0895; -10.9)

34.0061

H2O2 (34.0054; +0.6)

18.0093

H2O (18.0105; -1.2)

34.0048

H2O2 (34.0054; -0.6)

147.0909

C6H13NO3 (147.0895; +1.3)

129.0820

C6H11NO2 (129.0789; +3.0)

147.0922

C6H13NO3 (147.0895; +2.6)

129.0777

C6H11NO2 (129.0789; -1.2)

18.0107

H2O (18.0105; +0.2)

18.0102

H2O (18.0105; -0.3)

42.0105

C2H2O (42.0105; -0.06)

307.0596 O-II

454.1502

O-I

488.1563 [M+H+] 470.1470 454.1515 341.0654 325.0695 307.0593

IDA

498.1748

A-I

369.0970 [M+H+] 351.0863 333.0761 291.0656

a

C24H24NO8+ (454.1496; +0.5) C18H13O6+ (325.0706; +0.9) C18H11O5+ (307.0600; -0.5)

C24H26NO10+ (488.1551; +1.1) C24H24NO9+ (470.1445; +2.4) C24H24NO8+ (454.1496; +1.8) C18H13O7+ (341.0655; -0.1) C18H13O6+ (325.0706; -1.1) C18H11O5+ (307.0600; -0.8)

C20H17O7+ (369.0968; +0.1) C20H15O6+ (351.0863; -0.01) C20H13O5+ (333.0757; +0.3) C18H11O4+ (291.0651; +0.4)

Calculated from elemental composition calculator.

in TIC. These observations in comparison to the results obtained with LC–MS analysis using buffer free mobile phase (as disclosed in Section 3.1) indicated that ionization of these products was not affected by the presence or absence of ionization facilitator in the mobile phase. Even the use of varied ionization potentials did not help in ionization of K-I and K-II. Hence, these products were suggested to remain absent in TIC due to their low concentrations and/or weak inherent ionization. The MS-TOF spectra of O-I, O-II and A-I are given in Fig. 5. The most probable molecular formula

corresponding to each parent and fragment ion was calculated by Elemental Composition Calculator Software and given in Table 3. The tolerance between the observed and calculated accurate masses of each ion was set at not beyond ±5%. 3.4.1. Product O-II It was detected as cluster of peaks at m/z 454.1502, 476.1326 and 492.1071 (Fig. 5C). Based on the mass differences among these peaks, these were assigned respectively, as the parent ion (M1),

130

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

Fig. 6. Proposed mechanism of formation of degradation products from IDA.

M+Na+ , and M+K+ adduct ions of O-II. Incidentally, the M1 was also noted as fragment ion in MS2 spectrum of IDA wherein it was proposed to form by the loss of CH3 CHO from the parent ion of IDA. Moreover, the accurate mass of this fragment in MS2 was found to match very closely with that of M1 (Fig. 4). Hence, O-II was proposed to be desacetylidarubicin which might be formed by oxidative deacetylation through Baeyer Villiger oxidation in the presence of H2 O2 [16]. This proposition was supported by fragments of m/z 325.0716 and 307.0596 in LC–MS-TOF spectrum of O-II which were formed from the proposed desacetylidarubicin by the loss of glucosamine moiety similarly as m/z 351.0850 and 333.0735 were formed from the parent drug ion (Fig. 4). 3.4.2. Product O-I It was recorded as the major heavy ion at m/z 488.1563 along with its M+Na+ and M+K+ ions at m/z 510.1374 and m/z 526.1122, respectively (Fig. 5B). Hence m/z 488.1563 was assigned as the parent ion (M2) of O-I. An even molecular mass of M2 peak suggested an odd number of nitrogen atom (reverse nitrogen rule) [17]. Hence the single nitrogen atom present in glucosamine component of the IDA was suggested to be intact in O-I. The M2 was 10.0185 Da less than the parent ion of IDA and this mass difference did not correspond to any neutral molecule. It indicated that O-I was not formed directly from the drug. However, M2 was 34.0048 Da heavier than M1 and this mass difference was found to correspond to a H2 O2 molecule (34.0055 Da) [18]. Further, M1 was also noted as a fragment ion in MS-TOF spectrum of O-I. Hence, based on this discussion, O-I was suggested to form from O-II by the addition of H2 O2 . This addition is possible across the ketone group generated on the ring D in O-II to form hydroxyhydroperoxide [19]. Hence, O-I was proposed to be desacetylidarubicin hydroperoxide which undergoes mass fragmentation (Fig. 4) in consonant with its MSTOF spectrum. The heaviest fragment of m/z 470.1470 (18.0093 Da less than M2) was possible to form by the loss of a water molecule (18.0105 Da) whereas the fragment m/z 454.1515 (O-II) was proposed to form by the loss of H2 O2 molecule. Fragment m/z 363.0474 was 147.0900 Da less than the M+Na+ ion of O-I. Hence, it was proposed to form by the loss of glucosamine from the M+Na+ ion of

