Photodegradation assessment of amisulpride, doxepin, haloperidol, risperidone, venlafaxine, and zopiclone in bulk drug and in the presence of excipients

Photostability of amisulpride, doxepin, haloperidol, risperidone, venlafaxine, and zopiclone in APIs and powdered tablets during exposure to UVA irradiation was studied. In order to evaluate the photodegradation process an ultrahigh performance liquid chromatography method coupled with tandem mass spectrometry (UHPLC-MS/MS) was developed and validated. It was found that the photodegradation of the studied compounds depends on the type of drug and co-existing excipients. Different percentage of photodegradation of the studied drugs was observed, and therefore amisulpride decomposed at 2.63% in bulk drug and at 5.74% in pharmaceutical preparation, doxepin: 29% and 72.38%, haloperidol: 3.71% and 26.20%, risperidone: 7.13% and 12.86%, venlafaxine: 38.59% and 4.22%, zopiclone: 18.62% and 31.42% respectively, after 114 days of UVA irradiation. In addition, kinetic evaluation of the photodegradation process was performed by determining the order of reaction, reaction rate constant k and time t0.1 and t0.5. The photodegradation products of studied drugs were identified, and their fragmentation pathways, derived from MS/MS data, were proposed. The photostability testing is an integral part of the drug stability assessment to ensure quality, efficacy and safety of the formulated products during manufacturing process, storage as well as normal use.


Introduction
One of numerous factors influencing stability of the drug, is light which initiates the process of photochemical conversion, leading to changes in the chemical structure of the exposed compound. The problem of interaction between light and pharmaceutical products is important in the drug formulation technology during both, technological processes and storage as well. Irradiation can lead to the reduction or loss of drug activity, as well as formation of toxic products 1 3 of photodegradation. Therefore, photostability testing is an integral part of the assessment of the drug stability, which is important at the stage of drug development and its registration aiming at ensuring its safety and efficacy [1][2][3][4][5].
Active pharmaceutical ingredients (APIs) may be directly or indirectly exposed to radiation. Direct photochemical reactions occur when API degradation is the result of direct light excitation of its molecules. The substance will only degrade if it comes in contact with radiation of an appropriate wavelength. In the case of an indirect (sensitized) photoreaction there is a transfer of excitation energy to a molecule (e.g. API) other than the compound which initially absorbs the radiation. The compound that absorbs the radiation, the photosensitizer (e.g. excipient, impurity), is initially transformed from its ground state (singlet state) into a relatively long-lived excited triplet state. The excited triplet state can mediate a photosensitized reaction due to the long lifetime and bi-radical nature with unpaired electron spins [6]. Therefore, both the excipients and the type of preparation could influence the photoreactivity of the drug substance.
The purpose of this work was to study the photostability of selected drugs in APIs and powdered tablets during exposure to UVA irradiation. Furthermore, we studied the impact of excipients present in tablets on the photodegradation process. In order to evaluate the photodegradation process we have developed and validated an ultrahigh performance liquid chromatography method coupled with tandem mass spectrometry (UHPLC-MS/MS) for the determination of APIs and their phodegradation products. In addition, the kinetic evaluation of the photodegradation process was carried out and photodegradation products have been identified. The investigations have been performed on the group of drugs belonging to the central nervous system agents representing different structural groups. The selection of drugs was based on their properties, availability and photodegradation process. In the study the following drugs were used: amisulpride, doxepin, haloperidol, risperidone, venlafaxine, zopiclone which undergo photodegradation process.
Photostability of some of the investigated drugs is a subject of a few research publications: and so, by UHPLC method coupled with mass detection, AMI degradation products in methanol solutions were identified under the influence of UVA radiation [7]. For the identification of photodegradation products of aqueous DOX solutions, TLC with densitometric detection was used [8,9]. Photostability of VEN in aqueous solutions was evaluated using LC/ MS method [10], and the effect of UVA on the stability of solutions containing ZOP was evaluated by capillary electrophoresis [11].
In the available literature, no data on the photodegradation studies in solid state of the drugs we selected were found.

