Evaluation of the influence of fluoroquinolone chemical structure on stability : forced degradation and in silico studies

*Correspondence: F. R. Paula. Laboratório de Desenvolvimento e Controle de Qualidade de Fármacos e Medicamentos, Universidade Federal do Pampa (UNIPAMPA), Campus Uruguaiana, 97500-970 BR 472 Km 592, Uruguaiana, RS, Brasil. Fax: 55 55 3414 1484 Tel: 55 55 3413 4321. E-mail: faveropaula@unipampa.edu.br Evaluation of the influence of fluoroquinolone chemical structure on stability: forced degradation and in silico studies


INTRODUCTION
Quinolones are a group of synthetic broad spectrum antibacterial drugs commonly used throughout the world.Nalidixic acid was the first compound introduced for therapeutics in 1962 for the treatment of urinary tract infections (Andersson, MacGowan, 2003;Petri, 2005).Different chemical changes to the basic structure of quinolones have been made over the years, aiming to improve the spectrum and antimicrobial activity and to minimize bacterial resistance.Currently, there are four generations of quinolones available for clinical use (Andersson, MacGowan, 2003;Petri, 2005), the main quinolones have a fluoro moiety attached to position 6 of the fused ring.The basic chemical structure of fluoroquinolones is shown in Figure 1.
The stability of drugs is a critical point of quality control and pharmacology for therapeutic agents, since a non-stable or reactive molecule may influence its structure, efficacy and safety.Stability tests, known as forced degradation studies, are important tools used to detect, quantify and identify the generation of degradation products (Alsante et al.;2007;Singh et al., 2013).Forced degradation tests are applied by the pharmaceutical industry, usually under extreme conditions (ICH, 2003;  Klick, 2005;ICH, 2006a;ICH, 2006b;Silva et al., 2009;Brasil, 2013), such as submitting the drug to acid, alkaline and oxidative media, high temperatures, and also to photolytic conditions.Among the cited conditions, alkaline and photolytic environments show a high capacity to promote degradation in the fluoroquinolone class.In fact, this compound class is unstable under photolytic conditions, i.e. fast degradation of these drugs occurs upon exposure to UV radiation (Burhenne, Ludwig, Spiteller, 1999;Sunderland et al., 2001;Torniainen, Askolin, Mattinen, 2007;Paim et al., 2010).This instability promotes the formation of degradation products and may cause photosensitivity dermatitis (allergic or non-allergic), cytotoxicity and genotoxicity (Shimoda, 1998;Tokura, 1998;Stahlmann, 1999;Paim et al., 2013).
Fluoroquinolones, however, present differences in their structure degradation when exposed to UV light and alkaline pH (Torniainen, Askolin, Mattinen, 2007;Paim et al., 2010;Babic´, Periša, Škoric, 2013;Ahmad et al., 2013).Different pathways of degradation have been shown, as well as photochemistry reactions among the various compounds.This indicates the necessity of understanding the differences in reactivity among the fluoroquinolones.
Molecular modelling is a set of computational tools based on theoretical methods of calculation largely used in drug design to understand the molecular structure and physicochemical properties of compounds with biological or chemical interest (Carvalho et al., 2003).Thus, in silico calculations are able to provide information about the behavior of molecules under specific circumstances, which may elucidate, theoretically, the stability of the quinolone class under certain conditions.These methodologies may be used to explain the degradation of quinolone chemical structures and also to identify some aspects which may influence their molecular reactivity (Kieffer et al., 2010;Kleinman et al., 2015).
The aim of this study was to explain changes to the chemical structure of five quinolones, including norfloxacin, ciprofloxacin, enrofloxacin, danofloxacin and gemifloxacin, under the influence of alkaline medium and UV light, using computational tools.These data may be used to understand the stability of these chemical structures and also to help the study of new fluoroquinolones drugs design.

