Degradation Characteristics of a Novel PAF Receptor Antagonist, SY0916, in Aqueous Solution

SY0916 has been proven to be a potent treatment agent against rheumatoid arthritis in preclinical studies and has been shown to be safe in phase I clinical trials. However, SY0916 is unstable in water, which is frequently used in pharmaceutical development processes. The degradation behaviour and stability of SY0916 in aqueous solutions were investigated at different pH levels, periods of time, and temperatures. Two degradation products (DPs) were successfully separated and characterized by liquid chromatography coupled to high-resolution tandem mass spectrometry (LC-HRMS/MS), liquid chromatography coupled to nuclear magnetic resonance with solid phase extraction (LC-SPE-NMR), and nuclear magnetic resonance (NMR). SY0916 decomposed to its α,β-unsaturated ketone in protonic solvents, and the α,β-unsaturated ketone further transformed into its alcohol form through a conjugate addition reaction in aqueous media. The results of this study indicate that the pH of the buffer solutions should be maintained between 3.0 and 3.6 for maximum SY0916 stability. Factors that affect degradation should be carefully controlled to mitigate or avoid drug decay.


Introduction
Platelet-activation factor (PAF) is considered a significant inflammatory mediator of different pathologies, including various types of inflammation [1,2], allergy [3], and immune diseases [4,5]. SY0916 ((E)-ethyl1-(5-(4-chlorophenyl)-3oxopent-4-enyl)piperidine-4-carboxylate, Figure 1), a firstin-class PAF receptor antagonist, has been proven to be highly effective for rheumatoid arthritis therapy in preclinical studies and has been shown to be safe in phase I clinical trials [6][7][8][9][10][11]. Unlike commercially available antiarthritis drugs, which have the side effect of causing gastrointestinal injury, SY0916 possesses the notable advantage of protecting gastric mucosa [6]. Moreover, in rats with collagen-induced type II arthritis, SY0916 significantly relieved inflammation in soft tissue around the middle of joints, as observed by X-ray and pathological examination [7], and greatly decreased the serum levels of IgG, TNF-α, and IL-β [8]. In a pharmacokinetic study in healthy humans [7], SY0916 was absorbed rapidly (t 1/2 � 0.5 h ± 0.1 h) and was quickly eliminated in its metabolite form through renal excretion. Furthermore, doses of up to 500 mg were well tolerated and presented no serious adverse events in the first-in-human study.
SY0916 is slightly soluble in water but easily decomposes into two unknown degradation products (DPs) in water. Because water and other aqueous solutions are frequently introduced during various pharmaceutical development processes, the underlying degradation mechanism of SY0916 in aqueous solutions must be elucidated. e aim of this study was to identify the DPs and clarify the degradation characteristics of SY0916 in water and other aqueous solutions. Accordingly, structural characterization of the DPs was performed using liquid chromatography coupled to high-resolution tandem mass spectrometry (LC-HRMS/ MS), liquid chromatography coupled to nuclear magnetic resonance with solid phase extraction (LC-SPE-NMR), and nuclear magnetic resonance (NMR). e stability of SY0916 in solution was controlled by using a specific pH range and time. Based on the structural identification and stability studies of the DPs, the degradation mechanism of SY0916 in aqueous solutions was elucidated, and relevant measures aimed at avoiding drug decay were proposed.

HPLC Analysis.
e separation of SY0916 and its DPs was achieved using a Shimadzu LC-20AD liquid chromatography system with an autoinjector, binary solvent manager, and photodiode array (PDA) detector (Shimadzu, Tokyo, Japan). e column was an Alltima Cyano 100A column (4.6 × 250 mm, 5 µm, ermo Scientific, MA, USA), and the oven temperature was set to 30°C. e mobile phase was a mixture of a 0.006 mol/L ammonium acetate solution with 0.22% acetic acid and acetonitrile (pH 3.9, 70 : 30, v/v). e flow rate was set to 1.0 mL/min, the injection volume was 20 μL, and the detector wavelength was 300 nm.