Method validation
UHPLC-MS/MS used in the photostability study of AMI, DOX, HAL, RIS, VEN, and ZOP was validated by evaluation of its specificity, accuracy, precision, linearity range, limit of detection, and limit of quantification. The developed method provides satisfactory separation of the tested pharmaceutical substances from their photodegradation products as evidenced by the values of resolution factor R s calculated based on equation R s = 2 (t R2 − t R1 )/(W 1 + W 2 ) where t R1 , t R2 -retention time for peak 1 and 2, respectively; W 1 , W 2peak widths at the base [s] for peak 1 and 2, respectively.
The main peak purity was examined with MS spectra using CODA algorithm (Waters Corporation, Milford, MA, USA). The investigated MS spectra uniquely contained signals corresponding to the examined substances and solvent.
Recovery percent of the tested components shown as average values for three concentration levels was satisfactory, and ranged between 97.08 and 104.17%. High precision determined for three concentration levels was confirmed by the RSD values ranging from 0.06 to 2.95%.
Linearity was maintained over a concentration range, from 0.306 to 2.040 mg cm −3 for AMI, from 0.300 to 2.00 mg cm −3 for DOX, from 0.321 to 1.926 mg cm −3 for HAL, from 0.214 to 2.140 mg cm −3 for RIS, from 0.204 to 1.836 mg cm −3 for VEN, and from 0.315 to 2.100 mg cm −3 for ZOP. Linear determination coefficient R 2 did not specify clearly the linearity of the calibration method [12]. Therefore, to assess the linearity, Mandel's test was used. Comparative results of linear and quadratic fit of the studied drugs indicate linear fit of calibration curves for HAL and RIS. The p-values indicate the significance of the slopes and nonsignificance of the intercepts. In other cases, quadratic fit was obtained. The results of Mandel's test (p = 0.685) indicate equal contribution of linear and quadratic fit for ZOP calibration, however due to the statistical significance of the intercept in the linear model, quadratic model was selected. The normality of residuals distribution was confirmed by the Shapiro-Wilk test. The results of the Durbin-Watson (D-W) test showed no significant autocorrelation of random components, except for DOX, RIS, and ZOP. For the abovementioned drugs, D-W test did not decide on the presence or absence of autocorrelation, therefore Lagrange multiplier test was performed. The results of this test indicate no significant autocorrelation of random components in the selected models. Constant variance of random components in the proposed fits was checked using Bartlett test. The obtained results indicate the presence of constant variance in all cases (p-value > 0.05) ( Table 1).

Photostability study of the analyzed substances in bulk drug and in powdered tablets
Studies of the effect of UVA on AMI, DOX, HAL, RIS, VEN, and ZOP in bulk drug and in powdered tablets demonstrated, that they undergo photodegradation, which proceeds at different speeds, depending on the type of API and excipients in drug dosage form. In the registered chromatograms, differences in the number of additional peaks as well as in retention times (t R ) were observed. Different decomposition rate of a given substance in bulk drug and in a pharmaceutical preparation was reported. In most cases, these differences also include the number of resulting products occurring due to the photodegradation process. However, no changes were observed during irradiation of dark control samples. As a result of AMI decomposition in both, bulk drug and pharmaceutical preparation, three products were formed: AP-1, AP-2, and AP-3, in the case of DOX eleven products (DP-1-DP-11) were identified in both irradiated samples. Under the influence of UVA six products were formed in the preparation containing HAL: HP-1-HP-6, five in the irradiated substance: HP-1, HP-2, HP-3, HP-5, and HP-6 ( Fig. 1).
Both irradiated RIS samples decomposed into three products: RP-1, RP-2, and RP-3. VEN in bulk drug decomposed into six products: VP-1-VP-6, whereas in the pharmaceutical preparation only into four VP-1, VP-3, VP-5, and VP-6. Six products of photodegradation were formed in the case of ZOP, both in bulk drug and in the pharmaceutical preparation: ZP-1-ZP-6 ( Fig. 2).
Different percentage of photodegradation of the studied drugs was observed, and therefore AMI decomposed at 2.63% in bulk drug and at 5.74% in pharmaceutical preparation, DOX: 29% and 72.38%, HAL: 3.71% and 26.20%, RIS: 7.13% and 12.86%, VEN: 38.59% and 4.22%, ZOP: 18.62% and 31.42% respectively, after 114 days of UVA irradiation with estimated standard deviation calculated as uncertainty  (12,26,44,56,79,91, and 114 days) on the assumption of first order kinetics (Fig. 3). In majority of the studied drugs, higher decomposition occurs in powdered tablets than in bulk drug, except for VEN, for which the situation is reversed: in bulk drug, photodegradation occurs faster than in the pharmaceutical preparation.
Differences in photodegradation between the API and the pharmaceutical preparation can be justified by the presence of excipients in the form of drug. The effect of UV radiation on these substances is diverse: they can be inhibitors of photochemical reactions, be susceptible to attack of free radicals and inhibit the decomposition of the active substance, they can undergo photodegradation, they can also be a source of peroxides and also participate in free radical reactions [13,14].
Increased decomposition in the tablets as compared to a bulk drug may be caused by the presence of excipients, particularly TiO 2 exhibiting photocatalytic properties [15] associated with absorption of radiation. As a result, the energy of photons absorbed by TiO 2 can be converted into chemical energy leading to the formation of hydroxyl radicals (e.g. from water vapor). In the case of RIS and ZOP, the irradiated tablet mass contained TiO 2 .
Among the tested drugs AMI, DOX, VEN, and HAL preparations do not contain TiO 2 in powdered mass of the tablet. In terms of AMI tablets, photodegradation occurs to a lesser degree. As mentioned above, VEN in bulk drug decomposes to a greater extent as compared to pharmaceutical preparation. Taking into account slight differences in the composition of excipients, which relate to the presence of the Eudragit RS 100 copolymer, one can assume that this substance may weaken the direct effect of UV radiation. In the checked literature, publications describing the research showing the protective action of Eudragit against UV-VIS radiation on pantoprazole [16,17], ranitidine [18], clotrimazole [19], and microorganisms Spodoptera frugiperda multiple nucleopolyhedrovirus [20,21] were found.
In addition, among the tested pharmaceutical preparations DOX, RIS, and ZOP contained in their composition talc, which has nucleating properties and can increase the light transmission of the powdered tablet mass, similar to nucleated polymers [22,23].
The observed differences in photodegradation processes occurring in the studied drugs may also be associated with diverse chemical structures of individual substances.