MATERIAL AND METHODS
All chemicals used were of analytical grade and all solvents were of chromatographic grade.Methanol, acetonitrile and triethylamine (TEA) were obtained from Merck ® (Darmstadt, Germany), while phosphoric acid was acquired from Vetec ® (Duque de Caxias-RS, Brazil).Purified water was prepared using a Milli-Q Plus ® system (Bedford, USA).Standard reference danofloxacin, enrofloxacin, norfloxacin and ciprofloxacin hydrochloride with purity of ≥ 98% were obtained from Sigma-Aldrich (St. Louis, USA).Standard reference gemifloxacin mesylate (99%) was acquired from Toronto Research Chemicals, Inc. (Toronto, Canada).

Apparatus
A Shimadzu liquid chromatograph (Kyoto, Japan) equipped with an LC-20AT pump, an SIL-20A auto sampler, a CTO-20A column oven, a SPD-20AT photodiode-array detector and LC Solution V. 1.24 SP1 manager system software was used.A Millipore ® nylon filter (13 mm x 0.45 μm) was used to filter the samples before injection.Separation was performed by an Agilent Zorbax ® C-18 reverse phase column (250 mm x 4.6 mm x 5 µm ID).Photodegradation studies were carried out in a photostability UV chamber (0.98 cm x 0.24 m x 0.25 m height) coated with mirrors and equipped with UV-A and UV-C lamps (Light Express ® , 352 nm, 30W and Ecolume ZW ® , 254 nm, 30W, respectively).

Chromatographic conditions
Each standard reference was performed according to the United States Pharmacopeia (USP, 2009), using the reported or validated method with minor modifications, in which each mobile phase was filtered through a 0.45 μm nylon filter and degassed in an ultrasonic bath before use and stored in a dark closed flask at 4 ºC.The chromatographic conditions are provided in Table I.

Preparation of quinolone standards, alkaline and photolytic degradation study Alkaline degradation
Ciprofloxacin hydrochloride, danofloxacin, enrofloxacin, gemifloxacin and norfloxacin were accurately weighed (10 mg) and dissolved in a 100 mL volumetric flask with 0.1 mol L -1 NaOH, obtaining a concentration of 100 µg mL -1 for each drug.The solutions were filtered through a 0.45 µm membrane filter.Subsequently, 1.0 mL of each solution was placed in a 10 mL volumetric flask and stored at room temperature (21-23 °C) in the dark for 1 or 3 h.Then, these solutions were removed, neutralized with 0.1 mol L -1 HCl and diluted with methanol to achieve a concentration of 15 µg mL -1 .To zero time results was measured immediately after dilution and they were considered 100%.Results determined from all times were compared to each other.

Photolytic degradation
Ciprofloxacin hydrochloride, danofloxacin, enrofloxacin, gemifloxacin mesylate and norfloxacin were accurately weighed (10 mg) and dissolved in a 100 mL volumetric flask with methanol, obtaining a concentration of 100 µg mL -1 for each drug.The solutions were filtered through a 0.45 µm membrane filter.Subsequently, 1.5 mL of each solution was placed in a semi-micro UVcuvette, sealed with parafilm (American Co. Brand, USA) and immediately placed in the UV chamber for the photodegradation study for 5, 15 or 30 minutes (three replicates per period).After the photolytic degradation study, the cuvette solutions were placed to a 10 mL volumetric flask and diluted with methanol (to 15 µg mL -1 ).Then, a 1 mL aliquot of the solution was placed in a vial and 20 µL were injected into the chromatographic system.To zero time results was measured immediately after dilution and they were considered 100%.Results determined from all times were compared to each other.

Statistical analysis
Data analysis was performed using GraphPad Prism 5 (GraphPad Software, La Jolla, USA).Two-way ANOVA was used to investigate the results of the photolytic degradation study and descriptive statistics are expressed as mean + standard deviation (SD).For better visualization of the results, the absolute area from the solution measured immediately after dilution was considered as 100% and it was compared to the results from 5, 15 and 30 minutes of treatment.The Bonferroni post-hoc test was used for multiple comparisons and a p value < 0.05 was considered statistically significant.