LC-HRMS/MS Analysis.
All LC-HRMS/MS data were collected using an Agilent 1290 Infinity UHPLC coupled to a 6540 UHD Accurate-Mass Q-TOF mass spectrometer with an electrospray ionization source (Agilent Technologies, CA, USA) operated in a positive ion mode. MassHunter Qualitative Analysis B.06.00 was used to control the system and to perform data acquisition and processing. HPLC was performed as described above, with the flow rate decreased to 0.5 mL/min. e optimized source conditions for the MS scans in positive ion mode (ESI+) were as follows: an ion spray voltage of 3500 V, nozzle voltage of 1000 V, carrier gas temperature of 300°C, carrier gas flow of 8 L/min, sheath gas temperature of 350°C, sheath gas flow of 11 L/min, fragmentor voltage of 135 V, and collision energy of 20 eV for D1 and 15 eV for D2. Due to its high concentration as well as the possibility of it contaminating the mass analyser, the HPLC fraction corresponding to SY0916 was diverted to waste by switching the instrument valve at a set time.

NMR Analysis of D1.
One-and two-dimensional NMR spectra ( 1 H NMR, 13 C NMR, heteronuclear single quantum coherence (HSQC), and heteronuclear multiple bond correlation (HMBC)) of synthesized D1 were recorded on a Bruker AVANCE III HD 600 MHz spectrometer (Bruker BioSpin, Rheinstetten, Germany). e sample was dissolved in CD 3 OD, and tetramethylsilane was used as the chemical shift reference standard.

LC-SPE-NMR Analysis of D2.
Twenty millilitres of a 1 mg/mL SY0916 solution in monopotassium phosphate buffer (pH 5.0) was heated in a water bath at 80°C for 24 h to obtain a stressed solution of D2. e stressed solution was extracted three times with 5 mL of ethyl acetate each time, dried under nitrogen at room temperature, and dissolved in 4 mL of acetonitrile; the resulting solution was used for LC-SPE-NMR analysis.
One-and two-dimensional NMR spectra ( 1 H NMR, 13 C NMR, HSQC, and HMBC) of D2 were obtained using the LC-SPE-NMR system. LC-SPE-NMR analysis was performed on a Bruker AVANCE III HD 600 MHz spectrometer (Bruker BioSpin, Rheinstetten, Germany) coupled with an Agilent 1260 system (Agilent Technology, CA, USA). e resulting solution was separated using the HPLC method described in Section 2.2. Based on its ultraviolet (UV) spectrum and peak retention time, the component of interest was collected by a HySphere Resin GP10 cartridge after the postcolumn addition of water using a Knauer K100 HPLC pump (Berlin, Germany). e trapped analyte was dried under N 2 gas for 55 min and eluted at 0.2 mL/min with CD 3 OD into a 3-mm NMR tube for NMR analysis. e chemical shifts were recorded in ppm (δ) downfield from the internal standard, tetramethylsilane (TMS). Bruker TopSpin software version 3.2 was used for data acquisition and processing.