Kinetics study
Changes in the concentration of the studied drugs occurring during UVA irradiation were determined in accordance with first-order reaction kinetics (Fig. 4) calculating the reaction rate constant k, degradation half-times t 0.5 and t 0.1 , after which 50% and 10% of initial substances undergo degradation, respectively.
The calculated values of reaction rate constants of photodegradation in the case of the most studied drugs have different values and vary from the lowest for AMI and HAL to the highest for DOX in bulk drug, and in pharmaceutical preparations from AMI and VEN to DOX, respectively ( Table 2).

Identification of degradation products
The structures of photodegradation products of AMI, DOX, HAL, RIS, VEN, and ZOP were proposed on a basis of UHPLC/MS analysis supported with fragmentation patterns obtained from MS/MS experiments. The proposed degradation products are shown in Figs. 1, 2 with parameters in Table 3.
In case of AMI it seems that the photodegradation mainly involved dealkylation of pyrrolidine moiety (AP-1) or hydroxylation of ethyl substituent of nitrogen atom of pyrrolidine moiety (AP-2). Additionally one product of dealkylation of nitrogen atom of amide moiety with subsequent rearrangement was observed (AP-3).
In case of DOX the CAD experiments suggest that the photodegradation process mainly involved hydroxylation of propylidenyl chain (DP-1-DP-4, DP-6) and/or hydratation of double bond between propylidene chain and oxepine ring (DP-1 and DP-2). In case of products DP-1, DP-2, and DP-4 it seems that the photodegradation involved also oxidation of oxepine ring in close vicinity to oxygen atom. It seems that only in case of DP-5 oxidation took place in one of the phenyl rings. Further oxidation of DOX led to oxepine ring opening (DP-7-DP-9), demethylation of amine moiety (DP-7) and degradation of propylidene chain (DP-8-DP-11). Product DP-10 seems to be intramolecular condensation product of DP-8. In case of HAL the photodegradation process seems to involve dealkylation of nitrogen atom of piperidine ring (HP-1) with subsequent dehydrogenation (HP-2), opening of piperidine ring (HP-3), its hydroxylation (HP-4), or hydroxylation of phenyl ring of 4-(4-chlorophenyl)-piperidin-4-ol moiety (HP-5 and HP-6). The photodegradation of VEN seems to involve oxidation of methylene group in close vicinity to the nitrogen atom (all major products, VP-1-VP-6) and C-C bond breaking between cyclohexyl moiety and ethylene chain (VP-1-VP-5). Additionally for product VP-2 further oxidation of ethylene chain was observed, and for products VP-1 and VP-4 demethylation of oxygen (VP-1) or nitrogen (VP-4) atom took place.

Conclusions
The developed UHPLC method meets the criteria of validation and may be used for the determination of AMI, DOX, HAL, RIS, VEN, and ZOP in the presence of their photodegradation products. Photodegradation of the studied substances follows first-order kinetics. It was found, that most of the investigated drugs decompose much faster in powdered tablets than in bulk drug substance, except for VEN, which decomposes tenfold faster in bulk drug than in the form of tablets. Calculated values of the kinetic parameters (k, t 0.1 , and t 0.5 ) show that the quickest decomposition is observed for DOX in both, substance as well as in pharmaceutical preparation, and the slowest decomposition is observed for AMI and HAL substances and powdered tablets containing AMI and VEN.