Computational tools
The computational analyses were performed using Spartan 08 ® version for Windows (Wavefunction, Inc., USA) and all initial structures were built using atoms and structural fragments from its molecular editor.Geometry optimization was performed in the gaseous phase using Density Functional Theory (DFT), and the Becke three-parameter hybrid (B3) associated with Lee-Yang-Parr (LYP) correlation functionals, i.e. the B3LYP method, with the 6.31G* basis set.Maps of HOMO and LUMO energies and electrostatic potential (MEPs) were also calculated.The isosurface of MEPs was obtained at the van der Waals contact surface, which represents the potentials superimposed onto a surface of constant electron density (0.002 e/au 3 ).The isoenergy contours were generated in the range of 35 to 60 kcal mol -1 .The color intensity indicates more or less positive or negative electrostatic regions in the molecule.All color-coded surfaces obtained from these calculations provided a measure of the overall size of the molecular fragment and the location of positive, characterized by blue, and negative, indicated by green (less negative) and red (more intensely negative) electrostatic potentials.HOMO and LUMO isosurfaces were calculated at 0.0032 au 2 .The constants used in these isosurface calculations were the default parameters in Spartan software.

RESULTS AND DISCUSSION
In general, quinolones shows fast degradation under photolytic conditions, demonstrating high susceptibility to drug degradation when exposed to UV radiation, as described in the literature (Burhenne, Ludwig, Spiteller, 1999;Sunderland et al., 2001;Torniainen, Askolin, Mattinen 2007;Paim et al., 2010).Alsante et al. (2007) suggested a degradation of 5-20% for establishing a stability method, since intermediate degradation products should not interfere in drug analysis.Liquid chromatography procedures showed adequate resolution for all quinolones and degradation products generated by photolytic and alkaline degradation.
The alkaline stress conditions used in this work were set in accordance with Paim et al. (2013).In this procedure, only gemifloxacin showed degradation at 1 and 3 hours in the presence of 0.1 mol L -1 NaOH among the five compounds evaluated.For norfloxacin, ciprofloxacin, danofloxacin and enrofloxacin, no significant degradation occurred in alkaline medium at 5 and 15 minutes (Table II).Similar results were found in the literature, despite the different conditions tested.Ciprofloxacin showed 2.9% of degradation in 0.5 mol L -1 NaOH at 50°C under reflux for 5 h (Vaghela, Rao, 2013) and no degradation was observed for danofloxacin and enrofloxacin when exposed to 0.1 mol L -1 NaOH at 80 °C for 24 h (Liu et al., 2011) and 5.0 mol L -1 NaOH at 70 °C for 1 h (Chakravarthy, Sailaja, Kumar, 2015), respectively.Nevertheless, 18.9% norfloxacin degradation was found when diluted in acetonitrile and degraded with 1.0 mol L -1 NaOH at 70 °C for 11 days.A strong alkaline solution and a long period of time was required for this fluoroquinolone to achieve these degradation values.
Two-way ANOVA was performed to verify if there were significant differences among the fluoroquinolones  According to the results, there were significant differences between gemifloxacin and the other fluoroquinolones at all time points tested (p<0.001).
Ciprofloxacin and enrofloxacin showed significant differences after 30 minutes of UV light exposure (p<0.01) when compared to norfloxacin.Enrofloxacin also demonstrate in 15 minutes (p<0.01 and p<0.001).Significant differences were showed by danofloxacin throughout the experiment when compared with norfloxacin (5 minutes, p<0.05; 15 minutes, p<0.01; 30 minutes, p<0.001).The results are shown in Figure 2.
Fluoroquinolone behavior in the photolytic assay (0-30 min) demonstrated the absence of chromatographic coelution of impurities and/or degradation products from ciprofloxacin, enrofloxacin, danofloxacin, norfloxacin and gemifloxacin in the conditions tested.Peak purity tools showed the absence of impurities and/or degradation products from ciprofloxacin, enrofloxacin, danofloxacin, norfloxacin and gemifloxacin peaks in the photolytic assay.In fact, only gemifloxacin showed visible degradation products in the chromatogram under the tested conditions.These results are shown in Figure 3.