Degradation and Stability
Studies. SY0916 is slightly soluble in water according to the USP classification of solutes. A bulk sample of SY0916 was dissolved in water. e solution was kept at room temperature for 24 h and was assessed by HPLC in discrete intervals. Two unknown DPs (D1 and D2) were observed ( Figure 2). SY0916 degradation was apparent at as early as 2h in water, at which time the major DP was D1 (7.6%). At 24 h, 23.2% of the DPs was D1, and 4.1% was D2. e stability of SY0916 in buffer solutions at pH 3.0∼5.0 was further studied using HPLC in discrete intervals. At room temperature, SY0916 was stable for 4h in buffer solutions at pH 3.0∼3.6 (99.76% ± 0.05%, calculated by a normalization method). At low temperature (10°C ± 1°C), samples in solutions at pH 3.0∼3.6 remained stable for a longer time period (24 h) than did the samples at room temperature.
As shown in Figure 3, more D1 was produced in the aqueous solutions with higher pH levels (pH ≥ 4.0 and water). Compared with D1 in the same sample solutions, the D2 formation rate was much slower, and the yield was lower. Similarly, aqueous solutions at higher pH favoured the production of D2.
e LC-HRMS/MS spectrum of D1 exhibited a protonated molecular ion at m/z 193, which is by m/z 157 mass units lower than protonated SY0916 and possibly results from the loss of the 4-carboxylate piperidine moiety. As shown in Figure 4 and Table 1, the MS 2 spectrum of the precursor ion at m/z 193 displayed a prominent product ion at m/z 158 due to the loss of the Cl radical (35 Da). e product ion at m/z 175 was proposed to be the product after the loss of H 2 O (18 Da) from m/z 193 [12]. e product ion at m/z 130 was assumed to be formed after the loss of the Cl radical (35 Da) and a molecule of ethylene (28 Da) from the precursor ion. e product ion at m/z 125 was considered a chlorinesubstituted tropylium ion due to the loss of C 4 H 4 O (58 Da) from the ion at m/z 193. e ion at m/z 115 was speculated to be an ylium, as illustrated in Figure 4 [13,14]. Moreover, the fragment ion at m/z 81 resulted from the loss of the 4-chlorophenyl moiety from the precursor ion. e ion at m/z 53 was due to the loss of one molecule of 1-chloro-4-ethylbenzene (140 Da) from the precursor ion. erefore, we tentatively deduced that the chemical structure of D1 is (4E)-5-(4-chlorophenyl) penta-1,4-dien-3-one (Figure 1). e proposed fragmentation pathway is shown in Figure 4. Synthesized D1 was validated by MS analysis under the conditions described in Section 2.3, and the acquired data were in agreement with those of D1 in the sample solutions.
(2) NMR Analysis of D1. e atom assignments for the 1 H and 13 C chemical shifts of D1 are summarized in Table 2   Journal of Analytical Methods in Chemistry
(2) LC-SPE-NMR analysis of D2. e atom assignments for the 1 H and 13 C chemical shifts are listed in Table 2. Similar to D1, the protons of D2 were assigned to three separate spin systems. e appearance of protons at 7.37 ppm and 7.58∼7.61 ppm with a coupling constant of 8.5 Hz indicated the presence of a para-substituted phenyl subunit. e   conclusive evidence for the atom assignments of C-1′ and C-4′. e LC-SPE-NMR results confirmed the structure of D2 that was proposed by LC-HRMS/MS.

Degradation Mechanisms of SY0916 in Aqueous Solutions.
Based on the unambiguous structures of the degradants, the degradation pathway of SY0916 in solution was proposed. SY0916 is a Mannich base that tends to decompose to its α,β-unsaturated ketone (D1) under alkaline conditions. Furthermore, the α,β-unsaturated ketone may not be stable in aqueous solutions and is likely to transform to its alcohol form (D2) through a conjugate addition reaction.
With a better understanding of the formation mechanisms of the degradants, special attention should also be paid to the primary factors affecting degradant levels. Water, solutions with high pH, and protonic solvents such as alcohols should be used with care to prevent the degradation of SY0916 to D1. A large quantity of aqueous solution, especially at high temperatures, should be avoided to prevent the introduction of D2. e stability of SY0916 in aqueous solution depends on the pH of the solution. As described above, the results of this study provide a useful reminder to consider the conditions during SY0916 development processes, such as the dissolvent choice for analytical methods in routine assays or for pharmacokinetic studies and the solvent use in formula preparations.

Conclusions
With the help of LC-HRMS/MS, LC-SPE-NMR, and NMR approaches, structures of the two degradants of SY0916 were rapidly characterized, and their degradation mechanism in sample solutions was elucidated. e drug substance decomposed to its α,β-unsaturated ketone in protonic solvents, and the α,β-unsaturated ketone further transformed to its alcohol form through a conjugate addition reaction in aqueous media. Systematic stability studies of sample solutions showed that the pH of buffer solutions should be between 3.0 and 3.6 for maximum SY0916 stability. Factors affecting the degradant levels should be carefully controlled during the procedures used in various pharmaceutical development processes.

Data Availability
All data included in this study are available upon request from the corresponding author.  Table 2. e supplementary data was not annotated in Table 2 of the submitted manuscript. Figure