Irradiation conditions and preparation of samples
On Petri dishes with a diameter of 5.7 cm, 30.0 mg of the examined substances and powdered tablet formulation containing 30.0 mg of active substance, was weighed. For each sample, dark control sample was prepared, which was protected from radiation by four layers of aluminum foil.  The samples were inserted into the climatic chamber KBF-ICH 240 APT.line (Binder GmbH, Tuttlingen, Germany) and exposed to UVA irradiation. Irradiation was carried out under the following conditions: range of radiation 320-400 nm with maximum emission at λ = 365 nm at 20 °C and relative humidity of 60%. The samples were placed at a distance of 13 cm from the radiation source. Dose of emitted radiation was determined using VLX-3 radiometer (Vilber Lourmat) with CX-365 sensor and each time was 5.09⋅10 -3 J cm min −1 . Every few days, the samples were thoroughly mixed. The exposure time of the samples was chosen experimentally. Irradiation was stopped when the degradation of the more labile samples reached about 30%. After irradiation for a particular number of days, substances were dissolved in methanol and the powdered tablet formulation was extracted into methanol and centrifuged to obtain solutions at concentrations of approximately 1.00 mg cm −3 . The obtained solutions were then analyzed by UHPLC-MS/MS method.

UHPLC-MS/MS analysis
The UHPLC-MS/MS system consisted of a Waters ACQUITY ® UPLC ® (Waters Corporation, Milford, MA, USA) coupled to a Waters TQD mass spectrometer (electrospray ionization mode ESI-tandem quadrupole). Chromatographic separations were carried out using the Acquity UPLC BEH (bridged ethyl hybrid) C 18 column; 2.1 × 100 mm, and 1.7 µm particle size, equipped with Acquity UPLC BEH C18 VanGuard pre-column; 2.1 × 5 mm, and 1.7 µm particle size. The column was maintained at 40 °C, and eluted under gradient conditions using from 95 to 0% of eluent A over 10 min, at a flow rate of 0.3 cm 3 min −1 . Eluent A: 0.1% v/v formic acid in water; eluent B: 0.1% v/v formic acid in acetonitrile.
Chromatograms were recorded using Waters eλ PDA detector. Compound concentration (%i) after photodegradation was calculated by using the internal normalization method according to formula: %i = A i /∑A⋅100%, where A i is peak area of the constituent being determined, and ∑A is the sum of peak areas for all constituents present in the chromatogram. Spectra were analyzed in 200-700 nm range with 1.2 nm resolution and sampling rate 20 points/s. MS detection settings of Waters TQD mass spectrometer were as follows: source temperature 150 °C, desolvation temperature 350 °C, desolvation gas flow rate 600 dm 3 h −1 , cone gas flow 100 dm 3 h −1 , capillary potential 3.00 kV, cone potential 20 V. Nitrogen was used for both nebulizing and drying gas. The data were obtained in a scan mode ranging from 50 to 1000 m/z in time 0.5 s intervals; 8 scans were summed up to get the final spectrum.
Collision activated dissociations (CAD) analyses were carried out with the energy of 30 eV, and all the fragmentations were observed in the source. Consequently, the ion spectra were obtained by scanning from 50 to 500 m/z range. Data acquisition software was MassLynx V 4.1 (Waters).

Validation
Validation of the method was performed by the determination of specificity, accuracy, precision, linearity, limit of detection, and limit of quantification [24].
Specificity of the method was assessed by comparing chromatograms of pure standards and chromatograms of tested solutions after the exposure to UVA radiation, taking into account retention time (t R ) and resolution factor (R s ) of peaks derived from the tested substances and neighboring peaks from degradation products.
Accuracy was determined based on sample analysis of known concentrations and comparing the results obtained by a validated method with true values, followed by calculation of the recovery percentage. Determinations were performed at three concentration levels: 80%, 100%, and 120%, for each of them three repetitions were done.
Precision of the method was determined at three levels of drug concentrations in reference solutions: 50%, 100%, and 150%. For each concentration level, three repetitions were done.
Evaluation of linearity was performed by plotting a curve between peak area and concentration of the tested substances. For each drug, two series of determinations were performed within the concentration ranges described in "Standard solutions" section.
Limit of detection and limit of quantification were determined from linearity in the concentration ranges: AMI from 0.204 to 1.224 mg cm −3 , DOX from 0.200 to 1.200 mg cm −3 , HAL from 0.214 to 1.284 mg cm −3 , RIS from 0.214 to 1.284 mg cm −3 , VEN from 0.204 to 1.224 mg cm −3 , ZOP from 0.210 to 1.260 mg cm −3 , using formulas: LOD = 3.3⋅S y /a, LOQ = 10⋅S y /a, where: S yestimation error, a-slope of the regression line.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.