Computational procedures
The DFT method using the B3LYP functional and the 6.31G* basis set was used to optimize the geometry of the chemical structures of the five fluoroquinolones in the gaseous phase.Chemical structures in space filling, tube and maps of energies of density of electrostatic potential (MEP), of HOMO and LUMO were generated in this step.The properties of E HOMO , E LUMO and GAP (E LUMO -E HOMO ) were calculated and used to study the reactivity of fluoroquinolones.All structures, maps and physicochemical properties are shown in Figure 4 and Table IV.
The main structural differences among the five fluoroquinolones are the moieties attached at positions 1 and 7 of the quinolone ring, and the nitrogen atom at position 8. Norfloxacin shows an ethyl moiety attached at position 1 in comparison to the other fluoroquinolones which present a cyclopropane ring attached at this position.The fluoroquinolones tube models showed a similar structure, in which the quinolone ring is not a coplanar structure with carboxylic acid and substituent groups are attached to positions 1 and 7.In the space filling models, it is possible to observe that the lateral chain is higher according to gemifloxacin>danofloxacin>enroflox acin>ciprofloxacin≈ norfloxacin.
MEPs, HOMO and LUMO maps were generated in order to understand the mechanism of degradation of fluoroquinolone chemical structures, since these properties are generally associated with the reactivity and stability of compounds.Fluoroquinolone MEPs showed the electronegative value range around the molecules, but with high sites of this property in the carboxylic acid and the quinolone ring as well as the nitrogen atom of the piperazine ring.Sites of positive electrostatic charge were located in the hydrogen atom of carboxylic acid and at the ethyl or cyclopropane ring.The HOMO and LUMO energy sites are shown in Figure 4. HOMO sites were almost evenly distributed around the structures; however, gemifloxacin and ciprofloxacin showed lower sites in the ring attached at position 7 in comparison with enrofloxacin, danofloxacin and norfloxacin.LUMO sites were located mainly at the quinolone ring.No differences were observed among the fluoroquinolones studied.