O-I. The fragment m/z 341.0654 was possible to form by loss of glucosamine moiety from M2 similar to the fragmentation of the parent drug ion to m/z 351.0850. The fragments of m/z 325.0695 and 307.0593 were possible to form from M1 as discussed in Section 3.4.1. 3.4.3. Product A-I It was detected as a minor peak at m/z 369.0970, a major ion peak at m/z 391.0790 along with heaviest ion peak at m/z 407.0545 (Fig. 5C). Based on the mass differences amongst these peaks, the m/z 369.0970, 391.0790 and 407.0545 were assigned as parent ion (M3), M+Na+ and M+K+ ions, respectively. An odd molecular mass of M3 suggested an even number or no nitrogen atoms (reverse nitrogen rule) [17]. Therefore, the single nitrogen atom in the glucosamine component of IDA was expected to be lost during degradation of IDA to A-I. The m/z 369.0970 was 129.0778 Da less than the parent ion of IDA and this mass loss corresponded to the glucosamine moiety. Hence, A-I was proposed to be de-glucosaminyl idarubicin which could be formed by cleavage of glycosidic linkage between the tetracycline ring and the glucosamine moiety. This proposition was supported by the fact that M3 was also noted as fragment in MS2 spectra of IDA (Fig. 3). Hence, based on this MS-TOF and MS2 spectral data, A-1 was characterized as deglucosaminylidarubicin, which undergo mass fragmentation (Fig. 4) in agreement with its LC–MS-TOF spectrum. The product ions of m/z 351.0868, m/z 333.0761 and m/z 291.0656 (Fig. 5) were formed from M3 due to the loss of a water molecule, two water molecules and two water molecules along with acetyl group, respectively. The m/z 373.0697 and m/z 355.0580 were proposed to be the Na+ adduct ion of fragments 351.0868 and m/z 333.0761, respectively. 3.5. Drug degradation mechanisms The most probable mechanisms of formation of O-I, O-II and A-I are outlined in Fig. 6. The O-II (desacetylidarubicin) was proposed to form by oxidative deacetylation of IDA through Baeyer Villiger oxidation in the presence of H2 O2 . The mechanistic explanation for

D. Kaushik, G. Bansal / Journal of Pharmaceutical and Biomedical Analysis 85 (2013) 123–131

this oxidation is as follows [16,19]: Addition of H2 O2 to carbonyl bond of the acetyl group on ring D forms a hydroperoxide intermediate which undergoes alkyl group migration. The secondary carbon (attached to hydroxyl group), which is actually a part of ring D, migrates in preference over the methyl group due to their differential migration aptitudes. It converts the acetyl group to acetate group, which is actually a part of hemiacetal. The ester is readily hydrolyzed in the aqueous medium to generate an extremely unstable hemiacetal intermediate. The latter is immediately converted to O-II as ketone. However, the ketonic O-II may undergo keto-enol tautomerism to form O-II as an enol. Hence, O-II can exist in both keto and enol form. Formation of O-I from O-II was possible through the well known addition of H2 O2 to ketone group, generated on the ring D in O-II, to form a hydroxyhydroperoxide [19]. However, this addition reaction is reversible and the hydroxyhydroperoxides (O-I) decomposes to generate the reactant (O-II). Based on this discussion it is proposed that O-I and O-II exist in equilibrium. The A-I, characterized as deglycosylidarubicin, was possible to form by the well known acid catalyzed hydrolysis of glycosidic linkage. Similar type of glycosidic cleavage is also reported to occur in other anthracycline antineoplastics such as doxorubicin, 4 -deoxydoxorubicin, 4 -O-methyldoxorubicin, 4 -epidoxorubicin, doxorubininol, daunorubicin and carminomycin under acidic conditions [13]. However, A-I was detected as a minute peak in drug solution hydrolyzed with 1 M HCl but it was the major peak when the acid strength was increased to 2 M. This difference in the rate of deglycosylation of IDA was attributed to the fact that acidcatalysed hydrolysis of glycosides bearing an aglycon moiety in axial orientation is much slower than that of glycosides bearing an aglycon moiety in equitorial orientation [20]. A 3D energy minimized structure of IDA has revealed that the glycone part exist in chair conformation and the aglycone part (anthrquinone ring) is axially oriented with respect to the glycone part. Further, the tertiary alcohol (geminal to the acetyl group) in ring D of IDA was expected to undergo dehydration readily in acid medium. But no acid hydrolyzed product with mass 18 Da less than the IDA or A-I was detected in the TIC which indicated that this tertiary alcohol was stable in the acid medium. This exceptional stability of the tertiary alcohol may be attributed to the intramolecular H-bonding between the geminal hydroxy and carbonyl groups in ring D which in turn decreases nucleophilicity of the alcoholic oxygen making it significantly less susceptible to protonation in the acid medium and hence stable to dehydration. 4. Conclusion Forced degradation studies on idarubicin were conducted under the ICH prescribed conditions. The drug was found extremely unstable in alkaline medium, susceptible to oxidation and acidic hydrolysis whereas stable to thermal and photolytic stress conditions. Separate LC-UV methods were developed for separation of oxidative and hydrolytic degradation products. MSn and LC–MSTOF studies were carried out to characterize the major degradation products. Two oxidative degradation products were characterized as desacetylidrrubicin and desacetylidarubicin hydroperoxide, whereas the single acid degraded product was characterized as deglucosaminylidarudicin. The most probable mechanisms of degradation of idarubicin were outlined and discussed.

131

Acknowledgements The authors are thankful to Strides Arcolabs Pvt. Ltd. (Bangalore, India) for providing IDA as generous gift sample and to Prof. Saranjit Singh, Head, Department of Pharmaceutical Analysis, National Institute of Pharmaceutical Education and Research (SAS Nagar, India) for extending the facilities to carry out photostability and mass spectral studies. References [1] B.A. Chabner, P.C. Amrein, B.J. Druker, M.D. Michalson, C.S. Mitsiades, P.E. Gos, D.P. Ryan, S. Ramachandra, P.G. Richardson, J.G. Supko, Chemotherapy of neoplastic diseases, in: L.L. Brunton (Ed.), Goodman Gilman’s The Pharmacological Basis of Theraputics, 11th ed., McGraw Hill, Medical Publishing Division, New York, 2006, pp. 1357–1359. [2] R.S. Young, P. Ran, K.Y. Seon, Y.J. Deog, Synthesis of idarubicin analogues containing a glucose or galactose moiety as a glycone, Bulletin of the Korean Chemical Society 31 (2010) 69. [3] J. Blasiak, E. Gloc, K. Wozniak, W. Mlynarski, M. Stolarska, T. Skorski, I. Majsterek, Genotoxicity of idarubicin and its modulation by vitamins C and E and amifostine, Chemico-Biological Interactions 140 (2002) 1–18. [4] United States Pharmacopoeia, USP, Convention, Rockville, MD, (2006) p. 1105. [5] ICH Q3A(R2), Impurities in new drug substances, in: International Conference on Harmonisation, IFPMA, Geneva, 2008. [6] ICH Q3B(R2), Impurities in new drug products, in: International Conference on Harmonisation, IFPMA, Geneva, 2006. [7] ICH Q1A(R2), Stability testing of new drug substances and products, in: International Conference on Harmonization, IFPMA, Geneva, 2003. [8] I. Badea, L. Laz˘ar, D. Moja, D. Nicolescu, A. Tudose, A HPLC method for simultaneous determination of seven anthracyclines, Journal of Pharmaceutical and Biomedical Analysis 39 (2005) 305–309. [9] O. Kuhlmann, S. Hoffmann, M. Weiss, Determination of idarubicin and idarubicinol in rat plasma using reversed phased high performance liquid chromatography and fluorescence detection, Journal of Chromatography B: Biomedical Sciences and Applications 728 (1999) 279–282. [10] M.C. Yang, Z.F. Guo, Study on HPLC determination of idarubicin hydrochloride and related substances, Chinese Pharmaceutical Journal 37 (2002) 945–954. [11] ICH Q1B, Stability testing: photostability testing of new drug substances and products, in: International Conference on Harmonization, IFPMA, Geneva, 1998. [12] R. Wall, G. McMahon, J. Crown, M. Clynes, R.O. Connor, Rapid and sensitive liquid chromatography–tandem mass spectrometry for the quantitation of epirubicin and identification of metabolites in biological samples, Talanta 72 (2007) 145–154. [13] J.H. Beijnen, G. Wiese, W.J.M. Underberg, Aspects of the chemical stability of doxorubicin and seven other anthracyclines in acidic solution, Pharmacy World and Science 7 (1985) 109–116. [14] L. Sleno, V.C. Slater, D.A. Volmer, Dissociation reactions of protonated anthracycline antibiotics following electrospray ionization-tandem mass spectrometry, International Journal of Mass Spectrometry 255-56 (2006) 130–138. [15] F. Lachatre, P. Marquet, S. Ragot, J.M. Gaulier, P. Cardot, J.L. Dupuy, Simultaneous determination of four anthracyclines and three metabolites in human serum by liquid chromatography–electrospray mass spectrometry, Journal of Chromatography B 738 (2000) 281–291. [16] J. March, Advance Organic Chemistry, Reaction, Mechanism and Structure, 4th ed., John Wiley & Sons, Asia, 1999. [17] S. Singh, T. Handa, M. Narayanam, A. Sahu, M. Junwal, R.P. Shah, A critical review on the use of modern sophisticated hyphenated tools in the characterization of impurities and degradation products, Journal of Pharmaceutical and Biomedical Analysis 69 (2012) 148–173. [18] R.M. Silverstein, F.X. Webster, Spectrometric Identification of Organic Compounds, 6th ed., John Wiley & Sons, New York, 1998, pp. 45. [19] Y. Ogata, A. Kawasaki, Equilibrium additions to carbonyl compounds, in: J. Zabicky (Ed.), The Chemistry of the Carbonyl Group, Interscience Publishers, London, 1970, pp. 11–14. [20] M. Miljkovic, Carbohydrates Synthesis, Mechanisms and Stereoelectronic Effects, Springer Science+Business Media, New York, 2010, pp. 377.