Evaluation of the influence of chemical structure on fluoroquinolone degradation
Gemifloxacin was the most unstable compound under both alkaline and photolytic conditions when this molecule was compared with other fluoroquinolones.
The presence of a five member ring attached at position 7 and a nitrogen atom instead carbon located at position 8 of the quinolone ring may influence the reactivity of this molecule.In this sense, observing the MEPs, the nitrogen atom of the quinolone ring showed more negative electrostatic potential located at this position, which was different from the other fluoroquinolones.
In the alkaline forced stress test, ciprofloxacin, enrofloxacin, danofloxacin and norfloxacin showed a similar profile of degradation (Table II).Considering the maps and the calculated physicochemical properties, it was observed that gemifloxacin and ciprofloxacin showed lower E HOMO sites in the rings attached at position 7.This property, however, appears to have no influence on the alkaline degradation of these compounds since ciprofloxacin did not undergo alkaline degradation.Gemifloxacin showed a higher value of E LUMO , -1.33 eV, in comparison with the other compounds, but this property, as well as the tube and space filling models and GAP energy, did not show characteristics which may help in understanding alkaline stability under the studied conditions.
In the photolytic degradation study, norfloxacin was the most stable compound compared to the other four fluoroquinolones (Figure 2).The presence of an ethyl moiety in relation to cyclopropane is the main difference and may suggest that the presence of a three membered ring promotes instability in the fluoroquinolone structure.This feature is in agreement with a study by Babic´, Periša and, Škoric (2013), which demonstrated cyclopropane ring cleavage in enrofloxacin when submitted to photolytic conditions.
Danofloxacin and gemifloxacin were found to be the most unstable compounds after five minutes of the photolytic assay.Paim et al. (2016) performed a photolytic degradation study of gemifloxacin in which a product from  I.  the original fluoroquinolone with a structure missing the ring attached to position 7 was isolated and identified.Danofloxacin has five and six member fused rings similar to the five member ring of gemifloxacin, which may explain the enhanced degradation observed in comparison with norfloxacin, ciprofloxacin and enrofloxacin.These three compounds have a piperazine ring attached to position 7 of the quinolone nucleus and which may make these molecules more stable under photolytic conditions.Babic´, Periša, Škoric (2013) suggested the oxidative photodegradation of enrofloxacin and ciprofloxacin may occur at the piperazine ring attached at position 7 by opening the ring.In another study, Hubicka et al. (2014) identified sixteen degradation products of danofloxacin by LC-MS/MS when it was submitted to photolytic tests.All of them showed alterations mainly at the fused rings of the lateral chain.Additionally, Sturini et al. (2012) showed fluoroquinolone degradation under photocatalytic conditions using TiO 2 ; they indicated five member rings may be more unstable than six member rings.These data confirm the influence of the substituent moieties attached at position 7 on the stability of fluoroquinolone structures.
Gemifloxacin presents a nitrogen atom on the quinolone ring with more negative electrostatic potential located at this position; this may indicate a more susceptible site for interactions with photolytic energy.Similar to alkaline conditions, the physicochemical properties such as E HOMO , E LUMO and GAP, the maps of the molecular orbitals (HOMO and LUMO), as well as the tube and space filling models do not show differences that explain the distinct photolytic degradation behavior under the studied conditions.It has been suggested to apply the concepts of Fukui indices or to determine bond dissociation enthalpies in order to predict the susceptibility of fluoroquinolone degradation in an alkaline or photolytic environment.These procedures may be important to understand the degradation process, but they are not within the scope of this study.
The differences in stability of the five studied fluoroquinolones should be considered because they provide an appropriate comparison among the drugs.Structural features of these compounds may be used in the design of effective and more stable fluoroquinolones.These results are also important since they may increase efficacy and establish the shelf life of the drug product.

CONCLUSION
These studies suggest greater instability of gemifloxacin in comparison with the other fluoroquinolones under all conditions evaluated.A photolytic environment causes degradation of this class of compound and they must be protected from UV light.Chemical structure of these compounds or, specifically, the moiety attached to the quinolone ring should be strongly considered according with the influence on stability.MEPs are an interesting tool for the description of fluoroquinolone reactivity and can be used to investigate degradation products in stress stability studies.

FIGURE 2 -
FIGURE 2 -Results of degradation of fluoroquinolones after UVC radiation under photolytic conditions (A) and expanded graphic to 100-75% peak axis area (B).Media and standard deviation are expressed in percentage.* means statistical difference p <0.05 compared to norfloxacin; ** means statistical difference p<0.01 to norfloxacin; *** means statistical difference p<0.001 to norfloxacin; # means statistical difference p <0.001 between gemifloxacin compared to other fluoroquinolones.

FIGURE 3 -
FIGURE 3 -Chromatograms of fluoroquinolones after treatment in photolytic conditions at time of 0 and 30 minutes.Chromatographic conditions describing in TableI.

TABLE II -
Forced degradation study for fluoroquinolones using 0.1 mol L -1 NaOH in room temperature(21-23 °C)

TABLE III -
Degradation forced studies for fluoroquinolones using UVC radiation (254 nm) deviation are expressed in percentage.RSD, relative standard deviation relative to degradation by the action of UVC 254 nm light.

TABLE IV -
Energy of HOMO, LUMO and GAP eigenvalues obtained from DFT/3BLYP-6.31G*calculations using Spartan'08 for Windows